experiments in ion traps

FORMATION OF SMALL HYDROCARBON IONS UNDER INTER- AND CIRCUMSTELLAR CONDITIONS:

EXPERIMENTS IN ION TRAPS

von der Fakultät für Naturwissenschaften der Technischen Universität Chemnitz

genehmigte Dissertation zur Erlangung des akademischen Grades

doctor rerum naturalium

(Dr. rer. nat.)

vorgelegt von mr Igor Savić, dipl. fiz.

geboren am 19.08.1970 in Priština (Jugoslawien)

eingereicht am 2. Juli 2004

Gutachter:

Prof. Dr. Dieter Gerlich

Prof. Dr. Gotthard Seifert

Dr. Juraj Glosík

Tag der Verteidigung, 26. August 2004

http://archiv.tu-chemnitz.de/pub/2004/

3

BIBLIOGRAPHISCHE BESCHREIBUNG

FORMATION OF SMALL HYDROCARBON IONS UNDER INTER- AND CIRCUMSTELLAR CONDITIONS: EXPERIMENTS IN ION TRAPS

Dissertation an der Fakultät für Naturwissenschaften der Technischen Universität Chemnitz, Institut für Physik, von Igor Savić Chemnitz, 2004

• 143 Seiten inkl. 1 Publikation • in englischer Sprache mit 13 Tabellen und 44 Abbildungen

Referat Unter Verwendung von zwei Speicherapparaturen wurden ausgewählte, astrophysikalische wichtige Ionen - Molekülreaktionen untersucht. Durch die Kombination einer Kohlenstoffquelle mit einem Ionenspeicher, in dem so Reaktionen zwischen Ionen und Kohlenstoffmolekülen oder -atomen untersucht werden können, wurde Neuland betreten. Es werden Ergebnisse vorgestellt für die Reaktion von D3

+ Ionen, die in einem Ringelektrodenspeicher gefangen sind, mit einem Strahl von heißen Cn (n = 1, 2, 3). Die gemessenen Ratenkoeffizienten sind nur halb so groß wie die Werte, die in astrophysikalischen Modellen verwendet werden. Um die Kenntnis über alle möglichen Reaktionen, bei denen drei C-Atome beteiligt sind, abzurunden, wurden zwischen 15 K und Zimmertemperatur die Reaktionen zwischen C3

+, C3H+ und C3H3

+ Ionen mit H2 und HD in vielen Details untersucht. Diese Experimente wurden in einer zweiten Apparatur durchgeführt, in der ein temperaturvariabler 22-Polspeicher das zentrale Element ist (VT-22PT).

Berichtet werden Ergebnisse zu reaktiven Stößen, zur Deuterierung von Kohlenwasserstoffen und zur Strahlungsassoziation. In der Diskussion bleibt offen, was - in Verbindung mit der von 300 K zu 15 K zunehmenden Lebensdauer - der Grund dafür sein kann, daß die Bildung des exothermen Produkts C3H

+ anwächst. Der Tunneleffekt scheidet aus. Bei der Reaktion C3

+ + HD wurde ein Isotopeneffekt beobachtet, das C3D

+ Produkt wird etwas häufiger gebildet als C3H+. Ein Vergleich der

Reaktion zwischen C3H+ Ionen mit HD bzw. H2 zeigt, daß das deuterierte Molekül

wesentlich reaktiver ist. Es wurden Ratenkoeffizienten für die Strahlungsassoziation von H2 Molekülen mit C3H

+ und erstmals mit C3+ Ionen gemessen. Die Auswertung der

Daten zeigt, dass der Prozeß langsamer abläuft, wenn der neutrale Stoßpartner deuteriert ist. Schließlich wurde experimentell die theoretische Vorhersage überprüft, dass C3H3

+ keinen H - D Austausch mit HD eingeht. Eine sorgfältige Analyse aller konkurrierenden Prozesse ergab, dass bei 15 K der Ratenkoeffizient kleiner als 4 × 10-16 cm3s-1 ist.

Schlagwörter Ionenfallen, Ionen - Molekülreaktionen bei tiefen Temperaturen, H-Atom-Transfer, Deuterierung, Strahlungsassoziation, kleine Kohlenstoffcluster, interstellares Medium, Laborastrochemie, interstellare Moleküle: CD+, C2D

+, C3+, C3H

+, C3D+, C3H2

+, C3HD+, C3H3

+, C3H2D+

4

Abstract Using ion-trapping techniques, selected laboratory experiments on ion-molecule reac-tions of astrophysical interest have been performed. For the first time a carbon beam source has been integrated into an ion trapping machine for studying collisions between ions and neutral carbon atoms and molecules. Results are presented for the interaction of D3

+ ions stored in a ring-electrode trap (RET), with a beam of hot neutral carbon molecules, Cn (n = 1, 2, 3). The measured reaction rate coefficients are up to a factor two smaller than values presently used in astrophysical models. In order to complete our knowledge about the ion chemistry involving three carbon atoms, detailed investiga-tions of reactions of C3

+, C3H+ and C3H3

+ with H2 and HD have been performed be-tween 15 K and room temperature. These studies have been performed in a second ap-paratus, a variable-temperature 22-pole trap machine (VT-22PT).

Results include reactive collisions, deuteration and radiative association. It is discussed in connection with the increase in lifetime of the C3

+ + H2 collision complexes with fal-ling temperature, what could be responsible for producing more C3H

+ at 15 K. Tunnel-ing is excluded. In C3

+ + HD collisions an isotope effect has been detected, the C3D+

product ions being slightly more abundant than C3H+. Comparison of the reaction of

C3H+ primary ions with HD and H2 gas revealed that the deuterated molecules are sig-

nificantly more reactive. The process of radiative association of C3H+ and for the first

time of C3+ with hydrogen molecules has been observed. An analysis of the data shows

that radiative association becomes slower, if the neutral reactant is deuterated. Finally, the theoretical prediction from ab initio calculations that C3H3

+ does not exchange an H for a D in collisions with HD, has been proven in an ion trap experiment. Careful analy-sis of all competing processes allows the conclusion that the rate coefficient is smaller than 4 × 10-16 cm3s-1 at 15 K.

Key words ion-trap, low and high temperature ion-molecule reactions, hydrogenation, deuteration, radiative association, small neutral carbon clusters, laboratory astrochemistry, interstel-lar medium, interstellar molecules, CD+, C2D

+, C3+, C3H

+, C3D+, C3H2

+, C3HD+, C3H3+,

C3H2D+

5

CONTENTS 1 INTRODUCTION 7

1.1 The interstellar medium 7 1.1.1 Physical conditions in the interstellar medium 7 1.1.2 Processes in the interstellar medium 7 1.1.3 Chemistry of the interstellar medium 9

1.2 Overview 11

2 CARBON IN ISM 13 2.1 Observations 13

2.2 Carbon chemistry in ISM 17

2.3 Deuterium fractionation 25

2.4 A model for growing pure carbon chains 27

3 EXPERIMENTAL: ION TRAPPING APPARATI 31 3.1 Ion molecule reactions at variable temperatures 31

3.2 Description of the machines used 34 3.2.1 General overview 35 3.2.2 The two ion traps: VT-22PT and RET 44 3.2.3 Selected tests and outlook 47

3.3 The carbon source 50 3.3.1 Properties of carbon 51 3.3.2 Technical description 55 3.3.3 Test measurements and outlook 57

4 SUMMARY, CONCLUSION AND OUTLOOK 61

APPENDIX

A Ion-trapping apparatus for studies on reactions between ions and neutral carbon species 63

B Reactions of Cn (n = 1 - 3) with ions stored in a temperature-variable ra-dio-frequency trap 71

C Low-temperature experiments on the formation of deuterated C3H3+ 87

D Temperature variable ion trap studies of C3+ and C3H

+ + H2 and HD 103

REFERENCES 125

SELBSTSTÄNDIGKEITSERKLÄRUNG 135

CURRICULUM VITAE 137

LIST OF PUBLICATIONS AND CONFERENCE CONTRIBUTIONS 139

ACKNOWLEDGEMENTS 143

6

7

1. INTRODUCTION

1.1. The interstellar medium (ISM) The space between the stars is not empty but filled with diffuse material, the so called interstellar medium (ISM). This material plays an important role, especially as it also provides the matter from which new stars are formed. In general it consists of atomic and ionized hydrogen, molecular gases, dust and electrons. Each of these constituents has quite complicated velocity distributions and the number density can change over a wide range of scales.

In the early expansion of the universe the diffuse matter became dominant. Due to grav-ity, gas clouds started to collapse and to coagulate. In almost completely inelastic gas cloud-cloud collisions, energy of motion was transformed into heat through shocks and then radiated into space. These shocks are highly compressive and lead to formation of cold and dense phases of the ISM. These phases are necessary for star formation which occurs when the densities become sufficiently high. In the early time of the universe, all of these processes lead to the formation and later to the evolution of galaxies. Today, the ISM in galaxies is in a dynamical equilibrium concerning mass and energy flow. One of the most important evolutionary processes in the universe is the transformation of the ISM into stars which in their later stages enrich the gas with heavy elements. A fraction of these elements can condense into cosmic dust (submicron- and micron-sized solid particles).

1.1.1. Physical conditions in the ISM

As is mentioned in [ger99a], the ISM in our galaxy has a total mass of 6 × 109 solar masses and it is characterized by an average density of 1 hydrogen atom per cm3. It con-sists of various phases with enormous differences in their thermodynamic properties. Most of the total mass of the ISM is contained in cold interstellar clouds (nebulae). There are two different types of such clouds: dense molecular clouds and diffuse clouds. Molecular clouds consist mostly of molecular hydrogen and they are characterized by a kinetic temperature of 10 – 15 K and gas densities between a few hundred and 107 hy-drogen molecules per cm3 in their dense cores. In such clouds hot molecular cores can be found which are associated with star formation having temperatures up to 200 K. Diffuse clouds consist mainly of atomic hydrogen and they have an average temperature of 80 K and number densities of 50 hydrogen atoms per cm3. In addition, there are pro-toplanetary disks and outflows from evolved stars which have even higher densities and temperatures.

Many molecular lines and continuum radiation are observed at infrared, millimeter and radio wavelengths. This information is used for determining physical conditions such as temperature, column density, particle density or magnetic fields of ISM.

1.1.2. Processes in the ISM All constituents of the ISM (atoms, molecules, ions and grains) play an active role in its evolution. All of them and also the radiation field interact with each other in a very complex manner. The variety of molecular processes can be divided in two groups. One group contains physical processes such as heating and cooling and the other group chemical processes. Heating and cooling is based on emission of energetic photoelec-

1. INTRODUCTION

8

trons, emission from excited rotation of fine structure states and interactions with grains. A few typical heating and cooling processes are summarized in Table 1.1 [dis99].

Table 1.1. Heating and cooling processes. Photoelectric heating Grain/PAH + hν → grain/PAH+ + e- Cosmic ray heating H2 + cosmic ray → H2

+ + e- CO line cooling CO(J) + coll. → CO(J’) + hν O line cooling O(3P2) + coll. → O(3P1) → O(3P2) + hν C+ line cooling C+(2P1/2) + coll. → C+(2P3/2) → C+(2P1/2) + hν Gas grain heating / cooling Gas + grain → gas’ + grain’

As can be seen from this table, carbon has a very important role in the physical evolu-tion of the ISM. In general, diffuse clouds are heated by the absorption of starlight. Car-bon is the main supplier of free electrons in diffuse clouds and contributes therefore to the heating of the interstellar gas. The energy of the interstellar radiation field also can be transferred to the gas via the photoelectric effect on dust grains. Far ultraviolet pho-tons ionize grains and the ejected fast electrons heat the gas via inelastic collisions. The heating rate cannot be directly measured, but can be estimated from observing the emis-sion from C+ (see below). The comparison of this emission feature with the far-infrared emission allows an estimate of photoelectric heating efficiency of ~0.029 for translu-cent, high-latitude clouds [juv03].

Diffuse clouds are cooled by emission of photons at far-infrared and submillimeter wavelengths. As can been seen from Table 1.1 one of the major cooling processes is due to a carbon containing molecule, carbon monoxide (CO). The most abundant form of carbon gas in cold neutral medium, the C+ ion, can be easily excited by collisions around 80 K. Besides the hyperfine state of the hydrogen atom, the 2P3/2 state of C+ is the first excited state of all gas-phase constituents of the cold neutral medium. Because of this, the 2P3/2 -

2P1/2 emission of C+ at 158 µm is consider to be the major cooling mechanism of the cold neutral medium [wol95], [ing02]. That atomic carbon can con-tribute significantly to the thermal budget of interstellar gas has also been discussed in detailed in [ger99b]. In the case of our galactic ISM, C cools more than CO [but92], [wri91]. In addition, low lying transitions of neutral (C) and multiply ionized (Cn+) atomic carbon are important cooling channels of the warm interstellar gases. Also, C is claimed to be the major coolant in intermediate-velocity clouds [hei01].

The chemical processes occurring in the ISM are usually separated into neutral reac-tions, ionic reactions and reactions occurring on grain surfaces. A more detailed group-ing can be found in Table 1.2 [dis99].

Table 1.2. Chemical processes. Ion-molecule reactions X+ + YZ → XY+ + Z Charge transfer X+ + YZ → X + YZ+ Neutral-neutral reactions X + YZ → XY + Z Radiative association X + Y → (XY)* → XY + hν Grain surface formation X + Y + g → XY + g Associative detachment X- + Y → XY + e- Photodissociation XY + hν → X + Y Dissociative recombination XY+ + e- → X + Y Collisional dissociation XY + M → X + Y + M

1. INTRODUCTION

9

1.1.3. Chemistry of the ISM Until today more than 120 interstellar molecules have been detected including neutrals and ions. Many of them may have rather complex structures, e.g. one assumes that polycyclic aromatic hydrocarbons (PAHs) survive. One can expect that many more will be found because of the wide variety of processes occurring in the ISM as well as the variety of physical conditions like temperature, number density, radiation fields, ener-getic cosmic particles, shocks, etc.

In general, ion-molecule reactions have no barriers ([ger93], [smi93]) and therefore their reaction rate coefficients are usually larger than those for neutral-neutral reactions. This is the reason why early reaction models (since the seventies, e.g. [her73], [bla73]) describing molecular clouds and accretion disks were already quite successful using mainly ion chemistry. At low densities and temperatures of the ISM, radiative associa-tion is also a very important reaction [ger92a]. An important class of reactions is the recombination of ions with electrons because they are the final step in the synthesis of neutrals coming from ion-molecule reactions. Today it is generally accepted that also neutral-neutral reactions can play an important role in the evolution of the ISM, espe-cially in the case of radicals. It has been shown experimentally that there are processes with large rate coefficients at low temperatures [bro97], [cha98], [kai98], [cha00], [cha01], [kai02].

In contrast to ion-molecule and neutral-neutral reactions which are, at least in general, reasonably well understood, surface reactions occurring on cold icy grains are very complex. Modern chemical networks, describing the ISM, take these processes into ac-count; however, using rather crude models in most cases. The most discussed example is the formation of hydrogen molecules via a catalytic reaction. It is commonly assumed that the two H atoms have to meet on a dust particle [hol71]; however, there is no real proof for that. In addition, it has been postulated that grain surfaces play an important role in forming some of the hydrogenated molecules [tie82], [her93].

The ion-molecule chemistry of clouds is very complicated. Chemistry of diffuse clouds begins with two different ionization processes, the ionization of H, H2 and He by cosmic rays and photoionization of C by stellar UV radiation [smi92]. In dense clouds, which are well shielded from the UV radiation, ionization is dominated by the cosmic rays. The initially formed H+, H2

+ and He+ cations are the start of a complex network of reac-tions leading to the formation of a variety of complex molecules – organic and inorganic ones. Fig. 1.1 illustrates a few steps of the chemistry occuring in diffuse and dense clouds. Inspection of the inventory list of detected interstellar molecules (see http://www.cv.nrao.edu/~awooten/allmols.html) reveals that almost 80 % of the mole-cules are carbon containing, most of them hydrocarbons and cyanopolynes. The largest molecule detected so far in the ISM is the thirteen atom HC11N. In addition, carbon is one of the most abundant non-volatile elements in the universe [hen98]. Its cosmic abundance is comparable to that of oxygen and nitrogen. Because of its high reactivity, high abundance (fourth most abundant element in ISM) and ability to form complex molecules with other reactants, carbon is of paramount interest for astrochemistry and astrophysics in general. Detailed studies of how the different forms of carbon ranging from atoms through complex molecules to grains are processed in ISM, will lead to a better understanding of many fundamental processes and of the evolution of the uni-verse. Some more details of the carbon and hydrogen based interstellar chemistry will be discussed below.

1. INTRODUCTION

10

For understanding in detail of the evolution of the matter of the ISM, a concerted effort of astronomers, physicists and chemists is required to coordinate astronomical observa-tion, astrophysical modeling, laboratory experiments, and basic theory. There exist al-ready very detailed astrophysical models where the parameters can be tuned to explain specific observational facts; however, they should be improved. They must be based on a detailed knowledge of ionization processes, chemical reactions and grain evolution processes. Basically, two different classes of gas-phase models exist: steady-state mod-els and time-dependent models. Steady-state models are depth dependent models. In these models abundances of the molecules do not change with time but with distances. Time-dependent models are depth independent. Molecular abundances are calculated as a function of time at a given position deep inside the cloud.

H3+

H2+

H+

He+

H2

He

C+

H2

CO

C

H2

CH+

CH2+

CH3+

H2cosmicrays

H2

H3+H2

+H2

He+Hecosmic rays

CO

C

H2

CH+

H2

CC+

u.v. radiation CH3

+H2CH2

+H2

CH CHH2

C2H2+C2H+C2

+

Diffuse clouds:

Dense clouds:

H2

Fig. 1.1. First steps of chemistry in diffuse and dense clouds.

Concerning gas-grain models, there exist basically two different approaches [dis98]: the “accretion-limited” regime and the “reaction-limited” regime. In the first approach, the time for a mobile species to scan the surface is much shorter than the accretion time of the other reactant. Therefore, reactions are limited by the accretion rate of new species. In the second case, a species is trapped in a site and can react only with a migrating molecule that visits that site. The next critical action in gas-grain models is how mole-cules return into the gas phase. If there would be no desorption included, then, for typi-cal dark cloud with molecular densities of about 104 cm-3, most molecules would disap-pear from the gas phase in less then 106 years. This is inconsistent with observations. Therefore there must be, also in dark clouds, some efficient desorption mechanisms, e.g. thermal evaporation by cosmic ray spot heating, explosive heating due to exothermic

1. INTRODUCTION

11

reactions between radicals which can be triggered by cosmic rays or by grain-grain col-lisions at velocities greater than 0.1 km/s. For details see [dis98] and references therein.

1.2. Overview In the following Section 2.1, an overview of observational facts is presented in order to emphasize more on the importance of carbon in the ISM. The Section 2.2 describes in more detail the gas phase interactions of hydrogen, carbon and hydrocarbons which are of importance for the physical and chemical evolution of the ISM. In Section 2.3, recent observations and models are mentioned which have shown that the incorporation of D instead of H atoms is also a universal tracer for the interstellar environments and there-fore also of importance for understanding the fundamental microscopic processes. Sec-tion 2.4 describes a model for growing pure carbon chains.

With the aim to improve our knowledge in these fields, selected laboratory experiments have been performed in this thesis. Several measurements have been made utilizing two different ion trapping machines which allow one, on one side, to simulate the conditions prevailing in cold interstellar clouds and to study isotope enrichment in collisions with HD, and on the other side, to provide a temperature variable environment and a carbon target for understanding the interaction of ions with carbon atoms or molecules under inter and circumstellar conditions.

After a short overview of the experimental methods used in the field of laboratory astro-chemistry given in Section 3.1, the common features of the two instruments used in this work are described in 3.2. Since it was the first time that a carbon beam source was in-tegrated into a trapping machine, the details of this combination are described in the separate Section 3.3 and separate publications. First results have already been published [cer02], and the paper is reproduced in Appendix A. A manuscript of a detailed publica-tion describing the apparatus and results for the deuteron transfer reaction D3

+ + C3 → C3D+ + D2 and reaction C3

+ + C3 can be found in the Appendix B.

Dedicated investigations of reactions between simple carbon ions with H2 and HD have been performed in the Variable Temperature 22-Pol Trap apparatus (VT-22PT). The rate coefficients for forming deuterated C3H3

+ are given in Appendix C. Interesting new results for the hydrogenation of C3

+ and C3H+ ions reacting with H2 have been obtained

at different temperatures. They are presented in Appendix D.

The papers printed in the appendices B, C and D are manuscripts which have been sub-mitted or are ready for submission.

12

13

2. CARBON IN THE ISM

As emphasized in the introduction, the detailed understanding of the evolution of ISM needs many interdisciplinary, concerted efforts, ranging from astronomical observa-tions, astrophysical modeling, laboratory experiments to basic theory.

It is certain, that in astrochemistry, all forms of carbon play a special and very important role and therefore this work concentrates on this subfield i.e. focuses on experimental studies of ion-molecule reactions involving carbonaceous structures. In order to prepare dedicated experiments such as those of this thesis in detail, it was important to collect the actual knowledge in this interdisciplinary field and, in a few special cases, it became necessary to get a deeper insight in order to recognize current problems and try to con-tribute to their solution. This is the reason why the following part contains a collection of facts and various aspects of carbon in ISM and its role in astrochemistry. The major aim of this thesis was to set up a new experiment which allows to perform, for the first time, gas-phase processes between a beam of neutral carbon molecules and stored ions over a wide range of temperatures. The following overview describes the proposed routes for formation and destruction of carbon containing molecules and molecular ions under the extreme conditions prevailing in space and gives some hints to basic theories.

2.1. Observations Pure carbon molecules, carbon containing molecules (hydrocarbons, polyaromatic hy-drocarbons – PAHs, polyynes), carbon nanoparticles, and hydrogenated amorphous car-bon – HACs, are proposed to be responsible for many absorption and emission features observed in ISM. Many carbon containing molecules and isomers have already been identified in the ISM. They have been detected using both emission and absorption spectroscopy. In general, the molecules usually have electronic transitions in the visible or the UV range, vibrational transitions in the infrared and rotational transitions in the far infrared and the radio frequency range.

Atomic carbon has an ionization potential of 11.3 eV. Since this is below the Lyman edge, C exists in the ionized state everywhere in the ISM with the exception of dense clouds, which are shielded from UV radiation.

Emission and absorption lines of atomic carbon which are commonly used for detection and for providing information about the physical conditions in particular regions, are summarized in the Table 2.1. A few typical regions of the ISM where these lines are detected are mentioned in the last column. As presented in [tat99], atomic carbon can be used as a temperature probe in dark clouds and this atom is, even today, still in the focus of many observations. For example, the fine-structure line of the neutral carbon atom (3P1-

3P0) has been used for mapping clouds [ben03]. High-quality absorption spectra of interstellar C have been recorded for deriving gas densities and temperatures of diffuse clouds [zsa03].


14

Table 2.1. Emission and absorption lines of atomic carbon commonly used for their detection and for deriving information on the physical conditions prevailing in particular regions.

Emission lines

C 370 µm (3P2-

3P1) 609 µm (3P1-

3P0) Deep in PDRs* [hol97], [but92]

C+

- 157.74 µm (2P3/2-2P1/2)

- 103.7 nm - radio recombination lines: C92α, C110α, C166α

PDRs, diffuse neutral ISM [hen98], [hol97] Low-density hot phase of ISM [hen98], [kor98] PDRs [hen98], [gar98]

C2+ 97.7 nm Low-density hot phase of ISM [hen98], [kor98]

C3+ 154.8 nm 155.1 nm

Low-density hot phase of ISM [hen98], [kor98], [mar90]

Absorption lines C Multiplets in:

115-120 nm range 127-129 nm range near 132.9 nm

Cool neutral gas

C+ - 133.5 nm resonant line - 232.5 nm (2P1/2-

4P1/2) Diffuse clouds

* PDRs - photodissociated regions

The carbon chains with various terminal groups are conspicuous because of their high interstellar abundance [gie01], [tha99]. They can be separated into two different forms, according to their electric dipole moment: those with a permanent electric dipole mo-ment and nonpolar chains. Polar species, for example unsaturated polyynes or linear carbon chain radicals (CnH, n = 2, ..., 8), can be easily detected by means of their rota-tional spectra, whereas pure carbon chain molecules are nonpolar and exhibit no pure rotational spectrum. For this reason, they are more difficult to observe. They can be detected through other infrared transitions [gie01]. For example many carbon chain molecules have low-energy bending vibrations emitting in the infrared. It is expected that all pure carbon chain molecules with the exception of C2 have bending vibrations in the far-infrared region (terahertz domain). In this way the linear C3 molecule has been detected [gie01].

The smallest pure carbon molecule is C2. It has been observed in the regions towards a number of reddened early-type stars. Various absorption lines from rotational levels of the electronic ground state of C2 have been detected. Because C2 has no dipole moment, the rotational states are populated via a competition between collisional excitation and de-excitation, and radiative excitation through absorption into excited electronic states followed by infrared cascades [cha80], [dis86]. Many of the early ([lut77], [sou77], [cha78], [hob79], [cha80], [hob81], [hob82], [dan83], [dis84], [dis86]) but also some of the more recent observations [bak96], [bak97], [bak98a], [cra00], [gre01] have been performed in the Phillips band system which are in the far-red and near-infrared and which can be observed from Earth. Also the Mulliken bands [sno78], [lam95] and the Swan bands [bak96], [bak97], [izu03] have been used for detecting C2, as well as ground-based optical spectroscopy [hri91], [hri95].

In the last decades, efforts have been made to detect the triatomic molecule C3 in the ISM. A negative result was reported in [cle82]. Later, an upper limit of 1010 for the col-umn density of C3 or about 10-11 cm-2 with respect to hydrogen was been established in [sno88] for a few diffuse clouds. In [haf95] and [ord95] C3 molecule have been tenta-tively identified. Using infrared spectroscopy, C3 was observed in circumstellar shells


15

[hin88] and by millimeter measurements of its low-frequency bending mode in a dense interstellar cloud toward Sagittarius B2 [cer00]. Clear proof of C3 in three diffuse clouds (ζ Ophiuchi, 20 Aquilae and ζ Persei) has been reported recently by observing the 405.2 nm electronic band system [mai01]. Observation of the same band toward HD 210121 and a detailed excitation model was reported in [rou02]. In addition, the absorp-tion bands of C3 toward 15 translucent sight lines has been seen very recently. The de-rived C3 column densities are well correlated with the corresponding C2 column densi-ties, with [C2]/[C3] ~40, indicating that their chemical generation pathways are probably closely related [oka03]. Very recent observations of C3 were reported in [ada03].

Only in the last years, laboratory gas phase electronic spectra of longer pure carbon chains like C4 and C5 became available enabling direct comparison with astronomical data [mai02]. While C4 has not yet been found in ISM, C5 has been detected in the infra-red spectrum being so far the longest pure carbon chain observed [ber89]. There were several attempts to detect C4 but all of them were negative [gal01], [mai02], [gal02]. Based on the sensitivity, upper limits for the C4 column density have been derived vary-ing between 4 × 1012 cm-2 and 1013 cm-2 and C5 column density of 2 × 1011 cm-2 [mai02]. In [gal02] a two times lower limit for the C5 column density than in [mai02] was re-ported.

In addition to pure carbon chain molecules, a large number of carbon-bearing molecules have been detected in space, including simple diatomics and polyatomics, complex un-saturated acetylenic radicals CnH, cyanopolyynes HCnN, carbynes and carbon chains of type CnHm (n<m) and PAHs. Many of these molecules were seen in different isotope forms. Of special importance for astrophysics and physics in general are the problems connected with deuterium and deuteration processes of carbon-bearing molecules in ISM. Deuteration of a few selected hydrocarbon ions is one of two major parts of this thesis (see 2.3 and Appendix). Carbon chains and rings are presently in the focus of many experimental and theoretical investigations. Since, in addition, detection capabili-ties of new observational instruments are becoming better and better, it is to expect that many new complex molecules will be detected in the near future [tha99], [tak00]. An excellent summary about carbon chemistry in interstellar clouds has been given very recently [ger03a]. If one plots the fraction of carbon atoms bound into molecules of various sizes as a function of the size, one obtains first a steep decrease for small mole-cules followed by a flatter dependence for molecules with 3 or more carbon atoms. From this it has been concluded that detection of large carbon clusters should be possi-ble.

In addition to the large number of molecules identified in ISM via their specific emis-sion and absorption features, there are many spectral features observed in ISM which still await assignment. These are the diffuse IS absorption bands (DIBs - ranging from the near ultraviolet to the near infrared) superimposed on the extinction curve, discrete visible emission bands (RRBs), and discrete unidentified IR emission bands (UIR). There are some arguments that these unidentified features are due to transitions in car-bon-containing molecules. For example it has been shown recently that some of the DIBs are strongly correlated to the presence of interstellar C2 molecules [tho03]. In the last decades, many papers on the DIBs and their possible carriers have been published, see for example the excellent review [her95]. A few years ago some almost convincing coincidences between the spectra from anions of pure carbon chains (C6

-, C7-, C8

- and C9

-) and DIBs have been found [tul98], the most prominent candidate C7-; however, this

hypothesis was rejected finally [mcc01].


16

Also PAHs are included in the manifold of molecules which are believed to be respon-sible for the UIR and DIBs (see for example [lat91], [her95] and [hen98]). It has been proposed that the major source of PAHs may be the mass-loss winds of asymptotic gi-ant branch (AGB) carbon stars. PAHs consist of fused six-member benzenoid rings of sp2-hybridized C atoms and a certain number of H atoms bound to the outer carbon at-oms of the molecule. The π electrons delocalized over the C skeleton in an aromatic way leads to the high photostability and to the assumption that these molecules survive in ISM. Many details about PAHs and ion-molecule reactions in interstellar and circum-stellar chemistry can be found in [boh92]. An interesting model of hydrogenation of PAHs and their charge states ranging from benzene up to a variety of species in diffuse clouds has been published recently [pag03]. In the diffuse medium, PAHs can undergo a variety of hydrogenation and dissociation processes. The most critical process influenc-ing the PAHs in the diffuse ISM is photodissociation. It has been found that the hydro-genation state strongly depends on the size of the molecule. Small PAHs with less then 15-20 carbon atoms are destroyed in most environments. Intermediate size PAHs with 20-30 carbon atoms may be able to survive in the ISM. Larger PAHs have primarily normal hydrogen coverage but they can have also additional hydrogens. Very large PAHs may be fully hydrogenated with every peripheral carbon atom bonding to two hydrogen atoms. In addition, it has been found that extremely dehydrogenated PAH neutrals and also CmHn

+ ions (15 ≤ m ≤ 30, n ≤ 2) can survive in ISM [pag03]. In PDRs, the charge state of the PAHs is determined by the balance of photoelectric ejection and recombination of the ions with electrons [bak98b], i.e., by the UV flux and the overall gas density. Calculations [bak98b] show that the presence of PAHs in PDRs strongly influences the gas phase abundance of neutral atomic carbon, sulfur, silicon and metals. For example, the peak carbon atom line intensity is increased by factor of 2 if PAHs are present.

The fact that cosmic dust particles range in size distribution from nanometer to mi-crometer has been concluded from a detailed analysis of the IS extinction and polariza-tion curves at UV, visible and NIR wavelengths and from thermal IR and submillimeter and millimeter radiation. For many years these interstellar grains have been in the focus of interest as can be seen from the large number of theoretical and observational inves-tigations. A very recent theoretical treatment of the structure and size distribution can be found in [cla03]. Other interesting studies of the composition of dust and gas, their structure and size distribution as well as elemental abundances and mass densities in local interstellar medium, has been published recently [kim03a] [kim03b].

Features which are believed to originate from carbonaceous solids are (i) the UV ab-sorption band at 217.5 nm in the galactic extinction curve; (ii) the UV absorption band between 240 nm and 250 nm in the spectra of hydrogen deficient objects; (iii) the ex-tended red emission (ERE) between 650 nm and 700 nm in the spectra taken from re-flection nebula, HII regions, planetary nebulae and diffuse ISM; (iv) the broad emission plateau between 6.9 µm and 9 µm; (v) the 3.4 µm absorption feature in diffuse ISM and hypergiant; and (vi) the broad feature between 11.0 µm and 11.5 µm in spectra of car-bon-rich stars and planetary nebulae [hen98]. It has been proposed that solid carbon material can exists in the ISM in forms of graphite, diamond, fullerite solids, carbyne and noncrystalline materials. In [all92] it has been suggested that the diamond-like structure is the carrier of the 3.47 µm band and that a few percent of the available cos-mic carbon is tied up in this material. There are arguments [dul01a] that the ERE is likely to be emitted from particles with ~50 carbon atoms and that these particles are not strong emitters at 3.3 µm. The presence of ~70 to ~120 carbon atom clusters can be


17

connected with recently detected 1.15 µm and 1.5 µm ERE bands. It has been shown in detail [dul01a], that one can construct a realistic model, in which the emission and ab-sorption in IS sources are basically determined by carbon nanoparticles. Accounting for ERE, aromatic infrared (AIR), and UV extinction (217.5 nm band), the energy balance is fulfilled by using absorption and emission just from this material. It is interesting to note that a broad continuum, similar to the plateau emission, is obtained from infrared emission of 1200 K HAC (hydrogenated amorphous carbon) in the region between 3.0 µm and 3.7 µm [dul01b].

Formation of nanodiamond material in interstellar clouds has been proposed to be initi-ated by hydrogenation of PAHs followed by exposure to ultraviolet radiation in the presence of oxygen [dul01b]. The presence of HACs in extreme carbon stars, proto-planetary nebulae and planetary nebulae is deduced in [gri01] from a comparison of observed and laboratory spectra with emission features at 21 µm, 27 µm, and ~33 µm. Recently, infrared absorption spectra of deuterated amorphous carbon (D:HAC and DAC) have been obtained which can be identified via a special feature near 4.6 µm [gri03]. In addition, it has been suggested [gri03], that it may be possible to detect deu-terium fractionation in carbonaceous dust via absorption features in the CDn stretching band.

2.2. Carbon chemistry in ISM Interstellar molecules are dominated by carbon-containing structures. The reason for this rich C-based chemistry and for the dominance of these atoms is, in general, its abil-ity to form three different types of hybridized orbitals which are presented in Table 2.2 [hen98]. In addition to these three simplified representations, more complex arrange-ments are possible, for example the three center - two electron bonding the simplest example being the hypercoordinated carbocation CH5

+. Therefore, all forms of neutral and ionic carbon atoms and carbon containing molecules, fullerenes and probably also forms which have yet to be observed play a role in astrochemical models.

Table 2.2. Carbon hybrid orbitals. Atomic orbitals Resulting hybrid orbitals

2s, 2px,y,z Four equivalent sp3 tetrahedral orbitals 2s, 2px,y, 2pz Three planar sp2 orbitals and one perpendicular p orbital (2pz) 2s, 2px, 2py,z Two linear sp orbitals and two perpendicular p orbitals (2py,z)

One of the basic problems in astrochemistry is to find chemical pathways leading from small molecules like C, C2, C3, C2H or C3H to those complex molecular structures which are suspected to be synthesized in inter- or circumstellar regions. Already in the first steps from C to CH or CH2 unusual reaction mechanisms such as radiative associa-tion had to be postulated. Recently it has been pointed out [ger03a], that also for small molecules such as C4H and C3H2 our present understanding of chemical networks is not sufficient and that new formation routes must be found to explain their large abun-dances observed in PDRs. Another important question is what the most probable struc-tures of medium-sized carbon containing clusters are [pas00].

To describe formation and destruction of molecules and ions in different environments of the ISM, ranging from dark, cold nebulas, PDRs, shock chemistry, chemistry of at-mosphere of carbon rich stars and stellar outflows, many complex chemical models have been developed. Despite their being already rather complex, all models describing particular regions contain a rather limited number of global ("characteristic") parameters


18

such as gas composition and number density, physical conditions like temperatures ranging from around 10 K in dark nebulas to several thousand K in stellar atmospheres, and one or two parameters describing the influence of the radiation field. It must be questioned whether all important processes and steps are really included or whether some fundamental pathways have been overlooked.

It has been generally accepted that elemental carbon is the first of the heavier elements which is (almost) exclusively formed in the interior of stars via the 3α process and dredged up from the helium-burning shell during the thermal-pulsing phase of the as-ymptotic giant branch (AGB). The standard inflationary cosmology model predicts, that the elemental composition of the matter of the earliest stars and galaxies consists of 1H and 4He with admixtures of deuterium 2H , lithium 7Li and 3He. Note, that there are some hints that traces of 12C may have been existing already at early times [har03]. This information has been derived from the fact that all quasars and galaxies (including those at very high redshift) show spectra of heavy elements as do the oldest known Galactic stars [har03]. If carbon or/and oxygen would have been existing already in the early universe, the formation of heavier elements in the first generation of massive stars might have been facilitated.

Fig. 2.1 is a sketch of what kind of typical carbon structures and molecules formed in the outer atmosphere of a carbon-rich star at two different temperatures, 2000 K and 5000 K. The carbon to hydrogen ratio was fixed to 1. This figure which has been slightly simplified, has been taken from Pascoli and Polleux [pas00]. In their simulation of the outflow, they include about thousand different species and varied the temperature from 2000 K to 7000 K, a range which is typical for stellar atmospheres. With one more additional parameter, the carbon to hydrogen ratio, a rich and quite variable carbon based chemistry has been obtained.

T = 2000 K

T = 5000 K

r = 1013 cm

Rings Fullerenes

Small PAHs (fully hydrogenated)

Medium sized PAHs(regular and strongly dehydrogenated)

Small diamond-like structures

Large PAHs (fully hydrogenated)

GrainsCarbon-rich star

v0

Chains

Fig. 2.1. Illustration of typical carbon molecules and clusters formed in outer atmosphere of a carbon-rich star [pas00].

Cool stellar outflows are efficient environments for producing complex carbon bearing structures like PAHs, soot particles and grains. The material moving outwards, cools from around 5000 K down to 50 K. During the expansion it is diluted from 1012 to 106 particles/cm3. At temperatures below 1000 K where the gas densities are still ~1010 - 108 cm-3, effective molecular growth and cluster formation occurs [pat95]. The


19

kinetics are complicated by the radiation pressure acting on dust grains in the outflow of cool stars. By absorbing UV or visible photons, they can gain enough momentum that atoms can be removed from the surface of dust grain in energetic collisions. This effect of sputtering in the outflows of cool stars has been modeled and the size distribution of dust grain has been presented in [cov00]. Supersonic protostellar outflows can also lead to shocks which compress and heat surrounding gas to temperatures higher than 2000 K. At such high temperatures many neutral-neutral reactions can become very efficient [cod01]. An interesting model of chemistry taking place in stellar jets which penetrate into molecular clouds, is given in [wil02]. The interaction leads to a clump which is generated in the jet-cloud shock and which is exposed to an intense radiation field.

In the case of AGBs, the elemental carbon C abundance which is produced and dredged up from the helium burning shell, can exceed that which is formed in oxygen and car-bon stars. The low temperature environment allows carbon based molecules to be formed. They are ejected as part of the stellar wind. This wind depletes the hydrogen envelope and causes the star to leave the AGB state and to evolve to higher tempera-tures. In this way stars enter the protoplanetary nebula (PPN) phase or post-AGB phase. When stars reach high enough temperatures, they emit a high flux of UV photons and their circumstellar envelope will be ionized. In this way, planetary nebulae (PN) are created [olo96a], [olo96b], [kwo99], [her01], [has03]. In these regions, despite the harsh conditions, carbon-based molecules are created and have been detected.

The coldest places in the ISM are dense molecular clouds. The knowledge of the origin and distribution of atomic carbon in such regions is crucial for understanding their com-plicated gas phase chemistry. The problem of the gas phase distribution of carbon in molecular clouds is unfortunately still not completely understood [wal93]. Here, only a brief overview of the most important observational results and some tentative explana-tions are given.

Steady state models predict that, deep inside molecular clouds, almost all gas phase atomic carbon is locked in CO. Therefore carbon exists only in the atomic form, if there are temporal or spatial variations in the cloud conditions. The steady state conditions can be perturbed by UV radiation from young stars ionizing and dissociating CO. In this way atomic carbon can be released; however, the fractional abundance of atomic carbon can be increased only on the surface of the cloud. Deeper inside the cloud, where the UV flux is much lower, one would expect only very small amounts of atomic carbon. In contrast to this, atomic carbon has been found deep inside clouds. Detecting carbon at-oms near the ionization front of M17, it has been found [kee85] that a peak abundance occurs more than 60 mag of visual extinction into the cloud. On the other hand, it has been found in the Ophiuchus molecular cloud that C is less abundant in its deep cores than at intermediate Av’s [fre89]. Its fractional abundance remains high with around 10 mag in a cloud. The column-averaged fractional abundance ratio of C and CO is slightly less than 1 from 2 to 4 mag and decreases then by a factor of 20 at 100 mag.

There are many models trying to explain the presence of carbon in the interior of the clouds. A very interesting related discussion can be found in [kee85]. One of most ac-ceptable [wal93] explanation is that molecular clouds are very clumpy permitting UV radiation to penetrate deeper than it would be possible if they would be uniform as as-sumed in earlier models [tie85], [dis88].

The model presented in [dis88] treats a large variety of clouds ranging from diffuse clouds through translucent clouds to dense PDRs. Especially studied has been the effect of the variation of the external UV field on the depletion of gas-phase carbon and on C,


20

C+ and CO abundances. Taking an UV field similar to the mean interstellar field, the model for translucent clouds gives C and CO abundances equal to about 2 mag. The fractional abundance of C peaks at a visual extinction between 2 and 3 mag. At a visual extinction greater of 5 mag, almost all gaseous carbon is in the form of CO. In these calculations it has been assumed that gaseous carbon is depleted to 0.4 of the solar sys-tem value. The authors also discuss results assuming an UV field which is a factor 10 more intense than the mean interstellar field. A comparison of observed results for the Ophiuchus molecular cloud complex and the model from [dis88] is presented in [fre89]. In addition, this paper gives a comparison with the model presented in [tar85] in which the collapse of a cloud and time-dependent chemical evolution are incorporated. In this model the C abundance is enhanced because clouds spend much of their lifetime in a diffuse state [fre89].

One possible explanation for the large observed C abundance proposed in [lan84] is that carbon is more abundant then oxygen in dense clouds. In addition, it has been shown in [gla85] that C fractional abundance is proportional to the square of the total carbon abundance. In order to avoid that most of the residual carbon is locked onto or in grains, a mechanism for removing at least a few percent of the carbon is required in order to explain what has been observed deep in the interior of clouds.

The explanation of spectroscopic observations from the M17 SW star formation region requires the molecular cloud near the interface to be clumpy or filamentary [stu88]. This suggestion for the structure is deduced from the observed level of extended C+ emission which is 20 times higher than that expected from a single molecular cloud interface ex-posed to an UV field typical of the solar neighborhood. The clumpy structure of the photodissociation interface implies that emission from C should follow the distribution of column density of warm CO at the interface. Correlation between the emission strengths for C+ vs. CO in star-forming molecular clouds and starburst galaxies has been established [sta91], [cra85]. Another model of clumpy clouds [tau90] shows that the neutral carbon is formed throughout the cloud. In addition, this model predicts that C emission should peak weakly near the surface of the cloud, whereas CO peaks near the core but these predictions are not consistent with observations for many clouds. For the dark cloud D 478 in the Local Group galaxy M 31, it has been shown in [isr98] that both the neutral carbon abundance and the ratio between CO and C lines can be ex-plained by a two-component model based on high density cores and low-density enve-lopes.

Other possibilities for formation of atomic carbon in the interior of clouds are through high speed shocks found in molecular outflows from young stellar objects [wal93], [hol89] or by high speed neutral protostellar winds [gla91].

The neutral carbon equilibrium in the envelopes of molecular clouds is determined by two competing processes, photodissociation of CO creating C and destruction of the neutral carbon by photoionization. Typically, the fine-structure line (3P1-

3P0) is used as a tracer for transition zones of the ISM [sta97]. In addition, enhanced intensity ratios of spectral lines of C and CO are taken as tracers for “low metallicity” regions, examples being the Small and Large Magellan Clouds [sta97], [bol00]. Interesting observations of molecular bands of the carbon stars in the Large Magellan Cloud and a discussion about their “metallicity” are given in [mat02].

With the aim to follow in detail the quite complicated chemistry and understand proc-esses under different conditions of ISM, many models have been developed. Concern-ing gas-phase models, two different classes exist: steady-state models that are depth


21

dependent and time-dependent descriptions. In gas-grain models two different ap-proaches exist [dis98], the accretion-limited regime and reaction-limited one. More de-tails about these approaches can be found in Chapter 1, Section 1.1.3. These models [her73], [her83], [leu84], [her89], [mil90], [pin91], [ter98], [tur98], [tur99], [aik99], [tur00], [vit00], [rob00a], [rob00b], [mil02], [she03], include many species and thou-sands of reactions. In general, results from models which have been developed for spe-cial regions of the ISM usually describe rather well the observational results. However, this cannot be taken as a proof for the completeness and usually new and more detailed observations require modifications or, in some cases, disproof details of the model. There are many chemical routes leading to formation and destruction of complex carbon based molecules. In the following only some of the routes leading to formation and de-struction of hydrocarbons and pure carbon molecules are mentioned. Their chemistry plays an extremely important role in all networks of chemical models.

In Fig. 2.2 the main routes for the production and destruction of carbon and hydrocar-bon molecules and ions are illustrated. The huge number of possible ion-molecule reac-tions leading to hydrocarbon synthesis, may be arranged in the following groups [her89]:

a) carbon insertion or carbon fixation reactions C+ + AH+ → CA+ + H

b) carbon addition reactions CnHm+ + CiHj → Cn+iHm+j-1

+ + H → Cn+iHm+j-2

+ + H2

c) association reactions CnHm+ + CiHj → Cn+iHm+j

+

Since hydrogen is the most abundant element in ISM, hydrogenation and deuteration reactions are very important. Hydrogenation reactions of hydrocarbons can be generally divided into the two groups:

a) hydrogen abstraction CiHj+ + H2 → CiHj+1

+ + H

b) radiative association CiHj+ + H2 → CiHj+2

+ + hν

C, C+,CH4, CH3

+,C2H4

+, C3H+

hν, He+,O, C+, CH3, C2H

hν, e-,He+

C2H,C2H2, C2H2

+,C2H3, C2H3

+,C2H4, C2H4

+,C4H2 hν, e-, He+, H2,

H+, H3+, H3O+, HCO+

Cn-1+, Cn-1

Cn-1Hm+, Cn-1Hm

Cn+, Cn

CnHm+, CnHm

Cn+1+, Cn+1

Cn+1Hm+, Cn+1Hm

Fig. 2.2. Main routes for production and destruction of carbon and hydrocarbon molecules and ions.


22

The major motivation for investigating these ion-molecule reactions at low energies is their need in astrochemistry, because accurate rate coefficients are required in the mod-els described above. For a large number of reactions, only theoretical estimates are util-ized in the simulations. They are often based on crude assumptions which need to be cross checked with suitable experiments. Unfortunately the majority of experiments has been performed at room temperatures or down to the temperature of liquid nitrogen, 80 K (see Chapter 3.1). The assumption that rate coefficients are temperature independ-ent or the extrapolation of results measured at high temperatures to 10 K can lead to the big errors. It is well known that at different temperatures, different mechanisms can dominate the dynamics of the reaction, which may change completely the reaction rates. One example is the process of radiative association, another is that isotope effects are oversimplified in many cases. In order to derive reliable abundances of specific species from models, correct reaction rate coefficients are necessary. This means that reaction rate coefficients have to be measured over a wide range of temperatures, at various den-sities ranging from very low ones to the regime where ternary processes prevail and for different isotopes and isomers.

The detailed understanding of reaction dynamics at low and very low energies are also very important from a fundamental point of view. In recent years, the expression ultra-cold chemistry has been introduced, mainly in the field of laser cooled atoms. It is an interesting question what happens if the total energy varies by a small amount near zero. One of the obvious consequences is that the number of partial waves contributing to the collision complex formation, becomes smaller and smaller and finally only s-waves need to be considered. In reactions where H atoms are involved this effect plays a role already at collision energies in the meV range. In this limit, quantum-mechanical calcu-lations or also statistical models become simplified. Also dynamical effects such as resonances may play a role both in the experiments and in interstellar collisions. In ad-dition, at very low collision energies, other effects can occur because excitation of rota-tional or fine structure states can be the dominant contribution to the total energy. It is a general statement that low total energies often lead to long collision complex lifetimes enabling very unlikely processes to occur like radiative association, rearrangement via internal tunneling, or non-adiabatic mixing. In reaction systems, where several identical nuclei are involved selection rules imposed by the exchange symmetry have to be taken into account. In deuteration, the resulting dynamic restrictions play an important role [asv04b].

A well-understood example for a strongly temperature dependent reactivity is the hy-drogen abstraction reaction NH3

+ + H2 → NH4+ + H. This reaction has been measured

by several groups and using different techniques in the range from 300 K to 10 K. It has been found that, starting from room temperature, the reaction rate coefficient falls with falling temperature. It reaches a minimum at about 100 K and than it increases signifi-cantly towards lower temperatures. Statistical space phase calculations confirm the hy-pothesis that this exothermic reaction is hindered by a barrier and that it can occur via tunneling at low temperatures [her91]. Very recently, a strong negative temperature dependence has been found experimentally [asv04a] for the isoelectronic hydrogen ab-straction reaction CH4

+ + H2 → CH5+ + H. This observation has been tentatively ex-

plained with the temperature dependence of the complex lifetime but it has been con-cluded that, in this case, tunneling is not responsible for forming more products at low temperatures. It is an open question whether one can classify reactions with non-standard temperature dependences with a few simple reaction mechanisms based on simple properties of the potential energy surface such as barriers or wells. In some spe-


23

cific cases, simple classifications such as "closed shell ions" work. For example it has been predicted [mal92] that scrambling in the weakly bound complex is very unlikely and therefore CH5

+ + HD does not lead to deuteration. This is in accordance with recent experiments [asv04b]. One very fundamental example illustrating a few other aspects of low temperature scrambling and dynamic restrictions is the important exothermic proton scrambling reaction H+ + H2(j = 1) → H+ + H2(j = 0) [ger90], which is most probably the only gas-phase reaction converting in the ISM ortho- into para-hydrogen.

Radiative association reactions are usually described by a mechanism involving two steps, the formation of a long-lived complex and the stabilization via emission of a pho-ton. This process can be efficient if electronic transitions between low lying excited states and the ground state are involved. Usually transitions between rovibrational states of the highly excited intermediate molecule are responsible for cooling the complex. The overall efficiency depends on the ratio between the complex lifetime and the mean time to eject enough energy via radiation for stabilization. In general collision com-plexes which correspond to strongly bound molecules, have long complex lifetimes whereas weakly bond structures decay much faster. Typical examples which have been studied experimentally include the growth of H3

+ towards H5+, H7

+, etc. up to H23+

[pau95] or the radiative and ternary association reactions C+ + H2 → CH2+ + hν and

C+ + H2 + He → CH2+ + He. Using stabilization with a third collision partner, e.g. He,

and based on some model assumptions, mean values for the collision complex lifetime and radiative lifetime can be derived from experimental data. In this way it has been concluded from a low temperature experiment [ger92a] that the radiative lifetime of the (CH2

+)* complex is 26 times longer than used in theoretical calculations. As is discussed in [ger92a], this raises immediately the question whether electronic transitions are in-volved. A nice example indicating the involvement of vibrational transitions is that the radiative lifetimes of CD5

+ and CH5+ scale with the mass (also the lifetime of the com-

plex, for details see [ger92a] and references therein). Electronic transitions are not mass dependent, to a first order approximation. In general, the lifetime of complexes increase with decreasing temperature; however, at really low temperatures, this trend levels off because the participating degrees of freedom such as rotation cannot participate any-more due to lack of energy. This onset of freezing has been observed in the association rate coefficients measured for the CH3

+ + H2 collision system [ger92a].

The initial rotational energy of H2 has a significant influence. At low temperatures, it is logical to expect that association reactions with H2(j = 0) are more efficient than with H2(j = 1). Using D2 instead of H2 leads to an increase of the complex lifetime because the density of states is higher. Moreover there are other important consequences which are due to the symmetry restrictions caused by ortho and para forms of the reactants. For example, in the case of the three-body association reaction C2H2

+ + 2 H2 → C2H4+ + H2,

the ternary rate coefficients show a weak temperature dependence for normal H2 but increase steeply with decreasing temperature for para H2. In addition, ion-trap experi-ments indicate that the formation of C2H4

+ under astrophysical relevant conditions via radiative association is more important than the bimolecular hydrogen abstraction reac-tion leading to the formation of C2H3

+ + H. Reactions which proceed via the collision complex C3H3

+ appear quite complicated. Low temperature ion-trap studies [sor94] of the reaction C3H

+ + H2 have shown a very interesting temperature dependence and strong competition between radiative association and hydrogen abstraction. In addition, this complex, containing only 3 C and 3 H atoms, seems to have already enough degrees of freedom that its lifetime can become longer than the radiative lifetime. On the other hand one should not expect that this is valid for all six-atomic systems which can form a


24

strongly bound molecule. A typical counter-example is the reaction CH3

+ + H2O → CH5O+ + hν where recent experiments [luc02] obtained a much smaller

reactivity than theoretically predicted and used in astrochemical models.

Another important aspect in forming complex molecules is the role of isomers. Since the collision complexes can be formed in various ways having different structures and binding energies, one can expect that also products appear in different isomeric forms. A nice example is the reaction C3H

+ + H2 +He which proceeds via the C3H3+ collision

complex. It is known from experiments and theory that this ion has two distinct isomers, the c-C3H3

+ and l-C3H3+. The cyclic ion formed via radiative association is most proba-

bly the path leading, after recombination with an electron, to the smallest cyclic mole-cule detected in space, the c-C3H2.

There are very few reactions which are known to be really endothermic but where the endothermicity is so close to zero that they still may be of importance at low tempera-tures. A very fundamental reaction which plays an important role in the synthesis of ammonia, is the hydrogen abstraction reaction N+ + H2 → NH+ + H [ger93]. This reac-tion behaves like having a endothermicity of about 10 meV; however, substitution of one or two of the H atoms with deuterium leads to changes in the reactivity which can-not be simply explained using the asymptotic differences in vibrational zero-point ener-gies. An additional uncertainty is the role of fine structure energy provided from the N+(3PI) ion. There are experimental proofs that rotational excitation of H2 helps to in-crease the reactivity in the same way like increasing the translational energy. Another very fundamental reaction system which is most probably weakly endothermic is the hydrogen abstraction reaction C2H2

+ + H2 → C2H3+ + H. The rate coefficient for this

reaction decreases steeply with falling temperature and, in contrast to the low tempera-ture free-jet reactor results, low temperature ion-trap experiments do not show any C2H3

+ products [ger92a], [ger93]. Since quantum chemistry is not yet able to reach the required accuracy for solving problems like the mentioned ones, the reverse reactions CnHm

+ + H should be studied for a variety of the parameters (n,m).

Since different isomers have different bonding and structures, their reactivity is differ-ent. An interesting reaction network which has been taken from [tur00] and which ac-counts for cyclic and linear species, is shown in Fig. 2.3. For species with more than four carbon atoms there is no distinction between cyclic and linear structure because nothing is known about their relative chemistry or abundance. It is clear that a realistic chemical model must consider all possible isomers which may exist in ISM. A well-studied system is the formation of the two isomers HOC+ and HCO+ which both can be formed at low temperatures, e.g. via proton transfer in H3

+ + CO collisions [smi02]. These two isomers are separated by a very high transition state and therefore the internal energy of the HOC+ which is rather abundant in certain regions of the ISM, can pre-served as internal energy.


25

l-C3H+

l-C3H2+

l-C3H

l-C3N

c-C3H+

c-C3H2+

c-C3H

c-C3N

C2H2

C2H3C2H4+

c-C3H3+

l-C3H3+

l-C3H2

c-C3H2

H2

C4H

C4H2+

C4+

C4H+

H2

C4H3+

H2C4

e

e

e

C

C

C

H2

e

NC+

C

C

C+

C+

H2

e

N

ee

C+

C

C

C

e

H2 H

e

ee

H2

C+

Fig. 2.3. Basic hydrocarbon chemical network including cyclic and linear species [tur00].

2.3. Deuterium fractionation Different isotopes of the same element are in principle chemically equivalent but they have different physical properties and they also can behave quite differently in chemis-try and biology. Due to the fact that rotational and vibrational energy levels of mole-cules depend on the mass, isotopic abundances are easy to observe in spectra [wan80]. The separation of isotopes by chemical methods is called chemical fractionation. Of great importance for astrochemistry is the process of deuterium fractionation via low energy ion-molecule reactions. In general, H - D exchange reactions between hydrogen containing, strongly bound molecules and HD are exothermic because zero-point energy is gained. Therefore, it was generally assumed, that deuteration reactions, which are not hindered by an activation barrier, proceed rapidly in the exothermic direction. But as discussed recently [ger02], there are several examples, one of the simplest being D+ + H2, which indicate that the situation is more complicated. As recently discussed for the systems CHn

+ + HD for n=3-5 in [asv04b], symmetry selection rules and the (ap-proximate) conservation of the nuclear spin can significantly reduced the efficiency of deuteration.

Since the first detection of deuterium in the molecule DCN [jef73] theoretical and ex-perimental understanding of deuterium isotope enrichment is continuously of great in-terest for predicting isotope fractionation occurring in cold regions of interstellar space. Many singly and three doubly deuterated isotopomers (NHD2, D2CO and CHD2OH see [rou00], [tur90], [cec98], [loi01], [par02]) have been detected in low-temperature inter-stellar clouds. Meanwhile even fully deuterated ammonia has been found [lis02], [tak02]. There are molecules, where the DX/HX abundance ratio is a factor 104 larger than the D/H elemental ratio which is typically 2 × 10-5 in our galaxy ([mil02], [loi01]).


26

Today's models describing deuterium fractionation in dark clouds include gas-phase and grain-surface mechanisms ([mil89], [rob00a], [rob00b], [mil02]). However, these mod-els have many problems to predict quantitatively the abundances, e.g. the large deute-rium fractionation of c-C3H2. Reactions which may contribute to the formation of this molecule are subject of this thesis. More information and the results are discussed in a manuscript which can be found in the Appendix.

The gas-phase pseudo-time-dependent model [rob00a], [rob00b] calculates the varying abundances of around 300 species (including deuterated species) linked by around 5500 reactions. The only grain-surface reactions which are included in this model, are those for forming H2 and HD. Today, most model calculations use the reaction rate coeffi-cients collected in the UMIST database [mil97], [let00]. The molecules considered in the calculation include mono-deuterated analogues of all hydrogen bearing species and a small number of doubly deuterated species. For hydrogen and deuterium bearing spe-cies it is simply assumed that the rate coefficients for reactions with the same reactant are equal. In cases where it is uncertain into which products the deuterium atom should be incorporated, simple statistical arguments are utilized for predicting branching ratios. Wherever it was possible results from experiments have been included into the model.

In cold dense clouds, gas phase species can collide with dust grains and tend to stick on their surfaces. In this way grains begin to accrete icy mantles leading to a dramatic change of the properties of clouds. Models accounting for this process are so called de-pletion models. In modeling the influence of gas-grain interaction on isotope fractiona-tion, special attention is paid to the removal of those gas phase species which influence the XD/XH ratios. In the model reported in [rob00a], the H2D

+/H3+ ratio is enhanced,

followed by destruction of these ions by proton or deuteron transfer to other neutral molecules. The dominant formation of H2D

+ is via the reaction between H3+ and HD.

When molecules start to deplete (with exception of H+, D+, H2+, HD+, H3

+, H2D+ and He

for which is assumed that they do not stick), reaction of H3+ with HD become dominant.

This leads to an increase of the H2D+/H3

+ ratio and also of the XD/XH ratio for a period before all heavy species have been frozen out. Fig. 2.4 shows a comparison of the frac-tional abundances of selected molecules from the gas phase model (left) and depletion model (right). Fig. 2.5 shows the ratios [rob00a]. From this comparison it can be seen that condensation of molecules leads finally to the disappearance of them; however, the abundance ratios of deuterated to normal species increase during this time.

Log Time (yrs) Log Time (yrs)3 4 5 6 7 83 4 5 6 7 8

Log

Fra

ctio

nal a

bund

ance

-12

-10

-8

-6

-4 -4

-6

-8

-10

-12

CO

H2O

HCO+

DCO+

CO

H2O

HCO+

DCO+

Fig. 2.4. Comparison of the fractional abundances of some species from gas-phase model (left) and depletion model (right) [rob00a]. Without any desorption mechanism, all molecules shown are frozen out after 106 years.


27

Log Time (yrs) Log Time (yrs)3 4 5 6 73 4 5 6 7

Log

X

(D)

/X

(H)

-3

-2

-1

0

-3

-2

-1

0

D2CO/H2CO

NH2D/NH3

HDCO/H2CO

H2D+/H3+

H2D+/H3+

HDCO/H2CONH2D/NH3

D2CO/H2CO

Fig. 2.5. Comparison of molecular D/H ratios of some species from gas-phase model (left) and depletion model (right) [rob00a].

From the point of view of laboratory astrochemistry, many reactions involving HD need to be studied at low temperatures. For such investigations, a variable temperature 22-pole ion trap machine, as described in next chapter, is well-suited. In all these ex-periments, the influence of o-H2 has to be examined in more detail. Recently it has been realized that not only HD molecules but also D2 may play a role in the overall deutera-tion processes, especially for forming fully deuterated molecules. It also may be that some surprising results will be found in studies of low temperature reactions with D atoms. Of course the largest uncertainties in understanding interstellar H-D fractionation are correlated to the interplay between the gaseous phase and the grain surfaces. It is desirable to measure sticking coefficients under realistic conditions and to find out how condensed molecules can be recycled back into the gas phase.

2.4. A model for growing pure carbon chains Until today, the reaction path of growing pure carbon clusters under astrophysically relevant conditions is not known. Various mechanisms to form carbon chains and rings have been proposed but not yet experimentally proven. In regions with partially ionized carbon, the chain growth from Cn to Cn+1 may proceed via the radiative association process

C+ + Cn → Cn+1+ + hν (2.1)

The rate coefficients for radiative association increase in general with increasing num-ber of atoms. In [suz83] it has been postulated that this mechanism is efficient enough to become the main growing path for n ≥ 4. Assuming in addition, that the step from C3 to C4 is the bottle neck, the total formation rate of chain molecules is proportional to the rate coefficient of reaction (2.1) for n = 3 [suz83]. These authors also point out that lower temperature, higher densities and lower intensities of the external radiation field favor the growth of carbon chains. Another interesting network also including reactions with H2 and electrons is shown in Fig. 2.6 [suz83].

The n-dependence for radiative association reaction (2.1) has been investigated in a sta-tistical model calculation in [fre82]. Within the assumptions it was shown that the reac-tions proceed efficiently with the Langevin rate for n ≥ 4. For n=1 or 2, reaction (2.1) is rather inefficient. This is due to fact that the time needed for the spontaneous emission of an infrared photon (typically 10-3 s) is much longer than the duration of the encounter between an atom and a diatomic molecule (typically 10-10 s). The energized (Cn+1

+)* can


28

not get rid of the excess energy before it decays, most likely back to C+ and Cn [fre82]. In order to understand the dynamical and statistical assumptions which lead to the re-sults plotted in Fig. 2.7 several facts about the collision system are briefly summarized.

Cn+1+

CnH

CnH3+CnH+

Cn

Cn+

H2 H2

e ee

C+C+

Fig. 2.6. Reaction paths from Cn

+ to Cn+1+ [suz83].

m = 50 m = 30 m = 50

m = 301 s

1 ms

n + 12 3 4 5 6 7 8 9

-10

-5

0

log

t

Fig. 2.7. Calculated lifetime of activated (Cn+1

+)* [fre82]. Calculations are performed for differ-ent model parameters (see text). Dashed curves represent results of calculations in which low-frequency bending modes have been ignored.

During the approach of C+ and Cn, the long range attraction speeds up the relative mo-tion. The C+ ion may be directed to the any part of the Cn chain. This initial collision can excite local vibrations of Cn in vicinity of the impact place. This transfer of energy may remove enough energy from the relative motion, that the Cn-C

+ complex cannot dissociate immediately. Since the most stable positions for adding C+ are at the end of


29

the chain it can be expected that C+ migrates from the location of initial impact to the chain end. During this migration close interactions lead to further excitation of vibra-tions of the Cn+1

+ ion, or, with other words, to a randomization of the available energy. Therefore one can assume that statistical models can describe this collision. In [fre82] the simple RRKM formula has been used for life-time estimation.

It is obvious from Fig. 2.7 that the lifetime of (Cn+1+)* grows rapidly with n. The reason

is the large number of vibrational modes which can take a portion of the total energy and which prevent in this way the back transfer of the energy into degrees of freedom, leading to dissociation. The results presented in the figure have been calculated for a typical set of values ν/c = 1500 cm-1, m = 30 (Ed = 5.6 eV, m is number of quanta in-volved in the Ed = mhν - dissociation energy) and m = 30 (Ed = 9.3 eV) for n + 1 = 2 - 9 [fre82]. Dashed curves represent results in which low-frequency bending modes have been ignored. The solid curves show the importance of these low-frequency bending modes. From this figure it can be seen that the complex lifetime can reach 1 ms for n = 4. It should be added that the authors pointed out that their calculated values are lower limits. There are a few critical remarks necessary. First, the RRKM formula de-scribes only unimolecular decay and for properly describing the collision process, a correct statistical model (e.g. phase space theory) should be used. Second, the crude approximation for the calculation of the accessible states may be questioned. The Ein-stein model is supposed to be reasonable for the stretching modes while the low fre-quency modes need to be counted separately and also the rotational degrees of freedom should be included, especially at low temperatures.

In addition to reaction (2.1), also

C + Cn+ → Cn+1

+ + hν (2.2)

has to be considered. However, it it is less efficient under astrophysical conditions be-cause Cn

+ is depleted via the reaction Cn+ + H2 → CnH

+ + H. In addition the charge transfer reaction C + Cn

+ → C+ + Cn is exothermic by more than 1 eV. This exothermic channel will lead to a very fast decay of the activated (Cn+1

+)* [fre82].

Another interesting scenario for growing carbon chains and for condensation reactions occurring in a radioactive supernova gas has been explored in [cla01]. The expanding supernova interiors are hydrogen free and chemistry is dominated by neutral-neutral carbon

C + Cn → Cn+1 + hν (2.3)

and oxygen chemistry. Fast particles produced by the radioactivity disrupt CO molecule producing free C and O. These C atoms can associate with other molecules leading to linear chains. The predicted abundance of chains Cn decreases with n because of de-struction mechanisms such as oxidation or fission reactions of type

C + Cn → Cm + Cn-m . (2.4)

It is important to note that thermal dissociation of Cn, can be neglected for T < 2000 K. For example, a C2 molecule can live about one month at T = 2000 K before is thermally disrupted. When linear chains Cn are becoming long enough, they may isomerize into ring structures. If this is achieved, association becomes faster and oxidation and fission reactions become slower. In this way, large sizes can be achieved and ring clusters nu-cleate up to macroscopic graphite grains [cla01].

A lot of experimental and theoretical effort has been devoted to investigate pure carbon clusters, ranging from theoretical calculations of stable isomers, to experimental spec-


30

troscopy of different size clusters [hel93], [hun93], [han95], [ord98], [mai98], [wak99], [jon99], [del99], [del00]. The reason for this is not only their role in astrophysics, but these molecules play also an important role in material science and in combustion proc-esses. Nonetheless, despite the huge efforts of experimentalists, there are no systematic studies of the growth of carbon structures under astrophysical conditions, ranging from the high temperatures of stars to the very cold conditions of dense interstellar clouds. That is especially the case for reactions of ions with neutral carbon, because of the ex-perimental difficulties to provide high enough number densities of the neutrals. Almost all reaction rates included in astrochemical models are based on theoretical estimates and not on experiments. To measure reaction rates under astrophysical conditions, the Chemnitz group for Laboratory Astrophysics has started to set up a dedicated trapping apparatus, which is described in the next chapter.

31

3. EXPERIMENTAL: ION TRAPPING APPARATI

3.1. Ion molecule reactions at variable temperatures During the last few decades several experimental techniques have been developed for investigating inelastic and reactive interactions between ions and molecules. The first steps in this field have been performed at the beginning of the last century by establish-ing the mass spectrometry technique. In the following, a brief overview over three types of experimental techniques is given, (i) swarm experiments, (ii) beam methods, and (iii) ion traps.

A first insight into the field of ion chemistry became possible with swarm techniques which have their origin in the flowing afterglow plasma technique developed by Fergu-son, Fehsenfeld and Schmeltekopf in 1963. These techniques are based on the fact that a flow of neutral gas can carry ions in the same direction. Many published results show that an afterglow plasma is a suitable medium for studying binary and ternary ion-molecule reactions at thermal energies. The reason is that - at the high number densi-ties - the reactants have a Maxwellian kinetic energy distribution and also the internal states are populated according to a Boltzmann distribution [smi79].

In the stationary afterglow technique [smi79], ions are produced by a short pulsed dis-charge, using an rf or dc electric field. Following this, the time evolution of the ions in the afterglow is monitored. This is done by mass spectrometric sampling of the ion composition using a suitable detector. Since the decay of primary ions is proportional to the partial pressure of the reactant neutral gas, reaction rates can be easily obtained. However, the reacting neutral gas is also exposed to the discharge and, therefore, inter-nal states can be populated significantly, leading to non-thermal rate coefficients. Moni-toring the ions in the afterglow at later times (with some delay after the discharge) when the excited states have been relaxed by collisions, can reduce these problems. For inves-tigating negative ion-molecule reactions, this method cannot be used because of simul-taneous presence of electrons and negative ions in the plasma, which leads to a high ambipolar field. Due to this field, a flow of negative ions into the plasma boundary is suppressed and the ion current in the plasma boundary is no longer proportional to the number density of ions in the plasma. This disadvantage can be overcome by adding a proper amount of an electron attaching gas. In this way the plasma is converted from a plasma dominated by electrons and positive ions into a plasma dominated by negative and positive ions.

Several disadvantages of the stationary afterglow technique are overcome by the flow-ing afterglow technique [smi79]. In this technique, primary or precursor ions are created upstream in a fast flowing carrier gas. This gas is usually He with a flow rate of several hundred mbar liter s-1 and a velocity of about 100 ms-1. Typical flow tube pressures range from 0.4 mbar to several mbar. The source gas can be added into the afterglow - downstream of the ionization region - and there the desired primary ions are created via reactions with ions, metastable atoms or electrons. The afterglow plasma is distributed along the tube and it is separated from the ionization source. Further down in the flow, neutral reactant gas is exposed only to the thermalized plasma and in this way internal excitation is further reduced. The positive or negative ions are sampled in regions to-wards the end of the flow. The typical number density of ions in the reaction region is 107 - 109 cm-3. The reaction zone is usually 25 cm to 75 cm long. The number of pri-mary ions decreases as a function of the flow rate of reactants and from this dependence


32

the reaction rate coefficients can be derived. Using this technique, rate coefficients from 10-9 cm3s-1 to 10-13 cm3s-1 can be accurately determined. Some temperature dependence of reaction rates can be obtained by varying the temperature of the flow tube walls or by preheating or precooling the carrier gas.

The first flowing afterglow setup which allowed a study of the energy dependence of reactions is the so-called flow-drift tube experiment [smi79]. This technique allows one to study ion-neutral reactions from thermal energies to a few eV. The principle of the drift tube is that a superimposed uniform electrical field accelerates the ions and leads, via collisions with the non-reactive buffer gas, to an averaged ion energy. In general, the drift tube setups have the same ion production section as the flowing afterglow setup. Between the ion production section and the drift and reaction section, an ion separation section can be integrated to separate the plasma positive ions from negative ions and electrons. Also an ion shutter can be introduced which is used to prevent any flow of ions into the drift reaction section. The shutter can be operated in a pulsed mode for determining the drift velocities of ions. It is held open during the determination of reac-tion rate coefficients. Using this technique, reactions rate coefficients larger than 10-12 cm3s-1 can be measured over an translational energy range from 0.05 eV to 3 eV, and at a gas temperature of 300 K.

The general disadvantage of the flowing afterglow is that the ions have to be generated in the flow tube. The presence of the neutral precursors in the afterglow complicates the evaluation and interpretation of the data since the reactions of ions with these molecules are in competition with the reactions with the reactant gas. In addition, the presence of electrons can cause deionization by dissociative recombination, which is in competition with the processes of interest. Another disadvantage is the presence of metastables or energetic photons. To avoid these disadvantages the selected ion flow tube (SIFT) has been developed in 1976. In the SIFT technique [smi79], [smi88], mass selected ions are injected into a flowing gas. An important difference from the afterglow technique is that the reactant gas is not an afterglow and therefore no electrons are present. In the flow, defining the reaction region, only the neutral reactant gas and a low current of ions are present. To investigate reactions at different temperatures, the variable-temperature se-lected-ion flow tube (VT-SIFT) apparatus [smi88] has been developed. This machine is very similar to the SIFT apparatus but, in addition, it allows to study reactions down to liquid nitrogen temperature (80 K). So far, there has been constructed only one drift tube apparatus which can be used at temperatures between 18 K and 420 K. This in-strument developed by Arnold and Böhringer [böh83], was based on a liquid helium cooled drift section. However, only a few results have been reported from this machine. More detailed information about swarm methods can be found in the relevant literature (e.g. [lin86]).

There have been several other attempts to study reactions at very low temperatures. One of the solutions is to use supersonic expansions. The well-known technique which be-longs to the this group is the CRESU apparatus (Cinétique de Réaction en Ecoulements Supersoniques Uniformes) developed by Rowe and coworkers [row84]. It uses a uni-form supersonic jet generated by specific Laval nozzles. Buffer and reactant gas are mixed in the nozzle reservoir and the mixture expands through the nozzle. The tempera-ture which is established in the uniform supersonic jet can be varied by changing the operating conditions and the nozzle geometry. The lowest temperature achieved is 15 K with a 300 K nozzle, and 7 K with precooling the nozzle and the gas reservoir using liquid N2 [smi00]. The primary ions which are created in the flow in various ways, e.g. by electron impact, react with the neutrals. The number of primary and product ions is


33

recorded by using a quadrupole mass spectrometer detector. There are two ways of measuring rate coefficients. The first one is to record the decrease of primary ions as a function of the reactant density at a fixed distance from the nozzle exit. The second one is to record the decrease of the primary ions with increasing distance from the nozzle exit, keeping the reactants number density constant.

Studies of reactions below 3 K have been made in a free jet flow reactor developed by Smith [smi98]. The very cold mixture of reactant and buffer gas is prepared via a free expansion through a small nozzle orifice starting with high densities. At a given dis-tance from the nozzle, specific neutral species are selectively ionized either by one-photon vacuum ultraviolet ionization or by resonance-enhanced multiphoton ionization. These primary ions react with the neutral reactants which are also present in the flow. At a variable distance, primary and product ions are extracted in transverse direction relative to the main flow. They are mass analyzed using the time of flight method.

Different setups based on the Guided Ion Beam (GIB) technique have been used to in-vestigate low energy ion-molecule collisions in detail. This method allows one to pre-cisely determine integral and also differential cross sections with high sensitivity. In general, primary ions are produced by electron bombardment and thermalized, e.g. in a radio-frequency storage ion source. They are mass and energy filtered by passing through a quadrupole with typically 0.1 eV of axial energy. At the end of this quadru-pole, ions can be selected in addition according to their time-of-flight by using a suit-able pulse on an electrostatic deflection system. In this way ions with a very narrow energy distribution are injected into the octopole which guides them through the scatter-ing cell. After reaction, ions are mass analyzed and detected. Advantages of the univer-sal GIB apparatus include the extension of laboratory energies of guided ions to below 10 meV and the combination of the method with supersonic beams in the crossed and merged beam arrangement. Also photons from chemiluminescent reactions can be measured, even coincidences between photons and mass selected product ions have been detected [ger92b]. Various laser methods have been applied for state selective preparation of reactants or for analysis of reaction products. For getting collisional en-ergies below 1 meV, a slow guided ion beam has been superimposed with a supersonic beam. This method is superior to the traditional method of Merged Beams (MB) where two fast beams, with keV laboratory energies, are superimposed in parallel for obtaining meV collision energies. More details about this method and detailed description of the techniques can be found in [ger03b] and references therein.

An approach for investigating ion-molecule reactions which is quite different from swarm and beam techniques is to use ion trapping techniques. A storage technique very popular in ion chemistry is the Ion Cyclotron Resonance (ICR) method, which is similar to the Penning trap. Confinement of ions is based on the superposition of a static electric and magnetic field. The rather complicated ion trajectories are a superposition of cyclo-tron rotation, magnetron motion and trapping oscillation. In the ICR technique, the time evolution of ion masses in the presence of neutral target gas is observed by exciting the cyclotron motion of specific ions by a suitable radio frequency (rf) signal and detecting the image currents. A nice overview about ICR technique can be found in [mar98] and in references therein. Well known is also the Paul trap which is based on the use of elec-tric quadrupole fields alternating in time. After the invention of the first linear quadru-pole trap, various higher order rf multipoles (e.g. octopoles, 22-pole) also were con-structed. Other rf based trapping arrangements have been developed such as the Ring Electrode Trap (RET), which will be discussed in more detail below.


34

In general, many advantages are obtained by using trapping techniques based on inho-mogeneous rf fields. The fundamental principle and many details of this technique are described in [ger92b]. Paul traps, i.e., quadrupole ion traps, have been and are used in many laboratories for investigating ion-neutral reactions; however, this method is not suitable for reaching low temperatures since the harmonic rf field continuously influ-ences the ion energy. Using traps with wide field-free regions which can be created by high order multipoles, ring electrodes or other field geometries, one can significantly reduce rf heating. The arrangement which is used meanwhile worldwide in at least ten machines is the 22-pole trap, but also 16 or 24 poles are used successfully as traps in many applications [ger04]. The integration of a multipole trap into the so-called Vari-able Temperature 22-Pole Trap apparatus (VT-22PT) is described in detail in the next chapter. In general, the main advantages of this technique are that the nominal tempera-ture can be varied from 10 K up to 300 K, the storage time can be varied from µs to minutes or longer. For studying reaction processes, the target number density can be varied from 109 cm-3 to 1015 cm-3. The overall sensitivity of the technique can be illus-trated by fact that rate coefficients as low as 10-20 cm3s-1 can be obtained [asv04b].

3.2. Description of the machines used The experimental results presented in this thesis have been obtained using two different apparati, the Carbon beam – Ring Electrode Trap (C-RET) machine and the Variable Temperature 22-Pole Trap (VT-22PT) apparatus. Since the fundamental principle of both trapping experiments is the same and many characteristics are very similar both setups are described in parallel.

In general, a trapping machine consists of an ion source, a first mass filter for ion prepa-ration, the ion trap, a second mass filter for analyzing the trap content and an ion detec-tor (see Fig. 3.1). Typically the experiments are performed in a pulsed mode. Primary ions are generated in the ion source and a bunch of mass filtered ions is injected into the trap. There the ions are cooled to the temperature of trap walls via inelastic collisions with the cold buffer gas. After relaxation the ions are allowed to interact with the target gas neutrals. After a given storage time, the content of the trap is extracted, mass ana-lyzed in the second mass filter and finally detected and counted.

ion

source1st mass

filtermass

analyzertrap detector

Fig. 3.1. Elements of a typical ion trap experiment.

In order to study reactions between ions and a beam of neutral carbon atoms and mole-cules under astrophysical relevant conditions, a special ion-trapping apparatus has been developed in this thesis work, combining a radio-frequency RET with a carbon source. This machine (called C-RET) already has been described briefly in [cer02] (see Appen-dix A). However, more experimental details, especially concerning the carbon beam source, will be given in the next section. In general, the C-RET machine consists of the carbon source, an electron bombardment ion source, the RET, a mass analyzer and a single ion detector of the Daly type. The scheme of this setup is given in the paper in the Appendix. The ions can be generated by electron bombardment either from a pulse of neutral gas injected to the trap or by ionizing the carbon beam. During the trapping time these ions are exposed to the beam of the neutral carbon. After a given storage time, the


35

content of the trap is analyzed. After each step, the flux of the carbon beam is monitored to correct for instabilities of the carbon source.

The C-RET machine was constructed for studying reactions of ions with neutral carbon atoms or molecules; however, it became clear that, from a point of view of interstellar chemistry, also information on related reactions, e.g. of carbon or hydrocarbon ions with hydrogen is required. In order to contribute more to the understanding of the chemistry involving C, H and D in various combinations, a few selected experiments have been performed in the VT-22PT machine which is shown in Fig. 3.2. The main modules of this machine are the storage ion source, the first mass filter, the variable temperature 22-pole trap, the mass analyzer and the Daly detector. Since this apparatus has been de-scribed in many details in the past [ger93] and recently [asv04b] and in a variety of pub-lications (see [ger03b] and [ger03c]) the next section contains only a brief overview.

Storage Ion Source

1st Quadrupole MassFilter

22-pole Ion TrapDaly

Detector

Quadrupole MassAnalyzer

He Refrigerator

Fig. 3.2. The VT-22PT apparatus.

3.2.1. General overview A very good vacuum is one of the crucial points in an ion-trapping experiment due to fact that the stored ions, in addition to the reaction with the neutral target gas, also react with the residual gas. In order to suppress such parasitic reactions partial pressures of impurities should be below 10-11 mbar. The vacuum system of the C-RET machine is presented in Fig. 3.3. The high vacuum needed in the carbon source has been obtained using the turbo-molecular pump TPU 170 (TP1) having a pumping speed of 170 l/s. A differential pumping stage, consisting of two apertures with diameters of 3 mm (A1) and 6 mm (A3) and the turbo-molecular pump TPU 062 (TP2, pumping speed 56 l/s), has been used to improve the purity of carbon beam. The aperture (A2) of 20 mm di-ameter has been used for protecting the valve (V) sealing from a direct exposure to car-


36

bon. The main chamber has been evacuated with two turbo-molecular pumps, TMU 260 (TP3) and TPU 240 (TP4), both having a pumping speed of 230 l/s. The turbo-molecular pumps are fore-pumped by the two rotary pumps PK8D (FP1) and PK D44 706 (FP2). They are equipped with a pair of zeolith adsorption traps. Pumping speeds of these rotary pumps are 2 l/s and 4.4 l/s, respectively. This pumping system provides an UHV with a typical background pressure of 3 × 10-10 mbar in the main chamber. Liquid nitrogen cooling of the RET also improves the pressure and, at such conditions, less than 1% of primary ions are converted through parasitic reactions with the residual gas at a storage time of 1 s.

Vacuuminterlock

Spinning rotorgauge MKS-SRG2

TP1TP2 TP3

TP4

FP1FP2

A1 A2 A3

V

Fig. 3.3. Vacuum system of the C-RET apparatus.

The vacuum system of the VT-22PT machine is shown in Fig. 3.4. It consists of two magnetically levitated turbo pumps, TMU 400M and TMU 200M. The pumping speeds are given in the figure. These two pumps are connected via a TPD 011 drag pump (10 l/sec) to a MVP015 diaphragm pump (15 l/sec). Using this vacuum system, typical background pressure of 2 × 10-10 mbar is obtained after heating the apparatus to a tem-perature of 60°C, usually for 2 days. Note that for really baking the system, the cold head must be disassembled. The residual gas is mostly H2 and H2O with a number den-sity of about 108 cm-3 and 107 cm-3 at 300 K. Operating at cryogenic temperatures, the number density of H2O inside trap decreases to 105 cm-3, the H2 number density remains unchanged.

The pressure in the chambers is measured by two ion gauges. On the VT-22PT machine, ion gauges are mounted onto the chamber containing the first mass filter and the trap On the C-RET machine, one ion gage is located at the carbon source, one on the central chamber containing the RET. The vacuum interlock systems are connected to the ion gauge controller monitoring the pressure in the main chambers and in the carbon source. By an interlock system, all sensitive power supplies are shut down as soon as the pres-sure exceeds 10-4 mbar.


37

Vacuuminterlock

Spinning rotorgauge MKS-SRG2

TMU 400M390 l/s

TMU 200M180 l/s

TPD 011drag pump

10 l/s

MVP 015diaphragm pump

15 l/s

Fig. 3.4. Vacuum system of the VT-22PT apparatus.

The gas inlet systems for the C-RET machine and the VT-22PT apparatus are shown in Fig. 3.5 and Fig. 3.6, respectively. They allow flexible mixing of gases or introducing pure gases into the machines. In the C-RET apparatus, gas is lead in via a leak valve or a fast piezo valve into RET. A third valve (magnetic, from General valve) leads gas into main chamber. In the VT-22PT machine, one part of the system supplies gas for the ion source and for a second fast piezo valve. Neutral target gas is lead into the 22PT through pre-cooled tubes reaching the trap temperature before entering the interaction region.

toPiezoGauge

to Pump

to Trap to GeneralValve

Fig. 3.5. Gas inlet system of the C-RET apparatus.


38

to Piezo to Trapto Ion Source

Gauge

to Pump

Fig. 3.6. Gas inlet system of the VT-22PT apparatus.

Since most other uncertainties can be eliminated or corrected for, the accuracy of the absolute value of the rate coefficients is directly determined by the accuracy of the num-ber density of the neutral target gas in the trapping region. In order to measure the pressure of the neutral reactant molecules, the spinning rotor gauge MKS SRG2 is con-nected via a stainless steel tube to the inside of 22PT (at VT-22PT machine) and to the main chamber of C-RET machine. The spinning rotor gauge is specified to be accurate within 5 %. Since it is very sensitive to vibrations, it can only be used for pressures above 10-7 mbar. For lower pressures the ion gauge, which is calibrated against the spinning rotor gauge, is used for routine measurements. For calibration, the gas of inter-est is introduced via the gas inlet and both readings are recorded (pressure reading of the ion-gauge pig, of the spinning rotor gauge psrg) at 300 K. Plotting psrg vs pig and using the simple linear function psrg = C⋅pig for fitting the data results in the calibration factor C. The calibration factors for the gases of interest, HD, H2 and He are: CHD = 352, CH2 = 278 and CHe = 1024 for the VT-22PT apparatus. Note that these factors are valid only for setting the ion gauge controller (AML) to the sensitivity "19" (corresponding to N2) and using an emission current of 0.1 mA. In this way, the calibration factors ob-tained are in respect to N2. In the case of the C-RET machine, a different ion gauge type has been used. Therefore the sensitivity “12” has been set at an emission current of 0.1 mA.

In the VT-22PT machine the neutral target gas is introduced directly into trap. Therefore the number densities inside the trap and in the chamber are different due to the pressure gradient which is determined by the conductance of the entrance and exit apertures and the speed of the main chamber pump. The different calibration factors C are due to this pressure drop, due to the gas dependence of the pumping speeds and due to the gas de-pendent sensitivities (ionization probability) of the ion gauge. The deviation of the cali-bration factor for CN2 from one by 30% is in accordance with the deviations of non-calibrated ion gages used.

Knowing the pressure inside the trap, pt, it is easy to calculate the number densities. In most experiments, the ion trap is at a different temperature, Tt, while the ion gauge is at Tc. In this case on obtains


39

c

tct T

Tpp = , (3.1)

where pc is the pressure in the chamber. The number density of neutrals inside the trap is then given by

ctB

ig

tB

tt

TTk

pC

Tk

pn == , (3.2)

where kB is the Boltzmann constant. Due to fact that the main chamber has usually a temperature of 300 K, the number density of neutrals in the trap can be obtained from the practical formula

3

t

ig17t cm

T

p)1018.4(Cn −×= (3.3)

In the last three equations, the temperature is in units of K and pressure in mbar. It has been estimated that the overall accuracy for determination of nt is in the range of 10 % to 15 %. This includes also the uncertainties caused by the density gradient from the center of the trap towards the exit.

In the C-RET machine, the number density of the neutral C3 species has been deter-mined as described in Section 3.3.3.

The ion sources used in the two machines are (i) a simple ionizer and (ii) the "standard" storage ion source. In both cases, ions are created by electron bombardment.

The ionizer in the C-RET machine has been used on one side for producing primary ions for transfer into the RET and on the other side for monitoring the flux of the carbon beam. It is based on a commercial construction originally developed by Extranuclear for quadrupole mass spectrometers. It consists of cylindrical electrodes with axial apertures. The open construction allows the neutral carbon beam to pass through the source into the RET. At very good vacuum conditions this source can be operated in a space charge mode, in which ions are stored inside the cylindrical anode grid. Using this mode, very high ionization efficiencies can be reached. By choosing suitable potentials, the elec-trons from the filament can be injected directly into the trap for in situ production of ions.

The storage ion source used in the VT-22PT machine is basically an ion trap coupled with an electron source. It consists of a stack of 8 rectangular rf electrodes with double-H shaped cut-outs. They are isolated from each other using ruby balls. The labyrinth like storage volume is used for separating the ionization region from the source exit. The plates are alternatively connected to the two phases of an rf generator. The top and bottom of the storage volume are closed by appropriate dc voltage applied to the end-plates which contain small slits for the electron beam. Electrons, emitted from the fila-ment are accelerated to a selected energy and ionize the gas inside the source. These ions are trapped and cooled to the temperature of the source (typically 350 K). They also can react with other neutral species forming other ions of interest. Choosing the rf frequency, the trapping voltage, pressure of precursor gas and/or gas mixtures, a precise control over the emerging ion composition can be achieved. In the exit region, ions are guided in a field which changes smoothly to that of an octopole. An electrode is situated very close to the exit and by applying adequate pulsed voltages to this element, stored ions are gently extracted from the source. The main advantage of this source is the pre-


40

thermalization of the ions. The construction of the electrodes is such that the source is transparent in axial direction for laser applications.

The VT-22PT apparatus is equipped with two mass filters. The first is operated both in the mass-selective and the low-mass band-pass mode depending on the applications. The second mass filter is used exclusively in the mass selective mode (quadrupole mass spectrometer). Both systems consist of 4 rods with a diameter of d = 18 mm, circum-scribing an inner circle of a radius of r0 = 7.8 mm. The first quadrupole is 245 mm long, the second one 260 mm. At the C-RET machine only one quadrupole has been used for analyzing the products. In both machines, the quadrupoles used for analysis are almost identical. The first quadrupole was driven by a home-built rf power supply, the second one by a commercial 1.6 MHz rf power supply.

The basics of operating rf quadrupoles is thoroughly explained in many books and pa-pers (e.g. [ger92b], [mar97] and references therein), therefore only a very short sum-mary is given here. As illustrated in Fig. 3.7, the two voltages ±(U0 - V0 cosΩt) are al-ternatively connected to the 4 rods. U0 is the dc voltage, V0 the amplitude of the rf volt-age, Ω = 2πf is the angular frequency, and t is the time. The trajectories of ions in the rf field is fully described by the Mathieu equation, the solutions of which can be classified with the help of the a2 - q2 stability diagram shown in Fig. 3.8. The two stability parame-ters, a2 and q2, are given by [ger92b]

20

20

2 rm

qU8a

Ω= , (3.4)

20

20

2 rm

qV4q

Ω= . (3.5)

r0d

( )tcosVU 00 Ω−+

( )tcosVU 00 Ω−−

Fig. 3.7. Quadrupole and supplying voltages.


41

m1 > ms >

m2

0.2

0.1

0.5 1

a2

q2

stable

unstable

low mass band pass

mass-selective

operating lines

Fig. 3.8. Stability diagram and different operation modes of a quadrupole mass spectrometer.

If one operates the system such that the two parameters are in the triangle-shaped re-gion, the trajectories are stable and injected ions pass through the quadrupole. Two spe-cial operation modes are marked in the figure, "mass-selective" and "low-pass". In the upper corner of the stability triangle, only a narrow mass range around ms is transmitted (note that in this thesis we are considering only single charged ions) the width of which depends on how close (a2, q2) is set to the tip (0.237.., 0.706..). For transmitting other masses, U0 and V0 must be changed in accordance with equations (3.4) and (3.5). Oper-ating the quadrupole in the low-mass band pass mode is based on the fact that all masses fulfilling the condition ½q2

2 > a2 are transmitted through the quadrupole [ger92b]. In this mode, all masses above a very sharp cut-off value are removed by the filter. This mode is well suited for preparing an ion beam with a narrow energy distribution because it operates in the region where the kinetic energy is an adiabatic constant of the motion. The mode U0 = 0, the so called rf-only mode, can be used for guiding a wide range of masses; however, the focusing properties also may be used for selection.

For preparing the primary ions needed in this work (36 u < m < 40 u, mainly C3Hn+) and

for avoiding high voltages, a single-electron tube oscillator has been modified to operate in the frequency range of 600 – 700 kHz . The basic construction of this generator the central element of which is the double tetrode Valvo QQE 06/40, is shown in Fig. 3.9. A similar push-pull arrangement for amplifying the oscillation of an LC circuit has been described in [jon97]. Choosing a coil with L = 54 µH and a capacitor with 1 nF, the feedback circuit leading to the gates g1 and g1’ forces the system to run at a frequency of 682 kHz. By connecting this circuit to the quadrupole rods, the frequency is lowered to 651 kHz due to the additional capacity of the cable and the rod system. A dc voltage 2 U0 (i.e., adding ±U0) is applied to the quadrupole rods by a circuit shown on the upper left corner of Fig. 3.9. In order to define the kinetic energy of the ions in the quadrupole, a voltage UFA of typically 0.3 V to 0.4 V defines the potential of the field axis of the quadrupole. For selecting a mass ms at the frequency of 651 kHz, the necessary voltages are V0 = 1.9 ms and U0 = (0.237/0.706) V0. In these formulas, ms is in u and voltages are in V.


42

22 kΩ / 10 W

10 mH 10 mH

20 nF / 3 kV 20 nF / 3 kV

47 nF / 400 V

L = 54 µH

1 nF / 3 kV

Output

Anode2 x 1.25 mH / 0.8 Ω

1 nF / 3kV

0.1 µF / 350 V 0.1 µF

0.1 µF / 350 V 0.1 µF

100 Ω

25 kΩ

3.3 mH 3.3 mH

10 nF / 3 kV

VALVO QQE 06/40 ~ 80 pF~ 80 pF

a a'

g2

g1g1'

k

f ffm

Filament

100

µH

100

µH

0.1

µF

0.1

µF

10 n

F10

nF

DC in10 kΩ

10 kΩ47 µF

680

µH

680

µH

UFA2U0

Fig. 3.9. The scheme of a home-built rf generator with dc supply circuit for the first quadru-pole.

Ions extracted from the trap (RET or 22PT) are mass selected and then detected using an ion detector of the Daly type. This detector is described in literature [dal60]. Fig. 3.10 shows the principle. The beam of positive ions is accelerated and deflected by the high voltage (typically -30 kV) applied to the electrode which converts the ions into secondary electrons. They are accelerated and penetrate with 30 keV into the scintilla-tion plate where many photons are produced. These "events" are amplified by a photo multiplier and registered by a counter.


43

Positive ion beam

High voltage convertor

Scintillator

Photo multiplierSecondaryelectron

beam

Fig. 3.10. Working principle of a Daly detector: the positive ion beam is accelerated and de-flected by a high voltage applied to the metal converter. The energetic ions eject secondary electrons which are accelerated towards the scintillator. There they are converted into photons and detected and counted by a photomultiplier and a counter.

A simple bimolecular reaction between an ion A+ and a neutral B producing an ion C+ and a neutral D, i.e.,

DCBA +→+ ++ (3.6)

can be described by a reaction rate coefficient k which is a measure of the efficiency of the reaction. The time dependence of the number of primary ions fulfill the differential equation

BnkAA ++ −=& (3.7)

where nB is number density of neutrals of kind B. Integration of this equation leads to the exponential decay of the number of primary ions A+.

In general the situation in chemical systems such as dense interstellar clouds, combus-tion processes etc. is more complicated. Also in the ion trap, there are almost always competing processes and the time dependence is not just a simple single-exponential but more complex. Initially stored ions A+ can form, in reactions with various neutrals, more than one product. These products can undergo further reactions with the target gas forming new ionic species. Also back-reactions may occur. In order to properly under-stand the evolution of the trapped ion cloud and to obtain the rate coefficients for the reactions involved, all primary and products ions have to be recorded as a function of storage time. These time dependent number of ions, Ni, are then fitted with functions which are solutions of the adequate system of coupled differential equations describing the time evolution of the system. In the case of a complex chemistry the simultaneous fit of a large set of analytical solutions to the experimental data can easily lead to errors. Note, however, that the mathematics for such systems, including sensitivity analysis is well-developed. In most practical cases of this work, numerical integration of the cou-pled rate equations is preferred since it can provide some practical insight into the de-tails of the interdependencies.

The numerical integration starts with the measured or an assumed initial distribution of all initial ions. Then, it uses the interval approximation of equation (3.7),

tnkAA Biii ∆−=∆ ++ m...,1i = . (3.8)


44

In small time steps ∆t, the converted number ∆Ai+ is subtracted from the number of the

primary ions Ai+. This number is added to the product ion. In cases where several prod-

ucts are formed, one has to distribute ∆Ai+ according to the branching ratio. The set of

reaction rate coefficients, ki, can be varied manually or automatically until a good fit to the measured values is obtained. In this thesis, all results obtained from the TV-22PT machine are fitted with QuickBASIC programs. In general, many experimental parame-ters such as the number density of the reactants or the buffer gas, the distribution of ini-tially injected ions, the rf amplitude or others are varied. This leads to a large set of data and allows us to understand in many details the physical and chemical processes in-volved.

3.2.2. The two ion traps: VT-22PT and RET In both machines used in this thesis the central part is an ion trap. Two types of traps have been used, a ring-electrode trap (RET) the temperature of which can go from liq-uid N2 to 600 K and a 22-pole trap mounted onto a closed cycle He refrigerator which can operate between 10 K and 330 K. One of the advantages of the RET in comparison to the 22PT is that the construction is less sensitive to coating with carbon (shielded insulators) and that it is easier to clean.

The first RET was constructed already at the end of the sixties [bah69]. The liquid ni-trogen cooled trap used here was first described in [ger89]. Since many details can be found in [ger92a], [ger92b], [luc01] only a short description is given here. A sketch of the arrangement including the rf circuit is presented in Fig. 3.11. The trapping region is defined by 20 electrodes alternatively connected to two metal bars. The gap between neighboring electrodes is 1 mm, they are 1 mm thick and have a circular hole with a radius of r0 = 5 mm. The metal bars are welded to metal tubes. These tubes are used both for electrical connections and for flowing liquid N2 through the whole system in-cluding the metal bars. In addition they are winded up to form a coil which is the induc-tance LT of the resonant circuit supplying the trap with rf. For excitation, this circuit it is connected to a home built rf generator (Gn). Using two external capacitors, C = 500 pF, the operating frequency has been set to f = 7.8 MHz. Typically, the trap was operated with amplitudes up to V0 = 250 V. The dc potential of the trapping region has been de-fined by applying the voltage VT to the coil via a low frequency band pass circuit con-sisting of the coil LT and the capacitor CT.

Gn

lN2

TE

xSE

x

SE

C CSE

nT

En

SE

x

SE

RET

LVTCTLT

SE

n

lN2

Fig. 3.11. The scheme of a ring-electrode ion trap (RET).


45

In radial direction, ions are confined by the effective rf potential. In axial direction the trap can be closed and opened by applying suitable voltages to the trap entrance, TEn, and trap exit electrode, TEx. For closing the trap, voltages of a few 100 meV are suffi-cient, after thermalization of the trapped ion cloud, much lower values can be used. For extracting the ions, a small negative pulse, typically -1 V is used. The trap is equipped with two steering electrodes, SE, the electric field of which penetrates into the interior. They are made from two graphite-coated ceramic bars, with a resistance of 25 kΩ each. This arrangement is used for creating a potential with a slight slope towards the exit. Applying a voltage of 200 V to the one end the local potential increases by about 100 meV. For fast ejection, e.g. for TOF-MS [luc01] suitable axial potential gradients have been applied during ion extraction. The electrodes can also be used for testing in-homogeneities of the trap potential caused by surface effects and for reducing their in-fluence. Important experimental parameters are the extraction voltage which, in general, should be chosen rather low, and the kinetic energy the ions have in the quadrupole. These parameters influence the extraction efficiency, the time distribution, as well as the mass resolution. This energy, which is basically determined by the difference of dc po-tential of the trap and the potential UFA applied to the quadrupole, was typically 2.3 eV.

The 22-pole ion trap is presented schematically in Fig. 3.12. It consists of 2 × 11 stainless steel rods of 1 mm diameter circumscribing a circle of 1 cm diameter. Rods are connected alternatively to the two phases of the rf power supply. The ion cloud is con-fined in radial direction by an effective potential which is created by the multipole field alternating in time. In the axial direction the trap is closed by small potential barriers created by a suitable voltage applied to the two gate electrodes. This system is enclosed by copper walls which are mounted on a closed cycle refrigerator that can be cooled down to 10 K. For electrical insulation but good thermal contact thin sapphire plates are used. This structure is surrounded by a second thermal shield held at about 50 K. Much more details can be found in other publications [ger93], [ger95], [pau94].

Fig. 3.12. The 22-pole ion trap consists of 2 sets of 11 stainless steel rods connected alterna-tively to the two rings which lead to the phases of the rf power supply. In radial direction the ion cloud is confined by an effective potential which is created by the multipole field. In the axial direction the trap is closed by small potential barriers created by a suitable voltage applied to the two gate electrodes.

Since the principles of ion confinement in inhomogeneous radio frequency fields have been explained thoroughly in [ger92b], only a brief overview is given here, focused on linear 2n-poles and ring electrode structures. The motion of an ion with charge q and

mass m in an electrical inhomogeneous rf field )tcos(E0 Ωr

can be decomposed in a

rapidly oscillating motion and a superimposed smooth trajectory. At high enough fre-quencies, the smooth trajectory can be calculated from the so-called effective potential


46

2

20

2

eff m4

EqV

Ω=

r

. (3.9)

The range of validity of this high-frequency approximation can be estimated using the adiabaticity parameter η ,

( )2

0

m

Eq2r

Ω

⋅∇=η

rr

, (3.10)

It is important to note that the effective potential is time independent which significantly simplifies the calculation of trajectories. The stability parameter η is usually a function of space and should be everywhere smaller than 0.3. Under such conditions the total kinetic energy is an adiabatic constant of the motion.

For 2n-poles, the effective potential, the adiabaticity parameter and the characteristic energy are:

( ) 2n2

0

20

eff r

rVq

8

1V

−

ε

= (3.11)

( )2n

0

0

r

rVq

n

1nr

−

ε

−=η (3.12)

20

22

rmn2

1 Ω=ε (3.13)

where V0 is the amplitude of the rf voltage and r0 is the inscribed radius of the multipole arrangement. The experiments in this work have been performed typically with a fre-quency f = 17 MHz and an rf amplitude V0 = 20 V.

For the RET, the effective potential and the adiabaticity parameter can be calculated in the way described in [ger92b]. The result is

)r(I

)z(cos)r(I)z(sin)r(I

zm4

)Vq(V

020

221

220

220

20

eff

+Ω

= (3.14)

))z(cos)r(I)z(sin)r(I()r(Iz

)z2(sin))r(I)r(I(V

m

2q22

122

0020

40

2221

20

20

2 +−

Ω=η (3.15)

In equations (3.14) and (3.15) r = r/z0, r 0 = r0/z0 and z = z/z0 are reduced variables, 2πz0 is the distance between middles of the two closest electrodes connected to the same phase of the rf generator and I0 and I1 are modified Bessel functions of zero and first order. Inspection of these functions shows that the effective potential and adiabaticity parameter of a RET are weak functions of the coordinate z in contrast to 2n-poles where they are independent on z.

Examples of calculated effective potentials for a quadrupole, an octopole, a ring elec-trode and a 22-pole trap are given in Fig. 3.13. Comparison of these four normalized curves reveals the advantage of using higher order multipoles or ring structures. They have a much wider field free region than the 4PT which is mandatory if one wants to minimize rf heating of the stored ions.


47

0,0 0,2 0,4 0,6 0,8 1,00,0

0,2

0,4

0,6

0,8

1,0

22PTRET

8PT

4PT

Vef

f(r)

/ Vef

f(r0)

r / r0

Fig. 3.13. Calculated normalized effective potentials for quadrupole, octopole, ring-electrode and 22-pole trap.

Adequate conditions for ion trapping can be estimated by using simple calculations. Usually one chooses the effective potential high enough for safely storing all ions in a volume of radius r = 0.8 r0, Combining the equations for the effective potential and the adiabaticity parameter and using η < 0.3 and r = 0.8 r0 leads to the minimum of the fre-quency and voltage needed.

3.2.3. Selected tests and outlook To determine absolute rate coefficients, the number density of the neutral target gas must be accurately known. The measuring and calibration procedures are described in detail in Section 3.2. Since parameters such as the sensitivity of the ion gauge, the pumping efficiencies, the conductance of differential walls, or the instruments used may change, it is necessary to check from time to time the calibration factors C. As an exam-ple Fig. 3.14 shows the relation between pressures, measured VT-22PT with two indi-cated gauges for H2. The linear fit provides the calibration factor for this machine. De-viations from the linear fit at pressures above 10-2 mbar are due to nonlinearities of the spinning rotor gauge in the transition region from free molecular flow to viscous gas flow.

In the VT-22PT machine, ions are extracted from the storage ion source and mass se-lected in the first quadrupole. In all experiments, performed in this thesis, the mass filter has been operated in the mass selective mode. In order to illustrate the resolution of this device, Fig. 3.15 shows two mass spectra which have been recorded by scanning the second quadrupole. In one case the first quadrupole has been operated in the rf-only mode (U0 = 0 V), in the second case, the mass-selective mode, the dc difference was set to U0 = 11 V. In both cases the rf amplitude was the same, V0 = 66.7 V. In such tests, the ion trap is used just for guiding the ions from first to the second quadrupole. This figure illustrates that the intensities vary through several orders of magnitude and that the suppression of the unwanted higher and lower masses is very high. It is necessary to mention that the plotted mass-spectra are an average over 10 scans and that a value of 0.1 has been added to the count rates to be able to plot the results on the selected loga-rithmic scale.


48

1E-10 1E-9 1E-8 1E-7 1E-6 1E-5 1E-4 1E-31E-8

1E-7

1E-6

1E-5

1E-4

1E-3

0,01

0,1

1

pig / mbar

p srg

/ mba

r

Fig. 3.14. Calibration of ion gauge against spinning rotor gauge (for location see Fig. 3.4) for H2 for VT-22PT machine.

32 34 36 38 40 42 44 46

0,1

1

10

100

1000

mass-selective mode(U

0=11 V, V

0=66.7 V)

rf-only mode(U

0=0 V, V

0=66.7 V)

Inte

nsity

/ io

n s-1

mass / u Fig. 3.15. Filtering properties of first quadrupole: mass spectra obtained using of rf-only (U0 = 0 V and V0 = 66.7 V) and mass-selective mode (U0 = 11 V and V0 = 66.7 V) for first quadrupole and mass scanning with a second quadrupole.

The measurement sequence for the C-RET machine is fully described in the two manu-scripts in Appendix A and B. In order to test this sequence and the overall performance of the machine, rate coefficients for the reaction

D3+ + N2 → N2D

+ + D2 (3.16)

have been measured.

The primary ions have been produced in the way described in A, leading predominantly to D3

+ ions and some HD2+. The number of these ions is two order of magnitude smaller

then that of D3+ ions. Neutral target gas N2 is introduced into the chamber. During the

storage time, ions react with the neutral reactants. After a given storage time, the trap content is extracted and mass analyzed. Experiments were performed without any buffer gas, at conditions where the trap was hold at room and at liquid nitrogen temperatures.


49

In Fig. 3.16 typical experimental results are shown. The trap was been hold at liquid nitrogen temperature. Plotted is the time dependence of the averaged number of trapped ions as a function of storage time. In addition to the reaction (3.16), the following proc-esses have been observed:

D3+ + 14N15N → 14N15ND+ + D2 (3.17)

HD2+ + N2 → N2D

+ + D2 (3.18)

HD2+ + N2 → N2H

+ + D2 (3.19)

The solid lines in Fig. 3.16 show solutions of an appropriate system of differential equa-tions describing the competing reactions indicated in equations (3.16) – (3.19).

0,0 0,5 1,0 1,5 2,010-1

100

101

102

103

Σ

14N15ND+

N2H+

HD2

+

N2D+

D3

+

t / s

Ni

Fig. 3.16. Typical experimental result showing the time dependence of an average number of trapped ions as a function of time. D3

+ ions are stored in a RET which have been hold at liquid nitrogen temperature. In reaction with N2 ([N2] = 2.06 × 109 cm-3) they produce N2D

+ and 14N15ND+. HD2

+ ions also react with N2 forming N2H+.

At room temperature the trap also contains traces of water and additional product ions are observed. These reaction channels complicate the data evaluation, especially since the rate coefficients for reactions between H2O and D3

+ and HD2+ are not known. There-

fore these rate coefficients have been measured by calibrating the number density of H2O using reactions with known rate coefficients. Despite these complications, rate coefficient for reaction (3.16) have been extracted by using an simplified model, de-scribing the reactions occuring in the trap. In this model it was assumed that primary D3

+ ions react either with the target N2 producing N2D+ (reaction (3.16)) or alternatively

with impurities (summarized in X) producing Y+ ions. Reactions with these neutrals X, i.e.,:

D3+ + X → Y+ + Z (3.20)

can be just treated as loss. In this way, the reaction rate coefficient for reaction (3.20), klos, accounts effectively for all reactions of D3

+ with parasitic neutrals from residual gas. With this simplification the decay of D3

+ ions via reactions (3.16) and (3.20) and the appearance of N2D

+ ions can be written as:


50

[ ] [ ] t

033los2N

los2N

eDDττ

τ+τ−

++ = (3.21)

[ ] [ ] )e1(DDNt

losN

los032

los2N

los2N

2

ττ

τ+τ−

++ −τ+τ

τ= , (3.22)

where [D3+] and [N2D

+] are the number of ions after a reaction time t, [D3+]0 is the ini-

tial number of D3+ ions, and [N2] and [X] are the number densities of N2 and the impuri-

ties, respectively. The time constants are

[ ] [ ]Xk1

,Nk1

loslos

21N 2

=τ=τ . (3.23)

In this way, the rate coefficient for reaction (3.16) has been determined to be k1 = (1 ± 0.25) × 10-9 cm3s-1. This value is in agreement with previously reported values of (7.49 ± 0.75) × 10-10 cm3s-1 and (1.1 ± 0.17) × 10-9 cm3s-1 measured with an ICR ap-paratus (see [ani03] and references therein). In addition to these test measurements, similar experiments were used for determining in the RET the number density of those neutrals which emerge from the carbon source. Results are shown in Section 3.3.3.

Both instruments used in this study would benefit from some improvements in the setup. Lower temperatures, down to 4 K, can be reached by mounting the 22-pole onto a new type of cold head. This new machine is close to completion.

In general, the C-RET machine needs some more improvements. Many of interesting ions produced by an electron bombardment of neutral precursors are fragments of the parent ion, leading to mixture of secondary ions. In order to simplify the chemistry in the ion trap, a first mass selection quadrupole between the ion source and the ion trap has to be introduced. Many carbon based molecules and ions have isomers which can be distinguished by chemical probing and therefore injection of probing gases through a piezo valve would be beneficial in the experiments. In order to prepare different ion isomers for studies of reactions with neutral carbon, an additional valve for introducing the gas into ion source could be used.

3.3. The carbon source The astrochemical importance of reactions involving carbon atoms, molecules, and clus-ters has been emphasized in Chapter 2. As a consequence, many attempts have been made in recent years to perform dedicated experiments with C-atoms, C2 or C3 mole-cules and with other forms of carbon. On one side, the preparation of a beam of C atoms or molecules from solid carbon is a technical challenge due to the very high enthalpy of formation of 716.49 kJ/mol. On the other side, there has been such a wide variety of technical and theoretical experience collected in more than hundred years that it is al-most impossible to get a full overview over the existing literature. Applications range from combustion studies, light sources, electron microscopy to passively cooled graph-ite target as muon collider [hai02]. Theoretical know-how includes classical data from standard tables, phase diagrams etc; however, there are still many unsolved problems [ker01].

In the preparation of the research project [cer99], this thesis is a part of, several carbon sources have been considered, based on laser ablation, sublimation of graphite and evaporation of suitable carbon containing compounds. One candidate was and still is the


51

very intense and cold pulsed Smally type carbon source using laser ablation of graphite at 266 nm [kai95], [kai01]. The source which became the central approach of this thesis, is based on vaporizing, by directly heating, graphite rods with a current of several hun-dred Amperes. The basic design which has been improved in some respects (see the paper in Appendix A) was developed in Heidelberg [cer98]. In order to understand the mass and velocity distribution and the internal temperature of the carbon beam, dedi-cated tests of the carbon source have been made. As will be discussed in more detail below, it became clear that the emitted carbon flow does not sublimates from the sur-face area of the rods but emerges in a pulsed mode from very hot gas or may be liquid bubbles enclosed in the contact area between the two rods pushed together.

The following chapter first reviews the most important physical and chemical properties of carbon, which are relevant for understanding all the processes which finally lead to emission of carbon from the contact region. However, it must be mentioned that it was not the primary aim of this thesis to completely characterize this complicated system. Despite all the dedicated tests described below, the details of the dynamics of carbon emission from the source used are not yet fully understood. Nonetheless, the results will be a good starting point for future improvements of a hot and intense C3 beam needed for understanding carbon condensation a high temperatures, e.g. in the outflow of car-bon rich stars.

3.3.1. Properties of carbon Carbon is in the focus of experimental and theoretical studies for many decades because of its fascinating properties arising from differences in electronic structure of the carbon bonds. In the solid form, carbon atoms prefer two forms of electronic bonding, the sp2 (typical for graphite) and sp3 types (typical for diamond). In addition to the best known forms of carbon, graphite and diamond, many other solid forms of carbon are known such as pyrolytic graphite, amorphous carbon, fullerite, glassy carbons as well as other structures. Graphite is black, greasy, highly compressible, good electrical conductor, has low density and crystallizes in hexagonal system.

Many experimental and theoretical efforts have been put into the investigation of the carbon phase diagram. Melting graphite is an old problem, which is still not yet com-pletely solved. The experimental investigations of liquid carbon are hampered by the very high temperature and high pressures which are required to reach the interesting regions.

A phase diagram for carbon is shown in Fig. 3.17. This figure which is slightly simpli-fied, has been taken from [ker01]. It shows several interesting features which are not yet completely understood. Several experimental data are included, the different symbols and the references are explained in the figure caption. The dotted line is an extrapolation of the experimentally determined border between graphite and diamond towards higher pressures and temperatures. An accurate experimental determination of the triple point would be of great help in determining the heat of formation for C3, C2 and C species. Results from theoretical calculations are presented as solid lines. The dashed border line between graphite and liquid represents a melting curve which is obtained if one assumes that only monatomic species C are present in the liquid. This curve increases monotoni-cally from the graphite-diamond-liquid triple point towards lower pressures. If, how-ever, the calculation also includes polyatomic molecules in the fluid phase, especially C2 and C3, the results (solid curve) match nicely the experimental data. Inspection of this region of the T(p) diagram reveals that the border between graphite and liquid has a maximum close to 5 × 104 bar. From this it can be concluded that, at any pressure below


52

this value, liquid carbon is less dense than graphite. The presence of polyatomic mole-cules lowers the density of the liquid phase and leads to significant changes in the com-pressibility [ker01].

p / 104 bar

5 10 15 20 2500

1

2

3

4

5

6

T /

103

K

Diamond

Graphite

Liquid

Fig. 3.17. Phase and transition diagram for carbon [ker01]. Experimental data are presented with symbols: × [tog94], + [ken76] and [mus98]. The dotted curve is an extrapolation of ex-perimental data measured from 1250 to 2900 K [bun80]. Results of theoretical calculations are presented with dashed and solid curves [ker01]. The dashed curve is obtained assuming that only atomic C is present in the liquid. The solid lines are obtained from the calculations when polyatomic carbon molecules are included in fluid phase.

Theory also indicates that the conductivity of liquid carbon changes significantly with pressure. At low pressure it is an insulator but it behaves like a metal at high pressures. Some recent experimental results [bun96] do not support the prediction that a non-conducting liquid phase exists. Another theoretical result which is important for this work is, that C3 is the most important species in the liquid phase, even at 6000 K while C4 and C5 are unimportant at all temperatures and pressures. Near the melting region, the concentrations of C and C3 in the liquid are comparable in weight percent [ker01].

Fig. 3.18 shows a pT phase diagram for graphite in the lower pressure range taken from [bun96]. The melting line at low pressures is well established; however, there are uncer-tainties of a about hundreds degrees in the region of the graphite-liquid-vapor triple point. The generally accepted value of the graphite-liquid-vapor triple point [bun96] is around (p,T) = (100 bar, 5000 K).

At temperatures below the graphite-liquid-vapor triple point, carbon does not evaporate, it can only sublimate. In order to estimate the amount of material which is transferred from a hot surface into vacuum, the sublimation pressures must be known as a function of temperature. An approximation for the saturation pressure which is sufficient for most estimates in the 1500 K to 3000 K range is given by p = 3.2 × 1010 exp(-94700 K/T)) atm . This formula has been used for example in a re-cent report, in which quantitative sublimation tests have been performed on a graphite rod which will be exposed to 35 kW in a neutrino factory [hai02]. In general more pre-cise calculations of the vapor pressure are possible (e.g. [ker01]) and lead to very good agreement with the available experimental data.


53

0.1 0.2 0.300

4

5

6

p / 104 bar

T /

103

K

Graphite

Liquid

Vapor

Fig. 3.18. pT phase diagram for graphite [bun96] at low pressures.

For the purpose of the present work, the formation of a carbon beam, one also has to account for the mass and velocity distribution of the material emerging from the carbon source. It is well-known that the ejected material consists only to a small fraction of carbon atoms, the rest are Cn molecules with n=3 being dominant. This is illustrated in Fig. 3.19.

2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 300010-9

10-8

10-7

10-6

10-5

10-4

10-3

10-2

10-1

100

C2

C1

C3

T / K

p / m

bar

Fig. 3.19. Dependence of partial pressures of Cn species (n = 1, 2, 3) on temperature. Thin and thick solid lines show experimental results from [dro59] and [gin94] respec-tively. The dashed curves are the theoretical results from [pal67].

The thin solid lines are results from mass spectrometric studies [dro59] for the C and the molecules C2 and C3. Species larger then C3 have been omitted due to their negligible abundance. In this study, the ions have been produced by electron bombardment using a


54

kinetic energy of 17 eV, at which energy dissociative ionization can be neglected (see below). Some newer experimental results are plotted in the same figure by thick solid lines [gin94]. In this study, a high temperature Knudsen effusion mass spectrometric method has been applied for determining the equilibrium partial pressures of C1 - C7 above pure graphite. Finally Fig. 3.19 also shows results from calculations [pal67] as dashed lines. Comparison reveals that there is good overall agreement between experi-ment and theory.

In order to emphasize the temperature dependence of the relative abundance of C, C2 and C3 in the carbon vapor, Fig. 3.20 presents the relative percentages of these three species.

2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 30000

10

20

30

40

50

60

70

80

90

100

C3

C2

C1

T / K

Per

cent

age

of C

1, C

2 an

d C

3

Fig. 3.20. Dependence of relative percentage of Cn species (n = 1, 2, 3) on temperature. Thin and thick solid lines show experimental results from [dro59] and [gin94], respectively. The symbols present experimental data from [bur64]. The dashed curves are the theoretical results from [pal67].

The dashed curves are the theoretical results from [pal67], thin and thick solid lines are the above discussed experimental data from [dro59] and [gin94], respectively. The points plotted at 2500 K are experimental data taken from [bur64]. In this experiment, carbon vapor from a Knudsen cell was ionized by electron impact and analyzed using a mass spectrometer. Electrons with an energy of 15 eV, 17 eV, and 20 eV have been used. In addition the kinetic energy distribution of the ions has been measured. From this it has been concluded that all ions have been formed directly from the neutral par-ents and that fragmentation is negligible. In this way, rather precise relative values for the ion intensities have been determined, C1

+: 11.8 %, C2+: 7.6 %, and C3

+: 80.6 %. In order to convert the ion intensities into the corresponding percentages of the neutrals, one should account for the relative ionization cross sections. However, electron impact ionization cross sections as function of electron energy are only known for C atoms [kim02]. These experimental data are in good agreement with theoretical cross sections which are also available for C2 and C3 molecules [deu00]. The calculations indicate that the three ionization cross sections of interest are very similar at the collisions energies used in the above mentioned experiment. Therefore the reported ion intensities repre-sents also the percentage of parent's neutrals. In summary, it can be concluded from


55

Fig. 3.20 that C3 molecules dominate with about 80 %. in the temperature range be-tween 2000 K and 3000 K with an increasing tendency towards higher temperatures.

3.3.2. Technical description The working principle of the carbon source, some details of which are presented in Fig. 3.21, is based on electrical resistive heating of graphite rods in a high vacuum environ-ment. This carbon source has been used already in an apparatus at the MPI-K in Heidel-berg for spectroscopic studies on matrix-isolated carbon molecules [cer98]. On the basic construction only minor changes have been made while the operation mode of the source has been changed and improved leading to a much better stability.

10 cmIG

TP2

TP1

A3A2A1carbonbeam

A/D1

SMD

cathode

SM

A/D2

ICS

D/A

CB

RS

PSU

VC

H2OH2O

H2O

anode

serial bus

PC

V

Fig. 3.21. The scheme of carbon source.

Two graphite electrodes (cathode 4.6 mm and anode 6.15 mm diameters) are pressed together and resistively heated. The electrodes used (RW4 type) are produced by SGL Carbon. They have less then 2 ppm impurities.

For heating the electrodes high current and low voltages have been used, in contrast to the plasma discharge arcs. The required electric power is provided from a power supply (PSU). The PSU consists of two pairs of 5V/200 A DC switching powers-supply units (Philips PE 1980), which are connected in a serial-parallel way. Using this scheme, the


56

PSU can provide current up to 400 A at a maximum voltage of 10 V. In order to exter-nally control and monitor the PSU and the overall operation of the carbon source, a con-trol computer (PC) has been used. The PSU is connected to the PC via a digital-analog converter (D/A) and an optically isolated serial bus. During operation, the potential of the contact area between the electrodes usually has been held at ground potential. By applying an external voltage Vc (see Fig. 3.21) this potential can be changed. At some operational conditions, one can expect ion emission from the contact area. Applying a suitable potential Vc ions can be efficiently suppressed or also extracted. Under the op-erational conditions used in this work, no ions could be detected at all, despite a very careful search.

A very sensitive characteristic of the operating conditions of the source is the rod resis-tance which is monitored continuously. The resistance is measured by using a four-point scheme. The voltage drop is determined using the analog-digital converter A/D1 con-nected to the outer ends of the electrodes. The supply current is measured with an induc-tive current sensor (ICS) which is connected to the PC via the analog-digital converter A/D2 and a serial bus. As ICS, the current transducer HA 400-SB (LEM Components) has been used. From the heating voltage and current, the control computer calculates the heating power and the rod resistance. To obtain a stable beam of carbon, the heating power needs to be regulated as well as the pressure by which the carbon rods are pushed against each other. The PC controls the PSU such that the heating power is stabilized even when the rods resistance is changed.

The resistance of the rods strongly depends on the temperature. For the electrodes used (see above), the resistance drops down to typically 20 mΩ during the heating. The nec-essary low-resistance connections are made by the two copper blocks CB which hold mechanically the graphite rods. These copper blocks are made from oxygen-free hydro-lytic copper. Each of them is connected to a pair of high-current electrical feedthroughs. The blocks, the feedthroughs and the whole chamber are water-cooled. The PSU is con-nected to the feedthroughs by 150 mm2 copper cables because the typical supply current is 250 A. The voltage is rather low, around 5 V. The power required to start vaporiza-tion is about 1 kW. For high carbon emission, powers up to 1.3 kW are used. It has been found that the polarity of the voltage applied to the electrodes does not play a significant role. There are some indications that the efficiency is slightly better if the thin electrode it the cathode, i.e. if the C+ ions hit the thinner of the two electrodes.

The huge power dissipated during operation leads to heating of the vacuum chamber and to gas desorption from the walls. To reduce this effect, a copper radiation shield (RS) with an aperture of 4.5 mm diameter has been used to screen radiation from the graphite rods.

For proper and stable operation of the source, constant mechanical stress between the electrodes has to be applied. In order to maintain this and to obtain maximal beam in-tensity, both electrodes can be moved in axial direction. The outer ends of the electrodes are mounted to electrical feedthroughs in the flanges which are attached to the source chamber via vacuum bellows of types ZLTM100 and ZLTM100W (Vacuum Genera-tors). In order to limit the mechanical stress, the cathode connection is equipped with a spring. The anode can be manually moved in order to center the contact place between electrodes and to obtain maximum available beam intensity. Since two different diame-ter graphite rods are used, only the thinner one (cathode) vaporizes. Therefore the cath-ode is moved by a stepping motor (SM) for keeping the electrodes in contact and to achieve constant stress between them. The remote control of the movement is per-


57

formed by a stepping motor driver (SMD) connected to the PC via serial bus. During the operation of the carbon source, the mechanical stress between the electrodes cannot be measured accurately because of the friction of the electrodes in the copper blocks. To estimate the stress between electrodes the steep dependence of the resistance is used as a measure.

The automated control scheme of the carbon source, which has been developed, is de-scribed in detail in paper A in the appendix. The mean value of the contact resistance and its variations are followed continuously and, from this information, the required motion of the stepping motor is calculated. Two control programs written in LabVIEW are used in parallel. It also was tried to record the temperature of the rod and to control the emission using this parameter (see below).

3.3.3. Test measurements and outlook After starting the carbon source, it takes some time until optimal working conditions are reached, i.e., a high number density of gas phase carbon species and reasonable stabil-ity. It has been found that the best way to reach such conditions is to increase the tem-perature of the rods by increasing the heating power in small steps. Each time, when a new electrode has been put in, a huge desorption of gas has been observed in the power range between 350 W and 700 W. This is due to the air which is captured in the porous electrodes, which comes out during the first heating. In order to clean the carbon rods, new electrodes have been heated under high vacuum conditions for typically 5 - 7 hours with 600 - 700 W of electric power. Only after this cleaning procedure, the electrodes can be used properly.

One of the important parameters, which has to be determined, is the carbon vapor com-position. In the present experiment the C, C2, and C3 fractions have been measured by ionization the carbon beam using electron bombardment in the ion source (see 3.2.1). The ions produced are mass selected and counted. During such measurements, the ring electrode trap was used just for guiding the ions. The electron energy has been selected either such that the beam is efficiently ionized or such that fragmentation of the carbon molecules is avoided as much as possible. In the evaluation, fragmentation of C3 or C2 leading to formation of C+ has been accounted for using the values reported in [mai00]. From a series of careful studies performed in the C-RET machine, it has been estimated from the measured ion intensities that 80 ± 8 % of the carbon beam are C3 molecules, 8 ± 2 % C2, and 12 ± 6 % are C atoms.

It has been mentioned in Section 3.2.1 that the number density of the neutrals in the trapping region must be measured accurately for determining absolute rate coefficients with low uncertainties. The determination of the neutral carbon species coming from the carbon source is obviously more complicated than the determination of the number den-sity of a permanent gas, let into the trap via a leak valve. In addition, due to possible, unpredictable instabilities of the carbon source, the number density has to be monitored rather often. In all experiments of this work, a test measurement has been made after each trapping sequence. In a first approach, from a variety of methods for determining number density, chemical probing described in appendix A, has been used. However, due to the large uncertainties in the published rate coefficients of the reactions used a different method described below has been used. This allows us in addition to determine better values for the astrochemically important reaction D3

+ + C3.

The method used to determine the carbon number density is based on electron bom-bardment using a calibrated ionizer. In situ calibration has been achieved by utilizing an


58

additional gas the number density of which has been determined precisely using another independent technique. Each gas can be introduced into the main chamber and ionized there by electron bombardment using an ion source. The number of ions, N, produced and detected in this way, is proportional to the electron current and the number density n of the neutrals. The proportionality factor ε, so called efficiency, includes the efficiency of the ion source, the electron-molecule ionization cross section, various ion transmis-sion properties of the machine and the detection efficiency. In the present study, N2 has been used and the number density has been determined using the procedure described earlier in Section 3.2.1. Then, following equations can be written:

222

NeNNnNN −+ ε= (3.24)

333

CeCCnNN −+ ε= (3.25)

In these equations, the indexes indicate different species. Assuming the same transmis-sion property and detection efficiency for N2

+ and C3+ ions, efficiencies for N2 and C3

depends only on the electron bombardment ionization cross sections σ at a given en-ergy:

2

3

2

3

N

C

N

C

σσ

=εε

(3.26)

From equations (3.24), (3.25) and (3.26) the number density of C3 is:

23

23

3

NC

N

e

CC

1N

Nn

εσσ

=−

+

(3.27)

In order to test the measurement sequence for C-RET machine described in Appendix A and B and to measure ε for N2 test measurements on reaction D3

+ + N2 → N2D+ + D2

described in 3.2.3 have been performed. During so-called "calibration" phase, at given number densities of N2 and electron current, the number of produced ions is followed. After many repetitions, the efficiency for N2 of (3.7 ± 0.3) × 10-21 cm3 is measured. In this way, knowing this efficiency, electron current and following number of C3

+ ions, number density of C3 can be measured, since the ratio of electron bombardment ioniza-tion cross sections have been calculated from data reported in [kim02] and [deu00]. The C3 number density has been determined directly while the number densities of C2 and C have been estimated from the carbon vapor composition. The ionization method also could be applied directly for C2 and C species if it would be necessary, e.g. if one wants to study reactions with C2 or C neutrals.

In the final estimate of the overall uncertainty in the determination of the target gas den-sity (± 50%), it has been also accounted for that the calibration gas is distributed isot-ropically while the carbon beam has a rather well-defined angular and spatial distribu-tion in the trap. It is also necessary to mention, that the flux of the emitted carbon spe-cies is not only a simple function of the heating power applied on the rods. It has been found that the applied stress between the rods and the location of the contact area has a strong influence of the emission characteristics. More details are discussed below.

From the measured change in length of the electrode, a crude estimation of carbon number density in RET can be estimated. Typically, 7.5 cm of the rod have been vapor-ized during 8 hours of operation. That corresponds to 2.6×10-3 mm/s. Using the rod di-ameter and the specific density of graphite, on obtains that a mass of 7.6×10-5 g/s is va-


59

porized. It has been found that only 25 % of vaporized graphite goes to the gas phase and therefore, the transferred into the phase per unit time is 1.9×10-5 g/s. Assuming that all carbon in gas phase is in form of C3 having a temperature of 3000 K, at distance of 40 cm (distance from RET to contact place between rods), a number density of 1 × 108 cm-3 has been obtained. The number density estimated in this way is in agree-ment with the measured typical values.

In an attempt to control the number density via surface temperature and to estimate the temperature of the emitted carbon vapor, a temperature profile of the rods has been measured using an optical pyrometer [dec03]. At the optimum working conditions (around 1.25 kW of applied heating power) a temperature of 2000 K has been measured at a distance of 0.1 mm from the hottest observed place - contact between the two elec-trodes. Using an optical filter (to protect eyes) in addition to our optical system which has been used for magnifying the area over which temperature should be measured, this really hot area can be seen as a very thin and bright stripe, much thinner then 0.1 mm. Due to limitations of the optical scheme used and the low spatial resolution of 0.1 mm, recording a more detailed map of the emitting area was not possible. From the net car-bon emission it must be concluded that the temperature inside the contact area must be much higher.

It has been found that an oscillation of the rod resistance, being in phase with a low am-plitude temperature oscillation is a clear sign for effective vaporization. During stable operation of the carbon source the resistance varies quasiperiodically by approximately 10 %. If one keeps the heating power and the position the of electrodes unchanged after the onset of vaporization, the amplitude of these oscillations increases, also the emitted flux becomes higher until, finally, the contact between the electrodes is interrupted. By stabilizing the heating power and moving the electrode such that the resistance ampli-tude remains constant, quite stable carbon emission can be reached.

These oscillations can be explained with a simple model. Increasing the temperature of the hottest place – the contact between the electrodes - vaporization of carbon rod be-comes more effective. At high enough temperatures, the carbon gas enclosed in the con-tact between electrodes reaches a pressure high enough to push the electrodes apart. This leads to an increase of the resistance since the contact area between the electrodes is reduced. Simultaneously, the temperature becomes higher further increasing gas pres-sure. Finally, at some pressure part of the gas finds a way to escape. This leads to a de-crease of the resistance and the temperature. This quasi-periodic oscillation explains also the observation that the number density at applied constant heating power decreases if the applied stress is increased.

To support the described model and estimate the temperature in the contact area, a sim-ple estimation can be used. Using the above mentioned typical value for mass trans-ferred to the gas phase per unit time and assuming that only C atoms are emitted, a total of 9.5 × 1017 carbon atoms are ejected per second. Assuming further that the mean thermal velocity of C atoms corresponds to a temperature of 3000 K, the number den-sity of carbon atoms is 3×1012 cm-3 at a distance of 1 mm from the rod surface. Calcu-lating the corresponding partial pressure and converting it in accordance with [hai02] into temperature, a value of 2500 K is obtained.

During the operation of the carbon source moving the thin electrode is necessary for stable carbon vaporization. The required mechanical pressure has not been measured; however, it is estimated to be in the order of 10 N/cm2. Assuming that the volume in the


60

contact area is filled to 50 % with carbon gas, a vapor pressure of 2 bar corresponding to 4000 is needed.

The carbon source described in the Chapter certainly needs more characterization and improvements; however, it can be foreseen that it will become an important tool for studying high temperature carbon reactions. With this goal, the next step will be to care-fully measure the translational and internal energy of the emerging carbon species by using spectroscopic methods. REMPI, LIF or may be even absorption spectroscopy close to the emitting areas will give a deeper insight into the complicated vaporization prozess.

In the present setup, where the RET is at a distance of 40 cm from the carbon source, a maximum of 2 × 108 neutral carbon species per cm3 has been reached. This number density is just at the lower limit for studying radiative association reactions between C3

+ ions and neutral C3. This is partly due to the distance; however, the source is also very inefficient since it emits carbon in 4π. In addition, a large fraction condenses and finally falls down. It is hoped that a more collimated beam can be obtained by additionally heating a spot in the contact area by using a focused laser. The overall stability of ejec-tion process also will be improved if the force that presses the two rods together, can be further stabilized by using a constant or modulated pneumatic control.

Alternatives for the carbon source used are laser vaporization based carbon sources. For studying reactions with cold carbon atoms, a Smalley type laser vaporization source will be used. In addition, seeding the carbon beam in various fast expanding carrier gases, the kinetic temperature of the carbon beam can be varied.

61

4. SUMMARY, CONCLUSIONS AND OUTLOOK

Carbon plays an extremely important role in the overall physical and chemical evolution of the ISM. To understand complex chemistry in particular regions of ISM occurring at extremely different temperatures, detailed studies over a huge temperature range are needed. With the aim to improve our knowledge in this field, selected laboratory ex-periments on ion-molecule reactions have been performed and the results are presented in this thesis.

The major aim of this research was originally to study gas phase interactions between neutral carbon atoms and small molecules with ions. In addition corresponding reactions of carbon ions with neutrals have been included into the experimental program.

Since high temperature ion-molecule reactions are important for understanding the chemistry in stellar outflows, shocks, stellar atmospheres, etc., a neutral carbon beam source, well suited for studies of reactions on hot and small neutral carbon species, was integrated into an ion-trapping machine (C-RET) allowing for the first time such stud-ies. First experimental results on reactions between hot Cn (n = 1, 2, 3) and D3

+ cations has been obtained at an effective temperature determined by the ions stored in 80 K trap and estimated 3000 K of translational temperature of Cn. Measured reaction rate coeffi-cients are roughly a factor of two smaller than values often used in astrophysical mod-els. In addition, an upper limit for the association reactions Cn

+ + C3 has been derived from new experiment.

The used trap, the RET can be operated over a huge temperature range from 80 up to 600 K or even higher. It is aimed to extend the experiments towards higher ion tempera-tures. Much higher ion kinetic temperatures can be reached by operating the machine in a guided ion beam arrangement. It would be of special importance to measure reactions of the stored ions having very high internal temperatures, e.g. prepared by CO2 laser heating. On the other hand, for understanding the chemistry for example in cold dark clouds, experiments performed at very low temperatures are necessary since it is well known that rate coefficients for some reactions are strongly temperature dependent. In addition, ion-molecule reactions at extremely low energies as well as using different isotopes give insight in collision dynamic and fundamental physics in general. There-fore, studies of reactions of carbon at low temperatures are extremely important. For studying reactions with cold carbon atoms, for example a Smalley type laser vaporiza-tion sources can be used.

In order to complete the knowledge about the ion chemistry involving three carbon at-oms, detailed investigations on reactions of C3

+ and C3H+ with H2 have been performed

between 15 K and room temperature by use of the VT-22PT machine. The temperature dependent reaction rates have been parameterized. For the reaction C3

+ + H2 it is un-known why the increase in lifetime of the C3

+ + H2 complex with falling temperature leads to more C3H

+ products at 15 K. For the first time, formation of C3H2+ ions via

radiative association of C3+ and H2 has been measured. It has been found that C3H

+ ions formed in the exothermic reaction of C3

+ with H2 as well as C3H+ and C3D

+ ions formed in the reaction C3

+ + HD are most probably excited. Using HD instead of H2 in reactions with C3

+ it has been found that they proceed with roughly the same reaction rate at 15 K. An isotope effect has been detected for this reaction. The C3D

+ product ions are slightly more abundant than C3H

+.

In collisions between C3H+ and H2, competition between hydrogen abstraction and ra-

diative association has been observed. Reaction between C3H+ and HD at 15 K is 3.3

4. SUMMARY, CONCLUSIONS AND OUTLOOK

62

times more efficient than reaction with H2. The dipole moment of HD may play a role in enhancing reaction efficiency. In addition it was shown that most probably HD and H2 scrambling occurs. The present data fully support a model [sor94] that assumes that there are potential barriers both in the entrance and the exit channel.

In the reactions of C3+ and C3H

+, use of HD follows pattern that radiative association reaction efficiency decreases with the level of deuteration of neutral reactant.

To get a deeper insight into the reaction dynamics involving hydrocarbons, experiments on reactions of H and D atoms with C3H

+, C3D+, C3H2

+, C3HD+ as well as reactions C3D2

+ + H and C3D+ + H2 are needed. It can be foreseen that soon first results on these

reactions will be obtained from experiments ongoing in the Chemnitz group for Labora-tory Astrophysics. In addition detailed theoretical studies are needed to explain the ob-servations, i.e., the enhanced reactivity when HD is used instead of H2.

One of important results of these studies is the investigation of the C3H3+ + HD collision

since it has been suggested in the literature that C3H3+ + HD → C3H2D

+ + H2 can be responsible for large deuterium fractionation of c-C3H2 in dark clouds observed. The experimentally obtained upper limit for the rate coefficient, smaller then 4 × 10-16 cm3s-1 corroborates the conclusion from ab initio calculations that the H - D exchange reaction is hindered by a very high barrier and that therefore this reaction should not to be used in astrophysical models. On other hand, if in the dark clouds are enough of C3H

+ ions, reaction C3H

+ + HD → C3H2D+ + hν which proceeds with a 2.3 × 10-11 cm3s-1 may in-

fluence deuterium fractionation of C3H2.

To get a deeper insight into collision dynamics, additional experiments with n-H2, p-H2 and D2 at different temperatures have to be performed. Investigations on the reverse reaction will lead to the better understanding of collision dynamics.

First experimental results on reactions of hot neutral pure carbon species are encourag-ing to continue further investigations with other ions and at a variety of temperatures. Most interesting reactions are collisions between pure carbons from one side and hydro-carbons and fullerene cations from other side. Extending investigations into the direc-tion towards nanoparticles, the proposed way of growth of carbon atoms towards micro-scopic structures can be understood. Use of different isomer enrichment will give fur-ther insight into this field.

63

APPENDIX A

ION-TRAPPING APPARATUS FOR STUDIES ON REACTIONS BETWEEN IONS AND NEUTRAL CARBON SPECIES

I. Čermák, I. Savić, and D. Gerlich

University of Technology, Chemnitz, Institute of Physics, 09107 Chemnitz, Germany.

Abstract We present a recently developed ion-trapping machine that combines the trapping tech-nique with a beam of neutral carbon atoms and molecules. A trapping apparatus consist-ing of a radio-frequency trap and a single-ion mass analyzer was equipped with a graph-ite sublimation source. Electrical heating is used to vaporize the graphite, producing an effusive beam of neutral carbon atoms and molecules. A control program regulates the operating conditions to maintain the source stability. A special measurement sequence was developed to produce primary ions directly in the trapping device, to perform a standard storage experiment, and to monitor the concentration of the carbon vapor tar-get. As an example, we describe the measurements on the reaction D3

+ + Cn → CnD+ + D2 used to calibrate the emission properties of the graphite sublima-

tion source.

Introduction Carbon is one of the most abundant non-volatile elements in the universe. This fact, together with its chemical variability, explains the large fraction of carbon-bearing molecules among the molecular compounds identified in the interstellar and circumstel-lar matter so far. Many of the detected interstellar molecules are chain molecules - mainly highly unsaturated hydrocarbons and polyines containing up to ten and more carbon atoms (see e.g. the data compiled by [Nummelin, 2001]). The mechanism of pro-duction of these long chain molecules is not known in detail. The available astrochemi-cal reaction networks (see e.g. [Le Teuff et al., 2000]) can model the growth of different species in various regions of the universe. Ion-molecule and neutral-neutral reactions with bare carbon species belong to the most important channels producing chain mole-cules in the interstellar medium. From the plenty of reactions with neutral carbon mole-cules involved in these networks, only a few have been studied experimentally.

Carbon-bearing molecules are also common in various terrestrial processes. Reactions of neutral carbon molecules are responsible for the creation of large molecules in flames and for the formation of fullerenes from carbon vapor in a helium atmosphere. Ion-molecule reactions can also play a dominant role in these processes and may be in-volved in the yet unknown reaction path leading from small linear carbon molecules to cage structures like C60.

To study reactions between ions and neutral carbon species, we developed an ion-trapping apparatus that combines a radio-frequency ion trap with a beam of neutral car-bon molecules. The usage of the ion trapping technique makes it possible to perform experiments on bimolecular collision processes under conditions that are similar to those in interstellar space or in the vicinity of stars (see e.g. the review [Gerlich et al., 1992]). The aim of our research is to perform detailed studies on reactions leading to growth as well as to destruction of chain molecules. With our apparatus, we can investi-

APPENDIX A

64

gate collisions between ions and neutral atomic, diatomic, and triatomic carbon. In this paper, we describe our experimental setup and show some typical results of such colli-sion processes.

Experiment The experimental setup is shown in Figure 1. Small neutral carbon species (mainly C, C2, and C3) are produced in the sublimation source by resistive heating of graphite elec-trodes. The effusive beam is led through the apertures of the differential pumping stage into the main chamber, where it passes through the electrodes of the ion source (IS) and, finally, reaches the ion-trapping device. In the ring-electrode trap (RET, for details see [Luca et al., 2001] and references therein), the neutral carbon clusters collide with ions stored in the applied radio-frequency field. The ionic reaction products are extracted after a given storage time, analyzed by the quadrupole mass analyzer (QMA), and counted with help of a Daly detector. The ion source is used to produce primary ions and, alternatively, to monitor the concentration of the carbon molecules in the beam.

Carbonevaporator

Daly detector

QMA RET

Differentialpumping

TP TP TP

TP

IS

Figure 1. Scheme of the apparatus. From right to left: carbon evaporator, differential pumping stage, main chamber with the ring-electrode trap (RET) and ion source (IS), quadrupole mass analyzer (QMA), and single ion detector (Daly detector). Four turbo-molecular pumps (TP) evacuate the experimental setup.

The storage time of trapping setups is limited by the amount residual gas molecules that deplete the number of primary ions by parasitic reactions. A clean vacuum system is a prerequisite to reach high experimental sensitivity. We use four turbo-molecular pumps connected to rotary pumps equipped with zeolith adsorption trap to evacuate the appara-tus. The pressure routinely reached in the main chamber amounts to 5×10-10mbar. The ring-electrode trap can be cooled by liquid nitrogen to temperatures down to about 80 K. The attached radiation shield held at about 80 K also helps to pump out condens-able gases (mainly H2O and CO2). At storage times around one second, the parasitic reactions convert roughly only one percent of stored ions.

APPENDIX A

65

The trap is equipped with two different gas inlets for introducing gaseous targets. An ultra-high-vacuum sapphire leak-valve (continuous operation) and a piezoelectric valve (pulsed or continuous operation) are used to control the gas fluxes through each inlet. The fast pulsed gas inlet is able to produce a high peak concentration in the ion trap with comparatively low gas amounts. This results in gas pulses of millisecond duration, which is determined from the transport time of the gas from the valve into the trap and by the gas flow conductance of the trap apertures.

Carbon sublimation source The source has already been described elsewhere [Cermak et al., 1998]. Here, we only briefly summarize its function: Small carbon-vapor molecules are produced by electrical heating of two graphite rods. The graphite electrodes are pushed together by a stepping motor drive and, in the vicinity of the contact area, the carbon evaporates continuously. Due to the different diameters of the rods and due to the polarity of the applied heating current (analogous to an arc discharge), only the thinner anode evaporates.

Recently, we carried out several changes in the construction and operation of the source to increase its total operation time and stability. We replaced the 3 mm diameter anode rod by a one of 4 mm diameter. This required increasing the power supplied to the rods to about 1.0 to 1.3kW. Currently, we use four 5 V/200 A power supply units connected in a serial-parallel way. The typical rod resistance during heating is around 20 mΩ. The electrical power required for their sublimation is reached at heating currents of roughly 250 A. The power supply units are externally controlled by a computer, which also monitors the operating conditions of the carbon source (heating current and voltage). A control program actively stabilizes the heating power (see Fig. 2).

As the electrodes are being consumed by sublimation, the contact resistance between them fluctuates (see graph in Fig. 2). The changes of the operating conditions would be accompanied with an instability of the carbon emission. To avoid this, an automated stepping motor drive is used to maintain a constant mechanical stress between the elec-trodes. For controlling the stepping motor, we developed a control procedure based on monitoring the contact resistance. The program follows the mean value of the contact resistance and its variations. According to these values, the movement of the stepping motor is automatically controlled. In contrast to the old operation mode [Cermak et al., 1998], this new control scheme results in stabilized carbon emission with variations of around 10%. From the speed of the stepping motor, i.e. from the velocity of the anode, we derive the mass consumption of the graphite and estimate the carbon concentration in the ion trap (see the indicator fields Velocity and Density in the upper window in Fig. 2). For this calculation, we assume an effusive expansion of the carbon vapor molecules with a temperature of 3000 K and a cosine angular distribution.

APPENDIX A

66

Figure 2. LabVIEW programs for controlling the carbon sublimation source. The main program (Oven-PSU-Control.vi) controls the power supply units (Power [%]), monitors the supply cur-rent and the voltage between the graphite rods, calculates the contact resistance, stabilizes the heating power, and regulates the electrode shifting according to the measured contact resistance. The graph in the right part of the program window shows the time dependence of the measured contact resistance, the mean value and the variations (interval of 2×RMS) of the contact resis-tance between the graphite rods. A second program (Stepping-Motor-Driver.vi) controls the movement of the anode and monitors the consumption of the graphite. The operation of the carbon sublimation source is accompanied with a release of para-sitic molecules (mainly H2O, CO, and CO2). To reduce the excessive heating and the gas desorption from the walls, the evaporator chamber and the electric feedthroughs to the graphite electrodes are cooled by water. The differential pumping stage also effi-ciently improves the purity of the carbon beam. Using the ion source in the main cham-ber, we checked the composition of the gaseous species released from the carbon source. Also trapping experiments using deuteron transfer from D3,

+ to these residual gas molecules indicate a very low background consisting mainly of traces of CO. Even during the operation of the carbon source, the partial pressure of the residual gas in the trap is lower than 10-9 mbar.

Measurements Trapping experiments are usually performed in a pulsed mode. Primary ions can be pro-duced by electron bombardment either in an external ion source or directly in the trap and stored for times varying from microseconds to minutes. During the storage time, the ions interact with the neutral reactant coming in form of a beam of carbon radicals from the graphite sublimation source. After a selectable storage time, the trap is opened by lowering the potential on the exit electrode and the extracted ions are mass analyzed.

APPENDIX A

67

Ion preparation Preparation of atomic and many molecular ions can be performed directly in the trap-ping device without the necessity of an external ion source and a mass filter. We use the piezoelectric valve to introduce a short pulse of suitable gaseous target directly into the trap. After a certain delay, an electron pulse of several hundred microsecond duration is injected into the trap. The neutral gas is ionized by electron-impact and the created ions undergo interactions with the neutrals still present in the trap. The operating conditions can be adjusted so that the chemical reactions between the ions and the neutral gas pre-fer the production of a specific molecular ion.

Figure 3 shows an example of such an operation scheme. We used deuterium (Deute-rium 2.7, Messer Griesheim) as the gas target. To ionize it, electrons with energy of 70 eV were injected. The duration of the electron pulse was 500 µs. Electron-impact ionization at such energies produces not only D2,

+ but leads also to fragmentation creat-ing D+. The collision of D2,

+ with neutral deuterium forms D3,+. The correspondent re-

action D2,+ + D2 → D3,

+ + D is fast, occurring at approximately the collision rate. We stored the product ions in the trap for additional 10 ms and measured the mass spectra at various delays between the gas and ionization pulses. At large delays when the number density of the introduced gas is already low, the reaction between D2,

+ and D2 provides only a minor variation of the ion composition and the most abundant ion is D2,

+ as the direct product of ionization. At short delays of 5-10 ms, however, when the peak num-ber density of D2 is reached, the conversion of D2,

+ to D3,+ is dominant and D3,

+ is the most abundant product ion.

Delay / ms0 20 40 60 80 100 120 140 160 180 200

Ion

Cou

nts

per

Cyc

le

10-3

10-2

10-1

100

101

102

103

104

105

D+, H2+

HD+

D2+, H2D

+

HD2+

D3+

D2+

H2D+

Figure 3. Mass composition of ions produced in the trap by electron-impact ionization of D2 gas introduced into the trap by the piezoelectric valve as a function of the delay between the gas and ionization pulses. The mass spectra were obtained by averaging of about twenty measurement cycles. The profiles of D2,

+ and H2D+ (both mass 4 u) could be separated because H2D

+ domi-nates at short delays (analogous to HD2,

+ and D3,+) while D2,

+ signal is the major signal at longer delays.

Due to the isotopical composition of the introduced gas and due to experimental effects, the measured mass spectra show also other features. The linearity of the ion detector influences the number of detected D3,

+ ions at high count numbers, thus affecting the

APPENDIX A

68

measured HD2,+/D3,

+ ratio at delays of about 5 to 70 ms. The operating conditions of the ion trap (5.6 MHz, 150 V amplitude) make the adiabaticity parameter for lighter ions like D+ and HD+ comparatively high. Therefore, these species are not stored safely and only trace amounts of them are detected. Since the mass analyzer cannot distinguish between D2,

+ and H2D+ (both mass 4 u), we calculated the number of H2D

+ ions as a fraction of the detected HD2,

+ ions. This fraction corresponds to the final isotopical abundance of the ionic products H2D

+/HD2,+ and was obtained from the experimental

data for short delays. For this calculation, we assume that the measurement of HD2,+

still occurs within the linear range of the detector, thus providing a representative pic-ture of the isotopically substituted molecular ion. The number of D2,

+ ions is obtained as the remaining part of the total mass-4 u signal. The measured isotopical abundance dif-fers from that given by the gas supplier. There are two main effects influencing this number: First, the isotope-exchange reactions like HD2,

+ + HD ↔ D3,+ + H2 and

H2D+ + HD ↔ HD2,

+ + H2 clearly favor the deuteration of the concerned species [Ger-lich et al., 2002]. Second, the thermal velocities of isotopically substituted species are different, thus the temporal profile of the gas density in the trap is different for each isotopomer.

Calibration reaction To calibrate the emission properties of the graphite sublimation source, i.e. to obtain the concentration of the neutral carbon species released from the source, we used the reac-tions D3,

+ + Cn → CnD+ + D2. These reactions have not yet been studied experimentally.

However, it is generally assumed that they are fast, taking place approximately at the collisional rate (see e.g. [Le Teuff et al., 2000] and references therein). The primary D3,

+ ions were produced as described in the previous paragraph; the delay between the gas and ionization pulses was adjusted to 20 ms. At such conditions, the dominating pri-mary ions were D3,

+, and about 2% HD2,+. At a constant storage time of one second, we

followed the number of the reaction products CnDm,+ as a function of the concentration of carbon in the beam. Since the main number of the C3D

+ product ions is formed at lager delays after the gas pulse where the number density of the introduced deuterium already substantially decreased, the subsequent reaction C3D

+ + D2 between the product ions and the neutral deuterium background plays only a minor role. Moreover, this reac-tion is relatively slow, taking place with a rate of roughly 10-10 cm3·s-1 [Sorgenfrei, 1993]. Therefore, the main reaction product ion stemming from C3 is C3D

+.

To monitor the carbon concentration, we introduced a dedicated calibration phase into the measurement cycle. After each storage period, we reconfigure the ion optics (change the potentials on the lens elements) so that the trap is used as an ion guide and the ions produced by help of the ion source from the carbon beam are analyzed in the detection system. During this calibration phase, the mass analyzer was adjusted to measure the number of C3,

+ ions. In our previous studies [Cermak et al., 1998], we showed that the measured flux of Cn,

+ ions produced by electron-impact ionization of the beam is well proportional to the total mass flux of the carbon beam. Here, we use the flux of C3,

+ ions as a measure for the concentration of the carbon species emitted from the source.

Figure 4 shows the linear dependence of the reaction products C3D+ as a function of the

calibration measurement (flux of C3,+ ions). From the rate of the reaction

D3,+ + C3 → C3D

+ + D2 of 1.5·10-9 cm3·s-1 (calculated from the data compiled by [Le Teuff et al., 2000]), from the used storage time, and from the number of primary ions,

APPENDIX A

69

the number density of C3 for each particular measurement point was calculated. This numbers were used to scale the horizontal axis of the plot in figure 4.

Calibration / counts per second101 102 103 104 105

Ion

num

ber

per

filli

ng

100

101

102

103

C3 concentration / cm-3

104 105 106 107 108

Figure 4. Dependence of the number of C3D

+ product ions on the concentration of carbon in the neutral beam. During the calibration measurement, the flux of C3,

+ ions created by elec-tron-impact ionization was followed. The concentration of C3 molecules was derived from the number of products of the deuteron-transfer reaction D3,

+ + C3. For details see text.

Conclusion and outlook We have developed an experimental setup for studies on reactions between ions and neutral carbon molecules. The apparatus combines the trapping technique with a source of neutral carbon atoms and molecules. The function of the carbon source was substan-tially improved so that a stable operation is possible now. Preparation of primary ions is performed directly in the trapping device. The function of such a scheme is discussed on the example of D3,

+. We demonstrated the performance of the machine on a test reaction D3,

+ + C3 → C3D+ + D2.

The aim of this experimental effort is to study reactions between ions and neutral carbon species that are relevant in the interstellar and circumstellar media and many terrestrial processes. In the near future, the apparatus will be used to follow ion-molecule reactions and charge-transfer reactions between neutral carbon species and atomic and molecular ions that have not yet been experimentally studied. To this category belong reactions with ions like CO+, N2,

+, O2,+, N+, O+, etc. The long-term goal of the project is to study

reactions leading to the growth of carbon chain and ring molecules.

Some technical improvements of our setup are still necessary to enlarge the spectrum of usable primary ions and to enhance the detection power. A mass-selective ion source will be used to prepare primary ions in cases where the scheme described here for D3,

+ does not prefer a single product ion. To distinguish between different isomers of the product ions, we aim to use suitable titration reactions. We also plan to couple ion-chromatographic methods with the ion-trapping technique to detect the structure of dif-ferent isomers.

APPENDIX A

70

Acknowledgments The work was supported by the Deutsche Forschungsgemeinschaft within Project 7: Growth Dynamics of Carbon-Containing Nanostructures of the Forschergruppe Labo-ratory Astrophysics.

References Čermák, I., Förderer, M., Čermáková, I., Kalhofer, S., Stopka-Ebeler, H., Mon-ninger, G., and Krätschmer, W. ‘Laser Induced Emission Spectroscopy of Matrix-Isolated Carbon Molecules: Experimental Setup and New Results on C3’, J. Chem. Phys. 24, 108 (1998) pp. 10129-10142.

Gerlich, D., ‘Inhomogeneous RF Fields: A Versatile Tool for the Study of Processes with Slow Ions, in: State-Selected and State-to-State Ion-Molecule Reaction Dynamics’, Part 1 (ed.: Ng, C.-Y. and Baer, M.) Adv. in Chem. Phys. Series, Vol. LXXXII (1992) pp. 1-176.

Gerlich, D., Herbst, E., and Roueff, E. ‘H3+ + HD ↔ H2D

+ + H2: Low-temperature labo-ratory measurements and interstellar implications’, Plan. Space Sci., special issue: "Deuterium in the Universe" (2002), in press.

Le Teuff, Y.H., Millar, T.J., and Markwick, A.J. 'The UMIST database for astrochemis-try 1999', Astron. Astrophys. Suppl. Ser. 146 (2000) 157, http://www.rate99.co.uk/.

Luca, A., Schlemmer, S., Čermák, I., and Gerlich, D. ‘On the combination of a linear field free trap with a time-of-flight mass spectrometer’, Rev. Sci. Instrum., Vol. 72, No. 7 (2001) pp. 2900-2908.

Nummelin, A., Observations of interstellar molecules (2001), http://www.chl.chalmers.se/~numa/astrophysics/molecules/molecules.html.

Sorgenfrei, A., Ph.D. Thesis, University Freiburg (1993).

71

APPENDIX B

Submitted to International Journal of Mass Spectrometry

Reactions of Cn (n = 1 - 3) with ions stored in a temperature-variable radio-frequency trap

I. Savić, I. Čermák and D. Gerlich

Department of Physics, Technische Universität Chemnitz, 09126 Chemnitz, Germany

Abstract A new experimental setup has been developed for studying astrochemically relevant collisions between small neutral carbon molecules Cn (n = 1 - 3) and stored ions. The ions are confined for seconds in a ring electrode trap (RET) the temperature of which can be varied over a wide range (presently 80 K - 500 K). There they interact with an effusive carbon beam, which is produced via high-temperature vaporization of a carbon rod. Due to the accessible temperature range and other features of the setup, rate coeffi-cients can be measured, which are of importance for understanding the chemistry occur-ring in the outflow of stars, the formation of hydrocarbons in stellar atmospheres, and the interaction of the stored product ions with radiation. Results are reported for the interaction of stored D3

+ with hot Cn. D+ transfer dominates over all other exothermic

product channels for n = 1 - 3. The reaction rate coefficients measured for forming CnD+

is almost a factor two smaller than values presently used in astrophysical models. An-other important class of reactions concerns the growth of pure carbon chains via asso-ciative Cm

+ + Cn collisions. First results indicate that the rate coefficients are slower than generally assumed in models. Due to the weak signal, only rough limits can be reported. For future studies, the number density of carbon penetrating the trap must be increased. This and the planned extension of the temperature range is briefly discussed in the outlook.

Key words laboratory astrochemistry, low and high temperature ion-molecule collisions, small neu-tral carbon molecules, ISM molecules, proton transfer, CD+, C2D

+, C3D+, Cn

+

APPENDIX B

72

1. Introduction Within the last 30 years, a variety of chemical models has been developed for describ-ing the formation and destruction of molecules in different regions of the interstellar medium (ISM) (see for example [her89], [tur00], [pas00], [nom04]). In early simula-tions of dense or diffuse clouds, accretion disks, etc, a reaction network using mainly ion-molecule reactions was sufficient for explaining the observed molecular abun-dances. Ions certainly play an important, if not the dominant role in the formation of molecules at low temperatures since many of their reactions have no barriers [smi93] and, due to the long range ion - induced dipole attraction, the reaction rate coefficients are often high. With the growth of observational details, e.g., spatial correlation between specific molecules such as C4H and C3H2 [ger03a], the progress of quantum chemistry and reaction dynamics and the development of new experimental tools dedicated to study the interaction of molecules under conditions in space, more and more informa-tion became available. In this way, some specific questions can be answered; however, it also has been realized that the overall understanding of how atoms and molecules are processed in space, is much more complex than described by the early simple models.

Today models for describing the evolution of the ISM are very sophisticated and spe-cialized. In chemical reaction networks neutral - neutral reactions, especially radicals, are routinely included. Also the surfaces of grains play an important role in catalysis, in freezing out gas or in providing products synthesized in the ice, e.g. via photons or cos-mic rays. Due to this huge complexity, one often has to restrict the number of processes which one includes in a model describing a specific object, otherwise one is not able to numerically solve the problem. A typical example is that coupling of a complex chemi-cal network with a dynamic global 2D model of protoplanetary disk is already beyond present-day computer capabilities [hen03]. For such attempts one needs, in addition to the help from mathematics and computer science, relevant and selected data from ex-perimental and theoretical astrochemistry. In some cases simplified rates for the in-volved physical and chemical processes are sufficient in other situations detailed tem-perature dependent and state specific rate coefficients are needed. In addition the evolu-tion of the matter from atoms to molecules and grains is not only determined by the overall elemental abundance or the number densities and chemistry of the formed mole-cules but it depends in a complicated way on the environment, e.g. the radiation field. Therefore dedicated measurements performed under conditions of astrochemical rele-vance are needed.

Studies of chemical processes between ions and molecules have been studied originally in plasmas, afterglows and later under more or less well defined room temperature con-ditions in flow systems and traps. Stimulated by the need of learning more about reac-tions at the very cold conditions prevailing in dense interstellar clouds, the last two dec-ades have seen several successful initiatives to develop special instruments for extend-ing the temperature range down to 10 K and even below [smi00], [smi98], [ger93], [ger03b]. Many interesting observations have been made in reactions which are due to tunneling, differences in zero point energies in the case of isotope substitution, the en-ergy provided by ortho-hydrogen etc. Recent examples from our systematic low tem-peratures studies of the chemistry of hydrocarbons include the formation of CH5

+ in CH4

+ + H2 collisions the rate coefficient of which increase by one order of magnitude going from room temperature to 15 K [asv04] or the new efficient routes found for the formation of C3Hn

+ and their deuterated analogues in cold environments [sav04a], [sav04b].

APPENDIX B

73

Low temperature studies are now becoming routine and it is time to develop innovative instruments which are able to operate at temperatures ranging from the cold interstellar medium to the very hot environment prevailing in stellar atmospheres. Many molecules are formed in stellar outflows, where the emitted material cools from 5000 K or higher down to 50 K and where the number density goes from about 1012 to 106 particles cm-3. It is concluded [pat95], that at 1000 K and below, where the expanding gas has a density of 1010 cm-3 - 108 cm-3, effective cluster formation and growth occurs; however, there are almost no experimental studies available, especially not of the important processes of radiative association under such conditions. One central problem is for example the understanding of the chemical pathways, starting from very small molecules C2, C3, C2H or C3H towards nanostructures and finally to large grains.

Temperatures of more then 2000 K can be produced by shocks occurring in supersonic protostellar outflows [cod01] leading to mostly unknown chemical changes. As dis-cussed in [wil02] the chemistry occurring in the interaction of a stellar jet with a clump in a molecular cloud must also include the intense radiation field generated in the jet-cloud shock. The increasing need to high temperature data has also been emphasized in [teu00]; it has been started to include reaction rate coefficients into the database which are valid over a wide temperature range.

Due to a variety of reasons carbon atoms and carbon-bearing molecules play a dominant role in all fields of natural science, ranging from physics via chemistry to biology. They are found in all environments, ranging from the hot outflow of a carbon rich star to cold interstellar medium where carbon atoms and ions are important coolers. A challenge for planning dedicated experiments is the recent observational hint, that C atoms may have been present already in the early universe [har03]. Many researchers are convinced to-day that carbon containing molecules are responsible for the diffuse interstellar bands or the extended red emission; however, there are not yet any systematic experimental strategies to get some order into the complex astrochemistry producing and processing the huge variety of carbon containing ions, neutrals and radicals. A recent simulation of the growth of molecular structures in stellar atmosphere [pas00] can be used to illustrate the complexity. In this study, the interaction of thousand different species has been modeled at temperatures between 2000 K and 7000 K. The sensitivity of the model to the microphysics can be seen from the fact, that already the modification of one single parameter, e.g. the carbon to hydrogen abundance, leads to significant changes. Calcu-lated carbon atoms sputtering yields impacting on amorphous carbon target are domi-nant at velocities less than ~ 35 km/s [cov00] and therefore can be important for sputter-ing in the outflows of cool stars.

A lot of experimental and theoretical effort has been devoted to investigate pure carbon clusters, ranging from theoretical calculations of stable isomers, to experimental spec-troscopy of different size clusters [mai98], [del99], [del00]. The reason for this is not only the importance of carbon compounds in astrophysics, but also in material science and in combustion processes. Also from a fundamental point of view carbon is of inter-est due its fascinating properties and chemical diversity leading not only to the well-known classical structures of aromatic rings, graphite, fullerite, diamond etc, but also to the two electron – three center configuration [ola95]. The above mentioned CH5

+ ion is the simplest member of this group; however, its structure is not completely understood.

It was only in the last two decades, that experimental efforts have been started to inves-tigate chemical reactions involving carbon atoms or molecules. The reason is that it is experimentally difficult to obtain an intense and well defined beam of carbon atoms or molecules needed for studying reaction dynamics. The development of the intense laser

APPENDIX B

74

vaporization carbon source lead to the discovery of fullerenes for which the Nobel price has been awarded to R. F. Curl, H. W. Kroto and R. E. Smalley. Significant contribu-tions to the understanding of neutral reactions involving carbon and hydrocarbons have been made in the group of Y.T. Lee using the crossed beam technique [kai02]. Accord-ing to the best of our knowledge, there have been no experiments performed so far on reactions between ions and neutral carbon molecules. In the following a new ion trap-ping - carbon beam instrument is presented. Various details of the machine are ex-plained, since it is a development dedicated to the study of reactions of astrochemical importance over a wide range of temperatures and densities. Also specific properties of the carbon vaporizator are discussed since they are not only important for controlling the composition, the velocity etc. of the beam but it also provides some information on carbon chemistry occurring at high temperatures. Results are given for the deuteron transfer from D3

+ to carbon and for the interaction of stored carbon ions with the neutral carbon beam.

2. Experimental Many experimental techniques have been developed in the last decades for investigating inelastic and reactive interactions between ions and molecules. The present apparatus is a typical ion trapping machine aiming to study chemical reactions between ions and neutral carbon. As can be seen from a variety of applications [ger03b], [ger04], [scl99] the method to trap ions in inhomogeneous rf field is very versatile. For special purposes, a variety of modules and lasers can be integrated into the systems.

2.1. Apparatus A dedicated ion-trapping apparatus has developed for combing an intense carbon beam with a special ion trap which can be operated both at low and very high temperatures. The most important modules of the experimental setup, some technical details have been described already in [cer02], are shown in Fig. 1. They are (from the left) (i) the carbon vaporization source, (ii) an electron bombardment ion source, (iii) the 80 - 600 K radio-frequency ring electrode trap (C-RET), and (iv) a mass spectrometer detector. A few additional hints are given in the following section with emphasis on the carbon evaporator.

A very good vacuum is crucial for this type of experiments in order to suppress parasitic reactions of stored ions with impurities, which, in the present case, mainly emerge from the hot source. Therefore a total of four turbo pumps have been integrated into the sys-tem two of which (pumping speeds 2 × 230 l/s) evacuate the trapping and the detector chamber and two of which are used for differentially pumping the main and the second chamber of the carbon source (pumping speeds 170 l/s and 56 l/s, respectively). The two source chambers and the main chamber are separated from each other by two apertures (diameters 3 mm and 6 mm, respectively). The source can be closed using a UHV gate valve. In the trapping chamber the pumping system sustains a background pressure be-low 3 × 10-10 mbar. Additional cooling of the RET with liquid nitrogen leads to addi-tional reduction of background gas. At such conditions, less than 1 % of primary ions are converted through reactions with the residual gas at a storage time of 1 s. The versa-tile gas inlet system allows for introducing pure or mixed gases into the trap or the ion source, either continuously via a leak valve or pulsed using a fast piezo driven valve. A spinning rotor gauge (MKS SRG2) is connected via a stainless steel tube to the main chamber for monitoring the pressure or calibrating the ion gauge.

APPENDIX B

75

As can be seen from Fig. 1 and explained in more detail in [sav04] the ionizer which is based on a commercial construction originally developed for quadrupole mass spec-trometers (Extranuclear), is located between the carbon source and the ion trap. It has several functions such as (i) preparation of primary ions and loading them into the trap (ii) monitoring the carbon flux or (iii) injecting electrons directly into the trap. It con-sists of cylindrical electrodes with axial apertures. The open construction allows the neutral carbon beam to pass through the source into the RET. Below 10-8 mbar this source operates in a space charge trapping mode leading to very high ionization and collection efficiencies. All ions produced are mass selected in a quadrupole (rod diame-ter d = 18 mm, 260 mm long, rf frequency 1.6 MHz) and detected using an ion detector of the Daly type [dal60].

2.2. Ring electrode trap The central part of this machine is the ring electrode trap (RET) also shown in Fig. 1. It is mounted onto a flange via a light stainless steel structure which has such a good heat insulation that the trap can be easily cooled with liquid nitrogen or heated to high tem-peratures. Without a heat shield, a temperature of 600 K has been reached just from the rf power dissipated in the coil. Besides the wide temperature range, an important advan-tage of the RET design is, in contrast to the 22-pole trap, that it is rather insensitive to coating with carbon since most insulators are protected by shields. In addition the com-plete setup is easy to clean.

The geometry of the electrode arrangement was first described in [ger89] and in more detail later [ger92b], [ger92], [luc01]. The trapping region is defined by 20 electrodes alternatively connected to two metal bars. The gap between successive electrodes is 1 mm, the electrodes themselves are 1 mm thick and have a circular hole with a radius of r0 = 5 mm. The metal bars are welded to metal tubes. These tubes are used both for electrical connections and for flowing liquid N2 through the whole system including the metal bars. In addition they are wound to form a coil which is the inductance of the resonant circuit supplying the trap with rf.

For excitation, this circuit is connected to an rf generator. Using two external capacitors the operating frequency has been set to f = 7.8 MHz. Typically, the trap was operated with amplitudes up to V0 = 250 V. The dc potential of the trapping region has been de-fined by applying a voltage to the center of the coil via a low frequency band pass cir-cuit. In the radial direction, ions are confined by the effective rf potential. Adequate conditions for ion trapping can be estimated by using simple calculations [ger92]. Usu-ally one chooses the effective potential high enough for safely storing all ions in a vol-ume of radius r = 0.8 r0, Combining the equations for the effective potential and the adiabaticity parameter and using η < 0.3 and r = 0.8 r0 leads to the minimal values for the frequency and the voltage. Since in the present case the parameters have been set for carbon clusters, the light ion H3

+ was already at the limit while safe storage of D3+ and

C3D+ posed no problem.

In the axial direction the trap can be closed and opened by applying suitable voltages to the trap entrance and trap exit electrode. For closing the trap, voltages of a few 100 meV are sufficient, after thermalization of the trapped ion cloud, even much lower values can be used. For extracting the ions, a small negative pulse, typically -1 V, is used. The trap is equipped with two steering electrodes, the electric field of which penetrates into the interior. For fast ejection, e.g. for TOF-MS [luc01] suitable axial potential gradients have been applied during ion extraction. The electrodes can also be used for testing in-

APPENDIX B

76

homogeneities of the trap potential caused by surface effects and for reducing their in-fluence.

2.3. The carbon source The astrochemical importance of reactions involving carbon atoms, molecules, and clus-ters has been emphasized in the introduction. As a consequence, several attempts have been made in recent years to construct suitable carbon sources, e.g. based on laser abla-tion of carbon rods, sublimation of graphite, or evaporation of suitable carbon contain-ing compounds. The source used in this work is based on vaporizing graphite rods in high vacuum by electric resistive heating with a current of several hundred Amperes. The construction used was developed in Krätschmer´s group in Heidelberg for spectro-scopic studies on matrix-isolated carbon molecules [cer98]. Only minor changes have been made on the basic design while the control of the source parameters has been im-proved leading to a better longtime stability [cer02], [sav04].

Two graphite electrodes (cathode 4.6 mm and anode 6.15 mm diameters, RW4 type, impurities < 2 ppm, SGL Carbon) are pressed together and resistively heated using in a serial-parallel way four switching power supplies (Philips PE 1980, 5 V, 200 A each). Note, that the polarity does not play a remarkable role. For cleaning, a new rod must be baked first for 5 - 7 hours with 600 - 700 W. Vaporization starts at about 1 kW. For sta-ble emission a power up to 1.3 kW can be used. In order to reduce excessive heating of the vacuum chamber, a cooper radiation shield and water cooling is used.

A sensitive characteristic of the operating conditions of the source is the rod resistance which is monitored continuously via a four-point scheme. For the electrodes used, the resistance drops with increasing temperature and it reaches finally a mean value of about 20 mΩ. A control program stabilizes the heating power when the rods resistance changes. Via a stepping motor the carbon rods are pushed against each other for main-taining a constant pressure which is estimated to be in the order of 10 N/cm2.

A pyrometer has been used for monitoring the temperature of the emitting area. How-ever, detailed tests have shown that the emitted carbon flow does not sublimate from an outside surface area of the rods but the material is ejected in a pulsed mode from the contact area between the two rods pushed against each other [sav04]. Whether these eruptions which occur a few times per second, are due to hot carbon gas or liquid bub-bles enclosed in the interface is still uncertain. Another problem which is correlated with this observation, is the condensation of carbonaceous material in the vicinity of the contact area, reducing the efficiency of the source by 75 %.

A simple way to determine the total amount of evaporated material is to measure the mass loss which occurs exclusively at the expenses of the 4.6 mm rod. Typically, 7.5 cm of the rod are vaporized in 8 h, corresponding to 2.6 µm/s. Only 25 % of this vaporized material, corresponding to 1.9 × 10-5 g/s, actually goes into the gas phase, the rest condenses in macroscopic structures and finally falls down. Assuming for simplic-ity that C3 is isotropically emitted with a temperature of 3000 K, one obtains in the trap which is 40 cm away from the rod a number density of 1 × 108 cm-3. The actual number density measured at the location of the trap by electron bombardment is in accordance with this value. Also the carbon cluster distribution has been determined in this way as well by chemical probing to be 80 ± 8 % C3, 8 ± 2 % C2, and 12 ± 6 % C. For monitor-ing the absolute carbon flux on a regular basis, the ionizer (see 2.1) has been calibrated in situ using nitrogen and the spinning rotor gauge.

APPENDIX B

77

2.4. Measuring procedure The primary ions are produced by electron impact (see 2.1) directly in the trap and stored for a given storage time. They are continuously exposed to the carbon beam. Af-ter a period which is varied between ms and min, the content of the trap is analyzed by extraction, mass selection and counting. Directly after this, the flux of the carbon beam is monitored to correct for drifts of the carbon source. In the present experiment meas-urements have been performed in two different ways.

For better understanding the operation of the carbon source, the trapping time was held fixed and the number density of the neutrals was changed by varying the heating power. The relative changes in temperature were less than 350 K and are negligible in compari-son with the uncertainty in the actual temperature of the carbon gas. As can be seen from Fig. 2, measurements have been performed at carbon number densities starting at values as low as 106 cm-3 and going up to 2 × 108 cm-3.

In the second group of experiments, the trapping time was changed while the number density of C3 was stabilized at a constant value. For determining the changes of the ion composition due to reactions, the sequence (i) ion formation, (ii) injection, (iii) relaxa-tion and reaction, (iv) analysis and (v) carbon flux determination is repeated many times for each product mass and typically for five to ten different storage times. More details of typical trapping measuring procedures and applications can be found in recent publi-cations [scl99], [ger02], [smi02].

3. Results and Discussion

3.1 Reactions of D3+ with Cn

Selected experimental results for forming CnDm+ in collisions between D3

+ and Cn are presented in Fig. 2 and Fig. 3. The D3

+ primary ions have been prepared by introducing a short gas pulse of D2 into the RET. With a certain delay a pulse of electrons with an energy of 70 eV is injected into the trap using the filament of the ionizer. The primary D+ and D2

+ ions produced by electron bombardment collide with the ambient D2 gas. D2

+ ions are almost completely converted to D3+. By changing the delay between the gas

and the electron pulse, the composition of the ions in the trap can be controlled. Some isotopic variants of D2

+ and D3+ are present in small amounts due to the isotopic compo-

sition of the introduced gas. In the present experiment, HD2+ and D2

+ / H2D+ could be

suppressed to 2 % and 0.1 % of D3+, respectively.

Fig. 2 shows the dependence of D3+, C3D

+ and C3D2+ ions as a function of the number

density of C3 after 1 s storage time. The most abundant product is C3D+ formed in the

reaction

D3+ + C3 → C3D

+ + D2 . (1)

It can be seen from the figure that more C3D+ and C3D2

+ products are formed at higher number densities; however there are some nonlinearities in the dependence. Especially the loss of D3

+ at high number densities is larger than the gain in C3D+ and C3D2

+. This is only to a minor fraction due to the fact that not all product ions are presented in this figure. The main reason is that at large fluxes, the trap or ion optics become coated non-uniformly with a carbon layer influencing the work functions and in this way the trap-ping or the transmission properties. A clear indication is that this process is not directly reversible but has a rather long time constant. Additional water cooled baffles and aper-tures limiting the beam will reduce this problem.

APPENDIX B

78

In the present evaluation of the data an empirical density dependent term for ion loss has been introduced into the standard kinetic model. Inspection of Fig. 2 reveals that the resulting fit (solid line) follows nicely the experimental data points. The main result from this evaluation procedure is the rate coefficient for reaction (1) which is given in Table 1. It is gratifying that the line following the 13C12C2D

+ data has been obtained without any additional assumption just by using this rate coefficient and taking in ac-count the natural abundance of 13C12C2 of 3.3 %. Formation of traces of C3D2

+ have also been detected; however, the data have not been fitted because the signal is too low.

A typical result from an experiment where the trapping time was changed is presented in Fig. 3. The number density of C3 was fixed at a mean value of 2 × 107 cm-3, fluctua-tions have been monitored and accounted for. The sum of all experimental data meas-ured at different storage times has been used for normalization. Also here the data have been fitted (solid lines) using the solutions of the appropriate differential equations. The resulting reaction rate coefficient, k1 = 7.5 × 10-10 cm3s-1 is in accordance with the final value given in Table 1. The dotted lines in Fig. 3 represent simulations for the lowest and the highest value of k1.

In the same way the dependence of ions with two and one carbon atom has been fol-lowed as a function of the carbon density. It must be mentioned that in these experi-ments, the number density of C2 and C has not been measured separately but derived from the measured C3 number density using the measured C2/C3 and C/C3 ratios (see 2.3). For example C2D

+ ions are most probably formed in the reaction

D3+ + C2 → C2D

+ + D2 . (2)

Also a minor part (< 20 %) of C2D2+ has been detected, which is either formed directly

via

D3+ + C2 → C2D2

+ + D (3)

or which has been produced in the secondary reaction C2D+ + D2 → C2D2

+ + D with the traces of D2 remaining in trap. In Table 1 the resulting rate coefficients for reaction (2) and (3) are presented. In the analysis, C2D3

+ ions could not be accounted for because they have the same mass as the COD+ ions. In addition it is also assumed that radiative association reactions of D3

+ with Cn are negligible at the high temperatures of the car-bon molecules.

The CD+, 13CD+, CD2+ and CD3

+ product ions are most likely formed via

D3+ + C → CD+ + D2 (4)

D3+ + C → CD2

+ + D (5)

D3+ + C → CD3

+ (6)

It is also possible that CDn+ (n = 2 and 3) ions are produced in the subsequent reactions

CDn-1+ + D2 → CDn

+ + D. All resulting rate coefficients can be found in Table 1.

For determining the overall accuracy of the rate coefficients, one has to account for the uncertainties of the fitting process (20 %), the statistical errors, and especially the uncer-tainties in determining in the trap the number density of the carbon beam and its over-lapp with the ions cloud. Additional uncertainties arise from the fact that with increasing temperature of the carbon rods and the source chamber the background gas increases. In addition the VUV radiation emitted from the hot surfaces can lead to ionization. A mi-nor problem is the formation of COD+ and CO2D

+ ions in collisions of the stored D3+

APPENDIX B

79

ions with CO and CO2 background gas. From various tests it has been concluded that these molecules are not evaporated directly from the carbon rods; however they origi-nate from outgassing the carbon coated surfaces of the beam chamber. Weighting the various influences it is estimated that the resulting rate coefficients haven an overall uncertainty of 50 %.

The measured data are compared in Table 1 with values derived from the UMIST data for Cn + H3

+ → CnH+ + H2 [teu00]. The values of 2 × 10-9 cm3s-1, 1.8 × 10-9 cm3s-1 and

2 × 10-9 cm3s-1 for n = 1, 2 and 3, respectively have been converted just by accounting for the reduced mass. It is noted that these values may have an error of up to factor 2. It is assumed that the values are independent on temperature, and therefore they are used at 10 K in dense interstellar clouds as well as in stellar atmospheres. The Langevin model predicts larger rate coefficients.

Stimulated by the experiments, also theoretical work has been started [sco03] including the calculation of the potential energy surface and molecular dynamics simulations of the H3

+ + C3 collision [fis04]. Formation of the most stable collision complex, c-C3H3+

has been determined to be 907 kJ mol-1 below the energy of the reactants. Both the pro-ton transfer C3 + H3

+ → l-C3H+ + H2 as well as the reaction C3 + H3

+ → C3H2+ + H is by

330 kJ mol-1 exothermic. More details on the energetics and the accuracy of the calcula-tions can be found in [fis04]. In accordance with our experimental results the trajectory calculations show that proton transfer is by far the dominant reaction channel in H3

+ + Cn collisions. The thermal reaction coefficients derived from the calculations are given in Table 2. The results are well below the Langevin value. Two reaction mecha-nisms have been identified, direct protonation Cn + H3

+ → CnH+ + H2 and passage

through an excited intermediate Cn + H3+ → (CnH3

+)* → CnH+ + H2. These findings are

in accordance with earlier ab initio quantum mechanical calculations performed for the collision system C + H3

+ [tal91] showing that proton transfer prevails while there is no pathway without a significant activation energy to lead to the CH2

+ + H product chan-nel. Also for the C2 + H3

+ it has been found that the main reaction channel is formation of C2H2

+ instead of C2H+ [sco03].

3.2 Reactions of Cm+ with Cn

One of the major aims of the newly constructed machine was and is to study the growth of pure carbon clusters in the trap over a wide range of temperatures. One of the basic questions is to find out the conditions under which a strongly bound collision complex just "sticks" together. Another interesting problem is the formation of isomers or transi-tions between them, e.g. from linear to ring structures. Unfortunately the number den-sity of the carbon target is not yet high enough or the sensitivity of the apparatus is not yet sufficient to answer such questions in detail. Nonetheless first results for the forma-tion of C4

+ and C5+ have been obtained.

In this experiment the neutral carbon beam was first used as the precursor for the ions. Ionized with 70 eV electrons, the ions are accumulated for typically 200 ms in the trap. In addition the stored ion cloud consisting of the C+, C2

+ and C3+ is exposed continu-

ously to the flux of carbon. Storage times have been varied between 0.5 s and 5 s. Since there are several loss mechanisms competing with the formation of reactive products, the largest number of products has been observed at a storage time of 3 s. Typically only a very few counts per filling are obtained and it takes averaging over many itera-tions until useful information on the rate coefficients can be extracted. In addition the carbon beam flux is measured periodically in order to account for fluctuations. System-atic studies of the loss of primary or product ions are impossible under these conditions.

APPENDIX B

80

Nonetheless it is rather safe to assume that parasitic reactions or loss from the trap af-fects all Cn

+ ions from n = 2 - 6 in the same way at the conditions the trap is operated.

In order to evaluate data from this experiment, since three different ionic reactants are allowed to react with three different neutral ones, certain simplification have been made. Knowing that the C3 is by far the dominant constituent of carbon vapor, it has been as-sumed that the three following reactions are dominant: C+ + C3 → C4

+, C2+ + C3 → C5

+ and C3

+ + C3 → C6+. In the further evaluation a simple linear approximation has been

used, k = N2 /(N1 [C3] t). N1 is the number of primary ions, N2 the number of products, [C3] the number density, and t is the trapping time. It is obvious that other reactions also can contribute to the formation of C4

+, C5+ and C6

+ and therefore only estimated upper limits of rate coefficients are presented in Table 3. For the association reaction C3

+ + C3 → C6+ it has been estimated that the rate coefficient is smaller than

1 × 10-11 cm3s-1.

The results can be compared only with rather crude model calculations [fre82]. In this simple theory the rate coefficients for C+ + Cn → Cn+1

+ + hν have been parameterized simply by k = f × kL where kL is the Langevine rate. The fraction f = 0 for n = 1 is in accordance with the experience, that radiative association is very unlikely in atom - atom collisions. The values for n = 2 and 3 are f = 10-7 and f = 10-3, respectively. For n > 3 arguments have been given, that the collision complex, formed with the Langevin rate, is stabilized with unit efficiency. This is based on the assumptions that in the C4

+

collision complex, the total number of degrees of freedom is already so large, that it lives longer than the time needed for radiative stabilization. The present C3

+ + C3 results are in contradiction to this, although one has to account also that the C3 molecules are rather hot in the experiment.

5. Conclusions and Outlook Reactions between stored D3

+ and Cn+ ions and carbon atoms and clusters have been

studied at an effective temperature which is determined by the ions stored at 80 K or 300 K and by the fast and hot beam of Cn molecules, with an estimated translational and internal temperature of 3000 K. The new experimental setup which is still under devel-opment aims at studying high temperature reactions, which are important for under-standing the chemistry in stellar atmospheres, stellar outflows, shocks etc. The instru-ment is equipped with a trap the temperature of which can be varied over a wide range. With minor changes such as adding a heat shield, trapping temperatures of 1000 K are no problem. In addition, higher ion velocities are accessible by operating the trapping machine as a Guided Ion Beam arrangement and accelerating the ions to the desired kinetic energy. A CO2 laser is already being tested for heating the stored ions in order to reach higher internal temperatures.

Additional work is needed to prepare a better beam of carbon atoms or clusters, to achieve higher number densities in the trap, and to avoid partial coating of the trap and the electrodes. On one side it is planned to use a laser ablation source similar to the one described in [kai95] in which the C / C2 / C3 ratio and the temperature can be modified by changing the laser characteristics and the expansion conditions. On the other side it is also intended to continue to utilize the present source, although carbon emission from the heated graphite rods is not yet fully understood. It is sure that carbon is not vapor-ized by sublimation from the surface but ejected from inner regions of the contact area where much higher temperatures prevail. As soon as this process is understood in more detail the source can be used more efficiently, especially for measuring high tempera-

APPENDIX B

81

ture rate coefficients. In the process of further testing the source, it is also planned to use the carbon evaporator as a flow system and to use lasers for probing (LIF, REMPI) or modifying (heating with Nd-YAG or CO2 laser pulses) the concentrations very close to the rods. It can be foreseen that such tests will provide additional information on high temperature carbon chemistry.

Acknowledgments The authors thank Stephan Schlemmer for many stimulating discussions and comments as well as Silvio Decker for help in characterization of the carbon source. Financial support by the Deutsche Forschungsgemeinschaft (DFG) is gratefully acknowledged, especially via the Forschergruppe FOR 388 "Laboratory Astrophysics".

APPENDIX B

82

References [asv04] Asvany, O., Savić, I., Schlemmer, S. and Gerlich, D.: Variable Temperature Ion Trap Studies of

CH4+ + H2, HD and D2: Negative Temperature Dependence and Significant Isotope Effect, Chem.

Phys. 298 (2004) 97-105.

[cer02] Čermák, I., Savić, I., and Gerlich, D. Ion-Trapping Apparatus for Studies on Reactions between Ions and Neutral Carbon Species, in: Safránková (ed): Proceedings of Contributed Papers, Part II, WDS'02, Matfyzpress (2002) 281-287.

[cer98] Čermák, I., Förderer, M., Čermákova, I., Kalhofer, S., Stopka-Ebeler, H., Monninger, G. and Krätschmer, W.: Laser Induced Emission Spectroscopy of Matrix-Isolated Carbon Molecules: Ex-perimental Setup and New Results on C3, J. Chem. Phys. 108 (1998) 10129-10142.

[cod01] Codella, C., Bachiller, R., Benedettini, M., Caselli, P., Giannini, T., Nisini, B., Saraceno, P., Spinoglio, L. and Walmsley, C. M.: Waiting for FIRST: The Evolution of Molecular Outflows, in: Pilbratt, G. L., Cernicharo, J., Heras, A. M., Prusti, T. and Harris, R. (eds.): Proc. Symposium The Promise of the Herschel Space Observatory, ESA SP-460, (2001) 381-384.

[cov00] Covatto, C. and Aannestad, P. A.: Sputtering in the Outflows of Cool Stars, Mon. Not. R. Astron. Soc. 318 (2000) 67-72.

[dal60] Daly, N. R.: Scintillation Type Mass Spectrometer Ion Detector, Rev. Sci. Instrum. 31 (1960) 264-267.

[del00] Deleuze, M. S., Giuffreda, M. G., François, J.-P. and Cederbaum, L. S.: Valence One-Electron and Shake-Up Ionization Bands of Carbon Clusters. II. The Cn (n = 4, 6, 8, 10) Rings, J. Chem. Phys. 112 (2000), 5325-5338.

[del99] Deleuze, M. S., Giuffreda, M. G., François, J.-P. and Cederbaum, L. S.: Valence One-Electron and Shake-Up Ionization Bands of Carbon Clusters. I. The Cn (n = 3, 5, 7, 9) Chains, J. Chem. Phys. 111 (1999), 5851-5865.

[fis04] Fischer, G., Barthel, R. and Seifert, G.: Molecular Dynamics Study of the Reaction C3 + H3+, sub-

mitted to EPJ.

[fre82] Freed, K. F., Oka, T. and Suzuki, H.: On the n-Dependence of the Reaction Rate for C+ + Cn → Cn+1

+ in Interstellar Space, ApJ. 263 (1982) 718-722.

[ger02] Gerlich, D., and Schlemmer, S.: Deuterium Fractionation in Gas-Phase Reactions Measured in the Laboratory, Planet. Space Sci. 50 (2002) 1287-1297.

[ger03a] Gerin, M., Fosse, D. and Roueff, E.: Carbon Chemistry in Interstellar Clouds, in: Curry, C. L. and Fich, M. (eds): Proceedings of the Conference Chemistry as a Diagnostic of Star Formation, University of Waterloo, August 2002, NRC Research Press (2003); http://arxiv.org/PS_cache/astro-ph/pdf/0212/0212058.pdf.

[ger03b] Gerlich, D.: Molecular Ions and Nanoparticles in RF and AC traps, Hyperfine Interactions 146/147 (2003) 293-306.

[ger04] Gerlich, D.: Applications of rf Fields and Collision Dynamics in Atomic Mass Spectrometry, J. Anal. At. Spectrom., 19 (2004) 581 – 590.

[ger89] Gerlich, D. and Kaefer, G.: Ion Trap Studies of Association Processes in Collisions of CH3+ and

CD3+ with n-H2, p-H2, D2 and He at 80 K, ApJ 347 (1989) 849-854.

[ger92b] Gerlich, D. and Horning, S.: Experimental Investigations of Radiative Association Processes as Related to Interstellar Chemistry, Chem. Rev. 92 (1992) 1509-1539.

[ger92] Gerlich, D.: Inhomogeneous rf fields: a Versatile Tool for the Study of Processes with Slow Ions, in C.–Y. Ng and M. Baer (eds.): State-Selected and State-to-State Ion-Molecule Reaction Dynam-ics, Part 1: Experiment, Adv. in Chem. Phys. LXXXII, J. Wiley & Sons (1992), 1-176.

[ger93] Gerlich, D.: Experimental Investigations of Ion-Molecule Reactions Relevant to Interstellar Chemistry, J. Chem. Soc. Faraday Trans. 89 (1993) 2199-2208.

[har03] Harwit, M., Spaans, M.: Chemical Composition of the Early Universe, ApJ 589 (2003) 53-57.

APPENDIX B

83

[hen03] Henning, T.: Astrophysical Modeling – The Chemical Evolution of Protoplanetary Disks, in: Gerlich, D. and Henning, T. (eds.): Structure, Dynamics and Properties of Molecules and Grains in Space, FOR 388, Report 2000-2003 (2003) 79-85.

[her89] Herbst, E. and Leung, C. M.: Gas-Phase Production of Complex Hydrocarbons, Cyanopolyynes, and Related Compounds in Dense Interstellar Clouds, ApJS 69 (1989) 271-300.

[kai95] Kaiser, R.I., Suits, A. G.: A High-Intensity, Pulsed Supersonic Carbon Source with C(3Pj) Ki-netic Energies of 0.08 - 0.7 eV for Crossed Beam Experiments, Rev. Sci. Instrum. 66 (1995), 5405 - 5411

[kai02] Kaiser, R. I.: Habilitationsschrift Chemnitz, 2002; Experimental Investigation on the Formation of Carbon-Bearing Molecules in the Interstellar Medium via Neutral-Neutral Reactions, Chem. Rev. 102 (2002) 1309-1358.

[luc01] Luca, A., Schlemmer, S., Čermák, I. and Gerlich, D.: On the Combination of a Linear Field Free Trap with a Time-of-Flight Mass Spectrometer, Rev. Sci. Instrum. 72 (2001), 2900-2908.

[mai98] Maier, J. P.: Electronic Spectroscopy of Carbon Chains, J. Phys. Chem. A 102 (1998) 3462-3469.

[nom04] Nomura, H. and Millar, T. J.: The Physical and Chemical Structure of Hot Molecular Cores, A&A 414 (2004) 409-423.

[ola95] Olah, G. A.: My Search for Carbocations and their Role in Chemistry (Nobel Lecture), Angew. Chem. Int. Ed. Engl. 34 (1995) 1393-1405.

[pas00] Pascoli, G. and Polleux, A.: Condensation and Growth of Hydrogenated Carbon Clusters in Carbon-Rich Stars, A&A 359 (2000) 799-810.

[pat95] Patzer, A. B. C., Köhler, T. M., and Sedlmayr, E.: Primary Astrophysical Dust Formation: Laboratory Desiderata, Planet. Space Sci. 43 (1995) 1233-1239.

[sav04] Savić, I.: Formation of Small Hydrocarbon Ions Under Inter- and Circumstellar Conditions: Experiments in Ion Traps, PhD Thesis.

[sav04a] Savić, I. and Gerlich, D.: Temperature Variable Ion Trap Studies of C3+ and C3H

+ + H2 and HD, submitted to PCCP (2004).

[sav04b] Savić, I., Schlemmer, S., and Gerlich, D.: Low-Temperature Experiments on the Formation of Deuterated C3H3

+, submitted to ApJ (2004).

[sco03] Scholz, R., Schreiber, M. and Seifert, G.: Molecular Dynamics Simulation of Structure, Forma-tion and Reactivity of Molecules and Clusters, in: Gerlich, D. and Henning, T. (eds.): Structure, Dynamics and Properties of Molecules and Grains in Space, FOR 388, Report 2000-2003 (2003) 53-68.

[scl99] Schlemmer, S., Kuhn, T., Lescop, E. and Gerlich, D.: Laser excited N2+ in a 22-Pole Trap, Ex-

perimental Studies of Rotational Relaxation processes, Int. J. Mass Spectrom. 185 (1999) 589-602.

[smi00] Smith, I. W. M., and Rowe, B. R.: Reaction Kinetics at Very Low Temperatures: Laboratory Studies and Interstellar Chemistry, Acc. Chem. Res. 33 (2000) 261-268.

[smi02] Smith, M. A., Schlemmer, S., von Richthofen, J. and Gerlich, D.: HOC+ + H2 Isomerization Rate at 25 K: Implications for the Observed [HCO+]/[HOC+] Ratios in the Interstellar Medium, ApJ 578 (2002) L87-L90.

[smi93] Smith, M.A.: Ion Molecule reactions Dynamics at Very Low Temperatures, in: Ng, C.Y., Baer, T. and Powis I. (eds.): Current Topics in Ion Chemistry and Physics, Vol. 2, Wiley, New York (1993), 183-251.

[smi98] Smith, M. A.: Low Temperature Rate Studies of Ions and Radicals in Supersonic Flows, Interna-tional Reviews in Phys. Chem. 17 (1998), 35-63.

[tal91] Talbi, D., DeFrees, D. J., Egolf, D. A. and Herbst, E.: Calculations Concerning the Reaction C + H3

+ → CH+ + H2, ApJ 374 (1991) 390-393.

[teu00] Le Teuff, Y. H., Millar, T. J. and Markwick, A. J.: The UMIST Databas for Astrochemistry 1999, A&AS 146 (2000) 157-168, and http://www.rate99.co.uk

APPENDIX B

84

[tur00] Turner, B. E., Herbst, E. and Terzieva, R.: The Physics and Chemistry of Small Translucent Mo-lecular Clouds. XIII. The Basic Hydrocarbon Chemistry, ApJS 126 (2000) 427-460.

[wil02] Williams, D. A., Lim, A. J. and Viti S.: The Chemistry of Stellar Jets, RevMexAA (Serie de Conferencias) 13 (2002) 114-118.

Tables Table 1. Measured reaction rate coefficients for D3

+ + Cn reactions. The UMIST data have been given for H3

+ and are converted by accounting for the reduced mass.

k / 10-10 cm3 s−1 Reaction

This work UMIST

D3+ + C3 → C3D

+ + D2

7.3 14.7

D3

+ + C2 → C2D+ + D

2 5.7 13.4

D3+ + C2 → C2D2

+ + D

1.0

D3

+ + C → CD+ + D2 4.7 15.5 D3

+ + C → CD2+ + D 2.2

D3+ + C → CD3

+ 0.4

Table 2. Thermal reaction rate coefficients for reactions H3+ + C3 → C3H

+ + H2 and D3

+ + C3 → CnD+ + D2 [fis04], [sco03].

k / 10-10 cm3 s−1

T / K C3 + H3+ → C3H

+ + H2 C3 + D3+ → CnD

+ + D2 100 5.1 4.1 200 6.7 300 7.7 5.9 400 8.5 500 9.1 7.1 1000 11.0 8.6

Table 3. Measured upper limits of rate coefficients for Cn+ + C3 (n = 1 – 3) association

reactions.

Reaction k

C+ + C3 → C4+ < 4 × 10-11 cm3s-1

C2+ + C3 → C5

+ < 4 × 10-10 cm3s-1 C3

+ + C3 → C6+ < 1 × 10-11 cm3s-1

APPENDIX B

85

Figures

Fig. 1. Schematic diagram of the combination of the carbon evaporator (left) and the ring electrode trap. The distance (40 cm) is not to scale. In the trapping volume C3 number densities up to 2 × 108 cm-3 have been reached. The electron bombardment ion source (in the center of the figure) is used both for creating reactant ions and for analyzing the composition and density of the effusive carbon beam. For analysis, the trapped ions are extracted (to the right), mass selected with a QPMS and detected with a Daly type detec-tor.

104 105 106 107 108 10910-1

100

101

102

103

104

D3

+ + C3 →

nC

3

/ cm-3

Ni

D3

+

C3D+

13C12C2D+

C3D

2

+

Fig. 2. Number of ions per filling, Ni, as a function of the number density of C3. The primarily injected D3

+ ions have been exposed to the carbon flux during a storage time of 1 s. The data obtained for D3+,

C3D+ and C3D2

+, are fitted (solid lines) using a model described in the text. For the results shown in this plot have k1 = 6.3 × 10-10 cm3s-1 has been obtained for C3D

+ production. Using the same reaction rate coef-ficient and accounting for the natural abundance of 13C, the fit for 13C12C2D

+ has been obtained. As can be seen from the open triangels, only traces of C3D2

+ are produced.

APPENDIX B

86

0.0 0.2 0.4 0.6 0.8 1.010-1

100

101

102

103

104

D3

+ + C3 →

t / s

Ni

D3

+

C3D+

C3D

2

+

Fig. 3. Number of ions per filling, Ni, as a function of storage time t. The C3 number density has been stabilized at 2 × 107 cm-3. The data obtained for D3

+, C3D+ and C3D2

+ ions are normalized for number density and total sum of ions. For details see the text.

87

APPENDIX C

Submitted to The Astrophysical Journal

Low-temperature experiments on the formation of deuterated C3H3+

I. Savić, S. Schlemmer, and D. Gerlich


Abstract Many deuterated molecules have been discovered in inter- and circumstellar regions. In some cases, the observed abundances can be explained with simple thermodynamic models; often, however, isotope enrichment is more complicated. This has been seen recently in detailed low temperature experiments performed for the fundamental sys-tems H3

+ / H2D+ and CHn

+ / CHn-1D+. An unsolved problem is to explain the large abun-

dance of C3H2 and larger hydrocarbons and their deuterated variants observed in cold, dark interstellar clouds. In this work a variable temperature 22-pole trap is utilized for closely scrutinizing various ion-molecule reactions which may contribute to the forma-tion of C3H2D

+ or C3HD2+ and, via dissociative recombination, to C3HD. The experi-

mental study of the promising candidate C3H3+ + HD → C3H2D

+ + H2 which already has been excluded by theory, corroborates that this exothermic H - D exchange does not occur at all. A careful analysis of the data reveals that the 15 K rate coefficient is smaller than 4 × 10-16 cm3s-1. In contrary, quite efficient routes have been found in the low temperature experiments, starting with C3

+ and proceeding via deuterated C3H+ and

C3H2+ to C3H3

+. Formation of C3D+ in C3

+ + HD collisions is six times faster than as-sumed in astrochemical models (k = 9.3 × 10-10 cm3s-1). Surprisingly, also direct produc-tion of C3HD+ via radiative association has been observed (kr = 6.0 × 10-11 cm3s-1). The partly or fully deuterated C3H

+ + H2 collision system is strongly dependent on tempera-ture and on the hydrogen ortho to para ratio. In addition it shows very complicated iso-tope effects. For example, in C3H

+ + HD collisions, formation of C3HD+ (k = 4.6 × 10-10 cm3s-1) dominates over H - D exchange (k = 5.6 × 10-11 cm3s-1) and ra-diative association C3H2D

+ (kr = 3.2 × 10-11 cm3s-1). The reactions involving C3H2+ are

all very slow. It is recommended that the new values are included into astrochemical databases since they are very fundamental for describing correctly the carbon chemistry in interstellar clouds.

Key words laboratory astrochemistry, low temperature collisions, deuteration, ISM: molecules, C3H2, C3HD, C3H2D

+

APPENDIX C

88

1. Introduction

Isotope fractionation in general Theoretical and experimental understanding of isotope fractionation is of central impor-tance for modeling deuterium enrichment in molecules occurring in cold regions of in-terstellar space. Many singly and several doubly deuterated isotopomers have been de-tected in low-temperature interstellar clouds and meanwhile even fully deuterated am-monia has been found. There are molecules, where the DX/HX abundance ratio is a factor 104 larger than the D/H elemental ratio which is typically 2 × 10-5 in our galaxy [mil02], [loi01], [rou00]. Important in producing deuterium-containing species are exo-thermic exchange reactions between ions and HD or D. In addition it is believed that D2 molecules play a non-negligible role in the deuteration chemistry of cold, dark molecu-lar clouds.

In general the well-known difference in zero-point vibrational energy of the involved molecules determines the exothermicity of an H - D exchange reaction. Until recently it was generally accepted that, in the absence of an activation barrier, the reactions pro-ceed much faster in the exothermic direction than in the opposite one and that the ther-modynamic DX/HX ratio can be reached after an adequate time. However, dedicated low temperature experiments have shown [ger02a], [asv04a], [asv04b] that there can be significant deviations from such simple models. This may be due to subtle details of the potential energy surfaces such as small barriers or bottle necks which are not yet known with the accuracy required for low temperature reactions. In addition it is rather sure - although not yet understood quantitatively with exception of simple systems [ger90] - that symmetry selection rules play a pivotal role in replacing in a group of identical at-oms just one by an isotope or vice versa. The conservation of the total nuclear spin can have very restrictive consequences [ger04]. In this context it must be mentioned that traces of o-H2 are of exceptional importance in deuteration since they do not contribute only rotational energy but also the total nuclear spin I = 1 [ger02b].

C3H2 abundance and C3HD / C3H2 ratio A variety of gas-phase and grain-surface mechanisms processing hydrocarbon mole-cules and their deuterated analogues have been included in models of interstellar clouds ([mil89], [rob00a], [rob00b], [mil02]). For many species including isotopomers quanti-tative predictions have been made; however, there are several unsolved problems. A rather basic example is the large observed abundance of C3H2 and C4H in photo domi-nated regions. In a discussion of recent progress of interstellar carbon chemistry, Gerin et al. [gri03] state that the presently known carbon inventory is not complete. The ob-served spatial correlations of various hydrocarbons are very critical tests of our under-standing of the carbon chemistry. From the fact that in clouds, in contrast to models, more hydrogenated carbon species are present than pure carbon molecules, it is con-cluded that important formation routes are ignored or that the used rate coefficients are inaccurate.

Another example is the large deuterium fractionation of c-C3H2 in TMC-1. Values measured for the C3HD / C3H2 ratio vary from 0.08 to 0.16 depending on the position [bel88]; Turner reports 0.048 [tur01]. Theoretical predictions are below 10-2 [mil89], [rob00a]. In order to increase the C3HD / C3H2 abundance ratio resulting from the mod-els it was suggested by Howe and Millar [how93] that the reaction

C3H3+ + HD → C3H2D

+ + H2 (1)

APPENDIX C

89

occurs rapidly at low temperatures and that the underrepresented C3HD is formed via dissociative recombination of C3H2D

+. This suggestion has been excluded recently by Talbi and Herbst [tal01] using quantum chemical calculations. No low energy path way for reaction (1) could be found. As a consequence of the high transition barrier, the rate coefficient for C3H2D

+ formation is negligible at room temperature and below. In this contribution, this conclusion is corroborated by low temperature trapping experiments.

New routes in forming hydrocarbons and deuterated variants Since deuteration via reaction (1) is completely negligible there must be other ways leading to deuterated C3H2 in the observed abundances. In this context, one has to ask the general question whether important pathways for producing hydrocarbons are un-known or ignored in general. Besides the gas phase ion chemistry, the subject of this paper, there may be rather efficient neutral reactions involving radicals [kai02]. In addi-tion there is plenty of room for speculations in the chemistry occuring in and on grain ice mantles; there is observational evidence that deuterated carbonaceous species are released from the solid phase into the gas phase [mar01], [mar02]. In order to explore in more detail astrochemically important reactions involving carbon atoms and molecules, a specific research program has been started in our laboratory based on ion trapping and atomic and molecular beams [cer02], [bor04]. In addition to reactions with neutral car-bons [sav04b], new experiments with C3

+, C3H+, and C3H3

+ interacting with hydrogen molecules have been performed [sav04a]. In this contribution selected experimental data are presented and discussed which are important for producing deuterium enriched hydrocarbons at low temperatures.

The exothermic reaction C3+ + H2 → C3H

+ + H is included into the UMIST database with a temperature independent rate coefficient of only k = 3 × 10-10 cm3s-1. Experi-ments reveal that this basic process is strongly temperature dependent and the reaction becomes fast below 50 K (k = 1.7 × 10-9 cm3s-1) [sav04a]. Also the rate coefficients for deuteration via

C3+ + HD → C3D

+ + H (2a) → C3H

+ + D (2b) → C3HD+ + hν (2c)

which are reported in this contribution are fast at low temperatures.

The collision system C3H+ + H2 is of central relevance for carbon chemistry, e.g. for

synthesizing c-C3H2 [ada87]. The question which products are directly formed has a long history. Early measurements which have been performed in a SIFT (selected ion flow tube) at 550 K, 300 K and 80 K, have been discussed in [her83] and reevaluated in [her84]. Difficulties in interpreting the results originated from the fact that, in the high pressure flow tube, most collisions lead to C3H3

+ via saturated three-body association reaction and, therefore, hydrogen abstraction is almost completely suppressed. This hy-pothesis has been proven experimentally in an 80 K ion trap experiment [ger92] which provided clear evidence that at low densities the C3H2

+ product prevails. The fact that in the flow tube less C3H2

+ was formed at lower temperatures has lead to the erroneous conclusion that formation of C3H2

+ + H is endothermic by 4 kJ mol-1 [smi84]. Several ab initio calculations later obtained the same value [won93], or an even higher one of 7 kJ mol-1 [mal93a].

Many of the conclusions and assumptions are in conflict with experiments performed in a 22-pole trap at temperatures reaching down to 10 K and at buffer gas densities below 1012 cm-3 [sor94]. The results obtained for a wide range of temperatures and densities

APPENDIX C

90

clearly show that the hydrogen abstraction reaction increases with decreasing tempera-ture down to 20 K, indicating an at least thermoneutral way to the products. The forma-tion of C3H3

+ via radiative association shows a stronger temperature dependence and dominates finally at 10 K. Additional information on the energetics has been obtained by comparing n-H2 and p-H2 as reactant in [sor94], and also rate coefficients for the isotope combinations C3H

+ + D2 and C3D+ + D2 have been reported. Unfortunately, all

these results have been ignored in the relevant literature [eva99] and are not included in the UMIST database. Due to the central importance of this collision system the C3H

+ + H2 measurements have been reproduced and extended [sav04a] and, in addition, the following reaction channels have been studied at 15 K

C3H+ + HD → C3D

+ + H2 (3a) → C3HD+ + H (3b) → C3H2

+ + D (3c) → C3H2D

+ + hν, (3d)

As will be outlined in the discussion section, reaction (3b) and the radiative association (3d) will play an important role in forming deuterated ions at low temperatures.

In contrast to the fast reactions measured for systems which are pretended to be endo-thermic, almost no product formation has been observed for C3H2

+ + H2 → C3H3+ + H

which is, without any doubt, exothermic for both isomers [won93]. Experiments per-formed with different isotope combinations, e.g.

C3H2+ + D2 → C3D2

+ + H2 (4a) → C3HD+ + HD (4b) → C3DH2

+ + D (4c) → C3H2D2

+ + hν, (4d)

have been found that these reactions are so slow that they can be ignored in astrochemi-cal models. An explanation for this is a barrier which has been calculated to be 9 kJ mol-1 high [won93]. Nonetheless finite rate coefficients have been measured at low temperature for various isotope combinations indicating that there must be some ways to circumvent the barrier or to tunnel through it.

In the next section a few short remarks concerning the experimental technology are made and three typical examples of raw data are given in order to provide evidence for the sensitivity, the reliability and in some cases the complexity of the 22-pole trapping method. The results which are put together in Table 1. are discussed in section 3, In this part more information on the reaction dynamics is added which is mainly based on speculative details of the potential energy surfaces or on symmetry selection rules af-fecting the branching ratios in H - D scrambling. In the conclusions the astrochemical consequences of our low temperature results are summarized.

2. Experimental and typical results

2.1. 22-pole ion trap All laboratory measurements presented in this paper and the related work [sav04a] have been performed in the variable temperature rf 22-pole ion trapping apparatus most de-tails of which have been described thoroughly in [ger92] and [ger95]. Many additional hints can be found in other publications [pau94], [scl99], [ger02a], and [asv04a]. In the trap the ion cloud is confined in radial direction by an effective potential which is cre-

APPENDIX C

91

ated by the electric multipole field alternating in time and space. In axial direction the storage volume is bordered by small potential barriers created by suitable voltages ap-plied to the gate electrodes located at the two ends of the trap. Each of the two sets of 11 electrodes is press-fitted into two copper electrode which is mounted onto a closed cycle refrigerator. Thin sapphire plates are used for electric insulation and good thermal con-tact. The trapping volume is enclosed by copper walls. In order to achieve low tempera-tures - the present experiments have been performed at 15 K in order to avoid condensa-tion, 5 K have been reached in a new trap - the trap is surrounded by a second thermal shield held at about 50 K. The temperature is usually measured using a carbon resistor, a calibrated diode or a hydrogen gas thermometer.

2.2. Measuring procedure Primary ions C3Hn

+ (n = 0 - 3) have been produced by electron bombardment in a dif-ferentially pumped rf storage ion source using different precursors. The extracted ions are mass selected in an rf quadrupole optimized for an adequate mass resolution in the range from 36 u to 39 u on one side and for transferring the selected ions efficiently into the trap on the other side. Already in the storage ion source, several methods have been employed for chemically quenching those ions which have the same mass as the chosen 12C3Hn

+ but contain 13C or D isotopes. Adding small amounts of hydrogen gas to the hydrocarbon used as precursor has been found to work in certain cases, most probably also for quenching excited isomers. The selected ions are injected into the trap where they are stored for times varying from a few ms to tens of s. Directly after injection they are cooled down rapidly to the wall temperature by using a short (≈ 10 ms) intense He pulse. The neutral reactant gas is let into the trapping region via pre-cooled tubes. The HD used (Cambridge Isotope Laboratories Inc.) has been tested in situ [asv04a] to have a purity of 98 % with 2 % of H2 /D2. After a given reaction time, the remaining primary ions and the formed product ions are extracted by a pulse applied to the exit electrode. They are mass analyzed in the second quadrupole mass filter and counted using a Daly type detector. The sequence of ion formation, injection, relaxation, reaction and analysis is repeated many times for each mass and typically for ten different storage times. As discussed previously, the error in determining absolute rate coefficients is mainly due to uncertainties in the number density of the target gas. If not otherwise stated, this results in an error of 20 % at maximum.

2.3. Preparation of primary ions, isomers Since C3H

+, C3H2+, C3H3

+ have cyclic and non-cyclic isomers, several tests based on reactions in the trap have been performed in order to determine the possible presence of excited isomers. In the case of C3H

+ the cyclic form is 220 kJ mol-1 higher than the lin-ear one and none of the tests indicated the presence of excited ions in the present ex-periment. It is a general experience with the storage ion source that ions are quenched quite efficiently to the ground state. The situation is less clear for C3H2

+; however, it has been found that these ions always react very slowly, and therefore, no additional tests have been performed. For preparing C3H3

+ ions, allene CH2=C=CH2 (Aldrich, 97%) has been used as neutral precursor. Also C3H3

+ exists in two different isomeric forms, the cyclic one being the ground state while the H2C3H

+ structure is 106 kJ mol-1 higher in energy [wys01]. It is known from the literature [smi87] that, under certain conditions, ionization of methyl-acetylene with 70 eV electrons results in ~ 65 % c-C3H3

+ ions while ~ 35 % are "l"-C3H3

+. Since the reactivity of the latter ones is much higher than that of c-C3H3

+ [aus81], [eva94], it is possible to quench them in suitable collisions. In most trapping studies performed in our lab with the ion storage source, only mono-

APPENDIX C

92

exponential decays have been observed, one exception being C3H2+. This is taken as an

indication that this type of ion source produces predominantly the lowest energy isomer. The chance that the reactivity of both isomers is the same over a wide range of tempera-tures is very low.

2.4. Typical results A set of data for reaction (1) which is typical for our trapping experiments is shown in Fig. 1. Every 8 s a well-prepared number of C3H3

+ ions was injected into the trap, in average about Ni=1 = 1000 per filling. Without any further analysis it becomes clear that the H - D exchange in C3H3

+ + HD collisions is extremely slow since the number of primary ions does not change from 1 s to 8 s although they undergo some 104 collisions with the abundant HD gas ([HD] = 3.5 × 1012 cm-3). A careful analysis of the time de-pendence indicates a minor decay of C3H3

+. Since this loss may not be only due reac-tions it is more reliable and also more sensitive to evaluate the increase of product ions. As can be seen from Fig. 1, traces have been found on masses 40 u - 42 u. Note that less than three ions per filling are produced in the 8 s from the initially injected C3H3

+. The ions found on mass 42 u are assigned to C3H4D

+ formed by radiative and ternary asso-ciation. The masses 40 u and 41 u can be associated with C3H2D

+ produced via reaction (1) and C3HD2

+ produced via a second deuteration step, respectively. A simple evalua-tion of the C3H2D

+ and C3HD2+ products formed within the storage time from the pri-

mary ions (less than 3 from 1000) leads immediately to a rate coefficient of 1 × 10-16 cm3s-1. This small value is already sufficient for completely excluding reaction (1) from astrochemical models; however, it is certainly a question of fundamental im-portance, whether the H - D exchange in reaction (1) is completely forbidden or whether there are paths to form the exothermic deuterated product, e.g. via tunneling.

In order to determine more precisely such small rate coefficients, a complicated analysis of the data is required. There are several other reactions possible, e.g. production of C3H4

+ + D. This channel and several others have been excluded to play a role by inject-ing externally produced and mass selected C3H4

+ ions into the trap and monitoring their reactions with HD under similar conditions. Other perturbations are due to minor traces of 13CC2H2

+, 13C2CH+, 13C3+ or C3HD+ mixed into the mass 39 u primary beam since

some of them can react with HD via H - D exchange or hydrogen abstraction with rate coefficients which are 106 times larger. In addition, one must account for H2, D2 (2%) and other impurities in the target gas. It is obvious that most of these effects actually reduce the reaction rate coefficient for reaction (1), derived from the measured data. Accounting quantitatively for a variety of such processes, the time dependence of the recorded masses has been modeled with an adequate system of coupled differential equations. A typical set of solutions is shown in Fig. 1 as solid lines. The resulting rate coefficients are included in Table 1. Selecting from various measurements performed under different conditions those with the best quality it is concluded that the rate coeffi-cient for deuteration of C3H3

+ in collisions with HD is smaller than 1 × 10-16 cm3s-1.

Fig. 2 shows the time dependence of a variety of ions which are produced from initially injected C3H

+ ions in sequential reactions with HD at 15 K. In comparison to Fig. 1 a two orders of magnitude lower target density has been used, [HD] = 2.6 × 1010 cm-3. Nonetheless, the primary ions disappear with a time constant of ~70 ms indicating that fast reactions play a role. As already mentioned above, the mono-exponential decay over two orders of magnitude provides evidence that only ground state isomers come out of the storage ion source. As can be seen from the various products indicated in the figure, a rich chemistry is started with the C3H

+ + HD collisions. Comparing the two

APPENDIX C

93

initially dominant channels, deuteron abstraction, C3HD+, and hydrogen abstraction, C3H2

+, shows that an incredible isotope effect has been discovered favoring deuteration. This is especially obvious since on mass 38 u the dominant product is C3D

+, formed via H - D exchange. All relevant primary and secondary reactions have been modeled with a system of coupled differential equations. The solutions are shown in Fig. 2 as solid lines; the resulting rate coefficients are included in Table 1.

A third example for reaction sequences occurring in the trap is shown in Fig. 3 for C3H2

+ reacting with D2 at 10 K. Various experiments have shown that the hydrogen abstraction reaction C3H2

+ + H2 → C3H3+ + H is very slow at room temperature [smi84],

[han89], [eva99] while it becomes faster at elevated ion kinetic energies. The conclusion of an endothermicity of 17 kJ mol-1 [smi84] is in contradiction to a rate coefficient of 6 × 10-12 cm3s-1 measured at 80 K [ger92] and to ab initio molecular orbital calculations [won93] which predict exothermicities of 8 kJ mol-1 and 91 kJ mol-1 for forming l-C3H3

+ and c-C3H3+ in collisions of l-C3H2

+ and c-C3H2+ with H2, respectively. Obvi-

ously the slow rates observed are due to a barrier of several kJ mol-1. This is also in ac-cordance with the results plotted in Fig. 3. Despite the high number density of deute-rium, [D2] = 1.7 × 1012 cm-3, only 30 % of the C3H2

+ primary ions have been converted into products. Solutions from numerical simulation of the reaction system are shown as solid lines, the resulting rate coefficients can be found in Table 1. Note that the domi-nant product, mass 40 u is first produced via an H2 - D2 switching reaction H - D ex-change and reactions leading to C3HmD3-m

+ are significantly slower.

3. Discussion and astrochemical consequences The chemistry of hydrocarbon ions has been the object of many experimental and theo-retical efforts until the early nineties; however, inspection of the relevant literature re-veals that there was a certain stagnation in the last ten years. This is in contrast to the progress made in observations and therefore it is not surprising that chemical reaction networks are not (or not any more) able to describe the details of measured abundances. For observed molecules such as C4H and C3H2, Gerin et al. [gri03] wonder whether im-portant reaction paths are missing or whether the relevant reaction rates are incorrect. The results reported in this work and for C3

+ and C3H+ + H2 in [sav04a] provide a clear

answer to these questions: our understanding of hydrocarbon reaction under interstellar conditions must be revised! In the following discussion several aspects of the hydro-genation or deuteration of C3Hn

+ are discussed going from n = 0 to 3. It already has been mentioned in the experimental section that all these ions have different isomers includ-ing cyclic, bent and linear arrangements of the three carbon atoms. There is experimen-tal evidence that the storage ion source produces ground state isomers; however, no in-formation has been obtained on the structure of the products although the method of chemical probing could be used.

3.1. C3+ + HD

Hydrogenation of carbon clusters has been studied rather often. Despite an exothermic-ity of 195 kJ mol-1 [han89] C3

+ reacts with H2 surprisingly slow. Published room tem-perature values vary between 1.8 × 10-10 cm3s-1 and 4.6 × 10-10 cm3s-1 indicating that the reactivity may depend quite strongly from the way the experiments are performed. McEwan et al. report a value of k = 2.4 × 10-10 cm3s-1 [eva99] in an overview summariz-ing the most recent knowledge concerning ion molecule reactions between CmHn

+ and hydrogen molecules and atoms. Our recent temperature dependent study [sav04a] has

APPENDIX C

94

shown that C3H+ formation occurs with a rate coefficient of 1.7 × 10-9 cm3s-1 at tem-

peratures below 50 K, a value which is, within our experimental error of 20 %, in agreement with the capture rate coefficient.

In addition experiments have been performed for C3+ reacting with HD at number den-

sity below 1010 cm-3. The time dependence of the composition of the trap content (see Fig. 4 of [sav04a]) is more complicated than shown for C3H

+ + HD in Fig. 2 since al-ready in the first step H - D exchange competes with H or D atom abstraction and radia-tive association. As discussed in detail in [sav04a], the evaluation of the kinetics is also complicated by the fact that the intermediate products formed in exothermic processes react with slightly different rate coefficients because of their internal excitation. In order to get precise information on the C3D

+ / C3H+ branching ratio extremely small amounts

of HD have been used with partial pressures similar to the background gas.

As can be seen from the rate coefficients given in Table 1, formation of C3D+ via deu-

teron abstraction is slightly faster than hydrogen abstraction. This is in contradiction to a simple classical picture, which is based on orientation of HD because of the anisotropy of the interaction potential. The separation of the center of charge and the center of mass orients the HD, especially if it is not rotating, such that the H-atom is preferentially pointing towards the ion. This would lead to a branching ratio C3D

+ / C3H+ smaller than

one while the measured value is 1.22. Therefore is can be concluded that hydrogen ab-straction does not occur in a direct way. This is corroborated by the temperature de-pendence of hydrogen abstraction which can be explained with the formation of a colli-sion complex the lifetime of which must be long enough for rearrangement. Looking at possible (C3H2

*)+ structures [won93] it can be presumed that first the hydrogen mole-cule is loosely attached to the linear C3

+ ion which must be bent in order to break the hydrogen bond. Later it has to rearrange again to form the energetically lower linear C3H

+ product. A hint that the initially formed intermediate state has a small binding energy is the fact that the reaction probability approaches unity only below 50 K. An-other hint that some of the collision complexes are really long-lived is that also radiative association contributes with more than 3%.

3.2. C3H+ + HD

In their paper on the competition between association and reaction for C3H+ + H2, Mal-

uendas et al. [mal93a] have discussed that their theoretical results are in serious discrep-ancy with measurements from an 80 K ring electrode trap experiment [ger92]. They proposed that studies between 10 K and 300 K should be undertaken to give some fur-ther guidance to ab initio calculations. This has been done utilizing the first version of our temperature variable 22-pole ion trap; however, the important results [sor94] have been ignored in later publications discussing the carbon chemistry in interstellar clouds [eva99]. They are also not included in the collection of the rate coefficients used in models [teu00].

The obvious difference between flow tube and trapping experiments is the density and temperature range, the interaction time and, may be, difficulties in preparing thermal-ized ions. In a trap most problems can be avoided or accounted for by systematically varying the pressure over many orders of magnitude and changing the storage times from ms to min. Thermalization of the primary ions can be achieved efficiently by in-tense buffer gas pulses filling the trap for short times. All experiments presented in [sor94] and the new ones published in [sav04a] clearly proof that hydrogen abstraction in C3H

+ + H2 collisions is not endothermic and that the temperature dependence ob-served in flow tubes is just due to saturated stabilization of the long lived collision com-

APPENDIX C

95

plexes as already pointed out in [ger92]. Also the effective rate coefficients reported in [eva99] have to be reconsidered.

Less obvious to us is why all high quality calculations [mal93a], [mal93b] [won93] pre-dict thermochemical values in contradiction to our experimental findings. It should be possible today to get energies with the required accuracy. More demanding and chal-lenging for ab initio calculations is to find the detailed features of the potential surface which are needed to explain experimental observations such as competition between radiative association and hydrogen abstraction and especially the very complicated iso-tope effects reported for reactions with D2 in [sor94] and for HD in this contribution. Note for example, that the rate coefficient for radiative association of C3H

+ in collisions with D2 is 50 times smaller then with H2. The model potential which has been used to understand qualitatively the data presented here and in [sav04a] is based on shallow wells in the entrance and exit channel, and barriers or bottle necks hindering the transi-tion to and from the strongly bound intermediate (see Fig. 4 of [sor94]).

The results obtained now for C3H+ + HD collisions, are very surprising. As can be seen

directly from Fig. 2 formation of C3HD+ is by far the dominant channel for this astro-chemically important isotope combination. Inspection of the rate coefficients given in Table 1, reveals that H - D exchange and radiative association leading to C3H2D

+ is al-ready ten times less efficient while the rate coefficient for C3H2

+ formation is hundred times smaller. One possible explanation for that is an efficient switching reaction, i.e., the addition of the HD target molecule on one side of the linear chain and the loss of the H on the other side. This mechanism is partly supported by the results obtained for C3D

+ + HD and C3H+ + D2 (see Table 1). However it also can be seen from the large rate

coefficients measured for C3D+ + D2, that in reality the reaction dynamics are more

complicated. Since in all the various isotope combinations of the C3H+ + H2 system the

same potential surface plays a role it must be assumed that either the zero point energies in the transition states are responsible for the remarkable differences or restrictions due to symmetry selection rules.

Most important for the production of deuterated hydrocarbons in interstellar chemistry is that C3H

+ + HD leads to deuterated products with almost 50 % of the collision rate.

3.3. C3H3+ + HD and C3H2

+ + D2 From a point of their reactivity, collisions of C3Hn

+ with hydrogen or deuterium can be ignored for n = 2 and 3 in astrochemical models. Nonetheless they are interesting from a fundamental point of view since in both cases exothermic reactions are possible. The potential surface of C3H3

+ + HD has been discussed in detail by Talbi and Herbst [tal01]. The ab initio results show a well in the entrance channel which is 1.7 kJ mol-1 below reactants and which corresponds to a long-distance c-C3H3

+ - HD van der Waals molecule. The exit channel C3H2D

+ - H2 lays 6 kJ mol-1 below reactants. As a conse-quence, long lived complexes can be formed only at very low energies, if at all. More-over, the calculations indicate that the reaction can only proceed via a transition state which can be associated to a complex formed from c-C3H2 and H2D

+. Since this barrier is 320 kJ mol-1 above reactants it is clear, without any further calculations, that no ex-change will become observable. The experimental result presented in Table 1 is in ac-cordance with this conclusion. The data indicate an overall deuteration rate coefficient of 10-16 cm3s-1, which is most probably due to radiative or partly due to ternary associa-tion. More conclusive results would require additional measurrements; however, they are not necessary for the present purpose. It is certain that reaction (1) does not play any role in low temperature astrochemistry.

APPENDIX C

96

The same statement holds for the deuterated variants of C3H2+ + H2. In this case it is

known from calculations [won93] that hydrogen abstraction is quite exothermic (91 kJ mol-1) but hindered by a barrier of 9 kJ mol-1. The results given for reaction (4) in Table 1 are in accordance with this; however, it is rather clear, also from Fig. 3, that deuterated products are really formed. The speculation that this may be due to tunnel-ling is supported a significantly larger rate coefficient of 5 × 10-13 cm3s-1 which has been measured for the reaction C3H2

+ + H2 → C3H3+ + H for both n-H2 and p-H2 [sor94].

4. Conclusions In the center of this contribution are new low temperature ion trap measurements which are of general relevance for the carbon chemistry of interstellar clouds. The presented experimental results have been selected according to simple hydrocarbon reactions which many contribute to the production of C3HD. It is concluded that the chemical reaction networks have to be revised since important reactions are either missing or they are included with erroneous rate coefficients, especially concerning their temperature dependence. The detailed data show that the use of simple statistical branching ratios in distributing the hydrogen or deuterium atom from an HD target molecule has to be checked in each case; for C3

+ + HD, the experimental reality is close to 1 : 1 while com-pletely different ratios are obtained for the competing products in C3H

+ + HD collisions.

It is rather sure that the new rate coefficients will increase the number density of hydro-carbons which can be produced under conditions of dense interstellar clouds. It can be foreseen that also more details such as deuterium fractionation and correlations between different hydrocarbons will become in better accordance with the observations [gri03]. If, however, some differences remain, other reaction paths have to be identified. From a fundamental point of view there are many reactions which are of basic interest and a challenge for low temperature trapping experiments. For example reactions starting with C+ colliding with deuterated variants of hydrocarbons, e.g. C+ + C2H3D C3H2D

+ + H, or reactions involving CH2D

+ or DCO+. Of importance are certainly collisions of H2D+

with carbon clusters or hydrocarbons, one example being the direct deuteration of C3H2 in collisions with H2D

+. For D3+ + C3 first experiments have been reported [sav04b].

The results confirmed the theoretical predictions [fis04] that deuteron transfer is by far the dominant reaction channel; nonetheless it may be possible at low temperatures that H2D

+ + C3 can associate via photo emission. An example which also still belongs to the class of C3Hn

+ reaction systems is the reaction H3+ + CH2CCH2 c-C3H3

+ + 2 H2. If hydrogenation of carbon molecules occurs at a reasonable rate via radiative association with H-atoms, this reaction can be regarded also as an efficient way in producing hy-drogen molecules.

The near thermoneutral reactions involving hydrocarbon ions indicate the need of very sensitive low temperature and low density experiments as realized with the trapping technique while the traditional methods of ions chemistry seem to fail to provide the necessary information. The result measured for C3H3

+ HD and similar observations made for CH5

+ HD [asv04b] proof of the sensitivity of the trapping method. The de-termination of rate coefficients smaller than 10-16 cm3s-1 are important for finding routes with low probabilities. In addition to the reactions with HD or D2 experiments with condensable molecule, radicals and atoms such H or D must be performed at low tem-peratures. For example all reactions CmHn

+ + D → CmHn-1D+ + H are exothermic; how-

ever it is not straight forward to predict the reactivity at low temperatures from 300 K

APPENDIX C

97

measurements [eva99]. In a new special trapping apparatus constructed in our labora-tory, such studies are in progress [bor04].

Finally it must be concluded from the failure of the related theory that, in addition to low temperature experiments, quantum chemical calculations are needed with "astro-chemical accuracy". It is supposed that this is more demanding than spectroscopic accu-racy in which case one needs to be precise only in the vicinity of the ground state con-figuration while in a chemical reaction a variety of asymptotic regions and critical tran-sition states have to be known with accuracies much better than 1 kJ mol-1. Very impor-tant, especially for deuteration, are zero point energies in all critical regions. Provided that all details of the potential energy surface are available, detailed dynamical calcula-tions have to be performed or, alternatively, statistical theories may be used. This is often applicable at low temperature ion-molecule reactions with long lived and strongly interacting collision complexes. In these cases special attention has to be paid to the restrictions imposed by symmetry selection rules, which are especially important in the case of D - H exchange and if o-H2 is participating.

Acknowledgments Financial support by the Deutsche Forschungsgemeinschaft (DFG) is gratefully ac-knowledged, especially via the Forschergruppe FOR 388 "Laboratory Astrophysics".

APPENDIX C

98

References [ada87] Adams, N. G., Smith, D.: On the Synthesis of c-C3H2 in Interstellar Clouds, ApJ 317 (1987) L25-

L27.

[asv04a] Asvany, O., Savić, I., Schlemmer, S. and Gerlich, D.: Variable Temperature Ion Trap Studies of CH4

+ + H2, HD and D2: Negative Temperature Dependence and Significant Isotope Effect, Chem. Phys. 298 (2004) 97-105.

[asv04b] Asvany, O., Schlemmer, S. and Gerlich, D.: Deuteration of CHn+ (n = 3 - 5) in Collisions with

HD Measured in a Low Temperature Ion Trap, ApJ (2004) accepted.

[aus81] Auslos, P. J. and Lias, S. G.: Discrimination of C3H3+ Structures on the Basis of Chemical Reac-

tivity, J. Am. Chem. Soc. 103 (1981) 6505-6507.

[bel88] Bell, M. B., Avery, L. W., Matthews, H. E., Feldman, P. A., Watson, J. K. G., Madden, S. C., Irvine, W. M.: A Study of C3HD in Cold Interstellar Clouds, ApJ 326 (1988) 924-930.

[bor04] Borodi, G., Luca, A., and Gerlich, D., Probing Protonated Methane, CH5+, with H and D-Atoms,

in preparation.

[cer02] Čermák, I., Savić, I., and Gerlich, D. Ion-Trapping Apparatus for Studies on Reactions between Ions and Neutral Carbon Species, in: Safránková (ed): Proceedings of Contributed Papers, Part II, WDS'02, Matfyzpress (2002) 281-287.

[fis04] Fischer, G., Barthel, R. and Seifert, G.: Molecular Dynamics Study of the Reaction C3 + H3+, sub-

mitted to EPJ.

[ger90] Gerlich, D.: Ortho-Para Transitions in Reactive H++H2 Collisions, J. Chem. Phys. 92 (1990) 2377-2388.

[ger92] Gerlich, D. and Horning, S.: Experimental Investigations of Radiative Association Processes as Related to Interstellar Chemistry, Chem. Rev. 92 (1992) 1509-1539.

[ger95] Gerlich, D.: Ion-Neutral Collisions in a 22-Pole Trap at Very Low Energies, Phys. Scr. T59 (1995) 256-263.

[ger02a] Gerlich, D., Herbst, E. and Roueff, E.: H3+ + HD ↔ H2D

+ + H2: Low-Temperature Laboratory Measurements and Interstellar Implications, Planet. Space Sci. 50 (2002) 1275-1285.

[ger02b] Gerlich, D. and Schlemmer, S.: Deuterium Fractionation in Gas-Phase reactions Measured in the Laboratory, Planet. Space Sci. 50 (2002) 1287-1297.

[gri03] Gerin, M., Fosse, D. and Roueff, E.: Carbon Chemistry in Interstellar Clouds, in: Curry, C. L. and Fich, M. (eds): Proceedings of the Conference Chemistry as a Diagnostic of Star Formation, Uni-versity of Waterloo, August 2002, NRC Research Press (2003); http://arxiv.org/PS_cache/astro-ph/pdf/0212/0212058.pdf.

[ger04] Gerlich, D.: Influence of Exchange Symmetry on Low Temperature Ion-Molecule Reactions, in Capozza, G. and P. Casavecchia, P. (eds.), Symposium on Atomic, Cluster and Surface Physics, (2004), IL-2

[han89] Hansel, A., Richter, R., Lindinger, W.: Reactions of C2 and C3 Hydrocarbon Ions with H, D, H2 and D2 at Near-Thermal Energies. Int. J. Mass Sprectrom. Ion Proc. 94 (1989) 251-260.

[her83] Herbst, E., Adams, N. G. and Smith, D.: Laboratory Measurements of Ion-Molecule Reactions Pertaining to Interstellar Hydrocarbon Synthesis, ApJ. 269 (1983) 329-333.

[her84] Herbst, E., Adams, N. G. and Smith, D.: Theoretical Reinvestigation of Hydrocarbon and Cyanoacetylene Abundances in TMC-1, ApJ. 285 (1984) 618-621.

[how93] Howe, D. A., Millar, T. J.: Alternative Routes to Deuteration in Dark Clouds, Mon. Not. R. Astron. Soc. 262 (1993) 868-880.

[kai02] Kaiser, R. I.: Habilitationsschrift Chemnitz, 2002; Experimental Investigation on the Formation of Carbon-Bearing Molecules in the Interstellar Medium via Neutral-Neutral Reactions, Chem. Rev. 102 (2002) 1309-1358

APPENDIX C

99

[loi01] Loinard, L., Castets, A., Ceccarelli, C., Caux, E., Tielens, A. G. G. M.: Doubly deuterated molecu-lar species in protostellar environments, ApJ 552 (2001) L163-L166.

[mal93a] Maluendes, S. A., McLean, A. D., Yamashita, K., and Herbst, E.: Calculations on the Competition between Association and Reaction for C3H

+ + H2, J. Chem. Phys. 99 (1993) 2812-2820.

[mal93b] Maluendes, S. A., McLean, A. D., Herbst, E.: Calculations Concerning Interstellar Isomeric Abundance Ratios for C3H and C3H2, ApJ.,417 (1993) 181-186.

[mar01] Markwick, A. J., Charnley, S. B., Millar, T. J.: Deuterium Fractionation along the TMC-1 Ridge A&A 376 (2001) 1054-1063.

[mar02] Markwick, A. J., Millar, T. J., Charnley, S. B.: CH2DCCH along the TMC-1 Ridge, A&A 381 (2002) 560-565.

[ewa94] McEwan, M. J., McConnell, C. L., Freeman, C. G., and Anicich, V. G.: Reactions of Isomeric C3H3

+ Ions: A combined Low Pressure - High Pressure Study, J. Phys. Chem. 98, (1994) 5068-5073.

[ewa99] McEwan, M. J., Scott, G. B. I., Adams, N. G., Babcock, L. M., Terzieva, R. and Herbst, E.: New H and H2 Reactions with Small Hydrocarbon Ions and Their Roles in Benzene Synthesis in Dense Interstellar Clouds, ApJ 513 (1999) 287-293

[mil02] Millar, T. J.: Modeling Deuterium Fractionation in Interstellar Clouds, Planet. Space Sci. 50 (2002) 1189-1195.

[mil89] Millar, T. J., Bennett, A. and Herbst, E.: Deuterium Fractionation in Dense Interstellar Clouds, ApJ. 340 (1989) 906-920.

[pau94] Paul, W. and Gerlich, D.: The Problem of Cooling Ions in a Low Density Ion Trap, in: Märk, T. D., Schrittwieser, R. and Smith, D. (eds.), Contributions of Simposium on Atomic, Cluster and Surface Physics ‘94, (1994) 197-199.

[rob00a] Roberts, H., Millar, T. J.: Modelling of Deuterium Chemistry and its Application to Molecular Clouds, A&A 361 (2000) 388-398.

[rob00b] Roberts, H., Millar, T. J.: Gas-Phase Formation of Doubly-Deuterated Species, A&A 364 (2000) 780-784.

[rou00] Roueff, E., Tiné, S., Coudert, L. H., Pineau des Forêts, G., Falgarone, E., Gerin, M.: Detection of Doubly Deuterated Ammonia in L134N, A&A 354 (2000) L63-L66.

[sav04a] Savić, I. and Gerlich, D.: Temperature Variable Ion Trap Studies of C3+ and C3H

+ + H2 and HD, submitted to PCCP (2004).

[sav04b] Savić, I., Čermák, I. and Gerlich, D.: Reactions of Cn (n = 1 - 3) with Ions Stored in a Tempera-ture-Variable Radio-Frequency Trap, submitted to Int. J. Mass Spectrom. (2004).

[scl99] Schlemmer, S., Kuhn, T., Lescop, E. and Gerlich, D.: Laser excited N2+ in a 22-Pole Trap, Ex-

perimental Studies of Rotational Relaxation processes, Int. J. Mass Spectrom. 185, (1999) 589-602.

[sco97] Scott, G. B. I., Fairley, D. A., Freeman, C. G., McEwan, M. J., Adams, N. G. and Babcock, L. M.: CmHn

+ Reactions with H and H2: An Experimental Study, J. Phys. Chem. A 101, (1997) 4973-4978.

[smi84] Smith, D. Adams, N. G., Ferguson, E.E.: The Heat of Formation of C3H2+, Int. J. Mass Spec-

trom. Ion Proc., 61 (1984) 15-19.

[smi87] Smith, D. and Adams, N. G.: Cyclic and Linear Isomers of C3H2+ and C3H3

+: The C3H+ + H2

Reaction, Int. J. Mass Spectrom. Ion Proc. 76, (1987) 307-317.

[sor94] Sorgenfrei, A. PhD thesis Freiburg 1994., Sorgenfrei, A. and Gerlich, D.: Ion-Trap Experiments on C3H

+ + H2: Radiative Association vs. Hydrogen Abstraction, in Nenner, I. (ed.), Molecules and Grains in Space, AIP Press, New York (1994) 505-514.

[teu00] Le Teuff, Y.H., Millar, T.J., and Markwick, A.J., The UMIST Database for Astrochemistry 1999, A&AS. 146 (2000) 157-168.

APPENDIX C

100

[tal01] Talbi, D., Herbst, E.: On the Deuteration of C3H2 in Dense Interstellar Clouds via the Reaction C3H3

+ + HD → C3H2D+ + H2, A&A 376 (2001) 663-666.

[tur01] Turner, B. E.: Deuterated Molecules in Translucent and Dark Clouds, ApJS 136 (2001) 579-629.

[won93] Wong, M. W. and Radom, L.: Thermochemistry and Ion-Molecule Reactions of Isomeric

C3H2•+ Cations, J. Am. Chem. Soc., 115 (1993) 1507-1514.

[wys01] Wyss, M., Riaplov, E. and Maier, J. P.: Electronic and Infrared Spectra of H2C3H+ and Cyclic

C3H3+ in Neon Matrices, J. Chem. Phys. 114, (2001) 10355-10361.

Tables Table 1. Selected reaction rate coefficients which are relevant for forming deuterated C3H3

+. The data have been measured in the 22-pole trap at 15 K with HD [sav04a].

Reaction k / cm3s-1 Remarks

C3+ + HD → C3D

+ + H 9.3 × 10-10

→ C3H+ + D 7.6 × 10-10

→ C3HD+ + hν 5.9 × 10-11

C3H+ + HD → C3D

+ + H2 5.6 × 10-11

→ C3HD+ + H 4.6 × 10-10

→ C3H2+ + D 3.0 × 10-12

→ C3H2D+ + hν 3.2 × 10-11

C3D+ + HD → C3HD+ + D 1.0 × 10-10 (1)

→ C3D2+ + H 8.3 × 10-11

→ C3HD2+ + hν 8.0 × 10-12

C3H+ + D2 → C3D

+ + HD 3.0 × 10-13 (2)

→ C3HD+ + D 1.0 × 10-11 (2)

→ C3D2+ + H 2.7 × 10-11 (2)

→ C3HD2+ + hν 4.0 × 10-12 (2)

C3D+ + D2 → C3D2

+ + D 1.7 × 10-10 (2)

→ C3D3+ + hν 1.3 × 10-10 (2)

C3H2+ + D2 → C3D2

+ + H2 1.7 × 10-14 (2)

→ C3HD+ + HD 3.0 × 10-15 (2)

C3D2+ + D2 → C3D3

+ + D 1.0 × 10-16 (2)

C3HD+ + D2 → C3HD2+ + D 1.0 × 10-14 (2)

C3H3+ + HD → C3H2D

+ + H2 < 4 × 10-16

→ C3H4D+ 1.5 ×10-16 (3)

1 C3D+ has not been injected from the source, the rate coefficients have been deduced from Fig. 2

2 Data taken from [sor94] 3 Effective association rate coefficient measured at 1012 cm-3

APPENDIX C

101

Figures

0 1 2 3 4 5 6 7 80,1

1

10

100

1000

C3H

4D+

C3HD

2

+C

3H

2D+

C3H

3

+

t / s

Ni

C3H

3

+ + HD →

Fig. 1. Average number of C3H3

+ and product ions as a function of storage time t (T = 15 K, [HD] = 3.5 × 1012 cm-3). Even after t = 8 s, the number of injected primary ions (~1000 per filling) is almost unchanged despite the rather large number density of target gas. Minor traces of products are de-tected on the masses 40 u - 42 u. Details concerning the assignment to association (C3H4D

+), deuteration (C3H2D

+) and sequential deuteration (C3HD2+) are discussed in the text. The solid lines are solutions of a

system of coupled differential equations describing the reaction system.

0.00 0.05 0.10 0.15 0.20 0.25 0.301

10

100

1000

C3H+ + HD →

13C12C2H

2D+, 13C12C

2D

2

+

C3HD2+

C3D

2

+, C

3H

2D+

C3HD+

, C3H

3

+

t / s

Ni

C3H+

C3D+, C

3H

2

+

Σ

Fig. 2. Time dependence of the average number of trapped ions as a function of storage time t (T = 15 K, [HD] = 2.6 × 1010 cm-3). The injected primary C3H

+ ions undergo a mono-exponential decay over two orders of magnitude. As can be seen from the indicated products a rich chemistry is started with C3H

+ + HD. The dominant product is deuterium abstraction leading to C3HD+. Hydrogen abstraction leads to C3H2

+ which has the same mass as H - D exchange leading to C3D+. Another important channel is for-

mation of C3H2D+ via radiative association. The solid lines are solutions of a system of coupled differen-

tial equations describing primary and secondary reactions resulting also in doubly deuterated products.

APPENDIX C

102

0 2 4 6 8 1010-4

10-3

10-2

10-1

100

C3H

2

+ + D2 →

C3D

3

+

C3D

2H+

C3HD+

C3D

2

+, C3DH

2

+

Ni /

ΣN

i

t / s

C3H

2

+

Fig. 3. Average number of C3H2

+ and product ions as a function of storage time t (T = 10 K, [D2] = 1.7 × 1012 cm-3). Despite the high number density and long storage times, only 30 % of the primary ions react. The dominant product, mass 40 u is most probably an H2 - D2 switching reaction. H - D ex-change is significant slower. Solutions from numerical simulation of the reaction system are shown as solid lines, the resulting rate coefficients can be found in Table 1.

103

APPENDIX D

To be submitted to Physical Chemistry Chemical Physics

(Bondybey Festschrift)

Temperature variable ion trap studies of C3+ and C3H

+ + H2 and HD

Savić, I. and Gerlich D.


Abstract Hydrogenation and deuteration of C3Hn

+ (n = 0 - 2) in collisions with H2 and HD has been studied in a variable temperature 22-pole ion trap from room temperature down to 15 K. Although exothermic, the formation of C3H

+ from C3+ + H2 is rather slow at room

temperature but becomes faster with decreasing temperature. This behavior is tenta-tively correlated to the increasing complex lifetime with decreasing temperature which allows for the isomerization steps needed for hydrogen abstraction. For C3

+ + HD it has been shown that production of C3D

+ is slightly favored over C3H+ formation. The con-

troversy which products are really formed in C3H+ + H2 collisions and deuterated vari-

ants has a long history. Previous and new ion trap results clearly proof that low density experiments are needed in order to detect the C3H2

+ + H product channel. Low tempera-ture results clearly proof that this reaction is not endothermic, in contradiction to ab initio calculations. Due to the influence of zero point energies, it shows very compli-cated isotope effects. For example abstraction reactions with HD are faster than with H2 while radiative association is slower. The most surprising result has been obtained for C3H

+ + HD, where C3HD+ formation is hundred times faster than C3H2+, most probably

the consequence of a switching reaction. The results are of fundamental importance for understanding the energetics, structures and reaction dynamics of the C3Hn

+ collision system. They indicate that quantum chemical calculations are presently not yet accurate enough for understanding such simple reactions at low energies. All results are of essen-tial relevance for the carbon chemistry of dense interstellar clouds, both for formation of small hydrocarbons and for deuterium fractionation.

Key words reaction dynamics. low temperatures, barriers, tunneling, nuclear spin restrictions, the role the isotopes D and 13C, zero point energies, laboratory astrochemistry; reactive col-lisions; hydrogenation; deuteration; ISM: molecules; C3

+; C3H+; C3D

+; C3H2+; C3HD+;

C3H3+; C3H2D

+

APPENDIX D

104

1. Introduction Low-temperature and low-density experimental studies of ion-molecule reactions are of key importance for understanding how interstellar molecules are formed. Precise reac-tion rate coefficients are needed for predicting the chemical evolution of interstellar clouds. For many reactions, the reaction rate coefficients are constant; however, meas-urements are needed since often unexpected changes occur at low temperatures. In the group of processes which are important for astrochemistry, belong those exothermic reactions which are hindered by a small barrier, near thermoneutral processes, isotope-exchange reactions, and radiative association reactions. Isotope exchange is slightly exothermic in one direction because of differences in zero-point energies. This leads to isotope fractionation, a very sensitive probe of the chemical and physical conditions. In the low densities of the interstellar medium, the bimolecular reaction, where the colli-sion complex is stabilized by emission of a photon plays a central role, one of the most important examples being C+ + H2 → CH2

+ + hν.

In addition to the astrophysical importance, ion-molecule collisions at low energies and involving cold reactants are of fundamental interest. They can be used as a good test for theories of reaction dynamics, because the quantum nature of processes at low tempera-tures becomes pronounced. These reactions are especially useful in probing of the po-tential energy surfaces if weak barriers or small endothermicities are involved or if the transition from reactants to products is hindered by dynamical constraints. One of the best-understood examples of a temperature dependent reaction is the hydrogen abstrac-tion reaction NH3

+ + H2 → NH4+ + H, which has been measured by several groups, with

different techniques, and in the temperature range from 11 K up to 800 K [feh75], [smi81], [lui85], [boh85], [ger93]. It has been observed that with falling temperature, the reaction rate coefficient decreases, reaches a minimum at about 150 K and signifi-cantly reincreases below this value. In measurements with isotope variants [kim75], [ada84], [bar87], a pronounced isotope effect has been found. Reaction of NH3

+ with D2 is slower than with H2 while ND3

+ reacts faster with H2 than NH3+. A statistical phase

space calculation of Herbst at al. [her91] confirmed the hypothesis from Luine and Dunn [lui85] and Böhringer [boh85] that at low temperatures, the mechanism of NH4

+ formation is dominated by the initial formation of a long-lived complex, from which tunneling through a small transition state barrier occurs. The rise of the reaction rate at high temperatures has been explained with an increasing probability of a classical pas-sage over the transition state barrier.

Other interesting examples include the isoelectronic reaction CH4+ + H2 leading to the

hypercoordinated carbocation CH5+ [asv04], the hydrogen abstraction reaction in

C2H2+ + H2 collisions [ger92], [haw92], and the collision system C3H

+ + H2 leading for example to the smallest cyclic ion, the c-C3H2 [ada87]. Common to all these systems is a complicated competition between various channels the intermediate collision complex can decay to. Especially at low temperatures often very long-lived collision complexes are formed. In addition to rare events such as tunneling etc. the lifetimes of µs or longer results in experimental difficulties such as saturated three-body association. An example is the C3H

+ + H2 collision system where formation of C3H3+ competes with hydrogen

abstraction as proven experimentally in an 80 K ion trap experiment [ger92]. The fact that in flow tube experiments less C3H2

+ was formed at lower temperatures has lead to the erroneous conclusion that formation of C3H2

+ + H is endothermic by 4 kJ mol-1 [smi84]. Several ab initio calculations later obtained the same value [won93], or an even higher one of 7 kJ mol-1 [mal93]. Many of the conclusions and assumptions are in con-

APPENDIX D

105

flict with experiments performed in a 22-pole trap at temperatures reaching down to 10 K and at buffer gas densities below 1012 cm-3 [sor94]. Due to the central importance of this collision system the measurements have been reproduced and extended in this work, In addition, special studies have been performed at 15 K related to the formation of deuterated hydrocarbons under interstellar conditions [sav04c].

It is well known that the neutral C3 molecule has a linear or quasi-linear structure [wel89], [mar95] in the ground state. Concerning the C3

+ cation, older theoretical stud-ies predicted a linear structure in analogy with the neutral species. Experimental data [fai87] and newer theoretical calculations indicate that C3

+ has in its ground state at least a bent, if not cyclic structure. For example, it has been shown in [gre90] (see also erra-tum [gre90a]), using high-level ab initio quantum mechanical methods that C3

+ is strongly bent (CCC bond angle about 70°) with a barrier to linearity of almost 17 kJ/mol. The ab initio studies in [tay91] come to the conclusion that the linear form of C3

+ lies 22 kJ/mol above the cyclic form while only 4 kJ/mol have been reported in [fur02]. In addition, it has been found that the transition state for C3

+ isomer intercon-version is only 5 kJ/mol above the less stable linear isomer. This is an indication of an extremely facile isomerization.

Since C3+ cations can be produced from many neutral precursors, it is probably also

possible to find ways to produce carbon cations in different structures. Comparing the reaction rate coefficients of C3

+ ions, produced by direct laser evaporation with the lit-erature values of C3

+ ions produced by a electron impact of different hydrocarbons, up to a factor of 3 difference in results have been measured [par89]. An open question is whether the observed differences are due to differences in structure or in internal excita-tion. In contrast to this, it has been noted in [lif93] that the properties of Cn

+ clusters prepared by a laser evaporation of graphite or by dissociative electron ionization are remarkably similar, suggesting similar structures and no dependence on the production scheme. In [fur02] dissociative ionization of several linear, cyclic and branched mole-cules was used to generate C3

+. Collisionally activated dissociation (CAD) mass spectra of C3

+ was independent on a wide variety of precursor molecules used and neutraliza-tion-reionization spectra indicate only one isomer. That is consistent with theoretical calculations that point to facile interconversion of C3

+ [fur02].

Reactions of C3Hn+ (n = 0 - 3) with H2 and isotope modifications already have been in-

vestigated experimentally but mainly at room temperature, in some cases at 80 K. Only a few experiments have been performed at temperatures below 80 K. In an early SIFT experiment [her83], C3

+ ions have been generated by electron impact ionization of me-thylacetylene C3H4. In 298 K collisions with H2, only the reaction

C3+ + H2 → C3H

+ + H (1)

has been observed. The reported rate coefficient, 3.0 × 10-10 cm3s-1, is in good accor-dance with the 300 K SIFD result [han89], 2.5 × 10-10 cm3s-1. In newer experiments per-formed at 296 K using the SIFT technique [boh90], C3

+ ions are generated by electron impact from diacetylene. In addition to reaction (1) ternary association C3

+ + H2 + He → C3H2+ + He has been observed in 5 % of the collisions indicating long

lived intermediates, even in this exothermic reaction.

Several experiments have been performed for the collision system C3+ + D2. In a 300 K

FTICR cell [mce87], only C3D+ products have been detected being produced with a rate

coefficient of 1.5 × 10-10 cm3s-1. The C3+ ions were generated by laser vaporization. In a

300 K SIFDT experiments [han89], a reaction rate coefficient of 1.8 × 10-10 cm3s-1 has

APPENDIX D

106

been measured for the same channel. It has been noted in [han89] that reactions of C3+

with H2 (D2) are fast exothermic H (D) atom abstraction reactions. In a 296 K SIFT study [boh90], two channels have been seen. Both C3D

+ (95 %) and C3D2+ (5 %) are

formed with a rate coefficient of 1.3 × 10-10 cm3s-1. The C3+ was produced from methy-

lacetylene and graphite. Early measurements performed in the temperature variable 22-pole ion-trap machine at a nominal temperature of 10 K, resulted in considerable higher rate coefficients in comparison to the results discussed above, 1.3 × 10-9 cm3s-1 [sor94]. This result especially showed the need to study the temperature dependence of reaction (1) in more detail.

2. Experimental The laboratory measurements have been performed in a variable temperature rf 22-pole ion trapping apparatus. A comprehensive description of the trapping technique used can be found elsewhere ([ger92], [ger92b], [ger95]). Ions are generated in a storage ion source, mass selected in a first quadrupole mass filter operated here exclusively in the mass selective mode and injected into a 22-pole ion trap. The ion cloud is confined in the radial direction by an effective potential created by applying two opposite phases of an rf generator (80 V, 17 MHz) to the two sets of 11 electrodes. The trap is closed in the axial direction by small potential barriers created by a suitable voltage applied to the two gate electrodes. Due to the wide nearly field free region inside the 22-pole trap, experiments at low collision temperatures are possible. The trapping volume is enclosed by copper walls mounted onto a closed cycle refrigerator. In order to achieve low tem-peratures - the present experiments have been performed at 15 K in order to avoid con-densation, 5 K have been reached in a new trap - the trap is surrounded by a second thermal shield held at about 50 K. Higher temperatures are obtained by electric heating of the trap holder. The temperature is usually measured using a carbon resistor, a cali-brated diode or a hydrogen gas thermometer. Buffer and reactant gases are introduced by cooled tubes and are in thermal equilibrium with the cold walls surrounding the trap. The translational and internal degrees of freedom of the ions are coupled to the envi-ronment by inelastic collisions with the buffer gas which can be introduced continu-ously or in pulsed way. The stored ions react with target gas the number density of which is determined with an ion gauge calibrated with respect to a spinning rotor gauge. After each trapping cycle, primary and product ions are extracted, analyzed by a second quadrupole mass filter and detected by a Daly-type detector.

The primary C3+ ions are produced by electron bombardment with an energy up to

40 eV in an external rf storage ion source. Allene (CH2=C=CH2) has been used as the neutral precursor (Aldrich, 97%). After mass selection in a quadrupole mass filter, the ions are injected into the 22-pole trap. All ions are cooled to the low temperatures of the walls by injecting a short intense He pulse (~10 ms) into the trap leading to typically 104 collisions in a few ms. The H2 target gas used in these experiments was purchased from Messer-Griesheim with a purity of 6.0. The HD target gas has been purchased from Cambridge Isotope Laboratories Inc. The specified purity is 97 %, the major contamina-tion being H2 and D2. In situ tests have shown that both molecules contribute less than 2 %.

Reaction rate coefficients are measured in the trap in an iterative mode. First, a pulsed bunch of ions is injected into the 22-pole ion trap with low kinetic energy. The ions are then cooled to the ambient temperature by the He buffer gas and stored for times vary-ing from milliseconds to seconds. For avoiding space charge effects and saturation of

APPENDIX D

107

the Daly detector, only several hundred primary ions are trapped per pulse. In order to improve the statistics and to reach a better precision in determining the reaction rate coefficients, the already described procedure of ion formation, trapping and reaction and analysis is repeated rather often for each mass of interest and for typically 10 different storage times.

3. Results

3.1. Reactions with H2 Fig. 1 shows a typical experimental result, the number of various trapped ions Ni as function of storage time t, measured at 15 K for a sequence of reactions. The number density of H2 was rather low, 8.5 × 109 cm-3. The number of C3

+ ions initially injected into the 22-pole decays exponentially by forming C3H

+ ions via the hydrogen abstrac-tion reaction (1). The C3H2

+ ions are partially formed by the radiative association reac-tion

C3+ + H2 → C3H2

+ + hν , (2)

in contrast to the high pressure SIFT technique [boh90] where ternary association, C3

+ + H2 + He → C3H2+ + He, completely dominates over reaction (2). At the low H2

number densities used in present experiments, ranging from 2.2 × 109 cm-3 to 8.5 × 109 cm-3, the probability for a ternary collision is negligible. Some of the C3H2

+ ions are also formed via the hydrogen abstraction reaction

C3H+ + H2 → C3H2

+ + H . (3)

In further steps the products C3H+ and C3H2

+ continue to react with H2 forming finally C3H3

+

C3H+ + H2 → C3H3

+ + hν , (4)

C3H2+ + H2 → C3H3

+ + H . (5)

The solid lines in Fig. 1 present the solutions of a rate equation system describing the chemical interaction of all trapped ions with H2. The rate coefficients used to fit the measured points are presented in Table 1.

Using the method described above and repeating the same experiment at different tem-peratures, reaction rate coefficients between 15 K and 300 K have been obtained. They are presented in Fig. 2 for reactions (1) and (2). The results for reactions (3) and (4) are shown in Fig. 3. The rate coefficient k5 is equal to 5 × 10-13 cm3s-1 in the temperature range between 15 K and 100 K and becomes smaller than 10-14 at higher temperatures. It has been set to zero in our kinetic model. For this reaction a value of 5 × 10-13 cm3s-1 has been reported at 10 K from earlier measurements [sor94]. This is in accordance with calculations [won93] which predict, despite the large exothermicity, a small barrier. In room temperature measurements reported in [her83] products from C3H2

+ + H2 have not been detected. The same holds for the C3H2

+ + D2 collision [han89]. It has been shown in [smi84] that products from C3H2

+ + H2 interactions appear only at elevated kinetic energy. Anyhow, since this reaction is much slower than the other ones observed, reac-tion (5) has no influence at all on the other rate coefficients. The errors for k1 are esti-mated to be around 10 %, mainly due to uncertainties in determining the effective H2 number density. Errors for k2, k3 and k4 are slightly bigger due to lower count rates. Errors in the determination of the temperature of the trap are assumed to be 10 K for high temperatures and not more then 5 K at low temperatures.

APPENDIX D

108

As can be seen from Fig. 2, the rate coefficient k1 has a large constant value below 50 K. With increasing temperature, k1 falls to 4.6 × 10-10 cm3s-1 at 300 K. In addition to the results from this work (filled circles) some 300 K literature values. Our 300 K val-ues are slightly higher then the values reported in [her83], [han89] and [boh90]. This can be explained by the ternary association process competing in the flow tubes with reaction (1). Only the rate coefficient k3 shows a trend to decrease with decreasing tem-perature below 30 K. This is can be explained by the fact that reaction (3) is in competi-tion with reaction (4). The measurements clearly show a rather steep increase of radia-tive association with decreasing temperature. Also reaction (1) is in competition with reaction (2) but due to the difference in reactivity which is more than one order of mag-nitude, this effect can not be seen. Above 70 K the reaction rate coefficient k2 and coef-ficient k4 become so small that their determination would require additional careful measurements, e.g. testing the contribution of ions containing 13C isotopes.

For applications in reaction networks of interstellar chemistry, e.g. in the UMIST data-base [teu00], the temperature dependence of rate coefficients is approximated using the function k = α (T / 300 K)β exp(- γ / T). In the present evaluation, the parameter γ which is mainly for endothermic reactions, has been set to zero. In order to get better agree-ment with the data, pairs of parameters (α, β) have been determined for selected tem-perature intervals. The results are listed in Table 1. Inspection of Fig. 2 and Fig. 3 re-veals that quite satisfying fits of the data can be obtained. The open symbols in Fig. 3 which go down to 10 K are results from earlier studies using the same 22-pol ion trap apparatus [sor94]. In these measurements the primary C3H

+ ions have been produced by electron impact dissociation of methyl acetylene. As can be seen from Fig. 3, the new results are in good overall agreement with the values reported [sor94].

3.2. Reactions with HD In addition to H2 also HD has been used as target under otherwise identical experimen-tal conditions. It is obvious from Fig. 4 where a variety of ions are plotted as a function of trapping time, that the situation is more complicated than in Fig. 1 due to the various isotopes. In the experiment also minor traces of mass 42 (most probably C3D3

+) has been observed; however, they are not presented in the figure since the signal is at the lower limit of the scale. The initially injected and relaxed C3

+ ions react with HD which is present in the trap with a number density of 7.9 × 109 cm-3. A detailed explanation of the complex chemical interplay is not easy since, in addition to H and D atom abstrac-tion and radiative association, also H - D exchange reactions are possible. Since there are a variety of different ions having the same mass, one has to be very careful in the interpretation of the data.

In order to explain the measured ion abundances, a couple rate model has been devel-oped. The processes included in this model are shown schematically in Fig. 5 and are also using information from the non-deuterated C3

+ + H2 reaction. Basically, C3+ ions

first react with HD and abstract an H or a D atom with an overall rate coefficient k1'. C3H

+ or C3D+ are formed with the probabilities p and 1 - p respectively. In addition, the

primary ions are allowed to form C3HD+ via radiative association (k2'), The C3H+ ions

undergo further reaction with HD by (i) H atom abstraction (k3') forming C3H2+, (ii) H-

D exchange reaction (k4') forming C3D+, (iii) D atom abstraction reaction (k5') forming

C3HD+, and (iv) radiative association (k6') forming C3H2D+. In analogy also the C3D

+ ions react with HD via (i) H atom abstraction (k7') leading to C3HD+, (ii) D atom ab-straction (k8') forming C3D2

+, and (iii) through radiative association (k9') leading to C3HD2

+. Since it is known that C3H2+ and C3H3

+ ions react with H2 very slowly, it has

APPENDIX D

109

been assumed that they do not react anymore. The same assumption has been made for their isotope equivalents C3HD+, C3D2

+, C3H2D+, and C3HD2

+. It has been shown ex-perimentally [sav04c] that H - D exchange in C3H3

+ + HD collisions is slower than 4 × 10-16 cm3s-1 and therefore negligible here.

Since in the model 10 free fitting parameters are in use, nine reaction rate coefficients and the probability p, it has been decided to reduce the number of free parameters by performing additional independent measurements. In order to get more insight into the reaction network, some of the intermediate products have been prepared externally and injected into the trap containing HD under similar conditions. In this way a variety of reaction rate coefficients have been obtained directly or with less competing processes. The results for the reaction system C3H

+ + HD are given in Table 2 and can be com-pared there with the reactions starting with C3

+. Since the fast reaction leading from C3+

to C3H+ and C3D

+ is followed by fast secondary reactions, careful experiments at very low HD number densities have been performed to determine p. The amount of HD used was so small that it was almost impossible to separate partial pressure from the back-ground gas, but high enough to see the formation of ions of interest. From these meas-urements the value p = 0.45 have been obtained.

It can be seen from Table 2 that corresponding reaction rate coefficients needed to de-scribe the C3

+ + HD network are different from those determined directly from C3H

+ + HD studies. Only reaction rate coefficients k3' and k8' for H atom abstraction of C3H

+ and D atom abstraction of C3D+, respectively, derived from the C3

+ + HD network are higher in comparison to those derived from C3H

+ + HD network.

It is interesting that the difference in reaction rate coefficients for radiative association of C3D

+ from C3+ + HD system is lower by factor of two from the value obtained in

C3H+ + HD system, while radiative association of C3H

+ in C3+ + HD is not observed at

all. The comparison of the results for these separate experiments leads to the conclusion that the intermediate products formed via C3

+ + HD reactions are formed with internal excitation and, therefore, react with different rate coefficients.

4. Discussion

4.1. Reactions of C3+

4.1.1. Abstraction Reactions Although reaction (1) has a high exothermicity of 197 kJ mol-1 [han89], it proceeds rather slowly at room temperature. The increase of the rate coefficient with decreasing temperature can be correlated to the necessity that the initially linear C3H

+ ions which lies 220 kJ/mol below the stable cyclic form [rag81] must be bent in order to break the H2 bound. In accordance with measurements from [han89] it is assumed that the pri-mary C3H

+ ions are only linear. In addition, it can be concluded from experimental [smi84] investigations of reaction (2) and reaction

C3H+ + H2 + He → C3H3

+ + He (6)

that various ways to prepare the ions always lead to ground state C3H+ ions.

In the reaction of C3+ with HD, it is rather certain that the C3H

+ products, formed via H atom abstraction, are excited. In Table 3. are summarized results of reaction rate coeffi-cients and C3H

+ branching ratio for these reactions at low and high temperatures. In this table are also given Langevin values for appropriate reactions. From room temperature measurements, the ratio of reaction rate coefficients for H2 and D2 is 1.39 [han89] and

APPENDIX D

110

1.38 [boh90], which can be attributed to the difference in the collision rate coefficients. Neglecting differences in reaction rate coefficients from data reported in [han89] and [boh90], efficiencies for H2 and D2 have been calculated, defined as fraction of the Langevin rate coefficient, i.e. as k/kL. The efficiencies calculated for H2 and D2 are the same. Taking data from [han89] efficiencies are 0.16 and from [boh90] 0.12. In [boh90], the ratio of reaction rate coefficients for reactions with H2 and D2, have been compared with predictions of AQO theory [su75], [su76], which gives the ratio of 1.39. From the efficiencies 0.11 and 0.12 for hydrogenation and deuteration, respectively, it has been concluded that no isotope effect is evident [boh90]. The same can be con-cluded from data reported in [han89].

4.1.2. Radiative Association Reactions It is mentioned above that C3H2

+ ions are partially formed via radiative association and partially via hydrogen abstraction. Assuming that given association reaction is ternary association with a second H2 partner, measured results of k2 = 8.5 × 10-11 cm3s-1 at low-est measured temperature of 15 K and highest used H2 number density will lead to the completely unreasonable ternary reaction rate of 1 × 10-20 cm6s-1. On similar way, at temperature of 72 K that is the highest temperature where association of C3

+ and H2 have been observed, will lead to the value of 1.2 × 10-21 cm6s-1. In case when HD was used instead of H2 similar value of 2.4 × 10-21 cm6s-1 have been obtained. These unusu-ally high values strongly support our assumption that association reactions of C3

+ with H2 or HD at conditions of experiment are radiative association reactions given by equa-tion (2) and indicated by k2' for H2 and HD respectively. As is already mentioned, tem-perature dependence of reaction (2) is given in Fig. 1. A summary for observed radiative association reactions of C3

+ with H2 and HD at 15 K temperature is given in Table 4. Inspection of this table leads to the conclusion that observed association reactions are rather fast at 15 K, reaching values of approximately 9 × 10-11 and 6 × 10-11 cm3s-1. It is interesting to note that their efficiencies are quite similar and they are 0.06 and 0.05 for H2 and HD respectively.

4.1.3. Model From the summary in Table 3 it can be seen that at room temperature, the total reaction rate coefficient significantly decreases when D2 is used instead of H2. The same is evi-dent at lowest measured temperatures. It is reasonable to expect the same for HD at 15 K. But, it is interesting that it has been found that at this temperature, total reaction rate coefficients for H2 and HD are equal. In contrast to this fact, Langevin rate coeffi-cients are temperature independent and the small differences are due to different re-duced mass. Their values are calculated simply the averaged polarizability and are pre-sented in Table 3. The contribution of the ion-dipole interaction for HD as target and the quadrupole term of H2 has been neglected in these calculations. On the other hand, since it has been concluded that isotope effect is not present, the dipole moment of HD may play a role in a slight favoring of formation of C3H

+ over C3D+ in contrast to the meas-

urement (0.55:0.45). Observed branching fraction is almost identical with unbiased probability to pick an H atom from HD, which is only half of that for H2, what is for expecting since total reaction rate coefficients are almost identical. It is also intriguing that at the lowest temperatures all reaction rate coefficients are higher than correspond-ing Langevin values.

Also can be seen that for reaction with H2 total efficiency at 15 K is 3.9 times higher than at room temperature. This means that a significant percentage of C3

+ + H2 decays back to the input channel at 300 K, implying that the transfer to the product channel may be hindered by presence of a potential barrier, a dynamical bottleneck or by a sig-

APPENDIX D

111

nificantly small phase space. The discussion can be simplified by making the assump-tion that the abstraction reaction (1) proceeds in two separated steps. As first, an inter-mediate collision complex, most probably long lived at low temperatures even if the H2 is only loosely attached to the C3

+ ion

C3+ + H2 → C3

+⋅H2 (1a)

An upper limit for the rate coefficient K1a has to be capture coefficient kL. Since at 44 K and below reaction rate coefficient is even larger than predicted kL, can be concluded that there is no substantial barrier which would have to be overcome for reaching this complex. This collision complex may be strongly bound since stable l-C3H2

+ and c-C3H2

+ ion exists. This collision complex can further redissociate into reactants

C3+⋅H2 → C3

+ + H2 (1b)

or find the way and proceed towards products via atom abstraction reaction

C3+⋅H2 → C3H

+ + H (1c)

or stabilized via photon emission

C3+ ⋅ H2 → C3H2

+ + hν (2a)

Three competing processes determine the dynamics of reaction system within this model, (1b), (1c), (2a) K1b, K1c and K2a. It is straightforward that then rate coefficient for reaction (1) can be approximated by

k1 = kLK1c / (K1b + K1c + K2a) (7)

Since have been shown that radiative association is extremely inefficient in comparison with atom abstraction, equation (7) can be simplified

k1 = kLK1c / (K1b + K1c) (8)

Inefficiency of radiative complex stabilization can be explained by assumption that ra-diative lifetime is longer then collision complex lifetime.

In Langevin model kL is rate coefficient at which the reactants can overcome the cen-trifugal barrier in the entrance channel to form the complex. Within this model kL is temperature independent. Then can be concluded that complex formation is not of im-portance for temperature dependence of reaction (1). Next important parameter in (8) is K1b. In this simple model it is rather safe to assume that lifetime of the complex in-creases with decreasing temperature and in accordance with this K1b have to decrease. Statistical models predict a Tn dependence. The most uncertain parameter is K1c. In case that tunneling play a dominant role K1c parameter will drop significantly with decreas-ing the temperature leading to contradiction to experimentally observed results.

In order to get a deeper inside into the collision dynamics it would be helpful to have reaction rate constants for reverse reactions. For some of reactions they are reported in [han89] but only at room temperature and therefore would be not discussed here.

4.2. Reactions of C3H+

Reactions (3) and (6) have been previously studied but mostly at room and temperature of 80 K [her83], [boh83], [rak83], [smi84], [smi87], [han89], [pro92] and [ger92]. In a SIFT studies at room temperature [smi84], [smi87], reaction channel (6) was dominant at 300 K and below while at 550 K reaction channel (3) prevails that led authors to the assumption of slight endothermicity (~ 4.18 kJ/mol) of reaction (3). In attempt to distin-guish between ternary and radiative association, at 80 K, fixed value of H2 number den-

APPENDIX D

112

sity of 5 × 1010 cm-3 and varying He number density between 5 × 1012 cm-3 and 1 × 1015 cm-3 experiments in ring-electrode ion trap have been performed [ger92]. In this experiment have been found that at low He density the remarkably favored reaction of C3H

+ is reaction (3) while at high number densities three body association (reaction 6) dominates. From one hand, this support the conclusion from [pro90] that formation of C3H2

+ is exothermic by 25.1 kJ/mol, but from other hand it is in contradiction with results from [smi84]. In addition in [han89] have been shown absence of significant increase of k3 with kinetic energy in the drift filed of the SIFDT (in range from 0.042 to 0.17 eV). In [ger92] has been shown how these discrepancies can be understood. Namely, with increasing He number density the effective rate coefficient for C3H2

+ formation decreases but this is simultaneously compensated by an increase in the rate coefficient of the C3H3

+ product. The sum of these two effective rate coefficients is al-most constant up to 1014 cm-3 of He number density.

Theoretical calculations predict slight endothermicity for reaction (3) of 7.3 kJ/mol [mal93] and 4 kJ/mol [won93]. It is necessary to note that these values are for formation of c-C3H2

+. For the same reaction in which is formed l-C3H2+ is predicted endothermic-

ity of 36 kJ/mol [won93]. Namely, it is well known that C3H2+ ions can have stable lin-

ear and cyclic forms. It is predicted that cyclic form of C3H2+ lies by 28 kJ/mol (at 0 K)

below linear form (or 31 kJ/mol at 298 K) and in [smi87] it has been shown that the lower energy isomer c-C3H2

+ is produced in reaction (3). Concerning C3H3+ products

from reaction (6) in [smi87] has been shown that at experiment conditions, both stable forms, linear and cyclic (cyclic form is more stable then linear for 142.26 kJ/mol [rad76]) are produced in almost identical amounts.

In order to explain competition between reaction (3) and (6) a simple two-complex model has been used in [ger92]. Namely, assuming that two distinct collision complexes (c-C3H

+⋅H2 and l-C3H+⋅H2) are formed and since in general cyclic structures have large

binding energies and linear are more weakly bound, significant lifetimes differences of the intermediates can be expected. Reactions from the long lived cyclic complex are dominant at low densities while from short lived complex at higher densities. It was also reasonable to assume that C3H2

+ products emerge only from long lived complex because is known that formation of l-C3H2

+ is endothermic. Using described model have been found that both rate coefficients for complex formation are similar but significantly smaller then Langevin value. It has been found unusually long lifetime of 50 µs for cy-clic complex and that both complexes are stabilized with high efficiency. From compe-tition between reaction and stabilization that lead to the long reaction time have been found indication for existence of barrier between c-C3H

+⋅H2 complex and c-C3H2+ prod-

ucts. From this model also appear that at low densities short lived l-C3H+⋅H2 complex

can be almost neglected while at high density both complexes contribute equally [ger92].

The most detailed study of reaction (3) and (4) has been performed by use of a variable temperature 22-pole trap cooled down to 10 K in [sor94]. They found that the radiative association channel (4) dominates at the lowest temperature. Using D2 instead of H2 significant isotope effects have been seen. For example, the rate coefficient for radiative association of C3H

+ with D2 is 50 times smaller than in collisions with H2. Ab initio minimum energy pathway calculations show that the C3H

+ and H2 can form directly the strongly bound propargyl ion which can subsequently undergo fast isomerization to the lower energy cyclic isomer c-C3H3

+ [mal93]. On one hand, calculations [mal93] predict minimum energy pathways which lead directly to the different isomers of collision complex and which directly connect c-C3H3 isomer with the c-C3H2 product. From the

APPENDIX D

113

other hand, to explain experimental results from [ger92] and from their work, [sor94] has introduced a multi-barrier model with potential barriers or dynamical bottlenecks, in the entrance and the exit channels. It has been assumed that a weakly bound entrance complex is formed which is connected via a barrier to the strongly bound intermediate. Competition between the slow transition towards a strongly bound intermediate and decay back to reactants leads to a rate coefficient for product formation that is a fraction of the Langevin value. In this way, dependence of lifetime of the weakly bound com-plex on total energy explains the temperature dependence of reaction and influence of rotational excitation in the case of normal hydrogen. A fraction of complexes, which have been converted to strongly bound complexes can then decay toward products, sta-bilize by association or redissociate. In this model fast isomerization between linear and cyclic intermediates also has been included by introducing a small barrier between them. Within this model in order to reproduce experimentally derived long complex and reaction lifetimes, the exit channel is hindered by a suitable mechanism.

In addition to very detailed study presented in [sor94], in this work are collected some additional interesting information. As first have to be shown that in this paper are re-ported reaction rate coefficients for radiative association (4) instead of ternary associa-tion like in majority of other experiments. Assuming equal stabilization efficiencies of He and H2, looking in number density dependence of effective reaction rate for reaction (6) presented in [ger92] from where can be seen that ternary association takes a domi-nant role above 1013 cm-3 of He and knowing that in present experiments have been used number densities of H2 up to 8.5 × 109 cm-3 can be concluded that observed asso-ciation of C3H

+ with H2 is due to radiative association. In addition, at lowest tempera-ture and highest used number density have been measured reaction rate coefficient of 6 × 10-11 cm3s-1 that will lead to the unreasonable ternary reaction rate of 7 × 10-21 cm6s-1. Therefore can be concluded that at used experimental conditions, asso-ciation of C3H

+ with H2 is described by radiative association reaction (4).

Results on reactions of C3H+ are summarized in Table 5. Concerning reaction (3) our

measured value k3 = 1.5 × 10-11 cm3s-1 at 300 K agree quite well with reported value of 2.6 × 10-11 cm3s-1 from [her83]. In [han89] and [pro92] are reported values of 6 × 10-12 cm3s-1 and 5.3 × 10-12 cm3s-1 respectively. Disagreement in factor of 2 can be also seen comparing estimated values at 80 K from present data of 5.6 × 10-11 cm3s-1 and value of 1.4 × 10-10 cm3s-1 reported in [her83]. From other hand, estimated value at this temperature is in excellent agreement with value of around 6 × 10-11 cm3s-1 [ger92] that have been measured at lowest He densities. In measured temperature interval, pre-sent data also agree quite well with temperature dependent measurements reported in [sor94] and presented on Fig. 3. by open circles, within maximal discrepancies of around 30 %. Discrepancies in results from this work, [ger92] and [sor94] from one side and [her83], [han89] [pro92] from the other side, should not be taken seriously: reason is evident, since in second group of references are used techniques where ternary association plays a dominant role.

The measured radiative association reaction rate coefficient k4 (in Fig.3. presented by solid diamonds) agrees well with reported data from [sor94] (in Fig.3. presented by open diamonds) in measured temperature interval, within error bars. Deviations of the present data from that reported in [sor94] as well as for reaction (3) are obviously due to incompletely thermalized C3H

+ ions in the present experiment, since they are formed through exothermic reaction (1). This also strongly indicates that data for fully the deuterated system measured in [sor94] may deviate from thermalized ones since C3D

+ is

APPENDIX D

114

formed in similar way. This fact also explaining discrepancies at 80 K, between value of 3.4 × 10-12 cm3s-1 [ger92] obtained from effective rate coefficient density dependence in low He pressure region, which is considerably higher than value can be estimated from present measurements.

Within this work are collected some new facts on reaction of C3H+ and C3D

+ with HD at 15 K temperature. At first, it is obvious that a very significant isotope effect exists. Sec-ond, from inspection of Table 5 it can be seen that reaction between C3H

+ and HD is far more effective reaction in comparison with others. Efficiency of this reaction is 0.44 at 15 K temperature, more than 3 times bigger than efficiency of reaction with H2. In addi-tion, this reaction is even 3.1 times more efficient than reaction with H2 at 10 K or 1.6 times more efficient than C3D

+ + D2 reaction. The dipole moment of HD may play a role in enhancing efficiency of reaction with C3H

+. Efficiency of reaction C3H+ with HD

is enhanced due to very efficient D atom abstraction or may be due to HD scrambling. From Table 5 it is obvious that H-D2 scrambling occurs in reaction C3H

+ with D2. This reaction channel is most efficient channel for this reaction. Then it is expected that scrambling occurs also for HD and H2, with may be significant effect in formation of C3HD+ and C3H2

+ respectively. If this assumption is correct, scrambling can signifi-cantly influence so called H and D atom abstraction reactions. Namely scrambling will lead to higher reaction rate coefficient for abstraction if one does not make difference between them. In reality, scrambling and abstraction reactions are two distinct proc-esses.

Assuming that efficiencies can not change much between 10 K and 15 K, it is interest-ing to mention that reaction of C3H

+ with D2 instead of H2 leads to a decrease in effi-ciency. The same holds for reaction of C3D

+ with D2 and HD, i.e. with HD efficiency is lower. Concerning radiative association reactions of C3H

+, with H2, HD and D2, channel efficiency decreases with the level of deuteration of neutral reactant.

Differences in collisions with H2 and HD can be interpreted as the result of reactions with two species at different energies since at 15 K only the lowest rotational state (j = 0) can be occupied in HD while in the present experiment we use normal hydrogen consisting of 75 % of o-H2 (j = 1) and 25 % of p-H2 (j = 1). In [sor94] in addition to experiments with n-H2 experiments with p-H2 are performed. It was shown that use of p-H2 leads to an increase of complex lifetime and therefore to a larger radiative rate coefficient that is in qualitative accordance with statistical calculations but also to a de-crease of hydrogen transfer reaction rate coefficients. This fact was been taken to sug-gest that hydrogen abstraction channel is slightly endothermic. Therefore, it must be mentioned that the present data fully supports conclusions made in [sor94]. Within their model, for reaction between C3H

+ and HD height of entrance barrier have to be very small.

5. Conclusions The present study belongs to a systematic series of measurements of chemical reactions of interstellar relevance, in which always three C atoms interact with hydrogen or deute-rium. The investigations have been performed in the variable temperature 22-pol ion trap (TV-22PT). The observed temperature dependences are tentatively explained on the basis of the energy dependence of the complex lifetime, barriers or bottle necks and rivalry between competing channels. For example the use of HD shows that abstraction of D atom is slightly preferred over H. In order to prove the explanations presented and to get a better understanding of the dynamics, detailed calculations are needed. The dy-

APPENDIX D

115

namical restrictions, e.g. also imposed by the nuclear spin statistics and symmetry selec-tion rules have not yet been discussed, since these problems have not yet been fully un-derstood for simpler systems such as the H3

+ + H2 system and isotopic variants [ger02] or reactions proceeding via a CHn

+ complex.

The results presented in this paper are of central importance for explaining processes in interstellar medium (ISM), for example isotope enrichment in cold dark molecular clouds which usually cannot be predicted from simple thermodynamics. One unsolved problem is the formation of deuterated c-C3H2. As discussed in a separate contribution [sav04c] the C3H3

+ + HD → C3H2D+ + H2 cannot lead to C3H2D

+ and, by dissociative recombination, to C3HD. The studies of this paper indicate that radiative association of C3H

+ and HD may be responsible, at least in part, for the observed C3HD / C3H2 abun-dance ratio.

Acknowledgments Financial support by the Deutsche Forschungsgemeinschaft (DFG) is gratefully ac-knowledged, especially via the Forschergruppe FOR 388 "Laboratory Astrophysics". We thank S. Schlemmer for many fruitful discussions and O. Asvany for some technical help with the TV-22PT.

APPENDIX D

116

References [ada84] Adams, N. G. and Smith, D.: A Study of the Reactions of NH3

+• and ND3+• with H2 and D2 at

Several Temperatures, Int. J. Mass Spectrom. Ion Proc., 61 (1984) 133-139.

[ada87] Adams, N. G., Smith, D.: On the Synthesis of c-C3H2 in Interstellar Clouds, ApJ 317 (1987) L25-L27.

[asv04] Asvany, O., Savić, I., Schlemmer, S. and Gerlich, D.: Variable Temperature Ion Trap Studies of CH4


[bar87] Barlow, S. E. and Dunn, G. H.: The Effects of Isotopic Substitution on Abstraction Reactions of Ammonia Ions with Hydrogen at Very Low Temperatures, Int. J. Mass Spectrom. Ion Proc., 80 (1987) 227-237.

[boh83] Bohme, D. K., Raksit, A. B., and Fox, A.: Carbene Chemistry of Cations: The Chemistry of C3H

+ in the Gas Phase, J. Am. Chem. Soc., 105 (1983) 5481-5483.

[boh85] Böhringer, H.: Third-Body-Assisted, Binary Ion-Molecule Reactions. The Reaction NH3

+ + H2 → NH4+ + H, Chem. Phys. Lett., 122 (1985) 185-189.

[boh90] Bohme, D. K. and Wlodek, S.: Hydrogenation of Carbon-Cluster Cations with Molecular Hy-drogen: Implications for the Growth of Carbon-Cluster Molecules, Int. J. Mass Spectrom. Ion Proc. 102 (1990) 133-149.

[fai87] Faibnis, A., Kanter, E. P., Tack, L. M., Bakke, E. and Zabransky, B. J.: Geometrical Structure of C3

+, J. Phys. Chem. 91 (1987) 6445-6447.

[feh75] Fehsenfeld, F. C., Lindinger, W., Schmeltekopf, A. L., Albritton, D. L., and Ferguson, E. E.: Energy dependence of the reaction NH3

+ + H2 → NH4+ + H, J. Chem. Phys., 62 (1975) 2001-

2003.

[fur02]Fura, A., Tureček, F. and McLafferty, F. W: Small Carbon Clusters (Cn0, Cn

+, Cn-) from Acyclic

and Cyclic Precursors: Neutralization-Reionization and Theory, Int. J. Mass Spectrom. 217 (2002) 81-96.

[ger92] Gerlich, D. and Horning, S.: Experimental Investigations of Radiative Association Processes as related to Interstellar Chemistry, Chem. Rev. 92 (1992) 1509-1539.

[ger92b] Gerlich, D.: Inhomogeneous rf Fields: A Versatile Tool for the Study of Processes with Slow Ions, in: Ng, C.-Y. and Baer, M. (eds): State-Selected and State-to-State Ion-Molecule Reaction Dynamics, Part 1: Experiment, Adv. in Chem. Phys. LXXXII (1992) 1-176.

[ger93] Gerlich, D.: Experimental Investigations of Ion-Molecule Reactions Relevant to Interstellar Chemistry, J. Chem. Soc. Faraday Trans., 89 (1993) 2199-2208.

[ger95] Gerlich, D.: Ion-Neutral Collisions in a 22-Pole Trap at Very Low Energies, Physica Scripta T59 (1995) 256-263.

[gre90] Grev, R. S., Alberts, I. L. and Schaefer, H. F. III: C3+ is Bent, J. Phys. Chem. 94 (1990) 3379-

3381.

[gre90a] Grev, R. S., Alberts, I. L. and Schaefer, H. F. III: Additions and Corrections: C3+ is Bent, J.

Phys. Chem. 94 (1990) 8744-8744.

[han89] Hansel, A., Richter, R., Lindinger, W. and Ferguson, E. E.: Reactions of C2 and C3 Hydrocarbon Ions with H, D, H2 and D2 at Near-Thermal Energies, Int. J. Mass Spectrom. Ion Proc. 94 (1989) 251-260.

[haw92] Hawley, M. and Smith, M. A.: The Gas Phase Reaction of C2H2+ and H2 below 3 K: The Re-

opening of the Bimolecular C2H3+ Channel at Low Energy, J. Chem. Phys. 96 (1992) 1121-1127.

[her83] Herbst, E., Adams, N. G. and Smith, D.: Laboratory Measurements of Ion-Molecule Reactions Pertaining to Interstellar Hydrocarbon Synthesis, ApJ 269 (1983) 329-333.

[her91] Herbst, E., DeFrees, D. J., Talbi, D., Pauzat, F, Koch, W.and McLean, A.D.: Calculations on the Rate of the Ion-Molecule Reaction Between NH3

+ and H2, J. Chem. Phys., 94 (1991) 7842-7849.

APPENDIX D

117

[kim75] Kim, J. K., Theard, L. P. and Huntress, Jr., W. T.: ICR studies of Some Hydrogen Atom Abstrac-tion Reactions: X+ + H2 → XH+ + H*, J. Chem. Phys., 62 (1975) 45-52.

[lif93] Lifshitz, C., Sandler, P., Grützmacher, H. F., Sun, J., Weiske, T. and Schwarz, H.: Application of Thermal Kinetics to Small Carbon Ion Clusters, J. Phys. Chem. 97 (1993) 6592-6597.

[lui85] Luine, J. A. and Dunn, G. H.: Ion-Molecule Reaction Probabilities Near 10 K, ApJ 299 (1985) L67-L70.

[mal93] Maluendes, S. A., McLean, A. D., Yamashita, K., and Herbst, E.: Calculations on the Competi-tion between Association and Reaction for C3H

+ + H2, J. Chem. Phys. 99 (1993) 2812-2820.

[mar95] Martin, J. M. L., El-Yazal, J. and François, J. P.: Structure and Vibrational Spectra of Carbon Clusters Cn (n = 2 - 10, 12, 14, 16, 18)Using Density Functional Theory Including Exact Ex-change Contributions, Chem. Phys. Lett. 242 (1995) 570-579.

[mce87] McElvany, S. W., Dunlap, B. I. and O'Keefe, A.: Ion Molecule Reactions of Carbon Cluster Ions with D2 and O2, J. Chem. Phys. 86 (1987) 715-725.

[par89] Parent, D. C. and McElvany, S. W.: Investigations of Small Carbon Cluster Ion Structures by Reactions with HCN, J. Am. Chem. Soc., 111 (1989) 2393-2401.

[pro90] Prodnuk, S. D., DePuy, C. H. and Bierbaum, V. M.: The Heats of Formation of Cyclic and Lin-ear C3H2

+, Int. J. Mass Spectrom. Ion Proc., 100 (1990) 693-699.

[pro92] Prodnuk, S. D., DePuy, C. H. and Bierbaum, V. M.: Gas-Phase Reactions of C3H+ Ions, Organic

Mass Spectrom., 27 (1992) 416-422.

[rad76] Radom, L., Hariharan, P. C., Pople, J. A. and Schleyer, P. v. R.: Molecular Orbital Theory of the Electronic Structure of Organic Compounds. XXII. Structures and Stabilities of C3H3

+ and C3H+

Cations, J. Am. Chem. Soc., 98 (1976) 10-14.

[rag81] Raghavachari, K., Whiteside, R. A., Pople, J. A. and Schleyer, P. v. R.: Molecular Orbital The-ory of the Electronic Structure of Organic Molecules. 40. Structures and Energies of C1-C3 Car-bocations, Including Effects of Electron Correlation, J. Am. Chem. Soc., 103 (1981) 5649-5657.

[rak83] Raksit, A. B. and Bohme, D. K.: Studies of Reactions of C3H+ Ions in the Gas Phase at

296 ± 2 K, Int. J. Mass Spectrom. Ion Proc., 55 (1983) 69-82.

[sav04c] Savić, I., Schlemmer, S. and Gerlich, D.: Low-Temperature Experiments on the Formation of Deuterated C3H3

+, submitted to ApJ. (2004).

[smi81] Smith, D., and Adams, N. G.: Some Positive Ion Reactions with H2: Interstellar Implications, Mon. Not. R. Astron. Soc., 197 (1981) 377-384.

[smi84] Smith, D. Adams, N. G., Ferguson, E.E.: The Heat of Formation of C3H2+, Int. J. Mass Spec-

trom. Ion Proc., 61 (1984) 15-19.

[smi87] Smith, D. and Adams, N. G.: Cyclic and Linear Isomers of C3H2+ and C3H3

+: the C3H+ + H2

Reaction, Int. J. Mass Spectrom. Ion Proc., 76 (1987) 307-317.

[sor94] Sorgenfrei, A.: PhD thesis Freiburg 1994., Sorgenfrei, A. and Gerlich, D.: Ion-Trap Experiments on C3H

+ + H2: Radiative Association vs. Hydrogen Abstraction, in Nenner, I. (ed.), Molecules and Grains in Space, AIP Press, New York (1994) 505-514.

[su75] Su, T. and Bowers, M. T: Ion-Polar Molecular Collisions: The Average Quadrupole Orientation Theory, Int. J. Mass Spectrom. Ion Phys. 17 (1975) 309-319.

[su76] Su, T. and Bowers, M. T: Erratum: Ion-Polar Molecular Collisions: The Average Quadrupole Orientation Theory, Int. J. Mass Spectrom. Ion Phys. 21 (1976) 424-424.

[tay91] Taylor, P. R., Martin, J. M. L., François, J. P. and Gijbels, R.: Ab Initio Study of the C3+ Cation

Using Multireference Methods, J. Phys. Chem. 95 (1991) 6530-6534.

[teu00] Le Teuff, Y.H., Millar, T.J., and Markwick, A.J., The UMIST Database for Astrochemistry 1999, A&AS. 146 (2000) 157-168.

[wel89] Weltner, W. Jr. and Van Zee, R. J.: Carbon Molecules, Ions, and Clusters, Chem. Rev. 89 (1989) 1713-1747.

APPENDIX D

118

[won93] Wong, M. W. and Radom, L.: Thermochemistry and Ion-Molecule Reactions of Isomeric C3H2•+

Cations, J. Am. Chem. Soc., 115 (1993) 1507-1514.

Tables

Table 1. Temperature dependence of rate coefficients, k = α (T / 300 K)β.

Reactions α / cm3 s-1 β Temperature range / K

C3+ + H2 → C3H

+ + H 1.7 × 10-9 (4.7 ± 0.4) × 10-10

0 -0.691 ± 0.006

15 - 44 44 - 300

C3+ + H2 → C3H2

+ + hν (5.7 ± 0.7) × 10-12 -1.02 ± 0.05 20 - 44 C3H

+ + H2 → C3H2+ + H (1.5 ± 0.1) × 10-10 *

(1.4 ± 0.2) × 10-11 * 0.09 ± 0.04 * -1.05 ± 0.08 *

15 - 44 44 - 300

C3H+ + H2 → C3H3

+ + hν (2.3 ± 1) × 10-12 * -1.1 ± 0.2 * 15 - 44 * These values deviate from results for thermalized ions. For details see the text.

Table 2. Measured reaction rate coefficients for C3+ + HD and C3H

+ + HD system at temperature of 15 K.

ki' / cm3s-1 i Reaction C3

+ + HD system C3H+ + HD system

1 C3+ + HD → C3H

+ + D → C3D

+ + H 1.7 × 10-9

2 C3+ + HD → C3HD+ + hν 5.9 × 10-11

3 C3H+ + HD → C3H2

+ + D 9.0 × 10-12 3.0 × 10-12 4 C3H

+ + HD → C3D+ + H2 4.0 × 10-11 5.6 × 10-11

5 C3H+ + HD → C3HD+ + H 2.7 × 10-10 4.6 × 10-10

6 C3H+ + HD → C3H2D

+ + hν - 3.2 × 10-11 7 C3D

+ + HD → C3HD+ + D 6.0 × 10-11 1.0 × 10-10 8 C3D

+ + HD → C3D2+ + H 2.3 × 10-10 8.3 × 10-11

9 C3D+ + HD → C3HD2

+ + hν 4.1 × 10-12 8.0 × 10-12

Table 3. Reaction rate coefficients (in units of cm3s-1) for H(D) atom abstraction reac-tion of C3

+ with H2, HD and D2 and branching fraction. kL denotes the Langevin reac-tion rate coefficient.

H2 HD D2 T (K)

k1 k1 C3H+ fraction k1

Ref.

10 1.3 × 10-9 [sor94] 15 1.7 × 10-9 * 1.7 × 10-9 * 0.45 % this work 296 1.7 × 10-10 ** 1.24 × 10-10 ** [boh90] 298 3.0 × 10-10 [her83] 300 4.6 × 10-10 this work

2.5 × 10-10 1.8 × 10-10 [han89] 1.5 × 10-10 [mce87]

kL 1.52 × 10-9 1.26 × 10-9 1.10 × 10-9 * Inclusion of values for radiative association will lead to effective reaction rate coefficients of 1.79 × 10-9 cm3s-1 and 1.76 × 10-9 cm3s-1 for H2 and HD respectively. ** Originally reported data are corrected for ternary association. Original values are 1.8 × 10-10 cm3s-1and 1.3 × 10-10 cm3s-1 for H2 and HD respectively

APPENDIX D

119

Table 4. Summary for observed radiative association reactions of C3+ with H2 and HD

(in units of cm3s-1) at 15 K temperature.

H2 HD

k2 k2' 8.5 × 10-11 5.9 × 10-11

kL 1.52 × 10-9 1.26 × 10-9

Table 5. Reaction rate coefficients (in units of cm3s-1) for of C3H+ and C3D

+ with H2, HD and D2, and efficiencies f for reaction channels. Efficiency for reaction can be easily calculated like Σki / kL. Results from [her83], [rak83], [smi84], [smi87], [mce87] [han89] and [pro92] are not included in this table since in this references are used tech-niques where ternary association plays a dominant role.

Reaction Channel kL / 10-9 T (K) k Ref. C3H

+ + H2 → C3H2+ + H 10 5 × 10-11 [sor94]

15 1.02 × 10-10 [sor94] 15 1.15 × 10-10 * this work 300 1.5 × 10-11 * this work

→ C3H3+ + hν 10 2 × 10-10 [sor94]

15 8.93 × 10-11 [sor94] 15 6 × 10-11 * this work

1.52

80 3.4 × 10-12 [ger92] C3H

+ + HD → C3H2+ + D 15 3.0 × 10-12 this work

→ C3HD+ + H 4.6 × 10-10 this work → C3D

+ + H2 5.6 × 10-11 this work → C3H2D

+ + hν

1.26

3.2 × 10-11 this work C3H

+ + D2 → C3HD+ + D 10 1 × 10-11 [sor94] → C3D2

+ + H 10 2.7 × 10-11 [sor94] → C3HD2

+ + hν

1.10

4 × 10-12 [sor94] C3D

+ + HD → C3HD+ + D 15 1 × 10-10 this work → C3D2

+ + H 8.3 × 10-11 this work → C3HD2

+ + hν

1.26

8 × 10-12 this work C3D

+ + D2 → C3D2+ + D 10 1.7 × 10-10 * [sor94]

→ C3D3+ + hν 1.10

1.3 × 10-10 * [sor94] *These values might slightly deviate from thermalized ones. For details see the text.

APPENDIX D

120

Figures

0.0 0.1 0.2 0.30.1

1

10

100

C3

+ + H2 →

C3H

3

+

C3H

2

+

Σ

C3H+

C3

+

t / s

Ni

Fig. 1. Typical experimental results showing the time dependence of the averaged num-ber of trapped primary and product ions, Ni as a function of storage time t. By collisions with an intense pulse of He buffer gas, the injected C3

+ primary ions are relaxed to the ambient temperature T = 15 K. They react with H2 (number density 8.5 × 109 cm-3) via several reactions (1) - (5) and form C3H

+, C3H2+ and C3H3

+. The solid lines are solutions of an adequate reaction rate equation system. The obtained rate coefficients can be found in Table 1.

APPENDIX D

121

10 1001E-12

1E-11

1E-10

1E-9

C3H

2

+ + hν

C3+ + H2 →

k / c

m3 s

-1

T / K

C3H+ + H

Fig. 2. Temperature dependence of measured rate coefficients for the reactions C3

+ + H2 → C3H+ + H (filled circles) and C3

+ + H2 → C3H2+ + hν (stars). The solid lines

are fits of the data using the function k = α ( T / 300 )β with different parameters de-scribing the low temperature (T < 50 K) and the high temperature dependence. At 300 K additional experimental points for C3H

+ formation are included (taken from the literature: triangle: [her83], square: [han89], triangle pointing down [boh90]).

APPENDIX D

122

10 1001E-12

1E-11

1E-10

k / c

m3 s

-1

T / K

C3H+ + H

2 → C

3H

2

+ + H

10 100

n - H2

p - H2

n - H2

p - H2

C3H+ + H

2 →

C3H

3

+ + hν

T / K Fig. 3. Temperature dependence of the reaction rate coefficients for the reactions C3H

+ + H2 → C3H2+ + H (filled circles) and C3H

+ + H2 → C3H3+ + hν ( solid diamonds).

Literature values are from [sor94] (open symbols). The solid lines are fits to the data using the function k = α ( T / 300 )β for the temperature intervals given in Table 1. The dashed lines are just for guiding the eyes. The differences measured for normal-hydrogen and para-hydrogen indicate that at low total energies, the lifetime of the colli-sion complex is finally longer than the time needed for emitting a photon.

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.20.1

1

10

100

C3

+ + HD →

C3HD

2

+

Σ

C3D

2

+, C3H

2D+

C3HD+, C

3H

3

+

C3D+, C

3H

2

+

C3H+

C3

+

t / s

Ni

Fig. 4. Reactions of C3+ with HD (7.9 × 109 cm-3) at T = 15 K (see also Fig. 1). In order

to explain the temporal changes of the indicated products a complex chemical model has been developed. The solid lines are the solutions of the corresponding rate equation system. The rate coefficients are included in Table 2 or mentioned in the text.

APPENDIX D

123

C3D+C3H+

k3'C3HD+

C3D2+

k8'

C3H2D+C3HD2

+C3+

k1', p

k1', 1 - p

k2'

k4'

k5'k6'

k7'

C3H2+

m = 36 m = 37 m = 38 m = 39 m = 40 m = 41

k9'

Fig. 5. Schematic illustration of the reactions used to model the chemical evolution of C3

+ ions in HD.

124

125

REFERENCES

[ada03] Ádámkovics, M., Blake, G. A. and McCall, B. J.: Observations of Rotationally Resolved C3 in Translucent Sight Lines, Astrophys. J. 595 (2003) 235-246.

[aik99] Aikawa, Y. and Herbst, E.: Deuterium Fractionation in Protoplanetary Disks, Astrophys. J. 526 (1999) 314-326.

[all92] Allamandola, L. J., Sandford, S. A., Tielens, A. G. G. M. and Herbst, T. M.: Infrared Spectros-copy of Dense Clouds in the C-H Stretch Region: Methanol and "Diamonds", Astrophys. J. 399 (1992) 134-146.

[ani03] Anicich, V. G.: An Index of the Literature for Bimolecular Gas Phase Cation-Molecule Reaction Kinetics, JPL Publication 03-19, National Aeronautics and Space Administration, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California, (2003) 1-1172.

[asv04a] Asvany, O., Savić, I., Schlemmer, S. and Gerlich, D.: Variable Temperature Ion Trap Studies of CH4


[asv04b] Asvany, O., Schlemmer, S. and Gerlich, D.: Deuteration of CHn+ (n=3-5) in Collisions with HD

Measured in a Low Temperature Ion Trap, submitted to Astrophys. J. (2004).

[bah69] Bahr, R. E.: Untersuchung der Eigenschaften Eines Hochfrequenz-Ionen-Speichers, Diplomar-beit, Physikalisches Institut der Universität Freiburg / Brsg. (1969) 1-57.

[bak96] Bakker, E. J., Waters, L. B. F. M., Lamers, H. J. G. L. M., Trams, N. R. and Van der Wolf, F. L- A.: Detection of C2, CN and NaI D Absorption in the AGB Remnant of HD 56126, Astron. Astro-phys. 310 (1996) 893-907.

[bak97] Bakker, E. J., van Dishoeck, E. F., Waters, L. B. F. M., and Schoenmaker, T.: Circumstellar C2, CN and CH+ in the Optical Spectra of Post-AGB Stars, Astron. Astrophys. 323 (1997) 469-487.

[bak98a] Bakker, E. J., and Lambert, D. L.: The Circumstellar Shell of the Post-Asymptotic Giant Branch Star HD 56126: 12C12C/12C13C Isotope Ratio and 12C16O Column Density, Astrophys. J. 508 (1998) 387-396.

[bak98b] Bakes, E. L. O. and Tielens, A. G. G. M.: The Effects of Polycyclic Aromatic Hydrocarbons on the Chemistry of Photodissociation Regions, Astrophys. J. 499 (1998) 258-266.

[ben03] Bensch, F., Leuenhagen, U., Stutzki, J. and Schieder, R.: [CI] 492 GHz Mapping Observations of the High-Latitude Translucent Cloud MCLD 123.5+24.9, Astrophys. J. 591 (2003) 1013-1024.

[ber89] Bernath, P. F., Hinkle, K. H. and Keady, J. J.: Detection on C5 in the Circumstellar Shell of IRC+10216, Science 244 (1989) 562-564.

[bla73] Black, J.H. and Dalgarno, A.: The Cosmic Abundance of Deuterium, Astrophys. J. 184 (1973) L101-L104.

[böh83] Böhringer, H. and Arnold, F.: Studies of Ion Molecule Reactions, Ion Mobilities and their Tem-perature Dependence to Very Low Temperatures Using a Liquid Helium Cooled Iion Drift Tube, Int. J. Mass Spectrom. Ion Processes 49 (1983) 61-83.

[boh92] Bohme, D. K.: PAH and Fullerene Ions and Ion/Molecule Reactions in Interstellar and Circum-stellar Chemistry, Chem. Rev. 92 (1992) 1487-1508.

[bol00] Bolatto, A. D., Jackson, J. M. and Kraemer, K. E.: First Detection of Submillimeter [CI] Emission in the Small Magellanic Cloud, Astrophys. J. 541 (2000) L17-L20.

[bro97] Brownsword, R. A., Sims, I. R., Smith I. W. M., Stewart, W. A., Canosa, A. and Rowe, B. R.: The Radiative Association of CH with H2: A Mechanism for formation of CH3 in Interstellar Clouds, Astrophys. J. 485 (1997) 195-202.

[bun80] Bundy, F. P.: The P, T Phase and Reaction Diagram for Elemental Carbon, 1979, J. Geophys. Res. 85 (1980) 6930-6936.

[bun96] Bundy, F. P., Bassett, W. A., Weathers M. S., Hemley, R. J., Mao, H. K. and Goncharov, A. F.: The Pressure-Temperature Phase and Transformation Diagram for Carbon; Updated Through 1994, Carbon. 34 (1996) 141-153.

REFERENCES

126

[bur64] Burns, R. P., Jason, A. J., and Inghram, M. G.: Evaporation Coefficient of Graphite, J. Chem. Phys. 40 (1964) 1161-1162.

[but92] Büttgenbach, T. H., Keene, J., Phillips, T. G. and Walker, C. K.: Submillimeter Detection of Extragalactic CI Emission: IC 342, Astrophys. J. 397 (1992) L15-L17.

[cec98] Ceccarelli, C., Castets, A., Loinard, L., Caux, E. and Tielens, A. G. G. M.: Detection of Doubly Deuterated Formaldehyde Towards the Low-Luminosity Protostar IRAS 16293-2422, Astron. As-trophys. 338 (1998) L43-L46.

[cer00] Cernicharo, J., Goicoechea, J. R. and Caux, E.: Far-Infrared Detection of C3 in Sagittarius B2 and IRC +102016, Astrophys. J. 534 (2000) L199-L202.

[cer02] Čermák, I., Savić, I. and Gerlich, D.: Ion-Trapping Apparatus for Studies on Reactions between Ions and Neutral Carbon Species, in: Safrankova, J. (ed.): WDS’02, Proceedings of Contribution Papers, Part II, Physics of Plasma and Ionized Media, MATFYZPRESS, Prague (2002) 281-287.

[cer98] Čermák, I., Förderer, M., Čermákova, I., Kalhofer, S., Stopka-Ebeler, H., Monninger, G. and Krätschmer, W.: Laser Induced Emission Spectroscopy of Matrix-Isolated Carbon Molecules: Ex-perimental Setup and New Results on C3, J. Chem. Phys. 24 108 (1998) 10129-10142.

[cer99] Čermák, I. and Gerlich, D.: Growth Dynamics of Carbon-Containing Nanostructures, in: Gerlich, D. and Henning, T. (eds.): Laboratory Astrophysics, Structure, Dynamics and Properties of Mole-cules and Grains in Space, DFG Proposal for a Forschergruppe, TU Chemnitz, FSU Jena, (1999) 193-221.

[cha00] Chastaing, D., Le Picard, S. D. and Sims, I. R.: Direct Kinetic Measurements on Reaction of Atomic Carbon, C(3P), with O2 and NO at Temperatures down to 15 K, J. Chem. Phys. 112 (2000), 8466-8469.

[cha01] Chastaing, D., Le Picard, S. D., Sims, I. R. and Smith, I. W. M.: Rate Coefficients for the Reac-tions of C(3Pj) Atoms with C2H2, C2H4, CH3C≡CH and H2C=C=CH2 at Temperatures down to 15 K, Astron. Astrophys. 365 (2001) 241-247.

[cha78] Chaffee, F. H. Jr. and Lutz, B. L.: The Detection of Interstellar Diatomic Carbon Toward ζ Ophiuchi, Astrophys. J. 221 (1978) L91-L93.

[cha80] Chaffee, F. H., Lutz, B. L., Black, J. L., van den Bout, P. A. and Snell, R. L.: Rotational Fine-Structure of Interstellar C2 Toward Zeta Persei, Astrophys. J. 236 (1980) 474-480.

[cha98] Chastaing, D., James, P. L., Sims, I. R. and Smith I. W. M.: Neutral-Neutral Reactions at Tem-peratures of Interstellar Clouds. Rate Coefficients for the Reactions of C2H Radicals with O2, C2H2, C2H4 and C3H6 down to 15 K, Faraday Disscus. 109 (1998) 165-181.

[cla01] Clayton, D. D., Deneault, E. A.-N. and Meyer, B. S.: Condensation of Carbon in Radioactive Supernova Gas, Astrophys. J. 562 (2001) 480-493.

[cla03] Clayton, G. C., Wolff, M. J., Sofia, U. J., Gordon, K. D. and Misselt, K. A.: Dust Grain Size Dis-tribution From MRN to MEM, Astrophys. J. 588 (2003) 871-880.

[cle82] Clegg, R. E. S. and Lambert, D. L.: On C3 Molecules in Diffuse Interstellar Clouds, Mon. Not. R. Astron. Soc. 201 (1982) 723-733.

[cod01] Codella, C., Bachiller, R., Benedettini, M., Caselli, P., Giannini, T., Nisini, B., Saraceno, P., Spinoglio, L. and Walmsley, C. M.: Waiting for FIRST: The Evolution of Molecular Outflows, in: Pilbratt, G. L., Heras, A. M., Prusti, T. and Harris, R. (eds.): Proc. Symposium The Promise of the Herschel Space Observatory, ESA SP-460, (2001) 381-384.

[cov00] Covatto, C. and Aannestad, P. A.: Sputtering in the Outflows of Cool Stars, Mon. Not. R. Astron. Soc. 318 (2000) 67-72.

[cra00] Crawford, I. A. and Barlow, M. J.: Ultra-High-Resolution Observations of Circumstellar KI and C2 Around the post-AGB star HD 56126, Mon. Not. R. Astron. Soc. 311 (2000) 370-376.

[cra85] Crawford, M. K., Genzel, R., Townes, H. and Watson, D. M.: Far-Infrared Spectroscopy of Gal-axies: the 158 Micron C+ Line and the Energy Balance of Molecular Clouds, Astrophys. J. 291 (1985) 755-771.

REFERENCES

127

[dal60] Daly, N. R.: Scintillation Type Mass Spectrometer Ion Detector, Rev. of Scientific Instruments 31 (1960) 264-267.

[dan83] Danks, A. C., and Lambert, D. L.: Interstellar C2 in the Ophiuchus Clouds, Astron. Astrophys. 124 (1983) 188-196.

[dec03] Decker, S.: Temperaturbestimmung von elektrisch- und lasergeheizten Kohlenstoff-Teilchen, Diplomarbeit, TU Chemnitz, Institut für Physik (2003) 1-93.

[del00] Deleuze, M. S., Giuffreda, M. G., François, J.-P. and Cederbaum, L. S.: Valence One-Electron and Shake-Up Ionization Bands of Carbon Clusters. II. The Cn (n = 4, 6, 8, 10) Rings, J. Chem. Phys. 112 (2000), 5325-5338.

[del99] Deleuze, M. S., Giuffreda, M. G., François, J.-P. and Cederbaum, L. S.: Valence One-Electron and Shake-Up Ionization Bands of Carbon Clusters. I. The Cn (n = 3, 5, 7, 9) Chains, J. Chem. Phys. 111 (1999), 5851-5865.

[deu00] Deutsch, H., Becker, K., Matt, S. and Märk, T. D.: Theoretical Determination of Absolute Elec-tron-Impact Ionization Cross Sections of Molecules, Int. J. Mass Spectrom. 197 (2000) 37-69.

[dis84] Van Dishoeck, E. and de Zeeuw, T.: Observation of Interstellar C2 toward χ Oph, HD 154368, 147889 and 149404, Mon. Not. R. Astron. Soc. 206 (1984) 383-406.

[dis86] Van Dishoeck, E. and Black, J. H.: Detection of the (3,0) Phillips Band of Interstellar C2 Toward Zeta Ophiuchi, Astrophys. J. 307 (1986) 332-336.

[dis88] Van Dishoeck, E. and Black, J. H.: The Photodissociation and Chemistry of Interstellar CO, As-trophys. J. 334 (1988) 771-802.

[dis98] Van Dishoeck, E. and Blake, G. A.: Chemical Evolution of Star-Forming Regions, Annu. Rev. Astron. Astrophys. 36 (1998) 317-368.

[dis99] Van Dishoeck, E. and Hogerheijde, M. R.: Models and Observations of the Chemistry Near Young Stellar Objects, in: Lada, C. J. and Kylafis, N.D. (eds): The Origin of Stars and Planetary Systems, Kluwer, Dordrecht (1999) 97-140.

[dro59] Drowart, J., Burns, R. P., DeMaria, G. and Inghram, M. G.: Mass Spectrometric Study of Carbon Vapor, J. Chem. Phys. 31 (1959) 1131-1132.

[dul01a] Duley, W. W.: The 3.3 Micron and Extended Red Emissions in Interstellar Clouds: Further Evidence for Carbon Nanoparticles, Astrophys. J. 553 (2001) 575-580.

[dul01b] Duley, W. W. and Grishko, V. I.: Evolution of Carbon Dust in Aromatic Infrared Emission Sources: Formation of Nanodiamonds, Astrophys. J. 554 (2001) L209-L212.

[fre82] Freed, K. F., Oka, T. and Suzuki, H.: On the n-Dependence of the Reaction Rate for C+ + Cn → Cn+1

+ in Interstellar Space, Astrophys. J. 263 (1982) 718-722.

[fre89] Frerking, M. A., Keene, J., Blake, A. and Phillips, T. G.: The Abundances of Atomic Carbon and Carbon Monoxide Compared With Visual Extinction in the Ophiuchus Molecular Cloud Complex, Astrophys. J. 344 (1989) 311-319.

[gal01] Galazutdinov, G. A., Musaev, F. A. and Krelowski, J.: On the Detection of the Linear C5 Mole-cule in the Interstellar Medium, Mon. Not. R. Astron. Soc. 325 (2001) 1332-1334.

[gal02] Galazutdinov, G. A., Petlewski, A., Musaev, F. A., Moutou, C., Curto, G. Lo and Krelowski, J.: An Upper Limit to the Interstellar C5 Abundance in Translucent Clouds, Astron. Astrophys. 395 (2002) 223-226.

[gar98] Garay, G., Gomez, Y., Lizano, S. and Brown, R. L.: VLA Observations of Carbon Radio Recom-bination Lines Toward the HII Region Complex S88B, Astrophys. J. 501 (1998) 699-709.

[ger02] Gerlich, D., and Schlemmer, S.: Deuterium Fractionation in Gas-Phase Reactions Measured in the Laboratory, Planetary and Space Science 50 (2002) 1287-1297.

[ger03a] Gerin, M., Fosse, D. and Roueff, E.: Carbon Chemistry in Interstellar Clouds, in: Curry, C. L. and Fich, M. (eds): Proceedings of the Conference "Chemistry as a Diagnostic of Star Formation" University of Waterloo, August 2002, NRC Research Press (2003); http://arxiv.org/PS_cache/astro-ph/pdf/0212/0212058.pdf.

REFERENCES

128

[ger03b] Gerlich, D.: Rf Ion Guides, in: Armentrout, P. B. (ed.): The Encyclopedia of Mass Spectrometry, Vol. 1, Elsevier Ltd., (2003) 182 - 194.

[ger03c] Gerlich, D.: Molecular Ions and Nanoparticles in RF and AC traps, Hyperfine Interactions 146/147 (2003) 293-306.

[ger04] Gerlich, D.: Applications of rf Fields and Collision Dynamics in Atomic Mass Spectrometry, J. Anal. At. Spectrom., 19 (2004) 581 - 590

[ger89] Gerlich, D. and Kaefer, G.: Ion Trap Studies of Association Processes in Collisions of CH3+ and

CD3+ with n-H2, p-H2, D2 and He at 80 K, Astrophys. J. 347 (1989) 849-854.

[ger90] Gerlich, D.: Ortho Para Transitions in Reactive H+ + H2 Collisions, J. Chem. Phys. 92 (1990) 2377-2388.

[ger92a] Gerlich, D. and Horning, S.: Expermental Investigations of Radiative Association Processes as Related to Interstellar Chemistry, Chem. Rev. 92 (1992) 1509-1539.

[ger92b] Gerlich, D.: Inhomogeneous rf Fields: a Versatile Tool for the Study of Processes with Slow Ions, in: Ng, C.-Y. and Baer, M. (eds): State-Selected and State-to-State Ion-Molecule Reaction Dynamics, Part 1: Experiment, Adv. in Chem. Phys. LXXXII (1992), 1-176.

[ger93] Gerlich, D.: Experimental Investigations of Ion-Molecule Reactions Relevant to Interstellar Chemistry, J. Chem. Soc. Faraday Trans. 89 (1993) 2199-2208.

[ger95] Gerlich, D.: Ion-Neutral Collisions in a 22-Pole Trap at Very Low Energies, Physica Scripta, T59 (1995) 256-263.

[ger99a] Gerlich, D. and Henning, T.: Laboratory Astrophysics; Structure, Dynamics and Properties of Molecules and Grains in Space, DFG Proposal for a Forschergruppe, (1999) 1-299.

[ger99b] Gerin, M. and Phillips, T. G.: Atomic Carbon in Galaxies, in: Ossenkopf, V., Stutzki, J. and Winnewisser, G. (eds): The Physics and Chemistry of the Interstellar Medium, GCA-Verlag Her-decke (1999) 58-65.

[gie01] Giesen, T. F., Van Orden, A. O., Cruzan, J. S., Provencal, R. A., Saykally, R. J., Gendriesch, R., Lewen, F. and Winnewisser, G.: Interstellar Detection of CCC and High-Precision Laboratory Measurements Near 2 THz, Astrophys. J. 551 (2001) L181-L184.

[gin94] Gingerich, K. A., Finkbeiner, H. C. and Schmude, W. Jr.: Enthalpies of Formation of Small Lin-ear Carbon Clusters, J. Am. Chem. Soc. 116 (1994) 3884-3888.

[gla85] Glassgold, A. E., Huggins, P. J. and Langer, W. D.: Shielding of CO From Dissociating Radiation in Interstellar Clouds, Astrophys. J. 290 (1985) 615-626.

[gla91] Glassgold, A. E., Mamon, G. A. and Huggins, P. J.: The formation of Molecules in Protostellar Winds, Astrophys. J. 373 (1991) 254-265.

[gre01] Gredel, R., Black, J. H. and Yan, M.: Interstellar C2 and CN Toward the Cyg OB2 Association. A Case Study of X-ray Induced Chemistry, Astronomy & Astrophys. 375 (2001) 553-565.

[gri01] Grishko, V. I., Tereszchuk, K, Duley, W. W. and Bernath, P.: The Far-Infrared Spectrum of Hy-drogenated Amorphous Carbon and the 21, 27, and 33 Micron Features in Carbon-Rich Proto-Planetary Nebulae, Astrophys. J. 558 (2001) L129-L132.

[gri03] Grishko, V. I. and Duley, W. W.: Infrared Absorption Spectra of Deuterated Amorphous Carbon, Astrophys. J. 583 (2003) L43-L45.

[haf95] Haffner, L. M. and Meyer, D. M.: A Search for Interstellar C3 in the Translucent Cloud Toward HD 147889, Astrophys. J. 453 (1995) 450-453.

[hai02] Haines, J. R. and Tsai, C. C.: Graphite Sublimation Tests for the Muon Collider/Neutrino Factory Target Development Program, ORNL/TM-2002/27 (2002) 1-16.

[han95] Handschuh, H., Ganteför, G., Kessler, B., Bechthold, P. S. and Eberhardt, W.: Stable Configura-tions of Carbon Clusters: Chains, Rings, and Fullerenes, Phys. Rev. Lett. 74, No 7 (1995) 1095-1098.

[har03] Harwit, M., Spaans, M.: Chemical Composition of the Early Universe, Astrophys. J. 589 (2003) 53-57

REFERENCES

129

[has03] Hasegawa, T. I. and Kwok, S.: Neutral Carbon in Post-Asymptotic Giant Branch Stars and Planetary Nebulae, Astrophys. J. 585 (2003) 475-481.

[hei01] Heithausen, A., Weiß, A. Kerp, J. and Fritz, T.: Neutral Atomic Carbon in Intermediate-Velocity Clouds, Astrophys. J. 561 (2001) 238-243.

[hel93] von Helden, G., Hsu, M.-T. Gotts, N. and Bowers, T.: Carbon Cluster Cations up to 84 Atoms: Structures, Formation Mechanism, and Reactivity, J. Phys. Chem. 97 (1993), 8182-8192.

[hen98] Hening, Th., Salama, F.: Carbon in the Universe, Science 282 (1998) 2204-2210.

[her01] Herpin, F., Goicoechea, J. R., Pardo, J. R. and Cernicharo, J.: Evolution of Carbon-Rich Proto-Planetary Objects, in: Pilbratt, G. L., Heras, A. M., Prusti, T. and Harris, R. (eds.): Proc. Sympo-sium The Promise of the Herschel Space Observatory, ESA SP-460, (2001) 249-252.

[her73] Herbst, E. and Klemperer: W.: The Formation and Depletion of Molecules in Dense Interstellar Clouds, Astrophys. J. 185 (1973) 505-534.

[her83] Herbst, E.: Ion-Molecule Syntheses of Interstellar Molecular Hydrocarbons through C4H: To-ward Molecular Complexity, Astrophys. J. Suppl. Ser. 53 (1983) 41-53.

[her89] Herbst, E. and Leung, C. M.: Gas-Phase Production of Complex Hydrocarbons, Cyanopolyynes, and Related Compounds in Dense Interstellar Clouds, Astrophys. J. Suppl. Ser. 69 (1989) 271-300.

[her91] Herbst, E., DeFrees, D. J., Talbi, D., Pauzat, F., Koch, W. and McLean, A. D.: Calculations on the Rate of the Ion-Molecule reaction between NH3

+ and H2, J. Chem. Phys. 94 (1991) 7842-7849.

[her93] Herbst, E.: Gas-Grain Chemistry in Quiescent Dense Interstellar Clouds, in: Millar, T. J. and Williams, D. A. (eds.): Dust and Chemistry in Astronomy, IOP, London (1993) 183-204.

[her95] Herbig, G. H.: The Diffuse Interstellar Bands, Annu. Rev. Astrophys. 33 (1995) 19-73.

[hin88] Hinkle, K. W., Keady, J. J. and Bernath, P. F.: Detection of C3 in the Circumstellar Shel of IRC + 10216, Science 241 (1988) 1319-1322.

[hob79] Hobbs, L. M.: Interstellar C2 Molecules Toward ζ Persei, Astrophys. J. 232 (1979) L175-L177.

[hob81] Hobbs, L. M.: Interstellar C2 Molecules Toward ο Persei, Astrophys. J. 243 (1981) 485-488.

[hob82] Hobbs, L. M. and Campbell, B.: Interstellar C2 Molecules Toward Zeta Ophiuchi, Astrophys. J. 254 (1982) 108-110.

[hol71] Hollenbach, D. J. and Salpeter, E. E.: Surface Recombination of Hydrogen Molecules, Astrophys. J. 163 (1971) 155-164.

[hol89] Hollenbach, D. and McKee, C. F.: Molecule Formation and Infrared Emission in Fast Interstellar Shocks. III. Results for J Shocks in Molecular Clouds, Astrophys. J. 342 (1989) 306-336.

[hol97] Hollenbach, D. J. and Tielens, A. G. G. M.: Dense Photodissociation Regions (PDRs), Annual Review of Astronomy & Astrophys. 35 (1997) 179-216.

[hri91] Hrivnak, B. J. and Kwok, S.: The Carbon-Reach Proto-Planetary Nebula IRAS 22272+5435, Astrophys. J. 371 (1991) 631-635.

[hri95] Hrivnak, B. J.: Optical Spectroscopy of Carbon-Reach Proto-Planetary Nebulae, Astrophys. J. 438 (1995) 341-349.

[hun93] Hunter, J. M., Fye, J. L. and Jarrold, M. F.: Annealing and Dissociation of Carbon Rings, J. Chem. Phys. 99 (1993), 1785-1795.

[ing02] Ingalls, J. G., Reach, W. T. and Bania, T. M.: Photoelectric Heating and [CII] Cooling of High Galactic Latitude Translucent Clouds, Astrophys. J. 579 (2002) 289-303.

[isr98] Israel, F. P., Tilanus, R. P. J. and Baas, F.: Detection of Neutral Carbon in the M 31 dark Cloud D 478, Astronomy & Astrophys. 339 (1998) 398-404.

[izu03] Izumiura, H.: Circumstellar C2 Swan Band in Optical Carbon Stars, The Astronomical Herald 96 No. 6 (2003) 328-339.

REFERENCES

130

[jef73] Jefferts, K. B., Penzias, A. A., Wilson, R. W.: Deuterium in the Orion Nebula, Astrophys. J. 179 (1973) L57-L59.

[jon97] Jones, R. M., Gerlich, D. and Anderson, S. L.: Simple Radio-FrequencyPower Source for Ion Guides and Ion Traps, Rev. Sci. Instrum. 68 (1997), 3357-3362.

[jon99] Jones, R. O.: Density Functional Study of Carbon Clusters C2n (2 ≤ n ≤ 16). I. Structure and Bonding in the Neutral Clusters, J. Chem. Phys. 110 (1999), 5189-5200.

[juv03] Juvela, M., Padoan, P. and Jimenez, R.: Photoelectric Heating and [CII] Cooling in Translucent Clouds: Results for Cloud Models Based on Simulations of Compressible Magnetohydrodynamic Turbulence , Astrophys. J. 591 (2003) 258-266.

[kai01] Kaiser, I. R.: Experimental Investigations of the Formation of Carbon-Bearing Molecules in the Interstellar Medium via Neutral-Neutral Reactions, Habilitationsschrift, submitted in Sept. 2001 to the Fakultät für Naturwissenschaften of the Technische Universität Chemnitz.

[kai02] Kaiser, I. R.: Experimental Investigations of the Formation of Carbon-Bearing Molecules in the Interstellar Medium via Neutral-Neutral Reactions, Chem. Rev. 102 (2002) 1309-1358.

[kai95] Kaiser, I. R. and Suits, A. G.: A High-Intensity, Pulsed Supersonic Carbon Source with C(3Pj) Kinetic Energies of 0.08-0.7 eV for Crossed Beam Experiments, Rev. Sci. Instrum. 66 (1995) 5405-5411.

[kai98] Kaiser, I. R., Ochsenfeld, C., Stranges, D., Haed-Gordon, M. and Lee, Y.T.: Combined Crossed Beams and Ab Initio Investigations of the Formation of Carbon Bearing Molecules in the Interstel-lar Medium via Neutral-Neutral Reactions, Faraday Discuss. 109 (1998) 183-204.

[kee85] Keene, J., Blake, G. A., Phillips, T. G., Huggins, P. J. and Beichman, C. A.: The Abundance of Atomic Carbon Near the Ionization Fronts in M17 and S140, Astrophys. J. 299 (1985) 967-980.

[ken76] Kenedy, C. S. and Kenedy, G. C.: The Equilibrium Boundary Between Graphite and Diamond, J. Geophys. Res. 81 (1976) 2467-2470.

[ker01] Kerley, G. I. and Chhabildas, L.: Multicomponent-Multiphase Equation of State for Carbon, San-dia Report, SAND2001-2619, (2001) 1-50.

[kim02] Kim, Y. K. and Desclaux, J. P.: Ionization of Carbon, Nitrogen and Oxygen by Electron Impact, Phys. Rev. A 66 (2002) 012708-1 - 012708-11.

[kim03a] Kimura, H., Mann, I. and Jessberger, E. K.: Composition, Structure, and Size Distribution of Dust in the Local Interstellar Cloud, Astrophys. J. 583 (2003) 314-321.

[kim03b] Kimura, H., Mann, I. and Jessberger, E. K.: Elemental Abundances and Mass Densities of Dust and Gas in the Local Interstellar Cloud, Astrophys. J. 582 (2003) 846-858.

[kor98] Korpela, J., Bowyer, S. and Edelstein, J.: Spectral Observations of Diffuse Far-Ultraviolet Emis-sion From the Hot Phase of the Interstellar Medium with the Diffuse Ultraviolet Experiment, As-trophys. J. 495 (1998) 317-327.

[kwo99] Kwok, S., Volk, K. and Hrivnak, B. J.: Chemical Evolution of Carbonaceous Materials in the Last Stages of Stellar Evolution, Astronomy & Astrophys. 350 (1999) L35-L38.

[lam95] Lambert, D. L., Sheffer, Y. and Federman, S. R.: Hubble Space Telescope Observations of C2 Molecules in Diffuse Interstellar Clouds, Astrophys. J. 438 (1995) 740-749.

[lan84] Langer, W. D., Graedel, T. E., Frerking, M. A. and Armentrout, P. B.: Carbon and Oxygen Iso-tope Fractionation in Dense Interstellar Clouds, Astrophys. J. 277 (1984) 581-604.

[lat91] Latter, W. B.: Large Molecule Production by Mass-Losing Carbon Stars: The Primary Source of Interstellar Polycyclic Aromatic Hydrocarbons?, Astrophys. J. 377 (1991) 187-191.

[let00] Le Teuff, Y. H., Millar, T. J. and Markwick, A. J.: The UMIST Database for Astrochemistry 1999, Astronomy & Astrophys. Suppl. Ser. 146 (2000) 157-168.

[leu84] Leung, C. M., Herbst, E. and Huebner, W. F.: Synthesis of Complex Molecules in Dense Interstel-lar Clouds via Gas-Phase Chemistry: a Pseudo Time-Dependent Calculation, Astrophys. J. Supp. Ser. 56 (1984) 231-256.

REFERENCES

131

[lin86] Lindinger, W.: Swarm Methods, in: Futrell, J. H. (ed.): Gaseous Ion Chemistry and Mass Spec-trometry, A Wiley-Interscience Publication, John Wiley & Sons, New York (1986), 141-154.

[lis02] Lis, D. C., Roueff, E., Gerin, M., Phillips, T. G., Coudert, L. H., van der Tak, F. F. S. and Schilke, P.: Detection of Triply Deuterated Ammonia in the Barnard 1 Cloud, Astrophys. J. 571 (2002) L55-L58.

[loi01] Loinard, L., Castets, A., Ceccarelli, C., Caux, E. and Tielens, A. G. G. M.: Doubly Deuterated Molecular Species in Protostellar Environments, Astrophys. J. 552 (2001) L163-L166.

[luc01] Luca, A., Schlemmer, S., Čermák, I. and Gerlich, D.: On the Combination of a Linear Field Free Trap with a Time-of-Flight Mass Spectrometer, Rev. Sci. Instrum. 72 (2001), 2900-2908.

[luc02] Luca, A., Voulot, D. and Gerlich, D.: Low Temperature Reactions Between Stored Ions and Con-desable Gases: Formation of Protonated Methanol via Radiative Association, in: Safrankova, J. (ed.): WDS’02, Proceedings of Contribution Papers, Part II, Physics of Plasma and Ionized Media, MATFYZPRESS, Prague (2002) 294-300.

[lut77] Lutz, B. L. and Souza, S. P.: A Serach for C2 in the Interstellar Spectrum of ζ Ophiuchi, Astro-phys. J. 213 (1977) L129-L130.

[mai00] Mair, C., Deutsch, H., Becker, H. and Märk, T. D.: Partial and Total Electron Impact Ionization Efficiencies for C, C2 and C3, in: Bassi, D. and Tosi, P. (eds.): SASP 2000, XII Symposium on Atomic and Surface Physics and Related Topics, Contributions, Stampato dal Servizio, Stamperia e Fotoriproduzione dell’ Università degli Studi di Trento, Trento (2000) PA12 1-4.

[mai01] Maier, J. P., Lakin, N. M., Walker G. A. H. and Bohlender, D. A.: Detection of C3 in Diffuse Interstellar Clouds, Astrophys. J. 553 (2001) 267-273.

[mai02] Maier, J. P., Walker, G. A. H. and Bohlender, D. A.: Limits to Interstellar C4 and C5 Toward ζ Ophiuchi, Astrophys. J. 566 (2002) 332-335.

[mai98] Maier, J. P.: Electronic Spectroscopy of Carbon Chains, J. Phys. Chem. A 102 (1998) 3462-3469.

[mal92] Maluendes, S. A., McLean, A. D., and Herbst, E.: Calculations of Ion-Molecule Deuterium Frac-tionation Reactions Involving HD, Astrophys. J. 397 (1992) 477-481.

[mar90] Martin, C. and Bowyer, S.: Discovery of High-Ionization Far-Ultraviolet Line Emission From the Interstellar Medium, Astrophys. J. 350 (1990) 242-261.

[mar97] March, R. E.: An Introduction to Quadrupole Ion Trap Mass Spectrometry, J. Mass Spec., 32 (1997) 351-369.

[mar98] Marshall, A. G., Hendrickson, C. L. and Jackson, G. S.: Furier Transform Ion Cyclotron Reso-nance Mass Spectrometry: A Primer, Mass Spectrometry Rev. 17 (1998) 1-35.

[mat02] Matsuura, M., Zijlstra, A. A., van Loon, Th., Yamamura, I., Markwick, A. J., Woods, P. M. and Waters, L. B. F. M.: Very Large Telescope Spectra of Carbon Stars in the Large Magellanic Cloud and Thei Metallicity Dependence, Astrophys. J. 580 (2002) L133-L136.

[mcc01] McCall, B. J., Thorburn, J., Hobbs, L. M., Oka, T. and York, D. G.: Rejection of the C7- Diffuse

Interstellar Band Hypothesis, Astrophys. J. 559 (2001) L49-L53.

[mil02] Millar, T. J.: Modeling Deuterium Fractionation in Interstellar Clouds, (in:) Deuterium in the Universe, a special issue of Planetary and Space Science 50 (2002) 1189-1195.

[mil89] Millar, T. J., Bennett, A. and Herbst, E.: Deuterium Fractionation in Dense Interstellar Clouds, Astrophys. J. 340 (1989) 906-920.

[mil90] Millar, T. J. and Herbst, E.: Chemical Modeling of Dark Clouds in the LMC and SMC, Mon. Not. R. Astron. Soc. 242 (1990) 92-97.

[mil97] Millar, T. J., Farquhar, P. R. A. and Willacy, K.: The UMIST Database for Astrochemistry 1995, Astronomy & Astrophys. Suppl. Ser. 121 (1997) 139-185.

[mus98] Musella, M., Ronchi, C., Brykin, M. and Sheindlin, M.: The Molten State of Graphite: An Ex-perimental Study, J. Appl. Phys. 84 (1998) 2530-2537.

REFERENCES

132

[oka03] Oka, T., Thorburn, J. A., McCall, B. J., Friedman, S. D., Hobbs, L. M., Sonnentrucker, P., Welty, D. E. and York, D. G.: Observations of C3 in Translucent Sight Lines, Astrophys. J. 582 (2003) 823-829.

[olo96a] Olofsson, H., Bergman, P., Eriksson, K. and Gustafsson, B.: Carbon Stars With Episodic Mass Loss: Observations and Models of Molecular Emission From Detached Circumstellar Shels, As-tronomy & Astrophys. 311 (1996) 587-615.

[olo96b] Olofsson, H.: The Neutral Envelopes Around AGB and Post-AGB Objects, in van Dishoeck, E. F. (ed.): Molecules in Astrophysics: Probes and Processes, abstract book IAU Symposium 178, (1996) 457-468.

[ord95] van Orden, A., Cruzan, J. D., Provencal, R. A., Giesen, T. F., Saykally, R. J., Boreiko, R. T. and Betz, A. L.: A Search for the C3 Carbon Cluster in the Interstellar Medium, in: Haas, M. R., Davidson, J. A. and Erickson, E. F (eds): Airborne Astronomy Symposium on the Galactic Eco-system, ASP Conference Series 73 (1995) 67-70.

[ord98] Van Orden, A. and Saykally, R. J.: Small Carbon Clusters: Spectroscopy, Structure, and Energet-ics, Chem. Rev. 98 (1998) 2313-2357.

[pag03] Le Page, V., Snow, T. P. and Bierbaum, V. M.: Hydrogenation and Charge States of Polycyclic Hydrocarbons in Diffuse Clouds. II. Results, Astrophys. J. 584 (2003) 316-330.

[pal67] Palmer, H. B, and Shelef, M.: Vaporization of Carbon, Chemistry and Physics of Carbon 4 (1967) 85-135.

[par02] Parise, B., Ceccarelli, C., Tielens, A. G. G. M., Herbst, E., Lefloch, B., Caux, E., Castets, A., Mukhopadhyay, I., Pagani, L. and Loinard, L.: Detection of Doubly-Deuterated Methanol in the Solar-Type Protostar IRAS 16293-2422, Astronomy & Astrophys. 393 (2002) L49-L53.

[pas00] Pascoli, G. and Polleux, A.: Condensation and Growth of Hydrogenated Carbon Clusters in Car-bon-Rich Stars, Astronomy & Astrophys. 359 (2000) 799-810.

[pat95] Patzer, A. B. C., Köhler, T. M., and Sedlmayr, E.: Primary Astrophysical Dust Formation: Labo-ratory Desiderata, Planetary and Space Science 43 (1995) 1233-1239.

[pau94] Paul, W. and Gerlich, D.: The Problem of Cooling Ions in a Low Density Ion Trap, in: Märk, T. D., Schrittwieser, R. and Smith, D. (eds.): Contributions of Simposium on Atomic, Cluster and Surface Physics ‘94 (1994) 197-199.

[pau95] Paul, W., Lücke, B., Schlemmer, S. and Gerlich, D.: On the Dynamics of the Reaction of Positive Hydrogen Cluster Ions (H5

+ to H23+) with Para and Normal Hydrogen at 10 K, Int. J. Mass Spec-

trom. Ion Processes 149/150 (1995) 373-387.

[pin91] Pineau des Forets, G., Flower, D. R. and Herbst, E.: Thermal and Chemical Evolution of Interstel-lar Clouds: Complex Molecule Formation, Mon. Not. R. Astron. Soc. 253 (1991) 359-363.

[rob00a] Roberts, H. and Millar, T. J.: Modelling of Deuterium Chemistry and its Application to Molecu-lar Clouds, Astronomy & Astrophys. 361 (2000) 388-398.

[rob00b] Roberts, H. and Millar, T. J.: Gas Phase Formation of Doubly-Deuterated Species, Astronomy & Astrophys. 364 (2000) 780-784.

[rou00] Roueff, E., Tine, S., Coudert, L. H., Pineau des Forets, G., Falgarone, E. and Gerin, M.: Detection of Doubly Deuterated Ammonia in L134N, Astronomy & Astrophys. 354 (2000) L63-L66.

[rou02] Roueff, E., Felenbok, P., Black, J. H. and Gry, C.: Interstellar C3 Toward HD 210121, Astron-omy & Astrophys. 384 (2002) 629-637.

[row84] Rowe, B. R., Dupeyrat, G., Marquette, J. B. and Gaucherel, P.: Study of the Reactions N2+ + 2N2

→ N4+ + N2 and O2

+ + 2O2 → O4+ + O2 from 20 to 160 K by the CRESU Technique, J. Chem.

Phys. 80 (1984), 4915-4921.

[she03] Shematovich, V. I., Wiebe, D. S., Shustov, B. M. and Li, Z.-Y.: A Coupled Dynamical and Chemical Model of Starless Cores of Magnetized Molecular Clouds. II. Chemical Differentiation, Astrophys. J. 588 (2003) 894-909.

[smi00] Smith, I. W. M., and Rowe, B. R.: Reaction Kinetics at Very Low Temperatures: Laboratory Studies and Interstellar Chemistry, Accounts of Chemical Research 33 (2000) 261-268.

REFERENCES

133

[smi02] Smith, M. A., Schlemmer, S., von Richthofen, J. and Gerlich, D.: HOC+ + H2 Isomerization Rate at 25 K: Implications for the Observed [HCO+]/[HOC+] Ratios in the Interstellar Medium, Astro-phys. J. 578 (2002) L87-L90.

[smi79] Smith, M.A. and Adams, N. G.: Recent Advances in Flow Tubes: Measurement of Ion-Molecule Rate Coefficients and Product Distribution, in: Bowers, M. T. (ed.): Gas Phase Ion Chemistry, Vol. 1, Academic Press, New York (1979), 1-44.

[smi88] Smith, D. and Adams, N. G.: The Selected Ion Flow Tube (SIF): Studies of Ion-Neutral Reac-tions, Advances in Atomic and Molecular Physics 24 (1988), 1-49.

[smi92] Smith, D.: The Ion Chemistry of Interstellar Clouds, Chem. Rev. 92 (1992), 1473-1485.

[smi93] Smith, M.A.: Ion Molecule reactions Dynamics at Very Low Temperatures, in: Ng, C.Y., Baer, T. and Powis I. (eds.): Current Topics in Ion Chemistry and Physics, Vol. 2, Wiley, New York (1993), 183-251.

[smi98] Smith, M. A.: Low Temperature Rate Studies of Ions and Radicals in Supersonic Flows, Interna-tional Reviews in Phys. Chem. 17 (1998), 35-63.

[sno78] Snow, T. P. Jr.: The Kinetic Temperature in the Interior of the ζ Ophiuchi Cloud From Coperni-cus Observations of Interstellar C2, Astrophys. J. 220 (1978) L93-L96.

[sno88] Snow, T. P., Seab, C. G. and Joseph, C. L.: A New Search for Interstellar C3, Astrophys. J. 335 (1988) 185-187.

[sor94] Sorgenfrei, A. and Gerlich, D.: Ion-Trap Experiments on C3H+ +H2: Radiative Association vs.

Hydrogen Abstraction in: Nenner, I. (ed.): Molecules and Grains in Space, AIP Press, New York (1994) 505-513.

[sou77] Souza, S. P. and Lutz, B. L.: Detection of C2 in the Interstellar Spectrum of Cygnus OB2 Number 12 (VI Cygni Number 12), Astrophys. J. 216 (1977) L49-L51.

[sta91] Stacey, G. J., Geis, N., Genzel, R., Lugten, J. B., Poglitsch, A., Sternberg, A. and Townes, C. H.: The 158 Micron [CII] Line: a Measure of Global Star Formation Activity in Galaxies, Astrophys. J. 373 (1991) 423-444.

[sta97] Stark, A. A., Bolatto, A. D., Chamberlin, R. A., Lane, A. P., Bania, T. M., Jackson, J. M. and Lo, K. -Y.: First Detection of 492 GHz [CI] Emission From the Large Magellanic Cloud, Astrophys. J. 480 (1997) L59-L62.

[stu88] Stutzki, J., Stacey, G. J., Genzel, R., Harris, A. I., Jaffe, D. T. and Lutgen, J. B.: Submillimeter and Far-Infrared Line Observations of M17 SW: a Clumpy Molecular Cloud Penetrated by Ultra-violet Radiation, Astrophys. J. 332 (1988) 377-399.

[suz83] Suzuki, H.: Synthesis of Chain Molecules in Regions with Partially Ionized Carbon, Astrophys. J. 272 (1983) 579-590.

[tak00] Takahashi, J.: Ab Initio Calculations for Detectable New Isomers of Interstellar Carbon-Chain Radicals CnH (n = 2-8), Publ. Astron. Soc. Japan, 52 (2000), 401-407.

[tak02] van der Tak, F. F. S., Schilke, P., Muller, H. S. P., Lis, D. C., Phillips, T. G., Gerin, M. and Roueff, E.: Triply Deuterated Ammonia in the NGC 1333, Astron. & Astrophys., 388 (2002), L53-L56.

[tar85] Tarafdar, S. P., Prasad, S. S., Huntress, W. T. Jr., Villere, K. R. and Black, D. C.: Chemistry in Dynamically Evolving Clouds, Astrophys. J. 289 (1985) 220-237.

[tat99] Tatematsu, K., Jaffe, D. T., Plume, R. and Evans II, N. J.: Atomic Carbon is a Temperature Probe in Dark Clouds, Astrophys. J. 526 (1999) 295-306.

[tau90] Tauber, J. A. and Goldsmith, P. F.: A Model for Clumpy Giant Molecular Clouds with External Ultraviolet Heating, Astrophys. J. 356 (1990) L63-L66.

[ter98] Terzieva, R. and Herbst, E.: The Sensitivity of Gas-Phase Chemical Models of Interstellar Clouds to C and O Elemental Abundances and to a New Formation Mechanism for Ammonia, Astrophys. J. 501 (1998) 207-220.

REFERENCES

134

[tha99] Thaddeus, P. and McCarthy, M. C.: Carbon Chains and Exotic Rings in the Laboratory and in Space, in: Ossenkopf, V., Stutzki, J. and Winnewisser, G. (eds): The Physics and Chemistry of the Interstellar Medium, GCA-Verlag Herdecke (1999) 338-346.

[tho03] Thorburn, J. A., Hobbs, L. M., McCall, B. J., Oka, T., Welty, D. E., Friedman, S. D., Snow, T. P., Sonnentrucker, P. and York, D. G.: Some Diffuse Interstellar Bands Related to Interstellar C2 Molecules, Astrophys. J. 584 (2003) 339-356.

[tie82] Tielens, A. G. G. M. and Hagen, W.: Model Calculations of the Molecular Component of Inter-stellar Grain Mantles, Astron. Astrophys. 114 (1982) 245-260.

[tie85] Tielens, A. G. G. M. and Hollenbach, D.: Photodissociation Regions. I. Basic Model, Astrophys. J. 291 (1985) 722-746.

[tog94] Togaya, M., Sugiyama, S. and Mizuhara, E.: Melting of Graphite, in: Schmidt, S. C., Shaner, J. W., Samara, G. A. and Ross, A. (eds.): High Pressure Science and Technology – 1993, AIP Con-ference Proceedings, 309 (1994) 255-258.

[tul98] Tulej, M., Kirkwood, D. A., Pachkov, M. and Maier, J. P.: Gas-Phase Electronic Transitions of Carbon Chain Anions Coinciding With Diffuse Interstellar Bands, Astrophys. J. 506 (1998) L69-L73.

[tur00] Turner, B. E., Herbst, E. and Terzieva, R.: The Physics and Chemistry of Small Translucent Mo-lecular Clouds. XIII. The Basic Hydrocarbon Chemistry, Astrophys. J. Supp. Ser. 126 (2000) 427-460.

[tur90] Turner, B. E.: Detection of Doubly Deuterated Interstellar Formaldehyde (D2CO): an Indicator of Active Grain Surface Chemistry, Astrophys. J. 362 (1990) L29-L33.

[tur98] Turner, B. E., Lee, H.-H. and Herbst, E.: The Physics adn Chemistry of Small Translucent Mo-lecular Clouds. IX. Acetylenic Chemistry, Astrophys. J. Supp. Ser. 115 (1998) 91-118.

[tur99] Turner, B. E., Terzieva, R. and Herbst, E.: The Physics adn Chemistry of Small Translucent Mo-lecular Clouds. XII. More Complex Species Explainable by Gas-Phase Processes, Astrophys. J. 518 (1999) 699-732.

[vit00] Viti, S.,Williams, D. A. and O'Neill, P. T.: Hydrocarbons in Diffuse and Translucent Clouds, Astron. & Astrophys., 354 (2000), 1062-1070.

[wak99] Wakabayashi, T., Momose, T. and Shida, T.: Mass Spectroscopic Studies of Laser Ablated Car-bon Clusters as Studied by Photoionization with 10.5 eV Photons Under High Vacuum, J. Chem. Phys. 111 (1999), 6260-6263.

[wal93] Walker, C. K., Narayanan, G., Büttgenbach, T. H., Carlstrom, J. E., Keene, J. and Phillips T. G.: The Detection of [CI] in Molecular Outflows Associated with Young Stellar Objects, Astrophys. J. 415 (1993) 672-679.

[wan80] Wannier, P. G.: Nuclear Abundances and Evolution of the Interstellar Medium, Ann. Rev. Astron. & Astrophys., 18 (1980), 399-437.

[wil02] Williams, D. A., Lim, A. J. and Viti S.: The Chemistry of Stellar Jets, RevMexAA (Serie de Con-ferencias) 13 (2002) 114-118.

[wol95] Wolfire, M. G., Hollenbach, D., McKee, C. F., Tielens, A. G. G. M. and Bakes E. L. O.: The Neutral Atomic Phases of the Interstellar Medium, Astrophys. J. 443 (1995) 152-168.

[wri91] Wright, E. L., Mather, J. C., Bennett, C. L., Cheng, E. S., Shafer, R. A., Fixsen, D. J., Eplee Jr., R. E., Isaacman, R. B., Read, S. M., Boggess, N W., Gulkis, S., Hauser, M. G., Janssen, M., Kel-sall, T., Lubin, P. M., Meyer, S. S., Moseley Jr., S. H., Murdock, T. L., Silverberg, R. F., Smoot, G. F., Weiss, R. and Wilkinson, D. T.: Preliminary Spectral Observations of the Galaxy with a 7o Beam by the Cosmic Background Explorer (COBE), Astrophys. J. 381 (1991) 200-209.

[zsa03] Zsargó, J. and Federman, S. R.: Nonthermal Chemistry in Diffuse Clouds with Low Molecular Abundances, Astrophys. J. 589 (2003) 319-337.

135

SELBSTSTÄNDIGKEITSERKLÄRUNG

Ich erkläre, dass ich die vorliegende Arbeit selbstständig und nur unter Verwendung der angegebenen Literatur und Hilfsmittel angefertigt habe. Ich erkläre, nicht bereits früher oder gleichzeitig bei anderen Hochschulen oder an dieser Universität ein Promotionsverfahren beantragt zu haben. Falls diese Erklärung nicht zutrifft, füge ich eine Stellungnahme diesem Antrag bei. Ich erkläre, obige Angaben wahrheitsgemäß gemacht zu haben und erkenne die Promotionsordnung der Fakultät für Naturwissenschaften der Technischen Universität Chemnitz vom 10. Oktober 2001 an.

Igor Savić

136

137

CURRICULUM VITAE

Personal Name: Igor Savić

Birth date and place: 19 August 1970, Priština, Yugoslavia

Citizenship Serbia and Montenegro

Status Married, one child

Education School: 1977-1985 Primary School “Sevce”, Sevce, Yugoslavia

1985-1987 Secondary (comprehensive) School“7. April”, Novi Sad, Yugoslavia

1987-1989 Secondary School of Electrical Engineering “M. Pupin”, Novi Sad, Yugoslavia

University: 1990-1995 Studies of physics at Faculty of Sciences and Mathematics, University of Novi Sad, Novi Sad, Yugoslavia

1995 Diploma thesis:

“Uticaj dinamike jona na pomeraj spektralnih linija neutralnog helijuma u plazmi”

(“The Influence of Ion Dynamics on Shift of the Spectral Lines of the Neutral Helium in Plasma”)

MSc studies 1995-2000 Studies of physics at Faculty of Physics, Univer-sity of Belgrade, Belgrade, Yugoslavia

2000 MSc thesis:

“Asimterija vodonikove Hβ linije emitovane iz plazme”

(“The Asymmetry of the Hβ Hydrogen Line Emitted from Plasma”)

PhD studies: 2001-2003 Ion trap studies in the group of Prof. D. Gerlich, Fakultät für Naturwissenschaften, Institut für Physik, TU Chemnitz, Chemnitz, Germany

138

139

LIST OF PUBLICATIONS AND CONFERENCE CONTRIBUTION

List of Publications Savić, I. i Dr. Lukić, S. R.: Kvalitativni sastav kristalne smeše u mulju Borske reke,

Zbornik radova "Naša ekološka istina", Borsko jezero (1995), 229-232.

Savić, I.: Gama spektrometrijsko merenje prirodne radioaktivnosti mulja Borske reke, Zbornik radova "Naša ekološka istina", Borsko jezero (1995), 267-270.

Mijatović, Z., Kobilarov, R., Djurović, S., Konjević, N. i Savić, I.: Merenje parametara spektralnih linija emitovanih iz plazme sa poboljšanom tačnošću, u : Vujičić, B. (ed.) IX kongers fizičara Jugoslavije, Zbornik radova, Petrovac na Moru (1995), 717-720.

Savić, I.: Uticaj dinamike jona na pomeraj spektralnih linija neutralnog helijuma u plazmi, diplomski rad, Univerzitet u Novom Sadu, Prirodno-matematički fakultet, Novi Sad (1995).

Djurović, S., Mijatović, Z., Pavlov, M., Vujičić, B., Kobilarov, R. and Savić, I.: Asym-metry of the Balmer Hβ Line in the Low DC Magnetic Field, in: Burakov, V. S. and Dimitrijević, M. S.: Proceedings of the first Belarussian-Yugoslavian sympo-sium on physics and diagnostic of laboratory & astrophysical plasma, Minsk, Publ. Obs. Astron. 53, Belgrade (1996) 109-112.

Savić, I., Vujičić, B., Djurović, S., and Pavlov, M.: Central Structure of Dβ Line from T-tube Plasmas, in: Vujičić, B.and Djurović, S. (eds.),18th SPIG, Kotor, Yugoslavia, Contributed Papers & Abstracts of Invited Lectures and Progress Reports, Faculty of Sciences, Institute of Physics, Novi Sad (1996), 290-293.

Savić, I., Vujičić, B., Djurović, S. and Pavlov, M.: Central Structure of Hβ Line from T-tube Plasmas, in: Vujičić, B.and Djurović, S. (eds.),18th SPIG, Kotor, Yugoslavia, Contributed Papers & Abstracts of Invited Lectures and Progress Reports, Faculty of Sciences, Institute of Physics, Novi Sad (1996), 294-297.

Savić, I., Vujičić, B., Djurović, S. and Pavlov, M.: Shifts of the central parts of Hβ line, in: Popović, L. C. and Ćuk, M. (eds.), Contributed paper of 2nd Yugoslav Confer-ence on Spectral Line Shapes, Bela Crkva, Publ. Obs. Astron. Belgrade, 57, (1997) 113-116.

Savić, I., Vujičić, B., Djurović, S. and Pavlov, M.: Measurements of the Hβ Central Part Shifts, in: Konjević, N., Ćuk, M. and Videnović, I. R. (eds.),19th SPIG, Zlatibor, Yugoslavia, Contributed Papers & Abstracts of Invited Lectures, Topical Invited Lectures and Progress Reports, Faculty of Physics, University of Belgrade, Bel-grade (1998) 377-380.

Vujičić, B., Djurović, S., Kaloci, Dj. and Savić, I.: Shock Front Velocity in Modified Magnetically Driven T-Tube, in: Dimitrijević, M. S. and Burakov, V. S., Contrib-uted paper of the 2nd Yugoslav-Belarussian Symposium on Physics & Diagnostics of Laboratory & Astrophysical Plasmas, Zlatibor, Publ. Obs. Astron. Belgrade, 61, (1998) 179-182.

Savić, I., Djurović, S., Vujičić, B. T. and Kobilarov, R.: The asymmetry of the Hβ Line Profile, 3rd Yugoslav Conference on Spectral Line Shapes, Brankovac, Journal of Research in Physics, 28, No. 3, (1999), 267-269.

LIST OF PUBLICATIONS AND CONFERENCE CONTRIBUTIONS

140

Vujičić, B., Maksimović, R., Krstić, S. i Savić, I.: Ekspanzija laserski proizvedene plazme u ranoj fazi razvoja, u: Milić, B. i Markušev, D. (ed.), 10. Kongres fizičara Jugoslavije, Zbornik radova, Vrnjačka Banja, (2000), 705-708

Vujičić, B., Savić, I. i Grnja, J.: Odredjivanje koeficijenta prozračnosti Zemljine atmos-fere na talasnim duzinama 464 nm i 528 nm, u: Milić, B. i Markušev, D. (ed.), 10. Kongres Fizičara Jugoslavije, Zbornik radova, Vrnjačka Banja , (2000), 945-948.

Savić, I., Djurović, S., Vujičić, B. and Kobilarov, R.: The Asymmetry of the Balmer Hβ Line Profile, in: Petrović, Z. Lj., Kuraica, M. M., Bibić, N. and Malović, G.: 20th SPIG, Zlatibor, Contributed Papers and Abstracts of Invited Lectures, Topical In-vited Lectures and Progress Reports; Institute of Physics, Faculty of Physics, Uni-versity of Belgrade, Institute of Nuclear Sciences “Vinča”, Belgrade (2000), 309-312.

Savić, I., Djurović, S., Vujičić, B. and Kobilarov, R.: Experimental Determination of the Hβ Line Asymmetry Parameter, in: Petrović, Z. Lj., Kuraica, M. M., Bibić, N. and Malović, G.: 20th SPIG, Zlatibor, Contributed Papers and Abstracts of Invited Lectures, Topical Invited Lectures and Progress Reports; Institute of Physics, Faculty of Physics, University of Belgrade, Institute of Nuclear Sciences “Vinča”, Belgrade (2000), 313-316.

Savić, I., Djurović, S., Vujičić, B. and Kobilarov, R.: Determination of the Balmer Hβ Line Asymmetry Parameter, in: Burakov, V. S. and Dimitrijević (eds.): Proceed-ings of the 3rd Belarussian-Yugoslavi symposium on physics and diagnostic of laboratory and astrophysical plasma, Minsk, Publ. Astron. Obs. Belgrade 68 (2000), 143-146.

Savić, I.: Asimetrija vodonikove Hβ linije emitovane iz plazme, magistarski rad, Univerzitet u Beogradu, Fizički fakultet, Beograd, (2000).

Čermák, I., Savić, I. and Gerlich, D.: Ion-Trapping Apparatus for Studies on Reactions between Ions and Neutral Carbon Species, in: Safrankova, J. (ed.): WDS’02, Pro-ceedings of Contribution Papers, Part II, Physics of Plasma and Ionized Media, MATFYZPRESS, Prague (2002) 281-287.

Asvany, O., Savić, I., Schlemmer, S. and Gerlich, D.: Variable temperature ion trap studies of CH4

+ + H2 , HD and D2: negative temperature dependence and signifi-cant isotope effect, Chem. Phys. 298, (2004), 97-105.

Conference Contributions Mijatović, Z., Kobilarov, R., Djurović, S., Konjević, N. i Savić, I.: Merenje parametara

spektralnih linija emitovanih iz plazme sa poboljšanom tačnošću, Poster, IX kon-gers fizičara Jugoslavije, Petrovac na Moru, Jugoslavija, (1995).

Djurović, S., Mijatović, Z., Pavlov, M., Vujičić, B., Kobilarov, R. and Savić, I.: Asym-metry of the Balmer Hβ Line in the Low DC Magnetic Field, Poster, 1st Belarus-sian-Yugoslavian symposium on physics and diagnostic of laboratory & astro-physical plasma, PDP I’96, Minsk, Belaruss, (1996).

Savić, I., Vujičić, B., Djurović, S., and Pavlov, M.: Central Structure of Dβ Line from T-tube Plasmas, Poster,18th SPIG, Kotor, Yugoslavia, (1996).


141

Savić, I., Vujičić, B., Djurović, S. and Pavlov, M.: Central Structure of Hβ Line from T-tube Plasmas, Poster,18th SPIG, Kotor, Yugoslavia, (1996).

Savić, I., Vujičić, B., Djurović, S. and Pavlov, M.: Shifts of the central parts of Hβ line, Poster, 2nd Yugoslav Conference on Spectral Line Shapes, Bela Crkva, Yugosla-via (1997).

Savić, I., Vujičić, B., Djurović, S. and Pavlov, M.: Measurements of the Hβ Central Part Shifts, Poster,19th SPIG, Zlatibor, Yugoslavia, (1998).

Vujičić, B., Djurović, S., Kaloci, Dj. and Savić, I.: Shock Front Velocity in Modified Magnetically Driven T-Tube, Poster, 2nd Yugoslav-Belarussian Symposium on Physics & Diagnostics of Laboratory & Astrophysical Plasmas, Zlatibor, Yugo-slavia, (1998).

Savić, I., Djurović, S., Vujičić, B. T. and Kobilarov, R.: The asymmetry of the Hβ Line Profile, Poster, 3rd Yugoslav Conference on Spectral Line Shapes, Brankovac, Yugoslavia, (1999).

Vujičić, B., Maksimović, R., Krstić, S. i Savić, I.: Ekspanzija laserski proizvedene plazme u ranoj fazi razvoja, Poster, 10. Kongres fizičara Jugoslavije, Vrnjačka Banja, Jugoslavija, (2000).

Vujičić, B., Savić, I. i Grnja, J.: Odredjivanje koeficijenta prozračnosti Zemljine atmos-fere na talasnim duzinama 464 nm i 528 nm, Poster, 10. Kongres Fizičara Jugo-slavije, Vrnjačka Banja, Jugoslavija (2000).

Savić, I., Djurović, S., Vujičić, B. and Kobilarov, R.: The Asymmetry of the Balmer Hβ Line Profile, Poster, 20th SPIG, Zlatibor, Yugoslavia, (2000).

Savić, I., Djurović, S., Vujičić, B. and Kobilarov, R.: Experimental Determination of the Hβ Line Asymmetry Parameter, Poster, 20th SPIG, Zlatibor, Yugoslavia (2000).

Savić, I., Djurović, S., Vujičić, B. and Kobilarov, R.: Determination of the Balmer Hβ Line Asymmetry Parameter, Poster, 3rd Belarussian-Yugoslavi symposium on physics and diagnostic of laboratory and astrophysical plasma, Minsk, Belaruss (2000).

Čermák, I., Savić, I. and Gerlich, D.: Reactions Between Stored Ions and Neutral Car-bon Species, Poster, TMR Astrophysical Chemistry Final Network Meeting, Pe-rugia, Italy (2002).

Čermák, I., Savić, I. and Gerlich, D.: Ion-Trapping Apparatus for Experiments on Reac-tions between Ions and Neutral Carbon Species, Talk, WDS’02, Prague, Czech Republic, (2002).

Asvany, O., Savić, I., Schlemmer, S. and Gerlich, D.: Variable Temperature Ion Trap Studies of CH4

+ + H2, HD and D2: negative temperature dependence and signifi-cant isotope effect, Poster, XX International Symposium on Molecular Beams, Lissabon, Portugal, (2003).

Asvany, O., Schlemmer, S., Savić, I., and Gerlich, D.: Isotope fractionation of small hydrocarbon ions, Poster, 4th Cologne-Bonn-Zermatt-Symposium, Zermatt, Swit-zerland (2003).


142

Decker, S., Savić, I., Čermák, I., and Gerlich, D.: Characterization of a Carbon Subli-mation Source for Studying the Growth of Small Carbon Cluster Ions Cn

+, Poster, DPG Frühjahrstagung, München, Germany, (2004).

143

ACKNOWLEDGEMENTS

At this point I would like to express my gratitude to all the people who supported me during the last 4 years (2001-2004):

• First of all I would like to thank Prof. Dr. Dieter Gerlich for the excellent super-vision.

• Thanks to Dr. Ivo Čermák and PD Dr. Stephan Schlemmer for their help during the research work and nice working atmosphere

• Thanks to engineers Adelheid Steinrücken, Doreen Kunte and Angelika Hiller for the excellent technical support.

• The help of VZD Annett Kurasch on solving for me unsolvable administrative problems is gratefully acknowledged.

• I also would like to thank all the group members and some of former group members, namely Dr. Alfonz Luca, Silko Barth, Silvio Decker, Dr. Oskar As-vany, Dr. Stefan Wellert, Michael Grimm, Dr. Martin Stübig, Dr. Christophe Nicolas, Gheorghe Borodi and Dr. Hans – Jürgen Deyerl for the nice and friendly atmosphere and their help.

Financial support by the Deutsche Forschungsgemeinschaft (DFG) is gratefully ac-knowledged, especially via the Forschergruppe FOR 388 "Laboratory Astrophysics".

Date post:	04-Feb-2017
Category:	Documents
Upload:	vungoc
View:	222 times
Download:	0 times

experiments in ion traps

Documents