Ion chemistry in XH4/allene (X = Ge, Si) gaseous mixtures - Formation of X-C bonds

Article abstract of DOI:10.1002/(SICI)1099-0682(200003)2000:3<505::AID-EJIC505>3.0.CO;2-I

The gas-phase ion chemistry of germane/allene and silane/allene mixtures has been studied, with the aim of obtaining information on the experimental conditions leading to the formation of clusters of increasing size containing Ge or Si bonded to carbon atoms. Mechanisms of ion/molecule reactions have been elucidated by ion-trap mass spectrometry using single and multiple isolation steps. Rate constants for the most important reactions have been determined experimentally and compared with collisional rate constants. The germane/allene mixtures display a low reactivity and the most abundant germanium and carbon containing ion is GeCH₃/⁺. However, chain propagation proceeds after the first nucleation step, even if rather slowly, with the formation of large clusters such as Ge₄C₃H₃/⁺ at low abundance. In contrast, the silane/allene mixtures are very reactive and many different processes are observed, with the formation of several silicon and carbon containing ions with appreciable efficiency. Chain propagation proceeds mainly through reactions of silicon-containing ions with allene molecules and the subsequent formation of large clusters such as Si₃C₃H₅/⁺ and Si₄C₃H₇/⁺.

Full text of DOI:10.1002/(SICI)1099-0682(200003)2000:3<505::AID-EJIC505>3.0.CO;2-I

FULL PAPER
Ion Chemistry in XH4/Allene (X ؍ Ge, Si) Gaseous Mixtures – Formation of
X–C Bonds
Paola Benzi,[a] Lorenza Operti,*[a] and Roberto Rabezzana[a]
Keywords: Ion-molecule reactions / Mass spectrometry / Allene / Germanium / Silicon
The gas-phase ion chemistry of germane/allene and silane/
allene mixtures has been studied, with the aim of obtaining
However, chain propagation proceeds after the first nucle-
ation step, even if rather slowly, with the formation of large
information on the experimental conditions leading to the clusters such as Ge4C3H+3 at low abundance. In contrast, the
formation of clusters of increasing size containing Ge or Si
bonded to carbon atoms. Mechanisms of ion/molecule reac- processes are observed, with the formation of several silicon
tions have been elucidated by ion-trap mass spectrometry us- and carbon containing ions with appreciable efficiency.
silane/allene mixtures are very reactive and many different
ing single and multiple isolation steps. Rate constants for the Chain propagation proceeds mainly through reactions of sil-
most important reactions have been determined experiment- icon-containing ions with allene molecules and the sub-
ally and compared with collisional rate constants. The ger-
sequent formation of large clusters such as Si3C3H+5 and
mane/allene mixtures display a low reactivity and the most Si4C3H+7.
abundant germanium and carbon containing ion is GeCH+3.
Introduction
methylgermane/hydrocarbon mixtures, including GeH4/al-
lene and CH3GeH3/allene systems, were also studied by
Fourier transform and chemical-ionisation mass spectro-
metry under different experimental conditions from the pre-
sent work.[4] More recently, systems containing silane with
propane or propene,[5] ethane or ethyne,[6] and ethene,[7]
have been studied in an ion-trap instrument with the aim
of investigating the formation of ion species containing Si–
C bonds.
In this paper we report on the mechanisms of ion/molec-
ule reactions in silane/allene and germane/allene mixtures
studied by ion-trap mass spectrometry. Self-condensation
processes in allene have also been investigated. The results
are interesting both for fundamental and applied chemistry,
as they give insights on pathways leading to the formation
of silicon-carbon or germanium-carbon containing ions as
possible precursors of amorphous silicon or germanium
carbides.
In the last few years, hydrogenated amorphous com-
pounds of elements of the Group 14 have been extensively
studied for their possible applications in electronic and op-
toelectronic devices.[1] Amorphous semiconductors based
on germanium carbides, a-GeC:H, or silicon carbides, a-
SiC:H, have recently been considered[2] in order to obtain
collectors of wide range light.
The preparation of these materials is generally carried
out by laser-assisted, X-ray-assisted, or other assisted chem-
ical-vapour deposition (CVD) techniques, starting from
suitable mixtures of gaseous hydrides containing german-
ium or silicon with hydrocarbons. It has been observed that
the amount of Ge and C, or Si and C in the final solid is a
function of the molar composition of the gaseous system,
but it does not depend on it in a direct or predictable way.
Therefore, it is very important to know which are the main
processes taking place in the mixtures under examination,
and which occur more rapidly, contributing to a larger ex-
tent to the growth of the polymer containing carbon bon-
ded to germanium or silicon atoms. These results provide
information on the relationship between the partial pressure
of the gaseous reactants and the abundance of the corres-
ponding elements in the solid material.
The rate constants of the first steps of the chain propaga-
tion have been experimentally determined and compared
with the collisional rate constants, calculated according to
the Langevin theory.[8] A comparison can be also drawn on
the gas-phase reactivity of germane and silane towards al-
lene.
In previous studies, the mechanisms and kinetics of ionic
reactions involved in radiolytically assisted CVD of volatile
silicon and germanium hydrides in different mixtures (SiH4/
NH3, SiH4/PH3, CH3SiH3/NH3, CH3SiH3/PH3, GeH4/
NH3, GeH4/PH3, CH3GeH3/NH3, CH3GeH3/PH3) were de-
termined by ion-trap mass spectrometry.[3] Germane or
Results and Discussion
Tables 1–6 report experimental and collisional rate con-
stants (Langevin theory) of the ion/molecule reactions tak-
ing place in the systems studied, and reaction efficiencies
calculated as their ratio. When heats of formation were
available,[9–17] enthalpies at 298 K were calculated to con-
firm the feasibility of the observed processes. The most
[a]
Dipartimento
Applicata,Universita di Torino,
Corso Massimo d’Azeglio 48, I-10125 Torino, Italy
di
Chimica
Generale
ed
Organica
`
Eur. J. Inorg. Chem. 2000, 505Ϫ512
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
1434Ϫ1948/00/0303Ϫ505 $ 17.50ϩ.50/0
505

P. Benzi, L. Operti, R. Rabezzana
FULL PAPER
Table 1. Reactions of C2Hnϩ,C3Hnϩ, C4Hnϩ and C5H5ϩ ions in allene
[a]
[b]
Product Ions and Rate Constants (kexp
)
Σkexp
kL
Efficiency[c]
Reactant Ions
C2H2ϩ
C2H3ϩ
C2H4ϩ
C3Hϩ
C3H2ϩ
C3H3ϩ
C3H4ϩ
C3H5ϩ
C4H3ϩ
C4H4ϩ
C4H5ϩ
C5H5ϩ
C3H4ϩ (10), C3H5ϩ (0.95)
C3H5ϩ (7.7)
11
13.76
13.60
13.46
12.46
12.37
12.29
12.21
12.14
11.54
11.49
11.44
10.98
0.80
0.57
0.69
0.72
0.90
0.090
0.56
0.63
0.42
0.44
0.16
0.076
7.7
9.3
9.0
11
C2H5ϩ (0.20), C3H3ϩ (7.3), C3H4ϩ (1.2), C3H5ϩ (0.58)
C2H3ϩ (0.87), C3H3ϩ (0.50), C4H3ϩ (7.6)
C3H3ϩ (6.2), C4H4ϩ (4.1), C6H5ϩ (0.78)
C4H5ϩ (0.22), C6H5ϩ (0.90)
1.1
6.8
7.6
4.8
5.0
1.8
0.84
C3H5ϩ (0.86), C6H7ϩ (5.9)
C4H5ϩ (0.75), C6H7ϩ (2.0), C6H8ϩ (4.8)
C5H5ϩ (4.8)
C5H4ϩ (0.26), C7H7ϩ (4.7)
C6H5ϩ (0.86), C7H7ϩ (0.72), C7H8ϩ (0.20)
C5H7ϩ (0.46), C6H7ϩ (0.38)
[a]
Rate constants are expressed as 10–10 cm3molecule–1s–1; experiments were run at 333 K; uncertainty is within 20%. – [b] Rate constants
have been calculated according to the Langevin theory calculating polarizability of C3H4 as in ref.[19]
as the ratio Σkexp/kL.
–
Efficiency has been calculated
[c]
Table 2. Reactions of C2Hnϩ,C3Hnϩ and C4Hnϩ ions with germane and of GeHnϩ ions with allene in a GeH4/C3H4 mixture
[a]
[b]
Product Ions and Rate Constants (kexp
)
Σkexp
kL
Efficiency[c]
Reactants
C2H2ϩ ϩ GeH4
C2H3ϩ ϩ GeH4
C2H4ϩ ϩ GeH4
C3Hϩ ϩ GeH4
C3H2ϩ ϩ GeH4
C3H3ϩ ϩ GeH4
C3H5ϩ ϩ GeH4
C4H3ϩ ϩ GeH4
C4H5ϩ ϩ GeH4
Geϩ ϩ C3H4
Geϩ (2.0), GeHϩ (1.5), GeH2ϩ (1.6), GeH3ϩ (2.6), GeCH3ϩ (0.36), GeC2H3ϩ (0.64)
8.7
5.1
3.0
4.6
1.8
0.40
1.2
1.9
0.52
2.8
8.9
4.9
11.83
11.66
11.51
10.44
10.34
10.25
10.08
9.42
0.74
0.44
0.26
0.44
0.17
0.039
0.12
0.20
0.056
0.26
0.83
0.46
GeHϩ (1.7), GeH3ϩ (2.6), GeCH3ϩ (0.78)
GeHϩ (0.81), GeH2ϩ (1.7), GeCH3ϩ (0.48)
Geϩ (1.1), GeHϩ (1.8), GeCH3ϩ (0.94), GeC3H3ϩ (0.72)
GeCH3ϩ (1.8)
GeCH3ϩ (0.40)
GeCH3ϩ (1.2)
GeCH3ϩ (0.96), GeC2H3ϩ (0.92)
GeCH3ϩ (0.52)
9.31
GeC3H3ϩ (2.8)
10.76
10.73
10.71
GeHϩ ϩ C3H4
GeH2ϩ ϩ C3H4
GeCH3ϩ (8.9)
GeCH2ϩ (0.65), GeCH3ϩ (0.84), GeC2H3ϩ (1.1), GeC3H3ϩ (1.1), GeC3H5ϩ (1.2)
[a]
Rate constants are expressed as 10–10 cm3molecule–1s–1; experiments were run at 333 K; uncertainty is within 20%. – [b] Rate constants
have been calculated according to the Langevin theory calculating polarizability of C3H4 as in ref.[19], and taking polarizability of GeH4
from ref.[20]
–
Efficiency has been calculated as the ratio Σkexp/kL.
[c]
stable isomer for every species was considered in the calcu- lesser extent, of an ethene molecule or a C3H3 fragment.
lations of the reaction enthalpies.
Elimination of a hydrogen atom or molecule is also ob-
served, but is limited to C3Hnϩ (n ϭ 2–5) and C4Hnϩ ( n ϭ
4, 5) ions.
Allene Self-Condensation
Efficiencies of reaction are rather high for all primary
ions and protonated allene, except for C3H3ϩ, which reacts
very slowly. Reactions of the C4Hnϩ (n ϭ 3–5) ion family
show a lower efficiency, which further decreases for the
stable C5H5ϩ ion.
The ion chemistry of allene has been previously studied
by different mass spectrometric techniques.[18] Overall rate
constants have been reported for reactions of the C3Hnϩ
(n ϭ 1–4) ions with neutral allene,[18f] and are generally
higher than those reported in the present paper (see
Table 1). Such discrepancy is probably due to the different
conditions of sample pressure at which the instruments op-
erate, and to the presence of a buffer gas in the ion-trap
mass spectrometric experiments, giving a lower population
of excited ions with respect to other methods.
Germane/Allene Mixture
Table 2 reports the most significant ion/molecule reac-
tions occurring in a germane/allene mixture. During ionis-
Ionisation of allene gives rise to the C3Hnϩ (n ϭ 1–4) and ation, the GeHnϩ (n ϭ 0–3) ion family is formed, besides
C2Hnϩ (n ϭ 2–4) ion families. Formation of C2H3ϩ and primary ions of allene as described above. The main reac-
C2H4ϩ involves hydrogen migration before the fragmenta- tion pathway shown by all the C2Hnϩ (n ϭ 2–4) ions, as well
tion process. If stored for suitable periods of time, all the as C3Hϩ, consists of charge transfer to GeH4 with success-
above-mentioned ions react with neutral allene to yield ive fragmentation to form ions of the GeHnϩ (n ϭ 0–3) fam-
heavier ion species such as protonated allene (C3H5ϩ), ily. It is worth noting that GeCH3ϩ ions are formed by al-
C4Hnϩ (n ϭ 3–5), and C5H5ϩ ions.
most all primary and secondary ions of allene, even if these
Many different reactions are observed (Table 1) taking processes are generally rather slow and only in two cases
place with elimination of different neutral moieties. Rather the reaction rate constants are in the 10–10 cm3molecule–1s–1
common paths involve loss of an ethyne molecule or, to a order of magnitude. The exceptions are C3H4ϩ, C4H4ϩ, and
506
Eur. J. Inorg. Chem. 2000, 505Ϫ512

Ion Chemistry in XH4/Allene
FULL PAPER
C5H5ϩ, which not only do not give the GeCH3ϩ product, but
are unreactive towards germane.
Efficiencies of reaction of the primary C2Hnϩ (n ϭ 2–4)
and C3Hnϩ (n ϭ 1–4) ion families decrease with increasing
number of hydrogen atoms within the same family. More-
over, ions containing two carbon atoms are more efficient
than the C3Hnϩ species. Besides GeCH3ϩ, which is unreactive
under the experimental conditions used here, only two ion
species, GeC2H3ϩ and GeC3H3ϩ, contain Ge–C bonds and
are interesting for the chain propagation towards german-
ium carbide. It is also worth noting that in previous studies
on GeH4/allene mixtures, performed by Fourier transform
mass spectrometry, only these three Ge/C containing ions
(GeCH3ϩ, GeC2H3ϩ, and GeC3H3ϩ) were observed.[4a]
Among the germane primary ions, GeH3ϩ is not included in
the table as it is unreactive in the reaction times examined
here. Germane self-condensation reactions are not reported
as they have been already shown in previous papers.[3f,21]
All reactions of GeHnϩ (n ϭ 0–2) ions with allene yield ion
products containing germanium and carbon together.
Again, GeCH3ϩ, GeC2H3ϩ, and GeC3H3ϩ ions are observed,
together with GeCH2ϩ and GeC3H5ϩ, which are not formed
from allene ions reacting with germane.
Unfortunately, mixed ions containing both Ge and C, as
well as two-germanium containing ions, generally originate
in very slow processes. It follows that, in the time range
considered here, their abundance is too low to perform se-
lective isolation with reproducible results. Moreover, as pre-
viously also observed in CH3GeH3/allene systems,[4b] in
which GeCH3ϩ is a primary ion displaying a very high
abundance, mixed Ge/C ions react very slowly and do not
seem to be the precursors of higher order clusters. In par-
ticular, they do not react with germane in the reaction time
considered, as all mixed ions show a unique germanium
atom and up to six carbon atoms in high pressure chemical
ionisation experiments.[4b]
However, as the main interest in these systems relies on
the catenation of germanium and carbon atoms together to
give mixed cluster ions of increasing size, experiments were
carried out with isolation and reaction of the whole
Ge2Hnϩ (n ϭ 2–4) ion family in the GeH4/C3H4 (4.7 ϫ
10–7/1.9 ϫ 10–7 Torr) mixture. The mass spectra recorded
immediately after the isolation (a), and after 0.5 s (b), 0.9 s
(c), and 2 s (d) of reaction are reported in Figure 1. These
data do not allow the calculation of the rate constants,
nevertheless they give valuable information about Ge2Hnϩ
ions reactivity with allene.
At zero reaction time, signals in the mass spectra cover
the range of m/z values due to Ge2H2ϩ, Ge2H3ϩ, and
Ge2H4ϩ germane secondary ions, the first one being the
most abundant. Increase of the reaction time up to 2 s al-
lowed us to detect several multiplets. Because of the low
abundance of some signals, the identification of the number
of hydrogen atoms contained in some of the ion species was
not straightforward. Therefore, these ions are written with
an ‘x’ as coefficient for hydrogen atoms. Anyway, it is inter-
Figure 1. Mass spectra obtained after isolation of the Ge2Hnϩ (n ϭ
2–4) multiplet ions at 0 s (a), 0.5 s (b), 0.9 s (c), and 2 s (d) reac-
esting to note that with increasing reaction time the tion times.
Eur. J. Inorg. Chem. 2000, 505Ϫ512
507

P. Benzi, L. Operti, R. Rabezzana
FULL PAPER
Ge2Hnϩ multiplet reduces to the one corresponding to the SiH3ϩ primary ion of silane, reacting with silane and allene
Ge2H3ϩ species only. It means that this ion is less reactive neutral molecules.
than the Ge2Hnϩ (n ϭ 2, 4) species, whose abundance de-
creases rapidly due to ion/molecule reactions occurring in
the mixture. Therefore, the products observed are mainly
formed from the Ge2H2ϩ and Ge2H4ϩ ion species reacting
with allene, and in particular from Ge2H2ϩ, which is the
most abundant ion at zero reaction time.
On the basis of the lower and upper limits and of the
most intense signal of each multiplet in the spectra,
GeC3Hnϩ (n ϭ 3–5), Ge2C2Hϩ, Ge2C3H3ϩ, and Ge3Hxϩ (x ϭ
2 for the most abundant species in the multiplet) ion species
are formed. Weak signals due to the Ge2C4Hϩx (x ϭ 5 for
the most abundant species in the multiplet) ions are also
observed. As it contains four carbon atoms, it cannot be
directly formed from Ge2Hnϩ, but it can only originate from
Ge2C2Hϩ or Ge2C3H3ϩ. After about 0.9 s reaction time, the
signals due to the Ge3Hϩx ions almost disappear from the
mass spectra, while the intensities of higher mass multiplets,
assigned to the Ge3CHϩx (x ϭ 0 and 2 for the most abund-
ant species in the multiplet), Ge3C3Hϩx (x ϭ 5 for the most
abundant species in the multiplet), and Ge4Hϩx (x ϭ 0 and
Scheme 1. Paths of ion/molecule reactions starting from SiH3ϩ in a
SiH4/C3H4 mixture.
2 for the most abundant species in the multiplet) ion species,
increase remarkably. If the reaction time is further in-
creased, a weak multiplet due to the Ge4C3Hϩx ions be-
comes detectable, the possible precursors being Ge3C3Hxϩ
and Ge4Hϩx ions.
In separate and successive experiments, some of these ion
families were in turn isolated and reacted, in order to deter-
mine their direct products. In particular, Ge2C2Hϩ and
Ge2C3H3ϩ ions were isolated together, and the main prod-
ucts of reaction observed after 2 s are GeC5Hϩx and
GeC6Hϩx . These ions were obtained by substitution of a Ge
atom by two or three carbons. Moreover, small amounts of
Ge2C4Hϩx and Ge3Hϩx ions are also found. Isolation and
reaction of the Ge3Hϩx ion family shows that it produces
Ge3CHϩx and Ge4Hϩx ions, but a partial contribution of
other species, such as Ge2C2Hϩ or Ge2C3H3ϩ, cannot be
ruled out. Finally, it was possible to isolate the Ge3CHxϩ
multiplet, which leads to the formation of the Ge3C3Hxϩ
ion species.
Silane/Allene Mixture
Figure 2. Variation of the total abundances of silicon and carbon
containing ions as a function of reaction time for the 1:5, 1:1, and
5:1 silane/allene mixtures.
Three different mixtures were considered, which consist
of silane and allene in the approximate ratios 1:5 (1.8 ϫ
10–7/8.5 ϫ 10–7 Torr), 1:1 (5.0 ϫ 10–7/4.6 ϫ 10–7 Torr), and
Figure 2 reports the variation of the total ion abundance
5:1 (8.6 ϫ 10–7/1.7 ϫ 10–7 Torr). Ionisation of the two re- of mixed ions as a function of reaction time for the three
acting molecules again produces the allene primary ions silane/allene mixtures. In all cases, the abundance rapidly
and the SiHnϩ (n ϭ 0–3) ion species from silane. They are increases to a maximum value, which remains constant up
trapped in the cell of the mass spectrometer and react with to two seconds of reaction. The highest yield is given by the
both neutral species with increasing reaction times up to 2 SiH4/C3H4 5:1 mixture, which is more than twice that of
seconds. During this time, several ion families are formed the 1:5 mixture. However, it is worth noting that the behavi-
in successive reactions, and products that are obtained after our of the 1:1 silane/allene system is rather similar to that
four reaction steps are detected. As an example, Scheme 1 of the 5:1 one. Such a trend can be due to the involvement
shows the paths of ion/molecule reactions starting from of allene as neutral reagent.
508
Eur. J. Inorg. Chem. 2000, 505Ϫ512

Ion Chemistry in XH4/Allene
FULL PAPER
reason, Siϩ is not shown. Allene primary and secondary
ions react with silane through different pathways, in which
loss of small hydrocarbon molecules (C2H2, C2H4, C3H4),
fragments (CH3, C2H3), or H atom, with or without a mole-
cule of hydrogen,occurs. Rate constants are generally low
for all reaction paths considered. As observed for the ger-
mane/allene mixtures, efficiencies of reaction decrease with
increasing number of hydrogen atoms in the families of
primary C2Hnϩ (n ϭ 2, 3) and C3Hnϩ (n ϭ 1–4) ions.
Table 4 shows the ion/molecule reactions of the SiHnϩ (n ϭ
1–3) and Si2Hnϩ (n ϭ 2–5) ions with allene. Self-condensa-
tion reactions of silane are not reported, as already discus-
sed in previous papers.[3a,21] Generally, many different prod-
ucts are detected and the reaction efficiencies are high.
However, rate constants of single reactions are rather low
and the high efficiencies are mainly due to the great number
of pathways observed. The only two exceptions are the
SiHϩ and SiH3ϩ ions, which yield SiCH3ϩwith high rate con-
stants, through loss of the stable neutrals, C2H2 and C2H4,
respectively.
Figure 3. Variation of the abundances with reaction time of the
SiCHnϩ, SiC2Hnϩ, SiC3Hnϩ, SiCmHnϩ (m ϭ 4–6), Si2CpHnϩ (p ϭ 1–3),
and Si3CpHnϩ (p ϭ 1, 3) with Si4CHnϩ ion families for the 1:1 silane/
allene mixtures.
As a general remark, reactions with loss of hydrocarbon
molecules are the most frequently observed paths. Addi-
Table 3. Reactions of C2Hnϩ, C3Hnϩ and C4Hnϩ ions with silane in a SiH4/C3H4 mixture
[a]
[b]
Reactant Ions
Product Ions and Rate Constants (kexp
)
Σkexp
kL
Efficiency[c]
C2H2ϩ
C2H3ϩ
C3Hϩ
C3H2ϩ
C3H3ϩ
C3H4ϩ
C3H5ϩ
C4H3ϩ
C4H4ϩ
SiCH3ϩ (2.8), SiC2H3ϩ (2.6), SiC2H5ϩ (1.7)
SiH3ϩ (2.6), SiC2H5ϩ (0.77)
7.1
3.4
9.5
2.9
1.4
0.38
3.0
1.8
0.98
12.85
12.72
11.75
11.68
11.61
11.54
11.48
10.98
10.94
0.55
0.26
0.81
0.25
0.12
0.033
0.26
0.16
0.090
SiCH3ϩ (5.3), SiC3H2ϩ (1.8), SiC3H3ϩ (2.4)
SiCH3ϩ (1.6), SiC2H3ϩ (1.3)
SiCH3ϩ (0.47), SiCH5ϩ (0.89)
SiCH4ϩ (0.38)
SiCH5ϩ (3.0)
SiCH3ϩ (0.40), SiC2H3ϩ (0.56), SiC2H5ϩ (0.89)
SiC2H4ϩ (0.38), SiC3H5ϩ (0.26), SiC4H7ϩ (0.34)
[a]
Rate constants are expressed as 10–10 cm3molecule–1s–1; experiments were run at 333 K; uncertainty is within 20%. – [b] Rate constants
have been calculated according to the Langevin theory taking polarizability of SiH4 from ref.[20]
the ratio ΣkexpkL.
–
Efficiency has been calculated as
[c]
Variation of the abundances of the SiCHnϩ, SiC2Hnϩ,
SiC3Hnϩ, SiCmHnϩ (m ϭ 4–6), Si2CpHnϩ (p ϭ 1–3), and
Si3CpHnϩ (p ϭ 1–3) together with Si4CHnϩ ion families are
reported in Figure 3 for the 1:1 silane/allene mixture. The
total abundance of ions formed in the first reaction step
(SiCHnϩ, SiC2Hnϩ, and SiC3Hnϩ) reaches a maximum value
after about 100–200 ms of reaction, and then decreases to
a rather constant value.
tions of C3H4 and losses of an atom of hydrogen together,
or without a molecule of hydrogen are very frequent pro-
cesses too, with rate constants that are generally lower than
the previous ones. Every Si2Hnϩ (n ϭ 2–5) ion shows at least
one process characterised by substitution of a neutral frag-
ment containing only silicon and hydrogen by a moiety ori-
ginating from allene. In fact, loss of a silyl radical is dis-
played by all the Si2Hnϩ (n ϭ 2–5) ions, except Si2H3ϩ, which
loses a silylene group. The Si2Hnϩ (n ϭ 4, 5) ions also un-
dergo substitution of a silane by an allene molecule.
The other ion species show a slower increase, which is as
high as 20% for the SiCmHnϩ ion family, whereas ions pro-
duced in successive reaction steps never exceed 4%. However,
In order to follow the growth of ions containing silicon
it is remarkable that the abundance of each ion family exam- and carbon already bonded together, reactions of the
ined here does not change substantially after 1 s of reaction. SiCHnϩ (n ϭ 2–5) ion family with both allene and silane are
The rate constants of ion/molecule reactions occurring in reported in Table 5. As already noticed for ions containing
the silane/allene system are displayed in tables 3–6. In only silicon and hydrogen, reactions of the SiCHnϩ (n ϭ 2–
Table 3, reactions of the C2Hnϩ (n ϭ 2, 3), C3Hnϩ (n ϭ 1–5), 5) species with allene proceed through loss of various neut-
and C4Hϩn (n ϭ 3, 4) ions with silane are listed. The C2H4ϩ ral molecules (H2, CH4, C2H2, C2H4, and C3H6) or by elim-
ion is not reported, as it has the same nominal mass of ination of a variety of hydrocarbon fragments or hydrogen
the Siϩ species formed from silane, and no experiment was atoms. In contrast, the reactivity of this ion family towards
possible that allowed its selection alone. For the same silane is remarkably low, as only SiCH2ϩ and SiCH4ϩ react
Eur. J. Inorg. Chem. 2000, 505Ϫ512
509

P. Benzi, L. Operti, R. Rabezzana
FULL PAPER
Table 4. Reactions of SiHnϩ and Si2Hnϩ ions with allene in a SiH4/C3H4 mixture
[a]
[b]
Reactant Ions
Product Ions and Rate Constants (kexp
)
Σkexp
kL
Efficiency[c]
SiHϩ
SiH2ϩ
SiH3ϩ
Si2H2ϩ
Si2H3ϩ
Si2H4ϩ
SiCH3ϩ (14), SiC3H3ϩ (0.50)
14
13.32
13.19
13.06
11.22
11.18
11.15
1.05
0.80
0.75
0.59
0.26
0.64
SiCH2ϩ (2.3), SiCH3ϩ (1.5), SiCH4ϩ (1.2), SiC2H3ϩ (3.9), SiC3H5ϩ (1.6)
SiCH3ϩ (7.4), SiCH5ϩ (2.4)
10.5
9.8
6.6
2.9
7.1
SiC3H3ϩ (1.4), Si2CH2ϩ (2.9), Si2C2H3ϩ (1.8), Si2C3H3ϩ (0.10), Si2C3H5ϩ (0.38)
SiC3H5ϩ (0.38), Si2CH3ϩ (1.4), Si2CH5ϩ (0.46), Si2C3H3ϩ (0.25), Si2C3H5ϩ (0.38)
Si2H2ϩ (0.23), SiC3H4ϩ (0.28), SiC3H5ϩ (2.8), Si2CH4ϩ (1.5), Si2C2H5ϩ
(1.1), Si2C3H6ϩ (0.39), Si2C3H7ϩ (0.79)
Si2H5ϩ
Si2H3ϩ (0.65), SiC3H5ϩ (1.1), SiC3H6ϩ (0.76), Si2CH5ϩ (2.3), Si2CH7ϩ (0.31), Si2C2H5ϩ
(0.36), Si2C2H6ϩ (0.40), Si2C3H7ϩ (0.53), Si2C3H8ϩ (0.21)
6.60
11.11
0.59
[a]
Rate constants are expressed as 10–10 cm3molecule–1s–1; experiments were run at 333 K; uncertainty is within 20%. – [b] Rate constants
have been calculated according to the Langevin theory calculating polarizability of C3H4 as in ref.[19]
as the ratio ΣkexpkL.
–
Efficiency has been calculated
[c]
Table 5. Reactions of SiCHnϩ ions with allene and silane in a SiH4/C3H4 mixture
[a]
[b]
Reactants
Product Ions and Rate Constants (kexp
)
Σkexp kL
Efficiency[c]
SiCH2ϩ ϩ C3H4 SiCH3ϩ (1.1), SiC2H2ϩ (3.4), SiC2H3ϩ (0.21), SiC3H2ϩ (2.7), SiC3H3ϩ (2.3), SiC4H5ϩ (0.45) 10
12.06 0.84
SiCH2ϩ ϩ SiH4
Si2CH4ϩ (1.5)
1.5
2.4
7.8
11.42 0.13
12.00 0.20
11.93 0.65
SiCH3ϩ ϩ C3H4 SiC2H3ϩ (1.4), SiC4H5ϩ (1.0)
SiCH4ϩ ϩ C3H4 SiCH3ϩ (1.0), SiCH5ϩ (2.3), SiC2H4ϩ (1.3), SiC2H6ϩ (0.52), SiC3H5ϩ (1.3),
SiC4H6ϩ (0.58), SiC4H7ϩ (0.84)
SiCH4ϩ ϩ SiH4
Si2CH6ϩ (0.41)
0.41
7.2
11.30 0.036
11.87 0.60
SiCH5ϩ ϩ C3H4 SiCH3ϩ (5.8), SiC2H5ϩ (0.44), SiC2H7ϩ (0.91)
[a]
Rate constants are expressed as 10–10 cm3molecule–1s–1; experiments were run at 333 K; uncertainty is within 20%. – [b] Rate constants
have been calculated according to the Langevin theory calculating polarizability of C3H4 as in ref.[19], and taking polarizability of SiH4
from ref.[20]
–
Efficiency has been calculated as the ratio ΣkexpkL.
[c]
Table 6. Reactions of Si3H5ϩ, Si4H7ϩ,SiC2Hnϩ, and SiC3Hnϩ ions with allene and silane in a SiH4/C3H4 mixture
[a]
[b]
Reactants
Product Ions and Rate Constants (kexp
)
Σkexp
kL
Efficiency[c]
Si3H5ϩ ϩ C3H4
4.4
Si3CH5ϩ (0.93), Si3CH9ϩ (0.90), Si3C3H5ϩ (2.6)
10.38
0.43
0.36
0.23
Si4H7ϩ ϩ C3H4
3.6
Si3C3H5ϩ (0.46), Si4CH7ϩ (0.80), Si4CHϩ11 (0.78), Si4C3H7ϩ (1.6)
9.96
SiC2H3ϩ ϩ C3H4
2.6
SiC3H3ϩ (0.50), SiC3H5ϩ (0.76), SiC5H5ϩ (1.3)
11.35
SiC2H5ϩ ϩ C3H4
5.5
SiC2H3ϩ (2.4), SiC3H5ϩ (0.90), SiC3H7ϩ (0.67), SiC5H7ϩ (1.5)
11.26
0.49
1.5
SiC3H5ϩ ϩ C3H4
SiC3H6ϩ ϩ C3H4
4.2
SiC4H5ϩ (1.1), SiC4H7ϩ (0.39)
10.85
10.38
0.14
SiC4H6ϩ (0.74), SiC4H8ϩ (0.32), SiC5H7ϩ (0.97), SiC6H9ϩ (2.2)
10.82
0.39
0.95
SiC3H6ϩ ϩ SiH4
SiC3H7ϩ ϩ C3H4
Si2C3H8ϩ (0.95)
0.092
SiC3H5ϩ (0.15), SiC4H7ϩ (1.6), SiC4H9ϩ (0.66), SiC5H7ϩ (1.1),
SiC6H9ϩ (1.3), SiC6H1ϩ0 (0.25)
10.80
5.1
0.47
0.38
SiC3H7ϩ ϩ SiH4
Si2C3H9ϩ (0.38)
10.36
0.037
[a]
Rate constants are expressed as 10–10 cm3molecule–1s–1; experiments were run at 333 K; uncertainty is within 20%. – [b] Rate constants
have been calculated according to the Langevin theory calculating polarizability of C3H4 as in ref.[19], and taking polarizability of SiH4
from ref.[20]
–
Efficiency has been calculated as the ratio Σkexp/kL.
[c]
with SiH4 through loss of a hydrogen molecule with rather give SiCH5ϩ. The procedure was repeated in this SiH4/C3H4
low rate constants.
system, isolating and reacting SiCH4ϩ in experiments with
The reaction of SiCH4ϩ to give SiCH5ϩ is an hydrogena- allene at a constant pressure and pulsing silane into the
tion process which can involve both silane and allene as trap. In this way, it was found that the SiCH5ϩ ion is formed
neutral reagent. Similarly, ions at nominal mass 58 and 59, from SiCH4ϩ and allene, as well as SiC2H6ϩ and SiC2H7ϩ
could be formed by SiCH4ϩ reacting with both silane[5] and the corresponding formation rate constants were
(Si2H2ϩ and Si2H3ϩ) and allene (SiC2H6ϩ and SiC2H7ϩ). The calculated.
hydrogenation process was observed in the SiH4/C3H6 sys-
The propagation of the processes leading to cluster ions
tem previously studied,[5] and the use of a pulsed valve al- of increasing size was further studied and results are re-
lowed us to state that both neutrals react with SiCH4ϩ to ported in Table 6. In this table, the reactions of the Si3H5ϩ,
510
Eur. J. Inorg. Chem. 2000, 505Ϫ512

Ion Chemistry in XH4/Allene
FULL PAPER
Si4H7ϩ, SiC2Hnϩ (n ϭ 3, 5) and SiC3Hnϩ (n ϭ 5–7) ions with
allene are shown, together with rate constants of reactions
of SiC3Hnϩ (n ϭ 6, 7) with silane.
Conclusions
In the comparison of the two mixtures studied in this
paper, the great difference of reactivity between silane and
germane towards allene is remarkable. In fact, neutral ger-
mane as well as ions originated by its ionisation are consid-
erably less reactive than the corresponding silicon con-
taining species. Primary and secondary ions of silane react
with allene through many different paths and with appre-
ciable reaction efficiencies to yield mixed ion clusters of
growing size, such as Si4C3H7ϩ or SiC6H1ϩ0. In contrast, ger-
mane primary ions react with allene to give only one prod-
uct for each ion, the only exception being GeH2ϩ which
gives several mixed ions but with low rate constants. More-
over, the distribution of the signal due to an ion species
containing two Ge atoms and the consequent overlapping
of different multiplets make the single selection procedures
difficult. Anyway, the isolation of the whole Ge2Hnϩ (n ϭ
2–4) ion family and its reaction (with reaction times as long
as 2 s) allowed the detection of mixed clusters, up to
Ge4C3Hϩx ions, although in low abundance.
The Si3H5ϩ and Si4H7ϩ ions were chosen as an example of
the behavior of tertiary and quaternary silane ions, as they
are the most abundant within their corresponding families.
Hence, it was possible to isolate and select them. Also the
reactivity of the Si3H4ϩ ion was studied, which reacts very
slowly to give the Si3CH4ϩ and Si3CH8ϩ products. Unfortu-
nately their abundance was too low to allow the determina-
tion of their formation rate constants. The Si3H5ϩ and
Si4H7ϩ ions show a very similar behaviour towards allene.
The SiC2Hnϩ (n ϭ 3, 5) ions react only with allene with
loss of hydrocarbon and H2 molecules with rather low rate
constants.
A similar behaviour is displayed by the
SiC3Hnϩ (n ϭ 5–7) ions. Within this family, the number of
reactions with allene and the reaction efficiencies increase
with increasing number of hydrogen atoms in the reacting
ion. Only the SiC3Hnϩ (n ϭ 6, 7) ions react with silane and
yield the large clusters Si2C3Hϩnϩ2 via loss of a hydrogen
molecule characterised by low rate constants.
Hydrocarbon ions, formed in self-condensation processes of
allene, mainly reacted with germane by charge-exchange
processes to give ion products containing Ge and H only,
together with a few mixed ion clusters. To the contrary, al-
lene ions react with silane to yield a great number of differ-
ent products containing Si, C, and H atoms. These ion spe-
cies react further, preferentially with allene, in successive
steps, to give large clusters with increasing number of car-
bon atoms.
Finally, it appears clear that the cationic mechanism does
not give a high contribution to the growth of materials,
starting from gaseous mixtures in the GeH4/C3H4 system,
while it is more efficient in the SiH4/C3H4 one. Moreover,
in the silane/allene mixture, the largest cluster ions are gen-
erally formed in reactions of silicon containing ions with
allene and the highest yield are observed in mixtures con-
taining silane in excess.
The present results may be compared with those obtained
in the SiH4/C3H8 and SiH4/C3H6 systems previously
studied.[5] The reactivity of the SiH4/C3H8 system was
found to be remarkably low, and hence the comparison may
be limited to the SiH4/C3H6 and SiH4/C3H4 mixtures. Sil-
ane primary ions show a similar behaviour towards propene
and allene. In fact, reaction efficiencies are quite high and
the most favourable neutral losses are almost the same in
the two systems. However, when allene is the neutral re-
agent, the number of different reaction paths is high. In
contrast, reactions of the SiHnϩ (n ϭ 1–3) ions with propene
proceed through a lower number of paths, but generally
with higher rates. A similar trend is displayed by the
Si2Hnϩ (n ϭ 2–5) ions in the comparison of the two systems.
Because of the several different processes starting from sil-
ane ions, the number of Si, C, and H containing ion species
formed in the SiH4/C3H4 system is higher than in the
SiH4/C3H6 one. Among these, some ions are common to
the two systems: SiCHnϩ (n ϭ 4, 5), SiC2H5ϩ and SiC3H7ϩ.
Their reactivity towards silane is supposed to be the same;
nevertheless, reactions of SiCH5ϩ and SiC2H5ϩ ions could
not be studied in the SiH4/C3H4 system, the reason being
the lower abundance of the reacting ions in the SiH4/C3H4
mixture. In fact, SiCH5ϩ is formed through fewer processes
and with lower rate constants in the presence of allene than
with propene. SiC2H5ϩ has a higher number of precursors
in the SiH4/C3H4 system, but formation rate constants are
Experimental Section
Allene and silane were commercially supplied in high purity and
germane was prepared and purified as described in the literature.[22]
Prior to use, each of the reagent gases was dried by means of an-
hydrous sodium sulfate. Helium was supplied at an extra-high de-
gree of purity, and no further purification was necessary. The ex-
periments were performed on a ITMS Finnigan mass spectrometer
and the pressure was measured by a Bayard Alpert ionisation
remarkably low. Mixed ions, containing silicon, carbon gauge. The read pressures were corrected for the relative sensitivity
of the ion gauge with respect to different gases[23,24] and for a calib-
and hydrogen, that are present in both the systems react
ration factor, calculated as previously reported,[21] in order to get
with comparable efficiencies with allene and propene, the
the real pressure in the trap. Helium buffer gas was admitted to the
number of reaction paths being higher in the silane/allene
mixture as already observed for silane ions. Finally, sev-
trap at a pressure of about 4 ϫ 10–4 Torr (1 Torr ϭ 133 Pa), and
allene, germane and silane at pressures in the 2.0 ϫ 10–7–10.0 ϫ 10–7
eral hydrocarbon ions are common to the two mixtures:
Torr range. The trap temperature was 333 K in order to avoid
C2Hϩ3 , C3Hnϩ (n ϭ 1–5), rate constants of their reactions
with silane being in good agreement among the two sys-
tems.
thermal decomposition of reagents and to obtain results compar-
able with systems studied previously. In all experiments, ions were
detected in the mass range of 14–400 u . In order to prevent side
511
Eur. J. Inorg. Chem. 2000, 505Ϫ512

P. Benzi, L. Operti, R. Rabezzana
FULL PAPER
Chem. 1996, 509, 151. – [3d] G. Cetini, L. Operti, R. Rabezzana,
G. A. Vaglio, P. Volpe, J. Organomet. Chem. 1996, 519, 169. –
reactions with water background, the manifold and the lines for
introduction of reagent gases and helium were frequently baked-
out.
[3e]
P. Benzi, L. Operti, R. Rabezzana, M. Splendore, G. A.
Vaglio, Int. J. Mass Spectrom. Ion Processes 1996, 152, 61. –
[3f]
The scan modes used to isolate selected ion species or ranges of
ions for the determination of reaction mechanisms and rate con-
stants, as well as the procedures for calculations, were previously
described in detail.[3,21,25] Ionic species that were selectively isolated
from silane contained the 28Si isotope, whereas those isolated from
germane contained the 70Ge (Geϩ, GeHϩ) or 76Ge (GeH2ϩ, GeH3ϩ)
isotope. Selection of ions at a specific m/z value was performed
by using dc voltages and by resonance ejection, as in previous stud-
ies.[25] The latter method, applying rf voltages only, is expected to
cause a lower excitation of the selected ions with respect to the
method in which a dc voltage is used. The rate constants obtained
with the two different methods were very similar and this behaviour
is consistent with the hypothesis that ions are thermalized by unre-
active collisions with the buffer gas. The single exponential decays
of the abundance of the reacting ions observed in the kinetic experi-
ments further confirm this hypothesis.
Formation of ions was obtained with an electron beam at about
35 eV for ionisation times in the range 1–20 ms. A reaction time
suitable to maximise the abundance of ions to be stored was applied
after the ionisation event. Isolation of the selected ions, their reac-
tions with neutrals present in the trap for convenient reaction times
(50 ms for rate constants calculation and up to 2 s for mechanisms
determination) and acquisition were the successive steps.
In the silane/allene mixture, where Si is isobaric with C2H4, some
experiments were performed by introducing one reagent gas via a
pulsed valve in order to determine the neutral species involved in
the ion/molecule reaction under examination and, as a con-
sequence, the composition of the product ion. A General Valve
Corporation’s Iota One pulse valve was used,[26] and the experi-
mental procedures have been already described.[4] Briefly, the intro-
duction into the ion trap of one of the reactants through a pulsed
valve allows us to react selected ions with only the other gas which
is continuously leaked into the trap. The pressure of the gas to be
pulsed was set behind the valve to about 1 ϫ 10–6 Torr, the valve
opening time was 0.1 ms and the time required for the pressure of
the pulsed gas to decrease to zero was about 1.5 s.
P. Antoniotti, L. Operti, R. Rabezzana, G. A. Vaglio, Int.
J. Mass Spectrom. 1999, 182/183, 63.
[4] [4a]
P. Benzi, L. Operti, G. A. Vaglio, P. Volpe, M. Speranza,
R. Gabrielli, Int. J. Mass Spectrom. Ion Processes 1990, 100,
[4b]
647. –
L. Operti, M. Splendore, G. A. Vaglio, P. Volpe, M.
Speranza, G. Occhiucci, J. Organomet. Chem. 1992, 433, 35.
P. Antoniotti, L. Operti, R. Rabezzana, G. A. Vaglio, P. Volpe,
Int. J. Mass Spectrom. 1999, 190/191, 243.
P. Antoniotti, C. Canepa, L. Operti, R. Rabezzana, G.
Tonachini, G. A. Vaglio, J. Organomet. Chem., 1999, 589, 150.
P. Antoniotti, C. Canepa, L. Operti, R. Rabezzana, G.
Tonachini, G. A. Vaglio, J. Phys. Chem. A 1999, 103, 10945.
M. T. Bowers, Gas Phase lon Chemistry, Academic Press, New
York, 1979.
[5]
[6]
[7]
[8]
[9]
S. G. Lias, J. E. Bartmess, J. F. Liebman, J. L. Holmes, R.
D. Levin, W. G. Mallard, J. Phys. Chem. ref. Data Suppl. 1
1988, 17.
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
B. Ruscic, M. Schwarz, J. Berkowitz, J. Chem. Phys. 1990, 92,
1865.
R. Bakhtiar, C. M. Holznagel, D. B. Jacobson, J. Phys. Chem.
1993, 97, 12710.
B. H. Boo, P. B. Armentrout, J. Am. Chem. Soc. 1991, 113,
6401.
B. H. Boo, P. B. Armentrout, J. Am. Chem. Soc. 1987, 109,
3549.
S. Wlodek, A. Fox, D. K. Bohme, J. Am. Chem. Soc. 1991,
113, 4461.
B. H. Boo, J. L. Elkind, P. B. Armentrout, J. Am. Chem. Soc.
1990, 112, 2083.
B. L. Kickel, E. R. Fisher, P. B. Armentrout, J. Phys. Chem.
1992, 96, 2603.
S. K. Shin, R. R. Corderman, J. L. Beauchamp, Int. J. Mass
Spectrom. Ion Processes 1990, 101, 257.
[18] [18a]
J. J. Myher, A. G. Harrison, J. Phys. Chem. 1968, 72,
1905. – [18b] M. T. Bowers, D. D. Elleman, R. M. O’Malley, K.
[18c]
R. Jennings, J. Phys. Chem. 1970, 74, 2583. –
A. Nato, M.
Niwa, K. Honma, I. Tanaka, Int. J. Mass Spectrom. Ion Phys.
[18d]
1980, 34, 287. –
C. Lifshitz, Y. Gleitman, S. Gefen, U.
Shainock, I. Dotan, Int. J. Mass Spectrom. Ion Phys. 1981, 40,
[18e]
1. –
C. Lifshitz, Y. Gleitman, Int. J. Mass Spectrom. Ion
Phys. 1981, 40, 17. – [18f] V. G. Anicich, G. A. Blake, J. K. Kim,
M. J. McEwan, W. T. Huntress Jr. J. Phys. Chem. 1984, 88,
[18g]
4608. –
Ion Processes 1984, 59, 49.
D. Carrier, J. A. Herman, Int. J. Mass Spectrom.
[19]
[20]
K. J. Miller, J. Am. Chem. Soc. 1990, 912, 8533.
E. R. Lippincott, J. M. Stutman, J. Phys. Chem. 1964, 68, 2926.
[21] [21a]
P. Benzi, L. Operti, G. A. Vaglio, P. Volpe, M. Speranza,
Acknowledgments
Authors wish to thank MURST and Universita di Torino for finan-
cial support to this research.
R. Gabrielli, J. Organomet. Chem. 1988, 354, 39. – [21b] L. Oper-
ti, M. Splendore, G. A. Vaglio, A. M. Franklin, J. F. J. Todd,
Int. J. Mass Spectrom. Ion Processes 1994, 136, 25.
P. Benzi, M. Castiglioni, P. Volpe, L. Battezzati, M. Venturi,
Polyhedron 1988, 7, 597.
`
[22]
[23]
[1]
M. Decouzon, J.-F. Gal, P.-C. Maria, A. S. Tchinianga, per-
sonal communication.
J. G. Ekerdt, Y.-M. Sun, A. Szabo, G. J. Szulczewski, J. M.
White, Chem. Rev. 1996, 96, 1499.
[24]
[25]
[2]
J. E. Bartmess, R. M. Georgiadis, Vacuum 1983, 33, 149.
P. Antoniotti, L. Operti, R. Rabezzana, M. Splendore, G.
Tonachini, G. A. Vaglio, J. Chem. Phys. 1997, 107, 1491.
T. J. Carlin, B. S. Freiser, Anal. Chem. 1983, 55, 571.
Received July 30, 1999
P. Hess, Electron. Mat. Chem. 1996, 127, 169.
[3] [3a]
J.-F. Gal, R. Grover, P.-C. Maria, L. Operti, R. Rabezzana,
[3b]
G. A. Vaglio, P. Volpe, J. Phys. Chem. 1994, 98, 11978. –
P.
[26]
Antoniotti, L. Operti, R. Rabezzana, G. A. Vaglio, J.-F. Gal,
[3c]
R. Grover, P.-C. Maria, J. Phys. Chem. 1996, 100, 155. –
L.
Operti, R. Rabezzana, G. A. Vaglio, P. Volpe, J. Organomet.
[I99284]
512
Eur. J. Inorg. Chem. 2000, 505Ϫ512