Kinetics of silylene insertion into N-H bonds and the mechanism and kinetics of the pyrolysis of dimethylsilylamine

Article abstract of DOI:10.1021/jp981436s

Product and kinetic data on the copyrolysis of silane with ammonia and of silane with dimethylamine are reported. The main products of the SiH₄/NH₃ reaction are disilane and silylamine, and the main products of the SiH₄/Me₂NH reaction are Si₂H₆ and Me₂NSiH₃. The silylamine products are produced by SiH₂ insertions into the N-H bonds of the amine reactants, and rate constants for this insertion lie in the k₂ ～ 2 × 10⁹ M^-1 s^-1 range, with no temperature dependence within the errors. Kinetic and product data on the pyrolysis of dimethysilylamine are also presented. Reactant loss followed reasonable first-order kinetics and produced rate constants consistent with Arrhenius parameters of log A(sec^-1) = 12.64 ± 0.98 and E = 48.02 ± 3.1 kcal/mol. A mechanism of reaction, based on the kinds of silylene intermediate reactions expected, is presented. Comparisons of mechanistic expectations with the product data, however, indicate a more complex reaction system.

Full text of DOI:10.1021/jp981436s

J. Phys. Chem. A 1998, 102, 8493-8497
8493
Kinetics of Silylene Insertion into N-H Bonds and the Mechanism and Kinetics of the
Pyrolysis of Dimethylsilylamine
H. E. O’Neal,* M. A. Ring, J. G. Martin, and M. T. Navio
The Department of Chemistry, San Diego State UniVersity, San Diego, California 92182
ReceiVed: March 10, 1998; In Final Form: June 9, 1998
Product and kinetic data on the copyrolysis of silane with ammonia and of silane with dimethylamine are
reported. The main products of the SiH4/NH3 reaction are disilane and silylamine, and the main products of
the SiH4/Me2NH reaction are Si2H6 and Me2NSiH3. The silylamine products are produced by SiH2 insertions
into the N-H bonds of the amine reactants, and rate constants for this insertion lie in the k2 ∼ 2 × 109 M-1
s-1 range, with no temperature dependence within the errors. Kinetic and product data on the pyrolysis of
dimethysilylamine are also presented. Reactant loss followed reasonable first-order kinetics and produced
rate constants consistent with Arrhenius parameters of log A(sec-1) ) 12.64 ( 0.98 and E ) 48.02 ( 3.1
kcal/mol. A mechanism of reaction, based on the kinds of silylene intermediate reactions expected, is presented.
Comparisons of mechanistic expectations with the product data, however, indicate a more complex reaction
system.
Introduction
This nonequilibration effect was not a problem in the slower,
lower temperature reactions, but did present some difficulties
The deposition of thin films of silicon nitride (Si3N4) is very
important to device fabrication in the microelectronics industry
in which Si3N4 films are used as a capping material. Although
most silicon nitride films are deposited via the reaction of SiH4
with NH3, it appears that the basic details of the gas-phase
chemistry between SiH4 and NH3 are not known. In this paper,
we present results from investigations the aim of which was to
provide some understanding of the gas phase thermal processes
occurring in the reaction of SiH4 with NH3 and higher
alkylamines, e.g., Me2NH. Silylamine is the initial product of
the SiH4/NH3 reaction, therefore it was also of interest to study
the decomposition of this substance. However, because sily-
lamine has never been isolated, from which one can conclude
that it is either thermally or reactively unstable under normal
temperature (T) and pressure (P) conditions, we studied instead
the decomposition of its closest stable analogue, silyldimethy-
lamine (Me2NSiH3).
in the data analysis of the faster reactions.
Mass spectra from authentic samples of most of the species
involved in these decompositions, namely of SiH4, Si2H6, Si3H8,
NH3, Me2NH, Me2NSiH3, and (Me2N)2SiH2, were obtained so
that mass spectra of the reaction products could be converted
into product concentrations via sensitivity studies. Because
authentic samples of silylamine were not available, its concen-
trations could not be so determined. Instead, silylamine was
followed relative to disilane via MS peak ratios.
Me2NSiH3 was prepared by reaction of SiH3Br and Me2NH
as described by Sujishi and Witz.2 (Me2N)2SiH2 was prepared
by the same procedure from SiH2Br2 and Me2NH. Both
compounds were purified by trap-to-trap distillations under
reduced pressure conditions and identified by their mass spectra
(Table 1). The bromosilane reactants were prepared by reaction
of SiH4 and AgBr as described in an earlier paper.3
Results
Experimental Section
The pyrolysis of SiH4 with NH3, of SiH4 with Me2NH, and
of pure Me2NSiH3 were carried out in a heated 70.4 cm3
cylindrical quartz reactor with direct capillary leak to a
quadrupole mass spectrometer (MS). Product concentrations
were determined from mass spectra obtained as a function of
time. Up to five peaks were sampled sequentially in any scan
and scan intervals were set at 7, 14, or 25 s depending on
temperature. Details of the apparatus and of the operational
procedures have been described.1
One of the problems peculiar to the Me2NSiH3 decomposition
was a slow equilibration of the amine reactant pressures on
expansion into the reaction cell. Thus, for example, although
MS peaks for Argon (the internal analytical standard of these
studies) reached maximum values within a few seconds, times
for the reactant dimethylsilylamine peaks to maximize were on
the order of 70 s. Selective absorption of the amine on the
walls of the mass spectrometer seems the likely explanation.
The Reaction of SiH4 with NH3. Because the MS sensitivity
of silylamine could not be determined, the reaction of SiH4 with
NH3 was only studied under one set of conditions: with a SiH4/
NH3 ) 1 reactant mixture at 400 °C and a total pressure of 100
Torr. Initial products produced three multiplets, which in
decreasing order of intensity (most intense in parentheses) were
assigned as follows: Si2H6, m/e ) 56-62 (58); SiH3NH2, m/e
) 42-47 (44); (SiH3)2NH or SiH3SiH2NH2, 70-77 (74); and
Si3H8 with peaks at the detection limit around m/e ) 90. The
silane and ammonia reactants were monitored at m/e ) 30 and
31, and m/e ) 16 and 17, respectively, silylamine was monitored
at m/e ) 44 and 47, and disilane was monitored at m/e ) 60.
A representative plot of the primary product data for this system,
namely the ratio of silylamine to disilane (measured by their
MS peak heights) as a function of time, is shown in Figure 1.
The reason for the signal noise and instability of measurements
made at long reaction times is unknown. Contributions of
10.1021/jp981436s CCC: $15.00 © 1998 American Chemical Society
Published on Web 07/31/1998

8494 J. Phys. Chem. A, Vol. 102, No. 44, 1998
O’Neal et al.
TABLE 1: Mass Spectra of (Me2N)2SiH2 and Me2NSiH3
TABLE 2: Reactant Loss and Product Formation Data of
the Dimethylamine/Silane Reactiona,b
relative abundance
reactant and product concentrations in M × 103
m/e
(Me2N)2SiH2
Me2NSiH3
ion
T (°C) time (s)
SiH4
Me2NH
Si2H6
Me2NSiH3
+
118
117
77
76
75
74
73
72
60
58
56
45
44
42
31
30
11.1
3.4
(Me2N)2SiH2
398c
421d
452e
0
200
400
600
800
1200
0
49
98
147
196
294
0
1.42
1.37
1.33
1.31
1.29
1.26
1.60
1.38
1.36
1.34
1.32
1.28
1.40
1.22
1.14
1.06
0.94
0.87
3.54
4.05
4.05
4.07
4.08
4.06
4.68
4.38
4.40
4.37
4.33
4.28
4.39
3.80
3.56
3.40
3.13
2.96
(Me2N)2SiH+
Me2NSi30H3
+
0.011
0.021
0.026
0.029
0.032
0.007
0.027
0.049
0.077
0.117
6.6
9.9
60.5
100.0
19.7
42.8
7.4
23.7
7.4
41.1
60.5
68.4
20.9
9.9
+
4.4
20.8
73.6
25.6
20.0
8.4
25.2
8.0
47.2
100.0
47.2
8.0
Me2NSi29H3
+
Me2NSiH3
Me2NSiH2
+
Me2NSiH+
Me2NSiH+
+
0.012
0.022
0.028
0.032
0.037
0.007
0.024
0.050
0.078
0.13
MeNSiH3
MeNSiH+
CH2NSi+
Me2NH+
Me2N+
C2H4N+
+
14
28
56
140
210
0.014
0.021
0.031
0.046
0.052
0.017
0.041
0.106
0.24
SiH3
SiH2
+
8.0
unidentified, higher molecular weight products, particularly to
the m/e ) 47 peak, is a plausible explanation. Fortunately, data
of the important regimes of reaction were far better behaved.
The Reaction of SiH4 with Me2NH. The reaction of SiH4
with Me2NH in an initial reactant ratio of SiH4/Me2NH ) 0.36,
was studied at three temperatures, 398, 422, and 453 °C, and at
total pressures between 148 and 350 Torr. Product mass spectra
again showed a series of multiplets which were assigned as
follows: Si2H6, m/e ) 56-62 (58); Me2NSiH3, m/e ) 73-75
(74); Si3H8, m/e ) 84-92 (89); SiH3SiH2NMe2 or possibly
0.27
a Reaction mixture composition was Me2NH/SiH4/Ar ) 3.84/1.39/
1. b Not all data are shown. Actual data intervals were 25 s at 398 °C
and 7 s at the other two temperatures. c Runs at 398 °C were done at
total pressures around 297 Torr. d Runs at 421 °C were done at total
pressures around 320 Torr. e Runs at 452 °C were done at total pressures
around 144 Torr.
420, and 450 °C, and at total pressures between 40 and 100
Torr. Data of a typical run are shown in Table 3. In decreasing
order of importance the major products were (Me2N)2SiH2 (m/e
) 118), H2 (m/e ) 2), and SiH4 (m/e ) 30,31). Reactant loss
followed reasonable first-order kinetics over most of the reaction
and least squares treatments of the data in the linear regions of
the first-order plots yielded the following: k ) 9.02 ( 1.51 ×
10-4 s-1 at 393 °C (4 runs); k ) 3.13 ( 0.64 × 10-3 s-1 at 422
°C (10 runs); k ) 1.39 ( 0.63 × 10-2 s-1 at 450 °C (4 runs).
A least squares treatment of these rate constants gives: log
A(sec-1) ) 12.64 ( 0.98 and E ) 48.02 ( 3.1 kcal/mol.
CH2SiH2CH2N-SiH3, m/e ) 101-105 (103); and (Me2N)2SiH2,
m/e around 118. Dimethylamine was monitored at m/e ) 45,
silyldimethylamine was monitored at m/e ) 75, and bis-
dimethylaminosilane was monitored at m/e ) 118. Selective
data for these studies are shown in Table 2; actual data collection
was at more frequent intervals: 25 s intervals for the 398 °C
runs and 7 s intervals for runs at the higher two temperatures.
The Me2NSiH3 Pyrolysis. The decomposition of Me2NSiH3
in a reactant mixture Me2NSiH3/Ar ) 5 was studied at 393,
Figure 1. Silylamine/disilane product ratios in time for the SiH4/NH3 reaction. ~ SiH3NH2 (m/e ) 44)/Si2H6 (m/e ) 60); + SiH3NH2 (m/e )
47)/Si2H6 (m/e ) 60).

Silylene Insertion into NsH Bonds
J. Phys. Chem. A, Vol. 102, No. 44, 1998 8495
TABLE 3: Representative Data of the Me2NSiH3 Pyrolysis
disilane steady state is of most interest here because in this stage
silylamine products continue to form without significant loss
via their decomposition processes. Consequently, silylamine/
disilane product ratios rise more or less linearly in time. That
both reaction systems (SiH4/NH3 and SiH4/Me2NH) go through
this stage is apparent in the plots of Figure 1 and data of Table
2. Only in the latter stages of reaction do the silylamine
formation rates begin to plateau. In addition, the degree to
which “plateauing” occurs seems to vary with temperature and
with system. Thus, a falloff in silylamine production rates is
very evident in the silane/ammonia system (see Figure 1 for
the late stage leveling and completion of the sigmoidal behavior
of the SiH3NH2/Si2H6 ratios in time), but is only apparent at
the highest temperature and at the highest silane conversions
in the silane/dimethylamine reaction system. By the mechanism,
silylamine formation rates slow when silylamine decompositions
(reactions 3 and 4) become important.
(T ) 421 °C)
concentrations in M × 105
a
t (s)
[A]exp
[A]calc
[H2]
[SiH4]
[(Me2N)2SiH2]
0
14
28
42
56
104
0
0
0
4.0
6.6
8.3
64.0b
77.4b
81.4b
80.9b
78.7
65.4
54.2
40.2
32.8
30.1
23.4
21.6
19.5c
98.5
93.2
88.2
83.5
79.1
63.5
51.0
41.0
32.9
29.5
23.7
21.3
17.1
2.0
2.9
3.4
3.6
3.8
5.6
5.2
6.0
6.8
7.0
7.6
7.9
8.3
0.7
1.0
1.2
1.3
1.3
1.6
1.6
1.6
1.6
1.6
1.6
1.6
1.5
9.7
70
10.9
14.2
16.4
17.5
18.0
18.0
17.4
17.3
16.2
126
182
238
294
322
378
406
462
a These reactant concentrations have been calculated from the least
Rate expressions for silylamine and disilane formations are
given by eqs 1 and 2, respectively.
squares evaluated rate constant of the reaction, namely k ) 3.91 ×
10-3 s-1 b These anamously low reactant concentration measurements
.
are attributed to absorption of the reactant on the walls of the mass
spectrometer. c The anomalously high reactant concentration measure-
ment is attributed to high molecular weight contributions to the MS
peak at m/e ) 118 used to monitor the reactant.
d[SiH3NR2]/dt ) k2[SiH2][NHR2] - (k3 + k4)[SiH3NR2]
(1)
d[Si2H6]/dt ) k5[SiH2][SiH4] - (k6 + k7)[Si2H6] (2)
Discussion
If it were possible to establish the initial flat region of the
silylamine/disilane product ratio, a direct measure of the desired
k5/k2 rate constant ratio would result. However, this reaction
stage is too short and product measurement errors too large to
analyze reliably. However, the subsequent reaction stage, i.e.,
the stage after disilane reaches steady state but before silylamine
products decompose significantly, can be meaningfully analyzed.
For this stage, eqs 1 and 2 give equation 3.
The SiH4/NH3 and SiH4/Me2NH Reactions. The reactions
of silane with ammonia and of silane with dimethylamine occur
in the same temperature range and at comparable rates as the
pure silane decomposition.4 It is therefore reasonable to expect
that both reactions are initiated by the silane decomposition and
that subsequent reactions are driven by silylene insertion/
elimination reactions of the “usual” type. We therefore propose
the mechanism of Scheme 1 to describe the gas-phase behaviors
Scheme 1: Possible Mechanism of a
Silane/Amine Copyrolysis
dQ ) [k2(k6 + k7)/k5]dt
(3)
where Q ) ([SiH3NR2]/[Si2H6]) × ([SiH4]/[NHR2]).
SiH4 + (M) 9
18 SiH2 + H2 + (M)
Thus, a plot of Q ) vs t for data past the disilane maximum
but prior to significant silylamine product decomposition should
be linear with an intercept of zero and a slope of k2(k6 + k7)/k5.
As predicted, Q vs t plots of the SiH4/Me2NH system data were
indeed linear for reactions at the lower two temperatures, but
as indicated above, these plots for runs at the highest temperature
showed deviations from linearity at high silane conversions, see
Figure 2.
2
SiH2 + NHR2 S SiH3NR2
9
48 products
R ) H, Me
9
78 SiH3SiH + H2
3
5
SiH2 + SiH4 S Si2H6
6
8
10
SiH2 + Si2H6 S Si3H8 S SiH3SiH + SiH4
Values of k2(k6 + k7)/k5 determined from slopes of the linear
regions of the Q vs t plots of the silane/dimethylamine system
are given in Table 4. Values for this same rate constant ratio
relative to the silane/ammonia system are also shown. The
latter, obtained from the slopes of the linear regions of the SiH3-
NH2 (m/e ) 47 or 44)/Si2H6 (m/e ) 60) product ratio vs time
plots and the average reactant ratios, could be in error by a factor
of 2 or more because of the uncertainties in the SiH3NH2 MS
peak sensitivities.
9
11
of these two systems.
Scheme 1 is consistent with the observed products and with
the time frame of their formation. Thus, for example, the
sigmoidal shape of the SiH3NH2/Si2H6 vs time plots of Figure
1 are completely consistent with Scheme 1. Initially, back
reactions 3 and 6 are unimportant and production rates of
silylamine and disilane are constant, hence a plot of their ratio,
i.e., [m/e ) 47 or 44]/[m/e ) 60], is flat. This stage is very
short-lived primarily because of an early onset of reaction 6. In
the pure silane decomposition,4 pseudo equilibrium conditions
in reactions 5 and 6 (as well as 8-11) are established in the
first 3-5% of silane conversion, with disilane reaching a
maximum concentration which then slowly decays in accord
with silane loss. In the present system, disilane also rapidly
reaches a pseudo maximum condition but its concentration then
gradually drifts up in time. This could be an artifact of higher
molecular weight product contributions to the m/e ) 60 mass
peak. The period immediately following establishment of the
Because Arrhenius parameters of k6 and k7, including the
falloff behavior of reaction 6, are known5,6 and rate constants
7
for k5 under our reaction conditions can be estimated from
the falloff of reaction 6, 5 values for k2/k5 and k2 for both systems
have been evaluated (Table 4 columns 5 and 7). Although the
data at 398 °C suggest an inverse pressure dependence on the
k2/k5 ratio, the pressure range and rather small falloff calculated
for k5 do not warrant the factor of 2 change determined in the
ratio. It seems likely that the k2/k5 ratio spread is simply due
to experimental error. Within the errors, k2/k5 ∼ 2 × 10-2 and
k2 ∼ 2 × 109 M-1 s-1 for the silane/dimethylamine system,

8496 J. Phys. Chem. A, Vol. 102, No. 44, 1998
O’Neal et al.
TABLE 4: Evaluations of Relative and Absolute Rate Constants of SiH2 Insertions into NsH Bonds
SiH4 + Me2NH Reaction System
slopea
k6/k6(∞)
k2/k5
k2/k5
k2 (M-1 s-1
)
b
T (°C)
P (Torr)
398
297f
148
193
310
285f
361f
320
230
209
278
262f
9.4 × 10-4
1.5 × 10-3
1.5 × 10-3
8.4 × 10-4
3.6 × 10-3
3.8 × 10-3
3.2 × 10-3
1.1 × 10-2
2.1 × 10-2
2.1 × 10-2
1.7 × 10-2
0.95
0.89
0.90
0.96
0.94
0.95
0.95
0.88
0.85
0.90
0.89
1.7 × 10-2
2.9 × 10-2
2.8 × 10-2
1.5 × 10-2
1.7 × 10-2
1.7 × 10-2
1.5 × 10-2
1.1 × 10-2
2.1 × 10-2
2.0 × 10-2
1.7 × 10-2
2.2 ( 0.6 × 10-2 g
2.7 ( 0.6 × 109
422
453
1.6 ( 0.1 × 10-2 g
1.7 ( 0.4 × 10-2 g
2.0 ( 0.1 × 109
1.9 ( 0.4 × 109
SiH4 + NH3 reaction system
slopea
k6/k6(∞)
k2/k5
k2/k5
k2 (M-1 s-1
)
b
T (°C)
P (Torr)
400
100d
100d
100e
100e
5.9 × 10-4
7.4 × 10-4
1.7 × 10-3
1.8 × 10-3
0.78
0.78
0.78
0.78
1.1 × 10-2
1.4 × 10-2
3.2 × 10-2
3.4 × 10-2
1.25 ( 0.15 × 10-2 g
1.3 ( 0.4 × 109
3.3 ( 0.1 × 109
3.30 ( 0.1 × 10-2 g
cal/RT
a Slopes of Q vs t plots ) k2(k6 + k7)/k5 where k6 ) k6(∞) × k6/k6(∞)
.
b k6(∞) ) 1015.75 × e-52200
s-1 (ref 5); k5(∞) ) 1.3 × 1011 M-1 s-1 (ref
7); k6(10 Torr) ) 1015.13 x e-50922 cal/RT s-1 (ref 5); k6(150 Torr) ) 1015.38 × e-51300 cal/RT s-1 (ref 5); k6(500 Torr) ) 1015.67 × e-51800 cal/RT s-1 (ref 5); k7 ) 1015.81
× e-56310
s-1 (ref 6). (k6/k6(∞)) and (k5/k5(∞)) obtained by interpolation from the above parameters. c Slopes of (SiH3NH2/Si2H6) vs t plots )
cal/RT
k2/k5(k6 + k7)(NH3/SiH4). d Values shown are based on (m/e ) 47)/(m/e ) 60) data. e Values shown are based on (m/e ) 44)/(m/e ) 60) data.
f These are the runs used in the Figure 2 plots. g Average values.
and possibly at comparable rates. Calculations by Melius and
Ho9 for the back reaction, i.e., the SiH3NH2 decomposition, also
support this form of reaction.
From the SiH3NH2/Si2H6 product behavior of the silane/
ammonia system (see Figure 1, for example), it is clear that the
steady-state condition for silylamine is reached at 400 °C and
high silane conversions. For this condition, eqs 1 and 2 lead
to eq 4.
(k3 + k4) ) [(k2/k5)(k6 + k7)(NH3/SiH4] × [Si2H6/SiH3NH2]
(4)
Values of the terms in the first bracket on the right-hand side
(RHS) follow from the slopes of the linear regions of silylamine/
disilane vs t plots and the ratio of the reactant concentrations
(see eq 3 or Table 4), and values of the second bracket on the
RHS follow from the disilane/silylamine product ratios after
they reach constant values, i.e., steady-state conditions for both
products. Our SiH4/NH3 system data give average values for
the terms of the first bracket of 6.7 ( 0.7 × 10-4 and 1.7 (
Figure 2. Q vs t plots of the SiH4/Me2NH reaction. Q ) (Me2NSiH3/
0.7 × 10-3 (m/e ) 47 or m/e ) 44 for silylamine, respectively),
Si2H6)/(SiH4/Me2NH). ~ Partial data for a representative run at 398
°C. 2 and .: Data from two representative runs at 422 °C. b: Partial
and respective values for the terms of the second bracket are
1.1 ( 0.1 and 0.5 ( 0.05. Thus, silylamine decomposes with
a composite rate constant (k3 + k4) of about 8.0 ( 0.6 × 10-4
s-1 at 400 °C. Because uncertainties in the silylamine MS
sensitivities cancel in the product of the two RHS bracketed
terms, this result is a fairly solid one. It appears, then, that
silylamine decomposes at rates very similar to those of
dimethylsilylamine. Thus, from the Arrhenius parameters for
the latter’s rate constant, k(Me2NSiH3) ) 11.1 × 10-4 s-1 at
400 °C.
data from a representative run at 453 °C. The t and Q scales are
respectively too short to show all the data points of the runs at 398
and 453. Nevertheless, beginning of curvatures in the 453 °C data plot
is apparent. Data used for the plots are from the runs in Table 4.
with a possible negative activation energy for k2. Similar values
were obtained for the silane/ammonia system rate constants
(Table 4), although these are subject to the uncertainties
previously mentioned.
The k2 values determined here are close to those found by
Walsh et al.8 for the reaction of Me2Si with MeOH, namely k
) 7.1 × 108 M-1 s-1. According to Walsh,8 the latter reaction
involves an oxygen nonbonding electron pair attack on the
empty π orbital of silylene to form a dative bonded complex
which then moves to a transition state involving H atom transfer
from O to Si. Amine substrates, with their single nonbonding
electron pair, should react with silylenes in a similar manner
The Me2NSiH3 Pyrolysis. The Table 3 data illustrate several
of the problems inherent to this system: slow equilibration of
the reactant amine in the early stages of reaction, tailing-off of
reactant loss in the final stages of reaction, and poor mass
balances overall. The former is most pertinent to establishing
the reaction kinetics and the latter is most pertinent to establish-
ing the reaction mechanism. We have attributed the former
behavior to reactant absorption on the mass spectrometer walls,

Silylene Insertion into NsH Bonds
J. Phys. Chem. A, Vol. 102, No. 44, 1998 8497
and it is reasonable to believe that the reactant loss tail-off effect
is an artifact of higher molecular weight product contributions
to the m/e ) 118 peak.
SiH2. This shift should be facilitated by the nonbonding electron
pair on nitrogen. Similarly facilitated disilane decompositions
are the Cl atom10 and 1,2 OMe group shift processes.11 The
latter are among the fastest disilane decompositions known.
Reaction 15 should therefore also be fast.
Reactions at the two lower temperatures were relatively slow;
the early stage nonequilibration of reactant was not a problem
in determining the kinetics. Thus, one can see that the middle
temperature run of Table 3 followed good first-order loss
kinetics over 60% of the conversion range (between 20 and
80%), long enough to obtain a fairly accurate measure of the
rate constant. Runs at the lowest temperature were even better
behaved and yielded correspondingly better results. However,
for runs at the highest temperature, reactant equilibration
required almost a full half-life of reaction. Consequently, their
rate constant errors are relatively large and this is reflected in
correspondingly large errors of the reaction’s Arrhenius param-
eters.
Reactions 14-17 constitute a chain, and this is one of the
major problems with Scheme 2. It predicts a 50% yield of
(Me2N)2SiH2 and a combined H2 and SiH4 yield equal to that
of (Me2N)2SiH2. The data never meet the latter prediction and
only meet the former (within the errors) in the very early stages
of reaction. At longer times, (Me2N)2SiH2 yields are well below
50%. It is interesting to note that the bis and silane product
concentrations rise through maxima and then slowly decay,
suggesting that both species have decomposition rates compa-
rable to that of the reactant. This is clearly true for silane,4
and we have found, through several test pyrolyses of the bis
product, that it is also true for that substance. Because the
expected decompositions of silane and the bis product both
produce hydrogen, the low hydrogen yields are particularly
difficult to understand. We can only offer the two surface
reactions below as a possible explanation.
As for the mechanism of the Me2NSiH3 pyrolysis, a typical
silane gas-phase reaction scheme like that shown in Scheme 2
Scheme 2: Possible Mechanism of the
Silyldimethylamine Pyrolysis
Me2NSi2H5
9
188 Me2NSi2H5 (w) f ?
H3SiNMe2
9
128 HSiNMe2 + H2
138 Me2NH + SiH2
148 Me2NSiH2SiH2NMe2
158 (Me2N)2SiH2 + SiH2
168 Si2H5NMe2
178 SiH4 + HSiNMe2
18 SiH2 + H2 + (M)
Me2NSiH2SiH2NMe2
9
188 Me2NSiH2SiH2NMe2 (w) f ?
9
Here, aminodisilanes absorb and react on the surface without
formation of gas-phase products. Similar processes appear to
occur in the Si3H8 decomposition.12
HSiNMe2 + H3SiNMe2
Me2NSiH2SiH2NMe2
SiH2 + H3SiNMe2
9
9
Acknowledgment. The authors wish to thank the National
Science Foundation for support of this research under grant no.
CHE-9202437.
9
References and Notes
Si2H5NMe2
9
(1) Walker, K. L.; Jardine, R. E.; Ring, M. A.; O′Neal, H. E. Int. J.
Chem. Kinet. 1998, 30, 69.
(2) Sujishi, S.; Witz, S. J. Am. Chem. Soc. 1954, 76, 4631.
(3) Hollandsworth, R. P.; Ingle, W. M.; Ring, M. A. Inorg.Chem. 1967,
6, 844.
(M) + SiH4
9
(4) Purnell, H. J.; Walsh, R. Proc. R. Soc. Ser. A 1966, 293, 543. White,
R. T.; Espino-Rios, R.L.; Rogers, D. S.; Ring, M. A.; O’Neal, H. E. Int. J.
Chem. Kinet. 1985, 17, 1029.
(5) Martin, J. G.; Ring, M. A.; O’Neal, H. E. Int. J. Chem. Kinet. 1987,
19, 715.
(6) Dzarnoski, J.; Rickborn, S. F.; O’Neal, H. E.; Ring, M. A.
Organometallics 1982, 1, 1217; Moffat, H. K.; Jensen, K. F.; Carr, R. W.
J. Phys. Chem. 1992, 96, 7695.
(7) Becerra, R.; Frey, H. M.; Mason, B. P.; Walsh, R.; Gordon, M. S.
J. Am. Chem. Soc. 1992, 114, 2751.
(8) Baggott, J. E.; Blitz, M. A.; Frey, H. M.; Lightfoot, P. D.; Walsh,
R. Int. J. Chem. Kinet. 1992, 24, 127.
(9) Melius, C. F.; Ho, P. J. Phys. Chem. 1991, 95, 1410.
(10) Chernyshev, E. A.; Kamalenkov, N. G.; Bashkerova, S. A. Zh.
Obshch. Khim. 1971, 41, 1175.
(11) Atwell, W. H.; Weyenberg, D. R. J. Organomet. Chem. 1966, 5,
594; J. Am. Chem. Soc. 1968, 90, 3438.
(12) Ring, M. A.; O’Neal, H. E. J. Phys. Chem. 1992, 96, 10848.
seems likely.
Paralleling the silane decomposition,4 the initial steps of the
silyldimethylamine pyrolysis (reactions 12 and/or 13) produce
silylenes. These in turn should rapidly insert into the Si-H
bond of the reactant to produce disilanes (reactions 14 and 16).
Ab initio calculations of Melius and Ho9 on the SiH3NH2
decomposition suggest that the enthalpy requirement for reaction
13 is considerably higher than that of reaction 12, hence barring
any major surprises in their back reaction activation energies,
it is likely that k12 > k13. If this is the case, then the very low
yields of H2 observed would require the overall decomposition
to be a long chain process. The unique feature of Scheme 2 is
reaction 15, a 1,2-Me2N shift disilane decomposition to produce
the main reaction product: bis dimethylaminosilane or (Me2N)2-