Studies of the thermal and photochemical decomposition mechanisms of Bu(t)2Se in the presence or absence of Me2Zn using deuterium labelled analogues

Article abstract of DOI:10.1039/a904270i

Thermal decomposition of Bu(t)₂Se in helium produces 2-methylpropene and 2-methylpropane (8:1) and no volatile selenium containing products, apart from the element. In hydrogen, the products are the same (8:1 ratio of 2- methylpropene: 2-methylpropane). However, the 2-methylpropane produced by decomposing d¹⁸-Bu(t)₂Se in H₂ contains significant amounts of hydrogen (d⁹:d¹⁰=2.6:1). In the codecomposition of d⁰- and d¹⁸-Bu(t)₂Se in helium, d⁰, d¹, d⁹ and d¹⁰-2-methylpropane are all formed, but d⁹- Bu(t)₂Se is not a product. Photolysis of d⁰- and d¹⁸-Bu(t)₂Se in helium at room temperature produces 2-methylpropane and 2-methylpropene (1:1) with the 2-methylpropane being d⁰, d¹, d⁹ and d¹⁰ · d⁹-Bu(t)₂Se is also formed along with small amounts of 2,2,3,3-tetramethylbutane and d⁰- and d¹⁸-Bu(t)₂Se₂. Thermal codecomposition of Me₂Zn with d¹⁸-Bu(t)₂Se in H₂ produces do and d₁-methane and d₈-2-methylpropene together with traces of d⁸-2-methylpropane. These results are interpreted as indicating that thermally, Bu(t)₂Se decomposes via homolytic cleavage of the Se-C bonds and that the free Bu(t ·) so formed initiates a radical chain reaction involving H(^.) as the chain carrier through its reaction with intact Bu(t)₂Se to give hydrogen, 2 x 2-methylpropene, Se and H^.. The chain is initiated by Bu(t ·) abstracting H^. from Bu(t)₂Se or from H₂. Semi-empirical calculations carried out with a variety of levels of theory have confirmed that E C bond cleavage is the first step in the decomposition of all the group 16 precursors studied and have been used to model the product distribution with time. They confirm the conclusions drawn on the basis of our experimental observations, except that the H^. abstraction reactions from intact Bu(t)₂Se probably do not occur by a concerted pathway, rather they are stepwise. The radical chain process also occurs in the presence of Me₂Zn, but the radical chain initiator in this case is Me^. from homolytic fission of the Zn-C bonds. Evidence is presented that these reactions occur on the ZnSe surface. The small amount of 2-methylpropane produced from reaction of Me₂Zn with Bu(t)₂Se in H₂ comes from hydrogenation of 2-methylpropene catalysed by ZnSe. Photochemically, homolytic cleavage of the Se-C bonds in Bu(t)₂Se again occurs, but Bu(t)Se^. has sufficient lifetime to recombine with (t)Bu^. and the major hydrocarbon products are formed from radical-radical reactions.

Full text of DOI:10.1039/a904270i

View Article Online / Journal Homepage / Table of Contents for this issue
J O U R N A L O F
Studies of the thermal and photochemical decomposition
t
mechanisms of Bu Se in the presence or absence of Me Zn
2
2
using deuterium labelled analogues
a
Nigel L. Pickett, Douglas F. Foster, Nick Maung and David J. Cole-Hamilton*
a
b
a
C H E M I S T R Y
a
School of Chemistry, University of St. Andrews, St. Andrews, Fife, Scotland, UK KY16 9ST.
E-mail: djc@st-aud.ac.uk
Advanced Materials Research Laboratories, North East Wales Institute, Plas Coch,
Mold Rd., Wrexham, Wales, UK LL11 2AW
b
Received 27th May 1999, Accepted 6th September 1999
t
Thermal decomposition of Bu Se in helium produces 2-methylpropene and 2-methylpropane (8 : 1) and no
volatile selenium containing products, apart from the element. In hydrogen, the products are the same (8 : 1
ratio of 2-methylpropene : 2-methylpropane). However, the 2-methylpropane produced by decomposing
2
1
8
d -Bu
t
9
10
0
2
Se in H
2
Se in helium, d , d , d and d -2-methylpropane are all formed, but d -Bu Se is not a product.
2
contains signi®cant amounts of hydrogen (d : d ~2.6 : 1). In the codecomposition of d - and
1
8
d -Bu
t
0
1
9
10
9
t
2
0
Photolysis of d - and d -Bu Se in helium at room temperature produces 2-methylpropane and 2-methylpropene
18
t
2
0
1
9
10
9
t
(
2
1 : 1) with the 2-methylpropane being d , d , d and d . d -Bu Se is also formed along with small amounts of
0
,2,3,3-tetramethylbutane and d - and d -Bu
18
t
18
Zn with d -Bu
t
2
produces d and d -methane and d -2-methylpropene together with traces of d -2-methylpropane.
2
Se
2
. Thermal codecomposition of Me
8
2
2
Se in H
2
0
1
8
t
2
These results are interpreted as indicating that thermally, Bu
Se decomposes via homolytic cleavage of the
t?
Se±C bonds and that the free Bu so formed initiates a radical chain reaction involving H as the chain carrier
?
t
?
through its reaction with intact Bu
t
2
Se to give hydrogen, 262-methylpropene, Se and H . The chain is initiated
2
Se or from H . Semi-empirical calculations carried out with a variety of levels
t?
by Bu abstracting H from Bu
?
2
of theory have con®rmed that E±C bond cleavage is the ®rst step in the decomposition of all the group 16
precursors studied and have been used to model the product distribution with time. They con®rm the
?
conclusions drawn on the basis of our experimantal observations, except that the H abstraction reactions from
t
intact Bu
process also occurs in the presence of Me Zn, but the radical chain initiator in this case is Me from homolytic
2
Se probably do not occur by a concerted pathway, rather they are stepwise. The radical chain
?
2
®ssion of the Zn±C bonds. Evidence is presented that these reactions occur on the ZnSe surface. The small
amount of 2-methylpropane produced from reaction of Me
t
2 2 2
Zn with Bu Se in H comes from hydrogenation of
t
-methylpropene catalysed by ZnSe. Photochemically, homolytic cleavage of the Se±C bonds in Bu Se again
2
2
t
?
t
?
occurs, but Bu Se has suf®cient lifetime to recombine with Bu and the major hydrocarbon products are
formed from radical±radical reactions.
t
mechanism of decomposition of Bu
with Me M, one cannot be certain that neither Bu EH nor H E
2
E, and of their reactions
t
Introduction
2
2
Since we ®rst introduced 2,2-dimethylpropanethiol(tert-butyl
t
thiol, Bu SH) as a sulfur precursor for growing ZnS by metal
is formed in situ during the decomposition. We have recently
1
reported that Bu EH react with Me M in the gas phase to give
6
t
2
1
,2
t
clusters, [MeMEBu ] , which further decompose via involatile
organic vapour phase epitaxy (MOVPE),
di-tert-butyl
E, and Bu EH (E~S or Se) have become
widely used for MOVPE of wide band gap II±VI semiconduc-
5
t
2
t
chalcogenides, Bu
t
M(EBu )
[ to give nanoparticles of ME. Thus the active
hydrogen atom is removed and may not be important as a
2
]
n
3
±10
tors, ME (M~Zn or Cd).
simplest system, that involving H
problems leading to the formation of nanoparticulate depos-
This is partly because the
passivator for dopant N.
There have been some studies which show that Bu S
2
E, is plagued by prereaction
t
2
t
decomposes cleanly to Bu SH and 2-methylpropene and thence
1
1±14
t
its,
partly because the Bu containing compounds are the
17±19
to H S and 2-methylpropene.
These are the classic
products of b-H abstraction (Fig. 1) but reactions in the
2
least stable of the simple metal alkyls and have suitable
volatilities and partly because attempts to dope the grown
layers p-type using ammonia or amines as the sources of
dopant N lead in many cases to nitrogen incorporation but
presence of cyclohexene have been interpreted as involving a
t?
®rst step that is S±C bond cleavage to give free radicals (Bu
and Bu S ).
t
?
18
1
5
much of it is not electrically active. There is considerable
evidence that the electrical inactivity arises because of
passivation and in some cases this has been traced back to
the presence of N±H which, being isoelectronic with S or Se will
t
For Bu SH, detailed studies of the kinetic order of the
reaction and the effect of cyclohexene have been interpreted in
6
,9,15
not be electrically active.
The source of the passivating
hydrogen could be the amine, especially if it contains N±H
initially, the ambient (growth is usually carried out in
hydrogen) or the group 16 precursor. In order to remove one
of the potential sources of H, researchers have tended to favour
t
t
t
Bu
2
E over Bu EH, although, without detailed studies of the
Fig. 1 b-abstraction pathway for the decomposition of Bu
J. Mater. Chem., 1999, 9, 3005±3014
2
S.
3005
This Journal is # The Royal Society of Chemistry 1999

View Article Online
diphenylketyl and degassed prior to use. Powdered lithium
(99% with 0.5% sodium) was purchased from Aldrich.
Amorphous Se powder of 99.9999% purity and particle size
9
t
v325 mesh was purchased from Johnson Matthey. d -Bu Cl
was purchased from CDN Isotopes Inc. and used as supplied.
Scheme 1 Free radical process proposed for the decomposition of
Bu SH.
3
0
t
31
t
20
Me
previously. NMR spectra were recorded on
Associates AM300 spectrometer operating in the Fourier
Zn and Bu
Se were prepared and puri®ed as described
2
2
a
Br uÈ ker
terms of initial S±C bond cleavage followed by a radical chain
t
process involving abstraction of H from the Bu group of intact
1
3
transform mode with (for C) noise proton decoupling. The
1
3
1
C{ H} and H NMR spectra were run in deuteriated solvents
1
t
Bu SH by either Bu or HS (Scheme 1). Later studies have
t?
?
20
1
8,21±24
for the lock signal, with chemical shifts quoted in ppm to high
frequency of tetramethylsilane (TMS) as the internal reference.
generally con®rmed this analysis,
2
although b-H abstrac-
1,24
tion has also been proposed.
t
For Bu Se, a marked effect of the nature of the surface over
2
2
5
18
Preparation of d -Bu
t
which the reaction is carried out has been noted, perhaps
suggesting that surface catalysed reactions may be important,
t
but, despite not having observed Bu SeH as a product, the b-H
Se
2
9
A solution of d -Bu Li (1.7 mol dm ) in pentane was prepared
t
23
t
from the Bu OMe catalysed reaction between d -Bu Cl and Li
9
t
abstraction mechanism was proposed. Against this, a possible
?
32,33
18
t
according to the literature procedure.
prepared as previously described for Bu Se
2
d -Bu Se was
reaction with H
2
or H generated from the carrier gas was
t
31
from the
2
noted and reactions in D
ways of exploring this type of reaction. They were not,
2
or He ambients were suggested as
9
dropwise addition over ca. 3 h of the pentane solution of d -
t
Bu Li (100 cm , 1.7 mol dm , 0.17 mol) to a stirred solution of
3
23
2
however, carried out. In IR studies, we tentatively identi®ed
5
23
3
2
`SeCl ' (0.28 mol dm , 0.084 mol) in THF (300 cm ). This
yielded 5.7 g (32%) of d -Bu Se after puri®cation. NMR
t
Bu SeH as a decomposition product of Bu
t
2
Se and this together
18
t
2
with preliminary theoretical calculations led us to propose that
2
(
JD±C~19.3 Hz, CD ), 41.6 C(CD ) .
2 2 2 2 D 3 C
20% in CH Cl ±CD Cl ); d 1.44 (s,CD ); d 33.2 (heptet,
6
b-H abstraction was important.
In view of the somewhat con¯icting data especially on the
3
3 3
t
2
decomposition pathway for Bu Se, we have carried out
detailed studies of the thermal and photochemical decomposi-
Decomposition studies
t
tion of Bu Se in a variety of different ambients (He, H or D )
2
2
2
2 2
The carrier gas (H , He or D ) was passed through ¯ow meters
1
8
t
as well as of d -Bu Se,{ again in various different ambients,
into glass bubblers containing the required precursors, which
were connected to a quartz pyrolysis tube (30 cm62 cm
diameter) held within a tube furnace so that mixing of the
gas streams emanating from the bubblers occurred within the
pyrolysis tube. The ef¯uent from the pyrolysis tube was passed
through stainless steel tubing to a gas sampling valve (0.5 or
2
0
sometimes in the presence of d -Bu Se and/or Me Zn. We have
t
2
2
previously shown that labelling studies of this kind can give
deep insight into the precise mechanisms involved in the
i
2
27
decomposition process of e.g. Pr Te. In that study, we
i
showed that Pr Te decomposes by Te±C bond cleavage to give
2
i?
i?
i
2
23
2
6Pr and Te. Pr then reacts with intact Pr Te to give
1 cm ) attached to the inlet (splitless injection mode) of a
?
propane, 26propene, Te and H , as shown in Fig. 2. Labelling
studies have been used by others to help in the study of the
mechanism of decomposition of primary and secondary
butanethiols. We also report the results of semi-empirical
calculations carried out for reactions involved in the decom-
capillary gas chromatograph (Hewlett Packard 5890) with
mass spectrometric (Hewlett Packard 5792) detection (GCMS).
Both the ef¯uent from the GCMS and that from the bypass of
the gas sampling valve were passed through activated carbon
scrubbers (Embacel) and then through propan-2-ol (to remove
2
4
t
2 4
any traces of Me Zn) and aqueous CuSO solution (to remove
position of Bu
these results have appeared.
2
Se in H
2
or He. Preliminary reports of some of
2
8,29
traces of Se compounds). The entire apparatus was placed in a
fume-cupboard. Flows were allowed to stabilise at a given
temperature for 30 mins before sampling and most experiments
were conducted from low to high and from high to low
temperature. The temperature programme of the GC oven was
2
1
as follows: 230 ³C for 2 min, then warming at 20 ³C min to
2
1
3
0 ³C, then held for 2 min followed by heating at 20 ³C min
to 150 ³C and holding for 6 min. An HP-PONA (cross-linked
methylsilicone gum) column (50 m, 0.2 mm internal diameter,
®lm thickness 0.5 mm, phase ratio 100) was employed with
helium as the carrier gas. Calibrations for 2-methylpropane
and 2-methylpropene were carried out using standard gas
mixtures{ in connection with several gas sampling loop
i
27
2
Fig. 2 Radical decomposition mechanism for Pr Te.
3
volumes (0.5±5 cm ) and for group 16 compounds using
Experimental
{CAUTION: We have used three different samples of calibrant gas
General procedures
because, alarmingly, the ®rst two gave different results. A sample
nominally containing 1000 ppm (Scott Speciality Gases) each of 2-
methylpropane and 2-methylpropene suggested that the MS detector is
1.6 times more sensitive to 2-methylpropene than to 2-methylpropane,
whereas a sample nominally containing 100 ppm of each (Scott
Speciality Gases) gave a sensitivity ratio of 1.05 : 1. In both cases, the
response of the MS to both compounds was linear over the whole
range. A third sample nominally containing 1000 ppm of each
compound (Air Products) gave a ratio of 1 : 1.05 so we have assumed
All preparations were carried out under dry oxygen free
nitrogen or argon (BOC) puri®ed by passing through two
2
z
consecutive columns (2.5 cm680 cm) packed with Cr
on
silica. Greaseless joints and taps were employed and manip-
ulations were carried out using standard Schlenk line and
catheter tubing techniques. The solvents, THF and pentane
(
Fisher Scienti®c UK), were dried by distillation from sodium
29
that the correct ratio is 1 : 1. In a previous paper, which was submitted
before we had carried out the third calibration, we assumed that the
ratio of 1.6 : 1 was correct. The amounts quoted in that paper are
n
Throughout this paper, the nomenclature d - is used to indicate that
Se is
29
{
there are n deuterium atoms in the given compound. Thus, d -Bu
18
t
2
therefore wrong, but the interpretation of them is not greatly affected at
least qualitatively.
4 9 2
(C D ) Se.
3006
J. Mater. Chem., 1999, 9, 3005±3014

View Article Online
3
4
published or measured vapour pressure data. In cases where
the total amount of hydrocarbon produced did not correlate
exactly with the initial amount of precursor, it was assumed
that the vapour pressure data for the precursor were in error (it
is very sensitive to small changes in temperature) and the
amount of precursor was adjusted to ®t the amount of
hydrocarbons produced.
Theoretical methods
Hartree±Fock and post Hartree±Fock ab initio molecular
3
orbital techniques and density functional theory (DFT)
5
3
6,37
methods
have been used to study the proposed reaction
mechanism from ®rst principles. The C±Se bond dissociation
energies for dialkyl selenides and alkylselenium radicals have
been computed using the ab initio (ROHF and MP2) and
density functional theory (DFT) methods (BHandH, BHandH-
LYP, B3LYP, B3P86, B3PW91, BLYP and BP86) with the 6-
Labelled compounds were measured quantitively using total
ion currents (2-methylpropene) or single ion scans in the mass
2
spectrum, as described previously. Masses used were: 2-
7
3
8
0
methylpropene, m/z: d 56 (M ), d 64 (M ); 2-methylpro-
z
8
z
311zG(2d,p) basis set. All calculations were performed using
3
9
the GAUSSIAN 94 implementation of the density functional
0
z
1
z
2
3 3
pane, m/z: d 43 ([M2CH ] ), d 44 ([M2CH ] ), d 45
([M2CH ] ), d 49 ([M2CD ] ), d 50 ([M2CD ] ). These
3 3 3
masses were chosen to avoid interferences. However, for d - 2-
4
0
z
9
z
10
z
theory (DFT) method while the SPARTAN 4.1.2 electronic
structure program was used to generate starting structures and
to animate vibrational frequencies. Geometry optimisation,
critical point characterisation (all calculated structures were
found to correspond to true energy minima), normal mode
analysis (vibrational frequencies) and thermodynamic calcula-
0
methylpropane, there is an interference at m/z 43 from
z
1
(
times the area of the peak at m/z 44 so the area of the peak at
3
[M2CH 2H] ) from d -2-methylpropane. This peak is 0.38
1
m/z 43 arising from d -2-methylpropane was calculated and
subtracted from the total area of the peak to give the area
4
1±43
tions
convergence criteria.
were carried out using the default GAUSSIAN
4,45
4
0
arising solely from d -2-methylpropane. If d -2-methylpropane
2
All of the geometries for the
molecules studied were fully optimised without using symmetry
or structural constraints. A detailed analysis of the computed
energies and comparison with experimental results along with
an assessment of the suitability of the DFT methods, for the
computational study of these systems, has been published by
is present, the peak at m/z 44 has interferences from it
z
z
[M2CH D] and ([M2CH 2H] ). We estimate that the
2 3
(
ratio of the peak at m/z 45 : 44 in pure d -methylpropane should
2
2
1
be 1.23 : 1. We can then use this to determine the d : d ratio.
Values obtained using single ion scans were checked for
3
8
0
1
2
8
9
10
one of the authors recently. Calculations were also performed
consistency of the (d zd zd ) : (d zd zd ) ratio using total
ion currents.
Baseline separations were obtained for all species apart from
3
using the SPARTAN 4.1.2 electronic structure program and
9
4
MOPAC93 programs. The PM3 (Parametric Method 3)
6
0
1
2
8
9
10
47
method a variant of the MNDO (modi®ed neglect of
48,49
differential overlap) method developed by Dewar et al.
d -, d -, d -2-methylpropane and d -, d -, d -2-methylpro-
n
pane. There was no evidence for d -2-methylpropene (n|0 or
8
scrambling of H and D in the methyl groups of 2-
) in any of the samples below 700 ³C. At 700 ³C extensive
was used. The standard semi-empirical program modules were
used both for geometry optimisation, critical point character-
isation (all calculated structures were found to correspond to
true energy minima), normal mode analysis (vibrational
frequencies) and thermodynamic calculations. Transition
states were located using the linear synchronous transit
methylpropene occurs, but this is also observed when 2-
methylpropene is passed through a clean tube at 700 ³C with D
as the carrier gas. Labelled methane was quanti®ed as
2
2
7
t
2
previously described. Searches for H Se and Bu SeH were
5
0
carried out using m/z 80 and 138, respectively.
(LST) technique. In all cases a transition state was identi®ed
a
Table 1 Gas-phase mechanism and reaction rate parameters
Reaction
A
m
E
a
Ref.
3
t.
15.3
t
t
.
Bu 2Se DA Bu Se zBu
10
10
10
10
10
0
0
225.9
229.9
0
R
4
.
t
15.34
16.53
16.1
t
.
Bu Se DA Bu zSe
R
.
t.
5
t
b
t
2
Bu DA C4H8 zBu H
21.5
21.5
62
62
63
6
t
2
Bu DA Bu 2
Bu DA C4H8zH
0
150.6
.
7
.
t
9.34
1.48
.
.
8a
18.74
2
2
H zH2 DA 2H2
H zHe DA H2zHe
10
10
10
21.3
21.0
0
0
64
64
R
8b
17.8
.
9
.
t
t
22.37
Bu zD2 DA Bu DzD
4.43
45
.
.
10a
.
6.4
H zH2 DA H2zH
10
10
10
10
10
10
10
2.29
2.1
37.6
29.3
0
65
66
10b
.
7.01
10.8
10.8
25.3
5.8
H zD2 DA HDzD
.
.
.
11a
t
t
Bu zH DA Bu H
20.15
20.15
5.2
62
.
11b
t
t
Bu zD DA Bu D
0
R,62
R,60
R,60
62
.
12
.
t
t
t
t
Bu zBu 2Se DA Bu HzBu SeCꢀCH3† CH
37.9
10.8
37.5
N/C
N/C
N/C
154.8
2
2
.
13
H zBu 2Se DA H2zBu SeCꢀCH3† CH
.
t
t
2.4
.
14
Bu zH2 DA Bu HzH
.
2
2
t
t
21.7
4.24
N/C
N/C
N/C
0
.
15
Bu zBu 2Se DA Bu HzC4H8zBu Se
.
t
t
t
t
N/C
N/C
N/C
R
.
16
H zBu 2Se DA H2zC4H8zBu Se
.
t
t
R
1
8
t
t
Bu 2SezSe DA Bu 2Se2
R
60
2
4'
13.0
10
t
C4H8zH2 DA Bu H
.
25
.
13.7
10
t
t
Bu SeCꢀCH † CH DA C4H8zBu Se
0
133.9
R,60
3
26
2
2
.
t.
12.79
12.7
5.8
C4H8zH DA Bu
10
10
10
0.25
0
6.1
0
63
62
62
.
.
27
t
Bu zH DA C4H8zH2
Bu HzH DA Bu zH2
.
28
.
t
t
2.4
10.8
a
62±66
Rate data for hydrocarbon and deuterated species were obtained from the literature.
The kinetic rate constant is of the form
respectively, where n is the order of reaction. Concentrations
m
k~AT exp(2E
60
/RT) . The units of E
21
and A are (kJ mol ) and (mol cm
23 12n 21
)
a
a
s
are in mol cm and time in s. All rate data for Se-centred species were obtained from DFT and canonical transition state theory (CTST) as
2
3
explained in the text. R indicates data estimated in this work. R,60 indicates data combined from R and ref. 60. N/C indicates not calculated.
b
4 8
C H ~2-methylpropene.
J. Mater. Chem., 1999, 9, 3005±3014
3007

View Article Online
t
a
2
Table 2 Gas-phase mechanism and reaction rate parameters for reactions not thought to be of signi®cance in the thermal decomposition of Bu Se
Reaction
A
n
E
a
Ref.
2
9
13.5
13.3
13.3
17.1
t
t
b
Bu 2Se ? Bu SeHzC4H8
10
10
10
10
10
10
10
10
10
10
0
0
0
0
0
0
0
234.3
242.7
225.9
292.9
41.8
85.8
0
R
3
0
t
Bu SeH DA C4H8zH2Se
Bu SeH DA Bu HzSe
Bu 2 DA 2Bu
R
3
1
t
t
R
32
.
t
t
60
60
60
60
60
60
R
.
33
.
t
t
11.3
12.40
13.7
13.30
15.0
11.3
Bu zC4H8 DA Bu HzCH2CꢀMe†CH2
.
34
CH2CꢀMe†CH2 zBu 2Se DA C4H8zBu SeCꢀCH † CH2
.
t
t
3
2
.
.
35
CH2CꢀMe†CH2 zH DA C4H8
.
.
36
CH2CꢀMe†CH2 zCH2CꢀMe†CH2 DA C8H14
21.5
20.75
0
0
.
.
37
Bu z CH2CꢀMe†CH2 DA Bu CH2CꢀMe†CH2
t
t
0
3
8
t
t
Bu 2Se DA Bu 2zSe
581.6
a
60 b
Calculations and units are as in Table 1. Rate data for allyl hydrocarbon species were obtained from the literature.
4 8
C H ~2-methylpropene.
when the Hessian yielded only one signi®cant imaginary
21
system of differential equations constructed by the rigorous
application of mass-action principles. This system contains an
equation for each chemical species in the model, and each
equation expresses the time derivative of this species in terms of
the usual mass-action products. With numerical values
assigned to all the rate constants, and a starting pressure/
concentration selected for the precursor, this becomes an initial
value problem in the solution of non-linear ordinary differ-
ential equations (ODEs). Numerical solution of these so called
`stiff systems' was accomplished using the built in solver for
stiff ODE systems using the semi-implicit extrapolation
2
1
frequency (400i cm ¡n ¡2000i cm ). Veri®cation that the
TS
coordinate corresponding to the imaginary frequency smoothly
connects reactants and products was undertaken by `animat-
ing' the coordinate whilst in ambiguous cases the intrinsic
reaction co-ordinate (IRC) method implemented in MOPAC
4
6
9
criteria were used.
3
was used. In all computations the default convergence
Rate constants for all reactions involving Se-centred species
were estimated using canonical transition state theory (CTST).
The CTST calculations are based solely on the statistical
thermodynamic properties of the reactants and the transition
state. Enthalpies of activation, moments of inertia, and
vibrational frequencies obtained from quantum chemical
computations were used to compute the partition functions
of the reactants and the transition state. No corrections for
quantum mechanical tunnelling were made from CTST since
such corrections would be expected to be small for MOVPE
reactions since relatively high temperatures and large molecules
are involved. The computed values of the rate parameters are
collected in Tables 1 and 2.
Rather than deduce the speci®c contributions of individual
reactions directly from experimental data, a procedure not yet
possible given the current state of knowledge, we chose to
de®ne all the reactions that are expected to occur, assigned
numerical values to the appropriate rate parameters and
numerically solved the resultant differential equations, com-
paring the numerical solutions with experimental data as a test
of the validity of the chosen reaction set.
5
1,52
modi®cation of the Burlisch±Stoer method in Mathcad 6.0.
Since the diffusion velocities of the gas species can be
estimated to be an order of magnitude less than the average
¯ow velocity, the conditions in the ¯ow tube were simulated by
specifying initial non-zero partial pressures for only the Bu Se
t
2
and the He/H
radicals to the walls of the furnace tube were estimated to be
2 2
(D ) carrier gas. The diffusion times of H and D
2
.2 s based on an estimated diffusion coef®cient of 2 cm s
21
0
and a furnace tube radius of 3 cm. Errors in the isothermal
assumption on which the simulations were performed are
signi®cant at temperatures above 500 ³C, because at higher
temperatures the reactions occur in the steep thermal gradients
at the inlet to the furnace.
Results
t
Thermal decomposition of Bu Se
2
t
Heating Bu
methylpropane in an 8 : 1 ratio [Fig. 3(a)]. Very careful analysis
2
Se in helium produces 2-methylpropene and 2-
The purpose of the present study was to perform a thorough
t
numerical simulation of Bu
2
Se pyrolysis using comprehensive
of the GCMS traces using total ion currents or suitable single
t
ion scans shows that neither Bu SeH nor H
reaction schemes with rate parameters evaluated from the
computational quantum chemistry calculations. It was not
initially our intention to establish the de®nitive `correct'
mechanism for the pyrolysis processes. This will obviously be
limited by the present status of knowledge of organoselenium
free-radical chemistry and the appropriateness of the chosen
theoretical model (e.g. accuracy and precision of rate para-
meters, types of reactions which occur, validity of hydrocarbon
experimental data). The present study was used in determining
speci®c reactions which may warrant closer study and in
planning speci®c pyrolysis experiments to test certain features
of the proposed mechanism. It should be emphasised that these
computations do not involve parameter ®tting and are
therefore independent of the experiment that is being modelled.
In establishing the matrix of chemical reactions upon which the
simulations were based, the approach adopted was to include
all conceivable species and processes. Comparison of the
simulated data with the experimental results readily indicated
which of the reactions were super¯uous.
2
Se is detectable as
a product at any temperature, although we have shown in
t
separate experiments that both Bu SeH (50% decomposition at
3
suf®ciently stable under the reaction conditions that they
50 ³C) and H Se (stable over the whole temperature range) are
2
53
would be detectable if they were formed.
Carrying out the same reaction in a hydrogen ambient
produces almost identical results with a slight decrease in
decomposition temperature and a ratio of 2-methylpropene : 2-
methylpropane of 8 : 1 [Fig. 3(b)]. The 2-methylpropene : 2-
methylpropane ratios at different extents of decomposition are
also almost identical in hydrogen or helium (Fig. 4). If the
decomposition is carried out in D
formed contains some D; (CH ) CH : (CH ) CD~10.
2
, the 2-methylpropane
3
3
3 3
1
Decomposition of d -Bu
8
t
2
Se
1
8
t
8
In helium, the decomposition of d -Bu Se produces the d -2-
2
1
0
methylpropene and d -2-methylpropane. In hydrogen, both 2-
methylpropene and 2-methylpropane are formed, but in this
case, there is substantial incorporation of H into the 2-
The decomposition experiment was approximated as an
isothermal constant-volume pyrolysis and is described by a
3008
J. Mater. Chem., 1999, 9, 3005±3014

View Article Online
Fig. 4 Ratio of 2-methylpropene : 2-methylpropane at different tem-
Se in helium (Y) or
t
peratures during the decomposition of Bu
hydrogen (à).
2
1
8
t
using d -Bu
CH D : CH ~3.4 : 1) and d -2-methylpropene. Small amounts
of CH are present even at room temperature, presumably from
reaction of Me Zn with moisture.
Carrying out the codecomposition of Me
in hydrogen does generate some 2-methylpropane (mainly d )
2 3
Se in this reaction gives mainly CH D,
8
(
3
4
4
2
1
8
t
2
2
Zn and d -Bu Se
8
in addition to the other products (the methane is mainly CH
2
4
(
reaction of Me
CH
4 3
: CH D~3.3 : 1); d -2-methylpropane is formed from the
0
Zn with d -Bu
t
2
2
Se in D
2
. The amount of 2-
methylpropane formed increases at higher temperatures and
t
after complete loss of Bu Se.
2
t
2
Photochemical decomposition of Bu Se
t
UV photolysis of Bu
2
Se in helium at room temperature causes
ca. 29% decomposition under our conditions and produces
equal amounts of 2-methylpropene and 2-methylpropane,
together with small amounts of 2,2,3,3-tetramethylbutane
t
0
18
t
and some Bu
ratio) produces d - and d -2-methylpropene together with d -,
2 2 2
Se . Co-photolysing d - and d -Bu Se (1 : 3
0
8
0
1
9
10
9
t
2
d -, d - and d - 2-methylpropane (1 : 1.6 : 3.5 : 4.8). d -Bu Se is
0
9
18
0
produced (®nal d : d : d ratio is 6.6 : 1 : 14.7) together with d -
t
Fig. 3 Reaction pro®les for the decomposition of Bu
2
Se. (a) In helium
18
and d -Bu
t
2
9
. There is no sign of d -Bu
t
t
Se
2
2
Se
2
.
(
(
(
&) Bu
2
Se, ( ) 2-methylpropene, (+) 2-methylpropane or hydrogen
t
%) Bu Se, (#) 2-methylpropene, (() 2-methylpropane. (b) In helium
2
t
?
&) Bu Se, (#) 2-methylpropene, (:) 2-methylpropane. (c) In
2
t
hydrogen (&) Bu
2
Se, (#) 2-methylpropene, (:) 2-methylpropane.
Discussion
t
Decomposition of Bu Se
Lines in (b) and (c) are values calculated from the theoretical
simulations based on the reactions of Scheme 4 and the parameters
of Table 1.
2
Thermal. The possible initial steps in the decomposition of
t
2
Bu Se, reductive elimination, b-hydrogen abstraction and
methylpropane [(CD
position, falling to 1.4 : 1 at full decomposition].§
3 3 3 3
) CH : (CD ) CD~2.6 : 1 at 50% decom-
homolytic cleavage are shown in Fig. 5. Since 2,2,3,3-
tetramethylbutane is not observed as a product, reductive
elimination can be ruled out as the ®rst step of the reaction. The
0
18
Codecomposition of d and d -Bu
t
2
Se
t
fact that neither Bu SeH nor H
under conditions where they would both be expected to be
2
Se is observed as products
0
18
The codecomposition of d - and d - Bu Se (1.2 : 1) in helium
t
2
0
produces only 2-methylpropene (d : d ~1.4 : 1) and 2-methyl-
8
stable rules out b-H abstraction as a possible decomposition
t
0
1
propane (d : d : d : d ~2.29 : 1 : 2.86 : 1.2). There is no evi-
9
10
pathway for Bu
decomposition must be Se±C bond cleavage to give Bu and
2
Se and suggests that the initial step in the
t.
9
dence for formation of any d -Bu
t
2
Se at any of the
t
Bu Se [eqn. (3), Scheme 2].
?
temperatures we have studied.
In order to distinguish unequivocally between b-hydrogen
abstraction and free radical pathways, we have carried out the
t
Codecomposition of Bu Se and Me Zn
2
2
0
co-decomposition of d and d -Bu Se (1.2 : 1). In principle,
18
t
2
t
2
t
since b-hydrogen abstraction to give Bu SeH followed by b-
hydrogen abstraction/reductive elimination is all intramolecu-
The codecomposition of Bu Se and Me Zn in helium produces
2
methane and 2-methylpropene as the major products, whilst
0
lar, the 2-methylpropane produced should be all d or d
10
0
1
whilst a free radical process will lead to a mixture of d , d , d
9
§
concentration does not alter with the extent of decomposition.
H
2
Se is observed as an impurity in the initial precursor, but its
1
and d since the free Bu can abstract H from species derived
0
t?
J. Mater. Chem., 1999, 9, 3005±3014 3009

View Article Online
9
The fact that d -Bu Se is produced on photolysis of a
t
2
0
mixture of d - and d -Bu Se at room temperature but not on
18
t
2
thermolysis of the same mixture indicates that recombination
t
of Bu Se with Bu does not occur to a signi®cant extent in the
?
t?
t
high temperature reactions. This must either be because Bu Se
is unstable, or because it reacts faster with other species than
?
t?
with Bu . The only other reactions available to it in helium
t
would be with Bu Se, presumably this would be by H
2
t
abstraction, but since the expected product (Bu SeH) is not
t
t
observed, we conclude that Bu Se decomposes rapidly once it
?
Fig. 5 Possible ®rst steps in the decomposition of Bu
elimination; (ii) b-H abstraction; (iii) homolytic ®ssion of the Se±C
bond to give free radicals.
2
Se. (i) reductive
forms. This decomposition must be either via Se±C bond
?
cleavage [eqn. (4), Scheme 2], or by loss of H to give 2-
methylpropene and Se, although calculations (see later) show
that this last reaction is thermodynamically inaccessible
?
because two high energy species, H and Se(g) would be
formed. The only other alternative would be by b-H
?
abstraction to give HSe , which would be expected to give a
t
route to H Se (not observed). Bu Se then decomposes by
2
2
t?
cleavage of the two Se±C bonds to give 26Bu and Se.
However, this does not explain the products, 2-methylpropane
and 2-methylpropene, nor their ratio.
It would be possible to explain the observed products and
t?
their ratio on the basis of simple reactions of Bu with one
t?
another. At low temperatures, Bu react together by dis-
proportionation and combination to give 2-methylpropane and
2
-methylpropene (1 : 1) together with 2,2,3,3-tetramethylbu-
5
4
tane (10%, invariant with temperature). However, above
80 ³C, in the temperature region of our studies, the
3
?
t?
t
unimolecular loss of H from Bu becomes competitive and
dominates at higher temperatures. The disproportionation and
combination reactions have zero activation energy, whereas the
Scheme 2 Reactions involved in the decomposition of Bu
only reactions of Bu are simple gas phase reactions with itself or with
2
Se if the
t?
the carrier gas (D
2
).
?
loss of H has a marked temperature dependence. This means
?
0 18
either from d - or d -Bu
t
2
0
Se. In practice, we observe d , d , d
1
9
that, at above 400 ³C, the loss of H completely dominates and
1
0
and d -2-methylpropane in the ratio 2.29 : 1 : 2.86 : 1.2, sug-
gesting that a free radical process is operating with kinetic
isotope effects favouring H abstraction. However, this is not
unequivocal proof because we have shown that the H directly
t?
-methylpropene is the only product arising solely from Bu ,
2
particularly as recombination of Bu with H is non-
t?
?
?
5
competitive with other reactions of H . Simple reactions of
4
t
?
Bu with one another, as shown in Scheme 2 can then be ruled
out as being the sole mechanism by which reactions subsequent
to Se±C bond cleavage occur.
5
3
attached to Se is readily exchanged so that C D SeD and
4
9
C
H
4 9
SeH will exchange to give a mixture of the starting species
together with the exchange species, C SeD and C SeH, so
that 2-methylpropane formed from them would also be d , d ,
i
55
H
4 9
4
D
9
For Pr
shown in Fig. 6) to that described in Scheme 2, except that it
2
Te, Hicks has proposed a similar mechanism
(
involves surface bound free radicals which undergo b-H
0
1
9
10
d and d . However, if we assume that b-hydrogen abstrac-
tion followed by reductive elimination [step (ii) followed by
step (i) in Fig. 5] is responsible for the formation of 2-methyl-
propane, we can assume that b±H abstraction gives d - and d -
t
Bu SeH in the ratio 0.8 : 1 (the same ratio as C H X : C D X,
?
abstraction to give propene and surface bound H . Propane
i?
?
is then formed by combination of Pr
t
with H
For Bu Se, a similar process would give 2-methylpropene, 2-
(Fig. 6).
(
ads)
(ads)
0
10
2
?
?
4
9
4
9
methylpropane and H
?
(ads)
(Scheme 2). The relative rates of H
addition to Bu and combination of
(ads)
t?
X~H or D in the ®nal product). If the SeH/D equilibrium
is then fully established and is followed by reductive elim-
abstraction, H
?
(
ads)
two H
the ratio of 2-methylpropene : 2-methylpropane and a value of
(
ads)
[eqns. (7), (11) and (8), Scheme 2] then determine
0
1
9
10
ination, the ratio of d : d : d : d -2-methylpropane would be
: 1.25 : 1.25 : 1.56, far from the observed 2.29 : 1 : 2.86 : 1.2.
However a free radical process in which there is a kinetic
isotope effect (k /k ) of ca. 2.3 fully accounts for the observed
ratios. The increased amounts of products derived from d -Bu
?
1
8 : 1 is possible with the residual H
surface to give H . It is possible, invoking this mechanism, to
ads)
combining on the
(
2
H
D
explain the observation that in H the ratio of 2-methylpro-
2
9
t?
pane : 2-methylpropene is unchanged relative to the value in
helium but also that the labelling studies clearly show that H/D
from the ambient H /D is incorporated to a very signi®cant
0
t?
0
relative to those from d -Bu arise because more of the d -
t
?
18
Bu
Bu
2
2
Se is consumed by H abstraction (see later) so more d -
Se is available towards the end of the reaction for the
2
2
t
extent (up to 70%). If incorporation occurred by direct reaction
t?
9
t?
t?
8
with D or d -Bu
generation of d -Bu .
of Bu
(ads) 2
with H [eqn. (9), Scheme 2],
(
ads)
2
i
55
Fig. 6 Possible surface catalysed mechanism for the decomposition of Pr
J. Mater. Chem., 1999, 9, 3005±3014
2
Te.
3010

View Article Online
the relative amount of 2-methylpropane would increase
dramatically. However, the reactions shown in Scheme 2,
eqns. (7)±(11) occurring on the surface would adequately
account for the ratio of 2-methylpropene : 2-methylpropane
.
Fig. 7 Concerted mechanism for the reaction of Bu with Bu
t
2
Se [eqn.
being the same in He as in H
H from the carrier gas into 2-methylpropane. These reactions
2
but also for the incorporation of
(12) of Scheme 3].
t?
?(ads)
require that Bu (ads) does not react with D
2
, but that H
unlikely that surface processes occur in the decomposition of
t
Bu Se.
?(ads)
does to give HD and D
recombination of Bu (ads) with H
. The 2-methylpropane is formed by
?(ads)
2
t?
t?
The only alternative fate for Bu is that it will react with
. In this scheme, the
is not altered by changing from He to
?(ads)
t
intact Bu Se, which, at least in helium, will be the major
reactive species present. It cannot react by addition, since this
concentration of H
2
H2 as ambient, only its labelling pattern so that, ignoring
isotope effects, the ratio of 2-methylpropane : 2-methylpropene
should be the same, but there is a pathway for incorporating H/
D from the ambient H /D . This mechanism requires that the
9
would provide a pathway to d -Bu Se, which is not observed
t
2
0
on codecomposition of d - and d -Bu Se; so it must react by
18
t
2
2
2
abstracting a b-H atom. There are two possible pathways that
?(ads)
t?
2
rate of reaction of H
rate of reaction of Bu (ads) with D
constants at various different temperatures (Table 3) and ®nd
with D
is very much faster than the
2
. We have calculated the rate
can occur subsequent to H abstraction. One is shown in
eqn. (12), Scheme 3 and Fig. 7 and is analogous to the process
that we have already proposed on the basis of labelling studies
t?
by Bu is
i?
i
27
?
for the reaction of Pr with Pr Te. Abstraction of H from
2
that the rate constant for abstraction of D from D
2
?
t
t?
lower than that for abstraction by H by a factor of 1000 or
more over the entire temperature range investigated here.
We cannot unequivocally rule out the reactions of Scheme 2,
Bu Se by Bu leads to 2-methylpropane, 262-methylpropene,
2
t
?
?
Se and H . H then reacts with intact Bu Se in a similar
2
?
manner to give H , 262-methylpropene, Se and regenerate H
2
occurring on the surface, as being wholly responsible for the
t
[eqn. (13), Scheme 3]. This is then a chain propagation reaction
? ?
decomposition of Bu
that glass surfaces do not strongly absorb Bu
2
Se, but other workers have suggested
t
with termination occurring by reaction of H with H (needs a
26
Se and related
t?
t
2
third body to remove excess energy) or with Bu to give Bu H
[eqn. (8) or (17) Scheme 3]. An 8 : 1 ratio of 2-methylpro-
pene : 2-methylpropane is then easily achieved since the chain
propagating reaction only produces 2-methylpropene whilst
the 2-methylpropane is formed in the initiation and termina-
5
alkyls and that gas-phase reactions dominate when glass is
5
2
the only exposed surface. In our reactions, Se sublimes from
6
t
the tube at all temperatures where decomposition of Bu
occurring and the glass always appears clean at the start and
nish of the reaction, con®rming that the precursors are not
2
Se is
t
?
tion steps. We have drawn the decomposition of Bu SeC H ,
4 8
®
?
t
t?
?
exposed to surfaces other than glass and that surface catalysed
processes are probably not important. We therefore think it
formed after abstraction of H from Bu Se by Bu or H , as a
2
concerted reaction in the gas phase. Calculations (see below)
suggest that this concerted reaction is very unlikely since it
Table 3 Calculated rate constants, k, for the gas-phase reactions of H
?
2
?
produces two highly energetic species, H and Se(g). The
t
?
a
with Bu or of H with H
2
reaction could occur on a surface, to produce Se(s) and H(ads)
both of which are of much lower energy than the gas phase
,
3
k/cm mol
H with H
21 21
s
t?
Bu with H
.
species, or it could occur stepwise in the gas phase with H being
?
T/³C
2
(k
H
)
(k
2 B
)
k
/k
H B
t?
lost from Bu , a reaction which is known to occur in this
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
1
1
6
3
3
3
3
3
3
4
4
4
4
4
5
a
00
20
40
60
80
00
20
40
60
80
00
1.3610
1.8610
2.3610
3.0610
3.7610
4.7610
5.8610
7.1610
8.6610
1.0610
1.2610
3.46106
3.96103
54
temperature regime (see above). The decomposition mechan-
ism is then as shown in Scheme 4, omitting eqn. (14).
5.1610
7.66107
3.5610
3.06103
6
3
t?
In hydrogen, it is expected that Bu will react with H
2
to give
2-methylpropane and H [eqn. (14), Scheme 4]. Whether the
1.1610
1.66107
2.7610
2.46103
7
3
?
t?
chain reaction is initiated directly by Bu [eqn. (15), Scheme 4]
2.2610
3.06107
2.2610
1.96103
7
3
t?
or by Bu reacting with H to give H [eqns. (14) and (16),
?
2
4.0610
1.8610
7
3
Scheme 4], the ratio of 2-methylpropene : methylpropane
should be identical, thus accounting for the similarity of
observations in helium and hydrogen ambient. Eqn. (14),
Scheme 4, however provides a mechanism for the incorpora-
5.46107
1.66103
7.1610
9.3610
1.5610
1.3610
7
3
m
Calculated using the equation k~AT exp(±E
A~1610 cm mol21
A~0.0185 cm mol
A
/RT ); for
~29.3 kJ mol ; for Bu
s , m~4.24, E ~37.8 kJ mol .
A
H
?
?
,
,
7
3
21 21
21
t
s
, m~2.1, E
A
3
21
21
t
Scheme 3 Reactions involved in the decomposition of Bu
?
2
Se assum-
Se [eqns. (12) and
t?
t
t
2
ing a concerted reaction of Bu or H with intact Bu
13)].
2
Scheme 4 Proposed mechanism for the decomposition of Bu
hydrogen or helium.
Se in
(
J. Mater. Chem., 1999, 9, 3005±3014
3011

View Article Online
tion of H /D from the ambient into the product 2-
2
methylpropane.
Scheme 4 provides a mechanism for the decomposition of
constant with one set of Arrhenius parameters is unable to
describe the temperature dependence of reactions as complex as
(15) and (16). Intuitively one can see this by examining the
concerted reactions in more detail. Such reactions are not
microscopically reversible, since it is extremely unlikely that
two isobutene molecules and one isobutane/dihydrogen
molecule could simultaneously collide with a selenium atom
2
t
Bu
observations that we have made. It also explains the larger
2
Se that is consistent with all of the experimental
0
9
1
10
amounts of d - and d -2-methylpropane (than of d - and d )
obtained on codecomposition of d - and d -Bu Se in helium
since there is likely to be a signi®cant isotope effect operating
0
18
t
t
67
and hydrogen atom to regenerate a molecule of Bu Se. [The
2
?
on the H abstraction reactions, eqns. (15) and (16), Scheme 4.
enthalpy change for the concerted reactions is prohibitively
21
endothermic (w400 kJ mol ), principally as a result of
the production of hydrogen radicals and gaseous Se. Just as
important however is the unfavourable A factor. The
entropy change for three molecules (two molecules of isobu-
0
This isotope effect also accounts for the larger amounts of d -2-
methylpropene produced.
Other possible reactions, which may become competitive at
t
high percentage decomposition of Bu
tion of H from 2-methylpropene or 2-methylpropane by Bu or
2
Se, include the abstrac-
t?
tene and one isobutane/dihydrogen molecule) forming a
60
?
H . For 2-methylpropane, H abstraction would occur from the
single transition state can be estimated
21 21
at ca.
23
. At a standard state of 1 mol dm , this
t?
tertiary C atom to give Bu , thus proceeding back into the cycle
2251 J mol
K
would become 2197 J mol
2
1
21
and not altering the ratio of 2-methylpropene to 2-methylpro-
pane. However, the low concentration of 2-methylpropane
except at high conversion makes this sequence unlikely. For 2-
methylpropene, a methyl H atom would be abstracted to give
the (allyl) stabilised 2-methylpropenyl radical. This reaction
would be thermodynamically favourable but would lead to 2,5-
K
, yielding an A factor of ca.
1
10 cm mol
0
6
22 21
s ].
We can, however, simulate the experimental data if we break
down the concerted reaction into a subset of reactions,
eqns. (15) or (16), (4) and (7) Scheme 4. The reaction
mechanism is very sensitive to the rate parameters chosen for
5
6
t
dimethylhexa-1,5-diene from coupling or to scrambling of the
H/D labels in the methyl groups of the 2-methylpropene
eqn. (3), the unimolecular decomposition of Bu Se, i.e. the
2
initiation step. The extent of reaction shifts dramatically
depending on the value of the activation energy for this step; a
0
18
t
2
formed from codecomposition of d - and d -Bu Se. No such
2
1
change of z21 kJ mol gives rise to virtually no decomposi-
scrambling is observed, nor is 2,5-dimethylhexa-1,5-diene, so H
abstraction from 2-methylpropene cannot be occurring. Our
calculations (see below) con®rm that, despite being thermo-
dynamically favourable, H abstraction from 2-methylpropene
is not a competitive reaction.
tion in the experimental temperature range. The value assigned
is the estimate computed from the bond dissociation energy for
t
Bu
2
Se at the B3LYP/6-311zG(2d,p)//B3LYP/6-311zG(2d,p)
level after making thermal corrections appropriate to 298 K.
The identi®cation of critical reaction pathways requires
accurate predictions of activation energies; `chemical accuracy'
Calculations. In order to try to substantiate the mechanism
t
2
1
that we are proposing for the decomposition of Bu
carried out ab initio and density functional theory (DFT)
2
Se, we have
generally refers to errors of 4±8 kJ mol . It is worthwhile
pointing out the likely accuracy of the computed thermo-
2
1
calculations on all of the Se-centred reactions shown in
Tables 1 and 2. Our new calculations show that the activation
chemical kinetic quantities, since a 6 kJ mol
activation energy causes a large uncertainty in k298; a factor
error in
t
2
parameters for a b-hydrogen abstraction within Bu Se are
such that this process can be ignored as a possible decomposi-
tion pathway, and that the lowest energy ®rst step is homolytic
of 10. Computed estimates of thermochemical quantities are
2
1
generally no better than 8±12 kJ mol for activation energies
via enthalpies of activation) and may well be worse. For
(
®ssion of the Se±C bond. This conclusion contradicts our
earlier deductions concerning the activation energies for the
entropies the situation is somewhat better, with differences
between computed estimates and experimental values typically
v5%. This is because an error of even 10% in the vibrational
frequencies leads to only a 1% error in the total entropy for a
single vibrational degree of freedom for a mode at 200 cm
and 300 K. In the present preliminary investigations such a
level of accuracy is acceptable.
2
6
b-hydride elimination reactions computed using the PM3
5
7
method. This discrepancy is discussed elsewhere.
The
conclusion is, however in agreement with those of Benson,
who has suggested that current kinetic interpretations of the
pyrolysis of alkyl sul®des, disul®des and mercaptans are all
21
5
8
seriously in error. The most serious errors have to do with the
9,60
Further insight into the reaction mechanism can be gained
by performing a quasi-stationary state analysis (QSSA) of the
free radical mechanism. In particular, chain length of the
radical chain has been estimated in order to gauge the
importance of possible radical non-chain processes and their
dependence on temperature relative to the radical chain. The
kinetic chain length l is de®ned as the rate of propagation
divided by the rate of termination or initiation, since the latter
5
presumed four-centre b-hydride elimination reaction
either H S from RSH or RSH from R S. He suggests that this
of
2
2
reaction may be too slow to be important in pyrolysis and this
indeed has been found from both DFT and MP2 calculations
for both sulfur and selenium centred alkyls and alkyl
3
8
hydrides, viz. the activation energies for the b-hydride
t
t
2
elimination reactions forming H Se from Bu SeH or Bu SeH
t
21
from Bu
2
Se are ¢230 kJ mol and of no importance under
60
two rates must be equal in the steady state. Hence:
standard growth and pyrolysis conditions (300±500 ³C). A
different situation appears to exist with the experimental data
on As compounds.
The free radical mechanism proposed to account for the
t
2
t
2
3=2
0
kobsꢀBu Se†
kobs
t
1=2
0
6
1
l0~
~
ꢀBu Se†
ꢀ1†
2
kiꢀBu Se†
is the initial concentration of Bu
ki
0
t
2
t
2
t
2
pyrolysis of Bu
Se in both dihydrogen and helium is listed
where (Bu
Se)
0
Se. Using the
along with the relevant kinetic data in Tables 1 and 2. Alkyl
radical gas phase chemistry is well documented and the kinetic
parameters for these steps were obtained from the combustion
literature. Rate constants for the reversible loss of
t?
hydrogen from Bu were taken from the paper of Gutman
QSSA computed value for kobs and the estimated values for k
i
we ®nd that at 700 K (in the middle of the experimental range)
l is ca. 3 for pyrolysis in helium but ca. 10 for pyrolysis in
dihydrogen. Therefore, the approximations made in using a
long chain length are valid (a long chain length is generally
6
2,63
6
3
and co-workers. The recombination of two H was taken
?
60
taken to be ¢10) for pyrolysis in dihydrogen but invalid for
6
4
from Tsang and Hampson,
hydrogen and deuterium and their radicals were taken from
the papers of Michael et al.
However, a problem arises because one reaction rate
while reactions involving
pyrolysis in helium). It is also important to note that while l
decreases with temperature (from 43 to 5) for pyrolysis in
dihydrogen, for pyrolysis in helium l increases with tempera-
ture (from 0.3 to 10). By using the computed values listed in
6
5,66
3
012
J. Mater. Chem., 1999, 9, 3005±3014

View Article Online
Scheme 5 Proposed mechanism for the photochemical decomposition
t
2
of Bu Se.
Table 1 it is possible to simulate the experimental results and
especially to reproduce the minimal effect of the decomposition
pro®le on changing the carrier gas from He to H
experimental and computed pro®les are shown together in
Fig. 3. Incorporation of H/D from the carrier gas, H or D
into 2-methylpropane is also rationalised because eqn. (14)
forms part of the mechanism.
On the basis of the experimental and theoretical observa-
2
. These
Scheme 6 Proposed mechanism for the growth of ZnSe by MOVPE
using Bu Se and Me Zn.
2
2
t
2
2
by 200 ³C if the studies are carried out in a tube coated with
2
5
ZnSe, compared with a glass tube. This is clear evidence that
surface catalysed processes are important for the decomposi-
tions carried out in the presence of ZnSe and hence for the
codecomposition reactions. If the reaction occurs on the
surface, reactions such as (12) or (13) in Scheme 3 should
become thermodynamically accessible since surface bound Se
and H are formed rather than the high energy species, Se and
tions, we conclude that the reactions occurring during the
t
thermal decomposition of Bu Se in helium or hydrogen are
those shown in Scheme 4.
2
t
2
Photochemical decomposition of Bu
Se
(
g)
There are marked differences in the products and product
distributions obtained from thermolysis and photolysis of
?
H
Me
(g). All of the observations for the codecomposition of
Zn with Bu Se can then be explained as shown in
2 2
t
t
2
9
Bu Se. In particular, the observation that d -Bu Se is
t
2
0
produced from photolysis of d - and d -Bu
18
t
Scheme 6, with all the reactions occurring on the surface.
2
Se shows that
? ?
Me abstracts H from surface bound Bu Se, as in eqn. (21),
2
t
t
Bu Se has a signi®cant lifetime at room temperature when it is
?
t
Scheme 6, to produce methane, 262-methylpropene, adsorbed
Se atoms and H(ads), which then starts a chain reaction
2
generated from Bu Se by Se±C bond cleavage. This and the
remaining observations suggest that radical±radical reactions
dominate over H abstraction from intact neutral molecules
(Scheme 5). Thus, the near 1 : 1 ratio of 2-methylpropene : 2-
methylpropane is as expected for reactions between tert-butyl
[eqn. (13), Scheme 6], from which the products are H
and 2-methylpropene. The labelled products are the expected
2
, Se(ads)
8
D and d -2-methylpropene if d -Bu
18
t
CH
3
2
Se is used in place of
0
d -Bu Se. Hydrogen is incorporated into the methane from the
t
2
radicals at room temperature [eqn. (5), Scheme 5], as is the
t
5
appearance of 2,2,3,3-tetramethylbutane. Curiously, Bu
4
2
carrier gas (H ) through eqn. (22), Scheme 6. The 2-methyl-
2
Se
2
?
t
propane formed if the reaction is carried out in hydrogen or
deuterium must arise from a sequential hydrogenation of 2-
methylpropene [eqn. (24), Scheme 6] since it contains 2H from
does not appear to be formed from combination of 26Bu Se
9
since no d -Bu
t
2
0
is produced when using d - and d -Bu
18
t
Se
2
2
Se.
These products presumably arise from a concerted reaction
t
between intact Bu Se and Se, perhaps photochemically
2
2
0
t
8
18
t
2 2 2
the carrier gas (d - from d -Bu Se in D or d - from d -Bu Se
in H ). This also accounts for the increased formation of 2-
2
initiated [eqn. (18), Scheme 5]. It is known that for ditellurides,
the reverse reaction, the thermal loss of tellurium, does occur
by a concerted mechanism [no scrambling of alkyl groups
methylpropane at higher temperature. Presumably, ZnSe
catalyses the hydrogenation reaction, since this reaction is
not observed in the absence of zinc.
6
8,69
occurs for RTe
2
R' (R, R' are different alkyls)].
t
Decomposition of Bu Se in the presence of Me Zn
Conclusions
2
2
Dimethylzinc has been shown to undergo reversible Zn±C
A detailed study of the products obtained from thermal
t
7
bond cleavage [eqns. (19) and (20), Scheme 6]. The products
0
decomposition of deuterium labelled and unlabelled Bu
allowed us to show that the decomposition proceeds by initial
Se has
2
t
from codecomposing it with Bu
methylpropene with only traces of 2,2-dimethylpropane. We
2
Se are methane and 2-
t
Se±C bond cleavage. Bu Se is not stable under thermal
?
7
1
t
have shown in studies of the codecomposition of Bu TeMe
with Me Cd that 2,2-dimethylpropane is a major product if free
Bu and Me are generated together. This means that few if
t
any free Bu } are generated during the codecomposition of
Me Se, and hence that a mechanism similar to that
?
of Scheme 4, but with initiation of the chain reaction by Me
reaction conditions but decomposes via a second Se±C bond
t?
?
cleavage. The Bu formed lose H or react with intact Bu Se
2
t
2
t
?
?
initiating a chain reaction to give 2-methylpropane, 262-
?
?
methylpropene, Se and H , which then acts as the chain carrier.
?
t
2
Zn and Bu
2
In the presence of Me
the chain is still carried by H . Although Se±H bonds are not
2
Zn, the radical chain initiator is Me , but
?
t
t
abstracting H from intact Bu
However, the labelling pattern for the products from Me Zn
2
Se cannot be operating.
formed in the decomposition of Bu
Me Zn, the fact that the chain carrier in the decomposition is
2
Se, nor in its reaction with
2
2
1
and d -Bu
8
t
2
1
8
Se in helium, d -methane and d -2-methylpropene
?
H means that passivation of e.g. N dopants may still be a
signi®cant problem in achieving active doping with nitrogen.
Photochemical initiation of the decomposition of Bu
proceeds by a free radical process, but this time neither Se±H
nor H are produced so this may present a better method for
obtaining active N doping in ZnSe.
suggest that H abstraction by Me radicals must be occurring.
The presence of methyl radicals is also suggested by the
observation that the methane produced in the codecomposition
t
2
Se again
1
8
t
?
of Me
by the reaction of Me with H . Gorochov and coworkers have
2
Zn and d -Bu
2
Se in hydrogen is largely CH
4
, formed
?
2
t
shown that the decomposition temperature of Bu
2
Se is reduced
Acknowledgements
}
Bu is stable towards loss of H .
At the temperature of the codecomposition reactions (%400 ³C),
t
?
?
54
We thank the EPSRC for Fellowships (N.L.P., D.F.F.).
J. Mater. Chem., 1999, 9, 3005±3014
3013

View Article Online
3
3
3
3
6
7
8
9
R. G. Parr and Y. Weitao, Density-Functional Theory of Atoms and
Molecules, Oxford University Press, Oxford, 1995.
J. M. Seminario and P. Politzer, Modern Density Functional
Theory: A Tool for Chemistry, Elsevier, Amsterdam, 1995.
N. Maung, J. O. Williams and A. C. Wright, J. Mol. Struct.
References
1
D. N. Armitage, H. M. Yates, J. O. Williams, D. J. Cole-Hamilton
and I. L. J. Patterson, Adv. Mater. Opt. Electron., 1992, 1, 43.
D. F. Foster, I. L. J. Patterson, L. D. James, D. J. Cole-Hamilton,
D. N. Armitage, H. M. Yates, A. C. Wright and J. O. Williams,
Adv. Mater. Opt. Electron, 1994, 3, 163.
2
(THEOCHEM), 1998, 453, 181.
GAUSSIAN 94 (Revision E.1), M. J. Frisch, G. W. Trucks,
H. B. Schlegel, P. M. W. Gill, B. C. Johnson, M. A. Robb,
J. R. Cheeseman, T. A. Keith, G. A. Petersson, J. A. Montgomery,
K. Raghavachari, M. A. Al-Laham, V. G. Zakrzewski, J. V. Ortiz,
J. B. Foresman, J. Cioslowski, B. B. Strefanov, A. Nanayakkara,
M. Challacombe, C. Y. Peng, P. Y. Ayala, W. Chen, M. W. Wong,
J. L. Andres, E. S. Replogle, R. Gomperts, R. L. Martin, D. J. Fox,
J. S. Binkley, D. J. Defrees, J. Baker, J. P. Stewart, M. Head-
Gordon, C. Gonzalez and J. A. Pople, Gaussian Inc., Pittsburgh,
PA, 1995.
SPARTAN version 4.1.2, Wavefunction Inc., 18401 Von Karman
Ave., #370, Irvine, CA92715, USA, #1995 Wavefunction, Inc.
W. J. Hehre, L. D. Burke, A. J. Shusterman and W. J. Pietro,
Experiments in Computational Organic Chemistry, Wavefunction,
Inc. Irvine, 1993.
W. J. Hehre, Practical Strategies for Electronic Structure
Calculations, Wavefunction, Inc. Irvine, CA, 1995.
W. J. Hehre, A. J. Shusterman and W. W. Huang, A Laboratory
Book of Computational Organic Chemistry, Wavefunction, Inc.
Irvine, CA, 1996.
Gaussian 94 (Revision D.1 and higher) User's Reference, Gaussian
Inc., Pittsburgh, PA, 1994±1996.
J. B. Foresman and A. E. Frisch, Exploring Chemistry with
Electronic Structure Methods, Gaussian Inc., Pittsburgh, PA, 2nd
edn., 1995±96.
J. J. P. Stewart, MOPAC 93, Fujitsu Limited, Tokyo, Japan, 1993.
J. J. P. Stewart, J. Comput. Chem., 1989, 10, 209.
M. J. S. Dewar and W. J. Thiel, J. Am. Chem. Soc., 1977, 99, 4899.
M. J. S. Dewar, E. G. Zoebisch, E. F. Healy and J. J. P. Stewart,
J. Am. Chem. Soc., 1985, 107, 3902.
T. A. Halgren and W. N. Lipscomb, Chem. Phys. Lett., 1977, 49,
225.
W. H. Press, S. A. Teukolsky, W. T. Vetterling and B. P. Flannery,
Numerical Recipes in FORTRAN±The Art of Scienti®c Computing,
Cambridge University Press, Cambridge, 1992.
Mathcad 6.0, Mathsoft Inc., 1996.
3
4
5
M. Danek, J. S. Huh, L. Foley and K. F. Jensen, J. Cryst. Growth,
1
994, 145, 530.
K. Nishimura, Y. Nagato and K. Sakai, Jpn. J. Appl. Phys., 1993,
2, L428.
3
W. S. Kuhn, A. Naumov, H. Stanzl, S. Bauer, K. Wolf,
H. P. Wagner, E. Gebhardt, U. W. Pohl, A. Krost, W. Richter,
U. Dumichen and K. H. Thiele, J. Cryst. Growth, 1992, 123, 605.
B. Bollig, M. Blauermel, W. Taudt and M. Henken, J. Cryst.
Growth, 1994, 145, 562.
W. Taudt, B. Wachtendorf, R. Beccard, A. Wahid, M. Henken,
A. L. Gurskii and K. Vakarelska, J. Cryst. Growth, 1994, 145, 582.
T. Obinata, K. Vesugi, G. Sato, I. Suemune, H. Machida and
N. Shimoyama, Jpn. J. Appl. Phys., 1995, 34, 4143.
6
7
8
9
4
4
0
1
W. S. Kuhn, R. Driad, H. Stanzl, A. Lusson, K. Wolf, B. Qu'Hen,
H. Sahin, L. Srob, C. Grattepain, X. Quesada, W. Grebhardt and
O. Gorochov, J. Cryst. Growth, 1994, 138, 448.
4
2
3
4
10
11
12
13
14
15
16
17
K. Wolf, H. Stanzl, A. Naumov, H. P. Wagner, W. S. Kahn,
B. Hahn and W. Gebhardt, J. Cryst. Growth, 1994, 138, 412.
P. J. Wright, B. Cockayne, P. J. Parbrook, P. E. Oliver and
A. C. Jones, J. Cryst. Growth, 1991, 108, 525 and references therein.
N. L. Pickett, D. F. Foster and D. J. Cole-Hamilton, J. Mater.
Chem., 1996, 6, 507.
N. L. Pickett, D. F. Foster and D. J. Cole-Hamilton, J. Cryst.
Growth, 1997, 170, 476.
N. L. Pickett, F. G. Riddell, D. F. Foster, D. J. Cole-Hamilton
and J. R. Fryer, J. Mater. Chem., 1997, 8, 1855.
A. Kamata, M. Mitsuhashi and H. Fujita, Appl. Phys. Lett., 1993,
4
4
4
5
46
47
48
49
6
3, 3353.
N. L. Pickett, D. F. Foster, W. G. Thomas, F. G. Riddell, D. J. Cole-
Hamilton and J. R. Fryer, J. Mater. Chem., 1998, 8, 2769.
H. Bock and S. Mohmand, Angew. Chem., Int. Ed. Engl., 1977, 16,
5
0
1
5
1
04.
1
1
8
9
G. Martin and N. Barroeta, Int. J. Chem. Kinet., 1980, 12, 699.
T. Hirabayashi, S. Mohmand and H. Bock, Chem., Ber., 1982,
5
5
2
3
1
15, 483.
C. G. Thompson, R. A. Mayer and J. S. Ball, J. Am. Chem. Soc.,
952, 74, 3284.
N. L. Pickett, D. F. Foster and D. J. Cole-Hamilton unpublished
observations.
R. N. Birrell and A. F. Trotman-Dickenson, J. Chem. Soc., 1960,
2
0
1
54
55
56
57
21
22
23
W. Tsang, J. Chem. Phys., 1964, 40, 1498.
4
218.
W. K. An, M. J. Kappers and R. F. Hicks, J. Cryst. Growth, 1997,
73, 386.
T. Bamkole, J. Chem. Soc., Perkin Trans. 2, 1977, 439.
M. Sugioka, T. Yotsuyangi and K. Aomura, Hokkaido
Daigaku Kugakuba Kenkyu Hokoku, 1970, 57, 191; Chem.
Abstr., 1973, 78, 29000w.
1
J. Stevenson, W. Bell, J. Ferry. D. J. Cole-Hamilton and J. E. Hails,
J. Organomet. Chem., 1993, 449, 141.
N. Maung, J. O. Williams and A. C. Wright, Chem. Mater., 1999,
submitted for publication.
S. W. Benson, J. Chem. Soc., Faraday Trans. 2, 1987, 83, 791.
S. W. Benson, Chem. Rev., 1978, 78, 23.
S. W. Benson, Thermochemical Kinetics, John Wiley & Sons, New
York, 1976.
D. F. Foster, C. Glidewell, G. R. Woolley and D. J. Cole-
Hamilton, J. Electron. Mater., 1995, 24, 1731.
W. Tsang, J. Phys. Chem. Ref. Data, 1990, 19, 1.
V. D. Knyazev, I. A. Dubinsky, I. R. Slagle and D. Gutman,
J. Phys. Chem., 1994, 98, 5279.
W. Tsang and R. F. Hampson, J. Phys. Chem. Ref. Data, 1986, 15,
1087.
J. V. Michael and J. R. Fisher, J. Phys. Chem., 1990, 94, 3318.
J. V. Michael, J. Chem. Phys., 1990, 92, 3394.
J. I. Stein®eld, J. S. Francisco and W. L. Hase, Chemical Kinetics
and Dynamics, Prentice-Hall, Englewood Cliffs, NJ, 1989.
A. E. D. McQueen, M. B. Parker, J. B. Mullin and D. J. Cole-
Hamilton, Chemtronics, 1989, 4, 264.
A. E. D. McQueen, P. N. Culshaw, J. C. Walton. D. V. Shenai-
Khatkhate, D. J. Cole-Hamilton and J. B. Mullin, J. Cryst.
Growth, 1991, 107, 325.
J. E. Hails and D. J. Cole-Hamilton, unpublished observations.
J. E. Hails, D. J. Cole-Hamilton and A. E. D. McQueen, J. Crystal
Growth, 1998, 183, 594.
24
25
26
M. Yamada, T. Kamo, J. Tang, Y. Oshima and A. Amano, Nippon
Kagaku Kaishi, 1985, 2283; Chem. Abstr., 1986, 104, 224462f.
W. S. Kuhn, R. Helbing, B. Qu'Hen and O. Gorochov, J. Cryst.
Growth, 1995, 146, 580.
G. H. Fan, N. Maung, T. L. Ng, P. F. Heelis, J. O. Williams,
A. C. Wright, D. F. Foster and D. J. Cole-Hamilton, J. Cryst.
Growth, 1997, 170, 485.
58
59
60
61
27
28
29
W. Bell, J. Stevenson, D. J. Cole-Hamilton and J. E. Hails,
Polyhedron, 1994, 13, 1253.
N. L. Pickett, D. F. Foster, D. J. Cole-Hamilton and J. E. Hails,
Phosphorus Sulphur Silicon, 1998, 136±138, 427.
N. L. Pickett, D. F. Foster and D. J. Cole-Hamilton, in
Fundamental Gas-phase and Surface Chemistry of Vapour-phase
Synthesis, eds. M. D. Allendorf, M. R. Zachariah, L. Mountziaris
and A. H. McDaniel, Electrochemical Society Proceedings,
Pennington, 1999, vol. 98±23, p. 307.
6
6
2
3
64
65
66
67
3
3
0
1
D. F. Foster and D. J. Cole-Hamilton, Inorg. Synth., 1997, 31, 29.
D. F. Foster, N. L. Pickett and D. J. Cole-Hamilton, Polyhedron,
6
6
8
9
1
999, 18, 1329.
32
33
34
R. C. Morrison, R. W. Hall, J. A. Schwindeman, C. W. Kamienski
and J. F. Engel, US Pat., 5340507, 1994.
R. C. Morrison, R. W. Hall, J. A. Schwindeman, C. W. Kamienski
and J. F. Engel, Eur. Pat. Appl., 0525881Al, 1993.
L. D. Stockton, T. L. Ng, N. Maung, I. B. Poole, J. O. Williams,
A. C. Wright, D. F. Foster and D. J. Cole-Hamilton, J. Cryst.
Growth, 1998, 183, 95.
7
7
0
1
3
5
W. J. Hehre, L. Radom, P. R. v. Schleyer and J. A. Pople, Ab Initio
Molecular Orbital Theory, John Wiley and Sons, New York, 1986.
Paper 9/04270I
3
014
J. Mater. Chem., 1999, 9, 3005±3014