IR laser-induced thermolysis of silacyclopent-3-ene: Extrusion of silylene and chemical vapour deposition of polycarbosilane phases via reactions of silylene, buta-1,3-diene and methylene

Article abstract of DOI:10.1016/S0022-328X(00)00314-4

Infrared laser-induced (SF₆) photosensitized decomposition and infrared laser multiphoton decomposition of silacyclopent-3-ene occur as extrusion of silylene, yielding butadiene as a major gaseous product and affording chemical vapour deposition of solid saturated polycarbosilane films. The involvement of H₂Si:, H₂C: and buta-1,3-diene in the formation of the films is revealed through quantification of the gaseous products and identification of H₂Si: and H₂C: by laser induced fluorescence.

Full text of DOI:10.1016/S0022-328X(00)00314-4

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 605 (2000) 202–208
IR laser-induced thermolysis of silacyclopent-3-ene: extrusion of
silylene and chemical vapour deposition of polycarbosilane phases
via reactions of silylene, buta-1,3-diene and methylene
a
Josef Pola a,*, Marke´ta Urbanova´ , Luis D´ıaz b, Magna Santos b, Zdene˘k Bastl c,
d
&
Jan Subrt
a Institute of Chemical Process Fundamentals, Academy of Sciences of the Czech Republic, 165 02 Prague 6, Czech Republic
b Instituto de Estructura de la Materia, CSIC, Serrano 121, 28006 Madrid, Spain
c J. Heyro6sky´ Institute of Physical Chemistry, Academy of Sciences of the Czech Republic, 182 23 Prague 8, Czech Republic
d
´
Institute of Inorganic Chemistry, Academy of Sciences of the Czech Republic, 250 68 Rez near Prague, Czech Republic
Received 18 April 2000; accepted 18 May 2000
Abstract
Infrared laser-induced (SF6) photosensitized decomposition and infrared laser multiphoton decomposition of silacyclopent-3-
ene occur as extrusion of silylene, yielding butadiene as a major gaseous product and affording chemical vapour deposition of
solid saturated polycarbosilane films. The involvement of H2Si:, H2C: and buta-1,3-diene in the formation of the films is revealed
through quantification of the gaseous products and identification of H2Si: and H2C: by laser induced fluorescence. © 2000
Elsevier Science S.A. All rights reserved.
Keywords: Silacyclopent-3-ene; Laser-induced thermolysis; Silylene; Chemical vapour deposition; Si/C/H polymers
of H2Si: and subsequent self-polymerization of this
species into rare poly(hydridosilanes). However, UV
laser photolysis of gaseous SCP, although being a neat
extrusion of silylene and yielding buta-1,3-diene as a
major gaseous product, affords chemical vapour depo-
sition of Si/C/H phases. The formation of these phases
is due to the occurrence of polymerization steps of both
H2Si: and the products of buta-1,3-diene photolysis [7].
It was our aim to compare SCP decomposition from
the electronically excited state and vibrationally excited
ground state. In this study we examined IR laser-in-
duced thermolysis of SCP in the absence and presence
of photosensitizing SF6 in an effort to elucidate the
mechanism and to find out whether silylene is involved
in reactions with other products of SCP thermolysis.
1. Introduction
Thermal or photolytic decomposition of silacy-
clopent-3-enes is regarded [1–3] as a reversible process
[4] occurring via transient 2-vinyl-1-silirane [5,6] and
affording products of rearrangement and silylene extru-
sion (Scheme 1).
These two pathways have different significance de-
pending on the substituents at the silicon [1,2,5,6]. The
decomposition of the parent silacyclopent-3-ene (SCP)
might have been expected to occur as a clean extrusion
2. Experimental
Scheme 1.
Laser thermolysis experiments were conducted with a
tunable TEA CO2 laser (Plovdiv University) operating
on the P(14) and P(18) line of the 0001“1000 transition
* Corresponding author. Tel.:+420-2-20390308; fax: +420-2-
20920661.
E-mail address: pola@icpf.cas.cz (J. Pola).
0022-328X/00/$ - see front matter © 2000 Elsevier Science S.A. All rights reserved.
PII: S0022-328X(00)00314-4

J. Pola et al. / Journal of Organometallic Chemistry 605 (2000) 202–208
203
(949.50 and 945.98 cm−1, respectively). The former line
was used for the irradiation of SCP (10 Torr) and the
latter for the irradiation of the equimolar mixtures of
SCP and sensitizing SF6 (total pressure 20 Torr). The
unfocused radiation at both lines was of the same
fluence (0.5 J cm−2). The wavelength and fluence were
checked by a model 16-A spectrum analyser (Optical
Eng. Co.) and by a pyroelectric detector (ml-1 JU,
Charles University).
duced by an N2-pumped dye laser (PRA LN107) with a
bandwidth of 1.6 cm−1 at 580 nm. The dye laser beam,
counter-propagating to the CO2 laser beam, was fo-
cused by a 50 cm quartz lens at the focus of the IR
laser. The SL and LIF signals were detected through
the lateral window and focused onto an R928 Hama-
matsu photomultiplier. The LIF excitation spectrum
was collected through a Schott RG610 glass filter and
the SL emission through the same filter for SiH2
(A1B1“X1A1) detection and through an interference
filter centred at 560 nm (10 nm FWHM) for the C2
(d3Pg “ a3Pu) Dw= −1 Swan band detection.
The LIF signal could be detected only with low
pressures of SCP (0.02–0.45 Torr). With higher SCP
pressures, an intense visible spontaneous luminescence
signal appears that makes detection of the induced
fluorescence impossible. The appearance of the sponta-
neous luminescence background poses limits on the
fluence that can be used in the irradiation of SCP for
LIF detection.
The photoelectron spectra were acquired using ESCA
3 Mk II (VG Scientific) and ESCA 310 (Gammadata
Scienta) electron spectrometers equipped with an Al–
Ka X-ray source (hw=1486.6 eV). The background
pressure during spectra acquisition was typically of the
order of 10−6 Pa. The spectrometers were operated in
the fixed analyser transmission mode. Detailed spectral
scans were taken over Si (2p), C (KLL) and Si (KLL)
regions. The XPS peak positions and areas were deter-
mined by fitting the unsmoothed experimental spectra
after subtraction of Shirley [9] background. The surface
concentrations of elements were determined using theo-
retical photoionization cross-sections [10].
Gaseous samples of SCP were irradiated in a Pyrex
vessel (45 mm i.d., 10 cm length) equipped with two
NaCl windows, a PTFE stopcock and a sleeve with a
rubber septum. The cell accommodated metal sub-
strates which after the irradiation of SCP were covered
with solid deposited materials and transferred for mea-
surements of their properties by SEM, FTIR and XPS.
Changes in the composition of the gaseous content of
the vessel after laser irradiation were monitored by an
FTIR (Nicolet, model Impact 400) spectrometer. The
depletion of SCP was followed using the diagnostic
band at 2163 cm−1. Some gaseous products (silane,
ethene, ethyne, methane and buta-1,3-diene) were de-
tected using the diagnostic absorption bands of these
compounds. Identification of the gaseous products was
also accomplished by GC/MS. The GC/MS and GC
analyses of the gaseous samples after irradiation were
performed on a Shimadzu QP 1000 mass spectrometer
and Shimadzu 14A chromatograph with FID detector
which was coupled with a Chromatopac C-R5A com-
puting integrator. Both instruments were equipped with
Porapak P and SE-30 columns and were operated at the
programmed temperature (20–150°C) using helium car-
rier gas. The quantitative GC analyses were based on
knowledge of the response factors for the identified
products, which were determined or taken from ref. [8].
Real time studies of the IR multiphoton decomposi-
tion of SCP were carried out using experimental set-ups
for monitoring laser-induced fluorescence (LIF) and
spontaneous luminescence emission (SL) of the formed
photofragments. For these studies, a Lumonics K-103
TEA CO2 laser tuned to the 10P(14) line was employed
to initiate the decomposition of gaseous SCP in Pyrex
cells similar to that mentioned above but with a lateral
quartz window to detect the LIF and SL emissions. At
the low fluences obtained with unfocused geometry, the
signal-to-noise ratio was not sufficient to collect the
spectrum; real time detection of the fragments was
therefore carried out at higher fluences using focused
geometry. The CO2 laser beam was focused at the
centre of the cell with a 24 cm NaCl lens. The fluence
was calculated as the ratio of the pulse energy, as
measured with a Lumonics 20D pyroelectric detector,
and the FWHM cross-sectional beam area, measured at
the cell position with a pyroelectric array Delta Devel-
opment Mark IV. The resulting fluence was ca. 20 J
cm−2. Fluorescence in the formed fragments was in-
SEM and TEM studies of the deposits were carried
out on an ultrahigh vacuum Tesla BS 350 instrument
and a Philips 201 microscope.
SCP was prepared following the procedure in ref. [11]
and its purity was better than 98% as checked by gas
chromatography.
3. Results and discussion
The TEA CO2 laser-induced decomposition of SCP
carried out by irradiating SCP (10 Torr) in l(SiH2)
mode (absorptivity of SCP at 950 cm−1, 6.8×10−3
Torr−1 cm−1) is a much slower process than decompo-
sition carried out by irradiating mixtures of SCP and
SF6 (each 10 Torr). In the former, IR multiphoton
decomposition (IRMPD, e.g. [12,13]), SCP is vibra-
tionally excited through direct absorption and a 30%
decomposition of SCP is accomplished with as many as
6000 pulses. In the latter, laser-photosensitized decom-
position (LPD, e.g. [14,15]), SCP is vibrationally excited
mostly through collisions with better absorbing SF6
(absorptivity of SF6 at 946 cm−1, 230 Torr−1 cm−1
)

204
J. Pola et al. / Journal of Organometallic Chemistry 605 (2000) 202–208
Fig. 1. Yield of gaseous products (in relative mole per cent (a) and in mole/mole of SCP decomposed (b)) at different stages of IRMPD of SCP.
1,3-C4H6 ); C2H4 (ꢀ); C2H4+C2H2 (×).
(
Fig. 2. Yield of gaseous products (in relative mole per cent (a) and in mole/mole of SCP decomposed (b)) at different stages of LPD of SCP.
1,3-C4H6 ); C2H4 (ꢀ); C2H2 ("); CH4 (ꢁ); C2H4+C2H2 (×).
(
and the same 30% decomposition of SCP can be
achieved with as few as 30 pulses of the same fluence.
(Fig. 1a). In LPD, the distribution of the volatile prod-
ucts (in relative mole per cent) is as follows: buta-1,3-
diene (40–83), buta-1,2-diene (B1), ethene (15–24),
ethyne (10–16), methane (2–6), propene (2–4), silane
(1–2), 1-silyl-1-silacyclopent-3-ene (1–2). These values
are rather different; buta-1,3-diene is still a major
product but it is accompanied with significant amounts
of ethene and ethyne (Fig. 2a).
The distribution of the volatile products in IRMPD
is similar to that observed in ArF laser-induced photol-
ysis of SCP carried out in excess helium [7]. The
dominant formation of buta-1,3-diene in the IRMPD
proves the extrusion of silylene, and the presence of
1-silyl-1-silacyclopent-3-ene indicates insertion of
silylene into SCP. The formation of silane can be
explained by the secondary photolysis of 1-silyl-1-sila-
3.1. Final products
Both processes result in the deposition of a faintly
yellow solid deposit and in the formation of hydrocar-
bons, silane and 1-silyl-1-silacyclopent-3-ene. The rela-
tive amounts of hydrocarbons for both processes are
different. In IRMPD, the distribution of the volatile
products (in relative mole per cent) is as follows: buta-
1,3-diene (80–96), buta-1,2-diene (B1), ethene (5–8),
ethyne (2–3), methane (B1), propene (1–2), silane
(2–3), 1-silyl-1-silacyclopent-3-ene (B1). These values
are hardly altered as the decomposition progresses, ca.
10–35%, buta-1,3-diene being the dominant product

J. Pola et al. / Journal of Organometallic Chemistry 605 (2000) 202–208
205
cyclopent-3-ene, since addition of H2 to silylene is
significantly slower than insertion of silylene into the
Si–H bond [16,17]. The lower yields of buta-1,3-diene
and higher yields of ethene and ethyne (and appar-
ently of H2) in LPD are in accordance with cleavage
of buta-1,3-diene [18]. The distribution of the products
(in mole/mole of SCP decomposed) reveals that the
absolute yields of buta-1,3-diene in both IRMPD and
LPD are rather small; thus only the 0.25–0.60 fraction
of buta-1,3-diene along with a 0.20–0.27 fraction of
C2 products (ethene+ethyne) persist in the gas phase
for LPD (Fig. 2b), and even smaller fractions of buta-
1,3-diene (0.36–0.40) together with C2 products (ca.
0.05) are observed in the gas phase for IRMPD (Fig.
1b).
These findings reveal that buta-1,3-diene (and per-
haps also the C2 products) must react with silylene
and be involved in agglomerization reactions leading
to the deposition of the solid product. They also show
that the decomposition of buta-1,3-diene is less impor-
tant and the agglomerization reactions of buta-1,3-di-
ene are more feasible in IRMPD compared to LPD.
The assumed intermediate products of SCP decom-
position are confirmed by the LIF spectra of the pho-
tofragments produced during IRMPD which was
conducted at similar incident fluences.
Fig. 3. Excitation spectrum in the 578–594 nm interval of the photofragments formed during IRMPD of 0.15 Torr of SCP.
Fig. 4. Luminescence signals produced during IRMPD of 0.4 Torr of SCP: first shot (—), average of twenty shots ( ).

206
J. Pola et al. / Journal of Organometallic Chemistry 605 (2000) 202–208
Fig. 6. Fitted photoelectron spectrum of Si 2p electrons of the films
obtained by IRMPD (1) and LPD (2).
Scheme 2.
sures and/or high fluences. This signal was dispersed by
placing a monochromator in front of the photomulti-
plier, but the background, probably originating from
scattering of the luminescence in the depositing solid
particles, did not allow resolution of the specific bands.
It was observed (Fig. 4) that the profile of the SL
changes becomes wider with more shots and that it is
recovered only after some time is allowed to elapse.
This behaviour is likely due to the different nature of
the emission in the first and subsequent shots arising
from an accumulation of products and their photolysis
in the viewing region. The addition of Ar to SCP
produces a large increase in the SL signal of H2C: and
an appreciable decrease in the SL signal of :SiH2. These
changes are consistent with the initial unimolecular
production of :SiH2 followed by collisional dissociation
of the formed products giving rise to :CH2.
3.2. Real time analysis of IRMPD of SCP
The LIF excitation spectrum in the 578–595 nm
region is shown in Fig. 3. The spectrum, collected at a
time delay between the pulses of CO2 and dye lasers of
900 ns and at a pressure of 0.15 Torr of SCP, shows
several vibro-rotational bands belonging to :SiH2, :CH2
and H2. These bands were assigned to :SiH2 [19,20],
:CH2 [21] and H2 [22,23]. The bands centred at 580.2,
585.0 and 593.4 nm belong to the A1B1’X1A1 Dw=2
transition of the :SiH2 fragment. For the :CH2 species,
several lines belonging to the P (591.6, 592.2 and 593
nm), Q (591 and 591.6 nm) and R (589.05, 589.25,
1
589.45 and 589.75 nm) branches of the B1(0,14,0)’
1A1(0,0,0) transition and to the P (580.6, 582.5 and
1
584.6 nm) branch of the B1(0,16,0)’1A1(0,1,0) transi-
tion were found. The lines at 581 and 582.9 nm can be
assigned to molecular hydrogen which also contributes,
together with :CH2, to the strong line at 589.05 nm.
A spontaneous visible luminescence signal is ob-
served with the experiments carried out at high pres-
3.3. Gas-phase chemistry
It is conceivable that the observed final and transient
products can be reconciled with chemical reactions
Fig. 5. Typical FTIR spectrum of the deposit from SCP.

J. Pola et al. / Journal of Organometallic Chemistry 605 (2000) 202–208
207
described in Scheme 2, where SCP undergoes decompo-
sition to silylene and buta-1,3-diene, and where the
latter initial product decomposes into ethene, ethyne,
:CH2 species and hydrogen. The observation of 1-silyl-
1-silacyclopent-3-ene and silane among the volatile final
products is in line with the insertion of silylene into the
Si–H bond of SCP, and with secondary photolysis of
1-silyl-1-silacyclopent-3-ene into silane and silacy-
clopent-3-enylsilylene (see [16,17]). Other plausible reac-
tions of silylene are those with unsaturated
hydrocarbons. All of these contribute to the formation
of Si–C–H polymers as confirmed by analysis of the
solid material.
3.4. Properties of the solid deposit
The inference about the involvement of buta-1,3-di-
ene and other unsaturated carbon species in the forma-
tion of the solid material is in line with the IR and XP
spectral analysis of the solid deposit which is shown to
contain Si, C and H atoms. The pattern of the IR
spectrum (wavenumber (cm−1) (normalized absorptiv-
ity)) is as follows: 850 (1), 1245 (ca. 0.075), 2150 (ca.
0.16) and 2930 (ca. 0.05). This does not differ if the
films are produced by LPD or IRMPD (Fig. 5) and it
can be assigned, in the given order, to w(Si–C),
l(CH3Si), w(Si–H) and w(C–H) vibrations. The ab-
sence of absorption bands belonging to the CꢀC or
CꢁC bonds indicates that the films can be described as
completely saturated polycarbosilanes. The absorptivity
at the w(Si–H) and w(C–H) vibrations is instructive of
the distribution of H atoms between the Si and C
centres [24]; the A[w(C–H)]:A[w(Si–H)] ratio of the
films of ca. 0.25–0.30 shows that the concentration of
H(C) and H(Si) atoms is almost equal.
Fig. 7. Fitted photoelectron spectrum of Si KLL Auger electrons of
the films obtained by IRMPD (1) and LPD (2).
XPS analysis of the topmost (ca. 4 nm) layers of the
deposit (Figs. 6 and 7) reveals that the films produced
by IRMPD and LPD differ. The former possess exclu-
sively silicon at 101.1 eV assignable to an Si–C–H
polymer, whereas the latter are composed of silicon in
the Si–C–H polymer (63%) and of silicon at 102.8 eV
in an Si–O bond (27%) [25,26]. This indicates that the
films produced by IRMPD are not oxidized in air when
they are transferred from the reactor to the spectrome-
ter. The development of the Si–O component can be
attributed to a small concentration of unsaturated SiꢀC
bonds in superficial layers (see for example refs. [27,28])
and to their oxidation in atmosphere by reaction with
molecular oxygen [29]. The different propensity of the
polycarbosilane to undergo oxidation indicates that the
Si–C–H films from IRMPD do not contain significant
concentrations of SiꢀC moieties, in contrast to those
from LPD. Comparison of the C KLL Auger spectra of
the deposits with those of diamond and graphite [30]
reveals that the deposits from LPD and IRMPD pos-
sess solely sp3 carbon (Fig. 8) and this supports the
conclusion drawn from the FTIR spectra.
Fig. 8. C KLL derivative Auger spectra of (1) diamond, (2) graphite,
(3) the deposit from LPD and (4) the deposit from IRMPD.
SEM images of the deposits obtained from IRMPD
and LPD have the same pattern (Fig. 9) and show that
Fig. 9. Scanning electron micrograph of the deposit.

208
J. Pola et al. / Journal of Organometallic Chemistry 605 (2000) 202–208
Czech Republic (grant no. 104/00/1294) and by Spanish
DGES (project no. PB96-0844-C02-02).
References
[1] E.A. Chernyshev, N.G. Kovalenkova, S.A. Bashkirova, J.
Organomet. Chem. 271 (1984) 129.
[2] D. Lei, P.P. Gaspar, Organometallics 4 (1985) 1471.
[3] M.G. Steinmetz, C. Yu, Organometallics 11 (1992) 2686.
[4] D.S. Rogers, H.E. O’Neal, M.A. Ring, Organometallics 5 (1986)
1467.
[5] P.P. Gaspar, in: H. Sakurai (Ed.), Organosilicon and
Bioorganosilicon Chemistry, Horwood, Chichester, 1985, p. 88.
[6] I.M.T. Davidson, in: E.R. Corey, J.Y. Corey, P.P. Gaspar
(Eds.), Silicon Chemistry, Horwood, Chichester, 1988, p. 173.
[7] J. Pola, A. Ouchi, M. Urbanova, Y. Koga, Z. Bastl, J. Subrt, J.
Organomet. Chem. 575 (1999) 246.
Fig. 10. Transmission electron micrograph of the deposit (magnifica-
tion 100 000×).
[8] W.A. Dietz, J. Gas Chromatogr. 49 (1967) 151.
[9] D.A. Shierly, Phys. Rev. B 5 (1972) 4709.
[10] J.H. Scofield, J. Electron. Spectrosc. 8 (1976) 129.
[11] R. Damrauer, R. Simon, A. Laporterie, G. Manuel, Y.T. Park,
W.P. Weber, J. Organomet. Chem. 391 (1990) 7.
[12] C.D. Cantrell (Ed.), Multiple-Photon Excitation and Dissocia-
tion of Polyatomic Molecules, Springer, Berlin, 1986.
[13] R.V. Ambartzumian, S.V. Letokhov, in: C.B. Moore (Ed.),
Chemical and Biochemical Applications of Lasers, vol. 3, Aca-
demic Press, New York, 1987.
these deposits have a particulate structure and consist
of fluffy agglomerates. TEM analysis (Fig. 10) reveals
that these agglomerates are composed of small chain-
like structures several tens of nanometres in size.
[14] D.K. Russel, Chem. Soc. Rev. 19 (1990) 407.
[15] J. Pola, Spectrochim. Acta A 46 (1990) 607.
[16] J.M. Jasinski, R. Becerra, R. Walsh, Chem. Rev. 95 (1995) 1203.
[17] J.G. Martin, H.E. O’Neal, M.A. Ring, Int. J. Chem. Kinet. 22
(1990) 613.
[18] J.H. Kiefer, K.I. Mitchell, H.C. Wei, Int. J. Chem. Kinet. 17
(1985) 225.
[19] I. Dubois, Can. J. Phys. 46 (1968) 2485.
[20] M. Fukushima, S. Mayama, K. Obi, J. Chem. Phys. 96 (1992)
44.
[21] G. Herzberg, J.W.C. Johns, Proc. R. Soc. London Ser. A 295
(1966) 107.
[22] G. Herzberg, Molecular Spectra and Molecular Structure I.
Spectra of Diatomic Molecules, Van Nostrand, New York, 1950.
[23] H.G. Gale, G.S. Monk, K.O. Lee, Astrophys. J. 67 (1928) 89.
[24] H.C. Low, P. John, J. Organomet. Chem. 201 (1980) 363 and
refs. therein.
4. Conclusion
The CO2 laser induced IR multiphoton decomposi-
tion and photosensitized (SF6) decomposition of
gaseous SCP proceeded as extrusion of silylene and
afforded buta-1,3-diene as a major volatile product.
The other products of SCP decomposition (silylene,
methylene and H−2 ) were identified by laser induced
fluorescence. The transient silylene does not self poly-
merize to poly(hydridosilanes), but takes part in ag-
glomerization reactions with buta-1,3-diene and
presumably with other unsaturated species such as
ethene, ethyne and :CH2. Buta-1,3-diene is utilized
more for the formation of the solid polycarbosilane
films in IRMPD, whereas in LPD it is more efficiently
cleaved into ethene and ethyne. The topmost layers of
the films obtained by LPD, but not those of the films
obtained by IRMPD, are oxidized upon contact to air.
[25] IST X-ray Photoelectron Spectroscopy Database, US Depart-
ment of Commerce, NIST, Gaithersburg, MD, 1997.
&
[26] M. Urbanova´, H. Morita, V. Dr´ınek, Z. Bastl, J. Subrt, J. Pola,
J. Anal. Appl. Pyrol. 44 (1998) 219.
&
[27] J. Pola, Z. Bastl, J. Subrt, R. Taylor, J. Mater. Chem. 5 (1995)
1345.
&
[28] J. Pola, Z. Bastl, J. Subrt, R. Abeysinghe, R. Taylor, J. Mater.
Chem. 6 (1966) 155.
[29] I.M.T. Davidson, C.E. Dean, F.T. Lawrence, J. Chem. Soc.
Chem. Commun. (1981) 52.
Acknowledgements
[30] J.C. Lascovich, R. Giorgi, S. Scaglione, Appl. Surf. Sci. 47
(1991) 17.
The work was supported by the Grant Agency of the
.