Laser pyrolysis studies of β-diketonate chemical vapour deposition precursors. Part 1: β-diketones

Article abstract of DOI:10.1039/b412486n

The thermal decomposition of the β-diketones pentane-2,4-dione, 1,1,1,5,5,5-hexafluoropentane-2,4-dione, and 2,2,6,6-tetramethylheptane-3,5- dione has been studied using the technique of IR laser-powered homogeneous pyrolysis combined with the ab initio calculation of activation energies of possible reaction pathways. Pentane-2,4-dione decomposes by two parallel molecular elimination pathways, the keto tautomer affording acetone and ketene via a 6-electron retroene reaction and the enol tautomer 2-methylfuran via H₂O elimination, followed by cyclisation. 1,1,1,5,5,5- Hexafluoropentane-2,4-dione eliminates HF with ring closure to yield 2,2-difluoro-5-(trifluoromethyl)furan-3(2H)-one. There is also a minor route of loss of one or two CF₃ radicals, followed by H-abstraction to yield 3,3-difluoroacryloylfluoride and carbon suboxide, respectively. 2,2,6,6-Tetramethylheptane-3,5-dione decomposes via central C-CO bond homolysis, followed by CO or ketene loss from the resultant radicals and subsequent disproportionation of t-Bu radicals to produce 2-methylpropene and 2-methylpropane. The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2005.

Full text of DOI:10.1039/b412486n

View Article Online / Journal Homepage / Table of Contents for this issue
P A P E R
Laser pyrolysis studies of b-diketonate chemical vapour deposition
precursors. Part 1: b-diketones
Douglas K. Russell* and Anna Yee
Department of Chemistry, University of Auckland, Private Bag 92019, Auckland, New Zealand.
E-mail: d.russell@auckland.ac.nz; Fax: +64 9 373 7422; Tel: +64 9 373 7599, ext: 88303
Received (in Montpellier, France) 9th August 2004, Accepted 21st September 2004
First published as an Advance Article on the web 19th January 2005
The thermal decomposition of the b-diketones pentane-2,4-dione, 1,1,1,5,5,5-hexafluoropentane-2,4-dione,
and 2,2,6,6-tetramethylheptane-3,5-dione has been studied using the technique of IR laser-powered
homogeneous pyrolysis combined with the ab initio calculation of activation energies of possible reaction
pathways. Pentane-2,4-dione decomposes by two parallel molecular elimination pathways, the keto tautomer
affording acetone and ketene via a 6-electron retroene reaction and the enol tautomer 2-methylfuran via H2O
elimination, followed by cyclisation. 1,1,1,5,5,5-Hexafluoropentane-2,4-dione eliminates HF with ring closure
to yield 2,2-difluoro-5-(trifluoromethyl)furan-3(2H)-one. There is also a minor route of loss of one or two
CF3 radicals, followed by H-abstraction to yield 3,3-difluoroacryloylfluoride and carbon suboxide,
respectively. 2,2,6,6-Tetramethylheptane-3,5-dione decomposes via central C–CO bond homolysis, followed
by CO or ketene loss from the resultant radicals and subsequent disproportionation of t-Bu radicals to
produce 2-methylpropene and 2-methylpropane.
(BOC) were purified before use by repeated freeze-pump-thaw
Introduction
cycles. Materials were handled on a Pyrex vacuum line fitted
with greaseless J. Youngs taps; before use, the line was pre-
conditioned by exposure to the vapour under study and re-
evacuation. Precursor and product identification and analysis
(FT-IR spectroscopy and GC-MS) were accomplished using
commercial instrumentation in conjunction with comparison
with data from authentic samples.
Transition metal complexes of the anions of b-diketone ligands
such as acetylacetone (acacH, pentane-2,4-dione, 1) have been
well-known for many years and exhibit a range of structural
motifs, reactivity and stability.1 More recently, neutral b-dike-
tone complexes have been exploited as volatile precursors to the
deposition of metals in materials as diverse as semiconductors
and high temperature superconductors.2 While there has been
considerable work of a developmental or exploratory nature,
rather little is known about the detailed mechanisms of thermal
decomposition of such complexes. We have shown that the
technique of infrared laser-powered homogeneous pyrolysis (IR
LPHP)3,4 is uniquely suited to the elucidation of thermal
decomposition reaction mechanisms.5,6 When allied with stan-
dard analytical methods and augmented by techniques for the
detection and identification of short-lived reaction intermediates
and ab initio evaluation of the feasibility of reaction routes, very
detailed information may be obtained. The value of this multi-
lateral approach has been demonstrated in systems as varied as
precursors to chemical vapour deposition,7–11 various cyclic
systems12,13 and small organic molecules.14
In the present work, we report IR LPHP studies of the three
common diketones acacH, 1,1,1,5,5,5-hexafluoropentane-2,4-
dione (hexafluoroacetylacetone, hfacH, 2), and 2,2,6,6-tetra-
methylheptane-3,5-dione (thdH, 3). Not only are complexes
derived from these used widely as deposition precursors, but
the compounds themselves exhibit a range of thermal decom-
position behaviour, resulting from subtle changes in the bal-
ance of keto-enol equilibria and bond strengths. In future
work, we shall report parallel studies of b-diketonate com-
plexes of the metals used in the production of the high
temperature superconductors, namely Cu, Ba and Y.15
Infrared laser-powered homogeneous pyrolysis
All static cell pyrolyses utilised the IR LPHP technique. Since
this method has been described in detail elsewhere, only a brief
description is given here.3–6 Pyrolysis is performed in a cylind-
rical Pyrex cell (length 100 mm, diameter 38 mm) fitted with
ZnSe windows. Although ZnSe is opaque to infrared radiation
below 500 cmÀ1, it has several distinct advantages over cheaper
materials, such as NaCl. ZnSe is strong and thermally stable,
and non-hygroscopic. Most significantly, ZnSe is highly trans-
parent to the CO2 laser radiation. The pyrolysis cell is filled
with between one and two torr (1 torr = 133.3 Pa) of the
vapour under study and approximately 10 torr of SF6.
The contents of the cell are then exposed to the output of a
free running CW CO2 laser operating at 10.6 mm. The laser
power level (i.e., temperature) is generally set to initiate
decomposition of the target compound, with an exposure time
sufficient to provide an analysable yield of products, typically a
few percent of decomposition. As shown elsewhere,3,6 SF6
strongly absorbs the laser radiation, which is then rapidly
converted to heat via efficient inter-molecular and intra-mole-
cular relaxation. The low thermal conductivity of SF6 ensures
that a strongly non-uniform temperature profile is produced in
which the centre of the cell may reach temperatures of the
order of 1500 K while the cell wall remains at room tempera-
ture.16 An effective mean temperature may be determined by
measuring the rate of disappearance of a compound of known
decomposition kinetics.
Experimental methods
Chemicals
All diketones and other compounds used were of analytical
grade quality and obtained commercially. These and SF6
IR LPHP has a number of well-documented advantages.
The first of these is that pyrolysis is initiated directly in the gas
T h i s j o u r n a l i s & T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 5
N e w J . C h e m . , 2 0 0 5 , 2 9 , 4 8 5 – 4 9 2
485

View Article Online
phase, thereby eliminating the complications frequently intro-
duced by competing surface reaction. The second is that the
primary products of pyrolysis are rapidly ejected into the
cooler regions of the cell, inhibiting their further reaction. In
favourable cases, these products may be accumulated for
further investigation. One disadvantage of IR LPHP is that
the temperature of the pyrolysis is neither well-defined nor
easily determined; comparison with more conventional meth-
ods of pyrolysis is thus difficult.
Under favourable circumstances, further mechanistic infor-
mation may be obtained by the judicious use of additives. In
particular, the formation of new products or a change in
product distribution on co-pyrolysis with a few torr of H2 or
(even more revealingly) D2 is strong indirect evidence for the
involvement of free radicals. This well-established method is
based on the rapid reaction of many radicals with H2: R + H2 -
RH + H.17–19 We have exploited this to demonstrate, for
example, the role played by thienyl radicals in the decomposi-
tion of thiophene.13
able from those of stable products by observing their disap-
pearance on careful annealing to a higher temperature
(dependent on stability).
In matrix isolation studies, the gaseous pyrolysis mixture
flowed from a vacuum line to the cold head through a ZnSe-
fronted reaction cell into which the laser radiation was direc-
ted. While moderately short-lived species reach the cold head,
extremely reactive species react further and are not observed.
Consequently, by employing a Pyrex tube of variable length
between the reaction cell and cold stage, it is possible to change
the flow time between generation and collection, and hence to
discriminate to some extent between short-lived intermediate
species based on lifetimes.
This IR LPHP matrix isolation apparatus could be modified
so that product gas flow along the path between the reaction
cell and cold stage was restricted by passing through a narrow
aperture. The use of an aperture to restrict gas flow elevated
the pressure of SF6 attainable in the IR LPHP cell. At a given
laser power, the effective temperature was therefore signifi-
cantly higher than that attained when the restriction was not
employed. This was reflected in the laser power required to
initiate decomposition; relatively stable compounds, which
decomposed when the gas flow was restricted, remained intact
even at the highest laser power available when the gas flow was
unrestricted. When used, this modification of the matrix isola-
tion assembly therefore allowed IR LPHP of compounds that
could not be decomposed without, but precluded the observa-
tion of very reactive species.
Analytical methods
The FTIR spectra presented in this work were recorded using a
Digilab FTS-60 Fourier transform spectrometer at a resolution
of 1 cmÀ1 and averaged over 32 scans per spectrum. Win-IR
software was used to acquire and process spectral data. The
sample compartment of the spectrometer, which contained the
pyrolysis cell, was purged with N2 to reduce/eliminate con-
tributions of CO2 and water from the ambient atmosphere. All
spectra were collected against a background spectrum of an
evacuated pyrolysis cell. FTIR spectra were used primarily to
characterise small gaseous products such as CO, methane or
ketene, where resolved vibration-rotation structure permits
unambiguous identification even from a single band; in such
cases, comparison with authentic samples was used. The
technique is less useful for larger molecules, in which similar
vibrations often have almost identical bands, and it also lacks
the sensitivity necessary for quantitative analysis.
GC-MS analysis of gas samples was conducted using a
Hewlett Packard 6890 series gas chromatograph interfaced
with a Hewlett Packard 5793 mass selective detector (MSD).
Base pressure of the MSD was maintained at 1 Â 10À5 torr by a
turbomolecular pump backed by a rotary pump. Gas samples
were extracted from the pyrolysis cell, via the septum port, with
a Hamilton 2.5 ml gas-tight syringe fitted with a lockable valve.
The contents of the cell were at relatively low pressure (12–15
torr); typically, therefore, 2.5 ml of the gas sample were
extracted and, after the syringe valve was engaged, the sample
was compressed to 0.25 ml to achieve a pressure close to
atmospheric. An ‘‘HP5-MS’’ cross-linked phenyl methyl silox-
ane gum capillary column (i.d.: 0.25 mm, length: 29.2 m, film
thickness: 0.25 mm) was employed as a multi-purpose column,
while a ‘‘GS GasPro’’ column (i.d.: 0.32 mm, length: 30 m) was
used for the detection and identification of low molecular
weight species. Helium was used as the carrier gas in this
GC-MS system.
Calculation methods
The feasibility of a proposed pyrolysis scheme could be eval-
uated by determining the relative activation energies involved.
For the most part, this involved calculating the activation
barrier to molecular elimination routes such as dehydration
or bond homolysis, the predominant decomposition routes,
and comparing these values with that of an alternative system,
whose actual pyrolysis temperature (i.e., the laser power
required to initiate decomposition) was known. It was found
that theory agreed well with experiment; a system calculated to
possess higher activation energy required a higher decomposi-
tion temperature. Of course, an initial step may be followed by
further processes if the latter also possess sufficiently low
activation energies. Observed end products may therefore
reflect a complex series of reactions, each of which must be
feasible under the prevailing conditions.
All theoretical calculations were executed using the PC
Spartan ‘02 molecular orbital program.20,21 PC Spartan ‘02
provides a full range of molecular mechanics and quantum
chemical methods, and it was found that the density functional
theory method at the B3LYP level with the 6-31G* basis set22
provided sufficiently accurate results. The one exception was
for systems containing strong hydrogen bonds (for example,
the enol forms of the diketones), for which it was found to be
necessary to include the diffuse functions of the 6-311+G**
basis set. Relative values of the activation energies of various
reaction pathways were considered to be of greater value in
predicting mechanism than absolute values. As a benchmark,
the activation energy for 1,2-elimination of H2O from ethanol
was calculated as 300 kJ molÀ1 with the 6-31G* basis set and
282 kJ molÀ1 with the 6-311+G** basis set, both very
comparable with the experimentally determined value of
Matrix isolation spectroscopy
Short-lived intermediates were detected using matrix isolation
FTIR spectroscopy. A two-stage closed-cycle helium refrigera-
tor was used for all matrix isolation experiments. The second
stage (heat station) of the cryostat, with which a copper sample
holder was in direct contact, was maintained at a temperature
of between 15 and 20 K for the duration of the experiment. All
matrices were collected at this temperature. The matrix isola-
tion assembly was constructed so that any reactive species,
generated through IR LPHP of an appropriate precursor,
would rapidly co-deposit with the matrix material on the cold
sample holder. Spectra of short-lived species were distinguish-
281 kJ molÀ1 23
.
Results
Acetylacetone (pentane-2,4-dione), acacH
Previous work. Like all b-diketones, acacH exists largely as
the internally enol form (1) at room temperatures, with the
486
N e w J . C h e m . , 2 0 0 5 , 2 9 , 4 8 5 – 4 9 2

View Article Online
fraction of the keto form (4) increasing as temperature is
raised. Thus, at room temperature, acacH is 81% enol, while
at 230 1C it is 60% enol.24 This equilibrium is likely to be
reflected in the decomposition chemistry.
The earliest investigation of acetylacetone pyrolysis was
conducted by Hurd and Tallyn in 1925 with the primary
objective of producing ketene.25 Liquid acetylacetone was
subjected to temperatures between 610 and 700 1C to obtain
varying yields of ketene; the optimum yield of 16.7% was
obtained at 635 1C. The accompanying product from the
decomposition of acetylacetone was acetone. Charles et al.26
subsequently pyrolysed liquid acetylacetone at temperatures
lower than those used by Hurd and Tallyn. The main gaseous
products found upon pyrolysis at 346 1C were acetic acid and
acetone, whilst at 462 1C substantial amounts of CO2, CO and
CH4 were also detected. The latter products resulted from
pyrolysis of acetic acid or acetone rather than from acetyl-
acetone directly.
Fig. 2 GC trace of acacH + SF6 following IR LPHP. 1 = SF6, 2 =
C3H4 (propadiene/propyne), 3 = buta-1,3-diene, 4 = buta-1,2-diene,
5 = but-2-yne, 6 = acetone, 7 = 2-methylfuran (5), 8 = acacH.
Both these examples of acetylacetone pyrolysis were con-
ducted using conventional ‘‘hot-walled’’ techniques (where
surface reactions are possible). To date, only two homogeneous
pyrolysis studies of acetylacetone have been reported. Choudh-
ury and Lin reported a high temperature (1120–1660 K) study
using shock waves.27 In that work, acetylacetone was observed
to undergo pyrolysis via two concurrent pathways: (1) elimina-
using the shortest possible transit time between pyrolysis and
matrix collection; co-pyrolysis with D2 resulted in the produc-
tion of a small amount of deuterated methane, but no other
changes. The direct synthesis of 5 from acacH has not been
previously reported; it is very likely that its formation in this
work results from an H2O elimination pathway similar to that
proposed by Choudhury and Lin,27 as discussed below.
The results described above suggest that, under IR LPHP
conditions, acacH undergoes decomposition via two concur-
rent pathways, one giving acetone and ketene, the other 2-
methylfuran (5). As will be shown in the following discussion,
the tautomeric nature of the b-diketone significantly influences
the route of product formation. Evidently, the different struc-
tural configurations of the keto and enol tautomers yield quite
distinct decomposition products.
tion of H2O to yield CH3CRCCOCH3 and/or CH
C
Q Q
2
CHCOCH3 and (2) a-b dissociation to produce CH3COd and
CH3COCH2 radicals.28 More recently, Al-awadi and co-
d
workers have investigated the kinetics of the elimination reac-
tion in a number of pentane-2,4-dione derivatives, including
acacH, and have concluded that a concerted 6-membered
elimination process is responsible for the production of acetone
and ketene, rather than the radical pathway.29,30
Formation of acetone and ketene. Although acacH exists
predominantly in the enol form (1), the keto (4) content does
increase on raising the temperature, reaching ca. 40% at
230 1C.24 Hence, the temperatures attained under IR LPHP
conditions afford a sufficient amount of the acacH keto
tautomer 4 to yield detectable pyrolysis products. Of the acacH
tautomers, the keto form 4 bears the molecular structure most
conducive to acetone formation. The decomposition of acacH
to acetone and ketene was originally envisaged to proceed by a
free radical mechanism, with a-b fission [reaction (1.1)] as the
initiating step.27 It is also conceivable that decomposition is
initiated via b-g bond fission [reaction (2.1)]; in either case,
reaction is then envisaged to proceed by analogues of the
Rice–Herzfeld mechanism for acetone,31,32 as given in reac-
tions (1.1)–(2.3):
Experimental results. In our work, IR LPHP of acetylace-
tone (acacH) was effected at a laser power of approximately
5 W, resulting in approximately 10% decomposition over a
period of 60 s. From the kinetic parameters reported by Al-
Awadi et al.,29 we may estimate an effective temperature of
700–750 K, considerably lower than that employed in other
studies. The FTIR and GC results presented in Figs. 1 and 2,
respectively, show that the major products arising from pyro-
lysis are 2-methylfuran (5), acetone and ketene (all in very
similar amounts), the proportion of the first of these increasing
slightly with laser power. The minor products observed were
propadiene/propyne (C3H4 isomers), buta-1,3-diene, buta-1,2-
diene, but-2-yne, methane and carbon monoxide. The observed
production of acetone and ketene is in agreement with the
study conducted by Hurd and Tallyn.25 No short-lived pro-
ducts were detectable using matrix isolation spectroscopy, even
d
d
CH3COCH2COCH3 - CH3COCH2 + COCH3 (1.1)
d
d
CH3COCH2 - CH + CH
C
O
(1.2)
(1.3)
Q Q
3
2
d
dCH3 + COCH3 - CH3COCH3
d
CH3COCH2COCH3 - CH3COCH2COd + CH3 (2.1)
d
CH COCH COd - CH CO + CH
C
O
(2.2a)
(2.2b)
(2.3)
Q Q
3
2
3
2
CH3COCH2COd - CH3 + COCH2COd
d
d
dCOCH3 - CH3 + CO
d
To establish whether radical processes do indeed occur, co-
pyrolysis studies of acacH with H2 and D2 (pressure: 2 torr)
were conducted. Under these conditions, the radicals produced
in the initial step [reaction (1.1)] above will be trapped with H2
to produce acetone and acetaldehyde. Similarly, those of
[reaction (2.1)] would yield methane and 3-oxobutanal (aceto-
acetaldehyde); the latter is known to be unstable, however, and
self-condenses to form 1,3,5-triacetylbenzene.33 In the presence
of D2, deuterated analogues of these products would be
Fig. 1 Infrared spectra of acacH + SF6 prior to (top) and following
(bottom) IR LPHP. 1 = acacH, 2 = methane, 3 = ketene, 4 = CO,
5 = acetone, 6 = 2-methylfuran (5), unmarked = SF6.
N e w J . C h e m . , 2 0 0 5 , 2 9 , 4 8 5 – 4 9 2
487

View Article Online
formed. The resultant FTIR and GC-MS analyses showed no
indication of the formation of acetaldehyde or 3-oxobutanal,
although acetone was produced. No evidence of deuterated
acetone or acetaldehyde was found upon co-pyrolysis of acacH
with D2. While some deuteration of methane was detected, this
is inconclusive, as it could well arise from the subsequent
decomposition of acetone, a well-known radical process.31,32
An alternative six-centred unimolecular retroene mechan-
ism34 was proposed by Blades and Sandhu to account for the
production of acetone from pyrolysis of acacH.35 The activa-
tion energy (Ea) obtained from calculated Arrhenius A factors
and experimentally determined rate constants was found to be
214 kJ molÀ1. This was revised downwards to 198.3 kJ molÀ1
in the subsequent studies of Al-Awadi et al.29
In order to investigate these reactions further, we have
performed calculations using PC Spartan ’02 at the DFT/
B3LYP/6-31G* level. The bond dissociation energies corres-
ponding to reaction (1.1) and (2.1) were found to be 380 and
377 kJ molÀ1, respectively, whereas the activation energy via
the retroene 6-centre transition state shown in Fig. 3 (left) was
found to be 230 kJ molÀ1; no 4-centre transition state leading
to ketene and acetone could be located. On the basis of the
experimental evidence that free radicals are not involved, the
much lower calculated activation energies for the molecular
elimination route, and the good agreement with observed
activation energies, we conclude that the formation of acetone
and ketene is an intramolecular, rather than a free radical
chain, process.
mechanisms were offered to account for the formation of the
rearrangement product, both involving a vinylidene intermedi-
ate and a 1,2-hydrogen shift and differing only in the sequence
of these two processes; it was not clear which of these mechan-
isms is more likely. The significant aromatic stabilisation of
furan may provide a sufficiently strong driving force to com-
pete successfully with C–H insertion reactions in the carbenes.
The work of Huntsman and Yin,40 together with the pyr-
olysis study by Choudhury and Lin,27 strongly suggest that IR
LPHP of the enol form of acacH occurs initially with H2O
elimination to yield 6 as an intermediate, which readily re-
arranges to form 5. Matrix isolation studies, which were
performed in an effort to identify any intermediate or short-
lived species, revealed no evidence of the pentadienone or any
species other than unreacted acacH and ketene. There is no
previous documentation of the synthesis of 5 from the direct
decomposition of acacH and it therefore appears that a novel
route to the production of 5 from IR LPHP of acacH has been
established. Calculations at the DFT/B3LYP/6-31G* level
showed that a simultaneous cyclisation and 1,2-hydrogen shift
requires only 202 kJ molÀ1, confirming the very ready nature of
this step under the prevailing conditions.
The minor products observed upon IR LPHP of acacH can
be accounted for by the unimolecular decomposition of 5,
which produces a number of products including propadiene/
propyne, buta-1,3-diene, buta-1,2-diene and but-2-yne. A re-
flected shockwave study by Liftshitz et al.41 showed that the
homogeneous unimolecular decomposition of 5 proceeds via
1,2-H atom migration from the C(5) to C(4) position. Stabili-
sation of the resultant intermediate is achieved by further 1,2-H
atom migration and rearrangement to form CO, buta-1,3-
diene, buta-1,2-diene and but-2-yne. Additionally, 5 may also
undergo C–C bond cleavage after the initial 1,2-H atom
migration process. This decomposition route yields C3H4
(propadiene/propyne) and ketene.
Formation of 2-methylfuran (5). Elimination of H2O from
acacH, as observed by Choudhury and Lin,27 was suggested to
occur via a four-centred transition state to produce pent-3-yn-
2-one or penta-1,2-dien-4-one (acetylallene, 6); the latter is
illustrated in Fig. 3 (right). This elimination is readily facili-
tated when acacH adopts the enol form; its activation energy
calculated at the DFT/B3LYP/6-311+G** level is 310
kJ molÀ1. The synthesis of acetylenic ketones from b-diketones
is generally achieved by dehydrating with a,a-difluoroalkyl-
amines.36 6 may be prepared from acacH by treating with
triphenylphosphine dibromide to produce E/Z-2-bromo-4-
oxo-2-pentene; dehydrobromination affords 6.37,38
All observed products are thus accounted for and the
corresponding routes are illustrated in Scheme 1.
1,1,1,5,5,5-Hexafluoropentane-2,4-dione
(hexafluoroacetylacetone), hfacH (2)
No direct evidence of either of these dehydration products
was observed using FTIR or GC-MS. However, their involve-
ment in the decomposition scheme of the enol form of acacH
cannot be discounted, as both are known to rearrange readily.
The pyrolytic rearrangement of pent-3-yn-2-ones (acetylenic
ketones) has been shown to be an effective route towards the
synthesis of cyclopentenones.39 If pent-3-yn-2-one, the acetyl-
enic ketone formed from dehydration of acacH, were to under-
go such a reaction, cyclisation of the acetyl(methyl)vinyl-
idene intermediate would produce 2-methylcyclobut-2-en-1-
one. There is no evidence in the literature for preparation or
isolation of the latter and no indication of its formation in this
work.
Previous work. There has been little work reported on the
thermal decomposition of 2. There is, however, a substantial
On the other hand, the pyrolysis of allenyl ketones, such as 6
and hexa-1,2-dien-4-one, has been shown to lead to a rearran-
gement analogous to that of acetylenic ketones to produce
furan derivatives. Huntsman and Yin40 showed that synthesis
of 5 was possible from the pyrolysis of 6. Two possible
Fig. 3 Six-centred (left) and four-centred (right) transition states for
the decomposition of acacH, leading to the formation of acetone +
ketene and H2O + penta-1,2-dien-4-one (6), respectively.
Scheme 1 Proposed reaction scheme for IR LPHP of acacH. Ob-
served products in bold italics; numbers given are calculated activation
energies in kJ molÀ1
.
488
N e w J . C h e m . , 2 0 0 5 , 2 9 , 4 8 5 – 4 9 2

View Article Online
body of work on CVD related investigations of hfacH over
metal and metal oxide surfaces,42–48 while the photolysis study
by Bassett and Whittle49 is the only account of the homo-
geneous decomposition of 2. Photolysis of the predominant
enol form of 2 occurred with a novel HF elimination to yield
2,2-difluoro-5-(trifluoromethyl)furan-3(2H)-one (DHTF, 7).
Subsequent photolysis of this product produced CF2 carbenes,
which then inserted into the enolic OH bond of another hfacH
molecule to produce 4-difluoromethoxy-1,1,1,5,5,5-hexa-
fluoro-3-penten-2-one (DMHP, 8). As part of a study on
LaF3 film formation from La(hfac)3 Á diglyme, Condorelli
et al.50 investigated the thermal decomposition of 2 utilising
a conventional hot-wall MOCVD system with a quartz reactor
body; the heterogeneous formation of 1,1,1-trifluoroacetone
and 1,1,1-trifluroacetaldehyde was observed.
Fig. 5 GC trace of 2 + SF6 following IR LPHP. 1 = SF6, 2 =
trifluoroacetone, 3 = DFAF (9), 4 = carbon suboxide, 5 = hfacH (2),
6 = DHTF (7), 7 = DMHP (8).
Experimental results. Appreciable decomposition of 2 was
effected upon IR LPHP at a laser power of approximately 6 W,
corresponding to temperatures between 750 and 800 K. Fig. 4
presents FTIR spectra collected prior to and following pyro-
lysis; peaks attributable to carbon suboxide (C3O2), CF2O,
Co-pyrolysis with H2 or D2 (2 torr) did not result in
incorporation of D into any products or the generation of
new products, suggesting that free radicals are not involved.
Matrix isolation studies did reveal a number of bands of
species in the 2100 to 2300 cmÀ1 region (characteristic of
cumulative double bonds or triple bonds), as shown in Fig.
7. Although the spectra are complex and overlapped, and
assignment is far from certain, a band identified as arising
from a transient species at 2194 cmÀ1 is consistent with the
gas-phase harmonic value of 2252 cmÀ1 calculated for trifluor-
omethylketene, adjusted by the usual corrections for anharmo-
nicity and matrix effects.54
CO, CF CF , CF H and SiF were identified. SiF4 arises
2
Q
2
3
4
from the reaction of HF with the Pyrex cell walls and it is
therefore difficult to quantify fully the products in this system.
GC-MS analysis of heavier fragments (Fig. 5) detected
unreacted 2, trifluoroacetone (trace), C3O2 (16%), and three
products with retention times of 1.43 (28%), 1.61 (32%) and
1.65 (23%) min. Analysis with the GasPro column aided in the
detection of the lower molecular weight compounds CF
Q
2
CF2, CF3H and CF3CCH. The mass spectrum of the product
eluted at 1.61 min was identical with the mass spectra obtained
from the decomposition of a range of oxovanadium(IV) hfac
complexes, UO2(hfac)2, and from UV photolysis of 2.49,51,52
This compound was identified as 7. The mass spectrum of the
compound eluted at 1.65 min was attributed, upon comparison
with mass spectral data obtained by Bassett and Whittle, to 8,
that is, the product resulting from CF2 carbene insertion into
the O–H bond of a hfacH molecule following UV photolysis.49
The product eluted at 1.43 min has not been previously
observed. Its mass spectrum (Fig. 6) exhibits a parent mass
fragment at m/z = 110. No product possessing such a mass
spectrum has been reported from previous decomposition
studies of 2 or complexes containing 2. FTIR spectra provided
no indication of the identity of this product. However, a very
likely candidate on the basis of the mass spectra of similar
Formation of 2,2-difluoro-5-(trifluoromethyl)furan-3(2H)-one
(DHTF, 7) and 4-difluoromethoxy-1,1,1,5,5,5-hexafluoro-3-pen-
ten-2-one (DMHP, 8). The results obtained indicate that the
primary decomposition products of 2 are C3O2, 7, 9 and 8 (and,
by implication, HF); the minor products COF2, C2F4, CF3H,
trifluoromethylacetylene and trifluoroacetone arise, as dis-
cussed below, from subsequent decomposition of the primary
products. The most likely route identified by calculation was
indeed the HF molecular elimination route, with an activation
energy calculated at 196 kJ molÀ1, thus accounting for the
formation of 7 and SiF4. 8 arises from insertion of a CF2
carbene into the O–H bond of the enol form of 2; the most
likely origin of the CF2 is DHTF itself, as observed by Bassett
and Whittle in their photolysis study.49 CF2 may also dimerise,
of course, yielding the observed tetrafluoroethene. All these
products and formation routes are illustrated in Scheme 2.
compounds is 3,3-difluoroacryloylfluoride, CF CH–CFO
Q
2
(DFAF, 9). For example, the mass spectrum of benzoyl
fluoride, C6H5COF, exhibits strong peaks at m/z = 124 (M),
105 (M À 19), 86 (M À 28), and 77 (M À 47), a pattern also
evident in Fig. 6.53
Formation of 3,3-difluoroacryloylfluoride and carbon subox-
ide. No dehydration routes analogous to acacH are possible.
Instead, the most likely secondary (higher temperature) path-
way involves semi-concerted loss of CF3 radicals, followed by
Fig. 4 Infrared spectra of 2 + SF6 prior to (top) and following
(bottom) IR LPHP. 1 = hfacH (2), 2 = carbon suboxide, 3 = CO,
4 = carbonyl fluoride, 5 = trifluoromethane, 6 = tetrafluoroethene,
7 = SiF4, unmarked = SF6.
Fig. 6 Mass spectrum of the product of IR LPHP of 2 eluted with a
retention time of 1.43 min, ascribed to DFAF (9).
N e w J . C h e m . , 2 0 0 5 , 2 9 , 4 8 5 – 4 9 2
489

View Article Online
carbonylketenes and similar compounds.59 Thus, each step
in this sequence is quite feasible under the prevailing condi-
tions; the matrix isolation detection of a vibration character-
istic of ketenes at 2194 cmÀ1 lends additional weight to this
hypothesis.
2,2,6,6-Tetramethylheptane-3,5-dione, thdH (3)
Previous work. The thermal decomposition of 3 under
MOCVD conditions has been studied by a number of workers,
largely to complement decomposition studies of the corre-
sponding metal complexes. Condorelli et al.60 investigated
the thermal decomposition of 3 under MOCVD conditions
as part of a study of the decomposition of Cu(thd)2. Their
observations suggested the presence of two different reaction
paths depending on temperature. Decomposition was observed
to commence at temperatures above 400 1C with cleavage of
the aC–bCO bond to afford 3,3-dimethylbutan-2-one (13). The
remaining C5H8O moiety was proposed to rearrange to form a
cyclic ether (3-methyl-2,5-dihydrofuran), the rationale being
that the observed n(CO) at 1075 cmÀ1 was more consistent with
that of furan rather than an aliphatic ether. In the high
temperature route (500–600 1C), Condorelli et al. suggested
that the initial stage of thdH decomposition involved dehydra-
tion to form (CH3)3CCOCHC5H8, which subsequently decom-
posed to yield 3-methylpenta-1,4-diene, methylketene and
ethene.
The elimination of H2O as the initial step of decomposition
was also proposed by Bykov et al.61 in their mass spectrometric
study of the thermolysis of 3. Decomposition occurred at
approximately 590 1C and, analogous to the work of Condor-
elli et al.,60 dehydration resulted in the production of a
(CH3)3CCOCHC5H8 moiety, which eliminated methyl and
ethyl radicals with increased temperature. Their observations
also indicated that the thermal stability of 3 was lowered
significantly in the presence of oxygen.
Fig. 7 Partial FTIR spectrum of matrix isolated products of IR
LPHP of 2 + SF6. Peaks 1, 3 and 4 are not affected by mild annealing
and are assigned to carbon suboxide, trifluoromethylacetylene and CO,
respectively; peak 2 disappears on annealing and is assigned to an
unstable ketene (see text).
abstraction of the labile central H from the parent hfacH to
give the observed CF3H. This pathway has an activation
energy of approximately 290 to 300 kJ molÀ1, based on our
calculations and the reported CF3–C bond strength in
CF3COCH3.55 Loss of one CF3H results in trifluoroacetyl-
ketene (10) and loss of a second CF3H yields the observed
relatively stable C3O2. 10 is unstable and has not been isolated
to date; it is therefore likely to undergo further reaction in the
IR LPHP system. By analogy with well-known pyrolytic path-
ways in ketenes, a possible sequence of events is: (i) decarbo-
nylation—an activation energy in the typical range of 200–240
kJ molÀ1 would allow this reaction to proceed very rapidly,
resulting in the carbene 11;56 (ii) the familiar Wolff rearrange-
ment (activation energies typically 10–50 kJ molÀ1) of the
resultant trifluoroacetylcarbene 11 to yield trifluoromethylk-
etene (12);57,58 (iii) a 1,3-F shift resulting in 9. This last step is
somewhat unexpected at first sight, since 1,3-shifts are gener-
ally assumed to be thermally forbidden, but our calculations at
the DFT/B3LYP/6-31G* level predict an activation energy of
just 180 kJ molÀ1 for this step. This migratory ease has also
been demonstrated in a study of 1,3-sigmatropic shifts in
The lower temperature decomposition route observed by
Condorelli et al. is also supported by similar results by other
workers. A GC-MS study by Lopaeva et al.62 found that the
gas phase decomposition of 3 at 400 1C produced 13, isobutene
(2-methylpropene, 14) and carbon monoxide. The formation of
3,3-dimethyl-2-butanone along with 2,2-dimethylpropanal
and 14 was also observed upon decomposition of 3 at 484 1C
by Obi-Johnson.63
In summary, literature reports suggest that the decomposi-
tion of 3 proceeds by two different pathways influenced
by temperature. Dehydration occurs at higher temperature
(600 1C) while at lower temperature (400–484 1C) aC–bCO
bond scission is effected.
Experimental results. Appreciable decomposition of 3 was
effected upon pyrolysis at a laser power of approximately 6 W
(700–750 K). The sole significant decomposition product of 3
detectable using GC-MS was isobutene (14); other products
observed in trace amounts included 2-methylbuta-1,3-diene
(isoprene), 2-methylbut-1-ene and but-1-en-3-yne. Use of the
GasPro column aided in the detection of 14, propene, 2-
methylpropane (isobutane) and trace amounts of propadiene,
1-butyne, methane and propyne. Because of the low vapour
pressure of 3, pressures were limited to 0.25 torr. Although
sufficient to permit IR LPHP, such a low reagent pressure
produces weak infrared spectra, which can hinder the monitor-
ing of reaction progress and the identification of products.
However, the FTIR spectrum obtained following pyrolysis of 3
did reveal peaks attributable to unreacted 3, ketene and CO, as
shown in Fig. 8. No bands due to 14 or isobutane were
detectable by FTIR.
Scheme 2 Proposed reaction scheme for IR LPHP of 2. Observed
products in bold italics; numbers given are calculated activation
These results at first sight suggest that the decomposition of
3 occurs via a mechanism that involves the production of
energies in kJ molÀ1
.
490
N e w J . C h e m . , 2 0 0 5 , 2 9 , 4 8 5 – 4 9 2

View Article Online
Fig. 8 Partial FTIR spectrum of products of IR LPHP of 3, confirm-
ing the formation of ketene and CO (compare insert of Fig. 1).
t-butyl radicals; disproportionation of t-butyl radicals affords
isobutane and 14. A second possible reaction route involves
homolysis to produce other radicals and subsequent loss of
carbon monoxide and ketene. Supplementary pyrolysis studies
were therefore conducted to investigate the involvement of
t-butyl radicals and provide further insight into the decom-
position of 3.
Scheme 3 Proposed reaction scheme for IR LPHP of 3. Observed
products in bold italics; numbers given are calculated activation
energies in kJ molÀ1. Also shown is the formation of products on
trapping initial radicals with H2.
Co-pyrolysis studies of 3 with H2 were undertaken in an
effort to confirm the formation and involvement of t-butyl
radicals. If t-butyl radicals are indeed produced, then reaction
with H2 should compete with disproportionation to facilitate
formation of isobutane. Hence, an increase in isobutane pro-
duction would be observed. In practice, the co-pyrolysis of 3
with H2 did not significantly affect the product distribution of
14 and isobutane. However, two additional products not
previously observed in the pyrolysis of 3 alone were detected
in significant amounts by GC-MS. These two compounds were
eluted with retention times of 3.12 (trace) and 4.13 min, and
were identified as 3-methylbutan-2-one (15) and 13, respec-
tively. The formation of these products would require that the
first step involve homolysis of the bond between the a-carbon
and b-carbonyl. Subsequent reaction of the resulting radical
species with H2 would then yield 3,3-dimethylbutan-2-one and
2,2-dimethylpropanal (pivalaldehyde, 16). No evidence of 2,2-
dimethylpropanal was found, but the ready isomerisation of
this compound does produce the observed 15.64 In the presence
of D2, there was clear evidence of incorporation of a single D
atom into these products. On the other hand, matrix isolation
studies did reveal features consistent with t-butyl radicals when
the shortest transfer time was used. Partially overlapped peaks
observed at 2942 (2936), 2825 (2825) and 1456 (1455) cmÀ1
disappeared on annealing (reported values in parentheses, the
reported peak at 1366 cmÀ1 being completely obscured by the
parent thdH65,66), and radicals are evidently formed at some
stage during the pyrolysis.
was conducted under MOCVD conditions, so it is likely that
heterogeneous reactions with hydrogen-containing species on
the reactor surface lead to the formation of these products. The
absence of 2,2-dimethylpropanal (or its isomer, 3-methyl-2-
butanone) in the work of Lopaeva et al.62 led these authors to
suggest a unimolecular decomposition route involving intra-
molecular hydrogen atom transfer to yield 3,3-dimethylbutan-
2-one, CO and 14. Although the reaction conditions utilised by
both workers did not differ significantly, the difference in the
GC-MS columns used in the studies by Obi-Johnson and
Lopaeva et al. may account for the disparity in the products
observed.
The formation of mono-deuterated 3,3-dimethylbutan-2-one
upon co-pyrolysis of 3 with D2 in this work provided evidence
of aC–bCO bond scission and thus precluded the involvement
of an intramolecular mechanism. The corresponding aldehyde
was not observed, but as mentioned previously, 2,2-dimethyl-
propanal is capable of isomerising to 15. Traces of the latter
were detected, in its mono-deuterated form, on co-pyrolysis
with D2.
The initially proposed route to thdH decomposition invol-
ving homolysis to produce t-butyl radicals as the first step must
therefore be reconsidered in light of the co-pyrolysis results.
The IR LPHP of 3 instead appears to be consistent with a
mechanism involving firstly homolysis of the aC–bCO bond to
produce (CH ) CC( O)d and (CH ) CC(OH) CHd radicals.
Q
Q
3 3
3 3
As observed from co-pyrolysis studies, the addition of hydro-
gen to these radicals produces 3,3-dimethyl-2-butanone and
3-methyl-2-butanone, respectively. In the absence of hydrogen,
Discussion. a-Carbon–b-carbonyl bond homolysis of the
enol form will produce the radicals shown in Scheme 3. Each
radical then loses either CO or ketene to yield t-butyl radicals;
these processes are likely to proceed very rapidly, their calcu-
lated activation energies being 110 and 190 kJ molÀ1, respec-
tively. On the other hand, trapping with hydrogen will produce
3-dimethyl-1-buten-2-ol, which is essentially the enol tautomer
of 3,3-dimethylbutan-2-one, and 16, a well-known precursor to
15.64 3,3-Dimethylbutan-2-one is significantly more stable than
it is proposed that the radicals (CH ) CC( O)d and
Q
3 3
(CH ) CC(OH) CHd undergo very ready further homolysis
Q
3 3
to produce t-butyl radicals and CO or ketene, as shown in
Scheme 3. A slight preference for this a,b homolysis is con-
firmed by calculation of the bond energies. Loss of a t-butyl
radical requires 328 kJ molÀ1, whereas loss of (CH3)3C–CO
requires 310 kJ molÀ1, a difference leading to a ratio in the
rates of approximately 40 at the pyrolysis temperature. No
competitive molecular routes were revealed by calculation, in
accord with the radical routes revealed by the co-pyrolysis
studies. All observed major products are accounted for in
Scheme 3; trace products may be ascribed to partial further
decomposition of the principal products.
67
its enol tautomer; its enol-keto equilibrium constant pKE of
68
8.33 is comparable to that of acetone (8.76).69
Obi-Johnson63 found that
3 decomposed to produce
3,3-dimethyl-2-butanone and 2,2-dimethylpropanal, with no
addition of a hydrogen source. This particular study, however,
N e w J . C h e m . , 2 0 0 5 , 2 9 , 4 8 5 – 4 9 2
491

View Article Online
27 T. K. Choudhury and M. C. Lin, Int. J. Chem. Kinet., 1990, 20,
491.
Conclusions
28 For consistency among the three systems, we have adopted the
common convention of lettering the central C atom in the –CO–
CH2–CO- unit as a, so that the carbonyl C is b, etc.
29 N. A. Al-Awadi, M. H. El-Nagdi and T. Mathew, Int. J. Chem.
Kinet., 1995, 27, 517.
30 N. A. Al-Awadi, M. H. El-Nagdi, H. A. Al-Awadhi and O. M. E.
El-Dusouqui, Int. J. Chem. Kinet., 1998, 30, 457.
31 F. O. Rice and K. F. Herzfeld, J. Am. Chem. Soc., 1934, 56,
284.
32 S. H. Mousavipour and P. D. Pacey, J. Phys. Chem., 1996, 100,
3573.
33 R. Kaushal, S. Sovani and S. S. Deshapande, J. Indian Chem.
Soc., 1942, 19, 107.
34 D. V. Banthorpe, Elimination Reactions, Elsevier, Amsterdam,
1963.
35 A. T. Blades and H. S. Sandhu, Int. J. Chem. Kinet., 1971, 3, 187.
36 T. Kitazume and N. Ishikawa, Chem. Lett., 1980, 1327.
37 T. Constantieux and G. Buono, Org. Synth., 2000, 78, 135.
38 G. Buono, Synthesis, 1981, 11, 872.
The three b-diketones investigated here show very different
decomposition pathways. Where feasible, low-energy molecu-
lar elimination mechanisms, such as the production of ketene
and acetone or dehydration in acacH, or the loss of HF from 2,
dominate; of course, the primary products of such pathways
may themselves be subject to further reaction and therefore not
directly observable. In other cases, where no low-energy mo-
lecular route is available, radical pathways are significant. In
each case, calculations of the proposed pathways support the
experimental observations and deductions.
Of course, the proposed pathways will have a profound
effect on the chemistry of decomposition of metal complexes.
For example, the production of HF will almost certainly lead
to the formation of metal fluorides, as has been observed using
many hfac complexes.64 These factors will be explored in our
future work.15
39 M. Karpf and A. S. Dreiding, Helv. Chim. Acta, 1979, 62, 852.
40 W. D. Huntsman and T. K. Yin, J. Org. Chem., 1983, 48, 3813.
41 A. Lifshitz, C. Tamburu and R. Shashua, J. Phys. Chem. A, 1997,
101, 1018.
42 S. R. Droes, T. T. Kodas and M. J. Hampden-Smith, Adv. Mater.,
1998, 10, 1129.
43 M. A. George, D. W. Hess, S. E. Beck, J. C. Ivankovits, D. A.
Bohling and A. P. Lane, J. Electrochem. Soc., 1995, 142, 961.
44 S. W. Kang, H. U. Kim and S. W. Rhee, J. Vac. Sci. Technol., B,
1999, 17, 154.
45 H. L. Nigg and R. I. Masel, Surf. Sci., 1998, 409, 428.
46 H. L. Nigg and R. I. Masel, J. Vac. Sci. Technol., A, 1999, 17,
3477.
47 J. E. Parmeter, J. Phys. Chem., 1993, 97, 11530.
48 F. Rousseau, A. Jain, T. T. Kodas, M. Hampden-Smith, J. D.
Farr and R. Muenchausen, J. Mater. Chem., 1992, 2, 893.
49 J. E. Bassett and E. Whittle, Int. J. Chem. Kinet., 1976, 8, 859.
50 G. G. Condorelli, S. Gennaro and L. Fragala, Chem. Vap.
Deposition, 2000, 6, 185.
51 D. A. Johnson and A. B. Waugh, Polyhedron, 1983, 2, 1323.
52 A. Ekstrom, H. Hurst and C. H. Randall, J. Phys. Chem., 1982,
86, 2375.
Acknowledgements
We thank the University of Auckland for assistance with the
purchase of equipment, and the Department of Chemistry for a
doctoral scholarship to AY.
References
1
R. C. Mehrotra, R. Bohra and D. P. Gaur, Metal b-Diketonates
and Allied Derivatives, Academic Press, London and New York,
1978.
2
(a) G. B. Stringfellow, Organometallic Vapor Phase Epitaxy:
Theory and Practice, Academic Press, Boston, 1989; (b) The
Chemistry of Metal CVD, eds. T. T. Kodas and M. J. Hamp-
den-Smith, Weinheim, New York, 1994.
3
4
5
J. M. Shaub and S. H. Bauer, Int. J. Chem. Kinet., 1975, 7, 509.
J. Pola, Collect. Czech. Chem. Commun., 1981, 46, 2856.
G. A. Atiya, A. S. Grady, S. A. Jackson, N. Parker and D. K.
Russell, J. Organomet. Chem., 1989, 260, 180.
53 ‘‘Mass Spectra’’, NIST Chemistry WebBook, NIST Standard
Reference Database no 69, eds. P. J. Linstrom and W. G. Mallard,
NIST Mass Spec. Data Center (Director: S. E. Stein), Gaithers-
burg, MD, March 2003.
6
7
D. K. Russell, Chem. Soc. Rev., 1990, 19, 407.
D. K. Russell, I. M. T. Davidson, A. M. Ellis, G. P. Mills,
M. Pennington, I. M. Povey, J. B. Raynor, S. Saydam and A.
D. Workman, Organometallics, 1995, 14, 3717.
54 W. Forner, H. M. Badawi, A. A. Al-Saadi and S. A. Ali, J. Mol.
¨
Model., 2001, 7, 343.
8
9
D. K. Russell, Adv. Mater. CVD, 1997, 2, 223.
R. A. Berrigan, J. B. Metson and D. K. Russell, Adv. Mater. CVD,
1998, 4, 23.
55 Handbook of Chemistry and Physics, ed. R. C. Weast, 60th edn.,
CRC Press, Boca Raton, FL, 1980.
56 K. M. Forsythe, S. K. Gray, S. J. Klippenstein and G. E. Hall,
J. Chem. Phys., 2001, 115, 2134.
57 L. Wolff, J. Liebigs Ann. Chem., 1902, 325, 129.
58 W. Kirmse, Eur. J. Org. Chem., 2002, 2193.
59 M. T. Nguyen, L. Landuyt and H. M. T. Nguyen, Eur. J. Org.
Chem., 1999, 401.
10 J. E. Everett, N. D. Renner and D. K. Russell, Chem. Commun.,
1998, 341.
11 I. M. T. Davidson, A. M. Ellis, G. P. Mills, M. Pennington, I. M.
Povey, J. B. Raynor, D. K. Russell, S. Saydam and A. D. Work-
man, Adv. Mater. CVD, 1998, 4, 103.
12 N. R. Hore and D. K. Russell, J. Chem. Soc., Perkin Trans. 2,
1998, 269.
60 G. G. Condorelli, M. L. Hitchman, A. Y. Kovalgin and S. H.
Shamlian, Proc. Electrochem. Soc., 1999, 98-23, 141.
61 A. F. Bykov, P. P. Semyannikov and I. K. Igumenov, J. Therm.
Anal., 1992, 38, 1463.
62 O. V. Lopaeva, M. L. Hitchman, S. H. Shamlian and D. R.
Watson, J. Phys. IV: Proceedings 12 (13th Eur. Conf. on CVD,
2001), 2002, 12, Pr4/85.
63 B. J. Obi-Johnson, PhD Thesis, Georgia Institute of Technology,
Atlanta, GA, 1999 (Chem. Abs., 2000, 133, 114055; Diss. Abs. Int.,
2000, 60, 3255).
13 N. R. Hore and D. K. Russell, New J. Chem., 2004, 28, 606.
14 H. Hettema, N. R. Hore, N. D. Renner and D. K. Russell,
Aust. J. Chem., 1998, 50, 363.
15 D. K. Russell and A. Yee, in preparation.
16 J. L. Lyman, J. Chem. Phys., 1977, 67, 1868.
17 S. Davison and M. Barton, J. Am. Chem. Soc., 1952, 74, 2307.
18 G. H. Miller and E. W. R. Steacie, J. Am. Chem. Soc., 1958, 80,
6486.
19 R. R. Baldwin, R. W. Walker and D. H. Langford, Trans. Faraday
Soc., 1969, 65, 2116.
64 S. M. Fine, P. N. Dyer, J. A. T. Norman, B. A. Murator and R. L.
Iampietro, Mater. Res. Soc. Symp. Proc., 1990, 204, 415 (Chem.
Abs., 1990, 115, 244964).
20 PC Spartan ‘02, Wavefunction, Inc.,, Irvine, CA, 2001.
21 Spartan ‘02 Tutorial and User’s Guide, Wavefunction, Inc., Irvine,
CA, 2001.
65 J. Pacansky and J. S. Chang, J. Chem. Phys., 1981, 74, 5539.
66 B. Schrader, J. Pacansky and U. Pfeiffer, J. Chem. Phys., 1984, 88,
4069.
22 J. J. P. Stewart, J. Comput. Aided Mol. Des., 1990, 4, 1.
23 N. I. Butkovskaya, Y. Zhao and D. W. Setser, J. Phys. Chem. A,
1994, 98, 10779.
67 In aqueous solution at 25 1C.
68 J. R. Keefe, A. J. Kresge and N. P. Schnepp, J. Am. Chem. Soc.,
24 E. Funcke and R. Mecke, in Hydrogen Bonding, ed. D. Hadzi,
Pergamon Press, London, 1959, p. 433.
1990, 112, 4862.
69 Y. Chiang, A. J. Kresge and N. P. Schnepp, J. Am. Chem. Soc.,
1989, 111, 9677.
25 C. D. Hurd and W. H. Tallyn, J. Am. Chem. Soc., 1925, 47, 1779.
26 R. G. Charles, W. M. Hickam and J. von Hoene, J. Phys. Chem.,
1959, 63, 2084.
492
N e w J . C h e m . , 2 0 0 5 , 2 9 , 4 8 5 – 4 9 2