A study of the vacuum pyrolysis of para-substituted diazoacetophenones with He(I) ultraviolet photoelectron spectroscopy

Article abstract of DOI:10.1139/v98-127

An ultraviolet photoelectron (PE) spectrometer apparatus that utilizes a tuneable 50 W CW CO₂ laser as a directed heat source was used to study the vacuum pyrolysis of diazoacetophenone (1a) and its p-methyl, p-methoxy, p- chloro, and p-nitro a

Full text of DOI:10.1139/v98-127

1
162
A s tudy of the va c uum pyrolys is of pa ra -
s ubs titute d dia zoa c e tophe none s with He (I)
ultra viole t photoe le c tron s pe c tros c opy
Nic k H. We rs tiu k, He id i M. Muc ha ll, J ia gong Ma , a nd Mic h a e l T.H. Liu
Abstract: An ultraviolet photoelectron (PE) spectrometer apparatus that utilizes a tuneable 50 W CW CO laser as a
2
directed heat source was used to study the vacuum pyrolysis of diazoacetophenone (1a) and its p-methyl, p-methoxy,
p-chloro, and p-nitro analogues 1b, 1c, 1d, and 1e. Analysis of the pyrolysate with He(I) ultraviolet PE spectroscopy
shows that at a laser power level of 26 W (500 ± 50°C) 1a, 1b, 1c, and 1d, cleanly yield the corresponding
phenylketenes 2a, 2b, 2c, and 2d, respectively, the products of the Wolff rearrangement of the incipient ketocarbenes.
Of this group of highly reactive ketenes, which cannot be isolated in the condensed phase at ambient temperature, only
2
a has been the subject of a previous PE spectroscopic study. But our work indicates that the sample of 2a prepared in
the earlier study was impure. The low volatility of p-nitrodiazoacetophenone (1e) thwarted our attempts to generate 2e
and obtain its spectrum. Calculations at semiempirical (AM1) and ab inito (HF/6-31G(d)) levels of theory established
that the diazoacetophenones prefer to adopt twisted syn conformations. That the calculated ionization potentials
(
HAM/3 and Becke3LYP/6-31+G(d)//HF/6-31G(d)) of 1a–1d and the synthesized PE spectra of 1a, 1b, and 1c
correlate well with the PE spectroscopic data supports this finding. Shifts observed in the three low-energy ionizations
of ketenes 2b, 2c, and 2d induced by the para-substitution can be related to the character of the corresponding
occupied molecular orbitals of phenylketene (2a).
Key words: diazoacetophenones, phenylketenes, He(I) photoelectron spectroscopy, thermolysis, quantum chemical calcu-
lations.
Résumé : On a fait appel à un spectromètre photoélectronique (PE) ultraviolet, qui utilise comme source de chaleur un
laser à CO de 50 W ajustable et à VC, pour étudier la pyrolyse sous vide de la diazoacétophénone (1a) et de ses
2
analogues p-méthyl-, p-méthoxy-, p-chloro- et p-nitro- (1b, 1c, 1d et 1e). L’analyse des pyrolysats par spectroscopie PE
ultraviolette à He(I) montre que, à une puissance de laser de 26 W (500 ± 50°C), les composés 1a, 1b, 1c et 1d
conduisent proprement aux phénylcétènes correspondants, soit les produits 2a, 2b, 2c et 2d respectivement, qui
correspondent aux produits de réarrangement de Wolff des cétocarbènes qui se forment. De ce groupe de cétènes
hautement réactifs qui ne peuvent pas être isolés sous forme condensée à la température ambiante, seul le composé 2a
avait fait l’objet d’une étude antérieure par spectroscopie PE. Les résultats obtenus au cours de ce travail indiquent
toutefois que le composé 2a préparé antérieurement n’était pas pur. La faible volatilité de la p-nitrodiazoacétophénone
(
1e) nous a empêché de générer le produit 2e et d’en obtenir un spectre. Des calculs théories aux niveaux
semiempirique (AM1) et ab initio (HF/6-31G(d)) ont permis de déterminer que les diazoacétophénones adoptent
préférentiellement les conformations syn déformées. Le fait que les potentiels d’ionisation calculés (HAM/3 et
Becke3LYP/6-31+G(d)//HF/6-31G(d)) pour les composés 1a–1d et que les spectres PE synthétisés des produits 1a, 1b
et 1c donnent une excellente corrélation avec les données de spectroscopie PE est en accord avec cette conclusion. Les
déplacements observés dans les trois ionisations de faible énergie des cétènes 2b, 2c et 2d sont induits par les
substituants en para et on peut les relier au caractère des orbitales molécules occupées correspondantes du
phénylcétène (2a).
Mots clés : diazoacétophénones, phénylcétènes, spectroscopie photoélectronique He(I), thermolyse, calculs théoriques
par chimie quantique.
[
Traduit par la Rédaction] Werstiuk et al. 1173
1
173
Received March 26, 1998.
1
1
N.H. Werstiuk, H.M. Muchall, and J. Ma. Department of Chemistry, McMaster University, Hamilton, ON L8S 4M1, Canada.
M.T.H. Liu. Department of Chemistry, University of Prince Edward Island, Charlottetown, PE C1A 4P3, Canada.
1Authors to whom correspondence may be addressed. Telephone: (905) 525-9140. Fax: (905) 522-2509. E-mail:
werstiuk@mcmaster.ca; muchall@mcmaster.ca
Can. J. Chem. 76: 1162–1173 (1998)
© 1998 NRC Canada

Werstiuk et al.
1163
Fig. 1. Optimized structures of (a) 1a–1e obtained with AM1 and (b) 1a–1d obtained with HF/6-31G(d).
methyldiazoacetophenone
(1b),
p-methoxydiazoaceto-
Fig. 2. Potential energy surface (AM1) of 1a for the torsion of
NϭC—CϭO and phenyl—CϭO dihedral angles.
phenone (1c), p-chlorodiazoacetophenone (1d), and p-
nitrodiazoacetophenone (1e) that are potential precursors of
ketocarbenes and the corresponding para-substituted
phenylketenes, which are the Wolff rearrangement products
(
7). Only a limited number of ketenes which cannot be iso-
lated in the condensed phase at ambient temperature have
been studied in the gas phase with PE spectroscopy (5, 8–
1
3), and there is comparatively little experimental data avail-
able on the vacuum pyrolysis of α-diazocarbonyl com-
pounds (see, for example, ref.14) in contrast to the large
body of work that has been accumulated on the
photochemistry of this class of compounds (14–21). Given
that a wide variety of diazocarbonyl compounds can be syn-
thesized (14) and, as our studies have shown, easily
pyrolyzed in the gas phase, we expected PE spectroscopic
and calculational studies to provide fundamental information
on aspects of the electronic structure and reactivity of this
group of compounds and the products derived from the in-
cipient ketocarbenes (4, 5). We undertook a study of 1a, 1b,
1
c, 1d, and 1e specifically (i) to gain information about the
nature of the frontier molecular orbitals (MOs), (ii) to study
the conformational properties of these compounds in the gas
phase, (iii) to establish whether our instrument is generally
useful for preparing and studying ketenes in the gas phase
To facilitate the preparation and detection of highly reac-
tive transients in the gas phase, we developed a He(I) ultra-
(
five new para-substituted phenylketenes were to be pre-
pared), and (iv) to determine the effect of the para-substitu-
ents on the ionization potentials of 2a with the goal of char-
acterizing its highest occupied molecular orbitals. This paper
documents the results of a PE spectroscopic study of the py-
rolysis of diazoacetophenones 1a, 1b, 1c, and 1d and a com-
putational study of the conformational properties and molec-
ular orbitals of this group of compounds, the nitro analogue
violet photoelectron (PE) spectrometer-CW CO laser appa-
2
ratus and used it successfully to study the vacuum pyrolysis
of 6,6-dihalobicyclo[3.1.0]hexanes (1), 1,2,3-benzotriazines (2),
4
-diazoisochroman-3-one (3), and 4-diazoisothiochroman-3-
2
,7
one (4), 11-oxatricyclo[6.2.1.0 ]-undeca-2,9-diene (5), and
2
3
,2-dimethoxy-5,5-dimethyl-∆ -1,3,4-oxadiazoline (6). To
widen our research in this field, we undertook a study of the
vacuum pyrolysis of diazoacetophenone (1a), p-
1
e, and the corresponding ketenes (Scheme 1).
©
1998 NRC Canada

1
164
Can. J. Chem. Vol. 76, 1998
–
1
Table 1. The AM1 heats of formation ∆H (kJ mol ) and HF/6-31G(d) total energies E of two conformations of
f
T
diazoacetophenones 1a–1e and of phenylketenes 2a–2d.
–
1
Compound
∆Hf
∆∆Hf
ET (hartrees)
∆E (kJ mol )
T
syn-1a
anti-1a
syn-1b
anti-1b
syn-1c
anti-1c
syn-1d
anti-1d
syn-1e
anti-1e
221.42
232.55
188.87
199.95
60.54b
72.68
192.67
205.06
240.75
253.05
72.22
–490.141 659a
–490.135 222a
–529.179 538
–529.173 039
–604.023 915
–604.017 046
–949.041 019
11.13
11.08
12.13
12.38
12.30
16.88
17.05
18.02
2
2
2
2
a
b
c
–381.274 201
–420.311 107
–495.153 149
–840.173 832
40.25
–85.40b
42.55
d
a
–6
The total energies obtained with the keyword TIGHT are identical within 10 hartrees to the values given.
b
The stereoisomer with the methoxy group orientated anti to the CϭO group is marginally higher in energy (1c: ∆H = 61.17 kJ
f
–
1
–1
mol ; 2c: ∆H = –85.23 kJ mol ) than the correponding syn isomer.
f
a
Table 2. Selected calculated dihedral angles (degrees) of two
conformations of diazoacetophenones 1a–1e.
Scheme 1.
Dihedral
Compound
NϭC—CϭO
phenyl—CϭO
b
c
b
c
c
syn-1a
anti-1a
syn-1b
anti-1b
syn-1c
anti-1c
syn-1d
anti-1d
syn-1e
anti-1e
a
4.1(2.2) (1.6)
31.6(17.0) (17.3)
b
c
b
179.9(165.2) (165.2)
4.1(2.1)b
58.6(36.8) (36.7)
30.0(15.0)b
53.9(35.1)b
b
173.4(164.6)
b
b
4.3(1.7)
31.3(10.3)
172.1(163.3)b
2.9
52.3(32.2)b
35.9
178.9
2.4
175.7
81.2
38.1
60.3
Unless stated otherwise the values listed were calculated with AM1.
Calculated at the HF/6-31G(d) level of theory.
Calculated at the HF/6-31G(d) level of theory with the keyword
b
c
TIGHT.
HAM/3//AM1 and HAM/3//HF/6-31G(d) indicate that HAM/3
calculations were carried out on AM1 and HF/6-31G(d) op-
timized equilibrium geometries, respectively. Synthetic
spectra were calculated from the MO results with a FOR-
TRAN program PESPEC (29). Spectra were synthesized
from the 10 lowest IPs (HAM/3), with a Gaussian convolu-
tion of 0.4 and the temperature set at 300 K. Graphical rep-
resentations of the computationally determined eigenvectors
(AM1//HF/6-31G(d)) were plotted from HyperChem 4.0.
Calculations
Calculations at the semiempirical level of theory were
carried out with AM1 of HyperChem 4.0 (22), AMPAC 2.0
(
23), or MOPAC 6.0 (24), with the latter two programs run-
ning on IBM RS/6000 model 320H, 350, and 530 comput-
ers. An RMS gradient of 0.02 was used to obtain optimized
geometries with HyperChem 4.0, and the keyword PRECISE
was used in the case of AMPAC and MOPAC calculations to
tighten the convergence criteria. Ab initio HF (25) and
Becke3LYP (26) calculations were carried out with
GAUSSIAN 94 (27) running on IBM RS/6000 model 350,
PE spectroscopic studies
The locally built PE spectrometer-CW CO infrared laser
2
apparatus (the laser is used as a directed heat source) em-
ployed in this study is described in our earlier publications
(1–5). To obtain the PE spectra of the diazoacetophenones
(1a, mp 51–52; 1b, mp 57–58; 1c, mp 87–88; 1d, mp 115–
116; 1e, mp 116–117°C), which were prepared by estab-
lished methods (30), a 20 mg sample was placed in a bottle-
shaped quartz reaction vessel that was inserted into the tip of
a heated probe. The temperature was slowly ramped up (a
530, 39H, and SP2 and SGI R10000 Impact computers. Opti-
mized geometries were shown to be minima by frequency
calculations; total energies are uncorrected for zero-point ener-
gies. The semiempirical method HAM/3 (28) was used to
compute the ionization potentials. The acronyms
©
1998 NRC Canada

Werstiuk et al.
1165
a
Table 3. Calculated vertical ionization potentials IP (eV) and orbital energies ε (eV) for diazoacetophenones 1a–1d and phenylketenes
2
1
a–2d.
b
c
a
HAM3
HAM3
IP
IP
–ε
IP
IP
IP
–ε
IP
IP
IP
–ε
IP
–ε
IP
IP
–ε
IP
IP
–ε
IP
IP
–ε
IP
–ε
IP
8.60
8.72
6.87
8.8
8.41
8.55
6.77
8.6
8.28
8.35
6.46
8.2
7.00
8.8
8.17
5.97
7.9
7.93
5.79
7.7
7.67
5.52
7.4
6.08
7.9
9.09
9.47
7.21
9.1
9.07
9.28
7.05
8.9
8.75
8.77
6.80
8.5
7.25
9.1
9.36
7.15
9.1
9.20
7.07
9.0
9.17
7.15
9.1
7.53
9.3
9.53
9.58
7.41
9.3
9.24
9.36
7.27
9.1
8.95
9.28
7.16
8.9
7.51
9.3
10.19
8.20
10.1
9.83
7.93
9.8
9.33
7.46
9.4
9.60
9.66
7.47
9.4
9.39
9.38
7.34
9.1
9.43
9.42
7.43
9.1
7.77
9.6
11.99
9.62
11.5
11.73
9.43
11.3
11.13
9.11
11.0
8.70
10.5
11.74
11.87
9.82
11.96
12.15
10.17
12.1
11.86
11.98
9.89
12.33
12.49
10.24
12.1
12.14
12.19
9.96
12.72
13.10
11.37
13.3
12.46
12.82
10.85
12.7
12.24
12.45
10.63
12.3
10.76
12.6
13.80
11.74
13.6
13.01
11.23
13.1
12.60
10.68
12.6
11.18
13.0
13.22
13.68
11.61
13.5
12.71
12.95
11.49
13.3
12.43
12.83
10.87
12.6
11.04
12.8
14.04
12.14
14.0
13.54
11.61
13.5
13.25
11.48
13.4
11.75
13.6
13.83
13.88
12.01
13.9
12.90
13.34
11.88
13.7
12.64
12.91
11.32
13.0
11.69
13.5
14.21
12.55
14.5
13.74
11.65
13.6
13.35
11.81
13.7
12.15
14.0
d
B3LYP
e
e
11.7
b
1
1
b
c
HAM3
HAM3
B3LYP
11.39
11.51
9.62
c
d
11.4
11.7
11.8
b
HAM3
HAM3
B3LYP
10.84
10.95
9.13
10.8
8.93
10.7
12.03
9.86
11.8
11.84
9.66
11.6
11.25
9.35
11.3
9.68
11.27
11.46
9.32
11.0
9.46
12.05
12.33
10.24
11.9
10.32
12.1
13.35
11.45
13.4
12.56
10.63
12.5
12.21
10.48
12.4
10.47
12.3
c
d
e
e
1
2
d
a
B3LYPd
11.3
HAM3b
12.27
10.47
12.4
12.10
10.18
12.1
12.15
10.03
11.9
10.18
12.0
d
B3LYP
e
e
2
2
2
b
c
HAM3b
d
B3LYP
HAM3b
d
B3LYP
e
e
d
B3LYPd
7.97
9.8
11.5
a
Twisted syn conformers.
At the HAM/3//AM1 level of theory.
b
c
At the HAM/3//HF/6-31G(d) level of theory.
d
At the Becke3LYP/6-31+G(d)//HF/6-31G(d) level of theory.
e
d
Orbital energies shifted uniformly, so that the HOMO energy equals the calculated first IP as given in Table 4.
v
maximum temperature of 80°C was used) until a suitable
–
3
–3
source pressure in the region of (3 × 10 )–(8 × 10 ) Torr (1
Torr = 133.3 Pa) was obtained. In the case of 1b, 1c, and 1d,
Calculations
. Conformational properties of diazoacetophenones and
which exhibited low volatilities, N was used as a pressure
2
2
make-up gas to improve the signal to noise. Consequently,
1
vertical expansions of the spectra were required in the re-
gion of 8–15 eV.
phenylketenes
Figure 1(a) shows the AM1 optimized geometrical struc-
tures of the syn forms of 1a–1e. The gas-phase torsional po-
tential energy surfaces of 1a, 1b, and 1c were calculated
with AM1 by varying the NϭC—CϭO and phenyl—CϭO
dihedral angles from 0° to 180° in 15° increments. At the
AM1 level, 1a prefers a twisted syn rather than an anti con-
formation of the NϭC—CϭO group, and the conformer
with the NϭC—CϭO and phenyl—CϭO dihedral angles
fixed at 90° is the highest energy species. Because the po-
tential energy surfaces of 1a–1c exhibit very similar topolo-
gies we report only the surface obtained for
diazoacetophenone (1a) as Fig. 2. That the surfaces of 1b
and 1c (not shown) are virtually identical to the surface of
1a indicates that 1a, 1b, and 1c have similar conformational
properties in the gas phase. That twisted syn conformers are
Spectra were obtained by averaging 20–35 scans and cali-
brated with the 15.6 eV band of N . Linearity of the scale
2
was ensured through calibrations with methyl iodide (9.54
and 10.16 eV) and O (12.30 eV) performed prior to the ex-
2
periments. Vertical ionization potentials are accurate to ±
.05 eV. Pyrolyses of the diazoacetophenones were carried
0
out in a hot zone 1–2 mm in length obtained by focusing the
laser beam (power levels of 26, 42, and 47 W were used) on
the tip (2–3 mm inner diameter) of the bottle-shaped quartz
sample vessel. Due to the very low volatility of 1e we were
not able to obtain its spectrum and study its pyrolysis even
in the presence of the pressure make-up gas. Hatched bands
in the PE spectra are due to traces of H O (12.6 eV) and a
2
shadow peak (13.6 eV) of N2.
2
Although He(I) PE spectra of 1a, 1b, and 1d already have been reported by Modelli et al. (31) they are presented again for two reasons:
i) the spectra reported earlier contain bands of nitrogen — and along with these, bands of further products — due to decomposition of the
(
compounds, and (ii) the different analyzer transmission functions of the spectrometers used in the previous and in this study lead to PE
spectra that differ in intensities of the bands. As we planned to record pyrolyses spectra, knowledge of the actual appearance of the spectra
of the starting compounds taken with our instrument was essential.
©
1998 NRC Canada

1
166
Can. J. Chem. Vol. 76, 1998
a
Table 4. Total energies (hartrees) of diazoacetophenones 1a–1d
and phenylketenes 2a–2d (N electrons) and their radical cationsb
(
Fig. 3. The He(I) PE spectra of (a) 1a, (b) the pyrolysate of 1a
obtained at a laser power level of 26 W, and (c) the pyrolysate
of 1a obtained at a laser power level of 42 W. The
N – 1 electrons) as well as derived first vertical ionization
potentials IP (eV).
HAM/3//AM1 (vertical solid lines) and HAM/3//HF/6-31G(d)
v
(
vertical broken lines) calculated IPs are given in (a) for 1a and
B3LYP/6-31+G(d)//HF/6-31G(d)
Number of
(b) for 2a.
electrons
Total energy
IPv
1
1
1
1
2
2
2
2
a
b
c
N
N–1
N
N–1
N
N–1
N
N–1
N
N–1
N
N–1
N
N–1
N
–493.149 931
–492.826 329
–532.468 968
–532.154 406
–607.677 699
–607.377 936
–952.745 171
–952.423 728
–383.668 080
–383.376 646
–422.986 065
–422.703 768
–498.192 693
–497.922 314
–843.263 383
–842.971 686
8.81
8.56
8.17
8.75
7.93
7.68
7.36
7.94
d
a
b
c
d
N–1
a
Twisted syn conformers.
b
Single-point energy calculations on the optimized molecule (N
electrons) geometries.
Table 5. Selected experimental vertical ionization potentials IP
eV) of diazoacetophenones 1a–1d and phenylketenes 2a–2d.
(
Compound
IP
1
1
1
1
2
2
2
2
a
b
c
d
a
b
c
9.10
8.78
8.71
9.14
8.23
8.22
7.91
8.29
9.6a
9.19
11.86
11.62
11.01
11.24
9.00
9.12
9.33
9.34
9.36
9.22
9.25
9.41
9.6b
9.59
10.41
10.17
9.56
9.78
11.78
11.38
11.09
10.55
d
9.74
a
Consists of three overlapping bands.
Consists of two overlapping bands.
approximately one-half the values calculated with AM1
b
(
Table 2). This is most likely a result of the fact that the ab
initio calculations more accurately recover the stabilizing
phenyl—CϭO–π interaction, which would act to decrease
the twist angle. When the geometries of the syn and anti iso-
mers of 1a were reoptimized — the geometries obtained
with the default convergence limits were used as input — at
the HF/6-31G(d) level of theory with the keyword TIGHT to
the low-energy structures in the cases of 1d and 1e as well
indicates that the diazoacetophenones which are the subject
of this study prefer to adopt twisted syn conformations in the
gas phase.
At the AM1 level, the twisted syn isomers of 1a–1e (the
NϭC—CϭO twist angles are 4.1, 4.1, 4.3, 2.9, and 2.4°; the
phenyl—CϭO dihedral angles are 31.6, 30.0, 31.3, 35.9,
and 38.1°, respectively) are 11.08–12.38 kJ mol– lower in
energy than the twisted anti isomers (Table 1). The syn and
anti isomers of 1a, 1b, and 1c were also studied at the HF/6-
tighten the convergence criteria, the total energies (E s) ob-
T
–
6
tained were identical within 10 hartree to E s calculated
T
with the default convergence limits, and the geometries dif-
fered little from the input geometries (Table 2). These re-
sults are in accord with published experimental data: 1,4-
bisdiazo-2,3-butanedione adopts a bis-syn conformation in
1
1
3
1G(d) level of theory. Selected geometrical parameters are
the solid state (32), and variable-temperature solution H
collected in Table 2, and the optimized equilibrium geome-
tries of the syn isomers of 1a–1d are displayed in Fig. 1(b).
At the HF/6-31G(d) level, we find that the twisted syn iso-
mers of 1a, 1b, and 1c are 16.88, 17.05, and 18.02 kJ mol
lower in energy than the twisted anti isomers, and the
NϭC—CϭO and phenyl—CϭO twist angles are
NMR studies in the temperature range of +30 to –40°C
showed that diazoketones of the type RCOCHN (RϭH,
2
CH , C H , t-C H , C H CH , CH O, and C H O) adopt syn
3
2
5
4
9
6
5
2
3
2
5
–
1
and anti conformations with the syn forms being preferred
(33). In the case of 1a, 1c, and 1e, single stereoisomers iden-
tified as the syn forms were detected down to –40°C, but
©
1998 NRC Canada

Werstiuk et al.
1167
Fig. 5. The He(I) PE spectra of (a) 1c, (b) the pyrolysate of 1c
obtained at a laser power level of 26 W, and (c) the pyrolysate
of 1c obtained at a laser power level of 42 W. The
HAM/3//AM1 (vertical solid lines) and HAM/3//HF/6-31G(d)
(vertical broken lines) calculated IPs are given in (a) for 1c and
(b) for 2c.
Fig. 4. The He(I) PE spectra of (a) 1b, (b) the pyrolysate of 1b
obtained at a laser power level of 26 W, and (c) the pyrolysate
of 1b obtained at a laser power level of 47 W. The
HAM/3//AM1 (vertical solid lines) and HAM/3//HF/6-31G(d)
(
(
vertical broken lines) calculated IPs are given in (a) for 1b and
b) for 2b.
especially important where a good agreement between
absolute energy values is needed, as is the case for highly
reactive or transient species which are generated in the PE
spectrometer itself. The semiempirical HAM/3 method gives
IPs directly, and both semiempirical and ab initio optimized
geometries can be employed. The HAM/3//AM1 and
HAM/3//HF/6-31G(d) results for diazoacetophenones 1a–1c
are given in Table 3. Although AM1 and HF/6-31G(d) ge-
ometries for 1a–1c are distinctly different, IP values from
the two levels of theory differ little. We therefore report
HAM/3//AM1 results only for phenylketenes 2a–2c. Be-
cause HAM/3 is not parameterized for chlorine, we also use
the routine we developed in previous publications (6, 34):
orbital energies calculated with the hybrid HF/DFT
NϭC—CϭO and phenyl—CϭO twist angles were not de-
termined (33).
Phenylketenes 2a–2e are found to adopt planar conforma-
tions with AM1, in accordance with previous findings for 2a
(
13). This also holds at the ab initio level, 2a–2e are planar
with HF/6-31G(d). The AM1 heats of formation and ab in-
itio total energies of 2a–2e are listed in Table 1. According
to AM1, the 90° conformation of 2a is destabilized by 9.04
–
1
kJ mol .
2
. Ionization potentials
We have repeatedly demonstrated that the use of calcu-
lated ionization potentials (IPs) instead of orbital energies
for the interpretation of PE spectra leads to excellent corre-
lations between theory and experiment (2–4, 6, 34). This is
3
Becke3LYP method are shifted uniformly so that the
3
The successful application of the Koopmans’ theory-like (35) interpretation of orbital energies obtained at the DFT level of theory in PE
spectroscopic studies of numerous compounds has been shown and commented on (6, 34, 36–38).
©
1998 NRC Canada

1
168
Can. J. Chem. Vol. 76, 1998
Fig. 6. The He(I) PE spectra of (a) 1d, (b) the pyrolysate of 1d
obtained at a laser power level of 26 W, and (c) the pyrolysate
of 1d obtained at a laser power level of 42 W. The
Fig. 7. (a) Synthetic partial PE spectrum (from HAM/3//AM1
IPs) and (b) torsional potential (AM1) of 2a.
Becke3LYP/6-31+G(d)//HF/6-31G(d) (vertical solid lines)
calculated IPs are given in (a) for 1d and (b) for 2d.
Photoelectron spectroscopic studies
1
. Experimental PE spectra
The spectrum of 1a (20 mg; source pressure 2.2 × 10
–
3
Torr) displayed as Fig. 3(a) was obtained at a probe temper-
ature of 50°C. The vertical solid lines and vertical broken
lines on the spectrum correspond to the 10 lowest
HAM/3//AM1 and HAM/3//HF/6-31G(d) IPs (values in Ta-
ble 3), respectively, and show that the experimental spec-
trum of this diazoacetophenone can be reproduced nicely
with HAM/3. Figure 3(b) is a display of the spectrum of the
pyrolysate of 1a obtained at a laser power of 26 W (~500 ±
5
8
0°C). Based on the 1:1:1 relative intensities of the bands at
.2, 9.3, and 10.4 eV and excellent correlation between the
experimental vertical IPs and the IPs of phenylketene (2a)
calculated with HAM/3//AM1 (vertical solid lines), the
conversion of 1a to 2a is complete at a laser power level of
26 W. The IPs for 1a and 2a calculated with Becke3LYP/6-
31+G(d)//HF/6-31G(d) (Table 3, compare to experimental
IPs in Table 5) give an equally good representation of both
spectra and therefore support this. That there was no signifi-
cant change in the spectrum of the pyrolysate (Fig. 3(c)) and
no CO (it exhibits a sharp band at 14.01 eV (39)) was
formed when the laser power was increased from 26 W to
42 W (~850 ± 50°C) indicates that the loss of CO from 2a is
HOMO energy equals that of the calculated first vertical
ionization potential IP . The IP is determined from the en-
v
v
ergy difference between the neutral (N electrons) and its
radical cation (N – 1 electrons, geometry of the neutral) ob-
tained using Becke3LYP/6-31+G(d)//HF/6-31G(d) (Ta-
ble 4). Orbital energies and ionization potentials derived
from the shift are reported in Table 3 for 1a–1d and 2a–2d.
This level of theory is sufficient as could be concluded from
4
virtually identical IP s and orbital energies of 2a–2d ob-
subject to a high activation energy. This observation is in
v
tained with Becke3LYP/6-31+G(d)//Becke3LYP/6-31+G(d)
keeping with the results of our previous calculational study
(
not given).
(5). The spectrum of 2a obtained in this study differs to
4
A comprehensive computational study of the decomposition of 1a, 1c, and 1e (loss of nitrogen) and 2a, 2c, and 2e (loss of CO) as well as
the rearrangement of the ketocarbenes derived from 1a, 1c, 1d, and 1e will be the subject of a following publication. As shown in
Scheme 1, the pyrolyses of the diazoacetophenones are expected to proceed via the respective intermediate ketocarbenes, even though
thermolysis spectra provide no evidence for this. Alternatively, extrusion of nitrogen and rearrangement could be a concerted process.
©
1998 NRC Canada

Werstiuk et al.
1169
Fig. 8. (a) Synthetic partial PE spectrum (HAM/3//AM1), (b) surface for the torsion of the phenyl—CϭO dihedral angle (AM1),
(
c) synthetic spectrum (HAM/3//HF/6-31G(d)), (d) torsional potential (HF/6-31G(d)), and (e) angular dependences of the four lowest
HAM/3//HF/6-31G(d) IPs of 1a.
some degree from the published spectrum of phenylketene
that was prepared by the zinc-induced dechlorination of 2-
chloro-2-phenylacetyl chloride (13). While we find the rela-
tive areas of the first three bands to be 1:1:1 (Fig. 3(b)) in
line with HAM/3//AM1 and Becke3LYP/6-31+G(d)//HF/6-
cleanly yields p-methylphenylketene (2b). The IPs for 1b
and 2b calculated using Becke3LYP (Table 3) agree with
this finding. Figure 4(c) displays the PE spectrum of the
pyrolysate of 1b obtained at a laser power of 47 W
(>900°C). That spectra in Figs. 4(b) and 4(c) are virtually
identical and no CO is formed indicates that the activation
energy for decarbonylating ketene 2b, as in the case of 2a, is
significantly higher than the activation energy for loss of ni-
trogen from the respective diazoacetophenones.
3
1G(d) (Table 3) results, the relative areas of the first three
bands of the previously published spectrum are
approximately 1:2:1, an indication that phenylketene was not
the sole product of the dechlorination reaction.
The spectrum of 1b (20 mg; probe temperature 65°C), ob-
tained with nitrogen added as a make-up gas to maintain the
total pressure in the source at 3 × 10 Torr, is displayed
along with the HAM/3//AM1 and HAM/3//HF/6-31G(d) IPs
as Fig. 4(a). Again, HAM/3 values mirror experimental IPs.
Figure 4(b) is a display of the spectrum of the pyrolysate of
Figure 5(a) displays the PE spectrum of 1c obtained at a
probe temperature of 76°C along with its HAM/3//AM1 and
–
3
HAM/3//HF/6-31G(d) IPs. As in the case of 1b, N was
2
used as a make-up gas to maintain the source pressure at 3 ×
–
3
10 Torr. HAM/3 once more correctly predicts the experi-
mental spectrum. The spectra displayed as Figs. 5(b) and
(c) were obtained when 1c was pyrolyzed at laser power lev-
els of 26 and 42 W, respectively. These spectra show that 1c
cleanly affords p-methoxyphenylketene (2c), which does not
undergo pyrolysis at a nozzle-tip temperature in the region
1
b obtained at a laser power of 26 W and the calculated IPs
of ketene 2b. As in the case of the unsubstituted ketene 2a,
the experimental IPs in this spectrum are accurately repro-
duced by the computed values for 2b, which shows that 1b
©
1998 NRC Canada

1
170
Can. J. Chem. Vol. 76, 1998
Fig. 9. The four highest occupied molecular orbitals of 1a.
Fig. 10. (a) Synthetic partial PE spectrum (HAM/3//HF/6-
3
1G(d)) and (b) surface for the torsion of the phenyl—CϭO
dihedral angle (HF/6-31G(d)) of 1b.
of 850 ± 50°C. Again, HAM/3 and Becke3LYP IPs for 1c
and 2c do not differ much, which shows the validity of our
Becke3LYP approach.
Figure 6(a) displays the PE spectrum of 1d obtained at a
probe temperature of 55°C with N added as a make-up gas
2
–
3
to maintain the source pressure at 3 × 10 Torr. Fig-
ures 6(b) and (c) display spectra of the pyrolysate of 1d
obtained at laser power levels of 26 and 42 W, respectively.
As HAM/3 is not parameterized for chlorine, Figs. 6(a) and
(
b) include the 10 lowest ionization potentials of 1d and 2d
as calculated with Becke3LYP/6-31+G(d)//HF/6-31G(d) and
described above. As expected from the encouraging results
of the pyrolyses of 1a–1c with respect to this approach, the
correlation between predicted spectra (vertical bars) and ex-
perimental spectra is excellent. These spectra show that 1d
yields p-chlorophenylketene (2d), which also appears to be
stable at a nozzle-tip temperature in the region of 850 ±
surfaces and the angle dependence of ab initio or
semiempirical orbital energies as well as IPs calculated with
HAM/3 using PESPEC correlate well with experimental
spectra and can be used to obtain information about the
conformational properties of molecules in the gas phase (29,
4
0–42). Additional support for this finding is seen in the ex-
cellent correlation between the experimental spectrum of
phenylketene (2a) given in Figs. 3(b) and (c) and the synthe-
sized spectrum displayed as Fig. 7(a). The synthetic spec-
trum of 2a (with a Gaussian convolution of 0.4, temperature
is 300 K) was obtained from the AM1 potential energy sur-
face for the rotation of the ketene group given in
Fig. 7(b) and the angle dependence of the 10 lowest
HAM/3//AM1 ionization potentials (not shown).
5
7
0°C. The spectrum in Fig. 6(c) also exhibits a feature at
.8 eV. A calculational study on p-chlorophenyloxirene,
which is a second rearrangement product from the
ketocarbene derived from 1d, at the Becke3LYP/6-
3
1+G(d)//HF/6-31G(d) level of theory established that the
five lowest IPs of this compound should be 7.5, 9.6, 9.9,
1.0, and 11.7 eV. While the calculated first IP for p-
1
chlorophenyloxirene is close to the first IP in the pyrolysis
spectrum, at this point we cannot say with certainty whether
While Modelli et al. (31) do not offer an interpretation of
the spectra of 1a, 1b, and 1d with respect to orbital interac-
tions or conformational considerations, for 1a Vorob’ev et
al. (43) report only weak π–π interactions both in the
heterobutadiene group due to a large energy gap between in-
teracting MOs and between diazo group and phenyl ring be-
cause of inefficient spatial overlap. Figure 8(a) is the PE
spectrum of diazoacetophenone (1a) synthesized from the
angle dependence of the HAM/3//AM1 IPs (not shown) and
the AM1 potential energy surface displayed as Fig. 8(b) that
corresponds to the 0–180° trajectory (the NϭC—CϭO di-
hedral angle is fixed at 0°) of Fig. 2. The synthetic spectrum
recovers the general features of the experimental spectrum.
Figure 8(c) is the synthetic spectrum of 1a obtained from the
5
it actually is a pyrolysis product of 1d.
The low volatility of 1e thwarted our attempts to obtain
its spectrum and study its pyrolysis. Even heating a rela-
tively large sample (40 mg) of 1e at 94°C near its melting
point of 116°C in a 4 mm outer diameter tube inserted into
the probe did not yield an acceptable spectrum. Selected ex-
perimental vertical ionization potentials of the
diazoacetophenones 1a–1d and the corresponding ketenes
are collected in Table 5.
2
. Synthesis of PE spectra
We have already demonstrated in several previous studies
that PE spectra synthesized from torsional potential energy
5
While from the first IP p-chlorophenyloxirene is a likely candidate, it will only be present in the pyrolysis gas mixture if there is a reason-
able barrier to the reverse reaction, the ring opening of the three-membered ring to give the ketocarbene. Calculations for the rearrangement
4
of ketocarbenes derived from diazoacetophenones 1a, 1c, and 1e to the corresponding oxirenes have shown no significant barrier.
©
1998 NRC Canada

Werstiuk et al.
1171
Fig. 11. (a) Synthetic partial PE spectrum (HAM/3//AM1), (b) surface for the torsion of the phenyl—CϭO dihedral angle (AM1),
(
c) synthetic spectrum (HAM/3//HF/6-31G(d)), (d) torsional potential (HF/6-31G(d)), and (e) angular dependences of the four lowest
HAM/3//HF/6-31G(d) IPs of 1c.
HF/6-31G(d) torsional potential (Fig. 8(d)) and the angular
dependence of 10 HAM/3//HF/6-31G(d) IPs of which four
are displayed in Fig. 8(e). Even though the rotational barrier
at this level of theory is substantially higher than the AM1
barrier and the twisted syn conformer is only marginally
lower in energy than the planar structure, the synthetic spec-
trum closely resembles the one obtained from the AM1
potential energy surface. This result derives from the fact
that the low-energy IPs show a low sensitivity to the twist
angle in both cases, as can be seen in Fig. 8(e). An explana-
tion for this is given by the corresponding molecular orbitals
According to Becke3LYP/6-31+G(d)//HF/6-31G(d),6 the
HOMO is a π-type orbital with coefficients mainly on the
diazo group (πCϭN), HOMO – 1 is a σ-type orbital, which
represents an oxygen lone pair (n ), and HOMO – 2 and
O
HOMO – 3 are mostly phenyl π-type MOs (π) with small
coefficients on the oxygen atom. This orbital sequence is in
accord with the one given by Modelli et al. (31), while
Vorob’ev et al. (43) give πCϭN > π > n > π from MNDO
O
calculations. Thus in 1a, which can be considered to be the
parent of this class of compounds, the low lying occupied
MOs of the carbonyl group effectively function as an insula-
tor for phenyl and diazo moieties.
(
Fig. 9), which show that phenyl and diazoketo groups are
present as separate units that do not interact at this energy
level, a fact that was hinted at by Vorob’ev et al. (43).
Figure 10(a) displays the synthetic spectrum of 1b ob-
tained from the torsional potential obtained at the HF/6-
6
Even though there is “no physical relevance” to Kohn–Sham orbitals, we have found eigenvectors obtained at the Becke3LYP/6-
1+G(d)//HF/6-31G(d) level of theory to be similar to those computed with HF/6-31G(d)//HF/6-31G(d) or AM1//HF/6-31G(d). Yet for a
3
number of diazoketones and diazo acyl esters containing an aromatic ring the predicted orbital sequence differs at the above-mentioned lev-
els of theory. In a following paper (44) we will show that in these cases the orbital sequence as determined with the Becke3LYP method
should be taken. For this reason, the orbitals displayed in Fig. 9 are arranged according to the Becke3LYP/6-31+G(d)//HF/6-31G(d) calcu-
lated sequence.
©
1998 NRC Canada

1
172
Can. J. Chem. Vol. 76, 1998
Fig. 12. The three highest occupied molecular orbitals of 2a.
support that diazoacetophenones 1a–1d prefer twisted syn
conformations in the gas phase.
Substituent effects on ionization potentials of
phenylketenes
Figure 12 displays the HOMO, HOMO – 1, and HOMO – 2
of phenylketene (2a) obtained at the AM1//HF/6-31G(d)
level. The HOMO is a π-type MO characterized by substan-
tial antibonding mixing of the ketene and phenyl π systems,
with the latter exhibiting a large p coefficient at C-4. The
z
HOMO – 1 is a pure phenyl π-type MO, which does not mix
with the ketene π system and which has no p component at
z
C-4. The HOMO – 2 is the bonding π-type MO that, like the
HOMO, has contributions from the ketene and phenyl
groups. Nevertheless, it has a larger p coefficient at C-4
z
than the HOMO. The nature of these frontier MOs indicates
that the HOMO and HOMO – 2 should be affected more
strongly than HOMO – 1 by para-substituents, which inter-
act with the phenyl ring via π conjugation. That this is the
case is seen in the PE spectra of 2a (Figs. 3(b) and (c)), 2b
3
1G(d) level (Fig. 10(b)) and the HAM/3//HF/6-31G(d) IPs
(
Figs. 4(b) and (c)), and 2c (Figs. 5(b) and (c)), and the data
(
not shown). The torsional dependence of the IPs as well as
in Table 5, which clearly show that HOMO and HOMO – 2
are destabilized by the introduction of π donors at C-4. That
HOMO – 2 appears to be more sensitive to substitution at C-
the corresponding molecular orbitals are similar to what we
found for 1a. On the whole, the general features of the ex-
perimental spectrum are reproduced in the synthethic spec-
trum with reasonable accuracy even though there is base-
line separation of one band from three overlapping bands.
Figure 11(a) is the synthetic spectrum of 1c obtained from
the AM1 potential energy surface shown as Fig. 11(b).
While the general features of band structure correlate with
the PE spectrum, the relative intensities of the peaks of the
low-IP band are at variance with the experimental data. On
the other hand, the synthetic spectrum of 1c (Fig. 11(c)) ob-
tained from the HF/6-31G(d) torsional potential (Fig. 11(d))
and the angular dependence of 10 HAM/3//HF/6-31G(d)
ionization potentials of which four are displayed in
Fig. 11(e) closely resembles the experimental spectrum. In
contrast to 1a and 1b, the HOMO of 1c is a phenyl π-type
orbital that is raised in energy above the πCϭN because of a
4
is a consequence of the size of the p component at C-4
z
and its proximity in energy to the other filled orbitals of
lower energy with which it interacts.
The gas-phase thermolysis of diazoacetophenones 1a–1d
at a laser power level of 26 W (500 ± 50°C) gave the corre-
sponding phenylketenes 2a–2d, which do not decompose at
higher laser power levels. A comparison of our PE spectrum
of 2a with the one published earlier indicates that the sample
prepared in the previous study was impure. The low volatil-
ity of 1e made the recording of its PE spectrum and the gen-
eration of 2e impossible. According to AM1 and HF/6-
3
1G(d) calculations, diazoacetophenones prefer twisted syn
substantial coefficent on the methoxy oxygen atom. This
antibonding combination (π ) is followed by π
and the corresponding bonding combination (π ). As a con-
sequence of this change in orbital sequence a phenyl π-type
–
conformations in the gas phase. This is supported by excel-
lent correlations between calculated ionization potentials of
nO, π,
CϭN,
+
1
a–1d at the HAM/3 and Becke3LYP/6-31+G(d)//HF/6-
–
31G(d) levels of theory as well as simulated PE spectra of
orbital (π ) and the π
are closer in energy in 1c than the
CϭN
1
a–1c and experimental spectra.
respective orbitals in 1a, which results in electron density on
the diazo group being mixed into the HOMO of 1c. Despite
the presence of electron density on diazo group and aromatic
ring, it is clear from Fig. 11(e) that the rotation of the ring
does not affect the HOMO energy significantly, which again
shows the insulating effect of the low-lying carbonyl
orbitals. Our study shows that when compared to spectra
synthesized on the basis of torsional potentials obtained at
the HF/6-31G(d) level, experimental PE spectra provide
We gratefully acknowledge the usage of the IBM SP2
computer at Queen’s University granted under the auspices
of an IBM Canada – Queen’s University collaborative pro-
ject as well as the usage of the SGI computing installation in
the Geochemistry Labs at McMaster funded by NSERC. We
thank Ms. O. Donini for assistance with operational aspects
of parallel GAUSSIAN 94 and the Natural Sciences and En-
gineering Research Council of Canada for financial support.
1
2
. N.H. Werstiuk, C.D. Roy, and J. Ma. Can. J. Chem. 72, 2537
1994).
. N.H. Werstiuk, C.D. Roy, and J. Ma. Can. J. Chem. 73, 146
1995).
(
(
©
1998 NRC Canada

Werstiuk et al.
1173
3
4
5
6
. N.H. Werstiuk, J. Ma, C.D. Roy, A.J. Kresge, and E.A. Jeffer-
son. Can. J. Chem. 73, 1738 (1995).
. N.H. Werstiuk, C.D. Roy, and J. Ma. Can. J. Chem. 74, 1903
26. A.D. Becke. J. Chem. Phys. 98, 5648 (1993); C. Lee, W. Yang,
and R.G. Parr. Phys. Rev. B: Condens. Matter, 37, 785 (1988).
27. M.J. Frisch, G.W. Trucks, H.B. Schlegel, P.M.W. Gill, B.G.
Johnson, M.A. Robb, J.R. Cheeseman, T. Keith, G.A.
Petersson, J.A. Montgomery, K. Raghavachari, M.A. Al-
Laham, V.G. Zakrzewski, J.V. Ortiz, J.B. Foresman, 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. Gon-
zalez, and J.A. Pople. Gaussian 94, Revision B.3. Gaussian,
Inc., Pittsburgh Pa. 1995.
(
1996).
. N.H. Werstiuk, J. Ma, C.D. Roy, A.J. Kresge, and E. Jefferson.
Can. J. Chem. 74, 2536 (1996).
. H.M. Muchall, N.H. Werstiuk, B. Choudhury, J. Ma, J.
Warkentin, and J.P. Pezacki. Can. J. Chem. 76, 238 (1998).
. L. Wolff. Justus Liebigs Ann. Chem. 394, 23 (1912).
. D. Colbourne, D.C. Frost, C.A. McDowell, and N.P.C. West-
wood. J. Chem Soc. Chem. Commun. 250 (1980).
. A. Hotzel, R. Neidlein, R. Schultz, and A. Schweig. Angew.
Chem. Int. Ed. Engl. 19, 739 (1980).
7
8
9
28. L. Åsbrink, C. Fridh, and E. Lindholm. HAM/3, QCPE Bull.
13, 393 (1981).
1
0. S. Mohmand, T. Hirabayashi, and H. Bock. Chem. Ber. 114,
595 (1981).
1. D.P. Chong, N.P.C. Westwood, and S.R. Langhoff. J. Phys.
Chem. 88, 1479 (1984).
2. D. Colbourne and N.P.C. Westwood. J. Chem. Soc. Perkin
Trans. II, 2049 (1985).
3. R. Sammynaiken and N.P.C. Westwood. J. Chem. Soc. Perkin
Trans. II, 1987 (1989).
4. The chemistry of the diazonium and diazo groups. Parts 1 and
29. N.H. Werstiuk, G. Timmins, G. Ma, and T.A. Wildman. Can.
J. Chem. 70, 1971 (1992).
30. M. Regitz and F. Menz. Chem. Ber. 101, 2622 (1968).
31. A. Modelli, G. Innorta, and S. Torroni. J. Electron Spectrosc.
Relat. Phenom. 18, 359 (1980).
32. H. Hope and K.T. Black. Acta Crystallogr. Sect. B: Struct.
Crystallogr. Cryst. Chem. 28, 3632 (1972).
33. F. Kaplan and G.K. Meloy. J. Am. Chem. Soc. 88, 950 (1966).
34. H.M. Muchall, N.H. Werstiuk, and B. Choudhury. Can. J.
Chem. 76, 221 (1998).
2
1
1
1
1
1
1
1
1
1
2
. Edited by S. Patai. John Wiley and Sons, New York. 1978.
5. W. Kirmse. In Advances in carbene chemistry. Vol. 1. Edited
by U.H. Brinker. JAI Press Inc., Greenwich, Conn. 1994.
6. B.M. Trost and P.L. Kinson. J. Am. Chem. Soc. 97, 2438
35. T. Koopmans. Physica, 1, 104 (1933).
36. A.J. Arduengo, III, H. Bock, H. Chen, M. Denk, D.A. Dixon,
J.C. Green, W.A. Herrmann, N.L. Jones, M. Wagner, and R.
West. J. Am. Chem. Soc. 116, 6641 (1994).
37. H.M. Muchall and P. Rademacher. J. Mol. Struct. 435, 157
(1997).
38. N.H. Werstiuk, H.M. Muchall, C.D. Roy, J. Ma, and R.S.
Brown. Can. J. Chem. 76, 672 (1998).
39. K. Kimura, K. Katsumata, Y. Achiba, T. Yamazaki, and S.
Iwata. Handbook of HeI photoelectron spectra of fundamental
organic molecules. Japan Scientific Societies Press, Tokyo,
and Halsted Press, New York. 1981.
40. N.H. Werstiuk, J. Ma, M.A. McAllister, T.T. Tidwell, and D.-
C. Zhao. Trans. Faraday Soc. 90, 3383 (1994).
41. D. Chmielewski, N.H. Werstiuk, and T.A. Wildman. Can. J.
Chem. 71, 1741 (1993).
42. H.M. Muchall, N.H. Werstiuk, J. Ma, T.T. Tidwell, and K.
Sung. Can. J. Chem. 75, 1851 (1997).
43. A.S. Vorob’ev, I.I. Furlei, Yu.Z. Ekov, A.M. Nazarov, G.A.
Yamilova, U.M. Dzhemilev, and V.A. Dokichev. Russ. Chem.
Bull. 42, 281 (1993).
(
1975).
7. R.J. McMahon, O.L. Chapman, R.A. Hayes, T.C. Hess, and
H.-P. Krimmer. J. Am. Chem. Soc. 107, 597 (1985).
8. A. Allen, A.J. Kresge, N.P. Schepp, and T.T. Tidwell. Can. J.
Chem. 65, 1719 (1987).
9. R.J. McMahon, C.J. Abelt, O.L. Chapman, J.W. Johnson, C.L.
Kreil, J.-P. LeRoux, A.M. Mooring, and P.R. West. J. Am.
Chem. Soc. 109, 2456 (1987).
2
0. C. Bachman, T.Y. N’Guessan, F. Debû, M. Monnier, J.
Pourcin, J.-P. Aycard, and H. Bodot. J. Am. Chem. Soc. 112,
7
488 (1990).
1. J. Andraos and A.J. Kresge. J. Photochem. Photobiol. A, 57,
65 (1991).
2
2
2
2
2
1
2. HYPERCHEM 4.0. Hypercube Inc., 419 Philip Street,
Waterloo, Ont. 1994.
3. Dewar Research Group and J.J.P. Stewart. Austin Model 1
Package 1.0 (AMPAC). QCPE, 506 (1986).
4. J.J.P. Stewart. MOPAC, a semiempirical molecular orbital pro-
gram. QCPE, 455 (1983).
5. W.J. Hehre, L. Radom, P.v.R. Schleyer, and J.A. Pople. Ab in-
itio molecular orbital theory. John Wiley & Sons, New York.
44. H.M. Muchall, N.H. Werstiuk, J.L. Pitters, and M.S.
Workentin. Tetrahedron. In press.
1
986.
©
1998 NRC Canada