Conformational stability of 3-fiuoropropene in rare gas solutions from temperature-dependent FT-IR spectra and ab initio calculations

Article abstract of DOI:10.1021/jp9835162

The infrared spectra (3500-400 cm^-1) of 3-fluoropropene (allyl fluoride), CH₂=C(H)CH₂F, dissolved in liquid argon, krypton, and xenon have been recorded at various temperatures ranging from -180 to -65 °C. From these studies, the enthalpy difference between the more stable cis conformer and the high-energy gauche rotamer has been determined to range from 60 ± 8 cm^-1 (718 ±96 J/mol) in liquid xenon to 81 ±1 cm^-1 (969 ±12 J/mol) in liquid argon. These values have been extrapolated utilizing a linear relationship between the Kirkwood function of the solvent and the enthalpy differences in the solvents to give a value of 130 ± 25 cm^-1 (1.56 ±.30 kJ/mol) for the vapor. From the experimental enthalpy value, the gauche dihedral angle, torsional transitions for both retainers, and better structural parameters, the potential function governing the conformational interchange has been recalculated. Ab initio calculations utilizing the 6-31G(d,p) and 6-311G(d,p) basis sets with electron correlation at the MP2 level predict the cis conformer to be the more stable rotamer, but from the MP2/6-311 ++G(d,p) calculation the gauche conformer is predicted to be more stable by 117 cm^-1 (1.40 kJ/mol). By combination of the ab initio predictions of the structural parameters with the previously reported microwave rotational constants for 11 different isotopic species of both conformers, complete r_o parameters have been obtained for both rotamers. The results of these structural parameter determinations are compared to those previously reported.

Full text of DOI:10.1021/jp9835162

1976
J. Phys. Chem. A 1999, 103, 1976-1985
Conformational Stability of 3-Fluoropropene in Rare Gas Solutions from
Temperature-Dependent FT-IR Spectra and ab Initio Calculations
B. J. van der Veken and W. A. Herrebout
RijksuniVersitair Centrum Antwerpen, Laboratorium Voor Anorganische Scheikunde,
Groenenborgerlaan 171, 2020 Antwerpen, Belgium
D. T. Durig,† Wayne Zhao, and J. R. Durig*
Department of Chemistry, UniVersity of Missouri-Kansas City, Kansas City, Missouri 64110-2499
ReceiVed: August 25, 1998; In Final Form: January 8, 1999
The infrared spectra (3500-400 cm-1) of 3-fluoropropene (allyl fluoride), CH2dC(H)CH2F, dissolved in
liquid argon, krypton, and xenon have been recorded at various temperatures ranging from -180 to -65 °C.
From these studies, the enthalpy difference between the more stable cis conformer and the high-energy gauche
rotamer has been determined to range from 60 ( 8 cm-1 (718 ( 96 J/mol) in liquid xenon to 81 ( 1 cm-1
(969 ( 12 J/mol) in liquid argon. These values have been extrapolated utilizing a linear relationship between
the Kirkwood function of the solvent and the enthalpy differences in the solvents to give a value of 130 (
25 cm-1 (1.56 ( 0.30 kJ/mol) for the vapor. From the experimental enthalpy value, the gauche dihedral
angle, torsional transitions for both rotamers, and better structural parameters, the potential function governing
the conformational interchange has been recalculated. Ab initio calculations utilizing the 6-31G(d,p) and
6-311G(d,p) basis sets with electron correlation at the MP2 level predict the cis conformer to be the more
stable rotamer, but from the MP2/6-311++G(d,p) calculation the gauche conformer is predicted to be more
stable by 117 cm-1 (1.40 kJ/mol). By combination of the ab initio predictions of the structural parameters
with the previously reported microwave rotational constants for 11 different isotopic species of both conformers,
complete ro parameters have been obtained for both rotamers. The results of these structural parameter
determinations are compared to those previously reported.
TABLE 1: Experimental and Theoretical Values of the
Energy Difference (cm-1) between the Conformers of
3-Fluoropropene
Introduction
The 3-fluoropropene molecule has been the subject of several
spectroscopic studies for 2 decades, including infrared, Raman,
microwave, and NMR spectroscopies.1-11 A number of studies
have been devoted to its structure and thermodynamic stabilities.
It has been shown that 3-fluoropropene exists as a mixture of
cis and gauche rotamers with the cis conformer being the
thermodynamically preferred form in the gas phase at ambient
temperature. The enthalpy difference, ∆H, between the cis and
gauche conformers has been determined experimentally in
different phases by different techniques. From microwave
relative intensity measurements1 the enthalpy difference was
determined to be 58 ( 23 cm-1 (694 ( 275 J/mol) with the cis
rotamer the more stable form. In a subsequent microwave and
far-infrared spectroscopic study2 this value of the enthalpy was
used, along with the torsional transitions, to obtain the potential
function governing conformational interchange. A study4 of the
temperature-dependent Raman spectrum of the vapor gave an
enthalpy difference of 263 ( 25 cm-1 (3.15 ( 0.30 kJ/mol),
again with the cis conformer the more stable form. This value
was about 4 times larger than the value of 58 ( 10 cm-1 (694
( 120 J/mol) obtained for the liquid.4
method
state
∆H (∆E)
ref
Raman
microwave
microwave/far-infrared
Raman
infrared matrix
liquid
vapor
vapor
vapor
vapor
58 ( 10
58 ( 23
108
263 ( 25
211 ( 25
143
4
1
2
4
8
5
5
8
8
8
8
8
RHF/STO-3G
RHF/6-31G(d)
RHF/6-31G(d,p)
MP2/4-31G
148
171
216
MP4/4-31G//MP2/4-31G
MP2/6-31G(d,p)
MP4/6-31G(d,p)//MP2/6-31G(d,p)
276
256
266
assuming that the gas-phase equilibrium between the conformers
is trapped upon deposition. Therefore, the experimental value
of ∆H for the vapor phase ranges from 58 ( 23 to 263 ( 25
cm-1, and it is clear from these data that the experimental value
of ∆H for 3-fluoropropene is still relatively uncertain (Table
1). The energy difference between the two conformers of
3-fluoropropene has been obtained from ab initio calculations
at various levels of theory. The values obtained are also listed
in Table 1 and range from a low value of 143 cm-1 to a high
value of 276 cm-1, all with the cis conformer the more stable
rotamer.
More recently, the energy difference of 211 ( 25 cm-1 (2.52
( 0.30 kJ/mol) was obtained from a matrix isolation study8
* Corresponding author. Phone: 01 816-235-1136. Fax: 01 816-235-
5191. E-mail: durigj@umkc.edu.
Considering the well-known advantages of low-temperature
spectroscopy and the advantages of the spectroscopic studies
† Studies carried out while on sabbatical leave from the Departments of
Chemistry and Physics, The University of the South, Sewanee, TN 37383.
10.1021/jp9835162 CCC: $18.00 © 1999 American Chemical Society
Published on Web 03/12/1999

3-Fluoropropene
J. Phys. Chem. A, Vol. 103, No. 13, 1999 1977
Figure 2. Atomic numbering and internal coordinates of 3-fluoropro-
pene.
with boiling liquid nitrogen. Once the cell is cooled to the
desired temperature, a small amount of sample is condensed
into the cell. The cell is then pressurized with the noble gas,
allowing the compound to dissolve. For each temperature
investigated 200 interferograms were collected at 0.5 cm-1
resolution, averaged, and transformed with a Happ-Genzel
apodization function. The frequencies obtained for the funda-
mentals for the cis and gauche conformers in the three different
liquified noble gases are listed in Tables 2 and 3, respectively.
Figure 1. Infrared spectra of 3-fluoropropene in liquified noble
gases: (A) in Kr and (B) in Xe.
Computational Methodology
of cryosolutions over the conventional solid matrix-isolation
technique,11 we have carried out variable temperature infrared
spectroscopic studies of 3-fluoropropene in several liquified
noble gases. Additionally, as previously pointed out,10 there exist
artifacts in the experimentally determined structural data.
Therefore, ab initio calculations employing larger basis sets with
full electron correlation by the perturbation method12 to second-
order Moller-Plesset (MP2) have been performed, with par-
ticular attention to the ∆H values between the conformers along
with the predicted structural parameters. It was expected that
the present study could provide a more definitive value for ∆H
of the vapor and better insight into the structural parameters
for the two conformers of 3-fluoropropene. Both the theoretical
and experimental results are reported herein.
Complete geometric optimizations of the cis and gauche
conformers of 3-fluoropropene were carried out using the
Gaussian-94 program14 up to the MP2/6-311++G(d,p) level of
calculation. The atomic numbering for this molecule is shown
in Figure 2. The energy minima with respect to the nuclear
coordinations were obtained by the simultaneous relaxation of
all the geometric parameters using the gradient method of
Pulay.15 The energies for the potential surface were obtained
with optimization at 30° increments for fixed dihedral angles
between 0° (cis) and 180° (trans). The transition states were
characterized by one imaginary frequency. The vibrational
frequencies were calculated by the classic Wilson GF matrix
method16 with the different basis sets and levels. Subsequently,
the scaled quantum mechanical force field was used to correct
existing systematic errors for computed vibrational frequencies.
Infrared and Raman spectra were calculated using frequencies,
infrared intensities, and Raman scattering activities determined
from the MP2/6-311G(d,p) and RHF/6-311G(d,p) calculations,
respectively. Infrared intensities were calculated on the basis
of the dipole moment derivatives with respect to the Cartesian
coordinates. The derivatives were taken from the ab initio
calculations transformed to normal coordinates by
Experimental Method
The sample of 3-fluoropropene was prepared by the reaction
of 3-bromopropene with potassium fluoride in ethylene glycol
in the manner previously described.13 Purification was carried
out by a low-temperature, low-pressure fractionation column.
The purity of the sample was checked by comparing the mid-
infrared spectrum of the vapor with the previously reported
spectrum.4
The mid-infrared spectra (Figure 1) of the sample dissolved
in liquified noble gases (Ar, Kr, and Xe) as a function of
temperature were recorded on a Bruker model IFS-66 Fourier
transform interferometer equipped with a Globar source, a Ge/
KBr beam splitter, and a TGS detector. High-purity noble gases
(Matheson, 99.995%) were used typically in ca. (500-3000)-
fold excess. The temperature studies, ranging from -65 to -180
°C with an accuracy of (0.3 °C, were performed in a specially
designed cryostat cell that consisted of a 4 cm path length copper
cell with wedged silicon windows sealed to the cell with indium
gaskets and attached to a pressure manifold to allow for the
filling and evacuation of the cell. The temperature was
monitored with two Pt thermoresistors, while the cell was cooled
∂µu
∂Qi
∂µu
∂Xj
)
Lij
∑
(
)
(
)
j
where the Qi is the ith normal coordinate, Xj is the jth Cartesian
displacement coordinate, and Lij is the transformation matrix
between the Cartesian displacement coordinates and normal
coordinates. The infrared intensities were calculated by
2
2
2
∂µx
∂Qi
∂µy
∂Qi
∂µz
∂Qi
Nπ
Ii )
+
+
2
(
[
)
(
)
(
)
]
3c
The Raman scattering cross section, ∂σj/∂Ω, which is
proportional to the Raman intensity, can be calculated from the

1978 J. Phys. Chem. A, Vol. 103, No. 13, 1999
van der Veken et al.
TABLE 2: Observed and Calculated Vibrational Frequencies (cm-1) for cis-3-Fluoropropene
ab initio
observed liquid
MP2/6- MP2/6-
noble gasese
311G- 311++G- fixed
IR
Raman
species νi
description
(d,p)
(d,p)
scaleda intensb act.c gasd Ar
Kr
54.9 3114 3111 3108 3104 3096 3108 99S1
5.4 154.8 3032 3029 3026 3023 3029 3026 37S2, 62S3
6.8 30.9 2998 3005 3002 2998 2994 2998 62S2, 37S3
36.9 156.5 2938 2941 2937 2933 2948 2945 100S4
Xe liqd solidd
PED
A′
ν1 dCH2 antisymmetric stretch 3307
3302
3206
3192
3091
1706
1520
1460
1436
1313
1145
1015
927
3138
3046
3032
2928
1625
1452
1386
1369
1246
1100
968
4.2
ν2 dCH2 symmetric stretch
ν3 dCH stretch
ν4 -CH2 symmetric stretch
ν5 CdC stretch
ν6 -CH2 deformation
ν7 dCH2 deformation
ν8 -CH2 wagging
ν9 dCH in-plane bending
ν10 CF stretch
ν11 dCH2 wagging
ν12 C-C stretch
3211
3196
3087
1713
1531
1461
1443
1313
1159
1020
932
2.4
3.1
11.6
10.1
0.1
52.2
28.8
2.3
37.0 1653 1654 1654 1654 1651 1649 69S5, 12S7, 8S9, 7S12
12.5 1468 1470 1469 1467 1463 1461 95S6, 5S8
7.6 1417 1414 1412 1412 1417 1416 57S7, 19S8, 11S12
4.1 1389 1385 1384 1383 1387 1385 65S8, 25S7
22.3 1292
1292 1284 60S9, 15S10, 16S5
2.3 1118 1115 1114 1112 1108 1109 58S10, 10S10, 10S12, 6S13
5.0 971 971 970 970 973 976 53S11, 15S9, 28S11
7.2 905 903 902 901 903 900 66S12, 8S5, 6S14
1.8 604 604 604 603 602 604 44S13, 15S10, 31S14
1.4 268
88.5 2960 2965 2962 2958 2960 2959 100S15
7.3 1253 1256 1244 1246 91S16, 8S19
2.9 1032 1023 1022 1020 1020 1024 77S17, 6S16, 14S17
0.1 996 993 992 990 997 997 57S18, 46S20
2.9 928 926 925 923
7.2 549 560 550 549 555 567 60S20, 29S17, 12S19
4.2 164 222 100S21
884
ν13 C-C-F bending
ν14 CdC-C bending
614
611
582
5.8
2.7
277
271
263
272 284 55S14, 42S13
A′′
ν15 -CH2 antisymmetric stretch 3139
3143
1283
1050
1015
888
2978
1227
1003
974
30.6
0.2
15.0
16.6
38.3
9.1
ν16 -CH2 twisting
ν17 -CH2 rocking
1293
1058
1026
924
ν18 dCH out-of-plane bending
ν19 dCH2 out-of-plane bending
ν20 dCH2 twisting
877
928 100S19
556
546
527
ν21 asymmetric torsion
173
149
173
2.9
a Calculated using the MP2/6-311G(d,p) level and scaled with scaling factors of 0.9 for stretches and bends and 1.0 for the torsion. b Infrared
calculated intensities in km mol-1 from MP2/6-311G(d,p) calculation. c Raman scattering activity in Å4 u-1 from RHF/6-311G(d,p) calculation.
d Reference 5. e This study.
TABLE 3: Observed and Calculated Vibrational Frequencies (cm-1) for gauche-3-Fluoropropene
ab initio
observed liquified
MP2/6-
MP2/6- 311++ fixed
noble gasesf
IR Raman
species νi no.a
description
311G** G** scaledb intensc act.d gase Ar
Kr
62.8 3100 3097 3094 3090 3096 97S1
6.0 110.6 3008 3034 3030 3027 3015 93S2
Xe liqe
PED
A
ν1
ν2
ν3
1
3
2
-CH2 antisymmetric stretch 3287
dCH stretch
dCH2 symmetric stretch
3285
3212
3184
3165
3102
1701
1522
1475
1414
1315
1291
1198
3119
3048
3023
2997
2937
1621
1450
1403
1350
1246
1231
1136
10.2
3213
3186
3159
3096
1709
1528
1479
1423
1313
1298
1197
7.2
26.3
70.2
63.0
2995 2992 2989
2965 2962 2958 2960 97S15
99S4
95S3
ν4 15 CH2 antisymmetric stretch
ν5
ν6
ν7
ν8
ν9
ν10
4
5
6
7
8
9
CH2 symmetric stretch
CdC stretch
35.9 116.8 2938 2901 2898 2894
0.4
1.2
36.5 1646
5.5
1630 1628
1462 1460 1458
7.4 1432 1429 1428 1425 1428 58S7, 26S8, 8S9, 6S12
5.8 1368 1364 1363 1362 1367 68S8, 23S7
68S5, 13S7, 8S9
98S6
-CH2 deformation
dCH2 deformation
-CH2 wagging
26.4
11.3
0.2
dCH in-plane bending
17.1
11.7 1243
1242 1240 1240
59S9, 15S5, 13S10
1244 83S16
ν11 16 -CH2 twist
ν12 12 C-C stretch
3.6
5.5
1.8 1160 1098 1097 1096
25S12, 11S9, 18S11,
11S13, 30S19
ν13 10 C-F stretch
1089
1029
1003
939
934
657
438
334
114
1057
1019
999
937
918
656
435
339
115
1033 104.7
5.2
3.4
2.1
2.2
3.3
1022 1022 1021
993 992 990
918 917
90S10
77S18, 22S20
918 25S12, 39S17, 16S20
16S11, 77S19
ν14 18 dCH out-of-plane bending
ν15 17 -CH2 rocking
976
951
891
886
622
416
317
114
36.3
7.0
30.4
6.8
6.9
2.4
ν16 19 dCH2 out-of-plane bending
ν17 11 dCH2 wagging
ν18 20 dCH2 twist
ν19 13 CdC-C bending
ν20 14 C-C-F bending
ν21 21 asymmetric torsion
936 935 934
39S11, 12S12, 29S19
4.7 641 640 640 638 644 14S12, 16S18, 55S20
4.3 422
2.7 332
6.5 108
438 51S14, 20S13
49S13, 32S14, 6S17, 9S20
100S21
6.5
1.1
a Vibrational quantum number of the equivalent vibration for the cis conformer. b Calculated using the MP2/6-311G(d,p) level and scaled with
scaling factors of 0.9 for stretches and bends and 1.0 for the torsion. c Infrared calculated intensities in km mol-1 from MP2/6-311G(d,p) level.
d Raman Scattering activity in Å4 u-1 from RHF/6-311G(d,p) level. e Reference 5. f This study.
scattering activities and the predicted frequencies for each
normal mode using the relationship18-21
cross sections, the polarizabilities are incorporated into Sj by
Sj[(1 - Fj)/(1 + Fj)] where Fj is the depolarization ratio of the
jth normal mode. The Raman scattering cross sections and
calculated frequencies, together with a Lorentzian function, were
used to obtain the calculated spectrum. Since the calculated
frequencies are approximately 10% higher than those observed,
the scaled quantum mechanical force fields were employed for
corrections. The simulated infrared and Raman spectra for each
conformer and their mixtures are shown in Figures 3 and 4,
respectively. These spectra can be compared to the experimental
(ν0 - νj)4
-hcνj
24π4
h
∂σj
∂Ω
)
Sj
2
45
(
)
(
)
8π cνj
(
)
1 - exp
[
]
kT
where ν0 is the exciting frequency, νj is the vibrational frequency
of the jth normal mode, and Sj is the corresponding Raman
scattering activity. To obtain the polarized Raman scattering

3-Fluoropropene
J. Phys. Chem. A, Vol. 103, No. 13, 1999 1979
Figure 3. Calculated infrared spectra of 3-fluoropropene: (A) mixture
of the cis and gauche conformers; (B) pure gauche; (C) pure cis.
spectra, and they provide support for the assignment of the
observed bands to the indicated fundamentals, as well as for
the enthalpy difference between the two conformers.
Structural Parameters
The optimized structural parameters and rotational constants
with different basis sets for each conformer are listed in Tables
4 and 5, respectively, along with previously reported rs and
refined ro structures. In comparing the computed results with
experimental values, it is clear that a significant improvement
has been reached when the ab initio calculations were performed
with the larger basis sets at the MP2 level. The computed
structural parameters at the MP2 levels give more realistic
results, particularly for the CdC, C-F, and C-H bond distances
than the RHF calculations. The CdC bond length calculated
with the largest basis set at the MP2 level, i.e., MP2/6-311++G-
(d,p), are 1.3376 and 1.3393 for the cis and gauche conformers,
respectively, which are about 0.005 Å longer than the rs and ro
distances for these parameters even though they are re values.
The remaining calculated parameters are in quite good agree-
ment with the reported ro parameters. The C-C bond (ro) in
the cis form is slightly longer by 0.005 Å than the corresponding
distance in the gauche form that is predicted by the ab initio
calculations. These results are sharply different from the
previously reported1 rs values in which the gauche form was
predicted to have two C-H distances more than 0.029 Å longer
than the corresponding distance for the cis form, although the
uncertainties for these parameters for the gauche conformer were
nearly this large. For the C-H bond distances, the present results
indicate that the C-H bonds on the vinyl group have similar
bond lengths, which was not true for the rs parameters. In
general, calculated carbon-hydrogen bond lengths for both
conformers have approximately the same values. The difference
between them is less than 0.001 Å, except for the C1-H5 bond,
which is predicted to be 0.0025 Å longer in the gauche form.
Comparing the calculated C-F bond length to the corresponding
ro values, this bond in the gauche form deviates 0.005 Å from
the ro structure. It is probable that the C-F distances should
differ between the two conformers by 0.008 Å, which suggests
that the ro C-F distance in the gauche form should be longer.
A striking difference for the structural parameters between
conformers is found for both the C-C-F and CdC-C
angles. Ab initio calculations predict that both angles in the cis
Figure 4. Comparison of experimental and calculated Raman spectra
of 3-fluoropropene: (A) gas; (B) solid; (C) calculated mixture of cis
and gauche conformers; (D) calculated pure cis form.
form are ca. 1.7° larger than the corresponding ones in the
gauche form. The ro values differ by only about 1.0°, but the
larger ro values for these two parameters for the gauche form
compared to the ab initio values undoubtedly depends on the
dihedral angle for the gauche conformer as well as on the C-F
distance. In conclusion, the ab initio predicted parameters show
little difference between conformers and give the closest
agreement between the ro values and ab initio structures from
the MP2/6-311G(d,p) and MP2/6-311++G(d,p) calculations,
with the former ones in better agreement with the microwave
rotational constants.
Conformational Stability
With the availability of the liquified noble gases technique,
the enthalpy difference between the two conformers of 3-fluo-
ropropene was determined. The mid-infrared spectra of the
sample dissolved in liquified krypton and xenon are shown in
Figure 1. Because only small interactions are expected to occur
between the dissolved molecules and the surrounding noble gas
atoms,11,22-24 these “pseudo-gas-phase” spectra show only small

1980 J. Phys. Chem. A, Vol. 103, No. 13, 1999
van der Veken et al.
TABLE 4: Comparison of Computed and Experimental Structural Parameters for cis-3-Fluoropropenea
MP2/6-31 MP2/6-31
G(d)
RHF/
MP2/
RHF/
MP2/
microwave
G(d,p)b 6-311G(d,p) 6-311G(d,p) 6-311++G(d,p) 6-311++G(d,p)
rsc
r0
adjustede
d
parameter
r(C1dC2)
r(C2-C3)
r(C3-F)
r(C1-H5)
r(C1-H6)
r(C2-H7)
r(C3-H8)
r(C3-H9)
C1C2C3
C2C3F
C2C1H5
C2C1H6
H5C1H6
C1C2H7
C2C3H8
C2C3H9
H8C3H9
H8C3F
1.334
1.494
1.393
1.083
1.084
1.089
1.096
1.096
123.7
111.0
121.1
121.3
117.6
120.9
111.0
111.0
108.0
107.9
107.9
0.0
1.334
1.493
1.392
1.078
1.079
1.083
1.092
1.092
123.5
111.0
120.8
121.2
118.0
120.9
110.8
110.8
108.6
108.0
108.0
0.0
1.315
1.496
1.365
1.0746
1.0754
1.0792
1.0848
1.0848
125.0
111.5
121.5
121.0
117.5
120.5
110.6
110.6
108.5
107.8
107.8
0.0
1.3357
1.4960
1.3833
1.0832
1.0837
1.0883
1.0951
1.0951
123.9
111.6
120.8
121.0
118.2
120.7
110.4
110.4
108.5
108.0
108.0
0.0
1.3171
1.4965
1.3657
1.0749
1.0754
1.0794
1.0848
1.0848
125.4
111.6
121.6
120.9
117.5
120.4
110.7
110.7
108.5
107.6
107.6
0.0
1.3378
1.4955
1.3886
1.0837
1.0839
1.0886
1.0948
1.0948
124.5
111.5
121.1
120.7
118.2
120.5
110.7
110.7
108.6
107.6
107.6
0.0
1.333 ( 0.007 1.333 ( 0.003
1.503 ( 0.010 1.495 ( 0.004
1.382 ( 0.010 1.388 ( 0.004
1.077 ( 0.007 1.083 ( 0.002
1.106 ( 0.007 1.084 (fixed)
1.090 assumed 1.089 (fixed)
1.098 ( 0.007 1.098 ( 0.002
1.098 ( 0.007 1.098 ( 0.002
1.333
1.494
1.390
1.087
1.087
1.092
1.100
1.100
123.6
112.4
120.7
119.6
119.7
122.6
111.7
111.7
107.4
106.7
106.7
0.0
124.6 ( 0.8
111.7 ( 0.7
120.9 ( 0.5
119.2 ( 0.5
119.9 ( 0.5
124.5 ( 0.3
111.5 ( 0.3
121.1 ( 0.2
121.3 ( 0.2
117.6 ( 0.2
119.0 assumed 120.9 ( 0.2
111.1 ( 0.5
111.1 ( 0.5
108.1 ( 0.5
107.4 ( 0.5
107.4 ( 0.5
0.0
111.1 ( 0.3
111.1 ( 0.3
108.2 ( 0.2
107.9 ( 0.2
107.9 ( 0.2
0.0
H9C3F
dihedral
angle, æ
A
B
C
17020
6092
4616
17107
6133
4622
17218
6043
4602
17818
6002
4618
17210
5997
4576
17400
5941
4556
17234
6007
4582
17238
6003
4580
-(E + 215)
1.524771
0.978871
1.684670
0.983840
1.694874
(hartree)
∆E (cm-1
)
-174
-191
-19
117
a Bond lengths in angstroms; angles in degrees; rotational constants in MHz. b Reference 8. c Reference 1; experimental rotational constants are
17236.6, 6002.9, and 4578.8 MHz for A, B, and C, respectively, for the normal species. d Reference 10. e Total of 22 independent parameters (12
for cis, 17 for gauche, including 7 overlapping) from MP2/6-311++G(d,p) structures are adjusted to fit all rotational constants of 19 isotopic
species from ref 1. The maximum error for rotational constants is 0.06%.
TABLE 5: Comparison of Computed and Experimental Structural Parameters for gauche-3-Fluoropropenea
MP2/6-
MP2/6-
RHF/6-
MP2/6-
RHF/6-
MP2/6-
microwave
adjustedf
31G(d) 31G(d,p)b 311G(d,p) 311G(d,p) 311++G(d,p) 311 ++G(d,p)c
rsd
r0
e
parameter
r(C1dC2)
r(C2-C3)
r(C3-F)
r(C1-H5)
r(C1-H6)
r(C2-H7)
r(C3-H8)
r(C3-H9)
C1C2C3
1.336
1.491
1.401
1.086
1.084
1.088
1.095
1.095
123.3
109.6
121.5
121.8
116.7
120.7
111.5
111.2
108.9
107.1
108.4
127.2
1.335
1.490
1.399
1.081
1.079
1.083
1.090
1.091
123.3
109.6
121.3
121.7
117.0
120.6
111.0
111.5
109.4
107.3
108.5
126.0
1.3164
1.4956
1.3713
1.0775
1.0756
1.0782
1.0841
1.0823
124.0
109.9
121.8
121.8
116.7
120.5
110.9
111.4
109.3
107.0
108.1
128.6
1.3373
1.4931
1.3911
1.0862
1.0842
1.0873
1.0942
1.0934
123.1
110.1
121.1
121.6
117.3
120.6
110.9
110.7
109.3
107.2
108.4
125.8
1.3182
1.4958
1.3731
1.0775
1.0756
1.0783
1.0840
1.0823
123.8
110.0
121.8
121.5
116.7
120.5
111.1
111.6
109.3
106.8
108.0
126.7
1.3393
1.4922
1.3987
1.0863
1.0844
1.0875
1.0937
1.0930
122.8
109.9
121.1
121.5
117.4
120.7
111.4
111.1
109.5
106.8
107.9
123.4
1.354 ( 0.015 1.335 ( 0.003
1.486 ( 0.015 1.490 ( 0.004
1.371 ( 0.015 1.394 ( 0.004
1.098 ( 0.015 1.086 ( 0.008
1.054 ( 0.015 1.084 ( 0.006
1.090 assumed 1.088 ( 0.020
1.137 ( 0.020 1.095 ( 0.015
1.127 ( 0.020 1.095 ( 0.008
1.334
1.491
1.400
1.089
1.087
1.091
1.099
1.098
123.8
109.6
119.7
120.5
119.8
122.7
109.0
110.8
110.6
108.4
108.4
125.7
121.6 ( 1.0
110.9 ( 1.0
119.2 ( 1.0
121.5 ( 1.0
119.3 ( 1.0
123.3 ( 0.5
110.4 ( 0.5
121.5 (fixed)
121.8 (fixed)
116.7 (fixed)
C2C3F
C2C1H5
C2C1H6
H5C1H6
C1C2H7
C2C3H8
C2C3H9
H8C3H9
H8C3F
119.0 assumed 120.7 (fixed)
107.4 ( 1.5
105.2 ( 1.5
111.4 ( 1.5
107.1 ( 1.5
114.7 ( 1.5
127.1 ( 3.0
111.5 (fixed)
111.2 (fixed)
108.9 (fixed)
107.1 (fixed)
108.4 (fixed)
124.6 ( 0.5
H9C3F
dihedral
angle, æ
A 27720.35c 27959
27496
4248
4112
27907
4260
4126
28630
4311
4166
27241
4267
4152
28036
4288
4149
27720.3
4263.8
4132.1
27714
4264
4132
B 4263.63c
C 4131.98c
-(E + 215)
(hartree)
4255
4107
0.978077
1.683802
0.983755
1.695406
a Bond lengths in angstroms; angles in degrees; rotational constants in MHz. b Reference 8. c Additional parameters: τ(H5C1C2C3) ) -1.6,
τ(H6C1C2H5) ) 179.8, τ(H7C2C1C3) ) 181.1. d Reference 1; experimental rotational constants for the normal species are A ) 27 720.4, B ) 4263.6,
and C ) 4132.0 MHz. e Reference 10. f Additional parameters: τ(H5C1C2C3) ) -3.5 ( 1.9, τ(H6C1C2H5) ) 179.4 ( 0.4, τ(H7C2C1C3) ) 180.7 (
0.3.
frequency shifts compared with the frequencies of similar bands
in the corresponding spectra in the gas phase. Also, small
frequency shifts can be observed in different rare gas solutions
(Figure 1).
Although the region near 3000 cm-1 in liquid Ar contains a
rather well-separated doublet, the presence of bands on both
sides makes the band profile analysis difficult. At the higher
temperature in both liquid Kr and Xe, the half-widths of the

3-Fluoropropene
J. Phys. Chem. A, Vol. 103, No. 13, 1999 1981
Figure 5. Temperature dependence of the infrared spectra in the region 950-900 cm-1: (A) Ar; (B) Kr; (C) Xe.
bands prohibit a useful analysis. A similar remark must be made
for the CdC stretching region where only at the lowest
temperature in liquid Ar can a shoulder be clearly distinguished.
Therefore, no band profile analysis has been performed for these
regions. For the doublet at 1462/1470 cm-1 there is a shoulder
near 1477 cm-1, which causes problems in obtaining peak areas
at higher temperatures. Similarly, for the 1430/1415 cm-1
doublet, a weak band on the low-frequency side of the high-
frequency component causes difficulty in obtaining band areas.
Therefore, the only doublet where little difficulty was encoun-
tered from nearby weak transitions is the one near 930 cm-1
.
The intensities of the conformational doublet, observed at
934 and 923 cm-1 (Figure 5) and assigned to the dCH2 out-
of-plane bend of the gauche and cis conformers, respectively,
were obtained as a function of temperature and their ratio Ig/Ic
measured. The curve-fitting procedure to obtain the areas under
the bands has been used to obtain reliable intensity ratios. By
application of the van’t Hoff equation,
Figure 6. Plot of -ln(I934/I923) vs 1000/T.
measured in that solvent.25,26 This model has been applied to
the allyl fluoride data obtained herein, and the graph resulting
from this plot is shown in Figure 7. In this figure, also the least-
squares straight line through the experimental points is given.
The linear regression of Figure 7 can be used to obtain a
value for the enthalpy difference for the vapor phase. For the
latter, the relative permittivity κ is approximated to be 1 so that
the vapor-phase ∆H is obtained as the intercept of the straight
line. It is clear from Figure 7 that this extrapolation must be
carried out over a relatively large distance, but the high degree
of linearity of the experimental points ensures that a reliable
value for the vapor-phase enthalpy difference is obtained. The
value of the intercept obtained is 130 ( 25 cm-1, i.e.,
somewhere in the middle of the range of the available
experimental values but lower than the ab initio values where
the cis conformer is predicted to be the more stable form and
lower than the Raman value for the vapor phase. We believe
this is a reasonable value to use for the determination of the
potential function parameters governing conformer interchange.
-ln(Ig/Ic) ) (∆H/(RT)) - ln(Rg/Rc) - (∆S/R)
where ∆S is the entropy change and ∆H can be derived from
the slope of a plot of -ln(Ig/Ic) versus 1/T. It is assumed that
∆H, ∆S, and the quantity ln(Rg/Rc) are independent of the
temperature over the experimental temperature range.
The van’t Hoff plots are shown in Figure 6, and the plots are
quite linear for the Kr and Ar solutions, whereas that for the
Xe solution is not as linear. The ∆H values of 81 ( 1, 72 ( 4,
and 60 ( 8 cm-1 have been determined from the Ar, Kr, and
Xe solutions, respectively. These values are slightly higher than
the value of 58 ( 23 cm-1 determined by Hirota1 but
considerably lower than the value of 263 ( 25 cm-1, which
was obtained from the variable temperature studies of the Raman
spectrum of the vapor.4 The value is also much smaller than
the value of 211 ( 25 cm-1 obtained from the infrared spectrum
of the matrix-isolated material.8 It should be noted that all
experimental results indicate that the cis conformer is the more
stable form in all three physical states.
Torsional Potential Function
By use of a reaction field approximation, it has been shown
that there is a linear relationship between the Kirkwood function
(κ - 1)/(2κ + 1) of the solvent and the enthalpy difference
The computed energies for the conformational isomers of
3-fluoropropene from the various basis sets are collected in
Table 6. From a comparison of the energy difference between

1982 J. Phys. Chem. A, Vol. 103, No. 13, 1999
van der Veken et al.
optimized structure were input into the B matrix program
together with the complete set of 24 internal coordinates (Figure
2). This set of internal coordinates was used to form the
symmetry coordinates with three redundancies.6 The symmetry
coordinates were used to determine the corresponding potential
energy distributions (PED). The resulting B matrix converts the
ab initio force field in Cartesian coordinates to a force field in
the desired internal coordinate. Finally, scaling factors of 0.9
for the stretching and bending modes, 1.0 for torsions, and the
geometric average of scaling factors for the interaction force
constants were used to obtain the fixed scaled force field and
resultant frequencies, along with the potential energy distribu-
tions (PED), which are summarized in Tables 2 and 3.
The predicted vibrational fundamentals obtained from the
MP2/6-311G(d,p) and MP2/6-311++G(d,p) calculations give
almost the same wavenumbers, and for most modes, the
differences are less than 10 cm-1. The assignments given in
Tables 2 and 3 have been made on the basis of the characteristic
group and calculated frequencies, the ab initio predicted
intensities for the infrared and Raman spectra, along with
support from changes in relative intensities of doublets obtained
from temperature-dependent infrared spectra of the rare gas
solutions.
Figure 7. Enthalpy difference for the conformational equilibria of
3-fluoropropene as a function of the Kirkwood function. The experi-
mental values for the different rare gas solutions are represented by
the dots. The straight line is the result of the linear regression on the
experimental points.
There is a problem in the assignment of the carbon-hydrogen,
-CH2(F), antisymmetric and symmetric stretches for the cis
conformer, since there are three bands at 2888, 2943, and 2963
cm-1 in the Raman spectrum of the solid with similar bands in
the corresponding infrared spectrum. Clearly, one of these bands
is due to Fermi resonance, but it is not clear whether the doublet
is the 2888/2943 cm-1 bands or the 2943/2963 cm-1 bands. In
our earlier study5 we chose the 2888/2943 cm-1 bands as the
Fermi doublet with the assumption that it was due to the
overtone of the CH2 deformation (1455 cm-1). Therefore, the
CH2 symmetric stretch was assigned to the 2943 cm-1 Raman
line, since this is the strongest Raman line of the three under
consideration. This assignment is consistent with the predictions
of the intensity of this line from the ab initio calculations. The
CH2 antisymmetric stretch is then assigned to the 2963 cm-1
band, which is in agreement with ab initio predictions as well
as with earlier proposed vibrational assignments.5,8
the conformers, one finds that these differences decrease with
the increase in the basis set size and electron correlation. When
the 6-311G(d,p) basis set was used, the enthalpy differences
are determined to be 174 and 192 cm-1 at the RHF and MP2
levels, respectively, favoring the cis conformer as being the more
stable rotamer. Furthermore, calculations at the RHF/6-311++G-
(d,p) level indicate that the energy separation reduces to 19
cm-1. Surprisingly, with electron correlation with this basis set,
the calculation reverses the energy ordering and predicts the
gauche conformer to be more stable than the cis form (Table
6) by 117 cm-1. The experimentally determined ∆H value
clearly shows this prediction to be in error. It is probable that
the MP2/6-311G(d,p) calculation gives better energy differences
between the conformers whereas the MP2/6-311++G(d,p)
calculation gives better structural parameters.
On the basis of our ab initio values, the torsional potential
energy curve can be constructed by a best-fit Fourier cosine
series as a function of torsional angle:
Another area where the assignments are questionable is the
region from 900 to 1100 cm-1. The ab initio calculations
indicate that the band near 1100 cm-1 is the C-F stretch for
the cis conformer, which should give rise to a very strong
infrared band as found at 1118 cm-1. The corresponding gauche
mode in this region (1160 cm-1) is a very mixed mode with
the major contribution (25%) from the C-C stretch, which
leaves the band at 1099 cm-1 as a combination or overtone
mode. The remaining bands in this region, which are mostly
doublets, were assigned on the basis of their relative changes
in intensity with changes in temperature of the rare gas solutions.
6
1
V(φ) )
V (1 - cos iφ)
i
∑
2i)1
The calculated potential curves at various levels of theory are
depicted in Figure 8. The potential coefficients and barriers
governing internal rotation of the -CH2F moiety against a
molecular frame are summarized in Table 6. The ab initio
predictions, with the exception of the MP2/6-311++G(d,p)
calculation, give results that are in reasonable agreement with
the experimental results regarding the existence and relative
stabilities of the cis and gauche conformers. Moreover, the
predominance of the calculations indicates that the cis conformer
is the thermodynamically preferred rotamer.
The low-frequency skeletal deformations, i.e., the CdC-C
and C-C-F bending modes, involve four heavy atoms with
nearly equal atomic weight. This results in a high degree of
mixing of these skeleton vibrational modes. An interesting
feature is that the higher frequency band at 604 cm-1 for the
cis rotamer is more CCF bend at 44% compared to 31% for
the CCC bend, whereas for the gauche conformer the higher
frequency mode at 422 cm-1 is 51% CCC bend and only 20%
CCF bend. The lower frequency bend at 332 cm-1 for the
gauche rotamer is made up of 49% CCF bend and 32% CCC
bend. Part of this difference is due to the lower symmetry of
the gauche conformer, which results, in general, in a significant
increase in the mixing of the modes.
Vibrational Analysis
To obtain a more complete description of the molecular
motions involved in the fundamental modes of this molecule, a
normal coordinate analysis was carried out. This analysis was
performed utilizing ab initio calculations and the Wilson GF
matrix method.16 The Cartesian coordinates taken from the

3-Fluoropropene
J. Phys. Chem. A, Vol. 103, No. 13, 1999 1983
TABLE 6: Barriers to Internal Rotation (cm-1), Enthalpy Difference (cm-1), and Dihedral Angles (deg) for 3-Fluoropropenea
MP2/6-
RHF/6-
MP2/6-
RHF/6-
MP2/6-
parameter
31G(d,p)a
311G(d,p)b
311G(d,p)b
311++G(d,p)b
311++G(d,p)b
ref 6
36 ( 9
296 ( 10
826 ( 15
47 ( 7
14 ( 8
-32 ( 5
1098
this studyc
V1
V2
V3
V4
V5
V6
20.9
325.6
841.0
3.7
-153.0
390.4
885.3
17.0
-15.1
242.0
839.9
15.2
-263.9
266.8
893.7
28.0
-174.6
15.1
800.9
-26.4
17.7
-45.7
759
-162 ( 9
333 ( 12
860 ( 5
65 ( 2
-13 ( 5
-35 ( 2
1117
5.7
28.5
37.8
26.3
-7.4
1103
612
-45.4
1162
585
-56.1
1038
670
-42.5
1055
636
cis-to-gauche barrier
gauche-to-gauche barrier
gauche-to-cis barrier
energy (enthalpy)
difference
torsional dihedral angle
(gauche)
769
588
526
847
256
988
174
846
192
1036
19
876
-117
810
263 ( 25
958
130 ( 25
126.0
128.6
125.8
126.7
123.1
124.3 ( 0.5
124.6
a Data taken from refs 8 and 16. b This study. c Calculated from the observed far-infrared transitions (Table 8) and the following kinetic terms:
F0 ) 2.65251, F1 ) 0.51078, F2 ) 0.22400, F3 ) 0.05046, F4 ) 0.01501, F5 ) 0.00386, F6 ) 0.00109, F7 ) 0.00031, and F8 ) 0.00008 cm-1
.
angles and torsional angles in the same set will keep their
differences in degrees. If one CH bond is 1% longer than another
CH bond by ab initio calculation, it will still be 1% longer after
the optimization. If a CCH angle is 1° larger than another CCH
angle from the ab initio calculation, it will still be 1° larger in
the final result. This assumption is based on the fact that the
errors from ab initio calculations are systematic, which is now
commonly accepted, and this technique is currently used in our
normal coordinate analysis. The program searches the minima
of the function F(k1,k2,...):
F(k1,k2,...) ) (100K )2 + (20K )2 + (0.1K )2 +
∑
∑
∑
l
i
j
l
i
j
(0.02K )2
∑
m
m
Figure 8. Potential function governing the internal rotation around
the C-C bond determined by ab initio calculations: MP2/6-311++G-
(d,p); MP2-6-311G(d,p); RHF/6-311++G(d,p); RHF/6-311G(d,p).
The adjustment factors K in the formula above are defined as
Rci - Roi
Ki )
Roi
Discussion
On the basis of the fact that the calculated structural
parameters with the larger basis sets with electron correlation
are in excellent agreement with the experimentally determined
ro parameters clearly indicates that the ab initio calculations
with such basis sets can be used to obtain the structural
parameters for these types of molecules. The major difference
in the structural parameters, in addition to the two CCC and
CCF angles between the gauche and cis conformers, is the C-F
distance. We have attempted to fit the microwave rotational data
by maintaining the difference between the C-F distances that
were predicted from the ab initio calculations, and the results
are very satisfactory.
A new computer program, AandM (ab initio and microwave),
has been developed. This program combines the information
from microwave and ab initio calculations and gives structural
parameters that fit the rotational constants with the structural
parameters remaining close to the ab initio values. To reduce
the number of independent variables, the structural parameters
are separated into several sets according to their type. For
example, three CH bond lengths may form one set and three
CCH angles may form another set. Each set uses only one
independent parameter in the optimization, and all structural
parameters in one set will be adjusted by the same parameter,
i.e., an adjustment factor. The information on the differences
between the similar parameters from ab initio calculations will
be retained in the final results by the following method: bond
lengths in the same set will keep their relative ratio and bond
Lcj - Laj
Kj )
Laj
Kl ) Acl - Aal
Km ) Tcm - Tam
where R, L, A, and T represent the rotational constants, bond
lengths, bond angles, and torsional angles, respectively. The
lower case letters c, o, and a indicate calculated (calculated here
means calculated by the AandM program, not by the ab initio
program), observed (rotational constants), and ab initio (bond
lengths) bond angles and torsional angles in degrees. Thus, the
symbols Tc and Ta indicate the torsional angles calculated from
the program and the torsional angles obtained from the ab initio
predictions, respectively. The subscript runs over all structural
parameters in the optimization. The real meaning of these
formulas is the following: 1% error of any rotational constant
contributes 1.0 to the function F, which is equivalent to a 5%
shift of a bond length from its ab initio value, a 10° shift of
bond angle from its ab initio value, or a 50° shift of a torsional
angle from its ab initio value. Only real torsional angles around
single bonds are considered as torsional angles in the calcula-
tions, whereas other angles defined by the third internal
coordinate in the ab initio input data are treated as bond angles
because they are not as flexible as real torsion angles around

1984 J. Phys. Chem. A, Vol. 103, No. 13, 1999
van der Veken et al.
single bonds. To avoid the possibility of falling into a local
minimum instead of the global minimum of F, the simplex
algorithm instead of the gradient method was used to optimize
the adjustment factors in searching for the minimum of F.
the value of 263 ( 23 cm-1 by a variable-temperature-dependent
experiment of the Raman spectrum of the vapor.4 The lower
value of 58 cm-1 for the liquid is attributed to dipole-dipole
interactions in this phase, whereas the larger value of 263 (
23 cm-1 determined from the gas probably reflects the
experimental uncertainties arising from intensity measurements
in the vapor by the Raman technique. First, there is the problem
of measuring the temperature of the gas where the cell must be
hotter on the side where the hot nitrogen gas initially contacts
the cell compared to the other side. Also, intensity measurements
are difficult to make for the Raman spectrum and, additionally,
the heights of the bands were previously used rather than areas
under the curve. Finally, it should be noted that there is another
fundamental at 928 cm-1 for the cis conformer that is close to
the 920 cm-1 band used for the ∆H measurement of the Raman
spectrum of the vapor. Although the 928 cm-1 line is an A′′
mode, which will not have a Q branch in the Raman spectrum,
it will have very broad R, S, P, and O branches that could
contribute to the intensity of the 920 cm-1 Q branch of the A′
mode. In our present determination, the conformational pairs
utilized are well separated with reasonable intensities and, by
using the curve fitting procedure, the intensity ratios at different
temperatures can be determined with high accuracy. Addition-
ally, the temperature of the sample is accurately measured, and
therefore, the ∆H obtained in this study can be considered as
more reliable than those previously reported.
This program adds ab initio structural parameters, although
with much smaller weights, to the rotational constants so that
the number of observables is always larger than the number of
unknown parameters. As a consequence, we can obtain unique
results without any arbitrary assumptions for the values of
structural parameters.
The structural parameters are grouped in this way: all CH
bond lengths in the ethylene group are grouped into one set
and the other two CH bond lengths in the molecule are in
another set. Two HCCF dihedral angles (not in ethylene group)
are grouped together. Additionally, sets of structural parameters
in two conformers are combined further to reduce the number
of independent parameters. Corresponding sets of bond lengths
as well as two CdCH angles in the two conformers are
combined. The reasons for combining the parameters in this
way are that the errors of ab initio calculations in bond lengths
are smaller and more systematic than in bond angles and that
the angles of the ethylene group are more stable than the other
angles. The CdCH angles in the position cis to the CC single
bond in the two conformers are not combined because this H
atom is more affected by the CF bond in the cis conformer.
Therefore, the independent parameters are reduced to the number
22, which is much less than the number of available rotational
constants of 11 isotopic species in the optimization. The final
fitting of the rotational constants is satisfactory with maximum
relative and absolute errors no more than 0.06% and 7.8 MHz,
respectively (Table VIIIS in Supporting Information). These
structural parameter results are listed in the last columns of
Tables 4 and 5 as microwave-adjusted parameters. It is difficult
to estimate the uncertainties in these structural parameters, but
they should be less than those obtained as ro parameters from
just the microwave rotational constants in the previous deter-
mination.
Spectroscopic evidence on the conformational stability can
also be found by a comparison of the calculated Raman spectrum
of the mixture with the previously reported experimental
spectrum.4,5 They closely match in both band positions and
intensities. By review of these theoretical and experimental
spectra, the conclusion can be drawn confidently that the cis
conformer is the thermodynamically preferred conformer with
an enthalpy difference of ∼100 cm-1
.
The simulated infrared spectrum closely resembles the
observed spectrum, which provides excellent evidence for the
quality of the ab initio calculations. The calculated frequencies
agree within about 1% of those obtained from the cryosolution
as well as with those obtained from the matrix isolation
spectroscopic results performed by Nieminen et al.8 Because a
significant number of vibrational modes of this molecule involve
extensive mixing, it results in some difficulties in the assign-
ments of the normal modes. Under the guidance of the normal
coordinate analysis, along with the relative intensities of the
doublets with changes of temperature of the rare gas solutions,
we believe correct assignments have been made for the
fundamentals for both conformers. As pointed out in the
previous study,4 there is some difficulty in ascertaining the
torsional fundamental for the gauche conformer (108.05 or 92.10
cm-1). Our calculation for the gauche conformer closely matches
the experimental results, which are based on the 1- r 0(
transition originating at 108.05 cm-1. Therefore, the calculated
114 cm-1 value for this mode seems to add support for the
assignment to the higher frequency band and leaves unexplained
why the prediction1 of 84.6 ( 2.8 cm-1 from the microwave
data has such a large error. An error of nearly 30% is much
larger than expected where frequently predictions better than
10% have been made by this technique. Therefore, it would be
of interest to repeat these measurements to see if, in fact, they
are in error or if the gauche torsional fundamental should be
assigned to the lower frequency band.
The ab initio calculations indicate that the C-F bond distance
should be 0.01 Å shorter in the cis conformer compared to that
found in the gauche rotamer. By use of this difference as well
as the predicted difference of the other parameters, it was
possible to fit all of the previously determined rotational
constants and obtain reasonable structural parameters without
fixing any of them. The previously reported10 ro parameters are
the same within the given uncertainty. However, it was not
necessary to fix any of the angle parameters for the gauche
conformer. It is readily apparent that many of the previously
reported1 rs parameters for the gauche form were in error (Table
5) whereas most of the rs parameters for the cis conformer were
satisfactory except for the C-H6, which is much too long (Table
4). Parameters obtained by this method should be closer to the
actual parameters than those obtained by holding some param-
eters fixed while obtaining the other from the microwave
rotational constants. This procedure should be useful for
obtaining structural parameters for many other molecules where
the number of rotational constants is significantly smaller than
the number of structural parameters in the molecule.
The value of ∆H has been determined rather accurately via
the cryosolution method in three different rare gas solutions,
which indicates that the cis form is more stable than the gauche
rotamer. We have extrapolated these data to the vapor and
obtained an enthalpy value of 130 ( 25 cm-1. This is much
higher than the value of 58 ( 10 cm-1 obtained from the Raman
spectrum of the liquid;5 however, it is significantly lower than
With the more accurate ∆H value, as well as more accurate
structural parameters, we have recalculated the potential function
governing the conformational interchange. We have used the

3-Fluoropropene
J. Phys. Chem. A, Vol. 103, No. 13, 1999 1985
TABLE 7: Torsional Transitions (cm-1) and Assignment of
Gaseous 3-Fluoropropene
scopic equipment utilized in this study. J.R.D. acknowledges
partial support of these studies by the University of Missouri-
Kansas City Faculty Research Grant program. The authors also
acknowledge NATO for a travel grant, which made it possible
to complete this study.
conformer
cis
transition
observed
calculateda
∆
1 r 0
2 r 1
3 r 2
4 r 3
1- r 0(
2- r 1(
3- r 2(
99.75
164.41
157.69
151.392
142.57
108.05
104.55
99.62
164.36
158.11
150.91
143.50
108.05
104.38
99.38
0.05
-0.42
0.49
1.32
0.00
0.28
-0.13
Supporting Information Available: Table VIIIS listing the
fit of the rotational constants (MHz) of cis and gauche
3-fluoropropene microwave-adjusted structural parameters (ro).
This material is available free of charge via the Internet at http://
pubs.acs.org.
gauche
0.24
a Calculated with the six-parameter potential function listed in Table
4.
References and Notes
dihedral angle obtained from the fit of the microwave data for
the gauche conformer to eliminate the strong correlation of the
V1 and V2 terms in the potential function. The fit of the torsional
fundamentals for the two conformers and the associated “hot
bands” is given in Table 7. The determined potential parameters
are listed in Table 4 and compared to those obtained in previous
studies of 3-fluoropropene, as well as those obtained from the
various ab initio calculations. Because of the much lower value
for ∆H, the V1 term is now negative (-162 ( 9 cm-1), whereas
it was previously reported as positive (36 ( 9 cm-1). The
negative value is consistent with the value of this parameter
from the ab initio calculations. The other parameters are very
similar to those previously obtained except for the V5 term,
which now has a small negative value (-13 ( 5 cm-1), whereas
it was previously reported as positive (14 ( 8 cm-1) from the
previously reported potential function governing the conforma-
tional interchange.
It is interesting to compare the enthalpy difference between
the two stable conformers when the hydrogen on the carbon
2-position is replaced by a methyl group or a chlorine atom.
With the substitution of the chlorine atom at the 2-position there
is stabilization of the cis conformer that is predicted by the ab
initio calculations.27 On the other hand, substitution of a methyl
group at the 2-position favors the gauche conformer, and the
ab initio calculations indicate that the two conformers have
nearly equal energies.28 The preliminary results for the molecule
substituted with a methyl group on the 1-position indicate that
the gauche conformer is favored. It would be interesting to see
if the ab initio calculations predict this trend for this molecule,
and we plan to carry out such calculations in the near future.
In general, the theoretical calculations with a relatively large
basis set at the MP2 level can provide accurate geometric
parameters and reasonable guidance for the vibrational assign-
ment. However, as shown in the present study, when calculating
the energy difference between molecular conformers with values
less than 1.0 kcal/mol, caution must be used on utilizing the
conformer stability information. At this time the results need
to be verified by an experimental investigation.29
(1) Hirota, E. J. Chem. Phys. 1965, 42, 2071.
(2) Meakin, P.; Harris, D. D.; Hirota, E. J. Chem. Phys. 1969, 51, 3775.
(3) McLachlan, R. D.; Nyquist, R. A. Spectrochim. Acta 1968, 24A,
103.
(4) Durig, J. R.; Zhen, M.; Little, T. S. J. Chem. Phys. 1984, 81, 4259.
(5) Durig, J. R.; Zhen, M.; Heusel, H. L.; Joseph, P. J.; Groner, P.;
Little, T. S. J. Phys. Chem. 1985, 89, 2877.
(6) Durig, J. R.; Geyer, T. J.; Little, T. S.; Durig, D. T. J. Mol. Struct.
1988, 172, 165.
(7) Santhanam, V.; Sobhanadri, J.; Subramaniam, S. Pramana 1988,
30, 51.
(8) Nieminen, J.; Murto, J.; Rasanen, M. Spectrochim. Acta 1991, 47A,
1495.
(9) Durig, J. R.; Lee, M. J.; Badawi, H. M.; Sullivan, J. F.; Durig, D.
T. J. Mol. Struct. 1992, 266, 59.
(10) Durig, D. T.; Little, T. S.; Costner, T. G.; Durig, J. R. J. Mol. Struct.
1992, 266, 277.
(11) Bulanin, M. O. J. Mol. Struct. 1995, 347, 73.
(12) Moller, C.; Plesset, M. S. Phys. ReV. 1934, 46, 618.
(13) Hoffman, F. W. J. Org. Chem. 1949, 14, 105.
(14) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T. A.; Petersson,
G. A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94, revision B.3; Gaussian
Inc.: Pittsburgh, PA, 1995.
(15) Pulay, P. Mol. Phys. 1969, 17, 197.
(16) Luppi, J.; Rasanen, M.; Murto, J.; Pajunen, P. J. Mol. Struct. 1991,
245, 307.
(17) Wilson, E. B.; Decius, J. C.; Cross, P. C. Molecular Vibrations;
McGraw-Hill: New York, 1955.
(18) Frisch, M. J.; Yamaguchi, Y.; Gaw, J. F.; Schaefer, H. F., III;
Binkley, J. S. J. Chem. Phys. 1986, 84, 531.
(19) Amos, R. D. Chem. Phys. Lett. 1986, 124, 376.
(20) Chantry, G. W. In The Raman Effect; Anderson, A., Ed.; Marcel
Dekker Inc.: New York, 1971; Vol. 1, Chapter 2.
(21) Polavarapu, P. L. J. Phys. Chem. 1990, 94, 8106.
(22) Bulanin, M. O. J. Mol. Struct. 1973, 19, 59.
(23) van der Veken, B. J.; DeMunck, F. R. J. Chem. Phys. 1992, 97,
3060.
(24) Herrebout, W. A.; van der Veken, B. J.; Wang, A.; Durig, J. R. J.
Phys. Chem. 1995, 99, 578.
(25) Herrebout, W. A.; van der Veken, B. J. J. Phys. Chem. 1996, 100,
9671.
(26) Wong, M. W.; Frisch, M. J.; Wiberg, K. B. J. Am. Chem. Soc.
1991, 113, 4776.
(27) Durig, D. T.; Guirgis, G. A.; Durig, J. R. Struct. Chem. 1992, 3,
347.
(28) Durig, J. R.; Qui, H. Z.; Durig, D. T.; Zhen, M.; Little, T. S. J.
Phys. Chem. 1991, 95, 2745.
(29) Mack, H. G.; Oberhammer, H. J. Mol. Struct.: THEOCHEM 1992,
258, 197.
Acknowledgment. W.A.H. thanks the Fund for Scientific
Research (FWO, Belgium) for successive appointments as a
Research Assistant and Postdoctoral Fellow. The FWO is also
thanked for financial help toward the purchase of the spectro-