Electron attachment and detachment: Electron affinities of isomers of trifluoromethylbenzonitrile

Article abstract of DOI:10.1063/1.1806418

Rate constants for electron attachment to the three isomers of trifluoromethylbenzonitrile [(CF₃)(CN)C₆H₄, or TFMBN] were measured over the temperature range of 303-463 K in a 133-Pa He buffer gas, using a flowing-afterglow Langmuir-probe apparatus. At 303 K, the measured attachment rate constants are 9.0×10^-8 (o-TFMBN), 5.5× 10^-8 (m-TFMBN), and 8.9 × 10^-8 cm ³ s^-1 (p-TFMBN), estimated accurate to ±25%. The attachment process formed only the parent anion in all three cases. Thermal electron detachment was observed for all three anion isomers, and rate constants for this reverse process were also measured. From the attachment and detachment results, the electron affinities of the three isomers of TFMBN were determined to be 0.70(o-TFMBN), 0.67(m-TFMBN), and 0.83 eV (p-TFMBN), all ±0.05 eV. G3(MP2) [Gaussian-3 calculations with reduced Moller-Plesset orders (MP2)] calculations were carried out for the neutrals and anions. Electron affinities derived from these calculations are in good agreement with the experimental values.

Full text of DOI:10.1063/1.1806418

Electron attachment and detachment: Electron affinities of isomers of
trifluoromethylbenzonitrile
Thomas M. Miller, A. A. Viggiano, Jeffrey F. Friedman, and Jane M. Van Doren
Citation: J. Chem. Phys. 121, 9993 (2004); doi: 10.1063/1.1806418
View online: http://dx.doi.org/10.1063/1.1806418
View Table of Contents: http://jcp.aip.org/resource/1/JCPSA6/v121/i20
Published by the American Institute of Physics.
Additional information on J. Chem. Phys.
Journal Homepage: http://jcp.aip.org/
Journal Information: http://jcp.aip.org/about/about_the_journal
Top downloads: http://jcp.aip.org/features/most_downloaded
Information for Authors: http://jcp.aip.org/authors
Downloaded 06 Jul 2012 to 136.159.235.223. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions

JOURNAL OF CHEMICAL PHYSICS
VOLUME 121, NUMBER 20
22 NOVEMBER 2004
Electron attachment and detachment: Electron affinities
of isomers of trifluoromethylbenzonitrile
Thomas M. Millera) and A. A. Viggiano
Air Force Research Laboratory, Space Vehicles Directorate, Hanscom Air Force Base, Bedford,
Massachusetts 01731-3010
Jeffrey F. Friedman
Department of Physics, University of Puerto Rico, Mayaguez, Puerto Rico 00681-9016
and Air Force Research Laboratory, Directed Energy Directorate, Kirtland Air Force Base, Albuquerque,
New Mexico 87117-5776
Jane M. Van Doren
Department of Chemistry, College of the Holy Cross, Worcester, Massachusetts 01610-2395
͑Received 13 July 2004; accepted 23 August 2004͒
Rate constants for electron attachment to the three isomers of trifluoromethylbenzonitrile
(CF )(CN͒C H , or TFMBN͔ were measured over the temperature range of 303–463 K in a
͓
3
6
4
133-Pa He buffer gas, using a flowing-afterglow Langmuir-probe apparatus. At 303 K, the measured
attachment rate constants are 9.0ϫ10Ϫ8 ͑o-TFMBN͒, 5.5ϫ10Ϫ8 ͑m-TFMBN͒, and 8.9
ϫ10Ϫ8 cm3 sϪ1 ͑p-TFMBN͒, estimated accurate to Ϯ25%. The attachment process formed only the
parent anion in all three cases. Thermal electron detachment was observed for all three anion
isomers, and rate constants for this reverse process were also measured. From the attachment and
detachment results, the electron affinities of the three isomers of TFMBN were determined to be
0.70͑o-TFMBN͒, 0.67͑m-TFMBN͒, and 0.83 eV ͑p-TFMBN͒, all Ϯ0.05 eV. G3͑MP2͒ ͓Gaussian-3
calculations with reduced Møller–Plesset orders ͑MP2͔͒ calculations were carried out for the
neutrals and anions. Electron affinities derived from these calculations are in good agreement with
the experimental values. © 2004 American Institute of Physics.
͓DOI: 10.1063/1.1806418͔
I. INTRODUCTION
movable Langmuir probe. The electron density was kept low
enough that electron-ion and ion-ion recombination were
negligible processes, and low enough compared to the
TFMBN concentration that first-order kinetics applied. A
mass spectrometer at the downstream end of the 1-m-long
flow tube was used to determine the ion products of attach-
ment. The reactants were used as obtained from the manu-
facturer aside from degassing with freeze-thaw-pump
cycles.4 The vapor pressures of the TFMBN are so low
͑30–70 Pa͒ that the degassing procedure was finished off by
simply pumping on the room-temperature sample for a
minute or two. An example of the data is given in Fig. 1, for
o-TFMBN at a concentration of nrϭ3.9ϫ1010 cmϪ3. The
fitted lines are solutions of the one-dimensional rate equa-
tions for the electron density along the axis of the flow tube.5
For the case shown, the electron detachment process keeps
the electron density from falling rapidly to zero. The early
portion of the attachment/detachment curve is most sensitive
to ka , and the level of the curve at long times is dependent
upon the ratio kd /ka , as the curve approaches a diffusion-
limited steady state. The ka are estimated accurate to Ϯ25%.
The kd are estimated accurate to Ϯ30%.
Chowdhury and Kebarle used ion equilibria to measure
the electron affinity (EA) of each of the three isomers of
trifluoromethylbenzonitrile (CF )(CN͒C H or TFMBN͔
and obtained 0.70Ϯ0.10 eV for o-TFMBN, 0.67Ϯ0.10 eV
for m-TFMBN, and 0.76Ϯ0.10 eV for p-TFMBN.1 In the
present work, we have studied the process
͓
3
6
4
eϪϩTFMBNꢀTFMBNϪ.
͑1͒
We measured rate constants ka for electron attachment to
TFMBN and rate constants kd for thermal electron detach-
ment from TFMBNϪ for each of the three isomers in the
temperature range of 303–463 K. From these results, the
EAs of the three isomers were determined.
II. EXPERIMENT
A flowing-afterglow Langmuir-probe apparatus was used
in this work. The method has been well described in the
literature,2 and the Air Force Research Laboratory apparatus
has also been described.3 The experiment took place in a
weak flowing plasma in a He buffer gas at 133 Pa pressure.
TFMBN was added to the electron-Heϩ, Arϩ plasma, at a
level of typically 1 part per million by volume, at a point
halfway along the length of a flow tube, and the subsequent
decay in the electron concentration was measured with a
III. ATTACHMENT AND DETACHMENT RESULTS
Electron attachment to all three isomers of TFMBN
yielded only the parent anion in the temperature range of the
present experiments ͑303–463 K͒. At 303 K, the measured
a͒
Electronic mail: thomas.miller@hanscom.af.mil
0021-9606/2004/121(20)/9993/6/$22.00
9993
© 2004 American Institute of Physics
Downloaded 06 Jul 2012 to 136.159.235.223. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions

9994
J. Chem. Phys., Vol. 121, No. 20, 22 November 2004
Miller et al.
samples with mixing ratios of 0.23% and 0.47% were used.
The Klots formula for the capture cross section for electron
collisions gives 3.6ϫ10Ϫ7 cm3 sϪ1 for all three isomers of
TFMBN at 303 K, meaning that attachment occurs on every
4–7 collisions.7
Thermal electron detachment, with the signature illus-
trated in Fig. 1, was observed for all three isomers of
TFMBN. The kd were too small to measure at 303 K ͑though
marginally so for the o- and m-TFMBN isomers͒, but in-
creased to 1500 sϪ1 at elevated temperatures, as tabulated in
Tables I–III and shown in Fig. 2. While it is customary to
plot the logarithm of the kd versus 1/T, we have noted in
earlier work that detachment is not an Arrhenius process, so
the fitted lines in Fig. 2 are simply to guide the eye.8,9
IV. DETERMINATION OF ELECTRON AFFINITIES
FIG. 1. Electron attachment/detachment data obtained at 403 K for
The kd are related to the ka via
o-TFMBN. The concentration of o-TFMBN was 3.9ϫ1010 cmϪ3. The fit to
kdϭkaL0 273.15/T͒exp Ϫ EA/kT͒Ϫ ⌬Sؠ/k͒
the data gave kaϭ5.5ϫ10Ϫ8 cm3 sϪ1 and kdϭ1150 sϪ1
.
͑
͓
͑
͑
Ϫ H ϪH ͒/kT .
͑2͒
͑
͔
T
0
attachment rate constants are 9.0ϫ10Ϫ8 ͑o-TFMBN͒, 5.5
ϫ10Ϫ8 ͑m-TFMBN͒, and 8.9ϫ10Ϫ8 cm3 sϪ1 ͑p-TFMBN͒,
estimated accurate to Ϯ25%. All of the data are given in
Tables I–III. The attachment rate constants declined some-
what with temperature. The attachment rate constants for o-
and p-TFMBN are similar. Those for m-TFMBN tended to
be only half as great. This observation is consistent with the
known inability of the meta-isomer of an aromatic ring with
electron withdrawing substituents to offer resonance stabili-
zation to the intermediate anion formed in nucleophilic sub-
stitution reactions.6 The difference cannot be attributed to an
error in preparing mixtures in He gas as two different mix-
tures were used in each case. In particular, for m-TFMBN,
In Eq. ͑2͒, k is Boltzmann’s constant, L0 is Loschmidt’s
number, EA is the electron affinity of TFMBN ͑at 0 K, by
definition͒, ⌬Sؠ is the entropy change due to electron attach-
ment at temperature T, and HTϪH0 is the thermal energy
correction needed to reduce the EA result from the measure-
ment temperature T to 0 K. Details regarding ⌬Sؠ and HT
ϪH0 have been given earlier.8,9 Thus, the measured values
of ka and kd allow us to determine EAs provided that entro-
pies and heat capacities are known. These quantities were
calculated using density-functional methods described in
Sec. V, and the results are given in Tables I–III along with
the EA values determined form Eq. ͑2͒. Entropies and heat
capacities for the electron were taken from the Joint Army-
TABLE I. Electron attachment rate constants ka, detachment rate constants kd, ambipolar diffusion decay
constants D, entropies Sؠ, integrated specific heats
͐
CpdT, and apparent electron affinities EA from the
v
present work with o-TFMBN. The calculated values of Sؠ and
͐ CpdT in plain typeface are for neutral
o-TFMBN, and those in italics are for the anion o-TFMBNϪ. Values of
may show variation because of
vD
deviations in the amount of Ar added and of the plasma velocity.
Sؠ
͐
CpdT
ka
kd
EA
͑eV͒
vD
T
͑K͒
a
a
b
c
͑sϪ1
͒
͑meV KϪ1
͒
͑meV͒
(10Ϫ8 cm3 sϪ1
)
͑sϪ1
͒
303
314
¯
¯
¯
8.9
9.0
9.0
͑210͒
͑155͒
͑125͒
¯
343
383
393
403
413
366
460
483
510
535
¯
8.0
8.0
͑238͒
͑270͒
¯
¯
0.69
0.69
4.763
4.893
413.7
433.4
6.2
6.2
697
720
4.781
4.947
433.1
453.7
6.4
6.4
993
1005
0.69
0.69
4.833
5.001
452.9
474.3
5.8
5.5
1240
1150
0.70
0.70
4.884
473.1
6.3
1830
0.70
5.055
495.4
aSؠ and
͐ CpdT were calculated using the B3LYP/6-311ϩϩG(3df,2p) functional and basis set, with vibra-
tional frequencies scaled by 0.989.
bEstimated accurate to Ϯ25%.
cEstimated accurate to Ϯ30%.
Downloaded 06 Jul 2012 to 136.159.235.223. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions

J. Chem. Phys., Vol. 121, No. 20, 22 November 2004
Electron affinities of trifluoromethylbenzonitrile
9995
TABLE II. Electron attachment rate constants ka , detachment rate constants
TABLE III. Electron attachment rate constants ka , detachment rate con-
k , ambipolar diffusion decay constants D , entropies Sؠ, integrated spe-
stants k , ambipolar diffusion decay constants D , entropies Sؠ, integrated
v
v
d
d
cific heats
work with m-TFMBN. The calculated values of Sؠ and
typeface are for neutral m-TFMBN, and those in italics are for the anion
m-TFMBNϪ. Values of
may show variation because of deviations in the
͐
CpdT, and apparent electron affinities EA from the present
specific heats
work with p-TFMBN. The calculated values of Sؠ and
typeface are for neutral p-TFMBN, and those in italics are for the anion
p-TFMBNϪ. Values of
may show variation because of deviations in the
͐
CpdT, and apparent electron affinities EA from the present
͐
CpdT in plain
͐
CpdT in plain
vD
amount of Ar added and of the plasma velocity.
vD
amount of Ar added and of the plasma velocity.
Sؠ
͐
CpdT
ka
kd
EA
͑eV͒
Sؠ
͐
CpdT
ka
kd
EA
͑eV͒
vD
T
͑K͒ ͑sϪ1
vD
T
͑K͒ ͑sϪ1
a
a
b
c
a
a
b
c
͒
͑meV KϪ1
͒
͑meV͒
(10Ϫ8 cm3 sϪ1
)
͑sϪ1
͒
͒
͑mev KϪ1
͒
͑meV͒
(10Ϫ8 cm3 sϪ1
)
͑sϪ1
͒
303
353
363
373
383
393
330
¯
¯
5.5
5.6
͑130͒
͑125͒
¯
¯
0.65
0.65
303
383
300
429
¯
¯
¯
¯
8.9
8.7
8.3
9.0
9.1
8.4
͑5͒
͑50͒
͑60͒
͑305͒
͑280͒
͑295͒
¯
¯
¯
¯
¯
¯
¯
¯
420
442
443
484
512
4.699
4.793
358.9
377.5
4.4
4.7
513
520
393
403
413
423
433
443
453
463
505
520
537
561
536
595
600
649
¯
¯
¯
¯
¯
¯
4.752
4.849
377.1
396.7
4.1
4.2
718
720
0.65
0.66
¯
¯
¯
¯
8.4
8.4
8.4
8.0
8.2
͑388͒
͑518͒
606
¯
¯
¯
4.805
4.905
395.7
416.4
3.8
4.2
990
970
0.66
0.66
¯
¯
¯
¯
¯
4.858
4.960
414.8
436.4
3.3
3.3
1120
1120
0.67
0.67
5.125
5.230
5.175
5.282
5.225
5.334
5.275
5.385
516.4
539.0
537.6
561.0
559.1
583.3
581.0
605.9
0.82
0.82
0.83
910
4.910
5.015
434.2
456.8
3.5
3.5
1580
1552
0.68
0.68
1110
aSؠ and
͐CpdT were calculated using the B3LYP/6-311ϩϩG(3df,2p)
8.2
8.6
1548
1490
0.84
0.84
functional and basis set, with vibrational frequencies scaled by 0.989.
bEstimated accurate to Ϯ25%.
cEstimated accurate to Ϯ30%.
aSؠ and
͐CpdT were calculated using the B3LYP/6-311ϩϩG(3df,2p)
functional and basis set, with vibrational frequencies scaled by 0.989.
bEstimated accurate to Ϯ25%.
Navy-Air Force tables.10 We have shown earlier that these
quantities are of sufficient accuracy that the uncertainty in
EA values is sub-meV due to the calculated contribution,
partly because the entropy and heat-capacity contributions
tend to offset, but also because the calculated contributions
are only a small part of the EA evaluation to begin with
͑5–21 meV in the present work͒. The final EA values were
taken from those data where two conditions are met for op-
timal determination of ka and kd : the attachment frequency
kanr is comparable in magnitude to kd and both are greater
cEstimated accurate to Ϯ30%.
needed to interpret the ka and kd data in terms of EAs. The
GAUSSIAN 03 program package11 was used. The calculations
were carried out on a personal computer and on an IBM
SP/RS6000 computer at the Maui High Performance Com-
puting Center. Becke’s three-parameter hybrid functional12
with the Lee, Yang, and Parr ͑B3LYP͒ correlation
functional13 was used, with the 6-311ϩϩG(3df,2p) Gauss-
ian basis set, to calculate the molecular geometries, zero-
point energies ͑ZPE͒, and total energies. In doing so, the
calculated vibrational frequencies were scaled by 0.989.14
Entropies and heat capacities are given in Tables I–III. Total
than the diffusion decay constant D . As shown in Eq. ͑2͒,
v
EA is only logarithmically dependent upon the ratio ka /kd
and, hence, more accurate than ka and kd . We find the fol-
lowing:
EA(o-TFMBN)ϭ0.70 eV,
EA(m-TFMBN)
ϭ0.67 eV, and EA(p-TFMBN)ϭ0.83 eV, all Ϯ0.05 eV.
These values agree well with those measured by Chowdhury
and Kebarle studying ion equilibria: 0.70, 0.67, and 0.76 eV,
respectively, all Ϯ0.10 eV.1 The precision of the present
work is sufficient to prove that the para-isomer has a signifi-
cantly larger EA than the ortho- and meta-isomers. Interest-
ingly, this relationship between the EAs of the isomers is not
what one would expect on the basis of resonance stabiliza-
tion arguments,6 and suggests that structural relaxation in the
anion following electron attachment contributes more to the
stabilization of the para-isomer and, hence, to the larger EA.
Measurements of the vertical attachment energies could shed
light on this issue.
V. COMPUTATIONAL METHOD AND RESULTS
Density functional theory ͑DFT͒ calculations were car-
ried out primarily to generate entropies and heat capacities
FIG. 2. Electron detachment rate constants measured for TFMBN anions.
The solid lines were drawn to guide the eye.
Downloaded 06 Jul 2012 to 136.159.235.223. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions

9996
J. Chem. Phys., Vol. 121, No. 20, 22 November 2004
Miller et al.
TABLE IV. Results of Møller-Plesset and DFT calculations for isomers of trifluoromethylbenzonitrile, in
hartree units except where indicated. Results for the neutrals are given in plain type, and those for the anions are
given in italics.
Quantity
HF ZPEa
o-TFMBN
m-TFMBN
p-TFMBN
0.10050
0.09552
0.10216
0.10043
0.09503
0.10223
0.10043
0.09609
0.10218
DFT ZPEb
0.09782
0.09753
0.09819
c
G3͑MP2͒ ͑0 K͒
Ϫ660.70203
Ϫ660.72651
Ϫ661.66147
Ϫ661.69397
0.884 eV
0.63 eV
Ϫ660.70551
Ϫ660.72879
Ϫ661.66611
Ϫ661.69698
0.840 eV
0.59 eV
Ϫ660.70560
Ϫ660.73434
Ϫ661.66636
Ϫ661.70366
1.015 eV
0.77 eV
d
DFT ͑0 K͒
DFT EAe
DFT adjustedf
G3͑MP2͒ EAe
0.666 eV
0.70Ϯ0.10 eV
0.70Ϯ0.05 eV
0.634 eV
0.67Ϯ0.10 eV
0.67Ϯ0.05 eV
0.782 eV
0.76Ϯ0.10 eV
0.83Ϯ0.05 eV
g
EA ͑ion equilibria͒
EA ͑present expt.͒
aHartree–Fock/6-31G(d) level of theory, scaled by 0.8929, for G3͑MP2͒ results.
bB3LYP/6-311ϩϩG(3df,2p) level of theory, scaled by 0.989.
cTotal energy at 0 K using the G3͑MP2͒ formalism.
dB3LYP/6-311ϩϩG(3df,2p)//B3LYP/6-311ϩϩG(3df,2p)ϩZPE.
eDifference between the total energy of the neutral at 0 K minus that of the anion.
fOn the basis of experience, we recommend subtracting 0.25 eV from EAs obtained using the DFT method
specified in footnote d.
gReference 1.
energies and EAs are given in Table IV. Total energies for
the TFMBN neutrals yield essentially equal energies for the
meta- and para-isomers ͑differing by 7 meV, far below the
capability of DFT͒. The ortho-isomer lies higher by 130
meV. More accurate G3͑MP2͒ calculations15 show the same
trend: the meta-isomer lies 2 meV higher than para ͑well
below the accuracy of the method͒, and the ortho-isomer lies
97 meV higher. The G3͑MP2͒ method is capable of much
greater accuracy than DFT for obtaining total energies, but
the DFT calculations should be better for entropies and spe-
cific heats because a much larger basis set could be used with
the very economical DFT method. Zhou and Liu,16 using a
similar DFT method and basis set, also found the ortho-
isomer to be less stable than the meta-isomer in the related
͑unfluoronated͒ methylbenzonitrile molecules.
Optimized geometries are shown in Figs. 3–5. The ge-
ometries of the o- and m-TFMBN found in this study are
very similar to DFT geometries of the unfluorinated analogs,
except for the difference in the methyl C–F versus C–H
bond lengths.16 To the best of our knowledge, the present
study is the first investigation of the TFMBN neutral and
anionic geometries. The C–C bonds in the ring tend to
lengthen upon electron attachment, by at most 0.06 Å. The
C–CN and C–CF3 bonds shorten upon attachment, by 0.03–
0.05 Å. The C–F bonds lengthen by at most 0.05 Å. And the
CF3 group moves out of the average plane of the molecule
by a few degrees, most noticeably for p-TFMBNϪ.
The G3͑MP2͒ and DFT EAs are given in Table IV,
along with experimental values. In our experience, EAs cal-
culated using the B3LYP/6-311ϩϩG(3df,2p) functional
FIG. 3. Geometries for neutral and anionic o-TFMBN, with views above and to the side of the average plane. The bond lengths ͑in angstrom͒ in plain type
are from B3LYP/6-311ϩϩG(3df,2p) optimizations, and those in italics are from MP2/6-31G(d) optimizations used in the G3͑MP2͒ calculations. The three
F atoms are at the top of the figure. The N atom is at the far left. The angle ЄC5-C2-C8 is 179.6° and 179.3° ͑DFT, neutral and anion͒ and 177.2° and 170.2°
͑MP2, neutral and anion͒. The dihedral angle C3-C2-C8-F is zero for both DFT and MP2 neutral and anion.
Downloaded 06 Jul 2012 to 136.159.235.223. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions

J. Chem. Phys., Vol. 121, No. 20, 22 November 2004
Electron affinities of trifluoromethylbenzonitrile
9997
FIG. 4. Geometries for neutral and anionic m-TFMBN, with views above and to the side of the average plane. The bond lengths ͑in angstrom͒ in plain type
are from B3LYP/6-311ϩϩG(3df,2p) optimizations, and those in italics are from MP2/6-31G(d) optimizations used in the G3͑MP2͒ calculations. The three
F atoms are at the top of the figure. The N atom is at the far left. The angle ЄC6-C3-C8 is 178.3° and 178.6° ͑DFT, neutral and anion͒ and 177.3° and 178.3°
͑MP2, neutral and anion͒. The dihedral angle C2-C3-C8-F is 15.0° and 0° ͑DFT, neutral and anion͒ and Ϫ24.7° and 0° ͑MP2, neutral and anion͒.
and basis set tend to be 0.2–0.3 eV too high. Thus, the
present DFT results given in Table IV are also displayed as
‘‘adjusted’’ by subtracting 0.25 eV. The adjusted DFT EAs
are expected accurate to Ϯ0.25 eV. G3͑MP2͒ EAs are good
on average within Ϯ0.056 eV, and it is unusual for them to
be wrong by more than 0.1 eV. The G3͑MP2͒ and adjusted
DFT EAs are in good agreement with experimental values.
A natural population analysis was carried out using the
B3LYP/6-311ϩϩG(3df,2p) charge densities in order to
determine where the attached electron resides in the anions.17
These results are given in Table V. The extra electronic
charge in the anions is shared generously among the atoms,
with the CN group ͑specifically the N atom͒ and the C1 and
C4 atoms in the ring receiving the largest fractions, 0.20,
0.14, and 0.18 e, respectively, as expected from resonance
considerations. In a related study of benzonitrile binding to
metal surfaces ͑through the N atom͒, Mrozek et al.18 also
found that the donated electronic charge is carried by the N,
C1, and C4 atoms, as well as the C2 and C6 atoms. The ring
carbon bonded to the CF3 group in the o- and m-TFMBN
anions also carry significant negative charge, although to a
lesser extent in the meta-isomer. Finally, the CF3 group ͑F
atoms in particular͒ carries some of the added charge with
the largest group charge found in the ortho- and para-
isomers, as expected from resonance considerations.
VI. CONCLUSIONS
Rate constants for electron attachment to the three iso-
mers of TFMBN were measured in the range of 303–463 K
͑Tables I–III͒. The ka for the meta-isomer tended to be half
as great as those for the ortho- and para-isomers. The meta-
FIG. 5. Geometries for neutral and anionic p-TFMBN, with views above and to the side of the average plane. The bond lengths ͑in angstrom͒ in plain type
are from B3LYP/6-311ϩϩG(3df,2p) optimizations, and those in italics are from MP2/6-31G(d) optimizations used in the G3͑MP2͒ calculations. The three
F atoms are at the top of the figure. The N atom is at the bottom. The angle ЄC1-C4-C8 is 178.1° and 175.7° ͑DFT, neutral and anion͒ and 177.2° and 170.2°
͑MP2, neutral and anion͒. The dihedral angle C4-C3-C8-F is Ϫ88.9° and Ϫ86.7° ͑DFT, neutral and anion͒ and Ϫ88.3° and Ϫ86.7° ͑MP2, neutral and anion͒.
Downloaded 06 Jul 2012 to 136.159.235.223. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions

9998
J. Chem. Phys., Vol. 121, No. 20, 22 November 2004
Miller et al.
TABLE V. Atomic charges determined from a natural population analysis of the TFMBN neutrals and anions using B3LYP/6-311ϩϩG(3df,2p) electron
densities. The ‘‘extra eϪ’’ numbers are the difference between neutral and anion charges. The charges on the individual H and F atoms are so similar that they
have been averaged for this table.
Molecule
C1
C2
C3
C4
C5
C6
C7a
C8b
N
H
F
o-TFMBN
Ϫ0.148
Ϫ0.278
0.130
Ϫ0.169
Ϫ0.310
0.140
Ϫ0.161
Ϫ0.299
0.138
Ϫ0.099
Ϫ0.232
0.133
Ϫ0.128
Ϫ0.134
0.006
Ϫ0.143
Ϫ0.206
0.063
Ϫ0.172
Ϫ0.170
Ϫ0.003
Ϫ0.142
Ϫ0.234
0.092
Ϫ0.171
Ϫ0.210
0.039
Ϫ0.172
Ϫ0.338
0.166
Ϫ0.151
Ϫ0.337
0.186
Ϫ0.126
Ϫ0.302
0.176
Ϫ0.181
Ϫ0.268
0.087
Ϫ0.189
Ϫ0.210
0.021
Ϫ0.171
Ϫ0.210
0.039
Ϫ0.140
Ϫ0.174
0.034
Ϫ0.133
Ϫ0.256
0.123
Ϫ0.143
Ϫ0.206
0.063
0.274
0.269
0.005
0.282
0.282
0.000
0.282
0.279
0.003
1.096
1.076
0.021
1.091
1.087
0.004
1.088
1.066
0.023
Ϫ0.279
Ϫ0.476
0.197
Ϫ0.294
Ϫ0.493
0.199
Ϫ0.293
Ϫ0.488
0.196
0.222
0.188
0.034
0.226
0.192
0.034
0.227
0.194
0.033
Ϫ0.355
Ϫ0.387
0.032
Ϫ0.357
Ϫ0.388
0.031
Ϫ0.357
Ϫ0.399
0.042
o-TFMBNϪ
extra eϪ
m-TFMBN
m-TFMBNϪ
extra eϪ
p-TFMBN
p-TFMBNϪ
extra eϪ
aThe C7–N cyanide group is bound to the ring atom C1 in each case.
bThe C8–F3 trifluoromethyl group is bound to C2 in the ortho-isomer, C3 in the meta-isomer, and C4 in the para-isomer.
isomer of an aromatic ring with electron withdrawing sub-
stituents cannot offer resonance stabilization to the interme-
diate anion formed in nucleophilic substitution reactions, a
process analogous to electron attachment.6 Only the parent
anion was observed as a product of the attachment process.
Rate constants were also measured for the thermal electron
detachment process with the TFMBNϪ anions. The
attachment/detachment equilibrium constant yielded EAs for
the three isomers: 0.70͑o-TFMBN͒, 0.67͑m-TFMBN͒, and
0.83͑p-TFMBN͒, all in eV and accurate to Ϯ0.05 eV, and all
in good agreement with those measured by Chowdhury and
Kebarle.1 The relative values of the adiabatic EAs are not
well explained by resonance stability arguments. Density-
functional calculations were carried out in order to determine
the geometry changes in going from neutral to anionic
TFMBN and to obtain entropy and specific heat quantities
needed to reduce the experimental equilibrium constants to
EAs at 0 K. These calculations also yielded values of EAs,
which lie about 0.25 eV above the experimental ones, con-
sistent with our experience in the past. The G3͑MP2͒ method
was applied and yielded EAs in good agreement with experi-
ment.
1 S. Chowdhury and P. Kebarle, J. Am. Chem. Soc. 108, 5453 ͑1986͒; P.
Kebarle and S. Chowdhury, Chem. Rev. ͑Washington, D.C.͒ 87, 513
͑1987͒.
2
ˇ
ˇ
D. Smith and P. Spanel, Adv. At. Mol. Phys. 32, 307 ͑1994͒.
3 T. M. Miller, A. E. S. Miller, J. F. Paulson, and X. Liu, J. Chem. Phys.
100, 8841 ͑1994͒.
4 The ␣,␣,␣-trifluoromethylbenzonitrile isomers were purchased from Ald-
rich Chemicals and stated to be 99% pure, except for the ortho-isomer
͑98%͒. Approximate vapor pressures we observed while preparing mix-
tures in He at 299 K were 31͑o-TFMBN͒, 69͑m-TFMBN͒, and 55 Pa
͑p-TFMBN͒. None proved sticky on surfaces, and all were easily handled,
though transfer of p-TFMBN required a few degrees of warming to liq-
uefy.
5 T. M. Miller, R. A. Morris, A. E. S. Miller, A. A. Viggiano, and J. F.
Paulson, Int. J. Mass Spectrom. Ion Processes 135, 195 ͑1994͒.
6 J. McMurry, Organic Chemistry ͑Brooks/Cole, Pacific Grove, California,
1988͒, pp. 548–550.
7 C. E. Klots, Chem. Phys. Lett. 38, 61 ͑1976͒; polarizabilities were taken
from the B3LYP/6-311ϩϩG(3df,2p) calculations described in Sec. V:
14.8͑o-TFMBN͒, 14.9͑m-TFMBN͒, and 15.1 Å3 ͑p-TFMBN͒.
8 T. M. Miller, J. F. Friedman, and A. A. Viggiano, J. Chem. Phys. 120,
7024 ͑2004͒; there is a misprint in Table II of this paper: the C–F bond
lengths in c-C4FϪ8 are all 1.410 Å for the G3͑MP2͒ method, which uses a
MP2(Full)/6-31G(d) geometry optimization.
9 T. M. Miller, J. M. Van Doren, and A. A. Viggiano, Int. J. Mass. Spectrom.
233, 67 ͑2004͒.
10 M. W. Chase, C. A. Davies, J. R. Downey, D. J. Frurip, R. A. McDonald,
and A. N. Syverud, J. Phys. Chem. Ref. Data Suppl. 14, 1 ͑1985͒.
11 M. J. Frisch, G. W. Trucks, H. B. Schlegel et al., GAUSSIAN 03, Revision
B.02, Gaussian, Inc., Pittsburgh, PA, 2003.
ACKNOWLEDGMENTS
12 A. D. Becke, J. Chem. Phys. 98, 5648 ͑1993͒.
13 J. P. Perdew, K. Burke, and Y. Wang, Phys. Rev. B 54, 16533 ͑1993͒; K.
Burke, J. P. Perdew, and Y. Wang, in Electron Density Functional Theory:
Recent Progress and New Directions, edited by J. F. Dobson, G. Vignale,
and M. P. Das ͑Plenum, New York, 1998͒.
The authors are grateful for the support of the Air Force
Office of Scientific Research for this work. This research was
supported in part by a grant of computer time from the DOD
High Performance Computing Modernization Program at the
Maui High Performance Computing Center. One of the au-
thors ͑J.M.V.D.͒ acknowledges support from the National
Academy of Science of Sciences Air Force Summer Faculty
Fellowship Program and the College of the Holy Cross. One
of the authors ͑T.M.M.͒ is under contract ͑F19628-99-C-
0069͒ with Visidyne, Inc., Burlington, MA.
14 C. W. Bauschlicher and H. Partridge, J. Chem. Phys. 103, 1788 ͑1995͒.
15 L. A. Curtiss, P. C. Redfern, K. Raghavachari, V. Rassolov, and J. A.
Pople, J. Chem. Phys. 110, 4703 ͑1999͒.
16 X. Zhou and R. Liu, J. Mol. Struct.: THEOCHEM 461–462, 569 ͑1999͒.
17 E. D. Glendening, A. E. Reed, J. E. Carpenter, and E. Weinhold, NBO
Version 3.1, Gaussian, Inc., Pittsburgh, PA, 2003.
18 M. F. Mrozek, S. A. Wasileski, and M. J. Weaver, J. Am. Chem. Soc. 123,
12817 ͑2001͒.
Downloaded 06 Jul 2012 to 136.159.235.223. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions