Unusual ionic hydrogen bonds: Complexes of acetylides and fluoroform

Article abstract of DOI:10.1021/ja000806z

Ion-molecule complexes of substituted acetylides, RCC^- (R = tert-butyl, H, phenyl, p-tolyl), and fluoroform, HCF₃, were studied using Fourier transform ion cyclotron resonance mass spectrometry. These complexes, RCC^-·HCF3, all have complexation energies of approximately -19 kcal/mol and are, therefore, hydrogen bonded. The acetylides vary in basicity over a 6 kcal/mol range, but all have the same complexation energy with fluoroform. The structure of these complexes was verified by deuterium isotopic exchange reactions and equilibrium fractionation experiments. The relationship between acid-base thermochemistry and hydrogen bond stability is discussed.

Full text of DOI:10.1021/ja000806z

J. Am. Chem. Soc. 2000, 122, 8739-8745
8739
Unusual Ionic Hydrogen Bonds: Complexes of Acetylides and
Fluoroform
Michael L. Chabinyc and John I. Brauman*
Contribution from the Department of Chemistry, Stanford UniVersity, Stanford, California 94305-5080
ReceiVed March 6, 2000
Abstract: Ion-molecule complexes of substituted acetylides, RCC- (R ) tert-butyl, H, phenyl, p-tolyl), and
fluoroform, HCF3, were studied using Fourier transform ion cyclotron resonance mass spectrometry. These
complexes, RCC-‚HCF3, all have complexation energies of approximately -19 kcal/mol and are, therefore,
hydrogen bonded. The acetylides vary in basicity over a 6 kcal/mol range, but all have the same complexation
energy with fluoroform. The structure of these complexes was verified by deuterium isotopic exchange reactions
and equilibrium fractionation experiments. The relationship between acid-base thermochemistry and hydrogen
bond stability is discussed.
Introduction
The connection between acid-base thermochemistry and
hydrogen bond stability is widely held to be intrinsic because
of the many successful correlations that have been observed.
We studied complexes of carbon acids as hydrogen bond donors
and carbanions as hydrogen bond acceptors to better understand
the correlation between acid-base chemistry and hydrogen bond
strength. We found that ion-molecule complexes of fluoroform
and aliphatic alkoxides have the structure RO-‚HCF3, even
when proton transfer to the alkoxide is favorable in the separated
reactants.11 In a study of methanol-acetylide complexes, we
found that the complexation energy had a different dependence
on acid-base thermochemistry depending on whether the donor
or acceptor was varied.12 This result suggested that the
dependence of the complexation energy on the acid-base
thermochemistry is not always related to the shape of the
potential energy surface. The results from studies of deuterium
exchange proton-transfer reactions also support these claims.13,14
These studies indicate that the apparent relation between
complexation energies of ionic systems and acid-base chemistry
is not a causal one.
Hydrogen bonds to ions have been widely studied.1-5
However, relatively little information, aside from theory6 is
known about the strength of the interaction between carbanions
and carbon acids in either the gas phase7,8 or in solution.9,10
We report here our results on the stability and structure of gas-
phase hydrogen-bonded ion-molecule complexes of substituted
acetylides with fluoroform. By studying these complexes we
can gain insight into the factors affecting hydrogen bond stability
as well as learn about intermediates in proton-transfer reac-
tions.11,12
Considerable effort has been directed at exploring the
connections between acid-base chemistry and hydrogen bond
stability in both the gas-phase and in solution. The structure of
hydrogen bonded complexes is typically assumed to reflect the
acidity difference of the molecules in the complex. That is, the
structure is A-‚HB if AH is a stronger acid than BH, and AH‚B-
if BH is the stronger acid. Linear relationships are observed
between the acidity of the hydrogen bond donor (or the basicity
of the hydrogen bond acceptor) and the complexation energy.2,3,5
These observations appear to hold for many anionic and many
cationic systems as well. Information about the shape of the
proton transfer potential energy surface is frequently inferred
from these linear relationships.
Here we have synthesized complexes of fluoroform with
several substituted acetylides, RCC-‚HCF3 (RdH, phenyl,
p-tolyl). Their equilibrium complexation energies were deter-
mined relative to the corresponding methanol complex,
RCC-‚HOCH3. Their complexation energies are all ∼-19 kcal/
mol and are moderately stronger than a typical ion-dipole
complex (∼-10-15 kcal/mol). Therefore, they are hydrogen
bonded. Bimolecular proton/deuterium transfer solvent exchange
experiments and ab initio molecular orbital calculations were
performed to verify the structure of the complexes. This work
clearly demonstrates the importance of charge distributions to
hydrogen bond stability for anionic complexes.
(1) Hibbert, F.; Emsley, J. In AdVances in Physical Organic Chemistry;
Bethell, D., Ed.; Academic Press: London, 1990; Vol. 26.
(2) Perrin, C. L.; Nielson, J. B. Annu. ReV. Phys. Chem. 1997, 48, 511-
544.
(3) Kebarle, P. Annu. ReV. Phys. Chem. 1977, 28, 445-476.
(4) Mautner, M. Acc. Chem. Res. 1984, 17, 186-93.
(5) Takashima, K.; Riveros, J. M. Mass Spectrom. ReV. 1998, 17, 409-
430.
(6) (a) Saunders, W. H. J. Am. Chem. Soc. 1994, 116, 3452-3458. (b)
Saunders, W. H.; Van Verth, J. E. J. Org. Chem. 1995, 60, 3452-3458.
(7) Caldwell, G.; Rozeboom, M. D.; Kiplinger, J. P.; Bartmess, J. E. J.
Am. Chem. Soc. 1984, 106, 4660-4667.
(8) Meot-Ner, M. J. Am. Chem. Soc. 1988, 110, 3858-3862.
(9) Cram, D. J. Fundamentals of Carbanion Chemistry; Academic
Press: New York, 1965.
Experimental Section
Materials. Acetylene and d2-acetylene were synthesized by the
standard literature procedure for the reaction of calcium carbide with
(10) Ahlberg, P.; Davidsson, O¨ . J. Chem. Soc., Chem. Commun. 1987,
623-624.
(11) Chabinyc, M. L.; Brauman, J. I. J. Am. Chem. Soc. 1998, 120,
10863-10870.
(13) Squires, R. R.; Bierbaum, V. M.; Grabowski, J. J.; DePuy, C. H. J.
Am. Chem. Soc. 1983, 105, 5185-5192.
(14) Grabowski, J. J.; DePuy, C. H.; Bierbaum, V. M. J. Am. Chem.
Soc. 1983, 105, 2565-2571.
(12) Chabinyc, M. L.; Brauman, J. I. J. Phys. Chem. A 1999, 103, 9163-
9166.
10.1021/ja000806z CCC: $19.00 © 2000 American Chemical Society
Published on Web 08/25/2000

8740 J. Am. Chem. Soc., Vol. 122, No. 36, 2000
Chabinyc and Brauman
Scheme 1
Scheme 2
Table 1. Gas-Phase Acidities for the Reaction AH h A- + H+
AH
∆G° (kcal/mol)
∆H° (kcal/mol)
CH3OH
CD3OH
CF3H
HCCH
p-CH3-C6H4CCH
C6H5CCH
375.1a
375.5b
370.5b
369.8c
365.3d
364.2b
381.5
381.9
378.0
377.8
372.8
371.7
water and deuterium oxide, respectively.15 The products were purified
by a trap-to-trap vacuum distillation. Dimethyl peroxide was synthesized
by a literature procedure.16 The product was purified by a trap-to-trap
vacuum distillation. d3-Methyl formate, HCOOCD3, was synthesized
by a literature procedure.17 The product was purified by distillation
and preparative gas chromatography using a Hewlett-Packard Series
6890 GC with a carbowax column. d3-Methyl nitrite, CD3ONO, was
synthesized in situ by Bartmess’s technique using CD3OH and
isoamylnitrite.18 Phenylacetylene and p-tolylacetylene (4-ethynyltoluene)
were obtained from Aldrich. They were purified before use by
preparative gas chromatography. Tetrafluoromethane, CF4, was obtained
from Matheson. All other chemicals were obtained from Aldrich
Chemical and used without further purification.
a Experimental values from ref 28. b Reference 29 re-anchored to
data in ref 30. c Reference 30. d Reference 31.
found to be independent of ejected ion. The equilibrium constant
obtained at several pressure ratios was found to be in good agreement.
The major source of error in these experiments is the accuracy of the
measurement of the absolute pressures. We have assigned the error in
the equilibrium constants based on our estimate for the error in the
absolute pressure readings ((20%). The relative values should be more
accurate because the relative pressure errors should be smaller ((10%).
-
Kinetics. The rate constants for proton transfer between CF3 and
Instrumentation. All experiments were performed using an IonSpec
Fourier transform ion cyclotron resonance (FT-ICR) spectrometer.
Details of the spectrometer have been given previously.19 An additional
Varian leak valve was added to perform experiments requiring five
reagent gases. The magnetic field strength was 0.6 T. The temperature
in the cell is estimated to be 350 K. Background pressures were on the
order of 2.0-5.0 × 10-9 Torr and operating pressures ranged from 0.7
to 3.0 × 10-6 Torr. Pressure measurements were made using an ion
gauge (Granville Phillips 330) that was calibrated against a capacitance
manometer (MKS 170 Baratron with a 315BH-1 sensor). We estimate
the absolute pressure measurements to have an error of (20%.
Ion-Molecule Chemistry. Acetylide-fluoroform complexes,
RCC-‚HCF3, were synthesized through a sequence of ion-molecule
reactions. The Riveros reaction was used to synthesize methanol-
methoxide, CH3O-‚HOCH3 (Scheme 1).20,21 Acetylene, phenylacetylene,
and p-tolylacetylene, RCCH, were observed to undergo a reversible
exchange reaction with methanol-methoxide to form RCC-‚HOCH3
complexes (eq 3, in Scheme 1). The RCC-‚HOCH3 complexes then
underwent a reversible solvent exchange reaction with fluoroform (eq
4, Scheme 1).
acetylene and phenylacetylene were measured. CF3- was generated from
electron ionization of CF4. The measured rate constants are an average
of at least three trials at several pressure ratios on different days. The
reaction was measured over at least two half-lives at each pressure.
The main error in the rate constants is due to the absolute pressure
readings and we have therefore assigned an error of (20% to the
measured rate constants.
Results
Complexation Energies. The complexation energies (eq 7,
Scheme 2) of the substituted acetylide-fluoroform complexes,
RCC-‚HCF3, were measured relative to substituted acetylide-
methanol complexes, RCC-‚HOCH3. We previously determined
these values in another study (see Tables 1 and 2).12 The scale
is anchored to the complexation energy of methanol-methoxide.
The absolute values of its free energy of binding, ∆G°complex
,
and the enthalpy of binding, ∆H°complex, were determined by
Meot-Ner22 and Kebarle.23 We used Meot-Ner’s value here for
consistency with our previous work. The free energies of
complexation of the acetylide-fluoroform complexes were
determined from the thermochemical cycle in Scheme 2.24
The complexation energy of fluoroform-acetylide was
determined relative to deuterated methanol, CD3OH, because
of the difficulty of overlapping mass peaks. We determined the
equilibrium complexation energy of deuterated methanol-
methoxide, CD3O-‚HOCD3, on a scale relative to methanol-
methoxide using our instrumentation. Our value for the com-
plexation energy of CD3O-‚HOCD3 is consistent with the value
obtained from kinetic measurements by Bierbaum, but is 0.5
kcal/mol larger (less stable).25 We used our value because it
Appreciable amounts of the trimeric complex, CH3O-‚(HOCH3)2
(m/z ) 95), were observed at long delay times (∼2 s) under the reaction
conditions used. We were unable to resolve this peak from the
HCC-‚HCF3 (m/z ) 95), peak using our instrumentation. Therefore,
CD3OH was substituted for CH3OH during the studies of HCC-
complexes. d3-Methyl nitrite was used as the CD3O- source and
deuterated methyl formate, HCOOCD3, was used in place of methyl
formate.
Equilibrium Measurements. All equilibrium measurements were
obtained as an average of at least five trials at several pressure ratios
on different days. At the pressures used, the system reached equilibrium
in ∼1-2 s. Several methods were used to test whether a true equilibrium
had been achieved. After a constant ratio of ion intensities was obtained,
one species was ejected and the reaction was followed in time until a
constant ratio was reached again. The final ratio of ion intensities was
(22) Meot-Ner, M.; Sieck, L. W. J. Am. Chem. Soc. 1986, 108, 7525-
7529.
(15) Vogel, A. I. Practical Organic Chemistry; John Wiley & Sons
Inc.: New York, 1956.
(16) Hanst, P. L.; Calvert, J. G. J. Phys. Chem. 1959, 63, 104-110.
(17) Stevens, W.; vanEs, A. Recl. TraV. Chim. Pays-Bas 1964, 83, 1287-
1295.
(18) Caldwell, G.; Bartmess, J. Org. Mass. Spectrom. 1982, 17, 19-20.
(19) Zhong, M.; Brauman, J. I. J. Am. Chem. Soc. 1996, 118, 636-641.
(20) Blair, L. K.; Isolani, P. C.; Riveros, J. M. J. Am. Chem. Soc. 1973,
95, 1057-1060.
(21) DePuy, C.; Grabowski, J. J.; Bierbaum, V. M.; Ingemann, S.;
Nibbering, N. M. M. J. Am. Chem. Soc. 1985, 107, 1093-1098.
(23) Paul, G. J. C.; Kebarle, P. J. Phys. Chem. 1990, 94, 5184-5189.
(24) Previous work demonstrated that deprotonation of the acetylenic
hydrogen is both kinetically and thermodynamically more favorable than
the benzylic hydrogens. Therefore, the anion in the p-tolylacetylene
complexes was assumed to be the acetylide isomer. Also, the complexation
energy of benzyl isomer is expected to be much lower, because the
structurally similar benzyl-methanol complexes have been calculated to
be weakly bound ∼12 kcal/mol. Chabinyc, M. L., Brauman, J. I. J. Am.
Chem. Soc. 2000, 122, 3371-3378.
(25) Barlow, S. E.; Dang, T. T.; Bierbaum, V. M. J. Am. Chem. Soc.
1990, 112, 6832-6838.

Unusual Ionic Hydrogen Bonds
J. Am. Chem. Soc., Vol. 122, No. 36, 2000 8741
Table 2. Measured Thermochemical Values for Equilibrium Reactions
reaction
∆G° (kcal/mol)
∆H° (kcal/mol)
CH3O-‚HOCH3 + CD3OH h CH3OH + CH3O-‚HOCD3
0.0
1.0
1.7
2.1
2.8
2.2
0.5
0.5
2.7
2.1
2.8
2.2
CH3O-‚HOCD3 + CD3OH h CH3OH + CD3O-‚HOCD3
CD3O-‚HOCD3 + HsCtCsH h CD3OH + HsCtC-‚HOCD3
HsCtC-‚HOCD3 + HCF3 h CD3OH + HsCtC-‚HCF3
p-CH3‚C6H5-CtC-‚HOCH3 + HCF3 h CH3OH + p-CH3-C6H5-CtC-‚HCF3
C6H5sCtC-‚HOCH3 + HCF3 h CH3OH + C6H5sCtC-‚HCF3
Table 3. Thermochemical Values for Complexation Energies for
the Reaction B- + HA h B-‚HA
Table 5. Deuterium Fractionation Factors for Fluoroform-Acetylide
Complexes RsCtC-‚ + HCF3 + DCF3 h RsCtC-‚DCF3 + HCF3
B-‚HA
∆G° (kcal/mol)
∆H° (kcal/mol)
RCCH
Keq
CH3O-‚HOCH3
-19.4a
-28.8a
HCCH
C6H5CCH
0.94 ( 0.1
0.93 ( 0.1
CH3O-‚HOCD3
-19.4,b -19.6c
-18.8,b -19.4c
-11.6d
-28.3,b -28.5c
-28.2,b -28.8c
-21.6d
CD3O-‚HOCD3
HsCtC-‚HOCH3
HsCtC-‚HOCD3
p-CH3‚C6H4sCtC-‚HOCH3
C6H5sCtC-‚HOCH3
-11.3b
-21.4b
maximum efficiency to be less than 1.0% based on our detection
limits and the pressures used.
-11.1d
-21.6d
-11.0d
-21.5d
a All values measured at 350 K. Reference 22. b This work. c Ref-
erence 25. d Reference 12.
RsCtC-‚HCF3 + RsCtCD N RsCtC-‚DCF3 +
RsCtCH (8)
Table 4. Complexation Energies for Fluoroform and Acetylides
for the Reaction RsCtC- + HCF3 h RsCtC-‚HCF3
where, R ) D, C6H5.
RCC-
∆G° (kcal/mol)
∆H° (kcal/mol)
In contrast, deuterated fluoroform, DCF3, was found to
reversibly exchange into the RCC-‚HCF3 complexes (eq 9). For
both phenylacetylene and acetylene, the process was efficient.
Only one proton exchanged on the time scale examined (∼3 s)
for the acetylide-fluoroform complex, HCC-‚HCF3. The
deuterium fractionation factors (the equilibrium constant for eq
9) for phenylacetylide and acetylide-fluoroform complexes
were measured. The ratio of DCF3/HCF3 was measured in situ
HCC-
-9.2
-8.3
-8.8
-19.3
-18.8
-19.3
p-CH3 C6H4CC-
C6H5CC-
a All values measured at 350 K. The absolute error in each values is
estimated to be (1 kcal/mol. The relative error should be less than
(0.5 kcal/mol.
+
by the ratio of DCF2+/HCF2 in a positive ion spectra. This
will have a consistent relative error with our other measurements
and because it relies on a direct equilibrium measurement rather
than measurements of multiple rate constants. The complexation
energy of RCC-‚HOCD3 was determined relative to our value
of the complexation energy of CD3O-‚HOCD3. All experimen-
tally determined values are listed in Tables 2-4.
method for ascertaining neutral ratios of DCF3/HCF3 has been
used previously and established to be reliable.11 The fraction-
ation values are listed in Table 5.
RsCtC-‚HCF3 + DCF3 h RsCtC-‚DCF3 + HCF3 (9)
To derive ∆H°complex from ∆G°complex, a value of ∆S°complex
is needed. Obtaining accurate estimates of the entropies for ion-
molecule complexes is difficult. The vibrational modes with
the greatest contribution to the entropy are the six low-frequency
modes created upon complexation. These soft modes are very
anharmonic and are generally not reproduced accurately by
standard ab initio harmonic frequency calculations.26 Conse-
quently, we chose to assume that ∆S° of eq 5 in Scheme 2 is
due only to the symmetry of the molecules.27 We expect that
this will introduce an error of no more than (3 eu in the value
of ∆S°complex and (1.0 kcal/mol in the derived value of
∆H°complex. The relative error between values should be smaller
because ∆S°complex for these structurally similar complexes
should be nearly equivalent. All values were calculated assuming
a temperature of 350 K. The resulting values are listed in Table
4.
Solvent Switching/Isotopic Exchange Experiments. Incor-
poration of a deuterated neutral molecule into the complex
followed by solvent switching or by proton exchange and loss
of a protiated neutral molecule would be indicated by a change
in mass. Isotopic exchange reactions were studied for both
deuterated fluoroform and deuterated acetylenes. Deuterated
acetylenes were not observed to exchange a deuterium with the
RCC-‚HCF3 complexes on the time scale and attainable
pressures of our experiments (eq 8). We can estimate the
where R ) D, C6H5.
We also examined the exchange reaction of fluoroform with
RCC-‚DOCH3. The only product for acetylide and phenyl-
acetylide complexes was protiated RCC-‚HCF3 (eq 10).
where R ) H, Ph, Me-Ph.
Kinetics. The rate constants for proton transfer from acetylene
and phenylacetylene to CF3 were measured (eq 11). The
capture rate constants were calculated using the Su-Chesnavich
model.33 The efficiency of the reaction was calculated by eq
12. All values are listed in Table 6.
-
(28) Meot-Ner, M.; Sieck, L. W. J. Phys. Chem. 1986, 90, 6687-6690.
(29) Bartmess, J. E. In NIST Chemistry WebBook, NIST Standard
Reference Database 69; Mallard, W. G., Linstrom, P. J., Eds.; National
Institute of Standards and Technology: Gaithersburg, MD, 1998.
(30) Ervin, K. M.; et al. J. Am. Chem. Soc. 1990, 112, 5750-5759.
(31) Chabinyc, M. L.; Brauman, J. I. J. Am. Chem. Soc. 2000, 122, 3371-
3378.
(26) East, A. L. L.; Radom, L. J. Chem. Phys. 1997, 106, 6655-6674.
(27) Benson, S. W. Thermochemical Kinetics; 2nd ed.; Wiley: New
York, 1976.
(32) Chabinyc, M. L.; Craig, S. L.; Regan, C. K.; Brauman, J. I. Science
1998, 279, 1882-1886.
(33) Su, T.; Chesnavich, W. J. J. Chem. Phys. 1982, 76, 5183-5185.

8742 J. Am. Chem. Soc., Vol. 122, No. 36, 2000
Table 6. Kinetic Data for the Proton Transfer Reaction
Chabinyc and Brauman
CF3 + RsCtCsH f CF3H + RsCtC-
-
µ-VTST RRKM
barrier height
(kcal/mol)
kobs × 10-9 kcap × 10-9
RCCH
HCCH
C6H5CCH 0.29 ( 0.06
(cm3s-1
)
(cm3s-1
)
efficiency Φ
0.20 ( 0.04
0.98
1.43
0.20
0.20
-8.2
-7.2
a kcap was calculated using the Su-Chesnavich collision capture
model (ref 33) with the following parameters: dipole moment in D,
molecular polarizability in Å3, geometric mean of moments of inertia
in amu Å2: HCCH (0.0, 3.3, 13.9), C6H5CCH (0.66, 13.0, 228.6).
CF3- + RsCtCH f RsCtC- + HCF3
(11)
(12)
Figure 1. Potential energy surface for the reaction of CF3- and HCCH.
Energies were calculated at the MP2/6-311++G**//HF-6-311++G**
level of theory. Zero point energies were obtained from ab initio
vibrational frequencies scaled by 0.89.
kobs
Φ )
kcap
Microcanonical variational transition state (µ-VTST) Rice-
Ramsperger-Kassel-Marcus (RRKM) calculations were per-
formed to derive the proton-transfer transition state barrier
height.34 Calculations were performed with the HYDRA pro-
gram.35 The reaction was modeled using a double well potential
and the steady-state approximation. The parameters for the
RRKM calculations were determined from ab initio calculations.
The height of the central barrier was adjusted until the RRKM
efficiency matched the experimental efficiency. The experi-
mentally derived and ab initio barriers are in good agreement
(see Table 6, Figures 1 and 2).
Theoretical Methods. Molecular orbital calculations were
performed using Gaussian94.36 Calculations on the acetylide-
fluoroform system were done at the MP2/6-311++G**//HF/
6-311++G** level of theory. All stationary points were
confirmed by vibrational analyses at the HF/6-311++G** level.
Calculations on the phenylacetylide-fluoroform system were
done at the MP2/6-311++G**//HF/6-31+G* level of theory.
Vibrational frequencies were calculated at the HF/6-31+G* level
of theory. Thermochemistry was calculated using standard
statistical mechanics with ab initio vibrational frequencies scaled
by 0.89.37 All results are summarized in Table 7 and in Figures
1 and 2.
Figure 2. Potential energy surface for the reaction of CF3- and C6H5-
CCH. Energies were calculated at the MP2/6-311++G**//HF-6-31+G*
level of theory. Zero point energies were obtained from ab initio
vibrational frequencies scaled by 0.89.
Table 7. Ab Initio Results for Complexation Energies
∆H°
Discussion
reaction
(kcal/mol)
Background. The hydrogen bond is one of the most widely
studied molecular interactions. A hydrogen bond is defined as
an intermolecular (or intramolecular) interaction specifically
involving a proton donor, A-H, and a proton acceptor, B.38
Ionic hydrogen bonds are observed in solvent-ion interactions,39
intermediates in proton-transfer reactions,40 and in enzymatic
active sites.41 Their existence is substantiated by characteristic
spectroscopic features, such as intense IR bands and unusual
NMR shifts, and structural features, such as short contact
distances (smaller than van der Waals radii).1,2 Because solvation
RsCtC- + HCF3 h HsCtC-‚HCF3
C6H5-CtC- + HCF3 h C6H5-CtC-‚HCF3
HsCtC- + HsCtCsH h HsCtC-‚HsCtCsH
CF3- + HCF3 h CF3-‚HCF3
-16.6a
-16.0b
-10.1a
-15.0a
a Energies calculated at the MP2/6-311++G**//HF/6-311++G**
level of theory with vibrational corrections from 0.89 scaled HF/6-
311++G** frequencies. b Energies calculated at the MP2/6-311++G**//
HF/6-31+G* level of theory with vibrational corrections from 0.89
scaled HF/6-31+G* frequencies.
has dramatic effects on the energetics and reactivity of ions,32
it is difficult to understand the properties of hydrogen bonds to
ions in the condensed phase.
(34) Gilbert, R. G.; Smith, S. C. Theory of Unimolecular and Recom-
bination Reactions; Blackwell Scientific: Oxford, 1990.
(35) Wladkowski, B. D.; Lim, K. F.; Allen, W. D.; Brauman, J. I. J.
Am. Chem. Soc. 1992, 114, 9136-9153.
(36) Frisch, M. J.; et al. Gaussian94 ReV C.3; Gaussian, Inc.: Pittsburgh,
PA, 1995.
B- + HA h B-‚HA
(13)
(37) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502-16513.
(38) Pimentel, G. C.; McClellan, A. L. The Hydrogen Bond; Freeman:
San Francisco, 1960.
(39) Ions and Ion Pairs in Organic Reactions; Szwarc, M., Ed.; Wiley-
Interscience: New York, 1972.
(40) Bell, R. P. The Proton in Chemistry; Cornell University Press:
Ithaca, NY, 1973.
(41) Jencks, W. P. Catalysis in Chemistry and Enzymology; Dover
Publications: New York, 1987.
Ionic gas-phase complexes provide a measure of the intrinsic
stability of hydrogen-bonded complexes. We focus here on
hydrogen bonds to anions.42 Anionic complexes have been
(42) Positive ion complexes have been widely examined as well. For
recent work, see: Meot-Ner, M.; Sieck, L. W.; Koretke, K. K.; Deakyne,
C. A. J. Am. Chem. Soc. 1997, 119, 10430-10438.

Unusual Ionic Hydrogen Bonds
J. Am. Chem. Soc., Vol. 122, No. 36, 2000 8743
studied by mass spectroscopy,3,7,22,43 by photodissociation
spectroscopy,44 by electron photodetachment spectroscopy,45,46
and by multiple photon infrared dissociation spectra.47 The
complexation energy (∆H° for eq 13) is generally referred to
as the hydrogen bond strength. However, this definition does
not provide any indication of how much extra stabilization the
hydrogen bond provides relative to a simple ion-dipole
complex. Simple ion-dipole complexes in the gas phase can
be bound by ∼10-12 kcal/mol, whereas hydrogen-bonded
complexes can be bound by ∼25-35 kcal/mol.
RO-‚HCF3, we found that the structure of the complex is not
determined by the acid-base chemistry.11 That is, the complex
remains RO-‚HCF3 even when proton transfer from HCF3 to
the alkoxide is favorable in the separated reactants. The results
from studies of deuterium exchange proton-transfer reactions
suggest that other carbanions may behave similarly.13,14 We
studied complexes of acetylides with methanol, RCC-‚HOCH3.
In this system the complexation energy is nearly constant with
varying acetylides over ∼8 kcal/mol range of basicities.12
However, the corresponding plot with a constant acetylide and
varying alcohols, C6H5CC-‚HOR, has a slope of ∼0.5. This
work indicates that the frequent assumption that the trend in
complexation energy is related to the shape of the potential
energy surface does not hold. These results suggested that
examination of the charge density of the separated ion and
neutral is more useful than acid-base properties to interpret
complexation energy data.
Our results on carbanions and carbon acids suggested that
we could synthesize strongly bound carbon acid-carbanion
complexes. Here we report our work on the synthesis of
complexes of substituted acetylides with fluoroform, RCC-‚HCF3.
We measured their stability and verified their putative structures.
These complexes are unusual because neither of the ho-
modimers, RCC-‚HCCR and CF3-‚HCF3, are observed to have
similar stability. Unlike alcohols, which are good donors and
whose conjugate bases are good acceptors, only one species in
each homodimer is a good hydrogen bond donor (or acceptor).
Structure of RCC-‚HCF3 Complexes. The structure of
hydrogen-bonded complexes is frequently assumed to be related
to the acid-base chemistry of the separated ion and neutral.
Because we have shown that this assumption does not always
hold, we investigated the structure of the assumed RCC-‚HCF3
complexes by deuterium exchange experiments. These experi-
ments are completely consistent with the RCC-‚HCF3 structure
for R ) H and C6H5 and also provide information about the
potential energy surface.
We performed deuterium exchange experiments using both
acetylenes and fluoroform as the deuterium source. The reactions
of RCCD (R ) H, C6H5) with the acetylide-fluoroform
complexes were examined. If the complex resembled
RCCH‚-CF3, we would expect a simple solvent switching
reaction with RCCD. We do not observe incorporation of
deuterium from RCCD.56 Therefore, the RCCH‚-CF3 structure
is unlikely because solvent switching reactions are generally
kinetically efficient.25 These results suggest that the complex
has the structure RCC-‚HCF3. To test this hypothesis, we
examined the exchange reaction with DCF3. We observe
efficient exchange to incorporate DCF3, suggesting a simple
solvent switching mechanism and supporting the RCC-‚HCF3
structure. Although we cannot rule out the possibility of proton
transfer in the collision complex, we suggest that it probably
does not occur. The lack of deuterium incorporation from the
RCCD exchange experiments suggests that the RCC-‚HCF3
complex must pass through an unfavorable intermediate to
incorporate deuterium.
The relation of the hydrogen bond strength to acid-base
properties has been widely studied in the gas phase.3,22,43 In
general, strong acids appear to be good H-bond donors and
strong bases appear to be good H-bond acceptors. However,
polarity is also important. For instance, toluene is as acidic as
methanol but is a poor hydrogen bond donor.48 Because
hydrogen-bonded complexes are intermediates in proton-transfer
reactions, their structure is assumed to reflect the thermodynamic
endpoints of the reaction. That is, the structure is A-‚HB if
AH is the stronger acid (compared to HB), and AH‚B- if BH
is the stronger acid.49 This assumption has been tested for several
systems by isotopic exchange,50,51 electron photodetachment,52
and computation.53 A linear relationship is frequently observed
between the difference in acidity of the two partners in the
complex and the complexation energy.54 These relationships
have been suggested to provide information about the potential
surface near the complexes or the degree of proton transfer in
the complex. A value of 0.5 has been suggested to indicate that
the proton is equally shared between the two molecules in the
complex.55 The connections between acidity and hydrogen bond
stability have generally derived using a limited range of
structural components. Therefore, it is uncertain how general
these empirical rules are.
We recently studied complexes with carbanions as H-bond
acceptors and carbon acids as H-bond donors in an attempt to
better understand the relation between acid-base properties, the
shape of the potential energy surface, and the stability of
hydrogen-bonded complexes. These systems demonstrate that
previous assumptions about hydrogen-bonded complexes are not
completely general. For complexes of alkoxides with fluoroform,
(43) Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc. 1987, 109 (9),
6230-6236.
(44) Ayotte, P.; Bailey, C. G.; Johnson, M. A. J. Phys. Chem. A 1998,
102, 3067-3071.
(45) Gatev, G. G.; Zhong, M.; Brauman, J. I. J. Phys. Org. Chem. 1997,
10, 531-536.
(46) Bradforth, S. E.; Arnold, D. W.; Metz, R. B.; Weaver, A.; Neumark,
D. M. J. Phys. Chem. 1991, 95, 8066-8078.
(47) (a) Peiris, D. M.; Riveros, J. M.; Eyler, J. R. Int. J. Mass. Spectrom.
Ion Proc. 1996, 159, 169-183. (b) Weiser, P. S.; Wild, D. A.; Bieske, E.
J. Chem. Phys. Lett. 1999, 299, 303-308. (c) Weiser, P. S.; Wild, D. A.;
Bieske, E. J. J. Chem. Phys. 1999, 110, 9443-9449.
(48) We have been unable to synthesize the benzyl-methanol complex
under our conditions, suggesting it is weakly bound. Ab initio calculations
suggest the complex is only bound by ∼12 kcal/mol. Gatev, G. G.; Zhong,
M.; Brauman, J. I. J. Phys. Org. Chem. 1997, 10, 531-536
(49) Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc. 1987, 110, 1087-
1093.
(50) Ellenberger, M. R.; Farneth, W. E.; Dixon, D. A. J. Phys. Chem.
1981, 85, 4-7.
(51) Wilkinson, F. E.; Peschke, M.; Szulejko, J. E.; McMahon, T. B.
Int. J. Mass. Spectrom. Ion Proc. 1998, 175, 225-240.
(52) Mihalick, J. E.; Gatev, G. G.; Brauman, J. I. J. Am. Chem. Soc.
1996, 118, 12424-12431.
The HCC-‚HCF3 complex is a more complicated case
because it has two possible sites of exchange. Incorporation of
DCF3 is facile. In the resulting complex, HCC-‚DCF3, the
remaining acetylenic proton could still be exchanged. One
plausible mechanism involves the isomerization of the putative
HCC-‚DCF3 complex to DCC-‚HCF3 followed by simple
(53) Wladkowski, B. D.; East, A. L. L.; Mihalick, J. E.; Allen, W. D.;
Brauman, J. I. J. Chem. Phys. 1993, 100, 2058-2088.
(54) These relationships are also observed in the condensed phase. For
recent work, see: Shan, S.-O.; Loh, S.; Herschlag, D. Science 1996, 272,
97-101.
(56) Proton transfer could occur between the acetylene and the acetylide
in the collision complex, followed by simple solvent switching. We are
unable to determine if this pathway is operative because there is no change
in mass of the resulting complex.
(55) Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc. 1983, 105, 2944-
2950.

8744 J. Am. Chem. Soc., Vol. 122, No. 36, 2000
Chabinyc and Brauman
solvent exchange of DCF3. We know that the solvent exchange
process should be facile. Therefore, if complexes of DCC-‚HCF3
were present, they would exchange with DCF3. We do not
observe DCC-‚DCF3, indicating that the isomerization-
exchange reaction does not occur during the collision event.
This result further suggests that processes involving the
HCCH‚-CF3 structure are not energetically favorable.
In addition to the results from the acetylide-fluoroform
complexes, the reaction of RCC-‚DOCH3 with HCF3 produces
RCC-‚HCF3 only. Plausible mechanisms for formation of the
nonobserved product, RCC-‚DCF3, involve activated intermedi-
ates with CF3-. All of our results are consistent with the
hypothesis that all intermediates containing CF3- are energeti-
cally unfavorable.
Stability and Potential Energy Surface (PES). All three
RCC-‚HCF3 complexes have complexation energies of ∼-19
kcal/mol. These complexes are the most stable complexes of
carbanions and carbon acids that have been observed.57 Note
that although we observed the heterodimeric complex of the
comparably acidic components, RCC-‚HCF3, we do not observe
formationofthesymmetric,homodimericcomplexes,RCC-‚HCCR
and CF3-‚HCF3, under our conditions. We estimate that these
homodimeric complexes have complexation energies that are
weaker than -16 kcal/mol.58 We performed ab initio calcula-
tions on HCC-‚HCCH and CF3-‚HCF3 to predict their stability.
They are calculated to have complexation energies of -10 and
-15 kcal/mol, respectively. Both are weaker than the calculated
complexation energies for RCC-‚HCF3 (see Table 7). This result
is markedly different from the behavior of alcohol-alkoxides,
where the homodimers and heterodimers have comparable
stabilities, and the behavior of CH3CN, where neither the
homodimer nor its heterodimers are exceptionally stable.8
proton-transfer reaction of delocalized carbanions, the efficiency
indicates that an energetic barrier exists on the potential surface.
Microcanonical variational transition state RRKM modeling of
the proton-transfer efficiencies was performed to determine the
height of the central barrier. For acetylene and phenylacetylene,
the barriers were calculated to be ∼8 and 7 kcal/mol below the
entrance channels, respectively. These values are in reasonable
agreement with the calculated barriers (see Figures 1 and 2).
WemeasuredthedeuteriumfractionationfactorsofRCC-‚HCF3
with HCF3/DCF3 (eq 9) to learn about the shape of the potential
surface near the complex. The fractionation factors are nearly
unity for both (see Table 5). ∆G° for the reaction is ap-
proximately equal to the difference in zero point energies
between the complexes and H/D fluoroform (eq 14). Our
isotopic exchange experiments demonstrate that the structure
of both complexes is RCC-‚HCF3. In this case, a fractionation
factor of ∼1 indicates that the C-H stretch in fluoroform is
not strongly perturbed in the complex.61,62 Using our measured
complexation energies for RCC-‚HCF3, the proton-transfer
barrier is ∼10 kcal/mol higher than the HCC-‚HCF3 complex
and ∼17 kcal/mol higher than the C6H5CC-‚HCF3 complex.
Both results suggest that the proton transfer coordinate looks
essentially like HCF3 at the RCC-‚HCF3 minima for both
systems.
∆G°eq ≈ (ZPERCC ‚DCF - ZPERCC ‚HCF ) - (ZPEDCF
-
-
-
3
3
3
ZPE
) (14)
HCF3
Charge Distributions and H-Bond Stability. The majority
of gas phase anionic complexes for which data are available
are structurally similar pairs. In these strongly bound systems,
the acids and their conjugate base are good hydrogen bond
donors and acceptors, respectively. For example, both alcohols,
ROH, and alkoxides, RO-, form strongly bound hydrogen
bonded complexes with a variety of partners.5 For weakly bound
systems, many acids and their conjugate base are poor hydrogen
bond donors and acceptors. For example, neither acetonitrile
nor its conjugate base have been observed to form strongly
bound hydrogen-bonded complexes.
We have also calculated the potential energy surfaces proton
transfer between CF3- and acetylene and phenylacetylene (See
Figures 1 and 2). Although the well depths of the RCC-‚HCF3
complexes are underestimated by ∼2 kcal/mol, the agreement
is still reasonable.59 The difference in energy between the
RCCH‚-CF3 complex and the tautomeric RCC-‚HCF3 complex
is predicted to be greater than the exothermicity of the overall
proton transfer reaction for both R)H and phenyl. The result
-
is especially striking for the CF3 + acetylene surface where
The substituted acetylide-fluoroform system presents a case
where the hydrogen bonding character of the acid and its
conjugate base are substantially different. Both HCCH and HCF3
have similar acidities (within <1 kcal/mol) and show efficient
proton transfer reactions such as those for the alcohols. However,
they are completely different in regard to their ability to
participate in hydrogen bonds. HCF3 is a good hydrogen bond
the difference in complexes is ∼9 kcal/mol while the reaction
is only calculated to be ∼2 kcal/mol exothermic. These results
agree with our isotopic exchange experiments that suggest that
-
complexes where CF3 is the hydrogen bond acceptor are
energetically unfavorable.
To further explore the stationary points on the potential energy
surfaces, we measured the rate constants for the two proton-
transfer reactions. The proton-transfer reactions of acetylene and
phenylacetylene with CF3- were both ∼20% efficient (See Table
6).60 Although the reactions are relatively rapid compared to
-
donor to both anions11 and neutrals.63 In contrast, CF3 is a
poor hydrogen bond acceptor. The stability of this ion is due to
polarization of the charge away from the carbon into the fluorine
-
atoms. On the basis of electrostatics alone, CF3 should be a
poor acceptor compared to an anion with highly localized
charge.64 Although acetylene, HCCH, has been observed to form
hydrogen bonds to neutrals in the gas phase,65 it is a poor
hydrogen bond donor toward anionssno complexes of the
(57) Only two carbanion-carbon acid complexes have been examined.
CH2CN-‚CH3CN has ∆Hcomplex ) -12.8 kcal/mol and C5H5-‚CD3CN has
∆Hcomplex ) -15.5 kcal/mol (ref 8).
(58) Estimates are based on the pressure range obtainable and typical
signal-to-noise ratios.
(59) The origin of the underestimation is uncertain. We found similar
underestimations for RCC-‚HOCH3 complexes. Calculations at the MP2/
aug-cc-p-VDZ//HF/aug-cc-p-VDZ level gave similar results, ruling out the
possibility of a basis-set-dependent error. One possibility is basis set
superposition error, but this error would tend to make the complexation
energies less exothermic.
(60) Both reactions have the same efficiency despite the higher exother-
micity of the CF3- + phenylacetylene reaction. We observed a similar result
with the proton-transfer reaction of methoxide with these acetylenes. It is
unclear whether the origin is a true thermodynamic barrier or dynamic effects
during the reaction. Chabinyc, M. L.; Brauman, J. I. ref 31.
(61) Huskey, W. P. J. Am. Chem. Soc. 1996, 118, 1663-1668.
(62) Kreevoy, M. M.; Liang, T. M. J. Am. Chem. Soc. 1980, 102, 3315-
3322.
(63) Fraser, G. T.; Lovas, F. J.; Suenram, R. D.; Nelson, D. D. N., Jr.;
Klemperer, W. J. Chem. Phys. 1986, 84, 5983-5988.
(64) Clearly, charge distributions are perturbed in these strongly bound
complexes and charge-transfer effects may be important. However, our data
suggest that the charge-dipole component to the interaction is the dominant
factor. See also: Weinhold, F. THEOCHEM 1997, 398-399, 181-197.
(65) Fraser, G. T.; Leopold, K. R.; Klemperer, W. J. Chem. Phys. 1984,
80, 1423-1426.

Unusual Ionic Hydrogen Bonds
J. Am. Chem. Soc., Vol. 122, No. 36, 2000 8745
structure HCCH‚B- have been observed experimentally and
calculations indicate that they would be weakly bound.66 In
contrast, acetylide, HCC-, is a good hydrogen bond acceptor.
The charge is highly localized in a sigma orbital on the terminal
carbon.67 The anion forms stable complexes with both alcohols
and fluoroform.
The results for the acetylide-fluoroform system demonstrate
clearly that matched acidities do not imply any special hydrogen
bonding interaction as has been suggested by some.68 Both of
the homodimers, HCC-‚HCCH and CF3-‚HCF3, are less stable
than the heterodimer, HCC-‚HCF3, based on experimental
evidence, and they are both calculated to be weakly bound. In
addition, all of our isotopic exchange data suggests that the
structure HCCH‚-CF3 is energetically unfavorable. All of the
weak complexes contain a good donor (or acceptor) with a poor
acceptor (or donor). However, when a good donor, CF3H, and
a good acceptor, HCC-, are brought together the complex is
strongly bound (see Figure 1). This suggests that the stability
of the homodimeric complex in no way determines the stability
of heterodimeric complexes. As a consequence, Marcus-type
additivity rules for complexation energies based on homodimers
and heterodimers are certain to fail for systems where the
hydrogen bonding properties of an acid and its conjugate base
differ. This result has been suggested previously based on
limited computational studies, but we believe our results are
the first to demonstrate this hypothesis experimentally.69,70
Figure 3. Complexation energies, ∆H°complex, of RCC-‚HCF3 vs the
gas-phase acidities, ∆H°acid, of RCCH.
We obtained the same result for a series of methanol-acetylide
complexes.12
Conclusions
The substituted acetylide-fluoroform complexes clearly
demonstrate the need to examine charge distributions rather than
acid-base chemistry to understand hydrogen bonding ability.
The acetylide-fluoroform system shows that the difference in
energy between the intermediate hydrogen-bonded complexes
can be large even when the proton transfer reaction between
the separated species is nearly thermoneutral. Few hydrogen-
bonded systems that have been studied so far have this type of
potential surface, but there are likely to be more.13,14,72
The relationships between the magnitude of the deuterium
fractionation factor and the difference in energy between wells
on a double-well proton-transfer surface have been verified.
Complexes with proton-transfer barriers that are large compared
to the energy of the complex produce fractionation factors near
unity. However, these results do not address how the magnitude
of the complexation energy affects the fractionation factor.
This work further demonstrates that the correlations between
acid-base thermochemistry and complexation energies observed
for many systems are not an intrinsic property of hydrogen-
bonded anionic complexes. These relationships are frequently
used to predict the strength of unknown hydrogen bonds and
appear to be correct in many cases.3,5 While these relationships
can be useful tools, caution should be used in applying them to
new systems. Careful analysis of the origin of these relationships
must be made before they can be applied to new systems.
The relation between the complexation energies and acid-
base properties also can be understood by examining charge
distributions.11,12 A plot of the complexation energy vs the
basicity of the acetylide shows almost zero slope (<0.1) over a
6 kcal/mol range (see Figure 3). The complexation energy shows
no extra stability for the acetylide-fluoroform complex where
the acidities are nearly matched. The complexation data show
that the substituent on the acetylide provides the same stabiliza-
tion to the free anion as it does to the complex. We believe
that the charge distribution in the acetylide determines the trend
in complexation energies. A simple model to explain this
behavior is to treat the substituent as a remote dipole.71 If the
charge distribution at the terminal carbon remained nearly
constant for all the acetylides, a dipolar substituent would
provide similar stabilization in both the free ion and the complex.
(66) Cybulski, S. M.; Scheiner, S. J. Am. Chem. Soc. 1987, 109, 4199-
4206.
(67) Wiberg, K. B.; Rablen, P. R. J. Am. Chem. Soc. 1993, 115, 9234-
9242.
Acknowledgment. We are grateful to the National Science
Foundation for support of this research. M.L.C. thanks the
National Science Foundation, the ACS Division of Organic
Chemistry, and the Stauffer Trust for fellowship support.
(68) Gerlt, J. A.; Kreevoy, M. M.; Cleland, W. W.; Frey, P. A. Chem.
Biol. 1997, 4, 259-267.
(69) Magnoli, D. E.; Murdoch, J. R. J. Am. Chem. Soc. 1981, 103, 7465-
7469.
(70) Scheiner, S.; Redfern, P. J. Phys. Chem. 1986, 90, 2969-2974.
(71) It is difficult to uncover the origin of substituent effects for acetylides
because there is no unambiguious theoretical method to separate π and σ
orbital effects. See: Schleyer, P. v. R. Pure Appl. Chem. 1987, 59, 1647-
1660.
JA000806Z
(72) de Beer, E.; Kim, E. H.; Neumark, D. M.; Gunion, R. F.; Lineberger,
W. C. J. Chem. Phys. 1995, 99, 13627-13636.