Electron-withdrawing trifluoromethyl groups in combination with hydrogen bonds in polyols: Bronsted acids, hydrogen-bond catalysts, and anion receptors

Article abstract of DOI:10.1021/ja4036384

Electron-withdrawing trifluoromethyl groups were characterized in combination with hydrogen-bond interactions in three polyols (i.e., CF ₃CH(OH)CH₂CH(OH)CF₃, 1; (CF₃) ₂C(OH)C(OH)(CF₃)₂

Full text of DOI:10.1021/ja4036384

Article
pubs.acs.org/JACS
Electron-Withdrawing Trifluoromethyl Groups in Combination with
Hydrogen Bonds in Polyols: Brønsted Acids, Hydrogen-Bond
Catalysts, and Anion Receptors
†
,‡
,†
Alireza Shokri, Xue-Bin Wang,* and Steven R. Kass*
†
Department of Chemistry, University of Minnesota, 207 Pleasant St SE, Minneapolis, Minnesota 55455, United States
‡
Chemical and Materials Sciences Division, Pacific Northwest National Laboratory, P.O. Box 999, MS K8-88, Richland, Washington
99352, United States, and Department of Physics, Washington State University, 2710 University Drive, Richland, Washington 99354,
United States
*
S Supporting Information
ABSTRACT: Electron-withdrawing trifluoromethyl groups were char-
acterized in combination with hydrogen-bond interactions in three
polyols (i.e., CF CH(OH)CH CH(OH)CF , 1; (CF ) C(OH)C-
3
2
3
3 2
(
OH)(CF ) , 2; ((CF ) C(OH)CH ) CHOH, 3) by pK measure-
3 2 3 2 2 2 a
ments in DMSO and H O, negative ion photoelectron spectroscopy
and binding constant determinations with Cl . Their catalytic behavior
2
−
in several reactions were also examined and compared to a Brønsted
acid (HOAc) and a commonly employed thiourea ((3,5-
(
CF ) C H NH) CS). The combination of inductive stabilization and hydrogen bonds was found to afford potent acids
3 2 6 3 2
which are effective catalysts. It also appears that hydrogen bonds can transmit the inductive effect over distance even in an
aqueous environment, and this has far reaching implications.
INTRODUCTION
■
In recent years the development and application of Brønsted
acids has emerged as a fast growing branch of chemistry. These
1
compounds can be broadly classified into two categories: those2
Figure 1. Trifluoromethyl group containing polyols.
with one acidic site, such as BINOL-derived phosphoric acids,
3
4
bis(sulfonyl)imides, N-triflyl phosphoric amides, and carbor-
5
6
anes, and those with two acidic hydrogens, such as thioureas,
including aqueous and DMSO pK determinations, chloride
a
7
8
biphenols, and TADDOL derivatives. These compounds are
often employed as catalysts but sometimes require large
loadings and are corrosive, harmful to plants and animals,
anion binding association constants in acetonitrile, and gas-
phase adiabatic electron detachment energies (ADEs) of their
conjugate bases. The catalytic abilities of these compounds
were also explored in a Friedel−Crafts alkylation and an
aminolysis of an epoxide. These experimental results were
supplemented with detailed computations as well.
9
and sensitive to heat. We recently introduced a promising new
variant that makes use of multiple hydrogen bonds to stabilize a
10
charged center and enhance the acidity. For example, 1,3,5-
pentanetriol (HOCH CH CH(OH)CH CH OH), a simple
2
2
2
2
aliphatic alcohol with three hydroxyl groups, was found to be
more acidic than 2,2,2-trifluoroethanol (TFE) in DMSO and
EXPERIMENTAL SECTION
General. All glassware, needles, syringes, and NMR tubes were
dried overnight in ovens at 110 °C and subsequently stored in a
desiccator containing rubber septa and charged with phosphorus
■
almost as acidic as phenol (i.e., the pK ’s of TFE, 1,3,5-
a
pentanetriol, and PhOH are 23.4, 19.7, and 18.0, respectively).
That is, the formation of two hydrogen bonds to the alkoxide
center in the conjugate base of the triol leads to a 10.6 pKa
acidification relative to isopropanol. A larger hydrogen-bonding
pentoxide. DMSO and DMSO-d were dried under vacuum (1.5 Torr)
6
over CaH at reflux for several hours and then were distilled under
2
these conditions. The resulting solvents were stored in dark vials over
Å molecular sieves that had been activated in a furnace at 320 °C for
1 day and then kept under an argon atmosphere for up to a few days.
Pentane was dried and distilled over P O and then used three times in
3
network in a heptaol ((HOCH CH CH(OH)-
2
2
CH CH ) COH) was found to result in a compound that is
2
2 3
2
5
more acidic than acetic acid (pK = 11.4 vs 12.3) and 21 orders
succession to rinse mineral oil away from a 30% suspension of
potassium hydride as part of the process for making the potassium salt
of dimsyl anion (i.e., CH SOCH K). Fresh solutions of dimsyl
a
1
1
of magnitude stronger than tert-butanol. Electron-with-
drawing groups can be incorporated into polyols of these
sorts and should lead to even stronger Brønsted acids. To
explore this possibility, three trifluoromethyl-containing polyols
3
2
Received: April 11, 2013
(1−3, Figure 1) were characterized by a number of means
©
XXXX American Chemical Society
A
dx.doi.org/10.1021/ja4036384 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
potassium were prepared daily by reacting KH with DMSO or DMSO-
d6 at room temperature over a 30 min period. All of the dry solvents
were routinely degassed immediately before use by bubbling dry argon
through them for ∼20 min. Chloroform-d was dried by storing it over
activated 4 Å molecular sieves, diol 2 was used as supplied (Matrix
Scientific), and triol 3 was prepared as previously described but was
purified by vacuum sublimation at 15 Torr and 70 °C rather than by
recrystallization. NMR spectra were recorded on Varian VI-300 and
VI-500 spectrometers at 295 K, and the chemical shifts are given in
parts per million (δ) relative to the residual solvent peak. Mass spectra
were obtained with a Bruker BioTof II electrospray ionization time-of-
flight mass spectrometer using polyethylene glycol 200 as an internal
standard.
solver add-on program for Excel to obtain the association equilibrium
constants. Representative data and graphical fits of the results are
provided in Tables S1 and S2 and Figures S1 and S2.
Aminolysis of Styrene Oxide. A solution of 0.093 g (1.0 mmol)
aniline, 0.12 g (1.0 mmol) styrene oxide, and 5 mol % catalyst (0.05
mmol) in a 3 dram vial was stirred under argon at 60 °C for the
indicated times. Reaction progress was monitored by TLC (6:1
hexanes/ethyl acetate) on 250 mm 60 F-254 silica gel plates, and upon
1
2
completion, 1 mL of CDCl was added to the vial. The resulting
3
1
mixture was placed in an NMR tube to obtain the H NMR spectrum.
Friedel−Crafts Reactions. β-Nitrostyrene (0.0074 g, 0.050
mmol), N-methylindole (0.020 g, 0.15 mmol), and 10 mol % of the
1
catalyst (0.005 mmol) were dissolved in 0.6 mL of CDCl , and the H
3
meso-2-Phenyl-4,6-bis(trifluoromethyl)-1,3-dioxane. In a 500 mL
NMR spectra of the reaction mixtures were recorded after 24 h.
Computations. Conformational searches were carried out using
the MMFF force field and AM1 semiempirical calculations with
round-bottomed flask, 4.0 g (19 mmol) of a 40: 60 mixture of meso-
13
and dl-1,1,1,5,5,5-hexafluoropentane-2,4-diol, 2.9 mL (3.0 g, 19
mmol) of benzaldehyde dimethylacetal, and 15 mg of para-
toluenesulfonic acid were dissolved in 200 mL of dry methylene
1
7
18
Spartan 08. Single-point B3LYP/6-311+G(d,p) and M06-2X/
1
9,20
maug-cc-pVT(+d)Z
energy computations were carried out on all
−1
chloride freshly distilled from CaH . After magnetically stirring this
of the resulting structures that were found to be within 3−5 kcal mol
2
2
1
solution at room temperature for 7 days, it was vigorously extracted
with water (5 × 50 mL). The aqueous layers were combined and set
aside for later use because they contain the racemic 1,1,1,5,5,5-
hexafluoropentane-2,4-diol. The organic layer was dried over MgSO4
and concentrated with a rotary evaporator at water aspirator pressure
to afford the crude ketal. It was then dissolved in ethanol, brought to a
boil, and water was added until the solution turned cloudy. Upon
slowly allowing the mixture to cool to room temperature, white
needle-like crystals of meso-2-phenyl-4,6-bis(trifluoromethyl)-1,3-diox-
ane formed. They were filtered away from the mother liquor to yield
of the most favorable species using Gaussian 09. Full optimizations
and vibrational frequency calculations were subsequently carried out
using the same two DFT methods and basis sets on the most stable
conformers of the acids and their conjugate bases.
2
2
The conductor-like polarizable continuum model was used to
predict pK values in DMSO using both computational approaches
a
noted above. In this work, liquid-phase geometry optimizations and
harmonic frequencies were computed in addition to single-point
energies on the gas-phase structures. Relative pK
obtained and converted to absolute values since pK
been measured.
values to TFE were
a
(TFE) = 23.5 has
a
1
6
2
.1 g (37% from the starting diol mixture and 92% when accounting
13a
1
for the diastereomeric ratio) of the title compound.
H NMR (300
Photoelectron Spectroscopy. Low-temperature photoelectron
MHz, DMSO-d ): δ 1.85 (2H, m), δ 4.42 (2H, m), δ 5.69 (1H, s), δ
spectra were recorded with a home-built variable-temperature
6
13
23
7
.20−7.45 (5H, m). C NMR (75 MHz, DMSO-d ) δ 22.6, δ 73.4 (q,
photoelectron spectrometer that has been previously described.
6
J = 33.2 Hz), 101.3, 126.3, 128.5, 129.9, 130.0 (q, J = 280 Hz), 136.2.
The conjugate bases of 1 and 2 were readily generated by electrospray
1
9
−3
F NMR (282 MHz, DMSO-d ) δ −80.3 (d, J = 5.9 Hz). HRMS-ESI:
ionization from ∼10 M methanol−water solutions and were trapped
6
+
+
calcd for C H F O (M + H) , 301.0658; found, 301.0665.
and cooled to 20 K over a period of 20−100 ms by blocking incoming
anions for the final 20 ms of a 100 ms acquisition. These ions were
then extracted into a time-of-flight mass spectrometer at a repetition
rate of 10 Hz. Photoirradiation of the mass selected anions with an
excimer laser at 193 nm (6.424 eV) operating at 20 Hz was carried out
to enable shot-to-shot background subtraction for all of the reported
spectra. Photoelectrons were collected at ∼100% efficiency and
analyzed with a 5.2 m long electron flight tube. This provided spectra
with a resolution (ΔE/kinetic energy) of ∼2% or 30 meV at 5 eV
binding energy.
12
11
6
2
dl-1,1,1,5,5,5-Hexafluoropentane-2,4-diol. The combined aqueous
material set aside earlier was vigorously extracted with diethyl ether (3
×
50 mL) and dried over MgSO . Removal of the ether at water
4
aspirator pressure with a rotary evaporator gave an enriched
diastereomeric mixture of dl-1,1,1,5,5,5-hexafluoropentane-2,4-diol
1
3a
(
∼95:5 of the desired dl isomer to the undesired meso compound).
Further enrichment of the dl diastereomer was done by reketalizing
the mixture as described above to afford 1.7 g of dl-1,1,1,5,5,5-
hexafluoropentane-2,4-diol as a white solid in >99: 1 diastereomeric
1
purity. H NMR (300 MHz, DMSO-d ) δ 1.65 (2H, m), 4.1 (2H, m),
6
6
.49 (2H, d, J = 6.9, OH). 13C NMR (75 MHz, CDCl ) δ 29.0 (s,
CH ), 66.7 (q, J = 32.9 Hz, CHCF ), 125.6 (q, J = 280 Hz, CF ).
3
RESULT AND DISCUSSION
19
■
F
2
3
3
NMR (282 MHz, DMSO) δ −78.6 (d, J = 5.6 Hz). HRMS-ESI: calcd
Hydrogen bonding is ubiquitous in biological systems and plays
a critical role in molecular recognition and catalysis. Designing
small molecules to mimic this behavior is a major challenge and
the subject of much ongoing research. Brønsted acids and
hydrogen bond catalysts can be exploited in this regard. We
recently reported that hydrogen-bond networks can be used to
delocalize a charge site and increase the acidity or basicity of a
−
−
for C H F O (M − H) , 211.0199; found, 211.0201.
5
5
6
2
pKa Determinations. Aqueous acidities were measured by
potentiometric titrations using a stock solution of NaOH (0.01 M)
as the titrant after calibrating the pH meter with standard buffer
solutions. DMSO pK ’s were measured by the overlapping indicator
a
1
method at 20−25 °C by UV and H NMR spectroscopy as previously
11,14
described.
Multiple measurements were performed for each
24
compound. Proof of concept computations on perfluoropo-
lyols, species that are apt to be stable only at cryogenic
temperatures, revealed that hydrogen-bond arrays in con-
junction with electron-withdrawing groups lead to very strong
compound using two of the following indicators as long as they
were within 2 pK units of the polyol being measured: 4-chloro-2,6-
a
dinitrophenol (pKa = 3.3), 2,4-dinitrophenol (pKa = 5.1, Sigma
Aldrich), 9-fluorenetriphenylphosphonium bromide (pK = 6.6), and
a
1
5,16
11
9
-thiophenylfluorene (pK = 15.1).
Ion-pairing and self-association
Brønsted acids. To test this prediction, three trifluoromethyl
a
−5
of the acids were minimized by working at low concentrations (10 −
M).
Binding Measurements. Diols 1 and 2 were mixed with CD CN,
group containing polyols were examined.
Electrospray ionization of 1,3-bis(trifluoromethyl)-1,3-pro-
−
3
1
0
3
panediol (1) and 1,1,2,2-tetra(trifluoromethyl)-1,2-ethanediol
and the resulting 2.5 mM solutions were placed in NMR tubes.
Carefully measured volumes of 100 mM tetrabutylammonium chloride
−
(
2) afforded their corresponding (M − 1) ions, and the
photoelectron spectra of these anions (1a and 2a) were
obtained at 20 K with an excimer laser at 193 nm (Figure 2).
in CD CN were sequentially added, and these titrations were
3
25
1
monitored by recording an H NMR spectrum at each point. The
downfield chemical shifts of the OH signals were followed and
nonlinear 1:1 fits of the binding isotherms were carried out using the
These broad spectra are similar to those of other deprotonated
polyols, and the top of the bands give the vertical electron
B
dx.doi.org/10.1021/ja4036384 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Table 2. Calculated and Experimental pK Values
a
expt
B3LYP/6-
311+G(d,p)
M06-2X/maug-
cc-pVT(+d)Z
cmpd (ROH)
DMSO
1
2
3
17.5
0.6
16.0
2.3
16.0 ± 0.1
4.8 ± 0.1
7.1 ± 0.3
9.4
6.8
a
(
CF ) COH
8.3
9.6
10.7
3
3
b
c
HOCH CH OH
24.7
23.9
20.1
18.2
13.7
24.1
23.3
18.3
19.4
12.8
28.0
2
2
b
HOCH CH CH OH
25.4 ± 0.3
19.7 ± 0.2
18.0
2
2
2
b
(
HOCH CH ) CHOH
2 2 2
b
PhOH
b
CH CO H
12.3
3
a
2
HCl
1.8
a
b
c
See ref 16. See ref 11. This value was obtained from a linear
correlation between ΔH°acid and pK , see ref 10.
a
Figure 2. Low-temperature (20 K) photoelectron spectra of
−
−
CF CH(OH)CH CH(O )CF (1a, top) and (CF ) C(OH)C(O )-
(
3
2
3
3
2
Diol 1 is the least acidic of the three but the two trifluoromethyl
CF ) (2a, bottom) at 193 nm (6.424 eV).
3 2
groups enhance its acidity by 9.4 pK units relative to 1,3-
a
propanediol. As a result, it is a stronger acid than phenol by one
hundred fold. Triol 3 is 8.9 orders of magnitude more acidic
detachment energies (VDEs), whereas a linear extrapolation of
the onset region affords the adiabatic electron detachment
than 1 and 12.6 pK units more acidic than the unsubstituted
a
26
energies (ADEs, Table 1). The resulting values are very large
for alkoxide ions and correspond to enhancements relative to
ethoxide and 2,2,2-trifluoroethoxide of up to 3.29 and 2.45 eV
polyol without any trifluoromethyl groups. It is also a little
5
more than 10 times stronger than acetic acid even though it is
a saturated compound. The most acidic compound of the series
−1
27
23
(
75.9 and 56.5 kcal mol ), respectively, for 2a. This is due to
is diol 2, which is 10 times more acidic than ethylene glycol,
the stabilization of 1a and 2a resulting from their strong
intramolecular hydrogen bond and the presence of the
electron-withdrawing trifluoromethyl substituents. The latter
7.5 pK units stronger than acetic acid, and within 3 pK units of
a
a
the value of HCl. These results reveal that the combination of
intramolecular hydrogen-bonding and electron-withdrawing
trifluoromethyl groups can lead to saturated polyols that are
quite acidic in DMSO.
In protic solvents intramolecular hydrogen-bond stabilization
of polyol conjugate bases is relatively unimportant because the
anions are stabilized by intermolecular hydrogen bonds with
−1
effect is worth 16 kcal mol per CF3 group and was
determined by comparing the ADEs of 1a and 2a to estimates
for the conjugate bases of 1,3-propanediol (2.59 ± 0.14 eV)
28
and 1,2-ethanediol (2.28 ± 0.14 eV). It also results in
electron-binding energies that are larger than those for the
conjugate bases of strong acids, such as CH CO H (3.47 ±
the solvent. As a result, the pK of ethylene glycol and ethanol
3
2
a
2
9
30
33
0
0
.01 eV), HCl (3.613577 ± 0.000044 eV), HNO (3.937 ±
.014 eV), and in the later case, even for H SO (4.75 ± 0.10
in water differ by only 0.5 pK units (i.e., 15.4 vs 15.9). The
3
a
3
1
acidities of 2 and 3 consequently were measured in water as this
presumably provides an opportunity to probe the effects of the
trifluoromethyl group in the absence of intramolecular
hydrogen-bond stabilization. These compounds were found
2
4
3
2
eV). B3LYP/aug-cc-pVDZ and M06-2X/maug-cc-pVT(+d)Z
calculations are in accord with these findings, but the latter
approach is more accurate. It reproduces the experimental
ADEs with an average error of 0.16 eV, which is the same value
that was previously reported for a different set of polyol anions.
The photoelectron spectra of 1a and 2a reveal that the
combination of intramolecular hydrogen bonds and electron-
withdrawing groups can lead to very stable anions in the gas
to be acidic (i.e., pK = 5.6 and 7.1, respectively) and are
a
stronger acids than 1,1,1,3,3,3-hexafluoro-2-propanol (HFP,
pK = 9.3), but neither one is as strong as acetic acid (pK =
a
a
4.8). The acidity of 2 is not surprising in that the four
trifluoromethyl groups are on adjacent carbons and exert a
1
0
phase. To assess the impact of this stabilization in solution, the
strong stabilizing inductive effect. That is, there is a 10 acidity
enhancement of 2 relative to ethylene glycol which leads on
average to a 2.5 pK unit effect per CF group. In contrast, the
1
pK ’s of 1−3 were measured in DMSO by H NMR and UV
a
spectroscopy (Table 2). The former values span from 4.8 to
a
3
16.0, which makes these compounds remarkably acidic alcohols.
four CF groups in 3 are separated by three intervening carbons
3
Table 1. Experimental and Computed ADE and VDE in eV for Deprotonated Alcohols and Diols
a
expt
calcd
cmpd (RO−)
ADE
VDE
ADE
VDE
−
b
b
CH CH O
1.7120 ± 0.0040
2.5541 ± 0.0043
4.00 ± 0.10
1.65 (1.58)
2.80 (2.79)
3.73 (4.09)
4.74 (4.82)
4.45 (4.64)
1.82 (1.72)
2.95 (2.88)
4.31 (4.56)
5.27 (5.37)
5.27
3
2
−
CF CH O
3
2
−
CF CH(OH)CH CH(O )CF (1a)
4.51 ± 0.10
5.51 ± 0.10
3
2
3
−
(
(
CF ) C(OH)C(O )(CF ) (2a)
5.00 ± 0.10
3
2
3
2
−
(CF ) C(OH)CH ) CHO (3a)
3
2
2 2
a
Computed values are at 0 K and correspond to B3LYP and M06-2X (in parentheses) energies. The aug-cc-pVDZ basis set was used for computing
b
VDEs and ADEs with the former functional. M06-2X ADEs were computed with the larger maug-cc-pVT(+d)Z basis set. See ref 27.
C
dx.doi.org/10.1021/ja4036384 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
so little, if any, effect was expected from one of the geminal
ionization site can vary with the dielectric constant of the
solvent. The energy differences here, however, are quite small,
and the most stable structure may vary with the computational
approach; M06-2X/maug-cc-pVT(+d)Z calculations predict
that the internal alkoxide anion is more stable in all three
cases but that the energy differences are 0.55 (gas phase), 0.71
(DMSO), and only 0.06 kcal mol− in water.
pairs of trifluoromethyl substituents. The 2.2 pK unit acidity
a
enhancement relative to HFP, however, suggests that hydrogen
bonds can transmit the inductive effect over distance even in an
aqueous environment. This would provide an important long-
range stabilizing mechanism that may have far reaching
implications in biological processes.
1
Liquid-phase DMSO pKa values were calculated for
Molecular recognition of anions via ion channels and
transporters is biologically important in stabilizing membrane
potentials and controlling cell volumes and is intimately
(
CF ) COH and 1−3. This was accomplished by computing
3
3
their gas-phase acidities (ΔG° ) and that of TFE (Table 3)
acid
34
connected to a number of debilitating diseases. An acidic
Table 3. Calculated and Experimental Gas-Phase Acidities
and flexible polyol (HOCH CH CH(OH)CH ) C-OH, pK =
2
2
2 3
a
−1
(ΔG°acid, in kcal mol )
11.4 ± 0.2) was recently reported to bind chloride in
acetonitrile with an association constant of 360 M− despite
being an aliphatic alcohol that can form intramolecular
1
B3LYP/6-
M06-2X/maug-cc-
pVT(+d)Z
cmpd (ROH)
CF CH OH
311+G(d,p)
expt
35
hydrogen bonds. Diols 1 and 2 bracket the acidity of this
a
a
351.4
321.0
330.1
302.8
309.6
354.0
325.6
330.6
307.6
310.8
354.1 ± 2.0
324.0 ± 2.0
3
2
heptaol (i.e., 1 and 2 are 5−6 pK units less acidic and more
a
(
CF ) COH
3 3
acidic than it), and consequently their binding constants were
1
2
3
also measured. The less acidic diol (1) was found to have an
−1
association constant with chloride of 3300 M in acetonitrile.
This is nine times larger than for the more acidic heptaol, 100
fold bigger than an α-D-ribose receptor in which the C1 and C5
a
See ref 27.
hydroxyl groups are protected as their methyl and trityl (Ph C)
3
and then calculating the solvation energies of the acids and
their conjugate bases with a polarized continuum model.
B3LYP/6-311+G(d,p) and M06-2X/maug-cc-pVT(+d)Z en-
ethers, respectively, and 14 times larger than for the
36
corresponding β-anomer. Our association constant is also
greater than for phenols, such as catechol (1,2-(HO) C H , ∼3
2
6
4
ergies were used for this purpose, and absolute pK ’s were
37
a
fold) and resorcinol (1,3-(HO) C H , 23 times). 1,1,2,2-
2
6
4
derived from the relative values to TFE and its experimentally
measured acidity of 23.5. Both methods do well but the average
unsigned errors for the gas-phase acidities (2.9 (B3LYP) and
Tetra(trifluoromethyl)ethylene glycol (2) has a binding
−
−1
constant with Cl of 6700 M , which is the largest measured
value for an aliphatic alcohol to date, but it is only about twice
as large as for 1 even though their acidities differ by 11 orders
of magnitude in DMSO. This clearly indicates that there is only
a loose relationship between the acidity of a compound and its
anion binding affinity undoubtedly because the cavity size of
the receptor, steric interactions, and the solvation of the bound
complex also play a role.
−1
0
.9 (M06-2X) kcal mol ) and the DMSO pK ’s (2.6 (B3LYP)
a
and 1.0 (M06-2X)) are noticeably smaller for the M06-2X
density functional. The largest deviations from experiment (3.0
kcal mol− and 4.2 pK units (B3LYP) vs 1.6 kcal mol and 2.5
1
−1
a
pK units (M06-2X)) are also better for the meta hybrid
a
generalized gradient approximation Minnesota 06 functional.
Both methods are least accurate for 2, and this may be a
Our physical characterization of polyols 1−3 suggests that
they could be good Brønsted acid catalysts, and since they are
all capable of forming multiple hydrogen bonds in advance of
proton transfer, they maybe more effective than other specific
acid catalysts. To explore this possibility, their catalytic behavior
reflection of the four CF groups being in close proximity to
3
each other.
Triol 3 has two ionization sites, and our experiments do not
indicate which one is more acidic. B3LYP/6-311+G(d,p)
computations were carried out to address this issue. In the
gas-phase deprotonation of the internal hydroxyl group is
energetically preferred over a terminal one, but only by 0.84
38
in a Friedel−Crafts reaction between β-nitrostyrene and N-
methylindole (eq 1) was investigated. All three polyols promote
−1
kcal mol . This can be attributed to the formation of two
direct hydrogen bonds to the charged center in the former case
as opposed to one, along with a secondary hydrogen bond, in
the latter instance (Figure 3). The same stability order is found
this transformation and the reaction rates, as indicated by the
percent conversion, correlate with their acidities not with the
number of hydroxyl groups (Table 4). That is, 10 mol % of the
strongest acid (2) leads to a 95% conversion at room
temperature in 24 h whereas it is 19% with the weakest acid
(1). Acetic acid does not catalyze this process, however, even
Figure 3. Isomeric conjugate bases of triol 3.
in DMSO, but solvation reduces the predicted energy
−1
difference to 0.36 kcal mol . As a result, the relative stabilities
were also computed in water since it has a higher dielectric
constant than that for DMSO (i.e., ε = 78.4 vs 46.8,
though it is four pK units more acidic than 1 in DMSO. This
a
indicates that general acid catalysis is involved and strongly
suggests that more than one hydrogen bond is involved in this
transformation (i.e., the reaction proceeds via hydrogen-bond
catalysis).
All three alcohols also catalyze the aminolysis of styrene
oxide with aniline at 60 °C under solvent-free conditions (SFC,
2
1
respectively). In this case there is a reversal, and now the
terminal alkoxide anion is found to be more stable than the
−1
internal one by 0.81 kcal mol . These results indicate that both
the inductive effect and hydrogen-bond stabilization are
sensitive to the medium, and as a result, the preferred
D
dx.doi.org/10.1021/ja4036384 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Table 4. Results for Acid-Catalyzed Transformations
Institute for Advanced Computational Research are gratefully
acknowledged. The photoelectron spectra work was supported
by the Division of Chemical Sciences, Geosciences, and
Biosciences, Office of Basic Energy Sciences, U.S. Department
of Energy (DOE), and was performed at the EMSL, a national
scientific user facility sponsored by DOE’s Office of Biological
and Environmental Research and located at Pacific Northwest
National Laboratory, which is operated by Battelle for DOE.
a
conversion (%)
b
catalyst
eq 1
eq 2
5:6
1
19
95
53
4
89
100
100
5
73: 27
88: 12
81: 19
43: 57
86: 14
2
3
no catalyst
HOAc
4
78
a
1
b
Determined by H NMR. Reaction time = 2.75 h except for when 2
REFERENCES
■
was used, then it was 20 min.
(
1) (a) Akiyama, T. Chem. Rev. 2007, 107, 5744−5758. (b) Taylor,
M. S.; Jacobsen, E. N. Angew. Chem., Int. Ed. 2006, 45, 1520−1543.
eq 2, Table 4). The reactivity order again is in accord with the
acidity of the three alcohols as is the selectivity. That is, the
(
c) Akiyama, T. Adv. Synth. Cat. 2006, 348, 999−1010. (d) List, B.;
Yang, J. W. Science 2006, 313, 1584−1586.
2) (a) Uraguchi, D.; Terada, M. J. Am. Chem. Soc. 2004, 126, 5356−
357. (b) Akiyama, T.; Itoh, J.; Yokota, K.; Fuchibe, K. Angew. Chem.,
Int. Ed. 2004, 43, 1566−1568.
3) Garcia-Garcia, P.; Lay, F.; Garcia-Garcia, P.; Rabalakos, C.; List,
B. Angew. Chem., Int. Ed. 2009, 48, 4363−4366.
4) (a) Nakashima, D.; Yamamoto, H. J. Am. Chem. Soc. 2006, 128,
9626−9627. (b) Cheon, C. H.; Yamamoto, H. J. Am. Chem. Soc. 2008,
30, 9246−9247.
5) Juhasz, M.; Hoffmann, S.; Stoyanov, E.; Kim, K. C.; Reed, C. A.
Angew. Chem., Int. Ed. 2004, 43, 5352−5355.
6) (a) Wenzel, A. G.; Jacobsen, E. N. J. Am. Chem. Soc. 2002, 124,
(
5
(
(
most acidic acid (2) catalyzes the reaction most efficiently and
leads to the most selective transformation. It is also a more
effective catalyst in this reaction than Shreiner’s thiourea ((3,5-
1
(
39
(CF ) C H NH) CS) by about an order of magnitude. Acetic
3 3 6 3 2
(
1
1
acid is 4 orders of magnitude more acidic than 1 in DMSO but
is a slightly less efficient catalyst. This indicates that general acid
catalysis is also important in this reaction.
2964−12965. (b) Joly, G. D.; Jacobsen, E. N. J. Am. Chem. Soc. 2004,
26, 4102−4103. (c) Okino, T.; Hoashi, Y.; Takemoto, Y. J. Am.
Chem. Soc. 2003, 125, 12672−12673. (d) Inokuma, T.; Hoashi, Y.;
Takemoto, Y. J. Am. Chem. Soc. 2006, 128, 9413−9419.
CONCLUSION
(
1
7) (a) Hine, J.; Linden, S. M.; Kanagasabapathy, V. M. J. Org. Chem.
■
985, 50, 5096−5099. (b) Kelly, T. R.; Meghani, P.; Ekkundi, V. S.
Diols 1 and 2 and triol 3 were found to be quite acidic in
DMSO and water. The conjugate bases of the most acidic and
least acidic compounds (i.e., 2 and 1, respectively) are also
remarkably stable in the gas phase as indicated by their ADEs,
which in the former case exceeds that of deprotonated sulfuric
acid. Taken together, these results suggest that the inductive
effect can be transmitted via hydrogen bonds. Diols 1 and 2
also bind chloride anion in acetonitrile with the largest binding
constants reported to date for an aliphatic alcohol (i.e., K =
Tetrahedron Lett. 1990, 31, 3381−3384.
(
9
2
8) (a) Huang, Y.; Rawal, V. H. J. Am. Chem. Soc. 2002, 124, 9662−
663. (b) Huang, Y.; Unni, A. K.; Thadani, A. N.; Rawal, V. H. Nature
003, 424, 146. (c) Unni, A. K.; Takenaka, N.; Yamamoto, H.; Rawal,
V. H. J. Am. Chem. Soc. 2005, 127, 1336−1337.
9) (a) Curran, D. P.; Kuo, L. H. Tetrahedron Lett. 1995, 36, 6647−
650. (b) Vakulya, B.; Varga, S.; Csmpai, A.; Sos, T. Org. Lett. 2005, 7,
(
6
1967−1969.
(10) Tian, Z.; Fattahi, A.; Lis, L.; Kass, S. R. J. Am. Chem. Soc. 2009,
131, 16984−16988.
−1
3
300 (1) and 6700 (2) M ) and are capable of acting as
(
2
(
2
11) Shokri, A.; Abedin, A.; Fattahi, A.; Kass, S. R. J. Am. Chem. Soc.
Brønsted acid and hydrogen-bond catalysts (i.e., specific and
general acid catalysts, respectively). As a result, hydrogen
bonding in conjunction with electron-withdrawing groups is a
promising avenue for developing molecular recognition hosts
and Brønsted acid and hydrogen-bond catalysts.
012, 134, 10646−10650.
12) Loeb, S. J.; Martin, J. W. L.; Willis, C. J. Can. J. Chem. 1978, 56,
369−2373.
13) (a) Murphy, K., Ph.D. Thesis, University of Minnesota, 2010;
(
pp 125−127. (b) Jeulin, S.; Paule, S. D. D.; Ratovelomanana-Vidal, V.;
Genet, J. P.; Champion, N.; Dellis, P. Angew. Chem., Int. Ed. 2004, 43,
320−325.
ASSOCIATED CONTENT
Supporting Information
■
*
S
(14) (a) Matthews, W. S.; Bares, J. E.; Bartmess, J. E.; Bordwell, F.
G.; Cornforth, F. J.; Drucker, G. E.; Margolin, Z.; McCallum, R. J.;
McCollum, G. J.; Vanier, N. R. J. Am. Chem. Soc. 1975, 97, 7006−
Computed geometries and energies are provided along with
experimental data for the binding constant determinations and
the complete citation to ref 21. This material is available free of
charge via the Internet at http://pubs.acs.org.
7
7
(
014. (b) Chu, Y.; Deng, H.; Cheng, J. P. J. Org. Chem. 2007, 72,
790−7793.
15) For the syntheses of the indicators, see ref 14a and (a) Benedikt,
G. M.; Traynor, L. Tetrahedron Lett. 1987, 28, 763−766. (b) Johnson,
AUTHOR INFORMATION
Corresponding Author
kass@umn.edu; xuebin.wang@pnnl.gov
■
A. W.; Lee, S. Y.; Swor, R. A.; Royer, L. D. J. Am. Chem. Soc. 1966, 88,
1
953−1958.
16) For DMSO pK ’s, see: Bordwell, F. G. Acc. Chem. Res. 1988, 21,
(
a
Notes
456−463.
(
17) Spartan ’08 for Macintosh; Wavefunction, Inc.: Irvine, CA.
18) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652. (b) Lee,
The authors declare no competing financial interest.
(
C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785−789.
ACKNOWLEDGMENTS
■
(
19) (a) Zhao, Y.; Truhlar, D. G. J. Phys. Chem. A 2008, 112, 1095−
We thank Dr. K. Murphy for preparing and separating the two
diastereomers of 1. Generous support from the National
Science Foundation, the Petroleum Research Fund as
administered by the ACS and the Minnesota Supercomputer
1
2
099. (b) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215−
41. (c) Zhao, Y.; Truhlar, D. G. Acc. Chem. Res. 2008, 41, 157−167.
(20) Papajak, E.; Truhlar, D. G. J. Chem. Theory Comput. 2010, 6,
597−601.
E
dx.doi.org/10.1021/ja4036384 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(
2
(
3
1
(
0
(
21) Frisch, M. J., et al. Gaussian 09; Gaussian, Inc.: Pittsburgh, PA,
009.
22) (a) Barone, V.; Cossi, M.; Tomasi, J. J. Chem. Phys. 1997, 107,
210−3221. (b) Cammi, R.; Mennucci, B.; Tomasi, J. J. Phys. Chem. A
998, 102, 870−875.
23) Wang, X. B.; Wang, L. S. Rev. Sci, Instrum. 2008, 79, 073108−1−
73108−8.
24) (a) Shokri, A.; Schmidt, J.; Wang, X. B.; Kass, S. R. J. Am. Chem.
Soc. 2012, 134, 2094−2099. (b) Tian, Z.; Fattahi, A.; Lis, L.; Kass, S.
R. Croat. Chem. Acta 2009, 82, 41−45.
(
25) The conjugate base of 3 (3a) was only examined computation-
ally because its ADE is predicted to be between those for 1a and 2a,
and the acidities follow the same trend (i.e., 1 (least acidic) < 3 < 2
(
most acidic)).
(
26) The higher binding energy feature in the spectrum of 1a is due
to a higher electronic state of the corresponding radical, presumably,
due to the loss of a lone pair electron from a fluorine atom or the OH
group.
(
27) (a) Ramond, T. M.; Davico, G. E.; Schwartz, R. L.; Lineberger,
W. C. J. Chem. Phys. 2000, 112, 1158−1169. (b) Mihalick, J. E.; Gatev,
G. G.; Brauman, J. I. J. Am. Chem. Soc. 1996, 118, 12424−12431.
(
28) Crowder, C.; Bartmess, J. J. Am. Soc. Mass Spectrom. 1993, 4,
7
(
(
23−726.
29) Lu, Z.; Continetti, R. E. J. Phys. Chem. A 2004, 108, 9962−9969.
30) Berzinsh, U.; Gustafsson, M.; Hanstorp, D.; Klinkmuller, A.;
Ljungblad, U.; Martensson-Pendrill, A. M. Phys. Rev. A 1995, 51, 231−
38.
31) Weaver, A.; Arnold, D. W.; Bradforth, S. E.; Neumark, D. M. J.
Chem. Phys. 1991, 94, 1740−1751.
32) Wang, X. B.; Nicholas, J. B.; Wang, L. S. J. Phys. Chem. A 2000,
04, 504−508.
33) Stewart, R. The Proton: Applications to Organic Chemistry;
Academic: New York, 1985; pp 313.
34) Davis, A. P.; Sheppard, D. N.; Smith, B. D. Chem. Soc. Rev. 2007,
6, 348−357.
35) Shokri, A.; Schmidt, J.; Wang, X. B.; Kass, S. R. J. Am. Chem. Soc.
012, 134, 16944−16947.
36) Kondo, S.; Kobayashi, Y.; Unno, M. Tetraheron Lett. 2010, 51,
2
(
(
1
(
(
3
(
2
(
2
(
(
512−2514.
37) Smith, D. K. Org. Biomol. Chem. 2003, 1, 3874−3877.
38) Herrera, R. P.; Sgarzani, V.; Bernard, L.; Ricci, A. Angew. Chem.,
Int. Ed. 2005, 44, 6576−6579.
39) Jakab, G.; Tancon, C.; Zhang, Z.; Lippert, K. M.; Schreiner, P.
R. Org. Lett. 2012, 14, 1724−1727.
(
F
dx.doi.org/10.1021/ja4036384 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX