Correlations of pKlgMe with Reduction Potentials

Article abstract of DOI:10.1021/jo961658r

The quantitative indices for the ability of leaving groups to depart from C atoms are pK_lg^Me. In the cases of methyl transfers from arenesulfonates, these parameters have correlated a large number of nucleophilic data. A new scale of these parameters has been defined from methyl transfer data between phenylmethyl thioethers. The pK_lg^Me data from both sets of compounds correlated with both experimental E_1/2 values and E_LUMO values from computational chemistry. These correlations support the SCM model of Shaik et al. which requires the leaving group to display some SET character in an S_N2 transition state.

Full text of DOI:10.1021/jo961658r

J . Org. Chem. 1997, 62, 853-856
853
Me
Cor r ela tion s of p Klg w ith Red u ction P oten tia ls
K. R. Fountain,*,‡ Kamlesh D. Patel, Timothy W. Dunkin, J anette A. Powers, and
Dean A. Van Galen
Division of Sciences, Truman State University, Kirksville, Missouri 63501
Received August 28, 1996X
The quantitative indices for the ability of leaving groups to depart from C atoms are pKlgMe. In the
cases of methyl transfers from arenesulfonates, these parameters have correlated a large number
of nucleophilic data. A new scale of these parameters has been defined from methyl transfer data
Me
between phenylmethyl thioethers. The pKlg data from both sets of compounds correlated with
both experimental E1/2 values and ELUMO values from computational chemistry. These correlations
support the SCM model of Shaik et al. which requires the leaving group to display some SET
character in an SN2 transition state.
In tr od u ction
The quantitative behavior of leaving groups is impor-
tant in modeling transition state behavior in the SN2
reaction.1 Quantitative modeling of transition state data
has led to important insights into the timing and detailed
bonding in many reactions.2,3 A recent transition state
(TS) model for the SN2 reaction emphasizes single-
electron transfer (SET) character as an important com-
ponent. The principal features of this model, the state
correlation diagram (SCD), or valence bond configuration
mixing (VBCM) model, is summarized in Figure 1 for a
thermoneutral SN2 reaction.
The transition state (TS) in this diagram is the result
of avoided crossing of the states N•(•:X)- and N:-(R-X).
These states represent valence bond (VB) wave functions
describing the intimate complex of the nucleophile and
the substrate (lower left corner) and the electronically
excited state (upper left corner) of this complex. Similar
considerations correlate the right side (products) of
Figure 1. One of the predictions of this diagram is that
leaving group behavior should involve some SET char-
acter. This research explores this prediction by correlat-
ing the leaving group parameters (pKlgMe) with the single
F igu r e 1. State correlation diagram (SCD) for a thermo-
neutral SN2 reaction.
Me
âlg values for nucleophilic displacements range from
0.44 to 0.57. A typical use of this parameter is as an
index of bond order (BO) in the transition state.6 For a
very early transition state where a nearly full bond exists
Me
between C and the leaving group, âlg is nearly 0.0; so
Me
electron parameter, E1/2
.
when the bond is cleaved at the TS âlg is 1.0. The BO
Me
is then 1.0-âlg
.
A recent quantitative expression of leaving group
Me
Me
Interpreting the âlg can be based on any model one
chooses for the SN2 reaction. It is a quantitative expres-
sion about structure-rate relationships, but has not yet
been investigated for single-electron transfer (SET) char-
acter or polar (transfer of two electrons in the TS)
behavior is the âlg parameter, due to Shankweiler and
Hofmann.4 This parameter is the slope of a plot of -log
Me
Me
knuc vs pKlg
.
The pKlg
are based on the methyl
exchange data of Lewis et al.,5 for methyl transfer
between arenesulfonate anions. They successfully cor-
related a wide variety of experimental and theoretical
data for nucleophilic reactivity and eliminations. Typical
character. We report in this paper a correlation of the
Me
pKlg
with a specific SET parameter, the reduction
potential, E1/2, for the positively charged methylating
agent, and ELUMO for the O-C bond for the neutral
methylating agent.
† E-mail sc18@truman.edu.
‡ Formerly Northeast Missouri State University.
X Abstract published in Advance ACS Abstracts, February 1, 1997.
(1) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in
Organic Chemistry Harper & Row Publishers: New York, 1987; 3rd
edition, Chapter 4.
(2) Shaik, S. S.; Schlagel, H. B.; Wolfe, S. Theoretical Aspects of
Organic Chemistry: The SN2 Mechanism; J ohn Wiley & Sons, Inc.:
New York, 1992.
(3) Pross, A. Theoretical & Physical Principles of Organic Reactivity
J ohn Wiley & Sons, Inc.: New York, 1995.
(4) (a) Hoffman, R. V.; Shankweiler, J . M. J . Am. Chem. Soc. 1986,
108, 5536. (b) McManus, S. P.; Smith, M. R.; Shankweiler, J . M.;
Hoffman, R. V. J . Org. Chem. 1988, 53, 141.
(5) (a) Lewis, E. S.; Hue, D. D. J . Am. Chem. Soc. 1984, 106, 3392.
(b) Lewis, E. S.; Vanderpool, S. R. J . Am. Chem. Soc. 1977, 99, 1946.
J . Am. Chem. Soc. 1978, 100, 6421. (d) Lewis, E. S.; Kukes, S.; Slater,
C. D. Ibid. 1980, 102, 1619. (e) Lewis, E. S.; Smith, M. J .; Christie, J .
J . J . Org. Chem. 1983, 48, 2527.
Exp er im en ta l Section a n d Resu lts
Red u ction P oten tia ls. The reduction potentials of a series
of substituted phenlydimethylsulfonium fluoroborates were
determined in acetonitrile, DMF, and DMSO. The apparatus
was a BAS 100B/W electrochemical analyzer, which was used
to determine all the cyclic voltametry. The working electrode,
(6) (a) Fountain, K. R.; Fountain, D. P.; Michaels, B.; Myers, D. B.;
Salmon, J . K.; Van Galen, D. A.; Yu, P. Can. J . Chem. 1991, 69, 798.
(b) Fountain, K. R.; Hutchinson, L. K.; Mulhearn, D. C.; Xu, Y.-B. J .
Org. Chem. 1993, 58, 7883.
(7) Andrieux, C. P.; Robert, M.; Saeva, F. D.; Save´ant, J .-M. J . Am.
Chem. Soc. 1994, 116, 7864.
S0022-3263(96)01658-1 CCC: $14.00 © 1997 American Chemical Society

854 J . Org. Chem., Vol. 62, No. 4, 1997
Fountain et al.
F igu r e 3. Log kij vs log Keq for reaction 1. This figure justifies
computation of Keq values from enthalpy.
F igu r e 2. Hammett Plot of E1/2 of substituted phenyldimeth-
ylsulfonium fluoroborates in acetonitrile (+), DMSO (o), and
DMF (4).
reference electrode, and auxiliary electrode were glassy carbon
(BAS), Ag/AgCl (0.1 M), and Pt wire. The E1/2 values were
obtained at a scan rate of 100 mV/s. The glassy carbon
electrode was polished after every run, to remove any absorbed
species, with a 0.3 µm OP alumina suspension for 30 s and
then rinsed with distilled water before being polished with 0.05
µm OP alumina suspension for 30 s. The electrode was once
again rinsed with distilled water and sonicated for 1 min with
a Branson 1200 Ultrasonic Cleaner. The electrode was once
again rinsed with distilled water before use.
The reductions of methyl arenesulfonate esters were done
in acetonitrile in the same manner as described above. The
E1/2 values are summarized in Table 2. A Hammett plot (not
shown: N ) 6; r ) 567 mV/s; R ) 0.9998) of these values
showed excellent linearity and correlation.
Ma t er ia ls. All of the phenyldimethylsulfonium fluoro-
borate salts were synthesized by known methods.5 The methyl
arenesulfonates were from samples previously reported.6a The
supporting electrolyte was tetrabutylammonium fluoro-
borate (used as received from Aldrich). The solvents, aceto-
nitrile (Fisher Scientific, HPLC grade), dimethylformamide
(American Burdick & Jackson), and dimethyl sulfoxide (Fisher),
were used as received. An internal reference of ferrocene is
307 mV vs the SCE electrode; thus values of E1/2 are vs the
SCE and not Ag/AgCl.
Me
F igu r e 4. Hammett plot of pKlg for substituted phenyldi-
methylsulfonium ions exchanging methyl groups with substi-
tuted phenylmethyl thioethers.
Ta ble 1. Exp er im en ta l a n d Com p u ted Equ ilibr iu m
Con sta n ts for Rea ction 1 a n d p Klg Va lu es for
Me
Rea ction 3
G
Keq (×106)
pKlg
pKlg computed
Me b
Me
H
4-Me
4-Cl
3-Cl
3,4-diCl
4-NO2
4-MeO
4-CN
7.8
20.3a
7.14a
0.98
0.13
0
0.579
-0.744
-0.900
-1.778
Table 2, in the experimental section, summarizes the data,
along with data from reduction of methyl arenesulfonates. A
Hammett plot (Figure 2) allowed computation of the experi-
mentally inaccessible E1/2 values for (4-nitrophenyl)dimethyl-
-3.532(σ-)
0.712
sulfonium fluoroborate.
Me
-2.479(σ-)
-1.683
The pKlg
values for the methyl arenesulfonates are
3,5-diCl
Me
reported in the literature.4 Calculation of pKlg for methyl
transfers between substituted phenylmethyl thioethers were
computed from the rate equilibrium data in ref 5d for reaction
1 from Lewis et al.5 In their paper the equilibrium constants
a
b
Computed from the equation from Figure 3. Computed by
-log KH + log KG.
off the line, as demonstrated. All the other points lie on a
fairly good line. A Hammett plot of the experimental and
computed log Keq values (not shown; 5 points, F ) -2.845, R
) 0.996) excluding 4-Cl also showed Cl plotting off the line.
The Hammett plot was used to calculate log Keq data for
4-ClC6H4 in Figure 3.
–
G1–SMe + Me3O+BF4
G1–S+MeBF4– + Me2O
Me
(1)
for G values were available for G1 ) 3,5-di-ClC6H3 , 3-Cl-C6H4,
and C6H5. Rate constants were available for G1 ) 4-MeC6H4,
4-MeOC6H4, C6H4, and 3-ClC6H4. Enthalpy values were avail-
able for all of these. The equilibrium constants for G ) 3-C6H4
and C6H5 were used to evaluate T∆S terms in the Gibbs-
Helmholtz equation, which were 3.03 and 1.86 kcal/mol,
respectively. These values display similarity, so equating ∆G
with ∆H and subtracting two different ∆H values to get
approximate δ∆G values will perhaps be an adequate ap-
proximation. The goodness of this approximation thus de-
pends on the constancy of the T∆S terms. Equation 2 shows
the formula for computation of δ log Keq. A log-log plot of
the rate constants for reaction 1 vs the available Keq were used
to check the internal consistency of this process, Figure 3. The
point for 4-ClC6H4, from enthalpy computations, plots severely
∆H - ∆H2
) δ log Keq
(2)
2.303(1.98)(308)
Me
The computations of pKlg for eq 2 were made by the
subtraction of the logs of the experimental and computed Keq
Me
from Figure 3 to give the pKlg
column in Table 1. A
Me-
Hammett plot of the values in the pKlg
appears in Figure 4. The equation is pKlg
-0.0121 (points ) 5; R ) 0.995). A complete set of pKlg
values is thus present in Table 1. This procedure is analogous
to that of Shankweiler and Hofmann, except that three Keq
column in Table 1
Me
) -2.784σ
Me

Me
Correlations of pKlg with Reduction Potentials
J . Org. Chem., Vol. 62, No. 4, 1997 855
F igu r e 7. Correlation of E1/2 with LUMO of substituted
phenyldimethylsulfonium ions.
Me
F igu r e 5. Plot of pKlg
in Table 1 for E1/2 values of
phenyldimethylsulfonium fluoroborates in acetonitrile (+),
DMF (4), and DMSO (O): A ) 4-NO2; B ) 3,5-diCl; C ) 4-CN;
D ) 3,4-diCl; E ) 3-Cl;F ) 4-Cl;G ) H,H ) 4-Me;I ) 4-MeO.
F igu r e 8. Correlation of E1/2 with LUMO of substituted
methylarenesulfonate esters.
F igu r e 6. Correlation of E1/2 values for methylarenesulfonate
Ta ble 2. E1/2 Va lu es for Su bstitu ted
Meth yla r en esu lfon a tes in Aceton itr ile vs SCE
Me
esters with pKlg
.
G
E1/2 (V)
G
E1/2 (V)
values were approximated by calculations from enthalpy to
obtain the Hammett plot for computations of the complete set.
3-CF3
3-Me
4-F
-1.910
-2.210
-2.258
-2.310
4-Br
4-MeO
3-MeO
-1.994
-2.430
-2.200
CH3
Me
G–C6H4–S–CH3 + C6H5–S+(CH3)2
G–C6H4–S+–CH3 + C6H5–SCH3
(3)
bond. Since we are comparing these two substrates, it is
necessary to compare the same types of orbitals, centered on
the CH3 groups and the leaving group. Visual inspection of
the LUMOs of these compounds via ChemPlus (a graphical
user interface, GUI) especially for this purpose) in HyperChem
showed the noncorrelating orbitals to be largely centered on
the aromatic rings. The LUMOs of the phenyldimethyl-
sulfonium ions, which were centered on the S-CH3 region,
correlated very well, and the methyl arenesulfonates orbitals
largely located between the C and O somewhat less well, but
both show a relationship between E1/2 and LUMO. This
correlation is actually surprising since studies have shown9
that in cyclic voltametry experiments a single electron most
likely initially resides on the S atom of the sulfonate portion,
giving products from cleavage of S-OCH3 and C-S bonds,
rather than O-C bonds. The quality of the E1/2 correlations
of these compounds are thus not surprisingly poorer. They
serve some purpose because they show the capability of the
SO3 group as a single electron acceptor.
Cor r ela t ion s of t h e E lect r och em ica l a n d Qu a n t u m
Me
Da ta . The correlations for the E1/2 values vs these pKlg for
the substituted phenyldimethylsulfonium salts are displayed
in Figure 5. The corresponding E1/2 values of the methylarene-
Me
sulfonate esters vs pKlg are in Figure 6. These correlations
for the substituted phenyldimethylsulfonium ions are very
consistent for the order of the substituent effects and have
nearly the same equations in all three solvents.
Figures 7 and 8 show correlations of E1/2 with the lowest
occupied molecular orbital (LUMO) of a series of calculated
structures of substituted methyl arenesulfonates and substi-
tuted phenyldimethylsulfonium salts Table 2. These computa-
tions were done by MOPAC6.0, MOPAC7.0, HyperChem, or
Spartan. All structures were inspected visually after conver-
gence, and vibrational studies confirmed all positive eigen-
values for the Hessian matrix. The energies of the unoccupied
molecular orbitals which had significant antibonding character
on the C atom of the methyl group in the methylarene-
sulfonates were plotted vs pKlgMe. There was no correlation
between the LUMOs and the E1/2 values for methyl arene-
sulfonates. Saeva et al.8 have pointed out the σ* nature of
the LUMO for arylmethylsulfonium salts to be largely centered
on the S-C bond with some contributions from the aryl-S
Discu ssion
Me
These correlations show that the parameters pKlg
from both methyl arenesulfonates and phenyldimethyl-
sulfonium ions are related to SET phenomena such as
E1/2 values. The behavior of these leaving groups as
(8) Saeva, F. D.; Breslin, D. T.; Martic, P. A. J . Am. Chem. Soc. 1989,
111, 1328.
(9) Yoesefzadeh, P.; Mann, C. K. J . Org. Chem. 1968, 33, 2716.

856 J . Org. Chem., Vol. 62, No. 4, 1997
Fountain et al.
single-electron acceptors has not been well studied, yet
the SCM model makes specific predictions that inclusion
of some SET character is present in both the attack of
the nucleophile and departure of the leaving group. The
correlations with the gas phase computations, LUMO vs
pKlgMe, and the solution phase E1/2 values indicates that,
for these substrates, solvation of the incipient molecule
or ion of the leaving group in the TS is not very
important. Save´ant et al. have shown that the phenyl-
dimethylsulfonium salts insert a single electron into the
C-S σ*bond and not into the π* orbitals of the ring.8,10
This insertion in the case of phenyldimethylsulfonium
salts is nearly concomitant with release of the CH3
radical.
Lund suggests that when the potentials are right and
there is significant steric hindrance a reaction path
intermediate between the classical SN1 path and SN2
involving a nonchain electron transfer is possible.11 In
the present case the departure of the leaving group is
putatively involving a certain amount of SET character
mixed into the wave function of the transition state, and
not a full SET transfer. It is this character that we claim
Me
to correlate with pKlg
.
The consistency of the phenyldimethylsulfonium salts
in the three solvents used here is not surprising because
Me
it shows that the use of pKlg to model leaving group
behavior for phenyldimethylsulfonium ions will not vary
much from solvent to solvent. Determinations of E1/2
values in protic solvents, such as methanol, are not
possible because the solvent background obscures the
potential for the solute. We have previously discussed
Me
the solvent dependency of the pKlg scale for methyl
arenesulfonates, which also shows solvent independence.6
Together these two scales show that the type of charge
disposition, positive charge vs neutral, of the substrate
does not much affect the use of pKlgMe. This makes this
Me
pKlg scale a good one for mapping transition states.
Me
Subsequent papers will explore the use of pKlg with
common nucleophiles, phenolate anions, and R-nucleo-
philes. The use of this set of parameters is justified by
the self-consistency they show in a variety of nucleophilic
reactions in various solvents.
Ack n ow led gm en t. We thank the NSF for RUI
Grant CHE-9626663 in support of this work.
(10) Andrieux, C. P.; Robert, M.; Saeva, F. D.; Save´ant, J .-M. J . Am.
Chem. Soc. 1994, 116, 7864.
(11) H. Lund, H.; Daasbjerg, K.; Lund, T.; Pederson, S. U. Acc. Chem.
Res. 1995, 28, 313.
J O961658R