β Effect of phophorus functionalities

Article abstract of DOI:10.1021/ja9537181

Diphenylphosphinoyl (-P(=O)Ph₂) and diphenylthiophosphinoyl (-P(=S)Ph₂) respectively are modest and strong β-effect functionalities. When these groups are antiperiplanar to mesylate, the substrates solvolyze through unimolecular ioni

Full text of DOI:10.1021/ja9537181

3156
J. Am. Chem. Soc. 1996, 118, 3156-3167
â Effect of Phosphorus Functionalities
Joseph B. Lambert* and Yan Zhao
Contribution from the Department of Chemistry, Northwestern UniVersity,
EVanston, Illinois 60208-3113
ReceiVed NoVember 3, 1995X
Abstract: Diphenylphosphinoyl (-P(dO)Ph2) and diphenylthiophosphinoyl (-P(dS)Ph2) respectively are modest
and strong â-effect functionalities. When these groups are antiperiplanar to mesylate, the substrates solvolyze through
unimolecular ionization. In contrast, substrates with the synclinal (gauche) arrangement react bimolecularly with
solvent. The anti/gauche rate ratio is 440 for the oxide and 3.2 × 106 for the sulfide at 25 °C. The antiperiplanar
sulfide in fact reacts 220 times more rapidly than the analogous cyclohexyl substrate at 25 °C, despite the strong
electron-withdrawing nature of diphenylthiophosphinoyl. The R-secondary hydrogen/deuterium kinetic isotope effects
of 1.21 for the anti oxide and 1.26 for the anti sulfide suggest high sp2 character in the transition state, as expected
for stabilization by hyperconjugation in a vertical mechanism. Energies calculated at the MP2 level show that the
geometry with the P-C bond parallel to the empty carbocation p orbital is more stable than the perpendicular geometry
by 22.94 kcal mol-1 for PH2, 8.80 kcal mol-1 for P(O)H2, and 10.54 kcal mol-1 for P(S)H2, confirming significant
hyperconjugation for these substituents. The global minimum, however, is the bridged structure (four-membered
rings for the oxide and sulfide, three-membered ring for the simple phosphine), so that the mechanistic choice between
a vertical mechanism (hyperconjugation) and a nonvertical mechanism (bridging) is not clear-cut.
The formation of positive charge on carbon by cleavage of a
Scheme 1
C-X bond may be accelerated by the presence of certain atoms
or groups (M) attached to the adjacent or â carbon (M-C-
C-X). Two mechanisms of interaction between M and the
developing positive charge on carbon have been considered
(Scheme 1). In path A the group M attacks the backside of the
C-X bond in an intramolecular displacement to form a ring.
This mode has variously been called internal displacement,
anchimeric assistance, neighboring group participation, bridging,
or nonvertical participation. A wide variety of M groups have
been implicated in this mode of reactivity, including halogens,
amines, sulfides, alkoxy/alkoxides (all forming three-membered
rings), and acetoxy (forming a five-membered ring). This field
is part of classic mechanistic organic chemistry.1 The nucleo-
philic entity on all these functionalities is a lone pair of electrons
(the analogous reactions of π electrons are not under consid-
eration here).
The examination of strong electron donors such as silicon
that lack lone pairs or π electrons led to serious consideration
of path B in Scheme 1, which has variously been called
hyperconjugation, vertical participation, nonbridging, open, or
σπ conjugation.2 A highly polarizable and electron rich M-C
bond can stabilize the developing positive charge with minimal
alteration of the geometry around the M-C bond. In contrast
to path A, the transition state for path B has strong sp2 character
for the carbon bearing positive charge, and the open carbocation
can lead to a range of elimination and rearranged substitution
products. Silicon was the first such â-effect heteroatom to be
studied extensively2 followed by germanium, tin, lead,3 and
mercury.3,4
The most effective criteria for distinguishing between these
two modes of involvement have been stereochemistry,2,5 kinetic
isotope effect,6 and product studies. The optimum geometry
for nonvertical participation (path A) is the antiperiplanar
arrangement of the M-C-C-X fragment. All other geometries
are ineffective, as the displacement must occur on the backside.
Hyperconjugation in the transition state for path B, however, is
modulated as the cosine-squared function of the M-C-C-X
dihedral angle, with favorable interactions occurring at both the
antiperiplanar (180°) and synperiplanar (0°) geometries.5 Path
A should have an R-secondary kinetic deuterium isotope effect
(kH/kD for H and D on the carbon attached to the leaving group
X) of 1.00-1.05, since the transition state resembles a trigonal
bipyramid. For path B the transition state has sp2 character at
this carbon, with the expectation of an isotope effect close to
1.20.6 The substitution products (M-C-C-Y if the nucleo-
phile is Y- or HY) are expected to be formed with retention of
configuration at carbon as the result of two consecutive
inversions in the bridging mechanism of path A. Carbocationic
intermediates in path B should not exhibit retention in substitu-
tion products.
X Abstract published in AdVance ACS Abstracts, March 15, 1996.
(1) Gould, E. S. Mechanism and Structure in Organic Chemistry; Holt,
Rinehart and Winston: New York, 1959; Chapter 14.
(2) For the case of the â effect of silicon, see: Lambert, J. B. Tetrahedron
1990, 46, 2677-2689 and references therein.
(3) Davis, D. D.; Gray, C. E. J. Org. Chem. 1970, 35, 1303-1307.
Jerkunica, J. M.; Traylor, T. G. J. Am. Chem. Soc. 1971, 93, 6278-6279.
Lambert, J. B.; Wang, G.-t.; Teramura, D. H. J. Org. Chem. 1988, 53, 5422-
5428.
Finally, two subtleties of these mechanisms should be noted.
The bridged and open carbocation intermediates may be in
(5) Lambert, J. B.; Chelius, E. C. J. Am. Chem. Soc. 1990, 112, 8120-
8126.
(6) Lambert, J. B.; Emblidge, R. W.; Malany, S. J. Am. Chem. Soc. 1993,
115, 1317-1320.
(4) Kreevoy, M. M.; Kowitt, F. R. J. Am. Chem. Soc. 1960, 82, 739-
745. Lambert, J. B.; Emblidge, R. W. J. Phys. Org. Chem. 1993, 6, 555-
560.
0002-7863/96/1518-3156$12.00/0 © 1996 American Chemical Society

â Effect of Phosphorus Functionalities
J. Am. Chem. Soc., Vol. 118, No. 13, 1996 3157
equilibrium, as noted in Scheme 1. Thus the rate-determining
step may lead to one intermediate, which equilibrates with the
other form, and products may be formed from either structure.
Secondly, there may be strong vertical stabilization of the open
carbocation, but the nonvertical pathway can still be kinetically
favored.
solvolyzes by solvent participation (ks), analogous to SN2 or
E2, but the biased trans form solvolyzes by a carbocation
pathway (kc), analogous to SN1 or E1, or by anchimeric
assistance (k∆). The rate acceleration of the biased trans form
over the cis form was 12.7 in 97% trifluoroethanol (TFE) at 25
°C and 19.4 in hexafluoro-2-propanol (HFIP) at 70 °C. This
modest acceleration means that the biased trans form solvolyzes
more slowly than cyclohexyl tosylate, by a factor of about 15,
a number that still is remarkably small for such a strongly
electron-withdrawing group. This study showed that the anti-
periplanar diethyl phosphonate group changes the mechanism
from the normal ks process for a cyclohexyl tosylate to a
unimolecular ionization process and that the effect has a
stereochemical dependence. Thus diethyl phosphonate joined
carbonyl8 as an electron-withdrawing group that exhibits weak
â-effect activity.
This study left many questions unanswered. Although
hyperconjugation (path B) could explain the results, the pos-
sibility that the phosphonate group could nucleophilically
participate to form a four-membered ring (path A) was not
eliminated.9 Product studies provided the only evidence favor-
ing hyperconjugation over bridging. Phosphonate in its most
favorable stereochemistry (1′) did not provide a net rate
acceleration with respect to cyclohexyl, so that it clearly was
at best a weak â-effect group: strong enough to alter the
mechanism to kc or k∆ but still too weak to provide a rate
acceleration. In contrast, trimethylsilyl causes an acceleration2
of at least 1012. Consequently, we have sought other potential
phosphorus-containing â-effect groups and report herein our
results with phosphine oxide and phosphine sulfide.
Our phosphonate work was inspired by the observation by
Mastalerz et al. that the phosphonic acid group (P(O)(OH)2)
undergoes a Wagner-Meerwein 1,2-shift intact.10 The group
must have some carbocation-stabilizing ability in order to retain
P-C bonding (bridging would lead to P-O bonding). Warren
et al. analogously observed that phosphine oxides undergo
Wagner-Meerwein rearrangement intact.11 Since phosphine
oxide is encumbered with fewer electron-withdrawing oxygen
atoms than phosphonate, we thought that it might be a better
â-effect group. By further analogy, phosphine sulfides should
have increased polarizability and provide even better hyper-
conjugation. We report herein the experimental studies that
demonstrate that these functionalities are modest to strong
â-effect groups. We have also examined P(O)(OH)2, P(O)H2,
P(S)H2, and PH2 calculationally in the system M-CH2-CHMe+
in order to assess their hyperconjugative and bridging abilities.
Few if any phosphorus functionalities have been examined
as potential â-effect groups. Phosphines with the requisite
saturated group (R2P-CH2CH2X) tend to be highly air and water
sensitive and probably have not been examined, primarily
because they are difficult to handle. In more stable higher
valencies, in which the lone pair is replaced by a heteroatom
such as oxygen or sulfur, the phosphorus functionality is
extremely electron withdrawing, making the cationic reactions
of Scheme 1 very slow. The ability of R-electron-withdrawing
groups to stabilize cations has been the subject of considerable
recent investigation.7 The only â-electron-withdrawing group
(by σp) that has been studied is carbonyl.8ab Appreciable rate
accelerations are observed only for γ- and δ-carbonyl groups
(ketones and esters), which form five- and six-membered rings
via path A (including Winstein’s famous acetoxonium ions).
Little or no rate acceleration is observed for closure of four-
membered rings by â-carbonyl groups (the oxygen is the
nucleophile). Path A to form a four-membered ring thus is not
expected to be a kinetically favorable mechanism, from the
carbonyl precedents. Destabilization of an open carbocation
by the polar effect of a strongly electron-withdrawing â group
(M), however, also militates against path B. Electron-withdraw-
ing groups on silicon usually shut down its â activity.8c Thus
prior work with electron-withdrawing â groups was not
encouraging.
We have recently evaluated diethyl phosphonate (P(O)(OEt)2)
as a potential â-effect group.9 Our approach was to compare
the systems in which the M-C-C-X (M ) P(O)(OEt)2, X )
OTs) fragment is part of a cyclohexane system. The geometry
of the cyclohexane ring permits two stereochemistries, in which
the M-C-C-X dihedral angles respectively are antiperiplanar
in the biased trans form 1 or 1′ and synclinal in the cis form 2.
(PO(OEt) )
(PO(OEt) )
(PO(OEt) )
3
1'
2
2
2
2
Experimental Results and Discussion
Synthesis. The synthetic manipulations are summarized in
Scheme 2. The unbiased trans phosphine oxide 3(POPh2)-OH
was prepared readily by ring opening of cyclohexene oxide with
diphenylphosphide anion followed by oxidation with hydrogen
peroxide. A mixture of cis- and trans-4-tert-butylcyclohexene
oxide was treated in the same fashion to give the biased trans
oxide 1(POPh2)-OH after purification. Its structure was con-
firmed by 1D and 2D NMR experiments. The phosphinoyl
group of the unbiased trans compound was equilibrated with
butyllithium to a cis/trans mixture, from which the cis oxide
2(POPh2)-OH was obtained by chromatography. For the
R-deuterated version of the biased trans oxide, 1(POPh2-d)-OH,
4-tert-butylcyclohexanone was reduced to the alcohol with
2
3
(POPh )
1(POPh )
(POPh )
2
2
2
(PSPh )
(PSPh )
2
(PSPh )
2
3
1
2
2
For the phosphonate case, we prepared 1′. In addition, we
examined the unbiased trans form 3, which is a conformational
mix of the two stereochemistries. We found that the cis form
(7) Creary, X. Chem. ReV. 1991, 91, 1625-1678.
(8) (a) Yoshida, M.; Takeuchi, K. J. Org. Chem. 1993, 58, 2566-2572.
(b) Perst, H. Oxonium Ions in Organic Chemistry; Verlag Chemie:
Weinheim, Germany, 1971; Chapter 6. (c) In addition, â silicon has been
made electron withdrawing by halogen substitution: Brook, M. A.; Neuy,
A. J. Org. Chem. 1990, 55, 3609-3616.
(10) Richtarski, G.; Mastalerz, P. Tetrahedron Lett. 1973, 4069-4070.
Richtarski, G.; Soroka, M.; Mastalerz, P.; Starzemska, H. Rocz. Chem. Ann.
Soc. Chim. Pol. 1975, 49, 2001-2005; Chem. Abstr. 1976, 85, 5576x.
(11) Cann, P. F.; Warren, S. J. Chem. Soc., Chem. Commun. 1970, 1026-
1027. Cann, P. F.; Howells, D.; Warren, S. J. Chem. Soc., Perkin Trans. 2
1972, 304-311.
(9) Lambert, J. B.; Emblidge, R. W.; Zhao, Y. J. Org. Chem. 1994, 59,
5397-5403.

3158 J. Am. Chem. Soc., Vol. 118, No. 13, 1996
Lambert and Zhao
Scheme 2a
Table 2. Rate Constants for Phosphine Sulfides (M ) P(S)Ph2)
substrate
solventa
temp, °C
rate, s-1
1-OMs (biased trans)
97% TFE
70.0b
35.1
30.0
25.0
35.1
95.0
90.0
85.0
80.0
70.0b
25.0b
65.0
60.0
55.0
25.0b
65.0
65.0
65.0
65.0
65.0
9.34 × 10-2
2.45 × 10-3
1.32 × 10-3
7.27 × 10-4
1.95 × 10-3
8.13 × 10-6
4.71 × 10-6
2.22 × 10-6
1.26 × 10-6
3.18 × 10-7
2.28 × 10-10
8.49 × 10-4
4.86 × 10-4
2.54 × 10-4
4.22 × 10-6
6.78 × 10-4
6.03 × 10-4
8.49 × 10-5
1.29 × 10-4
1.49 × 10-4
1-1-d-OMs (biased trans) 97% TFE
2-OMs (cis)
97% TFE
3-OMs (trans)
97% TFE
80% TFE
60% TFE
80% EtOH
70% EtOH
60% EtOH
a (a) Ph2PLi; (b) H2O2; (c) BuLi; (d) H3O+, chromatography; (e)
HSiCl3/pyridine; (f) sulfur; (g) crystallization; (h) p-nitrobenzoyl
chloride/pyridine; (i) K2CO3/MeOH.
a See footnote a in Table 1. b Calculated from rates at other
temperatures.
modified) of Rickborn and Quartucci12 was used to obtain the
pure product by conversion to the p-nitrobenzoate, recrystalli-
zation, and hydrolysis. The cis sulfide 2(PSPh2)-OH was
obtained from the analogous phosphine oxide by treatment with
pyridine and trichlorosilane to form the phosphine followed by
treatment with sulfur. The structures of the cis sulfide and the
biased trans sulfide were determined by X-ray crystallography
as p-nitrobenzoate derivatives, the details of which are given
in the Experimental Section. The deuterated biased trans sulfide
1(PSPh2-d)-OH was obtained from 4-tert-butylcyclohexene-1-d
as with the undeuterated version.
Kinetics. The alcohols were converted to the mesylates and
solvolyzed in aqueous mixtures of ethanol (EtOH), trifluoro-
ethanol, and hexafluoro-2-propanol. The rates for the oxides
are given in Table 1 and those for the sulfides in Table 2.
Experiments were carried out as a function of temperature in
97% TFE, and the activation parameters are given in Table 3.
Raber-Harris plots were carried out for the biased trans oxide
(Figure 1), the unbiased trans oxide (Figure 2), and the unbiased
trans sulfide (Figure 3). The rates of the cis oxide and sulfide
were too slow to provide the ethanol data for such a plot, and
the plot was deemed unnecessary for the biased trans sulfide.
Product studies were carried out in both 97% TFE and 80%
EtOH. The products include structures 4-8, whose proportions
are given in Table 4.
Table 1. Rate Constants for the Phosphine Oxides (M ) P(O)Ph2)
substrate
solventa
temp, °C
rate, s-1
1-OMs (biased trans)
97% TFE
70.0
65.0
55.0
47.0
45.0b
25.0b
65.0
65.0
65.0
65.0
45.0
55.0
95.0
90.0
85.0
80.0
70.0b
25.0b
75.0
65.0
55.0
25.0b
65.0
65.0
65.0
65.0
65.0
70.0c
65.0c
45.0c
25.0c
45.0
5.37 × 10-4
3.12 × 10-4
9.95 × 10-5
3.79 × 10-5
2.95 × 10-5
2.05 × 10-6
3.86 × 10-4
7.58 × 10-5
1.33 × 10-4
1.90 × 10-4
5.74 × 10-4
8.17 × 10-5
8.04 × 10-6
4.83 × 10-6
3.07 × 10-6
2.09 × 10-6
7.68 × 10-7
4.62 × 10-9
1.25 × 10-4
3.69 × 10-5
1.11 × 10-5
1.58 × 10-7
4.88 × 10-5
6.48 × 10-5
2.08 × 10-5
2.74 × 10-5
5.16 × 10-5
1.79 × 10-4
1.21 × 10-4
2.22 × 10-5
3.24 × 10-6
1.69 × 10-4
80% TFE
80% EtOH
70% EtOH
60% EtOH
97% HFIP
1-1-d-OMs (biased trans) 97% TFE
2-OMs (cis)
97% TFE
3-OMs (trans)
97% TFE
80% TFE
60% TFE
80% EtOH
70% EtOH
60% EtOH
97% TFE
cyclohexyl mesylate
97% HFIP
4
5
a Percentages are wt/wt for trifluoroethanol (TFE) and hexafluoro-
2-propanol (HFIP) and v/v for ethanol (EtOH). b Calculated from rates
at other temperatures. c Calculated from ref 5.
lithium aluminum deuteride, the alcohol was converted to 4-tert-
butylcyclohexene-1-d by treatment with mesyl chloride and
collidine, the alkene was epoxidized, and the epoxide was ring
opened to the desired product on purification.
The unbiased trans phosphine sulfide 3(PSPh2)-OH was
prepared in the same fashion but with treatment of the
intermediate phosphine with elemental sulfur instead of hydro-
gen peroxide. An analogous procedure also produced the biased
trans sulfide 1(PSPh2)-OH, although the final isomeric mixture
could not be separated. As a result, the method (locally
6
7
8
Kinetic Comparisons. Both anchimeric assistance and
hyperconjugation should be manifested by a rate enhancement.
We have made comparisons in two ways. First, the ratio of
the functionalized systems 1-3 to cyclohexyl for the same
leaving group X gives an overall measure of the effect of M,
(12) Rickborn, B.; Quartucci, J. J. Org. Chem. 1964, 29, 2476-2477.

â Effect of Phosphorus Functionalities
J. Am. Chem. Soc., Vol. 118, No. 13, 1996 3159
Table 3. Activation Parameters in 97% TFE at 25 °C
Ea,
∆Hq,
∆Sq,
∆Gq,
substrate
kcal mol-1
25.2
log A
kcal mol-1
cal deg-1 mol-1
kcal mol-1
rb
1(POPh2)-OMs
(biased trans)
2(POPh2)-OMs
(cis)
3(POPh2)-OMs
(trans)
1(PSPh2)-OMs
(biased trans)
2(PSPh2)-OMs
(cis)
12.76
24.5
22.5
26.8
21.4
32.0
26.0
17.3
-2.3
-21.3
0.3
25.2
28.8
26.7
21.7
30.6
24.8
24.9
1.000
23.2
27.5
22.0
32.7
26.6
18.0
8.65
13.35
12.96
14.33
14.15
7.72
0.997
0.9998
0.9999
0.998
-1.3
4.7
3(PSPh2)-OMs
(trans)
4.0
0.9994
cyclohexyl mesylatea
-25.5
a Parameters calculated from ref 5. b Correlation coefficient.
Table 4. Product Studies (%)
the biased trans phosphonate 1′ the tert-butyl group is at a
different location from 1 for the oxide and sulfide, but the
change has no bearing on these comparisons). The cis forms
should provide a measure of the polar effect of the â substituent.
At 25 °C, the cis phosphonate solvolyzes about 180 times more
slowly than cyclohexyl in 97% TFE, the cis oxide 700 times
more slowly, and the cis sulfide 20 000 times more slowly.
These results indicate that the phosphine sulfide is the most
electron-withdrawing group. The polar effects are discussed
more extensively in the section on calculational results. In each
case, however, the slow rates of the cis forms document polar
inhibition of the solvolysis process for the synclinal stereo-
chemistry. The mechanism probably involves rate-determining
reaction with solvent.
Again at 25 °C, the biased trans phosphonate solvolyzes only
14 times more slowly than cyclohexyl. We characterized this
reduced ratio as the result of a weak but significant â effect of
diethyl phosphonate.9 The biased trans phosphine oxide is only
1.5 times as slow as cyclohexyl, and in the more ionizing solvent
HFIP (Table 1) the oxide is 3.4 times faster than cyclohexyl.
Thus the â effect of diphenylphosphine oxide is just sufficient
to overcome the strong polar effect of the group. Finally, the
sulfide clearly qualifies as a strong â-effect group, as the biased
trans form reacts 220 times faster than cyclohexyl.
In terms of the biased trans to cis ratios, the figures for the
phosphonate, oxide, and sulfide respectively are 12.8, 450, and
3.2 × 106. Table 5 also shows the values for these ratios at 70
°C, for which the trends are the same. Thus kinetic extrapola-
tions do not influence the interpretations. These results show
that the phosphine sulfide is a very strong â-effect group despite
its large polar effect, the phosphine oxide is respectable, and
the previously studied phosphonate is weak.
Molecularity. We have used the method developed by
Raber, Harris, and their co-workers for the assessment of
molecularity under these pseudo-first-order conditions.13 These
authors constructed plots of reaction rates in aqueous ethanol
and aqueous TFE, with the substrate on the y coordinate and
1-adamantyl bromide on the x coordinate. A substrate reacting
by an ionization mechanism (either carbocation (kc) or anchi-
meric assistance (k∆)) gives a single straight line, since the rates
on both coordinates are responding to the same factors.
Increased water content in aqueous ethanol enhances solvent
ionizing power but has little effect on nucleophilicity. In
contrast, increased water content in aqueous TFE enhances
solvent nucleophilicity but has little effect on ionizing power.
Raber and Harris found that reactions in which solvent is present
97% 80%
TFE EtOH
substrate
product
1(POPh2)-OMs
(biased trans)
4 (R ) t-Bu)
81
11
80
9
6
5 (R ) t-Bu)
6 (X ) OEt)
6 (X ) OCH2CF3)
6 (X ) OH)
4
4
44
47
5
15
78
3
2(POPh2)-OMs (cis)
4 (R ) H)
5 (R ) H)
3(POPh2) (X ) OEt)
3(POPh2) (X ) OCH2CH3)
3(POPh2) (X ) OH)
5
4
81
12
7
4
90
10
3(POPh2)-OMs (trans) 4 (R ) H)
5 (R ) H)
2(POPh2) (X ) OCH2CF3)
7 (R ) t-Bu)
1(PSPh2)-OMs
(biased trans)
100
100
2(PSPh2)-OMs (cis)
7 (R ) H)
8
42
58
100
11
89
100
3(PSPH2)-OMs (trans) 7 (R ) H)
Table 5. Rate Ratios in 97% Trifluoroethanol
substrate
cyclohexyl
1′(PO(OEt)2) (biased trans) 0.069
25 °C
70 °C
1.0a
1.0a
0.13
2(PO(OEt)2) (cis)
1′/2(PO(OEt)2)
1(POPh2) (biased trans)
2(POPh2) (cis)
0.0054
0.038
12.7b
3.5b
0.63
0.0014
3.0
0.0043
1/2(POPh2)
1(PSPh2) (biased trans)
2(PSPh2) (cis)
440
700
220
520
7.0 × 10-5
0.0018
1/2(PSPh2)
3.2 × 106
2.9 × 105
a Rate ratios in this column are with respect to cyclohexyl tosylate
(1.70 × 10-6 s-1 at 25 °C, 2.77 × 10-4 s-1 at 70 °C) for the diethyl
phosphonate and with respect to cyclohexyl mesylate (3.24 × 10-6
s-1 at 25 °C, 1.79 × 10-4 s-1 at 70 °C) for the oxide and sulfide. b Rate
ratios in this column are for the biased trans form 1 to the cis form 2
for each system.
which may include steric effects, induction, and the two modes
of â-effect activity. Secondly, the ratio of the biased trans
(∼180°) to cis (∼60°) forms provides a measure of stereo-
chemically dependent components of the â effect. The bridge-
forming nonvertical process (path A) should be possible only
in the antiperiplanar (180°) stereochemistry. Hyperconjugation
(path B) should have a cosine-squared dependence, with maxima
at 0° and 180° and a minimum at 90°. The â effect thus should
be vastly reduced in the cis stereochemistry by either mecha-
nism. If only an inductive effect is operating, the cis and biased
trans forms should have about the same rate.
(13) Raber, D. J.; Neal, W. C., Jr.; Dukes, M. D.; Harris, J. M.; Mount,
D. L. J. Am. Chem. Soc. 1978, 100, 8137-8146. Harris, J. M.; Mount, D.
L.; Smith, M. R.; Neal, W. C., Jr.; Dukes, M. D.; Raber, D. J. J. Am. Chem.
Soc. 1978, 100, 8147-8156.
Table 5 contains these rate comparisons for the phosphonate,9
oxide, and sulfide (it should be noted from Scheme 2 that in

3160 J. Am. Chem. Soc., Vol. 118, No. 13, 1996
Lambert and Zhao
Figure 1. Raber-Harris plot for r-4-tert-butyl-trans-2-(diphenylphos-
phinoyl)cyclohex-cis-1-yl mesylate (1(P(O)Ph2)-OMs).
Figure 2. Raber-Harris plot for trans-2-(diphenylphosphinoyl)-
cyclohexyl mesylate (3(P(O)Ph2)-OMs).
in the transition state (SN2 or E2, also called ks) generally give
two separate, diverging lines for the two solvents. The slope
for the TFE points is greater than that for ethanol because the
rates in TFE are changing faster for the y than the x axis.
In the case of the diethyl phosphonate,9 we found that the
cis substrate 2(PO(OEt)2) produced a two-line plot that is typical
of a bimolecular (ks) process, but the biased trans substrate 1′-
(PO(OEt)2) produced a one-line plot that is typical of an
ionization process (kc or k∆). Interestingly, the unbiased trans
substrate 3(PO(OEt)2) produced an intermediate plot, in which
there were two lines but with little difference in slope. The
appearance of this plot suggested that the mechanism was kc or
k∆ in the more ionizing TFE solvents but a mixture of bi- and
unimolecular processes in ethanol. In ethanol some of the
diequatorial form must be contributing to the rate expression
with some ks contribution.
The situation with the phosphine oxide is similar to that with
the phosphonate. The biased trans form 1(POPh2) produces a
one-line Raber-Harris plot (Figure 1) that is clearly indicative
of an ionization process (kc or k∆). The unbiased form
3(POPh2), like the analogous phosphonate, produces an inter-
mediate plot (Figure 2), with the TFE points bunched in a kc
fashion but two separate lines in a ks fashion. The correlation
coefficient for all six points is 0.82, the same as for the
phosphonate.9 We were not able to construct a Raber-Harris
plot for the cis substrate 2(POPh2) because the rates were too
slow to be measured in ethanol below its boiling point.
For the phosphine sulfide, the unbiased trans form 3(PSPh2)
produces a one-line Raber-Harris plot (Figure 3, correlation
coefficient of 0.95). Even in ethanol this substrate appears to
be reacting by an ionization process (kc or k∆). As with the
oxide, we were not able to construct a Raber-Harris plot for
the cis substrate because of the exceedingly slow rates in ethanol.
The Raber-Harris test of molecularity is in consonance with
the kinetic comparisons. The faster rate of the biased trans form,
in comparison with both cyclohexyl and the cis form, accom-
Figure 3. Raber-Harris plot for trans-2-(diphenylthiophosphinoyl)-
cyclohexyl mesylate (3(P(S)Ph2)-OMs).
panies a change in molecularity from a solvent-assisted mech-
anism (ks) to a unimolecular ionization mechanism (kc or k∆).
Products. Several pertinent comments may be made con-
cerning the product structures listed in Table 4. Firstly, the
products from the cis oxide are consistent with a solvent-assisted
reaction. Over 90% of the products result from elimination of
one of the antiperiplanar protons (positions 2 and 6 with respect
to the leaving group). The small amount of substitution occurs

â Effect of Phosphorus Functionalities
J. Am. Chem. Soc., Vol. 118, No. 13, 1996 3161
with inversion. Secondly, the products from the trans oxide
substrates require a carbocation mechanism. There is only one
antiperiplanar proton, but a significant amount of the products
come from elimination of a formally syn proton to give 5. The
only mechanistically viable route that avoids a syn elimination
is through a carbocation intermediate that can lead to either 4
or 5. Alternatively, phosphine oxide must be considered to favor
syn elimination by an unspecified steric effect. Thirdly, the
substitution products of the trans oxide substrates are formed
with inversion. Such a stereochemistry is common for reaction
within an ion pair. In the rate-determining step, the C-X bond
is broken to form an ion pair, which shields the front side of
the broken bond. Reaction of the carbocation then takes place
preferentially with inversion. This last observation is pertinent
to the question of vertical vs nonvertical pathways. The
nonvertical pathway leads to the bridged intermediate (path A
in Scheme 1) (M ) -PPh2-O-). Solvent would react with
the ring in a second inversion to form the substitution product
with retention. Retained stereochemistry has been a hallmark
of neighboring group participation. Thus the products do not
support a nonvertical mechanism, although substitution repre-
sents <10% of the product, and need not be a definitive indicator
of the major mechanism. Fourthly, the sulfides react exclusively
to form elimination products. The cis sulfide gives both
regiochemistries (7 and 8), but the trans sulfides give exclusively
7. Strong stabilization of the carbocation by the axial phosphine
sulfide group may inhibit elimination of any but the 6-axial
proton. Fifthly, neither the oxide nor the sulfide yielded any
fragmentation product, in which both the nucleofuge (OMs) and
the electrofuge (POPh2 or PSPh2) are eliminated to form a
hydrocarbon. Such a product might be expected when the
electrofuge bears an appreciable positive charge through hy-
perconjugation. Hagen and Mayr,14 however, obtained similar
results in elimination reactions involving the X-C-C-Si
bonds. They found that the amount of desilylation decreased
when the silicon atom bore bulky substituents. Attack of the
nucleophilic species then was directed more to carbon than to
silicon. The bulky phenyl rings may fulfill a similar role in
the present cases.
Chart 1
e
b
c'
c
Table 6. Hyperconjugative Ability as Measured by the Energy
Difference Between the Parallel (b) and Perpendicular (e) Forms
(∆Ebe, kcal mol-1) of M-CH2-C+HCH3
M
RHF/6-31G(d)
MP2/6-31G(d)
PH2
P(O)H2
P(S)H2
14.35
4.72
3.93
4.27
16.02
22.94
8.80
10.54
7.98
P(O)(OH)2
a
SiH3
22.20
a Data from ref 18.
We observed a kH/kD of 1.21 ( 0.01 for the phosphine oxide
at 55.0 °C and 1.26 ( 0.04 for the phosphine sulfide at 35.1 °C
in 97% TFE. These figures are strong evidence for a transition
state with high sp2 character, as expected for the vertical and
not the nonvertical process. Formation of a five-membered ring
in the k∆ reaction of MeO(CH2)3CHMeOBs exhibited a reduced
kH/kD of 1.08.16 To the extent that these secondary isotope
effects are valid indicators of transition state structure, these
observations are consistent with an open carbocation (path B)
rather than a bridged structure (path A).
Calculational Results and Discussion
The calculational work of Jorgensen and his co-workers17,18
provided considerable insight into the â effect of silicon.
Consequently, we have carried out analogous calculations on
the effects of four phosphorus groups (M ) PH2, P(dO)H2,
P(dS)H2, and P(dO)(OH)2). In addition to the hydrogen-
substituted analogues of the present and previous9 study, we
include the phosphine group, which has not been studied
experimentally because of its high reactivity with air and water.
As a measure of the hyperconjugative â effect, we compared
the energies of two conformations, in which the P-C bond is
either parallel to the empty orbital (the form Jorgensen called
bisected or b)17,18 or perpendicular (Jorgensen’s eclipsed form
or e), as seen in Chart 1. Hyperconjugation is optimal in the
parallel or b form and minimal in the perpendicular or e form.
In addition, we calculated the energies of the bridged form
(called cyclic or c by Jorgensen), which is a three-membered
ring for the phosphine (c in Chart 1, M ) PH2) and a four-
membered ring for the oxide, sulfide, or phosphonate (c′ in Chart
1: when M is P(O)(OH)2, R ) OH and X ) O; when M is
P(O)H2, R ) H and X ) O; when M is P(S)H2, R ) H and X
) S).
Product analysis thus is a mixed bag. The presence of
inverted rather than retained substitution products is consistent
with a vertical rather than a nonvertical mechanism, but these
products are observed only for the oxide and then in <10%.
The elimination products are consistent with a carbocation
process in the case of the trans substrates.
Isotope Effects. R-Secondary deuterium kinetic isotope
effects have been used widely in solvolysis reactions as a probe
for the hybridization of the carbon atom bearing the nucle-
ofuge.15 In a carbocation-forming reaction such as the vertical
(kc) process (path B in Scheme 1), the carbon atom changes
from sp3 in the ground state to sp2 in the intermediate, with a
transition state moving toward sp2. The R-hydrogen deuterium
isotope effect is normally in the range 1.15-1.25 for such a
process. For an SN2 reaction, with either direct displacement
by the solvent (ks) or internal displacement of the nucleofuge
by the electrofuge to form a ring (path A, k∆), the R-isotope
effect is small or even inverse (0.95-1.05). We applied this
method to the â-effect role of silicon and found an isotope effect
of 1.17 in TFE at 25 °C, experimentally the same as the 1.18
for 2-adamantyl tosylate, in which the process clearly is
carbocation formation, under the same conditions.6
Table 6 contains the difference in energy (∆Ebe) between the
parallel and perpendicular forms at two levels of calculation
for the four phosphorus functional groups as well as Jorgensen’s
calculations for the silicon group18 (M ) SiH3). The quantity
∆Ebe is a measure of hyperconjugation, which is maximal in
the parallel (b) form and absent in the perpendicular (e) form.
(14) Hagen, G.; Mayr, H. J. Am. Chem. Soc. 1991, 113, 4954-4961.
(15) Isaacs, N. Physical Organic Chemistry; Longman: Essex, U.K.,
1987; pp 265, 394-395, 599-600. Shiner, V. J., Jr. In Isotope Effects in
Chemical Reactions; Collins, C. J.; Bowman, N. S., Eds.; Van Nostrand
Reinhold: New York, 1970; pp 104-137. Harris, J. M. Prog. Phys. Org.
Chem. 1974, 11, 103-108.
(16) Eliason, R.; Tomicˇ, M.; Borcˇic´, S.; Sunko, D. E. Chem. Commun.
1968, 1490-1491.
(17) Wierschke, S. G.; Chandrasekhar, J.; Jorgensen, W. L. J. Am. Chem.
Soc. 1985, 107, 1496-1500.
(18) Ibrahim, M. R.; Jorgensen, W. L. J. Am. Chem. Soc. 1989, 111,
819-824.

3162 J. Am. Chem. Soc., Vol. 118, No. 13, 1996
Lambert and Zhao
Table 7. â-Effect Stabilization (M-CH2C+HCH3 + CH4 ) H-CH2C+HCH3 + M-CH3)
-E,a au
∆E,b kcal mol-1
systemc
RHF/6-31G(d)
MP2/6-31G(d)
RHF/6-31G(d)
MP2/6-31G(d)
H-CH3
40.195 17
117.381 16
381.486 26
458.690 61
458.667 73
458.692 41
456.360 85
533.542 83
533.535 32
533.580 98
779.015 84
856.190 51
856.184 26
856.237 62
606.172 36
683.356 73
683.349 92
683.387 56
156.420 80
407.481 63
407.456 10
407.470 70
40.332 44
117.745 52
381.720 91
459.169 51
459.132 95
459.182 77
456.779 25
534.197 80
534.183 79
534.241 67
779.369 78
856.788 77
856.771 98
856.846 98
606.947 49
684.368 13
684.355 40
684.407 89
156.917 58
407.932 89
407.897 52
407.926 57
H-CH2CH+CH3
PH2-CH3
0
0
b-PH2-CH2CH+CH3
11.52
-2.83
12.66
22.29
-0.65
30.61
d
e-PH2-CH2CH+CH3
c-PH2-CH2CH+CH3
P(O)H2-CH3
b-P(O)H2-CH2CH+CH3
-2.51
-7.23
21.43
3.44
-5.36
30.96
d
e-P(O)H2-CH2CH+CH3
c-P(O)H2-CH2CH+CH3
P(S)H2-CH3
b-P(S)H2-CH2CH+CH3
-7.10
-11.03
22.46
3.71
-6.83
40.24
d
e-P(S)H2-CH2CH+CH3
c-P(S)H2-CH2CH+CH3
P(O)(OH)2-CH3
d
b-P(O)(OH)2-CH2CH+CH3
-1.02
-5.29
18.33
4.05
14.83
-1.19
7.97
4.74
-3.24
29.70
6.58
22.13
-0.07
18.16
d
e-P(O)(OH)2-CH2CH+CH3
c-P(O)(OH)2-CH2CH+CH3
e
b-CH3-CH2CH+CH3
e
b-SiH3-CH2CH+CH3
d,e
e-SiH3-CH2CH+CH3
d,e
c-SiH3-CH2CH+CH3
a Calculated energy for a single species. b Difference in energy between the two sides of the title equation. c The prefix b signifies the parallel
form in Chart 1, e the perpendicular form, and c the bridged or cyclic form. d Not an energy minimum; calculated with a constraint. e Data from
ref 18.
Geometry was fully optimized at the lower level. Energies then
were calculated at the higher level for the geometries found at
the lower level. The MP2 calculations show uniformly higher
â effects, and we shall consider only these in the discussion.
The same trends are present in all the lower level calculations.
The phosphine â-hyperconjugative effect is comparable to the
well-known silicon â effect (22-23 kcal mol-1), yet it has never
been observed or discussed in the literature to our knowledge,
presumably because of its difficult accessibility. The other three
phosphorus functionalities show a strong but smaller â-hyper-
conjugative effect, in the general range 8-11 kcal mol-1, which
translates to a rate acceleration of about 1 000 000. These
calculations, like Jorgensen’s second set,18 are on secondary
carbocations, which correspond to the experimental structures.
Unfortunately, the substituents on phosphorus (phenyl or ethoxy)
cannot be modeled in a reasonable amount of calculation time,
so that the limitations of these calculations must be appreciated
not only in terms of absence of solvent effects but also in terms
of simplification of the substituents.
b) due to hyperconjugation. Thus hyperconjugation must
compensate for the polar effect to provide a rate acceleration
in comparison with the hydrogen case (which is cyclohexyl for
our experimental studies). The polar effects of these substituents
are well documented to be strongly electron withdrawing, as
measured, for example, by σm (0.55, 0.38, and 0.29, respectively,
for P(O)(OEt)2, P(O)Ph2, and P(S)Ph2) or σI (0.33, 0.26, and
0.15).19a Other measurements of σ values reverse this order,
making P(S)Ph2 the most electron withdrawing: σm ) 0.30,
0.44, and 0.45, and σI ) 0.21, 0.30, and 0.40.19b,c The
perpendicular or e forms were found not to be energy minima,
so the perpendicular geometry was set via the CH3-C+-C-P
dihedral angle, and then all other geometric factors were
optimized at the lower calculational level. The difference
between the value of ∆E for the b and e forms is ∆Ebe in Table
6, e.g., 3.71 - (-6.83) ) 10.54 for the sulfide.
The energies of the cyclic forms (c) of all phosphorus
functionalities were calculated to be lower than for the parallel
open forms (b), by about 8 kcal mol-1 for the phosphine, 28
kcal mol-1 for the oxide, 36 kcal mol-1 for the sulfide, and 25
kcal mol-1 for the phosphonate. Thus the bridged forms are
the global minima for all cases, although the effect of the omitted
ethyl or phenyl substituents is ignored. These huge stabilizations
clearly are not reflected in the kinetics. Thus the cyclic
phosphonate is stabilized by 29.70 kcal mol-1 compared with
the H system (H-CH2CH+CH3), but the diethyl derivative
solvolyzes more slowly than the H comparison (cyclohexyl
tosylate). Moreover, studies of acyclic carbonyl systems showed
no kinetic evidence for four-membered ring formation.8 The
kinetics more closely follow the hyperconjugation trends.
For each parallel conformation, the energies in Tables 6 and
7 correspond to the lowest energy form. Rotation about the
P-C bond, however, can generate a family of conformations.
We calculated the energy of the parallel (b) form as a function
of rotation around the P-C bond at 15° intervals for the
phosphine and the oxide. Figure 4 shows the results for the
phosphine. As seen from the structures on the figure, the lowest
In order to make direct comparisons between the individual
forms, thereby bringing in the bridged structures (c), we used
the isodesmic eq 1:
M-CH2-CHCH3+ + CH4 )
H-CH2-CHCH3+ + M-CH3 (1)
The energy difference between the left and right sides of this
equation represents the difference in the ability of hydrogen
and the â-substituent M to stabilize the positively charged vs
the neutral systems. This approach removes effects present in
the neutrals, so that the energy difference (∆E) is a measure of
the â effect of M vs H for a given geometry. The results are
given in Table 7, along with the absolute energies of all the
species.
For the phosphine and the silane,18 the perpendicular form
(e) is slightly destabilized by the â group. For the oxide, sulfide,
and phosphonate, the polar effect of the new atoms destabilizes
(19) (a) Hansch, C.; Leo, A. Substituent Constants for Correlation
Analysis in Chemistry and Biology; John Wiley & Sons: New York, 1979.
(b) Prikoszovich, W.; Schundlbauer, H. Chem. Ber. 1969, 102, 2922-2929.
(c) Johnson, A. W.; Jones, H. L. J. Am. Chem. Soc. 1968, 90, 5232-5236.
the perpendicular form by several kcal mol-1
. All four
phosphorus substituents show a clear stabilizing effect when
the M-C bond is parallel to the empty carbocation orbital (form

â Effect of Phosphorus Functionalities
J. Am. Chem. Soc., Vol. 118, No. 13, 1996 3163
+
Figure 5. Energy of H2P(dO)-CH2-CHCH3+ as a function of rotation
around the P-C bond. Structures are illustrated at the dihedral angles
38°, 158°, 188°, 248°, and 398°, which correspond to local minima or
maxima.
Figure 4. Energy of PH2-CH2-CHCH3 as a function of rotation
around the P-C bond. Structures are illustrated at the dihedral angles
72°, 177°, 252°, 327°, and 432°, which correspond to local minima or
maxima.
energy form (dihedral angle of about 180°) has the lone pair
antiperiplanar to the C-C+ bond, i.e., adverse to bridge
formation. There is a higher energy minimum at about 0°, when
the lone pair approximately eclipses the C-C+ bond, favorable
for bridge formation. For the oxide (Figure 5), the lowest energy
form is close to 0°, with the PdO bond nearly eclipsed with
the C-C+ bond and favorable for bridging. The minimum
energy angle of about 30° probably results in order to avoid
H-H eclipsing. There is a second, higher energy minimum at
about 180°, when the PdO bond is antiperiplanar to the C-C+
bond and unfavorable for bridging.
sulfide and 0.080 Å for the phosphine, compared with 0.132 Å
for the C-Si bond in the silane.18 The diminution of the C-C+
bond ranges from 0.018 Å for the phosphonate to 0.052 Å for
the phosphine, compared with 0.075 Å for the silane.18 As the
M group leans toward the carbocation, the M-C-C+ angle is
diminished by 5.7° for the oxide up to 22.6° for the phosphine,
compared with 19.1° for the silane.18 Thus in all cases
hyperconjugation is clearly manifested by geometric distortions
in the parallel form (b) that are absent in the perpendicular form
(e).
The conformations of these carbocations therefore are deter-
mined by two major factors: (1) Polar effects influence the
arrangement around the C-P bond, so that the negative end of
the dipole is close to the positive charge; and (2) hyperconju-
gation controls the arrangement around the C-C+ bond, so that
the empty p orbital is parallel to the C-P bond. Eclipsing and
nonbonded interactions also contribute, particularly when the
bulky phenyl groups are present.
Concluding Remarks
The experimental results indicate that phosphine oxide is a
modest â-effect group and phosphine sulfide is a strong â-effect
group. This conclusion is supported by anti/gauche rate ratios
provided by the biased trans and cis substrates 1 and 2 in 97%
TFE. For the oxide the ratio is 440 at 25 °C, and for the sulfide
it is 3.2 × 106. Compared with cyclohexyl, the biased trans
oxide reacts 0.63 times as fast and the biased trans sulfide 220
times faster. Both functionalities have compensated for a strong
rate-retarding polar effect (σm ) 0.44 for diphenylphosphine
oxide and 0.45 for diphenylphosphine sulfide19b). These rate
ratios indicate that the electronic effects responsible for the â
effect are extremely strong in the antiperiplanar stereochemistry
for the sulfide and modest for the oxide.
As is the case for the silane substituent,17,18 hyperconjugation
influences the bond lengths and angles of the phosphorus
carbocations (Table 8). The valence bond structure for hyper-
conjugation in the parallel form (b) is depicted by 9b, in which
One explanation for these observations that may be eliminated
is polar destabilization of the ground state. Such a rationale
has been put forward to explain rate accelerations for electron-
withdrawing R-effect groups.20 When one carbon carries both
the nucleofuge X and the perturbing group M (M-C-X),
leading to a carbocation M-C+<, the influence of M is referred
9b
9a
the C-M bond is lengthened, the C-C+ bond is shortened,
and the M-C-C+ angle is diminished in comparison with the
perpendicular form (e), in which hyperconjugation is absent.
The elongation of the C-P bond in the b form over the e form
ranges from 0.039 Å for the phosphonate to 0.078 Å for the
(20) Wu, Y.-D.; Kirmse, W.; Houk, K. N. J. Am. Chem. Soc. 1990, 112,
4557-4559.

3164 J. Am. Chem. Soc., Vol. 118, No. 13, 1996
Lambert and Zhao
Table 8. Geometric Comparison of the Parallel (b) and Perpendicular (e) Forms (M-C1H2-C2H+CH3)
P-C1 (Å)
C1-C2 (Å)
P-C1-C2 (deg)
M
b
e
b - e
b
e
b - e
b
e
b - e
PH2
P(O)H2
P(S)H2
P(O)(OH)2
1.971
1.917
1.951
1.873
2.070b
1.891
1.866
1.873
1.834
1.938b
0.080
0.051
0.078
0.039
0.132b
1.405
1.444
1.435
1.447
1.386
1.457
1.469
1.464
1.465
1.461
-0.052
-0.025
-0.029
-0.018
-0.075
93.7
105.1
109.3
105.1
100.0c
116.3
110.8
115.1
112.6
119.1c
-22.6
-5.7
-5.8
-7.5
a
SiH3
-19.1c
a Data from ref 18. b Si-C1 bond. c Si-C1-C2 angle.
to as an R effect. Rate accelerations (or smaller than expected
decelerations) have been interpreted in terms of transition state
stabilization through resonance delocalization, +MdC<.21 Wu
et al.20 pointed out that the rate changes also can be interpreted
in terms of ground state destabilization, since the repulsive
geminal interaction between M and X in the ground state is
eliminated on loss of X. In the case of the â-effect accelerations
in the present study, this argument is reversed. In the anti-
periplanar arrangement, the C-M and C-X dipoles are faVor-
ably arranged, so that loss of X would cause a deceleration if
anything. The observed accelerations do not support ground
state repulsion effects.
the Raber-Harris plots for the biased trans oxide and the trans
sulfide are one line, so they do not resemble the example of
sulfur participation.
The calculations summarized in Tables 6-8 clearly demon-
strate that all these phosphorus functionalities can provide
hyperconjugative stabilization for a â carbocation, in terms of
both energetics and structural perturbations. The effect of
phosphine is as strong as the well-studied effect of silane. More
surprising is that the strongly electron-withdrawing groups can
provide hyperconjugation valued at 8-11 kcal mol-1 at the MP2
level (Table 6). The hyperconjugated valence structures 9b
would have the structures 10-12 respectively for the diethyl
Thus the positive â effects of phosphine oxide and sulfide
must result from transition state-stabilizing effects such as those
depicted in Scheme 1. The strongest case for bridging (path
A) comes from the calculations. The figures in Table 7 indicate
that the three- and four-membered ring intermediates (c and c′
in Chart 1) represent the global energy minimum for all four
cases studied (phosphine and phosphonate as well as phosphine
oxide and sulfide). The effects of substituent introduction
(phenyl or ethyl) might dilute this preference.
10
11
12
There are three lines of experimental evidence, none entirely
convincing, that disfavor the bridging mechanism for the biased
trans substrates (path A). (1) A bridging mechanism should
produce a substitution product with retention via two inversions.
The substitution products observed for the oxide (Table 4) show
inverted stereochemistry, which is more in line with an ion pair
carbocation mechanism. Other products from the biased trans
form require the intermediacy of a carbocation to avoid a syn
elimination. (2) The R-hydrogen deuterium secondary isotope
effects of 1.21 for the oxide and 1.26 for the sulfide support a
transition state with carbocation character (path B) rather than
the internal displacement transition state expected for anchimeric
assistance (path A). (3) The Raber-Harris plots (Figures 1 and
3) clearly indicate a transition state lacking solvent involvement.
Such one-line plots are consistent with either a carbocation (kc,
path B) or an anchimeric assistance (k∆, path A) mechanism.
In their original work, Raber and Harris13 offered several
examples of both kc and k∆ reactions that gave one-line plots.
None of their early k∆ reactions, however, involved a heteroatom
displacement as in Path A. Instead they were assistance by σ
or π bonds. The only example of a Raber-Harris plot for a k∆
reaction with a heteroatom is for systems of the type RSCH2-
CH2Cl, which would form a three-membered episulfonium ion
by path A.22 The authors observed, however, that the reaction
gave a two-line plot, seemingly indicative of solvent assistance.
They eventually interpreted the results not in terms of nucleo-
philic solvent assistance for removal of the leaving group but
instead in terms of electrophilic solvent assistance.23 In contrast,
phosphonate, the diphenylphosphine oxide, and the diphen-
ylphosphine sulfide. Table 7 shows that, at the MP2 level,
inductive destabilization is about 3-6 kcal mol-1 (values for
the perpendicular or e form) but that stabilization in the parallel
or b form is more than enough to make up for this problem.
Thus hyperconjugation of phosphonate, phosphine oxide, and
phosphine sulfide is a calculational reality.
Whether hyperconjugation of these groups is an experimental
reality is less clear. Although there is some evidence against
the bridging pathway, there is no evidence that is compelling
for the vertical pathway, the secondary isotope effect being its
strongest argument. These interesting phosphorus functionalities
thus require further examination.
Experimental Section
trans-2-(Diphenylphosphinoyl)cyclohexanol (3(P(O)Ph2)-OH).24
An oven-dried, N2-flushed, 400 mL, round-bottomed flask was equipped
with a Claisen adapter to which a rubber septum and a condenser had
been fitted. The vessel was charged with a magnetic stirring bar, Li
(1.12 g, 0.16 mol), PPh3 (21.12 g, 0.08 mol), and 160 mL of anhydrous
tetrahydrofuran (THF). During 1 h of stirring, an exothermic reaction
took place and the color turned deep red and then intensified to black.
tert-Butyl chloride (7.46 g, 0.08 mol) in dry THF (20 mL) was
introduced slowly into the solution via a syringe. The reaction mass
was stirred for another 15 min before it was heated to reflux for 30
(21) Gassman, P. G.; Talley, J. J. J. Am. Chem. Soc. 1980, 102, 1214-
1216.
(22) McManus, S. P.; Neamati-Mazraeh, N.; Hovanes, B. A.; Paley, M.
S.; Harris, J. M. J. Am. Chem. Soc. 1985, 107, 3393-3395.
(23) Harris, J. M.; McManus, S. P.; Sedaghat-Herati, M. R.; Neamati-
Mazraeh, N.; Kamlet, M. J.; Doherty, R. M.; Taft, R. W.; Abraham, M. H.
In Nucleophilicity; Harris, J. M., McManus, S. P., Eds.; American Chemical
Society: Washington, DC, 1987; pp 247-254.
(24) Bridges, A. J.; Whitham, G. H. J. Chem. Soc., Chem. Commun.
1974, 142-143. Aguilar, A. M.; Beisler, J.; Mills, A. J. Org. Chem. 1962,
27, 1001-1011.

â Effect of Phosphorus Functionalities
J. Am. Chem. Soc., Vol. 118, No. 13, 1996 3165
min. The mixture was cooled to 0 °C with an ice bath, and cyclohexene
oxide (4.05 g, 0.04 mol) in THF (20 mL) was added dropwise to the
solution through a syringe. The solution was allowed to warm to room
temperature and stirred overnight. Addition of 5% acetic acid (20 mL)
changed the color of the solution to pale green. Then 10% H2O2 (60
mL) was introduced slowly into the system with cooling by an ice
bath. The solution was stirred at 0 °C for another 4 h and extracted
with CHCl3 (3 × 50 mL). The combined organic solution was dried
(MgSO4) and concentrated, and the residue was crystallized from
acetone-petroleum ether to yield 9.80 g (79%) of a white solid: mp
150-152 °C; 1H NMR (CDCl3) δ 0.86-1.50 (m, 4H), 1.58-1.82 (m,
3H), 2.12 (m, 1H), 2.54 (m, 1H), 5.42 (s, 1H), 7.46-7.54 (m, 6H),
7.56-7.84 (m, 4H); 13C NMR (CDCl3) δ 24.0, 25.5 (d, JCP ) 13.1
Hz), 26.6, 35.3 (d, JCP ) 10.2 Hz), 44.2 (d, JCP ) 69.2 Hz), 69.5 (d,
JCP ) 5.4 Hz), 128.3, 128.4, 128.7, 128.9, 131.0, 131.1, 132.1, 132.2,
132.4, 132.5; 31P NMR (CDCl3) δ 42.0; MS (EI) m/z 301 (3), 300
(M+, 11), 278 (11), 277 (27), 273 (12), 272 (61), 271 (10), 257 (29),
230 (24), 229 (100), 215 (12), 203 (20), 202 (100), 201 (57); HRMS
(M+) calcd 300.1279, obsd 300.1275. Anal. Calcd for C18H21PO2:
C, 71.99; H, 7.05. Found: C, 71.24; H, 6.83.
38.7, 42.0, 42.9, 76.4 (d, JCP ) 5.9 Hz), 128.6, 128.7, 128.8, 129.0,
130.7, 130.9, 131.1, 131.2, 131.8, 132.0, 132.1; 31P NMR (CDCl3) δ
31.3.
r-4-tert-Butyl-trans-2-(diphenylphosphinoyl)cyclohex-cis-1-yl Me-
1
sylate (1(P(O)Ph2-OMs)): H NMR (CDCl3) δ 0.71 (s, 9H), 1.19-
1.39 (dq, 1H), 1.59-1.74 (m, 3H), 1.87-1.98 (m, 2H), 2.31-2.44 (tt,
1H), 2.91 (s, 3H), 3.05-3.13 (m, 1H), 4.99-5.05 (m, 1H), 7.40-7.55
(m, 6H), 7.72-7.80 (m, 2H), 7.90-7.98 (m, 2H); 13C NMR (CDCl3)
δ 20.6, 21.8, 26.9, 29.8, 32.3, 38.2, 39.2, 40.1, 41.6, 78.3 (d, JCP ) 9.7
Hz), 128.5, 128.7, 128.8, 128.9, 130.7, 130.9, 131.0, 131.1, 131.7,
132.06, 132.11, 132.3; 31P NMR (CDCl3) δ 31.9.
r-4-tert-Butyl-trans-2-(diphenylthiophosphinoyl)cyclohexan-cis-
1-ol-1-d (1(P(O)Ph2)-OH-d): prepared from 4-tert-butylcyclohexene-
1
2-d oxide in the same fashion as above; H NMR (CDCl3) δ 0.68 (s,
9H), 1.25-1.40 (dq, 1H), 1.51-1.72 (m, 4H), 1.82-1.94 (m, 1H),
2.28-2.42 (dt, 1H), 3.05-3.15 (m, 1H), 7.37-7.48 (m, 6H), 7.81-
7.90 (m, 2H), 7.91-8.01 (m, 2H); 13C NMR (CDCl3) δ 20.2, 22.4,
27.0, 30.8, 32.5, 39.9, 40.7 (d, JCP ) 22.4 Hz), 65.5 (t, JCP ) 19.9 Hz),
128.25, 128.34, 128.4, 128.5, 128.7, 131.1, 131.2, 131.3, 131.4, 131.6,
132.1, 133.1; 31P NMR (CDCl3) δ 45.7.
trans-2-(Diphenylthiophosphinoyl)cyclohexan-1-ol (3(P(S)Ph2)-
OH). Lithium diphenylphosphide was generated and reacted with
cyclohexene oxide (2.13 g, 21.7 mmol) to give the corresponding
phosphine in the same fashion as in the synthesis of the trans oxide.
Elemental sulfur (2 equiv) was then added, and the mixture was stirred
for about 4 h at room temperature before the reaction was quenched
with water. Purification by column chromatography and crystallization
gave 2.54 g (37%) of a pale yellow solid: mp 176-178 °C; 1H NMR
(CDCl3) δ 1.08-1.80 (m, 7H), 1.97-2.08 (m, 1H), 2.65-2.97 (m, 2H),
4.03-4.17 (m, 1H), 7.37-7.48 (m, 6H), 7.75-7.84 (m, 2H), 7.94-
8.03 (m, 2H); 13C NMR (CDCl3) δ 24.4, 25.5, 25.7, 25.9, 34.9, 35.1,
44.7, 45.4, 70.7 (d, JCP ) 4.5 Hz), 128.3, 128.4, 128.5, 128.6, 131.2,
131.3, 131.8, 131.9; 31P NMR (CDCl3) δ 48.1. Anal. Calcd for C18H21-
PSO: C, 68.33; H, 6.69. Found: C, 68.41; H, 6.43.
cis-2-(Diphenylthiophosphinoyl)cyclohexan-1-ol (2(P(S)Ph2)-OH).
A dry 100 mL, round-bottomed flask was charged with cis-2-
(diphenylphosphinoyl)cyclohexan-1-ol (0.80 g, 2.67 mmol), pyridine
(1.20 mL, 14.8 mmol), trichlorosilane (0.70 mL, 6.94 mmol), anhydrous
benzene (50 mL), and a magnetic stirring bar. The mixture was heated
to reflux for 3 h, and sulfur (0.40 g, 12.5 mmol) was added to the
mixture. The reaction mixture was stirred overnight, the reaction
quenched with 2 N NaOH until the solution became clear, and the
mixture extracted with ether. Column chromatography over silica gel
with acetone-hexane (1/3) as eluent gave 0.15 g (18%) of a waxy
solid: 1H NMR (CDCl3) δ 1.22-1.68 (m, 4H), 1.76-1.93 (m, 3H),
2.09-2.28 (m, 1H), 2.56-2.67 (m, 1H), 4.10-4.16 (m, 1H), 4.42-
4.48 (s, 1H), 7.42-7.54 (m, 6H), 7.82-7.98 (m, 4H); 13C NMR (CDCl3)
δ 19.1, 19.7, 26.0, 33.8 (d, JCP ) 10.4 Hz), 41.3 (d, JCP ) 54.5 Hz),
65.2, 128.6, 128.7, 128.77, 128.84, 129.6, 130.3, 131.1, 131.2, 131.3,
131.4, 131.7; 31P NMR (CDCl3) δ 45.6. Anal. Calcd for C18H21PSO:
C, 68.33; H, 6.69. Found: C, 68.42; H, 6.72.
r-4-tert-Butyl-trans-2-(diphenylthiophosphinoyl)cyclohex-cis-yl p-
Nitrobenzoate (3(P(S)Ph2)-OPNB). A dry 100 mL, three-necked,
round-bottomed flask was charged with triphenylphosphine (5.91 g,
22.5 mmol), Li (0.28 g, 40 mmol), THF (25 mol), and a magnetic
stirring bar. The flask was fitted with a stopper, condenser, and rubber
septum. The reaction mixture was stirred for 1 h, and tert-butyl chloride
(1.89 g, 20 mmol) in THF (6 mL) was introduced slowly. The mixture
was stirred for 15 min and heated to reflux for 30 min. The reaction
mixture was cooled to 0 °C with an ice bath, and 4-tert-butylcyclo-
hexene oxide (1.54 g, 10 mmol) in THF (6 mL) was added through a
syringe. The reaction mixture was stirred overnight and the reaction
quenched with 5% acetic acid (10 mL). Sulfur (1.22 g, 38 mmol) was
added, and the solution was stirred for 3 h more. The reaction mixture
was extracted with ether. The ether layer was separated, dried, and
concentrated. The residue was combined with p-nitrobenzoyl chloride
(2.76 g, 15.1 mmol) and pyridine (5 mL) and stirred for 1 h. The
solid was filtered off and crystallized from CHCl3 and CH3OH. The
first batch of crystals (0.63 g, 12%) was the pure desired compound:
1H NMR (CDCl3) δ 0.76 (s, 9H), 1.38-1.66 (m, 2H), 1.68-1.80 (m,
2H), 1.86-1.95 (m, 1H), 2.24-2.36 (m, 1H), 2.58-2.74 (m, 1H), 3.34-
3.44 (m, 1H), 5.43-5.49 (m, 1H), 7.40-7.53 (m, 6H), 7.90-8.00 (m,
2H), 8.05-8.17 (m, 4H), 8.22-8.27 (d, 2H); 13C NMR (CDCl3) δ 21.0,
cis-2-(Diphenylphosphinoyl)cyclohexanol (2(P(O)Ph2)-OH).
A
100 mL, round-bottomed flask charged with a magnetic stirring bar
and 2.35 g (7.8 mmol) of the trans alcohol 3(P(O)Ph2)-OH was pumped
under vacuum for several hours to remove any moisture. The flask
then was fitted with a rubber septum and an inlet needle for N2.
Anhydrous THF (50 mL) was added to the flask with a syringe, and
the solution was cooled to -78 °C. On dropwise addition of BuLi
(2.5 M, 10 mL) through a syringe, the solution turned red instanta-
neously. The solution was stirred at dry ice temperature for another
30 min before the reaction was quenched with saturated aqueous NH4-
Cl solution. The solution was extracted with CHCl3 (3 × 25 mL).
The combined organic portions were dried (MgSO4) and concentrated.
The residue was chromatographed on silica gel with acetone and
petroleum ether as eluents to yield the cis isomer (0.50 g, 21%; the
trans isomer was not isolated): 1H NMR (CDCl3) δ 1.20-1.48 (m,
4H), 1.72-1.94 (m, 3H), 1.96-2.24 (m, 2H), 4.22 (m, 1H), 4.65 (m,
1H), 7.46-7.54 (m, 6H), 7.68-7.86 (m, 4H); 13C NMR (CDCl3) δ
19.18, 19.28, 27.6 (d, JCP ) 12.9 Hz), 32.9 (d, JCP ) 10.5 Hz), 40.4
(d, JCP ) 70.3 Hz), 64.9 (d, JCP ) 7.3 Hz), 128.6, 128.7, 128.8, 128.9,
130.6, 130.7, 130.8, 131.0, 131.9; 31P NMR (CDCl3) δ 39.4; HRMS
(M+) calcd 300.1279, obsd 300.1276. Anal. Calcd for C18H21PO2:
C, 71.99; H, 7.05. Found: C, 71.81; H, 6.96.
r-4-tert-Butyl-trans-2-(diphenylphosphinoyl)cyclohexan-cis-1-ol
(1(P(O)Ph2)-OH). A 100 mL, two-necked, round-bottomed flask, fitted
with a condenser and a rubber septum, was charged with Li (0.28 g,
40 mmol), PPh3 (5.31 g, 20 mmol), dry THF (40 mL), and a magnetic
stirring bar. Lithium diphenylphosphide was generated as above for
3(P(O)Ph2)-OH and allowed to react with 4-tert-butylcyclohexene oxide
(1.54 g, 10 mmol, mixture of trans and cis isomers). The same
oxidation and workup procedures were used as for 3(P(O)Ph2)-OH.
The crude product was crystallized from acetone-hexane to yield a
white powder (0.93 g, 26%), mp > 240 °C. The 1H, 31P, and 13C NMR
spectra all indicated a single product; 1H-13C COSY, DEPT, and
DQCOSY showed the product to be the desired compound: 1H NMR
(CDCl3) δ 0.69 (s, 9H), 1.24-1.39 (dq, 1H), 1.52-1.63 (m, 2H), 1.68-
1.78 (m, 3H), 2.22-2.35 (tt, 1H), 2.67-2.75 (m, 1H), 4.16-4.24 (m,
1H), 7.38-7.52 (m, 6H), 7.72-7.88 (m, 4H); 13C NMR (CDCl3) δ
20.4, 21.7, 27.0, 31.6, 32.5, 40.2, 41.1, 42.4, 65.6 (d, JCP ) 5.4 Hz),
128.5, 128.6, 128.7, 128.9, 130.8, 130.9, 131.0, 131.1, 131.46, 131.51,
131.55, 131.66; 31P NMR (CDCl3) δ 33.6. Anal. Calcd for C22H29-
PO2: C, 74.13; H, 8.20. Found: C, 73.59; H, 7.97.
trans-2-(Diphenylphosphinoyl)cyclohexyl Mesylate (3(P(O)Ph2)-
OMs). Mesylates were prepared according to standard procedure5 in
70-95% yields: 1H NMR (CDCl3) δ 1.14-1.28 (m, 1H), 1.48-1.61
(m, 2H), 1.64-1.83 (m, 4H), 2.45-2.56 (m, 1H), 2.59 (s, 3H), 2.66-
2.76 (m, 1H), 4.92-5.03 (m, 1H), 7.48-7.56 (m, 6H), 7.58-7.68 (m,
2H), 7.82-7.92 (m, 2H); 13C NMR (CDCl3) δ 23.3, 24.4, 24.6, 24.8,
33.7, 33.8, 38.1, 39.7, 40.6, 80.6, 128.6, 128.67, 128.73, 128.8, 130.7,
130.8, 130.9, 131.6, 131.7, 131.9, 134.0; 31P NMR (CDCl3) δ 33.3.
cis-2-(Diphenylphosphinoyl)cyclohexyl Mesylate (2(P(O)Ph2)-
1
OMs): H NMR (CDCl3) δ 1.23-1.38 (m, 1H), 1.48-2.04 (m, 6H),
2.34-2.53 (m, 2H), 2.93 (s, 3H), 5.22-5.28 (m, 1H), 7.42-7.56 (m,
6H), 7.73-7.82 (m, 4H); 13C NMR (CDCl3) δ 19.8, 21.4, 25.3, 32.3,

3166 J. Am. Chem. Soc., Vol. 118, No. 13, 1996
Lambert and Zhao
23.0, 27.0, 28.0, 32.6, 37.8, 38.5, 40.3, 71.6 (d, JCP ) 8.6 Hz), 123.5,
128.45, 128.55, 128.6, 128.7, 130.6, 131.2, 131.4, 131.5, 131.6, 132.3,
135.9, 150.4, 163.4; 31P NMR (CDCl3) δ 44.8. Anal. Calcd for
C29H32PSNO4: C, 66.78; H, 6.18; N, 2.69. Found: C, 65.37; H, 6.12;
N, 2.63.
Table 9. Structural Parameters for
2-(Diphenylthiophosphinoyl)cyclohexyl p-Nitrobenzoate
biased transa cisb
Bond Distances (Å)
O1-C6
1.467(4)
O1-C18
C18-C13
C13-P
P-S
1.475(3)
1.528(4)
1.841(3)
1.965(1)
r-4-tert-Butyl-trans-2-(diphenylthiophosphinoyl)cyclohexan-
1-ol (1(P(S)Ph2)-OH). A 50 mL, round-bottomed flask was charged
with the benzoate (0.63 g, 1.2 mmol), K2CO3 (0.70 g, 5.1 mmol), H2O
(1 mL), MeOH (15 mL), and a magnetic stirring bar. The mixture
was heated to reflux for 3.5 h. It was then diluted with H2O and
extracted with CHCl3. The organic layer was separated, dried (MgSO4),
and concentrated to give 0.43 g (96%) of a white solid: 1H NMR
(CDCl3) δ 0.68 (s, 9H), 1.25-1.40 (dq, 1H), 1.51-1.72 (m, 4H), 1.82-
1.94 (m, 1H), 2.31-2.45 (tt, 1H), 3.05-3.15 (m, 1H), 4.23-4.31 (m,
C6-C1
C1-P
P-S
1.530(5)
1.834(4)
1.952(2)
Bond Angles (deg)
C5-C6-O1
O1-C6-C1
C6-C1-P
108.8(3) C17-C18-O1
104.3(3) O1-C18-C13
111.7(3) C18-C13-P
114.1(3) C14-C13-P
102.7(2) C13-P-C1
106.1(2) C13-P-C7
116.5(1) C13-P-S
114.4(3) C17-C18-C13
112.0(3) C18-C13-C14
88.1(4) C17-C18-C13-P
153.2(2) O1-C18-C13-P
41.4(4) C17-C18-C13-C14
77.3(4) O1-C18-C13-C14
110.0(2)
104.8(2)
111.8(2)
117.2(2)
110.8(1)
106.6(1)
110.67(10)
111.8(2)
111.0(2)
171.0(2)
69.8(2)
C2-C1-P
C1-P-C18
C1-P-C24
C1-P-S
C5-C6-C1
C6-C1-C2
C5-C6-C1-P
O1-C6-C1-P
C5-C6-C1-C2
O1-C6-C1-C2
1H), 7.37-7.48 (m, 6H), 7.81-7.90 (m, 2H), 7.91-8.01 (m, 2H); 13
C
NMR (CDCl3) δ 20.3, 22.5, 27.1, 31.1, 32.6, 40.1, 40.9 (d, JCP ) 14.1
Hz), 66.2 (d, JCP ) 5.9 Hz), 128.3, 128.5, 128.6, 128.8, 131.1, 131.3,
131.4, 131.6, 131.7, 132.1, 133.1; 31P NMR (CDCl3) δ 45.8.
cis-2-(Diphenylthiophosphinoyl)cyclohexyl Mesylate (2(P(S)Ph2)-
1
OMs): H NMR (CDCl3) δ 1.24-2.08 (m, 7H), 2.38-2.52 (m, 1H),
56.1(3)
63.0(3)
2.56-2.68 (tt, 1H), 2.83 (s, 3H), 5.41-5.47 (m, 1H), 7.42-7.54 (m,
6H), 7.71-7.80 (m, 2H), 7.89-7.97 (m, 2H); 13C NMR (CDCl3) δ
19.4, 22.5, 25.9 (d, JCP ) 12.0 Hz), 32.8 (d, JCP ) 9.7 Hz), 38.9, 45.0
(d, JCP ) 53.3 Hz), 76.4 (d, JCP ) 4.7 Hz), 128.3, 128.5, 128.8, 129.0,
129.7, 131.1, 131.3, 131.8, 132.2, 132.4, 132.5; 31P NMR (CDCl3) δ
45.7.
a 1(P(S)Ph2)-OPNB; see Figure 6. b 2(P(S)Ph2)-OPNB; see Figure
7.
diffractometer with graphite monochromated Mo KR radiation. Cell
constants and an orientation matrix for data collection were obtained
from a least-squares refinement using the setting angles of 25 carefully
centered reflections in the range 20.0° < 2θ < 23.8°. The compound
has a formula weight of 521.61, belonging to the primitive monoclinic
space group P21/n (No. 14) with a ) 9.598(5) Å, b ) 20.372(4) Å, c
) 13.890(3) Å, â ) 93.76(3)°, Z ) 4, V ) 2710(1) Å3, and a calculated
density of 1.278 g cm-3. Diffraction intensities were collected at -120
( 1 °C using the ω - θ scan technique to a maximum 2θ of 45.9°. Of
trans-2-(Diphenylthiophosphinoyl)cyclohexyl Mesylate (3(P(S)-
1
Ph2)-OMs): H NMR (CDCl3) δ 1.08-1.84 (m, 7H), 2.34 (s, 3H),
2.48-2.60 (m, 1H), 2.94-3.05 (m, 1H), 5.04-5.15 (m, 1H), 7.36-
7.53 (m, 6H), 7.73-7.82 (m, 2H), 8.01-8.12 (m, 2H); 13C NMR
(CDCl13) δ 24.0, 24.8, 25.0, 25.1, 34.6 (d, JCP ) 8.6 Hz), 38.2, 41.2
(d, JCP ) 55.4 Hz), 81.6 (d, JCP ) 4.5 Hz), 128.3, 128.4, 128.5, 128.7,
131.1, 131.2, 131.3, 131.4, 131.6, 132.5, 133.5; 31P NMR (CDCl3) δ
48.5.
the 5542 reflections that were collected, 3908 were unique (Rint
)
.
r-4-tert-Butyl-trans-2-(diphenylthiophosphinoyl)cyclohex-cis-yl Me-
0.098). The linear absorption coefficient for Mo KR was 2.1 cm-1
1
sylate (1(P(S)Ph2)-OMs): H NMR (CDCl3) δ 0.72 (s, 9H), 1.28-
An empirical absorption correction was applied that resulted in
transmission factors of 0.95-0.98. Lorentz and polarization corrections
were applied. A correction for secondary extinction was applied
(3.1756 × 10-7). The structure was solved by direct methods and
expanded with Fourier techniques.25 Non-hydrogen atoms were refined
anisotropically. Hydrogen atoms were included in fixed positions but
not refined. The final cycle of full-matrix least-squares refinement was
based on 2113 observed reflections (I > 3.00σ(I)) and 326 variable
parameters and converged with unweighted and weighted agreement
factors of R ) 0.040 and Rw ) 0.036. The maximum and minimum
peaks on the final difference Fourier map corresponded to 0.19 and
-0.18 e/Å3, respectively. All calculations were performed using the
teXsan (Version 5.0) crystallographic software package of Molecular
Structure Corp.26 Additional crystallographic details are included in
the Supporting Information. Important structural parameters are given
in Table 9 with reference to Figure 6.
Crystal Structure of cis-2-(Diphenylthiophosphinoyl)cyclohexyl
p-Nitrobenzoate (2(P(S)Ph2)-OPNB). A colorless crystal of C25H24-
SPNO4 was obtained from the slow evaporation at room temperature
of a solution of CH3OH and CHCl3. A colorless, columnar crystal
with approximate dimensions of 0.46 × 0.17 × 0.15 mm was mounted
on a glass fiber using oil (Paratone-N, Exxon). All measurements were
made on an Enraf-Nonius CAD4 diffractometer with graphite mono-
chromated Mo KR radiation. Cell constants and an orientation matrix
for data collection were obtained from a least-squares refinement using
the setting angles of 25 carefully centered reflections in the range 21.9°
< 2θ < 23.3°. The compound has a formula weight of 465.50,
belonging to the primitive monoclinic space group P21/n (No. 14) with
a ) 10.081(1) Å, b ) 12.830(2) Å, c ) 17.657(6) Å, â ) 94.31(2)°,
1.44 (dq, 1H), 1.60-1.76 (m, 3H), 1.92-2.02 (m, 1H), 2.14-2.25 (m,
1H), 2.61-2.74 (tt, 1H), 2.87 (s, 3H), 3.46-3.56 (m, 1H), 5.03-5.08
(m, 1H), 7.44-7.55 (m, 6H), 7.89-8.02 (m, 2H), 8.08-8.16 (m, 2H);
13C NMR (CDCl3) δ 20.4, 22.5, 26.9, 29.3, 32.4, 38.2, 39.0, 39.6, 40.1,
78.5 (d, JCP ) 10.7 Hz), 128.5, 128.6, 128.8, 130.6, 131.3, 131.5, 131.6,
131.8; 31P NMR (CDCl3) δ 44.3.
r-4-tert-Butyl-trans-2-(diphenylthiophosphinoyl)cyclohex-cis-1-yl-
1-d Mesylate (1(P(S)Ph2)-OMs-d): 1H NMR (CDCl3) δ 0.71 (s, 9H),
1.27-1.42 (dq, 1H), 1.58-1.74 (m, 3H), 1.90-2.00 (m, 1H), 2.12-
2.23 (m, 1H), 2.58-2.72 (dt, 1H), 2.85 (s, 3H), 3.44-3.54 (m, 1H),
7.40-7.54 (m, 6H), 7.88-7.96 (m, 2H), 8.05-8.14 (m, 2H); 13C NMR
(CDCl3) δ 20.3, 22.5, 26.9, 29.2, 32.4, 38.3, 38.9, 39.6, 40.1, 128.5,
128.6, 128.8, 130.6, 131.3, 131.5, 131.6, 131.7, 131.8; 31P NMR
(CDCl3) δ 44.2.
Kinetic Method. Rates in aqueous solvents were determined
conductometrically with a YSI Model 32 conductance meter. The
conductance cell had a 40 mL capacity and Pt electrodes. The
temperature was kept constant with a Precision H8 heating bath.
Usually the concentration of the substrates was ca. 10-4 M. All linear
coefficients were 1.000 with one exception of 0.999 obtained for
solvolysis of the cis oxide in 97% TFE at 80.0 °C.
Product Studies. A 0.2-0.5 M solution of the substrate in 80%
ethanol or 97% trifluoroethanol (0.7 mL) was prepared in a NMR tube.
The tube was sealed at room temperature and put into a constant
temperature bath. After at least 10 half-lives, the sample was cooled
and the tube was opened. Water was added to the sample, and the
solution was extracted with ether. The organic solution was dried
(MgSO4) and concentrated under vacuum. The structures were assigned
by 1H and 31P NMR. Quantifications were based on the 31P peak
heights.
(25) Sheldrick, G. M. SHELX86. In Crystallographic Computing 3;
Sheldrick, G. M., Kruger, C., Goddard, R., Eds.; Oxford University Press:
Oxford, 1985; pp 175-189. Beurskens, P. T.; Admiraal, G.; Beurskens,
G.; Bosman, W. P.; Garcia-Granda, S.; Gould, R. O.; Smits, J. M. M.;
Smykalla, C. The DIRDIF Program System. Technical Report of the
Crystallography Library, University of Nijmegen, The Netherlands, 1992.
(26) teXsan: Crystal Structural Analysis Package. Molecular Structure
Corp., 3200A Research Forest Dr., The Woodlands, TX 77381. 1985 and
1992.
Crystal Structure of r-4-tert-Butyl-trans-2-(diphenylthiophosphin-
oyl)cyclohex-cis-1-yl p-Nitrobenzoate (1(P(O)Ph2)-OPNB). A color-
less crystal of C29H32PSNO4 was obtained from the slow evaporation
at room temperature of a solution of CH3OH and CHCl3. A colorless,
transparent, columnar crystal with approximate dimensions of 0.51 ×
0.20 × 0.08 mm was mounted on a glass fiber using oil (Paratone-N,
Exxon). All measurements were made on an Enraf-Nonius CAD4

â Effect of Phosphorus Functionalities
J. Am. Chem. Soc., Vol. 118, No. 13, 1996 3167
Figure 6. Crystal structure of r-4-tert-butyl-trans-2-(diphenylthio-
phosphinoyl)cyclohex-cis-1-yl p-nitrobenzoate (1(P(S)Ph2)-OPNB) with
the atomic numbering system.
Figure 7. Crystal structure of cis-2-(diphenylthiophosphinoyl)cyclo-
hexyl p-nitrobenzoate (2(P(S)Ph2)-OPNB) with the atomic numbering
system.
with no constraints in C1 symmetry. For all carbocations, three
geometric forms were considered as follows: parallel or bisected (b),
perpendicular or eclipsed (e), and bridged or cyclic (c). Since the
parallel forms usually are lower in energy, a constraint was made in
the calculations of the perpendicular forms, in which the P-C-C-C
dihedral angle was fixed at 180°. In the case of the â-phosphonate
carbocation, the parallel form was not an energy minimum and closed
to give the cyclic form. The parallel structure was optimized with the
P-C-C bond angle fixed at 105.1°, which was the optimized bond
angle in the â-phosphine oxide carbocation. For the phosphine and
phosphine oxide carbocations, in order to test the effect of the rotation
of the phosphorus group on â stabilization, the energy was calculated
as the P-C bond was rotated every 15° with all other geometric
parameters kept at the optimized values (Figures 4 and 5).
Z ) 4, V ) 2277.3(7) Å3, and a calculated density of 1.36 g cm-3
.
Diffraction intensities were collected at -120 ( 1 °C using the ω -
θ scan technique to a maximum 2θ of 49.9°. Of the 4329 reflections
that were collected, 4198 were unique (Rint ) 0.046). The linear
absorption coefficient for Mo KR was 2.4 cm-1
. An analytical
absorption correction was applied that resulted in transmission factors
of 0.96-0.97. Lorentz and polarization corrections were applied. A
correction for secondary extinction was applied (2.051 92 × 10-7). The
structure was solved by direct methods and expanded with Fourier
techniques.25 Non-hydrogen atoms were refined anisotropically. Hy-
drogen atoms were located in the Fourier map but included in fixed
positions and not refined. The final cycle of full-matrix least-squares
refinement was based on 2548 observed reflections (I > 3.00σ(I)) and
290 variable parameters and converged with unweighted and weighted
agreement factors of R ) 0.038 and Rw ) 0.033. The maximum and
minimum peaks on the final difference Fourier map corresponded to
0.32 and -0.26 e/Å3, respectively. All calculations were performed
using the teXsan (Version 5.0) crystallographic software package of
Molecular Structure Corp.26 Additional crystallographic details are
included in the Supporting Information. Important structural parameters
are given in Table 9 with reference to Figure 7.
Acknowledgment. This work was supported by the National
Science Foundation (Grant No. CHE-9302747). We thank
Charlotte L. Stern for carrying out the X-ray crystal structure
determinations.
Supporting Information Available: Listing of the crystal
structure of r-4-tert-butyl-trans-2-(diphenylthiophosphinoyl)-
cyclohex-cis-1-yl p-nitrobenzoate (1(P(O)Ph2)-OPNB) and cis-
2-(diphenylthiophosphinoyl)cyclohexyl p-nitrobenzoate (2(P(S)-
Ph2)-OPNB) (44 pages). This material is contained in many
libraries on microfiche, immediately follows this article in the
microfilm version of the journal, can be ordered from the ACS,
and can be downloaded from the Internet; see any current
masthead page for ordering information and Internet access
instructions.
Computational Methods. The ab initio calculations were performed
with GAUSSIAN-9227 on IBM RS/6000 workstations. Full geometry
optimizations were carried out by gradient methods for all neutral
molecules and ions with the 6-31G(d) basis set,28 which includes a set
of d orbitals on all non-hydrogen atoms. The resulting geometries and
basis sets were used in calculations that incorporated second-order
Møller-Plesset perturbation theory to include the effects of electron
correlation.29 In most cases the geometry optimizations were performed
(27) Frisch, M. J.; Trucks, G. W.; Head-Gordon, M.; Gill, P. M. W.;
Wong, M. W.; Foresman, J. B.; Johnson, B. G.; Schlegel, H. B.; Robb, M.
A.; Replogle, E. S.; Gomperts, R.; Andres, J. L.; Raghavachari, K.; Binkley,
J. S.; Gonzalez, C.; Martin, R. L.; Fox, D. J.; Defrees, D. J.; Baker, J.;
Stewart, J. J. P.; Pople, J. A. Gaussian 92, Revision C; Gaussian Inc.:
Pittsburgh, PA, 1992.
JA9537181
(28) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213-
222. Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J. S.; Gordon, M.
S.; DeFrees, D. J.; Pople, J. A. J. Chem. Phys. 1982, 77, 3654-3665.
(29) Møller, C.; Plesset, M. S. Phys. ReV. 1934, 46, 618-622. Hehre,
W. J.; Radom, L.; Schleyer, P. v. R.; Pople, J. A. Ab Initio Molecular Orbital
Theory; Wiley: New York, 1986.