J. Am. Chem. SOC.

1984, 106, 7825-7831

Stereoelectronic Effects on the Basicity and
Nucleophilicity of Phosphites and Phosphates. Ab Initio
7825
Molecular Orbital Calculations and the a-Effect
Kazunari Taira and David G. Gorenstein*
Contribution from the Department of Chemistry, University of Illinois at Chicago, Chicago,
Illinois 60680. Received J anuary 9, 1984. Revised Manuscript Received J une 8, 1984
Abstract: Ab initio molecular orbital calculations on phosphites, protonated phosphites, and protonated phosphates reveal
important stereoelectronic effects. In the phosphites, an antiperiplanar lone pair on oxygen to the phosphite lone pair raises
the energy of the molecule by 3.3 kcal/mol relative to a phosphite conformation with no antiperiplanar lone pairs to the phosphite
lone pair. Upon phosphorus protonation of the phosphite the relative energy difference between the conformations reverses.
The conformation with lone pairs on oxygen antiperiplanar to the P-H bond is now more stable than the conformation without
this antiperiplanar lone-pair interaction. Finally, the origin of the a-effect, the enhanced nucleophilicity of a base possessing
a heteroatom with an adjacent unshared electron pair, is suggested to arise fromthe stereoelectronic effect. This is attributed
to a transition-state stereoelectronic effect. Whereas oxygen lone pairs antiperiplanar to the P-H+in the ground-state protonated
phosphite stabilize the structure by only 1 kcal/mol at a P-H distance of 1.4 A, this stereoelectronic stabilization rises to >12
kcal/mol, at a P-H distance of -3 A, falls off again at even longer P-H bond distances, and finally reverses energies for
the two confirmations of the neutral phosphites.
The role of orbital orientation in organic and enzymatic re-
actions has been of considerable current interest.l-l0 Des-
longchamps and co-workers2 in studying tetracovalent carbon
species have demonstrated selective cleavage of bonds which are
trans-antiperiplanar (app) to lone pairs on directly bonded oxygen
and nitrogen atoms. Molecular orbital calculations have provided
theoretical justification for these stereoelectronic effects in tet-
racovalent carbon and phosphorus species and pentacovalent
phosphoranes.5-11 Thus, as has been shown in molecular orbital
calculations on the XI-Y-X2 (X =0, N; Y =P, C) structural
fragments, the XI-Y bond is strengthened (as indicated by an
increase in the Mulliken overlap population) while the Y-X2 bond
is weakened when the XI atom lone pair is app to the Y-X2
(Structure A, Figure 1) bond. In the gauche, trans (g, t ) con-
formation of dimethoxymethane (Structure A, Figure 1, X =
OCH,, Y =CH,) the overlap population for the trans C-0 (Y
=C, X, =0) bond is 0.022 electron lower than the overlap
population for the gauche C-0 bond. In the g,t-dimethoxy-
methane one lone pair (shaded in A) on the gauche bond oxygen
is app to the trans C-0 bond, while no lone pairs on the trans
bond oxygen are app to the gauche bond. Thus, the Y-X2 bond
is weaker than the XI-Y bond because it has one app lone pair
on X, and no lone pairs on X2 app to the XI-Y bond.8 Lehn5s6
and Pople and co-workers have shown similar overlap population
differences in related systems.
Molecular orbital calculations on phosphate esters (Figure 1,
XI , X2 =OCH,, Y =PO2-) and tetrahedral carbon intermediates
(1) Kirby, A. J . The Anomeric Effect and Related Stereoelectronic Ef-
fects at Oxygen; Springer-Verlag: Berlin, 1983; pp 1-149.
(2) Deslongchamps, P.; Taillerfer, R. J . Can. J. Chem. 1975.53, 3029 and
references cited therein. Deslongchamps, P. Stereoelectronic Effects in
Organic Chemistry; Pergamon Press: Oxford, 1983.
(3) Storm, D. R.; Koshland, D. E., J r. J. Am. Chem. Soc. 1972,94,5815.
(4) Mock, W. L. Bioorg. Chem. 1975, 4, 270.
(5) Lehn, J . M.; Wipff, G. J. Chem. SOC., Chem. Commun. 1975, 800.
(6) (a) Lehn, J . M.; Wipff, G. J. Am. Chem. Soc. 1974, 96,4048. (b) Ibid.
1976, 98, 7498. (c) Helu. Chim. Acta 1978, 61, 1274.
(7) (a) Radom, L.; Hehre, W. L.; Pople, J . A. J. Am. Chem. SOC. 1972,
94, 2371. (b) J effrey, G. A.; Pople, J . A,; Radom, C. Carbohydr. Res. 1972,
25, 117.
(8) (a) Gorenstein, D. G.; Findlay, J . B.; Luxon, B. A.; Kar, D. J . Am.
Chem. SOC. 1977, 99, 3473. (b) Gorenstein, D. G.; Luxon, B. A.; Findlay,
J . B.; Momii, R. Ibid., 1977, 99, 4170. (c) Gorenstein, D. G.; Luxon, B. A,;
Findlay, J . B. Ibid. 1977, 8048.
(9) Gorenstein, D. G.; Luxon, B. A,; Goldfield, E. M. J . Am. Chem. SOC.
1980, 102, 1757. Gorenstein, D. G.; Rowell, R.; Taira, K. In Phosphorus
Chemistry; American Chemical Society: Washington, DC, 1981; ACS
Symp. Ser. No. 171, p 69.
(10) Gorenstein, D. G.; Taira, K. Biophys. J. , in press.
(11) Gorenstein, D. G.; Luxon, B. A.; Findlay, J . B. J. Am. Chem. SOC.
1979, 101, 5869.
(Figure 1, X1 =NH,, X2 =OH, Y =CHOH) have also provided
confirmation for the stereoelectronic effect.6q8-10 Ab initio mo-
lecular orbital calculations on the reaction profile for the
base-catalyzed hydrolysis of dimethyl phosphate in various ester
conformations have provided support for this stereoelectronic
theory.
(CH,O),POF +-OH - (CH30)2P03HZ- -
(CH,O)PO,H- +CH30- ( I )
Separate transition states were observed for the first addition step
and the second elimination step. A metastable pentacovalent
intermediate [(CH30)2P03H2-] was established along the reaction
coordinate. Significantly, for the methoxide elimination step the
transition state which has an antiperiplanar lone pair to the
methoxide leaving group is ca. 11 kcal/mol lower in energy than
the transition state without this app lone pair.
Phosphite as well as phosphate esters undergo marked changes
in properties (basicity, nucleophilicity, rate of hydrolysis, spectra)
upon inclusion of the phosphorus atom into a monocyclic or a
bridgehead bicyclic systern.l2-l6 We demonstrate both theo-
bicyclic
phosphite
retically and experimentally in this and the accompanying paper
that stereoelectronic effects play an important role in these
properties.
Finally, the origin of the a-effect, the enhanced nucleophilicity
of a base possessing a heteroatom with an adjacent unshared
electron pair, has been much debated. We emphasize that the
a-effect may also be considered a stereoelectronic effect.
~~~~~~ ~~ ~~
(12) Hudson, R. F.; Verkade, J . G. Tetrahedron Lett. 1975, 37, 3231.
(13) Verkade, J . G. Eioinorg. Chem. 1974, 3, 165. Verkade, J . G. Phos-
phorus Sulfur 1976, 2, 251.
(14) (a) Hodges, R. V.; Houle, F. A,; Beauchamp, J . L.; Montag, R. A.;
Verkade, J . G. J. Am. Chem. SOC. 1980,102,932. (b) Vande Griend, L. J .;
Verkade, J . G. ; Pennings, J . F. M; Buck, H. M. J . Am. Chem. SOC. 1977, 99,
2459. (c) Cowley, A. H.; Lattman, M.; Montag, R. A.; Verkade, J . R. Inorg.
Chim. Acta 1977, 25, 151.
(15) Westheimer, F. H. Acc. Chem. Res. 1968, 1 , 70.
(16) (a) Zverev, V. V.; Villem, J .; Arshinova, R. P. Dokl. Akad. Nauk
SSSR 1981, 256, 1412. (b) Cowley, A. H.; Goodman, D. W.; Kuebler, N.
A,; Sanchez, M.; Verkade, J . G. Inorg. Chem. 1977, 16,854. (c) Yarbrough,
L. W., 11; Hall, M. B. J. Chem. SOC. Chem. Commun. 1978, 161.
0002-7863/84/ 1506-7825$01.50/0 0 1984 American Chemical Society
7826 J. Am. Chem. SOC., Vol. 106, No. 25, 1984 Taira and Gorenstein
Figure 3. Structures for different conformations of trimethyl phosphites.
This two-electron interaction of these two orbitals is stabilizing.
In contrast, it is destabilizing to have an oxygen lone pair orbital
app to the phosphorus lone pair orbital. This is a four-electron,
lone-pair-lone-pair interaction which is overall destabilizing.'*
Thus 4 is 3.3 kcal/mol lower energy than 5. The conformations
differ in rotation about only one of the three P-0 bonds. In 4
both lone pairs on 0- 3 are app to the adjacent polar P-0 bonds
while in higher energy 5, one of the lone pairs on 0- 3 is app to
the phosphorus lone pair. 5 has one more destabilizing four-
electron orbital interaction and one less two-electron stabilizing
orbital interaction. Similarly 2 is 3.0 kcal/mol higher energy than
3. Again 3 has two more n - u* favorable interactions and two
less no - np (0, oxygen; P, phosphorus) unfavorable interactions.
An exception to this analysis is seen in comparing 1 and 3 which
differ in conformation only about P-0,. Phosphite 1, which has
the same conformation as the bicyclic phosphite, is higher energy
than 3 in spite of the more favorable lone pair orientation on 0- 3
in 1 relative to 3. Actually, as has been shown earlier for phosphate
and tetrahedral carbon specie^,^,^a secondary "counterbalancing"
stereoelectronic effect probably is responsible for some of these
energy differences (particularly between 1 and 3). Thus, inter-
action of a lone pair app to an adjacent acceptor bond is favored
if the acceptor bond oxygen is not participating in any stereoe-
lectronic interactions with an adjacent polar bond. Note in 3,O-3
has one fewer oxygen lone pair app to adjacent polar bonds than
in 1. Thus the P-0-3 bond is a more effective acceptor bond.
Bond length differences (Table I) and population analyses and
atomic charges (Table 11) support this counterbalancing ster-
eoelectronic effect.
This difference could also be attributed to either an unfavorable
1,3-steric interaction between the hydrogens or unfavorable dipole
interaction in "bicyclic" 1. Note in Table I, (see Figure 3 for
structures) large 1,3-steric interactions between the methyl groups
of the trimethyl phosphite esters are presumably responsible for
the much higher energy of 6 (comparable conformation to 1)
relative to 7 (comparable conformation to 2).
Indeed in spite of this unfavorable steric or dipole interaction
in 1, it is still 1.4 kcal/mol lower energy than 2. 1 has three
additional favorable two-electron interactions relative to 2 and
three fewer unfavorable lone-pair-lone-pair interactions. Elim-
inating the unfavorable steric or dipolar interactions in 1, and
assuming a similar 3.3 kcal/mol stabilization per orbital interaction
as in the 4/5 pair, 1 should be 3.3-9.9 kcal/mol rather than just
1.4 kcal/mol more stable than 2. The actual stereoelectronic
stabilization of 1 relative to 2 is thus at least 1.4 kcal/mol and
possibly much greater. In fact, at the 6-21G level of calculation,
1 is 5. 5 kcaI/mol lower in energy than 2.
Stereoelectronic Effects in Protonated Phosphites. This relative
energy differences between conformations 1 and 2 reverses upon
protonation of the phosphite. As shown in Table I using both
STO-3G and 6-21G (Figure 4) basis sets, 9b (comparable con-
formation about the P-0 bonds to 2) is now 4.6 kcal/mol lower
energy than 8b (comparable conformation to 1) at a P-H+distance
of 1.8 This reversal is graphically depicted in Figure 5 . This
is exactly the stereoelectronically expected result since 9 has three
oxygen lone pairs app to the polar P-H+bond (analogous to a
polar P-0 bond). A similar energy difference was observed for
the protonated phosphates (10 and 11),20 where the protonated
"acyclic" phosphate 11 is 3.9 kcal/mol more stable than the
(19) The proton and phosphorus distance was arbitrarily chosen to be 1.8
A, and the geometries of the 'bicyclic" and protonated phosphates were not
optimized.
(20) All 0-H bond distances were assumed to be 1 .O A.
Figure 1. Conformations g,t (structure A) and g,g (structure B). Di-
hedral angles about the XI Y and X2Y bonds are defined by the -XI-Y-
X2- structural fragment and are gauche (g, dihedral angle *6Oo) or trans
( t , dihedral angle 180'). For XI =X2 =divalent oxygen, the sp'-hy-
bridized lone pairs are also shown, with the antiperiplanar lone pairs
being shaded.
n n
Figure 2. Structures for different conformations of phosphites.
Methods of Calculation
The SCF LCAO-MO ab initio calculations utilized the GAUSSIAN 70
and 80 series of programs generally with a STO-3G minimal basis set"
(although a 6-21G basis set also was used for single point calculations).
Except for some of the partially protonated phosphites, standard scale
factors were used where indicated using the geometry optimized at the
STO-3G level. I n the other structures, which would not converge with
the standard scale factors, the valence shell scale factors were optimized
and then sequentially in three groups: phosphorus 3sp, ester oxygen 2sp,
and proton 1s. The molecular geometries were optimized either by
sequentially varying a set of geometrical parameters until the total energy
had been effectively minimized (<O. 1 kcal/mol) or utilizing the geometry
optimization routine of GAUSSIAN 80 (both procedures yielded identical
geometries where tested). The final geometries for phosphites and pro-
tonated phosphite and phosphate esters are given in Table I.
Results and Discussion
Ground-State Stereoelectronic Effects in Phosphites. Shown
in Table I are the relative energies for various conformations of
phosphites (see Figure 2 for structures). The results are basically
consistent with the ground-state anomeric, stereoelectronic effect.','
Similar results are obtained with STO-3G or 6-21G basis sets.
Although other interactions are operativela and it is not our intent
to discuss them here, in general it is energetically more favorable
to have a conformation about the P-0 bond which allows an
oxygen lone pair to be trans, antiperiplanar (app) to an adjacent
polar bond (P-OR). This ground-state stereoelectronic effect is
presumed to arise from the favorable orbital interaction of the
oxygen lone pair orbital, n, with the antibonding P-0 u* orbital.'
All calculations were carried out on an I BM 4341 computer.
(17) Gaussian 70: Hehre, W. J .; Lathan, W. A.; Ditchfield, R.; Newton,
M. D.; Pople, J . A., Quantum Chemistry Program, No. 236; GAUSSIAN 80:
Binkley, J. S.; Whiteside, R. A.; Krishnan, R.; Seeger, R.; DeFrees, D. J .;
Schlegel, H. B.; Topiol, S.; Kahn, L. R.; Pople, J . A. QCPE 1981, 13, 437.
Corrections by A. B. Buda, E. Osawa and T. D. Bouman.
(18) Epiotis, N. D.; Cherry, W. R.; Shaik, S.; Yates, R. L.; Bernardi, F.
"Structural Theory of Organic Chemistry"; Springer-Verlag: New York,
1977.
MO Calculations on Phosphates and Phosphites J. Am. Chem. SOC., Vol. 106, No. 25, 1984 7821
i
d
i
ci
3
3
9
m
0
3
9
i
3
3
[:
vi
2
2
3
3
?
3
3
3
r:
W
2
09
vi
2
2
N
I
r:
N
i
3
-?
i
i
N
2
3
3
B
8
8
W
00
i
9
0
W
9
0 m
9
0
W
9
0
W
9
0
m e
9
0
00 3
9
- -
il
8
m
3
9
0
W
9
0 m
3
n
Lyl
U
- 0
099
&
0
O b
i - * b W w 00
h h O O *
m w
'4'9? ? 09Y Y
3 3
0 0 cam U
m m o o o o
3.-d
414444
7828 J. Am. Chem. SOC., Vol. 106, No. 25, 1984 Taira and Gorenstein
Table 11. Population Analyses and Atomic Charges for Phosphates, Protonated Phosphites, and Protonated Phosphates
Conformation
W ?
n ? ? ?
I ! n n I
w W
1 2 6 I 8a 8b 9a 9b 12 13
1 2 6 I Sa 8b 9a 9b 12 13
Population Analyses
P-0 0.3910 0.3756 0.3680 0.3786 0.4074 0.3970 0.4200 0.4124 0.3966 0.4182
0-H 0.4854 0.4932 0.5196 0.5236 0.5116 0.5080 0.5148 0.5146 0.5114 0.5122
H+-P 0.6762 0.5020 0.6026 0.4576 0.5058 0.4506
Atomic Charges
P +0.8652 +0.7924 +0.8956 +0.8357 +1.1784 +1.1219 +1.1800 +1.1306 +1.1457 +1.152
0 -0.4271 -0.4199 -0.4106 -0.3896 -0.3295 -0.3355 -0.3318 -0.3347 -0.3246 -0.317
H (C) +0.1387 +0.1558 -0.0756 -0.0764 +0.2470 +0.2400 +0.2632 +0.2608 -0.0700 -0.073
H+ +0.0693 +0.1647 +0.0256 +0.0913 +0.1269 +0.069
H
I
H
8a, b
-
H
10
-
11
- -
98, b
Figure 4. Structures for different conformations of protonated phosphites
and protonated phosphates.
H
, :o
.p\\
H
O P
0 - H
Figure 5. Graphical depiction of relative energies of phosphites, pro-
tonated phosphites, and solution phase transition states.
protonated bicyclic phosphate 10 presumably for the same
stereoelectronic argument.
This energy difference between the two protonated geometries
would have been greater if methyl groups could have been used
in place of hydrogen in both bicyclic and acyclic phosphites
(and also phosphates). The basis for stabilization of the acyclic
molecules is the stereoelectronic (anomeric-type) two-electron
stabilizing interaction of the vicinal nonbonding orbitals with the
antibonding orbital of the incipient P-H+bond. On the basis of
one-electron molecular orbital theory, the twc-electron stabilization
resulting from the interaction of a doubly occupied Mot{) with
a vacant nondegenerate MO(a*) may be approximated as being
Table 111. Calculated HOMO Ionization Potentials for Phosphites
structure IP, eV structure IP, eV
CH,
6
-
6.94
Calculated, this work. bCalculated, ref 16c, based upon full bi-
cyclic phosphite structure.
inversely proportional to the energy separation of the two MOs
and directly proportional to the square of their overlap
In the above example the doubly occupied MO(s) is a nonbonding
orbital ({=n) on oxygen, however, {can be o ~ ~ , u ~ - ~ ,
UC-C, etc, and it is expected that SE(n,u*) >SE(UO-H,U*), and
SE(n,a*) >S E ( U ~ - ~ , U * ) >SE(CT~_~,U*).~~ This implies that
electron donation from the u,-+~ and u ~ - ~ bonds into the u* bond
are better than the uGC and uc-c, respectively.
In spite of the fact that the ueH bonds in the bicyclic analogue
1 can effectively donate electrons to the incipient u* bond upon
protonation, the protonated bicyclic analogue 8b is still calculated
to be 4.6 kcal/mol higher energy than the protonated acyclic
analogue 9b. Therefore, the energy difference between protonated
bicyclic and acyclic geometries is expected to be greater if hy-
drogens are replaced by alkyl groups because there is much less
electron donation from the OC bonds in the bicyclic geometry.
Ionization Potentials. The stereoelectronic effect provides an
explanation for the difference in ionization potentials of the
phosphites. Assuming the validity of Koopmans theorem, each
ionization potential (IP,) is simply equal in magnitude to an orbital
energy (e,): IPj =-eJ . Calculated IPs for bicyclic (1 and 6 )
and acyclic (2 and 7) phosphites are listed in Table I11 and some
experimental values for a few phosphites in Table IV.I6 Earlier
ab initio MO calculations16b (see also ref 14c) with similar basis
set as ours (but including 3d functions on phosphorus) indicated
(21) Cieplak, A. S. J . Am. Chem. SOC. 1981, 103, 4540.
MO Calculations on Phosphates and Phosphites
Table IV. Experimental Vertical Ionization Potentials and Proton
Affinities of Phosphites
no. compd IP PA
14 P(OCH3)j 196 (8.50) 229.9
l5 F Y P \ O M e
192 (8.34) 227.5
0
0 Ma 200 (8.69) 224.9
I
16
209 (9.06) 215.6
P o 7 p :
17 OM8
-0
19
D
217 (9.42) 215.6
224 (9.72) 200.4
*O <?
From ref 14b in kcal/mol. Values in parentheses in eV.
that the phosphorus al orbital is the HOMO orbital and is assigned
to the lowest energy photoelectron band (lowest IP). Our own
calculations agree with this assignment, although with a slightly
poorer basis set our calculated IPs are slightly lower than Yar-
borough and Halls and the experimental values. Qualitatively,
however, there is good agreement between the relative energy
differences.
Basicity of Phosphites. Measurements of gas-phase proton
affi ni ti e~, ~ and solution-phase basicitiesI3 of phosphites have
revealed both bond angle and conformational effects. The gas-
phase proton affinity (PA), which is defined as the hetereolytic
bond dissociation energy for loss of a proton from the acid BH+
B H++B +H+ A H=P A
for various phosphites is listed in Table IV. As pointed out by
Hodges et al.,14a the proton affinity decreases as the phosphorus
lone pair ionization potential increases, a not uncommon rela-
ti 0n~hi p.I ~~ Indeed the proton affinity is thermodynamically
related to the base ionization potential, with
PA =D(B+H) +I P(H) - IP(B)
where D(B+H) is the homolytic bond dissociation energy and
I P(H) is the ionization potential of the hydrogen atom. As dis-
cussed previously phosphorus-lone-pairygen-lone-pair repulsion
in the phosphite will raise the energy of the phosphorus lone pair
orbital and hence decrease the I P and increase the PA.12-14 In
the protonated phosphite, oxygen lone pair overlap with the u*
orbital ( u * ~- ) will similarly affect the PA. Comparison of the
experimental values in Table IV confirms these predictions.
Clearly (0-P-0 and P-0-C) bond angle is one factor (compare
17 vs. 15 and 20 vs. 18) and is likely attributable to a Hinge
Effect offered by Verkadelz-I3 where hybridization changes with
bond angles. Similarly as shown by our ab initio calculations and
earlier CND0/2 calculation^,'^^oxygen-lone-pair-phosphorus-
lone-pair repulsion will vary with P-O torsional angles and hence
orbital orientation. Note in particular that 15 has a lone pair on
each of the ring oxygens which is app to the phosphorus lone pair.
In 16 none of the oxygen lone pairs is app to the phosphorus lone
pair, and significantly the phosphorus lone pair orbital is lower
in energy (higher I P Table IV). In contrast to the destabilizing
interaction of an oxygen app lone pair to tho phosphorus lone pair
in the neutral phosphite, in the protonated phosphite, the oxygen
lone pair will now be app to the P-H+bond and this stereoelec-
tronic effect will be stabilizing.
As supported by measured proton affinities (Table IV), the PA
for bicyclic phosphite 1 is calculated to be 2.4 kcal/mol higher
J. Am. Chem. Soc., Vol. 106, No. 25, 1984 1829
energy than that of the acyclic phosphite 2. While a portion of
the calculated PA difference is attributed to the I P difference of
the neutral phosphites (likely also a reflection of the stereoelec-
tronic effect), a portion is also due to the ground-state stereoe-
lectronic effect in the protonated phosphites. Note that the bicyclic
protonated phosphites 8a is calculated to be 1 .O kcal/mol higher
energy than the acyclic protonated phosphite 9a, whereas the
bicyclic phosphite 1 is 1.4 kcal/mol lower energy than the acyclic
phosphite 2.
Nucleophilicity and the a-Effect. As described in the accom-
panying paper, these results provide an explanation for the poor
nucleophilicity of the bicyclic phosphite: The poor nucleophilicity
ka
ZE!O-S-@
7
/
\ x *
<;;P:
/
<;;P: \ v
H A #
*
k C
Et OHp Et 0,320
OEt
0Et
of the bicyclic phosphite is likely due to the stereoelectronic effect.
Verkade, while acknowledging the importance of these orbital
interactions,I2 has also attributed these properties to a Hinge
Effect: ring constraint produces hybridization changes at the
alkoxy oxygens which result in alteration of the ?F donating ability
on oxygen lone pairs to the d-orbitals on phosph~rus. ~- ~ Our
own analysis emphasizes the stereochemistry of proper orbital
overlap between alkoxy oxygen lone pairs and the incipient u*
orbital of the newly forming bond to phosphorus, besides lone-
pair-lone-pair repulsion in the neutral phosphite. Protonation at
phosphorus will increase the 0-P-0 angle and the resulting in-
crease in ring strain in the bicyclic transition state is also possibly
partially responsible for the low nucleophilicity of the bicyclic
phosphite.22
Similar considerations can also rationalize the resistance to
alkylation of the bicyclic phosphate observed in this lab23 and
others.24 (Note protonated bicyclic phosphate 10 is 3.9 kcal/mol
higher energy than the protonated acyclic conformation of
phosphate 11.)
OCH3 -
H3CO
P- 0
H$O I OCH3
OCH3
Me
F S O ~ C H ~ SO^,,,,, N R. H\
/ lso2 -
< > p a 0 \
Not formed
The a-effect, the enhanced reactivity of nucleophiles pos-
sessing a pair of electrons a to the nucleophilic atom, can beviewed
as a stereoelectronic effect. Although numerous explanations have
(22) Ernsley, J.; Hall, D. The Chemistry of Phosphorus; Wiley: New
(23) In preparation.
(24) Finley, J . H.; Denney, D. Z. ; Denney, D. B. J. Am. Chem. SOC. 1969,
York, 1976.
91, 5826.
7830 J. Am. Chem. SOC. , Vol. 106, No. 25, 1984
been provided for the a-effect (see ref 25-27 and ref therein),
one explanation that comes closest to the stereoelectronic orbital
interaction picture is the one developed by Hudson2s and Klop-
man.29 Lone-pair-lone-pair orbital mixing (such as oxygen and
phosphorus lone pairs in phosphites) will raise the energy of the
HOMO and mixing with the lowest unoccupied molecular orbital
(LUMO) will be enhanced. The major emphasis in these MO
a-effect theories has been lone-pair-lone-pair i nt er a~t i on ~~~~~~~
rather than our stereoelectronic interpretation of transition-state
magnification of adjacent lone-pair-a* interaction^.^^^ Indeed,
as pointed out by Heaton30 these ground-state destabilization
arguments do not accord with the calculated and actual HOMO
energies. His ab initio MO calculations instead emphasize the
importance of the antibonding characteristics and higher polar-
izability of the a-effect HOMOS. All of these studies (including
Heatons) suffer from their emphasis on ground-state properties
of a-effect nucleophiles. As we have emphasized the stereoe-
lectronic effect (and likely the a-effect) is largely a transition-state
phen~menon.~ Significant mixing of n and u* is possible only
in the transition state. I t is this specific lowering of the transi-
tion-state energy by a-effect lone pair electrons that presumably
is responsible for a significant fraction of the enhanced reactivity
of these a-effect nucleophiles.
Our calculations strongly support the transition-state stereoe-
lectronic effect basis for the a-effect. Although the strongly
exoergic protonation of the phosphite does not proceed via a
transition state in these gas-phase calculations (Figure 6A), we
can still address the importance of the stereoelectronic effect on
a structure that would likely resemble the solution phase transition
state for electrophilic attack on a phosphite (Figure 5) . Since the
ionization potential of the neutral hydrogen atom is 13.6 ev,j3 and
the ionization potentials for the phosphites are 8.5-9.7 eV (cf.
Table IV) in order to prevent premature electron transfer from
the phosphite to the proton, the scale factor on the 1s orbital of
the proton was reduced at longer P-H distances. In this way we
achieved convergence to a ground state with charge still largely
on the proton at large (>4 A) separation of the proton from the
phosphites. In solution, of course, solvation of the proton would
ensure that charge transfer would not occur until the approach
of the proton was close enough. This distance depended upon the
choice of the hydrogen orbital scale factor and for a hydrogen
scale factor of 0.84, charge transfer and hence the transition state
occurred around 3-4 A. Again, in the gas-phase calculations no
true transition state is found in this region (3-4 A, Figure 6A).
Note, however, that the.stereoe1ectronic effect is maximized in
this region (Figure 6B) which we are defining as the transition
state. I t is likely significant that in other phosphorus reaction
surfaces which wehave calculated (ref 8, 9, and unpublished) true
transition states were obtained at P-X internuclear separation of
- 3 A.
It is most important that the energy difference between the
partially protonated bicyclic and acyclic phosphites reaches
a maximum at a P-H distance of 3.0 A ((Figure 6A,B), geometries
in Table I ) . Again, the lowest energy conformation has the
phosphite oxygen lone pairs app to the translating P-H bond. Thus
in the ground state of the protonated phosphite oxygen lone pairs
app to the P-H bond stabilizes the molecule by only 1 kcal/mol
(compare 9a vs. 8a). In contrast, conformation 9 (corresponding
to the acyclic structure) is at least 12 kcal/mol (with a STO-3G
Taira and Gorenstein
(25) DePuy, C. H.; Della, E. W.; Filley, J .; Grabowski, J . J .; Bierbaum,
(26) Grekov, A. P.; Veselov, V. Ya. Rum. Chem. Reu. (Engl . Transl.)
V. M. J. Am. Chem. SOC. 1983, 105, 2481.
1978. 47. 631
--. .. 7
-.
(27) Fina, N. J. ; Edwards, J . 0. Int . J. Chem. Kine?. 1973, 5, 1 .
(28) Filippini, F.; Hudson, R. F. J. Chem. SOC., Chem. Commun. 1972,
(29) Klopman, G.; Evans, R. C. Tetrahedron 1978, 34, 269.
(30) Heaton, M. M. J. Am. Chem. SOC. 1978, 100, 2004.
(31) We are aware of equilibrium a- ef f ect~~~ as well. However, these are
generally small (<loo, a-effects measured by rate constant ratio, k N H I N H 2 /
(35f Dixon, J. E.; Bruice, T. C. J. Am. Chem. SOC. 1971, 93, 3248.
( 33) Levine, I . N. Quantum Chemistry, 2nd ed; Allyn and Bacon:
522. Hudson, R. F.; Filippini, F. Ibid. 1972, 726.
kglycyl I cine),
Boston, 1974; p 106.
A I
c ,
-20 1
* ~ A C Y C
Figure 6. (A) Relative energies (STO-3G basis set) of protonated bi-
cyclic, 8 and acyclic, 9 phosphites as a function of P-H bond distance.
(0, 0) Open symbols, acyclic conformation 9, closed symbols, bicyclic
conformation 8, standard scale factors at dP-H 5 3 A. (A, A) Scale
factors optimized at 3.5 A. (0, U) Scale factors used at 6.0 A (see legend
to Figure 6B). (O, W) Scale factors adjusted for acyclic system at dP-H
=10.0 A, phosphite geometry. (O, 0 ) Standard scale factors, phosphite
geometry, no proton. (B) Energy difference between protonated phos-
phite conformations 8 and 9 as a function of P-H bond distance. (0)
dFH+ =1.8 A geometry. Standard scale factors: =1.90, 02sp =
2.25, HI , =1.24. (0) Phosphite geometry (dp-H+ =a), Standard scale
factors. (A) Phosphite geometry, scale factors optimized at dP-H+ =3. 5
A for bicyclic system: P3sp =1. 885, 02,p =2.229, HI , =0.842 (for the
translating proton). (m) Phosphite geometry, scale factors adjusted for
both systems at dP-H+ =6.0 A: acyclic PpSp =1.853, 02sp =2.226, HI ,
=0.44; bicyclic P3sp =1.866, Obp =2.219, HI , =0.55. (m) Phosphite
geometry, scale factors adjusted for acyclic system at dp- ~+ =10.0 A.
(0) Phosphite geometry, standard scale factors. Dashed line is drawn
only for visualization. ( C) HOMO/LUMO energies for protonated
phosphite conformations 8 and 9 as a function of P-H bond distance. See
captions to Figure 6 for structures and scale factors used.
basis set) lower in energy than 8 (corresponding to the bicyclic
phosphite) at a P-H bond distance of 3.0-4.0 A (depending on
the scale factors and basis set used in the calculations; the energy
difference is 8.2 kcal/mol at 3-A separation and calculated with
the 6-21 basis set).
According to eq 2, the stereoelectronic effect provides the
greatest stabilization energy at this putative transition state P-H
bond distance with the greatest no - u* (HOMO/LUMO)
mixing. Indeed, as shown in Figure 6C, at a P-H distance of 3-4
A, the LUMO for both cyclic 8 and acyclic 9 has dropped
to about its lowest energy and the HOMO for bicyclic and
acyclic phosphite conformations has risen correspondingly. Thus,
by 3-4 A, major mixing of the HOMO and LUMO will occur.
Significantly in the ground-state protonated phosphite, the LUMO
of the acyclic 9 conformation is higher energy than that of the
bicyclic 8 conformation. By 3-4 A, the LUMO of the acyclic
J. Am. Chem. SOC. 1984, 106, 7831-7835 7831
phosphite has dropped -2 eV below the LUMO of the bicyclic
phosphite, and thus oxygen lone pair orbitals which are app to
the antibonding P-H orbital can more effectively mix in this
LUMO. This supports the HOMO/LUMO (n - u * ) mixing
interpretation for the stereoelectronic effect and the a-effect.
I t should be stressed that the total and HOMO/LUMO orbital
energies for the partially protonated phosphites are strongly de-
pendent on the scale factors chosen (see Figure 6A,B). The
STO-3G minimal basis set with standard scale factors did not
provide enough flexibility to allow convergence in the SCF pro-
cedure. This is quite reasonable since during protonation of the
phosphite a large amount of charge transfer occurs. At infinite
separation of proton and neutral phosphite no electron density
is permitted on the proton. As the proton moves closer, electron
density from the phosphite (particularly from the phosphorus and
oxygen lone pair orbitals) moves onto the hydrogen. To permit
this the 1s orbital on hydrogen must be allowed to become less
diffuse and thus the hydrogen scale factor must increase with
decreasing P-H bond distance. At the stationary point for the
protonated phosphite nearly one electron (Table 11) has transferred
from the phosphite to the proton 1s orbital (the hydrogen has only
a small positive charge). Obviously the same optimized scale
factor for this cannot be used for unprotonated, protonated, and
partially protonated structures. By 3-4 A (again the
transition-state distance) most of the charge transfer has oc-
curred. Only at shorter P-H distances ( <3 A) are standard scale
factors permissible. Using standard scale factors and the STO-3G
basis set at a P-H distance of 3 8, the acyclic phosphite is 12
kcal/mol more stable than the bicyclic phosphite (Figure 6B).
When optimized scale factors are used for the higher energy
structure, this energy difference increases to 25 kcal/mol.
Finally, completing the picture for proton transfer, at longer
P-H bond distances the energy difference between conformations
9 and 8 again decreases and as discussed previously at infinite
separation of the phosphite and proton, the relative energies of
the two conformations reverses.
Experimental support for these ideas is provided in the ac-
companying paper.
Acknowledgment. Support of this work by NSF (CHE
8205353), NI H (GM 17575), and DoD (FED CHEM 83 K0098)
is greatly appreciated. The generous allocation of computing time
by the UI C Computer Center and the help of Shigeyuki Urano
(UIC) with the implementation of GAUSSIAN 80 is also appreciated.
Experimental Tests of the Stereoelectronic Effect at
Phosphorus: Nucleophilic Reactivity of Phosphite Esters
Kazunari Taira, William L. Mock, and David G. Gorenstein*
Contribution f rom the Department of Chemistry, University of Illinois at Chicago, Chicago,
Illinois 60680. Received January 9, 1984. Revised Manuscript Received June 8, 1984
Abstract: Triethyl phosphite rapidly reacts with ethyl benzenesulfenate or diethyl peroxide to yield pentaethoxyphosphorane.
In contrast, l-methyl-4-phosha-3,5,8-trioxabicyclo[2.2.2]~tane (1) fails to react with either electrophile to yield the expected
bicyclic phosphorane 5. The poor reactivity of the bicyclic phosphite 1 is due to a kinetic rather than a thermodynamic barrier
because 5 is formed smoothly from an equimolar mixture of P(OEt), and the triol 1,1,1-tris(hydroxymethy1)ethane. This
result is interpreted in terms of the stereoelectronic effect. The order of nucleophilic reactivity of trialkyl phosphites wi th
3-benzylidene-2,4-pentanedione is also shown to beconsistent with the stereoelectronic effect. The bicyclic phosphite 1 reacted
750 times slower than the pseudoequatorial 2-methoxy ester of hexahydrobenzo-l,3,2-dioxaphosphorinane in a Michael addition
reaction with 3-benzylidene-2,4-pentanedione.
Stereoelectronic effects have been suggested to significantly
influence the rates, products, and stereochemistry of reactions of
organophosphorus compounds.- In contrast to the large body
of experimental and theoretical work supporting the role of orbital
orientation (the stereoelectronic effect) in carbon chemistry,12-14
(!) Taira, K.; Gorenstein, D. G. J . Am. Chem. SOC. , preceding paper in
(2) Gorenstein, D. G.; Findlay, J . B.; Luxon, B. A,; Kar, D. J . Am. Chem.
(3) Gorenstein, D. G.; Luxon, B. A,; Findlay, J . B.; Momii, R. J. Am.
(4) Gorenstein, D. G.; Luxon, B. A,; Findlay, J . B. J. Am. Chem. SOC.
this Issue.
SOC. 1977, 99, 3473.
Chem. SOC. 1977, 99, 4170.
1977. 99. 8048.
(5) Gorenstein, D. G.; Luxon, B. A,; Goldfield, E. M. J . Am. Chem. SOC.
1980, 102, 1757.
(6) Gorenstein, D. G.; Rowell, R.; Taira, K. In Phosphorus Chemistry;
American Chemical Society: Washington, DC, 1981; ACS Symp. Ser. No.
171. D 69.
~ ~ . = -
(7) Gorenstein, D. G.; Luxon, B. A.; Findlay, J . B. J. Am. Chem. SOC.
1979, 101, 5869.
(8) Gorenstein, D. G.; Taira, K. J. Am. Chem. Soc. 1982, 104, 6130.
Taira, K.; Fanni, T.; Gorenstein, D. G. J. Am. Chem. SOC. 1984, 106, 1521.
(9) Gorenstein, D. G.; Rowell, R.; Findlay, J . J. Am. Chem. SOC. 1980,
102, 50.
(10) Lehn, J . M.; Wipff, G. J . Chem. SOC., Chem. Commun. 1975, 800.
(11) (a) Lehn, J . M.; Wipff, G. J . Am. Chem. SOC. 1974, 96, 4048. (b)
Ibid. 1976, 98, 7498. (c) Helv. Chim. Acta 1978, 61, 1274.
no direct experimental evidence has previously existed to support
this hypothesis in the reactions of organophosphorus compounds.
As described in the preceding article and in others i n this
series,-7 our ab initio molecular orbital calculations have suggested
that the orientation of lone pairs on directly bonded oxygen or
nitrogen atoms can significantly affect the reactivity of organo-
phosphorus compounds. In phosphate esters this stereoelectronic
effect involves activation of a P-0 ester bond by antiperiplanar
(app) interaction with oxygen or nitrogen electron lone pairs.
Calculations have suggested that orientation of a lone pair an-
tiperiplanar to a scissile bond can lower the energy of a transition
state by as much as 11 kcal/mol relative to a corresponding
transition state without this app lone air.^^^ Unfortunately,
attempts to experimentally confirm this effect have been frustrated
by conformational flexibility in the relatively unconstrained
phosphate ester systems earlier ~tudi ed.~
(12) (a) Radom, L.; Hehre, W. J .; Pople, J . A. J . Am. Chem. SOC. 1972,
94, 2371. (b) J effrey, G. A,; Pople, J . A,; Radom, C. Carbohydr. Res. 1972,
25, 117. (c) Cieplak, A. S. J. Am. Chem. SOC. 1981, 103, 4540. (d) Anh,
N. T.; Eisenstein, 0. Tetrahedron Lett. 1976, 155.
(13) Kirby, A. J . The Anomeric Effect and Related Stereoelectronic
Effects at Oxygen; Springer-Verlag: Berlin, 1983; pp 1-149.
(14) Deslongchamps, P.; Taillerfer, R. J . Can. J . Chem. 1975, 53, 3029
and references cited therein.
0002-7863/84/1506-7831$01.50/0 0 1984 American Chemical Society