Comparison of conformations of diesters of stabilized phosphonium ylides in solution and in the crystal

Article abstract of DOI:10.1080/10426500802077242

Computed bond lengths and angles of methyl ethyl and dimethyl triphenyl phosphonium ylidic diesters, 1b, c, respectively, are similar to those in the crystal, as for the diethyl ester, 1a, where both acyl oxygens are anti to phosphorus. The ¹H

Full text of DOI:10.1080/10426500802077242

This article was downloaded by: [Florida State University]
On: 05 October 2014, At: 07:30
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,
UK
Phosphorus, Sulfur, and Silicon
and the Related Elements
Publication details, including instructions for
authors and subscription information:
http://www.tandfonline.com/loi/gpss20
Comparison of Conformations
of Diesters of Stabilized
Phosphonium Ylides in Solution
and in the Crystal
Fernando Castañeda a , Paul Silva a , M. Teresa
Garland a , Ata Shirazi b & Clifford A. Bunton b
a University of Chile , Santiago , Chile
b University of California , Santa Barbara ,
California , USA
Published online: 19 Dec 2008.
To cite this article: Fernando Castañeda , Paul Silva , M. Teresa Garland , Ata Shirazi
& Clifford A. Bunton (2008) Comparison of Conformations of Diesters of Stabilized
Phosphonium Ylides in Solution and in the Crystal, Phosphorus, Sulfur, and Silicon and
the Related Elements, 184:1, 19-33, DOI: 10.1080/10426500802077242
To link to this article: http://dx.doi.org/10.1080/10426500802077242
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the
information (the “Content”) contained in the publications on our platform.
However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness,
or suitability for any purpose of the Content. Any opinions and views
expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the
Content should not be relied upon and should be independently verified with
primary sources of information. Taylor and Francis shall not be liable for any

losses, actions, claims, proceedings, demands, costs, expenses, damages,
and other liabilities whatsoever or howsoever caused arising directly or
indirectly in connection with, in relation to or arising out of the use of the
Content.
This article may be used for research, teaching, and private study purposes.
Any substantial or systematic reproduction, redistribution, reselling, loan,
sub-licensing, systematic supply, or distribution in any form to anyone is
expressly forbidden. Terms & Conditions of access and use can be found at
http://www.tandfonline.com/page/terms-and-conditions

Phosphorus, Sulfur, and Silicon, 184:19–33, 2009
Copyright © Taylor & Francis Group, LLC
ISSN: 1042-6507 print / 1563-5325 online
DOI: 10.1080/10426500802077242
Comparison of Conformations of Diesters of Stabilized
Phosphonium Ylides in Solution and in the Crystal
1
1
1
˜
Fernando Castaneda, Paul Silva, M. Teresa Garland,
Ata Shirazi,2 and Clifford A. Bunton2
1University of Chile, Santiago, Chile
2University of California, Santa Barbara, California, USA
Computed bond lengths and angles of methyl ethyl and dimethyl triphenyl phos-
phonium ylidic diesters, 1b, c, respectively, are similar to those in the crystal, as
for the diethyl ester, 1a, where both acyl oxygens are anti to phosphorus. The 1 H
and 13 C NMR spectra of the methyl ethyl diester, 1b, where one acyl oxygen is syn
and the other anti to phosphorus, are as expected in terms of the conformation in
the crystal, but the dimethyl ester, 1c, in the crystal is an equimolar mixture of
conformers. For a given ylidic diester the different conformers have similar ener-
gies from B3LYP//6-31G(d) computations, interconversions of conformers should
not be slow at ambient temperatures and 1 H and 13 C NMR signals in solution
are sharp. Estimation of Natural Atomic Charges indicates significant cationoid
character on phosphorus and the acyl carbons, and anionoid character on the ylidic
carbon and the ester oxygens depending on orientations towards phosphorus.
Keywords Phosphonium ylides; conformational analysis; structural computations;
NMR spectra
INTRODUCTION
Conformations of phosphonium ylides stabilized by two electronic with-
drawing substituents, e.g., ester or keto groups, depend on the bal-
ance between ylidic resonance and non-bonding interactions, including,
for the solid, packing forces in the crystal.1,2 Electronic delocalization
should favor near planar ylidic moieties with favorable interactions be-
tween acyl oxygens and cationoid phosphorus, and the bonds between
phosphorus and the ylidic carbon, and between it and the acyl carbons,
Received 7 January 2008; accepted 25 March 2008.
The authors thank the CEPEDEQ, Facultad de Ciencias Qu´ımicas y Farmace´uticas,
Universidad de Chile for instrumental facilities.
Address correspondence to Fernando Castan˜eda, Departamento de Qu´ımica Orga´nica
y Fisicoqu´ımica, Facultad de Ciencias Qu´ımicas y Farmace´uticas, Universidad de Chile,
Casilla 233, Santiago, Chile. E-mail: fcastane@ciq.uchile.cl
19

20
F. Castan˜eda et al.
SCHEME 1 R = alkyl, alkoxy.
should have orders between one and two. In solution NMR spectroscopy
indicates that these stabilized ylides exhibit free rotation about the
P C bond and the sharp signals indicate that one conformer is domi-
nant, or that conformers readily interconvert.3,4 With some ylides sta-
bilized by a single acyl group, geometrical isomers exist and the two
conformers can be identified, as shown in Scheme 1 for a triphenyl
phosphonium ylide.5
The conformers are designated as syn and anti, depending on the ori-
entation of the acyl oxygen relative to phosphorus. There are significant
energy barriers to interconversions of these monostabilized conformers
because ylidic resonance is lost on rotation out of the ylidic plane.6 The
situation is different for distabilized ylides, as shown in Scheme 2 for a
symmetrical diester, where rotation of one ester group out of the ylidic
plane should not affect resonance interaction involving the other.3,4
In Scheme 2, structures are written for simplicity, with a P-C double
bond and single bonds between the ylidic and acyl carbons; although
this classical structure is inadequate, does not represent bond lengths
(C–C, C–P) in the crystal. The phenyl groups are equivalent on the
NMR time scale, due to rotation about the bond between phosphorus
and the ylidic carbon. At the simplest level, one would expect a syn-syn
conformer with a planar ylidic moiety to be preferred, as in mono- sta-
bilized ylides, but with diesters there should be interference between
the two ester alkoxy groups anti to phosphorus. Despite expected fa-
vorable electrostatic interactions between acyl oxygens and cationoid
phosphorus there is a deviation from planarity of the ylidic moiety in
SCHEME 2 R = alkyl.

Diesters of Stabilized Phosphonium Ylides
21
SCHEME 3
the crystalline diethyl ester, 1a, R = Et, and both acyl oxygens are anti
to phosphorus, but there are favorable intramolecular interactions be-
tween phosphorus and the alkoxy groups.7 Inter- and intramolecular
interactions in the solid are important, butab initio computations at
the B3LYP//6-31G(d) level indicate that the independent conformers
should have similar energies and low barriers to interconversion, for
example, in solution the diethyl ester, 1a, may be a mixture of equili-
brating conformers.8 A difference in conformation in solution and the
solid in some ylidic keto compounds could be indicated by differences
in the 13C NMR spectra in these conditions, and packing forces should
be important in the solid. Changes in the alkyl groups in the diesters
might therefore control conformation in the crystal, and possibly in so-
lution, and in the crystalline methyl ethyl diester, 1b, the ethoxy group
and the acyl oxygen of the carbomethoxy group are syn to phosphorus,
but the crystalline dimethyl diester, 1c, is a 1:1 mixture of the syn-anti
and anti-anti conformers,9 Scheme 3.
In view of the estimated low energy barriers between different con-
formers of isolated stabilized phosphonium ylides, 1b and 1c may have
different conformations in solution and in the crystal.
In the present work, we used ab initio geometrical optimization
to predict geometries of isolated molecules, and 1H and 13C NMR
spectroscopy, plus estimation of Natural Atomic Charges, to examine

22
F. Castan˜eda et al.
TABLE I Selected Geometric Parameters (Å, ◦) for Ethyl Methyl
2-(Triphenylphosphoran ylidene)malonate, 1b, from Computation
and X-ray Crystallographya
P1-C1
C1-C2
C1-C3
C2-O2
C3-O4
C2-O1
C3-O3
P1-C1-C2
P1-C1-C3
1.75 (1.74)
1.45 (1.43)
1.46 (1.44)
1.21 (1.22)
1.19 (1.20)
1.32 (1.34)
1.35 (1.36)
113 (111)
C2-C1-C3
C1-C3-O4
O3-C3-O4
C1-C3-O3
O1-C2-O2
C1-C2-O2
O1-C2-C1
P1-C1-C2-O2
P1-C1-C3-O4
122 (124)
126 (127)
121 (121)
113 (112)
121 (121)
123 (123)
116 (116)
10 (10)
125 (125)
158 (154)
aValues in parentheses are for the X-ray crystallographic structure. Positions are
designated as in the X-ray crystallographic experiments.9
intramolecular interactions. The diester conformations are shown in
Tables I and II with positions designated as in the X ray crystallo-
graphic results.9 Positions of the phenyl groups are not specified be-
cause rapid rotation about the P C bond assures their equivalence in
solution on the NMR time scale, although interactions involving the
phenyl groups could influence conformations.
Estimation of energies of the various conformers of several stabilized
triphenylphosphonium ylides indicates that for isolated molecules the
syn-syn conformer should be preferred, but energy differences between
it and the syn-anti or anti-anti conformers should be small and much
lower than energy differences for monostabilized ylides. Rotation of
one ester group out of the ylidic plane increases the energy only mod-
estly, and interconversion of conformers in solution should be rapid
at ca., 25oC. These small energy differences indicate that in solution
interconversion of conformers should be fast on the NMR time scale.
However, NMR signal shapes and multiplicities are generally consis-
tent with one conformer being dominant.

Diesters of Stabilized Phosphonium Ylides
23
TABLE II Selected Geometric Parameters (Å, ◦) for
Dimethyl 2-(Triphenylphosphoranylidene) Malonate,
1c, from Computation and X-ray Crystallographya
P1-C1
C1-C2
C1-C3
C2-O2
C2-O1
C3-O3
C3-04
P1-C1-C2
P1-C1-C3
C2-C1-C3
O1-C2-O2
O2-C2-C1
O1-C2-C1
C1-C3-O4
O3-C3-O4
O3-C3-C1
P1-C1-C2-O2
P1-C1-C3-O4
1.74 (1.74)
1.45 (1.43)
1.46 (1.44)
1.19 (1.20)
1.35 (1.36)
1.36 (1.37)
1.19 (1.20)
123 (121)
119 (119)
118 (119)
121 (121)
127 (128)
112 (111)
128 (128)
121 (121)
111 (111)
−167 (−162)
163 (158)
P2-C6
C6-C7
C6-C8
C7-O6
C8-O8
C7-O5
C8-O7
P2-C6-C7
P2-C6-C8
C7-C6-C8
O5-C7-O6
O6-C7-C6
O5-C7-C6
C6-C8-O8
O8-C8-O7
O7-C8-C6
P2-C6-C8-O8
P2-C6-C7-O6
1.75 (1.74)
1.45 (1.44)
1.45 (1.43)
1.19 (1.21)
1.21 (1.22)
1.35 (1.36)
1.32 (1.35)
125 (119)
113 (117)
122 (123)
120 (121)
127 (129)
112 (110)
122 (125)
121 (121)
116 (114)
11 (9)
167 (155)
aValues in parentheses are for the X-ray crystallographic
structure. Positions are designated as in the X-ray crystallographic
experiments.9
In a phosphonium ylidic diester with both acyl oxygens oriented to-
wards phosphorus there could be, in the crystal, interference between
the alkoxy groups of the bulky trigonal ester groups and neighboring
molecules. Carboxylic esters take up the Z-conformation,10 with a small
preference over the E conformation and ab initio computation indicates
that this preference is maintained in ylidic esters. Rotation of the bulky
trigonal groups in esters should occur readily at ambient temperatures
but would involve significant interference between adjacent groups in
an ylidic diester with both acyl oxygens syn to phosphorus. Rotational
interconversion of Z- and E- ester conformers in an ylidic diester with

24
F. Castan˜eda et al.
both alkoxy groups anti to phosphorus will, therefore, be disfavored
by steric interference disfavoring these conformers. These unfavorable
interactions are neglected when single point energies are calculated
from ab initio geometrical optimizations of static structures, which
limits their efficacy in predicting conformations of the ylidic diesters.
This limitation is less important in treating conformations of monos-
tabililized ylides, or those stabilized by a linear cyano group where
Z E interconversions should not be sterically hindered, and syn acyl
oxygens in mono keto and ester derivatives are preferred.11 However,
so far as we know syn-syn conformers of diesters are not preferred in
the solid or in solution, and the question of their probable conforma-
tion is discussed later. The computations on static structures, as noted,
neglect intermolecular interactions and possible rotational mobility of
trigonal groups.
RESULTS AND DISCUSSION
Geometrical Optimizations
Ab initio computations were made with geometries from HF 6-31G(d)
or 6-311G(d) basis sets,12 and the predicted structures were indistin-
guishable by eye. Relative energies are from B3LYP computations with
the HF geometries, and for all the diester and diketo systems that we
have examined the syn-syn conformers are predicted to be preferred,
but energy differences between them and anti conformers are typically
within 3 kcal.mole−1and moving one ester group into an orthogonal
position has only a small energy effect, typically < 2 kcal.mole−1, con-
sistent with rapid conformational equilibration in solution.8,11b
These small energy differences indicate that in solution intercon-
version of conformers should be fast on the NMR time scale. However,
signal shapes and multiplicities are generally as expected in terms of
structures of the crystalline ylides.
Bond lengths and angles for conformers, with positions numbered
9
as in the crystal, are in Tables I and II, and for the dimethyl es-
ter, 1c, we give values for the anti-anti and syn-anti conformers. Bond
lengths and angles of distabilized ylides, but not necessarily torsional
angles, estimated by ab initio optimization, are typically similar to
those in the crystal where conformations, are apparently sensitive to
packing constraints. Computed distances and angles are given to the
second decimal place, and angles to the first whole number, but for the
crystal structures full numerical values are given in Reference 9. Crys-
talline 1c is an equimolar mixture of two interacting conformers and

Diesters of Stabilized Phosphonium Ylides
25
computations are for isolated molecules which complicates comparisons
of calculated and observed geometries.
Some bond angles at the acyl carbons differ from the expected 120◦for
trigonal carbons, probably because of steric effects, or strong interac-
tions between anionoid oxygens, and we see this effect in the crystal
and in the computed geometries. However, the sum of the bond angles
at these carbon atoms is close to the value of 360◦for a planar car-
bon, consistent with the strong electronic delocalization over the ylidic
system.
The computations neglect non-bonding interactions in mobile struc-
tures in solution as well as intermolecular interactions in the solid
and probably in solution. These phosphonium ylides have molecular
weights of ca. 400 and at the concentrations, 0.1−0.2 M , used in the
NMR measurements in solution, there may be interference between
bulky trigonal anti-acyl groups of neighboring molecules, as well as be-
tween these groups in the same molecule which complicates comparison
of NMR data with computation results. These unfavorable non-bonding
interactions in diesters are apparently more important than favorable
interactions between cationoid phosphorus and syn-acyl oxygens which
should be partially offset by interactions between syn alkoxy groups
and phosphorus, as shown later from consideration of Partial Atomic
Charges.
In the crystalline methyl-ethyl diester 1b the bulkier ethoxy group
is in the syn- position with respect to phosphorus, and the methoxy
group is anti, and later we consider evidence for this conformation in
solution.
NMR Spectroscopy
1
The H and 13C (1H decoupled) NMR spectroscopic signals were ex-
amined on a 300 MHz instrument and also on a 500 MHz instrument
where the 1H coupled 13C signals were also examined. All experiments
were in acid-free CDCl3.13
1
For all the stabilized phosphonium ylides sharp H and 13C signals
show that in solution rotation about the P-C bond is rapid on the NMR
time scale and the 31P chemical shifts indicate significant single bond
character, as in the crystal.
For the methyl ethyl and the dimethyl diesters, 1b, c, the marked
upfield 13C signals of ylidic carbons confirm their carbanionoid charac-
ter, consistent with the suggestion that these ylides can be written as
1
zwitterions, rather than with classical P C bonds.14 The H and 13C
signals shown in Table III are in the regions expected in terms of earlier

26
F. Castan˜eda et al.
TABLE III 1H-, 13C-, and 31P NMR Data for Methyl Ethyl Diester, 1b,
Ph3P C(CO2Me)CO2Et, and Dimethyl Diester, 1c, Ph3P C(CO2Me)2
Ylide
1b
1H NMR
13C NMR
31P NMR
20.87
0.75 (t,3H, J = 7Hz, O CH2 CH3) 13.94 (O CH2 CH3)
3.79 (q,2H, J = 7Hz;O CH2 CH3) 58.56 (O CH2 CH3)
3.30 (s,3H, O CH3)
50.22 (O CH3)
53.10(d, J = 122.6 Hz, P C)
168.10(d, J = 13.5Hz, CO2)
126.58(d, 1Jp−c = 94.1Hz, Cipso
7.4−7.8 (m,15H, aromatic)
)
128.49(d, 2Jp−c = 12.5Hz, Cortho
)
131.82(d, 4Jp−c = 3.0Hz, Cpara
)
)
133.26(d, 3Jp−c = 9.7Hz, Cmeta
1c
3.34 (s, 3H, O CH3)
49.15 (O CH3)
20.79
7.4−7.8 (m, 15H, aromatic)
52.30(d, J = 123.8Hz, P C)
167.40(d, J = 13.5Hz, CO2)
125.3(d,1Jp−c = 94.2Hz, Cipso
)
127.5(d,2Jp−c = 12.5Hz, Cortho
)
130.9(d,4Jp−c = 2.9Hz, Cpara
)
)
132.3(d,3Jp−c = 9.7Hz, Cmeta
NMR spectra in CDCl3 at 25◦C with TMS as internal reference for 1H and 13C, and
H3PO4, as external reference for 31P.
observations. Values of chemical shifts and multiplicities generally do
not provide evidence on preferred conformations, except that evidence
of π-shielding15 of terminal methyl hydrogens of alkoxy groups indi-
cates that in solution the group is oriented towards phosphorus, as
in 1b. Although in the crystal the dimethyl ester, 1c, is an equimolar
mixture of anti-anti and syn-anti conformers we saw only one set of
1H and 13C NMR signals and only one 31P NMR signal in CDCl3. The
signals, with the expected multiplicity, do not distinguish between the
possibilities that one conformer is dominant in solution, or that con-
version of conformers is fast on the NMR time scale, as expected in
view of small estimated energy differences between conformers stabi-
1
lized by two ester groups.8 The H chemical shifts and coupling con-
stants for both diester ylides are consistent with the other structural
evidence.
The 13C NMR spectrum of the methyl ethyl diester, 1b, in CDCl3,
provides conformational information by comparison of signals with,
and without, 1H decoupling, and some results concerning the ylidic
groups, are considered separately. The 13C signals of the phenyl groups
are as expected, in view of the equivalence of these groups on the NMR
time scale and the spectrum with and without 1H spin-spin decoupling
shows that 13C chemical shifts decrease in the sequencemeta >para >

Diesters of Stabilized Phosphonium Ylides
27
ortho, which accords with the generally accepted assignments based on
the magnitude of the coupling constants with 31P.2 These assignments
differ from those given earlier with the chemical shifts decreasing in
1
the sequence ortho > para> meta,16 but the H coupled 13C signals
do not fit this sequence because the signal at 128.49 ppm of the ortho
1
carbon is a doublet of doublets due to coupling with both H and 31P,
(2Jp−c = 12.53 Hz). The signal of the para carbon is identified by its
1
area, and is split by H and weak 31P coupling (4Jp−c = 3.0 Hz), but
without 1H decoupling part of the signal is under part of the meta
1
signal at 133.3 ppm, which itself is also split by H and 31P coupling.
In the 1H coupled 13C NMR spectrum the splitting patterns of themeta
and para carbons are similar, except that there is a single central peak
of the latter because there are two neighboring CH groups, and differ
from those of the ortho carbon with its single adjacent neighboring
CH group. However, only a single central peak was observed in the
1
para 13C signal. In the H coupled spectrum the 13C signal of the ipso
carbon becomes a doublet of triplets, J = 7.5 Hz, due to coupling with
1
31P and the ortho H. These generalizations regarding the 13C NMR
spectra probably also apply to 1c, although here only the 1H decoupled
spectrum was examined.
Ylidic Group in the Methyl Ethyl Diester, 1b
In crystalline 1b the ethoxy group is syn and the methoxy group is anti
to phosphorus. In solution (Table III) π-shielding15 of CH2CH3 indicates
that this geometry of the ethoxy group is maintained in solution. The
1
13C signals of acyl groups of stabilized phosphonium ylides in the H
decoupled spectra are sometimes broad singlets, which may overlap,
and they should be split by coupling with 31P, but they are well defined
in 1b. In CDCl3 at 125 MHz we saw doublets, J = 12.8 and 12.7 Hz
at 167.59 and 168.54 ppm, respectively, in the1H decoupled spectrum,
indicating different orientations of the acyl groups, and from the chem-
ical shifts we assume that the upfield signal is that of the ethoxy ester
1
acyl carbon. Removal of the H decoupling increased the line width of
each signal of the doublet at 167.59 ppm, due to long range 1H coupling,
although the signals were not resolved. Signals of the doublet at 168.54
ppm are split, and appear to become a somewhat distorted doublet of
quartets, and not all the signals are resolved, J = 3 Hz (Figure 1).
1
There is long range coupling with H of the methoxy group which in
the crystal is anti to phosphorus.9 These observations indicate that the
conformation of the methyl ethyl diester, 1b, is the same in the crystal
and in solution, at least as regards the syn- position of the ethoxy group,

28
F. Castan˜eda et al.
FIGURE 1 13C NMR signals of the carbonyl groups of 1b with long-range
coupling with 31P and removal of the 1H decoupling.
and later we show evidence, from estimation of Partial Atomic Charges,
indicating that the methoxy group is probably anti in solution.
The 13C NMR signals of the methoxy and ethoxy groups of 1b are
split by 1H coupling. The OCH2CH3 signal becomes a quartet, J = 126
Hz, and the OCH2CH3 signal becomes a triplet, J = 146 Hz, with weak
long range coupling with 31P, J = 5 Hz. The OCH3 signal becomes a
quartet, J = 145 Hz. It is generally convenient to monitor 13C NMR
spectra with 1H decoupling, but with some of these ylides 1H coupling
provides additional useful information.
Partial Atomic Charges
The assignment of charge to individual atoms is convenient in consid-
ering some molecular properties, although it is essentially an empirical
concept. Various approximate treatments have been developed for esti-
mating Partial Atomic Charges,12 and Natural Charges for some posi-
tions in the phosphonium ylides are given in Table IV from B3LYP//HF
6-31G(d) computations on conformers as in the crystal. The syn-anti
conformation of the methyl ethyl diester, 1b, in the crystal is probably
also that in solution, but the syn-anti- and anti-anti-conformers of the
dimethyl diester, 1c, are associated in the crystal9 and partial charges
were estimated for both conformers. Partial Atomic Charges for given

Diesters of Stabilized Phosphonium Ylides
29
TABLE IV Fractional Atomic Chargesa for Methyl Ethyl Diester, 1b,
and Dimethyl Diester, 1c
Methyl ethyl diester, 1b, syn-anti conformer
Position
P1
C1
C2
C3
O1
O2b
O3b
O4
Atomic charge
Dimethyl diester, 1c, anti-anti and syn-anti conformers.
Position
Atomic charge
Position
1.72
−0.83
0.79
0.80
−0.53
−0.68
−0.58
−0.61
P1
1.72
P2
C1
−0.85
C6
C2
0.79
C7
C3
0.80
C8
O1b
−0.58
O5b
O2
−0.60
O6
O3b
−0.58
O7
O4
−0.61
O8b
Atomic charge
1.72
−0.83
0.79
0.79
−0.58
−0.61
−0.52
−0.68
aAtoms are numbered as in Tables I and II. bOxygens oriented towards phosphorus.
groups are similar for these diesters and it appears that substitution
of an ethyl for a methyl group has little effect on the charges.
The ylidic carbon is predicted to be strongly anionoid, consistent
with the strongly upfield 13C NMR signal, typical of these stabilized
ylides, and the partial charges on phosphorus and the acyl carbons are
as expected in terms of ylidic resonance, but the alkoxy oxygens are
also significantly anionoid and in the syn position should interact fa-
vorably with phosphorus, as in the crystal. Estimated atomic charges
are similar for the anti-anti and syn-anti conformers of the dimethyl
ester, consistent with bond lengths and angles being similar for the
two conformers (Table IV). Differences in the anionoid characters of
acyl or alkyl oxygens are modestly sensitive to conformation, and, for
an acyl group, the anionoid character is consistently slightly higher for
oxygens oriented towards, rather than away from, phosphorus, which
corresponds to lower carbon-oxygen bond orders and slightly longer
bonds differing by ca. 0.01 Å, (Tables I and II). This generalization re-
garding atomic charge and location relative to phosphorus also applies
to hypothetical conformers with structures different from those in the
crystal (data not shown).
The higher negative partial charge on an alkoxy oxygen syn, rather
than anti, to phosphorus (Table IV) should shift the 13C NMR signal
of the CH3 group slightly upfield, and this hypothesis can be tested
by considering these chemical shifts in 1b and 1c which are 50.22 and
49.15 ppm., respectively, in CDCl3(Table III).In crystalline 1c the OCH3
groups are largely oriented towards phosphorus, and the more anionoid
character on oxygen, as compared to that in 1b (Table IV), is consis-
tent with the difference in 13C chemical shifts of the methoxy groups
(Table III). This analysis involves the assumption that the preferred
orientation of OCH3 in the crystal (Tables 1 and 2) is maintained in
solution.

30
F. Castan˜eda et al.
Electronic and Steric Effects on Conformation
Electronic delocalization is important in these stabilized phosphonium
ylides with favorable interactions between cationoid phosphorus and
anionoid acyl oxygens in ester substituent groups, and to a lesser ex-
tent in alkoxy groups. This hypothesis is supported by ab initio compu-
tations which indicate that syn-syn conformers should be energetically
preferred, but so far as we know, both acyl oxygens are not syn to phos-
phorus in the crystalline diesters. Differences in computed energies
of conformers of diester ylides are small, as are barriers to intercon-
version, so that non-bonding interactions are probably very important,
and estimation of Partial Atomic Charges indicates that alkoxy oxygens
should interact with phosphorus, but less readily than acyl oxygens. Mi-
nor changes in the alkyl groups, e.g., from Me to Et affect structure in
the crystal and probably in solution, and are consistent with small ener-
getic differences between conformers. In hypothetical diesters with two,
freely rotating, bulky alkoxy groups anti to phosphorus there should
be in solution considerable interference between these groups. These
interactions are more important than those of an acyl oxygen with
phosphorus, although they are neglected in our computations on static
structures. In the crystal, where packing is all important, bulky alkoxy
groups anti to phosphorus are disfavored sterically by interaction with
neighboring ylides, and in diesters the bulkier alkoxy group is oriented
towards phosphorus. For example, in the methyl ethyl diester, 1b, the
bulkier ethoxy group is oriented towards phosphorus, in the solid and
probably in solution, despite the more negative character of the acyl
over the ethoxy oxygen. Bond angles at the trigonal ylidic C1 in the
crystal are larger towards C3 of the ethyl ester group, than towards C2
of the methyl ester group, although the sum of the bond angles is ca.,
360◦(Table I). Deviations from the 120◦ bond angles at trigonal carbons
are also evident for the acyl carbons C2 and C3 in the crystal and for
computation on an isolated molecule (Table I).
For the dimethyl ester, 1c, which has associated anti-anti- and syn-
anti- conformers in the crystal, there is reasonable agreement between
computed bond angles and those in the crystal, and observed deviations
from the expected 120◦ bond angles are qualitatively as predicted for
isolated molecules by ab initio geometrical optimization, but the sum of
the bond angles is close to the expected value of 360◦. This association
of anti-anti and syn-anti conformers in the solid is consistent with the
melting point of the dimethyl ester, 192◦C, being much higher than
those for most triphenyl phosphonium ylides.
For both ylides, 1b and 1c, predicted bond lengths are similar to
those in the crystal. This generalization is reasonable, because energies

Diesters of Stabilized Phosphonium Ylides
31
for changes in bond lengths are much higher than for changes in bond
angles. The inability of geometrical optimization for isolated molecules
to provide relative energies of stabilized ylides consistent with the crys-
tal structure is probably due to neglect of packing forces in the solid
and the assumption of static structures in the computations, but exten-
sive evidence indicates that energy differences between conformers is
small, and their interconversion involves changes in torsional angles
rather than in covalencies.
CONCLUSIONS
The ester groups have the Z-conformation in the crystal, and, based
on geometrical optimization, also preferentially for isolated molecules.
Conformations of various crystalline ylidic diesters differ, depending
1
on the alkoxy groups. In solution the sharp H and 13C NMR signals
are as expected for stabilized ylides in terms of electronic effects of
substituents and rapid interconversion of conformers.
Interference between two trigonal groups anti-to phosphorus proba-
bly precludes the existence of syn-syn conformers in the solid and prob-
ably also in solution. The situation is different with monostabilized
ylides, or when one of the stabilizing groups is linear, as in phospho-
nium ylides stabilized by cyano and keto or ester groups, where the
syn conformers are preferred in both the solid and in solution.11 Ester
groups generally exist as Z-conformers about the acyl carbon-oxygen
bond,10 as a result in a hypothetical syn-syn diester with a static struc-
ture there would be an unfavorable interaction between the electron
lone pairs of the alkoxy groups and if the ester groups take up the dis-
favored E-conformation there would be steric interference between the
two alkyl groups. We know of no experimental evidence for these ylidic
diesters taking up the syn-syn conformation in the crystal.
EXPERIMENTAL
Diethyl 2-(triphenylphosphoranylidene) malonate, 1a, was prepared
by reaction of dibromotriphenylphosphine and diethylmalonate in the
presence of triethylamine as described by Horner and Oediger.17
Methyl ethyl and dimethyl 2-(triphenylphosphoranylidene)
malonate, 1b and 1c, were obtained by transylidation
reaction of (methoxycarbonylmethylene) triphenyl phosphorane,
Ph3P CH CO2CH3, with ethyl and methyl chloroformate, respec-
9
˜
tively, as reported by Castaneda et al.

32
F. Castan˜eda et al.
1H, 13C, and31P NMR for 1a, 1b, and 1c were monitored on Bruker
DRX 300 or Varian Inova 500 spectrometers and were referenced to
TMS or external 85% H3PO4. 13C NMR spectra were obtained with and
without 1H decoupling.
Selected geometric parameters and molecular structures for 1b and
1c were provided by a Bruker SMART APEX CCD area-detector diffrac-
9
˜
tometer as published by Castaneda et al.
Computations for structural optimization were made with Spartan
06 for Windows (Wavefunction) software, and optimized with HF 6-
31G(d) basis set or occasionally with the HF 6-311G(d) basis set with
almost identical results. Energies were estimated by density functional
B3LYP and HF geometries.
REFERENCES
[1] H. J. Cristau and F. Plenat, In The Chemistry of Organophosphorus Compounds, F.
R. Hartley, Ed. (Wiley, New York, 1994), Vol. 3, Ch. 2, pp. 108–110.
[2] S. M. Bachrach and C. I. Nitsche, In The Chemistry of Organophosphorus Com-
pounds, F. R. Hartley, Ed. (Wiley, New York, 1994), Vol. 3, Ch. 4, pp. 273–299.
[3] (a) A. D. Abell and R. A. Massy-Westropp, Aust. J. Chem., 35, 2077–2087 (1982); (b)
A. D. Abell, B. M. Clark, and W. T. Robinson, Aust. J. Chem., 42, 1161–1167 (1989);
(c) A. D. Abell , J. Trent, and W. T. Robinson, Aust. J. Chem., 41, 1243–1249 (1988);
(d) A. D. Abell and R. A. Massy-Westropp, Aust. J. Chem., 35, 2277–2287 (1982); (e)
M. P. Cooke and K. P. Biciunas, Synthesis, 283–285 (1981).
[4] (a) F. Castan˜eda, C. A. Terraza, C. A. Bunton, N. D. Gillitt, and M. T. Garland,
Phosphorus, Sulfur, and Silicon, 178, 1973–1985 (2003); (b) F. Castan˜eda, G. I.
Recabarren, J. Hu, N. D. Gillitt, and C. A. Bunton, Phosphorus, Sulfur, and Silicon,
178, 2505–2518 (2003).
[5] (a) F. J. Randall and A. W. Johnson, Tetrahedron Lett., 2841–2845 (1968); (b) I. F.
Wilson and J. C. Tebby, J. Chem. Soc. Perkin I, 31–34 (1972); (c) D. M. Crouse, A.
T. Wehman, and E. E. Schweizer, Chem. Commun., 866–867 (1968); (d) I. F. Wilson
and J. C. Tebby, Tetrahedron Lett., 3769–3770 (1970).
[6] H. I. Zeliger, J. P. Snyder, and H. J. Bestmann, Tetrahedron Lett., 2199–2202 (1969).
[7] F. Castan˜eda, C. Aliaga, C. A. Bunton, M. T. Garland, and R. Baggio, Acta Crystal-
logr., C61, 496–499 (2005).
[8] F. Castan˜eda, C. Aliaga, M. T. Garland, J. Hu, N. D. Gillitt, and C. A. Bunton,
Phosphorus, Sulfur, and Silicon, 182, 151–160 (2007).
[9] F. Castan˜eda, C. Jullian, C. A. Bunton, M. T. Garland, and R. Baggio, Acta Crystal-
logr., C63, 267–270 (2007).
[10] (a) E. L. Eliel and S. H.. Wilen, Stereochemistry of Organic Compounds (Wiley,
New York, 1994), Ch. 10, pp. 618–619; (b) C. E. Blom, H. H. Gu¨nthard, Chem. Phys.
Lett., 84, 267–268 (1981); (c) K. B. Wiberg and K. E. Laidig, J. Am. Chem. Soc., 109,
5935–5948 (1987).
[11] (a) F. Castan˜eda, C. A. Bunton, R. Baggio, and M. T. Garland, Acta Crystallogr., C63,
65–67 (2007); (b) F. Castan˜eda, C. Acun˜a, M. T. Garland, N. D. Gillitt, A. Shirazi,
S. J. Yunes, and C. A. Bunton, Phosphorus, Sulfur, and Silicon, 182, 2541–2554
(2007).

Diesters of Stabilized Phosphonium Ylides
33
[12] (a) W. J. Hehre, A Guide to Molecular Mechanics and Quantum Chemical Cal-
culations (Wavefunction, Irvine, 2003); (b) C. J. Cramer, Essentials of Quantum
Chemistry, (Wiley, New York, 2002).
[13] H. J. Bestmann, H. G. Liberda, and J. P. Snyder, J. Am. Chem. Soc., 90, 2963–2964
(1968).
[14] M. Kobayashi, F. Sanda, and T. Endo, Macromolecules, 33, 5384–5387 (2000).
[15] (a) M. Nishio, Y. Umezawa, M. Hirota, and Y. Takeuchi, Tetrahedron, 51, 8665–8701
(1995); (b) M. Nishio and M. Hirota, Tetrahedron, 45, 7201–7245 (1989).
[16] (a) G. A. Gray, J. Am. Chem. Soc., 95, 7736–7742 (1973); (b) T. A. Albright, M. D.
Gordon, W. J. Freeman, and E. E. Schweizer, J. Am. Chem Soc., 98, 6249–6252
(1976).
[17] L. Horner and H. Oediger, Chem. Ber., 91, 437–442 (1958).