The Stereoselectivity of the Wittig Reaction

Discovered in the 1950s, the Wittig reaction is one of the premier methods for forming
carbon-carbon bonds, constituting just the third general method for forming such bonds that you
have seen thus far in your study of organic chemistry (the other two being the Diels-Alder reaction
and addition of Grignard reagents to carbonyl groups). In fact, the Wittig reaction is of such
widespread use and value to society in areas as diverse as pharmaceutical synthesis and petroleum
research that its namesake received the 1979 Nobel Prize in Chemistry for its discovery. The
following commentary is meant to supplement your McMurry text by offering a formalized
presentation of the reaction and its power in selectively uniting two compounds to form olefinic
products with high selectivity.
A generalized picture for the net transformation is shown below for an aldehyde substrate,
though ketones work equally well. In the event, the carbonyl group reacts with a phosphorouscontaining reagent known as an ylide, ultimately yielding an alkene product along with
triphenylphosphine oxide (O=PPh3). High selectivity in terms of the geometry of the newly formed
alkene is always observed, dependent only upon the nature of the group Y attached to the anionic
carbon of the ylide reagent. To understand this statement, we must look at the mechanism of this
reaction in detail.
R

Ph3P

Ph3P

Y
Z-alkene

Y
Ylide

or

O
+

PPh3

E-alkene

In the initial step, the resonance form of the ylide in which an anionic charge resides on
carbon attacks the electrophilic carbon of the aldehyde to form a new carbon-carbon bond. Though
this overall addition is similar to many other reactions you have seen, the unique (and most salient)
feature of this process is that it occurs via a transition state that minimizes the steric interactions
between the R group on the aldehyde and the Y group on the ylide (i.e. placing them as far apart as
possible). Once this attack leading to a betaine intermediate is complete, rotation around the newly
formed C-C bond then makes it possible to form a four-membered ring known as an oxaphosH

Ph3P

Ph3P

Y
R

H
R

[Rotate around
central bond]

PPh3

Betaine
Y = alkyl group

O
PPh3

+
R

Z-alkene

[Rate
determining
step]

H
R

PPh3
O

H
Oxaphosphatane

phatane. As you can see in the structure in the lower right-hand corner, the R and Y groups are now
quite close together; in fact, they possess a high degree of steric interaction, imparting strain that
can be relieved only by breaking the relatively weak phosphorous-oxygen bond and reverting back
to the original betaine partner, a structure which is always in equilibrium with the closed fourmembered ring alternative. Over time, however, this mix will funnel entirely via LeChateliers
principle to an alkene product and triphenylphosphine oxide in a final step that mechanistically

pushes two pairs of electrons (in the form of two sigma bonds) concurrently as drawn. As you might
expect, given the orientation of the groups in the oxaphosphatane structure a cis-, or Z-disposed
alkene results, and it is this concluding operation that is the rate-determining step for the Wittig
reaction (i.e. the slowest one in the entire sequence). Overall, this picture is an accurate reflection
for the reaction when Y is an alkyl group.
However, if Y is something other than this type of substituent, an opportunity exists for a
slightly different outcome in terms of product olefin geometry due entirely to the fact that the final
step is so slow. Indeed, if Y is a functional group capable of delocalizing a neighboring anionic
charge via resonance, a trans-, or E-disposed olefin will be formed instead. To understand this
scenario, we need to look at the mechanism below. As you hopefully see upon initial inspection,
the opening operations are exactly the same as that on the first page of this handout, with
nucleophilic attack occurring in such a way that steric interactions are minimized, yielding exactly
the same oxaphosphatane that places the Y and R groups on the same side of the molecule. Now
comes the key difference. Because the final breakdown of the oxaphosphatane is so slow, and
because carbon-phosphorous bonds tend to be fairly weak, when group Y can support an anionic
charge via resonance (i.e. if it is a C=O, cyanide, NO2, or SO2Ph), then that bond breaks prior to
alkene formation to give a new, and stabilized, carbanion intermediate, one in which all chiral
information has been lost since it possesses sp2 hybridization. When this new intermediate reforms
an oxaphosphatane, it now does so in such a way that it minimizes the steric interactions between
the R and Y groups by orienting them as far apart as possible on opposite sides of the closed ring.
This oxaphosphatane is the most stable thermodynamically, so, in the final rate-determining
elimination, this oxaphosphatane will be the only one in solution, and thus only a trans-olefin
product will be observed.
H

Ph3P

Ph3P

Y
R

H
R

[Rotate around
central bond]

PPh3

H
R

PPh3
O

H
Oxaphosphatane

Betaine
Y = group that can stabilize
an anion via resonance

O
PPh3

+
R

E-alkene

[Rate
determining
step]

H Y

PPh3

Y
R

PPh3

Oxaphosphatane

The four examples below offer a more concrete view of this reaction and its stereochemical
outcome. The only new key feature to take note of is when multiple groups are present on either the
carbonyl fragment or the ylide, the product that results is still the one that has the expected E- or Zstereochemistry based on the identity of the larger substituent of flavor Y and/or R, giving an
answer that follows the priority rules for alkene labeling that you learned last semester. Convince
yourself of the veracity of this statement.
A

C
O

EtO
O

Ph3P

O
O

Ph3P

OMe
D

NO2
O

Ph3P

Ph3P

NO2