Some time ago, we learned how to turn alkenes into carbonyls via ozonolysis.
But we haven’t yet learned how to go in the opposite direction – from carbonyls to alkenes.
I like to think of functional groups like airports, and reactions as being like flights. In any reaction map you care to draw, carbonyls and alkenes are big hubs. So only knowing how to go from carbonyls to alkenes is a bit like being able to book a one-way flight from New York to Chicago… but not a return!
Let’s address this important gap in our synthetic toolbox.
Below is a useful reaction called the Wittig reaction that achieves this transformation. It won its inventor, Georg Wittig, the 1979 Nobel Prize in Chemistry (along with the father of hydroboration, H.C. Brown).
The two components of this reaction are:
- a carbonyl compound (aldehydes and ketones both work, but not esters or amides)
- a rather strange-looking species known as an ylide. (specifically, a “phosphonium ylide”, because there are also ylides of nitrogen and sulfur).
The technical definition of an ylide is a species with opposite formal charges on adjacent atoms. Although we drew the ylide, above, with a double bond between C and P, it also has an important resonance form with a positive charge on phosphorus and a negative charge on carbon:
The carbon of this ylide therefore behaves in many ways as a carbanion, and can readily act as a nucleophile.
The Mechanism of the Wittig Reaction
If you look above to the bonds that form and break in the Wittig reaction, you’ll see that it essentially swaps C=P and C=O bonds for C=C and O=P bonds.
So how does it work?
The version of events described in most introductory textbooks follows below. [In this footnote, I describe a slightly modified account of the mechanism that is generally more accurate. ]
We’ve already seen many examples of how carbonyl carbons are excellent electrophiles. reacting with nucleophiles such as Grignard reagents, metal hydrides, organolithiums and many other species. [e.g. The Simple 2-Step Pattern For 7 Reactions of Aldehydes and Ketones]
So one can imagine the first step of the Wittig reaction as being the attack of the nucleophilic ylide carbon on the electrophilic carbonyl carbon, providing a species with a negative charge on oxygen and a positive charge on phosphorus. This is the classic “addition” (sometimes called “1,2-addition”) mechanism to carbonyls.
The second step of the Wittig, then, is the attack of the resulting oxygen at phosphorus, giving a 4-membered ring. [Cocktail-party worthy fun fact: this is called an oxaphosphetane ]
The 4-membered ring is very short-lived and quickly breaks down, via a process called a reverse [2+2] cycloaddition, to give the final products: a phosphine oxide (“triphenylphosphine oxide” in this case), and the new alkene.
[In many cases, step 1 and step 2 essentially happen simultaneously, but this mechanism is fine for our purposes. ]
Although the example above is fairly simple, the Wittig reaction can be readily extended to more complex reaction partners, as we’ll see below. Before we dive into that, though, it might be worth a few moments for a brief digression.
How Ylides Are Made: A Quick Primer
Ylides might look a little exotic, but their synthesis is actually quite straightforward and involves no unfamiliar chemistry. They are usually made through just two familiar reactions: nucleophilic substitution reaction (SN2) followed by an acid-base reaction.
We start by treating an alkyl halide (another functional group “hub” in our airport analogy) with the excellent nucleophile triphenylphosphine (PPh3), which displaces the leaving group (via SN2) to give a phosphonium salt.
[When planning a Wittig, it’s generally best to use a primary alkyl halide (or alkyl sulfonate) here, as secondary alkyl halides don’t work as well. ]
The C-H bond adjacent to the phosphorus is relatively acidic [Note 2] and can be deprotonated with strong base to give the ylide shown. A common base to use is the readily available n-butyllithium (n-BuLi ). Sodium amide (NaNH2) can also be used.
The resulting ylide is then ready to go. No need to isolate it – just add an aldehyde or ketone, and the reaction should proceed nicely.
Some Examples Of The Wittig Reaction
For example, here’s the above ylide in a Wittig reaction with cyclohexanone:
The Wittig can be used to convert a wide variety of ketones and aldehydes to alkenes. [click for an image with yet more examples]
It can even be used to form rings. Here, we form a double bond between C-1 and C-6:
In summary, the Wittig is a very important reaction for several reasons:
- it’s a carbon-carbon bond forming reaction, which allows for extension of the carbon chain
- the components (carbonyls and ylides) are readily available and/or easily synthesized from readily available precursors
- the resulting alkenes can be further transformed into a large variety of functional groups – too many to list here, but for some inspiration, check out this reaction map of alkenes.
Here’s a quick example in synthesis: extending the carbon chain and incorporating an alcohol at the end, via hydroboration.
For many students, that’s all you need to know about the Wittig for now. For those insatiably curious about what can go wrong when we move beyond some simple examples… read on.
[Advanced] When Stereochemistry Rears Its Ugly Head
The examples above rigorously avoided any situation where a mixture of E and Z alkenes could be obtained.
What happens when we try to combine an aldehyde (or unsymmetrical ketone) with an unsymmetrical ylide?
We won’t wade too deeply into the topic of stereochemistry here, but for your average aldehyde reacting with your average ylide prepared by the methods above, the major alkene stereoisomer tends to be Z. [note 4]
The ratio of the Z isomer decreases as electron-withdrawing groups are added to the ylide. These species are called, “stabilized ylides”, as they are less basic (and less reactive).
This is simple enough and probably enough for most purposes. We don’t have time here to get into the excellent Horner-Emmons-Wadsworth reaction, which bears many similarities to the Wittig, and provides excellent E:Z selectivity.
Note 1. We’ve shown the mechanism occurring in a stepwise process, but a detailed study of the Wittig mechanism [see here] strongly suggests that it mainly through a [2+2] cycloaddition followed by a reverse [2+2] cycloaddition.
That is not nearly as complicated as it sounds. Following the central figures in this square dance will give you the main idea:
[adapted from the original on youtube: also invoked to to explain the Chauvin mechanism of olefin metathesis]
Here’s what it looks like. Note that instead of an initial addition step (leading to a betaine intermediate), the four-membered ring is just formed directly. The second step (reverse [2+2] cycloaddition) is the same.
Note 2. Called a “betaine”.
Note 3. The Evans pKa table gives a pKa for Ph3P–CH3 of 22 (in DMSO), making it more acidic than a terminal alkyne. Note that PPh3 is a good choice because it doesn’t have any potentially acidic carbons adjacent to the phosphorus. If we used P(CH3)3 as the nucleophile, for example, then using n-BuLi could lead to a mixture of of ylides in many cases. We don’t want that! [Nor, for most purposes, do we want P(CH3)3, which is among the more foul-smelling and toxic liquids you will ever encounter in a chemistry lab, but I digress…. ]
Note 4. Very interesting to note that the identity of the base matters greatly: using a lithium base, such as n-BuLi, (and added lithium iodide) results in a mixture of products with a Z:E ratio of 58:42 , whereas sodium bases give more of the Z. In the presence of added lithium salts, it’s likely true that a betaine intermediate is present, and there is more equilibration between the starting materials and the intermediate oxaphosphetane.