Williamson Ether Synthesis

The big idea: build an ether one bond at a time

Ethers — molecules with the general formula R–O–R′ — are everywhere, from diethyl ether (the classic anesthetic and solvent) to the methyl ethers that decorate natural products and pharmaceuticals. They are remarkably stable and unreactive once formed, which makes them excellent solvents but also means you cannot simply heat two alcohols together and expect them to fuse cleanly. Acid-catalyzed dehydration of alcohols works only for symmetric ethers from simple primary alcohols and gives messy mixtures otherwise.

The Williamson ether synthesis solves the problem with surgical precision. Instead of relying on equilibrium, it uses a one-way reaction: a strongly nucleophilic alkoxide (R–O⁻) attacks the electrophilic carbon of an alkyl halide, and the new C–O bond forms while the C–halide bond breaks in the same motion. There is no intermediate, no carbocation, no chance for the pieces to scramble. You decide exactly which carbon connects to the oxygen, and the reaction delivers that ether and nothing else.

The mechanism: a single concerted SN2 step

The reaction is a textbook example of bimolecular nucleophilic substitution, SN2. Two species meet in the rate-determining step, so the rate law is second order: rate = k[alkoxide][R–X]. The sequence has two conceptual parts.

Step 1 — make the nucleophile. An alcohol has a pK_a around 16–18, so it is only weakly acidic. To convert it fully to its conjugate base, you need a base stronger than the alkoxide itself. Sodium hydride (NaH) is ideal: it deprotonates the alcohol irreversibly, releasing H₂ gas and leaving a clean, salt-free alkoxide. Sodium or potassium metal works the same way (R–OH + Na → R–O⁻Na⁺ + ½H₂), and for the very hindered tert-butoxide, potassium tert-butoxide is bought ready-made. Phenols, with pK_a ≈ 10, are acidic enough that a mild carbonate base such as K₂CO₃ generates the phenoxide.

Step 2 — the SN2 displacement. The lone pair on the alkoxide oxygen attacks the carbon bearing the leaving group from the side directly opposite the leaving group — a backside attack. As the new O–C bond forms, the C–X bond stretches and breaks. At the transition state the central carbon is sp²-like and trigonal-bipyramidal, with the incoming oxygen and departing halide on the axis. The carbon's three other substituents flip through, like an umbrella in the wind, so a stereocenter is inverted (Walden inversion). The products are the ether and a halide ion.

Because everything happens at once, the energy profile has a single hill — the SN2 transition state — and no valley in between. The activation energy is governed by how crowded that transition state is, which is why the structure of the electrophilic carbon matters so much.

Why the halide must be primary

SN2 transition states are sensitive to steric bulk because the nucleophile must squeeze past the substituents already on the carbon. Each added alkyl group raises the activation barrier sharply. Relative SN2 rates (methyl set to 1) make the point vividly:

Substrate	Example	Relative SN2 rate	Williamson outcome
Methyl	CH₃–I	~30	Excellent — fastest
Primary (1°)	CH₃CH₂–Br	1	Excellent — the sweet spot
Secondary (2°)	(CH₃)₂CH–Br	~0.03	Poor — E2 competes, mixtures
Tertiary (3°)	(CH₃)₃C–Br	~0	Fails — E2 gives only alkene

The lesson is built into how chemists plan the reaction. The alkoxide can be as bulky as you like — it does not sit at the crowded carbon. But the primary halide partner must be unhindered. When the two halves of a target ether differ in bulk, the strategic disconnection always assigns the hindered group to the alkoxide and the simple, unbranched group to the electrophile.

The textbook illustration is tert-butyl methyl ether, (CH₃)₃C–O–CH₃ (a relative of the fuel additive MTBE). There are two ways to imagine assembling it:

Route	Alkoxide	Halide	Result
Right way	(CH₃)₃C–O⁻ (tertiary, but it's the nucleophile)	CH₃–I (methyl, the electrophile)	Clean ether — SN2 at methyl
Wrong way	CH₃–O⁻ (methoxide)	(CH₃)₃C–Br (tertiary electrophile)	Failure — only isobutylene via E2

Same atoms, same connectivity in the product, but only the first route works. The tertiary carbon can never be the SN2 target.

Leaving groups, solvents, and the numbers that matter

Three knobs control how fast and how cleanly the displacement runs.

Leaving group. A good leaving group is a weak base — it leaves with the bonding electrons and is happy to do so. For alkyl pseudohalides the order is iodide > bromide > tosylate (OTs) ≈ mesylate (OMs) > chloride ≫ fluoride. Iodides and tosylates are the workhorses; fluorides essentially never undergo SN2. Alcohols themselves cannot be electrophiles directly because hydroxide is a terrible leaving group — that is precisely why we convert one alcohol to a halide or tosylate first.

Solvent. Polar aprotic solvents — DMSO, DMF, acetone, acetonitrile, THF — are the right choice. They solvate the sodium or potassium cation strongly but leave the alkoxide anion "naked," with its full nucleophilic punch. In a protic solvent like water or ethanol, hydrogen bonding would cage the alkoxide and slow it dramatically — SN2 rates can rise by factors of 10³–10⁶ on switching from a protic to an aprotic medium. Temperatures are mild, typically 25–80 °C.

Base. Match the base to the alcohol's acidity. NaH and Na metal give clean alkoxides from ordinary alcohols (pK_a ≈ 16–18); KOtBu supplies the bulky tert-butoxide; K₂CO₃ is enough for the far more acidic phenols (pK_a ≈ 10). Using a base much stronger than necessary wastes reagent and can promote elimination.

The competition: substitution versus elimination

The single biggest threat to a Williamson synthesis is that the alkoxide is both a good nucleophile and a strong base. When it cannot easily reach the carbon — because the carbon is hindered or the temperature is high — it instead grabs a β-hydrogen and triggers an E2 elimination, producing an alkene and the alcohol. The two pathways share the same starting materials but diverge at the transition state.

Factor	Favors substitution (ether)	Favors elimination (alkene)
Substrate	Methyl, primary	Secondary, tertiary, β-branched
Nucleophile/base	Small alkoxide (methoxide, ethoxide)	Bulky base (tert-butoxide)
Temperature	Lower (25–50 °C)	Higher (entropy favors making more molecules)
Leaving group position	Few/no β-hydrogens	Many accessible β-hydrogens

This is why the substrate rule is non-negotiable. With a primary, β-unbranched halide and a modestly sized alkoxide at moderate temperature, substitution dominates and the ether is the clean major product.

Concrete reactions and where it shows up

Anisole (methyl phenyl ether). Phenol + NaOH → sodium phenoxide; add iodomethane → anisole. The aromatic ring's oxygen is the nucleophile and the methyl group is the electrophile. This aryl alkyl ether route is run on industrial scale to make fragrance and pharmaceutical building blocks; you cannot put the leaving group on the aromatic carbon, so the aryl piece must always be the alkoxide.

Diethyl ether, the symmetric case. Sodium ethoxide + bromoethane → CH₃CH₂–O–CH₂CH₃. Both halves are primary, so the reaction is fast and clean. Symmetric ethers are the easiest Williamson targets because either disconnection works.

Epoxides by intramolecular ring closure. Treat a halohydrin — a molecule carrying an –OH and a halide on adjacent carbons — with base. Deprotonation gives an alkoxide that immediately performs an intramolecular SN2 on the neighboring C–X carbon, snapping shut a strained three-membered ring. Ring closure to a 3-membered ring is unusually fast because the nucleophile and electrophile are tethered close together (high effective molarity), more than compensating for the ~27 kcal/mol of ring strain that ends up in the product. This is the standard way to make epoxides from chlorohydrins, and it is exactly how industrial ethylene oxide and propylene oxide chemistry is conceptualized.

Crown ethers and macrocycles. Charles Pedersen's Nobel-winning crown ethers — rings of repeating –O–CH₂CH₂– units that trap metal cations — are assembled by Williamson reactions between glycol-derived alkoxides and dihalide chains. The template effect of the captured cation pre-organizes the chain, helping the final macrocyclic SN2 close instead of polymerizing.

Why it endures

Nearly two centuries after Williamson's discovery, this is still the first method an organic chemist reaches for to install an ether. It is general, predictable, stereospecific (clean inversion at the electrophilic carbon), and tolerant of a huge range of functional groups elsewhere in the molecule. The mechanistic constraints — keep the electrophile primary, keep the bulk on the alkoxide, use a good leaving group in an aprotic solvent, and beware elimination — are not limitations so much as a clear recipe. Master those four rules and you can join almost any two fragments through oxygen, exactly where you intend.