Stir diethyl ether with sodium methoxide and nothing happens, even after days. Ethylene oxide is also an ether, yet under the same conditions its ring opens readily and gives 2-methoxyethanol, CH3O–CH2–CH2–OH.
The difference is the ring. An epoxide is an ether whose oxygen sits in a three-membered ring with two carbons, and ethylene oxide is the simplest one. Ether chemistry showed that ordinary ethers hardly react at all. Squeeze the C–O–C unit into a three-membered ring and it becomes one of the most useful electrophiles in synthesis.
This section explains why the ring is so reactive, then how epoxides are named, made and opened. Opening an epoxide puts a new group on one carbon and an OH on the carbon next to it, with predictable stereochemistry. The rules that decide where the new group goes are the SN2 and SN1 rules from Substitution & Elimination, applied to a strained ring.
Ring strain makes epoxides reactive
Look at the ring angles in ethylene oxide. Three atoms joined in a triangle leave each angle close to 60°, yet every ring atom is sp³ and would like its bonds 109.5° apart. The same squeeze makes cyclopropane strained, and Cyclohexanes puts its ring strain at 27.6 kcal/mol. An epoxide carries a similar amount, about 27 kcal/mol. It comes from angle strain and from eclipsing, because the ring holds the C–H bonds on its two carbons in line with each other.
Opening the ring releases that strain, and the energy released drives the reaction forward. That explains the two flasks at the top of the page. In both ethers the leaving group would be an alkoxide, which is a poor leaving group. With diethyl ether nothing pays that cost. With ethylene oxide the released strain does.
Strain also makes the reaction faster. The orbitals that each ring atom uses for its ring bonds point much further apart than 60°, so they cannot point straight at each other. They overlap off to one side of the line between the two nuclei, and the electron density of each ring bond bulges outside the triangle. These are bent bonds, or banana bonds, the same kind cyclopropane has. Overlap off the line is weaker overlap, so the ring C–O bonds are weaker than ordinary ones and break more easily. Strain makes the product more favorable, and it also lowers the barrier on the way there.
Naming them
Treat 2-methylpropene with mCPBA and you get an epoxide with both methyl groups on one carbon. Its name is 2,2-dimethyloxirane.
Oxirane is the name of the ring itself. Its numbering starts at the oxygen, so the oxygen is atom 1 and the two carbons are C2 and C3. Substituents then take the lowest numbers, as on any ring, so here both methyl groups go on C2.
A second style names the oxygen as a bridge across a chain. The prefix epoxy carries the numbers of the two carbons the oxygen joins, and it goes in alphabetical order with the other prefixes. So the epoxide of 2-methylbut-2-ene is 2,3-epoxy-2-methylbutane, with e before m. As an oxirane it is 2,2,3-trimethyloxirane.
The two simplest epoxides also keep older common names. Oxirane itself is usually called ethylene oxide, and 2-methyloxirane is called propylene oxide.
Nothing chemical depends on which name you use. The numbering matters in one way, though. In an oxirane, atom 1 is the oxygen, so an answer that puts a nucleophile on C1 describes an attack on the oxygen, which does not happen.
Making an epoxide
From an alkene and a peroxyacid. This is the route from Oxidative cleavage & dihydroxylation. A peroxyacid such as mCPBA (meta-chloroperoxybenzoic acid) hands one oxygen to the C=C in a single concerted step, through the butterfly transition state drawn there. Both new C–O bonds form on the same face of the alkene, so the addition is syn. The reaction is also stereospecific: the alkene’s geometry carries straight into the ring. cis-But-2-ene gives the cis epoxide, and trans-but-2-ene gives the trans epoxide.
From a halohydrin and a base. A halohydrin has a halogen and an OH on neighboring carbons, and Addition reactions makes one from an alkene with Br2 in water. Start from trans-but-2-ene. Bromine and water add anti to give 3-bromobutan-2-ol, and NaOH then closes the ring to trans-2,3-dimethyloxirane.
The ring closure is an SN2 reaction inside one molecule. NaOH first removes the proton from the OH, which leaves an alkoxide, O−. That oxygen then attacks the carbon that holds the bromine, and bromide leaves. Like any SN2 it needs a backside attack, so the O− must reach that carbon from the side opposite the C–Br bond.
That requirement means the C–O− and C–Br bonds must point in opposite directions, anti, across the two carbons. An open chain simply turns about its C–C bond until they do, as in the figure below. A cyclohexane ring cannot turn freely, so there the O− and the Br must both be axial, one up and one down, which only a trans halohydrin can manage. Bromine and water add anti, so the halohydrin they give on a ring is trans, and it closes.
The two routes give the same epoxide. The anti addition and the backside ring closure cancel each other out, so trans-but-2-ene gives the trans epoxide here, just as it does with mCPBA.
Base-catalyzed opening: SN2 at the less hindered carbon
Start with 2-methyloxirane and sodium methoxide in methanol. Methoxide bonds to the CH2 carbon of the ring, and the product is 1-methoxypropan-2-ol.
The opening is a single SN2 step. Methoxide approaches the CH2 carbon from the back, opposite that carbon’s bond to the ring oxygen, and that C–O bond breaks as the new C–O bond forms. The attacked carbon is inverted, as in any SN2. The ring oxygen leaves as an alkoxide that is still attached to the other carbon, and it picks up a proton from the methanol solvent.
Why the CH2? Nothing has made either ring carbon more positive than the other, so only sterics decide. The nucleophile reaches the carbon with fewer groups around it more easily, just as SN2 prefers a primary carbon to a secondary one.
Any strong nucleophile opens an epoxide this way: hydroxide, alkoxides, cyanide, acetylide ions, amines and thiolates (RS−). Hydroxide gives a 1,2-diol, a compound with OH groups on two neighboring carbons. Every other nucleophile gives the new group on one carbon and an OH on the next.
When there is no methanol or water to supply a proton, the alkoxide stays an alkoxide until the workup. That is the separate last step of a procedure, where you add something, usually dilute aqueous acid, to neutralize charged species and get a neutral product.
With a carbon nucleophile the opening also makes a new C–C bond. Ethylene oxide and the acetylide ion of propyne, followed by workup, give pent-3-yn-1-ol. The chain has grown by exactly two carbons, with the OH on the far one. So when a chain needs two more carbons ending in an OH, ethylene oxide is the reagent to reach for.
Preview. Two reagents from later chapters open epoxides the same way. Hydride from LiAlH4 gives an alcohol (Oxidation & Reduction). A Grignard reagent attacks the less hindered carbon by SN2 from the back, and after workup the OH sits on the carbon next to the one the reagent bonded to (Organometallic Bonding & Grignard Reagents). Both are taught in those chapters, and nothing on this page depends on them.
Acid-catalyzed opening: attack shifts to the more substituted carbon
Now put 2,2-dimethyloxirane in methanol with a little sulfuric acid. Methanol bonds to the carbon that carries the two methyl groups, the more hindered one, and the product is 2-methoxy-2-methylpropan-1-ol. Sodium methoxide would have attacked the CH2 instead and given 1-methoxy-2-methylpropan-2-ol.
The acid acts on the ring oxygen first and adds a proton to it. The leaving group is now a neutral OH instead of an alkoxide, the same trick that activates an alcohol in Alcohol reactions. The protonated ring is a much stronger electrophile, strong enough for a weak nucleophile such as methanol or water.
Protonation weakens both ring C–O bonds, but not equally. The bond to the more substituted carbon stretches further, because that carbon carries positive charge best (3° > 2° > 1°, as in Carbocations). In the transition state it holds a large partial positive charge, as if it were on its way to becoming a carbocation, and the nucleophile goes where the positive charge is.
That C–O bond has not fully broken when the nucleophile arrives, so no free carbocation forms. The oxygen still covers one face of the carbon, and the nucleophile still attacks from the back. Only the choice of carbon changes. A last proton transfer to the solvent then gives the neutral product.
How strongly acid favors the more substituted carbon depends on how much positive charge that carbon can hold. With a tertiary carbon on one side, or a benzylic one (a carbon attached to a benzene ring), the preference is strong and one product forms. Between a secondary and a primary carbon it is weaker, and mixtures can form. Between two similar carbons, sterics can still win. Exam questions usually give one clearly more substituted carbon, so the rule holds there. Real molecules can fall in between.
Same substrate, opposite regiochemistry
2,3-Epoxy-2-methylbutane (2,2,3-trimethyloxirane). One ring carbon carries two methyl groups; the other carries one methyl group and an H.
With NaOCH₃ in methanol: basic conditions, so sterics decide. Methoxide attacks the less substituted carbon, the one with the H. The product is 3-methoxy-2-methylbutan-2-ol, with the methoxy group on C3 and the OH on C2.
With CH₃OH and a little H₂SO₄: acidic conditions, so the positive charge in the protonated ring builds up on the more substituted carbon, and methanol attacks there. The product is 3-methoxy-3-methylbutan-2-ol. The methoxy group and the OH have swapped carbons.
The same epoxide gets the same methoxy group at opposite ends of the ring, and only the conditions decide which end. It is the same kind of choice that hydration and hydroboration–oxidation gave for an alkene in Markovnikov / anti-Markovnikov. Each product has just one stereocenter, so these two cannot show that the opening was anti. The attack is still from the back in both; the next section uses a ring where the result can be seen.
| Basic / neutral | Acidic | |
|---|---|---|
| Nucleophile | strong, anionic | weak, often solvent |
| Attacks | less substituted C | more substituted C |
| Controlled by | sterics | cation stability |
| Stereochemistry | anti | anti |
Both openings are anti
Cyclohexene oxide shows the stereochemistry. Open it with water and acid, or with hydroxide, and the product is trans-cyclohexane-1,2-diol. The cis diol does not form.
The reason is the backside attack. The ring oxygen sits on one face of the two carbons, so the nucleophile has to arrive from the other face. The new group and the old oxygen end up on opposite faces, which makes the opening anti. Oxidative cleavage & dihydroxylation used this to turn alkenes into anti diols, and the bromonium ion in Addition reactions forces anti addition for the same reason: a three-membered ring on one face blocks that face. Acid changes which carbon is attacked, but not which face.
On a cyclohexane ring there is one more detail. The nucleophile must line up with the C–O bond it breaks, and on a chair that line is axial. So the diol first forms with both OH groups axial, one up and one down, an arrangement called trans-diaxial. A ring flip then moves both groups to the roomier equatorial positions. A flip changes the shape of the ring, not which face each group is on, so the diol stays trans.
Why this matters beyond the exam
Epoxides are electrophilic enough to react with the nucleophilic nitrogen and oxygen atoms in DNA. They alkylate DNA, which means they attach carbon groups to it, and that is why many epoxides are mutagens: they cause changes in DNA.
The body makes some of these itself. Liver enzymes add oxygen to aromatic rings to make them water-soluble enough to excrete. The first product is often an arene oxide, an epoxide across two carbons of the ring, and it is an electrophile looking for a nucleophile. This is how the body’s own handling of aromatic compounds can lead to cancer.
The same reaction is also put to use. Several anticancer drugs are designed to alkylate DNA on purpose, which damages fast-dividing cancer cells.
One more name is worth recognizing: the Sharpless asymmetric epoxidation. It works on an allylic alcohol, an alcohol whose OH carbon sits next to a C=C. The reagents are a titanium alkoxide, one enantiomer of diethyl tartrate and tert-butyl hydroperoxide. It delivers the oxygen to one chosen face of the C=C and so makes a single enantiomer of the epoxide from an achiral alkene. K. Barry Sharpless shared the 2001 Nobel Prize in Chemistry for this work. Treat it as a label; the machinery behind it is beyond this course.
What carries forward
Opening an epoxide with a carbon nucleophile is a standard way to lengthen a chain in synthesis, and the acid/base switch in regiochemistry is a favorite exam question. This section also closes Alcohols, Ethers & Related Chemistry with a general point: reactivity often depends on strain and activation as much as on the functional group. The same C–O bond is inert in diethyl ether and eager to react in ethylene oxide.