What a coupling reagent does

A carboxylic acid and an amine do not spontaneously form an amide at any useful rate; mix them and they just pair up as a salt. They need a coupling reagent to react, which works in three stages. First it activates the acid — a carbodiimide to an O-acylisourea, an onium salt to an acyluronium or acyloxyphosphonium. Then an additive (HOBt, HOAt, Oxyma, NHS) traps that reactive intermediate as a milder active ester — the stereochemistry-protecting step, which reacts cleanly with the amine while holding the side reactions back; in the onium reagents the additive is built in as the leaving group, so one reagent does both. Finally the amine attacks and the amide forms. Which cation carries which leaving group is mapped out with the reagent families below.

The two failure pathways. A stereocentre-bearing activated acid can cyclise to an oxazolone, which racemises and then couples with scrambled stereochemistry — the origin of most epimerisation, and a risk under any activation method. Separately, a carbodiimide's O-acylisourea can rearrange to an unreactive N-acylurea, consuming the acid and capping the yield. The two are different kinds of failure: the oxazolone loses stereochemistry, while the N-acylurea — and, for uronium and aminium salts, capping of the amine as a guanidino derivative — loses yield. Those, with a coupling that is simply slow, are addressed by the small set of levers in the map below.

Treat the scheme as a first pass, not a rule: base, solvent, water content, pre-activation time and the amine's own nucleophilicity can decide a coupling as much as the choice of reagent does.

Carbodiimides: the foundational activators

The carbodiimides are where coupling chemistry began and they remain the economical backbone of large-scale and routine work. All three common members activate the acid the same way — through the O-acylisourea — and differ mainly in the by-product they leave behind.

The essential rule with any carbodiimide: on its own it both racemises a stereocentre-bearing acid and loses material to N-acylurea, so for peptide work it is always combined with an active-ester additive. The additives in the next section are precisely the partners a carbodiimide such as DIC or EDC is run with.

Additives: why you never use a carbodiimide alone

An additive converts the carbodiimide's reactive intermediate into a defined active ester, which is what suppresses racemisation and N-acylurea formation. The historical benchmarks are the benzotriazoles.

HOBt (1-hydroxybenzotriazole), introduced by König and Geiger, was the first widely adopted additive and remains the reference point. HOAt (1-hydroxy-7-azabenzotriazole), introduced by Carpino, places a nitrogen at the 7-position; that nitrogen provides a neighbouring-group effect that accelerates the coupling and suppresses racemisation more effectively than HOBt. Both are excellent — and both carry a problem.

The benzotriazoles are also energetic compounds: HOBt, HOAt and the uronium and aminium salts derived from them carry recognised explosion and thermal-decomposition hazards. That hazard was the explicit motivation for the non-benzotriazole alternatives below; what it means for handling and shipping is covered under Hazard profile and storage.

Uronium and aminium salts: the solid-phase workhorses

These reagents do activation and active-ester formation in a single step: add the reagent and a base to the acid, and the active ester is generated in situ. They are the default for routine SPPS and for much of solution-phase medicinal-chemistry coupling because they are fast and convenient.

A naming point worth knowing: although they are universally drawn as uronium (O-bound) species, crystallographic studies show that the common reagents in fact exist as the aminium (guanidinium N-oxide, N-bound) form in the solid state. The "uronium" label is historical; the distinction matters when you are recording an accurate structure, identifier or InChIKey for one of these reagents.

One side reaction defines how you use them. Used in excess, or left to activate too long before the amine is added, a uronium/aminium salt can cap the amine terminus as an unreactive guanidino derivative. The practical guard is to keep the reagent close to stoichiometric, keep the base modest, and pre-activate only briefly. The members differ mainly in their built-in leaving group:

Oxyma-based "third-generation" reagents

These reagents take the one-pot convenience of the onium salts and the lower-hazard profile of Oxyma, combining them in a single reagent. They are the safety-forward route to HATU-class performance.

Classic phosphonium salts

The phosphonium reagents activate the acid as an acyloxyphosphonium species and, like all phosphonium reagents, do not guanidinylate the amine. The family has an instructive safety history.

BOP, the original benzotriazolyloxy-phosphonium reagent, is highly effective but releases hexamethylphosphoramide (HMPA) — a carcinogen — as a by-product. PyBOP was developed by Coste, Le-Nguyen and Castro to solve exactly that: replacing the dimethylamino groups with pyrrolidino gives equivalent coupling performance and low racemisation, with a non-carcinogenic tris(pyrrolidino)phosphine oxide by-product in place of HMPA. PyBOP is a strong choice for macrolactamisation and head-to-tail cyclisation, and for hindered couplings where the absence of amine capping is an advantage. For the most demanding ring closures, PyAOP — the HOAt-based member of the same family — keeps that advantage while raising activation to HOAt level, giving faster, higher-yielding couplings than PyBOP on difficult substrates, at a premium price.

Those three onium families — uronium and aminium, Oxyma-based, and phosphonium — are easier to hold together as one picture. Each reagent is simply an activating cation paired with a leaving group that becomes the active ester:

Carbonyl-transfer reagents

For general amide, ester, carbamate and urea formation — particularly in small-molecule synthesis rather than peptides — the azole carbonyl-transfer reagents offer a mild, often aqueous-workup-friendly route.

CDI (1,1′-carbonyldiimidazole) reacts with a carboxylic acid to give an acyl-imidazole that an amine or alcohol then displaces, with carbon dioxide and imidazole as the only by-products — a clean, non-phosgene way to make an amide or ester. CDT (1,1′-carbonyldi(1,2,4-triazole)) is a more reactive analogue: 1,2,4-triazole is a weaker base than imidazole and therefore a better leaving group, so CDT is used to activate less reactive substrates while keeping the same mild, phosgene-free profile.

DMAP: the acyl-transfer catalyst

DMAP (4-dimethylaminopyridine) is not a stand-alone coupling reagent but a nucleophilic catalyst that you add to one. It is the classic partner for carbodiimide-mediated esterification (the Steglich esterification) and it accelerates difficult or hindered acylations by forming a highly reactive acylpyridinium intermediate. The caveat for peptide work: racemisation at the activated residue is base-catalysed, proceeding largely through the oxazolone, and DMAP — a strong nucleophilic base — can accelerate it. Use it catalytically, and avoid it in stereochemically sensitive segment couplings where racemisation is the main risk.

Buying and handling: practical notes

A few things that are obvious to specialists but trip up people new to these reagents.

Hazard profile and storage

Coupling reagents span a wide range of handling demands, and getting this wrong wastes reagent or creates a genuine hazard. Roughly:

Quick-reference selection summary

The honest logic, as with any reagent class, is robustness first, then escalate for difficulty — and the right answer still depends on the acid, the amine, the base, the solvent and the temperature, not the acronym alone.

The same logic, as a decision tree

If you would rather follow a path than scan a grid, the same recommendations route as a decision tree. Start at the top and take the first branch that fits — the special cases (ester, bioconjugate) peel off first, then the main spine routes an amide coupling by ring closure, difficulty, and phase.

© 2026 SynFinn Discovery Oy / NorrChemica · All rights reserved · Content may not be reproduced without written permission