Most amide bonds in a medicinal chemistry route are made with a carbodiimide or a uronium salt. But for a substantial class of transformations – ureas, carbamates, cyclic carbonates, oxazolidinones, and many peptide couplings – the reagent of choice is 1,1′-carbonyldiimidazole (CDI, CAS 530-62-1), because it leaves behind nothing that has to be filtered, extracted or chromatographed.
This guide is about the practical chemistry: how the acyl imidazole forms, how to run the two-step coupling without racemising the stereocentre, what the by-products mean for work-up, and where CDI beats carbodiimides and where it does not.
The two-step mechanism
CDI does not couple an acid to an amine directly. It runs in two distinct stages, and understanding that is the key to running it well.
- Activation. The carboxylic acid attacks CDI, displacing one imidazole and releasing CO2, to give an acyl imidazole – a reactive intermediate that is stable enough to accumulate in solution.
- Aminolysis. The amine attacks the acyl imidazole, displacing the second imidazole and giving the amide.
Both steps are fast at ambient temperature, but they have different requirements. Activation needs dry conditions and benefits from a brief hold (typically 15-60 minutes at 0-25 deg C) so that all the acid is converted. Aminolysis needs the amine present and, for hindered systems, a little heat.
| Stage | Reagent | By-product | Practical consequence |
|---|---|---|---|
| Activation | Carboxylic acid + CDI | Imidazole + CO2 | Vent the vessel; CO2 evolution is immediate |
| Aminolysis | Acyl imidazole + amine | Imidazole | Water-soluble, removed in the acidic wash |
Why the by-products matter more than the yield
Amide couplings usually work. What decides whether a route scales is what you have to do afterwards.
| Reagent | Main by-product | Removal | Scale implication |
|---|---|---|---|
| CDI | Imidazole, CO2 | Aqueous wash; CO2 leaves as gas | Minimal – often no chromatography |
| DCC | Dicyclohexylurea (DCU) | Filtration (partially soluble), chromatography | Significant – DCU is hard to remove completely |
| EDC | Water-soluble urea | Aqueous wash | Moderate – good, but EDC is costly at scale |
| HBTU / HATU | Tetramethylurea, guanidinium salts | Chromatography or crystallisation | Significant – and HATU has a cost and safety history |
This is the single strongest argument for CDI in process chemistry: imidazole is water-soluble and volatile enough to be chased off in drying, and CO2 removes itself. A CDI coupling frequently ends with a phase cut, an acid wash and a crystallisation.
Racemisation: the real risk, and how to control it
For peptide couplings the question is not yield but stereochemical integrity. Acyl imidazoles can racemise through oxazolone formation, and the rate depends on temperature, solvent and the absence of an additive.
| Factor | Effect on racemisation | Control |
|---|---|---|
| Temperature during activation | Higher temperature increases oxazolone formation | Activate at 0-10 deg C, warm only after the amine is in |
| Additive | HOBt or HOAt suppress oxazolone formation | Add 0.1-1.0 equivalent of HOBt for sensitive residues |
| Base | Strong base accelerates enolisation | Use the minimum of tertiary amine; avoid excess DIPEA |
| Solvent | Polar aprotic solvents favour clean acyl imidazole formation | DMF, NMP, THF or DCM depending on solubility |
| Reaction time | Long activation holds increase exposure | Complete activation, then add the amine without delay |
For non-peptide amides – where the acid has no alpha-stereocentre, or racemisation is irrelevant – none of this matters and CDI can be run hot and fast.
Practical procedure outline
| Parameter | Typical practice | Notes |
|---|---|---|
| Solvent | DMF, NMP, THF or DCM | Dry; CDI and water compete for the acid |
| CDI equivalents | 1.05 – 1.3 on the acid | More if the acid is hindered or the CDI assay is at the low end |
| Activation | 0-25 deg C, 15-60 min | Vent for CO2; confirm by HPLC/TLC that the acid is consumed |
| Amine charge | 1.0 – 1.2 equivalents | Added after activation, not together |
| Coupling | 20-60 deg C, 1-16 h | Hindered amines need the higher end |
| Work-up | Dilute, wash with dilute acid (removes imidazole), then base wash | Imidazole is water-soluble; the amide usually crystallises |
Ureas, carbamates and carbonates
The same reagent builds the carbonyl linkages that carbodiimides cannot:
- Symmetrical ureas – two equivalents of amine with one of CDI, often in one pot.
- Unsymmetrical ureas – amine first to give the carbamoyl imidazole, then a different amine.
- Carbamates – an alcohol or amine plus CDI, then the second partner.
- Cyclic carbonates and oxazolidinones – intramolecular versions with a diol or amino alcohol.
These transformations are the reason CDI is common in pharmaceutical process patents: the carbonyl source is a solid, easy to charge, and its by-products are benign.
Compatibility and limitations
| Situation | CDI suitability | Alternative |
|---|---|---|
| Solution-phase peptide coupling | Good, with HOBt for sensitive residues | EDC/HOBt or HBTU |
| Solid-phase peptide synthesis | Poor – heterogeneous conditions favour other reagents | HBTU, HATU, DIC/HOBt |
| Hindered secondary amines | Moderate – needs heat and excess | HATU, or acid chloride |
| Water-containing or aqueous media | Not suitable – CDI hydrolyses | EDC/NHS in aqueous buffer |
| Urea / carbamate / carbonate targets | Best in class | Phosgene equivalents (hazardous) |
| Large-scale amide formation | Good – simple work-up | EDC, or enzymatic coupling |
Frequently asked questions
Can CDI and the amine be charged together? It works, but you lose control: the acid, CDI and amine compete, and carbamoyl imidazole or urea by-products can form. Activation first, then the amine, is the reliable order.
Why did my coupling stall? Usually moisture. CDI hydrolyses to imidazole and CO2, so a wet solvent or a wet lot consumes the reagent before it activates the acid.
Do I need a base? Often a small amount of tertiary amine helps, but excess base promotes racemisation in sensitive systems. Imidazole released in the reaction already provides some buffering.
How do I remove imidazole? A dilute acid wash takes it into the aqueous phase. If the product is water-soluble, use an acidic ion-exchange or a crystallisation instead.
Does CDI work with alcohols? Yes – the acyl imidazole reacts with alcohols to give esters, though less readily than with amines. Acylation catalysts such as DMAP can help.
Contact PolyBlueChem for CDI samples, COA and coupling support.