How to tell a real saffron extract from a fake one

The test for a real saffron extract is picrocrocin. It is the compound that makes saffron taste bitter, it is present in the stigmas of Crocus sativus and in almost nothing else, and the flower most often substituted for saffron does not contain it at all. If an extract has no picrocrocin in it, it is not saffron.

That is a laboratory test, not a kitchen one, but you do not need a laboratory. You need to know that the test exists and to ask whether it has been run.

Why saffron gets faked

Saffron is the dried stigma of a crocus flower. Each flower produces three of them, and they are picked by hand. It takes somewhere between 90,000 and 170,000 flowers to make a single kilogram of dried stigmas, which is why saffron is the most expensive spice in the world by weight.

Anything that expensive attracts substitution. In the spice aisle that means safflower, turmeric or dyed plant fibre sold as threads, and an experienced cook can often spot it. In the supplement aisle the saffron has already been ground and extracted into a powder, so there is nothing left to look at. The powder is anonymous, and that is the point.

Colour proves nothing

Three compounds do most of the work in saffron. Crocins make it red. Safranal gives it the aroma. Picrocrocin gives it the bitter taste.

The trouble with judging an extract by its colour is that gardenia fruit also contains crocins. A gardenia extract is a deep orange powder, it looks like saffron, it dissolves like saffron, it will pass any test based on colour, and it costs a fraction of the price. What gardenia does not contain is picrocrocin.

So picrocrocin is the compound that settles the question. Crocins tell you something orange is present. Picrocrocin tells you it came from a crocus.

What happened when someone checked

In 2023 a team of analytical chemists published a study in the journal Foods in which they measured the crocins and the picrocrocin in commercial saffron extracts by high performance liquid chromatography, alongside raw saffron stigmas for comparison. (Mena-Garcia et al., Foods, 2023;12:1634.)

Raw stigmas measured 89 to 119.9 milligrams of picrocrocin per gram. The commercial saffron extracts measured 20.5 to 39.7 milligrams per gram, which is the range you would expect once stigmas have been extracted and standardised. Batches of one extract, affron, measured 30.7 to 35.3 milligrams per gram.

Two of the commercial products did not fit the pattern. In one product sold as a saffron extract, no picrocrocin was detected at all. It was gardenia. A second had been mixed with ground stigma rather than being the pure extract it was sold as, which the picrocrocin figure also exposed.

Both of those products would have passed a look, a smell and a colour check. Neither passed the chemistry.

Four questions worth asking

• What compound is it standardised to, and at what percentage? A real answer names a marker and a number. No answer, or a vague reference to potency, means there is nothing being measured.
• Is the assay run on every production batch, or once? These are very different promises. Per-batch testing is the one that means the material in front of you was checked.
• Can I see a certificate of analysis for this batch? A specification sheet is the promise a manufacturer makes about every batch. A certificate of analysis is the measured result for one batch. The second is the one that proves anything.
• Where was it grown and where was it extracted? A supplier who can name a region and a facility is a supplier with a supply chain. A supplier who cannot is buying powder on an open market.

What to take from it

Saffron is expensive enough that the incentive to fake it will never go away, and once it is a powder you cannot see the difference. The good news is that the analysis that catches it is routine, cheap for a manufacturer to run, and published.

Ask which marker, ask how often, and ask for the certificate. A brand doing the work will have the answers ready. A brand that does not is telling you something too.

Studies conducted on individual ingredients, not on any finished product. Individual results may vary.