A saffron lab report has three numbers: crocin (colour), picrocrocin (taste) and safranal (aroma). Good saffron measures roughly 250–280 crocin, 95–110 picrocrocin, and 30–40 safranal. Higher is better for the first two. For safranal — counter-intuitively — higher is not better, and above 50 it is a defect.
Almost every premium saffron brand mentions these three words somewhere. Very few explain what the numbers mean, and the one that is most often used as a selling point is the one you should read most sceptically. Here is how to read a report properly.
What do crocin, picrocrocin and safranal actually measure?
All three are measured the same way. A ground sample is dissolved in water, light is passed through the solution, and a spectrophotometer records how much light the solution absorbs at three specific wavelengths under ISO 3632-2.
| Compound | Wavelength | What it governs | Good range |
|---|---|---|---|
| Crocin | 440 nm | Colouring power — how much gold one gram delivers | 250–280 |
| Picrocrocin | 257 nm | The bitter, hay-sweet taste | 95–110 |
| Safranal | 330 nm | Aroma — but see the warning below | 30–40 |
What does a number like 270 mean?
It is an absorbance value, written formally as E1%1cm. It describes how strongly a one-gram-per-100-millilitre solution absorbs light in a one-centimetre cell.
Three things follow from that, and they are worth understanding:
It is not a percentage. Crocin 270 does not mean 27% crocin. People routinely read it that way and it is wrong.
It is proportional. Because it is normalised to a fixed weight, the numbers are directly comparable. Crocin 260 really is twice the colouring power per gram of crocin 130. You would need half as much saffron for the same colour.
It is corrected for moisture. Results are reported on a dry basis, which means a seller cannot inflate the number by drying the product harder. Two identical saffrons at 4% and 11% moisture report the same value. This is a genuinely well-designed part of the standard.
Why is high safranal a bad sign?
This is the part almost nobody explains, and it inverts how the number is usually sold.
ISO 3632 does not set safranal as a minimum to be exceeded. It sets it as a range — roughly 20 to 50 — and the same range applies to every grade. It is not a quality-discriminating measure at all. That is unusual enough to be worth asking why.
The reason is chemical. Safranal is not primarily something the flower produces. It is largely a breakdown product of picrocrocin, created by heat and by hydrolysis during drying and afterwards during storage. As saffron is over-dried or simply ages, picrocrocin converts into safranal.
So a rising safranal number is not a signal of more aroma. It is a signal of conversion. High safranal alongside low picrocrocin means the flavour precursor has already been cooked out of the spice. In one Greek study of a full annual harvest, samples sat comfortably within specification when fresh — but after a year in storage, nearly a quarter had drifted past 45 units, heading out the top of the ISO band.
The pairing to look for is high picrocrocin with mid-range safranal. That combination is the signature of gentle processing and a recent harvest. A brand advertising "exceptionally high safranal" as a mark of superiority is, in standards terms, advertising a defect.
Can you trust the safranal number at all?
The 330 nanometre reading is the weakest measurement in the standard, and it is fair to know this before you weigh it heavily.
Crocin also absorbs light at 330 nm. The safranal reading is therefore contaminated by the very compound being measured at 440 nm, along with other breakdown products. Published comparisons against chromatography — a far more precise method — have found the spectrophotometric reading overstating true safranal content by an order of magnitude or more.
The practical consequence is that saffron sometimes fails ISO on aroma alone while being perfectly good spice. In one multi-method study, most of the samples that could not be classified had entirely adequate colour and taste and failed only on this one reading. Treat crocin and picrocrocin as the reliable numbers. Treat safranal as directional.
What else should be on the report?
The three headline numbers are the ones brands quote. The rest of the report is where adulteration actually shows up.
- Moisture. ISO permits up to 12%. Well-handled saffron lands between 6% and 9%. High moisture is both weight you are paying for and a shorter shelf life.
- Acid-insoluble ash. This is the mineral test, and it is the one that catches weight fraud. Sand, barium sulphate and similar loading agents show up here. Category I permits a maximum of 1.0%; a clean result is well under half that.
- Total ash. Maximum 8% under ISO. A broad measure of inorganic content.
- Synthetic dyes. Not part of the core ISO panel, so it must be requested. This matters: an EU regulatory survey of 141 saffron samples found twelve containing non-authorised synthetic dyes, some of them carcinogens. A report that says "absent" here is doing real work.
- Extraneous matter. Petals, stamens, broken style. Extra class allows 0.25%; Category I allows 0.5%.
What is missing from nearly every saffron certificate?
Something worth knowing as a buyer: the standard panel tests potency and basic cleanliness. It does not test for contaminants at all.
Heavy metals, pesticide residues, aflatoxin and microbiological safety are separate tests that have to be commissioned separately. As far as we can establish, no saffron brand selling to United States consumers currently publishes measured results for any of them. Not one — including the brands that publish genuine potency certificates.
For reference, the American Spice Trade Association places saffron in the "flower pistil spices" category with a guidance level of 1.0 ppm for lead. If you want to ask a seller a question they almost certainly cannot answer, ask for a measured lead result against that figure.
How do you check a certificate in sixty seconds?
When you are handed a saffron certificate, run it through five questions:
- Is the laboratory named? "Lab tested" is not a laboratory.
- Is there a batch or lot number, and does it match the product you are buying?
- Is there a sampling date? Saffron degrades; an undated certificate has no shelf life.
- Are all three numbers shown, or only the flattering one?
- Is the picrocrocin figure plausible? A useful trick: if a report lists picrocrocin as "257," that is the wavelength, not a result. It happens more often than you would expect.
Where we stand on this ourselves is on our lab report page, stated plainly, including what has not been done yet. Related reading: how to spot fake saffron and why Super Negin is not a lab grade.
