Lab reports·7 min read

NMR vs HPLC Olive Oil Testing: The Case Against NMR

The short answer

The case against NMR is that it reads higher than HPLC and is not the International Olive Council's method. Part is true: the IOC's declared methods are HPLC, and NMR numbers must not be compared with HPLC numbers. But methods count differently, and HPLC methods also disagree with each other.

By Remy Assir ·

What is the case against NMR?

A common argument, made by some sellers and labs that test by HPLC, runs like this: NMR “can read 2–3× higher”, is “not recognized by the IOC”, and is a tool for research, not for the number on a label.

It is a fair worry. If you chose an oil because its label said HPLC, and then see a bigger number measured by NMR on someone else's label, you want to know whether that number means what it seems to mean. The argument deserves a straight answer, and part of the straight answer is that the critics are right about some things.

We make olive oil ourselves, so we are not a neutral party. That is why everything below is cited to published research or to the International Olive Council's own document, and why we start with what the argument gets right.

Where are the critics right?

On two points, in full: NMR and HPLC numbers differ and must never be compared as if they sat on one scale, and the International Olive Council's declared methods for phenolic compounds are HPLC methods.

The IOC's document COI/T.20/Doc. No 29/Rev.2 (June 2022) declares two methods for phenolic compounds, and both are HPLC. The pages of it we read do not mention NMR at all. If your rule is to trust the methods the IOC has declared, that rule points to HPLC, and we would not argue with it.

On the first point, the most careful source we know of is a 2019 cross-laboratory study by Tsimidou and colleagues, nine authors from institutions including the Instituto de la Grasa (CSIC) and the universities of Perugia, Barcelona and Bologna, published in Molecules. They analysed the same oils by five liquid-chromatography protocols and by ¹H NMR, in five labs, and the absolute values differed. The authors favour their own UHPLC protocol and stress correction factors in the calculation.

That supports the critics' first point: one method's number is not another method's number. A seller who sets an NMR figure against an HPLC figure is comparing two different scales, and that applies to us as much as to anyone.

Does NMR read higher, and why?

NMR figures can differ from HPLC and Folin figures for the same oil, because each method counts against a different reference and some compounds change during chromatography, so a higher figure can mean counted differently rather than overstated.

Start with the reference compound. A Folin total is expressed as caffeic acid. The IOC's first HPLC method expresses its result as tyrosol. NMR reports the named compounds themselves. Each is a different yardstick, so the same oil returns a different number on each, and none converts into another.

The IOC is candid about that first method. Its document says Method 1 “quantifies using tyrosol and therefore may underestimate the true value”, gives it a range of 30 to 800 mg/kg, and says it “may be used to perform a quick determination”. The newer IOC method, SPE-HPLC-DAD, is the one it says “can be used to determine the real concentration of phenols to fulfil an EFSA claim”.

A 2014 University of Athens paper in the Journal of Agricultural and Food Chemistry adds a second reason. It showed that contact with a chromatography stationary phase artificially forms some aldehyde isomers of oleuropein and ligstroside aglycon. In plain terms, some compounds can change form during chromatography, which NMR avoids by measuring the extract directly. That does not make HPLC wrong. It is one more reason the two methods report different numbers.

Set side by side, the three methods start from different reference compounds, count different things and state different limits, which is why one oil's numbers cannot be swapped between them.

Folin IOC HPLC (Method 1) NMR
Reference compound
Caffeic acid Tyrosol The compounds themselves
What is counted
One total, by colour A total, read as tyrosol Named compounds, one by one
Range or limit stated
None in our source 30 to 800 mg/kg None in our sources
What the source says of its limits
One total only “May underestimate the true value” No limit in the abstracts read

The table holds no oil's figures, on purpose: the point is the scales, not any one result.

Do HPLC methods agree with each other?

Not always: in a 2019 cross-laboratory study, five liquid-chromatography protocols gave absolute values that differed, and the IOC's own collaborative tests found differences between its two HPLC methods.

Against the study's UHPLC reference, the other protocols showed “slight or statistically significant differences” (Tsimidou and colleagues, Molecules, 2019). The IOC's document says of its collaborative tests that “differences were observed when comparing the results obtained by the different methods”.

None of that makes HPLC unreliable. It means that “HPLC tested” on a label names a family of methods rather than one scale, and two HPLC numbers made by different protocols, or in different labs, deserve the same care as an HPLC number beside an NMR one.

What about “2–3× higher”?

We searched on 14 November 2026 and found no published study behind the “2–3× higher” figure; the page where we saw it cites none, and the published cross-method comparison we found reports differences, not a multiple.

That does not prove the figure wrong. It means nobody reading it can check it: there is no study, no method detail and no oils named. A number without its method is the kind of number this site does not print about anyone's oil, ours included.

So we will not offer a multiple of our own. The published comparison reports differences, and the reasons on record are the reference compound, the method and the lab. A single factor would hide all three.

Is NMR published and validated?

Yes: quantitative ¹H NMR for olive oil polyphenols is published in peer-reviewed journals, was validated by its University of Athens authors in 2014, and is run for certificates at the National and Kapodistrian University of Athens.

Karkoula, Skantzari, Melliou and Magiatis set it out in the Journal of Agricultural and Food Chemistry in 2012, measuring oleocanthal and oleacein directly in 175 Greek and Californian single-variety oils. The 2014 paper validated its extension across 340 oils from 23 varieties over three years, and a 2020 paper in Molecules used it to follow phenols through harvest and malaxation.

Certificates from the World Olive Center for Health state that the analysis follows “the method that has been submitted to EFET and published in” those papers; EFET is the Greek food authority. So the fair version of “NMR is for research” is that NMR is a published research method that is also used for certificates. What it is not is one of the IOC's declared methods. For the case in NMR's own terms, read why NMR is a valid way to test olive oil polyphenols.

How do you compare two oils fairly?

Compare numbers made by the same method, ideally in the same laboratory, from samples taken close together, with the method and report number printed beside every figure.

If one oil reports by HPLC and the other by NMR, the fair comparison is not between their numbers; it is between how open each is about its method, its lab and its report. For what each line on a report means, here is how to read an olive oil lab report, and for what a polyphenol total tells you in the first place, what a polyphenol count actually means.

Here is what we do. Every batch is tested by an independent laboratory, and we publish the results — the laboratory's name and its report number included. The figures themselves are on the batch report for each pressing.

Who is this page not for?

Common questions

Does the IOC say HPLC can underestimate polyphenols?
For its first method, yes. The International Olive Council's document COI/T.20/Doc. No 29/Rev.2 (2022) says Method 1 “quantifies using tyrosol and therefore may underestimate the true value”, and that it “may be used to perform a quick determination”. Its newer SPE-HPLC-DAD method is the one it says can be used to “determine the real concentration of phenols to fulfil an EFSA claim”.
Can I compare an HPLC number on one brand with an NMR number on another?
No. They are counted against different references: the International Olive Council's first HPLC method reports a total as tyrosol, NMR reports the named compounds themselves, and Folin reports a total as caffeic acid. Even chromatographic protocols differed among themselves in a 2019 five-lab study. Compare HPLC with HPLC and NMR with NMR, and ask each brand for its method, laboratory and report number.

Remy Assir

Founder, KALTA

Writing about early-harvest, high-polyphenol olive oil — what the numbers mean, why it tastes the way it does, and how to use it.

Explore KALTA →