Material guide: 2,3,5-trimethylpyrazine — one of nine materials asking for 0.01% or less, and by far the easiest to buy

CAS 14667-55-1, 99 suppliers. Its strength field reads high, evaluate at 0.01% or less, and of the 2,451 records in this database carrying a strength field, only nine land in that bracket. Most of the nine are specialist items: maple furanone at 67 suppliers, caramel furanone at 41, propyl mercaptan at 24, o-cresol at 13, and even denatonium benzoate, a bittering agent whose odor field is empty. Trimethylpyrazine's 99 suppliers make it the outlier on that list. The family disagrees with itself, too: 236 records carry pyrazine in the name, 87 of them have odor data, and the recommended dilutions for the eight best-supplied run from 0.01% to 10%. Stranger still, 2-acetylpyrazine is labelled very high and recommended at 0.10%, while this one is labelled high and recommended at 0.01%, so the adjective and the number point opposite ways. Two papers: a 2013 stable isotope dilution study quantified 12 alkylpyrazines in coffee, ranking this one seventh, with total alkylpyrazines in commercial ground coffee at 82.1 to 211.6 mg/kg and decaffeinated samples at 0.3 to 0.7 times regular; and a 2025 study heating 18 amino acids at 160 °C for an hour found pyrazines forming even without reducing sugars.

About a 4 minute read.

Too long, didn't read

Name: 2,3,5-trimethyl pyrazine, CAS 14667-55-1, 99 suppliers
Odor family nutty; descriptors: nutty, nut skin, earthy, powdery, cocoa, baked potato, roasted peanut, hazelnut, musty
Strength field: high, evaluate at 0.01% or less. Only nine materials in the database land in that bracket
Longevity 21 hours, vapour pressure 1.724 mmHg, boiling point 171–172 °C, flash point 54.44 °C
Listed in all four regulatory indexes: JECFA, FEMA GRAS, CoE, FLAVIS
Every natural occurrence has been roasted or cooked: roasted almond, roasted barley, beef, wheat bread, Emmental cheese, cocoa, coffee

The 0.01% bracket holds nine materials

The previous piece measured the strength field's distribution: of 2,451 records carrying it, only nine give 0.01% as the recommended dilution. They are:

Material	Suppliers	Odor family
2,3,5-trimethylpyrazine	99	nutty
Maple furanone	67	caramellic
Caramel furanone	41	caramellic
(Z)-4-heptenal	28	green
Propyl mercaptan	24	alliaceous
Isopropyl mercaptan	16	alliaceous
o-Cresol	13	phenolic
Denatonium benzoate	9	(empty)
m-Cresol	8	phenolic

The shape of that list is clear: mercaptans, phenolics, furanones, and one pyrazine. Apart from trimethylpyrazine and maple furanone, none has more than 41 suppliers.

The second from bottom is worth a note. Denatonium benzoate's odor field is entirely empty because it is not a fragrance material at all; it is a bittering agent, the thing added to antifreeze and cleaning products to stop people swallowing them. It appears on this table because the strength field records "how far to dilute before smelling it", and the catalogue keeps it in the same data structure as the fragrance materials.

99 suppliers makes this one the outlier here. The other eight are specialist items; this is a commodity.

The family spans a thousandfold internally

The database holds 236 records with pyrazine in the name, 87 of them with odor data. The eight best supplied:

Material	Suppliers	Odor family	Strength field
2,3,5-trimethylpyrazine	99	nutty	high, 0.01% or less
2-Acetylpyrazine	80	popcorn	very high, 0.10% or less
2,3,5,6-Tetramethylpyrazine	75	nutty	(empty)
2,5-Dimethylpyrazine	60	chocolate	(empty)
Filbert pyrazine	54	nutty	(empty)
2,3-Dimethylpyrazine	52	nutty	high, 0.10% or less
2-Isobutyl-3-methoxypyrazine	49	green	(empty)
2,6-Dimethylpyrazine	48	chocolate	medium, 10.00% or less

One chemical family, and the recommended dilution runs from 0.01% to 10%, a thousandfold.

The top two rows are where to stop. 2-Acetylpyrazine is labelled very high and recommended at 0.10%; this one is labelled high and recommended at 0.01%. The adjective says the first is stronger; the number says the second needs diluting ten times further. The two fields point opposite ways.

That is not an error in the data. The strength field was never a threshold. The adjective and the recommended dilution are two judgements entered by different people at different times, never calibrated against each other. To compare, read the number, not the word.

In coffee, it ranks seventh

In 2013 Pickard and colleagues at the University of Kaiserslautern developed a stable isotope dilution analysis (SIDA-GC-MS) for the quantitative analysis of 12 alkylpyrazines in commercially available coffee (PMID 23745606).

One methodological aside is worth recording: comparing extraction solvents, water proved superior to dichloromethane for this quantitation.

The results in the paper's own order. The distribution patterns of alkylpyrazines in different roasted coffees were quite similar. The most abundant in every sample was 2-methylpyrazine, followed by 2,6-dimethylpyrazine, 2,5-dimethylpyrazine, 2-ethylpyrazine, 2-ethyl-6-methylpyrazine, 2-ethyl-5-methylpyrazine, and then 2,3,5-trimethylpyrazine. The lowest concentrations were 2,3-dimethylpyrazine and three ethyl-substituted pyrazines.

For scale: in commercially available ground coffee, total alkylpyrazine concentrations ranged between 82.1 and 211.6 mg/kg.

The last sentence is the paper's most interesting. Decaffeinated coffee samples contained alkylpyrazines at 0.3 to 0.7 times the level of regular coffee, which the authors suggest might be a result of the decaffeination procedure.

Ranking seventh is worth setting against the strength field. It is not the most abundant alkylpyrazine in coffee, yet it is the easiest to buy and the one the catalogue asks you to dilute hardest. Abundance ranking and market ranking are different things.

Pyrazines form without reducing sugars

The textbook route to a pyrazine is the Maillard reaction: a reducing sugar plus an amino acid, heated. In 2025 Li and colleagues at Anhui Agricultural University pulled those variables apart (PMID 41267239).

The design: 18 proteinogenic amino acids heated at 160 °C for one hour under low-moisture conditions, tracking volatile formation with and without sugar.

The results come in layers. The structure of amino acid side chains strongly influenced volatile profiles and reaction pathways, with serine and phenylalanine producing the highest levels of volatiles.

The key sentence: even in the absence of reducing sugars, amino acids with aliphatic hydroxyl groups in their side chains (serine and threonine) and with amide groups (glutamine and asparagine) generated pyrazines.

Beyond that, branched chain amino acids (alanine, isoleucine, leucine) and valine produced significant Strecker aldehydes without sugars, and introducing a sucrose-rich sugar mix simulating the composition of tea leaves shifted the volatile profiles towards furans, aldehydes and acids while modulating pyrazine formation. Epigallocatechin gallate (EGCG) suppressed it.

What that means for blending: pyrazines are not confined to the sugar-plus-amino-acid combination. Given the right side chain and enough heat, an amino acid gets there on its own. The Maillard meat notes covered in the sulfurol piece are the other half of the same chemical world.

Buying and storing

99 suppliers, easy to source and not expensive. The care goes into handling:

The 0.01% recommendation is not politeness. Build a two-step dilution (1%, then 1% of that). Do not try to take a small amount from the neat material. The routine in weighing and dilution is a precondition here.
One sniff of the neat material costs you the next half hour of your nose. This family has low thresholds and lingers. Olfactory fatigue is easy to trigger here.
Longevity 21 hours, vapour pressure 1.724 mmHg. Not a base note; a middle with presence.
Flash point 54.44 °C, higher than most esters and terpenes, so shipping and storage constraints are looser.
No GHS data and no shelf life in the record. Ask the supplier for the SDS.

→ 2,3,5-Trimethylpyrazine in the database: full physical data, 99 suppliers and suggested blends.

What this article does not establish

I point out that the adjective and the number disagree without deciding which is right. There is no record of who filled either field.
The coffee paper measured concentrations of alkylpyrazines, not their contribution to aroma. Seventh is an abundance ranking, not an OAV ranking.
The 0.3–0.7 times figure for decaf is the authors' observation and suggestion (possibly a result of the process), not a demonstrated mechanism.
160 °C for one hour is a model system, not the actual conditions of roasting coffee or baking bread.
Denatonium benzoate appearing in the 0.01% list is my inference that the catalogue keeps non-fragrance items in the same data structure; I have no other evidence.