Material guide: sotolon — the database calls it caramel furanone, which is why I could not find it last round

Short version: one of nine nuclear-grade materials, where concentration decides whether it is curry or maple syrup. It can be bought (41 suppliers). It just is not filed under this name. And whether what went dull after you spilled it is taste or smell has to be separated first.

The fields

Field	Value
Database name	caramel furanone
Common name	Sotolon (also sotolone)
Systematic name	4,5-dimethyl-3-hydroxy-2,5-dihydrofuran-2-one
CAS	28664-35-9
Formula	C6H8O3, MW 128.13
Database category	Flavoring agents
Appearance	White to pale yellow, liquid to solid (est)
Melting point	25–29 °C
Boiling point	93–95 °C @ 2 mmHg
logP	0.40 (est)
Water solubility	222,800 mg/L @ 25 °C (est)
Odour type	Caramellic
Odour	Sweet, caramellic, maple, burnt sugar, coffee, brown sugar, cotton candy
Strength	Very high; evaluation recommended at 0.01% solution or less
Substantivity	> 400 hours at 3% in propylene glycol
Suppliers	41 (plus 37 selling the solution)
Regulatory	JECFA 243, FEMA GRAS 3634, CoE 11834, FLAVIS 10.030
Toxicity	Oral, dermal and inhalation all not determined

One of nine

This series measured the database's distribution of recommended evaluation strengths in article #43: of 42,225 records, only 802 carry the field, of which 527 say 10%, 204 say 1% and 45 say 0.1%.

And only nine in the whole database say 0.01% or less.

Sotolon is one. The other eight on that list are 2,3,5-trimethyl pyrazine, maple furanone (abhexon), (Z)-4-heptenal, propyl mercaptan, isopropyl mercaptan, o-cresol, denatonium benzoate and m-cresol.

Mercaptans, cresols, a pyrazine, and two furanones. That is the entire list of materials this database considers unsafe to assess above one part in ten thousand.

The substantivity field reads > 400 hours, and it was measured at 3% in propylene glycol. By this series' rule, > is a lower bound: nobody has reported how long it actually lasts, only that even diluted to 3% it is still there past sixteen days.

Concentration decides what it is

The database's notes field quotes this in full, which is unusually helpful:

"Sotolon (also known as sotolone) is a lactone and an extremely powerful aroma compound, with the typical smell of fenugreek or curry at high concentrations and maple syrup, caramel, or burnt sugar at lower concentrations."

One molecule, and at two different concentrations it is two different foods.

The two odour description records show both ends of the same thing:

• Luebke (tgsc, 1998), at 0.01% in propylene glycol: "extremely sweet strong caramel maple burnt sugar coffee"
• Mosciano (Perfumer & Flavorist, 1997): "Sweet, caramellic, maple, brown sugar, cyclotene and cotton candy"

Note Luebke's condition: 0.01%. One part in ten thousand, and at that concentration his description is still "extremely sweet, strong".

The notes field also carries a practical line: "Can get some extraordinary blends if you shade it well." What "shade it well" means here is unambiguous: get the amount right.

What it is in

The natural occurrence field:

coffee; fenugreek seed; buckwheat honey; maple; pineapple fruit; sake; sherry; virginia tobacco; rice wine

And the notes add: "Key flavour compd. found in raw cane sugar, aged sake, coffee, fenugreek, lovage and wines."

There is a pattern in that list: ageing and heat. Sake, sherry, rice wine and tobacco are ageing; coffee, maple and cane sugar are heat. Sotolon is what those two processes have in common.

In 2014, Gabrielli, Fracassetti and Tirelli published a UHPLC method for quantifying sotolon in white wine in the Journal of Agricultural and Food Chemistry. That fact alone says something about its standing: in oenology, sotolon concentration is an indicator worth building a dedicated analytical method for. (In white wine it is generally regarded as a marker of oxidation or ageing.)

It gets into you

The US National Institutes of Health's Drugs and Lactation Database (LactMed) writes, in the adverse effects section of its fenugreek entry:

"Perhaps its most unusual side effect is the imparting an odor of maple syrup to the urine, sweat, feces, and possibly breastmilk by the sotolon in fenugreek."

Urine, sweat, faeces, and possibly breastmilk.

That is a referenced official surveillance database listing an adverse effect of a plant, with sotolon named as the cause.

To be precise: that sentence concerns ingesting fenugreek. What it establishes is that the mechanism exists — sotolon leaves the body in fluids and carries a recognisable odour with it — not any conclusion about skin contact. This round's beginner article deals with a forum user's related situation.

That name

This is the section worth remembering, because it is what made me publish something wrong last round.

In this site's database, this material's name field reads caramel furanone.

• The word sotolon sits in the notes field
• The chemical name 4,5-dimethyl-3-hydroxy-2,5-dihydrofuran-2-one sits in the head synonym field
• The name field is a marketing name

Last round, checking whether caramel molecules could be bought, my query was "name field contains sotolon". It returned zero. I wrote that result up as "no record in the database."

It has 41 suppliers.

The same query missed two more:

The name I searched	The database's name	CAS	Suppliers
sotolon	caramel furanone	28664-35-9	41
furaneol	strawberry furanone	3658-77-3	101
maple lactone / cyclotene	cyclotene	765-70-8	72

This database files flavour materials under marketing names. caramel furanone, strawberry furanone, maple furanone: a naming convention, not three coincidences.

(Cyclotene I could have found by searching cyclotene directly. I had assumed my way into "maple lactone" and "methyl cyclopentenolone", neither of which matches.)

This series devoted an entire article, #41, to the claim that the database indexes by chemical name while formulas use trade names, so you cannot look things up. I measured that phenomenon and then fell into the same hole the following round, from the opposite direction.

The relevant section of last round's article has been rewritten with a correction date, and its conclusion reversed.

Whether to buy it

• For maple, caramel, curry or aged-liquor directions, this is the core molecule of that position, and 41 suppliers make it easy to obtain.
• Start at 0.001%, not 0.01%. The database recommends 0.01% or less, and only nine materials in the whole database reach that class.
• Buy the solution. Thirty-seven suppliers sell caramel furanone solution, which for a material of this strength is usually the more sensible entry.
• Wear gloves. A substantivity floor past 400 hours, and it carries its odour out through body fluids. Together those mean getting it on you is not solved by washing.
• Its database category is flavoring agents, not fragrance. All three toxicity fields read "not determined", which is common for a FEMA GRAS flavour material (food safety runs through a different assessment system), but if you are making something to put on skin, those three blanks are worth noticing.
• To look it up, use caramel furanone or CAS 28664-35-9.

What this doesn't establish

• All three toxicity fields read "not determined", and I did not go to the toxicology literature separately. This article makes no claim about its dermal safety.
• The > 400 hours substantivity is a lower bound, and measured in a 3% solution. The actual value is unknown.
• "Evaluation recommended at 0.01% or less" is the database's rating, not an odour threshold. The database does not document how that field is set, and only 1.9% of materials carry it.
• Gabrielli 2014 is a method paper. I cite it to establish that sotolon is an indicator worth quantifying in oenology, and cite none of its figures.
• The LactMed sentence concerns ingesting fenugreek, not contact with neat sotolon. Doses and routes differ.
• I did not find it in the corpus. No identifiable sotolon entry appears in the 954 published formulas, which is consistent with its classification as a flavour material, but I did not further confirm that no other spelling escaped me — given this article's subject, that sentence deserves a second reading.