Emulsifiers
Comedogenic charts print one number against each emulsifier name. Fulton's table prints different numbers for related materials a few rows apart — glyceryl stearate NSE at 1 of 5 against the soap-containing SE grade at 3 of 5, PG monostearate at 0-3 in 1989 against propylene glycol monostearate at 4 in his 1984 paper. In this family the number belongs to the material and the conditions that were tested, not to the name on the label.
What they are
This family is defined by a JOB, not by a chemistry, and that is unusual on this site. An emulsifier is whatever keeps an oil phase and a water phase from separating, so the nine members share a function and very little else. Glyceryl stearate is an ester of glycerin and a fatty acid. Polyglyceryl-3 diisostearate is an ester of a glycerol oligomer. Propylene glycol monostearate is an ester of a different alcohol again. Sorbitan oleate is a sugar-alcohol ester; polysorbate 20 and polysorbate 80 are ethoxylated sorbitan esters, which is to say they are chemically part of the ethoxylate story and we file them here instead, under what they do. Lecithin is a mixture of phospholipids. And triethanolamine is not independently an emulsifier in every use — it is an alkanolamine, most often a pH adjuster, that participates in emulsification where it forms a soap with a fatty acid already in the formula. So "these are all emulsifiers" is a statement about a formulator's shelf, not a claim about molecular kinship, and we would rather say that plainly than dress the grouping up as chemistry. Several of the names are not even single compounds: commercial polysorbates are mixtures of ethoxylated species around an average ethoxylation value, commercial glyceryl stearate is a distribution of esters, and lecithin's composition depends on its source and processing. That matters here more than anywhere, because this page is about scores attaching to samples. Two boundaries, because the ones this site keeps crossing are exactly here. FIRST — the members named on this page are the nine in the table below, and nothing else. Several of the rows quoted in the next section are NOT members of this family. Cetearyl alcohol we file as a fatty alcohol, ceteareth-20 as an ethoxylate, stearic acid as a fatty acid; emulsifying wax NF we do not rate at all. They are quoted as rows in a paper, not as members of a family, and the distinction is the one the oil hub had to be corrected on. SECOND — we cannot tell you which of Fulton's numbered sections his emulsifier rows sat in. Our transcription of that table records the heading for exactly two blocks, "VI. Oils*" and "X. Vitamins and herbs", and no other section number or title in that paper is recorded anywhere in this project. What we can say is narrower and checkable: no member row of this family carries the source-dependence asterisk. The one asterisked row quoted below — stearic acid, at 2-3* — is not a member of this family; it is quoted as context. And the asterisk in Fulton's 1984 table is a different mark with a different meaning — that paper's own footnote uses it for materials diluted in mineral oil or an alcohol gel rather than used as-is — and the two must not be read as one notation.
What the evidence actually says
Print the key first, because a dozen numbers read as one series otherwise, and this family's numbers come from two papers with two different keys. Fulton's 1989 key, in his words: "a minimal grade of 0 to 1 is not considered significant. Grade 2 to 3 is borderline. However, a grade of 4 to 5 is uniformally reproduceable and considered positive." His 1984 paper draws the line somewhere else: "a minimal grade of 1 to 2 is not considered significant. However, a grading score of 3 or above is uniformly reproducible and considered positive." Both scales run 0 to 5. A 3 is borderline in one paper and positive in the other. Nothing below may be compared across the two without carrying each paper's own key. Now the thing this page exists for. The popular charts give an emulsifier one number. Fulton's table shows, four different ways, that the number attaches to the sample in the jar he opened — a grade, a formulation, a nominally matching name — and not to the name a chart prints. ONE — TWO RELATED GRADES, TWO NUMBERS. Fulton has no row reading "Glyceryl stearate". He has "Glyceryl stearate NSE" at 1 of 5 and "Glyceryl stearate SE" at 3 of 5. Those are related commercial grades, not two spellings of one test material: the Cosmetic Ingredient Review's final 2020 monoglyceryl-monoester assessment defines Glyceryl Stearate SE as a self-emulsifying grade of glyceryl stearate containing some sodium and/or potassium stearate — a soap component the NSE name does not carry. By Fulton's 1989 key the 1 is not significant and the 3 is the top of the borderline band. Fulton recorded no supplier, no lot and no composition for either row, and he presented no controlled comparison of the grades. So what the pair establishes is material-specific disagreement: two related, compositionally undercharacterised grades answered two points apart, and the table does not identify why. It is tempting to say the soap did it. Fulton did not say so, we found no study isolating the soap's contribution, and both of our member pages stop at the same place. So does this one. One more thing the pair establishes: plain "Glyceryl Stearate", the unqualified INCI name on most labels, has no printed row of its own. Our Clear for it reads the NSE row across — a qualified legacy-name crosswalk, which its page says in those words, not an exact printed-name match. TWO — THE COLD-CREAM ENTRY. Fulton prints "Triethanolamine" at 2 of 5 and, in the row immediately below it, "Stearic acid: TEA" at 3 of 5. The second row is not a clean assay of an isolated TEA-stearate salt. It reports a formulation experiment, in his words: "In testing different ratios [4:1, 1:1, 1:4] of stearic acid to triethanolamine (stearic acid:TEA) in a cold cream base, all combinations were found to be comedogenic." That is evidence about mixtures in an incompletely described cold-cream base — different ratios, an unstated degree of neutralisation — and its 3 transfers neither to the amine above it nor to TEA-Stearate as a defined ingredient. The amine's own 2, borderline under the 1989 key, is a separate row and stays separate. As for the salt itself: the CIR's 1995 final report on TEA Stearate states that "published data on TEA Stearate as an individual ingredient were not available" at that review — a safety assessment, a different endpoint, and no comedogenic reading either. No document we hold contains an isolated-material measurement of TEA-stearate at all. An earlier version of this page called these rows an amine, an acid and the salt they make together, "each measured on its own". That was wrong, and the external fact-check caught it. THREE — THE PAIR FILED IN TWO FAMILIES. Fulton scores "Sorbitan oleate" 3 of 5, "Polysorbate 20" 0 of 5 and "Polysorbate 80" 0 of 5. The materials are related — the CIR's polysorbate report gives polysorbate 80's chemical name as polyoxyethylene (20) sorbitan monooleate, the sorbitan ester with ethylene oxide on it — but commercial polysorbates are mixtures of ethoxylated species around an average value, and Fulton's sorbitan oleate was not their compositionally controlled parent sample. So these three rows are a cross-material observation, not an ethoxylation experiment. The direction they gesture at — more ethoxylation, lower score — runs across several of Fulton's series and belongs to the ethoxylate hub, where it is reported as a consistent direction and not as a mechanism. Nothing in these rows establishes that ethoxylation caused the zeros. FOUR — THE SAME NAME, TWO SURVEYS. Propylene glycol monostearate is the one member Fulton scored under a matching name twice, seven years apart. His 1989 table prints "PG monostearate" at 0-3 of 5; his 1984 table prints "Propylene glycol monostearate" at 4 of 5. Read with each paper's own key, the first never reaches positive and the second is positive. These are two nominally matching entries from the same investigator lineage, with supplier, lot and composition unreported in both — not two readings of one demonstrated sample. The conditions differed too: the 1984 row is unasterisked, which under that table's preparation footnote means the material was used as-is, while the 1989 survey ran its routine procedure at roughly 10% in propylene glycol. A genuine conflict at the level of the printed name, flagged as a conflict in our ratings file, and unresolved. THE FULL INVENTORY, because a hub that hides its family's worst number is doing the thing this site exists to refuse. Polyglyceryl-3 diisostearate — 4 of 5, positive under the 1989 key, an exact name-level match and the family's strongest legacy result, read once. Sorbitan oleate — 3 of 5, borderline. Glyceryl stearate SE — 3 of 5, borderline, a grade's number. Triethanolamine — 2 of 5, borderline. Glyceryl stearate NSE — 1 of 5, not significant, crosswalked to the plain name. Propylene glycol monostearate — 0-3 of 5 in 1989 against 4 of 5 in 1984. Polysorbate 20 and polysorbate 80 — 0 of 5 each. Lecithin — no row in either Fulton paper, or anywhere else in the corpus we hold. So: Fulton's emulsifier-related rows show why a copied ingredient number is incomplete. Related commercial grades, formulation systems and nominally matching samples produced different legacy rabbit-ear results. But these were undercharacterised materials in small historical surveys, not controlled composition studies, and they do not establish a general rule about emulsifiers, soaps, mixtures or ethoxylation. The score attaches to the material and the condition tested. That is as far as the evidence reaches, and for five of the nine members — each with a single row under its own name, or none — nothing here moves their numbers at all.
What the evidence doesn’t tell you
Start with what would break any tidy rule drawn from the pairs above, because it is in the same table. THE ROWS DO NOT RUN IN ONE DIRECTION, AND MIXTURES ARE NOT RELIABLY WORSE. "Emulsifying wax NF", a self-emulsifying blend, reads 0 of 5. A few rows earlier "Cetearyl alcohol + ceteareth 20" reads 4 of 5, while "Cetearyl alcohol" and "Ceteareth-20" each read 2 of 5 on their own. Neither is a member of this family — we file cetearyl alcohol as a fatty alcohol and ceteareth-20 as an ethoxylate, and emulsifying wax NF carries no rating here at all — but they are the rows that stop "the blend scores higher" from becoming a rule. One blend row sits above both its named components; another blend row sits at the floor. These are contextual formulation rows in one survey, and you cannot predict a mixture's row from its parts, or the reverse, from anything on this page. IT IS STILL A RABBIT'S EAR. Every number above comes from a rabbit-ear assay: the material at roughly 10% in propylene glycol, applied to the inner ear of New Zealand albino rabbits, three per assay, five days a week for two weeks, follicular keratosis read under a micrometer — or, in the 1984 paper, two rabbits per test with the opposite ear as control. Fulton called his 1989 survey "not at all definitive but simply designed to stimulate research." Draelos and DiNardo's 2006 paper tested finished products on small human upper-back panels and argued that an isolated ingredient's rabbit grade is a poor guide to the product it ends up in; a 2025 review in JAAD Reviews found that no standardised comedogenicity test exists at all. Those objections apply to polysorbate 80's 0 exactly as hard as to the SE grade's 3. THE LATER RABBIT SURVEY HAS NONE OF THEM. Nguyen, Dang and Maibach's 2007 rabbit study, the one post-Fulton primary survey we hold, includes none of this family's nine members among its named test materials. That is a bounded absence we checked in the paper, and it means no member of this family has a reading from the only survey conducted after 1989. WE IDENTIFIED NO HUMAN COMEDOGENICITY STUDY OF ANY MEMBER OF THIS FAMILY. Each of the nine member pages says the same thing in its own words, and we checked them rather than assuming it. That is a statement about what we searched, not about what exists. THE LABEL IS NOT USELESS HERE, AND WE ARE NOT GOING TO PRETEND IT IS. "Glyceryl Stearate" and "Glyceryl Stearate SE" are two separate INCI names, and TEA-Stearate is a third. A label that prints the SE name is naming the soap-containing grade, and that resolves which of Fulton's grade rows is the nearer one. But the resolution only goes so far: no row prints the unqualified name "Glyceryl stearate", so the plain label name resolves to a crosswalk rather than to a printed row, and no label resolves whether the material in the jar matches the sample Fulton tested — he recorded no supplier, no lot and no composition for any row in this family. THE SECOND LABORATORY EXISTS, AND IT HAD THE SAME PROBLEM. Two members carry a genuinely non-Fulton reading. Morris and Kwan, in 1983, in a different laboratory and on the same 0-5 scale, scored Glyceryl Stearate 2 at 50% in mineral oil, and Polysorbate 20 0 at 1% aqueous. That is real independent same-name evidence for two of nine — and neither reading replicates Fulton's conditions: different concentrations, different vehicles, and their glyceryl stearate row records no SE or NSE designation, so it cannot adjudicate the grade difference this page is built on. Equal digits under different preparations are not same-condition replication. The second laboratory recorded the grade no better than the charts do. MEMBERS THIS PAGE DOES NOT EXPLAIN. Lecithin has no row in the corpus we searched — its Clear is an editorial inference with no legacy assay behind it, and its own page says so. Polyglyceryl-3 diisostearate's 4 is the highest number in the family, it has one row under its exact name and no pair, and nothing above softens it; the idealised name also does not establish the polyglycerol or ester distribution of whatever sample was tested. Sorbitan oleate has one row and no pair. Polysorbate 20 and polysorbate 80 are the family's zeroes and stay zeroes. AND THE OBVIOUS ONE. A family-level generalisation is precisely the kind of claim this site exists to distrust, including when we are the ones making it. What is established is that Fulton printed these particular rows and that related names carry different numbers. Everything past that — that grade generally decides an emulsifier's number, that a chart's single figure is generally the wrong one — is an extrapolation from a handful of rows in one table, and we have not tested it against the rest of the family, because for the rest of the family there is only one row each, or none.
All 9 emulsifiers we rate
The legacy score is printed exactly as it appears in the paper, including the ranges, the asterisks and the parentheses. Those marks are not decoration. They are the paper telling you what its own number is worth.
Sources
- Fulton JE. "Comedogenicity and irritancy of commonly used ingredients in skin care products." J Soc Cosmet Chem 40:321-333 (1989) The primary source for nearly every comedogenic number in circulation, read off our verbatim transcription of Table I rather than via a list quoting it. A rabbit-ear assay: the ingredient at roughly 10% in propylene glycol, applied to the inner ear of New Zealand albino rabbits, three per assay, five days a week for two weeks, follicular keratosis measured with a micrometer, graded 0-5. Everything this page argues comes out of that table's rows read exactly as printed — the two glyceryl stearate grade rows, the "Stearic acid: TEA" entry his text ties to a cold-cream formulation experiment, the polysorbate zeroes — together with Fulton's own key, which the republishers of his numbers do not print. He recorded no supplier, lot or composition for any row quoted here, and described the survey as "not at all definitive but simply designed to stimulate research." Not human data, and not a measurement of any finished product except where it says so.
- Fulton JE Jr, Pay SR, Fulton JE III. "Comedogenicity of current therapeutic products, cosmetics, and ingredients in the rabbit ear." J Am Acad Dermatol 10(1):96-105 (1984) Fulton's earlier rabbit-ear survey, two rabbits per test, the opposite ear as control, on a 0-5 scale with a DIFFERENT key from the 1989 paper - 1-2 not significant, 3 or above considered positive. Cited for one row and two cautions. The row is "Propylene glycol monostearate 4", against the 1989 table's "PG monostearate 0-3", the only member of this family the investigator scored under a matching name twice. The first caution is the key itself: a 3 from this paper and a 3 from the 1989 paper do not mean the same thing. The second is the preparation footnote: this paper's asterisk marks materials diluted in mineral oil or an alcohol gel, the monostearate row is unasterisked and so was used as-is, and neither fact makes it comparable to the 1989 row's roughly-10%-in-propylene-glycol procedure.
- Morris WE, Kwan SC. "Use of the rabbit ear model in evaluating the comedogenic potential of cosmetic ingredients." J Soc Cosmet Chem 34:215-225 (1983) A genuinely independent rabbit-ear assay - a different laboratory, six years before Fulton, on the same 0-5 scale, and one that contradicts him in print on other materials. It is the only non-Fulton reading we hold for this family, and it covers two of the nine members: Glyceryl Stearate 2 at 50% in mineral oil, and Polysorbate 20 0 at 1% aqueous. Read from our transcription of its ingredient table. Cited as much for its limit as for its numbers: it records the glyceryl stearate grade no better than a consumer chart does, so it cannot adjudicate the SE-against-NSE difference, and its concentrations and vehicles differ from Fulton's, so its digits are not same-condition replications of his. We do not hold this paper's own key to its scale, so its numbers are reported with their scale, concentration and vehicle and with no band label borrowed from Fulton.
- Nguyen SH, Dang TP, Maibach HI. "Comedogenicity in rabbit: some cosmetic ingredients/vehicles." Cutan Ocul Toxicol 26(4):287-292 (2007) The one post-Fulton primary rabbit survey we hold, on a 0-4 scale — a different scale from every other number on this page. Cited for what it does NOT contain: none of this family's nine members appears among its named test materials, so the newest primary document in the corpus offers this family no reading at all. A bounded absence, checked in the paper rather than assumed.
- Cosmetic Ingredient Review. "Final Report on the Safety Assessment of TEA Stearate." J Am Coll Toxicol 14(3):240-248 (1995) An industry-funded but independently paneled toxicological review, in the FINAL report state, of the triethanolamine salt of stearic acid. We read its abstract verbatim from the publisher's page; we did not read the full text. Cited for exactly one recorded absence, quoted here: "published data on TEA Stearate as an individual ingredient were not available" at that review. It assesses SAFETY, not comedogenicity — a different endpoint — and no comedogenic inference is drawn from it in either direction; it is cited to document that no isolated-material reading of the salt existed for that panel to review.
- Fiume MM, Bergfeld WF, Belsito DV, et al. "Safety Assessment of Monoglyceryl Monoesters as Used in Cosmetics." Int J Toxicol 39(3 Suppl):93S-126S (2020) The Cosmetic Ingredient Review's FINAL peer-reviewed safety assessment of the monoglyceryl monoester family, replacing the 2015 public-comment draft an earlier version of this page cited for the same fact. Cited for one definition and nothing else: it defines Glyceryl Stearate SE as a self-emulsifying grade of glyceryl stearate containing some sodium and/or potassium stearate. That definition is what establishes that SE and NSE are related but compositionally different commercial grades rather than one test material printed twice, which is the hinge of this page. It is a safety document and contains no comedogenicity finding.
- Draelos ZD, DiNardo JC. "A re-evaluation of the comedogenicity concept." J Am Acad Dermatol 54:507-512 (2006) A Current Issues and Opinion piece that tested FINISHED PRODUCTS on small human upper-back panels and found products testing non-comedogenic despite containing ingredients the rabbit assay had graded comedogenic. It reports no comedogenic score for any individual ingredient, and it is not a same-material validation of the rabbit ear against human skin. Cited because it cuts against this family's zeroes exactly as hard as against its threes, and a page that used it in one direction only would be choosing its evidence.