Starting out · 4
Reading a raw material datasheet: what each number changes about your decision
· 22 min read
Molecular weight, boiling point, flash point, specific gravity, logP. Most guides tell you what these fields are. This one tells you which decision each one changes — with the distribution across 42,225 materials, so you know what a normal value looks like. Includes the unit trap that will make you misread flash point by 100 °C.
Look up any raw material and you get a table: a molecular weight, a boiling point, a flash point, a specific gravity, a logP. Most introductions explain what those fields are. That is not the problem — you can read a definition anywhere.
The problem is that a number you cannot act on is decoration. So this page is organised the other way round: for each field, what decision does it change, and what does a normal value look like?
The "normal" part matters more than it sounds. A boiling point of 247 °C means nothing to you until you know that it is almost exactly the median of every material on this site. The distributions below are computed over all 42,225 materials in the database. Coverage varies by field — not every entry has every number — so the count is stated each time.
Molecular weight — the volatility budget
Coverage: 20,713 materials (49%). Median 204 Da; the middle 80% runs 124 to 465.
You cannot smell a material that does not reach your nose, and reaching your nose means leaving the bottle as vapour. Heavier molecules do that less readily. That is the whole reason this field is on the sheet.
73.9% of the materials here fall between 30 and 300 Da. That band gets quoted a lot, and it is worth being clear about what it is not. It is not a test for whether something smells. Molecules below 30 Da can be odorous, molecules above 300 are not automatically odourless, and sitting inside the band guarantees nothing — vapour pressure, functional group and detection threshold all still decide. Treat it as the range most common organic odorants happen to fall in, not as a boundary.
It is also worth remembering that this database includes solvents, fixatives and materials sold for flavour work, so not every row is here to be smelled.
What it changes for you: molecular weight is your first guess at where a material will sit in time, before you have smelled anything. It is a guess, not a verdict — the evaporation curve is the thing that actually decides it, and it does not follow molecular weight cleanly.
Boiling point — a better guess, with a condition attached
Coverage: 11,879 materials (28%), atmospheric values only. Median 259 °C; the middle 80% runs 155 to 463.
Boiling point is the more direct measure of the same thing, and it is the field most people reach for when they want to know whether something is a top note.
But look at how the source actually stores it:
434.00 to 437.00 °C. @ 760.00 mm Hg | 250.00 °C. @ 9.00 mm Hg
Two numbers for one material, about 185 °C apart. Neither is wrong. A boiling point is only defined at a pressure, and 760 mm Hg is atmospheric while 9 mm Hg is a vacuum distillation. The second number tells you how the material is purified, not how it behaves in your formula.
What it changes for you: when you compare two materials' boiling points, check they were measured at the same pressure. If one is a vacuum figure you are comparing a smell to a manufacturing step.
We had to take our own advice here. A first pass at this distribution simply read the first number in each entry, which for 745 materials meant a reduced-pressure figure. That pulled the median down to 247 °C. Restricted to entries actually measured at 760 mm Hg, it is 259 °C — so the rule above is not hypothetical, and a distribution that ignores it is quietly wrong by about twelve degrees at the middle.
Flash point — a shipping field, not a smell field
Coverage: 14,277 materials (34%). Median 100 °C; the middle 80% runs 29 to 215.
This one is not about odour at all, which is why it gets skipped — and it is the field most likely to actually stop you doing something.
21.4% of the materials here with a recorded flash point are at or below 60 °C, which is the basic upper limit in the UN Model Regulations for a Class 3 flammable liquid. Below that line you may be looking at dangerous-goods paperwork, a courier who declines the shipment, and a surcharge that can exceed the price of the material.
Two things that figure is not. It is not a count of Class 3 materials — that classification applies to liquids, and this database contains solids like carbon black that have a flash point field anyway. And crossing the line in either direction settles nothing on its own: above 60 °C a material is not automatically an ordinary parcel, and below it a material is not unshippable, it is a material that has to be classified and shipped under the dangerous-goods rules. Closed-cup is the preferred test method — the TCC in the data is Tag Closed Cup — but open-cup results are also accepted, subject to a correction.
Now the trap. The source writes flash point in Fahrenheit first, with Celsius in brackets:
180.00 °F. TCC ( 82.22 °C. )
Read the first number and you will take this for a material that flashes at 180 °C. It flashes at 82 — nearly a hundred degrees lower, in the direction that matters. The fields either side of it, boiling point and specific gravity, are metric-first, so nothing warns you that this one changed units. When I first computed the distribution for this article I made exactly this mistake and got a median of 200 rather than 100.
While we are being careful: a flash point is the temperature at which a material gives off enough vapour to form an ignitable mixture in a specific controlled test. It is not a ceiling below which the material is safe, and it is one input to a flammability assessment rather than the assessment.
What it changes for you: before you order, check the flash point, not the price. A cheap material that needs dangerous-goods handling is not cheap.
Do not make a shipping decision from a single row. 1,149 of these entries carry a
<or>qualifier that a parser strips, turning a bound into an apparent measurement. In 23 entries the Fahrenheit and Celsius values disagree by more than a degree, and in 13 by more than five — one reads1250 °F (52.1 °C), which cannot be both. The distribution is sound; any individual number should be confirmed against the supplier's safety data sheet.
Specific gravity — the reason you weigh instead of measure
Specific gravity is how much a given volume weighs, relative to water. It is the least glamorous field on the sheet and the one that will quietly ruin a formula.
The article on weighing already works this field properly, over the 3,863 entries that are liquid and have a usable value: median 0.941, middle 90% spanning a factor of 1.38, full range 2.21. It also declares the cleaning step that makes those figures trustworthy — 31 of the raw values have a shifted decimal point, and left in they inflate the range to a spurious 7.8×.
The reason it matters here is short. A formula is written in mass percent. If you build it by pipetting volumes — because volumes are easier — every material away from the middle of that range is wrong by a margin you cannot see.
What it changes for you: weigh everything. This is the entire argument of weighing and dilution, and specific gravity is the number that proves it rather than asserts it.
logP — how it partitions, which is not the same as how long it lasts
Coverage: 19,580 materials (46%). Median 2.8; the middle 80% runs 0.1 to 6.6.
logP is the logarithm of how a material distributes itself between octanol and water. High means it prefers oil; low means it prefers water.
47.1% of the materials here sit at logP 3 or above. You will see that figure of 3 treated as the line above which materials become "substantive" — things that hang on to skin and fabric. Before leaning on it, know where it comes from: it appears mainly in formulation patents rather than peer-reviewed work, it is usually a calculated value (ClogP) rather than a measured one, and the patents generally pair it with a second condition — a boiling point at or above about 250 °C — rather than using logP alone.
So treat it as an orientation, not a threshold. It tells you which phase a material would rather be in, which is genuinely useful when deciding whether something will survive in an alcoholic base, a lotion or a wash-off product. It does not tell you how long you will smell it.
What it changes for you: logP predicts where a material goes, not how long it stays. If your product is water-based and you pick a logP of 6, you are going to have a solubility problem before you have a longevity one — see solvents.
What the sheet cannot tell you
Every number above is a physical constant. None of them is a smell.
Two materials can share a molecular weight, a boiling point and a logP and be nothing alike in the bottle. The datasheet narrows the field and rules things out; it does not choose. That work is done by smelling in a controlled way and writing down what you found — which is what an evaluation note is for, and why "smells nice" is not data.
A reading order that works
- Flash point first. It is the only field that can stop the purchase outright. Check the unit.
- Specific gravity, if you are going to formulate with it. It tells you the scale you need.
- Molecular weight and boiling point together for a first placement in time — and check the boiling point's pressure.
- logP when the base is not neat ethanol.
- Then smell it, at a dilution you decided in advance rather than one the bottle suggested. Odour threshold is why that dilution matters more than the material.
Start with your first ten materials if you have not bought anything yet — and note that the fragrance wheel will not help you here, because it arranges finished perfumes and every field on this page belongs to a raw one.