SDS Section 9: Physical and chemical properties

Measurements, not adjectives

Section 9 is the only part of a safety data sheet made almost entirely of numbers. Everything around it is a judgement: a classification, a control measure, an instruction. Section 9 is the evidence those judgements were drawn from, and it is the one place a reader can go to check them.

That gives it two jobs, and they fail separately. The first is disclosure — eighteen named properties, each present, each in units, each with the conditions it was measured under. The second is corroboration: the flash point has to support the flammable-liquid classification in section 2, the decomposition temperature has to support the conditions to avoid in section 10, and the vapour pressure has to make sense of the airborne limit in section 8. A section 9 that is internally tidy but disagrees with the rest of the sheet has failed at the second job, which is the one that matters on the floor.

What Regulation (EU) 2020/878 rebuilt

No other section of the EU sheet was reorganised as thoroughly in the 2020 amendment. It is worth knowing exactly what moved, because a great many sheets in circulation still carry the old shape.

The properties that left the list

Under Regulation (EU) 2015/830, subsection 9.1 ran to twenty lettered items, (a) to (t). It began with appearance and odour and odour threshold as three separate entries and ended with explosive properties and oxidising properties. In between sat evaporation rate, vapour density, relative density and a bare viscosity.

Regulation (EU) 2020/878 cut that to eighteen, (a) to (r):

• Appearance was split into two items — physical state and colour — because “appearance” invited a sentence where a classification wanted a word.
• Odour threshold was folded into odour.
• Evaporation rate was removed from the basic list altogether. It now appears only in the 9.2.2 list of other safety characteristics.
• Explosive properties and oxidising properties were deleted as lettered items. The underlying test data moved into 9.2.1, where it sits beside the hazard class it belongs to.
• Viscosity became kinematic viscosity, in mm²/s — a dynamic value in mPa·s cannot be read against the aspiration hazard criteria without the density.
• Vapour density became relative vapour density, which names its reference; relative density became density and/or relative density, which forces the author to say which one is on the page.
• Particle characteristics was added.

The order changed too. The fire cluster — flammability, explosion limits, flash point, auto-ignition temperature, decomposition temperature — now runs consecutively at (f) to (j), so a reader scanning for the combustion data finds it in one block rather than scattered from (g) to (q).

The dates are the part that bites. The Regulation applied from 1 January 2021, and safety data sheets not complying with it could continue to be provided only until 31 December 2022. A twenty-item section 9 is not merely dated; it is the wrong annex.

9.2.1 and 9.2.2 — the half that did not exist

The old 9.2 was a single line inviting other information. The new one is two named subsections.

9.2.1, information with regard to physical hazard classes, is organised by the classification itself. Where the product is classified as an explosive, a self-reactive substance, a pyrophoric solid, a water-reactive, an oxidiser, an organic peroxide, corrosive to metals or a desensitised explosive, this is where the test results behind that classification go — and where data that was relevant to a physical hazard but did not produce a classification can go as well. UN GHS Annex 4 sets out in table A4.3.9.2 what each class calls for, class by class, from the Koenen test for explosives to the corrosion rate on steel or aluminium for metals corrosives.

9.2.2, other safety characteristics, is broader than its UN counterpart. Where GHS Annex 4’s table A4.3.9.3 names four characteristics — mechanical sensitivity, self-accelerating polymerisation temperature, formation of explosible dust/air mixtures, acid/alkaline reserve — the EU subsection adds evaporation rate, miscibility, conductivity, corrosiveness, gas group, redox potential, radical formation potential and photocatalytic properties. Two of those matter constantly and are almost never filled in: the dust explosion characteristics for any handled powder, and the acid or alkaline reserve for anything at the pH extremes, which is the figure that explains why a weak acid at pH 2 and a strong one at pH 2 do not do the same thing to skin.

Conditions, units and method

Three rules run underneath every number in the section, and each one is a common failure.

Conditions. Annex II sets standard conditions of 20 °C and 101,3 kPa unless the sheet says otherwise; GHS Annex 4 gives the same figures as 20 °C and an absolute pressure of 101.3 kPa. Any value measured somewhere else has to carry the conditions with it. A vapour pressure with no temperature is not data.

Units. Appropriate units throughout — and, where a value relates to a hazard class, the units the criteria for that class are written in. This is why kinematic viscosity is asked for in mm²/s and not in centipoise.

Method. Where the method changes how the number should be read, name it. Annex 4’s own example is open-cup versus closed-cup for flash point. The same applies to a calculated value: say that it was calculated rather than measured, which matters most for the partition coefficient, where a QSAR estimate and a measured log Kow are routinely printed as if they were the same kind of fact.

For mixtures there is a fourth rule. Where valid data exists for the mixture as a whole, give it. Where it does not, data for the most relevant ingredient may be given instead — but the sheet has to say which ingredient it applies to. An unattributed flash point on a mixture sheet is unusable, because the reader cannot tell whether it describes the product or its most volatile component.

The US list after 2024

The OSHA table used to look nothing like this. The 2024 Hazard Communication final rule replaced it, and Appendix D now asks for the same eighteen lettered items in the same order — physical state through particle characteristics. OSHA’s stated reason for adding particle characteristics was that it identifies exposure issues the previous format did not address, nanomaterials in particular.

Two differences survive. Item (o) is “Vapor pressure (includes evaporation rate)”, so the US keeps a property the EU moved out of 9.1. And there is no 9.2 at all — a US sheet has eighteen headings and stops, which means the physical-hazard test data and the supplementary safety characteristics have no prescribed place to go. In practice, sheets written for both markets carry the EU structure, because the EU structure has room for everything the US structure asks for and the reverse is not true.

How section 9 connects to the rest of the sheet

Almost nothing in section 9 is read for its own sake:

• The flash point and the explosion limits are what section 5 is reasoning from when it chooses extinguishing media and tells the crew what the fire will do.
• The vapour pressure decides whether the airborne limit in section 8 is a live concern or a formality — a substance that cannot reach its limit by evaporation is a different problem from one that can.
• The decomposition temperature and the oxidising and explosive test data are the measured basis for the conditions to avoid and the incompatibilities in section 10.
• The physical state, the flash point and the temperature limits feed the storage conditions in section 7 — and in the US, the boiling point does so through an explicitly separate route.
• Kinematic viscosity is the property that decides the aspiration hazard classification appearing in section 2 and discussed in section 11.
• Water solubility and the partition coefficient reappear in section 12, where they drive mobility and bioaccumulation.

What a good section 9 looks like

Three tests, all answerable by reading:

1. Are there eighteen lettered entries and no blanks? Every item present, and each one either a value, “not applicable” or “not available” — with a reason where one exists. A blank is the single most common defect in this section and the easiest to see.
2. Can each number be reproduced? Units, conditions and — where it changes the reading — the method. A flash point without open-cup or closed-cup, a pH without a concentration and a vapour pressure without a temperature are all unverifiable.
3. Does it agree with section 2? The physical hazard classes claimed at the top of the sheet have to be supported by the data here, and the data here has to be reflected up there. Where the two disagree, the sheet has not been reviewed as a document.