A 1,000 m3 parcel of white mineral oil is gauged at 20 °C at the loading terminal and again at 35 °C on discharge. The second gauge reads 1,012 m3. Nothing was added and nothing leaked: the twelve cubic metres are thermal expansion. Whether they reach the invoice depends on one line of the contract — the base temperature — and that line is frequently left to whoever writes the shipping documents last.
Volume is not a property of a chemical. It is a property of a chemical at a temperature. Bulk trading is therefore less about chemistry than about agreeing which temperature a number belongs to, and the disputes that reach arbitration are almost never about quality. They are about which of two correct numbers was written on the ticket.
Three base temperatures, three invoices
Three bases are in simultaneous use, and they are not interchangeable:
| Base | Where it comes from | Typical use | 1,000 m³ at 20 °C, expressed at this base |
|---|---|---|---|
| 20 °C | GB/T 1885 | China domestic trade, most Asian chemical contracts | 1,000.0 m³ |
| 15 °C | ASTM D1250 Table 54B, ISO 91 | European and much international practice | 996.3 m³ |
| 60 °F (15.56 °C) | API MPMS Chapter 11.1 | US petroleum and marine custody transfer | 996.7 m³ |
The arithmetic is one multiplication. For a mineral oil with a volumetric expansion coefficient of 0.00075 per °C, moving from a 20 °C base to a 15 °C base multiplies the volume by 0.99625: 1,000 m³ becomes 996.3 m³. At USD 1,200 per cubic metre, 0.375% of the invoice is USD 4,500 on a parcel that has not changed by a gram.
Note also the smaller gap inside the 15 °C family. A 60 °F base sits 0.56 °C below a 15 °C base, and on 1,000 m³ that is another 0.4 m³ — 0.04%. Small, but it does not vanish, and it stacks on top of every other basis choice in the chain.
VCF is an exponent, not a lookup table
Since the 2004 edition, API MPMS Chapter 11.1 / ASTM D1250 has been a calculation procedure rather than a stack of printed tables. The temperature correction is exponential:
CTL = exp[ −α60 · Δt · (1 + 0.8 · α60 · Δt) ]
where Δt is the observed temperature minus the base temperature and α60 is the volumetric expansion coefficient at 60 °F. That coefficient is not a constant — it is derived from the base density through three commodity-specific constants, α60 = K0 / ρ60² + K1 / ρ60 + K2:
| Commodity group | K0 | K1 | K2 |
|---|---|---|---|
| Generalized crude oil | 341.0957 | 0 | 0 |
| Gasolines and naphthas | 192.4571 | 0.2438 | 0 |
| Jet fuels and kerosenes | 330.3010 | 0 | 0 |
| Fuel oils and diesel | 103.8720 | 0.2701 | 0 |
| Lubricating oils | 0 | 0 | 0.34878 |
The standard covers −50 °C to 150 °C and 0 to 1,500 psig, with α60 running from 0.000230 to 0.000930 per °F across the product range. A worked example published with the standard: 35 °API crude at 849.0 kg/m³, 10,000 barrels measured at 85 °F, gives α60 = 0.0004732 per °F and CTL = 0.98813 — 9,881.3 barrels at 60 °F.
Two consequences matter commercially. First, the commodity group is a commercial decision, not a technical footnote: applying the diesel curve to a product that behaves like a jet fuel shifts α60, and the corrected volume moves with it. Second — and this is the useful one for our own range — for lubricating oils K0 and K1 are both zero, so α60 collapses to a single constant independent of density. The entire white mineral oil ladder from 5# to 100# therefore shares one temperature curve, which is why one table has served mineral oils for forty years while gasolines needed several.
Expansion coefficients differ sevenfold across a normal product list
| Liquid | β, per °C | Volume change over 15 °C |
|---|---|---|
| Water (at 20 °C) | 0.00021 | 0.32% |
| Glycerol | 0.00050 | 0.75% |
| Ethylene glycol | 0.00057 | 0.86% |
| Vegetable oils | 0.00068–0.00072 | 1.02–1.08% |
| White mineral oil / paraffin oil | 0.00070–0.00076 | 1.05–1.14% |
| Diesel | 0.00083–0.00085 | 1.25–1.28% |
| Gasoline | 0.00095–0.00110 | 1.43–1.65% |
| Ethanol | 0.00109–0.00112 | 1.64–1.68% |
| Ethyl acetate | 0.00138 | 2.07% |
| Acetone | 0.00143 | 2.15% |
The spread between water and acetone is a factor of nearly seven. The old cargo rule of thumb — fats and oils gain about 1% for every 14 °C of warming — is the same number seen from the other side: 0.00071 per °C.
Two operational consequences follow directly. An ISO tank filled to 98% at 15 °C and carried across a deck that reaches 50 °C gains 0.00075 × 35 = 2.6% and is above 100% on arrival: it either vents product or it arrives with a full ullage and a claim. Acetone over the same 35 °C swing gains 5.0% and must be loaded at 92% or less. Fill limits are not conservatism; they are arithmetic.
Two errors that cancel — and why that is not reassurance
A published worked example compares the two chains on a single jet fuel parcel: 10,000 m³, 15 °C density 0.7869, actual temperature 45 °C. The ASTM chain returns VCF 0.9705 against the 15 °C density. The GB/T 1885 chain first converts the density to 0.7831 at 20 °C and returns VCF 0.9758.
The correction factors differ by 0.55%, which is unsurprising — one of them is correcting from a base 5 °C higher. But the densities differ by 0.48% in the opposite direction: 0.9758 × 0.7831 = 0.76415 against 0.9705 × 0.7869 = 0.76368, a residual of 0.06%.
The temptation is to read that 0.06% as proof the two standards agree. They do not. Two half-percent errors happen to point in opposite directions at that particular temperature and density; move the parcel to 90 °C, or to a heavier product, and the cancellation is only partial. The correct discipline is to pick one chain, name it in the contract with its edition year, and correct densities only inside that chain. Converting a 15 °C density to 20 °C with a rule of thumb and then feeding it to the other table is how 0.06% becomes an argument.
Where the money actually leaks
| Source | Size of the error | Cost on an 855 t parcel |
|---|---|---|
| Air buoyancy (vacuum mass vs mass in air) | 0.0011 g/cm³, flat | 0.13% — about 1.1 t |
| Thermometer, 1 °C | 0.00075 × 1 = 0.075% | about 0.64 t |
| Hydrometer, one division (0.0005 g/cm³) | 0.058% | about 0.50 t |
| Table lookup without interpolation | ≈0.07% combined | about 0.60 t |
Three observations. First, air buoyancy is not an error but a convention: the 0.0011 g/cm³ subtracted in Chinese practice converts vacuum mass to apparent mass in air, and the two answers differ by 0.13% forever. A contract that does not say which basis it uses contains 0.13% of unspecified value.
Second, a one-degree thermometer error costs more than a one-division hydrometer error — 0.075% against 0.058%. Buyers spend money on density meters and use whatever thermometer is in the drawer. If the parcel is billed by volume, that priority is backwards.
Third, the no-interpolation rule in GB/T 1885 is deliberate and it is not free: density is taken at the nearest tabulated row and temperature at the nearest degree, so half a row of density and half a degree of temperature are simply discarded. Each is small; together they are comparable to the instrument errors they sit on top of.
Density is a method, not a number
| Method | Uncertainty | Sample | Limitation |
|---|---|---|---|
| ASTM D1298 / GB/T 1884, glass hydrometer | ±0.0005 g/cm³ | 500 mL | Viscosity must let the float settle |
| ASTM D941 / GB/T 2540, pycnometer | ±0.0001 g/cm³ | 25–100 mL | Slow, but the reference method for viscous products |
| ASTM D4052 / SH/T 0604, oscillating U-tube | ±0.0001 g/cm³, repeatability ≤0.0001 | 1–2 mL | Viscosity below about 15 mm²/s at test temperature, no bubbles |
The gap between the hydrometer and the other two is a factor of five, and it is larger than every other error in the chain. A certificate reading "0.855 g/cm³" therefore specifies nothing: it does not say at what temperature, by which method, or against which base. Ask for the three together and the number becomes reproducible; ask for the number alone and you have bought a rounding.
One detail that catches laboratories: a glass hydrometer is calibrated at 20 °C and the glass itself expands. The correction coefficient is 0.000023 per °C, which at 60 °C amounts to about 0.09% — larger than the instrument's own ±0.0005. Reading a hydrometer hot without the glass correction spends the method's precision before the sample is even weighed.
The six lines a volume contract needs
- Base temperature and standard, with edition. "20 °C per GB/T 1885-1998" or "15 °C per ASTM D1250-19". A base temperature without a standard is half a clause.
- Density: method and temperature. ASTM D4052 at 20 °C is a specification; "0.855" is not.
- Measurement point. Shore tank before loading, ship tank after loading, or meter — three different numbers, and line fill belongs to someone.
- Free water and sediment. Total observed volume minus free water equals gross observed volume, times CTPL equals gross standard volume, minus sediment and water equals net standard volume. Only the last number is the cargo.
- Mass basis. Vacuum mass or apparent mass in air — 0.0011 g/cm³, or 0.13%.
- Temperature measurement. Point, number of readings and averaging rule. A 20 m tank is not isothermal after three days of loading.
Where temperature is also a phase change
For liquids the argument is a percentage. For the higher-melting grades it stops being a percentage at all: lauryl alcohol (CAS 112-53-8) melts around 24 °C, cetyl alcohol (36653-82-4) around 49 °C, and paraffin wax 60# (8002-74-2) around 60 °C. A drum that is pourable in Guangzhou in August is solid in Tianjin in January, and discharge temperature becomes a commercial term rather than a handling detail. Cetyl palmitate (540-10-3), myristyl myristate (3234-85-3) and isocetyl myristate (83708-66-1) sit in the same band and are quoted with a melting point precisely because the buyer needs to know what the drum will look like on arrival.
White mineral oil is the easy case and therefore the useful reference. Its expansion coefficient barely moves across the viscosity ladder — 5# and 100# share one temperature curve — so it is the product on which a buyer can verify their own VCF arithmetic before applying it to esters that expand faster. Butyl oleate (142-77-8) and butyl stearate (123-95-5) sit slightly above the mineral oils; caprylic/capric triglyceride (65381-09-1) and isopropyl myristate (110-27-0) are in the same neighbourhood, and every one of them reads low on a 35 °C dock if the ticket was cut at 20 °C.
What we document
Every bulk and IBC shipment leaves with a COA carrying density by ASTM D4052 at 20 °C, the base temperature used for the volume figure, and the edition of the standard behind it — plus, for solid grades, the melting range from the same batch. If your contract already fixes a base other than 20 °C, tell us at quotation stage and the certificate will be issued on your base rather than converted afterwards.
Related reading on the same delivery chain: how packaging interacts with the product over a sea voyage in chemical packaging compatibility, how a sample becomes evidence in sampling and retained samples, and how the three oxidation indices on a COA are read in peroxide value, anisidine value and TOTOX. The grades themselves are described in white mineral oil purity grades.
Available for quotation: white mineral oil 5# industrial, 22#, 100# and 100# food grade (all CAS 8042-47-5); cosmetic grades 5# and 32# (8012-95-1); esters butyl oleate (142-77-8), butyl stearate (123-95-5), GTCC (65381-09-1), IPM (110-27-0), cetyl palmitate (540-10-3), myristyl myristate (3234-85-3) and isocetyl myristate (83708-66-1); and the melting grades lauryl alcohol C12 (112-53-8), cetyl alcohol C16 (36653-82-4) and paraffin wax 60# (8002-74-2).
Send us the volume, the base temperature you bill on and the discharge port, and we will tell you what the ticket should read at each end.