Why Engine Oil Can Leave the Sump Thinner Than It Went In

Last verified: 2026-07-29

Suppose a sample is drawn from a sump that has been running a wide-spread multigrade, and it reads thinner at operating temperature than the grade on the drum — with no overheating, no coolant, nothing unusual in the way the machine has run. The film every bearing and gear tooth is riding on is now thinner than the one the machine was designed around, and it will stay that way. The oil did not wear out. Something dissolved in it broke. That something is the viscosity-index improver, the long polymer stirred into the base oil so a single blend can satisfy both a cold-crank rating and a hot-running grade — and sustained shear can snap it in half.

What actually thinned?

Heat, wear, and a bad batch are the three suspects a low viscosity result usually draws. Each is reasonable. None of them is what happened here.

Heat does change viscosity, substantially and predictably — but reversibly. That response is precisely what an oil’s viscosity index describes, and it disappears the moment the oil cools; a sample measured cold on a bench has already surrendered any thermal explanation. Wear puts metal into the oil without altering the oil’s own molecules. A bad blend would have read low on day one rather than arriving gradually. What is left is the polymer, and the polymer is the only ingredient in the sump with a backbone long enough to be torn apart by the machine it is lubricating.

Here is the part that catches specifiers out. The one shear-related figure that routinely appears on a finished-oil technical data sheet — high-temperature high-shear viscosity, or HTHS — does not measure this at all. HTHS reports what the oil is doing while shear is being applied to it, a reading that recovers the instant the shear stops. A data sheet can carry a perfectly healthy HTHS value for an oil whose polymer population has already been permanently reduced. The gap between the number that gets published and the damage that actually accumulates is the whole of this subject.

Why does a dissolved polymer snap?

A viscosity-index improver works because it is long. In cold oil the chain sits coiled and compact and contributes little bulk. Carried through a pump gap, an injector orifice, or a loaded bearing clearance, that same chain is pulled open in the flow, and a stretched chain carries tension along its own backbone.

Two different things can then happen, and conflating them is the error this article exists to correct. In the first, the chain stretches, partially disentangles, measures thinner while the shear field acts on it, and relaxes back to its coiled shape the moment shear stops. Nothing is lost. That is temporary shear loss, and HTHS is built to catch it. In the second, the tension exceeds what the backbone can carry and the chain physically ruptures. Molecular weight is now permanently lower; a shorter fragment thickens the oil less per unit of mass than the parent chain did; and the oil’s resting, low-shear viscosity — the figure reported as KV100 — comes back permanently reduced. Nothing relaxes back from this one. The Society of Tribologists and Lubrication Engineers illustrates the scale of it with a worked example in its Lubrication Fundamentals column: permanent loss sufficient to leave a nominal 10W-30 testing as a 10W-20 or lower.

Where along the chain it breaks is not random, and it explains why the effect is so blunt. A fully extended linear chain in a strong flow carries its maximum tension at its own midpoint rather than at either end, so scission clusters there and leaves two fragments of roughly half the parent molecular weight — the midpoint-scission finding published in Nature in 1984. Chain scission does not shave a little off a molecule here and there. It tends to cut it in two.

Underneath sits a trade-off no formulation escapes. A longer, heavier chain thickens oil more efficiently per unit of mass added, which is exactly why it gets chosen; and the same length gives drag-induced tension more backbone to concentrate on, and reaches the stretched, vulnerable configuration more readily than a short chain does.

Thickening efficiency and shear vulnerability are not two properties of a viscosity-index improver. They are bought together, in the same design decision.

It is worth noting how recently that mechanism was tested. The textbook account — a coil that expands as the oil warms and thickens the blend more at temperature — traces to a single 1958 paper by Selby, which argued the case from polymer-solvent theory without physical measurement data behind it. Peer-reviewed work by Covitch and Trickett in 2015 found that coil expansion with temperature is not necessary to achieve significant elevation of viscosity index, though polymers that do expand contribute more than those that do not.

Which rig measures which kind of loss?

On a data sheet, an HTHS value reads as a result from one of three instruments — a tapered bearing simulator (ASTM D4683), a tapered plug viscometer (ASTM D4741), or a multicell capillary viscometer (ASTM D5481) — with the engine-oil condition taken at 150 °C while the sample is under shear. D4741 also has a 100 °C mode, so the temperature belongs in the specification, not just the method number. D4683’s own title is instructive: it covers new and used engine oils, because the same instantaneous reading is taken on both to check whether an oil that has separately degraded still clears its HTHS floor. HTHS is a floor check. It is not a diagnosis of what moved the floor.

The rig that measures permanent damage is a different machine, and its result does not normally reach a data sheet. ASTM D6278 forces the oil through a European diesel injector apparatus for 30 cycles and reports percent viscosity loss at KV100, comparing the sample before and after — measured at rest, with no shear field acting during the measurement itself. ASTM D7109 uses the same injector apparatus and reports loss at both 30 and 90 cycles, the 90-cycle point being the more severe condition.

One rig is regularly mistaken for the other, and the mistake matters when a specification is being written. The KRL test is not a diesel injector method at all: it is a tapered roller bearing rig, run to the DIN 51350-6 and CEC L-45-A-99 designations. Two different machines, both measuring permanent loss, and a number from one is not a number from the other — so a data sheet or approval document should state which rig produced it.

Sitting on top of the D6278 result is ASTM D6022, the standard practice for calculating Permanent Shear Stability Index. PSSI expresses what fraction of the polymer’s own contribution to viscosity was permanently lost, not what fraction of the oil’s total viscosity disappeared — which is why it isolates polymer performance in a way a raw percent-loss figure does not. And STLE states the commercial reality plainly: PSSI is not a typical performance data result provided on product technical data sheets. If a specifier wants it, someone has to ask for it by name.

Does the grade on the drum stay true in service?

No — and that is a documented choice rather than an oversight. SAE J300, the standard that defines every multigrade designation an engine oil carries, states in its own scope that it defines the limits for a classification of engine lubricating oils in rheological terms only, and that other oil characteristics are not considered or included. The current edition, J300_202405 of May 2024, superseding J300_202104, carries that same scope language. J300 tells you which viscosity band a sample falls into at the moment it is measured. It is silent on whether the oil holds that band in service.

Set that beside a sister standard from the same publisher. SAE J306, the automotive gear lubricant viscosity classification — current edition J306_202502, February 2025 — builds a shear-stability requirement directly into its classification scheme, using the KRL test.

J300 could have required a stay-in-grade test. It does not. J306 does.

Enforcement for engine oils therefore sits one layer above J300 rather than inside it. API 1509 Annex F, the API guidelines for SAE viscosity-grade engine testing, names ASTM D6709 — the Sequence VIII test on a CLR spark-ignition engine, which evaluates stay-in-grade viscosity behaviour alongside bearing corrosion — as the reference method, with the 30-cycle D6278 injector test accepted as a bench alternative. The ACEA sequences carry shear stability as a requirement line of its own, listed separately from the HTHS line. A grade classification and a shear-stability requirement are two different documents answering two different questions, and checking the grade has only ever answered one of them.

What decides how much a given oil shears?

Not every multigrade is equally exposed.

Grade spread is the first lever. A wide-spread multigrade has to start from a thinner base blend and be thickened harder at the hot end, and that thickening is the polymer’s work — so a wide spread typically carries a higher polymer treat rate than a narrow one. More dissolved polymer, by the trade-off already described, means more scission-vulnerable material in the sump to begin with. This is consistent engineering reasoning drawn from formulation literature rather than a treat-rate table published by a standards body, and it is worth holding at that strength.

Architecture is the second, and it is the one that overturns the obvious assumption. A star, or radial, polymer distributes its total molecular weight across several shorter arms radiating from a compact core instead of one long backbone. No single arm reaches the fully extended, high-tension configuration that makes a linear chain of equivalent total weight fail at its midpoint. That decouples two things a linear-chain mental model treats as one: total molecular weight, which governs thickening efficiency, and single-segment extension length, which governs vulnerability to scission. Xue, Agarwal and Lemstra reported that resistance for star polymers in elongational flow in Macromolecules in 2005; the full paper sits behind a paywall, so the finding is carried here at the level its abstract supports.

The base stock is the third. A sufficiently high-viscosity-index base oil — Group III+, gas-to-liquids, or PAO — can carry a multigrade on a much lower polymer treat rate, and for some narrower grades that treat rate can reportedly approach zero. Remove the high-molecular-weight polymer population and the population that can be broken goes with it. The route is more readily achievable for narrow spreads than for the widest ones, so it is a formulation strategy rather than a property that comes free with the word synthetic.

None of which turns one low reading into a conclusion. If a used-oil result shows KV100 below the grade’s stated band, that is consistent with permanent shear — and equally consistent with fuel dilution, and with the sump having been topped up using a lighter oil than it was charged with. Contamination that raises viscosity, such as soot or oxidation, points the other way and does not belong in this set. A single viscosity figure cannot separate the three causes that move it downward; only a corroborating result can, and reading the same number harder will not produce one.

How does a specifier pin this down in writing?

Begin from what the data sheet will not carry. PSSI, and the raw D6278 or D7109 percent-loss figure, do not normally sit beside KV100 and HTHS. Where shear stability genuinely governs the duty — extended-drain service, high-pressure injection hardware, a wide-spread grade doing work at both ends of its range — request the D6278 or D7109 result, or the calculated PSSI per D6022, as a named line item in the approval documentation, and require the test method and cycle count to be stated alongside the number. Asking for the qualification data a supplier already holds to license the grade is a different and far more answerable request than asking for a new test.

Name the reference method rather than accepting a bare pass. Ask whether stay-in-grade performance was demonstrated against ASTM D6709 or the accepted D6278 bench alternative under API 1509 Annex F. For an oil built on a high-VI base stock, ask how much of the grade’s high-temperature body comes from the base oil and how much from polymer; that single question tells you more about shear exposure than the grade number ever will. Suppliers of automotive lubricants on the Altonex Global platform can be asked for those figures at the request-for-quotation stage, where a named method is a far more answerable ask than a general enquiry about quality.

And when a sample does come back thin, specify the corroborating measurement before the interpretation. The grade on the drum was a statement about the oil that went in. Whether it still describes the oil in the sump is a separate question, and it has its own test.

This article is general technical information for specifiers and buyers. It is not a recommendation of any product, and Altonex Global does not test, inspect, certify or approve any lubricant or supplier claim.

Frequently asked questions