BRENNEN

Industries / Mining and materials / Chemicals and petrochemicals / Refining / Lubricants

Additives

Additives are the chemicals blended into base oil to keep engines clean, cut wear, slow the oil's breakdown and stop it from thinning too much as it heats.

6

Milestones achieved, from 1937 to 2019.

1

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1937

    A polymer holds viscosity

    Oil thins as it heats, yet an engine needs it to flow on a cold start and still hold a film when hot. Herman Bruson of Rohm and Haas filed a patent in 1934, granted in August 1937, for adding small amounts of oil-soluble polymerized esters of the acrylic acid series to lubricating oil. It reports that about 1% of the polymer "raises the viscosity index about 30 points", and that the polymers also lower the pour point, the temperature at which oil stops flowing.

    patents.google.com

    Credit

    Herman A. Bruson, Rohm and Haas

  2. 1944

    Zinc salts protect bearings

    Engine oil oxidizes in service and forms acids that attacked the cadmium-silver and copper-lead bearings of the time. Herbert Freuler of the Union Oil Company of California filed a patent in 1941, granted in December 1944, for oil containing zinc salts of octyl thiophosphates, which it says greatly improve the oil's resistance to oxidation and inhibit the conditions that corrode those bearings. These zinc dialkyldithiophosphates (ZDDP) were later found to lay down films on rubbing metal that cut wear. Hugh Spikes's 2004 review of ZDDP dates its first use in engine oils to the late 1930s, and concludes that much about how its anti-wear films form is still unknown.

    patents.google.comdoi.org

    Credit

    Herbert C. Freuler, Union Oil Company of California

  3. 1965

    Sludge held in suspension

    Engines driven on short, cold, stop-and-go trips collect water and dirt that form sludge. William Le Suer and George Norman of the Lubrizol Corporation filed a patent in 1959, granted in March 1965, for products of large succinic acids or anhydrides reacted with ethylene polyamines, useful "as a dispersing agent" that keeps sludge from settling out, especially in crankcase oils. The patent names metal phenate and sulfonate detergents as the earlier answer; the new dispersants contain no metal, so they leave no ash when burned, and are known as ashless dispersants.

    patents.google.com

    Credit

    William M. Le Suer and George R. Norman, Lubrizol

  4. 1990s

    Phosphorus capped for catalysts

    The phosphorus in ZDDP has a cost. ILSAC, a committee of carmakers, capped phosphorus in its first passenger car oil standard, GF-1 (withdrawn in 1997), at 0.12% by mass, because "engine-oil-derived phosphorus poisons emission control devices" and no test for that poisoning existed. Later standards lowered the cap to 0.10% and then, in GF-4, to 0.08%, with a floor of 0.06%. API's licensing manual still notes that some carmakers are concerned that oils with more than 800 parts per million (0.08%) of phosphorus can harm catalytic converters in gasoline cars.

    LessonAn additive that protects the engine can damage the exhaust system downstream of it.

    apidam.widen.net

    Credit

    ILSAC and the American Petroleum Institute

  5. 2012

    Detergent linked to pre-ignition

    Small turbocharged gasoline engines with direct fuel injection can suffer low-speed pre-ignition (LSPI), in which the mixture fires before the spark at low speed and high load. Earlier studies had suggested that droplets of oil from the gap between piston and cylinder wall ignite on their own. Kazuo Takeuchi and colleagues at Toyota reported in 2012 that the oil's formulation had large effects on LSPI in a prototype engine: the temperature at which each oil ignited by itself, measured by high-pressure differential scanning calorimetry, tracked how often LSPI occurred, and both base oils and metal-containing additives could make it better or worse. In 2014 a Nihon University team, Tomomi Miyasaka and colleagues, found in a single-cylinder engine that adding more of a calcium salicylate detergent, of the kind typically mixed into engine oil, made autoignition come earlier and abnormal combustion more severe.

    LessonAn additive chosen for one generation of engines can cause a new problem in the next.

    doi.orgdoi.org

    Credit

    Kazuo Takeuchi, Kosuke Fujimoto, Satoshi Hirano and Minoru Yamashita (Toyota); Tomomi Miyasaka and colleagues (Nihon University)

  6. 2019

    An additive plant burns

    On the night of 26 September 2019 a fire destroyed part of Lubrizol's chemical plant in Rouen, France, and the warehouses of its neighbor Normandie Logistique. Lubrizol makes lubricant additives for engine and industrial oils, among other chemicals. No one was killed or directly injured, but the smoke plume spread far beyond the zones planned around the site. A French Senate inquiry reported in June 2020 that the authorities could not get real-time information on where products were stored at the plant, an upper-tier Seveso site (the strictest class under the EU's rules on major chemical hazards), or what they were, sometimes because operators did not know their own stocks precisely; it called a mandatory, accurate and current stock inventory a priority. Court-appointed experts said in December 2020 that the fire most likely started on Lubrizol's site, though they could not locate its origin precisely.

    LessonA chemical site has to know exactly what it stores and where, in a form responders can use the same night.

    senat.frsenat.frtendanceouest.comberkshirehathaway.com

    Credit

    the French Senate commission of inquiry, rapporteurs Christine Bonfanti-Dossat and Nicole Bonnefoy

Watch for

  1. Oils for hydrogen engines

    Engines that burn hydrogen make water, much of which ends up in the oil, and some are prone to pre-ignition. Afton Chemical says that in its single-cylinder tests gasoline LSPI results did not predict hydrogen pre-ignition (the oil that did worse on gasoline did better on hydrogen), that a seven-day emulsion stability test predicts hydrogen engine performance better than the usual 24-hour test, and that water separating from the oil can leave too little lubricity to stop wear. Infineum says hydrogen engines will need optimized or purpose-made oils, and has extended the industry's water emulsion test, ASTM D7563, to 168 hours without its ethanol fuel component. An internal project at Southwest Research Institute, run from July to November 2025, found no statistically significant difference in pre-ignition among three oils of similar volatility and different calcium content in its hydrogen engine.

    Achieved whentests outside the additive companies, on several oils in more than one hydrogen engine, show repeatable differences in pre-ignition or water handling that follow the oil's formulation.

    lubesngreases.cominfineuminsight.comswri.org

    Claimed by

    Afton Chemical, April 2026; Infineum, May 2026

Open

  1. 01

    Additive test results, pooled

    An open, machine-readable table of published bench test results for lubricant additives. Each row gives the additive's chemical structure in SMILES (a text code for molecules), its concentration, the base oil, the test machine and its load, speed, temperature and time, and the measured friction coefficient and wear scar. Rows are pulled from papers by an AI document reader and every value is checked by a person. Add a baseline model that predicts wear from structure and conditions, scored on additive families it has not seen.

    The brief
    Credit · unclaimed
  2. 02

    Seven-day water test

    A public test of how retail engine oils hold water over a week, the property additive makers now say matters for hydrogen engines. Run ASTM D7563 without its ethanol fuel component and extend it to 168 hours, as Infineum describes, on oils bought off the shelf. Photograph each sample on a timer and measure any free water layer with a short image-analysis script. Publish which oils separate, and when.

    The brief
    Credit · unclaimed

Open 01 · the brief

Additive test results, pooled

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

An open, machine-readable table of published bench test results for lubricant additives. Each row gives the additive's chemical structure in SMILES (a text code for molecules), its concentration, the base oil, the test machine and its load, speed, temperature and time, and the measured friction coefficient and wear scar. Rows are pulled from papers by an AI document reader and every value is checked by a person. Add a baseline model that predicts wear from structure and conditions, scored on additive families it has not seen.

Why it is still open

The largest literature-built set found, by Jianfang Liu and colleagues in 2024, holds 779 results for 79 anti-wear additives, all in rapeseed oil; its appendix tables the results, but the structures appear only as images and no data file was found. A 2025 set built the same way by Oussama Cherguy and colleagues holds about 4,100 points, but for dry friction of diamond-like carbon coatings, with no lubricant. Tribocollect, an older web database of tribological measurements, charges for full access, and researchers at the Karlsruhe Institute of Technology wrote in 2022 that it was unclear whether it was still maintained. Searches of GitHub, Hugging Face and Zenodo in November 2026 found single-study data sets and no pooled one covering mineral, PAO and ester base oils.

pmc.ncbi.nlm.nih.govdoi.orgdoi.org

Built with what exists

Crossref and OpenAlex to list papers; an AI model to read results tables, each value checked by hand (the Note "Seven more flagged" covers an AI reviewer of long documents); RDKit to check structures; scikit-learn or LightGBM for the baseline; Zenodo to publish under CC BY.

  1. First step

    Take 20 papers on ZDDP and the ashless additives proposed to replace it, extract their results by hand, and settle the field list.

  2. Done looks like

    At least 5,000 results from at least 200 papers across mineral, PAO and ester base oils, a random 5% re-checked by a second person at 98% field accuracy or better, and the baseline's error on held-out additive families published.

Open 02 · the brief

Seven-day water test

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public test of how retail engine oils hold water over a week, the property additive makers now say matters for hydrogen engines. Run ASTM D7563 without its ethanol fuel component and extend it to 168 hours, as Infineum describes, on oils bought off the shelf. Photograph each sample on a timer and measure any free water layer with a short image-analysis script. Publish which oils separate, and when.

Why it is still open

Afton and Infineum say week-long tests separate oils that pass at 24 hours, but their articles name no products and give no per-oil results. A 2025 study by Nur Aisya Affrina Mohamed Ariffin and colleagues mixed three commercial oils with 5 to 40% water, but checked the emulsions when fresh and after three months, not over a week. Searches in November 2026 found no per-product 168-hour results.

infineuminsight.comdoi.org

Built with what exists

The ASTM D7563 method; a laboratory blender, graduated cylinders and deionized water; a phone or webcam on a timer; OpenCV to measure layer heights in the photos.

  1. First step

    Test five oils at 24 and 168 hours, twice each, to check that the method gives the same answer both times.

  2. Done looks like

    At least 30 retail oils of different brands and viscosity grades, each tested three times at 24 and 168 hours, with photos, layer heights and the method's repeatability published.