Industries / Mining and materials / Chemicals and petrochemicals / Refining / Lubricants
Blending and grease making
Blending mixes base oils and additives into finished oils, cooks base oil with a thickener to make grease, and fills the result into bottles, drums and bulk tanks for sale.
4
Milestones achieved, from 1909 to 2022.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1909
Castrol goes on sale
Charles Cheers Wakefield founded the Wakefield Oil Company in London in 1899 to sell lubricants. In 1909 it began making Castrol, a motor lubricant named for the castor oil it was made with, and it went on to develop oils for each new use of the internal combustion engine: cars, motorcycles and aircraft. Castrol oil was used on pioneering flights across the Atlantic in 1919.
Credit
Charles Cheers Wakefield, C. C. Wakefield and Co
1984
Oil in plastic bottles
Quaker State became the first motor oil sold in plastic bottles in 1984, according to a company timeline that the Associated Press published in 1997.
Credit
Quaker State Corp
1985
Calcium sulfonate complex grease
Ron Muir and William Blokhuis of Witco Chemical filed a patent in September 1983, granted in December 1985, for a grease built on overbased calcium sulfonate (a detergent that carries fine particles of calcium carbonate) with calcium borate and a calcium soap. Their example grease did not turn fluid at temperatures approaching 600 °F (about 315 °C) and resisted water, and they offered it as a multipurpose rival to lithium complex, aluminum complex and polyurea greases. NLGI's production survey for 2024 found calcium sulfonate grease output still growing year on year, led by North America, China and Europe.
Credit
Ron Muir and William Blokhuis, Witco Chemical Corp
2022
Batteries outbid grease for lithium
Lithium-ion batteries became the largest use of lithium, and the price of lithium hydroxide, the raw material for lithium soap, rose about sevenfold from 2021 to 2022 before easing back toward 2021 levels, Lube Report found. Lithium soap greases, patented by Clarence E. Earle in 1942, had held more than 70% of the world grease market for many years, by STLE's account. Grease makers moved away: in the National Lubricating Grease Institute's (NLGI) production survey, lithium and lithium complex greases fell from 70% of world output in 2020 to 58% in 2023, while calcium greases rose from 15% to 24% and polyurea greases from 7% to 10%. Gareth Fish, a technical fellow at the additive maker Lubrizol, said approving a new grease for one component can cost more than $250,000 and take three to five years, which slows any switch. NLGI warns that falling participation makes its survey a rough gauge.
LessonA raw material shared with a faster-growing industry can be priced out of its older use, and the cost of approvals makes the change slow.
Credit
NLGI's production survey; analysis by Lube Report
Watch for
Immersion cooling in Singapore
Keppel and Shell Singapore planned a six-month pilot, starting in May 2026 at one of Keppel's operating data centers, in which servers sit in Shell's gas-to-liquids (GTL) based immersion fluid, an oil that carries heat away but does not conduct electricity. Shell says such cooling can cut energy use by up to 48% and capital and operating costs by up to 40% compared with air cooling; the figures come from Shell's own brochure. Keppel said its data centers are mostly designed for air cooling.
Achieved whenmeasured energy use and fluid condition from the pilot are published and checked by someone outside Shell, or Keppel builds immersion cooling into a new data center.
Claimed by
Keppel and Shell, April 2026
Open
01
Grease families tested blind
A public, blind bench comparison of commercial multipurpose greases from each main thickener family: lithium, lithium complex, calcium sulfonate complex, polyurea and anhydrous calcium. An accredited lab runs the same standard tests on each: dropping point (ASTM D2265, the temperature at which the grease drips), worked penetration (ASTM D217, how stiff it is), water washout (ASTM D1264), roll stability (ASTM D1831), four-ball wear (ASTM D2266) and oxidation stability (ASTM D942). Each product's measured values are also set beside its own data sheet.
The briefCredit · unclaimed02
Open blend viscosity data
An open dataset of small lab blends, made from a few base oils of API Groups I to IV with a viscosity index improver and a pour point depressant at stepped treat rates, with kinematic viscosity at 40 °C and 100 °C and density measured on every blend. Use it to test how well the standard blending equations (ASTM D7152) and a machine-learning model trained on the data predict blends they have not seen.
The briefCredit · unclaimed
Open 01 · the brief
Grease families tested blind
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public, blind bench comparison of commercial multipurpose greases from each main thickener family: lithium, lithium complex, calcium sulfonate complex, polyurea and anhydrous calcium. An accredited lab runs the same standard tests on each: dropping point (ASTM D2265, the temperature at which the grease drips), worked penetration (ASTM D217, how stiff it is), water washout (ASTM D1264), roll stability (ASTM D1831), four-ball wear (ASTM D2266) and oxidation stability (ASTM D942). Each product's measured values are also set beside its own data sheet.
Why it is still open
NLGI's survey records only production shares, which shifted fast after 2020, and users who change thickener test privately, at up to $250,000 and three to five years per component approval by Lubrizol's estimate. The open grease data found in November 2026 cover lab-made vegetable greases (Kozdrach, Zenodo, 2026) and a 10-row table of typical "archetype" greases on GitHub (August 2026). The YouTube channel Project Farm compared three retail greases in April 2022 for film strength, water resistance, adhesion, corrosion and cold-weather performance, on its own rigs and not by the standard methods. In 2014 Chuck Coe of ExxonMobil reported tests of nine commercial greases, identified only by number, that set the top temperatures claimed on their data sheets against dropping point, oxidation and two bearing-life tests, and found that claims resting on different tests varied widely; it covered heat resistance only, and only lithium-complex and polyurea greases. Web searches found no independent test of commercial greases across all the thickener families by the standard methods.
lubesngreases.comstle.orgzenodo.orggithub.comlube-media.ambrit.combangshift.com
Built with what exists
Greases bought from retailers and distributors; an ISO/IEC 17025 accredited grease lab; blind sample codes; a spreadsheet; Zenodo to publish the data with a data card.
First step
Buy two NLGI grade 2 greases from each family and get quotes from two accredited labs for the six tests.
Done looks like
At least 40 greases, at least 8 from each of the five families, each tested on all six methods, with at least 10% of the tests repeated at a second lab, published under an open license together with the data-sheet comparison.
Open 02 · the brief
Open blend viscosity data
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An open dataset of small lab blends, made from a few base oils of API Groups I to IV with a viscosity index improver and a pour point depressant at stepped treat rates, with kinematic viscosity at 40 °C and 100 °C and density measured on every blend. Use it to test how well the standard blending equations (ASTM D7152) and a machine-learning model trained on the data predict blends they have not seen.
Why it is still open
Blending by computer model is industrial practice. ExxonMobil patented a system that relates a blend's properties to its components and adjusts the recipe from lab results and tank sensors (priority 2006, granted 2010), and Chemspeed sells a workstation, FORMAX LUBES, that blends up to 60 or more ingredients and measures viscosity unattended. Zhmud and colleagues described an AI platform for formulating crankcase oils from component data in a 2022 SAE paper. The measurements stay with the companies. Searches of Zenodo and GitHub in November 2026 found no open set of measured lubricant blend viscosities; the closest GitHub set lists typical values, not measured blends.
Built with what exists
An automated Stabinger viscometer that follows ASTM D7042, such as the Anton Paar SVM 3001, at a university lab or rented; a precision balance; optionally a low-cost robot arm taught by demonstration to move vials, such as the SO-101 with Hugging Face's LeRobot software (see the Note "Five demonstrations per task"); Python with scikit-learn; Zenodo.
First step
Blend 20 samples by hand from two base oils and one viscosity index improver, measure them, and compare the results with ASTM D7152's predictions.
Done looks like
At least 1,000 blends with both viscosities and density, published under an open license, and a table of prediction error, in percent of measured viscosity at 100 °C, for ASTM D7152 and the trained model on at least 100 held-out blends made from base oils the model never saw.