OCP Material Compatibility Workstream
Through its participation in the OCP Material Compatibility Workstream, fluidiQ by Magnalytix helped build and continues to help maintain the open-source Material Compatibility Database, covering wetted material compatibility across single-phase and two-phase dielectric fluid families.
The work at a glance
- Compatibility records
- 570
- Across single-phase and two-phase dielectric fluid families
- Component profiles
- 95
- Application and material combinations from the restructured appendix
- Fluid families
- 6
- GTL and PAO synthetic hydrocarbons, synthetic and natural esters, biobased renewable hydrocarbon, PFCs
- Failure mechanisms
- 8
- Six direct and two indirect interaction mechanisms
Research question
Which wetted materials are compatible with which dielectric immersion fluids, and by what failure mechanism do incompatible pairs fail?
Method
Compatibility observations for seals, gaskets, metals, plastics, cables and other wetted components are collected against dielectric fluid families and classified by risk level and failure mechanism. The source of record for the data and its direct and indirect failure-mechanism classifications is the published OCP whitepaper, Material Compatibility in Immersion Cooling, Version 1.0.
Findings
- The workstream identifies six direct and two indirect interaction mechanisms by which wetted components fail in immersion service, including dissolution, absorption and swelling, and environmental stress cracking.
- The database is openly available to the community under CC-BY-4.0 and is searchable, filterable and exportable through the OCP Material Compatibility Dashboard.
Assessed risk across the dataset
All 570 records in v3.0.0. "Not tested" is reported rather than hidden, coverage is part of the picture.
- Low risk35763%
- Moderate risk6812%
- Elevated risk81%
- High risk376%
- Not tested10018%
570 records total.
Limitations
- The workstream is actively revising the whitepaper toward Version 2.0, so classifications and test conditions may evolve with that release.
- Failure-mechanism tags were backfilled only where remarks unambiguously named a mechanism, and are a starting point pending maintainer review rather than an authoritative annotation.