The Docket · Brief · Working level
Union Carbide v. Commissioner: plant-scale trials and the limits of supply QREs
The Tax Court's 2009 Union Carbide decision held that experiments run on production-scale equipment can be qualified research, but disallowed most claimed supply costs because materials consumed in ordinary production are not research expenses.
Union Carbide Corp. v. Commissioner, T.C. Memo 2009-50, aff'd 697 F.3d 104 (2d Cir. 2012), is the leading case on research conducted inside a working factory. After one of the longest trials in Tax Court history, the court held that experiments run on production-scale chemical plant equipment could satisfy the Section 41(d) four-part test — but it allowed only a sliver of the claimed supply costs, because materials the plants would have consumed in ordinary production were not expenses of the research.
The dispute
Union Carbide claimed research credits for the early 1990s built substantially on experimental programs at its chemical plants — trials of new catalysts, process conditions, and anti-fouling techniques run on full-scale production units because no pilot plant could replicate commercial conditions. The claimed qualified research expenses included enormous quantities of raw-material supplies fed through the units during the trial periods. The IRS argued the projects were routine production and quality control, and that in any event the feedstocks were production costs, not research costs.
The holding
The Tax Court held that several of the sampled projects were qualified research. Where Union Carbide's engineers formed hypotheses about process behavior, designed trials to test them, and evaluated results systematically, the work was a process of experimentation even though it occurred during — and simultaneously produced — commercial output. Other projects failed as routine data collection or quality control excluded under Section 41(d)(4).
On supplies, the taxpayer lost almost entirely. The court allowed only the extra supply costs the experiments caused — additional catalyst, materials lost to experimental variation — and disallowed the base quantities of feedstock that would have been consumed producing sellable product regardless of any experiment. The Second Circuit affirmed.
The reasoning that matters
Two threads carry the decision's weight. First, the court refused to treat "production environment" as disqualifying. Treas. Reg. §1.41-4 asks whether substantially all of the activities constitute elements of a process of experimentation; a trial does not stop being an experiment because the reactor is also filling orders. What mattered was the record: Union Carbide could show hypotheses, protocols, measured variables, and evaluation — the anatomy of a scientific method.
Second, the court applied a but-for discipline to supplies under Section 41(b)(2). Expenses qualify only to the extent the research caused them. A plant running a catalyst trial while producing 100 tons of ethylene amine it would have produced anyway cannot convert the whole feedstock bill into a QRE. The research credit rewards the cost of experimentation, not the cost of operations that host an experiment.
What it means for claims today
Manufacturers remain the direct beneficiaries: first-run trials, process-improvement experiments, and scale-up work on commercial lines can qualify, and examiners generally concede the point when the experimental record exists. But Union Carbide is also the authority behind the standard exam adjustment to supply QREs — the demand that taxpayers isolate incremental trial costs from ordinary production consumption. Claims that sweep entire production-run material costs into supplies invite the same result Union Carbide got: qualified projects, decimated dollars. The IRS research credit page and Form 6765 instructions reflect this project-and-expense discipline.
The case rewards contemporaneous engineering documentation more than any other single factor. Trial plans, run sheets, and evaluation memos won the qualification battle; their absence on the supply-allocation side lost the dollar battle.
Related cases on the site
Union Carbide anchors the four-part-test line mapped in the research credit case law map. For the test itself, see the four-part test explained. Suder applies the same framework to product development, and Siemer Milling shows the outcome when the experimental record Union Carbide had is missing. For handling the supply-cost adjustment on exam, see research credit audit defense.
Frequently asked questions
- Can production trials qualify for the research credit after Union Carbide?
- Yes. Union Carbide v. Commissioner, T.C. Memo 2009-50, held that experiments conducted on production-scale equipment during commercial operation can be qualified research when the taxpayer is systematically testing hypotheses about a process improvement. The experimentation, not the setting, is what matters under Section 41(d).
- Why did Union Carbide lose most of its supply expense claim?
- The Tax Court allowed only the incremental supply costs attributable to the experiments. Raw materials the plant would have consumed in ordinary production anyway were not expenses of the research, so hundreds of millions of dollars of claimed supplies were disallowed even where the underlying trials qualified.