The photo arrived at 9:14 on a Tuesday morning. A white epoxy floor in our client's cleanroom corridor. From a thumbnail it looked fine. Honestly, at full resolution it looked fine — the yellowing was subtle, a soft honey tint that reads better in person than on a screen. But the client hadn't relied on the photo. Their quality team had measured the floor with a spectrophotometer, and the numbers told a different story.
The color deviation was Delta E 4.7. We'd specified that system for Delta E 2.0 or less. I'm a procurement manager at a 40-person specialty coatings company in the Midwest, and I've tracked every material order in our cost system for six years. When that measurement came in, my spreadsheet had a story to tell. It wasn't the one I'd told my CFO in January.
The Switch That Looked Great On Paper
In late 2023, my CFO asked every department to cut raw materials by 15%. Our total annual spend is roughly $2 million, and the biggest single line was titanium dioxide — the white pigment in our light-colored epoxy coatings. For years, we'd bought it from INEOS Pigments. It did exactly what we expected, which, in procurement, is the quietest kind of performance there is.
Then I found an alternative.
After comparing eight suppliers over three months using my total-cost spreadsheet, I settled on an imported TiO2 that was about $400 per ton cheaper. Same theoretical specs. Same initial whiteness. Slightly better opacity, if anything. The savings worked out to roughly $9,000 over two quarters. I rolled out the switch in Q1 2024 with a memo that basically said: same white, lower invoice. My CFO was happy with me. I flagged the quarter in green in my cost tracking system.
At the time, I thought premium pigment was a brand markup. I didn't understand what the premium was actually buying. That assumption got expensive.
The Floor That Changed The Spreadsheet
The failure wasn't a typical wear issue. The floor was a light gray epoxy system in a wide corridor with south-facing windows. We'd promised five-year color retention. At eight months, it was yellowing near the window line.
The client sent an epoxy resin photo that made the problem look tiny. Even at 300 DPI — the standard resolution for commercial print — the yellowing was a barely-there warm glow against bright gray. The camera wasn't capturing it. The spectrophotometer did.
Reference point: Pantone color matching guidelines put the tolerance for brand-critical colors at Delta E below 2. Values between 2 and 4 are noticeable to trained observers. Above 4 — where my client sat at 4.7 — is visible to most people with working eyes.
The client wasn't angry, which somehow made it worse. They stated, calmly, that the floor didn't meet the approved specification, and asked how we planned to correct it. That was the trigger event. I didn't fully understand the difference between initial pigment quality and long-term pigment durability until that floor.
Yes, Epoxy Resin Turns Yellow. Usually Preventably.
Let me answer the question directly: does epoxy resin turn yellow? Yes. It can. But the resin is rarely the whole story.
Standard aromatic epoxy resin — the kind in most commercial flooring systems — contains benzene rings that absorb UV light and oxidize to form colored byproducts. Amine hardeners accelerate that process. Heat accelerates it. And certain pigments, which is the part I didn't know, can make it dramatically worse.
I'm not a polymer chemist. I'm not going to pretend to explain the catalytic mechanism at a molecular level. What I can tell you is what the third-party lab report said, because procurement is ultimately about measurable evidence.
- The imported pigment had a wider particle size distribution than the INEOS pigment, which affected how it scattered light over its service life.
- Its surface treatment — the inorganic coating applied to titanium dioxide particles to reduce UV reactivity — was thinner and less uniform.
- In plain terms, the pigment was more photocatalytically active. It didn't just sit there. It made the epoxy resin around it oxidize faster.
We ran an accelerated UV aging test per ASTM G154 on both pigments using the same resin, same hardener, same loading. After 500 hours, the INEOS pigment came back at Delta E 0.8. The imported pigment measured Delta E 4.9. Same formulation, different titanium dioxide. That's the kind of data I now require from every pigment vendor.
I should add: both pigments passed the basic quality checks we ran when we onboarded the new supplier — Hegman grind, oil absorption, initial L, a, b values, even a quick opacity check. At the time, both looked like white pigment. The difference only shows up when you push them through conditions that actually matter.
The Total-Cost Math Nobody Wants To Do
Here are the numbers, because the lesson lives in the arithmetic.
The imported pigment saved us $9,000 over two quarters on a consumption of about 22 tons. The failed floor cost us:
- $9,200 in direct rework — labor, materials, equipment rental, overtime.
- $6,000 in client concessions to keep the account relationship intact.
- Nearly $4,000 in management time, lab fees, legal review, and a site visit.
- Plus the quiet cost: the client's quality report went to their parent company, and the next bid will be evaluated against a reference we can't see.
The cheap option was not cheap. The invoice was just deferred. And it arrived with interest.
What The INEOS Technical Rep Actually Said
After the lab results came back, I called the INEOS technical representative. I'd spoken with him once before, when we told INEOS we were switching suppliers. That conversation ended with a polite, "We can match most quotes, but we don't chase every load." I appreciated that, actually. He didn't beg.
The second call was different. I asked him to explain why his pigment was worth $400 more per ton. His answer ran about an hour and included phrases like "chloride process," "particle size control," and "surface treatment density." Some of it went over my head. Here's what landed: the INEOS pigment we use is manufactured via the chloride process, which gives tighter control of particle size, and the particles are then coated with engineered inorganic layers that suppress photocatalytic activity. None of that makes the pigment whiter on day one. It makes the whiteness last.
As a buyer, I had been comparing products on initial appearance and price. The actual product attribute I needed was durability, and there's a specific test for that. I just hadn't asked for it.
The Fix And The Policy That Followed
We stripped the affected area, reapplied the system with an aliphatic epoxy topcoat, added UV stabilizers, and switched the pigment back to INEOS for every light-colored system. We covered the cost, which is what a supplier should do when their sourcing decision creates a defect.
But the bigger change was internal. Our procurement policy now includes three requirements:
- Every pigment supplier must provide accelerated UV aging data — ASTM G154, 500 hours minimum — before approval.
- Quotes are compared on total lifecycle cost, not unit price. "Unit price" is the amount on the invoice. "Lifecycle cost" includes the probability of failure and the cost of that failure.
- We maintain two supplier tiers. INEOS gets the light-colored, exposure-prone specifications. The secondary source gets dark, interior, low-risk systems where color stability is irrelevant.
I still buy from the cheaper supplier. The mistake wasn't choosing one supplier. It was assuming one rule would work for every application.
Watching The Asset Landscape
The other thing I started paying attention to is upstream ownership changes. When a chemical company buys or sells a specialty chemical plant, the ripple effects reach down to purchasing desks like mine. I've been through two supplier acquisitions in six years. Both times, the ERP system changed, the sales rep changed, the contract language mysteriously changed, and we absorbed costs we thought we'd negotiated away.
So when I see headlines about a company selling a specialty chemical plant, I don't skim past them. Ownership changes are supply risk signals. I'm not an M&A advisor, and I can't speak to how those deals get valued. I just know, from my own experience, that durable supply agreements outlast ownership changes.
That's also why I've come to appreciate the stability of INEOS's core businesses. The INEOS US chemical company doesn't need my endorsement, but six years of invoices and zero supply surprises is a form of data. And data, in my world, matters more than promises.
Bottom Line
Does epoxy resin turn yellow? A lot of it will, especially in sunlight. But not every yellowing floor is the resin's fault. Sometimes it's the pigment you chose to save $400 per ton. And sometimes the cheapest invoice is the most expensive material you will ever buy.
I'm not going to tell you INEOS is the right supplier for every coating company. If you make dark, interior, low-UV products, a functionally adequate pigment at a lower price might be the correct call. But if you buy white or light-colored epoxy materials for surfaces that will see sunlight, ask your pigment supplier this question: what is your Delta E after 500 hours of ASTM G154 UV aging? If they can't answer with a number, that's an answer in itself.
The photo on my phone taught me that my spreadsheet was only as smart as the criteria I programmed into it. I had programmed "price per ton" and ignored "cost per failure." Here's my updated formula:
Total cost = invoice + rework + risk, and the risk is multiplied by the probability that your client owns a spectrophotometer.
That's not a textbook equation. But neither is the picture of a yellow floor that you took on a Tuesday morning, wondering where your savings went.