What a $4,800 Batch of Cracked Parts Taught Me About Material Selection

The Phone Call That Started It All

Back in early 2018, I was handling material sourcing for a mid-sized medical device manufacturer. We were developing a new portable diagnostic unit, and the housing needed to withstand repeated sterilization cycles. I thought I had it figured out. I'd worked with polypropylene before. How different could it be?

Pretty different, as it turned out.

The First Mistake: Assuming 'Plastic' Is Just 'Plastic'

People outside the industry often assume plastic is plastic. You pick a color, you pick a price, and you're done. That assumption cost me about $4,800 on my first big order.

We'd specified a commodity-grade polypropylene for the housings. It was cheap—$0.08 per cubic inch. The supplier's sample looked fine. We approved it, placed an order for 500 units, and waited. When the parts arrived, they looked great in the box. But after three autoclave cycles? Cracks. Everywhere. Not just cosmetic—full-on structural failures.

From the outside, it looked like a simple spec. The reality is that material selection in medical applications involves trade-offs you can't see on a data sheet. Things like creep resistance, chemical compatibility with disinfectants, and long-term fatigue under cyclic loading. I learned that the hard way.

Trying to Fix It the Wrong Way

My first instinct was to just switch suppliers. Find someone who could mold the same PP at a tighter tolerance. Maybe the problem was just bad workmanship, right?

Wrong.

I wasted another two months and roughly $1,200 in prototyping fees before I realized the material itself was the issue. Commodity PP wasn't designed for repeated steam sterilization. I had assumed 'standard' meant the same thing to every vendor. Turned out each had slightly different interpretations—but none of them could make PP work for this use case.

That's when I started looking seriously at engineered thermoplastics. I'd heard of Celanese before—they had a reputation for technical support and a broad portfolio. But I'd always assumed they'd be way more expensive. In my head, 'specialty' meant 'luxury.'

The Turning Point: A Conversation That Changed My Approach

I scheduled a call with a Celanese applications engineer, expecting a sales pitch. Instead, I got a tutorial. They didn't just push their most expensive product. They asked about the actual use case:

• How many sterilization cycles per year?
• What chemicals are used in cleaning?
• What's the expected product lifespan?
• Are there any regulatory certifications needed?

Based on that, they recommended a Celanese EVA-modified formulation that balanced impact resistance with cost. Not the cheapest option they offered, but not the most expensive either. They explained that the total cost of ownership would be lower because (a) scrap rate would drop, (b) the product would last longer, and (c) downstream testing costs would be avoided.

I'll admit—I hesitated. The quote was 40% higher per part than the commodity PP. But I remembered the $4,800 mistake. And I did the math.

Doing the Math: TCO vs. Sticker Price

Here's the rough breakdown I scribbled on a napkin that day:

Commodity PP:
Part cost: $1.20 each
Scrap rate from cracking: ~15%
Additional testing per batch: $800
Warranty claims (estimated): $2.50 per unit
Effective cost per good part: $2.85

Celanese EVA formulation:
Part cost: $1.68 each
Scrap rate: ~2%
Additional testing: $100 (virtually no need)
Warranty claims: negligible
Effective cost per good part: $1.78

The cheaper material was actually costing us more per usable part. And that didn't even include the time lost to troubleshooting and delayed product launches. As the engineer put it: “The value of guaranteed performance isn't the material cost—it's the certainty.”

What I Learned (and What I Now Teach New Hires)

That experience fundamentally changed how I evaluate materials. Now, before I even look at a price list, I ask:

1. What is the application's real performance requirement? Don't assume—test. If sterilizable housings are needed, look at materials like acetal (also known as Celcon from Celanese) or PPS for high-heat applications.
2. What are the hidden failure modes? Commodity HDPE might be fine for a water bottle. But for a medical device? Probably not. (Speaking of which, if you're wondering 'what is HDPE plastic used for'—think bottles, cutting boards, and industrial containers. Not sterilization-chamber duty.)
3. What is the total cost of ownership? Factor in scrap, testing, rework, and warranty risk. The lowest quoted price is rarely the cheapest in the long run.

I also started keeping a checklist. After that incident, I documented every material specification error I encountered. Over the next two years, I logged 47 distinct pitfalls. Some were obvious in hindsight; others would have caught anyone. That checklist is now part of our onboarding for new engineers.

Bottom Line: Don't Learn This the Hard Way

If I could rewind to 2018, I'd still probably start with the commodity PP—because I had to see the failure to truly believe it. But if you're reading this before making a similar mistake, take the shortcut: look beyond the sticker price.

For medical devices, check materials with proven track records—like Celanese's medical-grade TPU or their specialty EVA formulations. For high-temperature environments, PPS or acetal are worth the premium. For prototyping, sure, you can use TPU prints to test fit and feel. But for production? Don't trust the sample until you've validated the full lifecycle.

And if you're still unsure? Call an applications engineer. Ask them the hard questions. It might feel like admitting weakness, but it beats a $4,800 pile of cracked plastic. Trust me on that one.