Why PEEK and Ultem blanks warp after cutting

A customer cuts a clean 12x12 in PEEK blank, checks it flat off the saw, and comes back the next day to find it's bowed by a few thousandths. Nothing was wrong with the cut. This is one of the most common support questions in this business, and it has a simple physical explanation: the material was never flat to begin with, internally.

Sheet stock like PEEK, Ultem, and PPS is extruded or compression-molded under heat and pressure, then cooled. That cooling doesn't happen perfectly evenly through the thickness of the sheet - the surfaces cool faster than the core. The result is a plate that looks flat and behaves flat, but is holding a real internal stress gradient, locked in place by the surrounding material.

Cut a piece free from that sheet, and you remove the surrounding material that was holding the stress in check. The freed piece is now free to relax toward whatever shape relieves that internal stress - usually a slight bow or twist. This isn't unique to plastics; it's the same underlying phenomenon that makes machinists pre-stress-relieve certain metal stock before finish machining. Plastics just show it faster and more visibly.

The practical number to know: a 12x12 in blank in one of these materials can bow somewhere in the range of 0.010 to 0.025 in within about 48 hours of being cut. It usually stabilizes after that - the stress that's going to release, releases, and the part settles into its final shape. This is why we disclose it automatically on every quote line for a residual-stress-flagged material, rather than waiting for someone to ask after the fact.

Whether this matters depends entirely on what the part needs to do. A spacer or a bracket that gets bolted flat against a rigid surface on final assembly usually doesn't care about a few thousandths of pre-assembly bow - it gets pulled flat when it's installed. A part that needs to hold its own flatness unsupported, or that's going into a precision fixture, is a different story.

That's what stress-relief annealing is for. It's a controlled heat cycle - a slow ramp up, a soak at temperature, and a slow cool back down - that lets the internal stress relax before the part ships, instead of after a customer has it in hand. It's not free: it adds oven time and a business day to the lead time, and it's blocked on same-day and next-day tiers because the cycle itself takes longer than that. But for a part that has to hold tight tolerance or sit unsupported, it's usually worth the day.

One thing worth being direct about: annealing reduces the tendency to warp, it doesn't eliminate it outright, and it doesn't turn a stressed material into an unstressed one. If a design absolutely cannot tolerate any post-cut movement, that's a conversation about tolerance and inspection, not just a checkbox at quote time.

If a part in PEEK, Ultem, or PPS is going to be inspected or assembled soon after it's cut, it's worth building the 48-hour settling window into a schedule rather than treating day-one dimensions as final. And if the application genuinely can't tolerate movement, add the anneal option when getting a price - it's the same page, one checkbox, and the tool will tell you the new ship date before you commit to anything.