What I am comparing, and on what terms
I have been the quality and brand-compliance manager at a fabric converter for six years. My job is to review incoming yarn and greige — roughly 200 lots a year — before any of it reaches a cutting table or a customer. In 2024 I rejected about 11% of first submissions, and most of those were not defects. They were spec drift: something sitting at the edge of tolerance that would have become a defect three processes downstream.
So when someone asks me whether recycled polyester is as good as virgin, I usually ask them to be more specific. Virgin polyester (PET) and rPET at the same denier, same intrinsic viscosity, same finish are close enough that the comparison gets boring fast. The comparison that actually matters is between a well-controlled rPET stream and a poorly controlled one — and that gap is much wider than the virgin-versus-rPET gap everyone argues about.
I am comparing five things: spec conformance and batch consistency, mechanical behavior, color consistency, compliance paperwork, and landed cost per approved meter. Every dimension assumes matched spec, because comparing a 0.64 dl/g virgin chip to an unverified rPET chip tells you nothing except that you bought two different things. (And if you searched for what fruit is high in fiber, this is the wrong aisle — textile fiber, not dietary fiber.)
1. Spec conformance and batch consistency
Virgin PET from an integrated producer is predictable for a structural reason: the PTA and MEG going in are controlled, so the intrinsic viscosity coming out sits in a narrow band. In our incoming data, the integrated chains — Reliance is the obvious example in our region, since they run polyester PET through to their own spinning — show a tighter lot-to-lot IV spread than a trader reselling the same nominal grade. That is not a knock on traders. It is just what vertical integration buys you.
rPET variance is a feedstock problem, not a chemistry problem. Bottle bales, post-industrial film, and fiber waste all carry different heat histories, and that shows up in IV spread, diethylene glycol content, and color. Swap one rPET supplier for another and you can see more variation between those two sources than between rPET and virgin overall.
Conclusion: rPET is not inherently inconsistent. It is inconsistently sourced. Ask for the IV standard deviation over the last twelve shipments, not just the certificate of analysis for the lot you are being shown.
I assumed for two years that matching IV meant matching behavior. Didn't verify it. Turned out two lots at 0.64 dl/g ran completely differently on the same texturizing line because the finish pick-up was 0.15% apart. Now the purchase order carries the OPU range, not just the polymer spec.
2. Mechanical behavior
At matched IV and denier, the average tenacity gap I see between virgin and rPET textile yarn is small — somewhere in the 0.2 to 0.5 g/den range, and I am not going to pretend that number holds across every supplier. The average is not where the problem lives. The tail is.
What surprises people is that rPET usually passes the mean and fails the variance. If your weaving or knitting process is set up around a consistent elongation at break, a slightly wider distribution means more ends down and more stop-mark defects — and none of that appears in a single-lot test report. The other place the difference shows up is heat-setting behavior: shrinkage in boiling water and crimp stability drift more in recycled streams, because the polymer's thermal history was never uniform to begin with.
Conclusion: for woven apparel and home textiles, rPET at matched spec is fine and I will approve it. For high-tenacity industrial yarn — geotextile, cord, strapping — it is not there yet at a comparable price, and I would rather say that plainly than approve a lot I will be explaining to a customer in six months.
3. Color consistency (the one that flips)
The assumption I hear most is that rPET is 'yellower.' That is the wrong thing to watch. The failure mode is lot-to-lot b* drift, not a fixed color offset. A fixed offset you can recipe around once. Drift means your dyer is adjusting the recipe every shipment and eventually one of those adjustments lands wrong.
We tracked this across roughly 60 lots in 2024. Recycled lots needed a shade adjustment about one in six times; virgin ran about one in twelve. Nothing catastrophic — but each adjustment is a color-matching hour and a re-dye risk, and one bad correction cost us 1,200 meters and about nine days.
They warned me to put a b* limit on the purchase spec. I didn't listen, because the first three rPET suppliers all came in clean. Then a grey lot arrived at b* 7 against our cap of 4, the dyer corrected for the chip instead of the target shade, and we ate the re-dye. Every rPET spec we write now has a chip color limit in it, and I ask for the last six months of b* data, not a single reading.
Conclusion: recycled polyester is not harder to dye. It is harder to dye predictably, and predictability is a spec problem you can solve with a color limit and a data request.
4. Compliance paperwork
This is the dimension where virgin and rPET are genuinely not comparable, and it is the one most sourcing teams under-resource.
Under the Textile Fiber Products Identification Act (16 CFR Part 303), fiber content must be disclosed by generic name and percentage. 'Recycled polyester' is part of that content statement — verify the current rule at ftc.gov before the hangtag art is signed off.
Then comes the claim itself. The FTC Green Guides (16 CFR Part 260) require recycled-content claims to be substantiated and properly qualified, and the trap is mass balance. If your supplier is buying credits rather than running physically segregated recycled feedstock, you generally cannot put 'made from recycled bottles' on a label. State-level labeling rules add another layer on top. I am a quality manager, not your lawyer, so confirm the current position with counsel — but design the claim and the chain-of-custody document at the same time, because a label reprint has cost me more than the fabric did.
Conclusion: budget QC hours for documentation, not just lab tests. Ask for the transaction certificates before the first shipment, not after the first marketing deck.
5. Landed cost and supply reliability
I am not going to quote you a price that will be wrong next quarter. What I will say is that the rPET premium compressed noticeably between 2021 and 2024 for standard apparel grades in the markets we buy from — take that with a grain of salt, it moves with bale collection rates.
The more useful point is that per-kilo price is the wrong number to compare. Compare landed cost per approved meter. That number includes the re-dye rate, the reject rate, the extra incoming inspection, and the color-matching hours. On two of our programs, rPET looked 6% more expensive per kilo and came out within 1% per approved meter. On one program it looked 4% cheaper and landed 11% worse, because we were buying a cheap stream through a trader with no segregation and no b* history.
Conclusion: virgin is more predictable on price because it tracks PTA, MEG, and crude, and an integrated chain absorbs shocks better than a trader can. rPET is cheaper than you think per meter and more expensive than you think per kilo. Price both correctly or neither comparison means anything. (Prices referenced here are for illustration; verify current rates before you build a budget.)
What I would actually choose
Neither material wins outright. The choice is contextual, and I would push back on anyone telling you otherwise.
- Choose rPET when the product is apparel or home textile, your brand has a published recycled-content target, and someone in legal or compliance has the capacity to substantiate the claim properly. Dual-source it, put a b* cap and an IV spread limit on the spec, and hold the paperwork to the same standard as the lab report.
- Choose virgin when you are running high-tenacity industrial yarn, when the colorway set is tight and the lead time is short, or when your documentation capacity is thin. A recycled claim you cannot defend is worse than no claim.
- Consider the blend if you are stuck. A fixed recycled percentage with virgin tolerances often gets you most of the marketing and none of the variance, and it is the quiet answer a lot of mills are running in 2025.
What was best practice in 2020 does not hold in 2025. Five years ago rPET was a marketing line item; now it is a spec line. But the fundamentals have not moved — IV, denier, finish, b*, shrinkage, and a certificate you can actually verify. The execution changed. The checklist didn't.
