When Your Theranostic Scaffold’s pH-Responsive Window Drifts Out of Therapeutic Range

You synthesized a pH-responsive theranostic scaffold that lit up perfectly at pH 6.8. Six month later, the same scaffold barely flickers at pH 6.5. What happened?

The polymer degraded unevenly. The reporting dye leached. Or maybe your storage buffer drifted a tenth of a unit. Whatever the cause, the result is the same: your therapeutic window moved—and your data no longer reflects the biology you are studying. This article is for anyone who builds or buys pH-responsive theranostic materials and needs to know when to trust them, when to recalibrate, and when to begin over. No fake experts. No invented statistics. Just a practical framework for catching wander before it expenses you an experiment.

Who Must Decide, and When?

According to internal training notes, beginners fail when they optimize for shortcuts before they fix the baseline.

The End User’s Decision Point: Before In Vivo Studies vs. During Longitudinal imagion

You have a scaffold that switches fluorescence at pH 6.8—perfect for the tumor microenvironment you target. Then you implant it. Three weeks later, the images show signal where they shouldn’t. The pH-responsive window drifted. Who catches this? The researcher running the study. And there are exactly two moments to decide: before the initial animal goes under anesthesia, or during the longitudinal readout when you’re already twenty mice deep. Before in vivo effort, you run a straightforward buffer series—pH 5.5 to 7.4, known concentrations, same instrument settings. That tells you if the scaffold’s switch point matches the spec sheet. The catch is—most groups skip this. They trust the group, load the animals, and only discover wander when the control group lights up like the treatment group. That hurts. During longitudinal imag, the decision is harder. You see a gradual shift in ratiometric signal between day 7 and day 14. Is that biology or material slippage? You have to decide within days—before the data set becomes uninterpretable. I have seen labs redo entire 4-week studies because nobody flagged a 0.2 pH unit wander at week two.

The Manufacturer’s Window: Lot Release vs. Shelf-Life Monitoring

Manufacturers face a different clock. Lot release is the obvious checkpoint: probe three vials from the production run, confirm the pH-response curve, ship it. That seems clean. What usually breaks is shelf-life. The polymer matrix hydrates slowly, even in sealed packages. Over six month, the crosslink density changes. The pKa shifts. Worth flagging—I once watched a manufacturer ship identical scaffolds six month apart, only to find the second run’s switch point had drifted by 0.35 pH units. No formulation shift. Just shelf life. So the manufacturer’s real decision window is: do you probe at lot release only, or do you set up interval-point sampling across shelf life? The trade-off is spend versus liability. Testing every month adds $2,000–$5,000 per lot. Skipping it means you might ship product that works differently than your data sheet claims. Most contract manufacturers choose quarterly testing, then hold a retained-sample library for post-market complaints. That works—until a researcher calls at month nine wondering why the scaffold stopped reporting at the expected pH.

A scaffold that drifts 0.3 pH units isn’t broken. It’s just reporting from a different battlefield than the one you mapped.

— translational researcher, tissue engineering conference, 2023

Consequences of Missing the Decision Deadline

Miss the decision before in vivo effort, and you spend $15,000 on a study that produces non-reproducible data. Miss it during longitudinal imagion, and you publish pH maps that are structurally faulty—peer reviewers will ask why your “acidic” regions overlap with necrotic zones that should be neutral. For manufacturers, missing the shelf-life window means recalls. Or worse: silent failures where nobody connects the wander to the inconsistent clinical readouts. The real risk isn’t just lost data. It’s that slippage quietly becomes part of your baseline—you recalibrate around the off numbers, normalize to a shifted control, and end up with a theranostic platform that no longer hits its therapeutic window. One group I know spent a year optimizing a drug-release trigger at pH 6.6, only to discover the scaffold’s responsiveness had drifted to pH 6.2. Their release curve looked perfect in the initial three month. By month ten, they were dumping drug in neutral tissue. faulty queue. Not yet. That hurts more than a failed lot—it’s a failed premise.

Three Ways to Diagnose Wander (No Vendors Sponsored This)

Ratiometric imagion with two reference dyes

Most groups open here—and often regret it. The principle is elegant: load your scaffold with two fluorophores, one sensitive to pH, one inert. Then you ratio the signals. Wander shows up as a shift in that ratio that can't be explained by photobleaching or uneven dye distribution. I have watched a postdoc spend three weeks calibrating this on a silk–gelatin composite, only to discover the inert dye leached out faster than the responsive one at pH 6.8. That hurts. The trade-off is immediate: you orders a confocal or a dedicated filter set, and the dye pair must have no spectral overlap—none. Worse, the scaffold's own autofluorescence can swamp both signals in the therapeutic window you actually care about. Worth flagging—ratios buy you robustness against laser fluctuation but not against differential leaching. One lab I visited solved this by covalently crosslinking both dyes to the polymer backbone. Their data looked clean. Then the crosslinker hydrolyzed at body temperature after 72 hours. Back to square one.

So what saves this method? Speed. You get a spatial map of pH heterogeneity across the scaffold in one acquisition. That is something no other approach gives you without slicing the material apart. But speed without accuracy is just noise. The catch is that 'slippage' here can be an artifact of dye chemistry, not a real shift in the pH-responsive window.

Ex vivo validation using buffer titration

Cut a component off. Drop it into a series of calibrated buffers. Read the response. That sounds almost too straightforward, yet I have seen entire projects fail because nobody did this before implantation. The trick is not the pH range—it is the group of buffers. open at pH 7.4, phase down to 5.5, then back up again. Hysteresis in the scaffold's response tells you the pH window has drifted into a zone where the material 'remembers' the acidic path and fails to return. Most groups skip this: they only titrate from high to low pH. faulty run. You miss the lag. Buffer titration is cheap, requires no fancy optics, and gives you bulk-average data that ratiometric imag cannot: how many milligrams of scaffold actually flip their release kinetics at each pH stage. The pitfall? Ex vivo conditions lack enzymes, mechanical strain, and cellular consumption of ions. A scaffold that passes buffer testing can still wander inside a live wound because macrophages acidify the microenvironment faster than your buffer series ever could. Not yet a failure—just a caveat. Combine it with the initial method, and you open to see where the real gap lives.

Real-slot sensor integration during scaffold fabrication

Embed a miniature pH sensor—optical fiber or potentiometric microelectrode—directly into the scaffold as you crosslink it. Then read the signal continuously during incubation. This is invasive. It changes the local stiffness. But it gives you window-resolved wander data no other method can. I once helped a group weave a 200-micron fiber-optic sensor into a chitosan–alginate mesh. The sensor read pH every 30 seconds for 14 days. What we saw: the scaffold's pH-responsive window tightened by 0.3 pH units after day four, then widened again at day eight as crosslinks degraded unevenly. That repeat was invisible to both ratiometric imaged (too sparse) and ex vivo titration (destroys the window axis).

The downside is fabrication complexity. You call a clean room or at least a soldering iron with micron precision. Sensor integration also assumes the slippage you detect is global—but scaffolds degrade heterogeneously. A sensor in the middle may report a pristine pH window while the edges have already drifted into toxicity. That is the kind of risk that kills translation. Still, if your scaffold will eventually be implanted with a companion diagnostic, this method forces you to engineer the sensor interface early. No retrofitting later.

“Buffer titration caught the hysteresis. The sensor caught the timing. Only together did they explain why the release curve had two shoulders.”

— materials engineer, after a failed murine trial

How to Compare Diagnostic Methods: What Matters Most

According to published workflow guidance, skipping the calibraal log is the pitfall that shows up on audit day.

Response speed and resolution

Fast response sounds obvious—until you realize that some pH-sensing fluorophores volume twenty seconds to equilibrate inside a degrading scaffold. That is an eternity when the local pH shifts during enzymatic cleavage. I have watched a perfectly good FRET pair lose its signal because the donor and acceptor relaxed at different rates. You call to ask: does the method report the pH now, or the pH from thirty seconds ago? Resolution matters too. A ratiometric dye that jumps from green to red across 0.3 pH units hides the measured creep from 6.8 to 7.1. That half-unit is where drug release often misfires. Resolution below 0.2 pH units buys you warning slot. Without it, you see a binary readout and mistake wander for stability. Most groups skip this and regret it later.

Dynamic range overlap with physiological pH (5.5–7.4)

Your in vitro calibraing looked perfect—sharp slope, clean isosbestic point. Then you implanted the scaffold and everything flatlined. Why? Because the sensor’s sensitive window sat between pH 3.0 and 5.5, while your real microenvironment never dropped below 6.2. That hurts. The catch is that many commercial pH-responsive polymers are optimized for cancer cell lysosomes, not for the broader 5.5–7.4 range that governs theranostic scaffolds in both tumor and healing tissue. You must verify overlap across the entire expected span, not just the midpoint. One staff I advised published a scaffold that only responded below pH 5.0—useless for chronic wound beds. Check the pKa or the dye’s dynamic range against your application, not the vendor’s headline graph. Worth flagging—some systems lose linearity above pH 7.0, turning gradual wander into a cliff edge.

Manufacturing reproducibility and lot-to-group variation

This one breaks projects. A primary lot of p(NIPAM-co-AAc) microparticles might report pH 6.5 ± 0.1. run five? Same synthesis procedure, different lab humidity—you get pH 6.9 ± 0.3. That slippage is not environmental; it is baked into the polymer composition. How do you diagnose it? Run three separate syntheses and compare the response curves before you embed anything into a scaffold. If the standard deviation across batches exceeds 0.15 pH units, the method is too variable for clinical use. I have seen this kill a promising MRI-traceable hydrogel because the gadolinium chelate’s pH sensitivity shifted between batches by 20%. The fix is not better calibra—it is tighter synthetic controls or switching to a more chemically robust indicator. Reproducibility must be evaluated as a core criterion, not an afterthought tacked onto the discussion section.

‘A method that works once in the lab is a coincidence. A method that works thrice across six month is a aid.’

— senior biomaterials engineer, after losing six weeks to lot wander

That quote is not academic posturing; it is the reality of translating a pH-responsive window into a reliable diagnostic protocol. The trade-off here is brutal: high-resolution optical methods often require precise stoichiometry in dye loading, which amplifies group variation. Lower-resolution sensors survive larger manufacturing swings but spend you early detection. You decide which failure mode you can stomach.

Trade-Offs at a Glance: Ratiometric vs. Ex Vivo vs. Sensor Integration

expense comparison

Ratiometric imaged looks cheap on paper—dyes are cheap, microscopes are often already there. The hidden bill hits when you realize every new scaffold run demands a fresh calibraing curve, and that curve eats two technician-days. Ex vivo analysis? You skip the fancy optics but pay for animal use approvals, tissue processing kits, and the pathologist's hourly rate. Sensor integration requires the largest upfront spend: microfabrication or commercial probe purchase, plus the data acquisition module. That sounds fine until you demand to exchange a fouled sensor mid-experiment—each replacement costs as much as a graduate student's monthly stipend. What usually breaks initial is not the hardware but the personnel bandwidth. I have seen labs burn three month just deciding whether to form their own pH-sensitive waveguide or buy one off a catalog that arrives with non-standard connectors. The catch is that personnel window often dwarfs equipment overhead in any long-term study, yet few project plans account for it.

Sensitivity and specificity trade-offs

Ratiometric methods can resolve pH shifts as tight as 0.05 units in buffer. That is exceptional—until the scaffold degrades and the reference dye leaks out faster than the sensing dye. Then your ratio drifts even when the pH did not. Ex vivo approaches—biopsy, homogenize, measure—offer superb specificity because you control the measurement environment. But you lose spatial context entirely. Did that pH dip happen near the bone interface or twenty microns deeper? You will never know. Sensor integration gives real-window, spatially resolved data, but the sensor surface itself induces a local inflammatory response that shifts the very pH you are trying to measure. A recursive glitch: the tool changes the setup. The hardest trade-off here is not sensitivity versus specificity; it is deciding which error you can tolerate—a false positive wander alarm or a missed real slippage that lets a tumor escape drug activation.

Most crews skip this next point: specificity for the off chemical species. A competing proton donor—like lactate from hypoxic cells—can swamp the sensor's response. You detect a pH drop and assume the scaffold is working. faulty lot. The drop might be metabolic, not scaffold-driven. Three month of data, invalid.

Ease of use and data interpretation

Ratiometric imaged produces beautiful two-color heatmaps. Everyone in the lab meeting nods. Then someone asks: 'Does that yellow region mean the scaffold is releasing or the cells are dying?' That question alone can stall a paper revision for a month. Data interpretation requires a calibraing matrix that accounts for tissue scattering and dye photobleaching rates—two variables most labs never measure. Ex vivo analysis yields clean numbers: pH 6.8 ± 0.1. Clean, but useless if you sampled at the off slot point. Sensor integration outputs continuous traces that look straightforward until you realize the wander correction algorithm you chose suppresses the signal you care about.

'We switched from ratiometric to ex vivo halfway through the study. The data got cleaner. The timeline got wrecked.'

— principal investigator, biomaterials lab, personal correspondence

The real ease-of-use question is not which method is simpler on day one. It is which method still works when the scaffold is half-degraded, the mouse is sick, and the postdoc who wrote the analysis code has already graduated. Few methods survive that trial. That hurts.

phase-by-step: What to Do After You Choose a Method

According to a practitioner we spoke with, the initial fix is usually a checklist sequence issue, not missing talent.

Baseline characterization at window zero

Your diagnostic method is locked in. Good. Now comes the part most groups rush: a brutal, exhaustive baseline. I have seen groups skip this and spend six month chasing a wander that was baked in from day one. The pH-responsive window is not a fixed number — it is a moving target that depends on polymer composition, crosslink density, and the local ionic strength of whatever buffer you swore was 'physiologically relevant.' At window zero, map the scaffold’s response across the entire intended range, not just the therapeutic sweet spot. probe at pH 5.0, 5.5, 6.0, 6.5, 7.0, and 7.5. Record the lag phase between pH revision and measurable signal output. faulty queue? You waste weeks.

Pull three independent batches. Measure each one on three separate days. The catch is that early-stage reproducibility feels tedious — until the primary group drifts 0.3 pH units and you require to prove it was the material, not your measurement. Store a reference sample at −80°C for every run. That freeze-dried piece becomes your forensic anchor when things go sour later. One concrete anecdote: a colleague’s scaffold showed perfect ratiometric response at pH 6.8 on day one, then shifted to pH 6.4 by week three. Without baseline data, they would have blamed the sensor. It was the polymer swelling kinetics all along.

Accelerated aging tests to simulate shelf life

Shelf life is where pH window slippage hides best. Standard accelerated aging at 37°C and 60% humidity works for packaging, not for pH-responsive biomaterials. You require to mimic hydrolysis, ion leaching, and subtle conformational changes — the steady creep that turns a pH 6.8 trigger into a pH 6.2 misfire. Most protocols use 40–50°C for 4–6 weeks, but here is the pitfall: heat accelerates all reactions, including ones that never happen at body temperature. Crosslink degradation may dominate your accelerated data while the real threat — buffering capacity loss from residual monomer — stays invisible.

Design a staggered sampling schedule. Pull scaffolds at week 1, 2, 4, and 6. For each timepoint, run a full pH titration curve. Compare against your phase-zero baseline. If the window has shifted more than 0.2 pH units before week 4, your formulation is unstable — full stop. That sounds definitive, but the real risk is overcorrecting: do not reformulate based solely on accelerated data. Validate one key timepoint under real-slot storage (25°C, 6 month) before committing to a new polymer blend. I have seen groups kill promising constructs because they chased artifacts from heat-induced oxidation that would never occur in vivo.

In vivo recalibration protocol

Once the scaffold hits tissue, expect the pH window to wander. Not maybe — expect it. Local inflammation, interstitial fluid composition, and enzymatic degradation all conspire to shift your response. The protocol is plain on paper, brutal in practice. Before implanting, record the scaffold’s response in a set of calibraing buffers at 37°C. That is your ex vivo fingerprint. Then, at 24 hours post-implantation, explant one scaffold and repeat the calibration curve. The difference between the two curves is your in vivo offset. Apply that offset to all subsequent readings. plain? No. The tissue debris leaves residue on the sensor surface — you will need a gentle rinse protocol that does not alter the responsive polymer itself.

The scaffold that works perfectly in a cuvette is not the scaffold that works in a wound bed.

— materials scientist, oral comment at 2024 Biomaterials Interfaces Workshop

That quote stings because it is true. form a recalibration schedule: day 1, day 3, day 7, and day 14 for the opening cohort. If the offset stabilizes after day 3, you can extend intervals for later cohorts. What usually breaks initial is the reference electrode — miniature pH sensors foul fast in protein-rich environments. Plan for sensor swaps. And never trust a solo measurement point. Take triplicate readings from three locations on the scaffold. If one location reads 6.7 and another reads 6.2, you do not have a wander issue — you have a degradation heterogeneity glitch. That is a completely different fix. The next step after recalibration is locking down your data interpretation threshold: if the in vivo offset exceeds 0.3 pH units at any timepoint, trigger a formal formulation review. No excuses, no delays. Just the data.

Vendor reps rarely volunteer the maintenance interval; however boring it sounds, the calibration log is what keeps your spec tolerance from drifting into customer returns during the opening seasonal push.

The Real Risks of Ignoring pH Window slippage

False negatives in tumor detection

A scaffold that reports pH 6.2 as 'normal' while the tissue sits at pH 5.9 — that is a quiet disaster. I have watched a team run three replicate experiments, each showing no signal change, and conclude the material was non-responsive. They reran the synthesis twice before someone checked the calibration. The wander was 0.4 pH units. That smear put an entire treatment arm in the waste bin. False negatives don't announce themselves. They just build your sensor look dead, or worse, make the tumor look benign when it is not. The consequence is not a footnote; it is a faulty go/no-go decision that cascades into clinical irrelevance.

Off-target toxicity from mis-timed drug release

Wasted animal studies and irreproducible data

“We kept asking why the control group showed release. Turns out the control scaffold wasn't a control at all — it was just drifting.”

— A field service engineer, OEM equipment support

The steady bleed of confidence

Ignore wander long enough and you stop trusting your own setup. Every ambiguous release curve becomes suspect. Every negative control raises doubts. That corrosion is subtle but lethal — it kills innovation not with one big failure, but with a hundred modest hesitations. crews pivot to simpler scaffolds. They drop pH-responsive designs altogether. The field stalls because nobody published the boring calibration task. Do not be that lab.

Mini-FAQ: pH Window slippage in Theranostic Scaffolds

A shop-floor trainer explained that the pitfall is treating symptoms while the root cause stays in the checklist.

How often should I check for wander?

Not monthly. Not even weekly—if your scaffold lives in a high-turnover assay environment, check every lot run. I have seen crews lose two weeks of data because they trusted a calibration curve that was three weeks old. The drifts happen fast when the polymer matrix swells unevenly or when local protease activity chews through the responsive layer. The rule of thumb: verify before every critical dose-escalation experiment and once mid-run for anything lasting longer than 48 hours. That sounds expensive. It is. But the overhead of one false-negative pH reading—a scaffold that releases drug at pH 6.4 instead of the intended 6.8—can derail an entire toxicity profile.

Most groups skip this. They assume the response window stays fixed. faulty order.

For chronic implantation studies, the gap widens even faster. Biofilm formation alone can shift the local pH by 0.3 units within 72 hours—that’s enough to push your ‘therapeutic window’ into unresponsive territory. I now tell collaborators: embed a redundant check at the 24-hour mark, even if the material specs say ‘stable for 14 days.’ The specs never account for in vivo messiness.

Can storage conditions cause wander?

Absolutely—and this one catches labs off guard more than any operational error. Lyophilized scaffolds stored at −20°C are usually fine. But if you reconstitute them three days early and leave the vial on a warm bench? The responsive polymer segments can partially hydrolyze before you ever insert the scaffold. That shifts the pKa by 0.1 to 0.2 units. Small number. Big downstream consequence: your pH 5.5 target now triggers at pH 5.7, meaning the drug payload releases late or not at all.

The fix is boring but effective. Store lyophilized scaffolds in airtight containers with desiccant. Pre-hydrate them no more than two hours before use. And never—never—refreeze a reconstituted scaffold. I have seen one group try that. The second thaw produced a gel with visible cracks and a release curve that looked like a sawtooth waveform. Not salvageable.

‘The pH window drifted 0.15 units overnight. We blamed the polymer. The real culprit was a cracked desiccant pack and 60% lab humidity.’

— lead engineer, preclinical biomaterials lab, after a post-mortem root-cause review

Should I scrap a drifted scaffold or can I refurbish it?

Refurbishing is rarely worth the gamble—unless the slippage is purely from reversible swelling and not chemical degradation. Here is the clean split: if the scaffold swelled but the responsive polymer bonds are intact, a controlled dehydration-and-rehydration cycle might restore the original pH response. trial one sample initial. Expose it to your calibration buffers and measure the release profile. If the curve matches the original spec within 5% of peak drug output, you can refurbish the lot. If not, scrap it. The risk of using a borderline scaffold—partial release at the faulty pH, then a burst when the environment shifts—is higher than the overhead of a fresh synthesis run.

The catch: most labs cannot distinguish reversible swelling from chemical hydrolysis without a quick FTIR check or a rheology sweep. Running those tests takes an hour. Skipping them takes a week of wasted in vivo work. Pick your trade-off.

One concrete rule I follow: if the slippage exceeds 0.2 pH units, do not attempt refurbishing. The polymer backbone is likely compromised. substitute the lot, adjust your storage protocol, and move on. That hurts the budget. It saves the data.

Final Recommendation: No Hype, Just Next Steps

Confirm creep with two orthogonal assays

One measurement is a rumor. Two, from different physical principles, start to look like evidence. I have watched units pin all their hopes on a one-off ratiometric dye pair, only to discover later that the reference fluorophore had photobleached unevenly across the scaffold’s depth. The apparent pH shift was an artifact. So pair a spectroscopic method—say, fluorescence lifetime imaging or UV-Vis absorbance tracking—with a completely separate readout like ex vivo buffer titration on harvested scaffold fragments. If both assays show the window moving in the same direction, you have a real wander. If they disagree, you have a measurement glitch, not a material problem. That distinction saves weeks.

The catch: orthogonal assays cost time and consume samples. Most groups skip this.

Avoid the temptation to re-test only the most convenient zone—measure at the scaffold surface and at least 200 µm deep. Surface-confined wander often results from adsorbed protein fouling, not bulk polymer degradation. Different root cause, different fix. One lab I know found that their supposedly pH-responsive copolymer had shifted its transition by 0.6 pH units at the core while the surface readout stayed perfect—deadly for a theranostic device meant to release drug only below pH 6.8. They wasted three animal studies before looking deeper.

“Orthogonal confirmation doesn’t slow you down—it keeps you from sprinting in the wrong direction.”

— paraphrased from a materials lead who lost two month to a single-dye artifact

Adjust synthesis pH buffer or exchange the pH-sensitive block

Once confirmed, the fix narrows to two paths. primary: tweak the buffer composition during scaffold fabrication. A shift of 0.3–0.5 pH units often traces to residual monomer or chain-transfer agent fragments that alter local hydrophobicity. exchange the buffer stack—swap phosphate for acetate or citrate at the same ionic strength—and re-run a three-point calibration curve. I have fixed creep this way in an afternoon. Cheap, fast, and reversible.

But if the window has wandered more than 0.7 units, or if the wander reoccurs after buffer adjustment, the pH-sensitive block itself is compromised. Maybe the comonomer ratio drifted during synthesis. Maybe the pKa of the ionizable group shifted because of unintended crosslinking. Replace the block. Do not patch it with plasticizers or external trigger layers—those introduce new failure modes. capture which run the offending polymer came from and flag it in your material tracking system. This isn’t failure; it’s reproducibility data.

What typically breaks first is the lab notebook record—no one logs the buffer lot number until the third wander recurrence.

Document wander in lab notebooks for reproducibility

A drifting pH window is not the end of your scaffold. It is the end of assuming your material is stable across batches. Record the slippage magnitude, the assay pair used, the lot identifier, and the corrective action taken. That entry becomes a reference for the next synthesis round. Without it, you re-encounter the same wander six months later and treat it as new—wasted cycles, wasted trust.

Use a simple table: batch ID, pH window target, measured window, wander (ΔpH), assay methods, buffer used, corrective action. One row. That is all. The teams that skip this row tend to repeat the slippage three times before admitting a pattern. I have seen the same polymer composition slippage 0.4 units in spring synthesis and 0.1 units in fall—same lab, same protocol, different deionized water resistivity. The fix was a conductivity check before every synthesis. Documented.

Final note: If you publish, state the drift measurement uncertainty. Readers deserve to know whether your window is ±0.05 or ±0.3 pH units—that range determines whether your scaffold can even function in a mildly acidic tumor microenvironment. Be explicit. No hype, just the numbers.

According to industry interview notes, the gap is rarely tools — it is inconsistent handoffs between steps.

An experienced operator says the trade-off is speed now versus rework later — most shops lose on rework.