Illustrative case study. Company names and identifying details are fictional. Technical details reflect real industry practice and current regulation.

Spray drying should kill Salmonella. The inlet air hits 180°C; the droplets flash-dry in seconds; the physics looks lethal. So when finished milk powder — destined for infant formula — started testing positive for Salmonella across multiple lots, the first assumption was survival: the dryer wasn’t hot enough, the feed was too contaminated. The investigation proved otherwise. The dryer was lethal. The powder leaving the dryer was sterile. And then it fell past cracks in the dryer housing — cracks harboring a persistent Salmonella strain in the insulation behind the panels — and was recontaminated on its way to the bag. The kill step worked perfectly. The equipment around it was the reservoir.

Background: a milk powder plant feeding infant formula

DairyDry Ingredients (fictional) employed about 60 people spray-drying pasteurized milk into powder, much of it sold as base powder to infant formula manufacturers — the most demanding customers in the dairy world, with zero tolerance for Salmonella and audit teams to enforce it. The process was conventional: pasteurized milk, evaporation, spray drying at 180°C inlet, powder handling, bagging. The plant ran dry — no water in the powder areas, ever — per the industry’s standard Salmonella control doctrine for low-moisture foods.

Challenge: positives in the finished powder

Three lots tested positive for Salmonella in finished-product testing — caught by the plant’s own hold-and-test program before shipment, which is the good news. The bad news was the pattern: the positives kept coming, lot after lot, while every process parameter looked correct. Dryer temperatures were on spec. The milk was pasteurized. The environment was dry. The plant chased the feed — raw milk Salmonella levels, pasteurizer performance — and found nothing wrong. The dryer was killing everything in the feed. Something was adding Salmonella back afterward.

Investigation: the teardown that found the reservoir

The breakthrough came from a full dryer teardown during a scheduled shutdown — panels off, insulation exposed, the kind of invasive inspection plants avoid because it costs downtime. Behind the outer panels, in the insulation voids around weld seams, the team found it: cracked welds creating crevices, powder infiltrated into the insulation, and Salmonella — swabbed, cultured, whole-genome sequenced — matching the finished-product strain exactly. A persistent strain, resident in the dryer housing for what the investigation estimated as years, aerosolizing into the powder stream as dried product fell past the harborage points.

The dry-cleaning doctrine had protected the reservoir. The plant’s “dry plant” rule — no water, ever, in powder areas — is correct Salmonella control for operations, but it had hardened into dogma: dry brushing was the only cleaning method, and dry brushing doesn’t eliminate Salmonella from a harborage — it spreads it. The organism is famously dry-tolerant; it survives desiccation for months, even years, in powder residues. Brushing moved contaminated dust around without killing anything. And because the harborage was behind panels — invisible during routine inspection, unreachable by brushing — it persisted indefinitely.

Environmental monitoring had missed it because the program swabbed accessible surfaces — floors, equipment exteriors, the powder handling areas — not the voids inside the dryer housing. The monitoring was looking where the light was, not where the organism lived.

Root cause: harborage protected by design and doctrine

1. Dryer housing harborage. Cracked welds and insulation voids created inaccessible niches where powder accumulated and Salmonella persisted. The sterile powder exiting the drying chamber was recontaminated by contact with — and aerosolization from — these reservoirs. Post-lethality contamination in its purest form.

2. Dry-cleaning-only doctrine. The no-water rule, correct as a general principle, prevented the periodic wet cleaning that could have eliminated the harborage. Dry brushing spread contamination without reducing it. Doctrine without exceptions becomes a vulnerability.

3. Monitoring that couldn’t see the problem. The environmental program sampled the accessible world. The harborage lived in the inaccessible one. Seek-and-destroy requires seeking — including teardown inspections that plants resist because they cost production time.

Corrective actions: opening the dryer, literally and procedurally

Immediate: the dryer was fully disassembled, the cracked welds ground out and re-welded to sanitary standards, contaminated insulation removed and replaced, and the entire housing sanitized. The affected lots were recalled — or rather, destroyed under hold, since the hold-and-test program had caught them before shipment. That program earned its keep many times over.

Within 30 days, controlled wet cleaning entered the program: quarterly wet wash of the dryer housing — with full dry-down verification, moisture checks, and environmental clearance before restart. The “dry plant” rule was rewritten as “dry operations, with controlled periodic wet cleaning under procedure” — doctrine updated with the exception the science required. The environmental monitoring program added annual teardown inspections: panels off, voids swabbed, harborages sought deliberately.

Within 90 days, the longer-term fix: dryer housing redesign to sanitary principles — no insulation voids, continuous welds, accessible for inspection and cleaning. The capital project took a year, but the specification was written the month after the incident. Infant formula customers were briefed transparently; two of them sent their own auditors, who left satisfied — the hold-and-test program plus the corrective actions told a story of a system that catches failures and fixes them.

Results: the reservoir eliminated

Twelve months later: zero Salmonella in 200 finished lots, zero in the expanded environmental program including teardown swabs. The persistent strain — the WGS fingerprint the lab had come to know — never appeared again. The quarterly wet cleans, initially feared as Salmonella risks themselves (water in a dry plant), proved safe under the controlled procedure: dry-down verified, clearance swabs negative, no moisture-related issues.

Lessons learned: what you’d do Monday morning

Schedule your next dryer teardown inspection — and when you do it, swab the voids: behind panels, around welds, in the insulation. If you’ve never looked inside your dryer housing, you don’t know what’s living there. The harborage in this case had likely been resident for years, through hundreds of negative finished-product tests, until the conditions aligned. Seeking beats assuming.

Then examine your “dry plant” doctrine for the exceptions it needs. Dry operations are correct. Dry-only cleaning forever is dogma. Controlled periodic wet cleaning — with verified dry-down — eliminates harborages that dry methods protect. Write the procedure, validate the dry-down, and stop treating water as the enemy when the actual enemy is the biofilm behind the panel.

Yeah, but actually — the most unsettling part of this case is how long it ran. Years of production, hundreds of lots, a persistent Salmonella strain living in the equipment, caught only when the contamination level crossed the detection threshold in finished product. The hold-and-test program was the hero — without it, this powder goes into infant formula. If you’re making low-moisture ingredients for sensitive applications and you’re not on hold-and-test for Salmonella, this case is your argument for starting. And remember the physics: the dryer was lethal, the powder was sterile, and none of it mattered — because food safety isn’t about the kill step, it’s about everything after it. The sterile powder falling past a contaminated crack is just expensive Salmonella delivery. Tear down the dryer. Look behind the panels. The reservoir is always where you haven’t looked.