An out-of-specification viscosity reading arrives as a single number on a sheet, and the temptation is to add thickener. That decision, taken in the first hour, is what turns a recoverable batch into a deviation report. This note walks the sequence: confirm the result, find the cause, then choose a disposition with the record to support it.
1. Quarantine before anything else
Hold the batch, label it, and stop the filling line. Nothing is added to the tank until the first three checks are complete. In most cases the batch is worth several thousand units, so an hour of investigation is cheap insurance. Record who held it, at what time, and what was already filled from that tank.
2. Retest the same sample
Many apparent failures are sampling artefacts: a surface skim, an unmixed top layer, or a sample taken hot. Let the retained sample rest 24 hours at 25 degrees, then measure again with the same spindle and speed. Repeat three times and compare. A result that moves more than 10 percent between readings is a method problem before it is a formula problem.
3. Check the instrument
Confirm calibration status, spindle number, rotation speed and guard leg. A spindle that is bent, or a spring that has drifted, produces a plausible but wrong number. Running a certified viscosity standard is the fastest way to rule the instrument out, and it takes about ten minutes if the standard is in the lab.
4. Trace the thickener lot
Pull the raw material lot record: supplier, batch, certificate, and the amount weighed. Carbomer and xanthan lots vary in molecular weight and hydration speed, and a lot change is the single most common cause of a shift. Compare the previous approved batch of the same product with the current one, ingredient lot by lot.
5. Read the process record again
Neutralisation pH, addition temperature, homogenisation time and cooling rate all move viscosity. A batch neutralised at 60 degrees instead of below 40 can read low for a day and then recover. Check the actual timestamps in the batch record against the process sheet; gaps between mixing steps often explain the whole deviation.
6. Reproduce at lab scale
Rebuild 500 grams using the same lots and the documented process. If the lab sample lands in specification, the formula is sound and the fault sits in the plant: temperature, shear, order of addition or a weighing error. If the lab sample also fails, the issue is upstream in raw material or specification.
7. Disposition options
Three routes exist: reject, rework, or blend into a later batch. Blending is usually capped at 10 percent of the new batch, and the blend needs its own full test set. Adding thickener to the original tank is a rework and needs a documented justification, because it changes the formula as filed and may affect stability data.
8. The paperwork that protects you
Write a deviation report with the failing result, the retest data, the root cause, the action and the batch disposition. Attach the instrument printout and the raw material certificates. If the batch ships, the buyer’s technical team will ask for this file within a year, and reconstructing it later is far harder than writing it now.
9. Prevention that costs little
An in-process viscosity check 30 minutes after neutralisation catches most drift before filling starts. A second check on the first filled unit closes the loop between tank and pack. Both take under fifteen minutes per batch and remove the most expensive class of rework, which is product already in a sealed jar.
10. Where it fits
Viscosity control belongs to the same quality system as microbial and packaging checks. Read the encyclopedia on rheology, or see cream formats where the specification usually starts.
