This article addresses In-Process Checks for PBT: Moisture, Melt Flow and How Much Regrind Is Too Much as a single front-of-press control loop. Polybutylene terephthalate (PBT) is a semicrystalline polyester whose ester repeat unit is susceptible to hydrolytic chain scission when melt-phase water content exceeds 0.02 wt%. In injection moulding and extrusion, that threshold is not a resin lot certificate value; it is a release gate that shifts melt viscosity, dimensional stability, weld-line strength, and part brittleness. The control loop begins at the dryer outlet and ends at the gravimetric blender. It is measured with standardised methods and production-scale equipment, not with subjective operator judgement.

Moisture below 0.02% is the release criterion, not a suggestion

As-supplied PBT pellets can carry 0.10–0.25 wt% moisture depending on warehouse humidity and packaging integrity. At 23 °C and 50% RH, equilibrium moisture for many unfilled PBT grades climbs toward the upper end of that range, and granulated regrind reaches it faster because of increased surface area. PBT typically enters the barrel at 240–270 °C. At those temperatures, the reaction between residual water and ester bonds is fast; there is no conditioning period long enough to compensate for wet pellets. Hydrolysis consumes water and cleaves the ester linkages along the polymer backbone. The resulting reduction in chain length appears first as an apparent improvement in flow and later as a drop in Charpy impact and tensile elongation.

The moisture check at the dryer outlet is commonly a halogen loss-on-drying measurement calibrated against ISO 15512 Karl Fischer titration. A 5 g sample is taken from the hopper inlet or screw feed throat, dried at 130 °C to constant mass, and compared with the melt-processing limit of 0.02 wt%. A reading above 0.03 wt% requires re-drying rather than release. The drying system should be a closed-loop desiccant-bed dryer delivering air at a dew point of -30 °C or lower and an inlet temperature of 120–130 °C. Residence time for virgin pellets is normally 4 h; for blends containing regrind, extension to 6 h is prudent because granulate does not flow through the hopper as a uniform bulk and retains moisture in folds and fines. At temperatures above 130 °C, oxidative yellowing and surface degradation can begin; below 120 °C, drying must be extended, especially when plant relative humidity exceeds 60%.

Dryer residence time is not the same as elapsed time. A machine consuming 35 kg/h through a 100 kg hopper provides an average residence time under 3 h, which is below the 4 h minimum. Hopper volume must be matched to consumption, or the dryer must be audited against the number of hopper turnovers per shift.

Regrind from sprues, runners and short shots enters the process with a different moisture history. The granulator cuts a larger specific surface area than virgin pellets, and fines below 2 mm absorb moisture rapidly when stored open at a press. In high-humidity plants, sealed regrind hoppers with dry-air purge are not optional if regrind content exceeds 10 wt%. The material should be blended with virgin before drying rather than after the dryer; otherwise wet regrind can enter the feed throat as unmixed pockets and produce local hydrolysis that is not reflected in an average hopper sample.

What Do Shifts in Melt Flow Indicate at the Press?

Melt volume-flow rate (MVR) is determined at 250 °C under a 2.16 kg load according to ISO 1133-1:2022; the same value may be reported as melt flow rate in g/10 min according to ASTM D1238-23. Unreinforced injection moulding grades typically exhibit MVR values of 12–30 cm³/10 min, while extrusion grades may run from 8 cm³/10 min to 15 cm³/10 min. Glass-filled and flame-retardant grades can fall outside these bands because additives and fillers alter apparent viscosity. The in-process limit is therefore not only the published grade datasheet value; it is the dried virgin reference MVR for each specific lot.

The measurement should be made only after pellets have been dried to 0.02 wt% moisture. Water is an effective processing plasticizer and hydrolysis agent; an undried sample can produce a melt flow shift that mimics molecular-weight loss. A standard die of 2.095 mm diameter and 8.000 mm length is used with a temperature stability of ±0.5 °C. The preheat time is 4 min, and piston travel is recorded over a segment appropriate to the flow rate. In production, a shift in MVR of more than 20% from the dried virgin reference is treated as an excursion. Possible causes include wet resin, excessive regrind, contamination by PET or other polyesters, additive lot variation, or unreported barrel temperature drift. A fall in MVR can indicate crosslinked gel formation or contamination with a higher-viscosity resin and is just as actionable as a rise.

The melt indexer itself is maintained with a calibration check against a reference resin. The capillary die is cleaned after each measurement; carbonized PBT deposits can reduce the effective die diameter and produce false low-flow readings. In a production laboratory, the MVR result is recorded alongside the sample moisture and the granulator batch number. For glass-fibre-reinforced PBT, the material may require a higher test load if it does not flow sufficiently under 2.16 kg, but the standard condition for PBT remains 250 °C and 2.16 kg. When a grade falls below 8 cm³/10 min, piston travel may be slow enough to increase measurement uncertainty; in such cases, equipment verification and operator technique become part of the release decision.

At a moulding cell with a barrier screw of 25 mm or larger diameter and a shot mass above 50 g, a drop in screw recovery time at constant backpressure and screw speed is an indirect signal that melt viscosity has fallen. When recovery time changes by more than 8–10%, the shift should be verified by an MVR measurement rather than by adjusting barrel temperatures. A cushion pressure fall at the end of hold can also indicate lower melt viscosity, but it is not a substitute for standardized flow testing because machine response includes check-ring wear, barrel band heating, and screw condition.

Regrind Inclusion and the 25% Front-of-Press Cap

Regrind is produced from sprues, runners, short shots and occasional edge trim using a granulator with a screen aperture of 4–6 mm. Oversized regrind can bridge in the hopper, while excessive fines below 2 mm can create feed instability and localized overheating. Dull granulator knives generate fluff and lower bulk density, making gravimetric feeding less stable. The granulator should be dedicated to PBT, or purged with virgin regrind, because polyester contamination from PET or polycarbonate can create phase separation and inconsistent properties.

In unfilled PBT, a regrind fraction up to 20–25 wt% is generally tolerated in industrial mouldings if the regrind is dry, clean, and the blend MVR remains within 20% of the virgin reference. For thin-wall connectors or parts that must meet notched Charpy impact requirements under ISO 179-1/1eA, a tighter cap of 15 wt% is common because even a small reduction in molecular weight shifts the brittle–ductile transition. For glass-fibre-reinforced PBT, the granulation process reduces fibre length; that damage is permanent and cannot be recovered in the extruder barrel. Published data for every combination of grinder speed, screen size and PBT glass content is limited, so conservative release limits are used when the part is specified for impact or fatigue.

The cap of 25 wt% is a warning limit, not an unconditional release. At 30 wt% regrind, unfilled PBT may still fill a simple thick-wall moulding but fail elongation or impact tests under ISO 527-2 or ISO 179-1. The in-process decision is made by measuring the MVR and, when the regrind source contains multiple heat histories, by checking a moulded tensile bar from the blend. If the tensile elongation at break falls below the part print requirement, the regrind ratio is reduced regardless of how well the material fills.

A separate issue is cumulative regrind. A blend containing 25 wt% of once-reground PBT is not equivalent to a blend containing 25 wt% of material that has passed through two or three heat histories. Each pass through drying and melt processing advances chain scission and additive depletion. The only dependable control is to record the number of heat histories on regrind bins and to place an upper limit of 2 heat histories for critical PBT components. If the lot MVR already approaches the upper control limit, adding regrind that increases it beyond 20% is excessive.

The interaction among these checks is not linear. A lot that enters the press at 0.01 wt% moisture but contains 30 wt% regrind can fail the MVR limit because the regrind has already lost molecular weight. Conversely, a lot at 0.03 wt% moisture with no regrind may pass MVR early in the shift but fail after barrel residence time increases. In a production line using a hot runner manifold, residence time can be 5–15 min depending on the number of cavities and manifold volume; the longer the residence time, the more damaging even slight excess moisture becomes. In-process checks are therefore recorded together, not as isolated data points.

When the Dryer Dew Point Rises Above -30 °C, Treat the Hopper as a Moisture Source

Desiccant dryer efficiency is not constant. If the return-air dew point rises above -30 °C, the drying bed is either saturated or leaking ambient air. A dew point of -20 °C may still appear acceptable on a plant audit but will slow moisture removal from PBT because the driving force for water desorption is reduced. Under these conditions, a moisture sample taken from the dryer outlet can read below the limit while the pellets reaching the screw feed throat are wetter: the hopper may be drawing in humid plant air through worn seals or the conveying line may be unheated. The sampling point therefore belongs at the hopper inlet to the screw, not at the dryer discharge.

If the dryer is started before the resin is loaded or the desiccant bed is not regenerated, the first 30–60 min of production can carry damp pellets into the barrel. A moisture check at start-up is mandatory, followed by a second check after 2 h of stable operation. The same applies after any interruption longer than 2 h. When plant relative humidity remains above 60%, hopper openings should be minimized, and transfer piping should be purged with dried air. PBT exposed to 80% RH overnight may need 8 h of drying to return to the 0.02 wt% limit.

If a moisture excursion is confirmed, the immediate press response is to stop pulling material from that hopper and evaluate the melt flow signature. Because hydrolysis during barrel residence is time-dependent, a short interruption may not fully degrade the lot; a prolonged hold at melt temperature will. The lot is not used for parts with critical weld lines or snap-fit arms unless it passes the same MVR and moisture checks as virgin material.