PVC

Our PVC line runs three suspension grades continuously—S-57, S-65, and S-70—plus one emulsion grade (P-72) in campaign mode, totaling roughly 180,000 metric tonnes annually across all grades. S-65 pipe-grade accounts for 65% of that volume, which tells you where our process optimization effort goes. After 15 years on this line, the single biggest lesson I can pass along is that grain porosity and residual VCM cause 80% of downstream problems that end up getting blamed on the stabilizer package. We run a two-stage countercurrent steam stripping column—first stage at 85°C under -0.7 bar vacuum for bulk monomer removal, second stage with live steam at 98°C and 25-minute residence time—which is slower than the single-pass stripping most plants use, but it gets our residual VCM consistently below 0.3 ppm across all suspension grades. The trade-off is energy cost: we burn about 12% more steam per tonne than a single-stage column would. I'll take that trade because we have not had a single VCM migration failure on a potable water pipe certification audit since 2014. Our drying system targets 0.12% moisture by Karl Fischer coulometric titration—not loss-on-drying, which picks up volatile organics and overestimates by 0.05-0.08 percentage points. Summer shipments (June through August) drift up to 0.18-0.22% because our plant is in a high-humidity region and the pneumatic conveying air picks up ambient moisture between the dryer and the silo. If your silo relative humidity exceeds 65% RH for more than 48 hours, expect bridging at the discharge cone regardless of which grade you bought. This is not a resin defect—it's a site condition problem that a desiccant dryer on your conveying air will fix in a day.

1. Rigid Pipe Extrusion — Counter-Rotating Twin-Screw, S-65 and S-70

This is where 70% of our S-65 ends up. The typical extrusion setup we see at customer plants is a conical counter-rotating twin-screw (KraussMaffei or Cincinnati type) running 25-35 RPM screw speed with barrel zones ramping from 160°C at the feed throat to 195°C at the die, processing dry blend from a hot-cool mixer system. Our S-65 runs a K-value of 65 ± 1 (Fikentscher, measured in cyclohexanone at 25°C), bulk density 0.52-0.56 g/cm³, and a mean particle diameter of 145-165 μm by laser diffraction. What matters downstream is the grain porosity—we run a higher ratio of secondary suspending agent (a low-hydrolysis PVA, ~72 mol% hydrolysis) at 18% of the total PVA charge specifically for S-65, which opens up the grain structure more than you'd get from a standard fixed-ratio recipe. The practical effect: our S-65 absorbs DOP in the hot mixer 15-20% faster than most competitors' equivalent K-65 grades. We measure cold plasticizer absorption at 26-29% DOP uptake (ASTM D3367, 24-hour soak). If you're switching from another supplier's S-65 to ours, drop your hot mixer discharge setpoint by 5-8°C on the first trial batch—our resin hits the dry point earlier, and if you run your old temperature profile you'll over-dry the blend, generate fines, and see feeding pulsation at the extruder throat. I've seen this happen twice in the last three years. The loss-in-weight feeder at the extruder intake should be sized with at least a 150-mm screw diameter for throughputs above 800 kg/h; our S-65 flows well at that rate, but if your feeder hopper has a steep cone angle (steeper than 60° from horizontal), the powder can bridge when the silo transfer line delivers material above 0.20% moisture. Install a vibratory pad on the hopper wall or spec a 70° cone—we learned this the hard way from a customer who lost 11 hours of production across a weekend because nobody noticed the feeder had starved and the extruder was running empty-screw for six of those hours.

Fish eye count on our S-65 is guaranteed at <5 per 100 cm², tested by dissolving 10 g of resin in 100 mL of cyclohexanone at 80°C, casting a 0.3-mm film, and counting gel defects at 100x magnification under transmitted light. We run this test on every 20-tonne lot. If you see fish eyes above 15 per 100 cm² in your extruded pipe, check these in this order:

1. Cross-contamination in your silo or conveying system. Even 0.5% polyethylene or polypropylene contamination produces visible gel particles because PE and PP do not fuse into the PVC matrix at PVC processing temperatures. One customer shared a pneumatic line between their PE silo and PVC silo with a diverter valve that leaked—cost them three days of scrap pipe and a valve rebuild. Flush your lines if you handle multiple polymers on the same conveying circuit.
2. Stabilizer dispersion. If your one-pack stabilizer wasn't pre-dispersed properly in the hot mixer, you'll get localised degradation spots that look exactly like fish eyes but are actually charred PVC. Check your mixer blade clearance—anything over 3 mm gap and you're not getting adequate shear on the stabilizer agglomerates.
3. Screw wear in the extruder compression zone. Worn flights reduce shear and leave unmelted grain cores that survive through the die. Measure your screw diameter at the compression zone—if you've lost more than 1.2 mm from the original diameter, the screw is done.
4. Moisture in the dry blend. Above 0.4% total volatiles in the dry blend, you get steam pitting in the melt that looks like fish eyes on the pipe surface. Test the blend, not just the resin—fillers and regrind bring their own moisture.

For pressure pipe applications (PN10 and above), S-65 is the minimum K-value we recommend. Below K-62, the long-term hydrostatic strength drops off enough that you'll fail the 1,000-hour hoop stress test at 42 MPa unless you significantly up the wall thickness, which erases any material cost savings. S-70 (K-value 70 ± 1) gives you an extra 8-12% in hydrostatic design basis at 50 years, but you'll run barrel temperatures 10-15°C higher and take a 15-20% throughput penalty on the same extruder. Most of our customers running large-diameter pressure pipe (250 mm OD and up) have settled on S-65 for the throughput advantage and meet the pressure rating through wall thickness rather than resin K-value. The S-70 is primarily going into window profile extrusion where impact resistance matters more than throughput.

Injection Moulding: Why S-57 Exists, and When Not to Use It

S-57 is our lowest-K suspension grade, K-value 57 ± 1, bulk density 0.55-0.58 g/cm³, with a deliberately narrow particle size distribution—90% between 100-180 μm—because injection moulding machines with small-diameter reciprocating screws (35-60 mm) are far less forgiving of particle size variation than large extruders. The melt viscosity at 190°C and 100 s⁻¹ runs about 1,800-2,200 Pa·s for S-57 versus 3,200-3,800 Pa·s for S-65 under the same conditions (capillary rheometer, L/D 20:1 die). That difference is why S-57 exists: if you try to injection-mould pipe fittings with S-65, you either raise the melt temperature to the point where thermal degradation starts (Congo red stability on our S-65 drops below 30 minutes above 210°C) or you live with incomplete cavity filling on thin-wall sections. We've seen processors push S-65 into injection moulding because they wanted to consolidate to one resin grade across their entire plant, and it works for about two shifts before the scorch marks start appearing on the parts and the mould cleaning frequency doubles. Don't do it.

The thing to watch with S-57 is its narrower processing window. Because the molecular weight is lower, the melt is more shear-sensitive—if your injection speed is too aggressive (above 80 mm/s ram speed on a 50-mm screw), you'll shear-heat the melt locally and burn it before the barrel thermocouples even register the spike. We recommend a fill time of at least 1.8 seconds for mould cavities under 500 g shot weight; faster than that and you're gambling with burn marks on the gate land. Also, S-57 has a slightly higher residual VCM than our other suspension grades—typically 0.4-0.6 ppm versus 0.2-0.3 ppm for S-65—because the lower-K resin requires a slightly different polymerization endpoint and the stripping efficiency drops off marginally at lower molecular weight. This is still well within the 1 ppm limit for general applications, but if you're moulding fittings for potable water contact, use S-65 or verify migration compliance with your specific moulding conditions.

Dry-Blend Calendering and Flexible Sheet: Plasticizer Uptake Timing

When customers run flexible PVC sheet on a calender line—typically a four-roll inverted-L configuration with roll temperatures between 165°C and 185°C—the rate-limiting step is almost always the dry-blend preparation in the hot-cool mixer, not the calender itself. Our S-70 is the preferred resin here, and the reason goes back to that PVA ratio adjustment I mentioned earlier. For S-70, we run the secondary suspending agent at 22% of total PVA, higher than the 18% we use for S-65, which produces an even more porous grain structure. DOP absorption reaches 90% of equilibrium in 11-14 minutes in a Henschel-type hot mixer running at 1,200 RPM with a jacket temperature of 110°C, versus 16-19 minutes for a typical competitor S-70 with a denser grain morphology. This sounds like a small difference, but on a line running 2,500 kg/h of dry blend, it's the difference between your mixer being the bottleneck or your calender being the bottleneck. If you want to exploit this, you need a hot mixer with a minimum 800-litre working capacity and a discharge gate that opens fully in under 2 seconds—we have seen installations where the slow discharge gate was adding 45 seconds per batch and erasing the entire benefit of faster plasticizer uptake.

One warning specific to flexible sheet: the higher porosity of S-70 means it also picks up moisture faster during storage. If your resin silo is outdoors and the ambient dew point swings above 15°C, the surface moisture on the grain can increase by 0.05% in 24 hours. In rigid extrusion this is an annoyance; in flexible calendering it produces bubbles in the sheet that show up as pinholes after the embossing station. We had a customer in a coastal location who couldn't figure out why their pinhole rate tripled every July—turned out their outdoor silo was hitting 80% RH overnight and the resin was hydrating. They installed a jacketed silo with dried air purge at 2 m³/h and the problem disappeared in 48 hours.

Paste PVC (P-72) for Coated Fabrics and Flooring

P-72 is an emulsion-polymerized PVC with a K-value of 72 ± 1 and a primary particle size of 15-25 μm, spray-dried into agglomerates that break down under shear when you disperse them in plasticizer to form a plastisol. We produce this grade in 8-week campaigns, roughly 6,000 tonnes per campaign, because switching the line between suspension and emulsion polymerization is a 4-day cleaning and turnaround process and we only do it when order backlog justifies it. The paste viscosity behaviour is what matters—our P-72 yields a low-shear Brookfield viscosity of 3,000-5,000 cP at 60 phr DOP loading (spindle #6, 20 RPM), and more importantly, it shows pseudoplastic behaviour with a shear-thinning index of 0.65-0.72 between 2 RPM and 20 RPM. This means it coats fabric smoothly under a knife-over-roll coater at line speeds up to 40 m/min without striation marks.

The biggest failure mode with P-72 is plastisol ageing. If you let a mixed plastisol sit for more than 72 hours at ambient temperature without agitation, the plasticizer continues penetrating the particle agglomerates and your viscosity will climb by 40-60%. By 96 hours, you're looking at a gel that won't coat at all. This is not unique to our P-72—it's inherent to emulsion PVC chemistry—but our slightly higher surface area (BET 8-12 m²/g versus the 6-10 m²/g typical for paste grades) accelerates the ageing by about 15%. If your process involves extended plastisol hold times, either spec a lower-surface-area paste grade or agitate continuously at 30-50 RPM in a jacketed tank kept below 25°C. We've also seen customers try to recover aged plastisol by adding more plasticizer—this ruins the fusion rheology and produces a coated fabric with 30% lower tear strength. Once the plastisol has gelled from ageing, it's scrap. No recovery tricks work; I've tried.