Polyvinyl Alcohol (PVA) Desizing Wastewater: COD Problem and Treatment

Polyvinyl alcohol (PVA) is a water-soluble synthetic polymer produced by controlled hydrolysis of polyvinyl acetate. In textile warp sizing, PVA functions as a film former that binds fibres and reduces yarn hairiness during weaving. PVA add-on in sizing is commonly between 8% and 15% of dry yarn weight, depending on yarn count and construction. After weaving, hot-water desizing transfers nearly all of the applied PVA into the aqueous phase. The resulting wastewater stream carries a chemical oxygen demand (COD) that is dominated by dissolved polymer rather than starch or particulate matter. The theoretical oxygen demand of fully hydrolysed PVA, calculated from the empirical formula C₂H₄O, is 1.81 g O₂/g. Commercial grades with residual acetate groups typically display measured COD values between 1.7 g/g and 1.8 g/g when digested by ISO 6060:1989. This polymer-derived oxygen demand is not readily accessible to conventional biological treatment, creating a specific COD problem in textile effluents.

Why Does PVA Desizing Effluent Exhibit a Low BOD5/COD Ratio?

Under ISO 5815-1:2003 5-day biochemical oxygen demand testing, PVA desizing wastewater frequently shows a BOD5/COD ratio below 0.10 for high-molecular-weight, fully hydrolysed grades. The low ratio reflects the polymer’s molecular weight, chain stereochemistry, and hydrogen-bonded structure. Aerobic bacteria cannot transport intact PVA chains through the cell wall; degradation depends on extracellular oxidase and hydrolase secretion, which is slower than starch hydrolysis. Commercial PVA with weight-average molar mass above 60,000 Da and hydrolysis above 98 mol% is especially resistant to ready biodegradation. In OECD 301B screening, high-MW PVA often fails the 60% theoretical oxygen uptake threshold within the 28-day test window, while low-MW fractions below 10,000 Da may mineralise after microbial acclimation. Consequently, an activated sludge plant receives a load that appears as COD but does not behave as readily available BOD.

Reported COD concentrations in PVA desizing wash water depend on liquor ratio, countercurrent washing efficiency, and size add-on. At a typical size add-on of 10% and a single-stage wash at a 20:1 liquor ratio, the polymer contribution alone can exceed 9,000 mg/L COD; high add-on or low-flow washing can push combined desizing liquors above 20,000 mg/L. Countercurrent rinse trains reduce final volume-specific COD by concentrating the same polymer mass into a smaller first-stage volume. Field sampling should therefore identify whether a quoted COD value represents first-stage desizing concentrate, mixed rinses, or end-of-pipe combined effluent.

In a conventional slasher sizing operation, PVA is cooked at 88 °C to 95 °C and applied at a size-box viscosity between 25 mPa·s and 120 mPa·s depending on grade. After weaving, desizing uses hot water, usually with non-ionic surfactant or oxidative auxiliary, at temperatures from 70 °C to 90 °C. Fully hydrolysed PVA can gel on cooling, so collection sumps must be kept above 40 °C or equipped with continuous agitation. If a quiescent sump cools below the gel point, viscosity rises sharply and transfer pumps with low net positive suction head margins can cavitate. Sampling from a cooled, unmixed tank can also produce non-representative COD results because the polymer concentrates near the bottom or surface depending on density and gelation.

Fenton Oxidation Variables and Iron Sludge Management

Advanced oxidation with Fenton’s reagent is widely evaluated for PVA desizing wastewater because hydroxyl radicals cleave the polymer backbone and convert soluble long-chain macromolecules into smaller organic fragments. Bench-scale optimisation for PVA-laden streams commonly evaluates pH between 2.8 and 3.5, using ferrous sulfate heptahydrate and hydrogen peroxide. Reported molar ratios frequently fall between 1.0 mol H₂O₂ : 0.10 mol Fe²⁺ and 1.0 mol H₂O₂ : 0.20 mol Fe²⁺, corresponding to a Fe²⁺:H₂O₂ molar ratio of approximately 1:10 to 1:5. The oxidant demand scales with initial COD; bench testing must establish the H₂O₂:COD mass ratio for a given mill effluent because size auxiliaries, starch carry-over, and surfactants consume oxidant.

The principal operational limitation is ferric hydroxide sludge production. Neutralisation of spent Fenton effluent generates iron-bearing sludge with high water content; filter press dewatering usually requires lime or polymer conditioning. For illustrative mass balance, a textile finishing plant treating 100 m³/day with an initial COD of 8,000 mg/L and a H₂O₂:COD ratio of 1.0 g/g would consume approximately 800 kg/day hydrogen peroxide on a 100% basis, not including parasitic reactions with carbonate. Under a Fe²⁺:H₂O₂ molar ratio of 1:10, the dry Fe(OH)₃ production would be approximately 250 kg/day. The wet sludge mass and disposal cost can be substantially higher depending on dewatering performance.

In continuous Fenton systems, a common failure mode is ferric hydroxide scaling on static mixers, level probes, and neutralisation pH sensors. Probe fouling can occur within one 8-hour shift if the neutralisation tank is not agitated or if acid dosing overshoots. Plant operators should install redundant pH sensors and use flow-paced chemical dosing rather than oscillation-prone feedback loops. Hydrogen peroxide at 50% concentration must be handled as a strong oxidiser and stored separately from ferrous sulfate, sodium bisulfite, and acids. For COD loads above 10,000 mg/L, reaction cooling may be required because radical initiation is exothermic.

When Membrane Separation Is Placed Upstream of Biotreatment

Ultrafiltration can concentrate PVA before COD is discharged because the polymer’s hydrodynamic radius is much larger than common membrane pore sizes. Spiral-wound UF modules with molecular weight cut-off (MWCO) between 20,000 Da and 100,000 Da have been evaluated for sizing recovery trials, with transmembrane pressure maintained below 5 bar and crossflow velocity above 3 m/s to limit gel-layer formation. In such systems, the concentrate stream may contain PVA at 40 g/L to 80 g/L, allowing partial recycle to the sizing kitchen, while the permeate COD decreases by 40% to 70%. However, PVA is a known membrane foulant. The polymeric gel layer increases transmembrane pressure, reduces flux, and requires chemically enhanced cleaning with sodium hypochlorite or enzyme-based cleaners. Membrane systems are economically constrained by concentrate reuse; if the recovered PVA is mixed with starch, wax, and fibre fragments, it cannot be returned to high-speed weaving without purification. Published data for closed-loop PVA recovery from desizing wash water are largely limited to vertical textile operations where size formulations are controlled and fabric construction is stable.

Activated sludge processes receiving PVA desizing effluent without pre-treatment show removal patterns consistent with extracellular polymer attack rather than direct uptake. The soluble COD fraction remains elevated even when hydraulic retention time is extended from 8 h to 24 h. Heterotrophic biomass can partially degrade low-molecular-weight PVA fractions and acetate side groups, but the polymer backbone may persist. Acclimated cultures containing Sphingomonas or Pseudomonas species have been reported to mineralise PVA when supplementary nitrogen and trace nutrients are maintained at a BOD5:N:P ratio near 100:5:1. The degradation rate is slow, with first-order rate constants typically two orders of magnitude lower than those for starch. A sequencing batch reactor may need a solids retention time above 20 days to retain slow-growing PVA degraders, but many textile-plant biotreatment units operate at 10–15 days SRT to control sludge age. This mismatch explains why final COD compliance is often met only after source separation or physicochemical pre-treatment.

Anaerobic contact processes have also been tested for PVA-containing desizing streams. Methanogenic granules do not directly hydrolyse high-MW PVA; the bottleneck is depolymerisation. Sulphate-reducing and fermentative organisms can degrade PVA only if the polymer is partially oxidised first. Anaerobic treatment alone is therefore not recommended as a stand-alone COD removal step; it may serve after Fenton or ozonation to convert fragmentation products into methane.

Physicochemical Precipitation and Electrocoagulation for High-Strength Streams

Because PVA is non-ionic and highly water-soluble, conventional aluminium or ferric coagulants often fail to destabilise PVA at neutral pH. Effective physicochemical removal requires either salting-out with high concentrations of sodium sulfate, borate complexation, or electrocoagulation with sacrificial anodes. Electrocoagulation using aluminium or iron electrodes at current densities from 20 A/m² to 100 A/m² and inter-electrode gaps below 20 mm can remove PVA through a combination of anode hydrolysis, polymer adsorption onto metal hydroxide flocs, and entrapment. The process is pH-sensitive; aluminium electrodes work best near pH 6.0–7.5, while iron electrodes are reported to perform adequately from 6.5 to 8.0. Energy consumption is typically reported as 0.5 kWh/m³ to 2.0 kWh/m³ for COD reductions up to 60%, but electrode passivation and sludge production remain constraints.

Salting-out with sodium sulfate at concentrations above 50 g/L can phase-separate concentrated PVA from strongly acidic or neutral streams, but the resulting brine creates its own disposal burden. Borate crosslinking can gel PVA, which is useful for analysis but undesirable in drains because the gel adheres to piping. These physicochemical routes are best reserved for high-strength desizing concentrates rather than end-of-pipe dilute flows.

COD reporting for PVA desizing wastewater should distinguish filtered and unfiltered samples. Filtration through 0.45 µm membranes followed by ISO 6060:1989 digestion gives a reproducible soluble COD value. Unfiltered total COD includes size film fragments and fibre lint, which complicates treatment benchmarking. The BOD5 measured by ISO 5815-1:2003 should not be used as a surrogate for PVA removal because the method measures oxygen uptake from readily degradable fragments; the residual PVA may have negligible 5-day oxygen demand but still contributes to COD. Discharge permits are site-specific, so compliance evaluations require pairing laboratory COD data with flow-proportional composite sampling. For biological effluent, the BOD5:N:P ratio should be maintained near 100:5:1; for oxidative pre-treated streams followed by biotreatment, nutrient dosing is adjusted after measuring residual ammonia and orthophosphate.