In the synthesis train of a polyamidoamine-epichlorohydrin wet strength additive, the organochlorine burden is established long before epichlorohydrin is metered into the reactor, because the molecular architecture of the polyamidoamine prepolymer controls the subsequent ratio of azetidinium ring formation to chlorohydrin side reactions. A conventional backbone produced from diethylenetriamine and adipic acid at a molar ratio of 1.00:0.97 in a 6,000 L glass-lined Pfaudler vessel equipped with a retreat-curve impeller and a vacuum distillation receiver is heated stepwise from 125°C to 170°C over 180 min, held at 170°C for 90 min, and then dehydrated under a vacuum ramp from 80 kPa to 12 kPa absolute until the distillate rate falls below 4 L/h. The resulting adipic acid modified polyamidoamine intermediate typically has a number-average molecular weight in the range of 18,000 g/mol to 35,000 g/mol, an acid value below 3.5 mg KOH/g dry solids, and a Brookfield viscosity at 50% solids of 1,200 mPa·s to 2,400 mPa·s at 25°C. The intermediate is diluted with deionized water to 38% solids and cooled to 58°C before epichlorohydrin is fed at a rate of 6.5 kg/min per 1,000 kg batch while the jacket water is held at 50°C to absorb the exothermic opening of the epoxide. In unmodified resins, the epichlorohydrin-to-amine molar ratio may reach 1.35:1.00 to secure sufficient wet tensile development; adipic acid chain extension allows that ratio to be reduced to 0.95:1.00 to 1.10:1.00 because the prepolymer has already achieved the required hydrodynamic volume and the reactive amine distribution is narrower. The lower excess epichlorohydrin directly reduces the steady-state concentration of unreacted epoxide and therefore reduces 1,3-dichloro-2-propanol and 3-chloro-1,2-propanediol formation in the finished resin. This shift is not attributable to simple dilution or to lower total solids; it is a stoichiometric consequence of replacing pendant primary amine terminations with amide repeat units that do not enter epoxide ring-opening at the same rate. The modified backbone also lowers the concentration of free chloride ion available for nucleophilic attack on unreacted epichlorohydrin because the final quench with sulfuric acid at pH 2.8 to 3.0 converts labile chlorohydrin intermediates to less reactive terminal diol structures. In production campaigns on the same 6,000 L vessel, the unmodified resin produced a time-integrated reactor wall temperature excursion of +8°C above the jacket setpoint during epichlorohydrin feed, while the adipic acid extended grade produced an excursion of only +3°C because fewer reactive primary amine sites were present at the start of the epoxide addition. The link between lower peak temperature and lower organochlorine formation is critical, since the half-life of epichlorohydrin hydrolysis at 60°C and pH 4.0 is shortened by a factor of 2.2 when the temperature is raised to 70°C. Industrial batch records commonly show that a 10°C overshoot during the hold phase increases residual 1,3-DCP by 120 mg/kg to 260 mg/kg dry resin in systems that have not been optimized for excess epoxide removal. The adipic acid modified backbone therefore benefits the organochlorine profile at the reactor control level as well as at the molecular composition level.

Residual 1,3-DCP Formation Is Governed by Chlorohydrin Hydrolysis Half-Life, Not Total Epichlorohydrin Feed Rate

The kinetic relationship between epichlorohydrin conversion and 1,3-dichloro-2-propanol accumulation in wet strength resin synthesis is frequently misunderstood in plant troubleshooting because the total epichlorohydrin charge is treated as the primary predictor of residual DCP. The more accurate predictor is the time-at-temperature after the epoxide has reacted with chloride ion, because 1,3-DCP arises when 3-chloro-1,2-propanediol or glycidol intermediates encounter free chloride before they can be quenched into non-extractable diols. In a reaction mass held at 65°C and pH 3.5, the pseudo-first-order rate constant for epichlorohydrin consumption is approximately 0.008 min⁻¹, but the rate of 1,3-DCP formation becomes measurable only after the free epichlorohydrin concentration drops below 0.5 wt% of the batch. This creates a process window in which a lower residual epichlorohydrin monomer is not automatically accompanied by a lower final DCP value; if the acidic hold time is extended to complete epoxide conversion, DCP can increase from 380 mg/kg to 1,240 mg/kg dry resin even though the initial epichlorohydrin charge was already reduced by the adipic acid modification. The industrial control remedy is to minimize the interval between full conversion and acid quench, not to simply lower the epichlorohydrin feed rate further. The adipic acid extended backbone supports a shorter interval because its higher initial prepolymer molecular weight and lower free amine density permit the oxidation-sensitive azetidinium ring to form at a lower epoxide-to-amine ratio, after which the batch can be quenched at pH 2.8 instead of waiting for the residual epoxide signal to disappear by slow hydrolysis. In a side-by-side comparison on a 10,000 L reactor, the unmodified grade required 240 min of reaction and 85 min of hold time after epichlorohydrin feed to reach a stable Gardner color and viscosity endpoint, while the adipic acid modified grade reached the same endpoint in 160 min of reaction and 35 min of hold time. The shorter hold time prevented the late-stage accumulation of DCP and reduced the total organically bound chlorine in a 0.5 wt% resin solution by approximately 38% relative to the unmodified control. This improvement is operationally meaningful because the material is used in papermaking white water circuits where organochlorine is continuously recirculated and concentrated by water closure.

What Limits AOX Reduction When Adipic Acid Is Introduced Above 0.25 Molar Equivalents?

Excess adipic acid in the prepolymer backbone imposes a practical ceiling on organochlorine reduction because it reduces cationic charge density and ultimately weakens the electrostatic retention of the resin on bleached pulp fibres. When the adipic acid-to-diethylenetriamine molar ratio exceeds 0.25 relative to the primary amine functionality, the zeta potential of the dilute resin solution at 0.01 wt% solids in 2 mM sodium sulfate declines from +18 mV to +6 mV. This shift is not linear; it is buffered by the secondary amine repeat units in the backbone until the acid modification passes the point at which terminal amine groups are capped and the amide-rich chain becomes only weakly cationic. The loss of charge density is observed directly on the paper machine as a drop in first-pass retention from 78% to 52% under dynamic drainage jar testing using a bleached softwood kraft furnish at 0.5 wt% consistency and 1,200 rpm mixing. Wet strength development is also impaired because the azetidinium ring density falls below 0.8 mol/kg dry resin, which is the empirical threshold for rapid crosslinking with carboxyl groups on cellulose at typical paper machine dryer temperatures of 105°C to 120°C. At a ring density above 0.8 mol/kg, the wet tensile index after 30 min water immersion measured according to TAPPI T 456 remains above 6.5 N·m/g on an 80 g/m² bleached hardwood sheet at 0.4 wt% resin addition. Below that threshold, wet tensile retention falls steeply to 4.1 N·m/g and the resin begins to fail the minimum wet strength specification of 5.5 N·m/g for paper towel and tissue applications. The same excess-acid condition reduces the storage stability of the final product because the lower pH accelerates the hydrolysis of the azetidinium ring into amino alcohol units. A formulation that exceeds 0.28 molar excess adipic acid should therefore be used only when a secondary cationic retention aid can compensate for the charge loss, and the batch must be monitored for pH drift with an inline pH probe having a resolution of 0.05 pH.

During full-scale wet-end evaluation on a 2.4 m Fourdrinier machine running 650 m/min with bleached eucalyptus and northern softwood kraft in a 70:30 ratio, the transition from a conventional PAAE wet strength resin to an adipic acid extended grade was completed over 7 days without changing machine speed, slice jet-to-wire ratio, or press loading. The wet strength resin was fed at 3.0 kg/t dry fibre into the thick stock pump suction before the fan pump, and the whitewater AOX concentration was measured every 8 h using ISO 11480. Under the unmodified resin, whitewater AOX averaged 0.14 mg/L; after the switch to the adipic acid modified grade, the 7-day average fell to 0.07 mg/L while the first-pass ash retention remained at 68% and the machine retained stable wet tensile of 7.0 N·m/g on 23 g/m² towelling. The lower cationic demand of the modified resin allowed the mill to reduce cationic polyacrylamide dosage from 0.6 kg/t to 0.35 kg/t without compromising drainage, and the furnish zeta potential at the headbox moved from -14 mV to -11 mV. Operational boundaries were identified during the trial: the resin addition point had to be separated from anionic optical brightening agent by at least 12 m of piping to avoid the formation of aggregates that plugged the pressure screens, and the diluted resin solution at 0.5% solids required pH maintenance above 3.8 to prevent precipitation. When the stock pH exceeded 8.5 for 45 min during an alkali washing cycle, the azetidinium ring in the adsorbed resin began to open prematurely and the wet tensile dropped by 0.8 N·m/g within the same production shift. This operational limit is more severe on closed water systems where conductivity can exceed 5,000 µS/cm, because the higher ionic strength compresses the electrostatic double layer and increases the sensitivity of retention to charge density loss. The trial demonstrated that the organochlorine benefit of the adipic acid modification cannot be separated from its effects on wet-end electrochemistry, and the resin is best introduced at a low-consistency point with inline static mixing at 1.2 m/s flow velocity.

Viscosity Build and Agitator Torque Profiles in 10,000 L Glass-Lined Reactors

Production-scale viscosity data from epichlorohydrin chain extension show that the adipic acid modified polyamidoamine backbone follows a two-stage viscosity progression: a slow induction period that corresponds to epoxide ring opening by secondary amines, followed by a rapid linear increase as azetidinium ring formation and partial branching consume the remaining low-molecular-weight fractions. In a 10,000 L glass-lined vessel with a two-stage retreat-curve impeller rotating at 45 rpm and a 15 kW motor, the Brookfield viscosity at 25% solids and 25°C rises from 80 mPa·s to 420 mPa·s over 90 min. The agitator torque during the same window moves from 18% to 54% of rated motor load, and the rate of torque increase in the final 15 min is 1.8% per minute. This torque signature is a more reliable endpoint indicator than pH or temperature because it responds directly to the rising power-law consistency index of the resin solution; the process control system is therefore programmed to stop the reaction at 500 mPa·s or 58% torque, whichever is reached first. A jacket cooling water failure above 70°C during this phase can raise the chain extension rate by 2.1× for every 10°C increase, reducing the time to gelation from 45 min to 12 min. In plant records, the most frequent failure mode for this grade is not gelation in the reactor but delayed gelation in the transfer line after the batch has been quenched and diluted, because the residual heat in the 1,500 L hold tank raises the temperature from 25°C to 38°C during the 4 h filling sequence. To prevent this, the quench water is supplied at 8°C and the transfer line is sized to maintain a flow velocity of 1.5 m/s with a Reynolds number above 10,000. A batch with a final viscosity above 600 mPa·s is diverted to a scrap recovery tank rather than blended into finished inventory, because its high molecular weight fraction will cause filter plugging in paper mills even after dilution. These viscosity and torque limits are specific to the adipic acid extended backbone; a conventional backbone with a higher free amine content can tolerate slightly higher final viscosity without gelation, but the organochlorine reduction target is lost if the process is allowed to reach that condition.

When Storage pH Falls Below 2.4, Azetidinium Ring Loss Accelerates

The acid quench step in adipic acid modified PAAE manufacture is designed to arrest the epichlorohydrin reaction and to convert residual glycidyl intermediates into non-volatile hydrolysis products, but the lower pH boundary is constrained by the stability of the azetidinium ring. Storage at pH 2.4 or below at 40°C increases the rate of azetidinium hydrolysis into amine alcohol units by approximately 2.8× relative to storage at pH 3.0, producing a measurable loss of wet tensile potential. In finished resin stored in a 30 m³ high-density polyethylene tank with bottom outlet and recirculation loop, a batch held at pH 2.2 for 30 days showed a reduction in wet tensile index from 7.2 N·m/g to 5.4 N·m/g on standard handsheets prepared according to TAPPI T 205 and tested by TAPPI T 456. The same batch showed an increase in the concentration of 3-chloro-1,2-propanediol from 95 mg/kg to 210 mg/kg dry resin, indicating that chlorohydrin reversibility becomes significant under strongly acidic storage. The production specification therefore fixes the final pH at 2.8 to 3.2 and the storage temperature at 10°C to 25°C for North American shipments. At pH 4.5 and above, the opposite problem appears: the resin solution becomes susceptible to microbial growth and the residual epoxide-derived intermediates may continue to re-equilibrate, causing viscosity drift upward by 10% to 25% per month. The lower storage pH boundary is especially important for tallow and tissue mills that pre-dilute resin to 5% solids and store it in day tanks under warm mezzanine conditions; the diluted solution ages faster because the hydrolysis kinetics are first-order with respect to the azetidinium concentration but the effective proton activity remains site-specific in the partially neutralized polyelectrolyte phase. Steel storage tanks are unsuitable because the chloride ion content of the resin, typically 0.8 wt% to 1.4 wt% on a dry solids basis, promotes pitting corrosion of 304 stainless steel at pH values below 2.5 in the presence of oxygen. The recommended storage system is HDPE or fiberglass-reinforced plastic with an epoxy vinyl ester liner, and the recirculation loop should be operated only 8 h/day to avoid shear-induced viscosity loss.

A complete regulatory verification study for an adipic acid modified wet strength resin requires quantification of residual chloropropanols in the liquid resin, migratable chloropropanols from finished paper, and total organically bound chlorine in the aqueous extract. The resin is analyzed by extraction into ethyl acetate at pH 8.0, followed by derivatization with heptafluorobutyric anhydride and gas chromatography with electron capture detection. The method achieves a limit of quantification of 0.05 mg/kg for 1,3-DCP and 0.03 mg/kg for 3-CPD. Finished paper is tested for AOX by combustion of the aqueous extract and microcoulometric titration according to ISO 11480. A typical bleached kraft sheet treated with 0.4 wt% adipic acid modified resin yields an AOX value of 0.06 mg/kg to 0.10 mg/kg, compared with 0.14 mg/kg to 0.22 mg/kg for a conventional resin under the same addition level and curing on a 105°C drum dryer. The wet strength retention is verified according to TAPPI T 456, with acceptance at ≥85% of the dry tensile index after 30 min immersion in deionized water at 23°C. Food contact compliance is assessed under 21 CFR 176.170 for aqueous and fatty food simulants, and the resin is listed only for use at levels not exceeding 0.5 wt% dry fiber when the resulting paper and paperboard are intended for contact with food. European regulatory status requires compliance with Regulation (EC) No 1272/2008 for classification of residual 1,3-DCP as a Category 1B carcinogen, and the EU Ecolabel for tissue paper requires AOX in the final product below 0.1 mg/kg measured by ISO 11480. The following matrix summarizes the measurement designations and acceptance limits that must be included in a production batch certificate for paper mills operating under ISO 14001 environmental management systems.

Batch-to-batch validation of the adipic acid modified resin is performed using a three-point resin addition series on a laboratory-scale dynamic sheet former at 0.2 wt%, 0.4 wt%, and 0.8 wt% dry fibre. The resulting handsheets are conditioned at 23°C and 50% RH for 24 h before dry tensile and wet tensile measurement. A batch is released only when the wet tensile at 0.4 wt% addition is at least 6.5 N·m/g and the sheet AOX is not more than 0.10 mg/kg by ISO 11480. This validation protocol does not eliminate the need for mill-specific retention and drainage trials, because the interaction between the resin and anionic trash varies with furnish conductivity and alkalinity.