The selection of purchased caustic soda for kraft pulping is governed less by the nominal sodium hydroxide concentration than by the impurity load that cycles through the chemical recovery loop and by the acid-consumption profile of the specific wood furnish. Effective alkali is customarily expressed as the sum of NaOH plus one-half Na₂S, calculated as Na₂O, and it is the primary control variable that determines the degree of delignification at a given H-factor and liquor-to-wood ratio. Hardwood species such as Eucalyptus globulus and Betula pendula contain acetylated xylan residues that hydrolyse during impregnation; the resulting acetic acid consumes sodium hydroxide before the main bulk delignification phase, reducing the hydroxide concentration available for lignin fragmentation. Softwood species such as Pinus taeda and Picea abies carry a higher lignin mass fraction and a more condensed guaiacyl-rich lignin structure, which demands a higher terminal hydroxide concentration to maintain delignification kinetics without excessive carbohydrate peeling. Purchased caustic soda is used both as make-up for sodium losses and as the primary alkali source in oxygen delignification, extraction, and specialty pulping stages; its specification must therefore address sodium chloride, sodium chlorate, iron, calcium, silica, and nickel as critically as NaOH concentration itself. Membrane cell grades are typically supplied at 49.5–50.5 wt% NaOH and diaphragm cell grades at 49.0–50.5 wt% NaOH; the operational distinction is not alkalinity but the chloride and chlorate burden that accumulates in closed-cycle mills. In a mill where sodium chloride is not effectively purged from the recovery cycle, chloride concentration in white liquor can rise to levels that increase molten-phase corrosion rates in recovery boiler superheaters, especially when flue gas temperatures exceed 550–600 °C. The kappa number of the resulting pulp is determined by ISO 302:2015, and the target range defines the alkali demand profile far more than a single total alkali charge.
| Parameter | Membrane cell caustic | Diaphragm cell caustic | Reference method |
|---|---|---|---|
| NaOH concentration | 49.5–50.5 wt% | 49.0–50.5 wt% | ASTM E291-18 |
| NaCl | 0.02–0.05 wt% | 0.8–1.2 wt% | ISO 979:1974 |
| NaClO₃ | 0.001–0.005 wt% | 0.05–0.15 wt% | ISO 981:1974 |
| Fe₂O₃ | 0.0005–0.002 wt% | 0.003–0.01 wt% | ISO 983:1974 |
| CaO | 0.0005–0.002 wt% | 0.002–0.005 wt% | ISO 986:1974 |
| SiO₂ | 0.001–0.003 wt% | 0.005–0.02 wt% | ISO 984:1974 |
| Nickel | 0.0001–0.0005 wt% | 0.0005–0.002 wt% | ASTM E291-18 or supplier lot certificate |
When eucalyptus and birch are cooked to bleachable-grade kappa values of 14–18, the effective alkali charge is strongly influenced by xylan acetyl substitution. For Eucalyptus globulus, acetyl content of 2.5–4.0 % on oven-dry wood is not uncommon; hydrolysis of each acetyl group releases acetic acid, and the resulting neutralisation consumes 1.3–2.0 kg NaOH per tonne of wood per weight percent of acetyl content. This consumption occurs early in the cook, before the heating plateau, and therefore reduces the alkali available for bulk delignification unless the charge is increased or the impregnation stage is prolonged at 110–130 °C. Sulphidity in hardwood cooking is usually held at 28–35 % to accelerate nucleophilic attack by hydrosulphide while limiting carbohydrate dissolution; the corresponding effective alkali for Eucalyptus globulus is often 14–16 % NaOH on oven-dry wood at H-factor 700–900, whereas Betula pendula may require 16–18 % effective alkali because of less reactive lignin and higher xylan mass. The selected caustic soda for hardwood lines should be membrane cell grade when the mill operates an oxygen delignification stage: iron in caustic accelerates hydrogen peroxide decomposition in the subsequent bleaching sequence, and the target is frequently a Fe mass fraction below 0.0005 % as Fe₂O₃, determined by ISO 983:1974. Chloride in purchased caustic is also restricted below 0.05 wt% NaCl because the hardwood bleach plant often uses hot acid stages on stainless steel equipment; chloride levels above 50 mg/L in the acid stage can initiate localised corrosion at welds and crevices if the pH falls below 2.0 and temperature exceeds 70–85 °C. A mill cooking hardwood to kappa 14 will observe higher residual alkali in blowline liquor than a mill cooking to kappa 18 if effective alkali is held constant; the residual alkali must remain above 8–10 g/L as Na₂O to avoid lignin condensation onto fibre surfaces, which manifests as low brightness and poor extractive bleachability. The caustic storage system for this grade should maintain tank temperature above 18 °C because 50 wt% sodium hydroxide begins to crystallise at approximately 12 °C, and viscosity at 10 °C can exceed the suction capability of metering pumps designed for low-viscosity fluids.
Mill-scale continuous digesters cooking radiata pine to kappa 28–32 for elemental chlorine-free bleaching frequently operate with a split alkali profile in which 55–65 % of the total effective alkali is added to the impregnation vessel at 115–130 °C and the remainder is introduced into the top of the cocurrent or countercurrent cooking zone. The high resin acid content of fresh pine chips, particularly when the furnish contains more than 5 wt% extractives-rich heartwood, consumes sodium hydroxide through saponification of abietic acid and pimaric acid during the initial liquor contact. Stoichiometrically, complete saponification of 1 kg of abietic acid of molar mass 302 g/mol consumes 132 g NaOH, so each 1 wt% increase in resin acid content adds at least 1.3 kg NaOH per tonne of oven-dry wood before any delignification demand is considered. This acid demand is overlooked when the caustic soda specification is based only on sodium hydroxide concentration, but it directly affects the residual effective alkali at the blowline and therefore kappa uniformity from chip to chip. Softwood continuous digesters of the hydraulic two-vessel type with a nominal capacity of 1,200–1,500 ADt/d have measured blowline kappa variation of ±1.5–2.0 kappa units when the effective alkali charge is held constant and the chip moisture varies by 2–5 percentage points, because wetter chips absorb less alkali at the initial impregnation and leave the circulation screens with lower hydroxide concentration. The use of membrane cell caustic with NaCl below 0.05 wt% in these machines is justified less by the digester metallurgy than by the recovery boiler: the chloride input from purchased caustic accumulates in the smelt and increases the chloride content of recovered white liquor, and a typical make-up rate of 8–12 kg NaOH/ADt with diaphragm-grade caustic at 1.0 wt% NaCl introduces 0.08–0.12 kg chloride/ADt directly into the closed liquor loop. Published data for this specific configuration is limited for proprietary mills, but public corrosion studies on recovery boiler superheater tubes indicate that chloride accelerates molten-phase corrosion when the tube surface temperature exceeds 450–500 °C, and this risk controls the allowable chloride input more tightly than the pulping chemistry itself.
| Species and product | Digester kappa target | Effective alkali, % NaOH on OD wood | Sulphidity, % | H-factor | Liquor-to-wood ratio, m³/t OD | Primary caustic purity driver |
|---|---|---|---|---|---|---|
| Eucalyptus globulus bleachable | 14–18 | 14–16 | 30–35 | 700–900 | 3.2–3.5 | Fe and chloride for ECF |
| Betula pendula bleachable | 14–18 | 16–18 | 30–35 | 700–900 | 3.0–3.3 | Fe control for peroxide |
| Pinus taeda bleachable | 28–32 | 18–22 | 25–30 | 1300–1600 | 3.5–4.0 | Chloride for recovery boiler |
| Pinus radiata linerboard | 55–65 | 15–18 | 25–30 | 900–1200 | 3.2–3.5 | Cost versus chloride tolerance |
| Acacia mangium bleachable | 16–20 | 15–18 | 25–30 | 600–800 | 3.0–3.5 | Fe and silica after caustic cleaning |
| Bambusa spp. bleachable | 18–22 | 18–22 | 25–30 | 1000–1300 | 3.0–3.5 | Silica and chloride in recovery |
Pinus taeda chips contain 28–30 wt% lignin on an extractive-free oven-dry basis, whereas Eucalyptus globulus typically contains 22–25 wt%. The softwood lignin is composed predominantly of guaiacyl units with a lower proportion of β-O-4 linkages accessible to hydroxide, and the condensed 5-5′ aryl-aryl structures resist fragmentation until the pH remains above 12.5–13.0 at the cooking temperature. Hardwood syringyl-rich lignin contains more β-O-4 linkages and fewer carbon-carbon condensations, so delignification proceeds at lower hydroxide concentration. In addition, the hemicellulose composition differs: softwood galactoglucomannan stabilises the fibre wall but does not supply the same acetyl-derived acid demand as hardwood xylan, so the pH in a softwood cook remains high only when the external effective alkali is sufficient to compensate for the lower carbohydrate dissolution. This is why a hardwood cook can delignify to kappa 16 with an effective alkali of 14–16 % on oven-dry wood, while a softwood cook to kappa 30 typically requires effective alkali of 18–22 %. The difference is not a linear function of lignin content alone; the effective alkali concentration must also offset extractive saponification and maintain residual alkali above 6–8 g/L as Na₂O at the end of a softwood cook to prevent reprecipitation of degraded lignin onto the exposed tracheid surfaces. The sulphidity of the softwood liquor is usually adjusted upward to 25–30 % to provide sufficient hydrosulphide for selective lignin cleavage, but hydrosulphide cannot replace hydroxide for neutralising acidic extractives or maintaining the phenolate ion concentration required for rapid fragmentation. In practice, mills select the same membrane cell caustic grade for pine and eucalyptus if an oxygen delignification stage follows the digester, but the pine line consumes more sodium hydroxide per tonne of pulp because the higher effective alkali is applied to a lower yield pulp and the sodium loss from the recovery cycle is proportionally larger. The relevant caustic specification for pine pulping therefore emphasises low chloride and low chlorate because the higher alkali charge amplifies the input of any impurity per tonne of pulp. A chloride specification of 0.05 wt% NaCl is appropriate if the recovery boiler operates at flue gas temperatures above 550 °C or if the mill reuses cleaned condensates for pulp washing without a dedicated chloride removal stage.
The chloride load introduced by purchased caustic soda is amplified in closed-cycle mills because sodium chloride remains soluble through green liquor clarification and causticising, and the precipitation of calcium carbonate does not carry chloride into the lime mud. Purge points such as electrostatic precipitator dust and dregs preferentially remove sulphate, carbonate, and potassium salts, while chloride tends to partition to the aqueous phase and recirculate. A mill replacing 10 kg NaOH/ADt with diaphragm-grade caustic containing 1.0 wt% NaCl adds 0.1 kg chloride per air-dried tonne; if the recovery cycle has a liquor inventory of 5 m³/ADt and no chloride purge, the theoretical steady-state increase is 20 mg/L per day before purging. This accumulation is not negligible because molten-phase corrosion of recovery boiler superheater tubes becomes aggressive when chloride-bearing deposits form on tube surfaces operating above 450–500 °C. Caustic soda specifications for recovery boiler protection therefore include a chloride limit below 0.05 wt% for membrane cell grade, a chlorate limit below 0.005 wt% to reduce oxidative attack on stainless steel in oxygen delignification and chemical recovery equipment, and a nickel limit below 0.0005 wt% because nickel catalyses the decomposition of hypochlorite and hydrogen peroxide in bleach stages. Iron in purchased caustic should be kept below 0.002 wt% Fe₂O₃ when the pulp is destined for viscose or acetate derivatives, and the analytical acceptance protocols are ISO 979:1974 for chloride, ISO 981:1974 for chlorate, ISO 983:1974 for iron, ISO 984:1974 for silica, and ISO 986:1974 for calcium. In mills that use purchased caustic only for sodium make-up and not for direct digester alkali, the same impurity limits are still applied because the make-up cation enters the recovery loop and any chloride or chlorate introduced with it eventually appears in the white liquor. The operational boundary is particularly tight in mills that close the bleach plant filtrates and send them to the recovery cycle; in such configurations, a chloride rise in white liquor above 5–10 g/L as NaCl can increase corrosion in the causticiser and lime kiln chains, and the only practical response is a stricter raw material specification or a dedicated chloride removal unit such as an electrostatic precipitator ash purge with modified dust removal.
When the digester kappa target for eucalyptus or acacia is pushed below 14 to support elemental chlorine-free bleaching sequences, the alkali profile must be shifted from a simple bulk delignification strategy to an extended delignification strategy with multiple alkali addition points. In this mode, the initial effective alkali is set to maintain the impregnation pH below 13.0 at 110–120 °C to protect glucomannan and xylan from peeling, while the final effective alkali is increased to hold the terminal residual alkali above 8–10 g/L as Na₂O. The total effective alkali for Eucalyptus globulus at kappa 10–12 may reach 16–18 % on oven-dry wood, an increase of 2–3 percentage points over bleachable-grade cooking. This higher caustic charge increases the sodium inventory in the recovery cycle and makes the impurity content of purchased caustic more impactful: every tonne of NaOH charged carries chloride, chlorate, and iron into the liquor loop, and at low kappa the oxygen delignification stage following the digester operates at 90–110 °C and 6–10 bar oxygen partial pressure, where chloride stress corrosion of austenitic stainless steel is a known failure mode if the chloride concentration exceeds the material-specific threshold. The caustic used in oxygen delignification is not recovered white liquor alone; mills frequently inject membrane cell caustic to control pH, and the material compatibility boundary for duplex stainless steel in the oxygen reactor is commonly stated as chloride below 0.05 wt% in the incoming caustic. At kappa values below 14, the residual lignin is more oxidised and less hydrophobic, so the fibre surface has an increased tendency to retain transition metals; iron from caustic or process water at concentrations above 1–2 mg/kg pulp can contribute to brightness reversion and hydrogen peroxide instability in the final brightening stage. The quality verification of low-kappa pulping caustic therefore includes not only a certificate of analysis but also periodic loop audits against ISO 979:1974, ISO 981:1974, and ISO 983:1974. Batch-to-batch variance in membrane cell caustic is usually smaller than 0.02 percentage points for NaCl and 0.001 percentage points for Fe₂O₃, but when a mill uses caustic from multiple suppliers or truck shipments are not heated, stratification can occur and the impurity profile at the tank bottom may differ from the certified top sample. For this reason, unloading lines should be fitted with in-line conductivity and density meters calibrated against ASTM E291-18, and the storage tank recirculation loop should operate at 30–60 min turnover to homogenise the concentration before dosing.
Prehydrolysis kraft pulping of eucalyptus for dissolving pulp targets kappa 8–10, alpha cellulose mass fraction above 92 %, and intrinsic viscosity in the range 450–550 mL/g determined by ISO 5351:2010. The prehydrolysis step removes xylan and acetyl groups, so the subsequent kraft cook has lower acetic acid demand but requires tight control of hydroxide concentration because the remaining cellulose is more susceptible to alkaline hydrolysis when lignin is nearly exhausted. The NaOH charge is usually split into an initial impregnation charge at 110–125 °C and a bulk delignification charge at 150–165 °C, with total effective alkali of 12–16 % on oven-dry wood and sulphidity of 25–30 %. The caustic soda used for dissolving pulp must be selected for low transition metals and low silicon because the final dried pulp enters viscose or lyocell processes where residual iron, nickel, and silicon alter xanthation or spinning performance. A conventional membrane cell caustic with Fe below 0.0005 wt% as Fe₂O₃ and SiO₂ below 0.002 wt% is normally adequate, but the certificate of analysis must report these elements against ISO 983:1974 and ISO 984:1974 at the lot level. Published data for this specific configuration is limited for proprietary dissolving pulp lines; however, the quality control principle is well established because residual transition metals in bleached dissolving pulp accelerate chain scission in the viscose process and reduce spinnability. The caustic storage and dosing system for this application should be isolated from aluminium alloys and from galvanised piping because sodium hydroxide corrodes both materials at concentrations above 5 wt%, and the reaction with aluminium generates hydrogen gas. Carbon steel is acceptable for 50 wt% sodium hydroxide storage if the tank is stress-relieved and the temperature is maintained above 18 °C to avoid caustic embrittlement at weld seams; aqueous sodium hydroxide at 50 wt% and ambient temperature has a vapour pressure close to that of water and does not require pressurised storage, but the vent should be fitted with a desiccant or caustic scrubber to prevent atmospheric CO₂ ingress and sodium carbonate formation. In closed-cycle dissolving pulp operations, chlorate input from purchased caustic must be limited to 0.005 wt% as NaClO₃ because chlorate can consume hydrosulphide in the white liquor and increase residual oxidant load in the recovery cycle. Sodium chloride is less damaging in this grade because the prehydrolysis liquor does not burn; however, chloride still exits through acid process streams and can attack stainless steel tanks used for acid hydrolysis of hemicellulose. The final selection of caustic soda for dissolving pulp therefore prioritises purity over delivered cost per dry tonne, with specification limits verified by ISO 979:1974 for chloride, ISO 981:1974 for chlorate, ISO 983:1974 for iron, ISO 984:1974 for silica, and ISO 986:1974 for calcium.