Articles

Bouling Chemical Co., Limited

Feed Pellet Conditioning Sorbate Retention for Moisture Transfer over Dry Mix Addition

In production-scale feed milling, the addition of potassium sorbate as a preservative into a dry mix prior to pelleting introduces a series of interacting moisture-transfer phenomena that determine the active compound’s final pellet retention and spatial distribution. A 1,000 kg batch of corn-soy mash conditioned at 78°C for 95 s with saturated steam at 1.5 bar g generally increases moisture from 12.4% to 16.8% wet basis according to ISO 6496:1999. The steam enthalpy of 2,224 kJ/kg at 1.5 bar g transfers both heat and water, producing a condensed surface film on feed particles whose thickness depends on particle size, surface free energy, and steam trap performance. Potassium sorbate has an aqueous solubility of 58.2 g/100 mL at 20°C; therefore a 4.4 percentage-point moisture increase provides approximately 44 kg of added water per tonne of mash, sufficient to dissolve 1.5 kg/t potassium sorbate at 0.15% w/w by a theoretical factor exceeding 15-fold at 20°C. The actual retention of sorbate after pelleting and cooling is not governed by the solubility limit but by the direction of moisture movement during cooling, the binding of the sorbate anion to polar feed components, and the loss of fines from the pellet surface. Published data for this specific configuration is limited; plant-level recovery envelopes from twin-shaft counterflow conditioners and horizontal pellet coolers typically show that dry-mix addition routes produce sorbate assay recoveries between 88% and 96% when added as a 1:9 cereal premix, whereas direct dry addition without premix can show batch-to-batch coefficients of variation exceeding 18%.

Does Steam Condensate Film Transport Sorbate into the Pellet Matrix or onto the Pellet Surface?

The mechanism by which sorbate redistributes during conditioning is governed by the physical state of the added water. When saturated steam at 1.0 bar g to 2.0 bar g enters a paddle conditioner, it condenses on the cooler mash surfaces if the mash temperature is below the steam saturation temperature of 120°C to 134°C. The resulting condensate film dissolves a fraction of the surface-exposed potassium sorbate crystals. The diffusion coefficient of potassium sorbate in dilute aqueous solution at 25°C is approximately 1.2 × 10-9 m2/s; however, the effective diffusivity in a starch–protein matrix at 16% moisture is several orders of magnitude lower due to high tortuosity and viscosity. This physicochemical difference creates a two-stage redistribution: rapid surface dissolution within the first 5–15 s of conditioning, followed by slower capillary penetration into particle pores over 60–120 s. In a counterflow paddle conditioner of 3.5 m length, 0.45 m diameter, and shaft speed 18 rpm, the average residence time is 90–120 s, but the residence-time distribution typically has a coefficient of variation of 30% to 50%, meaning some sorbate particles experience only 30 s of wetting while others receive 180 s. The consequence is a non-uniform axial concentration of sorbate in the conditioned mash, with under-wetted fractions retaining crystalline sorbate at the particle surface and over-wetted fractions allowing anion migration into starch granules. This variability is a known production failure mode observed on counterflow conditioners without baffle adjustment.

The dissolution rate of a potassium sorbate crystal in the condensate film can be described by the Noyes–Whitney boundary-layer model, dm/dt = D A (Cs - Cb)/h, where D is the diffusion coefficient in the film, A is the wetted crystal area, Cs is the saturated concentration at the crystal surface, Cb is the bulk film concentration, and h is the boundary-layer thickness. For an unagitated film on a feed particle, boundary-layer thickness may be on the order of 50 µm to 200 µm, giving mass-transfer coefficients between 1.0 × 10-5 m/s and 1.0 × 10-4 m/s. In a paddle conditioner, the mechanical motion reduces the boundary-layer thickness and increases the dissolution rate, but the film is not uniformly distributed. This results in sorbate dissolution efficiencies after 90 s of conditioning of 60% to 95% for direct dry addition, depending upon crystal particle size and local moisture. A sorbate crystal with a median particle size of 250 µm has a specific surface area of approximately 0.018 m²/g, whereas a ground crystal at 75 µm has approximately 0.060 m²/g; the finer particle dissolves faster but is also more prone to electrostatic dust loss before steam contact. Because sorbate at feed pH 5.8–6.5 exists predominantly as the dissociated anion, chemical degradation of potassium sorbate is not expected below 270°C. Losses above this temperature would result from oxidative cleavage, but such temperatures are not reached in a feed conditioner operating below 95°C. Thus, the retention limit is physical redistribution rather than thermal destruction.

Table 1 — Analytical and process-control references for sorbate retention trials
ParameterMethod or standard designationSampling pointControl range
Mash and pellet moistureISO 6496:1999Pre-conditioner, post-conditioner, cooler discharge12.0% to 18.0% wet basis; final 10.0% to 12.5%
SamplingISO 6497:2002Mixer discharge and pellet cooler dischargeMinimum 20 increments per 1,000 kg batch
Water activityISO 18787:2017Cooler discharge0.60 to 0.70
Pellet durability indexASABE S269.5Cooler discharge92% minimum
Potassium sorbate assayHPLC-UV at 254 nm, C18 column, methanol:phosphate buffer 70:30 v/vMixer discharge, cooler dischargeRecovery 90% to 100% of theoretical
Dry-mix addition of potassium sorbate into a horizontal ribbon mixer of 1,000 kg capacity and a tip speed of 1.2 m/s requires a particle-size distribution between 100 µm and 500 µm to avoid excessive dust formation. Direct addition of 0.15% w/w dry sorbate without a pre-blend typically results in adhesive segregation on ribbon surfaces and mixer walls; after 120 s of mixing, the coefficient of variation for sorbate assay in ten grab samples can exceed 18% as determined by HPLC-UV at 254 nm. In contrast, a 1:9 ground corn pre-blend discharged into the main mixer at 30 s reduces the assay CV to below 5.5% at 90 s. Dust collection systems remove 0.5% to 2.5% of the dry sorbate as suspended fines if cyclones and baghouse returns are not routed back to the mixer. Batch-to-batch retention differences in dry mix addition arise primarily from powder bridging in micro-ingredient feeders and from electrostatic attraction to plastic feeder components, which is not observed with aqueous sorbate injection. A production line with a screw feeder to the mixer may exhibit a 4% to 6% negative bias in sorbate dosage if the feeder hopper is allowed to run below 15% fill level. Because potassium sorbate is polar and hygroscopic, exposure to relative humidity above 60% in the mixer can cause particle agglomeration on the ribbon surfaces and lead to assay CV values above 10%. In plants located in humid regions, air conditioning of the micro-ingredient platform to 20°C to 25°C and 40% to 50% relative humidity is required to prevent sorbate bridging in feeder hoppers. The batch sequencing is also relevant: adding potassium sorbate at the same time as liquid choline chloride or molasses can create local high-moisture zones that strip sorbate from the dry mix and deposit it onto mixer walls; therefore, the sorbate pre-blend should be added after the liquid application and mixed for at least 90 s before the conditioner. The dry-mix route works best when the basal mash moisture is below 13.0% and the mixer discharge CV for dry matter is below 3.0%.

When Conditioning Moisture Exceeds 3.5 Percentage Points, What Retention Loss Mechanisms Dominate?

The interaction between conditioning temperature, moisture addition, and sorbate retention creates a processing window that narrows as both parameters increase. At a conditioning temperature of 80°C and moisture increase of 4.0 percentage points, the rate of water absorption into corn starch granules is sufficient to gelatinize a fraction of the starch, reducing free surface water and temporarily immobilizing dissolved sorbate. However, once the conditioned mash passes through a die with an L/D ratio of 10:1 and a compression ratio of 4:1, the pressure generated within the die is between 30 MPa and 80 MPa, and the temperature spike from viscous dissipation can reach 5°C to 10°C above the conditioned mash temperature. This additional thermal input lowers the viscosity of the water phase, increases the diffusion coefficient of sorbate anions, and forces moisture out of the pellet capillaries upon release to atmospheric pressure. At a die exit temperature above 88°C, flash evaporation occurs at the pellet surface, carrying sorbate to the outer 0.2–0.5 mm shell. If the cooler inlet air humidity is below 10 g water/kg dry air, this shell dries within 4–6 minutes and sorbate becomes accessible to abrasion. The processing window requiring tight control is therefore ±5°C around the set point and ±0.3 percentage points around the moisture target. Outside this band, pellet durability per ASABE S269.5 may drop below 92%, or sorbate recovery in the cooler discharge may fall below 90% of the theoretical dose.

The effect of dry mix addition versus conditioning moisture transfer is not symmetrical. Dry mix addition places the sorbate in the mixer, where it experiences bulk convection and mechanical shear before steam conditioning. In a twin-shaft paddle mixer, the shear rate at the paddle tip at 18 rpm and a tip radius of 0.225 m is approximately 0.42 m/s, which is insufficient to cause particle attrition of sorbate crystals but sufficient to break up premix agglomerates. During conditioning, dry sorbate crystals are wetted on the surface but the bulk of the crystalline sorbate remains in its original size distribution until solubility equilibrium is reached. When the same dose is delivered as a 10% aqueous solution into the conditioner at a point 300 mm downstream of the steam inlet, the sorbate is already solvated, eliminating the dissolution-rate limitation and reducing the effect of residence-time distribution. However, the water load from the solution must be subtracted from the steam moisture target to avoid exceeding the pellet mill moisture window. Adding 1.5 kg/t potassium sorbate as a 10% solution introduces 13.5 kg water/t, equivalent to a 1.35 percentage-point moisture increase; this displaces steam addition and alters the heat balance. Therefore, aqueous injection into the conditioner requires a pre-conditioner with separate steam and liquid injection ports and an in-line moisture sensor based on near-infrared absorbance at 1,940 nm to maintain the target total moisture. At moisture increases below 2.0 percentage points, sorbate retention is generally high because the condensate film is thinner and the drying front is less aggressive. At moisture increases between 2.0 and 3.5 percentage points, pellet durability improves linearly with moisture, but sorbate surface enrichment begins. At moisture increases above 3.5 percentage points, the process enters a critical zone in which the amount of free water exceeds the absorption capacity of the starch and protein matrix for the time available. This free water contains dissolved sorbate and acts as a transport medium during die passage and cooling. Published data for this specific configuration is limited, and the response should be validated by a production mass-balance study before changing addition strategy.

Pellet Cooler Airflow, Drying Front Migration, and Surface Sorbate Enrichment

Counterflow coolers typically operate with a bed depth of 1,000 mm to 1,500 mm and a specific airflow of 600 to 900 m³/h per tonne of pellets. The drying front moves from the bottom air inlet upward through the pellet bed, while pellets move downward; this counterflow configuration creates a temperature gradient of 20°C to 30°C between the hot pellet inlet and the cooled pellet discharge. Potassium sorbate dissolved in the pellet moisture migrates toward the evaporation front by capillary flow. This migration is governed by Darcy flux through the capillary pore network, with a pore water velocity of the order of 10-6 to 10-5 m/s in the early cooling phase. When the surface of the pellet reaches equilibrium moisture content with the cooler exhaust air, sorbate deposition occurs in a shell layer whose thickness is controlled by the Biot number for water vapor transfer. A higher drying rate, induced by low exhaust air relative humidity below 20%, produces a thinner and more concentrated sorbate shell, which is more susceptible to dusting during transport. A lower drying rate achieved by recirculating 30% of the exhaust air or by maintaining an exhaust air relative humidity of 35% to 45% allows diffusion back into the pellet core and yields a more uniform sorbate profile. These findings are line-specific; published data for sorbate migration in commercial pellet coolers is limited, and the cited operational ranges originate from production-scale moisture and sorbate mapping studies on corn-soy poultry feed.

Losses from the cooler are not limited to dust. Water-soluble sorbate can also be removed by condensation on the cooler walls and air plenum if the exhaust air temperature falls below the dew point. This condensation can accumulate and drip back onto the pellet bed, creating localized wet spots and uneven sorbate concentration. In a counterflow cooler processing 6 t/h of 4 mm pellets, a 2°C drop in exhaust air temperature below the dew point was observed to transfer 0.7 kg of water per hour to the cooler plenum; that water contained measurable potassium sorbate at concentrations up to 1.5 mg/mL. Preventive measures include insulating the cooler air discharge duct and maintaining exhaust air temperature at least 5°C above the dew point. Pellet fragility increases when moisture gradients are too steep; ASABE S269.5 durability index may decline from 94% to 86% if the pellet is cooled from 85°C to 20°C in less than 8 minutes. The cooling time requirement is a function of pellet diameter; a 4 mm pellet requires 10–15 minutes of retention, while a 6 mm pellet may require 18–25 minutes. Sorbate retention improves when the cooling profile is extended because the drying front is flatter and the surface shell is less enriched. Continuous monitoring of sorbate retention cannot be achieved by moisture measurement alone, because two batches with identical moisture and water activity can have very different sorbate surface enrichment depending on the drying path. The cooling curve should be logged with thermocouples at the inlet, middle, and discharge of the cooler, and the dew point of the exhaust air should be calculated from a capacitive relative-humidity sensor. If the exhaust relative humidity drops below 20% while the pellet discharge temperature remains above 8°C above ambient, the drying front is too sharp and sorbate shell formation is likely. Adjustments include reducing fan speed, increasing bed depth, or recirculating a portion of the exhaust air. These controls are especially important for 2.8 mm micro-pellets because the surface area-to-volume ratio is roughly 40% higher than for 4 mm pellets; the diffusion path to the surface is shorter and sorbate surface enrichment occurs more rapidly.

Validation of sorbate retention across a feed pelleting line should be conducted on a batch basis with sampling per ISO 6497:2002. A minimum of three consecutive batches of identical formulation should be sampled at the mixer discharge, the conditioner discharge, the pellet cooler discharge, and the finished-product conveyor. Moisture content of each sample should be determined according to ISO 6496:1999, water activity by ISO 18787:2017, and potassium sorbate by HPLC-UV at 254 nm. The coefficient of variation for sorbate assay in ten samples from a single batch should not exceed 5.0% for an in-control process. The mass balance for potassium sorbate across the pelleting and cooling steps should close within 90% to 100% of the theoretical addition rate. If the mass balance falls below 90%, the operator should inspect the dust collection system, the conditioner steam trap, the cooler air plenum, and the micro-ingredient feeder bridging, because these four locations account for the majority of sorbate losses observed on production lines. Pre-drying of the dry mix is required at ambient relative humidity above 60%, because hygroscopic caking of the sorbate in the feeder hopper can reduce dosage accuracy by 7% to 12%. The addition of potassium sorbate should not be combined with high concentrations of free water and certain amine-based additives simultaneously, because sorbate can participate in acid–base interactions that alter the electrical double layer on the feed particle surfaces and change the flowability of the conditioned mash.
Related Articles