Dry fractionation separates plant material into protein-rich fine fraction and starch-rich coarse fraction based on physical differences between liberated protein bodies and starch granules. Separation efficiency is defined by two core indicators: protein purity in fines and overall protein recovery rate. Many interconnected raw material, pretreatment, grinding, classification and operational factors determine final separation performance. This article systematically sorts all critical influencing factors for pulse dry processing, aligned with technical content on protein-mill.com.
1. Raw Material Characteristics
(1) Natural particle size gap between protein bodies and starch granules
Protein bodies are typically 3–10 μm. Starch granule size varies by crop.
- Larger starch granules (pea, fava bean: 15–65 μm) create a clear size boundary, enabling high separation efficiency.
- Small starch granules (chickpea) overlap in size with protein bodies, causing unavoidable mixing and lower separation sharpness.
(2) Baseline composition and inherent tissue structure
Higher native protein content in dehulled cotyledons improves enrichment potential. Thick cell walls, tightly embedded protein–starch matrices require stronger grinding for full dissociation. Oil-containing raw materials may form sticky agglomerates and interfere with sorting.
(3) Raw material moisture content
Optimal range: 8–10.5%.
- Too high: particles agglomerate; fine protein adheres to starch surfaces; airflow cannot separate individual particles.
- Too low: excessive static charge; starch is easily fractured into fine debris that contaminates protein fractions.
(4) Dehulling completeness
Seed coat contains crude fibre. Unremoved hull fragments report to both fine and coarse streams. Fibre dilutes protein purity and disrupts aerodynamic sorting. Residual hull content should be controlled below 1.5%.
2. Pretreatment Performance
- Incomplete cleaning introduces stones, foreign grains and immature beans, creating inconsistent feedstock hardness and particle behaviour.
- Uneven conditioning leads to moisture variation batch-to-batch, destabilising separation results.
3. Grinding Process Parameters (One of the Most Critical Stages)
The core objective of grinding: rupture cell walls to liberate protein and starch, without shattering intact starch granules.
(1) Grinding intensity and rotor speed
- Under-grinding: protein remains bonded with starch as composite particles. These composites behave like coarse particles and exit into the starch by-product → low protein recovery.
- Over-grinding: large starch granules break into micro-fine starch (<10 μm). Fine starch enters the protein stream → significantly reduces protein purity.
(2) Particle Size Distribution (PSD) after milling
Ideal PSD shows a bimodal distribution: distinct peaks for fine protein bodies and intact starch granules. A merged single peak indicates poor liberation or excessive starch fragmentation.
(3) Heat generation during grinding
High shear raises powder temperature. Severe heat triggers protein denaturation and promotes particle agglomeration, weakening separation efficiency. Low-temperature, low-shear grinding design is preferred.
4. Air Classification System Settings
Separation inside the classifier depends on balance between centrifugal force and air drag force.
(1) Classifier wheel rotational speed
- Higher speed: tighter cut-point; rejects more mixed composite particles → higher protein purity, lower recovery.
- Lower speed: wider cut-point; more protein composites pass into fines → higher recovery, lower purity.
(2) Main airflow and secondary air volume
Secondary air disperses agglomerated powder. Insufficient dispersion causes clusters of protein + starch to move unpredictably, increasing cross-contamination. Improper air balance blurs the separation cut-point.
(3) Feeding rate
Overfeeding leads to particle crowding inside the classification zone. Inter-particle collisions reduce sorting precision. Moderate, stable feed rate improves separation sharpness.
(4) Cut-point accuracy
The aerodynamic threshold that separates fines and coarse fractions must match the particle characteristics of the milled powder. Mismatched cut-point causes serious material mis-sorting.
5. Particle Surface Phenomena: Static and Agglomeration
After dry milling, particles carry static electricity. Fine protein bodies electrostatically adhere to large starch granule surfaces.
These attached protein particles cannot be separated by size and density alone, and are lost into the starch by-product. Anti-static conveying design and optimised secondary air dispersion help mitigate this issue.
6. Circuit Configuration & Material Recirculation
- Single-pass classification delivers limited separation effect. Multi-stage classification with middling recovery improves overall efficiency.
- Improper middling loop design: accumulation of hard-to-separate micro-composites builds up inside the circuit, gradually destabilising separation.
- Excessive recirculation load increases the proportion of fractured starch, negatively impacting protein purity.
7. Operation Stability & Batch Consistency
Frequent fluctuations in feedstock hardness, moisture, or feeding rate force the classifier to work under continuously changing conditions. Well-tuned parameters cannot maintain optimal separation if raw material quality drifts without real-time adjustment.
8. Supplementary Separation Equipment (If Deployed)
When size overlap exists between fine starch and protein bodies, pure air classification hits a performance ceiling. Integration of triboelectrostatic separation improves efficiency by separating particles based on surface charge difference, breaking the limitation of aerodynamic-only sorting.
Separation efficiency is jointly governed by:
- Raw material properties (starch granule size, moisture, dehulling quality);
- Grinding effectiveness (full liberation vs starch fragmentation);
- Air classifier tuning (wheel speed, airflow, feed rate, cut-point);
- Particle dispersion and electrostatic adhesion;
- Production circuit layout and closed-loop recirculation strategy.
Achieving high separation efficiency requires balancing protein purity and recovery. Optimising upstream grinding and pretreatment is more cost-effective than relying only on classifier adjustment downstream.
