How to measure D90 in protein powder

Particle size distribution (PSD) is one of the most critical quality indicators for dry-fractionated plant protein powder. D90 means 90% of particles by volume are smaller than the stated diameter, directly influencing protein-starch separation efficiency, powder flowability, dispersion behaviour and final protein enrichment performance. For dry protein fractionation lines on protein-mill.com, stable D90 control ensures consistent liberation of protein bodies without excessive starch fragmentation. This article introduces standard sampling, testing procedures, interpretation and common pitfalls for measuring D90 of protein powder.

1. Definition Recap: D90

When laser particle size analysis generates a volume-based cumulative distribution curve:

• D10: 10% of particles are smaller than this size
• D90: 90% of particles are smaller than this size
  The remaining 10% consists of larger particles, typically undissociated protein–starch composites or agglomerates. In dry protein processing, operators monitor D90 to judge grinding effectiveness.

2. Recommended Instrument: Laser Diffraction Particle Size Analyser

Laser diffraction is the industry standard for pulse protein, starch and middling samples.
Alternative methods (sieve analysis) are not suitable: sieves cannot accurately measure fine powder below 45 μm, which covers protein bodies (3–10 μm) and fractured starch particles.

Two measurement modes

1. Wet dispersion (liquid phase)
   Disperse protein powder in a suitable solvent (deionised water or ethanol).
   Advantages: good de-agglomeration, stable test results.
   Caution: Some plant proteins hydrate and swell in water; ethanol is preferred to avoid particle swelling distortion.
2. Dry dispersion (gas phase)
   Powder is dispersed by compressed air inside the instrument.
   Best match for dry fractionation process conditions, reflects real particle state inside air classifier mills.
   Key challenge: Find optimal air pressure to break soft agglomerates without fracturing intact starch granules.

For dry protein processing R&D and factory QC: dry dispersion mode is preferred to guide grinding parameter tuning.

3. Step-by-Step Standard Testing Procedure

Step 1: Representative Sampling

Poor sampling is the top cause of unreliable D90 readings.

• Collect samples from steady running points: mill outlet, classifier fines, coarse starch stream, middling return line.
• Avoid grabbing powder from static pile surfaces (segregation occurs; fine protein floats, coarse starch settles).
• Use a sample thief, take multiple sub-samples over 5–10 minutes, blend to form a composite sample.
• Seal immediately to prevent moisture uptake; moisture causes agglomeration and shifts PSD results.

Step 2: Instrument Calibration & Baseline

1. Warm up laser analyser following manufacturer protocol.
2. Run background baseline to eliminate dust and optical interference.
3. Verify instrument with certified reference standard powder periodically.

Step 3: Dispersion Parameter Optimisation (Critical for Protein Powder)

Protein powder easily forms electrostatic agglomerates. You must distinguish:

• Soft agglomerates: particles stuck together by static; should be separated
• Hard composites: protein chemically/physically bonded to starch; real structural particles that should NOT be broken

Dry dispersion guidance:

• Start with medium dispersion air pressure; run replicate tests while gradually increasing pressure.
• If D90 keeps dropping significantly as pressure rises → agglomerates still exist.
• Once D90 stabilises at higher pressure → further pressure will start shattering starch granules, creating false fine fractions. Stop at this optimal pressure.

Step 4: Sample Loading & Measurement

• Feed a small, consistent powder quantity to maintain appropriate obscuration (follow instrument recommended range; too high = multiple scattering error; too low = poor signal).
• Run 3–5 repeated measurements on the same composite sample.
• Accept results only if D90 variation between replicates is below ±3–5%. Discard unstable data.

Step 5: Data Output

Extract volume-based PSD report and record D10, D50, D90 and full distribution curve.
For dry fractionation monitoring, focus on D90 and whether a bimodal distribution (protein peak + starch peak) exists.

4. How to Interpret D90 for Dry Protein Fractionation

Typical target for dehulled pulse powder after grinding:

• D90 range: 30–60 μm (varies by bean type)

1. D90 too high
   Large proportion of undissociated protein–starch composites. Cell walls not fully ruptured. During classification, composites go into coarse starch stream → low protein recovery. Need higher grinding rotor speed.
2. D90 too low
   Excessive over-grinding. Many starch granules shattered into fine fragments (<10 μm). Fine starch enters protein fines → protein purity cannot reach high levels (limits ability to hit 70% protein). Reduce grinding intensity.

5. Common Errors That Distort D90 Measurement

1. Moisture absorption before testing
   Powder absorbs humidity, forms hard agglomerates → measured D90 artificially higher than actual primary particle size.
2. Over-dispersion in dry test
   High air pressure breaks native starch granules in the analyser; PSD no longer represents real powder leaving the mill.
3. Under-dispersion
   Agglomerates counted as large single particles → D90 falsely elevated. Operators mistakenly increase grinding speed unnecessarily.
4. Using intensity-based instead of volume-based distribution
   Always select volume distribution; number distribution heavily over-represents tiny protein bodies and misleads process decisions.
5. Non-representative sampling due to particle segregation inside silos or conveying pipes.

6. Alternative Quick Screening Method (For On-Site Rough Reference)

If laser equipment is unavailable temporarily: air jet sieve can measure coarse fractions, but cannot capture fine protein particles. This method cannot generate accurate D90 values and is only for auxiliary trend checking, not formal QC.

7. Practical Application in Dry Protein Processing

• Track D90 continuously to stabilise grinding operation.
• Pair D90 data with offline protein content testing of fines and starch by-product.
• If protein purity declines while D90 decreases: clear sign of starch over-grinding.
• If protein recovery drops while D90 increases: insufficient liberation of protein-starch composites.

D90 of protein powder is reliably measured via laser diffraction, preferably with dry dispersion to mimic actual dry fractionation conditions. The core of accurate testing lies in representative sampling and properly set dispersion energy to separate static agglomerates without destroying native starch granules. Trend monitoring of D90 enables operators to adjust grinding intensity, optimise protein-starch separation, and move closer to target purity such as 70% protein in fines.