Isolated soy protein is the most concentrated form in which soy protein is sold: 90% crude protein on a dry basis, made by alkaline extraction of defatted soy flour followed by isoelectric precipitation, neutralisation and spray drying. It is also the most misunderstood, because “90% protein” describes only the composition – and composition is not the property that decides whether a batch works in your plant. Two isolates can carry the same certificate and behave in opposite ways: one disperses to a thin, pumpable liquid at 15% solids, the other thickens into a gel that binds four to five times its own weight in water and holds fat in an emulsion.
This article covers the manufacturing route and what each step sets, the two functional families (dispersion/extrusion and meat systems), the basis arithmetic that decides how much protein you actually buy, the dispersion chemistry behind viscosity, and the microbiological logic behind a statistical sampling plan.
PolyBlueChem supplies both families: the extrusion grade (soluble, very low viscosity) and the meat type (high water and fat binding).
1. How isolated soy protein is made
| Step | What happens | Property it sets |
|---|---|---|
| Defatting | solvent extraction of oil from soy flakes | residual fat (typically 1% max in the isolate) |
| Alkaline extraction | protein dissolved at pH 8-9, fibre and insolubles removed | protein yield and the removal of hull and cell-wall material |
| Isoelectric precipitation | pH lowered to about 4.5, where the protein is least soluble, and the curd separated | purity, and part of the functional character |
| Washing | water wash of the curd | removal of sugars and phytate; flavour neutrality |
| Neutralisation | pH raised with sodium or potassium hydroxide | ash content – and therefore the sodium load |
| Spray drying | drying to a powder, with heat history controlled | dispersibility, viscosity and solubility |
Two of these steps explain most of the differences between commercial grades. The neutralisation step determines ash: an isolate neutralised with sodium hydroxide carries sodium into the ash figure, so a grade specified at 6% ash is not simply “cleaner” than one at 8% – it was neutralised differently, and if you run a low-sodium claim that difference is yours to test. The drying and heat history step determines how much the protein has aggregated, which is the practical origin of the viscosity difference between an extrusion grade and a standard isolate.
2. Two functional families, one certificate
| Property | Extrusion / dispersion grade | Meat type |
|---|---|---|
| Protein (dry basis) | 90% min | 90% min |
| Ash | 6% max | 8% max |
| Moisture | 7% max | 7% max |
| pH | 6.5 – 7.5 | 7.0 +/- 1.0 (5% slurry) |
| Particle size | 95% through 100 mesh | 96% through 100 mesh |
| Dispersion viscosity at high solids | low – designed for pumping and extrusion | higher; gels on heating |
| Water holding | moderate | high – the defining property |
| Fat binding / emulsification | moderate | high |
| Microbiological plan | single limit, 20,000 cfu/g | two-class plan n=5 c=2 m=5,000 M=10,000 |
The specification differences look small and the functional difference is not. A meat processor who buys on price and receives a low-viscosity dispersion grade will see it as poor water binding and a wet cook; a bar manufacturer who receives the meat type will see it as an unworkable, sticky mass. Neither failure looks like a supplier problem from the inside, which is why the application has to be stated before the grade is quoted.
3. Basis: what 90% really means
| If 90% is quoted… | At 7% moisture | Protein in 25 kg |
|---|---|---|
| Dry basis | 83.7% as-is | 20.9 kg |
| As-is | 96.8% dry basis | 22.5 kg |
| Difference | about 7 percentage points | 1.6 kg per bag |
The nitrogen factor is the analytical cross-check: 90% crude protein corresponds to 90 / 6.25 = 14.4% nitrogen. A Kjeldahl result of 13.8% N is 86.3% crude protein. If a certificate reports both protein and nitrogen, divide the protein by 6.25 and see whether the two agree – a mismatch usually means one of the numbers is a template value.
There is a second, subtler issue. The 6.25 factor is a convention, not a measurement: it assumes protein is 16% nitrogen. Soy protein is closer to 5.7 in some calculations, and the amino acid sum is the more honest number. For labelling in most markets the 6.25 convention is what applies, so use it – but do not be surprised when a laboratory’s amino acid sum differs from the crude protein figure by a few percent.
4. Dispersion chemistry: why viscosity is a specification
When you disperse an isolate in water, viscosity rises with concentration, and the shape of that curve is what separates an extrusion grade from an ordinary one.
| Variable | Effect on viscosity | Practical use |
|---|---|---|
| Concentration | rises steeply above 10-12% solids for a standard isolate | the reason high-protein beverages and extrusion melts need a low-viscosity grade |
| pH | minimum solubility at about pH 4.5 (isoelectric point); viscosity falls as the protein is more soluble away from it | formulations run at pH 6.5-7.5 for solubility; acidified drinks need stabilisers |
| Ionic strength | salts screen charges and can raise or lower viscosity depending on the ion | why a brine behaves differently from plain water |
| Heat history | heating above the denaturation temperature gels the protein irreversibly | the basis of meat binding – and the reason a UHT beverage needs a heat-stable grade |
| Shear | most dispersions are shear-thinning | pumping and mixing feel easier than the bench viscosity suggests |
That table is also the diagnosis guide. A dispersion that is unexpectedly thick usually has: too much solids, a pH near 4.5, a high calcium or salt load, or a grade that has been heat-damaged. A dispersion that will not dissolve at all usually has: powder added too fast into a vortex-poor tank, or powder that has absorbed moisture in storage and formed case-hardened lumps.
The standard dispersion procedure is worth stating because so many failures come from it: add the powder into water in a thin stream under high shear, at 40-50 degC rather than hot, keep the pH above 6.5, and let it hydrate for 20-30 minutes before adjusting. Powder dumped into a tank in one charge forms lumps whose interiors never hydrate, and no amount of subsequent mixing recovers them.
5. Meat systems: the water arithmetic
In processed meat, isolate is used for yield. The arithmetic that matters is the brine calculation, and it is worth doing in full before any trial:
| Parameter | Example | Note |
|---|---|---|
| Raw meat | 1,000 kg | – |
| Injection rate | 30% | typical for cooked ham |
| Brine volume | 300 kg | 30% of raw weight |
| Isolate in brine | 6% | 18 kg of isolate |
| Water:isolate ratio achieved | 282 : 18 | about 15:1 water to isolate in the brine itself |
| Typical isolate hydration capacity | 4-5 times its weight | so 18 kg can hold about 72-90 kg of the 282 kg water |
That last line is the honest way to think about it: the isolate does not “hold the whole brine”, it holds its own share and the salt, phosphate and processing do the rest. Buying more isolate than the system can hydrate wastes money and produces a rubbery texture; buying too little gives purge in the pack. The functional ratio is a starting point for trials, not a guarantee, because comminution energy, pH, salt level and cook schedule all move it.
Application detail – injection versus tumbling, the order of addition, and the failure signatures – is in Soy Protein Water Binding in Processed Meat.
6. Extrusion and high-protein bars
Where the meat type builds viscosity, the extrusion grade avoids it. In a cooker-extruder, the protein phase competes with the starch for water and contributes to melt viscosity; a grade that thickens early limits how much protein you can put through the die before the torque or pressure limit stops you. A low-viscosity grade raises that ceiling, which is the technical reason protein crisps and high-protein cereals can reach 60-80% protein in the finished piece.
In bars the same property shows up as workability: a low-viscosity protein phase mixes into a syrup binder without climbing the mixer, and the bar sets by binder crystallisation rather than by protein aggregation, so the texture after six weeks is closer to the design intent. Combined with a crisp for crunch, it is the standard architecture of a modern high-protein bar – see Soy Crisp vs Isolate vs Whey Crisp.
7. Microbiology and sampling plans
The two grades carry different microbiological instruments, and the difference is deliberate:
| Grade | Plan | Assumption |
|---|---|---|
| Extrusion grade | single limit: 20,000 cfu/g max | the product will be extruded, baked or otherwise heat-treated |
| Meat type | two-class: n=5, c=2, m=5,000, M=10,000 cfu/g; coliform n=5, c=1, m=10, M=100 | a tighter, consistency-based control for products where the protein enters a cooked but complex system |
Read n, c, m and M as: five samples per lot, up to two may exceed the m level, none may exceed the M level. It controls the distribution, not just the average – which is what matters when one high-count bag can spoil a production run. If your process has no lethality step after the protein is added, specify the two-class plan regardless of grade.
8. Certification and origin
Both grades are non-GMO, purely vegetable in origin, with no animal derivatives, and are produced under an identity-preserved chain. The document set – IP, ISO, BRC, HALAL, KOSHER, HACCP – is what allows a customer to onboard the material without a plant visit. At the purchasing desk, the two that matter most are the IP certificate (because soy is the most widely planted genetically modified crop) and the BRC certificate (because it is the audit evidence most retail programmes accept).
9. Purchasing guide
- State the application before asking for a price: extrusion, beverage, bar, or meat system. The grade follows from it.
- Fix the protein basis and the moisture limit together.
- Ask for the ash figure, and for sodium if you have a low-sodium target.
- For meat use, ask for the water-holding result the plant measures, and how it is measured.
- For extrusion and beverages, ask for a dispersion viscosity at a stated solids level and temperature – without a method the number is meaningless.
- Specify the microbiological plan, not just a limit.
- Require the IP and BRC documents with the first shipment, and confirm the certificate dates have not expired.
10. Frequently asked questions
Is isolated soy protein the same as soy protein concentrate?
No. Concentrate is typically 65-70% protein and retains more of the original fibre and carbohydrates; isolate is above 90% protein on a dry basis and is made by aqueous extraction and isoelectric precipitation. They are not interchangeable in high-protein applications.
Why does one grade gel and another not?
Heat history and the degree of protein aggregation during manufacture, plus the neutralisation route. Gelation is a functional property to be specified, not a defect – the meat type gels on heating, which is how it binds; the extrusion grade is designed not to gel early.
Can I use the meat type in a high-protein beverage?
Usually not successfully: it thickens the dispersion and can sediment or gel on sterilisation. Use the low-viscosity grade and confirm heat stability with a trial.
Is it suitable for vegan and vegetarian products?
Yes – it is purely vegetable in origin with no animal derivatives, and HALAL and KOSHER certificates support that position. Vegan claims are still the brand owner’s responsibility to substantiate for the finished product.
What is the shelf life?
Typically 12 to 24 months unopened, dry, below 25 degC. Once opened, hygroscopicity and caking are the practical limit, so plan consumption rather than storage.
References and scope note
The arithmetic above – nitrogen conversion, basis conversion, brine composition and hydration ratios – is reproducible from the numbers shown. Hydration capacity (4-5 times weight) is an indicative industry range for orientation, not a specification; confirm it on your own system. No health, nutrition or therapeutic claims are made for this ingredient.
Technical and commercial orientation only. Not a substitute for product development, process validation or regulatory review.
Grade selection is set out in Extrusion Grade vs Meat Type. Contact PolyBlueChem at polybluechem@163.com for samples of both grades, certificates, or a specification written against your process.