Select soy-gel ingredients against the finished structure and the setting sequence. Glucono-delta-lactone (GDL) is an acidification route to screen for an undisturbed gel when its setting time and acidity fit the product. For pressed salt-set tofu, compare calcium sulfate and magnesium chloride across suitable dose ranges, judging curd formation, cutting integrity, and cooking loss within the required moisture range. Suspension or solution dosing determines feasibility; it cannot establish the preferred texture. Add carrageenan or starch only for a defined remaining requirement, and separate any added-salt effect from the performance of a commercial structuring blend.

This guide covers heated soymilk and hydrated soy protein isolate (SPI) systems for tofu and related gels. It excludes extrusion and enzyme-led structuring, which require different selection criteria.

Confirm That the Soy Base Can Form the Required Network

For soymilk, control source, protein concentration, total solids, filtration, and heating history. Keep oil and retained insoluble material comparable across trials.

For SPI, confirm hydrated protein concentration, dispersion, hydration time, and response to heating. Powder protein content alone cannot establish gel functionality.

Research on a specially prepared cold-gelling SPI showed that protein concentration during prior heating affected later network formation, even after reconstitution to the same concentration. Processing history therefore matters, while the appropriate heating schedule remains grade-specific. pubmed.ncbi.nlm.nih.gov

Keep soymilk and SPI trials separate, with a working reference for each base.

Shortlist the Coagulant Against the Production Sequence

Calcium and magnesium salts initiate protein association through changes in electrostatic interactions and ion–protein association. GDL hydrolyzes to gluconic acid, reducing electrostatic repulsion as pH falls.

A matched magnesium chloride–GDL study found similar precipitation trends with different pH responses. Its tofu-like precipitate model does not establish a common pH target or equivalent finished texture. pmc.ncbi.nlm.nih.gov

Use the following branches to organize trials. They are conditional selection guidance, rather than a literature-derived ranking of finished textures.

Establish the pressed product’s moisture range before comparing hardness. If coagulation prevents uniform distribution, test feasible changes in dilution or addition; exclude the route if distribution remains impractical.

A tofu study found benefits from a calcium sulfate–magnesium chloride blend under its tested heating and pressing conditions. It supports screening a blend without establishing a universal optimum ratio. pmc.ncbi.nlm.nih.gov

Calcium Sulfate and Magnesium Chloride: Compare Delivery and Product Response

Calcium sulfate is sparingly soluble. Specify its hydrate form and suspension preparation, then check delivery after the expected holding period. Settling that changes the first and last doses must be resolved before evaluating texture.

Magnesium chloride can be delivered as a solution. Check dilution, addition position, and mixing against the onset of aggregation. Granules near the inlet of a larger batch justify a distribution check before changing the ingredient.

For both salts, compare completion of distribution with aggregation onset. Then use finished moisture, cutting, and cooking results to judge the product.

GDL: Qualify the Whole Acidification Path

Record the time between dissolving GDL and contacting the soy base, because hydrolysis begins in water. Follow pH and temperature from addition through the required setting period.

Confirm that filling finishes before disruptive setting and that final texture and acidity are acceptable. Endpoint pH alone cannot establish this sequence.

Choose a Structuring Ingredient by the Function Still Missing

Evaluate structuring ingredients in separate batches against the selected coagulant-only control. Incorporate them at the appropriate dispersion and heating stage; this development sequence does not mean adding them to an already formed gel.

Carrageenan: Separate Gelling Types from Thickening Types

Carrageenan type and ion response guide screening; the soy composite requires confirmation.

FAO documents these basic distinctions; they do not rank soy formulations.

Research on individual and mixed κ/ι solutions found that calcium and potassium changed gelation and stiffness. This supports controlling ions in soy trials without predicting their exact response.

Identify added salts and standardization ingredients in a commercial blend. Where its calcium or potassium contribution could materially change the trial, and the composition is known, use a conditional comparison:

• The selected coagulant-only reference.
• The reference with the same additional salt contribution, without the added carrageenan.
• The complete carrageenan blend at the proposed dose.

Match the added salt identities and amounts and, where feasible, other carriers; keep the soy base, coagulant, water, and processing conditions consistent. This proposed trial distinguishes an effect achievable with the added salts alone from the performance of the complete blend. Any remaining difference can include polymer–ion interactions, so the effects should not be treated as simply additive.

If a meaningful matched control cannot be prepared, qualify the blend as a complete ingredient system and keep claims at that level.

Confirm hydration and evaluate texture after cooling and the intended reheating or storage exposure.

Starch: Accept the Consequences of a Composite Structure

For cook-up starch, swelling and gelatinization contribute to body and water distribution. In SPI–corn starch research, changing starch concentration altered gel strength, rheology, and water holding. The reported response belongs to that tested composite system.

Screen starch when the product brief allows a protein–starch structure and the process can achieve the grade’s required cooking conditions. Specify botanical source, modification status, and whether it is pregelatinized.

Resolve visibly uneven coagulation before screening starch. Reject additions that create unacceptable pastiness, flavor release, or protein-to-solids changes.

At fixed solids, replacing SPI with starch reduces protein; adding starch without replacement increases solids. Separate these trial designs and control water.

Build a Trial That Separates Ingredient Effects from Process Effects

Screen suitable dose ranges under a common, production-relevant procedure. Optimize promising candidates within permitted line adjustments, then compare their best workable conditions against the same finished-product targets.

Define Dose and Timing Consistently

State each dose basis: mass of finished mixture, volume of soymilk, or mass of protein. For salts, include hydrate form, assay basis, and solution or suspension concentration. Calculate delivered calcium or magnesium where needed to explain powder-dose differences.

Equal molar additions of different ions still do not establish functional equivalence. GDL needs a separate dose–acidification assessment.

Use first coagulant contact as the timing origin. Record completion of addition, completion of mixing, end of filling, and the start of undisturbed holding.

For pH, use matched sacrificial containers or a validated monitoring arrangement that avoids repeatedly disturbing the test gel. Compare readings at defined times and a consistent measurement temperature.

Define the Setting Observation Before Testing

For bench screening, use matched samples at predefined times. Specify tilt-test conditions and record persistent loss of flow alongside visible flocculation; loss of flow alone does not establish a continuous protein network.

With rheology, follow storage modulus (G′) under fixed temperature and small-deformation conditions. Predefine the operational onset rule; a modulus rise cannot replace finished cutting or spoonability assessment.

Keep Finished-Product Measurements Separate

Measure texture at a fixed age and temperature using consistent geometry and test settings. Record moisture alongside hardness.

For pressed tofu, hold curd handling, pressing, and drainage conditions constant during the first comparison. If yield matters, measure recovered protein as well as block mass.

Package drainage, centrifugal water holding, and cooking loss represent different exposures. A favorable centrifugal result alone cannot establish storage stability.

Use the Results to Decide What to Change

Interpret texture, moisture, and handling together:

• Higher hardness with lower moisture: investigate whether water loss explains the increase before concluding that the network has improved.
• Higher block mass with unchanged recovered protein: identify the additional retained water and judge whether it meets the finished-product specification.
• Uneven curd appearing during addition: test distribution and timing before introducing a structuring ingredient.
• A uniform initial gel that releases water during storage: compare a targeted structuring addition against the unchanged reference under that storage condition.
• Improved water retention with unacceptable gumminess or pastiness: reject the formulation against the product brief despite the numerical gain.

Challenge the shortlisted formulation against the mixing, temperature, or holding variation that matters most to the actual line.

ChemicalCell’s Food Additives directory provides a starting point for browsing ingredient categories. Food-grade suitability and permitted use require confirmation for the exact material and destination market.

Advance the formulation whose setting sequence fits production and whose texture, moisture, and relevant storage or cooking performance remain acceptable across the permitted process variation.