Reb M Sweetener Selection for Short Dough and Cake Matrices

Replacing sucrose with rebaudioside M in short dough matrices cannot be reduced to a direct sweetness substitution. The base dough system, with fat-to-flour ratios of 0.5–0.7, added water of 10–15 wt%, and sucrose at 18–25 wt% of the flour-plus-sugar mass, relies on undissolved sugar crystals to interrupt gluten hydration, restrict starch swelling, and create a glassy, friable structure after baking. Rebaudioside M at 150–400 mg/kg of finished dough delivers the required sweetness potency but does not occupy the solid phase, bind water, or participate in Maillard browning. Formulation work therefore proceeds by fixing rebaudioside M as the sweetness source and then selecting co-formulants that restore bulk, water activity, spread behaviour, and surface colour. The spread ratio measured according to AACC International Method 10-50D and water activity measured by ISO 18787:2017 are the two primary release criteria in short dough work, because deviations greater than ±1.0 cm in spread or ±0.03 in a_w relative to the sucrose control alter packaging compatibility, rotary mould release, and shelf life. In rotary moulding, a dough that becomes more tacky or more cohesive after polyol addition will not release cleanly from die cavities, and in wire-cut lines, a lower-viscosity dough will produce piece-weight variability across the depositor. Production-scale records from rotary moulding units show that dough temperature at the die is a critical variable; when erythritol-containing blends cool the dough below 15 °C, the dough contracts and sticks to engraved cavities, producing misshapen units. The selection of rebaudioside M therefore sits inside a larger solids-replacement problem rather than a sweetener-only problem.

How Do Low-Molecular-Weight Polyols Alter Short Dough Structural Set During Baking?

In short dough, erythritol, allulose, inulin, and isomalt are not interchangeable because their thermal and water-binding behaviour diverges under the same oven conditions. Erythritol, with a molecular weight of 122 g/mol, has a strongly negative heat of solution and crystallizes during cooling; this can produce a fractured surface and a rapid hardness increase after baking, but it does not retain water, so a_w can fall below 0.45 if erythritol replaces more than 30 wt% of the original sucrose solids. Allulose is a reducing monosaccharide and participates in Maillard reactions; it reduces the loss of crust browning and limits the cooling effect, but it lowers the glass transition temperature of the baked matrix and can create a chewy texture at high substitution. Inulin-type oligofructose adds soluble fibre and contributes to water binding, but under acid pH and conventional tunnel-oven air temperatures above 180 °C, partial hydrolysis can release fructose, increasing browning and increasing the caloric contribution. Isomalt combines low hygroscopicity with resistance to acid hydrolysis and provides bulk close to sucrose, but it does not match sucrose solubility and can leave a coarse mouthfeel if the particle size distribution exceeds 150 µm. Practical formulations therefore use binary or ternary co-formulant systems, with the ratio adjusted by differential scanning calorimetry, three-point bend texture analysis, and spread testing. Pilot reports indicate that a blend containing 55–65 wt% erythritol and 35–45 wt% allulose on the original sucrose mass produces spread ratios within the target window in many wire-cut doughs, but published data for this specific configuration is limited. Processors should verify on their own equipment because the heat transfer rate in a 60 m tunnel oven varies with band load, air velocity, and product moisture release. The optimum addition level is not universal: it depends on whether the short dough is rotary-moulded, wire-cut, or sheeted, and on whether the baked piece is packaged in rigid trays or flexible film with moisture-barrier properties.

High-ratio cake batters present a different constraint set because sucrose at 20–30 wt% of the batter mass elevates starch gelatinization onset, delays egg protein denaturation, and stabilises the air cells formed during fat-sugar creaming. Shifting to rebaudioside M without a matched bulking system creates a batter that sets too early in the oven, traps less carbon dioxide and steam, and yields a collapsed, tunnelled, or gummy crumb. Specific gravity measurement with a tared 100 mL cup is the standard in-line control; values above 0.95 after mixing indicate insufficient air incorporation, while values below 0.80 indicate over-aeration that may destabilise the foam at the depositor. In cakes formulated with rebaudioside M and sugar alcohols, batter viscosity at 20 °C measured on a Brookfield RVDV-II+ viscometer with spindle 6 at 20 rpm must be kept within the range of the sucrose control; deviations larger than ±15% produce visible layer volume differences in high-ratio production. The native starch gelatinisation behaviour of the cake flour is also shifted because bulking agents do not uniformly mimic the plasticising effect of sucrose. Differential scanning calorimetry of model high-ratio batters shows that sucrose elevates the starch endotherm onset relative to water alone; lower-molecular-weight polyols can narrow this shift, causing the crumb to set before the gas cells have reached maximum expansion. This is the main mechanistic risk in reduced-sugar cakes: the failure is not sweetness loss but premature structure formation and gas retention failure. Rebaudioside M remains chemically stable through this process, but its presence does not arrest the physical changes.

When High-Ratio Cake Batter Falls Below pH 5.0 After Acid-Sensitive Bulking Agent Selection

The pH boundary matters because acidic batters catalyse the hydrolysis of some bulking agents and influence the sensory stability of rebaudioside M. Rebaudioside M is stable in dry storage at 25 °C and 60% RH for at least 24 months under sealed conditions, but aqueous systems at pH 4.0 subjected to extended holding at 80 °C can show measurable hydrolysis to steviol or other steviol glycosides. In cakes, the crumb pH after baking usually remains between 5.0 and 7.0; below 5.0, acid-catalysed inversion of added isomalt or oligofructose may increase reducing sugar content, and the residual rebaudioside M concentration should be confirmed by HPLC. If phosphoric acid is used as a leavening acid in a reduced-sugar formulation, the local pH before mixing may be below 4.5, and pre-dissolving rebaudioside M in an aqueous phase with prolonged residence time should be avoided. Temperature logging across the oven zones is also necessary because rebaudioside M in a high-moisture batter experiences a different thermal load than in short dough; the time above 120 °C at the crumb centre is typically less than 5 min for a 500 g cake, while crust regions can reach 180 °C within 8 min. If the formulation includes acid-thinned starch or fruit concentrates, the pH drop during baking is greater and the sweetener system should be buffered with sodium citrate or disodium phosphate to hold the batter above 5.2. In low-pH systems, sensory panels should be used to monitor the development of bitter degradation notes, because rebaudioside M degradation products are not necessarily visible by colour or volume measurement. The operational boundary is clear: at pH below 5.0 and high moisture, the formulation must be validated for sweetener stability rather than assumed stable because the dry powder was stable.

Because rebaudioside M has a slower sweetness onset and a shorter lingering aftertaste than rebaudioside A, sensory evaluations for cake and short dough require time-intensity profiling with trained assessors under ISO 8586:2023. Panel data from paired comparison tests show that bitterness is reduced by combining rebaudioside M with 2–5 wt% allulose or 1–2 wt% erythritol relative to finished product mass, but the dose-response curve depends on fat level and flavour system. In short dough, fat masks bitterness, permitting higher rebaudioside M levels without objectionable off-notes; in cakes, vanilla, cocoa solids, and dairy proteins bind hydrophobic off-flavour compounds and shift the detection threshold. Temporal dominance of sensation procedures following ISO 13299:2016 should be used because a clean sweetness profile at first bite does not predict aftertaste after 30 s. In a wire-cut chocolate chip format, the inclusion of cocoa powder at 8–10 wt% reduces the perceived metallic note associated with steviol glycosides; in plain shortbread, the same note becomes detectable above 350 mg/kg rebaudioside M among a subset of tasters, indicating that the application-specific ceiling must be set by panel data rather than by supplier recommendation. Moisture migration and storage hardness are also sensory defects. A short dough with a_w below 0.40 becomes brittle and may show surface cracking after 4 weeks at 25 °C and 50% RH; above 0.55 the risk of mould growth increases even with sorbate unless the pH remains below 5.0. Rebaudioside M does not bind water, so humectant co-formulants must be selected to keep a_w in the design window. Glycerol at 1–3 wt% is effective but softens texture and reduces spread; sorbitol at similar levels resists moisture uptake but increases caloric contribution relative to erythritol. These choices are matrix-specific and must be confirmed by accelerated storage at 30 °C and 65% RH for 8 weeks with both instrumental and sensory evaluation.

Thermal Load and pH Stability Boundaries for Reb M in Tunnel Ovens

Thermal degradation of rebaudioside M in baked matrices is governed by the combination of crumb moisture, local pH, and time-temperature history rather than by the maximum oven air temperature alone. In short dough, the low water content limits hydrolysis even when the product surface reaches 170–200 °C; in high-ratio cake batters, the crumb remains above 80% moisture during the first half of baking, which increases the hydrolytic vulnerability of the sweetener if the crumb pH drops below 5.0. Rebaudioside M is considered stable under typical baking conditions at neutral or mildly acidic pH, but published kinetic data for rebaudioside M in complex cake crumb is limited, and the processor should not extrapolate from aqueous buffer studies. Practical risk control uses oven zone profiling with thermocouples placed at the product centre and at the base, and the data are used to verify that the crumb does not remain between 60 °C and 90 °C for more than 10 min at pH below 5.0; this is the window in which hydrolysis may become measurable. The addition of buffering salts is often more effective than reducing Reb M dose because the target sweetness threshold of 250–400 mg/kg is narrow and cannot be lowered without losing sweetness quality. If the formulation uses calcium carbonate or sodium bicarbonate for leavening, the local pH may exceed 8.0 in the early mixing stage, and rebaudioside M can undergo alkaline degradation if held for prolonged periods before baking; this is rare in continuous mixing lines but can occur in batch systems with delayed ovens. For this reason, the dry rebaudioside M should be added with the flour fraction rather than dissolved in the alkaline liquid phase, and the batter holding time at room temperature should be less than 45 min before the depositor.

Production-scale troubleshooting records describe a characteristic failure mode when rebaudioside M-polyol blends are introduced into short dough without adjusting the dough temperature. The negative heat of solution of erythritol can reduce dough temperature at the rotary mould die from 19 °C to 14 °C, and this changes dough viscosity enough to produce piece-weight variance above ±5%. The defect is corrected by conditioning the dry blend to 18–20 °C before mixing and by reducing the first-zone oven temperature by 10 °C to compensate for lower surface moisture and altered browning. In high-ratio cake production with a continuous batter mixer, the same sweetener system can reduce batter specific gravity from 0.90 to 0.80 at the depositor, causing over-expansion and cell elongation; adjusting the emulsifier-to-sugar-alcohol ratio from 2.5% to 3.5% on flour mass restores air bubble stability without changing the rebaudioside M dose. These observations are not universal; they reflect specific equipment configurations and are provided as operational boundaries rather than as predictive models. Batch-to-batch variance in polyol particle size, ambient humidity above 60% RH, and flour protein differences all shift the optimum co-formulant ratio. Formulators must verify the system on the actual production line because the interaction between rebaudioside M, co-formulants, and dough rheology cannot be fully captured in benchtop mixing studies.