Sucrose substitution in cake systems is not a piece of cake

The study addresses the challenge of replacing sucrose in cakes, where sucrose contributes not only sweetness but also key functional roles in rheology, aeration, and thermal setting. Cakes are categorized as foam (e.g., sponge), batter/emulsion (cream, pound), and chiffon types, with air incorporation and stability during mixing being critical for volume and crumb texture. Prior work linked batter viscosity to air incorporation and foam stability, with sucrose increasing the viscosity of the aqueous phase. The authors hypothesized that, when replacing sucrose, the quantity and quality of the batter’s aqueous phase (batter liquor, BL) – specifically its fraction, viscosity, and water proton mobility – govern air incorporation and batter density. They evaluated sucrose reduction and substitution with maltitol, mannitol, oligofructose, and inulin (and mixtures) in sponge cakes, and examined whether findings extend to cream and pound cakes. Because cake structure sets during baking through starch gelatinisation and protein denaturation, the timing of these transitions was also investigated to understand impacts on cake texture and quality.

The paper situates its work within literature showing: (1) batter viscosity correlates with efficient air incorporation and foam stability in cake systems; (2) sucrose increases aqueous-phase viscosity and stabilizes foams; (3) in sponge cakes, air is incorporated in the batter’s aqueous phase (BL), highlighting BL properties as critical; and (4) during baking, structure setting depends on overlapping thermal transitions of starch gelatinisation and protein denaturation, which are influenced by sucrose and water. Prior studies also note functional and health aspects of sucrose substitutes (polyols, fibers) and their solubility and sweetness profiles, motivating examination of different substitutes and their mixtures.

• Cake systems: Foam-type sponge cakes for primary mechanistic study; cream and pound cakes to test generalizability.
• Formulations: Sucrose control; sucrose reduction; full replacement with maltitol, mannitol, oligofructose, or inulin; mixtures of mannitol with oligofructose or inulin in varying ratios (recipe sets detailed in Tables 1 and 4). Water content standardized across batters.
• Batter preparation: Standardized mixing protocols for each cake type (sponge, cream, pound). Batter density (BD) measured by weighing 100 ml aliquots.
• BL isolation and properties: Sponge cake batters ultracentrifuged at 165,000 g (25 °C, 65 min) to isolate BL. BL viscosity measured with a Brookfield DV-III Ultra rheometer (spindle SC4-13R, 10 rpm) at 25 °C. Time-domain 1H NMR (CPMG sequence) quantified proton populations (A–D) and T2 relaxation times at 25 °C to assess water mobility; emphasis on population D (exchanging protons of water and dissolved sugars/substitutes). Proton distribution derived via inverse Laplace transform. Effective volumetric hydrogen bond density calculated per van der Sman & Mauer using literature values of molar properties and hydroxyl-group counts.
• Thermal analysis: Differential scanning calorimetry (DSC, TA Instruments Q2000) on wheat starch and freeze-dried egg white in media simulating BL compositions and on entire batters. Heating from 0 to 130 °C at 4 °C/min; peak temperatures recorded for wheat starch gelatinisation (T_gel,STA), egg white denaturation (T_den,EW), and batter transition (T_batter). Note: DSC not feasible for some high-mannitol/inulin formulations due to overlapping dissolution peaks.
• Structure setting/viscosity development: Rapid viscosity analysis (RVA Super 4) monitored batter viscosity during mixing and thermal profile; final viscosity (FV) and temperature at which structure setting starts (T_R) determined.
• Baking and analysis: Standardized baking conditions per cake type. Cake density (CD) measured from weight and volume; texture profile analysis (TPA) on crumb slices measured firmness, springiness, cohesiveness, and resilience. Statistics via Tukey’s test (p < 0.05); correlations via polynomial fits.

• BL quantity and water mobility govern air incorporation: High correlations observed between BL viscosity/quantity and functional outcomes. Specifically, BL fraction correlated strongly with T2,D RT (R2 = 0.93) and with batter density (BD) (R2 = 0.84). A minimum BL viscosity of approximately 60–80 mPa·s was necessary for appropriate BD (adequate aeration).
• Water mobility reflects intermolecular interactions: T2,D RT increased with reduced sucrose content and with poorly soluble substitutes, indicating higher water mobility and weaker water–solute interactions. T2,D RT was positively related to effective volumetric hydrogen bond density of the solute system.
• Solubility matters: • Maltitol (highly soluble) behaved similarly to sucrose: comparable BL fraction and viscosity, low T2,D, efficient air incorporation, and appropriate structure setting; cake density and texture close to control. • Mannitol (poorly soluble) reduced BL quantity, increased T2,D RT (higher water mobility), and led to higher BD (denser batter), worse aeration, and significantly poorer sponge cake texture (high crumb firmness, reduced springiness/cohesiveness). • Oligofructose (soluble, heterogeneous) provided adequate BL and low BD (good aeration), but DSC/RVA showed mismatched timings: starch gelatinisation was delayed relative to protein denaturation, raising T_batter and yielding poor texture despite good aeration. • Inulin did not fully dissolve; partial replacement of mannitol with inulin did not match sucrose performance and higher inulin fractions degraded quality markedly.
• Mixture strategy: Combining mannitol with oligofructose (e.g., 25% mannitol/75% oligofructose) overcame individual shortcomings: improved aeration (vs mannitol) and more suitable biopolymer transition timing (vs oligofructose alone), yielding batter and cake properties closer to sucrose reference.
• Extension to cream and pound cakes: Trends broadly held. In emulsion-type cakes, gas cells are also stabilized in the lipid phase, moderating the impact of BL properties on BD and CD. Maltitol again approximated sucrose; mannitol had limited effect on CD in cream/pound cakes but still impaired texture; oligofructose in mixtures could improve BD/CD but slightly delayed structure setting.
• Overall criteria for successful sucrose substitution: Substitutes or mixtures must (1) provide sufficient BL with low water mobility (low T2,D and adequate BL viscosity) to ensure air incorporation/stability and (2) ensure appropriate overlap/timing of starch gelatinisation and protein denaturation during baking for desirable texture.

The work demonstrates that sucrose’s key functional roles in cakes are twofold: it structures the batter’s aqueous phase (BL), promoting air incorporation and stability during mixing, and it modulates the thermal transitions during baking, governing when the matrix sets. Replacing sucrose successfully therefore requires replicating both functions. TD 1H NMR of BL provides a molecular-level descriptor (T2,D RT) that, together with BL viscosity and fraction, predicts aeration (BD) and relates to an effective hydrogen bond density metric. DSC and RVA establish whether substitutes align starch gelatinisation and protein denaturation at suitable temperatures to set structure at the right time. Maltitol best mimicked sucrose across these dimensions. Mannitol’s low solubility reduced BL and impaired aeration, damaging texture; oligofructose supported aeration but mistimed biopolymer transitions, also harming texture. Mixtures, particularly mannitol with oligofructose, balanced these trade-offs, approaching sucrose-like outcomes. In cream and pound cakes, lipid-phase stabilization reduces sensitivity to BL properties for BD/CD, but texture still depends on proper thermal transitions, so the two-criteria framework remains relevant.

This study identifies mechanistic criteria for sucrose replacement in cakes: successful substitutes or mixtures must generate sufficient batter liquor with low water mobility (captured by low T2,D and adequate BL viscosity) and ensure appropriate timing/overlap of starch gelatinisation and protein denaturation during baking. Maltitol closely reproduces sucrose functionality; mannitol and inulin perform poorly alone due to low solubility, while oligofructose, although enabling aeration, mistimes biopolymer transitions. Mixtures (e.g., mannitol with oligofructose) can overcome individual shortcomings and approach sucrose-like cake quality. These principles extend from sponge to cream and pound cakes. The findings provide practical, measurable targets (BL viscosity range, TD-NMR T2,D behavior, DSC/RVA transition profiles) to guide formulation and optimization of reduced-sucrose cakes.

• DSC constraints: Thermal transitions could not be resolved for several mannitol/inulin-rich formulations (e.g., recipes 4, 6, 7, 9) due to overlapping dissolution endotherms, limiting complete comparison across all substitutions.
• Solubility/heterogeneity effects: Oligofructose’s compositional heterogeneity broadened NMR relaxation distributions; inulin and mannitol’s incomplete dissolution introduced batter heterogeneity, complicating interpretation.
• Some observations lacked mechanistic explanation (e.g., certain cake density or texture deviations with mannitol) and were noted as not fully understood.
• Not all substitution ratios or combinations were tested (e.g., high inulin fractions were excluded after poor performance), and sensory evaluation was not reported.
• Sample sizes were modest (typically n = 3 replicates), which may limit generalizability.
• BL-based analyses were performed on ultracentrifuged fractions, which may not perfectly replicate in-batter microenvironments during processing.