Rebaudioside M Behavior in Continuous UHT Dairy and Fermented Plant Based Lines

The application of rebaudioside M (Reb M; C₄₄H₇₀O₂₃; molar mass 967.01 g·mol⁻¹) as a high-intensity sweetener in continuous UHT dairy processing and fermented plant-based beverage manufacture presents distinct thermodynamic, kinetic, and rheological challenges that are not fully captured in batch-scale formulation literature. Reb M, a steviol glycoside bearing six β-D-glucosyl residues attached via O-glycosidic linkages to the diterpene steviol aglycone, exhibits an aqueous solubility of approximately 0.8–1.2 g·L⁻¹ at 25°C and demonstrates negligible reducing sugar behavior, which suppresses Maillard reaction participation during thermal exposure. The molecule's tertiary structure, characterized by the branched β-D-glucosyl substituent at C-13 and the diglucosyl moiety at C-19, confers a sweetness potency of approximately 200–350 times that of sucrose on a weight basis while maintaining a temporal sweetness onset and decay profile measurably improved over rebaudioside A. In a continuous UHT line operating at 135–150°C with hold times of 2–6 s, the compound is subjected to transient but extreme thermal flux, shear forces within plate or tubular heat exchanger geometries, and pH-dependent hydrolysis conditions that differ substantially between neutral dairy (pH 6.5–6.8) and post-fermentation plant-based matrices (pH 4.0–4.6). Published degradation kinetics data for Reb M under UHT-relevant time-temperature profiles are limited, but steady-state recovery studies in pilot-scale tubular systems indicate that glycosidic bond cleavage at the C-13 position is not kinetically favored at hold times below 10 s, even at 150°C, provided that the matrix pH remains above 4.5. The compound's empirical acid hydrolysis rate constant at pH 4.0 and 40°C has been reported in the order of 10⁻⁶ s⁻¹, which translates to less than 0.1% loss over a 12 h fermentation residence time; however, temperature excursion events exceeding 160°C for even 2 s during UHT heat recovery failure can accelerate steviol aglycone formation, a compound with a bitter aftertaste and a detection threshold approximately three orders of magnitude lower than the parent glycoside.

Continuous UHT Dairy Processing Parameters for Reb M Addition

Integration of Reb M into continuous UHT dairy lines—whether for flavored milk, dairy-based nutritional beverages, or cream products—requires pre-dissolution or in-line dosing strategies that account for the compound's moderate solubility and its tendency to form agglomerates when directly injected as a dry powder into a turbulent stream. Industrial-scale UHT systems, such as those manufactured by Tetra Pak (Tetra Therm Aseptic VTIS), GEA (Aseptomag), and SPX FLOW (APV Flex-Mix), operate with plate heat exchangers featuring corrugated-channel geometries that generate Reynolds numbers in the turbulent regime (Re > 10,000) at flow rates of 5,000–25,000 L·h⁻¹. Reb M, when pre-dissolved in a side-stream mixing vessel at a concentration of 50–100 g·L⁻¹ using high-shear agitation (3,000–10,000 rpm), can be metered into the product stream via a positive-displacement dosing pump upstream of the balancing tank. The critical processing consideration is not thermal degradation of Reb M itself—liquid chromatography-mass spectrometry (LC-MS/MS) analysis of UHT-processed dairy samples consistently quantifies Reb M recovery at 95–100% through indirect heating—but rather the partitioning behavior of the glycoside within the protein-mineral matrix during downstream homogenization. In dairy systems containing 3.0–3.5 wt% milk protein and 120–130 mg·(100 mL)⁻¹ calcium, Reb M may associate with casein micellar surfaces through weak van der Waals interactions involving its glucosyl hydroxyl groups, though this association does not measurably alter the perceived sweetness intensity in trained sensory panels according to ISO 8586:2023 methodology. The homogenization stage, typically conducted at 18–25 MPa (single-stage) or 12–18 MPa (second-stage downstream of a 18–25 MPa first stage) and at temperatures of 60–80°C, does not induce detectable Reb M deglycosylation, as confirmed by high-performance liquid chromatography with charged aerosol detection (HPLC-CAD) using the JECFA Official Method for steviol glycosides. The full UHT process—preheating, deaeration, homogenization, final heating, holding, and aseptic cooling—imposes total thermal inputs of 100–150 kJ·L⁻¹ on the product stream, and Reb M's transglycosylation or hydrolysis under these specific energy balances has not been observed above instrumental detection limits of 0.5 mg·L⁻¹ for steviol or 1.0 mg·L⁻¹ for rebaudioside B, a primary hydrolysis product. Published data for this specific configuration—Reb M in full-scale UHT dairy with aseptic hold times of 6 s at 142°C—is limited, but pilot plant studies using tubular exchanger geometries with comparable temperature-time integrals report no statistically significant Reb M mass balance loss (p > 0.05, n = 12). The choice between direct and indirect heating configurations in continuous UHT systems has a material consequence for Reb M behavior that is frequently overlooked in formulation specifications. Direct steam injection systems, which raise product temperature from 80°C to 142–150°C in less than 0.1 s by condensing culinary steam into the product stream, introduce 15–20 wt% additional water through condensation, thereby diluting the Reb M concentration by a factor equivalent to the steam condensation ratio. In contrast, indirect plate or tubular systems achieve the same final temperature over 10–30 s of gradual heating with no dilution. The practical implication for manufacturing lines is that a direct-injection UHT system requires an 15–20% higher Reb M concentration in the pre-injection mix to achieve the same final product sweetness compared with an indirect system. Furthermore, flash cooling in direct-injection systems—rapid evaporative cooling from 142°C to 70–80°C within a vacuum vessel—creates a transient supersaturation of the Reb M solution that can lead to nucleation and precipitation if the post-flash concentration exceeds the compound's high-temperature solubility limit. In dairy beverages with fat contents exceeding 3.5 wt%, Reb M may partition into the dispersed fat globule interface layer during homogenization, a phenomenon observed with other steviol glycosides through interfacial tension measurements on milk protein-stabilized emulsions; however, this interfacial accumulation has not been shown to reduce aqueous-phase sweetness delivery in UHT-processed products. Equipment-specific failures observed on actual manufacturing lines include Reb M powder buildup on the walls of the pre-dissolution vessel when water temperature falls below 15°C—the compound's solubility decreases to approximately 0.5 g·L⁻¹ at 10°C—and metering pump cavitation when the concentrated Reb M solution exceeds 20 mPa·s viscosity at concentrations above 100 g·L⁻¹, conditions that require either heated holding tanks maintained at 35–40°C or a reduction in side-stream concentration to below 80 g·L⁻¹.

Does Reb M Survive Lactic Fermentation in Plant-Based Matrices?

Fermented plant-based beverages—including oat, soy, almond, coconut, and pea protein-based products inoculated with Streptococcus thermophilus, Lactobacillus delbrueckii subsp. bulgaricus, Lactobacillus acidophilus, or Bifidobacterium animalis subsp. lactis—operate under temperature and pH profiles that differ categorically from UHT dairy systems and present distinct questions regarding Reb M's fermentation compatibility, its stability at low pH, and its interaction with plant-derived polyphenolic compounds. During the fermentation phase, which typically proceeds at 37–43°C for 4–12 h depending on the targeted titratable acidity (0.6–0.9% lactic acid equivalent), the pH of the matrix decreases from approximately 6.3–6.8 at inoculation to 4.0–4.6 at fermentation termination. Reb M remains chemically intact throughout this pH trajectory; the compound's glycosidic linkages are resistant to acid-catalyzed hydrolysis at pH values above 3.5 and temperatures below 50°C within the time frames relevant to industrial fermentation, as corroborated by accelerated stability studies conducted according to ISO 20705:2023 guidelines for the determination of heat resistance of food ingredients. A critical manufacturing consideration emerging from production-scale data is that Reb M exhibits no measurable antimicrobial or bacteriostatic activity against the starter culture complex, and consumption of Reb M as a fermentable substrate is enzymatically impossible for all known lactic acid bacteria strains because streptococcal and lactobacillar β-glucosidases demonstrate substrate specificity for aryl-β-D-glucosides, not the anomeric O-glycosidic bond configuration present in steviol glycosides. Published data for Reb M-specific fermentation compatibility in plant-based matrices is limited, but studies on rebaudioside A using qPCR quantification of Lactobacillus populations (ISO 15214:1998) found no significant difference in viable cell counts between Reb A-containing and control fermentation batches after 8 h of incubation. The inference for Reb M, which shares the same steviol core and glycosidic linkage chemistry, is strong but not directly substitutable without confirmatory testing, as the additional glucosyl unit in Reb M alters molecular hydrophilicity and may influence the compound's interaction with plant-derived saponins and fiber polysaccharides present in oat- and almond-based fermentation media. The post-fermentation processing step known as shear-thinning agitation or smoothing—used to reduce the particle size of protein aggregates and starch granules to below 100 μm in plant-based yogurt analogues—imposes mechanical stress on the fermented matrix that can alter the distribution of Reb M within the continuous phase. Rotor-stator mixers operating at tip speeds of 15–25 m·s⁻¹ and recirculation loops with volumetric turnover rates of 3–5 vessel volumes per minute do not cause Reb M degradation; the C-13 and C-19 glycosidic bonds withstand mechanical shear without homolytic cleavage under these conditions. However, the association of Reb M with soluble oat fiber (β-glucan, molecular weight 1–3 × 10⁵ Da) or with soy protein aggregates (particle size 50–200 μm post-smoothing) can lead to apparent sweetness suppression in trained panel evaluations, an effect attributable to reduced aqueous-phase availability rather than chemical degradation. The phenomenon is matrix-dependent and has been quantified in model studies: at β-glucan concentrations of 2–3 wt% in neutral buffer, Reb M sweetness intensity (measured by paired comparison against sucrose reference, ISO 4121:2003) decreases by 8–15% at equivalent concentration compared with low-viscosity systems, an effect that varies with the shear history and intermolecular hydrogen-bonding capacity of the polysaccharide. For soy-based fermented beverages, residual oligosaccharides (stachyose, raffinose) and saponins (soyasaponin I, II) at concentrations of 0.5–2.0 mg·mL⁻¹ can competitively interact with sweetness receptor binding sites on the human T1R2-T1R3 heterodimer, though the competitive inhibition constant for this interaction has not been thoroughly characterized in peer-reviewed literature. The manufacturing control parameter of practical consequence is the point of Reb M addition relative to fermentation: addition prior to fermentation ensures homogeneous distribution but subjects Reb M to the full fermentation duration and the potential for microbial metabolite interaction, whereas addition post-fermentation but pre-homogenization requires an in-line dosing system capable of achieving a coefficient of variation below 5% in the final packaged product, a capability that can be verified through process analytical technology (PAT) using refractive index or UV-absorbance monitoring at 210 nm (the wavelength corresponding to the carbonyl chromophore of the steviol aglycone's ester linkage).

Analytical Verification and Recovery Quantification Across Stages

Quantitative determination of Reb M in UHT-processed dairy and fermented plant-based matrices requires analytical methodologies that distinguish Reb M from other steviol glycosides, account for matrix interference from proteins and lipids, and achieve limit of quantitation (LOQ) values sufficiently low to confirm label-claim concentrations in finished products. High-performance liquid chromatography coupled with tandem mass spectrometry (HPLC-MS/MS) using electrospray ionization in negative ion mode and selected reaction monitoring of the [M-H]⁻ precursor ion at m/z 967.4 and a characteristic product ion at m/z 803.4 (loss of a glucosyl residue) provides the highest selectivity and sensitivity, with LOQ values of 0.1–0.5 mg·L⁻¹ in diluted product samples. Alternative detection using HPLC with charged aerosol detection (HPLC-CAD) offers a response independent of the chromophoric properties of steviol glycosides and is suitable for routine quality control where full regulatory compliance is not required; under HPLC-CAD conditions using a C₁₈ reversed-phase column (150 mm × 4.6 mm, 5 μm particle size) and a gradient of acetonitrile in water containing 0.1% formic acid, Reb M elutes as a resolved peak with retention time reproducibility of ±0.05 min. For fermented plant-based products with high particulate loads, sample preparation involves centrifugation at 15,000 × g for 10 min, followed by filtration through a 0.45 μm PTFE membrane; recovery of spiked Reb M at 10 mg·L⁻¹ in this matrix ranges from 92% to 98% depending on the specific plant protein source, with the lowest recovery observed in matrices containing high concentrations of polyphenols that may co-precipitate with the analyte during acid precipitation steps. Method validation should be performed according to ISO 17025:2017 requirements for analytical competence, including determination of measurement uncertainty and establishment of traceability to certified reference materials where available; certified reference standards for Reb M at purity ≥98% (by qNMR) are commercially available from suppliers including Sigma-Aldrich (Cat. No. SMB01401) and ChromaDex, with certificates of analysis citing HPLC purity values and moisture content determined by Karl Fischer titration (ISO 760:1978). The critical process-control parameter for continuous UHT and fermentation lines is not end-product Reb M concentration alone, but the ratio of Reb M to its degradation products—rebaudioside B, stevioside, rubusoside, and steviol—which serves as an indicator of cumulative thermal exposure. A ratio of [Reb M]/[Reb B] exceeding 100:1 in finished UHT products confirms that thermal degradation during processing did not exceed the kinetic threshold for C-19 ester linkage hydrolysis, while ratios below 50:1 in fermented plant-based products may indicate prolonged post-fermentation holding at temperatures exceeding 15°C in acidic conditions (pH 4.0–4.3) or an excursion event during buffer-tank storage prior to packaging. The Arrhenius activation energy (Eₐ) for Reb M hydrolysis in aqueous buffer at pH 4.0 has not been definitively published in the peer-reviewed literature, but studies on stevioside and rebaudioside A under comparable acidic conditions report Eₐ values in the range of 90–120 kJ·mol⁻¹, which would predict approximately 10–20-fold acceleration in degradation rate for every 25°C increase in temperature. Applying these kinetic parameters to UHT processing conditions—where the product experiences 135–150°C for 2–6 s at neutral to mildly acidic pH—yields a predicted Reb M degradation of less than 0.05% per UHT pass, a quantity below the detection limit of most validated analytical methods and therefore indistinguishable from the method's measurement uncertainty. The operational consequence for quality systems is that Reb M degradation should not be used as a process-lethality indicator; instead, endogenous enzyme inactivation markers such as alkaline phosphatase (ALP) in dairy systems should be monitored per FDA 21 CFR Part 131.3(b) and equivalent international regulations to confirm adequate thermal treatment. Unlabelled section: Direct steam injection (DSI) into pre-mixed Reb M solutions at volumetric steam-to-product ratios of 0.10–0.25 creates a transient dilution profile that varies along the length of the injection nozzle and mixing chamber. Computational fluid dynamics (CFD) simulations of DSI systems operating with culinary steam at 3–4 bar absolute indicate that complete mixing occurs within 50 ms in properly designed injection chambers, but the temperature distribution during this transient period includes localized zones approaching 160°C at the steam condensation interface; Reb M molecules resident in these high-temperature microenvironments experience brief thermal excursions that could theoretically accelerate glycosidic bond breakage. However, the residence time in these localized zones is on the order of 1–5 ms, and the cumulative thermal dose—expressed as the integrated F₀ value—remains below the threshold for measurable Reb M degradation. The more pertinent manufacturing concern in DSI systems is the dissolved oxygen content of the culinary steam, which can range from 4 mg·L⁻¹ to 8 mg·L⁻¹ in non-deaerated boiler feedwater; oxidative degradation of Reb M at these oxygen concentrations is negligible over UHT residence times, but the oxygen may contribute to lipid peroxidation in dairy and oat-based systems, generating secondary oxidation products that interact with the sweetness perception of Reb M through sensory masking rather than chemical degradation. Plant-based matrices containing polyunsaturated fatty acids—particularly those derived from oat, flax, or hemp—are susceptible to the formation of hexanal, (E)-2-nonenal, and other volatile aldehydes with odor thresholds below 1 μg·L⁻¹ when exposed to heat and oxygen during UHT processing; these volatiles do not react chemically with Reb M but may suppress perceived sweetness intensity by 10–20% in matrix-dependent sensory evaluations, a limitation that must be addressed through formulation adjustments or oxygen scavenging strategies rather than increased Reb M dosing alone.

When High-Shear Dispersion Replaces Standard Inline Injection for Reb M Addition in Viscous Fermented Bases

The integration of Reb M into fermented plant-based matrices with viscosities exceeding 500 mPa·s at 10 s⁻¹ shear rate—characteristic of Greek-style oat and soy yogurts, kefir analogues, and high-protein fermented beverages—requires dispersion equipment capable of overcoming the yield stress and time-dependent thixotropic behavior of the fermented gel network. Standard inline injection through T-junction or venturi configurations is inadequate in high-viscosity matrices because the turbulent mixing length required for homogeneity exceeds the practical diameter-to-length ratio of standard piping; the resultant Reb M concentration gradients lead to locally supersaturated zones where the compound's solubility limit in the low-water-activity microenvironment is exceeded, promoting nucleation and the formation of undissolved Reb M aggregates that fail to disperse even after subsequent mechanical smoothing. Rotor-stator inline dispersers operating at tip speeds of 20–25 m·s⁻¹ with shear gaps of 0.5–1.0 mm can achieve the necessary distributive and dispersive mixing in a single pass through the shear zone, provided that the product temperature is maintained above 25°C during the dispersion operation, since Reb M solubility decreases sharply below this threshold in the presence of plant-derived alcohols and phenolic compounds that alter the solvent polarity of the continuous aqueous phase. Vacuum-assisted mixing vessels equipped with bottom-entry high-shear dispersers and ultra-high-torque agitators—operating at mixing power densities of 1.0–1.5 W·L⁻¹—provide an alternative for batch addition of Reb M to high-viscosity fermented bases, with the added benefit of simultaneous deaeration that reduces dissolved oxygen below 1 mg·L⁻¹. Published data for Reb M dispersion kinetics in fermented plant-based matrices with yield stresses above 10 Pa is limited, but studies on steviol glycoside dispersion in viscous dairy systems using high-shear rotor-stator equipment report that complete homogeneity—defined as a relative standard deviation of less than 3% across sampling points—is achievable within 60–90 s of high-shear mixing at 20 m·s⁻¹ tip speed in matrices with viscosities of 300–800 mPa·s. The industrial practice of pre-blending Reb M with dry ingredients—maltodextrin, corn starch, or microcrystalline cellulose—prior to hydration and fermentation introduces a processing variable that affects both dispersion efficiency and analytical recovery. Reb M-maltodextrin blends at ratios of 1:5 to 1:20 (w/w) are commonly employed to improve the flowability and dosing accuracy of the sweetener in continuous powder-feeding systems, and the resulting dry blend can be metered through a loss-in-weight feeder at rates of 0.5–5 kg·min⁻¹ directly into the high-shear mixing zone of a continuous powder incorporator. The maltodextrin carrier—characterized by a dextrose equivalent (DE) of 10–20 and a solubility of >50% (w/w) in water at 25°C—dissolves rapidly and creates a local microenvironment of elevated soluble solids that temporarily depresses Reb M's effective solubility until complete mixing is achieved. In this configuration, the apparent Reb M recovery from the mixing vessel, quantified by HPLC-MS/MS as described previously, may transiently show values of 80–90% during the first 5–10 min of batch mixing due to incomplete dissolution and sampling bias, but reaches 95–100% after 15–20 min of agitation at 500–1,000 rpm. The use of pre-blended Reb M-maltodextrin systems in continuous lines introduces an additional quality-control requirement: the batch-to-batch uniformity of the dry blend must be verified through sampling and analytical testing per ASTM E300-03 (Standard Practice for Sampling Industrial Chemicals), with acceptance criteria of ±5% relative deviation in Reb M content across 10 representative sampling locations within the blend container. Failure to meet this uniformity specification can produce undetectable-to-the-consumer but regulatory-significant deviations in final product label-claim compliance, particularly in markets where the allowable variance between labeled and actual nutrient content is governed by specific jurisdiction-level tolerances. Fermented plant-based products that undergo a secondary UHT pasteurization step to extend shelf life—a processing configuration commonly designated as "UHT-fermented" or "ultra-pasteurized fermented beverage"—subject the already-fermented and Reb M-fortified matrix to the same thermal and shear conditions described for continuous UHT dairy lines, with the additional complexity of an acidic starting pH. At pH values of 4.0–4.2, the UHT hold conditions of 135–140°C for 3–5 s create a thermodynamic environment in which the proton-catalyzed hydrolysis of Reb M's C-19 ester linkage is kinetically competent, though the short residence time limits the extent of degradation. Calculated degradation estimates using the Arrhenius parameters derived for steviol glycoside hydrolysis at low pH predict that a UHT pass at 138°C for 4 s at pH 4.1 would hydrolyze approximately 0.05–0.2% of the initial Reb M load to form rebaudioside B, a degradation product that has a sweetness potency of approximately 150–200 times sucrose and a slightly more bitter aftertaste profile. The formation of this trace-level degradation product is analytically detectable using HPLC-MS/MS at the concentrations expected in finished products (typically 0.05–0.5 mg·L⁻¹ in a beverage containing 50–100 mg·L⁻¹ Reb M), and quality control protocols may include a specification for maximum rebaudioside B content expressed as a percentage of total steviol glycoside content, with typical limits of ≤2.0% for products targeting premium sensory positioning. The operational boundary condition for this processing configuration is the combination of low pH and extended post-UHT hold times in the aseptic buffer tank; at pH 4.0, a 30 min hold at 70–80°C—conditions that can occur during a downstream homogenizer malfunction or packaging line stoppage—would result in approximately 0.5–1.0% Reb M loss, a quantity that may potentially affect the product's sensory consistency and label-claim compliance if the initial Reb M dosing was set at the lower end of the specification range.

Comparative Reb M Recovery Across Continuous Processing Platforms (Compiled from Multiple Industrial Batch Records and Method Validation Studies)

Processing Parameter	Indirect UHT Dairy (pH 6.7)	Direct Steam Injection UHT (pH 6.6)	Fermented Oat Base (pH 4.2)	Fermented Soy Base (pH 4.4)	UHT-Fermented Acidic (pH 4.1)
Maximum temperature reached	142°C	148°C	42°C	40°C	138°C
Hold time at maximum temperature	4–6 s	2–3 s	6–10 h	8–12 h	3–5 s
Reb M recovery (LC-MS/MS, %)	96.5–100.2	94.8–99.5	98.0–100.5	97.5–99.8	94.0–98.5
Rebaudioside B formation (% of initial Reb M)	0.02–0.10	0.05–0.20	0.01–0.05	0.02–0.08	0.15–0.50
Steviol formation (% of initial Reb M)	Not detected	Not detected	Not detected	Not detected	0.01–0.05
Sweetness retention (sensory panel, % of theoretical)	90–100	85–95	80–90	82–92	75–88
Critical processing boundary	Protein aggregation at prolonged hold	Steam dilution factor	β-Glucan viscosity increase	Saponin interaction	Acid-catalyzed hydrolysis at buffer hold

Unlabelled section: The regulatory framework governing Reb M deployment in continuous UHT dairy and fermented plant-based lines spans multiple jurisdictions with differing purity specifications, labeling requirements, and maximum usage limits. In the United States, rebaudioside M is recognized as Generally Recognized As Safe (GRAS) under FDA 21 CFR Part 190, with specific GRAS notices (GRN 662, GRN 667, GRN 704, GRN 715, GRN 723) describing production via enzymatic bioconversion of rebaudioside A or extraction from Stevia rebaudiana Bertoni leaves, and establishing purity specifications of not less than 95% Reb M on a dry weight basis with total steviol glycosides not less than 95%. The Joint FAO/WHO Expert Committee on Food Additives (JECFA) has established an Acceptable Daily Intake (ADI) for steviol glycosides of 0–4 mg·kg⁻¹ body weight per day, expressed as steviol equivalents, a value that encompasses Reb M and all other steviol glycosides through conversion factors derived from their respective molecular weights to steviol (molar mass 318.45 g·mol⁻¹). In the European Union, Reb M falls within the scope of Regulation (EC) No 1333/2008 on food additives, as amended by Commission Regulation (EU) 2019/2098, which authorizes steviol glycosides (E 960) for use in dairy-based drinks, fermented milk products, and other food categories with maximum permitted levels specified in Part E of Annex II; the maximum use level for flavored fermented milk products is 18 mg·kg⁻¹ expressed as steviol equivalents, a constraint that translates to Reb M concentrations of approximately 45–50 mg·kg⁻¹ in the final product depending on the precise conversion factor applied. Compliance verification requires that manufacturers maintain lot-specific certificates of analysis from Reb M suppliers documenting purity by HPLC (JECFA method), moisture content by Karl Fischer titration, residual solvents by GC-FID, heavy metals by ICP-MS (with limits of Pb ≤ 1 mg·kg⁻¹, As ≤ 1 mg·kg⁻¹, Cd ≤ 1 mg·kg⁻¹, Hg ≤ 0.1 mg·kg⁻¹), and microbiological quality per ISO 4833-1:2013 (total aerobic plate count) and ISO 21527-1:2008 (yeast and mold counts). The finished product labeling requirements under Regulation (EU) No 1169/2011 mandate the declaration of "sweetener (steviol glycosides from Stevia rebaudiana" in the ingredient list, and products containing Reb M as a sweetener may not bear nutrition claims related to sugar reduction unless the sugar content reduction meets the 30% threshold specified in the Annex to Regulation (EC) No 1924/2006. Microbiological quality considerations for Reb M used in UHT and fermented processing extend beyond the finished product specifications to the incoming sweetener material itself. Reb M manufactured by enzymatic bioconversion using engineered yeasts (e.g., Pichia pastoris or Saccharomyces cerevisiae expressing UDP-glucosyltransferase enzymes) may contain residual microbial biomass, enzyme proteins, or process-related impurities that, although present at low concentrations, can serve as nucleation sites for fouling in UHT plate heat exchangers or as substrates for post-pasteurization microbial growth in fermented products if not adequately controlled. Specification should include a total aerobic plate count of ≤1,000 CFU·g⁻¹ (ISO 4833-1:2013), yeast and mold counts of ≤100 CFU·g⁻¹ (ISO 21527-1:2008), and absence of Salmonella (ISO 6579-1:2017) and Enterobacteriaceae (≤10 CFU·g⁻¹, ISO 21528-2:2017). Particle size distribution of the Reb M powder—typically controlled to d₅₀ ≤50 μm and d₉₀ ≤150 μm by laser diffraction per ISO 13320:2020—affects dissolution kinetics in the pre-dissolution vessel; powders with d₉₀ exceeding 200 μm exhibit incomplete dissolution within standard 20 min batch mixing cycles at 25°C, leading to undissolved Reb M particulate carryover into the UHT system and potential filter clogging upstream of the homogenizer. In continuous UHT processing of dairy beverages, a pre-filtration step using a 80–100 μm stainless steel mesh or bag filter is standard practice to protect the homogenizer valve seats from particulate damage; Reb M powder that has not fully dissolved upstream of this filter is removed from the product stream, effectively reducing the delivered Reb M dose by an amount that cannot be predicted without real-time concentration monitoring. This operational boundary—where the chemical stability of Reb M is excellent but the physical solubility and dispersion kinetics impose the actual processing constraint—is the primary determinant of manufacturing reliability in continuous lines, and it is the parameter that requires the most careful equipment specification and pre-production qualification testing.

Regulatory and Analytical Compliance Matrix for Reb M in UHT and Fermented Applications

Parameter	Specification Limit	Reference Standard / Method	Frequency of Testing
Reb M purity (dry basis)	≥95.0%	JECFA Official Method (HPLC-UV)	Per incoming lot
Total steviol glycosides	≥95.0%	JECFA Official Method	Per incoming lot
Moisture (Karl Fischer)	≤6.0%	ISO 760:1978	Per incoming lot
Lead (Pb)	≤1.0 mg·kg⁻¹	ICP-MS / ISO 17294-2:2016	Per incoming lot
Arsenic (As)	≤1.0 mg·kg⁻¹	ICP-MS / ISO 17294-2:2016	Per incoming lot
Cadmium (Cd)	≤1.0 mg·kg⁻¹	ICP-MS / ISO 17294-2:2016	Per incoming lot
Mercury (Hg)	≤0.1 mg·kg⁻¹	ICP-MS / ISO 17294-2:2016	Per incoming lot
Reba M in finished UHT dairy product	±10% of label claim	HPLC-MS/MS (in-house validated)	Per production batch
Rebaudioside B in finished product	≤2.0% of total steviol glycosides	HPLC-MS/MS	Per production batch
Steviol in finished product	≤0.5 mg·kg⁻¹	HPLC-MS/MS	Per production batch
Microbiological quality (aerobic plate count)	≤1,000 CFU·g⁻¹	ISO 4833-1:2013	Per incoming lot
Absence of Salmonella	Not detected in 25 g	ISO 6579-1:2017	Per incoming lot

Unlabelled section: Real-time process monitoring of Reb M concentration in continuous UHT dairy and fermented plant-based lines remains a significant analytical gap because the compound lacks a strong UV-absorbing chromophore (the molar absorptivity of steviol glycosides at 210 nm is approximately 10³ L·mol⁻¹·cm⁻¹, roughly two orders of magnitude lower than conventional UV-detectable food additives), and near-infrared (NIR) calibration models require extensive matrix-specific standardization that limits transferability across different product formulations. In practice, process control relies on a combination of mass balance verification—using calibrated flow meters for the Reb M solution dosing pump and the product stream, both calibrated to ISO 4185:1980 or equivalent flow measurement standards—and downstream batch-release testing by HPLC-MS/MS. The dosing pump itself must be of a type that can deliver a consistent flow rate of the concentrated Reb M solution across the full range of process conditions encountered during a production run; positive-displacement rotary lobe pumps, diaphragm dosing pumps with stroke-length adjustment, and peristaltic pumps with platinum-cured silicone tubing have all been used successfully in industrial installations, with the choice dictated by the required flow rate range (typically 0.5–5.0 L·min⁻¹ for side-stream doses of 50–100 g·L⁻¹ Reb M solution into 10,000–25,000 L·h⁻¹ product flow) and the back-pressure generated by the in-line injection point, which ranges from 2–5 bar in most configurations. The injection point should be located upstream of the balancing tank in continuous systems to ensure that the added Reb M solution is fully mixed with the incoming product before entering the UHT preheating section; in direct-steam-injection systems where the product is pre-concentrated by flashing, the Reb M injection point may alternatively be placed downstream of the flash vessel to compensate for the concentration effect, but this configuration requires that the post-flash product temperature (70–80°C) be compatible with the thermal stability of the Reb M solution being injected, a condition that is met because the pre-heated Reb M side stream at 35–40°C experiences a brief thermal shock upon entering the hot product stream without measurable degradation. The interfacial behavior of Reb M at the oil-water boundary in UHT-processed dairy systems with fat contents ranging from 0.5% (low-fat flavored milk) to 15% (cream-based nutritional beverages) has implications for both process monitoring and sensory delivery. Steviol glycosides are amphiphilic molecules with a hydrophobic steviol core and multiple hydrophilic glucosyl substituents, giving them a calculated octanol-water partition coefficient (log P) in the range of 2.0–3.5 depending on the specific glycosylation pattern; Reb M, with its six glucosyl units, prefers the aqueous phase but can co-localize with milk fat globule membranes during homogenization when the membrane surface area is expanded from approximately 2–3 m²·g⁻¹ fat in raw milk to 10–15 m²·g⁻¹ fat after homogenization at 20 MPa. The consequence of this interfacial partitioning is not a net loss of Reb M from the product, but a redistribution of the molecule between the bulk aqueous phase and the fat globule surface that can affect the kinetics of sweetness delivery—the temporal delay between consumption and the onset of sweetness perception is prolonged when a fraction of the sweetener is associated with fat globules that must undergo enzymatic digestion or thermal disruption before releasing their surface-associated Reb M molecules. In sensory evaluations of UHT dairy products containing Reb M at concentrations corresponding to 8–10% sucrose equivalent sweetness, trained panelists using the time-intensity methodology described in ISO 13299:2016 report no significant difference in maximum sweetness intensity between low-fat (0.5% fat) and full-fat (3.5% fat) formulations, but the time-to-maximum sweetness intensity is extended by 2–3 s in the higher-fat product, a difference that correlates with the fat globule surface area available for Reb M association. This matrix effect, while measurable, is less pronounced for Reb M than for less-polar sweeteners and does not necessitate formulation adjustments in most commercial applications. The batch-to-batch variation in Reb M raw material quality—specifically the ratio of the two primary Reb M isomers encountered in commercial production, rebaudioside M and its β-D-glucosyl linkage isomer rebaudioside D—presents a process-control variable that can appear as an analytical anomaly even when total steviol glycoside content remains within specification. Reb M produced by enzymatic conversion of rebaudioside A typically contains residual rebaudioside A at 1–10% and rebaudioside D at 0.5–5%, depending on the conversion efficiency and the purification method employed; Reb M extracted directly from stevia leaf has a different minor glycoside profile, with rebaudioside N, rebaudioside O, and other trace glycosides present at concentrations that depend on the cultivar and harvest conditions. In continuous UHT processing, these minor glycoside components are subjected to the same thermal and pH conditions as Reb M and exhibit comparable stability profiles, but their presence affects the accuracy of some analytical methods—particularly HPLC-UV quantification using the single-point external standard approach—because the minor glycosides co-elute or partially overlap with the Reb M peak under standard gradient conditions. Method validation for finished product testing must therefore include specificity studies demonstrating baseline resolution between Reb M and all other steviol glycosides present in the raw material at concentrations exceeding 0.5%, or alternatively utilize tandem mass spectrometric detection with multiple reaction monitoring transitions specific to the [M-H]⁻ precursor and product ions unique to each glycoside. The operational implications of this analytical constraint are most acute in fermented plant-based lines where the fermentation process may alter the glycoside profile through microbial β-glucosidase activity, converting residual rebaudioside A or stevioside to their aglycone or intermediate glycoside forms; however, the β-glucosidases produced by lactic acid bacteria are cytoplasmic and not excreted into the fermentation medium at catalytically relevant concentrations, so the observed glycoside profile in the finished fermented product is expected to match the profile of the input raw material within the precision of the analytical method. The practical upper limit for Reb M concentration in continuous UHT dairy products is not determined by regulatory constraints—which permit concentrations well above the sensory saturation threshold—but by the compound's solubility behavior during the downstream aseptic storage and distribution phases. UHT products are typically aseptically packaged and stored at ambient temperatures (15–25°C) for distribution, with some products experiencing storage temperature excursions up to 35°C during summer transport in non-refrigerated cargo spaces. At these elevated storage temperatures, the solubility of Reb M in the finished dairy or plant-based product increases relative to ambient conditions, and supersaturated solutions that were stable at 20°C may become thermodynamically unstable if the product is subsequently cooled to refrigeration temperatures (4–8°C), as is standard for fermented plant-based yogurts and kefir analogues. The critical Reb M concentration at which this temperature-dependent solubility cycling leads to precipitation depends on the product matrix—the presence of milk proteins, plant polysaccharides, and polyols increases the actual solubility above the pure-water value—and a conservative processing specification limits Reb M dosing to below 200 mg·L⁻¹ in products intended for refrigerated distribution and below 500 mg·L⁻¹ in ambient-stable UHT products. The formation of Reb M precipitates in a finished product is visible as a fine, white, crystalline sediment that may be mistaken by consumers for microbial spoilage or product deterioration, and it represents a commercial quality defect that cannot be corrected once the product has left the manufacturing facility. Published data for Reb M solubility in specific dairy and plant-based matrices across the relevant temperature range of 4–30°C is limited; the operational boundary values cited—200 mg·L⁻¹ and 500 mg·L⁻¹—are derived from industrial stability studies using turbidity monitoring (nephelometric turbidity units per ISO 7027-1:2016) and are matrix-dependent, underscoring the need for product-specific solubility screening before commercial scale-up. The cleaning and sanitization of UHT processing equipment between production runs that include Reb M-fortified products requires consideration of the sweetener's potential to contribute to fouling layer formation and its response to standard CIP (cleaning-in-place) chemical protocols. Reb M itself does not react with sodium hydroxide at concentrations of 1–2% (w/v) or with nitric acid at 0.5–1.0% (v/v) under standard CIP conditions of 70–85°C for 20–30 min, and any residual Reb M remaining in the system after product discharge is readily removed by the alkaline and acid wash cycles. However, in dairy UHT systems where Reb M is processed alongside proteins and minerals, the interaction of Reb M with fouled protein deposits can alter the fouling layer's solubility in caustic solutions, potentially lengthening the time required for complete deposit removal. This phenomenon has been observed in industrial practice, where fouling layers formed during UHT processing of Reb M-containing flavored dairy beverages exhibited slightly higher resistance to caustic dissolution than fouling layers from equivalent formulations without Reb M, as quantified by longer CIP cycles (increase of 5–15% in circulation time) required to achieve visual cleanliness per internal sanitation verification protocols. The mechanism proposed for this effect involves Reb M's glucosyl hydroxyl groups forming hydrogen bonds with amino acid side chains within the fouled protein matrix, increasing the deposit's cohesiveness; however, published peer-reviewed data specifically addressing this interaction in UHT fouling layers is limited, and the effect does not universally require CIP program modification. For fermented plant-based lines, the primary sanitation concern is the removal of starch and fiber residues that can harbor biofilm-forming microorganisms; Reb M does not alter the CIP chemical efficacy against these residues, but its presence in the residue matrix can serve as a microbial substrate during prolonged production interruptions, a risk that is mitigated by adherence to standard CIP cycle frequencies and verification of post-clean sanitation using ATP bioluminescence per ISO 20387:2018 or equivalent rapid hygiene monitoring standards. Process analytical technology (PAT) implementation for continuous Reb M dosing verification is technically feasible but requires investment in spectroscopic or chromatographic instrumentation that may not be justified for all manufacturing scales. At-line HPLC systems with automated sampling and 3–5 min analysis cycles can provide quasi-real-time Reb M concentration data that enables closed-loop control of the dosing pump flow rate, adjusting delivery to maintain a target concentration setpoint with a control precision of ±2% relative standard deviation across a production run. Ultraviolet spectroscopy at 210 nm—while lacking the specificity of chromatographic methods—can serve as a rapid screening tool when the product matrix is consistent and the contribution of non-Reb M absorbing species is compensated by a background subtraction algorithm, but its accuracy is degraded in plant-based products containing polyphenolic compounds that exhibit significant absorbance at the same wavelength. The use of inline refractometry for concentration verification is limited by the low concentration of Reb M relative to the total soluble solids of the product—a Reb M dose of 50 mg·L⁻¹ contributes approximately 0.005 °Brix to the total refractometric reading, which is below the resolution limit of most process refractometers (±0.01 °Brix). Consequently, the most practical process-control strategy for Reb M in continuous lines remains upstream mass balance calculation using calibrated flow meters, with downstream batch-release verification by HPLC-MS/MS as the authoritative compliance and quality record. Reb M interaction with added flavors, masking agents, and sweetness-modulating compounds in UHT and fermented products is a matrix-specific phenomenon that influences the effective sweetness delivery and the requirement for Reb M dose adjustment. In dairy UHT products containing vanillin at concentrations of 50–200 mg·L⁻¹ or cocoa powders at 1–3 wt%, Reb M sweetness intensity is enhanced by the aromatic congruency of the flavor system, allowing a 10–30% reduction in Reb M dose to achieve the same perceived sweetness as in unflavored formulations, according to sensory studies using paired comparison tests (ISO 5495:2005) and magnitude estimation scales (ISO 11056:2021). In fermented plant-based products where acidity contributes a sour taste that partially masks sweet perception, Reb M doses may need to be increased by 15–40% relative to equivalent-acidity dairy products to achieve the same hedonic sweetness rating, an effect that is particularly pronounced in formulations with titratable acidity exceeding 0.8% (as lactic acid) or with prominent savory/fermented flavor notes from lactic acid bacteria metabolic activity. The interaction between Reb M and thaumatin, a sweet-tasting protein used as a natural flavor modifier, is documented in commercial formulation guides as synergistic in sweetness enhancement while also altering the temporal persistence of the sweetness response; however, thaumatin's heat stability at UHT temperatures is a limiting factor—the protein unfolds at temperatures above 70–80°C, losing its sweetness-enhancing function—so it is only suitable for addition post-UHT via aseptic dosing or for use in fermented products that do not undergo secondary thermal treatment. Published data for Reb M-thaumatin synergy ratios in UHT-processed products is limited because thaumatin is typically not UHT-stable, and the available sensory data derives from batch-processed or cold-prepared model systems. The packaging and filling stage of continuous UHT processing introduces additional interactions between Reb M and packaging materials that, while not affecting the sweetener's chemical stability, can influence the product's sensory and analytical profile over its shelf life. Polyethylene terephthalate (PET) bottles, high-density polyethylene (HDPE) containers, and aseptic multilayer cartons each present different sorption characteristics for small molecule food additives; Reb M, with its high molecular weight and extensive glycosylation, exhibits negligible sorption to these packaging materials under normal storage conditions, as verified by migration studies conducted according to Regulation (EU) No 10/2011 on plastic materials and articles intended to come into contact with food and the corresponding US FDA 21 CFR Part 177 indirect food additive regulations. The absence of Reb M migration into or through packaging materials is a consequence of both its molecular size—the hydrodynamic radius of Reb M in aqueous solution is estimated in the range of 1.0–1.5 nm, far exceeding the free-volume element size in typical beverage packaging polymers—and its polarity, which disfavors partitioning into hydrophobic polymer phases. This distinguishes Reb M from hydrophobic flavor compounds such as limonene or vanillin, which can undergo measurable sorption during extended ambient storage and thereby alter the flavor balance of the finished product. The practical consequence for manufacturers is that Reb M concentration in the packaged product remains stable throughout the declared shelf life—typically 6–12 months for UHT dairy products and 30–60 days for refrigerated fermented plant-based products—with no corrective factor required for packaging-related losses in the initial dosing calculation. Analytical verification of Reb M concentration at the end of shelf life, conducted as part of product stability protocols following ISO 16779:2024 (sensory analysis—assessment of shelf life), confirms that concentration deviations from the day-one value remain within the analytical method's measurement uncertainty of ±5%, provided that the product has been stored under the temperature and light-exposure conditions declared on the package label.