Reb M Heat & Acid Stability — Shelf-Life Data by pH, Temperature and Time
Reb M retains 94 % of its sweetness after two hours at 120 °C in pure water.[2] Dissolve it in a cola formula at 80 °C — forty degrees cooler — and only 4.46 % survives 48 hours.[2] That is not an inconsistency. It is the single most important fact about steviol glycoside stability: the formulation determines shelf life far more than the processing temperature.
Every data sheet that calls Reb M “heat-stable” without specifying pH, acid type, co-solutes and storage duration is telling you nothing actionable. Until October 2025, there was no published long-term stability dataset for Reb M specifically. Formulators extrapolated from Reb A and stevioside studies — or trusted supplier claims backed by internal data you could not verify. That changed with the first peer-reviewed 26-week Reb M stability study, which mapped retention across four pH levels and four temperatures in a mock citric-acid beverage.[1]
One disclosure. We supply rebaudioside M, rebaudioside A and rebaudioside D — the steviol glycosides discussed in this article. Every number below comes from published studies, JECFA evaluations, FDA GRAS notices or peer-reviewed kinetic analyses. Read with that supply-side context in mind.
This guide covers a quick-reference table for nine common processing scenarios, the full 26-week Reb M stability dataset with calculated rate constants and shelf-life estimates, why Reb M’s degradation products are all sweet and what that means for your label, processing survival rates from hot-fill through UHT, a head-to-head stability comparison with Reb A and stevioside, the formulation variables that accelerate or slow degradation — including two that most guides miss — and a practical specification checklist for procurement.
01 · The one-minute answer
Will Reb M Survive My Process?
If you are scanning this article for a pass/fail answer on your specific thermal process, start here. The evidence behind each row is developed in Sections 02–06.
| Scenario | Conditions | Reb M result | What to do |
|---|---|---|---|
| Hot-fill, citric acid | 85 °C / 2 min, pH 3.0 | Stable — matches Reb A pattern | Standard hot-fill is safe[13] |
| HTST pasteurisation | 95 °C / 15 s | 97–98 % | No dosage adjustment needed[7] |
| UHT, dairy milk | 140 °C / 5 s | ~80 % | Over-dose by 20–25 %[7] |
| UHT, soy drink | 140 °C / 5 s | ~101 % (no loss) | No adjustment needed[7] |
| Yogurt pasteurisation | 88 °C / 5 min → ferment | No measurable loss through 6 weeks at 4 °C | Safe[3] |
| Pure-water thermal | 120 °C / 2 h, pH 3–8 | 94 % at 120 °C; > 96 % at 80 °C | Heat alone does not destroy Reb M[2] |
| Room-temp storage, 1 year | 24 °C, pH 3.8+, citric acid | Essentially zero decomposition | Safe[6] |
| Refrigerated, 26 weeks | 5 °C, pH 2.8–4.2 | 99.4–100.9 % | Cold chain eliminates concern[1] |
| Cola matrix, accelerated | 80 °C / 48 h, pH ~2.5 | 4.46 % | Matrix, not heat, is the threat[2] |
The cola number in context
80 °C for 48 hours is not a realistic production scenario — it is an accelerated stress test used to rank matrices by aggressiveness. No bottler runs cola through 48 hours at 80 °C. The 4.46 % figure tells you that the combination of phosphoric acid, caramel colour (E150d) and low pH creates one of the most hostile environments for any steviol glycoside. At 26 weeks ambient storage, Reb M in a citric-acid beverage at pH 3.8+ does not measurably degrade.
02 · The 26-week dataset
The First Published Long-Term Reb M Stability Study
Before 2025, every Reb M shelf-life claim in the industry traced back to one of two sources: Prakash et al.’s 2014 application study, which tested Reb M in cola, yogurt and chewing gum but reported stability in qualitative terms (“similar to Reb A”),[3] or Cargill’s GRAS Notice 882, which tested hot-fill at pH 3.0 but released only summary-level data.[13] Neither gave formulators a pH × temperature kinetic matrix they could use for shelf-life modelling.
In October 2025, the first complete dataset appeared: a 26-week accelerated stability study of Reb M in a mock carbonated beverage (citric acid, potassium sorbate, sodium benzoate) at four pH levels (2.8, 3.2, 3.8, 4.2) and four temperatures (5, 20, 30, 40 °C), starting from ~463 mg/L.[1] The retention values at the extremes — the two data points most useful for shelf-life estimation:
| pH | 5 °C, 26 wk | 40 °C, 26 wk | k40 (day−1) | t10 % at 40 °C |
|---|---|---|---|---|
| 2.8 | 99.4 % | 34.1 % | 0.0059 | 18 days |
| 3.2 | — | 57.1 % | 0.0031 | 34 days |
| 3.8 | 99.6 % | 83.4 % | 0.0010 | 106 days (~15 wk) |
| 4.2 | 100.9 % | 92.2 % | 0.00045 | 236 days (~34 wk) |
The rate constants (k) assume pseudo-first-order kinetics — an assumption validated by the mass balance of 97.2–100.4 % across all conditions, meaning degradation products are fully accounted for and no material disappears to unknown pathways.[1] The t10 % column is the time to reach 90 % retention, calculated as ln(0.9)/−k. That is the number you hand to your shelf-life team.
What this table actually says
Dropping from pH 4.2 to pH 2.8 at 40 °C multiplies the degradation rate by 13×. But shifting from 40 °C to 5 °C at pH 2.8 multiplies it by 179× in the other direction. Temperature is a bigger lever than pH — but pH is the one you set in your formula, and temperature is the one the supply chain decides. For any product stored at ambient or refrigerated conditions, pH 3.8 or above gives you a t10 % that comfortably exceeds 12 months.
Photostability
The same study tested Reb M under light exposure at pH 3.8 / 25 °C and found minimal additional degradation.[1] This is consistent with Clos et al. (2008), who showed that steviol glycoside degradation in carbonated beverages is acid-promoted, not light-promoted — light- and dark-stored samples produced statistically identical degradation products.[9] One exception: in citrate buffer at pH 3, light exposure accelerated the Reb A degradation rate approximately 8×.[5] If your product uses citric acid and has UV-transparent packaging, the light × acid interaction can matter. Section 06 covers this.
03 · What Reb M degrades into
Every Degradation Product Is Sweet
When sucralose degrades under heat, it can form chlorinated furans. When aspartame degrades, it loses sweetness entirely and produces diketopiperazine. Steviol glycoside degradation is fundamentally different: the products are other steviol glycosides, and they are all sweet.
Prakash and colleagues identified the three principal Reb M degradation products at pH 2.0 / 80 °C:[14]
| Product | Mechanism | Glucose units | Sweet? |
|---|---|---|---|
| C-15 isomer (compound 12) | Alkene migration from C-16/17 to C-15/16 | 6 (same as Reb M) | Yes — sweeter than Reb M, similar taste[3] |
| 16β-hydroxy (compound 13) | Hydration of the exocyclic double bond | 6 (same as Reb M) | Yes |
| Reb B (compound 2) | C-19 ester hydrolysis, loses the C-19 sugar chain | 3 | Yes |
The C-15 isomer is the dominant product under mild acid conditions. It retains all six glucose units, has an identical molecular weight to Reb M, and — critically — Prakash’s sensory panel rated it as sweeter than Reb M with a similar taste profile.[3] This creates what you might call a graceful failure mode: as Reb M degrades, the primary product is another, comparably sweet molecule. The perceived sweetness of a beverage drops more slowly than the Reb M concentration on paper.
None of these products are novel food-safety concerns. The C-15 isomer, the 16β-hydroxy derivative and Reb B are all known steviol glycoside structures covered under existing JECFA and FDA safety evaluations. No steviol aglycone was detected as a degradation product in any study — not in the IJPSR 26-week data,[1] not in Wölwer-Rieck’s commercial soda tests,[10] and not in the JECFA compilation.[6]
Reb M does not degrade into off-flavours. It degrades into other sweeteners.
04 · Processing survival
Condition-by-Condition: What Survives and What Does Not
The quick-reference table in Section 01 gives the verdict. This section gives the evidence, organised by process type so you can jump to the one that matters for your product.
Hot-fill (85 °C / 2 min)
Cargill’s GRAS Notice 882 tested Reb M in a hot-fill simulation at 185 °F (85 °C) for 2 minutes in citric acid at pH 3.0. The degradation pattern matched Reb A (RA95) under identical conditions.[13] Prakash confirmed stability in lemon-lime and cola formats through 26 weeks of ambient storage after hot-fill processing.[3] No over-dosing is necessary.
HTST pasteurisation (72–95 °C / 15–30 s)
Amery and colleagues at KU Leuven tested steviol glycosides in a fermented milk drink at 95 °C / 15 s: Reb A retained ~98 %, stevioside ~97 %.[7] Reb M-specific HTST data are not published, but given that Reb M and Reb A share the same C-19 ester bond that drives thermal degradation and Prakash describes their stability as “similar,”[3] the same range applies. We routinely advise customers that HTST pasteurisation does not require dosage compensation.
UHT (140 °C / 5 s) — matrix-dependent
This is where the matrix starts to dominate. In the KU Leuven data:[7]
- Dairy milk (140 °C / 5 s): Reb A ~80 %, stevioside ~91 %. A 10–20 % loss that requires dosage compensation.
- Soy drink (140 °C / 5 s): Reb A ~101 %, stevioside ~101 %. No measurable loss at all.
- Ice cream (85 °C / 15 s): Reb A ~108 %. The over-recovery is an analytical artefact, but the message is clear: no degradation.
The 21-point gap between milk and soy at identical UHT conditions is startling. The most likely explanation is that milk’s casein and whey protein participate in Maillard reactions at 140 °C, generating reactive intermediates that accelerate steviol glycoside hydrolysis. Soy protein behaves differently at UHT — isoflavones and soy melanoidins may provide antioxidant protection similar to what Gelinas observed with coffee melanoidins.[4] Jooken et al. tested across additional food matrices (yogurt, dry biscuits, jam, fermented milk drink) and found “no sign of steviol glycoside decomposition in any sample.”[8]
At UHT, test in your matrix. A result from milk does not predict a result in soy — or vice versa.
Yogurt (pasteurisation + fermentation)
Prakash tested Reb M in yogurt pasteurised at 190 °F (88 °C) for 5 minutes, followed by bacterial fermentation. No significant sweetness loss was detected, and the product was stable through 6 weeks at 40 °F (4 °C).[3] The combination of moderate heat, near-neutral initial pH and rapid cooling makes yogurt one of the most forgiving matrices for steviol glycosides.
Dry powder (25–40 °C storage)
GRAS 882 tested Reb M powder at 25 °C / 60 % RH and 40 °C / 75 % RH for 3 months. The purity moved from 86.8 % to 85.65 % at ambient and to 85.0 % at accelerated conditions — both well within specification.[13] Dry Reb M is effectively stable in powder form under standard warehouse conditions.
Data gap: retort and dry baking
No peer-reviewed study has tested Reb M under retort conditions (121 °C / 15–30 min, pressurised). Kim et al. showed 94 % retention in water at 120 °C / 2 h,[2] which suggests retort-temperature exposure of typical duration (< 30 min) should be survivable — but the pressurised, low-pH conditions of a canned product are not the same as pure water. For dry baking, the only published data are for stevioside: ~91 % loss at 105 °C / 48 h dry heat,[6] an extreme condition unlikely to map to a standard oven cycle. If your application involves retort or prolonged dry heat above 100 °C, request a bench trial rather than relying on extrapolated data.
05 · Glycoside comparison
Reb M vs Reb A vs Stevioside: Stability Is Not the Differentiator
The honest answer to “Is Reb M more heat-stable than Reb A?” is: marginally, under extreme conditions. Not meaningfully in real formulations.
Kim et al. (2026) ran the only published head-to-head acid stability test at identical conditions. At pH 2.0 / 80 °C / 2 hours, Reb M retained 88.24 % vs stevioside at 83.22 % (p < 0.05).[2] The proposed mechanism is steric: Reb M’s bulkier trisaccharide at C-13 partially shields the acid-labile C-19 ester bond. At pH 3.0–8.0, both glycosides retained > 96 % — the steric advantage disappears when the acid challenge is moderate.
| Metric | Reb M | Stevioside | Reb A |
|---|---|---|---|
| pH 2.0, 80 °C, 2 h | 88.24 % | 83.22 %[2] | — |
| pH 3.0–8.0, 80 °C, 2 h | > 96 % | > 96 %[2] | > 96 % (est.) |
| 120 °C, 2 h, pure water | 94.07 %[2] | — | — |
| Mock beverage, pH 2.8 / 40 °C, 26 wk | 34.1 %[1] | 38.8 %[12] | — |
| Mock beverage, pH 3.8 / 40 °C, 26 wk | 83.4 %[1] | 87.7 %[12] | — |
| Commercial soda, 80 °C, 72 h | — | — | Up to 70 % loss[10] |
| Stability described as | “Similar to Reb A”[3] | Baseline | Baseline |
Note the 26-week mock beverage numbers. The Reb M data (IJPSR 2025[1]) and the stevioside data (Prakash 2012[12]) come from the same research group using the same mock beverage protocol. Reb M’s retention at pH 2.8 and 3.8 is slightly lower than stevioside’s — not higher. The differences are small enough to fall within inter-study variability, but they certainly do not support a marketing claim that Reb M is more shelf-stable in finished beverages.
You choose Reb M for its taste profile, not for shelf-life advantage. Our Reb A vs Reb M comparison covers the sensory evidence.
06 · Formulation variables
The Six Factors That Actually Control Degradation
If temperature alone predicted steviol glycoside stability, the Kim data — 94 % at 120 °C in water, 4.46 % at 80 °C in cola — would make no sense. What sits between those two numbers is formulation. These are the variables that move the needle, ranked by the size of their published effects.
1. Acid type: phosphoric acid is the problem
JECFA compiled the most complete acid-type comparison. At 500 mg/L and room temperature over 4 months:[6]
- Phosphoric acid (pH 1.6): 75 % loss.
- Citric acid (pH 2.1): 20 % loss.
- Acetic acid (pH 2.6): greatest stability of any acid tested.
The pH values are not matched, but the ranking holds across concentrations: phosphoric acid is consistently the most aggressive food acid for steviol glycoside hydrolysis. Cola’s use of phosphoric acid as its primary acidulant is one reason the cola matrix is so hostile. If you are formulating a carbonated soft drink and have any latitude on acid choice, citric acid buys you a large stability margin at equivalent pH.
2. Caramel colour: +83–89 % acceleration
Gelinas et al. (2022) quantified a factor that industry formulators had suspected but never measured: caramel colour (E150d, the type used in cola) increases the acid-catalysed degradation rate of Reb A by 83–89 % relative to caramel-free controls at identical pH and temperature.[4] The mechanism likely involves caramel’s reactive carbonyl intermediates participating in general acid catalysis. If your formula includes E150d and you need 12-month steviol glycoside shelf life, the 26-week data at 40 °C in the IJPSR study[1] — which did not include caramel colour — will overestimate your actual stability.
3. Coffee melanoidins: −80–90 % protection
The same Gelinas study found the mirror image in coffee: melanoidins and polyphenols from roasted coffee reduced the formation of the epoxide degradation product by 80–90 %.[4] Our Reb M in coffee guide discusses melanoidins from a sensory perspective (they do not shield steviol glycosides from bitter-receptor activation). From a stability perspective, the protection is real and substantial. Coffee melanoidins act as radical scavengers and general antioxidants, slowing the acid-catalysed pathways that break the C-19 ester bond. If you are formulating an RTD coffee with Reb M, the melanoidin matrix is working for you on shelf life, even as it works against you on taste.
4. Phosphate buffer at neutral pH: the counter-intuitive one
Most formulators assume that neutral pH is safe. It usually is — unless the formula contains phosphate. Gong (2013) measured Reb A degradation kinetics in phosphate and citrate buffers across pH 3, 5 and 7 at 20–40 °C.[5] The results at pH 7 / 40 °C in 0.1 M phosphate buffer: t10 % = 41 days. That is faster than pH 3 / 40 °C in 0.02 M phosphate (t10 % = 182 days). Phosphate acts as a general acid–base catalyst at neutral pH, accelerating hydrolysis through a mechanism that does not require free protons.
The practical implication: protein beverages, meal replacements and functional drinks at pH 6–7 that use phosphate-containing ingredients (sodium phosphate, potassium phosphate, disodium phosphate — common buffering and emulsifying salts) may see faster steviol glycoside degradation than their pH alone would predict. Replace phosphate with citrate where your formula permits it.
5. Light × acid interaction
Steviol glycosides are generally photostable. Clos et al. showed that Reb A and stevioside in carbonated beverages were within 3–7 % of dark-stored controls after 3 000 langleys of sunlight exposure.[9] But Gong found an 8× acceleration of degradation rate under light in citrate buffer at pH 3.[5] The distinction: it is not direct photolysis. Light generates trace radicals in acidic citrate solutions, and those radicals catalyse the same acid-promoted degradation pathways. In phosphoric acid or at higher pH, the effect is minimal.
Bottom line: if your product uses citric acid at pH < 3.5 in a clear bottle, UV-blocking packaging or opaque containers are worth the cost. Above pH 3.5 or in opaque packaging, light is not a significant concern.
6. Carbonation: the pH shift nobody budgets for
Dissolved CO2 forms carbonic acid, lowering the equilibrium pH of a sealed carbonated beverage by approximately 0.2–0.3 units relative to the flat (uncarbonated) formula at the same acid concentration. At pH 3.8, that shift to 3.5–3.6 roughly doubles the degradation rate constant at 40 °C, based on the IJPSR kinetics. When you measure the pH of your formula during development, measure it carbonated and sealed at target volumes, not flat. The flat-formula pH overestimates your stability margin.
Erythritol as a stabiliser
Santana et al. (2025) studied the thermal decomposition of stevioside by TGA/DSC and found that blending with erythritol eliminates the autocatalytic decomposition behaviour observed in pure stevioside above 150 °C.[11] This is relevant for dry applications (tableting, baking premixes) where steviol glycosides are exposed to sustained heat in the absence of water. In aqueous systems, the autocatalytic pathway is not the dominant mechanism and the erythritol effect is less significant — but erythritol co-formulation still helps for other reasons (solubility and taste).
07 · Specification guide
What to Specify When You Buy Reb M for a Heat-Processed Product
The stability data above feed directly into procurement. Here is what to ask for and why.
| Parameter | Recommendation | Rationale |
|---|---|---|
| Purity grade | RM95 for most applications; RM98 for clear, cold, long-shelf-life beverages | RM95 can contain 3–7 % Reb D, which crystallises at low temperature. Our RM95 vs RM98 guide has the thermal-cycle test protocol. |
| Production method | Leaf-extracted, enzymatic or fermentation — equivalent for stability | The IJPSR sensory panel (n = 14, trained, 0–15 scale, tested at 100 and 300 ppm) found no significant difference between the three production types.[1] |
| Storage conditions | Sealed container, ≤ 25 °C, away from direct light | Powder stability is excellent at ambient conditions (86.8 → 85.65 % over 3 months at 25 °C / 60 % RH).[13] No refrigeration required for dry powder. |
| Max use concentration | < 0.5 % (5 000 ppm) in aqueous systems | Kim et al. could not dissolve Reb M above 0.5 % for aqueous storage testing.[2] At beverage dosages (150–500 ppm), solubility is not a constraint. For tabletop or concentrate formats, pre-dissolve in warm water or co-formulate with erythritol. |
| COA request | Ask for HPLC degradation profile, not just purity assay | A COA that reports only “Reb M ≥ 95 %” does not tell you how much Reb D, Reb A, stevioside or isosteviol is present. These impurities have different solubility and stability profiles. Request a full glycoside breakdown. |
08 · Frequently asked questions
FAQ
Is Reb M more heat-stable than Reb A?
At extreme acid conditions (pH 2.0 / 80 °C), yes — Reb M retained 88 % vs stevioside’s 83 %, and the difference was statistically significant.[2] The mechanism is steric shielding from Reb M’s bulkier C-13 sugar chain. But at pH 3.0–4.2 — the range of virtually all beverages — the stability of Reb M, Reb A and stevioside is comparable. Prakash describes Reb M stability as “similar to Reb A.”[3] You are not gaining a meaningful shelf-life advantage by switching from Reb A to Reb M. You are gaining a taste advantage. See Reb A vs Reb M.
Can Reb M survive retort processing?
Probably, but no peer-reviewed study has tested it directly. Kim et al. showed 94 % retention in water at 120 °C for 2 hours,[2] which is a more severe thermal challenge than a standard retort cycle (121 °C / 15–30 min). But the Kim test was in neutral water, not in an acidic canned product. At pH < 3.5 under pressure, the combination of temperature and acid may produce more degradation than the water-only number suggests. Our recommendation: run a bench retort trial in your actual matrix before committing to Reb M for a retort application.
Does carbonation reduce Reb M shelf life?
Indirectly, yes. Dissolved CO2 forms carbonic acid, lowering the equilibrium pH by roughly 0.2–0.3 units in a typical CSD. At pH 3.8, that shift to 3.5–3.6 approximately doubles the degradation rate at 40 °C. When you measure your formula’s pH for shelf-life prediction, measure it carbonated and sealed at target volumes — not from the flat batch. The flat pH will overestimate your stability.
Does the Reb M production method affect stability?
No. The IJPSR study compared leaf-extracted, bio-converted and fermentation-derived Reb M in a trained sensory panel (14 panellists, 0–15 scale, at 100 and 300 ppm). No significant sensory differences were detected between the three types.[1] The degradation chemistry is driven by the C-19 ester bond and the exocyclic double bond on the steviol aglycone, which are identical regardless of how the Reb M molecule was produced. Choose your production method on cost, purity and supply-chain criteria, not stability.
What shelf life can I expect for a Reb M beverage at room temperature?
At pH 3.8 or above in a citric-acid-based beverage, the JECFA compilation reports “essentially no decomposition” after 1 year at 24 °C.[6] The 26-week IJPSR data at 5 °C / pH 3.8 shows 99.6 % retention, and the calculated t10 % at 40 °C / pH 3.8 is 106 days (∼15 weeks).[1] A standard 12-month ambient shelf life at pH 3.5 or above is well within the kinetic envelope. Below pH 3.0 at ambient temperature, run your own accelerated stability study — the rate constants rise steeply.
Does light degrade Reb M?
In most formulations, no. Clos et al. showed that steviol glycosides in carbonated beverages were photostable, with light- and dark-stored samples showing identical degradation products.[9] The one documented exception is in citrate buffer at pH < 3.5, where Gong measured an ~8× acceleration under light.[5] This is not direct photolysis — it is light-generated radicals catalysing the acid-promoted degradation pathway. Practical guidance: above pH 3.5 or in opaque packaging, do not worry about light. Below pH 3.5 in a clear bottle with citric acid, choose UV-blocking packaging.
Why does Reb M degrade faster in cola than in water at a lower temperature?
Three factors stack. First, cola uses phosphoric acid, the most aggressive common food acid for steviol glycoside hydrolysis — JECFA data show 3.75× more loss in phosphoric acid than citric acid at room temperature.[6] Second, cola contains caramel colour (E150d), which Gelinas showed accelerates acid-catalysed degradation by 83–89 %.[4] Third, the test duration (48 hours at 80 °C vs 2 hours at 120 °C) gives the synergistic matrix effects far more time to compound. The 4.46 % retention number[2] is a stress-test result, not a production scenario — but it demonstrates that formulation beats temperature as the predictor of shelf life.
Need Reb M Samples with Stability Data?
Tell us your product format, pH range, thermal process and target shelf life. We’ll recommend the right glycoside grade, provide stability guidance from comparable matrices, and ship samples with full COA.
References
References
Reb M stability data
[1] Prakash I, Bunders C, Engles S, Tao Y, Sandoval J. Stability of rebaudioside M in a citric acid mock beverage at pH 2.8–4.2 and 5–40 °C over 26 weeks. Int J Pharm Sci Res. 2025;16(10). (First published long-term Reb M stability study. Starting concentration ~463 mg/L; mass balance 97.2–100.4 %. Sensory panel: n = 14, no difference between leaf-extracted, bio-converted and fermentation-derived Reb M.)
[2] Kim S-K et al. Synthesis and characterisation of rebaudioside M analogues with improved solubility and acid stability. Foods. 2026;15(x). (Pure thermal: 120 °C / 2 h, Reb M 94.07 %. pH 2.0 / 80 °C / 2 h: Reb M 88.24 % vs stevioside 83.22 %, p < 0.05. Coca-Cola matrix 80 °C / 48 h: Reb M 4.46 %.)
[3] Prakash I, Dubois GE, Clos JF, Wilkens KL, Fosdick LE. Development of rebiana, a natural, non-caloric sweetener. Food Chem Toxicol. 2008;46(suppl 7):S75–82; Prakash I et al. Development of next generation stevia sweetener: Rebaudioside M. Foods. 2014;3(1):162–175. (Yogurt: 190 °F / 5 min pasteurisation, stable 6 weeks at 40 °F. Cola / lemon-lime: 26 weeks. Degradation products all sweet; C-15 isomer sweeter than Reb M. Stability described as “similar to Reb A.”)
[13] FDA GRAS Notice 882 (Cargill). Rebaudioside M from stevia leaf extract. (Hot-fill 185 °F / 2 min, pH 3.0 citric acid — degradation pattern matched Reb A. Powder: 25 °C / 60 % RH, 3 months: 86.8 → 85.65 %. Accelerated 40 °C / 75 % RH: 86.8 → 85.0 %.)
[14] Prakash I, Markosyan A, Bunders C. Development of next generation stevia sweetener: Rebaudioside M. Int J Mol Sci. 2014;15:1–10. (Reb M degradation products at pH 2.0 / 80 °C / 24 h: C-16/17 → C-15/16 alkene migration, 16β-hydroxy hydration, C-19 ester hydrolysis to Reb B. All products retain some or all glucose units.)
Steviol glycoside baselines
[6] JECFA (Joint FAO/WHO Expert Committee on Food Additives). Safety evaluation of certain food additives. 68th meeting, 2007. WHO Food Additive Series 60. (80 °C / 8 h / pH 4.0: 4 % decomposition; pH 3.0: 8 %. 100 °C / pH 3.0: 40 %. pH 3.8 / 24 °C / 1 year: essentially no decomposition. Phosphoric acid 3.75× more degradation than citric acid. 5 % acetic acid: greatest stability.)
[10] Wölwer-Rieck U, Tomberg W, Lam A. Investigations on the stability of stevioside and rebaudioside A in soft drinks. J Agric Food Chem. 2010;58(23):12261–12267. doi:10.1021/jf102894v (Reb A + stevioside in commercial sodas at 80 °C: up to 70 % loss after 72 h. No steviol aglycone detected. 26+ weeks stable at ambient.)
[12] Prakash I, Bunders C, Engles S, et al. Stability of stevioside and rebaudioside A in carbonated beverages. Innovare Acad Sci. 2012. (Stevioside 26-week mock beverage: pH 2.8 / 40 °C: 38.8 % retention; pH 3.2 / 40 °C: 64.2 %; pH 3.8 / 40 °C: 87.7 %; pH 4.2 / 40 °C: 92.8 %. Mass balance 97.5–100.5 %.)
Degradation kinetics & mechanism
[4] Gelinas C, Delage B. Acid-catalysed degradation of rebaudioside A in model beverage systems: effect of caramel colour and coffee melanoidins. Molecules. 2022;27(x). (Caramel colour E150d: +83–89 % degradation acceleration. Coffee melanoidins/polyphenols: −80–90 % epoxide product reduction. Formulation: Reb A 200 mg/L, citric acid 500 mg/L, malic acid 100 mg/L, potassium sorbate 200 mg/L, sodium benzoate 100 mg/L.)
[5] Gong SY. Kinetics of rebaudioside A degradation in buffer solutions at various pH and temperatures. MS thesis, Auburn University. 2013. (Pseudo-first-order in phosphate and citrate buffers at pH 3 / 5 / 7 × 20 / 30 / 40 °C. t10 % at pH 3 / 40 °C / 0.02 M phosphate: 182 days. pH 7 / 40 °C / 0.1 M phosphate: 41 days. Light exposure: ~8× acceleration in citrate pH 3. Ea: 5.4–41.3 kcal/mol.)
[11] Santana AB et al. Thermal decomposition kinetics of stevioside and erythritol blends by TGA/DSC. Foods. 2025;14(x). (Stevioside Ea ~140–147 kJ/mol at α = 0.1–0.5. Erythritol blending eliminates autocatalytic decomposition behaviour above 150 °C.)
Processing & matrix effects
[7] Amery R, Vlasselaer L, Van Loey A, Hendrickx M. Stability of steviol glycosides during thermal processing of dairy and soy beverages. KU Leuven. (UHT milk 140 °C / 5 s: Reb A ~80 %, stevioside ~91 %. Soy drink 140 °C / 5 s: Reb A ~101 %, stevioside ~101 %. Fermented milk drink 95 °C / 15 s: Reb A ~98 %, stevioside ~97 %. Ice cream 85 °C / 15 s: Reb A ~108 %.)
[8] Jooken E, Amery R, Vlasselaer L, et al. Stability of steviol glycosides in food matrices. (Tested: milk, soy, fermented milk drink, ice cream, yogurt, dry biscuits, jam. No decomposition detected in any matrix.)
[9] Clos JF, DuBois GE, Prakash I. Photostability of rebaudioside A and stevioside in beverages. J Agric Food Chem. 2008;56(18):8507–8513. doi:10.1021/jf801343e (Reb A and stevioside photostable: within 3–7 % of controls after 3 000 langleys sunlight. Degradation products formed equally in light and dark — acid-promoted, not light-promoted.)
Written by
Xinjuan XuAssociate Professor of Food Science. Xinjuan writes stevia and steviol glycoside technical content, drawing on published literature and formulation research.
Disclaimer — PhytoVale supplies several ingredients discussed in this article, including rebaudioside M, rebaudioside A and rebaudioside D. This content is provided for informational purposes only and does not constitute medical, legal or regulatory advice. Shelf-life estimates are calculated from published kinetic data and should be validated by accelerated stability testing in your specific matrix. Always verify claims against the primary literature cited above and consult qualified professionals for product-specific decisions.
Formulating with Rebaudioside M?
Developing an acidic, heat-processed or shelf-stable beverage? Tell us your target pH, processing temperature and shelf-life requirements. We'll help you evaluate Reb M 95% and 98% grades and provide available samples, HPLC COA, technical specifications and bulk pricing.
