Formulating a Zero-Sugar Carbonated Soft Drink: A Sweetener Blend Guide
Replacing sugar in a carbonated soft drink is harder than it looks, because sucrose does three jobs at once. It sweetens, it builds body and mouthfeel, and it balances the acid bite. A full-sugar CSD carries roughly 100 to 110 g/L of sugar, about 10 to 11 degrees Brix. High-intensity sweeteners replace the sweetness at a few hundred mg/L and contribute none of the body.
If you produce a diet or zero-sugar CSD, a good one is a blending exercise, not a one-for-one swap. Here is the data, the mechanism, a worked formula, and the failure modes.
The sweetener data table
Relative sweetness is versus sucrose at 1.0. Typical use level is the approximate single-sweetener dose to reach cola-level sweetness, before blending. Regulatory caps vary by country, so confirm your market.
| Sweetener | Rel. sweetness | Typical use (single) | Acid/shelf stability | Profile note |
|---|---|---|---|---|
| Aspartame | ~180 to 200x | ~500 to 550 mg/L | Poor: hydrolyzes over months | Most sugar-like, clean |
| Acesulfame K | ~200x | ~300 to 500 mg/L | Excellent | Fast onset, metallic tail |
| Sucralose | ~600x | ~200 to 300 mg/L | Excellent | Clean, slight lingering |
| Steviol glycoside (Reb A) | ~200 to 300x | ~300 to 500 mg/L | Good | Licorice, lingering tail |
| Monk fruit extract | ~150 to 250x | dose to taste | Good | Clean, slight aftertaste |
Mechanism: why you blend, and why aspartame fades
Why synergy works. Sweet taste is triggered by the T1R2/T1R3 receptor. Different high-intensity sweeteners bind at different sites on that receptor complex. When two are present, the combined receptor activation is greater than the linear sum of each alone, so a blend tastes sweeter than the two doses added together. In practice an Aspartame plus Acesulfame K blend shows roughly a 20 to 40 percent synergy boost, so you use less of each and cut cost. The three single options are compared in sucralose vs acesulfame-K vs aspartame.
Why the temporal profile matters. Ace-K has a fast sweetness onset and an early peak. Sucralose and aspartame are rounder and slower. Sugar itself has a smooth rise and clean finish. Pairing a fast-onset sweetener with a rounder one rebuilds that sucrose-like curve instead of leaving a sharp front peak or a long sweet tail.
Why aspartame loses sweetness on the shelf. Aspartame is a dipeptide methyl ester. In the acidic, watery CSD environment (pH below about 3.5) the ester and peptide bonds slowly hydrolyze, and the molecule also cyclizes to diketopiperazine. The breakdown products are not sweet, so a 100 percent aspartame diet drink measurably loses sweetness over months, faster when warm. This is the single reason aspartame is blended with Ace-K or replaced by sucralose for long-shelf-life product.
The missing mouthfeel, and how to get it back
This is what separates a flat-tasting diet drink from a good one. Sugar at 100 g/L gives viscosity and a coating body that 300 mg/L of sweetener cannot.
Levers to rebuild body: a touch of soluble fiber such as Resistant Maltodextrin at 5 to 20 g/L (also supports a fiber claim); a very low dose of a high-clarity hydrocolloid; or in reduced-sugar (not zero) versions, a small allulose or erythritol contribution that adds genuine bulk (see allulose vs erythritol vs monk fruit). Carbonation level (typically 3.5 to 4.0 volumes CO2 in cola) and serving temperature also modulate perceived body and bite.
The acid system
Cola relies on Phosphoric Acid (pH around 2.5 to 2.7). Citrus and fruit CSDs use Citric Acid Anhydrous, often with L-Malic Acid for a rounder, longer sourness (see citric vs malic vs tartaric acid). When you remove sugar, the perceived sweetness that masked some acidity is gone, so re-tune total acid down slightly to keep the same balance. Sodium Citrate or potassium citrate buffers hold pH steady against flavor-acid drift.
A worked diet-cola formula (per liter)
| Component | Amount | Role |
|---|---|---|
| Acesulfame Potassium | ~150 mg | fast onset, stability |
| Sucralose | ~120 mg | clean body, stability |
| Phosphoric Acid (85%) | ~0.5 to 0.6 mL | cola acid, pH 2.5 to 2.7 |
| Sodium Citrate | ~0.2 g | buffer |
| Caffeine (optional) | ~100 to 150 mg | bitterness, lift |
| Caramel color, cola flavor | to spec | color, flavor |
| Resistant Maltodextrin (optional) | 5 to 10 g | body, fiber claim |
| CO2 | 3.5 to 4.0 vol | carbonation |
This Ace-K plus sucralose base is the stable, long-shelf-life standard. Where shelf life is shorter and the most sugar-like taste is the priority, an Aspartame plus Ace-K blend is the alternative.
Troubleshooting
| Problem | Likely cause | Fix |
|---|---|---|
| Sweetness fades over shelf life | Aspartame hydrolysis at low pH | Switch to Ace-K plus sucralose, or raise pH slightly within flavor limits |
| Metallic or bitter top note | Acesulfame K alone, or overdose | Blend Ace-K down with sucralose or aspartame for synergy |
| Lingering sweet tail | Sucralose or stevia overdose | Reduce dose, rebalance blend toward faster-onset Ace-K |
| Thin, watery body | No bulk phase | Add Resistant Maltodextrin or low-dose hydrocolloid, tune carbonation |
| Drink tastes too sharp or sour | Acid unchanged after sugar removed | Lower total acid, add citrate buffer |
| Licorice off-note (natural version) | Steviol glycoside profile | Blend stevia with erythritol and monk fruit to round it |
Choose by what you produce
- If you produce a long-shelf-life canned or PET cola, use the Ace-K plus sucralose base and avoid relying on aspartame alone.
- If your finished products carry a natural claim, build on steviol glycosides plus monk fruit, rounded with erythritol, and accept a harder flavor-masking job.
- If you produce a reduced-sugar (not zero) line, keep some sugar or allulose for body and top up with a high-intensity blend.
We supply the full range of carbonated soft drink ingredients, including Aspartame, Acesulfame Potassium, Sucralose, Monk Fruit Extract, the acid system, and preservatives, in bulk with CoA and documents. Tell us your target sweetness, shelf life, and natural versus synthetic positioning, and we will propose a blend in mg/L and quote it.





