ZL Botanicals

Buying citrus fibre for beverages: choose the grade and the process together

Application & Formulation Guides

Application & Formulation Guides 2026-09-28
Citrus fibre Beverages Material selection

Citrus fibre can be bought to build body, hold water or keep particles suspended—not just to raise a fibre declaration. Tate & Lyle’s current NUTRAVA product page explicitly positions the ingredient for beverage mouthfeel and stabilisation, as well as fibre enrichment. It also lists thickening and emulsification in condiments and soups. These are supplier-described applications, not a single performance specification that applies to every grade or recipe. [S4]

For a beverage developer, the important commercial question is whether the supplied fibre will deliver the required texture on the equipment available. A powder that looks well dispersed after mixing may behave differently after more intensive processing. Conversely, increasing processing intensity need not improve every stability measurement. Two primary studies make that distinction useful for sample selection: one compared named commercial fibres under ultrasound, while the other varied citrus-fibre concentration and homogenisation pressure in cocoa milk. [S1, S2]

Citrus fibre is not simply another name for citrus pectin

The 2023 ultrasound study used two Herbafood citrus fibres: Herbacel AQ Plus Citrus N, abbreviated CF AQ, and Herbacel Classic CF 02, abbreviated CF Cl. Their measured soluble and insoluble fibre contents differed substantially. [S2, Methods 2.1 and 2.3.1; Table 2]

Study fibre composition — S2, Table 2; mean ± SD, % dry matter
Study materialInsoluble dietary fibre, % dry matterSoluble dietary fibre, % dry matter
Study materialCF AQInsoluble dietary fibre, % dry matter74.56 ± 0.34Soluble dietary fibre, % dry matter14.57 ± 0.83
Study materialCF ClInsoluble dietary fibre, % dry matter39.80 ± 0.16Soluble dietary fibre, % dry matter23.30 ± 0.22

These are the study’s measurements using AOAC 991.43, not current batch specifications. Values are means ± standard deviations; the methods describe duplicate experiments unless otherwise specified. [S2] They describe two mixed fibre materials, rather than two purified pectins with different esterification levels.

The material with more soluble dietary fibre was not the one showing the larger viscosity response to ultrasound. The authors reported a pronounced increase for CF AQ, whereas CF Cl’s response was much smaller. Their interpretation involved the fibre structure, particle interactions and processing history—not pectin content alone. [S2, Results, discussion of Figure 7]

This was a laboratory comparison: 1 wt% fibre suspensions, adjusted to pH 7, were treated for 150 seconds at a reported energy density of 225 kJ/L, with cooling so that the suspension did not exceed 40°C. It does not establish equivalence between ultrasound and a factory homogeniser. It does establish why soluble-fibre percentage alone is a poor basis for predicting the process response of another citrus-fibre material. [S2, Methods 2.2]

For procurement, keep three identities separate: a purified citrus pectin selected for a particular gel or protein-stabilisation role; a mixed citrus-fibre ingredient; and the exact commercial grade within that fibre family. The existing citrus-pectin selection guide addresses the first task. Here the buying decision concerns the second and third.

Cocoa milk shows why pressure cannot be specified on its own

A 2025 study used NUTRAVA Citrus Fiber Zest, donated by CP Kelco, in milk containing 3% fat and 3% protein. The formulation also included 6% sugar, 1% strongly alkalised cocoa, 1% dipotassium phosphate and 0.05% potassium sorbate. The mixture was heated to 75°C before homogenisation, followed by 95°C for four minutes and cooling. These are the reported research conditions, not a production formula or a validated safety process for another drink. [S1, Methods 2.1–2.2]

Four rows of the original experimental table show an important contrast:

Dose × pressure subset — S1, Table 1; mean ± SD
Fibre in the study formula, %Pressure, barSedimentation, %Creaming index, %Viscosity, mPa·s
Fibre in the study formula, %0.100Pressure, bar100Sedimentation, %17.96 ± 0.53Creaming index, %59.15 ± 2.07Viscosity, mPa·s77.99 ± 0.30
Fibre in the study formula, %0.100Pressure, bar300Sedimentation, %3.96 ± 0.29Creaming index, %29.31 ± 0.11Viscosity, mPa·s94.36 ± 0.79
Fibre in the study formula, %0.400Pressure, bar100Sedimentation, %0.00 ± 0.00Creaming index, %0.00 ± 0.00Viscosity, mPa·s161.10 ± 0.66
Fibre in the study formula, %0.400Pressure, bar300Sedimentation, %0.75 ± 0.12Creaming index, %2.46 ± 0.22Viscosity, mPa·s201.60 ± 1.08

Source: S1, Table 1, runs 7, 1, 11 and 5. Values are means ± SD; the table says analyses were performed at least in triplicate. For viscosity, the methods explicitly specify three measurements on the same sample, not three independent manufacturing batches. Sedimentation and creaming methods specify 21 days at 4 ± 1°C. Viscosity was measured at 25°C and 100 rpm with the stated R-2 spindle. The percentages describe layer-height ratios, not masses of lost cocoa or fat. [S1, Methods 2.4–2.7]

At 0.100% fibre, the higher-pressure treatment had lower sedimentation and creaming values. At 0.400%, higher pressure increased viscosity but did not improve those two separation results. Water-holding capacity nevertheless increased at 0.400% fibre, from 92.36 ± 0.23% at 100 bar to 95.35 ± 0.49% at 300 bar. Higher pressure therefore improved some measured properties while worsening the two separation means. These are observed means, not a new pairwise statistical analysis. They do not support a blanket instruction to maximise—or minimise—pressure.

These four rows are the matched low/high-pressure comparisons at the two stated fibre levels, not the entire design. The full table contains 11 runs, including axial points and three 0.250%/200-bar centre runs. Those centre runs all recorded zero sedimentation but different creaming means (0.00%, 2.46% and 0.69%). The data describe responses under a particular design; they are not a universal dose–pressure map.

The paper’s prose says sedimentation was zero at fibre concentrations of 0.250% and above and pressures of 100 bar and above. Its own 0.400%/300-bar row is a counterexample. The table should remain visible rather than turning the prose into a universal minimum-dose rule. [S1, Results 3.1; Table 1]

No visible sediment does not mean unchanged drinking texture

The authors selected 0.400% fibre and 272.67 bar as their model-derived optimum. Table 2 reports experimental viscosity of 197.22 ± 0.91 mPa·s, with sedimentation and creaming recorded as “nd”—not detected. That is an optimum for the study’s chosen objectives and formula, not a recommended setting for all cocoa drinks. [S1, Results 3.2.1; Table 2]

The subsequent storage table is more commercially informative than the word “stable” alone. Sedimentation remained recorded as nd, but viscosity increased from 204.95 ± 6.29 mPa·s in month 1 to 305.10 ± 6.08 mPa·s in month 9. The first two months share significance letter a; month 3 carries b. Thus, measured viscosity had already changed while sedimentation still went undetected. [S1, Table 5]

The complete monthly series also shows why sedimentation should not stand in for every stability measure:

Nine-month physical measurements — S1, Table 5; mean ± SD (n = 3)
MonthSedimentation, %Creaming index, %WHC, %Viscosity, mPa·s
Month1Sedimentation, %ndCreaming index, %ndWHC, %95.80 ± 0.42aViscosity, mPa·s204.95 ± 6.29a
Month2Sedimentation, %ndCreaming index, %ndWHC, %95.45 ± 0.49aViscosity, mPa·s213.45 ± 3.46a
Month3Sedimentation, %ndCreaming index, %ndWHC, %94.80 ± 0.14aViscosity, mPa·s225.90 ± 4.53b
Month4Sedimentation, %ndCreaming index, %ndWHC, %93.75 ± 1.06aViscosity, mPa·s238.43 ± 3.80bc
Month5Sedimentation, %ndCreaming index, %0.01 ± 0.0008aWHC, %93.18 ± 0.32aViscosity, mPa·s244.38 ± 4.61c
Month6Sedimentation, %ndCreaming index, %0.01 ± 0.0011aWHC, %92.85 ± 0.21aViscosity, mPa·s259.70 ± 3.11cd
Month7Sedimentation, %ndCreaming index, %0.02 ± 0.0016aWHC, %91.03 ± 0.32aViscosity, mPa·s276.69 ± 3.44d
Month8Sedimentation, %ndCreaming index, %0.02 ± 0.0021aWHC, %90.30 ± 0.14bViscosity, mPa·s290.08 ± 2.52e
Month9Sedimentation, %ndCreaming index, %0.03 ± 0.0016aWHC, %89.65 ± 0.35bViscosity, mPa·s305.10 ± 6.08e

Source: S1, Table 5. Values are means ± SD (n = 3); nd means not detected, not a measured zero. Different letters within the same measurement column denote significant differences by Tukey’s test (p < 0.05); shared letters do not. The letters are retained as published and must not be compared across columns. Creaming was reported as nd only in months 1–4; small nonzero values appear in months 5–9. WHC fell from 95.80 ± 0.42% to 89.65 ± 0.35%, with months 8–9 in group b rather than a. This is evidence of several physical changes, not an unchanged drink. The table does not measure how consumers perceived the stored drink.

Do not use this result to claim a nine-month commercial shelf life. The dedicated nine-month storage-method section does not explicitly state its storage temperature, and the table reports physical properties rather than microbial safety. The separate 21-day refrigerated method does not justify silently assigning that temperature to the entire nine-month experiment. [S1, Methods 2.4–2.5 and 2.14]

Buy a workable texture, not the highest fibre number

The evidence supports a more specific sample request than “citrus fibre for a stable beverage.” State whether the main job is suspending cocoa, replacing lost body, retaining water or adding a measured amount of dietary fibre. Those goals can lead to different acceptable textures and different inclusion levels.

Ask for the exact grade and the mixing or homogenisation conditions behind its application data. Then assess the candidate on the intended line, retaining the current formulation as a reference. Compare both suspension and drinking texture after processing and during storage; neither can stand in for the other. These are development recommendations drawn from the studies, not completed ZL trials.

A supplier’s family-level beverage positioning is a useful starting point, but it does not prove that today’s proposed grade is identical to the Zest material studied in 2025. Nor does a successful cocoa-milk result validate a clear acidic drink, a plant-based protein beverage or an arbitrary pectin replacement.

The practical purchasing outcome is a defined fibre material paired with an achievable process and an acceptable texture over time. Grade availability, destination-market suitability, specifications, batch documents and commercial terms still require project-specific confirmation; the named examples here do not imply that ZL stocks or distributes those brands.

For related distinctions between extract, fibre and pectin, read the apple material guide.

Sources