Process-Engineering Mechanics of High-Dilution Intermediate Trituration Matrices for Microgram-Scale Actives

Manuscript 4 of 8 in the BARC Protocol specification series · Release B-2.1.0 · September 2026

Author. Prasanna M, Growlrr Foods Pvt Ltd, Chennai, India. Correspondence: support@growlrr.com. ORCID: 0009-0003-6812-347X.

DOI. 10.5281/zenodo.22647614 — the concept DOI, which always resolves to the current version of this record · Published. 8 September 2026 · Licence. CC BY-SA 4.0 · Status. Not peer-reviewed as of the date of publication, 8 September 2026.

Competing interests and funding. The author is the founder of Growlrr Foods Pvt Ltd, a pioneer and proponent of the BARC Protocol for pet foods. BARC stands for Biologically Appropriate, Residually Corrected, and is conceptualised as an open protocol under a Creative Commons Attribution–ShareAlike licence, for like-minded collaborators around the world to formulate and develop. Growlrr Foods Pvt Ltd develops and commercialises a BARC-based residual-correction product, BowlBalancer. The work was funded entirely by the company; no external funding was received. No independent laboratory, institution or reviewer has verified the results reported here. Readers should weigh these facts alongside the analysis, which is published in full — including its limitations and open items — precisely so that it can be checked.

Framework. This work applies BARC — Biologically Appropriate, Residually Corrected — an open protocol for correcting a biologically appropriate fresh base to a published nutrient standard, released under a Creative Commons Attribution–ShareAlike licence (CC BY-SA 4.0) at growlrr.com/barc-protocol (concept DOI 10.5281/zenodo.22648597) so that others may formulate and develop against it. BARC is the protocol; the correction specified and validated here is one implementation of it.

Terms. A glossary of the statistical and nutrition vocabulary used across this series — requirement, floor, safe upper limit, Wilson bound, N-limited, draw, stress multiplier, retention factor, trituration and the rest — is in the series cover note (10.5281/zenodo.22647606), section 4.1.

Computational tooling and AI assistance. The model is implemented in TypeScript under Node.js with no external numerical dependencies. Claude (Anthropic) was used as a coding and analysis assistant in developing this work and preparing these records; it was not used to generate results, all of which were produced by executing the model. See the series cover note (10.5281/zenodo.22647606), section 6.1.

Plain-language material. A non-technical account of the unit-dose format is at growlrr.com/pet-home-cooked-recipes-blog/the-5-gram-revolution. The public reference list behind these claims is at growlrr.com/pet-nutrition-references.

Disclosure boundary for this paper. The method is disclosed: the dilution architecture, the carrier concept, the acceptance criterion, the four-view bill-of-materials standard, and the principle of characterising the palatant as a nutritional input. Per-unit active masses, carrier fractions, premix quantities per production lot, ingredient grades and supplier specifications are withheld, being formulation rather than method.

Abstract

The principal process-engineering risk in a low-inclusion micronutrient corrector is the mass imbalance between its smallest actives and its bulk. Several actives in a 5 g unit — vitamin D₃, vitamin B₁₂, selenium, iodine, folate — are present at microgram scale, yet must be distributed uniformly through a matrix composed largely of calcium salts and a food-form palatant. Weighing such actives neat into a bulk mixer invites gravity segregation and particle-size mismatch, producing hot spots and voids across sequential packing runs.

This paper describes the mitigation: a multi-stage geometric trituration ladder of at least fifty-fold per dilution stage, in decades as the neat mass requires, onto a shared inert carrier pool in an isolated compounding stream before the intermediate enters the bulk blend. It sets out a four-view bill-of-materials standard that keeps raw weighed material, process requirement, acceptance criterion and measured result in separate columns so that a design target is never mistaken for an analytical result — and it states plainly that the measured column is pending the production performance-qualification phase at the date of this record.

It also documents why the food-form palatant is treated as a characterised nutritional input rather than an inert carrier, and why that characterisation had to be made by the authors: commercial palatants are proprietary formulations whose suppliers do not disclose nutrient composition, so a validation programme waiting on such a disclosure would wait indefinitely.

Finally it reports in-pack stability, which is where the paper's quantitative result sits. A single blended unit has no compartments, so protection for labile actives is particle-level and environmental rather than locational, and rests on four shared controls. Because the potency ceiling of vitamins A and D₃ is a toxicity limit, their decay must be addressed by form, never by adding more.

Modelling twelve months of decay against literature retention factors, both adult corrections clear the acceptance criterion at expiry as well as at manufacture: at 100,000 draws per species, the feline Wilson-95 lower bound is at or above 99.973% at manufacture and 99.983% at twelve months, and the canine at or above 99.972% and 99.950%. Differences of that size are run-to-run variation rather than a decay effect in either direction, and the correct reading is that twelve months of modelled decay does not move the gate. The reason is structural: the binding nutrient is copper, a mineral that does not decay, which accounts for the majority of the residual failing tail in both species, while the labile vitamins sit far enough above their floors that a 5 to 12% loss does not bring them near. Decay's only signature is compositional — riboflavin's share of that (already negligible) tail grows. That overlay was run on the release-A-1.0.0 formulation at the species reference animals only; requirements throughout are the NRC recommended allowances [6].

Keywords: BARC Protocol · residual correction · homemade dog food · homemade cat food · home-prepared diet · companion animal nutrition · geometric dilution · trituration · blend uniformity · content uniformity · microgram-scale actives · relative standard deviation · bill of materials · palatant characterisation · vitamin stability · pharmaceutical release standards

1. Introduction — the high-dilution homogeneity problem

The actives at issue are present at microgram scale per unit dose in a matrix of several thousand milligrams. Two failure modes follow directly.

Segregation. A minor mechanical dead zone or an unmatched particle-size distribution allows gravity separation of a dense, fine active from a coarser bulk, producing units that carry a multiple of the intended dose alongside units that carry a fraction.

Undetectability by the statistical model. The stochastic validation of Manuscript 3 models potency and fill as lognormal coefficients of variation. That representation cannot see a discrete misweigh — a decimal slip on selenium, copper or vitamin A — because such an event is not a draw from the modelled distribution. The fat tail intrinsic to weighing micrograms of neat active is a process risk, not a statistical one, and must be controlled by process.

2. Methods — the geometric dilution ladder

Unweighable actives are expanded through sequential dilution stages — at least fifty-fold per stage, arranged in decades (1:10 → 1:100 → 1:1,000) according to the neat mass involved — onto a shared inert carrier pool, inside an isolated compounding stream, before the resulting intermediate enters the bulk blend.

Neat weighing (rejected)
   µg neat active → bulk mixer → clumping and segregation → hot spots and voids

Geometric dilution (specified)
   µg neat active → ≥50× → ≥50× → ≥50× → shared carrier pool → uniform dispersion

The carrier pool is a conventional pharmaceutical excipient set — microcrystalline cellulose with a pre-gelatinised starch and a maltodextrin. Carrier demand is a small fraction of the unit's existing excipient mass, so the finished unit weight and the dose ladder are unchanged; the specific fractions are withheld.

Sequenced compounding. Each labile water-soluble vitamin is introduced as its own coated dilution at the final blend, never pre-mixed with its specific antagonist. The antagonist relationships that govern the sequence are nutrient-specific:

3. The four-view bill-of-materials standard

To keep raw weighed material distinct from delivered nutrient — and a design target distinct from an analytical result — the formulation is maintained in four linked views.

View 4 is empty at this record: no blend-uniformity or finished-unit assay result is claimed. A specification that reported a design target in the measured column would be claiming an analysis that has not been performed. Release requires assayed dilute premixes and a per-lot finished-unit assay; the stochastic model sizes and reassures, and the assay is the control.

4. Iodine: a purified active, and the stable salt

Iodine is supplied as calcium iodate — a purified active with a declared assay — and deliberately not as a marine botanical such as kelp.

The reason is variance, not preference. Botanical iodine content varies with species, harvest site, season and processing, and that variation is largely invisible to the formulator because it is not declared per lot. Substituting a variable input for a fixed one is always a loss of control, and it matters most in a nutrient with a narrow band between requirement and ceiling — which iodine is, being one of the two nutrients that bind first under deterministic over-dosing [Manuscript 6].

And within the purified actives, the more stable salt. Calcium iodate is used in preference to a potassium salt. Potassium iodide and potassium iodate are more hygroscopic and appreciably less stable in a trace-mineral premix, where moisture, oxidation and catalysis by neighbouring mineral particles all drive loss across shelf life; calcium iodate is the conventional premix choice for that reason. The distinction matters more here than it would elsewhere, because the entire argument for block-borne iodine is that the delivered amount does not drift. A source that decays in the pack would reintroduce over months exactly the variability the botanical was rejected for on day one.

The consequences for the model are direct. Because iodine is block-borne at a declared assay in a stable form, its contribution enters the simulation with the purified-active coefficient of variation of 0.05 rather than the grocery coefficient of 0.10 or worse, and it carries no shelf-life retention discount (§6.1). Constancy is also the clinically relevant property: feline thyroid disease has been associated with fluctuating rather than simply low dietary iodine, which argues for a fixed-assay, shelf-stable source independently of the mean delivered.

5. The palatant as a characterised nutritional input

The food-form palatant fraction — a spray-dried animal-digest blend — is the unit's largest single component by mass and is not chemically inert. Treating it as a carrier would mis-state the formulation. It is therefore characterised as a nutritional input, with three contributions mapped by the engine:

• Phosphorus, dominated by the meat-powder line. This is the residual that remains after the calcium re-architecture of Manuscript 2 and is why correction-borne phosphorus falls sharply but not to zero.
• Sodium. Animal digests are natively sodium-rich. The process specifies desalted grades and applies an iterative trim so that unit sodium remains below 90% of the NRC allowance.
• Incidental copper, zinc and cobalamin from the liver-digest line, which provide an organic trace cushion that measurably buffers the low-adequacy tail under copper stress.

Basis of the characterisation. Commercial palatant suppliers do not disclose nutrient composition, so the characterisation used throughout this series is derived by the authors from the palatant's constituent breakdown. Every validation figure in Manuscript 3 is computed on that basis, and any reader reproducing the method should treat the palatant contribution as a characterised input rather than a supplier-certified one.

6. Results — in-pack stability

A single blended unit has no compartments. Protection for labile actives is particle-level and environmental, not locational: "keep apart" is a compounding-sequence instruction, never a physical partition. Four shared controls carry it.

1. Coating and encapsulation — a universal barrier that protects even when an antagonist particle is an immediate neighbour, because the loss reaction requires molecular contact across a moisture or air bridge that the coat denies.
2. A dry matrix at water activity ≤ 0.60, in an opaque, oxygen-barrier laminate — removing the shared accelerants of moisture, oxygen and light.
3. Chelated, non-ionic minerals rather than free sulfate salts. The classic premix antagonism is a free-ionic-salt problem this formulation does not have.
4. Sequenced compounding as in §2.

Fat-soluble vitamins and long-chain omega-3 fail by oxidation and light and are beadlet- or micro-encapsulated with an antioxidant; riboflavin is light-degraded and is protected by the opaque pack.

The vitamin A and D₃ rule. Their potency ceiling is a toxicity limit, so decay in those two must be solved by form, never by increasing the input. This is the single most consequential line in the stability specification.

Shelf life is a target of 12 to 18 months, at a stated retention of ≥ 0.90 of requirement at expiry for the labile actives. The binding control is measurement — accelerated and real-time stability pulls under the conventional protocol conditions [1], with per-lot finished-unit assay — not the model. Terminal oxidation control by inert-gas flush or an equivalent packaging process is to be established with the contract manufacturer and is not asserted here.

6.1 The modelled ageing result, and why it is reassuring for a structural reason

Adequacy at manufacture and adequacy at expiry are different failure modes, and are run as two passes of the same validation model [Manuscript 2]. In the second pass the correction's labile-nutrient content is multiplied by a lognormal 12-month retention factor — literature-based, deliberately conservative, with means of 0.94 to 0.96 for the vitamins and long-chain omega-3 at a coefficient of variation of 0.05, and 1.00 for minerals, amino acids and taurine, which do not decay [2,3,4,5]. Only the correction decays; the owner's food is bought fresh.

On the release-A-1.0.0 formulation at 100,000 draws per species, both adult corrections pass the gate at 12 months: cat Wilson-95 lower bound ≥ 99.973% at manufacture and ≥ 99.983% at 12 months; dog ≥ 99.972% and ≥ 99.950%. The two passes draw independent random streams, so differences of this order — a hundredth of a percentage point at 100,000 draws — are run-to-run variation and not a decay effect in either direction. The reading these figures support is that twelve months of modelled decay does not move the gate, not that reliability improves with age.

The reason is structural, and it is what running the real model shows rather than a per-nutrient calculation. The binding nutrient is copper — a mineral that does not decay — which accounts for 90% of the dog failing tail and 61% of the cat's, while the vitamins sit at roughly 116 to 127% of their floors on average, so a 5 to 12% loss never brings them near. Decay's only signature is compositional: riboflavin's share of the (already tiny) failing tail grows from 32% to 43% in cats. On this overlay it is riboflavin, rather than vitamin A, D or B₁₂, that would repay monitoring.

A stress check confirms the ageing step is biting. Unencapsulated forms at roughly 0.88 retention still pass; the gate breaks only at an unrealistic 50% loss on every labile nutrient, where thiamine binds first.

The consequence for overage. Adequacy holds with no overage. Any overage is therefore an optional stability margin for the potency floor, sized from measured data — not an adequacy requirement.

7. Limitations and open items

• View 4 of the bill-of-materials standard is empty. Blend-uniformity and finished-unit assay results are pending the production performance-qualification phase (§3). No analytical yield is claimed.
• Retention factors are a-priori literature values, not measured (§6.1). The binding control is the measured-retention pass, which supersedes the provisional one when it lands.
• Decay is correlated within a production lot — one vendor's encapsulation quality, one pallet's thermal history — so the real risk is a whole-lot excursion rather than scattered unlucky units. That is caught by per-lot assay, not by the model.
• The ageing overlay was run on the release-A-1.0.0 formulation and at the species reference animals only. A re-run on the current formulation and a body-weight sweep are outstanding. The copper-bound conclusion is expected to carry, because the binding nutrient is unchanged.
• Growth and reproduction configurations are expected to follow the adult ageing result, sharing the architecture and the same non-decaying binding nutrient. Each will be confirmed on the same model in turn.
• Terminal oxidation control is not established (§6).
• Not peer-reviewed as of the date of publication (8 September 2026). Should it later be reviewed or independently replicated, that will be recorded in a subsequent version of this record.

8. Conclusion

The governing risk in a low-inclusion corrector is the mass ratio between its smallest actives and its bulk, which places uniformity — not potency — at the centre of the process design. Geometric dilution in stages of no more than fiftyfold, with the four-view bill of materials keeping raw weighed mass, process requirement, acceptance criterion and measured result in separate columns, is the method adopted to manage it.

Two input choices follow the same logic rather than a nutritional one. Iodine is supplied as calcium iodate, a purified active of declared assay, in preference both to marine botanicals whose content varies by species, season, site and lot, and to potassium iodate on shelf-life grounds. The palatant is treated as a characterised nutritional input rather than an inert carrier, because a material present at that mass fraction cannot be assumed to contribute nothing.

The modelled twelve-month ageing overlay is bound by copper, a mineral that does not decay, which is why decay does not move the acceptance criterion. That overlay was run on the release-A-1.0.0 formulation at the species reference animals only; a re-run on the current formulation across a body-weight sweep remains outstanding, and no blend-uniformity or finished-unit assay result is claimed at this record.

References

1. International Council for Harmonisation. Q1A(R2): Stability Testing of New Drug Substances and Products. Step 4 version, 6 February 2003. https://database.ich.org
2. DSM Nutritional Products. Optimum Vitamin Nutrition (OVN) guidelines for companion animals [industry guideline]. Kaiseraugst: DSM Nutritional Products. https://www.dsm-firmenich.com/anh/products-and-services/tools/ovn.html (accessed 7 September 2026). Manufacturer guideline, cited as industry practice and not as an independent standard.
3. Gadient M. Effect of pelleting on nutritional quality of feed. In: Proceedings of the 1986 Maryland Nutrition Conference for Feed Manufacturers. College Park (MD); 1986. p. 73.
4. Frye TM. The performance of vitamins in multicomponent premixes. In: Proceedings of the Roche Technical Symposium. Jefferson (GA); 1994.
5. Coelho MB. Vitamin stability in premixes and feeds: a practical approach. In: BASF Technical Symposium. Bloomington (MN); 1991. p. 56–71.
6. National Research Council (US), Committee on Animal Nutrition. Nutrient Requirements of Dogs and Cats. Washington (DC): The National Academies Press; 2006. doi:10.17226/10668

How to cite

Prasanna M. (2026). Process-Engineering Mechanics of High-Dilution Intermediate Trituration Matrices for Microgram-Scale Actives (Manuscript 4 of 8, Release B-2.1.0). Growlrr Foods Pvt Ltd. Zenodo. https://doi.org/10.5281/zenodo.22647614

The series

This manuscript is one of eight deposited together under release B-2.1.0, with a series cover note carrying the concept DOI and the full DOI list. All eight apply the BARC Protocol, published openly and separately at growlrr.com/barc-protocol.