Why does direct compression fail when the powder’s D90 exceeds 250 µm?

In direct compression of Changqing Powder Veterinary Grade API into veterinary tablets, blend uniformity is controlled more by particle size distribution than by total blend time. When the supplied powder retains particles above 250 µm and a bulk density below 0.45 g/mL, segregation occurs during hopper discharge into the feed frame. A typical low-to-medium dose tablet core is formulated with 10–25% w/w active powder, 55–75% w/w microcrystalline cellulose NF, 10–20% w/w lactose monohydrate, 2–4% w/w croscarmellose sodium, and 0.5–1.0% w/w magnesium stearate. The API is first pre-blended with microcrystalline cellulose at a 1:4 ratio in a 600 L IBC bin blender for 15 min at 10 rpm to break loose agglomerates. The remaining filler is added through a 0.8 mm conical screen, followed by croscarmellose sodium and final lubrication with magnesium stearate for 3 min. Prolonged lubrication beyond 5 min or batch moisture above 2.5% w/w increases tablet sticking to D-tooling steel and causes chipping at the lower punch edge.

Compression is carried out on a 45-station rotary tablet press with D-tooling, a forced feeder set at 30 rpm, precompression of 8 kN, and main compression between 12 kN and 18 kN. In production batches, tablet hardness is maintained at 80–120 N; friability is controlled under USP 1216 with a limit of not more than 1.0%. Disintegration is tested under USP 701 and must not exceed 15 min. Content uniformity follows USP 905 or EP 2.9.40 with an acceptance value not greater than 15.0. Dissolution is evaluated under USP 711 using an immediate-release criterion of Q = 75% at 30 min. If the powder has a Hausner ratio above 1.35, direct compression is not recommended without a pre-granulation step; die fill variance produces weight RSD above 3.0% and content uniformity failures in tablets below 20 mg active per unit. Finished tablets are film-coated with an HPMC-based moisture barrier and packaged in Alu/Alu blisters to limit hydrolytic degradation. Regulatory release is governed by 21 CFR 210/211 and EU GMP Part II; residual solvent control follows VICH GL18 with USP 467 as the analytical platform. Open-container stability at 40°C/75% RH is relevant for tropical markets under VICH GL1 and GL2.

Typically, aqueous parenteral manufacture of the powder begins with solubility mapping in Water for Injection at 20°C and 37°C across pH 3.0–7.0, because a pH shift at the dilution front can cause local precipitation even when bulk solubility is adequate. A representative injectable solution is formulated at 1–10% w/v active, adjusted to pH 4.5–6.5 with citrate or acetate buffer, and rendered isotonic at 280–320 mOsm/kg with sodium chloride or dextrose. If the active is oxygen-sensitive, sodium metabisulfite is added at 0.1% w/v only after compatibility testing under 25°C and 60/4000 lux photostability; nitrogen overlay in the mixing vessel reduces headspace oxygen to below 2.0%.

The solution is prepared in a 316L stainless steel jacketed vessel with a side-entry agitator operated at 150–250 rpm. The API is added slowly through a vacuum transfer lance to avoid powder clumping. For heat-stable products, terminal sterilization is performed at 121°C for 15 min; for heat-labile products, the solution is filtered through a 0.45 µm polyethersulfone depth filter followed by two 0.22 µm sterilizing-grade PVDF filters in series and filled aseptically in a grade A isolator. Glass vials are Type I borosilicate, washed and depyrogenated in a tunnel at 250°C for 30 min. Stoppers are bromobutyl rubber, sterilized by steam and dried to moisture below 0.5%. In-line monitoring includes pH, conductivity, and filter integrity testing before and after filtration; bubble point values below the manufacturer’s water-based threshold indicate membrane damage and require batch rejection.

Release testing is performed under USP 1 Injections, USP 788 particulate matter, and USP 85 bacterial endotoxins; subvisible particle counts must not exceed 6000 particles per container at ≥10 µm and 600 particles per container at ≥25 µm. For veterinary parenteral products, the endotoxin limit is derived from the maximum dose and route under USP 85; for intravenous large-volume products, the limit can be as low as 0.5 EU/mg when the maximum expected dose is high. Process control must comply with EU GMP Annex 1 and 21 CFR 210/211; environmental monitoring in filling areas is assigned alert and action limits of 1 CFU/m³ and 5 CFU/m³ for grade A cleanrooms under ISO 14644-1. If the API forms insoluble salts with phosphate or calcium, phosphate buffers are replaced with acetate, and calcium-containing excipients are excluded from the formulation.

Low-fill capsule uniformity and the 20 mg potency threshold

If the API is potent and the capsule dose is below 20 mg, geometric dilution before encapsulation becomes the main content-uniformity lever. Capsule manufacture typically uses a 1:5 pre-blend of the API with lactose monohydrate NF, passed through a 0.8 mm conical mill to reduce agglomerates, then diluted to final fill weight with microcrystalline cellulose or pregelatinized starch. Formulation ranges are 2–20% w/w active, 60–85% w/w filler, 2–5% w/w disintegrant such as croscarmellose sodium, 0.5% w/w colloidal silicon dioxide, and 0.5–1.0% w/w magnesium stearate. Hard gelatin capsule size 3 or 4 is filled on an intermittent-motion dosing-disc machine; powder bed height is held at 30–40 mm and dosator compression ratio is set at 2:1 to achieve a fill weight RSD below 3.0%.

Because low-fill capsules are highly sensitive to active agglomeration, the pre-blend is sampled at 10 thief locations after 15 min in a 500 L bin blender and analyzed for active content by a validated HPLC method following ICH Q2(R1) or VICH GL49. The acceptance criterion is RSD ≤5.0% before lubrication. After encapsulation, content uniformity is determined under USP 905 or EP 2.9.40, and disintegration is tested under USP 2040 with a limit of not more than 15 min. Dissolution of filled capsules follows USP 711 with the same immediate-release specification used for tablets. Humidity control during encapsulation is critical: if process air exceeds 40% RH, the gelatin shell absorbs moisture and becomes brittle, while active hydrolysis may increase above the stability specification. Therefore, primary packaging is a PVC/PVDC/Alu blister or HDPE bottle with silica gel desiccant.

Dry powder formulations for oral administration are not compressed, so particle size distribution and bulk density control the final mixing behavior. The API is diluted to 5–20% w/w with a carrier blend of lactose monohydrate NF and maltodextrin, and 0.5% w/w colloidal silicon dioxide is included to reduce electrostatic adhesion to stainless steel surfaces. All materials are screened through a 425 µm sieve and discharged into a 1000 L tumble blender operating at 10 rpm for 20 min. Because the API powder may be cohesive at ambient moisture, the processing room is held below 40% RH and the product is sampled for homogeneity after mixing. Homogeneity is accepted when 10 powder samples taken from defined top, middle, and bottom locations show active content within 95.0–105.0% of label and RSD ≤5.0%.

Unit-dose sachets are filled on vertical form-fill-seal equipment with auger dosing; fill weight is monitored every 15 min with an acceptance range of ±2.0% around target. Bulk multi-dose jars are fitted with a desiccant closure because the powder can cake when stored above 25°C and 60% RH. Release testing includes particle size distribution by analytical sieving under USP 786 or EP 2.9.12, loss on drying under USP 731 with a typical limit of NMT 3.0%, and microbial enumeration under USP 61 and USP 62. For non-sterile oral powders, total aerobic microbial count is generally controlled at ≤1000 CFU/g and total combined yeasts/molds at ≤100 CFU/g. If the powder is intended for reconstitution in milk replacer or water for neonatal animals, microbial contamination risk is addressed under ICH Q6A because the finished dose is not sterile.

When roller compaction raises granule porosity risk, dissolution fraction at 30 min becomes the controlling release parameter

Because the API is heat-labile, roller compaction is selected over wet granulation when the powder degrades above 40°C or in contact with granulation fluid. The pre-blend is composed of 10–25% w/w active powder, 55–75% w/w microcrystalline cellulose, 10–20% w/w lactose monohydrate, and 2–4% w/w croscarmellose sodium as intragranular disintegrant. Roller compaction is performed at roll pressure 4–8 MPa, roll gap 2–3 mm, and roll speed 5–10 rpm. The compacted ribbons are milled through a 1.0 mm screen using an oscillating granulator to produce granules with a bulk density of 0.55–0.75 g/mL. Extragranular croscarmellose sodium is then added at 1–2% w/w, followed by magnesium stearate at 0.5% w/w and final blending for 5 min.

The main process conflict is granule porosity: roll pressure below 4 MPa creates friable granules that segregate during subsequent filling or compression, while pressure above 8 MPa densifies the granules and suppresses disintegration. Dissolution is therefore monitored as the key granulation endpoint; immediate-release granules are required to release not less than 75% of the active at 30 min under USP 711 or EP 2.9.3. In roller-compacted batches, granule particle size distribution is measured by USP 786 or EP 2.9.12; the target is 100% passing 1.4 mm and NMT 35% below 150 µm. Granules can be filled into sachets or compressed into tablets, and the same release specification applies to the finished dosage form. Moisture in the granulation area must remain below 40% RH because compacted granules absorb water quickly and may produce tablets that cap during compression.

Compliance for granulation processes is derived from 21 CFR 210/211 and EU GMP Part II; process validation follows ICH Q8(R2) principles where a design space is established for roll pressure, screen size, and final blend time. Stability testing is performed according to VICH GL1 and GL2 with the same zone-specific humidity conditions used for tablets. Published data for this specific API configuration may be limited at the upper roll-pressure boundary; therefore, a roll-pressure tolerance of ±0.5 MPa is recommended during commercial validation until a terminal dissolution curve is generated with production-scale batches.

A staged dilution sequence is established before the first production batch is weighed. In feed-mill premix operations, the arithmetic of dilution changes because the active level in the finished feed may be below 1 ppm, which is beyond the sensitivity of direct blending of technical API. The premix is prepared at 2–10% w/w active on a carrier of wheat middlings or calcium carbonate, with 0.5–1.0% w/w mineral oil added to reduce dust and electrostatic carryover. The first dilution from technical powder to premix is made at 1:10 in a 500 L horizontal ribbon mixer; after 10 min of mixing, the batch is discharged and diluted again at 1:100 in a 2000 L plow mixer. The final feed mill then incorporates this premix into complete feed at an additional 1:100 to 1:500 dilution depending on species and label dose.

Validation of premix homogeneity is performed by sampling at 10 fixed points and analyzing active content; the acceptance criterion for the first stage is RSD ≤5.0%, and for the second stage RSD ≤10.0%. If the carrier has a mean particle size below 200 µm, dust formation increases; if above 800 µm, segregation occurs during transport and bin discharge. Therefore, carrier particle size is controlled to 200–800 µm with not more than 5.0% passing 75 µm. Production equipment must be cleaned between batches because carryover into non-target feed can produce tissue residues in food-producing animals. Under EU 2019/4, maximum permitted carryover for certain active substances is limited to 1% in feed for target species and 3% in feed for non-target species; the exact limit depends on the regulatory classification of the API. In the United States, medicated feed premixes are handled under 21 CFR 558 and associated GMP regulations. Finished premix is packed in 25 kg multi-wall paper bags with an inner PE liner and stored below 30°C and 60% RH.

For drinking water and drench solutions, the stock concentration is determined by the solubility of the powder in farm water at the target pH. A concentrated stock solution is prepared at 10–20% w/v active, with citric acid or lactic acid added to maintain pH 3.5–4.5; if the API is poorly soluble, propylene glycol is included at up to 20% v/v. The target drinking water concentration is typically 0.01–0.05% w/v active after in-line dilution. The stock is mixed in a 200 L high-density polyethylene tank with a 100–200 rpm stainless steel agitator for 30 min after full addition of the powder. Acidification is performed before the API is added; adding the acid after the API can create a low-pH gradient that causes premature precipitation or degradation.

For field administration, the stock solution is metered through a membrane dosing pump at 1–5% into the drinking water main. Water hardness above 300 mg/L CaCO₃ can reduce solubility if the active forms a calcium salt; therefore, a chelator such as disodium EDTA is added at 0.1% w/v in hard-water regions. The solution is not sterile, so microbial quality is controlled by the preserved stock system or by freshly prepared batches that are used within 24–48 h. If the stock solution contains organic acids, contact with galvanized steel, copper, or aluminum is avoided because corrosion can release metal ions and catalyze oxidation. Distribution lines are flushed after each medication cycle to prevent residue accumulation and biofilm formation.

Release and stability tests for the stock solution are performed under USP 795 for nonsterile preparations where applicable, or under 21 CFR 210/211 for commercial manufactured stock. Assay is typically 95.0–105.0% of label, pH is 3.5–4.5, and total aerobic microbial count is ≤100 CFU/mL with absence of Salmonella and E. coli in 10 mL. The final drench solution must not be mixed with alkaline chlorinated cleaners, because hypochlorite can oxidize the active and generate chlorinated degradation products. Data for this specific water matrix can vary widely by farm; therefore, bracketing studies at pH 2.5–7.0 and hardness 0–400 mg/L CaCO₃ are indicated for registration in markets with variable water quality.