A silica anti-caking agent for feed works by coating host particles with fine, porous silica so they cannot form crystal bridges where moisture collects at contact points. For premixes and trace-mineral blends, the grade that wins is the one that gives continuous coverage at the lowest addition rate — proven in a multi-week storage trial, not in a flow check on fresh blend.
Caking is one of the most expensive problems in a premix plant precisely because it shows up late. A blend that pours perfectly at the mixer discharge can arrive at a customer's dosing station as a solid block, or arch in a silo weeks after it was made. Stopped dispensing lines, broken-up bags, rework and complaints all trace back to the same few mechanisms. This article explains those mechanisms, what a silica flow aid actually does about them, and how purchasing, QC and formulation teams can choose and validate a grade without guesswork.
Most premixes contain at least one component that does not tolerate moisture well. Choline chloride is strongly hygroscopic; trace-mineral sulphates, some organic acids and various salts behave in a similar way. Under ordinary warehouse conditions these components take up water vapour at the points where grains touch. A thin film of solution forms, dissolves a little of the surface, and when temperature or humidity shifts it recrystallises into a solid neck between the two grains. Repeat that at millions of contact points and the powder becomes a bonded mass.
Three conditions accelerate the process, and none of them is present on the day the blend is made:
This is why a funnel test on fresh blend says very little about the warehouse risk. The material that caused trouble last season usually passed its release checks.
Precipitated silica controls caking from the outside of the host particles. Fine silica particles distribute over the surface of the premix grains during mixing and act as spacers: they reduce the direct contact area between grains, and their porous structure takes up some of the free moisture that would otherwise form liquid films at contact points. With fewer and weaker contact points, the crystal bridges that form are thinner and break easily under normal handling.
It helps to separate this from the other job silica does in feed. When a liquid additive such as choline chloride solution or an organic-acid blend has to be turned into a powder, silica acts as a carrier: the liquid is drawn into the particle's internal pore network and held there. Anti-caking is the opposite geometry — the silica works on the outside of other particles. The two duties favour different particle properties, which is why the most economical answer is often a paired approach rather than one grade doing everything. Our technical note on sizing carrier load against premix flowability covers the carrier side in more detail.
When you compare candidate grades, look at the properties that govern how well a small mass fraction can cover a large host surface and resist consolidation. All of them are reported on a grade's technical data sheet (TDS); check the controlled document for the actual values rather than brochure figures.
| Property | Typical test method | Why it matters for anti-caking |
|---|---|---|
| Particle size (D50) | ISO 13320 | Finer particles spread over more host surface at the same dosage |
| Tapped density | ISO 697 | Low density means a light coating that resists collapsing under bed pressure |
| Loss on drying | ISO 787-2 | The water the flow aid itself brings into a moisture-sensitive blend |
| BET surface area | ISO 9277 | Indicates how readily the silica takes up surface moisture |
| Sieve residue | ISO 2591-1 | Oversize particles that can segregate during handling |
No single property predicts the result. A grade with excellent surface area but high moisture content can still make a hygroscopic blend worse, and a very fine grade may raise dust at the intake. The composition of the host powder matters just as much: where hygroscopic chlorides and sulphates dominate the recipe, continuous fine coverage at a low addition rate generally beats a larger dose of something coarse.
Censil puts three precipitated silica powders forward for the feed additives track, stepped by how much liquid they need to carry and how much flow is required:
The base grade for this track. It is the place to start when the liquid load is moderate and the real problem is flow and caking rather than raw absorption capacity — for example, keeping premixes and trace-mineral blends free-flowing.
The middle route. It takes more liquid than 260P while keeping the powder free-running, and is the usual answer when 260P holds the load but the flow is not there. It is positioned for vitamin, enzyme and antioxidant liquids at moderate loading.
The high-loading candidate for choline chloride, organic acids and other demanding liquids, where target loading is the binding constraint. Where a plant powders a liquid with 280P and still sees caking in the finished premix, the anti-caking duty may be better handled separately.
The product selector maps each of these grades to its application and the metrics to validate. Censil does not publish typical-value tables on the site; the TDS, COA and SDS for each grade are sent with the sample, so your evaluation runs on the documents that will govern supply.
A useful trial answers one question: will this blend still discharge cleanly after it has spent the longest realistic time in your storage chain? A sequence that works in practice:
Record the current flow aid, its addition rate, the recipe's hygroscopic components and where caking appears — silo, bag, or customer site. Without a baseline, a trial cannot show improvement.
Start at the rate the plant uses today and step down in fixed increments over successive runs. Incumbent dosages often survive for years simply because no one has run the descending series.
Check fresh-powder flow, blend uniformity across the mixer, dosing repeatability at the feeder, and bag behaviour, dust and dosing accuracy at packing.
Hold samples for four to six weeks at representative fill height, temperature and humidity, then measure flow again and look for setting, bridging or lumping. Week-one results routinely disagree with week-five results.
Test a lead grade alongside at least one alternative, each with its own COA, so the trial is a real A/B comparison rather than a single data point.
If bench and plant results diverge, the cause usually sits in one of three places: the material, the dosage, or the process step (mixing time, order of addition, or how the liquid components are introduced). Isolate them one at a time.
The fastest way to a relevant recommendation is a short, specific brief. Include the premix type and the components you consider hygroscopic, the current anti-caking agent or carrier and its addition rate, any liquid additives and their target loading, where and after how long caking appears, and the storage conditions in your warehouse and your customers'. With that, a lead grade and contrast samples can be proposed for a trial that answers the question in one round.
The controlled TDS, COA and SDS for CENSIL 260P, 265P and 280P are available through the Download Center, and our team can be reached through the contact page to discuss a trial plan. Final performance and application fit always follow the latest TDS, the lot COA and your own test results.
Related: Feed Additive Flow & Carrier · Request a sample · All news