The drum is opened in the customer's warehouse after five weeks at sea, and the powder inside has become one piece. A scoop will not go in. The QA person takes a chisel to the top layer, and a lump the size of a fist comes out whole, with the shape of the liner pressed into it.
The material passed every test at the factory and nothing was added in transit. The powder crossed its glass transition.
The water was already inside
A spray-dried extract is mostly an amorphous solid. The drying step removed the water fast enough that the sugars, organic acids and polysaccharides in the extract did not crystallise; they froze in place as a glass, with a small amount of water trapped in it.
A glass is stable only below its glass transition temperature. Above that temperature the same material becomes rubbery: it deforms under its own weight, and where two particles touch, they fuse. A lump is a weld at the contact points rather than a sticky surface.
Water is the reason the transition temperature moves. Water plasticises an amorphous solid, and each additional percent of moisture lowers the transition temperature substantially. A powder at 3% moisture and a powder at 6% moisture are not the same material with different water contents; at warehouse temperature, one may be a glass and the other may already be rubber.
Three numbers describe the powder, and one is usually missing
Loss on drying is on nearly every certificate. Water activity is on almost none, and it is the figure that predicts what happens inside a sealed drum. Loss on drying measures how much water is present. Water activity measures how available it is, and availability is what drives the fusion at a particle contact.
Particle size is the second. In the trade, 80 mesh is about 180 micrometres and 100 mesh about 150. A finer powder dissolves faster and cakes faster, because it has more contact points per gram and more surface for the same moisture to sit on. A customer who asks for a finer mesh for mouthfeel has also asked for a shorter shelf life in the drum.
Bulk density is the third, and it is usually reported as one number when it should be two: loose and tapped. The ratio between them, the Hausner ratio, describes how much the powder compacts under its own handling. A powder with a high ratio is already partway to a brick before it leaves the factory.

A drum is a small climate
The headspace above the powder is not empty. It holds air, and that air holds whatever moisture the powder gives up at the temperature of the day.
Temperature swings move water back and forth. A container on a wharf warms during the day and cools at night; the air in the headspace loses the ability to hold water as it cools, and the water condenses on the coolest surface, which is the liner, and then runs down or is reabsorbed unevenly. The top layer of the drum is the layer that sees this first.
Loading warm powder into a cool container is the classic version of the same failure. The cargo warms the air in the container, the warm air carries water, the container walls are cold, and the water condenses there and drips. The trade calls it container sweat. At the factory the drum was fine; at the destination the top of the drum is wet and the rest of it has been curing for five weeks.
Stacking finishes the job. The bottom drum carries the weight of everything above it, and compression accelerates the fusion that moisture started.
The carrier decides the transition
Two suppliers can quote the same extract ratio, the same assay and the same carrier percentage, and ship powders that behave differently in the same container. The difference is often the carrier's dextrose equivalent.
Maltodextrin is a family, not a substance. A low-DE maltodextrin is a longer chain molecule, with a higher glass transition temperature and less affinity for water. A high-DE maltodextrin is closer to a syrup solid: shorter chains, lower transition temperature, more hygroscopic. The same weight of the second one will pull more water out of the air and lose its glass at a lower temperature, which is exactly the combination that produces a brick.
None of that appears on a certificate that says maltodextrin. It is worth one line in the specification, because it is the difference between a powder that survives a summer crossing and one that does not.
The other route to the same failure is crystallisation. An amorphous powder is not at equilibrium; it wants to crystallise, and it does so slowly, faster at higher humidity and temperature. When part of it crystallises, the water that was trapped in the glass is released into the rest of the powder, which then plasticises, and the drum sets. The lump that forms this way is denser and harder than the one formed by simple moisture uptake, and it is the one that arrives after a long voyage rather than a short one.

What stops it
The order of operations matters more than any single control.
Dry to a moisture that keeps the powder below its transition temperature at the highest temperature it will meet in a container, then let it cool before packing. Packing warm is how a correct moisture becomes an incorrect one, because the powder that goes in warm will release water into the headspace as it cools.
The liner is the second control. A polyethylene liner inside a fibre drum, folded and tied, with a desiccant sachet in the headspace for long sea freight, is the ordinary arrangement. A foil-only bag is not; it is an excellent barrier in both directions, which means it keeps the water in as effectively as it keeps it out.
An anticaking agent is the third option, and it works by coating the particle surface so that two particles cannot weld. Silicon dioxide is the usual one. It also adds mass that is not extract, which changes the assay arithmetic and has to appear on the label, and it does not fix a powder that was packed too wet.
The test that predicts it
A storage trial answers the question better than a certificate does. Seal samples in the same liner the customer will receive, hold them at 40 °C and 75% relative humidity for a defined period, and check flow with a test rather than with an opinion.
A compression test is the second: powder in a cylinder under a known load for a known time, then the force needed to break the column. It is a crude number and it reproduces the condition that actually causes the lump, which is weight.
And the cheapest test is the one that matches the shipment. Fill a drum, seal it, leave it at ambient for the length of the sea freight, and open it at the other end. A powder that survives that has a specification that probably works.
What to write in the specification
Six lines, and the first two decide the other four.
- Moisture and water activity. Both, with the method, because loss on drying alone does not describe a glass.
- Particle size. Mesh, with the sieve standard, because 80 mesh means different things in different countries.
- Bulk density, loose and tapped. Two figures, plus the ratio, so that compaction can be compared between lots.
- The flow or caking test. With the temperature, the humidity, the duration and the liner.
- The packaging. Liner material, seal, desiccant, and headspace, stated rather than assumed.
- Storage and transport conditions. A temperature range, and a note on whether the drum may be loaded warm.

The stability in formulation article covers what the same moisture does to the marker inside a finished product, the production lines page shows where drying and packing sit, and the contact page reaches the people who set the packing specification.
Ding Shaohua, Foreign Trade Manager, takes enquiries for this line. The same person answers on WhatsApp, TikTok, LinkedIn and Facebook.
FAQ
Does caking mean the powder was poor quality?
Not necessarily, and the distinction matters for a claim. Caking is a relationship between moisture, temperature and time, so a material that met its specification at the factory can cake in a container that spent three weeks on a hot wharf. The powder was not adulterated and nothing was added; the glass passed its transition temperature. A supplier can still be at fault, but the fault is usually in the packing moisture or the loading temperature rather than in the extract.
Can the lumps be broken up and used?
Mechanically yes, and the result is a different powder. A lump is a weld, so breaking it produces hard fragments with fresh surfaces, a wider particle size distribution and a slower, less even dissolution. In a capsule it may not matter. In a drink mix it usually does, because the fragments settle differently from the original particles.
Does a desiccant sachet solve the problem?
It manages the headspace, not the particles. A desiccant will absorb the water that the powder releases as it cools, which helps, and it cannot remove the water already inside the glass. A drum packed at 7% moisture with a desiccant is still a drum packed at 7% moisture.
Why does the same product cake in one shipment and not in another?
Because the two shipments did not have the same temperature history. Loading temperature, the position of the container on the vessel, the route, the season and the time on the wharf all change the maximum temperature the powder sees, and the transition temperature is what decides whether it fuses. That is why the specification has to describe transport conditions rather than only the material.
