How does powder selection change in anhydrous and solid-cleanser formats?
With no continuous water phase, the powder stops being a filler and becomes part of the structure, so the selection logic changes with it.
First, do not chase high oil absorption. A foundation wants absorbency for oil control; an anhydrous paste does not, because an absorbent powder pulls oil out of the continuous phase and leaves the stick hard, draggy and prone to a white smear. Grades in the 30-60 range behave better here. Second, go fine and narrow: with no water to buffer it, a coarse fraction reads directly as grittiness and as drag during application, so keep D90 under about 25 µm. Third, the treatment has to match the oil phase. SI (PDMS-type) and AS (triethoxycaprylylsilane) disperse best in oils, while a hydrophilic treatment tends to clump in this environment.
A solid cleanser adds one requirement: it has to break down quickly in water without forming lumps. Hydrophilic or semi-hydrophilic treatments, amino-acid based for example, or a water-soluble powder blended in to trigger disintegration, are the usual answers. Without one of them a syndet bar simply does not dissolve on the face.
One behaviour is easy to overlook: in an anhydrous system the powder does not settle, it compacts. The stick hardens on standing and becomes difficult to pick up. Fix it through the thixotropic structure — a little fumed silica or organoclay — rather than by raising the total powder load, which hardens the skin feel along with the stick.
A last note on stability: for an anhydrous format, test hardness and payoff after three months at 45 °C, not just whether the product separated.