How Fillers (Silica vs. Quartz) Affect Silicone Mold Shrinkage and Viscosity Rates

Fumed silica and ground quartz are both silicon dioxide. They do not do the same job in a silicone mold compound.

Fumed (pyrogenic) silica is a high-surface amorphous powder, the classic reinforcing and thixotropic filler in platinum RTV, tin RTV, HCR, and many LSR grades. Ground quartz (silica flour, crystalline SiO2) is an extending filler: more volume for less viscosity penalty, more abrasion, and a shrink lever because mineral does not shrink with the polymer. Stirring “some silica” into a platinum jacket is not a spec. Untreated fume can gel a pour. Quartz can sand a mill. Neither loading belongs as a shop-invented weight percent. RTV pours, millable HCR, and injectable LSR do not share one packing limit.

Two SiO2 powders, two mechanisms

Fumed silica is made in a flame. Particles are small, chained, and covered in silanol groups. In a polysiloxane they build a temporary network (hydrogen bonding, polymer bridging). That is why a little untreated fumed silica turns a thin RTV into a non-slump paste, and why a treated grade (HMDS or silane) can reinforce with a milder viscosity spike. Tear (ASTM D624) and green strength are why compounders use it. Clarity can survive better than with quartz flour, which is why transparent mold rubbers are not “quartz-filled clear.”

Precipitated silica is also amorphous and reinforcing, with a different structure and moisture story. It shows up in millable rubber more than in a hobby RTV cup. Treat it as a cousin of fumed silica, not as quartz.

Quartz flour is crushed crystalline silica. Surface area is low compared with fume. Viscosity rises mainly because you are replacing liquid polymer with hard particles (volume fraction, packing, particle shape). You can put more quartz in before the mix refuses to pour or inject. The cured stock is stiffer, heavier, more opaque, usually lower elongation, and harder on equipment. Respirable crystalline silica is a handling and housekeeping issue in the mill room. Amorphous fumed silica is a different dust conversation. Do not bag them as one “silica.”

Fused silica (amorphous crushed glass) sometimes appears as an extender. It is still not fumed silica. If the TDS says only “silica,” ask which.

Viscosity is a network problem, then a packing problem

RTV pours live at low shear. You do not get LSR’s shear-thinning gift from a paddle. Untreated fumed silica in a cup is a thixotrope: it looks thick at rest, may yield if you stir, then rebuilds. That is useful on a vertical master and hostile in a deep, fine cavity if the yield stress keeps the rubber off a letter. Surface-treated fume is how many pourable grades stay pourable and still tear. Datasheet cPs numbers are snapshots at one spindle speed (see the viscosity article in this cluster). A fumed-silica paste and a quartz-extended paste can share a cup number and fill differently because one is a network and one is a slurry.

Quartz at extending levels makes a heavier, more Newtonian slurry until packing gets high, then it is a paste with air in the particle bed. Degassing that slurry is slower. Bubbles do not rise. Vacuum still works if the pot life allows.

LSR injection is different. Gate shear can thin a filled grade, then the network rebuilds in the cavity. Quartz-filled LSR still wants larger gates, more pressure, and abrasion-aware steel. Fumed-silica LSR is the usual tear-and-clarity path; viscosity is managed by the supplier’s treatment and polymer viscosity, not by a shop adding powder at the press.

HCR on a two-roll mill is the natural home for high mineral extenders. Shear wets quartz. Fumed silica on a mill is a different technique (it dusts, it bites, it needs incorporation time). Do not assume a mill recipe pours as RTV.

Abrasion is the quartz tax

Quartz is hard and often angular. It wears:

• Two-roll mills, sigma blades, and LSR screws and barrels
• Aluminum prototype cavities and soft coatings
• Shut-offs if flash is carrying filler
• The master, if you are pouring a heavily quartz-extended RTV jacket over a soft pattern and then grinding on it during demold

Fumed silica is not gentle in a dusty mill room, but it is not a crystalline cutting grit at the same particle scale. If tool life or a polished cavity is the constraint, quartz-extended production LSR/HCR is a steel-and-maintenance conversation, not a resin-tool conversation.

A quartz-filled mold rubber also changes the jacket: higher hardness, lower elongation, less forgiveness on undercuts, more weight in a large glove. That is sometimes the point (a rigid mother-mold face, a dimensionally lazy block). It is the wrong powder if you needed a soft platinum skin for a deep undercut.

Abrasive castings wearing an unfilled jacket are a different wear path. Here the filler is inside the silicone.

Shrinkage follows filler volume, not a slogan

The polymer network shrinks. Quartz and silica particles do not. Raising mineral volume fraction generally lowers linear shrink of heat-cured HCR and LSR. That is why extending fillers show up in dimensional grades. Quartz is the usual volume lever because you can load more of it before viscosity dies. Fumed silica, used at reinforcing levels, is there for tear and rheology. Its effect on shrink is real but secondary. Do not treat Shore A as a shrink calculator, and do not invent a quartz percent that “takes 3% down to 1%.” Ask the compounder for a shrink range on that grade in that cure, then coupon it (ISO 3302-1 still cares about process class; some silicones’ shrink makes tight classes hard).

Process split:

• Platinum RTV jackets already sit near 0.1% or less after a complete cure. Dumping quartz into a platinum mold rubber to “kill shrink” is usually about stiffness and cost, not about a CAD scale factor. If the casting resin shrinks, compensate the master for the resin, not with a fake mold-rubber percent.
• Tin-cure RTV still loses condensate and keeps moving in storage. Filler does not freeze that time-dependent shrink. A quartz-extended tin block is still a tin block.
• LSR and HCR steel tools are where filler-driven shrink actually belongs in the cavity scale. More extender, often less shrink, also more anisotropy risk if particles orient, and more post-cure movement if volatiles and bound water differ. Measure the filled grade, not the unfilled cousin.

Constraint still wins. Rubber bonded to a metal insert cannot shrink the way a free lip can. Thick and thin sections do not always agree. None of that is fixed by guessing a silica number.

Which powder belongs in which shop job

Use fumed (or precipitated) silica when the compound must tear, hang on a vertical master, or stay closer to unfilled mechanicals. That is most quality RTV mold rubbers and tear-rated LSR.

Use quartz flour when the compounder is extending a heat-cured grade for hardness, lower shrink, cost, or a heavier section, and the process can take the abrasion (mill or production LSR steel). Prototype aluminum and printed tools are the wrong home.

Do not mix a bag of construction-grade silica sand into platinum RTV. Particle size, moisture, and crystalline content are uncontrolled. Moisture is a tin-cure variable and a void source. Some dirty minerals inhibit platinum. A compounding house wets, treats, and screens. A bucket does not.

Clarity, food, and medical are separate gates. Quartz-extended stock is not a food-contact argument. FDA 21 CFR 177.2600 and ISO 10993 attach to the finished formulation, not to the word silica. Optical molds want treated fume or unfilled gum, not flour.

Write the filler as a function, not as a percent

“Compound: [platinum RTV / tin RTV / platinum LSR / peroxide or addition HCR]. Filler family: reinforcing amorphous silica and/or crystalline quartz extender as supplied in the named grade. Do not add powder at the press unless this is a documented let-down. Targets: shrink range from supplier coupon after [cure/post-cure], viscosity or rheology method as on the TDS (not a lone cPs), hardness, ASTM D624 tear, and abrasion/tool-steel notes if quartz-extended. Process: pour / mill / LIM. Ban shop percents.”

Refuse a PO that says “add silica to reduce shrink” with no family, no process, and no coupon. That sentence has wrecked more cavities than it has saved.

FAQ

Can I add quartz flour to platinum RTV to make the mold shrink less?

Platinum RTV mold rubber is already in the near-zero shrink class after a proper cure. Quartz will mainly raise hardness, density, and abrasion, and it may ruin pourability. If the casting shrinks, resize the master for that resin. Do not invent a quartz percent to “correct” a platinum jacket.

Why does a little fumed silica thicken a mix more than a lot of quartz?

Fumed silica builds a silanol particle network. Quartz is a low-surface extender; viscosity follows volume fraction and packing. Untreated fume is a thixotrope. Treated fume is milder. A cup cPs number can match while the two pastes fill a letter differently.

Is precipitated silica the same as quartz flour?

No. Precipitated silica is amorphous and reinforcing, closer to fumed silica than to crystalline quartz. Quartz flour is crushed crystalline SiO2, used to extend, stiffen, and cut shrink at higher mineral volume. The mill-room dust controls are not the same.

Will quartz-filled LSR wreck an aluminum prototype tool?

It can. Quartz is a hard, angular mineral. Production quartz-extended LSR belongs in steel with abrasion in the maintenance plan. Printed resin tools and soft aluminum tryouts are the wrong cavity for that slurry. Fumed-silica grades are still not a license to ignore wear, but they are not the same grit.

Does more filler always mean less shrink on a tin-cure jacket?

Not in a way you can CAD. Tin-cure keeps moving as condensate leaves. Mineral volume can change how a block feels on day two; it does not turn tin into platinum. For heat-cured LSR/HCR, extending fillers are a real shrink lever, and you still coupon the named grade rather than guessing a percent.