Dry Ice Blasting vs. Ultrasonic Cleaning for Industrial Silicone Production Molds

The cavity still prints a dull film after three solvent wipes. That film is not shop dust. It is baked release, pigment, siloxane residue, or a smear of flash that has cooked onto the steel.

Dry ice blasting and ultrasonic cleaning both remove that film without grit. They are not interchangeable. One is an in-press, dry, line-of-sight process. The other is an off-press bath that reaches what a nozzle cannot see.

What each method is doing to the dirt

Dry ice blasting. Solid CO₂ pellets or shaved dry ice hit the surface in a compressed-air stream. The ice is soft (Mohs ~1.5-2), so it does not cut hardened tool steel the way glass bead does. Cleaning comes from kinetic impact, a local cold shock that embrittles the film, and rapid sublimation that helps lift residue. The ice disappears as gas. The only waste is the dirt you just removed, plus a CO₂ ventilation problem.

Ultrasonic cleaning. The tool, or more often its inserts, sits in a bath. High-frequency sound grows and collapses cavitation bubbles. Those micro-jets undercut residue in lettering, ejector holes, and cooling-circuit dead legs. The bath is water plus a detergent or a specified solvent. You then rinse, dry, and rust-protect. Miss the dry step and you trade residue for flash rust.

Neither method is a rubber-mold wash. These are processes for metal production tools that mold HCR or LSR. Do not dry-ice blast a soft RTV silicone mold to “refresh” it. You will erode edges and drive ice into tears. Do not drop a glued-up RTV block into an ultrasonic tank.

Choose by the residue, not by the brochure

Baked release and light polymer film on an open cavity. Dry ice, on the press, while the tool is still warm, is usually faster. Warm steel releases the film more easily. You avoid cool-down, pull, and a two-hour dry.

Pigment packed in texture, or LSR flash in a micro-vent. Dry ice is line-of-sight. It will not reliably empty a 0.02 mm vent or the back side of a slide. Ultrasonic (or a localized ultrasonic horn, if you have one) plus a brush on the pulled insert is the thorough path.

Rust, scale, or water-side sludge in cooling channels. Dry ice will not clean the inside of a gun-drilled water line. Ultrasonic with the right chemistry, or a dedicated water-circuit flush, is the job. Then dry the circuit.

Heavy carbonized rubber after a scorch event. Dry ice may need several passes. A soak plus ultrasonics plus a wooden or plastic pick on the worst islands is slower but more complete. Do not jump to bead blast on a polished LSR cavity unless you have already accepted a re-polish.

Food or medical tools. The cleaner itself becomes a contaminant. Dry ice is attractive because it leaves no detergent. You still have to capture the blasted debris and keep oil from the air line off the cavity. Ultrasonic detergents must be rinsed to a validated blank. Do not use a general rust-removing acid on a food cavity and then guess.

Downtime is the real cost

A mid-size two-plate rubber mold that can stay in the press: dry ice is often tens of minutes of blasting plus vacuuming residue, versus a half shift to pull, tank, rinse, dry, and reset. That is why tire and rubber shops adopted dry ice for scheduled PM.

A 16-cavity LSR cold-runner tool with valve gates, slides, and a cold deck: blasting the parting face in place does not clean the needle guides or the water jacket. Plan an insert-level ultrasonic on the PM that actually restores fill balance. Mixing the two is normal: dry ice between PMs, ultrasonic when the tool is already apart for a valve rebuild.

Damage modes people skip

Dry ice is “non-abrasive” relative to sand. It is not harmless. Too high a pressure, too close a standoff, or a concentrated pass on a sharp polished edge can frost a SPI A-2 finish or peen a soft aluminum prototype tool. Start conservative (many shops begin around 4-5.5 bar / 60-80 psi and a generous standoff) and test on a runner block, not on the logo. Moisture in the air line makes mud. CO₂ in a closed cell needs ventilation; the occupational limit is a concentration problem, not a smell problem.

Ultrasonic cavitation can etch a high polish if the chemistry is too aggressive or the cycle is too long. Dissimilar metals in one bath invite plating attack. Water in ejector holes rusts overnight. Heat-treat labels and witness marks can fade in a hot alkaline tank. Torque the tool back to spec after a full strip; ultrasonics do not realign a plate you dropped in the tank.

A practical split for a silicone press bay

• Daily/shift: dry-ice or wipe the parting face if you are printing haze and the geometry is open.
• Weekly or when vents pack: pull the vented inserts, ultrasonic, rinse, dry, inspect under raking light.
• After a material or color change: dry-ice the face, then qualify the first shots for contamination. Pigment ghosts are not always gone because the steel looks shiny.
• Never: grit blast a cosmetic LSR cavity to save 20 minutes. Re-polish costs more than a proper clean.

If you only buy one process, buy the one that matches how often you can pull the tool. High-uptime HCR and LSR cells lean dry ice. Precision insert tools with blind features lean ultrasonic, with dry ice as the in-press assist. Write both into the PM sheet rather than arguing which one “wins.”

FAQ

Can I dry-ice blast a soft RTV silicone mold to refresh it?

No. These methods are for metal production tools that mold HCR or LSR. Ice will erode RTV edges and drive into tears. Do not drop a glued RTV block into an ultrasonic tank either.

When is ultrasonic the better clean than dry ice?

When residue hides from line-of-sight: pigment packed in texture, LSR flash in a micro-vent, the back of a slide, needle guides, or sludge in a gun-drilled water line. Dry ice will not empty a cooling circuit. Pull inserts, tank, rinse, and dry, or you trade residue for flash rust.

Why is downtime the real cost, not the brochure method?

A two-plate tool that can stay in the press may take tens of minutes of warm dry-ice versus a half shift to pull, tank, rinse, dry, and reset. A 16-cavity LSR cold-runner with valves still needs insert-level ultrasonic when you are already apart. Mix them: dry ice between PMs, ultrasonic on rebuilds.

Can dry ice frost a high SPI polish even though it is "non-abrasive"?

Yes, relative to sand. Too much pressure, too close, or a concentrated pass on a sharp polished edge can frost A-2 or peen aluminum. Start conservative on a runner block, not on the logo. Wet air makes mud; CO2 in a closed cell is a concentration problem.

What split belongs on a silicone press PM sheet?

Daily or shift: dry-ice or wipe an open parting face if you are printing haze. When vents pack: pull, ultrasonic, inspect under raking light. After a color change: dry-ice then qualify first shots; shiny steel can still ghost pigment. Never grit-blast a cosmetic LSR cavity to save minutes.