Mineral Deposit and Etch Treatment Technique

Why this matters

By the time you are running a buffer, the decision is already made: the diagnostic said etch or chemistry-resistant deposit, and the customer signed off on polishing. What happens next is a mechanical process that removes a real layer of glass with a spinning pad, and it fails in exactly two directions. Run it too hot or too dry and you put a thermal stress crack through the pane you were hired to save. Run it timid or uneven and you leave a wavy optical distortion that reads worse under evening light than the haze did in the morning. This article is the equipment and the hand-technique that sits between those two failures: how to set up the polisher, read the glass while you work it, and know when the pane has given you all it is going to give. The diagnostic that gets you here (deposit versus etch, the wet test, when to walk away from tempered or coated glass) belongs to a sibling reference card; do not re-run that decision here, confirm it was already made.

Precondition: confirm before the pad touches glass

Two facts have to be true before you plug in a polisher, and both are checks you perform, not assumptions you carry over from the phone quote.

The glass is not tempered. Tempered glass carries internal stress locked in during manufacture, and mechanical polishing concentrates heat and removes surface material unevenly enough to trigger spontaneous breakage in a pane that was already close to its stress limit. Look for the manufacturer's etch mark in a corner (a small logo or a series of numbers and letters, usually bottom corner). If you find it, the pane is tempered and mechanical polishing does not proceed; hand the job back to the chemistry-and-referral path the deposit/etch diagnostic already lays out. If you cannot find a mark, do not treat "no mark found" as "not tempered." Ask whether the glass type is documented (a spec sheet, a permit set, a prior restoration record); if nobody can confirm it, treat the pane as tempered and do not polish it.

The etch is confirmed, not assumed. A pane that still responds to the chemistry sequence is not a polishing job yet. Polishing removes real glass and is not reversible; running it on a pane that would have cleared with one more chemistry cycle is wasted glass thickness and wasted time. If the wet test or the diagnostic cycle from the chemistry procedure was not documented for this specific pane, run it before the buffer comes out.

Equipment and consumables

• Variable-speed polisher, rated for glass work, speed control down to roughly 1,000 to 1,500 RPM. A fixed-speed buffer built for automotive paint runs too hot for glass and cannot be throttled back when the pane starts to warm.
• Felt or wool polishing pads, one per grit stage, bagged and labeled by stage. A pad that has picked up grit from a coarser stage and gets reused on a finishing pass scratches instead of polishing.
• Cerium oxide compound, mixed to a thin slurry in a squeeze or spray bottle. Powder applied dry burns the glass and loads the pad; always mix wet.
• A continuous water feed: a spray bottle with an assistant, a drip line clipped above the work area, or a wet sponge within reach for constant re-wetting. Not optional; see the heat-buildup stop rule below.
• Masking tape and plastic sheeting for the frame, gasket, and any adjacent glass.

Setup and masking

1. Mask the frame and gasket line with tape run flush to the glass edge, then a second strip overlapping it about half an inch onto the frame. Cerium slurry dried on anodized aluminum or vinyl leaves a chalky residue that is its own cleanup job, and slurry under a rubber gasket can work into the weep channel. On a painted or decorative finish, add plastic sheeting over the tape rather than relying on tape alone.
2. Mask the adjacent pane in a multi-lite unit or storefront run. A slurry drip on clean glass one pane over is easy to miss until it dries.
3. Confirm the power circuit is GFCI-protected. You are running a corded electric tool inches from a continuous water feed. Plug into a GFCI-protected outlet, or use a portable in-line GFCI adapter ahead of the polisher's cord. If neither is available on site, the job does not proceed until one is; a standard outlet next to running water is not an acceptable substitute, and if the GFCI trips during the work, that is the device doing its job, not a fault to bypass.
4. Inspect the cord and pad backing for cuts, fraying, or a pad that is coming loose from its backing plate before the first pass. A pad that separates mid-polish under load can whip.
5. Put on safety glasses or goggles before the first pass. A pad running at working speed flings fine slurry droplets back toward the operator, and cerium oxide is a mild abrasive that irritates on eye contact even though it is not a caustic chemical. If slurry does reach an eye despite the glasses, flush immediately with clean water for several minutes and stop the work until vision is clear and free of grit.

The polishing procedure

1. Start the polisher off the glass, bring it up to the working speed setting away from the pane, then lower it onto a pre-wetted section. Starting a spinning pad already in contact with dry glass is what puts the first heat spike into the pane before you have a feel for it.
2. Work in overlapping passes across a section roughly the size of a sheet of paper, keeping the pad flat and moving continuously. A pad held stationary in one spot for even a few seconds concentrates heat and removes glass unevenly, which is exactly the wavy distortion you are trying to avoid.
3. Re-wet before the slurry starts to dry. Dried slurry turns abrasive and stops lubricating, so the pad starts grinding instead of polishing. If the slurry on the glass has gone from wet-and-milky to a dry film in under 60 seconds, your water feed is too slow; increase it before continuing.
4. Check glass temperature by touch every 60 to 90 seconds of continuous work on one section. Press the back of your bare hand (not fingertips, which tolerate more heat before you notice) flat against the glass just outside the area you are working. If you cannot hold it there comfortably for a slow three-count, the glass is too hot to continue on that spot. Lift the pad off, let the section sit for a full 5 minutes before returning to it, and increase your water feed rate when you resume. This is the stop rule for thermal buildup, and it is not optional at any point in the procedure, including inside the worked example below.
5. Check optical clarity every two to three minutes by wiping the section with a clean microfiber and viewing it at a low angle in direct or raking light, wet and dry. Deposit clears quickly; genuine etch clears slowly and unevenly, and this is where patience matters more than pressure.
6. Move to the next section only once the current one has stopped visibly improving between checks, not once it has reached perfect clarity. A pane with true sub-surface etch plateaus at "much improved" rather than "invisible," and continuing to grind past that plateau removes glass thickness for a diminishing return while raising the crack risk with every additional pass.
7. When a section plateaus, mark it (a small piece of tape on the frame, not the glass) and move on. Do not spend more than about 15 minutes of continuous polishing time on one section in one visit. If it has not plateaued by then, stop, let it cool fully, and reassess whether this pane belongs on the "improves but will not fully restore" list already disclosed to the customer.

Reading the glass while you work

Residual deposit clears fast under the pad, usually within the first minute on a section, and evenly across the whole area at once. True etch clears slowly, unevenly (shallow parts of the pattern first), and the rate visibly slows the longer you work a section; that slowing is the plateau signal in step 6, not a sign of bad technique. If a section shows zero improvement after 90 seconds of correct technique, stop: either the compound has loaded with debris and needs replacing, or the etch is deeper than cerium oxide addresses in one visit and belongs in the referral conversation, not further grinding.

Finishing and the loop you have to close

1. Final rinse the polished section thoroughly, removing all slurry residue before it dries on the glass or migrates to the masking edge.
2. Remove the masking, working from the glass edge outward so you are not dragging a slurry-loaded tape edge back across clean glass.
3. Inspect the frame and gasket you masked, not just the glass. This closes the isolation set up in Setup and Masking: confirm no slurry worked past the tape onto the frame or into the gasket channel, and clean any that did before it dries there. Skipping this check turns dried cerium residue on anodized aluminum into a second job you did not quote.
4. Wipe and dry with the standard squeegee-and-microfiber pass, then inspect under the same lighting used for the baseline photo, and photograph the finished section for the record.

Worked example

A storefront panel with three lower-frame lites shows haze the diagnostic already confirmed as etch on two of the three, with the wet test flagging the bottom four inches of the third as etch and the rest of that pane as residual deposit chemistry did not fully clear.

Setup: Both fully-etched panes and the bottom strip of the third are masked at the frame. The circuit is confirmed GFCI-protected through a portable adapter, since the storefront's exterior outlet is not itself GFCI. Polisher set to 1,200 RPM with a fresh felt pad, cerium slurry mixed thin in a spray bottle, second bottle of plain water for the continuous feed.

Pane one (fully etched): The first temperature check at 90 seconds is comfortable, so work continues. At 3 minutes the haze has thinned but not cleared, consistent with true etch. At 8 minutes the temperature check fails; work stops per the stop rule, the section sits 5 minutes, water feed is increased, and work resumes. By 13 minutes two consecutive clarity checks 2 minutes apart show no further change: the section has plateaued. Work stops there rather than continuing toward the 15-minute ceiling, and the crew marks it and moves on.

Pane three (mixed): The top two-thirds, correctly diagnosed as residual deposit, clears within 90 seconds per section, matching the fast-and-even signature. The masked bottom strip, the confirmed etch portion, follows the same slow-and-uneven path as pane one and plateaus around 11 minutes, improved but with a faint shadow still visible at a raking angle.

Finish: All masking is removed and the frame and gasket on all three panes are inspected; a small bead of dried slurry is found at one corner of pane one's gasket channel and cleaned out with a plastic pick before it can migrate into the weep path. Final photos match the baseline angle and light. The customer walk-through explains, pane by pane, which areas fully cleared and which improved but retain a shadow, matching the disclosure signed before work began.

Verify you got it right

• Every masked frame and gasket has been inspected and cleaned of any migrated slurry, not just the glass.
• Every section that plateaued was marked and left alone rather than pushed past 15 minutes chasing full clarity.
• No section was worked while it failed the back-of-hand temperature check.
• The finished result, photographed at the same angle and light as the baseline, matches what was verbally disclosed before work started: which panes cleared fully and which improved but retained a shadow.
• The GFCI protection was in place for the entire polishing operation, confirmed working (test button, if the adapter has one) before the first pass.

References

• See related: hard-water-etch-vs-deposit-decision (diagnostic, wet test, chemistry ranking, and when to walk away from a job entirely).
• See related: calcium-stain-removal-chemistry (the chemical removal sequence that precedes mechanical polishing).
• ASTM C1036, Standard Specification for Flat Glass, surface defect classification.
• ANSI Z97.1, Glazing Materials Safety Performance, tempered glass identification.
• IWCA International Window Cleaning Association restoration technical guidance.