The space looked finished on Friday. By Monday there is a gray film on the sills, the new glass and the conference table, and nobody has been in the building. Two things cause that, and often both are in play at once: some of the dust was moved rather than removed, and the rest is still sitting in reservoirs above eye level, waiting for the first slammed door or fan start to bring it down. Neither is a matter of scrubbing harder. Both come down to what fine gypsum dust does in a room once it is airborne, and once you know where it goes, a clean that did not hold is easy to diagnose.
Why does drywall dust stay in the air so long?
Because sanded joint compound breaks into particles light enough that air resistance beats gravity. The coarse fraction, the gritty part you can feel between your fingers, drops out in seconds and lands where it fell. The fine fraction behaves less like sand and more like smoke. It drifts on air currents too weak to feel: a door swinging, someone walking a corridor, a temporary heater cycling, warm air rising through an open stair. Particle size is what governs how long a solid stays suspended, a point the EPA sets out in its overview of particulate matter (opens in a new tab). Activity in the room restarts the clock every time.
That has a consequence people underestimate. A clean performed while another trade is still cutting or sanding anywhere on the same air path is a clean performed into a headwind. The crew is not failing. They are wiping surfaces that are being reloaded from the air behind them. A genuine stop in the drywall work does more for the finished result than an extra body on the crew.
Where the dust is actually hiding
Above sightline, on horizontal surfaces nobody looks at from standing height. These are the reservoirs, and a clean that skipped them was never finished, no matter how good the floors looked.
- Door tops, the top edge of every door frame, and the head of sidelights and borrowed lights
- Ceiling grid flanges, the top edge of tiles at grid intersections, and any tile that was lifted for cable or ductwork
- Supply diffusers and return grilles, including the vanes and the cavity behind them
- Light lenses, troffer housings and the top of pendant fixtures
- Sprinkler escutcheons, smoke detectors, exit sign tops and speaker rings
- The top of wall cabinets, millwork, lockers and toilet partition rails
- Window heads, blind valances, the top rail of storefront glass and every track
- Exposed structure in an open ceiling: conduit, cable tray, duct wrap, joists and beam flanges
The working order falls out of that list, since anything dislodged from height lands on whatever sits below it. That is why a finish clean runs ceiling to floor as defined passes, which the post-construction cleanup page covers in full. Several of these surfaces are not reachable from a ladder at all, which makes them a high angle dusting question instead.
Why a shop vac can spread more dust than it collects
The broom is the obvious offender, which is why filtered vacuuming replaces sweeping once anything in the building is finished. The shop vac is the problem people do not expect, because it looks like the right tool.
It was built for wood chips, offcuts and water, not for gypsum fines. General purpose filter media passes a share of the finest particles straight through, and the motor then blows that air back out at speed, frequently at floor level. Dust gets collected in one corner and jetted across the rest of the slab. The failure modes stack up from there: a bag fills and bypasses, a lid gasket sits crooked, a filter gets knocked loose and put back dusty. The machine that looks like it is cleaning is quietly recharging the air.
The fix is unglamorous. Vacuum with a sealed machine and fine filtration, keep the exhaust pointed away from cleaned surfaces, run a brush and crevice tool along every ledge instead of a floor head everywhere, and fold and turn microfiber rather than shaking it out. Compressed air and blowers have a place on a rough slab with nothing finished in the building, and no place after that.
Why wiping too early leaves a film you cannot buff out
Gypsum dust plus water is not dirty water. It is a thin slurry of joint compound, and a wet rag dragged across a dusty surface spreads a skim coat of it instead of lifting it. On glass it flashes into a haze that will not buff out dry. On new stainless it leaves a streak field. On a new hard floor it dries into a chalky bloom that appears to come back after every mopping, because each mopping is moving the same material around instead of taking it out of the room.
So dry first, wet second, and change the rinse water before it goes milky. A bucket that has turned gray is a tool for spreading compound, not for removing it. What you wipe with counts as much as when: flat weave microfiber folded into fresh faces, and a real dry pass over a surface before anything damp touches it. Which product is then safe on a newly installed finish is a separate question, and the commercial floor care page answers it material by material.
What a Western New York winter adds to the problem
A sealed building. Through the heating season the windows stay shut and the air in the space recirculates rather than exchanging, so what the build put into the air has fewer ways out of it.
Dry heating-season air also adds static. Fines cling to painted walls, glass and the underside of ledges instead of falling, which makes a room look cleaner than it is and then releases later when air movement or humidity changes. A renovation inside older building stock brings a second dust source with it, since opening up plaster, ceiling cavities and old duct runs sheds material that was settling long before the new drywall went in. And as soon as the doors are in service, boots carry in road salt and grit that mix with residual fines and grind against brand new flooring. Chloride residue is hygroscopic, so it pulls moisture out of the air and keeps a floor looking damp and hazy well after the traffic has gone.
How to tell whether the dust was removed or just moved
Check it before you accept it, and check under the lighting the owner will walk with. Run a clean dark cloth along the top of a door, the top edge of a frame and one window head. Look at the vanes of a supply diffuser and the face of a return grille. Lift the corner of one ceiling tile near a grid intersection. Push a finger along the top rail of a toilet partition. If those come back gray, the space was cleaned from the floor up and it is going to dust itself again.
The stronger test is to have the air handling run, walk the space, then come back to the flat surfaces. A fresh film after the system has moved air means a reservoir somewhere is still loaded and nothing below it will stay clean. That is a fixable problem, and it is worth finding before handover rather than after. It does not get fixed by cleaning the same table a third time.
If the first clean did not hold
When a clean does not hold, look at the order before you look at the crew. A second attempt only works if it starts at the ceiling and ends at the floor, and that holds for spaces where a clean has already been paid for once. Send the details of the space, including the square footage, the ceiling type and what has already been done, and RTJ can scope it against what the building actually needs rather than against its floor area.