Casting sand is recycled inside the foundry: the same sand is cleaned, cooled, topped up with new sand and binder and molded again many times, and only the share that wears out leaves the plant, for a landfill or for beneficial reuse in construction and manufactured soil. Each pass coats the grains with burnt carbon, residual binder and dust that has to come off before the sand goes back into molds.
A foundry needs about a ton of sand for every ton of iron or steel castings it pours, so the cost of new sand and of disposal pushes almost every plant to reuse it. EU rules set the benchmark at more than 90% reuse for iron foundries and more than 80% for steel foundries.
In 2007, about 10 million tons of spent foundry sand left US foundries, and only 26% of it went to beneficial use; most of the rest was landfilled.
Step 1: Separate the Sand From the Castings at Shakeout
At shakeout the mold is broken apart on a vibrating grate or in a rotary drum: the sand falls through, the castings and flasks stay on top, and the sand is carried off to be reprocessed.
Surplus sand beyond what the loop needs is drawn off after the shakeout screen or from the storage silos and leaves the plant as spent sand.
Step 2: Remove Tramp Metal, Core Lumps and Oversize
Magnets come first and pull out flash, pouring drops, sprue pieces and shot.
The remaining lumps are then broken down and the sand is screened, so oversize lumps and debris are rejected and only grains in the specified size range go on. Green sand lumps have to be broken without crushing the grains.
Uncured cores and cores that did not break at shakeout should be picked out before the sand goes back.
A plant that runs incompatible binder systems should keep their sands apart at shakeout so each can be reclaimed on its own.
Step 3: Cool the Returned Sand
Green sand should be cooled to about 40–45 °C (104–113 °F) before it is mulled again, and reclaimed no-bake sand has to come back to the binder’s working range, 15–25 °C (59–77 °F) for most acid-cured and urethane systems.
Fluidized-bed and evaporative coolers bring the temperature down and remove excess fines at the same time.
For chemically bonded sand, the temperature is held as constant as possible. The EU’s reference document on best available techniques for foundries gives a separate range for each binder system.
Step 4: Strip the Spent Binder From the Grains
Green sand only needs reconditioning, because its clay bond is restored by adding back water and bentonite, while chemically bonded sand needs reclamation: a mechanical, thermal or wet process that removes the residual binder from each grain.
When to Use Mechanical Reclamation
Mechanical reclamation is the standard route for no-bake resin sand, such as furan and phenolic urethane, and it also handles cold-box, hot-box and shell sand once pouring has left their binder brittle enough to break off.
Furan no-bake sand comes back from a mechanical reclaimer at more than 90% and goes into the same molding line with small additions of new sand. The process does not clean each grain completely, so the reclaim mostly goes back into sand with the same binder it came from.
When to Use Thermal Reclamation
Thermal reclamation is used on chemically bonded sand, such as cold-box and shell core sand and resin no-bake sand, and on mixed sand when the share of core sand in it is high enough.
Thermally reclaimed grain comes out close to new sand, and German plants reclaiming cold-box and furan sand report about 95% recovery, with the sand going back into cores and molds.
Bentonite is dead-burned in the furnace, so a stream that contains green sand needs efficient mechanical cleaning alongside it to remove the clay. Binders that leave inorganic salts on the grain, such as alkaline phenolic and sodium silicate, can sinter the grains, so thermal reclamation may not suit them.
When to Use Wet Reclamation
Wet reclamation is used only on green sand and sodium silicate (CO2- or ester-cured) sand, and few foundries choose it.
Recovery is only up to about two thirds of the sand, and each ton needs on the order of 2,000 gallons of water, which has to be treated before it is reused or discharged.
By 1999, US equipment makers had stopped building wet systems. Wet reclamation suits a central reclamation plant, because it copes with variable incoming sand better than mechanical or thermal units.
Step 5: Screen and Classify the Grains
Reclaimed sand is screened to take out the remaining lumps and classified to remove the fines the reclaimer knocked off, so its grain distribution comes back to what the molds were set up for. Most molding sand runs at an AFS grain fineness number of 50–60.
Reclamation keeps wearing the grains, so the grain fineness of the system sand has to be measured and corrected over time. In green sand, plants keep the fines removed in the cooler to a minimum, because those fines carry bentonite that can be reactivated.
Step 6: Add New Sand and Binder
Fresh binder goes into the sand at every mix, and new sand replaces the sand lost or drawn off from the loop, in an amount no standard sets. In green sand molding without cores, European plants renew 2–5% of the sand on average.
Cores add sand to the system every time they break up at shakeout, and some green sand foundries add no new sand beyond their core sand. Core sand is not free make-up, though. When a ductile iron foundry stopped adding new sand and relied on its core sand, the coarser core sand dropped the system’s grain fineness from 60 to 54 and made the sand more friable.
Green sand gets water and bentonite added in the muller at every mix. Reclaimed no-bake sand gets a fresh charge of resin and catalyst, for furan about 0.9–2.0% resin and 0.2–1% acid catalyst by weight of sand.
Step 7: Test the Reclaimed Sand Before It Goes Back Into Molds
Reclaimed resin sand is tested for what is left on the grain and for grain size, against limits that depend on the binder system and the reclamation route.
- Loss on ignition (LOI): the weight burned off at high temperature, a measure of the binder and carbon still on the grain. In one published trial, thermally reclaimed resin sand stayed under a 0.3% target taken from the German VDG guideline for reclaimed sand; mechanically reclaimed sand has been judged against an allowable LOI and a grain size close to new sand.
- Grain size and fines: the AFS grain fineness number and fines content, checked against the sand the system was built on.
- Acid demand value: for acid-cured furan and phenolic sand, how much acid the sand itself reacts with; plants set the limit against the range their binder supplier gives.
- pH and conductivity: for alkaline phenolic sand.
- Sulphur: for furan sand cured with sulphonic acid, above all when the castings are ductile iron.
Green system sand is not passed or failed on LOI. It is controlled by daily tests on the prepared sand, with compactability and moisture as the primary checks and green compression strength, permeability and methylene blue clay alongside them.
Step 8: Send Spent Sand to Beneficial Reuse
Sand the foundry can no longer reclaim leaves the plant as spent foundry sand and goes either to a landfill or, where the state allows, to beneficial use. In 2002, 98% of US foundry sand was classed as non-hazardous, so disposal is normally in an ordinary non-hazardous landfill.
What Spent Foundry Sand Can Be Used For
Most reused spent foundry sand goes into construction: fill, asphalt and concrete took the largest tonnages in the 2007 AFS survey, ahead of manufactured soil and road construction.
EPA’s 2014 risk assessment evaluated silica sand from iron, steel and aluminum foundries in manufactured soil and extended the finding to potting media and road subbase. Asphalt, concrete and fill are covered by state programs and engineering specifications.
As an aggregate, spent sand can carry enough organic residue, especially from sea coal in green sand, to rule it out of Portland cement concrete, and some spent sands are too fine for fine-aggregate specifications.
Which Spent Sands Qualify
Silica-based spent sand from iron, steel and aluminum foundries is the sand regulators accept for reuse; sand from leaded brass and bronze foundries is often regulated as hazardous waste because it can carry lead, cadmium and other heavy metals.
Under the federal TCLP leach test, a sand whose extract reaches 5.0 mg/L lead or 1.0 mg/L cadmium, or any other listed limit, is hazardous waste and cannot go to beneficial use.
Wisconsin’s higher grade of reusable foundry sand includes resin-bonded molding and core sand as long as it passes the state’s leach limits, but excludes uncured chemically bonded sand.
Other foundry wastes, such as baghouse dust, slag, broken cores and core-room sweepings, do not count as spent foundry sand. Sand mixed with them falls outside EPA’s findings.
Approval and Testing for Off-Site Reuse
The state approves off-site reuse, not EPA: the foundry first proves the sand is non-hazardous, then meets the state’s own leach and composition limits and the engineering specification for the intended use.
In EPA’s 2002 review of 18 states, every state required a first characterization showing the sand is non-hazardous, most often by TCLP, and five also required an analysis of the sand’s total composition. The two tests do not substitute for each other.
What the state grants depends on the use and the state:
- Exemption: the sand is exempt from waste rules for the listed use and can be reused freely.
- Prior notification: the foundry notifies the state before reuse, sometimes with a review and approval.
- General permit: one permit covers a type of reuse for any qualified applicant.
- Case-by-case approval: a determination for one sand and one use, as Minnesota has required for foundry sand.
Wisconsin requires both a bulk analysis and an ASTM D3987 water leach test, and repeats the characterization at least every four years or sooner if the process changes.
Before shipping, the sand is screened and run past magnets to clear debris and metal, and stockpiled in lots large enough to make a uniform product. Using one recognized lab and following its sampling instructions keeps results comparable from test to test.