Silica Sand Grades: What Actually Differs Between Filtration, Foundry and Glass Sand

One mineral, three completely different specifications

Silica sand is chemically simple — silicon dioxide, SiO₂, the same mineral as quartz. What makes it commercially complicated is that three of its largest markets specify it on almost entirely non-overlapping parameters. A sand that is excellent for a rapid gravity filter may be useless in a glass batch, and a premium glass sand may be a waste of money in a mould.

This matters because sand is often bought casually — "0.5 mm washed sand" — and then underperforms for reasons the purchase order never addressed. The table below shows what each market actually cares about.

Filtration sand: a bed, not a powder

In a rapid gravity or pressure filter, sand does not strain particles the way a sieve does. Filtration happens through depth — contaminants are captured on grain surfaces throughout the bed by adsorption and physical interception. That reframes what you are buying: you want a bed with predictable porosity, predictable head loss development, and clean fluidisation on backwash.

Two numbers govern this:

• Effective size (D₁₀) — the sieve aperture through which 10% of the sand passes. It sets the pore scale, and therefore the trade-off between filtration efficiency and run length. Smaller effective size filters finer but clogs sooner.
• Uniformity coefficient (D₆₀/D₁₀) — how tightly graded the sand is. Target 1.5 or below. A wide distribution lets fine grains pack into the voids between coarse ones, which drives head loss up, concentrates all the filtration into the top few centimetres, and stratifies badly after backwash.

Acid solubility is the third parameter and the one most often overlooked. Sand with carbonate or feldspathic contamination dissolves slowly into treated water, contributing dissolved solids and, in aggressive waters, degrading the bed over time. A ceiling of 2.0% in hydrochloric acid is a reasonable commercial specification. Where the sand contacts drinking water, AS/NZS 4020 is the relevant Australian standard for products in contact with drinking water.

Foundry moulding sand: permeability against finish

In green-sand and resin-bonded moulding, sand is the structural material of the mould. Every property trades against another:

• AFS grain fineness number summarises the size distribution in one figure. Coarser sand (lower GFN) is more permeable, letting mould gases escape and avoiding blows and pinholes, but leaves a rougher casting surface. Finer sand (higher GFN) gives a better finish but traps gas. Most ferrous work lives in the 50 to 70 GFN range.
• Grain shape — rounded grains flow and pack well, need less binder and give higher permeability. Angular grains interlock for green strength but demand more binder. Sub-angular is the usual compromise.
• Clay and fines content — in green sand the bonding clay is added deliberately and controlled; incoming fines in the base sand are an uncontrolled variable that disturbs the mix.
• Refractoriness — the sand must not fuse or sinter against the metal. Alkali and iron impurities lower the fusion point, causing burn-on and penetration defects in ferrous casting.

A narrow distribution, typically spread across three or four adjacent sieves, gives consistent permeability. A wide distribution packs densely and unpredictably.

Glass batch sand: the iron ceiling

Glass sand is specified overwhelmingly on chemistry, and one impurity dominates. Iron oxide is the primary colourant in glass, and because silica makes up 70 to 75% of a typical batch by weight, the iron in the sand is the dominant contributor to iron in the finished product.

• Standard clear container glass: the industry typically accepts Fe₂O₃ in the silica source up to 0.03 – 0.05%.
• Low-iron / optiwhite float glass: silica Fe₂O₃ must be below 0.015%.
• Amber and green container glass: more tolerant, since the glass is coloured deliberately.

Above roughly 0.05% Fe₂O₃ in the finished glass, the green tint is clearly visible at thicknesses over 6 mm. Chromium, titanium and cobalt matter at trace levels too. Size consistency is the secondary parameter: oversize grains fail to dissolve fully and leave stones and cords, while excessive fines dust out of the batch and segregate during handling. Our glass manufacturing guide goes deeper on batch chemistry.

Buying without overpaying

The most common purchasing error is buying purity that the application cannot use. Ultra-high-purity quartz at 99.9%+ SiO₂ with electrical conductivity below 5 µS/cm is a genuinely expensive material, justified for crucible feedstock and solar silicon. Putting it in a filter bed is money set on fire — a filter cares about grain size and acid solubility, not about parts-per-million trace elements.

Conversely, buying a cheap construction sand for a glass batch will produce tinted glass and a furnace campaign nobody enjoys.

State the application, state the two or three parameters that actually control performance for it, and ask for those on the Certificate of Analysis. PIME separates its high-purity quartz grit from its 0.1 to 1 mm graded silica sand for exactly this reason — they are the same mineral doing different jobs at very different prices.

One Australian-specific note: silica sand may not be used as an abrasive blasting medium in Australia. That is covered in our companion article on Australia's crystalline silica rules.

Frequently asked questions

What is uniformity coefficient and why does it matter in filtration sand?
Uniformity coefficient is D60 divided by D10 — the ratio of the sieve size passing 60% of the sand to the size passing 10%. It measures how tightly graded the sand is. A low coefficient, ideally 1.5 or below, means grains are close to the same size, which gives an open, evenly packed bed with predictable head loss and clean backwash behaviour. A high coefficient means small grains pack into the voids between large ones, raising head loss, causing premature clogging in the top few centimetres, and making the bed stratify badly after backwash.

What is AFS grain fineness number in foundry sand?
AFS GFN is a single number summarising the grain size distribution of a moulding sand, calculated from a sieve analysis as a weighted average. Roughly speaking it approximates the mesh number of a hypothetical uniform sand with the same surface area. Lower numbers mean coarser sand with higher permeability and rougher casting surfaces; higher numbers mean finer sand with better surface finish but lower permeability and greater risk of gas defects. Most ferrous casting sits in the 50 to 70 GFN band, with finer sands for non-ferrous and detailed work.

Why does iron content matter so much in glass sand?
Iron oxide is the dominant colourant in glass and silica is 70 to 75% of a typical batch by weight, so the iron in the sand overwhelmingly determines the iron in the finished glass. Even 0.01 to 0.02% Fe2O3 imparts a visible green tint. Standard clear container glass typically accepts Fe2O3 in the silica source up to 0.03 to 0.05%; low-iron or optiwhite float glass requires the silica to be below 0.015%. Above about 0.05% the tint is clearly visible in glass thicker than 6 mm.

Can I use the same sand for filtration and foundry work?
Sometimes, but usually not economically. The specifications optimise for different things. Filtration sand is bought on effective size, uniformity coefficient and acid solubility, and grain shape matters less. Foundry sand is bought on AFS grain fineness, grain shape (rounded grains flow and pack better) and clay content, with refractoriness mattering for ferrous work. A sand meeting both specifications exists but you will typically pay for parameters one application cannot use. Buy to the application.