The US produces about 12 million tons of waste glass annually, and only a third gets recycled. Meanwhile, we lay roughly 350 million tons of asphalt per year on roads that crack, rut, and require resurfacing every 10-15 years. What if we sintered crushed glass cullet into highway-grade pavement instead?

The physics is more interesting than "glass road = slippery death." Soda-lime glass has a compressive strength of ~1,000 MPa — roughly 30× that of typical asphalt concrete (35 MPa) and 3× Portland cement concrete. Its Mohs hardness of 5.5 means it laughs at studded tires that grind asphalt to dust. So why don't we already do this?

The sintering problem

You can't just pour molten glass on a highway. Bulk fusion requires ~1,500°C — energetically absurd for road construction. The trick is vitrified aggregate: crushed glass particles bonded by partial surface melting at ~720°C, similar to how foam glass insulation is made. Energy budget per square meter of 20 cm-thick road:

Volume:    0.2 m³/m²
Mass:      ~500 kg (density 2500 kg/m³)
ΔT:        700 K
Cp glass:  840 J/(kg·K)
Energy:    500 × 840 × 700 ≈ 294 MJ/m²
         ≈ 82 kWh/m²

At industrial natural gas prices (~$0.04/kWh-thermal), that's $3.30/m² of energy cost. A one-mile, four-lane highway (~22,000 m²) needs ~1.8 GWh and costs ~$73,000 in fuel — comparable to the asphalt binder it replaces. Manageable.

The friction problem

Wet glass has a coefficient of friction around 0.2 — terrifying for highway speeds, where we want μ ≥ 0.5. Solution: etched macrotexture. During the sintering pass, embed coarse silicon carbide grit (Mohs 9) at 15-20% volume fraction in the top 5 mm. The SiC stays proud as the glass matrix wears, maintaining microtexture indefinitely. Think of it as permanently chip-sealed road.

The thermal shock problem

This is where it gets dicey. Glass has a thermal expansion coefficient of ~9×10⁻⁶/K and modest tensile strength (~50 MPa). A 40°C diurnal swing in a constrained slab generates:

σ = E·α·ΔT = 70 GPa × 9e-6 × 40
  = 25 MPa

That's half the tensile strength — no margin for ice expansion or load stress. We'd need expansion joints every 3-4 meters (vs. 30+ for concrete) and a borosilicate-style formulation with α ≈ 3×10⁻⁶/K. Borosilicate triples raw material cost.

What we get

• Service life: 60-80 years vs. 15 for asphalt. Vitrified surfaces don't oxidize, don't rut under heat, don't strip in rain.
• Albedo: ~0.35 vs. asphalt's 0.05 — meaningful urban heat island mitigation.
• Recyclability: 100% closed-loop. Mill it up, re-sinter.
• Waste sink: Paving 10% of US roads would absorb every bottle ever thrown away.

The catch: in-place sintering requires towing a 2 MW induction or microwave kiln behind the paver at ~1 m/min. That's a 30 GJ/hour machine crawling down I-95. Doable, but it's a fundamentally different construction logistics chain — closer to railroad welding than asphalt paving.