What if roads healed themselves? The science of self-healing asphalt

Asphalt already heals, just too slowly to beat a monsoon. Add steel wool and a magnet, or oil-filled spores, and the clock changes. Here is the maths.

A potholed road outside Kolkata airport, with broken asphalt and puddles across the carriageway
Photo: Slyronit, CC BY-SA 4.0, via Wikimedia Commons

Roads already heal themselves a little: bitumen is a very thick liquid, and on a hot day it slowly flows back into hairline cracks. Self-healing asphalt speeds that up, either by mixing in steel wool and heating it with an induction coil, or by hiding oil-filled capsules that burst when a crack opens. It works for microcracks, and could roughly double a surface's life, but no road can heal a pothole once the stones have gone.

Key takeaways
  • 1,856 people died in pothole-related crashes in India in 2022, in 4,446 accidents (MoRTH).
  • About 10,000 times faster: how much quicker VG-30 bitumen flows at 85 °C than at 25 °C (my estimate).
  • 80 °C gave the best crack healing in a 2018 induction and microwave study.
  • 25% more material cost for steel-wool asphalt, against a claimed doubling of life.
  • About 4.7 MW would be needed to heal one lane at 1 km/h (my estimate).

This is a tier-2 what-if, so I treat it as an engineering problem. Why do roads crack? What does "healing" mean physically? How fast can it happen, how much energy does it take, and would it pay for itself on an Indian highway? The bitumen numbers come from the Indian standard IS 73; the healing data from laboratory studies; the field result from the Dutch motorway trial that started it all.

Why do roads crack in the first place?

Because asphalt is a glue that ages. The bitumen binding the stones oxidises and stiffens over the years, and every passing axle flexes it a tiny amount.

  • Fatigue: millions of small bends start microcracks in the thin films of bitumen between stones.
  • Ageing: oxidation hardens the binder, so it cracks instead of stretching.
  • Water: in the monsoon, water gets into the cracks. Tyres pump it in and out under pressure, stripping bitumen off the stones.
  • Ravelling: loose stones pop out, the hole fills with water, and the next truck makes it a pothole.

For riders this is not an abstract problem. MoRTH's report for 2022 counted about 4,446 accidents, 1,856 deaths and 3,734 injuries attributed to potholes. A two-wheeler hitting one at night has far less margin than a car.

Close view of a network of cracks in an old asphalt road surface
Fatigue cracking in old asphalt. Each of these began as a microcrack too fine to see; this is the stage self-healing tries to prevent.Photo: Tony Webster from Minneapolis, Minnesota, United States, CC BY 2.0, via Wikimedia Commons

What does it mean for asphalt to heal?

It means the bitumen flows back across a crack and sticks the two faces together again. Bitumen is a viscoelastic material: solid-like when cold or loaded fast, liquid-like when warm or given time.

The theory: capillary flow

A thin crack draws liquid into itself by surface tension, the way a paper towel soaks up tea. The Lucas-Washburn equation gives how long a liquid takes to fill a narrow gap. For a slit of half-width \( h \) and depth \( L \):

\[ t = \frac{3 \eta L^2}{\gamma \, h \cos\theta} \]
  • \( \eta \) is the viscosity: thicker liquid, slower filling.
  • \( \gamma \) is the surface tension, and \( \theta \) the contact angle (bitumen wets its own crack faces, so \( \cos\theta \approx 1 \)).
  • The key point: healing time is directly proportional to viscosity, and bitumen viscosity collapses as it warms.

How viscosity falls with heat

India's standard paving binder is VG-30 bitumen. IS 73:2013 requires an absolute viscosity of 2,400 to 3,600 poise at 60 °C (240 to 360 Pa·s), at least 350 cSt at 135 °C, and a softening point of at least 47 °C.

Asphalt engineers fit viscosity against temperature with the Walther relation used in ASTM D2493: \( \log\log\eta = A + \text{VTS}\,\log T_R \), with \( \eta \) in centipoise and \( T_R \) in degrees Rankine. Fitting it to VG-30's two specification points (my working):

  1. Points. 240 Pa·s at 60 °C (599.7 °R); about 0.35 Pa·s at 135 °C (734.7 °R).

    Calculated in this post from the VG-30 limits in the Bharat Petroleum datasheet to IS 73:2013: 2,400 poise is 240 Pa·s, and 350 cSt is taken as about 0.35 Pa·s (my approximation, since bitumen is close to 1 g/cm³). Temperatures convert as (°C + 273.15) × 1.8.

  2. Slope. \( \text{VTS} = \dfrac{\log\log(350) - \log\log(240{,}000)}{\log 734.7 - \log 599.7} = \dfrac{0.406 - 0.731}{0.0882} = -3.69 \)

    Calculated in this post from the two VG-30 points above (Bharat Petroleum datasheet to IS 73:2013) with the log-log relation given in the paragraph before; 240 Pa·s is 240,000 cP and 350 cSt is taken as 350 cP.

  3. Intercept. \( A = 0.731 + 3.69 \times 2.778 = 10.98 \)
  4. Result. About 13 Pa·s at 85 °C, 5,800 Pa·s at 40 °C, and 130,000 Pa·s at 25 °C.
\[ \frac{\eta_{25\,°C}}{\eta_{85\,°C}} \approx \frac{1.3\times10^{5}}{13} \approx 10{,}000 \]

These are extrapolations from a two-point fit, so treat them as order-of-magnitude figures. Real binders also stiffen as they age, which pushes every number up.

How long would a crack take to heal by itself?

Weeks to months at normal road temperatures, and minutes when heated. That gap is the whole case for self-healing technology.

I take a crack 0.1 mm wide (\( h = 50 \) µm) and 5 mm deep, and a bitumen surface tension of 0.03 N/m (my assumption, typical of oils and binders). Then \( 3L^2/(\gamma h) = 50 \) seconds per Pa·s of viscosity:

Road temperatureVG-30 viscosity (my fit)Time to fill the crackWhen you'd see it
25 °Cabout 130,000 Pa·sabout 75 daysMonsoon night
40 °Cabout 5,800 Pa·sabout 3 daysHot afternoon surface
60 °C240 Pa·sabout 3 hoursPeak summer surface
85 °Cabout 13 Pa·sabout 11 minutesOnly under an induction coil
Hotter bitumen, faster healingTime to fill a 0.1 mm × 5 mm crack (log scale)25 °C40 °C60 °C85 °C75 days3 days3 hours11 minLucas-Washburn: time ∝ viscosity. VG-30 fit to IS 73 points.
Each 15 to 25 °C of warming cuts healing time by roughly ten to a hundred times. My calculation, using the viscosity fit above.

How does induction-healing asphalt work?

You mix fine steel fibres into the asphalt, then drive a machine with an electromagnetic coil over the road. The coil's alternating field induces currents in the steel, which heat up and warm the bitumen around them.

The physics

  • Induction: by Faraday's law, a changing magnetic field drives eddy currents in a conductor; their resistance turns them into heat. Stone and bitumen don't conduct, so only the fibres warm directly.
  • Fibre content: recent work uses steel wool at 2 to 6% of the bitumen's volume.
  • Depth: heating is fast but shallow, which suits a surface layer where cracks start.
Induction healing, before and afterinduction coil, 30–120 kHzMicrocrack, 25 °CBitumen at 80 °C flows inShort lines: steel wool fibres, the only part that heats.Hot fibres soften the bitumen around them; the crack closes.
The steel is a hidden heating element. My diagram, after the mechanism described in the studies below.
Heating Characteristics and Induced Healing Efficiencies of Asphalt Mixture via Induction and Microwave HeatingQuantao Liu et al. · Materials 11(6) · 2018

The team heated AC-13 asphalt beams containing steel wool (6% of binder volume) with an 8.3 kW, 123 kHz induction unit and a 5 kW microwave, then broke and re-tested them at −10 °C. Induction warmed the surface at about 17.5 °C a minute against about 3 °C for microwaves, but only to a useful depth of about 2 to 5 cm. The upper half of induction-heated beams recovered 30.8%, 42.8% and 65.7% of their strength at 60, 80 and 100 °C, while the lower half barely healed; microwaves healed evenly, best at 80 °C, at about 66 to 71%. The limit: one mix, few samples, and broken stones can't heal, so true binder healing may be higher.

Read the paper (free)
Strength recovered after healing, by heating temperature
See the data as a table
Induction, upper beam (%)Induction, lower beam (%)Microwave, upper beam (%)Microwave, lower beam (%)
60 °C30.8030.336.2
80 °C42.8065.870.9
100 °C65.712.662.956.9

Source: Liu Q et al., Materials 11(6), 2018, Tables 2 and 3 · Healing ratio is the fracture force after healing divided by the original fracture force, in three-point bending at −10 °C.

Induction Heating Optimization for Efficient Self-Healing in Asphalt ConcreteMarina Penalva-Salinas et al. · Materials · 2024

The authors made AC16 surface mixes with steel wool at 2% and 4% of bitumen volume (6% clumped badly), using steel slag or porphyry stone, and heated slabs with a 15 kW, 30 to 100 kHz induction unit. Mixes with 4% steel wool reached 100 °C in under 4 minutes at full current; 2% mixes needed up to 9.5 minutes and reached it only at full current. Frequency barely mattered, and a two-turn coil heated most evenly. The limit: heating only, in the lab; the study did not measure crack healing, moisture or traffic.

Read the paper (free)
Close-up of a tangle of fine steel wool fibres
Steel wool, the same stuff in a kitchen drawer. Chopped short and mixed in at a few per cent of the bitumen, it turns the road into an induction heater.Photo: Johan, CC BY-SA 3.0, via Wikimedia Commons

Has anyone built a road that heals?

Yes, in the Netherlands, more than fifteen years ago. It is still the reference case.

The story

Around the end of 2010, Dutch road crews laid a 400 m test section of porous asphalt in the southern Netherlands, mixed by Erik Schlangen's team at Delft University of Technology. Less than 1% of the asphalt's volume was steel wool. The plan was simple: every couple of years, a truck-mounted induction machine would pass over and warm the fibres, so the bitumen would seal microcracks and re-grip loose stones before they became holes. Porous asphalt normally needs its top layer replaced after about eight years. Schlangen told Reuters in 2013 that his tests suggested the new mix could double that, and that the steel added no more than 25% to material costs.

GMA News / Reuters: Self-healing roads under way (2013)

Notice the catch in the story. The road heals, but only when a machine visits. Truly autonomous healing needs a different trick.

Can a road heal with no machine at all?

Partly, with capsules. Instead of heat, you store a healing liquid inside the road and let the crack itself release it.

  • The idea: tiny porous particles filled with a rejuvenator, an oil that softens aged bitumen, are mixed into the asphalt.
  • The trigger: when a crack runs through a particle, it breaks and the oil seeps into the crack, softening the binder so it can flow and close.
  • The 2025 version: researchers at King's College London and Swansea University, with Chilean collaborators, use plant-based spores smaller than a hair, filled with recycled oils. In lab tests a surface microcrack healed fully in under an hour. The work was still unpublished as a full paper when announced.
  • The limit: each capsule works once. When the oil is used up, that spot can't heal again, whereas a steel-wool road can be reheated many times.
An asphalt paver laying a fresh surface course, followed by a roller compacting it
A paver and roller laying a new surface. Self-healing additives go into the same hot mix, so the paving process barely changes.Photo: Peulle, CC BY-SA 4.0, via Wikimedia Commons

How much energy does it take to heal a road?

About 0.7 kWh per square metre in the asphalt, and roughly double that from the machine. That makes healing cheap per kilometre but slow, unless the machine is enormous.

Calculated in this post from the density and specific-heat assumptions listed below (Penalva-Salinas et al., Engineering ToolBox) and these example inputs: a 20 mm depth, a rise from 30 °C to 85 °C and 50% efficiency. 0.67 kWh/m² divided by 0.5 is about 1.3 kWh/m² from the machine.

Assumptions

  • Heat the top 20 mm, about the useful depth Liu's team measured after two minutes of induction.
  • Asphalt density 2,400 kg/m³ (Penalva-Salinas measured 2.38 to 2.71 g/cm³) and specific heat 920 J/kg·K (Engineering ToolBox).
  • Warm from 30 °C to 85 °C, so \( \Delta T = 55 \) K; coil-to-road efficiency 50% (my assumption).
  • One lane is 3.5 m wide, an example input.
  1. Heat per square metre.
    \[ q = \rho d c \Delta T = 2{,}400 \times 0.02 \times 920 \times 55 = 2.43\ \text{MJ/m}^2 = 0.67\ \text{kWh/m}^2 \]
  2. Per lane-kilometre. \( 0.67 \times 3{,}500 / 0.5 = 4{,}700 \) kWh.
  3. Cost. At ₹8 per kWh (my assumption), about ₹38,000 per lane-km per treatment.
  4. Machine power. To treat one lane at 1 km/h, 3,500 m² an hour: \( 0.67 \times 3{,}500 / 0.5 = 4.7 \) MW. Liu's lab unit was 8.3 kW.

    Calculated in this post with these example inputs (a 3.5 m lane treated at 1 km/h, so 3,500 m² an hour) from the 0.67 kWh/m² in the first step, which rests on the density and specific heat sourced above, and 50% efficiency.

That last line is the engineering wall. Either the machine crawls along, or it carries the power of a small diesel generating station.

Would it pay for itself on Indian roads?

Only if it extends the surface's life by more than its extra cost. That rule falls straight out of the arithmetic.

Let a normal surface cost \( C \) and last \( T \) years. A self-healing one costs \( kC \) and lasts \( nT \). Ignoring the heating passes, it is cheaper per year when:

\[ \frac{kC}{nT} < \frac{C}{T} \quad\Longleftrightarrow\quad n > k \]
  • Schlangen's figures: \( k = 1.25 \), \( n = 2 \). The annual surface cost falls to \( 1.25/2 = 0.63 \) of normal, a 37% saving before heating.

    Calculated in this post from the 25% material premium and the doubled surface life Schlangen gave in GMA News / Reuters (2013): 1.25 ÷ 2 ≈ 0.63.

  • Break-even: if the life only rises by 25%, the saving is zero, and the heating passes make it a loss.
  • Indian twist: many Indian potholes come from water and poor drainage, not slow fatigue. Healing microcracks doesn't fix a road that floods.

In 2024 NHAI was reported to be considering steel-fibre self-healing asphalt for national highways. No trial results have been published that I could find.

Use this to check your own crack

Use this to check what you just read. Change the temperature and the crack size, and see how viscosity, healing time and the energy bill move.

Self-healing road calculator

How sensitive is the answer?

Temperature dominates everything. Crack size matters a lot too, because time rises with the square of depth.

Change from base (85 °C, 0.1 × 5 mm)Healing timeEnergy per lane-km
Base caseabout 11 minabout 4,700 kWh
Heat only to 70 °Cabout 1 hourabout 3,400 kWh
Crack twice as deep (10 mm)about 45 minabout 4,700 kWh
Crack twice as wide (0.2 mm)about 5.5 minabout 4,700 kWh
Aged binder, 3 times stifferabout 35 minabout 4,700 kWh

Where my model breaks down: real healing also needs the crack faces to touch and the molecules to diffuse across, and aged binder is stiffer than fresh VG-30. A wide crack fills faster in this model, but in reality a wide crack may be too open to close at all. That is why the method is for microcracks, not potholes.

Verdict: could roads heal themselves?

Verdict: possible with changes

Microcracks, yes: bitumen flows back into a 0.1 mm crack in about 11 minutes at 85 °C against about 75 days at 25 °C, and lab beams recover up to about 70% of their strength. The changes are steel wool at around 4% of the binder plus a periodic induction pass (about 4,700 kWh and ₹38,000 per lane-km by my estimate), or one-shot oil capsules. It pays only if the surface life rises by more than the roughly 25% material premium. Potholes, no: once stones are gone and water has the base, there is nothing left to heal.

A road doesn't need to be taught to heal. It needs to be kept warm for ten minutes before the monsoon finds the crack.From the working above

More on what monsoon roads do to riders: monsoon onset dates for riders, why Himalayan roads collapse, two-wheeler road deaths in India and how ghat roads are designed. Whyblog explains why Indian cities flood, which is where many potholes really start.

What if next?

Questions people ask

Is self-healing asphalt real?

Yes. Steel-wool asphalt healed by induction has been tested in laboratories and, since about 2010, on a 400 m Dutch motorway section. Capsule-based asphalt that releases oil into cracks is still mainly a laboratory technology.

How does self-healing asphalt work?

Heat or a released oil lowers the bitumen's viscosity so it flows back into microcracks by capillary action. Induction versions heat steel fibres in the mix; capsule versions break open when a crack passes through them.

Can self-healing roads fix potholes?

No. They are designed to close microcracks before potholes form. Once stones have come out and water has reached the layers below, the road needs a conventional patch.

How much more does self-healing asphalt cost?

Reports put the material premium at up to about 25%, plus the cost of an induction machine. It saves money only if the surface lasts more than 25% longer.

Is India using self-healing roads?

In 2024 NHAI was reported to be considering steel-fibre self-healing asphalt for national highways. I found no published Indian trial results as of November 2026.

Sources

  1. Citizen consumer and civic Action Group: India's neglected roads, citing MoRTH Road Accidents in India 2022 (4,446 pothole accidents, 1,856 deaths, 3,734 injuries)
  2. Bureau of Indian Standards: IS 73:2013, Paving bitumen specification (VG grades)
  3. Bharat Petroleum: Bitumen VG-10, 30, 40 datasheet to IS 73:2013 (VG-30: 2,400–3,600 P at 60 °C, 350 cSt minimum at 135 °C, softening point 47 °C minimum)
  4. Liu Q et al.: Heating Characteristics and Induced Healing Efficiencies of Asphalt Mixture via Induction and Microwave Heating, Materials 11(6) (2018; 17.5 °C/min, healing ratios at 60–100 °C, 2–5 cm depth)
  5. Penalva-Salinas M et al.: Induction Heating Optimization for Efficient Self-Healing in Asphalt Concrete, Materials (2024; 2% and 4% steel wool, under 4 minutes to 100 °C, densities 2.38–2.71 g/cm³)
  6. GMA News / Reuters: Self-healing roads under way (2013; 400 m Dutch section, under 1% steel wool, 8-year life, up to 25% cost premium)
  7. King's College London: Self-healing potholes pave the way for more sustainable roads (Feb 2025; spores, recycled oils, under one hour)
  8. Engineering ToolBox: Specific heat of solids (asphalt concrete 0.92 kJ/kg·K)
  9. NBM&CW: NHAI plans self-repairing road technology (May 2024; steel fibre and bitumen)
  10. Finshots: Goodbye potholes, hello self-healing roads? (May 2024; about 25% costlier, 7–10 year normal life)
  11. PNNL: Conserving asphalt resources, rejuvenator performance through peptizing efficiency (Apr 2026)
self-healing asphaltinduction heatingsteel woolbitumen viscosityLucas-Washburn equationVG-30potholesTU Delft
Sapta

Sapta (Saptarshi Sarkar) is a mechanical engineer turned marketer in Bengaluru and a solo touring rider. He writes the seven Saptaverse blogs. Work with him or read more of RoadHook Rides.

Keep reading in RoadHook Rides

From the other blogs