Lubricant-infused surface cuts ice adhesion to 9.9 kPa

A lubricant-infused surface on fluorinated porous polyurethane held ice at 9.9 kPa, about one sixteenth of what bare glass needed. The membrane comes out of a phase inversion bath, gets fluorinated to knock its surface energy down, and is then filled with a low surface tension fluid. Every infused surface in this work pushed freezing past 700 seconds and stayed under 20 kPa of ice adhesion through six freeze-thaw cycles.

Why superhydrophobic surfaces keep failing in the field

Self-cleaning, anti-icing, drag reduction, water separation. Superhydrophobic surfaces are sold on all four, and in a lab they do deliver. Then they go outside. High surface tension, weak mechanical stability, poor pressure tolerance and a short life arrive bundled together, and fixing a worn texture usually means pulling the part out of service. Any coating that runs on air trapped in a texture hits that wall sooner or later.

The SLIPS coating, short for slippery liquid-infused porous surface, was invented to get past it. Fill the texture with a hydrophobic lubricant instead of air and you get a stable, mobile interface. Repellency, ice resistance and fouling resistance all improve. What is left for anyone building one is a substrate question: which skeleton holds the fluid long enough to be worth the effort?

Polyurethane answers reasonably well. It tolerates oily media, keeps its mechanical behavior across a wide temperature range, and bonds to the parts you would actually coat. Those three things made it the skeleton here, and the pore structure became the design problem worth solving.

Test conditions at a glance

How the porous polyurethane membrane is built: what phase inversion controls

Dissolve thermoplastic polyurethane granules under stirring, spread the solution on a 200 micrometer carrier, drop the coated carrier into a coagulation bath. Solvent leaves, polymer precipitates, and once it dries at 40 degrees Celsius you have a porous film. Two knobs decide what you get: how much dimethylformamide sits in the bath, and how long the film stays in it.

Uniformity sits in the middle of that grid. At 10 percent the pores come out too small and gas permeability drops. At 40 percent they are bigger but disordered, so permeability suffers again. The condition carried forward was 20 volume percent dimethylformamide for 2 minutes, which gave the regular dense through-hole pore structure a reservoir actually needs.

Fluorination: what six hours of treatment buys

Fluorination is what makes an open membrane hold on to its fluid. The film goes into a room temperature bath of 2 g perfluoroalkyl silane (1H,1H,2H,2H-perfluorododecyltriethoxysilane, FOTS) and 5 g ammonia in 100 mL anhydrous ethanol, then cures at 120 degrees Celsius for 10 minutes. Perfluoroalkyl chains sit at very low surface energy, so grafting them onto the pore walls drags the energy of the whole skeleton down and lets the lubricant wet the surface instead of beading off it.

Six hours is where to stop. Contact angle climbs with treatment time, reaches 117 degrees at 6 hours, and then flattens, so going out to 12 hours only adds chemical exposure and process time. Every infusion test ran on the 6 hour membranes.

Choosing the lubricant: three fluids with three different failure modes

Three fluids went into the same fluorinated substrate: a silicone oil, a fatty acid ester, and a perfluoropolyether lubricant. Each membrane measured 65 mm by 25 mm and took 1 mL, left to spread and soak before the excess was removed by spin coating. Where the three diverge is durability far more than wetting.

Oil retention over 20 days: where ethyl oleate breaks down

A lubricant-infused surface dies when the oil runs out. Mass was logged daily for 20 days at 20 degrees Celsius in a desiccator, with contact angle and sliding angle measured alongside. Ethyl oleate is the outlier and the cautionary tale. Mass fell steadily, the measured angle went from 6 degrees to 23 degrees within 15 days, and lubricity was gone by day 18.

The PDMS lubricant and perfluoropolyether held mass and angles flat across the full 20 days. A stable fluid plus a fluorinated porous structure is what makes the retention work, and with two viable fluids left standing the decision moves on to whatever else the surface has to do.

Static icing delay: 36 seconds versus 865 seconds

A 30 microliter droplet at 25 degrees Celsius and 18 percent relative humidity sat on each surface on a plate held at -10 degrees Celsius, and freezing was called when the droplet surface bulged. Bare glass to the best infused surface is a gap of more than twenty fold.

Two mechanisms stack here. Air caught in the porous micro and nano texture conducts heat badly, and on its own that already takes the fluorinated membrane from 36 seconds to 230 seconds. Infusion adds a second one, since the lubricant forms a continuous film smooth to about 1 nm, near molecular scale, and a droplet finds almost no solid impurity to nucleate on.

Ice adhesion strength on a lubricant-infused surface: the number that matters

Delaying ice helps, but on an icephobic surface the number that decides whether a coating gets installed is the force it takes to get the ice off. One mL of deionized water was frozen against each surface in an inverted tube, giving 100 square millimeters of ice contact, and a force probe pushed at 2 mm per second until the ice let go, with the surface at -10 degrees Celsius. Adhesion strength is peak force over contact area, three or four runs per sample.

Three readings come out of that table. Every infused surface landed under 20 kPa. The fluorinated polyurethane membrane with nothing in it managed only 98.6 kPa, so the fluid does the work while the texture mostly holds it in place. Within the infused group the order follows surface tension, because the lowest tension fluid leaves the weakest interface for ice to grab.

The counterintuitive result: the best water repellency is not the best anti-icing

This is the finding that should change how an anti-icing coating gets specified. The fluid reported as the most hydrophobic of the three is also the one that evaporated out of the membrane and lost lubricity by day 18, so a first day wetting number pointed straight at the earliest failure. Perfluoropolyether, the lowest surface tension fluid of the three, gave both the lowest ice adhesion at 9.9 kPa and 20 day stability.

None of this is mysterious. Low interfacial friction and a mobile liquid film leave very little dynamic friction at the ice interface, so ice slides away rather than shearing against a solid. Rank candidates by contact angle alone and the volatile fluid floats to the top of the list while the durable one drops down it.

Six freeze-thaw cycles: what survives and what does not

Each lubricant-infused surface went through six cycles, with ice adhesion remeasured after every one, and all three stayed close to where they started. That result matters more than the single best figure, because it shows infusion is not a one-shot effect that washes out on the first thaw.

Six cycles is short for anything meant to live outdoors, so take it as evidence of mechanism rather than a lifetime claim. The source gives no per-cycle numbers, only the end-state comparison, so a cycle-by-cycle trend cannot be reconstructed from it.

Practical takeaways for coating engineers

• Fix the pore structure first. Across the grid that was screened, 20 volume percent dimethylformamide in the coagulation bath with a 2 minute immersion gave the most uniform through-hole network.
• Fluorination time saturates. Six hours reached 117 degrees and more time bought nothing, so there is no reason to leave the film in the bath for another six.
• Pick the fluid on volatility before picking it on contact angle. Ethyl oleate fails by evaporation, not by wetting.
• Plan on topping the reservoir up. A lubricant-infused surface is a consumable store rather than a permanent finish, and pore volume sets how big that store can be.
• Quote the contact area and the probe speed with any ice adhesion number. A 100 square millimeter contact at 2 mm/s is not the same measurement as a lap shear geometry.
• Compare candidates on adhesion after cycling, not on adhesion taken on day one.

Limitations

One laboratory study, one substrate, one test geometry, so treat these as working values rather than a validated specification. The abstract and the results section describe the ethyl oleate surface differently, and the figures used here follow the results section. The 20 day retention test sat at 20 degrees Celsius in a desiccator with no ultraviolet exposure, rainfall, abrasion or shear, none of which a coating outdoors gets to avoid. Ice adhesion depends on the measurement geometry and on the 2 mm per second probe speed, so comparisons across studies need matched conditions. No wear data were produced, the molecular weight of the perfluoropolyether was not reported, and the 30 percent dimethylformamide condition was screened but never described in the source.

Bottom line

Fluorinated porous polyurethane filled with a perfluoropolyether holds ice at 9.9 kPa and still holds it after six freeze-thaw cycles, against 156.3 kPa on bare glass. A lubricant-infused surface is only ever as good as its grip on the lubricant, so write retention into the specification before you write wetting into it.

FAQ

What is a lubricant-infused surface and how does it differ from a superhydrophobic coating?

A lubricant-infused surface is a porous solid whose pores are filled with a low surface tension fluid, so water and ice rest on a mobile liquid film rather than on trapped air. A superhydrophobic coating instead holds an air layer inside its texture. Once that air layer is displaced or the texture is worn the effect is gone, while an infused surface works as long as the fluid stays in the pores.

How low can ice adhesion strength go on a slippery surface?

On the fluorinated porous polyurethane tested here, ice adhesion strength fell to 9.9 kPa with a perfluoropolyether lubricant, against 156.3 kPa on bare glass. The other two fluids reached 14.6 kPa and 12.2 kPa, so all three infused surfaces finished below 20 kPa. At that level ice can often be shed by its own weight or a light mechanical nudge rather than by heating or chemicals.

Why does ethyl oleate lose its slip while PDMS and PFPE hold up?

Ethyl oleate is far more volatile than the other two fluids, so it steadily evaporates out of the membrane during the test. On the lubricant-infused surface it was carried into, the measured angle climbed from 6 degrees to 23 degrees within 15 days and lubricity was gone by day 18. The other two fluids showed no meaningful mass change over 20 days and held their angles flat.

What DMF content in the coagulation bath gives the most uniform pores?

A bath carrying 20 volume percent dimethylformamide gave the most uniform pore structure, with regular densely packed pores and many through-holes. At 10 percent the pores were too small and gas permeability suffered. At 40 percent the pores were much larger but disordered, which degraded permeability again, so 20 percent for 2 minutes became the chosen condition.

How long should the fluorination step run?

Six hours is enough in this system. The water contact angle rose with treatment time and reached 117 degrees at 6 hours, then leveled off, so extending the treatment to 12 hours bought no further gain. Longer immersion only adds processing time and chemical exposure, which is why the 6 hour membranes were used for every infusion test that followed.

Does a slippery surface stop ice from forming or only make ice easier to remove?

It does both, through two separate mechanisms. Infusion delayed freezing of a 30 microliter droplet from 36 seconds on glass to between 733 and 865 seconds depending on the fluid, largely because the film is smooth to about 1 nm and offers few nucleation sites. It also cuts the force needed to break ice free, which is the property that governs de-icing energy and downtime.

How many freeze-thaw cycles can a lubricant-infused surface survive?

In this work all three infused surfaces held a low ice adhesion strength through six freeze-thaw cycles with no meaningful rise. That is a short cycle count for a coating meant for outdoor service, so it should be read as evidence of stability rather than as a lifetime figure. No per-cycle numeric values were reported, only the end-state comparison.

Which lubricant gives the lowest ice adhesion on a porous polyurethane membrane?

Perfluoropolyether did, at 9.9 kPa, then ethyl oleate at 12.2 kPa and polydimethylsiloxane at 14.6 kPa. The ranking follows the surface tension of the fluid, since the lowest tension liquid leaves the weakest interface for ice to grip. Where long service life matters more than the last few kPa, polydimethylsiloxane and perfluoropolyether are the safer pair, since both retained oil for 20 days.

Why does a lubricant layer delay freezing for so long?

The infused film replaces a rough solid with a continuous liquid surface whose roughness is about 1 nm, close to molecular scale, leaving almost no solid impurity for ice to nucleate on. The porous membrane underneath also traps air, which conducts heat poorly and slows the droplet on its way to the freezing point. Together they pushed freezing past 700 seconds on all three infused surfaces.

Can a porous polyurethane membrane be made without phase inversion?

Phase inversion is the route used in this study, though routes such as porogen leaching and electrospinning also produce open porous polyurethane. What matters for infusion is a connected pore network plus a low surface energy skeleton rather than the method used to make it. This work compares no alternative routes, so phase inversion is demonstrated here, not proven best.

What limits the service life of a lubricant-infused surface?

Lubricant loss sets the main limit, and it arrives through evaporation, drainage under gravity or shear, and displacement by contaminants. In the 20 day test, evaporation alone ended the life of the ethyl oleate surface by day 18, while the two less volatile fluids were unaffected. Mechanical wear of the porous skeleton is the second limit, and it was not measured in this study.

Where does anti-icing drag reduction actually matter?

Low ice adhesion paired with low surface friction suits parts that cannot be heated, inspected often, or dosed with chemicals, such as aircraft skins, turbine blades and polar infrastructure. In those settings the value is avoided de-icing energy and downtime rather than the absolute adhesion number. Field performance still has to be proven, since every number here comes from laboratory tests.