The Unsinkable Iceberg: How Britain Nearly Built an Aircraft Carrier Out of Frozen Sawdust

In the winter of 1943, a small crew of men on a frozen lake in the Canadian Rockies built a strange box. They put up a wooden cabin on the ice, threaded it with cold-air ducts, wrapped it in insulation and tar, and kept piling ice blocks around it until the thing weighed about 1,000 tons. Then they sawed it free of the lake and let it float. The workers were conscientious objectors doing alternative service, and nobody told them what they were making.

They were building a scale model of what would have been the largest ship in history: an aircraft carrier made mostly of frozen water. Its code name was Habakkuk. Winston Churchill wanted it. Lord Louis Mountbatten carried a sample of its building material into the prime minister’s bath. By one famous account, Allied commanders at a summit in Quebec ducked a bullet that ricocheted off a block of it. And one of the men testing that material spent the war in a freezing basement under a London meat market, before going on to win a Nobel Prize for something else entirely.

The hole in the middle of the Atlantic

The problem Habakkuk was supposed to solve was a patch of ocean. In 1942, German U-boats were sinking Allied merchant ships faster than they could be replaced, and they did much of their hunting in a stretch of the central North Atlantic that land-based patrol planes from Britain, Iceland, and North America couldn’t reach. Sailors called it the Black Pit, or simply the air gap. Inside it, a convoy had only its escort ships for protection, and wolf packs could shadow it for days.

Aircraft carriers could have closed the gap, but carriers take steel, and steel was the thing Britain had least of. Every ton was spoken for by tanks, guns, merchant hulls, and the warships already on the slipways. If someone wanted a floating airfield in the middle of the ocean, they would have to make it out of something else.

An escaped prisoner with an iceberg idea

That someone was Geoffrey Pyke, one of the stranger characters to pass through the British war effort. Pyke was a journalist by training. In the First World War he had talked his way into Germany on a false passport, been caught, and been sent to the civilian internment camp at Ruhleben outside Berlin. In July 1915 he and a fellow prisoner hid under tennis nets until dark, climbed the fences, and made their way across Germany to the Netherlands. His account of the escape became one of the big newspaper scoops of that war.

By 1942, Pyke had landed at Combined Operations Headquarters, the department Mountbatten ran to dream up raids and new kit. Mountbatten liked unusual thinkers and regarded Pyke as a genius. Pyke’s first big scheme there, Project Plough, called for a commando force riding tracked snow vehicles into occupied Norway. When Churchill heard it, he wrote in the minutes: “Never in the history of human conflict will so few immobilize so many.” Plough eventually turned into the American-Canadian First Special Service Force and a tracked vehicle called the Weasel, though Pyke himself was pushed out after quarrelling with his American partners.

While he was in the United States working on the Plough vehicles, Pyke turned his attention to the air gap. His answer was ice. Ice was free, there was an ocean of it in the Arctic, and making a ton of it took about 1 percent of the energy needed to make a ton of steel. Take an iceberg, natural or artificial, level the top for a runway, hollow out the inside for hangars, and you would have an unsinkable airfield. A torpedo might blow a hole in it, but a hole in a 2-million-ton block of ice is just a dent.

He sent the proposal back to London in a diplomatic bag, labelled so that nobody except Mountbatten could open it. Mountbatten passed it to Churchill, who was enthusiastic. Pyke wasn’t actually the first to think of a floating ice island. A German scientist named A. Gerke had experimented with the idea on Lake Zurich in 1930, and in 1940 a similar proposal circulated at the Admiralty, where officers, including the future novelist Nevil Shute, treated it as a joke and passed it around collecting sarcastic comments. Pyke’s version went further because he had Mountbatten, and Mountbatten had Churchill.

Pyke chose the code name from the Old Testament book of Habakkuk, verse 1:5: “Behold ye among the heathen, and regard, and wonder marvellously: for I will work a work in your days, which ye will not believe, though it be told you.” In many official papers it was misspelled Habbakuk, a mistake that may have been Pyke’s own.

Why plain ice wasn’t good enough

To test the idea, Pyke and the crystallographer J. D. Bernal brought in Max Perutz, a young Austrian chemist at Cambridge. Perutz was an unusual recruit. He had come to Cambridge from Vienna in 1936, and when the war started he was rounded up as an enemy alien and shipped across the Atlantic to an internment camp before being allowed back to Cambridge. But before the war he had studied how snow turns into glacier ice in the Swiss Alps, which made him one of the few people in Britain who could be called an ice expert.

Perutz quickly found two problems with natural icebergs. First, most of an iceberg sits underwater, so the flat part above the waves is far too small for a runway. Second, icebergs melt unevenly and can suddenly roll over, which is not a quality you want in an airfield. Plain ice also has a deeper flaw: it’s brittle. It shatters when you hit it, and cracks run through it easily. A ship made of it would be a very large, very cold pane of glass.

The project might have died there, except for a report Pyke handed Perutz that he said he couldn’t follow. It came from Herman Mark, a polymer chemist who had been Perutz’s own professor in Vienna. Mark had been arrested by the Gestapo after Germany annexed Austria in 1938, bribed his way back to his passport, and driven his family out of the country with a Nazi flag on the radiator and skis strapped to the roof. He ended up at the Polytechnic Institute of Brooklyn. There, he and his assistant Walter P. Hohenstein had tried stirring a little wood pulp or cotton wool into water before freezing it, and found that the fibres made the ice dramatically tougher.

It’s the same trick builders use with concrete and steel rods, or that fibreglass uses with resin. The fibres don’t make the ice much harder; they stop cracks. A crack that starts in plain ice keeps running until the block breaks. In fibre-filled ice, the crack hits a strand of wood, has to go around it or pull it out, and runs out of energy.

A secret lab under a meat market

Perutz advised his superiors to drop the pure-ice experiments and set up a lab for reinforced ice. Combined Operations requisitioned a meat store five floors underground beneath Smithfield Market in the City of London, issued electrically heated suits of the kind worn by airmen, and lent him some young commandos as lab technicians. The work went on behind a protective screen of frozen animal carcasses, so that anyone who wandered in would see nothing but meat.

They built a wind tunnel to freeze slurries of wet wood pulp, sawed the results into blocks, and broke them. The tests confirmed Mark and Hohenstein’s results. Perutz later wrote that blocks with as little as 4 percent wood pulp were, weight for weight, as strong as concrete. The best overall mix turned out to be about 14 percent wood pulp to 86 percent water. In honour of the man who started it all, they named the material pykrete.

Pykrete had other useful habits. You could machine it like wood and cast it like metal. It floated. And it melted remarkably slowly, because when the surface thawed, it left behind a soggy layer of wood fibre that insulated the ice underneath. In one of their tests, a rifle bullet shattered a block of plain ice two feet square and a foot thick. Fired into pykrete, the same kind of bullet made a small crater and stuck.

But Perutz also found the problem that would eventually sink the whole project. Ice, even reinforced ice, flows. Under a steady load it slowly deforms, the same way a glacier creeps downhill. His tests showed that a pykrete hull would gradually sag under its own enormous weight unless it was kept colder than about −15 °C (5 °F). That meant the ship couldn’t just be a big frozen block. It would need an insulating skin, a refrigeration plant, and a network of pipes or ducts running cold air or brine through the hull for its whole working life.

A block in the prime minister’s bath

Mountbatten loved a demonstration. Years after the war, he told an after-dinner audience how he had first shown pykrete to Churchill, as recorded by Pyke’s biographer David Lampe: “I was sent to Chequers to see the Prime Minister and was told he was in his bath. I said, ‘Good, that’s exactly where I want him to be.’ I nipped up the stairs and called out to him, ‘I have a block of a new material which I would like to put in your bath.'” Once the thin outer film melted in the hot water, the exposed wood pulp protected the rest of the cube, and it sat there refusing to dissolve. Churchill suggested Mountbatten take it to the Quebec Conference.

That August, Churchill, Franklin Roosevelt, and their military chiefs met at the Citadel in Quebec City to plan the next phase of the war. According to the best-known version of the story, Mountbatten brought two blocks into a meeting of the senior commanders, one of plain ice and one of pykrete, and drew his pistol. The ice shattered. The bullet fired into the pykrete bounced off and grazed the trouser leg of the American admiral Ernest King before lodging in the wall. The British army chief, Sir Alan Brooke, wrote in his diary that it “rebounded out of the block and buzzed round our legs like an angry bee.”

Perutz told it differently. In his memoir, he said the shooting happened earlier at Combined Operations Headquarters in London, that the gun was fired by a naval officer named Douglas Adshead-Grant rather than by Mountbatten, and that the ricochet hit Brooke in the shoulder. Either way, the point landed: the stuff stopped bullets. At Quebec, the project won the backing of both Churchill and Mountbatten and was handed to the National Research Council of Canada, which had the cold winters and the ice experience to make it.

The model on Patricia Lake

The Canadians had already been testing natural ice at Lake Louise, and in January 1943 they chose Patricia Lake, near the town of Jasper in Alberta, for a prototype. The spot was cold and remote, it had a railway connection at Jasper, there was already military training in the area to act as cover, and nearby Mennonite and Doukhobor communities included conscientious objectors who could do the work.

The model was about 60 feet long and 30 feet wide and weighed around 1,000 tons. It was made of plain ice, partly natural and partly frozen by machine, around a wooden cabin packed with refrigeration gear and six-inch cold-air ducts, with tar-like bitumen in the insulation. The whole thing was kept frozen by a motor of about one horsepower. Through March 1943 the crew built the ice walls higher, and as the weight grew they sawed the bottom loose so the model floated free. It was an insulation-and-cooling experiment rather than a warship: the point was to learn whether a big body of ice could be held frozen through a summer.

It could. The refrigeration was switched off in June, and the model still lasted the whole summer before it finally melted and sank. On paper the Canadians were confident they could build a full ship for 1944, and they had worked out the shopping list: about 300,000 tons of wood pulp, 25,000 tons of fibreboard insulation, 35,000 tons of timber, and 10,000 tons of steel, at an estimated cost of £700,000.

The ship that kept growing

Then the requirements arrived. The Admiralty wanted the ship to be torpedo-proof, which meant a hull at least 40 feet thick. It had to have a range of 7,000 miles and survive the biggest waves ever recorded. The Fleet Air Arm wanted heavy bombers to be able to take off, which meant a deck about 2,000 feet long. The navy insisted on a rudder, and nobody ever worked out how to mount and steer a rudder more than 100 feet tall on a ship made of ice.

By the end, the design known as Habakkuk II had a displacement of about 2.2 million tons, roughly twenty times that of the largest aircraft carriers afloat today. Steam turbogenerators were to supply 33,000 horsepower to 26 electric motors mounted in pods outside the hull, because ordinary engines inside would have melted their own ship. It would have carried 40 twin 4.5-inch gun turrets and up to 150 twin-engined aircraft. By one critic’s estimate, its top speed would have been about 6 knots.

The cold-flow problem also got worse as the tests went on. By May 1943 it was clear the design needed much more steel reinforcement and a better insulating skin, and the cost estimate jumped to £2.5 million. The Canadians said building it that coming season was impractical, and Bernal and Pyke had to concede that no ship would be ready in 1944. Pyke was also cut out of the planning, partly to win American participation; his clashes with the Americans on Project Plough had not been forgotten.

Critics inside the Admiralty piled on. Sir Charles Goodeve, the Admiralty’s assistant controller of research, later pointed out that each ship would need enough wood pulp to dent Britain’s paper supply, plus about 40,000 tons of cork insulation, thousands of miles of steel tubing for the brine that kept it cold, and four power stations. All that, he noted, to get a vessel that could manage about 6 knots.

The gap closed without it

In the end, Habakkuk wasn’t defeated by melting. It was defeated by the war moving on. In 1943, Portugal gave the Allies permission to use airfields in the Azores, right in the middle of the gap. Very-long-range Liberator patrol bombers with extra fuel tanks could now stay over the convoys far longer. And small, cheap escort carriers, built on merchant-ship hulls, were coming out of American yards in large numbers. Together they did what Habakkuk was meant to do, years sooner and for a fraction of the cost.

Perutz, who had moved to Washington while U.S. Navy engineers reviewed the scheme, summed up the verdict neatly: “The U.S. Navy finally decided that Habakkuk was a false prophet. One reason was [that] the enormous amount of steel needed for the refrigeration plant that was to freeze the pykrete was greater than that needed to build the entire carrier of steel, but the crucial argument was that the rapidly increasing range of land-based aircraft rendered floating islands unnecessary.” The last meeting of the Habakkuk board took place in December 1943. Its official conclusion was that the big pykrete ship had “been found to be impractical because of the enormous production resources required and technical difficulties involved.”

Some of the thinking carried over. Pyke had also proposed pykrete breakwaters and landing stages for an invasion, and the Allies did build floating harbours for D-Day, the Mulberries, out of concrete and steel instead. Bernal was one of the people behind them.

What became of the people, and the pykrete

Pyke never got another project on that scale. He died in London in 1948. Perutz went back to Cambridge, briefly returned to glaciology to show how glaciers flow, and then went back to the protein he’d been working on before the war. In 1962 he shared the Nobel Prize in Chemistry for using X-ray crystallography to work out the structure of haemoglobin. Mark became one of the founders of American polymer science and set up the first polymer research institute in the United States, in Brooklyn.

The Patricia Lake model is still there. Divers in 1985 found its remains on a steep slope just off the shore, about 85 to 140 feet down: the wooden hull walls, a tangle of cold-air ductwork, and a lot of bitumen. An underwater plaque now marks the site, and recreational divers still visit it.

Pykrete itself has stayed a crowd-pleaser more than a building material. In 1985 it was considered and rejected for a quay in Oslo harbour. In 2011 engineers from the Vienna University of Technology built a 10-metre pykrete ice dome in the Austrian village of Obergurgl, and in 2014 students from Eindhoven University of Technology raised a pykrete dome in Juuka, Finland, by spraying the mix over an inflated balloon. Their dome came down within days of opening; no one was inside at the time.

Television has had a go too. In 2009, the MythBusters built a small boat from a version of pykrete made with wet newspaper, named it Yesterday’s News, and drove it around Alaskan waters at more than 20 miles per hour. It began leaking within about 20 minutes, and they called the idea “plausible, but ludicrous.” A year later, the presenters of the BBC’s Bang Goes the Theory launched a 5-tonne hemp-fibre pykrete boat in Portsmouth Harbour, and it began taking on water almost immediately.

You can try the basic experiment at home. Freeze one tray of plain water and another with a spoonful of fine sawdust or shredded paper stirred into each cell, then compare how they break and how long they take to melt on a counter. The fibre version usually wins both contests, and you’ll see the furry, insulating crust Perutz described forming on the outside as it thaws.

Why it’s worth remembering

It’s easy to laugh at a battleship made of frozen sawdust, but Habakkuk wasn’t stupid. It was a reasonable answer to a real shortage, and the science behind it held up: pykrete really is much tougher than ice, and the Canadian model really did survive a summer. What killed it was ordinary engineering arithmetic. Keeping a 2-million-ton ice ship cold turned out to need about as much steel as just building a steel ship, and by the time anyone could have finished one, cheaper tools had already done the job.

If you like stories about wildly ambitious engineering, the CIA’s attempt to lift a Soviet submarine from three miles down has a similar mix of genius and absurdity. Galloping Gertie is another case where a material behaved in a way its designers didn’t fully expect. And the Halifax Explosion shows how much the earlier world war’s Atlantic convoys depended on a single harbour. For more strange chemistry in everyday materials, see why cast iron turns nonstick.

Featured image: Roche Noire reflected in Patricia Lake, Jasper National Park, where the Habakkuk model was built and sank. Photo: Bhasker Thodla / Wikimedia Commons (CC BY 3.0).

Further reading