Bergships and Pykrete

— Notes from a 1946 lecture by Max Perutz —

In November 1946, the man who did most of the original research on Pykrete (and who bestowed its name), Max Perutz, gave a talk to the British Glaciological Society in which he detailed its physical properties and its possible use for building "Bergships". (Interestingly he never mentions the name "Habakkuk".) The paper wasn't published in the Society's journal until a year and a half later, but reading it now demolishes a lot of the mythology that has grown up around Pykrete and Project Habakkuk.

The Concept

In his introduction he makes clear that the original idea was to carve either an arctic iceberg or an ice floe into a suitable shape to provide a runway and hangar space for planes. However, icebergs would not provide room for a big enough runway, and ice floes would be too thin (and would break up in waves). The ship would have to be built from scratch.

Unfortunately ice turns out to be pretty crap as a constructional material. It has a very low, and unpredictable, 'Modulus of Rupture' — measured by supporting a beam of the material at its ends and applying increasing pressure in the middle until it breaks. Ice's average modulus is something like 22.5 kg/cm2, but can be as low as 5 kg/cm2! (Compare this to pine wood, with a modulus of about 800 kg/cm2.)

The Promise

Pykrete — discovered originally by Hermann Mark and Walter Hohenstein at Brooklyn Poly — improved on this considerably. Adding any proportion from 4-15% [presumably by weight, though Perutz doesn't specify] of wood pulp to water before freezing gives a result with definitely superior (though not fantastic!) properties. Its modulus of rupture is up to 3 times that of ice, but it's also much more consistent (about 25% variability). Further — as those of us who have experimented have also discovered — unlike ice, Pykrete doesn't shatter when hit with a hammer, or even small arms fire. According to Perutz, Pykrete is so ductile it can be machined on a lathe.

The percentage of pulp doesn't seem to be particularly critical. Less than 4% is not satisfactory, but anything from 7-14% gives about the same strength of material. The compression strength of 14% Pykrete is given as 77.3 kg/cm2, which is only about twice that of ice. In passing, this means that the often quoted figure of "3000 lbs/sq in" is off by a factor of nearly three! (Converting units comes out to only about 1100 lbs/in2.) Its performance in tension is rather better — 4 to 5 times that of ice. The type of wood used in the pulp was also found to have an effect. In creep tests (which we'll come back to in a moment) Canadian Spruce was much more suitable than Scotch Pine.

Overall, the properties of Pykrete were thought to be adequate to embark on the design of a bergship. And that design was pretty spectacular! Hollow, but with 9m thick hull walls, the craft would have to provide a 600m long runway. The calculated displacement was 2,200,000 tons — twenty-six times heavier than the liner Queen Elizabeth! And all this from a totally novel material, and one that would have to be kept well below freezing. Well below, as we'll see...

The Problems

Pykrete, just like ice, has a flaw. It creeps under pressure. If the creep is too great, the ship would slump under its own weight. It was determined, though, that if the temperature could be kept below -15 deg.C creep would be negligible after a few weeks of sagging. Not that easy when the environment might be +15 deg.C or warmer.

To keep the material sufficiently cold, the ship would need an insulating skin, and numerous (steel tube) refrigerating ducts carrying compressed air at -30 deg.C. This would be supplied by sixteen refrigerating plants. For propulsion, at a required minimum of 7 knots, over twenty 1100 BHP motors would be in external nacelles, powered by 32000 BHP of turboelectric generation. For some reason, steering by varying the speed and direction of these motors was rejected, but a rudder would add too much extra weight. They never solved this one.

The fact that the bergship was hollow was also a problem. The walls could well creep inward unless some sort of reinforcement could be designed. And a crack — possible even with a 9m thick wall — could flood the whole thing.

What really finally did the whole project in was its scale. There was no accessible place where "Nature could do the job" of freezing, so it would have to be done artificially. They intended to use a place in Newfoundland which could provide about 100 days of average -5 deg.C, but they would have to build a plant for manufacturing Pykrete that would cover 100 acres and would actually place a heavy drain on the whole of North America's engineering facilities! [And though Perutz doesn't mention it, I'd imagine that the hundreds of thousands of tons of wood pulp needed would also be a difficulty.]

In any case, the environment of the war changed. Flight ranges increased markedly, while on the other hand 600m was no longer sufficient for handling the planes that the carrier would be expected to service. So the project came to a quiet, and not very glorious, end.

Another interesting omission in the paper is the matter of melt-rate. Which is strange, given the tales that have since arisen regarding Pykrete's almost magical resistance to melting! (He does make passing reference to it forming a "soggy insulating layer" in his 1998 essay.) Whatever the reality, it doesn't seem to have been an important factor in the saga of the bergship.

M. F. Perutz [J. Glaciol. Vol.1 No.3 March 1948 pp 95-104]

Go to Pykrete main page

Have a comment or query? Please use the feedback page