I’m still unclear if it’s actually feasible to build on the ice. Even if most of the mass remains frozen for hundreds of years won’t the surface of the ice be constantly melting such that creating building foundations is almost impossible?
you may be able to construct huge stacked layers about 50-100 meters tall made of reinforced pykrete with a whole city or park per layer and very convenient vertical transport between them. These layers will need a very slight amount of active cooling and a sensor network to monitor temperature
Doesn’t this prove too much? If pykrete is such a cheap strong material, why don’t we use it for regular buildings?
Yes, creep is a concern but I think it can be kept at bay with geometry and low temperatures. Creep rate is exponential in temperature so keeping ice at −30 degrees or so makes a huge difference. Though I would really need to research this more.
If pykrete is such a cheap strong material, why don’t we use it for regular buildings?
Probably because we have the option of tunnelling into regular ground instead for subways. But tunnelling in general is underutilized and overpriced, likely due to lack of innovation, risk aversion, high time preference, property rights problems, NIMBYism etc etc etc. Elon is absolutely right that Tunnels would solve traffic, and tunnelling into rock is better than making a pykrete megastructure.
But you certainly could make a pykrete structure on rock foundations (which are not everywhere!) and have a pretty economically valuable and affordable megastructure in the form of a multilayer city. I tentatively suspect that tunnelling the lower layers and then building skyscrapers (like Manhattan Island) is the superior option though.
In general, megastructures are underbuilt for political economics reasons. Things like The Hoover Dam and the US Highway System would probably be illegal to build today because of NIMBYism. Most of Manhattan Island would be illegal today too.
Also we have aging and shrinking populations in The West so there’s immense social pressure to run things as a form of managed decline rather than innovate and build. Old people have lots of capital and limited life left, so they want to make everything quiet and static and peaceful, and they throw their monetary and political capital around to keep things that way.
If pykrete is such a cheap strong material, why don’t we use it for regular buildings?
Also, this is sort of obvious, but it’s only good for large structures like huge columns. Anything small or intricate will have a lot of surface area and will be hard to keep cool. So this means that it is only useful for a very particular set of applications.
Pykrete is much weaker than concrete. You need to look up actual numbers for both tension and compression, in MPa. Concrete with >50 MPa compressive is common now.
Pykrete is not 50+ times cheaper than concrete. Sawdust is now something like $40/ton + say $20/ton delivery. At 14%, that’s $8.40/ton for the sawdust alone. Portland cement is like $130/ton and you only need 15% for concrete, so around $20/ton for the concrete component. Then you need aggregate, but with Pykrete you also need to deal with freezing it, and that’s relatively difficult because it has low thermal conductivity.
Pure freshwater ice has a compressive strength that varies as |temperature| in Celsius to the power 0.78. So once it gets to say −40 Celsius it has a compressive strength of around 60 MPa
...do you think all concrete has the same strength? Here is a paper with “concrete” that has 800 MPa compressive.
So once it gets to say −40 Celsius it has a compressive strength of around 60 MPa
If you care about creep, ice at −20 C shouldn’t have >1 MPa on it.
Measured compressive strength of Pykrete was much lower. It took me 10 seconds to find this paper, there’s some data for you. 20 MPa with 14% sawdust, but creep would obviously happen at lower stress.
If you care about creep, ice at −20 C shouldn’t have >1 MPa on it.
So what is your criterion for caring about creep? How does it vary with temperature? From the limited reading I have done it looks like you basically eliminate creep in practice at −100°C and you can put something like 30MPa on ice.
But I am not clear how sawdust/fiber affects this.
Here is a paper with “concrete” that has 800 MPa compressive.
Sure, but this material is presumably impractical. Practical, cheap concrete seems to be in this 20-40 MPa bracket.
But really, I don’t think this idea is particularly limited by the compressive strength of ice. I think the biggest threat is creep of the ice under load, and the difficulty of making an upper layer that covers the ice with good strength and also low thermal conductivity, and is also cheap at huge scale. I think it’s doable but it looks like this is the hard part.
I think agricultural waste products like straw can be substituted for sawdust so maybe you are paid to take it off their hands. So you mostly just have to pay for shipping which is between $10 and $20. Then you mix with water so you’re down to $1 per cubic yard. Pure Ice is up to 25MPa under compression so maybe you can get to 50MPa with the right Pykrete
with Pykrete you also need to deal with freezing it, and that’s relatively difficult because it has low thermal conductivity.
yes, this may be a problem. I think you’d want to deposit it in layers in the cold Antarctic. A 50cm layer per 24 hours might work?
Concrete is listed here at $170 per cubic yard and UHPC is about 10x that but much of the cost drivers could be eliminated.
A major consideration is how these numbers change when you’re talking about making 10 gigatons of the stuff. Water scales well and so does agri waste. I think you might struggle with concrete.
Hmm. But I suspect you could bring the cost of that down since the ingredients are basically just ordinary rock. So maybe there is something to be said for UHPC concrete megastructures if you can scale up and vertically integrate the manufacturing.
But that’s a very expensive building! $5000 per square meter of floor area. And it’s not very efficient because lots of its volume is taken up with elevators.
Maybe we should take a step back and consider what your higher-level goals are here. I’m not sure what they are. The possibilities that come to mind are:
build large stuff out of pykrete because it’s novel
build large stuff out of ice because it’s large
establish a community separate from existing governments (seasteading)
What exactly is the difference that’s needed from current large cruise ships? Is it size per se? Independent production of food and fuel? Production of trade goods?
It’s physical size, permanence (lasts for 200+ years with extensions possible) and cost per unit area (cheap enough for middle class people), safety, no sea motion, pleasant land to live on, strong foundations for large buildings, robustness (not sensitive to one small mistake)
You don’t need to produce food, you don’t even need to produce physical goods. But you do need a population of 1 million people who are there permanently and call it home.
In general, things with a higher value per mass have less price variation across countries, because transport costs are less important, but less competition and price transparency, because they’re more specialized and lower-volume.
No, fiber reinforcement mostly affects the tensile strength.
I’m a bit confused about this but that is plausible. I have seen some sources giving quite low strength values for ice such as 3MPa but then the same source will report double that value for pykrete. This is probably because of the temperature, purity, grain size etc dependence.
Still, 25 MPa is pretty good and implies a self-support height of 2.6km. Since concrete has 2.4 times the density of pykrete this means you can make taller pykrete structures under pure compression.
Also regarding creep rate I am still somewhat confused. Firstly, there is a claim that fibers drastically reduce creep. Secondly this diagram seems to indicate that to avoid almost all the creep in pure ice you have to cool it to just below −100 degrees Celsius.
Now this is bad news for pure ice but still cooling to −100 degrees Celsius is doable, you just have a 10x greater insulation requirement and you need to run a massive cooling plant. But what about Pykrete under load? I cannot find the data.
Well, yes, but on a per cubic kilometer basis it’s probably not that bad. Still, I wonder what can be achieved with smaller amounts of fiber and less extreme temperatures. I am particularly thinking of the initial thick layer of ice that you want to put on top of an iceberg. You don’t want that to creep.
14% may be too much for this. You may get most of the relevant benefits at 5%.
Most importantly, you can’t get stuff collected for free, or shipped for $10.
Well, shipping and collection may scale better for large projects. Shipping via ocean is $10/ton for 4000 miles but it may be possible to knock that down in a large operation.
Rice husk and/or straw may be perfect after some processing/shredding.
Also there might be other ingredients that are beneficial on a cost/utility basis, such as basalt fibers, shredded plastic waste, etc.
I think there’s some underlying misunderstanding of material science here. You have some fibers, OK, and they have some amount of net strength. Distributing them in ice or whatever at a really low concentration doesn’t increase the total amount of strength those fibers have. It’s not better than putting those fibers together with less filler, unless:
you specifically want only a little bit of extra tensile strength for a lot of material
you want to keep viscosity of a thermoplastic or resin sufficiently low during processing
you’re getting much better dispersion at lower loading
Adding sawdust to ice, you’re not going to get more additional strength per wood than just using good lumber or plywood.
Yes I get that the increase in tensile strength is probably going to be proportional to the percentage of fiber added, but there’s no need to go to 14% just because that was the original pykrete formula. Maybe you only need 1⁄3 of the strength. Or maybe limiting creep is more important. Or maybe you just want to make it less brittle.
I just haven’t got the data for the properties of pykrete at lower temperatures and I don’t understand how it affects creep rates.
Since the water is almost free you might be able to choose between X meter thick pykrete or 3X meter thick 1/3-strength material.
I also don’t fully understand understand what properties you’re actually going to need.
I’m still unclear if it’s actually feasible to build on the ice. Even if most of the mass remains frozen for hundreds of years won’t the surface of the ice be constantly melting
This is addressed in the original post, please search for “Finally you want to finish your top surface(s) with something that doesn’t melt but that is also a good thermal insulator.”
Lightweight aggregates like expanded/foamed glass will form that layer, though likely with a honeycomb of basalt fiber-reinforced concrete and a final reinforced concrete topping layer.
Ok thanks. So will the top layer of concrete on foamed glass be floating of a layer of melted ice? Won’t it gradually sink as more ice melts into denser water?
It’s unclear to me how thick this layer is supposed to be. Will building foundations go though it and be anchored in the pykrete below? Presumably it’s not possible to build building foundations in foamed glass?
will the top layer of concrete on foamed glass be floating of a layer of melted ice?
No, because ice has a thermal conductivity of 2.22 W/(m K) whereas foamed glass aggregate has a thermal conductivity of about 0.08 W/(m K) .
So the ice/foamed glass interface will stay quite cold and most of the temperature drop will happen across the foamed glass layer, assuming that it is a few meters thick or something. One issue is that in the very long-term you want to actively cool the upper layers of ice because the ice will start to develop a nonuniform temperature distribution.
The insulation may also be a type of concrete with foamed glass as the solid and a cement to fill the voids. This will then be topped with a much stronger layer of concrete. That concrete can be built on. Additionally that top concrete layer can be hollow for added strength and insulation. Think of a square grid of cells for the strength layer and a solid insulation layer underneath of the foamed glass ultralight concrete.
If you want to build a heavy building it may be better to just thicken that concrete topping layer than to penetrate into the ice, as the ice will be very cold (say, −40 degrees C) and you don’t want to melt it. But Maybe there are special materials that will allow this to work. Certainly NOT metals though.
This is a very interesting idea.
I’m still unclear if it’s actually feasible to build on the ice. Even if most of the mass remains frozen for hundreds of years won’t the surface of the ice be constantly melting such that creating building foundations is almost impossible?
Doesn’t this prove too much? If pykrete is such a cheap strong material, why don’t we use it for regular buildings?
I’m reminded of Project Iceworm. Even massive Greenland icesheets still creep and flow quite a bit on building-relevant timescales.
Yes, creep is a concern but I think it can be kept at bay with geometry and low temperatures. Creep rate is exponential in temperature so keeping ice at −30 degrees or so makes a huge difference. Though I would really need to research this more.
Probably because we have the option of tunnelling into regular ground instead for subways. But tunnelling in general is underutilized and overpriced, likely due to lack of innovation, risk aversion, high time preference, property rights problems, NIMBYism etc etc etc. Elon is absolutely right that Tunnels would solve traffic, and tunnelling into rock is better than making a pykrete megastructure.
But you certainly could make a pykrete structure on rock foundations (which are not everywhere!) and have a pretty economically valuable and affordable megastructure in the form of a multilayer city. I tentatively suspect that tunnelling the lower layers and then building skyscrapers (like Manhattan Island) is the superior option though.
In general, megastructures are underbuilt for political economics reasons. Things like The Hoover Dam and the US Highway System would probably be illegal to build today because of NIMBYism. Most of Manhattan Island would be illegal today too.
Also we have aging and shrinking populations in The West so there’s immense social pressure to run things as a form of managed decline rather than innovate and build. Old people have lots of capital and limited life left, so they want to make everything quiet and static and peaceful, and they throw their monetary and political capital around to keep things that way.
Also, this is sort of obvious, but it’s only good for large structures like huge columns. Anything small or intricate will have a lot of surface area and will be hard to keep cool. So this means that it is only useful for a very particular set of applications.
Pykrete is much weaker than concrete. You need to look up actual numbers for both tension and compression, in MPa. Concrete with >50 MPa compressive is common now.
Pykrete is not 50+ times cheaper than concrete. Sawdust is now something like $40/ton + say $20/ton delivery. At 14%, that’s $8.40/ton for the sawdust alone. Portland cement is like $130/ton and you only need 15% for concrete, so around $20/ton for the concrete component. Then you need aggregate, but with Pykrete you also need to deal with freezing it, and that’s relatively difficult because it has low thermal conductivity.
Engineering toolbox lists it as 20-40 MPa
https://www.engineeringtoolbox.com/concrete-properties-d_1223.html
Pure freshwater ice has a compressive strength that varies as |temperature| in Celsius to the power 0.78. So once it gets to say −40 Celsius it has a compressive strength of around 60 MPa
https://ffden-2.phys.uaf.edu/311_fall2004.web.dir/heike_merkel/Slide4.htm
So sufficiently cold pure freshwater ice is in fact stronger than concrete, and I suspect that pykrete is even stronger but I cannot find good data.
...do you think all concrete has the same strength? Here is a paper with “concrete” that has 800 MPa compressive.
If you care about creep, ice at −20 C shouldn’t have >1 MPa on it.
Measured compressive strength of Pykrete was much lower. It took me 10 seconds to find this paper, there’s some data for you. 20 MPa with 14% sawdust, but creep would obviously happen at lower stress.
So what is your criterion for caring about creep? How does it vary with temperature? From the limited reading I have done it looks like you basically eliminate creep in practice at −100°C and you can put something like 30MPa on ice.
But I am not clear how sawdust/fiber affects this.
Sure, but this material is presumably impractical. Practical, cheap concrete seems to be in this 20-40 MPa bracket.
But really, I don’t think this idea is particularly limited by the compressive strength of ice. I think the biggest threat is creep of the ice under load, and the difficulty of making an upper layer that covers the ice with good strength and also low thermal conductivity, and is also cheap at huge scale. I think it’s doable but it looks like this is the hard part.
I think agricultural waste products like straw can be substituted for sawdust so maybe you are paid to take it off their hands. So you mostly just have to pay for shipping which is between $10 and $20. Then you mix with water so you’re down to $1 per cubic yard. Pure Ice is up to 25MPa under compression so maybe you can get to 50MPa with the right Pykrete
yes, this may be a problem. I think you’d want to deposit it in layers in the cold Antarctic. A 50cm layer per 24 hours might work?
https://www.concretenetwork.com/concrete-prices.html
Concrete is listed here at $170 per cubic yard and UHPC is about 10x that but much of the cost drivers could be eliminated.
A major consideration is how these numbers change when you’re talking about making 10 gigatons of the stuff. Water scales well and so does agri waste. I think you might struggle with concrete.
would need to be finely ground
No, straw / corn stover / etc is not free. Sugarcane bagasse is only worth about its value as fuel, tho.
check your math there
Not if you care about creep.
No, fiber reinforcement mostly affects the tensile strength.
that’s poured, not raw material costs
No. Maybe for overpriced proprietary mixes with >100 MPa, but we were talking about 50 MPa concrete.
So how expensive is 30-50MPa concrete in terms of pure raw materials at large scale?
For 50 MPa concrete you basically need to add 1% of additives that are maybe $3500/ton.
For much higher strengths you start needing expensive stuff, eg silica fume.
Hmm. But I suspect you could bring the cost of that down since the ingredients are basically just ordinary rock. So maybe there is something to be said for UHPC concrete megastructures if you can scale up and vertically integrate the manufacturing.
it seems some people agreed with you about that https://www.youtube.com/watch?v=tIQrGfV9oA8
But that’s a very expensive building! $5000 per square meter of floor area. And it’s not very efficient because lots of its volume is taken up with elevators.
Maybe we should take a step back and consider what your higher-level goals are here. I’m not sure what they are. The possibilities that come to mind are:
build large stuff out of pykrete because it’s novel
build large stuff out of ice because it’s large
establish a community separate from existing governments (seasteading)
make more land so there’s more land overall
Seasteading, which means making new land at scale
What exactly is the difference that’s needed from current large cruise ships? Is it size per se? Independent production of food and fuel? Production of trade goods?
It’s physical size, permanence (lasts for 200+ years with extensions possible) and cost per unit area (cheap enough for middle class people), safety, no sea motion, pleasant land to live on, strong foundations for large buildings, robustness (not sensitive to one small mistake)
You don’t need to produce food, you don’t even need to produce physical goods. But you do need a population of 1 million people who are there permanently and call it home.
Does this vary on market at large scale as it does for medium scale? USA vs Asia, for example was 2-3x difference in price in concrete 10 years ago.
In general, things with a higher value per mass have less price variation across countries, because transport costs are less important, but less competition and price transparency, because they’re more specialized and lower-volume.
Gold is high value per mass, but has a lot of price transparency and competition.
I’m a bit confused about this but that is plausible. I have seen some sources giving quite low strength values for ice such as 3MPa but then the same source will report double that value for pykrete. This is probably because of the temperature, purity, grain size etc dependence.
Still, 25 MPa is pretty good and implies a self-support height of 2.6km. Since concrete has 2.4 times the density of pykrete this means you can make taller pykrete structures under pure compression.
Also regarding creep rate I am still somewhat confused. Firstly, there is a claim that fibers drastically reduce creep. Secondly this diagram seems to indicate that to avoid almost all the creep in pure ice you have to cool it to just below −100 degrees Celsius.
https://engineering.stackexchange.com/questions/19692/would-creep-have-been-a-major-problem-for-pykrete-designs
Now this is bad news for pure ice but still cooling to −100 degrees Celsius is doable, you just have a 10x greater insulation requirement and you need to run a massive cooling plant. But what about Pykrete under load? I cannot find the data.
Cooling at −100C is a lot more expensive.
Well, yes, but on a per cubic kilometer basis it’s probably not that bad. Still, I wonder what can be achieved with smaller amounts of fiber and less extreme temperatures. I am particularly thinking of the initial thick layer of ice that you want to put on top of an iceberg. You don’t want that to creep.
If you pay $10 per ton for rice husk or rice straw and you dilute it 90% with water, you’re left with a cost of $1.
I saw 14% for Pykrete. $10 to $20 a ton * 14% is $1.40 to $3.80.
Rice husk is very different.
Most importantly, you can’t get stuff collected for free, or shipped for $10.
14% may be too much for this. You may get most of the relevant benefits at 5%.
Well, shipping and collection may scale better for large projects. Shipping via ocean is $10/ton for 4000 miles but it may be possible to knock that down in a large operation.
Rice husk and/or straw may be perfect after some processing/shredding.
Also there might be other ingredients that are beneficial on a cost/utility basis, such as basalt fibers, shredded plastic waste, etc.
I think there’s some underlying misunderstanding of material science here. You have some fibers, OK, and they have some amount of net strength. Distributing them in ice or whatever at a really low concentration doesn’t increase the total amount of strength those fibers have. It’s not better than putting those fibers together with less filler, unless:
you specifically want only a little bit of extra tensile strength for a lot of material
you want to keep viscosity of a thermoplastic or resin sufficiently low during processing
you’re getting much better dispersion at lower loading
Adding sawdust to ice, you’re not going to get more additional strength per wood than just using good lumber or plywood.
Yes I get that the increase in tensile strength is probably going to be proportional to the percentage of fiber added, but there’s no need to go to 14% just because that was the original pykrete formula. Maybe you only need 1⁄3 of the strength. Or maybe limiting creep is more important. Or maybe you just want to make it less brittle.
I just haven’t got the data for the properties of pykrete at lower temperatures and I don’t understand how it affects creep rates.
Since the water is almost free you might be able to choose between X meter thick pykrete or 3X meter thick 1/3-strength material.
I also don’t fully understand understand what properties you’re actually going to need.
This is addressed in the original post, please search for “Finally you want to finish your top surface(s) with something that doesn’t melt but that is also a good thermal insulator.”
Ok thanks. So will the top layer of concrete on foamed glass be floating of a layer of melted ice? Won’t it gradually sink as more ice melts into denser water?
It’s unclear to me how thick this layer is supposed to be. Will building foundations go though it and be anchored in the pykrete below? Presumably it’s not possible to build building foundations in foamed glass?
No, because ice has a thermal conductivity of 2.22 W/(m K) whereas foamed glass aggregate has a thermal conductivity of about 0.08 W/(m K) .
So the ice/foamed glass interface will stay quite cold and most of the temperature drop will happen across the foamed glass layer, assuming that it is a few meters thick or something. One issue is that in the very long-term you want to actively cool the upper layers of ice because the ice will start to develop a nonuniform temperature distribution.
Thanks. So will the building foundations be going through several meters of foam glass to the ice below?
The insulation may also be a type of concrete with foamed glass as the solid and a cement to fill the voids. This will then be topped with a much stronger layer of concrete. That concrete can be built on. Additionally that top concrete layer can be hollow for added strength and insulation. Think of a square grid of cells for the strength layer and a solid insulation layer underneath of the foamed glass ultralight concrete.
If you want to build a heavy building it may be better to just thicken that concrete topping layer than to penetrate into the ice, as the ice will be very cold (say, −40 degrees C) and you don’t want to melt it. But Maybe there are special materials that will allow this to work. Certainly NOT metals though.
https://www.sciencedirect.com/science/article/abs/pii/S2352710223008562