Rebuilding Science in a Magic World

The process of separating air is quite complicated, and I have multiple major hurdles to clear before I can do so.  The first step is making a functional cryocooler.  This can be accomplished using a sterling engine, but I'll have to do a significant amount of fine tuning to make that work.  My existing sterling freezers aren't that cold, and likely aren't even freezing carbon dioxide.  That means there are significant improvements that I need to make before that can work.  I already know of a few areas where significant improvements can occur.

The second step is making metal dewar flasks.  I've already managed to pull a pretty good vacuum using stone shaping before, so if I can devise a way to utilize that before sealing off the walls of the flask, we'll have the design for containers for holding liquid air, as well as the distilled components.  I'll also want to use similar technology we'll develop here to help insulate the distillation column we'll make in the next step.

In air separation processes on earth, air liquification and separation is done continuously.  Here, however, we don't have the means to produce facilities to the degree necessary to do so.  I also lack multiple bits of critical information to do so.  For instance, I don't actually know what the atmospheric pressure is here, which critically affects the boiling points of liquid air components.  I also don't actually know what the total air composition is here on the planet.  If, for instance, nitrogen makes up far less of the atmosphere than on earth, I would potentially imbalance the distillation column and waste all the construction effort.  I also can't remember off the top of my head the boiling points of the gases in air, so I'll have to rediscover that.

So, for the third, and hopefully final step of the process, I'll build a batch distillation column.  Using a batch column, I'll be able to start separations, despite being unaware of the exact contents of the column.  Since we're dealing with liquid air, the heated bottom portion of the column can be substituted with a metallic heat sink to boil off the liquid air.  The condenser at the top of the column will be cooled using a large cryocooler.

If I have enough layers, then I should be able to pull of mixtures from the column tops repeatedly, with different batches being different compositions of the air.  Whichever gas boils first should be largely the composition of the first batch, and so on through the different gases.  This won't get us to 100% of any gas, but it will give me multiple different samples that I can use to start determining compositions from.

For instance, if the first liquid I pull off encourages flames, then it's liquid oxygen.  If it doesn't, then it's some other atmospheric gas, likely nitrogen or argon.  I probably won't be able to achieve temperatures cold enough to condense neon, hydrogen, or helium, nor do I expect the atmosphere to contain a large amount of it.  I also don't expect to be condensing enough material to recover much xenon or krypton at this stage, as they, along with neon and helium, will likely be too rare to recover without processing very large amounts of air. 

Since we're basically limited to using sterling cryocoolers, we'd need an extreme amount of them to produce enough liquid air to recover a meaningful amount of xenon and krypton, and we'd need a closed air system to isolate neon and helium.  So I'm really hoping that Argon is our target gas for mana crystals.  Argon is the best candidate though, followed by helium, just based on expected abundances, so that's encouraging.

The first round of improvements was fairly obvious for the sterling engine design.  The regenerator needed to be completely redesigned first.  I hadn't put much thought into the regenerator other than the fact it needed to be somewhat porous.  To reach cryogenic temperatures though, all components of the engine are going to have to be as efficient as possible.  For testing purposes, that meant I wanted an engine that I could easily resize the regenerator, and fill it with different packing material with high porosity.

First, I worked with Karsh to make dies for making metal wools.  They're basically like the wire dies we made before, but instead of being perfectly round, they have a sharp edge that cuts into the hole, which causes a section of the wire to peel off.  Those strands are then just bunched together to make the metal wool.  Since we have a few different metal candidates, I gave the task to one of the goblin smiths to make a bunch of wools of each metal, and provide me with three different strand sizes of wool of each type.  

Initially, I'm going to just use steel wool in the regenerator, since it should be good enough for testing other components until we get to cryogenic temperatures.  Once I actually get a stirling engine down to cryogenic temperatures, then I'll experiment with the different metal wools, since at those temperatures, I expect the metals to behave differently than they do at ambient temperatures.

The second improvement that I knew I could make was replacing the air in the engine with hydrogen.  With the one fluorite crystal we bought which produces hydrogen, I can collect a large amount of it, then dry it (since the act of bubbling it through water will evaporate water into the gas mixture), and fill the stirling engine with it.  Hydrogen should be a significantly more effective gas for two reasons.  Hydrogen should be a better thermal conductor than air, and I know that it liquefies at a much lower temperature than air.

The third, and final, improvement that I knew I could make off the top of my head had to do with the general shape and design of the stirling engines I had made before.  In short, they need better thermal designs to reduce general losses.  I need to make more portions from insulative components like dried lightstone, while also adding in better heat exchangers on the hot side, to help remove the excess heat. 

Since I want to produce liquid air on the scale of gallons, I'll either need a lot of small stirling cryocoolers, or a few big ones.  The general issue with larger coolers is being able to dissipate heat fast enough.  I plan on building all the final components at the hydroelectric facility to provide consistent power to the setup long term.  That being the case, I realized that I can utilize the mountain stream as liquid cooling to remove heat faster, so I decided to design a larger cryocooler.

Though for testing purposes, I started with a smaller design, to get a feel for the general mechanics, since it's much easier to replace and redesign a smaller cooler over a large one.  I took a month just tinkering with various designs, and running them to check temperatures.  By the end of that month, I'd gotten to the freezing point of the ethanol thermometers, and was able to make a small amount of solid CO2 from pre-dried air which had deposited on the cold side head. 

Unfortunately, below the freezing point of ethanol, I don't have a great way of checking just how cold the stirling engine was getting, and without a dewar flask to hold liquid air without it evaporating quickly, I couldn't really tell if I was actually at that cold of a temperature.  So that was the next thing on my agenda.  If I have a dewar flask, and we can get to the point where air liquefies, then I can test different regenerator materials by measuring the total volume of liquid air in the flask after a set period of time.

Actually making a dewar flask wasn't quite as difficult as I initially thought it would be, and I was able to make five half-gallon flasks in ten days with Karsh's help.  A dewar flask is just two containers, with a vacuum between them.  You want as little physical contact between the flasks as possible, to minimize any conductive heat loss.  I'm sure the method I settled on probably reduces it's effectiveness by a small amount.  We were able to mold steel into the shape we wanted for a double walled container and left a small hole at the bottom.  Then, using stone shaping, I slowly pulled a vacuum through that hole, and then sealed it with a large dried lightstone plug. 

The downside of this method is that the lightstone plug is somewhat large on both the inside and outside, so that it can resist the vacuum causing it to break.  Thus the flasks are a bit bulkier than I'd like, and the properties of the lightstone probably make it less effective than true scientific grade dewar flasks would have.  I did make special lightstone stands for them though, so that the bottoms of the flasks aren't in contact with the ground as much, to help reduce losses a small amount.  Though next year, once the reservoir is refilled, I want to try electroplating flasks with what little silver we have, to further reduce losses due to radiation.

The flasks seemed to work well, and kept boiling hot water hot for a considerable amount of time.  With a few flasks ready, I can finally try to make some liquefied air.