Rebuilding Science in a Magic World

I had a few months left before winter, so I started work on the batch distillation tower.  A major problem I'm going to run into with the tower is pressure.  The tower will be condensing and evaporating air, forming layers of different concentrations of liquid, meaning equilibrium might actually occur at a higher pressure than I'd normally want.  Unfortunately, with our technology at the stage it is at, I can't really control that without doing a few years worth of research.  I already don't know what the air's overall composition is, nor do I know the exact ambient pressure.  Both would play a critical role in being able to properly size the batch distillation process to avoid those sorts of issues.

So instead, I'm going to aim for overkill of the process.  A batch distillation tower is basically a bunch of layers in a tower, where the bottom evaporates the material that we're trying to separate, and the top has a condenser to liquefy it again.  Then, we need a way to pull material from somewhere in the batch process, in this case, I'll be using the top of the tower near the condenser, with a secondary collector near the bottom, in case I need it if something breaks.

I had initially planned on just using a few small cryocoolers for the condenser portion, but I'm now inclined to make the second large cryocooler, and have its cold piston feed the condenser.  If I could thermally isolate the distillation tower effectively, then I'd be more confident in using a smaller cryocooler, but we'll likely be dealing with some amount of heat gain throughout the tower that the main condenser is going to be dealing with.

The bottom of the tower needs an evaporator, but given the cryogenic temperatures involved, I actually don't think I'll need a heat source.  Instead, I plan on just making a copper heat sink to transfer heat from the ground directly into the distillation tower at the base.  I also need to figure out what volume I want to process at a time, as that determines the size of the heat sink and the tower.

On one hand, a larger batch process ultimately means I can extract purer product, and it should also be more efficient.  As the column increases in size, it's surface area to volume decreases, meaning the outside of the process has a lower percentage of influence on the temperatures in the process.  However, the higher surface area also means that there is an overall increase in total thermal energy leaking in, which needs to be removed from the process somehow.

Ultimately, the current cryocooler produces about a gallon of liquid air every three or four days.  Meaning that realistically, I could maybe process five to ten gallons in a batch.  While I could try to go higher, I'd need to design a larger dewar flask to store more liquid, and give that flask it's own small cryocooler to help prevent loses over the time it takes to fill another container.  Given I don't know that much information about the atmospheric composition, overkilling it with a ten gallon process might be a bad idea.

With distillation processes, the more independent layers you have, the better your separations are, at the cost of needing more thermal energy and a longer time to reach equilibrium.  Layers I make aren't going to be perfect either.  In high quality separations, layer design is highly important to getting good efficiencies.  Each layer will house some amount of condensed liquid, and some amount of gas passing over top.  Each tray in a layer needs a weir to hold some amount of liquid, and a way for gas to pass through.

From memory, having raised holes in the tray in each layer that the gas can rise through is a somewhat effective method.  The inside of the column will be metal, with metal trays, but it will be encased with dried lightstone to give it some degree of pressure resistance.  Then, I want to have a partial vacuum gap, similar to the dewar flask, and a second layer of dried lightstone.  Then, outside that, I want an air gap, which is cooled with a stirling cooler.  All of this is to try to remove as much thermal influence as possible from the outside world from the column.

Five gallons of liquid isn't actually that much volume either.  I wanted to try to make twenty layers in the column, but that might be unrealistic given the total volume of material I'm working with.  The column is going to end up being quite small, but that's probably for the best.  Ultimately, I want to have a bunch of different tray designs made before I start testing, so I can swap them out quickly, and then make requests for new trays once I have an idea of what I might want.

I'm going to have to do a lot of fine tuning to make it work, and I'll need to come up with some testing methods to determine how well I'm actually separating products.  The best test I can think of is that I can put dewar flasks of equal mass on a balance, and extract liquid from the top and bottom of the column of the same volume, and measure the mass difference.  The larger the mass difference, the higher the purity of the the gases on either side.  The biggest flaw with that is being precise with the volume of liquid gas in each flask, but if I make the flasks tall and thin, it should minimize that problem.

Building out the column took a bit longer than I had hoped, because I kept running into multiple different problems as I built different parts.  First, because I wanted the partial vacuum layer and an outside cooling layer, I ended up having to partially compromise on that design so that I could have easy access to changing designs within the column itself.  So, about 80% of the column is surrounded by that design, and twenty percent is designed as three removable wedges that are re-attached with stone shaping after any modifications are done.  Meaning modifications will have an extra hour or so of work added in any time I want to do so.  

Second, Rather than have to change the design of the building we're in, I ended up doing quite a bit of excavation into the ground to put the column in.  The top with the condenser sits at ground level where the large cryocooler piston is.  The bottom sits about fifteen feet underground, with copper being used as the heat exchanger, embedded into the ground to try to improve heat exchange on that side.  

Third, designing new dewar flasks as well as the valves to pull material out of the column took some time.  Getting everything right for a mass balance was relatively easy.  Thanks to thermal hands, I can activate it along the rim of a dewar flask to attempt to evaporate a little extra of the liquid in one side of the balance if they aren't quite even.

After all was said and done, we only have 20 days of winter left, meaning we have about 30 days until I can start using the dam as a power source and start testing.  Since I have just about everything ready, I'm going to go do my agreed upon leveling for the next year again.  I'm both excited and afraid to finish this final step of the process.  In the best case, we can make a ton of massive mana crystals.  In the worst case, I've wasted a few years making a liquid nitrogen facility that we barely have any use for, if I can get the column working.  If I can't, then we have a liquid air facility, which I'm not sure what uses we have for.

Initial testing results were awful.  I expected bad results, but everything was basically mixed throughout the whole process, and that was after fixing multiple leaks and breaks that occurred during the first few tests.  The biggest problem, by far, is that it takes a little over a day to modify the column for another test, since I have to wait for it to warm up some, then disassemble and reassemble everything.  I am thankful that the production of liquid air at least keeps up with the column's losses of gases from extraction and mass balancing, so I wasn't waiting for more liquid air production before I had to run another test.

After a month of making various changes, I started to get the most minute of results, and it was because I had made a really dumb mistake.  The larger condenser was pretty much overwhelming the passive heat sink, and the rock that the heat sink was attached to wasn't able to keep up.  While I initially considered trying to use water, I had a brief epiphany, and exposed the heat exchanger to air, and cut a bunch of the larger fluorite heat crystals to fit the existing design, and attached them to the heat sink, and that worked, giving small results.

Unfortunately, I then had to take another month of time restarting on the tray experiments.  However, I did start to see marginal improvements through various testing.  Following testing results, I ultimately added more space to the evaporator at the bottom, and a larger tank for condensation, and charged the column with more total liquid, which helped.  I also started narrowing down the tray design necessary for the whole process at that point.

After another month and a half, I'd gotten to a point where I was confident that I was separating the liquid gases to a high degree with the column.  Meaning I could start testing on those purer gases.