Accelerating Gallium Ions to 0.056% light speed

there's basically one predominant method to move a spacecraft one booster ignition throw something out the back that might be combustion gases from a liquid or solid rocket or it might be individual atoms from an ion Thruster but generally the formula is the same take some mass and throw it real hard the opposite direction today I'd like to show you one of those methods a homemade Liquid Metal Field emission electric Thruster now that sounds like a lot of buzzword salad but in principle it's a pretty simple device but as you can see from my numerous prototypes I built it's been far from easy to implement in practice certainly a lot harder than I expected I spent about 2 months testing six or seven prototypes using three different variants of the technology and test fired the device over 150 times I have about half a terabyte of test footage alone this project came real close to Breaking Me Like quit everything and go work on a farm at Upstate New York level of burnout but I finally got it working and I think it'll be a pretty interesting video okay so complaining done what do we building well it's a Thruster which is designed to move small spacecraft around think like Cube sets or small sets these things don't generate a lot of thrust something on the order of micro to Mill Newtons so you're never going to launch a rocket with one of these but they're super efficient meaning they can accelerate their reaction Mass to very high velocities so once you're in space these give you a very efficient method to maneuver or gain Delta V specific impulse is the metric that we care about here for some context the Raptor engine that SpaceX uses has a specific impulse around 300 330 something like that uh in contrast electric propulsion methods have specific impulse anywhere from a th000 to 10,000 feep the thing that I'm building usually clocks in around 3,000 to 9,000 the fuel is a liquid metal usually indium or gallium and it's drawn up a sharp tip by capillary action in an intense electric field typically on the order of 7,000 to 10,000 volts at the end of this tip the electric field is so strong that it starts to pull the metal away from the tip itself this forms what is known as a tailor cone it's a dynamic balance between the surface tension of the liquid trying to keep it adhered to the tip and the electric field trying to pull it away towards the extractor electrode if the potential difference is large enough the very end of the cone forms an aex that is only a few nanometers wide here the field is so concentrated that individual metal atoms begin to ionize and are ripped free from the cone this field induced emission of ions is often just called field emission in the literature but confusingly it's not the same thing as field emission where electrons are emitted from a sharp tip like used in some electron microscopes so I don't know scientists are just bad at naming things sometimes anyway once free the ions are attracted by the exract our electrode and Achieve huge velocities on the order of 20 to 40 km/s since these are fired in vacuum they ions don't collide with anything and are simply ejected out the back and that's how it works by accelerating and ejecting the ions you can generate a tiny amount of thrust but do it very efficiently it's a very commercialized technology too there's a well-known company called impulsion that makes a variety of liquid metal feep thrusters for actual use in space anyway that's enough background for now grab your favorite beverage buckle up because we're going to start building some prototypes and frankly I really have no idea how long this video is going to be it could be a minute there are three main variants that have been studied in the literature categorize basically how the liquid metal gets to the tip there's capillary externally wedded and porous emitters my first prototype used a porous emitter array made from centered glass this is basically the glass filters you might find in like a chemistry lab but CNC machine to generate the SP the idea is that once saturated the small pores in the material help efficiently Wick propellant up to the tips impulsion uses this technique uh but they use Center tungsten for their emitters instead of glass this entire project came about because a few years ago I machined some glass emitters for Michael Bry over at applied Ion systems Michael is an amazing engineer that has been working on a variety of different electric propulsion modules everything from vacuum Arc thrusters to micro hall effect he reached out asking if I could machine some of these porous glass emitters for an ionic liquid electr spray system that he was building this uses a different propellant an ionic liquid but it's a very close cousin to liquid metal thrusters anyway ever since then I've been enamored with both electri spray and field emission and wanted to build my own since I knew how to machine these glass emitters already I figured that'd be a good place to start so I machined up an array and began trying to load it with propellant I chose to use gallon Stan which is a

mix of gallium and indium unlike the two pure Metals individually this alloy is liquid at room temperature so I won't need to use heaters to liquefy the metal it just makes Logistics a little simpler for me gallon stand is really quirky stuff though like on one hand it's super reactive and will wet out most materials but on the other hand it forms a thin skin of oxide almost immediately and this oxide has incredibly high surface tension to actually wet out a material you have to puncture this oxide skin which ends up being really challenging for example it's relatively easy to coat a glass slide you just rub the gallon stain around a little bit and you end up wetting it out but it's almost impossible to do the same for a porest glass emitter I found you can't rub the tips because they're too fragile and the trapped air inside of the pores prevents it from infiltrating any deeper I tried a few methods to wet the array and none of the techniques I tried were particularly safe so we won't be showing that footage the best I could do was a partial wedding where some of the surface and a few of the tips were coated I figured I was probably good enough to at least test it so I made up a Teflon base then a second electrode was aligned over the top of the tips and it was placed inside the vacuum chamber for testing we pump down to 2 * 105 mbars and begin applying high voltage not a whole lot happens as the voltage is slowly raised until suddenly we get a bunch of fireworks and then nothing else the uh device seems to have died but how cool is that right I mean this was my first attempt and the results looked super promising already I tried reloading the propellent a few different times but it quickly became clear that the pores glass array was probably not the right approach there's some prior art in the literature and it seems like most people use extremely high pressure to force liquid metal into centered glass and the centered tungsten folks do it under high vacuum and high temperature neither of those seemed like a lot of fun so I decided to explore some of the other variants instead I ended up building three types of capillary thrusters and two types of externally wetted thrusters but the development timeline is really confusing since I kept jumping around between all the different prototypes as my frustration grew and my mental Wellness kind of slowly evaporated so I'm going to rearrange some things in the edit to make it a little more coherent but if you happen to see like the test stand or whatever changing in the background just know that it might not be entirely chronological so the capillary type thrusters use some kind of enclosed tube or wall structure to Wick the metal upwards the first one I built was called an anular slot Thruster there's basically an inner section and an outer section which are separated by a very small slot I machined this prototype out of stainless steel and then carefully lapped the diameter so that the two parts are separated by a 12 Micron Gap the outer section is bolted down Precision shims are placed around the circumference and then the intersection is secured in place the shims are 12 microns thick so in theory we should have an equal spacing all the way around in practice the slot varies a little due to manufacturing issues on my end it ranges from 9 to about 30 microns depending on location but I figure that's probably acceptable gallium is loaded into the reservoir at the Bas and then assembled with the electrode on top and it's into the vacuum chamber for testing the first few tests were really energetic but not necessarily in a good way unfortunately I don't have much footage of this because apparently this thing puts out so much electromagnetic interference that it corrupted the video files on my camera's hard drive I have three recordings that are completely corrupted and I think it actually ate a few unrelated clips that were just sitting on the drive once I realized this uh I switch to using an internal SD card which seemed to fix the issue but not an expected problem to have the main issue with this design is the bridge section between the inner and the outer electrode because it's directly over the slot the electric field is highest here and the Gallum prefers to emit from that location it also sometimes spews out droplets which then connect to the bridge and shorts the top to the bottom which effectively ends the test and you have to vent it all to Atmosphere to fix I tried reducing the size of the bridge here it's only5 mm wide but it still was too big and we had the same problems So eventually I remove the inner electrode entirely the impulsion thrusters actually only have an outer ring so in theory it should work I tried a ton of variations in the size and the distance and thickness and just could never get it working either the electrod was too far away and nothing happened or it was too close and it would start to Arc the arcs look really cool but it's not stable emission or anything and

doesn't actually generate thrust it's just fireworks although at one point I did accidentally make an Lees a light emitting screw this was both hilarious and terrifying at the same time the screws are nylon which is a good insulator but it's also very hygroscopic under vacuum nylon slowly outgasses water vapor and I guess it had outgas enough that kind of the local area around the screw became sufficiently conductive that it could form a plasma so yeah that was uh exciting my final iteration on this was to put the central electrode back but connect it by a long wire instead of bridges so this should hopefully avoid the bridging issue because there's nothing directly above the slot anymore but keep the field strength High by having both an inner and an outer electrode unfortunately I couldn't get this to work either for much the same reasons either nothing happened or it would start to Arc okay so anular slot thrusters are known to be relatively finicky to get working so let's back it up and try something just a little simpler so we can get some kind of prototype working this is a spike array for an externally wetted variant it's a lot like the porous array we tried earlier except it's solid instead of porous so basically there's the same set of pointy spikes and the liquid metal clings to the outside of the spikes and is drawn up upwards along the external Surface by capillary action cating with metal is challenging here too but a little easier than the glass gallon stand just kind of sucks to work with honestly uh I found the easiest way to coat it is to put it in a chamber with some hydrochloric acid fumes which helps accelerate the wedding process the fumes dissolve and help prevent the oxide from reforming on the surface of the gallon in and this helps the liquid metal wet out other metals a lot easier with some addition mechanical encouragement I got most of the spikes covered then the spiky spatula Nestles into a PTFE base gets capped with another piece of PTFE and an array of extractor holes and you know the drill by now into the vacuum chamber pump it down and ramp the voltage for testing this variant turned out to be very unstable uh I tended to get big bursts of activity and then nothing at all since the spikes are all kind of in the same cavity I think even the smallest Arc or emission would generate enough plasma to lower the conductance inside the chamber and then that would trigger more spikes to Emit and the effect would kind of cascade across the whole array if the spikes were loaded with too much gallium sometimes a blob would fly off and short out the array I tried a variety of whole diameters and spacings and you know just could never get a stable configuration to work not to worry though the engineering department here at breaking Taps is fullon ADHD so let's go work on those slot thrusters again there's a variant that's much simpler than the annular Thruster and that's just a single linear slot it's the same thing in principle but it's much simpler to manufacture and ensure a consistent spacing one side is flat and the other has a reservoir for propellant in a very shallow passage for the liquid metal to flow in the two halves are carefully lapped so that the Gap is only 10 to 12 microns wide and the whole thing is assembled it's loaded into a PTFE fixture and it's vac chamber time this setup happened to find a whole bunch of unrelated logistical issues there was arcing between the aluminum test stand and the wire I fixed that with some capton tape and better spacing but then there was arcing between the Thruster back and the capton tape at one point the insulation on one of the magnet wires just totally came off I don't know how well this shows up on camera but the insulation on The Wire uh started to Arc inside the vacuum chamber and I realized the vacuum pressure had dropped uh two orders of magnitude so I machined a full PTFE back plate and swap the wires over to Peak insulation but like all the other thrusters this one suffered from excessive arcing it was made worse by the aspect ratio over the extractor electrode because it was so long and thin and unsupported down the middle it would actually Flex towards the emitter once it would get close enough it would Arc the force would temporary vanish and the Ro would spring back up to the top and then the process would repeat and basically started vibrating at really high frequency okay so not going to lie around this point I was strongly considering quitting and just throwing everything into the ocean this whole project was supposed to be an easy diversion from The Bird robot project I had recently let the magic smoke out that's not good oh no and hadn't really worked up the willpower to get started again I was hoping to just knock out a cool little plasma Thruster and

then get back to work on the robot instead it's been like a month of pure frustration each prototype takes a few days to design a machine and then each test cycle is a minimum of an hour or two waiting for the vacuum chamber to pump down run the test and then vent back to Atmosphere then more time to review the footage and debug what went wrong before setting up the next test it's been a long month however I am more stubborn than I am smart so I decided to keep going and make a new design one that honestly I should have started with from the beginning a single capillary emitter this is how the first original thrusters were made they use small metal needles but I opted to use thin glass capillaries because they can be pulled to a very sharp tip to make micro pipets I even have a pipet puller but of course when I pulled it out of the storage it decided to stop working I don't know I think like the transformer died or something this thing is probably older than I am but not to be deterred I hooked it up to a big DC power supply and just controlled the temperature manually it's all sorts of janky but it got the job done once pulled the capillary is cut down to size and filled with gallium you can actually pull tips that are crazy small like only a micron or two wide uh these are used in electrophysiology all the time as patch pipets but at some point the pressure required to fill it with gallium gets so high that you just explode the capillary so there's a practical limits to how small you can make it a nice feature about these glass capillaries is that they're entirely insulating except for where the Gallum pokes out the top this should in theory help prevent arcs and accidental shorts I also changed the filming setup so that we could get a really close-up view for debugging purposes and boy I am glad I did that check this out okay first of all I think we can appreciate just how cool that looks right I wish I had it in slow motion but even at this frame rate like wow that's really cool to be clear this isn't working at all this was an abject failure of a test but diagnostically this was huge first we can see the liquid metal being pulled towards the extractor and then we can see some tailor cones starting to form on the surface then there are blasts of plasma erupting from those spots they're also very obvious droplets shooting around and landing on the electrode so the process is almost working but we're generating big droplets instead of controlled ion emission we can also see an issue here that I've been suspecting for a while the oxide skin in a different clip there's very obvious skin of oxide that's resisting the formation of a tailor cone remember I said the oxide has a really high surface tension and so it's going to take more voltage to form a tailor cone than fresh gallium eventually it explodes outwards and is replaced with fresh shiny gallium because this is in vacuum that oxide skin doesn't reform which means the new gallium has a much lower surface tension and is far more likely to form a tailor cone so my working hypothesis is that the tests are going over voltage due to that oxide skin erupting and then the excess voltage it starts ripping whole droplets off of the fresh gallium this would explain a lot of the beh havior we've seen so far in the tests around this time I was talking to Michael at applied ion and he suggested adding some current limiting resistors by limiting how much current can jump between the electrodes we should hopefully control the arcs and prevent that droplet formation honestly I should have thought of this sooner my power supply has a current limit on it but I guess it's just not fast enough to Tamp down on Pulses from arcs and I don't know I suck at e so this isn't very surprising so I got some 1 Mega ohm High voltage resistors and try it again the results were not perfect but encouraging on multiple tests I managed to get a very faint blue glow and no droplets or arcs which is fantastic I'll take that any day at this point the glow is gallan plasma that's being extracted from the capillary now the trick is just to increase that glow and make it more stable I tried to reduce the aperture size but this brought back some arcs and shorting and I wasn't able to get it stable at a larger aperture I decided I needed a sharper tip than I could get with the capillary tubes I thought sharper tips would reduce the onset voltage because you know the sharper you get it the more intense the field is at that location I later learned this really wasn't the case for ION field emission but oh well that's what I did at the time so I etched some tungsten wire electrochemically like I did in the DIY image sensor video uh these were then coated in gallium and tus fired and some tenative success we can see the gallium blob moving around on the Spike

oh is that it I think that's it and then it starts to glow from the sides a bit of Gallum can start to be seen building up on the edge of the aperture as well if we shift the camera position a little bit we can see the glow is actually a tiny little plasma Spike there it is how about that just the cutest little plasma plume oh man it's adorable look at it is so cute that's it that is exactly what we wanted I let this run for a few minutes and it seems stable I could also turn on and off the power and basically plasma as well you have no idea like no idea how much relief I felt at this point after so much work we' finally found some success of course I couldn't really stop there I want to know how much thrust this was generating uh to do that we need to start using a faraday cup basically this is a big copper cap that's connected to a Pico a meter the ions hit the cup and generate a current which we can read out then we make some mathematical assumptions and we can get a value for the force so how much force was this generating well of course it's not going to be that easy is it like nothing's ever that easy after getting the Faraday cup setup I just couldn't replicate the plasma it would not relight again I also let the magic smoke out of a whole bunch of resistors while testing which is to be fair my fault they aren't actually rated up to 10,000 volts like I thought so they had a habit of dying on me I ordered some new resistors and began working on the last major problem getting the gallium to nicely coat the spikes gallium is sitting on these tips but they aren't really like wetting out correctly when you look closely it's full of holes and there's lots of oxide junk kind of sitting around the test that worked may have just been a lucky fluke another issue is the surface roughness of these tips the electrochemical process makes a very sharp tip but it also makes a very smooth tip and apparently from reading the literature you need a certain amount of roughness to help Wick metal up to the very top this is done usually with some kind of chemical etching agent but those are really quite gross for Tungsten so I decided just to start making these on the Belt Sander it's almost as sharp as the electrochemical approach but leaves a nice rough surface incidentally focused ion beam microscopes use a gallium ion source with basically the same structure a sharp tungsten tip so how are those made you ask well the tungsten is heated to a white hot heat in ultra high vacuum and then dunked in gallium which is also been heated and allowed to react for a while and I don't know that sounded like a lot of work so how about a blowtorch instead it took some fiddling but if you make sure the tip is entirely submerged in the Gallum blob and then torch the whole thing until it glows that usually works it forms like a really nasty crust on the outside of the gallium but when you pull the Tung out it will have a nice clean coating then we pop it into the vacuum chamber and finally we can get some good measurements got a test running in the vacuum chamber right now and I thought I would give you just a little tour of how the whole setup is well how it's set up and you can see this is basically it uh forgive the messiness uh it kind of evolved and yeah it's not great uh but so we've got the chamber here obviously uh the test stand is on the inside aluminum Extrusion the Thruster under test has the feraday cup which is used to collect current got that camera monitor in it turbo molecular pump controller there valves Etc turbo pump uh this is the high voltage feed through with a kind of a custom Teflon and copper uh connector cuz it didn't come with one and then over here we've got some electron ICS there's the vacuum uh pressure gauge the high voltage power supply and can't really see that because of the glare but that is a Pico M meter which is going to a feed through right there and that's being used to collect the current or measure the current from the F cup and then obviously lights camera and a display showing everything that happening on the camera so I can keep an eye on things it's not quite as pretty as the earlier one but it's very stable and I've relid it a few times cycling in and out of the vacuum chamber current on this one maxed out around 114 microamps which if we run the numbers this works out to around 11 microns this includes a few assumptions like beam Divergence and ion charge and ignoring any secondary electrons generated the Faraday cup and you know stuff like that but as a rough measurement it should be pretty close so there you go two months of work and I've

generated about a milligram of force other than the Gallum Cod issue I found that the spike was much happier with a 100 megaohm resistor seeing that the spike liked a higher resistor I decided to retry the glass capillary and found it actually worked really well with 50 megaohms of resistance but thrust is just one of the measures and I was curious what the efficiency the specific impulse was to do that we need to measure the mass before and after a test fire so that we can calculate the mass flow rate the spike unfortunately just doesn't really carry enough propellant to measure easily with my equipment so I loaded up a capillary massed it and ran a test for 30 minutes afterwards we can calculate the change in mass and from that the specific impulse this test run actually came out to 15 microns of thrust and about 3100 seconds of specific impulse which is it's pretty mediocre for a Thruster of this variety but not at all awful comparable to like a chemical Thruster and you know for a DIY thing I'm happy I haven't Revisited the array or slot designs yet but I think I could probably get them working by carefully tuning the current limiting resistor in retrospect I think a lot of the arcing problems was having was just from the lack of having a resistor there seems to be a sweet spot where enough current flows to maintain the plasma but not so much that it generates arcs or droplets that said I think the spike arrays and slots will be harder to keep stable just because any differences in you know manufacturing or assembly of the different tips or section of the slot will change the onset voltage of that location and if you get one starting to fire it'll affect how the other ones are firing and I think it'll just be a more challenging problem so what's next you ask well as noted earlier this is a lot like a focused ionb microscope historically FIB is actually a descendant of space thrusters which is cool I always thought it was the other way around where the microscopes came first but actually it was the work on the liquid metal capillary thrusters that led to the tungsten tipped emitters which then went on to be used for the ion beam microscopes so I don't know I thought it was just kind of cool that like a space application ended up building this whole new field of ultra Precision micro Machining and microscopy it's just neat how stuff like that works anyway I could continue to refine the setup add a bunch of ion Optics and make a really lowquality focused ion beam microscope that seems kind of cool or alternatively I've worked up some designs for a better space Thruster there are features like Serpentine Pathways to prevent short circuiting from the sputtered material and like a focusing electrode to improve the thrust and you know just better logistical manufacturing layout I don't think I'll touch any of this anytime soon I'm super burned out on it but maybe someday no sponsor this week so if you'd like to support this channel I have a patreon I usually write a monthly write up about what I'm working on and recently started doing video updates with additional context and footage this video for example generated so much footage and side tangents that there's no way it's going into the main video on YouTube I always tell myself I'll make like a secondary engineering cut video with more details but it never happens I've said that like three or four times now so I'm going to start using patreon as an outlet for all of that technical detail just so it can land somewhere other than my notebook okay I think that's all I got for you I hope you enjoyed it and I'll see you all next time