That also means the math should be a division instead of a multiplication right? otherwise the units don't check out. I think that factor of about 500 on the final result does affect the conclusion of the calculation. Still a really great video, and let's be honest most people come here for watching the experimental work. Thanks so much for making this really intersting series!
22.4 L/mol is the value for a temperature at 0°C, while the value at 25°C is 22.4 x (273K+25K)/(273K) = 24.5 L/mol. For a temperature in the tube furnace of 1000°C, the molar volume of an ideal gas is 22.4 x (273K + 1000K)/273K = 104.5 L/Mol.
Just to add what the others said, instead of using a rather derived value for one temperature and pressure, just use the ideal gas equation directly. Ideal gas is fairly accurate for these conditions,
@@testingmaterials3242 ⚠ God has said in the Quran: 🔵 { O mankind, worship your Lord, who created you and those before you, that you may become righteous - ( 2:21 ) 🔴 [He] who made for you the earth a bed [spread out] and the sky a ceiling and sent down from the sky, rain and brought forth thereby fruits as provision for you. So do not attribute to Allah equals while you know [that there is nothing similar to Him]. ( 2:22 ) 🔵 And if you are in doubt about what We have sent down upon Our Servant [Muhammad], then produce a surah the like thereof and call upon your witnesses other than Allah, if you should be truthful. ( 2:23 ) 🔴 But if you do not - and you will never be able to - then fear the Fire, whose fuel is men and stones, prepared for the disbelievers.( 2:24 ) 🔵 And give good tidings to those who believe and do righteous deeds that they will have gardens [in Paradise] beneath which rivers flow. Whenever they are provided with a provision of fruit therefrom, they will say, "This is what we were provided with before." And it is given to them in likeness. And they will have therein purified spouses, and they will abide therein eternally. ( 2:25 ) ⚠ Quran
@@leocurious9919 hi. when you multiply it by 1/10000 do you take into account that you have to convert it to pa? or should you? I was trying to make similar calculations but my vacuum pulls 3.9996 Pa
Can one of the big chip manufacture please please get this lad sponsored and offer him tours through a production facility it would be great to see how much a guy like this gets out of such a hands on tour.
@@projectsinflighta number of universities have small scale chip fab labs, try emailing them perhaps? The small scale might be closer to your process and teachers often love to teach someone who has shown as much interest as you have.
I think you might be getting a stress riser where the end of the tube meets your metal plug. The atmosperic pressure ends up bearing on a very small area. You could put an o ring there as padding. And make sure there is tolerance for the metal to expand. This is a 1 bar pressure seal so you could leave the o rings a bit proud and it will still seal. If the metal plug heats up enough to put two opposing point loads on the inside of that glass hoop - you're hooped!
In the final design, the glass is contacted only by the o-rings, which squeeze around the tube and prevent the metal flange from pushing against the glass.
I was actually looking at the final photo and a little worried about having all that mass out on the two ends. Are you at all worried about the tube snapping at the point it meets the furnace?
The tube seems pretty strong and i don't currently expect it to snap from the weight. plus this is how they do it in industry so it's *probably* fine until i make a new video and break the tube again
Adding a vacuum reservoir tank and valve inline with the pump would allow you to "precharge" the vacuum system while firing the furnace, and save you those 2-3 minutes of time it takes to draw everything down, perhaps thinning the oxide layer a bit more.
Sure, but you said you didn't want the expense of having to source welding gas. Reusing an old tank is pretty cheap and effective and never needs refilling.
This kind of project (making your own silicon chips) absolutely isn't in my wheelhouse, but I'm so down for all this problem solving you do on the channel. These videos are such a good demonstrations on how to tackle complex problems and how to be resilient in the face of constantly "failing" towards success.
You are a great storyteller! It's one thing to make something, it's completely something else to record it in a way thats engaging for a general audience.
Man, this project is sick! I can't believe it doesn't get more views, especially with how well laid-out and well-explained it all is. Keep up the good work!
Small note, although probably insignificant: EDIT: this might be a larger problem because depending on the wires used, you end up measuring the furnace temperature in reference to the flange. See sub-comments. Usage of thermocouple extension wire is recommended in this case as the termination point temperature is for sure above the board reference temperature. On the other hand if you are fine with having 50K uncertainty (I'm assuming the flange is 50C above ambient), it should be fine. In presence of temperature gradients (which in case of a furnace end cap will for sure be the case) you should not solder thermocouple leads as it creates a unpredictable set of thermocouples at the soldering point. Depending on how precise you want to go, the optimal solution is to use a specially designed thermocouple wire (for given thermocouple type) and spot welding it to the thermocouple tail, a mechanical connection can also work. At the board, the connection point is less critical if you can guarantee that there are no thermal gradients, but type-compensated connectors are also available.
I was about the say the same. Because you soldered wires at the flange, you will be subtracting the temperature of the flange from the measured temperature, so the more the flange heats up the more error you will have. For this reason they make so called "Thermocpuple extension wire" it is made out of the same material as the thermocpuple, so that it has no thermal EMF, this brings your cold junction all the way to the thermocouple connector (Where you can measure the connectors temperature to compensate for it)
Thinking about it more, you are right. If we simplify the model down to only copper from the flange, then the termocuople measures temperature of (Furance - Flange) but the calculation right now probably assumes (Furnace - Ambient/uC board).
The temperature of the solder joint is roughly ambient temperature, even with the furnace cranked to it's max temp. I can double check again the exact temp, but i'm pretty sure it's cold enough not to be a huge problem.
Yes, you built another pair of thermocouple junctions! Yea, it will add a temperature error to that (roughly the temperature difference between the ends of the copper involved). Thermocouple extension wire recommended to avoid this
Suggestion if you ever rebuild the thermocouple assy: instead of using an epoxy paste like the JB weld, use an epoxy liquid. There won't be as many entrained bubbles (you can improve on this a lot more if you use your vacuum pump to pull the micro bubbles out too), but you will have a much smoother surface finish. This is important because oxygen will be diffusing from the surface of the epoxy in a good vacuum. The liquid epoxy also reduces the risk of a leak along the solid wires. If you end up using stranded wire for something the liquid can flow between the strands and still give a pretty good seal (though you might want to spread the strands a bit). Keep up the good work. BTW: I have gotten that HPLC going and I still plan on running that PCB resist you sent me on it to see what happens, but I will want to get a bad column to run it on first so I don't risk a good one.
yeah, that might be a better method. i like JB weld because it *seems* very compatible with vacuum, but it is true that it's a lot more viscous than i'd like
JB Weld works WAY better with vacuum than most off-the-shelf epoxies, especially the low viscosity ones. It's the right choice here, and it tends to cure very clean, without bubbles. Hysol EA 1C is better than JB Weld, but a bit more expensive(2x).
A lot of excellent problem-solving again. I loved your vacuum development process. I was sad that you got caught out by the scam-like nature of modern glass, specifically that of faux borosilicate. Pyrex especially (which I don't know if you were using but is relevant) built a business of trust in the 20th century that their glassware was made of the high-quality material, but around the turn of the millennium, started allowing more use of their name on regular, low-quality glass. Even today, people are surprised when their Pyrex dishes shatter under normal, medium-temperature conditions because they'd owned other Pyrex dishes from 25+ years ago that had survived for so long. It's no doubt that most lab equipment companies now are more than happy to allow consumers to assume they're made of materials that had become standard but are instead no more robust than a $3 glass coffee carafe. But enough of my ranting. I love watching your processes.
It could have been borosilicate glass. Borosilicate glass should only used up to temperatures of 490°C. At 817°C it already softens and 1000°C is just overkill. Only quarz glass can survive these temperatures
@@hantrio4327 This is what I was going to say. The annealing temperature for clear boro is around 1050 F, or about 560 C. Anything above that is soft enough to deform, given enough time, and the vacuum inside will just make it deform more quickly than it would under just gravitational force. The softening point is listed online as about 820-825 C, but that's just the point where it reaches a specific arbitrarily specified viscosity. It's still well below 1000 C, though.
Damn, you are such an inspiring person man! Really loving what you do and breaking the limits of what a single person can as hobby! Please don't get discouraged because of less views and following, you will surely get the fame you deserve one day.
The way you explain your steps and assumptions, along with mentioning failures and successful results, is very helpful and interesting. Thanks for sharing such videos
Fascinating work. Ive used quartz rod for a clock pendulum due to its exceptionally low thermal expansion, nice work on the end caps. Hope the heatshring on the thermocouples doesn't leak vacuum, you can get ceramic feedthroughs but also glue lined heatshrink .
oh the heatshrink isn't doing any sealing- the epoxy takes care of that so it's not a problem. that being said i do want to learn more about non porous ceramics
20:15 I think someone else mentioned this, but depending on how hard of a vacuum you intend to pull this *could* be an issue. There are Vacuum Rated Epoxies/RTV Silocones for Vacuum Use if i remember correctly…granted for a price. That “Faraday Wax” someone on-RUclips made may be worth a look, that and/or a less “witches brew” and moreso just “this fraction of paraffin wax” or something type recipe may be neat. GRANTED “if it ain’t broke don’t fix it”, and the build is already coming along great! Just thought I’d throw this out there in case you didn’t already know about that and “virtual leaks” and all that. Keep up the great work!
It will never be as accurate, and it might suffer from confounding circumstances, but you should be able to run a series of heating and cooling cycles, collecting temperatures from both thermocouples, and develop a calibration curve that makes the permanently installed thermocouple useful. As long as this tube survives this means you could potentially make it workable, while still having access to both ends, if say you wanted a setup with gas inlet at one end, and the vacuum pump on the other, or you wanted to develop a cap that somehow lets you push and pull the boats while they're already under modified atmosphere
@@projectsinflightIf you're really interested in temperature accuracy and removing the oxygen from the tube, i would move the thermal couple from one end and put it in the center, but cut a hole in the center quartz tube on one side to fit it. This will allow you to keep as many nichrome loops around the tube in the center. Submerge the tube under water to allow for heat and stress dissipation without cracking the tube and cut a hole with a glass bit in the center, seal it off with JB weld extreme heat which should be fine for 1,200C. I would then add a tube and a flage to the other end so you can purge the tube with inert gas after pulling a vacuum, then pulling the vacuum again. Argon can be obtained relatively cheaply (less than $20) from wine preserver bottles. It also gives you the option to use the other end for other projects like if you want to record the doping process with a camera
Thanks for sharing all of these DIY semiconductors content! I couldn't find any source on internet, which shows whole process keeping it simple and at the same time comprehensively discussing the topic. Just watched all videos on your channel and now patiently looking forward for updates on project. I also hope for some mosfets, because they're what's really used because of their energy effieciency. You could even make SSD like memory with NANDs if you will be able to grow floating gate on SiO2 to hold charge and therefore store data and another SiO2 layer on top of that for insulation and finally gate contact on top to change state of memory:)
Hey, great video again! Few suggestions below, but definitely include lathe work videos and keep up the amazing work! You can create an offset calibration curve for the external thermocouple, this is how many lab furnaces operate, as the hot spot is not usually very large and you're planning on putting the sample in the middle of it, giving a temperature offset anyway! I do this in my lab furnace which reaches 1500C so those radiation issues are massive. For your thermocouples, just buy thermocouple extension wire of whatever type you want, then use an acetylene torch (or spot welder) to melt the two wires together. Bam, now you have any length you like without any issues, and can shorten very easily (cut and remelt). Either a vacuum charger, or nitrogen/argon flush of the tube while loading might help remove some oxide growth. You could use your blank end to provide the flush?
I thought about doing that, but I was wary of adding any more sources of noise in the thermocouple measurement, especially since an offset of 10 or 20 degrees C can be pretty devastating to some processes. i like the idea of the nitrogen flush while loading- i may look into that
@@projectsinflight I would say that in my experience getting temperature readings that precise in a furnace is possible, but requires careful measurement what ever way you approach it. We use an internal probe like you do to calibrate our carbolite furnaces, but our atmosphere would destroy them if we left them in, so we rely on the external sensor (used by the carbolite controller) to read and create a compensation curve for the internal conditions. This temperature offset is also a function of position along the tube so we take samples along the length to find the hot spot and it's size/shape, AND we even build test geometries of our crucibles and powders with the thermocouple buried inside to verify the final performance, as just putting a boat/crucible in causes a significant change in the radiation flux compared to no boat. If you're getting noise, this is probably due to electrical issues, I would definitely recommend borrowing a proper thermocouple meter to just verify your Arduino setup is right too
The JB weld should definitely be fine outgassing-wise at these temperatures. I'm not sure if you said so, but make sure those threaded connections are sealed with either teflon tape (if they are NPT) or epoxy. Looks really good!
Hi. Verry nice video. I have some experience using tubular funaces in our lab. Though we are not using it for Silicone manipulation but for chemical syntheis. We are using a self made compression flange as well and I think our design is much simpler to make. If you are still interested I can share the design with you. Some other comments. If you start using gas flow I suspect that your O-Rings will melt . Even with massflowcontrollers, you have a significant heat transfer to the outgoing end at lest when you are using the furnace for days or weeks like we do regulary. I even had FFKM O-Rings melt on me! Also be aware that the glowing quarz tube creates significant UV-Radiation over 900°C wich will go into your flanges and heat them up eventually. Also if you wand to use Argon or other gases sure you will need a primary reduction valve but after that a needle valve and a buble counter with silicone oil on the other end will give you relativly fine controll over the gas flow. It might be enough for you. Finaly please!!! do not heat you quarz tube over 1100°C for longer than an hour. If you go longer you will reach the strain point of the Quarz wich basically creats a silicone oxide cristal growth chamber. Basically the glass will start to realine into its natural cristal structure on the surface of your tube wich creates strain over time, wich will eventually brake the tube. I have lost so many tubes because unknowing scientists were doing yust that.
this is very useful information! if you don't mind, would you send me an email? projects in flight at gmail dot com? i'd love to talk about the flange design
Even the cleanest welding argon 5.0 can have impurities in some bottles. I had troubles with arc stability during aluminum TIG welding. I was prepared to send the machine for servicing, but after receiving advice to change the bottle, I was able to weld properly again.
@@projectsinflight I'm no expert, but I wonder if something like a molecular sieve filter would help clean up the argon (and maybe vacuum if ever needed). No idea the cost to buy or fab a filter for your purposes, it might not be practical. But I've been wondering where I'd see practical uses for them since Tech Ingredients used 3A sieves to concentrate alcohol a few years back and from what I find online, tube furnaces sometimes use them.
In my lab we have to work with oxygen free atmospheres in tube furnaces, though in my case the furnace tube is alumina so that we can reach 1500 degC. One of the simplest ways to go is to use what is called a test plug. It is an expandable rubber stopper with a gas feedthrough. That works fine, but oxygen does still get in, so we would place a dish of titanium sponge in a hot, but not too hot zone of the tube upstream of the sample. This did a nice job reacting with any oxygen that would get in. But in the end, I bought two weldable quick connect flanges and used a liquid engine gasket compound to make an airtight seal between the flange and the tube. This holds vacuum very well. But with all that said, flowing argon at a slight positive pressure is not very expensive really and works way better than vacuum to keep oxygen out. The $1,000 for argon and a regulator is way overpriced unless you are talking about buying a large gas cylinder.
yeah, i may have overestimated the cost a tiny bit. also, interesting that i didn't try a simple rubber stopper. i guess it just didn't come to mind. not sure how i'd have attached the flare fitting though
Great math around the residual oxygen at the vacuum level, but you can flush the tube with Argon and then pull the vacuum, which means there shouldnt be more than 1E-5 oxygen that was in the Argon, which should be much less. edit: P.S. great project and thanks for sharing.I wish they were longer and more detailed, I just love the combination of engineering, DYI and physics!
Argon does not need 2 regulators, all those for welding already include what you call secondary regulator. Also, since this is so wildly used for welding, it is dirt cheap. You can also use any hose you want, since you are not working with pressures. Note that you can turn on the vacuum pump seconds/minutes/hours before you actually need to use it. You do not need to turn it on right as you seal the vessel.
standard welding regulator dispenses too much gas. i only need a tiny bit, like on the order of a few cc/min. that being said i probably overstated the cost associated with argon a bit and i'll likely add a purge input for inert gas to the setup
@@projectsinflight Measuring that low will not be possible with a normal regulator (with rotameter), correct. But for those low values just bubble it through water to set it to what you need. If you bubble the output through water you not only meter the flow but also make sure to always have a bit of positive pressure in the system.
Wow, such a nice video and so well presented! Thanks for the insights on glass seals. Now that you have a vacuum tube furnace, you could try putting some H2 inside, that would actually reduce any oxides, and even if your substrate was not clean enough, you could "clean" it in-situ! Also, for the external thermo-couple, yes, it doesn't see the same amount of infrared radiation, but doesn't it get due it heat diffusion at some point up to the temperature? If not, you could actually calibrate it, knowing what you get from the internal one, and not throw it away. Anyways, very fond of your channel, best regards.
If you make heatsinks and clamp them on the tube right before your flanges, then you can prevent the heat being conducted all the way to the O-rings. Just in case if it ever becomes a problem.
if it doesn't happen at 1200C i doubt it'll ever be a problem. those rings are rated at about 200C and it was only about 100C. Also, if it is a problem i'll rebuild the tube since it had to be shortened and the extra inch would probably lower the temp another 50C or so
An even better method of sealing the quartz tube is a simple cap with a rubber washer in it. It's easy to remove when off and the stronger the vacuum, the harder it pulls the cap on and compresses the tube end to end.
9:15 you could test it at a variety of temperatures, measuring both its relation and its lag. if the lag is slow enough, the temperature difference isn't an issue, since you can fit a function to the relationship between them, to get one from the other. if it's mostly radiative, i don't think it would lag all that much, would depend on what contribution conduction makes to it.
I was thinking, once sealed with gas fittings, if this could be used as a Catalytic Reactor for things like Ethanol to Ethane (Similar to @HyperSpacePirate ) , CO2 Methanation, The Haber Process, and *maybe* Fischer Troph? Granted pressure/temp requirements are probably different so this may be a bit of a…PIPE DREAM, but I figured using an existing, and well made, device would be easier!
tube furnaces are very versatile in chemistry for controlling the atmosphere of reactions. i'm sure you could use it- however, you might want to use stainless fittings instead of aluminum due to reactivitiy
You can calculate the oxide formation rate by first calculating amount of Oxygen partial pressure in the tube and then looking at the ellingham diagram and checking the co-responding silicon line and its oxygen partial pressure needed at a given tempereature to form an oxide. You probably need to investigate how to read the ellingham diagram, it looks complicated but is very simple, there are some great youtube videos explaining it.
18:25 - Interested on making these compression tube fittings. Can you explain in more detail how this compression implementation for the fittings works? Thanks
You need to be using thermocouple extension wire of the same type as your thermocouple. Using soldered copper wire for your thermocouple cable will likely throw off your temps fairly significantly, as right now your cold junction is at the end of the flange. You'll have an error roughly equivalent to the temperature different between the flange and your thermocouple input, which is probably somewhere in the range of 20-80 kelvin of error
the flange is actually roughly at room temperature believe it or not. i was surprised at how quickly the heat dies off as you move away from the insulated part of the furnace
@@projectsinflight Oh wow, yeah I wouldn't have expected the temps to decrease so quickly that's pretty impressive insulating power. Still, probably good practice for accuracy's sake. Wouldn't want to be pulling your hair out at project randomly showing inconsistent results because of an uncontrolled +/-10C error
Did you use thermocouple wire for that lead? If you used regular wire that means your cold junction is at the end of the tube, and that will mess up the cold junction compensation.
the cold junction is just outside the tube and is roughly at room temperature. i was surprised it wasn't super hot but i guess the tube actually doesn't conduct heat very well
Another factor affecting the external thermocouples is the conducted heat loss through the thermocouple wires. The internal TC in nearly isothermal near the tip, so no lead conduction loss from the measuring point. A chunky gauge TC traversing a steep temperature gradient, with only the tip touching is the worst case scenario. Choose an R, S, or B type thermocouple which can survive an oxidizing atmosphere, which would permit using thinner wires, and then wrap the tip at least 1/4 or 1/2 around the tube, so the tip is not involved in any TC lead conduction gradients. My $0.02
No, go with the argon method. I hope he changes his mind, i must watch the rest. Well, i think i would have gone the argon method but it seemed to work well with vacuum also. Interesting and good video, thanks for sharing this stuff.
Idk why people keep saying borosilicate glass has such a high "melting point" It is higher than normal glass. But much lower than quarz glass. You even showed in the video that you can melt it with a normal torch ;D
The difference of the softening temperature is only 100K and the difference of the max working temperature only 30K. The advantage of borosilicate glass lies in the lower thermal expansion and higher chemical resistance
He can do that easily with the current setup, with a valve connected between the end cap and the vacuum pump: First pull a vacuum, then open the valve connected to argon to flhsh the tube with argon, then close it and pull another vacuum. The leftover oxygen should now be reduced by a few orders of magnitude!
I can only second this. One or two purge cycles and you can guarantee that the amount of oxygen molecules left in that thing is too low to even bother measuring.
Nice oven upgrade, was kinda hoping for a flow setup but going vacuum/batch is still plenty good in 90% of cases. As others have mentioned the math for V(O2) is off, also im pretty sure the unit is written mol. I have never seen Mol before.
yeah youtube removed the ability to add annotations to the video to correct things like this sadly. i pinned a comment which is the best i can do to correct this
Purge the pipe with a tiny trickle of argon gas. You can get the Oxygen content into the ppb atmospheric equivalent range with argon purge and simple vacuum.
You could calibrate the thermocouple result, by comparing it to the inside thermocouple. I believe it should correlate to the inside temperature linearly, so calibration would be easily doable
You could measure internal temperature by knowing relation between temperatures of internal to external couple. Should be something similar to multiplying your external thermal couple temperature by 2 point something and then adding couple of degrees, a simple linear regression or can be any other kind of regression. Or it doesn't have to be linear you could just record temperature readings plot the results and see the actual relation yourself then just use the interpolation.
isnt the nicely homemade compression flange on backwards? it looks like the o-ring sealing part wont make much difference? whats keeping the seal between the end of the glass tube and the big open gap between the halves of the compression fitting?
sorry, my diagram may not have done a good job showing it. the front oring (closest to the flange) does all the heavy lifting and seals against the tube, the front flange piece, and the coupling piece
You can use argon for much cheaper. I live in the US and got an argon setup for my lab. I have a 10 ish lb cylinder and I got it from airgas for like $200 ish and the regulator setup for like another $200 because I have both a flow regulator and a pressure regulator for pressurized inert atmospheres. It’s totally doable for not that much. Also if you want to get even cheaper, look up argon wine preservation. There are setups with decent sized cylinders for around 200$ for everything. One more thing, what vacuum pump are you using?
I'm concerned that the amount of argon required is high. i think you need at least 1cc/s flow rate, and i have no idea how much i'd need to refill the argon if i have to run several hour long diffusion steps
Could you add some material with a high surface to volume ratio that gets oxidized alongside the wafer and consumes the oxygen ? I'm thinking of copper dendrites
This is fascinating, would it be possible to add a material that sublimates into an inert gas either with temp rise or pressure drop to further remove oxygen ?
so, it's actually really interesting that you bring this up specifically. you see, the most common dopant for silicon is phosphorus pentoxide, which sublimates at 600C and reacts with the silicon to form a phosphorus-glass layer on the surface. what's interesting is that the dopant actually contributes small amounts of oxygen, so the need to totally purge the reaction chamber of oxygen is not really necessary
When pumping down to pressures much below about 1% of atmospheric pressure, the vast majority of the gas is water vapor and not residual nitrogen or oxygen unless there is a significant leak in the vacuum system.
that is probably true for systems at room temperature. i am not sure if it is true in this case, given the interior of the tube is extremely hot. it is possible that the water vapor is being boiled off aggressively and desorbed from the inner wall of the tube as a result. i suppose not all of the interior is at hundreds of degrees C though.
The end of the tube opposite from the pumping connection is near room temperature and so will be a continuing source of water vapor flowing through the tube and over the silicon wafers.
Now you could just flush the oxygen out with argon and then pull a vacuum, should work but might depend on how you plan to add the dopants. I've had some experience with tube furnace while making activated carbon using KOH and this flushing -> sealing method worked.
could you combine the two appraoches - flush with non-oxygen, and create a moderate vacuum - and do you need to use oxygen - could nitrogen be a cheaper flush?
Never use a sodium silicate as separating layer between heating wire and quartz, because fused silica is not immune and will react with it on surface creating some kind of soda glass with much lower melting point than a quartz. As heating for operating 1200deg C only option is Kanthal A1( and equivalents ) in ribbon form , not a wire . Another thing is quartz (or fused silica ) can work up to max 1000 deg C for exdended period of time , higher temperature cause recrystalization of quartz , tube became opaque and very fragile.
i believe that it's the improved flexibility of the glass fibers that increases the tensile strength of fiberglass. the fibers have some ability to move without cracking in a way that solid glass does not. Add to that the strength of the epoxy resin, and the fact that crack in a single glass fiber won't propagate to others like it would in a piece of solid glass.
wouldn't making a thread around the whole inner/outer piece be better for this? then you wouldn't have to worry about tightening the 4 bolts too much one by one, or not equally tightening them
@@projectsinflight therefor it didnt happen. lol my bad.. also i made that comment mid way.. i was thinking you should make external claps. great work man!
Why didnt you get another attachemt and conected argon, then you pull a vacum again, then you pump arogon again. You do this few times and most of the air in tube is argon.
Very cool! I'm wondering how your design stops air getting in between the front and back flange? I assume the front flange (in red on the diagram at 18:10) makes contact with one of the O rings and that's what creates a seal there, but that's probably one area worth looking at if you need a stronger vacuum in the future
a titanium sublimation pump would definitely work to remove air but i'd be better off using a turbo pump instead because of concerns over contamination
@@projectsinflight That fine brother. i was just asking can you do it? if you can what would i need that all iam asking brother. i know it melts silicon. i ask you becase your the best ON RUclips for this to ask.
i'd probably start with an aluminum oxide or graphite tube and i'd have to replace the sodium silicate with something else to contain the windings. speaking of- nichrome melts below 2000c so it would have to be replaced by something else- not sure what though.
Can I have your help? This weekend I melted my expensive high borosilicate tube. Can you please point me to the aliexpress seller that you are using. I am from Portugal (europe) so buying from McMaster-carr is too expensive to ship to Portugal.
3:43 I made an error: It's 22.4L/Mol, not 22.4Mol/L. Sorry for the confusion.
That also means the math should be a division instead of a multiplication right? otherwise the units don't check out. I think that factor of about 500 on the final result does affect the conclusion of the calculation. Still a really great video, and let's be honest most people come here for watching the experimental work. Thanks so much for making this really intersting series!
22.4 L/mol is the value for a temperature at 0°C, while the value at 25°C is 22.4 x (273K+25K)/(273K) = 24.5 L/mol. For a temperature in the tube furnace of 1000°C, the molar volume of an ideal gas is 22.4 x (273K + 1000K)/273K = 104.5 L/Mol.
Just to add what the others said, instead of using a rather derived value for one temperature and pressure, just use the ideal gas equation directly. Ideal gas is fairly accurate for these conditions,
@@testingmaterials3242 ⚠ God has said in the Quran:
🔵 { O mankind, worship your Lord, who created you and those before you, that you may become righteous - ( 2:21 )
🔴 [He] who made for you the earth a bed [spread out] and the sky a ceiling and sent down from the sky, rain and brought forth thereby fruits as provision for you. So do not attribute to Allah equals while you know [that there is nothing similar to Him]. ( 2:22 )
🔵 And if you are in doubt about what We have sent down upon Our Servant [Muhammad], then produce a surah the like thereof and call upon your witnesses other than Allah, if you should be truthful. ( 2:23 )
🔴 But if you do not - and you will never be able to - then fear the Fire, whose fuel is men and stones, prepared for the disbelievers.( 2:24 )
🔵 And give good tidings to those who believe and do righteous deeds that they will have gardens [in Paradise] beneath which rivers flow. Whenever they are provided with a provision of fruit therefrom, they will say, "This is what we were provided with before." And it is given to them in likeness. And they will have therein purified spouses, and they will abide therein eternally. ( 2:25 )
⚠ Quran
@@leocurious9919 hi. when you multiply it by 1/10000 do you take into account that you have to convert it to pa? or should you? I was trying to make similar calculations but my vacuum pulls 3.9996 Pa
Can one of the big chip manufacture please please get this lad sponsored and offer him tours through a production facility it would be great to see how much a guy like this gets out of such a hands on tour.
you have NO IDEA how many un-google-able questions i need answered by someone in the industry
@@projectsinflighta number of universities have small scale chip fab labs, try emailing them perhaps?
The small scale might be closer to your process and teachers often love to teach someone who has shown as much interest as you have.
I think you might be getting a stress riser where the end of the tube meets your metal plug. The atmosperic pressure ends up bearing on a very small area. You could put an o ring there as padding. And make sure there is tolerance for the metal to expand. This is a 1 bar pressure seal so you could leave the o rings a bit proud and it will still seal. If the metal plug heats up enough to put two opposing point loads on the inside of that glass hoop - you're hooped!
In the final design, the glass is contacted only by the o-rings, which squeeze around the tube and prevent the metal flange from pushing against the glass.
I was actually looking at the final photo and a little worried about having all that mass out on the two ends. Are you at all worried about the tube snapping at the point it meets the furnace?
The tube seems pretty strong and i don't currently expect it to snap from the weight. plus this is how they do it in industry so it's *probably* fine
until i make a new video and break the tube again
Adding a vacuum reservoir tank and valve inline with the pump would allow you to "precharge" the vacuum system while firing the furnace, and save you those 2-3 minutes of time it takes to draw everything down, perhaps thinning the oxide layer a bit more.
this might work, though it's likely easier to use a nitrogen or argon purge to get the O2 out quicker
Sure, but you said you didn't want the expense of having to source welding gas. Reusing an old tank is pretty cheap and effective and never needs refilling.
This kind of project (making your own silicon chips) absolutely isn't in my wheelhouse, but I'm so down for all this problem solving you do on the channel. These videos are such a good demonstrations on how to tackle complex problems and how to be resilient in the face of constantly "failing" towards success.
Can't wait for the next one, really happy I found your channel :)
hopefully i'll get the next one cranked out in less than a month!
You are a great storyteller! It's one thing to make something, it's completely something else to record it in a way thats engaging for a general audience.
Man, this project is sick! I can't believe it doesn't get more views, especially with how well laid-out and well-explained it all is. Keep up the good work!
slowly working my way through the youtube algo lol
I would really enjoy a full video about machining the end cap fittings :)
And as always, really well made video. Keep up the great work!
i might get around to making a more detailed video about it
You're doing much hard, accurate and smart work...
doing my best! i occasionally make errors and i appreciate people for being patient with me as well :)
Small note, although probably insignificant:
EDIT: this might be a larger problem because depending on the wires used, you end up measuring the furnace temperature in reference to the flange. See sub-comments. Usage of thermocouple extension wire is recommended in this case as the termination point temperature is for sure above the board reference temperature. On the other hand if you are fine with having 50K uncertainty (I'm assuming the flange is 50C above ambient), it should be fine.
In presence of temperature gradients (which in case of a furnace end cap will for sure be the case) you should not solder thermocouple leads as it creates a unpredictable set of thermocouples at the soldering point. Depending on how precise you want to go, the optimal solution is to use a specially designed thermocouple wire (for given thermocouple type) and spot welding it to the thermocouple tail, a mechanical connection can also work.
At the board, the connection point is less critical if you can guarantee that there are no thermal gradients, but type-compensated connectors are also available.
I was about the say the same.
Because you soldered wires at the flange, you will be subtracting the temperature of the flange from the measured temperature, so the more the flange heats up the more error you will have.
For this reason they make so called "Thermocpuple extension wire" it is made out of the same material as the thermocpuple, so that it has no thermal EMF, this brings your cold junction all the way to the thermocouple connector (Where you can measure the connectors temperature to compensate for it)
Thinking about it more, you are right. If we simplify the model down to only copper from the flange, then the termocuople measures temperature of (Furance - Flange) but the calculation right now probably assumes (Furnace - Ambient/uC board).
The temperature of the solder joint is roughly ambient temperature, even with the furnace cranked to it's max temp. I can double check again the exact temp, but i'm pretty sure it's cold enough not to be a huge problem.
Yes, you built another pair of thermocouple junctions!
Yea, it will add a temperature error to that (roughly the temperature difference between the ends of the copper involved).
Thermocouple extension wire recommended to avoid this
I came to the comments to say this too
Suggestion if you ever rebuild the thermocouple assy: instead of using an epoxy paste like the JB weld, use an epoxy liquid. There won't be as many entrained bubbles (you can improve on this a lot more if you use your vacuum pump to pull the micro bubbles out too), but you will have a much smoother surface finish. This is important because oxygen will be diffusing from the surface of the epoxy in a good vacuum. The liquid epoxy also reduces the risk of a leak along the solid wires. If you end up using stranded wire for something the liquid can flow between the strands and still give a pretty good seal (though you might want to spread the strands a bit).
Keep up the good work. BTW: I have gotten that HPLC going and I still plan on running that PCB resist you sent me on it to see what happens, but I will want to get a bad column to run it on first so I don't risk a good one.
yeah, that might be a better method. i like JB weld because it *seems* very compatible with vacuum, but it is true that it's a lot more viscous than i'd like
JB Weld works WAY better with vacuum than most off-the-shelf epoxies, especially the low viscosity ones. It's the right choice here, and it tends to cure very clean, without bubbles.
Hysol EA 1C is better than JB Weld, but a bit more expensive(2x).
@@projectsinflight You can try acetone to thin it out.
@@abdullahanton Isopropyl alcohol works well, but any solvent will cause outgassing from the seal.
A lot of excellent problem-solving again. I loved your vacuum development process.
I was sad that you got caught out by the scam-like nature of modern glass, specifically that of faux borosilicate. Pyrex especially (which I don't know if you were using but is relevant) built a business of trust in the 20th century that their glassware was made of the high-quality material, but around the turn of the millennium, started allowing more use of their name on regular, low-quality glass. Even today, people are surprised when their Pyrex dishes shatter under normal, medium-temperature conditions because they'd owned other Pyrex dishes from 25+ years ago that had survived for so long.
It's no doubt that most lab equipment companies now are more than happy to allow consumers to assume they're made of materials that had become standard but are instead no more robust than a $3 glass coffee carafe.
But enough of my ranting. I love watching your processes.
yeah, that was unfortunate. I guess i need to be more careful with my suppliers in the future. glad you liked the video!
Not sure if it’s related, but the Pyrex brand includes regular soda lime glass in addition to borosilicate. That might be part of the confusion.
@@GRBtutorialsYes, soda-lime, that's it. I couldn't remember the name
It could have been borosilicate glass. Borosilicate glass should only used up to temperatures of 490°C. At 817°C it already softens and 1000°C is just overkill. Only quarz glass can survive these temperatures
@@hantrio4327 This is what I was going to say. The annealing temperature for clear boro is around 1050 F, or about 560 C. Anything above that is soft enough to deform, given enough time, and the vacuum inside will just make it deform more quickly than it would under just gravitational force. The softening point is listed online as about 820-825 C, but that's just the point where it reaches a specific arbitrarily specified viscosity. It's still well below 1000 C, though.
Damn, you are such an inspiring person man! Really loving what you do and breaking the limits of what a single person can as hobby! Please don't get discouraged because of less views and following, you will surely get the fame you deserve one day.
working my way through the youtube algorithm lol
The way you explain your steps and assumptions, along with mentioning failures and successful results, is very helpful and interesting.
Thanks for sharing such videos
thank you! it's been fun sharing one of my projects with everyone :)
Fascinating work. Ive used quartz rod for a clock pendulum due to its exceptionally low thermal expansion, nice work on the end caps. Hope the heatshring on the thermocouples doesn't leak vacuum, you can get ceramic feedthroughs but also glue lined heatshrink .
oh the heatshrink isn't doing any sealing- the epoxy takes care of that so it's not a problem. that being said i do want to learn more about non porous ceramics
Man I have no real idea what this project is really for but I am impressed with your build skills and problem solving. Subbed!
goal is to build an integrated circuit eventually
20:15 I think someone else mentioned this, but depending on how hard of a vacuum you intend to pull this *could* be an issue.
There are Vacuum Rated Epoxies/RTV Silocones for Vacuum Use if i remember correctly…granted for a price.
That “Faraday Wax” someone on-RUclips made may be worth a look, that and/or a less “witches brew” and moreso just “this fraction of paraffin wax” or something type recipe may be neat.
GRANTED “if it ain’t broke don’t fix it”, and the build is already coming along great!
Just thought I’d throw this out there in case you didn’t already know about that and “virtual leaks” and all that.
Keep up the great work!
I've actually used JB weld in high vacuum setups without issue. I've also got some corning high vacuum grease too
It will never be as accurate, and it might suffer from confounding circumstances, but you should be able to run a series of heating and cooling cycles, collecting temperatures from both thermocouples, and develop a calibration curve that makes the permanently installed thermocouple useful. As long as this tube survives this means you could potentially make it workable, while still having access to both ends, if say you wanted a setup with gas inlet at one end, and the vacuum pump on the other, or you wanted to develop a cap that somehow lets you push and pull the boats while they're already under modified atmosphere
right now i care more about accuracy than convenience, but i will continue to try to find a way to get the pesky thermocouple outside the tube
@@projectsinflightIf you're really interested in temperature accuracy and removing the oxygen from the tube, i would move the thermal couple from one end and put it in the center, but cut a hole in the center quartz tube on one side to fit it. This will allow you to keep as many nichrome loops around the tube in the center. Submerge the tube under water to allow for heat and stress dissipation without cracking the tube and cut a hole with a glass bit in the center, seal it off with JB weld extreme heat which should be fine for 1,200C. I would then add a tube and a flage to the other end so you can purge the tube with inert gas after pulling a vacuum, then pulling the vacuum again. Argon can be obtained relatively cheaply (less than $20) from wine preserver bottles. It also gives you the option to use the other end for other projects like if you want to record the doping process with a camera
Thanks for sharing all of these DIY semiconductors content! I couldn't find any source on internet, which shows whole process keeping it simple and at the same time comprehensively discussing the topic. Just watched all videos on your channel and now patiently looking forward for updates on project. I also hope for some mosfets, because they're what's really used because of their energy effieciency. You could even make SSD like memory with NANDs if you will be able to grow floating gate on SiO2 to hold charge and therefore store data and another SiO2 layer on top of that for insulation and finally gate contact on top to change state of memory:)
This was super cool to watch! Excited to see where the experiments will go.
thank you! i am having a lot of fun making these :)
Hey, great video again! Few suggestions below, but definitely include lathe work videos and keep up the amazing work!
You can create an offset calibration curve for the external thermocouple, this is how many lab furnaces operate, as the hot spot is not usually very large and you're planning on putting the sample in the middle of it, giving a temperature offset anyway! I do this in my lab furnace which reaches 1500C so those radiation issues are massive.
For your thermocouples, just buy thermocouple extension wire of whatever type you want, then use an acetylene torch (or spot welder) to melt the two wires together. Bam, now you have any length you like without any issues, and can shorten very easily (cut and remelt).
Either a vacuum charger, or nitrogen/argon flush of the tube while loading might help remove some oxide growth. You could use your blank end to provide the flush?
I thought about doing that, but I was wary of adding any more sources of noise in the thermocouple measurement, especially since an offset of 10 or 20 degrees C can be pretty devastating to some processes.
i like the idea of the nitrogen flush while loading- i may look into that
@@projectsinflight I would say that in my experience getting temperature readings that precise in a furnace is possible, but requires careful measurement what ever way you approach it. We use an internal probe like you do to calibrate our carbolite furnaces, but our atmosphere would destroy them if we left them in, so we rely on the external sensor (used by the carbolite controller) to read and create a compensation curve for the internal conditions. This temperature offset is also a function of position along the tube so we take samples along the length to find the hot spot and it's size/shape, AND we even build test geometries of our crucibles and powders with the thermocouple buried inside to verify the final performance, as just putting a boat/crucible in causes a significant change in the radiation flux compared to no boat.
If you're getting noise, this is probably due to electrical issues, I would definitely recommend borrowing a proper thermocouple meter to just verify your Arduino setup is right too
The JB weld should definitely be fine outgassing-wise at these temperatures. I'm not sure if you said so, but make sure those threaded connections are sealed with either teflon tape (if they are NPT) or epoxy. Looks really good!
i like teflon or corning high vacuum grease for the task
Jb weld will most likely outgas under vacuum. Torr seal is a vacuum grade epoxy.
Outstanding video as always. Great job on the clamps. 👍
thanks! i very much enjoyed machining them :)
days when you upload are always good days
I wish i could upload every day ;)
I can see the future. 😊
Welcome back... Last video I made flanges. In today's video we will be adding a turbo molecular pump.
I *probably* won't need to do this but i do actually have a turbo pump ready to go
@@projectsinflight stop it! Wait you built it didn't you.
Hi. Verry nice video. I have some experience using tubular funaces in our lab. Though we are not using it for Silicone manipulation but for chemical syntheis. We are using a self made compression flange as well and I think our design is much simpler to make. If you are still interested I can share the design with you. Some other comments. If you start using gas flow I suspect that your O-Rings will melt . Even with massflowcontrollers, you have a significant heat transfer to the outgoing end at lest when you are using the furnace for days or weeks like we do regulary. I even had FFKM O-Rings melt on me! Also be aware that the glowing quarz tube creates significant UV-Radiation over 900°C wich will go into your flanges and heat them up eventually. Also if you wand to use Argon or other gases sure you will need a primary reduction valve but after that a needle valve and a buble counter with silicone oil on the other end will give you relativly fine controll over the gas flow. It might be enough for you. Finaly please!!! do not heat you quarz tube over 1100°C for longer than an hour. If you go longer you will reach the strain point of the Quarz wich basically creats a silicone oxide cristal growth chamber. Basically the glass will start to realine into its natural cristal structure on the surface of your tube wich creates strain over time, wich will eventually brake the tube. I have lost so many tubes because unknowing scientists were doing yust that.
this is very useful information! if you don't mind, would you send me an email? projects in flight at gmail dot com? i'd love to talk about the flange design
Super impressive build and explanation along the way! Great work!😊
Glad you liked the new flanges!
Even the cleanest welding argon 5.0 can have impurities in some bottles. I had troubles with arc stability during aluminum TIG welding. I was prepared to send the machine for servicing, but after receiving advice to change the bottle, I was able to weld properly again.
yeah i heard there is even some water in welding argon, which would be a huge problem for oxide growth (worse than o2 gas)
we use welding grade argon in the lab and dry it with a column of indicating silica gel. @@projectsinflight
@@projectsinflight I'm no expert, but I wonder if something like a molecular sieve filter would help clean up the argon (and maybe vacuum if ever needed). No idea the cost to buy or fab a filter for your purposes, it might not be practical. But I've been wondering where I'd see practical uses for them since Tech Ingredients used 3A sieves to concentrate alcohol a few years back and from what I find online, tube furnaces sometimes use them.
@@straphyrthat is how it's done I believe even for fairly high purity argon it can still be quite wet.
Great work, keep it up!
thanks!
a vacuum muffle furnace.. BRILLIANT Thanks for sharing this with us
Thanks" Glad you liked the new furnace tube :)
In my lab we have to work with oxygen free atmospheres in tube furnaces, though in my case the furnace tube is alumina so that we can reach 1500 degC. One of the simplest ways to go is to use what is called a test plug. It is an expandable rubber stopper with a gas feedthrough. That works fine, but oxygen does still get in, so we would place a dish of titanium sponge in a hot, but not too hot zone of the tube upstream of the sample. This did a nice job reacting with any oxygen that would get in.
But in the end, I bought two weldable quick connect flanges and used a liquid engine gasket compound to make an airtight seal between the flange and the tube. This holds vacuum very well. But with all that said, flowing argon at a slight positive pressure is not very expensive really and works way better than vacuum to keep oxygen out. The $1,000 for argon and a regulator is way overpriced unless you are talking about buying a large gas cylinder.
yeah, i may have overestimated the cost a tiny bit. also, interesting that i didn't try a simple rubber stopper. i guess it just didn't come to mind. not sure how i'd have attached the flare fitting though
Great math around the residual oxygen at the vacuum level, but you can flush the tube with Argon and then pull the vacuum, which means there shouldnt be more than 1E-5 oxygen that was in the Argon, which should be much less.
edit: P.S. great project and thanks for sharing.I wish they were longer and more detailed, I just love the combination of engineering, DYI and physics!
thank you! i plan to try the argon purge soon
Beautiful machining, very impressive
Argon does not need 2 regulators, all those for welding already include what you call secondary regulator. Also, since this is so wildly used for welding, it is dirt cheap. You can also use any hose you want, since you are not working with pressures.
Note that you can turn on the vacuum pump seconds/minutes/hours before you actually need to use it. You do not need to turn it on right as you seal the vessel.
standard welding regulator dispenses too much gas. i only need a tiny bit, like on the order of a few cc/min. that being said i probably overstated the cost associated with argon a bit and i'll likely add a purge input for inert gas to the setup
@@projectsinflight Measuring that low will not be possible with a normal regulator (with rotameter), correct. But for those low values just bubble it through water to set it to what you need. If you bubble the output through water you not only meter the flow but also make sure to always have a bit of positive pressure in the system.
Wow, such a nice video and so well presented! Thanks for the insights on glass seals. Now that you have a vacuum tube furnace, you could try putting some H2 inside, that would actually reduce any oxides, and even if your substrate was not clean enough, you could "clean" it in-situ!
Also, for the external thermo-couple, yes, it doesn't see the same amount of infrared radiation, but doesn't it get due it heat diffusion at some point up to the temperature? If not, you could actually calibrate it, knowing what you get from the internal one, and not throw it away. Anyways, very fond of your channel, best regards.
"it was slightly colder than before and the o-rings were firmer as a result"
NASA management: "Write that down! write it down!"
That region is DOPE!!
Wow nice, you are a smart boy my friend! 🙌🏻🙏🏻 love to see what people are up to! You getting a new subscriber!👍🏻
much appreciated!
Youve now got a good vacuum system, and a hot furnace. I hear the sounds of CVD.
i'm scared of the toxic chemistry involved right now but i might try it
Can you check if termal diference is constant if it is you can just add the the diference
it may work, but i don't want to introduce any unnecessary sources of noise in the measurement
If you make heatsinks and clamp them on the tube right before your flanges, then you can prevent the heat being conducted all the way to the O-rings. Just in case if it ever becomes a problem.
if it doesn't happen at 1200C i doubt it'll ever be a problem. those rings are rated at about 200C and it was only about 100C. Also, if it is a problem i'll rebuild the tube since it had to be shortened and the extra inch would probably lower the temp another 50C or so
Can't wait to see the diffusion!
same here! it's been a long road to get here
An even better method of sealing the quartz tube is a simple cap with a rubber washer in it.
It's easy to remove when off and the stronger the vacuum, the harder it pulls the cap on and compresses the tube end to end.
I like what you are doing, please continue.
9:15 you could test it at a variety of temperatures, measuring both its relation and its lag. if the lag is slow enough, the temperature difference isn't an issue, since you can fit a function to the relationship between them, to get one from the other. if it's mostly radiative, i don't think it would lag all that much, would depend on what contribution conduction makes to it.
i thought about doing that but i don't really trust myself to not introduce a significant (>20C) error
@@projectsinflight fair enough
Man i love this channel! keep going !
thanks! I plan to :)
I was thinking, once sealed with gas fittings, if this could be used as a Catalytic Reactor for things like Ethanol to Ethane (Similar to @HyperSpacePirate ) , CO2 Methanation, The Haber Process, and *maybe* Fischer Troph?
Granted pressure/temp requirements are probably different so this may be a bit of a…PIPE DREAM, but I figured using an existing, and well made, device would be easier!
tube furnaces are very versatile in chemistry for controlling the atmosphere of reactions. i'm sure you could use it- however, you might want to use stainless fittings instead of aluminum due to reactivitiy
It may even be easier to add a internal metal pipe for the process chemistry but the tube furnace is a great way to heat the reaction.
orings must rattle on the shaft, glass filled teflon so cool!
You can calculate the oxide formation rate by first calculating amount of Oxygen partial pressure in the tube and then looking at the ellingham diagram and checking the co-responding silicon line and its oxygen partial pressure needed at a given tempereature to form an oxide. You probably need to investigate how to read the ellingham diagram, it looks complicated but is very simple, there are some great youtube videos explaining it.
18:25 - Interested on making these compression tube fittings. Can you explain in more detail how this compression implementation for the fittings works? Thanks
Can you do a workshop tour? I would be really interested in how you use your mill and lathe...
i'll do that just as soon as i get a scanning electron microscope ;)
You need to be using thermocouple extension wire of the same type as your thermocouple. Using soldered copper wire for your thermocouple cable will likely throw off your temps fairly significantly, as right now your cold junction is at the end of the flange.
You'll have an error roughly equivalent to the temperature different between the flange and your thermocouple input, which is probably somewhere in the range of 20-80 kelvin of error
the flange is actually roughly at room temperature believe it or not. i was surprised at how quickly the heat dies off as you move away from the insulated part of the furnace
@@projectsinflight Oh wow, yeah I wouldn't have expected the temps to decrease so quickly that's pretty impressive insulating power. Still, probably good practice for accuracy's sake. Wouldn't want to be pulling your hair out at project randomly showing inconsistent results because of an uncontrolled +/-10C error
Did you use thermocouple wire for that lead? If you used regular wire that means your cold junction is at the end of the tube, and that will mess up the cold junction compensation.
the cold junction is just outside the tube and is roughly at room temperature. i was surprised it wasn't super hot but i guess the tube actually doesn't conduct heat very well
Great adjustments learned a little bit good video
thanks!
This is amazing, man!
Fascinating. Thank you for the great effort.
thanks!
Another factor affecting the external thermocouples is the conducted heat loss through the thermocouple wires. The internal TC in nearly isothermal near the tip, so no lead conduction loss from the measuring point. A chunky gauge TC traversing a steep temperature gradient, with only the tip touching is the worst case scenario. Choose an R, S, or B type thermocouple which can survive an oxidizing atmosphere, which would permit using thinner wires, and then wrap the tip at least 1/4 or 1/2 around the tube, so the tip is not involved in any TC lead conduction gradients. My $0.02
Any ideas on how to get compression fittings without a lathe? Or are you pretty much just stuck having to buy them
i don't really think it's possible to make your own without a lathe, but you can always use the PVC pipe caps method instead
@@projectsinflightthe kf flanges have 24/40 glassware adapters which is what i need, pvc cap could just use a hose to the glassware I guess
No, go with the argon method. I hope he changes his mind, i must watch the rest.
Well, i think i would have gone the argon method but it seemed to work well with vacuum also.
Interesting and good video, thanks for sharing this stuff.
Usually most welding supply shops sell small Argon bottles for hobbyist that doesn't cost that much. Argon setup for this purpose is 200 dollars max.
i've been contemplating an argon purge before the vacuum as an economical fix
Idk why people keep saying borosilicate glass has such a high "melting point" It is higher than normal glass. But much lower than quarz glass. You even showed in the video that you can melt it with a normal torch ;D
that may be true, but after double checking these tubes are actually not even borosillicate at all
The difference of the softening temperature is only 100K and the difference of the max working temperature only 30K. The advantage of borosilicate glass lies in the lower thermal expansion and higher chemical resistance
You should try to make an end cap with 2 tubes, one goes to the vacuum pump and the other goes to an argon tank. Just saying
He can do that easily with the current setup, with a valve connected between the end cap and the vacuum pump: First pull a vacuum, then open the valve connected to argon to flhsh the tube with argon, then close it and pull another vacuum.
The leftover oxygen should now be reduced by a few orders of magnitude!
@@trabladorr You're right 👍, my bad fam
i've definitely been considering adding an argon purge for this reason
@@projectsinflightyou read my mind. Argon purge for the win!
I can only second this. One or two purge cycles and you can guarantee that the amount of oxygen molecules left in that thing is too low to even bother measuring.
Nice oven upgrade, was kinda hoping for a flow setup but going vacuum/batch is still plenty good in 90% of cases.
As others have mentioned the math for V(O2) is off, also im pretty sure the unit is written mol. I have never seen Mol before.
yeah youtube removed the ability to add annotations to the video to correct things like this sadly. i pinned a comment which is the best i can do to correct this
Purge the pipe with a tiny trickle of argon gas. You can get the Oxygen content into the ppb atmospheric equivalent range with argon purge and simple vacuum.
I saw a Video from breaking taps about copper metalization with a Diode laser you can use that to make the conductive traces on the silicon
You could calibrate the thermocouple result, by comparing it to the inside thermocouple. I believe it should correlate to the inside temperature linearly, so calibration would be easily doable
i thought about it but i right now am not confident i can achieve the same temperature accuracy with such a method. perhaps in the future
You could measure internal temperature by knowing relation between temperatures of internal to external couple. Should be something similar to multiplying your external thermal couple temperature by 2 point something and then adding couple of degrees, a simple linear regression or can be any other kind of regression. Or it doesn't have to be linear you could just record temperature readings plot the results and see the actual relation yourself then just use the interpolation.
yeah i didn't trust myself not to introduce another significant (>20C) temperature error so i didn't try that
isnt the nicely homemade compression flange on backwards? it looks like the o-ring sealing part wont make much difference? whats keeping the seal between the end of the glass tube and the big open gap between the halves of the compression fitting?
sorry, my diagram may not have done a good job showing it. the front oring (closest to the flange) does all the heavy lifting and seals against the tube, the front flange piece, and the coupling piece
You can use argon for much cheaper. I live in the US and got an argon setup for my lab. I have a 10 ish lb cylinder and I got it from airgas for like $200 ish and the regulator setup for like another $200 because I have both a flow regulator and a pressure regulator for pressurized inert atmospheres. It’s totally doable for not that much. Also if you want to get even cheaper, look up argon wine preservation. There are setups with decent sized cylinders for around 200$ for everything. One more thing, what vacuum pump are you using?
I'm concerned that the amount of argon required is high. i think you need at least 1cc/s flow rate, and i have no idea how much i'd need to refill the argon if i have to run several hour long diffusion steps
Could you add some material with a high surface to volume ratio that gets oxidized alongside the wafer and consumes the oxygen ? I'm thinking of copper dendrites
any copper in the tube would risk contaminating the silicon. Copper ions are a bad dopant and ruin its electrical properties
@@projectsinflight any material that suits your constraints !
Nice work on the flanges. Any concerns about their weight causing the tube to deflect?
so far it seems to be just fine. i was considering adding some supports to the tube near the flanges though
Great series. Please show the lathe work next time.
if people are interested i can make a whole video and making those flanges
Very interesting. Nice work
This is fascinating, would it be possible to add a material that sublimates into an inert gas either with temp rise or pressure drop to further remove oxygen ?
so, it's actually really interesting that you bring this up specifically. you see, the most common dopant for silicon is phosphorus pentoxide, which sublimates at 600C and reacts with the silicon to form a phosphorus-glass layer on the surface. what's interesting is that the dopant actually contributes small amounts of oxygen, so the need to totally purge the reaction chamber of oxygen is not really necessary
@@projectsinflight that's really cool! I look forward to seeing where this goes!
sir can i have the schematic diagram for the wiring please
This “vacuum gauge digital reader” wher did you bay this one?
Uniweld UVG. Not sure if they are still making them but the company was super helpful when i troubleshooted it (wasn't the gauges fault it turned out)
@@projectsinflight thanx bro🙌🏻🙏🏻
what kind of glue are you using to stick coil heater element? thank you
refractory cement
When pumping down to pressures much below about 1% of atmospheric pressure, the vast majority of the gas is water vapor and not residual nitrogen or oxygen unless there is a significant leak in the vacuum system.
that is probably true for systems at room temperature. i am not sure if it is true in this case, given the interior of the tube is extremely hot. it is possible that the water vapor is being boiled off aggressively and desorbed from the inner wall of the tube as a result. i suppose not all of the interior is at hundreds of degrees C though.
The end of the tube opposite from the pumping connection is near room temperature and so will be a continuing source of water vapor flowing through the tube and over the silicon wafers.
I would add rubber grommets to the wire passthru holes, where you used glassfibre tubing. Timestamp 8:00
yeah i thought about that. i might try since there is a possibility of abrasion
Now you could just flush the oxygen out with argon and then pull a vacuum, should work but might depend on how you plan to add the dopants. I've had some experience with tube furnace while making activated carbon using KOH and this flushing -> sealing method worked.
i'll have to look into this. it seems like a good idea
Also more machining videos please
if people want them i'd be happy to provide
Well, now I know what I need to add to my (mental) design of a tube furnace lol
eventually i'll iterate through everything and make a single video on the best way to build the furnace lol
What's the resistance of a wafer that's just freshly out of the HF bath?
about 10-30k
For oxide thickness have you tried a weigh strip weigh method?
not sure what you mean
could you combine the two appraoches - flush with non-oxygen, and create a moderate vacuum - and do you need to use oxygen - could nitrogen be a cheaper flush?
Never use a sodium silicate as separating layer between heating wire and quartz, because fused silica is not immune and will react with it on surface creating some kind of soda glass with much lower melting point than a quartz. As heating for operating 1200deg C only option is Kanthal A1( and equivalents ) in ribbon form , not a wire . Another thing is quartz (or fused silica ) can work up to max 1000 deg C for exdended period of time , higher temperature cause recrystalization of quartz , tube became opaque and very fragile.
if glass sucks at tension, then how come fiberglass is so strong?
i believe that it's the improved flexibility of the glass fibers that increases the tensile strength of fiberglass. the fibers have some ability to move without cracking in a way that solid glass does not. Add to that the strength of the epoxy resin, and the fact that crack in a single glass fiber won't propagate to others like it would in a piece of solid glass.
wouldn't making a thread around the whole inner/outer piece be better for this? then you wouldn't have to worry about tightening the 4 bolts too much one by one, or not equally tightening them
it would be an appropriate solution, but i chose to make it this way because i saw it as a slightly easier machining operation
hey man hit the groves on the brass bung with some Emory board to slightly champher the material so it doesn't cut your o rings!
i hit them with a chamfer tool but didn't show that clip in the video
@@projectsinflight therefor it didnt happen. lol my bad.. also i made that comment mid way.. i was thinking you should make external claps. great work man!
if you are going to cut glass you need to make sure the glass is wet.
Why didnt you get another attachemt and conected argon, then you pull a vacum again, then you pump arogon again. You do this few times and most of the air in tube is argon.
For now, the vacuum alone seemed ok, but i want to upgrade to argon in the future
what if you flush the tube with an enert gas and then pull the vacuum?
i plan to try that soon actually
@@projectsinflight can't wait to see the result!
Very cool! I'm wondering how your design stops air getting in between the front and back flange? I assume the front flange (in red on the diagram at 18:10) makes contact with one of the O rings and that's what creates a seal there, but that's probably one area worth looking at if you need a stronger vacuum in the future
Sorry my diagram kinda sucks. The front o-ring does all the heavy lifting and seals in three places: the front flange, the back flange, and the tube
love you videos you have gaven me alot of adias
thanks! glad to be able to make interesting things
This is a pretty small volume. Thermal sputtering some titanium wire might get you a much better vacuum.
a titanium sublimation pump would definitely work to remove air but i'd be better off using a turbo pump instead because of concerns over contamination
@@projectsinflight I was just thinking a bit of wire and a high voltage is a bit easier to come by.
O-rings and low temperatures, I think I’ve heard that before
thank god lives don't depend on my engineering
can you made one for 2000c?
no point for silicon or silicon dioxide as it melts well below that. for those temps you'd probably want an aluminum oxide tube i think
@@projectsinflight That fine brother. i was just asking can you do it? if you can what would i need that all iam asking brother. i know it melts silicon. i ask you becase your the best ON RUclips for this to ask.
i'd probably start with an aluminum oxide or graphite tube and i'd have to replace the sodium silicate with something else to contain the windings. speaking of- nichrome melts below 2000c so it would have to be replaced by something else- not sure what though.
@@projectsinflight can i paypal you for this
Can I have your help? This weekend I melted my expensive high borosilicate tube. Can you please point me to the aliexpress seller that you are using. I am from Portugal (europe) so buying from McMaster-carr is too expensive to ship to Portugal.
please email me and i'll try to help. my email is projectsinflight (at gmail dot com). its also shown at the end of the lithography video
@@projectsinflight thanks for your reply. I just sent it to you
15:18 Wait I’ve seen this one before…
i cry every time