There's definitely been a small number of people doing this over the years, but a lot of components have become more accessible. I doubt TEOS was widely available in the past
If I knew it was you getting the TEOS, I would have let you have it for free. Whenever you need more, let me know and I'll donate it to you absolutely free of charge.
I knew that was your name!! You really are the man for some unique and quality stuff. I'm happy that we're acquainted at least in a limited manor. I hope this gets a heart and a pin.
well if i'm understanding correctly i must have been the only one to buy it in the last couple months so a sharp uptick would be like... 2 more orders lol
I would love if you sent the wafers over to breaking taps or some other youtuber with an electron microscope, I think it would be really intresting to see how the glass layer changes at every step.
@@projectsinflight Im a college student with access to a lab that can conduct various surface characterization techniques, if you're curious on any other surface properties :) I've part of a research project that involves etching silicon wafers so we have experience with it.
Thomas Edison once said "Genius is 1% inspiration and 99% perspiration." If the amount of effort that you put in and the perseverance that you demonstrate when you find a dead end are any indicator, you sir are a genius.
A perhaps more relevant Edison story is that he (and his lab) spent most of their time making better vacuum pumps when he began experimenting with incandescent bulbs. They got several patents on that before they had any useful results on the incandescent bulbs themselves.
When you heat boric acid with alcohols you get the corresponding borate esters. So you could make trimetyl borate first add that instead of boric acid and balance the pH with nitric acid. Also the boric acid is way less acidic that the phosphoric acid so you should have already been adding another acid to get the same pH.
at 33:18 you can seem my Trimethyl Borate I made just for this purpose :) It works well, but I didn't want to overcomplicate the process in this video, so it's gonna have to be in the next one
it's so crazy- all the people who know how to make the stuff using these methods are 70+ years old. These days it's all ion implantation and vapor deposition
I have worked extensively with dispersants for silicon nanoparticles; I am unsure of how well this would work in general but the long and short of it is that toluene and silicone oil are both effective dispersants for fumed silica. the ineffective behavior of the isopropanol dispersant could have been due to the solvent being too thin, and thus the nanoparticles started to settle too quickly on the spinning plate. Miscibility of the aqueous phosphoric acid into such a solution would still require experimentation, but I do know isopropanol is miscible with silicone oil?
As someone who work in a lab, I approve the 2:07 Sigma Aldridge tagline. And the 'dead-ends and complications that didn't made it' just resonates too well with me. Great video as always, keep them coming.
6:08 Small correction: the diagram is labeled with a flow of protons (H+) instead of holes (o+), which is definitely wrong even in a simplified explanation
Great work. The editing and cadence was amazingly easy to watch/listen to. Like a smooth train ride. I look forward to what you come back with in a few to several months!
I find it funny that with a reasonable amount of background knowledge, some basic properties, and a bit of testing, it's possible to reverse engineer most proprietary products
Sadly, modern CPUs are practically immune to reverse engineering. For example, no one knows if rdrand instruction of Intel CPUs really works as advertised.
there is an emerging technology of super high resolution CT scans and i suspect in a few years or decades the current generation of nanometer cpus will be able to be scanned and their hardware turned into schematics by some kind of program
awwww but dad! mom promised we could turn our garage into a superfund site! i already told everyone at school, now they're all gonna think i'm a dink and cindy's gonna dump me before prom!
I love the jank factor, I’d bet going the low pressure cvd route would probably be something you could pull off on this scale, especially since specialty gases like silane are reasonably priced at this scale
Just finsihed this video, this is such an incredible work! You're making semiconductor science more accessible, hopefully I will have agarage to join this hobby soon because it fascinates me!
@@projectsinflight I wonder that, instead of using furnace. Is it possible to use solar energy from big Fresnel len? This can also penetrate to the glass which will also protect the item from O2 as well.
@@prachayaputtapanasub1113 I'm certain that solar energy, when condensed will be able to reach high temperatures. However, I don't know how you would manage to both insulate the tube while allowing sufficient thermal energy to penetrate that insulation to keep the temperature up. I'm sure someone could get creative with that.
Note that your resistance calculation assumes straight field lines and wide contacts, but your probes are teeny tiny, and not particularly well bonded to the chip. The resistance of a "point" contact is dominated by the local resistance under the contact point, and basically all the material inbetween the points acts in parallel so has little effect on the total. Now, ideal points, would have a width of zero and thus infinite resistance, so clearly that's not the case here, but what then is the contact area? It depends on the probe (tip) curvature, contact pressure, and elastic modulus of the materials. What exact relation is involved, I don't know... but just an FYI, there should be a rough (but better) calculation somewhere, as this is a fairly typical probing method.
you are correct- the better way to do this is with a 4-point-probe and it'll be the subject of the follow-up video when i go into more detail about evaluating doped layers. i used this quick and dirty method because i only needed a very rough estimate
@@projectsinflight I was wondering about this measurement method as well. I have some familiarity with the concept of "resistant-per-square" as is used with things like ITO-coated glass/polymers. In the off chance you're not aware of it, the way the measurement is performed is that two relatively low resistance conductors which can span the entire edge length of a square are put at opposite ends and are used as the probing points, and thus for a material rated at, say, 100 ohms-per-square, it doesn't matter if the sample measured is 1sq cm or 100, the resistance would always be 100 ohms. would it not be the same in doped semiconductors? Recently came across this channel and am grateful for the content. Your video on the fundamentals of semiconductors was particularly excellent! I hope you keep the great videos coming!
This is some of the most fascinating content I have seen on the internet in a while. This is incredible work with amazing editing. I am so excited to see this channel grow!
Nothing stops you this is amazing ! As I have none of your knowledge, the solutions you find really feels like magic to me, and this whole quest is epic !
It would be interesting if you had chemistry youtubers synthesize the necessary molecules you need that aren't provided. It'd make for a nice colab. There's around 3-6 of them and at least more than half would be eager to help. Some i know are explosions and fire, chemiolis and rest i cant remember.
there is actually one or two chemicals that i'd love to have help with. Making a chemical called diazonapthaquinone is out of my reach. it's like a 6 step organic synthesis. No idea if anyone on youtube would be up for that.
@@projectsinflight You bought the TEOS, though? Or are you just wanting a way to make it even more accessible by making it from starting materials that are easily available?
Thank you for producing some of the most fascinating science content on youtube. I continue to be surprised not to see you with nebula mentions everywhere. Also I would like to join the other voices here in gratitude for making it feel like home semiconductor fabrication is within reach.
@@projectsinflight I've no personal experience of producing videos for either platform. It does seem to have been attracting a number of the more informative creators out there though.
This is excellent! I especially liked that you guided us through both the motivation and your process. Seeing rainbow films after you explained them extensively in previous videos was also a big moment! And I also like that you made a point of not relying on connections to source your materials. Too many DIY R&D guys do it, and while it is cool, I also envy them a bit...
Nice video. For a homemade spin coater we use a variac and a drimmel with a tupperware bowl. If you want to measure the speed you can put a small magnet on the edge of the sample mount and use a hall sensor nearby to measure rotation rate.
i planned to make a better one but this one that i threw together actually seems to be holding up pretty well even without any fancy features. Ideally in the future i'll have variable speed, and a vacuum chuck instead of tape
i love watching you videos every time! i can only imagine how frustrated you must have felt many times throughout the project. you dedication, knowledge and skills are very impressive. i love to see where this will go in the future!
This is absolutely fantastic work. Your channel popped up a few weeks ago and I eagerly devoured this whole series. I love the explanation of the theory behind what you are doing at each step followed by your trials to get to a workable solution. Eagerly awaiting your next video when ever it happens to come out :D
Another absolutely incredible video in this series, you should write a paper on this! These amateur accessible methods would be a really interesting addition to the world of published literature, especially when so much knowledge is safeguarded by the silicon industry.
What you’re doing is amazing and I want to thank you so much for making hobby chips a possibility within our lifetimes. I can’t wait to see where the project goes next! One future improvement could be finding a way to making the doping more consistent. There’s honestly so many possibilities for improvement, and many more for what this tech can help us achieve. Thank you again!
I applaud you for persevering through the challenges, I've been wondering just how possible it was for small scale or home lab silicon tinkering. seems a home lab can make simple chips and in theory solar cells as well. Efficiency and time costs aside, this is amazing!
Thanks so much for sharing the fruit of your hard work! I'd love to see you expound on how N and P regions, properly combined, produce a transistor or diode like effect.
perfect timing, im only just getting interested in peltier effect cells and have been wondering about silicon carbide semiconductors but wondered if they could be doped similarly to usual silicon
I think by the hot probe test you are roughly measuring the Seebeck coefficient, which is voltage difference generated by a heat gradient. This is negative for n type and positive for p type, which means positive voltage differential for n type and vice versa
I belive I have read a soviet book regarding this process 1970-s edition. There was a section on this topic. They added a dopand as a gas and heated a wafer in furnace. They propoused to have a separate sealed furnace to avoid contamination from already doped glass. There was a big section regarding crosscontamination and perfomance drop. I think you might have observed this effect.
@@projectsinflight I think it mostly refers to same heating system, but separate chambers to avoid dopant cross-contamination. The book was about industrial setup. In your setup it might include having separate glass tubes for diferent dopant. But hopefully you will not need it and everything would be fine. Good luck.
Once you've perfected the whole process, what sort of thing will you be able to make? Do you have your sights set on integrated circuits, or individual transistors/diodes? If ICs, and given the resolution of your photolithography, what scale/complexity of circuit do you think would be attainable?
Wow!!❤❤❤ I don't know very much on this subject, just a hair above full naivety, but enough to know how to appreciate a cheap alternative! Good job and I hope you get lots of views so you can buy more stuff to make fun things. ❤❤❤
Excellent content! Subscribed. One small correction. At around 28 minutes you mention that salts are more soluble in water as temperature increases. While this is mostly true, there are exceptions. For instance, plain table salt (sodium chloride) - its solubility in water doesn't really change with temperature over the liquid domain of water (0C to 100C). The speed of dissolution does increase, but the overall solubility and final concentration of a saturated solution is more or less the same.
@@projectsinflightthat's good to hear, unfortunately I just have my shed that is 6*10 feet (and is also my laundry&tool storage). Oh the joys of living in a city in the Netherlands!
First off, amazing work. The months you put into this project are much appreciated in the DIY community Can you tell me more about what you used for the silicone oil on that attempt? While this didn't work for doping, I didn't know that silicone when heated will decompose into glass. Any details on what you used would be Greatly appreciated. I'm researching laser pumped remote phosphors and embedding them in silicone to turn into glass sounds like what I've been looking for
Yeah - it makes me a little uneasy that an educational video has to make their content a little wrong, in order not to be offensive. At least the note lets us know this has happened (and tells us what the actual shape probably is, haha)
haha, actually at 33:18 you can see my bottle of trimethyl borate i have for just this purpose. it did produce better results but it is going to have to be in the next video on this subject
if the wafers you are using were already lightly p-type doped, would that effect the eventual electric potential after you re-doped them in the phosphorus (n-type)? also thank you for teaching so many things! this is very cool
It'd be awesome to see you turn a couple of these into some diodes and then build something out of them, like diode-rectifier AM receiver. Like the old whisker radios, but with a better diode. I mean, what does it take to turn two doped chunks into a diode? Squish them together? With heat? Without heat? Seems like yet another area full of interesting wrinkles to solve.
It's not too hard to make a diode, you just need to diffuse a P, then N region and then make some electrical contacts. I've gotten them to work already, but i don't yet have a video together on the topic
I don’t remember subscribing to you. But damn is this cool stuff. It’s feeling more and more possible to use a diy DLP wafer stepper to make custom silicon in the garage, with only a few thousand dollars of equipment.
if ethanol is leaving when the TEOS polymerizes, doesn’t that mean the reaction is prevented while there’s a large excess of ethanol present? once the ethanol has evaporated, the reaction can proceed. it’s the same as water based acrylic polymer paints. the polymerization reaction is a dehydration, a ketone and a hydroxyl condensing to link them. the paint is stable in the tube, where there’s an excess of water. once it’s used and has a chance to dry, there is no longer that excess of water and the paint polymerizes.
i'm not going to pretend that i'm an expert on the subject, but my impression is that the ethanol produced by the reaction does not work to significantly limit the speed of the reaction. Also, the ethanol is required to keep the water and teos miscible, so it cannot be entirely removed
Excellent work, I know the kind of work that can go into this kind of development, hats off! I do have one question though. I'm quite squeamish when it comes to HF, can I substitute another acid to etch off the oxide layer? In my case I'm looking at replicating your work here to produce PV cells so a layer of glass is actually helpful.
unfortunately there is not really an acceptable substitute for HF. really the only way around it would be doing a physical etch instead of a chemical etch, which would involve making a plasma etching chamber. Sorry about that.
Amazing and inspiring work. One question. You determined the doping level by means of measuring resistance on the surface. I assume that is the etched side? Do you get consistent measurements on the other side (untreated) surface?
Ah yeah, i possibly should have made that more clear. The dopant only diffuses in to form a thin layer, so the side that had the glass is the side that is doped. There is a small amount of dopant that enters the air so the other side might be doped a little, but it should be orders of magnitude less.
You are measuring the resistivity of the wafer directly, but as far as I know if the wafer is lightly doped (which obv has to be) you are not making an ohmic contact with your probes. Therefore the resistivity measurement cant be fine, right? Great video by the way, there is not enough content con semiconductor manufacturing and devices on TH-cam. Congrats!
Yeah the minimally doped wafers definitely don't provide very good contacts. This is actually why I was so adamant about the S-O-D reaching the heavily doped region because that's what you need to make good ohmic contacts
The image at around 26:40 where the cracks form reminds me of the eyeball. There's a nice depiction by DrWealz on the anatomy of the eyeball which has a nice match. If you look at the images there you may recognise a few things. If you Google the difference between hydrocarbons and carbohydrates and compare the Wikipedia images of the Sphynx of Egypt, and the Serpent of the pyramid in Chechen Itza and the Clapping in front of the pyramid, you may recognise a few more things which may be of some interest. 🤔
First of all, thank you for the excellent video: entertaining and very interesting.🤩👍 But I have a question: if you have a pre-doped wafer, say n-doped, can you make it p-doped? I mean, if you add n-dopant to an already n-doped wafer, it seems obvious that it will become even more n-doped; but what happens if you add p-dopant? Will the new dopant sort of cancel out the existing doping and finally overwhelm it? Or will it “push away” the existing dipping and replace it? Or do you have to remove the existing doping first? How does this works? In the video you don’t say which wafer you used for each recipe, but I guess that if you want a pn junction, you’ll have to have both dopings on the same wafer, so somehow an n-doped wafer must become partially p-doped or vice versa.
Yes absolutely! if you have a region that is n doped at say 10^16 then you can dope it with P at 10^18 and basically overwrite that region. it becomes tricky when you make devices because you have to add from the surface, so generally you try to make the lightly doped layer as the deepest, then overwrite with a shallow layer with heavier coping, etc. ion implantation can implant layers below the surface so it can get around this problem.
great video, me and a few friends were attempting to do SOD, with various results, but we managed to make some, with maybe not the best shell life, but kind of acceptable. What's the shell life for your SOD ?
At room temperature, some recipes go bad in a few days, others seem to last for much longer. I don't currently have an estimate for the shelf life at cold temperatures, but it's at least a month if not months. I'll definitely provide updates on this in the future!
Your thick layer P test looked like when it crackled it detached from the wafer pretty nicely (26:39). Have you tried a slightly slower transitioning mix and tested it in the furnace? It seems beneficial if you had a dopant where the glass detached cleanly from the wafer after the diffusion step is finished. If you could find an optimal point where the stresses during cooling made it lift off, that would be amazing.
The glass in 26:39 had been on the hotplate but not the furnace, so it only reached about 500C tops. Also, if the glass doesn't acutally bond well to the surface of the silicon then the dopant (probably) wont be as effective at both delivering the impurities needed, and keeping out the impurities that arent.
Are you planning to try diodes (or even BJTs) soon? I love the quality of these videos, and I feel that a few more videos pushing the boundaries could really lead to an explosion of hobbyist semiconductor experimentation.
I'd imagine making a glycol based borate (poly?) ester is probably the safest bet for actually keeping the boron from crashing out at room temperature causing bad film structure, while still keeping it from evaporating when heating it up in the kiln (which may be a risk with the simple alcohol borate esters). it would need external pH balancing (nitric acid?), and the glycol will add carbon which might mess with the glass structure at elevated temperatures and cleanup afterwards. worth trying though.
So, i've had some successes with borate esters, and no success with polyethylene glycol or ethylene glycol mixtures (ruins the spinnability). Apparently some commercial S-O-D formulations use polyvinyl boronic acid esters but i don't know how to synthesize them
@@projectsinflight I was mostly thinking it should be possible to use something like propylene glycol (readily vape juice ingredient) as the alcohol group, the resulting massive molecule should have a much higher likelihood of staying in or even fusing to the glass matrix during heating. can't perform tests myself though as almost all chemicals involved are illegal unless you own a registered lab here in sweden.
I have a question. Have you at any point considered waterglass (sodium silicates) during this process? If you haven’t, would you consider testing it? And if you have, what issues have you ran into using the stuff? Likewise, sodium aluminate can be used to make aluminium hydroxide hydrogels which can be calcined to form alumina at 1100C. And if memory serves me right, alumina and silica are quite compatible with each other. Thoughts?
@@projectsinflight Hadn’t considered using Al as a dopant, interesting idea. Silicates are available bound to other atoms like Ca, and there’s also colloidal silica (also made by Cab-o-sil) which was used to glue together the insulation fibers on the space shuttle heat shielding tiles due to having a lower melting point than bulk silica.
Very interesting, but can not be did something similar with completely homemade semiconductor? For example with thin layer ZnO, it is easy to make at home but i was not able to make something more difficult than diode. But i read some papers and with good equipment is possible to make transistor, maybe even an integrated circuit, they said.
![SAFEPLANET - ดาวเคราะห์ [ THE PLANET ]](http://i.ytimg.com/vi/4X1XL3I_k3c/mqdefault.jpg)
the fact that were actually living in a time where home-made semiconductors are something somebody can consider is crazy.
There's definitely been a small number of people doing this over the years, but a lot of components have become more accessible. I doubt TEOS was widely available in the past
The first time we see scientific hobbies “
Homemade is a bit of a stretch but yeah😅
Stupidity.
@@itzhexen0 it's not stupid if the proper procedures are taken, calm down.
If I knew it was you getting the TEOS, I would have let you have it for free. Whenever you need more, let me know and I'll donate it to you absolutely free of charge.
I knew that was your name!! You really are the man for some unique and quality stuff. I'm happy that we're acquainted at least in a limited manor. I hope this gets a heart and a pin.
That is super generous of you! I really appreciate it! and thank you so much for doing what you are doing and making stuff available to hobbyists :)
Haha, I wonder if your TEOS orders are about to see a sharp uptick from all the people watching the video :P
well if i'm understanding correctly i must have been the only one to buy it in the last couple months so a sharp uptick would be like... 2 more orders lol
@@projectsinflight Hah, guess we'll see if that number goes up any more
Lots of research, cost saving, and problem solving. Science in its purest form. This will be a critical series of videos for many future DIYers
Thank you! I hope to be a part of making this hobby more accessible!
Agreed!
This man is kind enough to call it “our dopant” instead of “my dopant”.
should get a hobbyist nobel prize for your work
I'll 3d print one and then spray paint it gold :P
Homebrew semiconductor doping is genuinely one of the most impressive things I've seen on youtube, huge props
thank you! i'm hoping to make this stuff more accessible to hobbyists
I would love if you sent the wafers over to breaking taps or some other youtuber with an electron microscope, I think it would be really intresting to see how the glass layer changes at every step.
i dont want to spoil anything but uhh.. don't worry i've got that covered lol
@@projectsinflight Im a college student with access to a lab that can conduct various surface characterization techniques, if you're curious on any other surface properties :) I've part of a research project that involves etching silicon wafers so we have experience with it.
Thomas Edison once said "Genius is 1% inspiration and 99% perspiration." If the amount of effort that you put in and the perseverance that you demonstrate when you find a dead end are any indicator, you sir are a genius.
Well, it does help knowing that what I am doing is definitely possible. I'd struggle a lot more if I were in truly uncharted territory
Edison was kind of awful, though, which always sours the delivery of that quote a bit for me.
I bet he just stole it from someone else anyway.
A perhaps more relevant Edison story is that he (and his lab) spent most of their time making better vacuum pumps when he began experimenting with incandescent bulbs. They got several patents on that before they had any useful results on the incandescent bulbs themselves.
When you heat boric acid with alcohols you get the corresponding borate esters. So you could make trimetyl borate first add that instead of boric acid and balance the pH with nitric acid. Also the boric acid is way less acidic that the phosphoric acid so you should have already been adding another acid to get the same pH.
at 33:18 you can seem my Trimethyl Borate I made just for this purpose :) It works well, but I didn't want to overcomplicate the process in this video, so it's gonna have to be in the next one
My idle daydreams about hobbyist fabrication of integrated circuits have taken a GIANT leap forwards thanks to your hard work. THANK you!!!
Glad to be of service :)
Semiconductor device manufacture is one of the most shrouded fields of engineering. Thank you for cutting one hole in that veil.
it's so crazy- all the people who know how to make the stuff using these methods are 70+ years old. These days it's all ion implantation and vapor deposition
I have worked extensively with dispersants for silicon nanoparticles; I am unsure of how well this would work in general but the long and short of it is that toluene and silicone oil are both effective dispersants for fumed silica. the ineffective behavior of the isopropanol dispersant could have been due to the solvent being too thin, and thus the nanoparticles started to settle too quickly on the spinning plate. Miscibility of the aqueous phosphoric acid into such a solution would still require experimentation, but I do know isopropanol is miscible with silicone oil?
i can't believe i didn't try mixing the fumed silica and silicone oil lol- thank you for the tips :)
As someone who work in a lab, I approve the 2:07 Sigma Aldridge tagline. And the 'dead-ends and complications that didn't made it' just resonates too well with me. Great video as always, keep them coming.
6:08 Small correction: the diagram is labeled with a flow of protons (H+) instead of holes (o+), which is definitely wrong even in a simplified explanation
Ah, dang it I didn't catch that. Well, hopefully I didn't confuse people too badly
Great work. The editing and cadence was amazingly easy to watch/listen to. Like a smooth train ride. I look forward to what you come back with in a few to several months!
thanks! i'm super excited to begin working on my next video. i can't spoil it yet but it's gonna be the coolest thing i've done on the channel :)
For all the trash YT throws in my recommendations when 90% of the channels I follow are STEM/project oriented, it's amazing I came across this.
That's truly fantastic to hear
I find it funny that with a reasonable amount of background knowledge, some basic properties, and a bit of testing, it's possible to reverse engineer most proprietary products
turns out keeping the secret family recipe in a safe doesn't protect you from a mass spectrometer
@@projectsinflight it sees right through anything (literally)
Sadly, modern CPUs are practically immune to reverse engineering.
For example, no one knows if rdrand instruction of Intel CPUs really works as advertised.
there is an emerging technology of super high resolution CT scans and i suspect in a few years or decades the current generation of nanometer cpus will be able to be scanned and their hardware turned into schematics by some kind of program
@@projectsinflight these two centuries have been insane
as somebody trying to develop their own type of transistor and semiconductor, this is really useful, thanks!
awwww but dad! mom promised we could turn our garage into a superfund site! i already told everyone at school, now they're all gonna think i'm a dink and cindy's gonna dump me before prom!
I love the jank factor, I’d bet going the low pressure cvd route would probably be something you could pull off on this scale, especially since specialty gases like silane are reasonably priced at this scale
possibly- i'm a little nervous though at the idea of using a pyrophoric gas
Just finsihed this video, this is such an incredible work! You're making semiconductor science more accessible, hopefully I will have agarage to join this hobby soon because it fascinates me!
thank you! i appreciate it :)
@@projectsinflight I wonder that, instead of using furnace. Is it possible to use solar energy from big Fresnel len? This can also penetrate to the glass which will also protect the item from O2 as well.
@@prachayaputtapanasub1113 I'm certain that solar energy, when condensed will be able to reach high temperatures. However, I don't know how you would manage to both insulate the tube while allowing sufficient thermal energy to penetrate that insulation to keep the temperature up. I'm sure someone could get creative with that.
@@projectsinflight Oh!! I see...... The exact temperature from solar is also hard to maintain without appropriate tools. Thank you!!
FINALLY! been waiting so long for this. Keep going!!!
Thank you! I appreciate everyone's patience
Note that your resistance calculation assumes straight field lines and wide contacts, but your probes are teeny tiny, and not particularly well bonded to the chip. The resistance of a "point" contact is dominated by the local resistance under the contact point, and basically all the material inbetween the points acts in parallel so has little effect on the total. Now, ideal points, would have a width of zero and thus infinite resistance, so clearly that's not the case here, but what then is the contact area? It depends on the probe (tip) curvature, contact pressure, and elastic modulus of the materials. What exact relation is involved, I don't know... but just an FYI, there should be a rough (but better) calculation somewhere, as this is a fairly typical probing method.
you are correct- the better way to do this is with a 4-point-probe and it'll be the subject of the follow-up video when i go into more detail about evaluating doped layers. i used this quick and dirty method because i only needed a very rough estimate
@@projectsinflight I was wondering about this measurement method as well. I have some familiarity with the concept of "resistant-per-square" as is used with things like ITO-coated glass/polymers. In the off chance you're not aware of it, the way the measurement is performed is that two relatively low resistance conductors which can span the entire edge length of a square are put at opposite ends and are used as the probing points, and thus for a material rated at, say, 100 ohms-per-square, it doesn't matter if the sample measured is 1sq cm or 100, the resistance would always be 100 ohms. would it not be the same in doped semiconductors?
Recently came across this channel and am grateful for the content. Your video on the fundamentals of semiconductors was particularly excellent! I hope you keep the great videos coming!
This is genuinely really really cool. Science youtubers for the win!!!
Thanks! I appreciate it
This years Nobel prize in DIY goes too...
This is some of the most fascinating content I have seen on the internet in a while. This is incredible work with amazing editing. I am so excited to see this channel grow!
Nothing stops you this is amazing ! As I have none of your knowledge, the solutions you find really feels like magic to me, and this whole quest is epic !
Thanks! It took me quite a while to get this far. I'm hoping to improve the S-O-D in the future too
It would be interesting if you had chemistry youtubers synthesize the necessary molecules you need that aren't provided. It'd make for a nice colab. There's around 3-6 of them and at least more than half would be eager to help. Some i know are explosions and fire, chemiolis and rest i cant remember.
there is actually one or two chemicals that i'd love to have help with. Making a chemical called diazonapthaquinone is out of my reach. it's like a 6 step organic synthesis. No idea if anyone on youtube would be up for that.
@@projectsinflight absolutely! These people are incredibly capable a 6 step process is nothing. Have you watched the cubane synthesis video?
@@projectsinflight Ouf, that's gonna be a pretty nasty synthesis. What's the other chemical? Also, fantastic video!
oh yeah the other one is TEOS. i'm looking for a way to synthesize it that doesn't require SiCl4
@@projectsinflight You bought the TEOS, though? Or are you just wanting a way to make it even more accessible by making it from starting materials that are easily available?
Thank you for producing some of the most fascinating science content on youtube. I continue to be surprised not to see you with nebula mentions everywhere. Also I would like to join the other voices here in gratitude for making it feel like home semiconductor fabrication is within reach.
I haven't really thought of nebula yet. I am not sure what you get for being on that platform- is it worth the effort?
@@projectsinflight I've no personal experience of producing videos for either platform. It does seem to have been attracting a number of the more informative creators out there though.
Thank you for sharing your knowledge.
I hope someday open source hardware will contain open source ICs as well.
I'm hoping that eventually making an IC is something that hobbyists have reasonable access to.
This is excellent! I especially liked that you guided us through both the motivation and your process. Seeing rainbow films after you explained them extensively in previous videos was also a big moment! And I also like that you made a point of not relying on connections to source your materials. Too many DIY R&D guys do it, and while it is cool, I also envy them a bit...
Nice video. For a homemade spin coater we use a variac and a drimmel with a tupperware bowl. If you want to measure the speed you can put a small magnet on the edge of the sample mount and use a hall sensor nearby to measure rotation rate.
i planned to make a better one but this one that i threw together actually seems to be holding up pretty well even without any fancy features. Ideally in the future i'll have variable speed, and a vacuum chuck instead of tape
i love watching you videos every time! i can only imagine how frustrated you must have felt many times throughout the project. you dedication, knowledge and skills are very impressive. i love to see where this will go in the future!
it was definitely frustrating at times, but it turns out a ton of research and effort goes a long way to solving problems :)
I haven't seen the video, but I am sure it was worth the wait. Keep it up!😁
Thank you and i appreciate the patience :)
This is the Best TH-cam-Science Project Ever!
You are doing great work in pushing the envelopes of homemade Science, absolutely love it! ❤
Thank you for the support!
This type of content is the future of YT, it needs to be to keep this platform alive! you got a sub in me along with the bell, cant wait to see more!
Thank you so much for following the channel!
This is amazing work. Makes me want to go work in a lab again.
You are getting so close to a transistor. I am so excited for you.
Incredible work, I'm learning so much about semiconductor manufacturing
Glad to help :)
Amazing, I would never have this kind of patience to see it through.
you and me both lol
This is absolutely fantastic work. Your channel popped up a few weeks ago and I eagerly devoured this whole series. I love the explanation of the theory behind what you are doing at each step followed by your trials to get to a workable solution. Eagerly awaiting your next video when ever it happens to come out :D
bro your logo is hilarious i love it
wow. just wow. this is maximally nifty.
Very nifty!
Great video and I can't wait to see more about the testing!
Thanks! I can guarantee that the next video is going to to be the most exciting one yet
You sir are an absolute legend 100 out of 10. I will carry a red carpet around from now on in case I ever meet you. LEGEND.
Another absolutely incredible video in this series, you should write a paper on this! These amateur accessible methods would be a really interesting addition to the world of published literature, especially when so much knowledge is safeguarded by the silicon industry.
i was thinking about better documenting some of the more boring details on my website or something if people are interested.
What you’re doing is amazing and I want to thank you so much for making hobby chips a possibility within our lifetimes. I can’t wait to see where the project goes next!
One future improvement could be finding a way to making the doping more consistent. There’s honestly so many possibilities for improvement, and many more for what this tech can help us achieve.
Thank you again!
The research you have done is impressive. Subscribed to see where this goes!
thanks for the sub! i promise you will like my next video ;) no spoilers yet
I applaud you for persevering through the challenges, I've been wondering just how possible it was for small scale or home lab silicon tinkering. seems a home lab can make simple chips and in theory solar cells as well. Efficiency and time costs aside, this is amazing!
Thanks so much for sharing the fruit of your hard work! I'd love to see you expound on how N and P regions, properly combined, produce a transistor or diode like effect.
what a great result from so much effort! Great work!
perfect timing, im only just getting interested in peltier effect cells and have been wondering about silicon carbide semiconductors but wondered if they could be doped similarly to usual silicon
I think by the hot probe test you are roughly measuring the Seebeck coefficient, which is voltage difference generated by a heat gradient. This is negative for n type and positive for p type, which means positive voltage differential for n type and vice versa
You could sell small samples on a merch shop. Might be a good way to fund research and provide cheap dopant to the hobby community.
I belive I have read a soviet book regarding this process 1970-s edition. There was a section on this topic. They added a dopand as a gas and heated a wafer in furnace.
They propoused to have a separate sealed furnace to avoid contamination from already doped glass. There was a big section regarding crosscontamination and perfomance drop. I think you might have observed this effect.
A furnace within a furnace? Wild. Sounds like quite the engineering feat
@@projectsinflight I think it mostly refers to same heating system, but separate chambers to avoid dopant cross-contamination. The book was about industrial setup.
In your setup it might include having separate glass tubes for diferent dopant.
But hopefully you will not need it and everything would be fine.
Good luck.
Once you've perfected the whole process, what sort of thing will you be able to make? Do you have your sights set on integrated circuits, or individual transistors/diodes? If ICs, and given the resolution of your photolithography, what scale/complexity of circuit do you think would be attainable?
Ah yea that does seem like it would take a while. Worth the wait, thanks for doing the tedious bits for us ♥
No problem! Glad to be of service to the community
Wow!!❤❤❤ I don't know very much on this subject, just a hair above full naivety, but enough to know how to appreciate a cheap alternative! Good job and I hope you get lots of views so you can buy more stuff to make fun things. ❤❤❤
Thanks!
Amazing work! can't thank you enough for sharing your experience
glad to be of service:)
Excellent content! Subscribed.
One small correction. At around 28 minutes you mention that salts are more soluble in water as temperature increases. While this is mostly true, there are exceptions. For instance, plain table salt (sodium chloride) - its solubility in water doesn't really change with temperature over the liquid domain of water (0C to 100C). The speed of dissolution does increase, but the overall solubility and final concentration of a saturated solution is more or less the same.
oops i must have confused NaCl for something else. probably sugar
@@projectsinflight It's okay, many salts behave in the intuitively correct manner, but some don't. Sodium chloride is one of the anomalous ones.
One day I have the space to do this. For now, this is an amazing substitute
Doesn't take too much space- but you definitely need to do it somewhere that has ventilation, and that isn't also used for food prep
@@projectsinflightthat's good to hear, unfortunately I just have my shed that is 6*10 feet (and is also my laundry&tool storage). Oh the joys of living in a city in the Netherlands!
First off, amazing work. The months you put into this project are much appreciated in the DIY community
Can you tell me more about what you used for the silicone oil on that attempt? While this didn't work for doping, I didn't know that silicone when heated will decompose into glass. Any details on what you used would be Greatly appreciated.
I'm researching laser pumped remote phosphors and embedding them in silicone to turn into glass sounds like what I've been looking for
swiss navy silicone lube was used for that test. i spun it onto the chip at a few thousand rpm (3k to 6k) and then baked it at 1000C for like 30m
You are an actual wizard
17:18 That note 😆
Yeah - it makes me a little uneasy that an educational video has to make their content a little wrong, in order not to be offensive. At least the note lets us know this has happened (and tells us what the actual shape probably is, haha)
yeah- I can't be 100% sure if youtube would flag it or not, but i put too much effort into this video to be slapped down by a soulless bot
@@projectsinflight Oh sure, I don't disagree with your decision, it just troubles me that this is the current state of things.
me too, me too
Awesome well researched video thanks for showing us your results
thank you for watching it :)
Bro has ProjectsInFlight so whenever he come back to earth then uploads a video 😅
Good thing the channel name isn't projects in freefall :P
Overall quality like nowhere, also that in-depth knowledge from book (I might some day try it at home), am so grateful for U
I am so thankful for people who take the time to write really good books on niche subjects like the one i showed :)
If youre having issues with the boric acid, maybe try make trimethyl borate or triethyl borate and use that instead.
haha, actually at 33:18 you can see my bottle of trimethyl borate i have for just this purpose. it did produce better results but it is going to have to be in the next video on this subject
Very nice work!
Awsome stuff! Can’t wait until the next video :D
oh i promise my next video is gonna have the coolest thing ever just you wait
if the wafers you are using were already lightly p-type doped, would that effect the eventual electric potential after you re-doped them in the phosphorus (n-type)?
also thank you for teaching so many things! this is very cool
It'd be awesome to see you turn a couple of these into some diodes and then build something out of them, like diode-rectifier AM receiver. Like the old whisker radios, but with a better diode.
I mean, what does it take to turn two doped chunks into a diode? Squish them together? With heat? Without heat? Seems like yet another area full of interesting wrinkles to solve.
It's not too hard to make a diode, you just need to diffuse a P, then N region and then make some electrical contacts. I've gotten them to work already, but i don't yet have a video together on the topic
My dream is to create desktop semiconductor fab machine....I think there is some amount of us that want this, all of us should join forces.
Please do
Have you had a look at push coating with silicone as a more accessible alternative to spin coating?
not sure what that is actually
amazing video, you are talented and can't wait for further videos!!!
thank you, i appreciate the enthusiasm:)
awesome!! i haven't seen anything like this since sam zeloof! i think about making chips myself all the time so to have more content on it is great!
I don’t remember subscribing to you. But damn is this cool stuff. It’s feeling more and more possible to use a diy DLP wafer stepper to make custom silicon in the garage, with only a few thousand dollars of equipment.
I think a couple youtubers have been successful in making that work. Sam zeloof comes to mind
if ethanol is leaving when the TEOS polymerizes, doesn’t that mean the reaction is prevented while there’s a large excess of ethanol present? once the ethanol has evaporated, the reaction can proceed.
it’s the same as water based acrylic polymer paints. the polymerization reaction is a dehydration, a ketone and a hydroxyl condensing to link them. the paint is stable in the tube, where there’s an excess of water. once it’s used and has a chance to dry, there is no longer that excess of water and the paint polymerizes.
i'm not going to pretend that i'm an expert on the subject, but my impression is that the ethanol produced by the reaction does not work to significantly limit the speed of the reaction. Also, the ethanol is required to keep the water and teos miscible, so it cannot be entirely removed
Excellent work, I know the kind of work that can go into this kind of development, hats off! I do have one question though. I'm quite squeamish when it comes to HF, can I substitute another acid to etch off the oxide layer? In my case I'm looking at replicating your work here to produce PV cells so a layer of glass is actually helpful.
unfortunately there is not really an acceptable substitute for HF. really the only way around it would be doing a physical etch instead of a chemical etch, which would involve making a plasma etching chamber. Sorry about that.
This is awesome I would love to get to a point where we can design and manufacture simple microprocessors in a home lab
that's definitely a very long way off for me, but hopefully in our lifetimes we make it that far
working with you will be fun
Amazing and inspiring work. One question. You determined the doping level by means of measuring resistance on the surface. I assume that is the etched side? Do you get consistent measurements on the other side (untreated) surface?
Ah yeah, i possibly should have made that more clear. The dopant only diffuses in to form a thin layer, so the side that had the glass is the side that is doped. There is a small amount of dopant that enters the air so the other side might be doped a little, but it should be orders of magnitude less.
Awesome work like always
Thanks!
You are measuring the resistivity of the wafer directly, but as far as I know if the wafer is lightly doped (which obv has to be) you are not making an ohmic contact with your probes. Therefore the resistivity measurement cant be fine, right?
Great video by the way, there is not enough content con semiconductor manufacturing and devices on TH-cam. Congrats!
Yeah the minimally doped wafers definitely don't provide very good contacts. This is actually why I was so adamant about the S-O-D reaching the heavily doped region because that's what you need to make good ohmic contacts
The image at around 26:40 where the cracks form reminds me of the eyeball. There's a nice depiction by DrWealz on the anatomy of the eyeball which has a nice match. If you look at the images there you may recognise a few things. If you Google the difference between hydrocarbons and carbohydrates and compare the Wikipedia images of the Sphynx of Egypt, and the Serpent of the pyramid in Chechen Itza and the Clapping in front of the pyramid, you may recognise a few more things which may be of some interest. 🤔
First of all, thank you for the excellent video: entertaining and very interesting.🤩👍 But I have a question: if you have a pre-doped wafer, say n-doped, can you make it p-doped? I mean, if you add n-dopant to an already n-doped wafer, it seems obvious that it will become even more n-doped; but what happens if you add p-dopant? Will the new dopant sort of cancel out the existing doping and finally overwhelm it? Or will it “push away” the existing dipping and replace it? Or do you have to remove the existing doping first? How does this works? In the video you don’t say which wafer you used for each recipe, but I guess that if you want a pn junction, you’ll have to have both dopings on the same wafer, so somehow an n-doped wafer must become partially p-doped or vice versa.
Yes absolutely! if you have a region that is n doped at say 10^16 then you can dope it with P at 10^18 and basically overwrite that region. it becomes tricky when you make devices because you have to add from the surface, so generally you try to make the lightly doped layer as the deepest, then overwrite with a shallow layer with heavier coping, etc. ion implantation can implant layers below the surface so it can get around this problem.
@@projectsinflight Thank you for the clear answer, I appreciate.👍
Congratulations - you deserve it!
thank you! i am very glad it worked out
great video, me and a few friends were attempting to do SOD, with various results, but we managed to make some, with maybe not the best shell life, but kind of acceptable. What's the shell life for your SOD ?
At room temperature, some recipes go bad in a few days, others seem to last for much longer. I don't currently have an estimate for the shelf life at cold temperatures, but it's at least a month if not months. I'll definitely provide updates on this in the future!
Bloody brilliant
thanks!
This is great!
thanks! I appreciate it
Your thick layer P test looked like when it crackled it detached from the wafer pretty nicely (26:39). Have you tried a slightly slower transitioning mix and tested it in the furnace? It seems beneficial if you had a dopant where the glass detached cleanly from the wafer after the diffusion step is finished. If you could find an optimal point where the stresses during cooling made it lift off, that would be amazing.
The glass in 26:39 had been on the hotplate but not the furnace, so it only reached about 500C tops. Also, if the glass doesn't acutally bond well to the surface of the silicon then the dopant (probably) wont be as effective at both delivering the impurities needed, and keeping out the impurities that arent.
1:45 can you point out where you explained why pure boron and phosphorus won't work? I don't see anything.
it was at 19:50 on the first video about doping- the one released a month ago
@@projectsinflight Ahh, that makes sense. Thank you! Missed that one after watching the band gap video.
great work.
Thanks!
Are you planning to try diodes (or even BJTs) soon? I love the quality of these videos, and I feel that a few more videos pushing the boundaries could really lead to an explosion of hobbyist semiconductor experimentation.
Idk why I read that as dildos
diodes will probably be the next next next video if i had to guess. i've got a very cool video coming next. no spoilers yet ;)
I'd imagine making a glycol based borate (poly?) ester is probably the safest bet for actually keeping the boron from crashing out at room temperature causing bad film structure, while still keeping it from evaporating when heating it up in the kiln (which may be a risk with the simple alcohol borate esters). it would need external pH balancing (nitric acid?), and the glycol will add carbon which might mess with the glass structure at elevated temperatures and cleanup afterwards. worth trying though.
So, i've had some successes with borate esters, and no success with polyethylene glycol or ethylene glycol mixtures (ruins the spinnability). Apparently some commercial S-O-D formulations use polyvinyl boronic acid esters but i don't know how to synthesize them
@@projectsinflight I was mostly thinking it should be possible to use something like propylene glycol (readily vape juice ingredient) as the alcohol group, the resulting massive molecule should have a much higher likelihood of staying in or even fusing to the glass matrix during heating. can't perform tests myself though as almost all chemicals involved are illegal unless you own a registered lab here in sweden.
I tried using ethylene glycol, and for some reason it ruined the spinnability of the dopant. no idea why though.
I have a question. Have you at any point considered waterglass (sodium silicates) during this process? If you haven’t, would you consider testing it? And if you have, what issues have you ran into using the stuff? Likewise, sodium aluminate can be used to make aluminium hydroxide hydrogels which can be calcined to form alumina at 1100C. And if memory serves me right, alumina and silica are quite compatible with each other. Thoughts?
The main problem is that sodium is a very serious contaminant to the silicon
aluminum is a potentially useful dopant though
@@projectsinflight Hadn’t considered using Al as a dopant, interesting idea. Silicates are available bound to other atoms like Ca, and there’s also colloidal silica (also made by Cab-o-sil) which was used to glue together the insulation fibers on the space shuttle heat shielding tiles due to having a lower melting point than bulk silica.
backyardscience2000 FTW!
for real- that dude saved me 100%
Very interesting, but can not be did something similar with completely homemade semiconductor? For example with thin layer ZnO, it is easy to make at home but i was not able to make something more difficult than diode. But i read some papers and with good equipment is possible to make transistor, maybe even an integrated circuit, they said.
Well my mind is blown
If you like this video you're gonna be very excited when you see the next one :P No spoilers though
are you gonna make a vid about how diodes are made?