Very good! You are becoming more and more involved in the interesting part of vacuum technology! Sputtering is one of the most interesting ways of depositing materials, specially when reactive gasses are involved. Few tips: 1) Dont use such high pressures, try and stay in the 1x10-2mBar (get a baratron), if you go to high, the plasma will be super hot and the particles cant reach your substrate because of too much gas particles blocking them. 2) Keep the distance from target to substrate to a maximum, dont let the plasma touch whatever your trying to coat, it will destroy the layer(exactly what your problem is with adhesion). 3) Evaporation rate is usually very low with sputtering (5A/s is considered very high), so keep the energy down, let things be cool and take your time. If the layer gets too thick, and the substrate is cool, from my experience, anything above 2000A will start to flake off... 4) Use high pumping speeds and mass flow controller is a must that has a loop with baratron, i have used MKS systems that controls everything, from the gate valve position to the the mass flow controller, basically a closed servo loop that allows for steady pressure. 5) Consider getting a RF supply and a matching box, you can then do dielectric depositions, with DC you cant.
Thank you very much for the helpful tips! Today I tested placing the substrate as far from the magnetron as my chamber allows. However, the layer was still very poor, especially where the layer thickness was too high. Interestingly, this problem did not happen when using thermal evaporation, even though the layer thickness was higher. Of course, adhesion to the glass was generally much worse, but at least it was a nice reflective surface. Other comments suggested applying a layer of chrome first to improve the adhesion of the copper. I will test this as soon as I get hold of a chrome target. Building a closed loop for the measuring device and the MFC is a very good idea! Sounds like a new project haha.
@@AdvancedTinkering Thin metal films are often under high stress, and in some cases this is the main limitation on achievable film thickness. The stress can be tensile or compressive, and depends on deposition rate, deposition process, substrate temperature and probably a few other things I don't remember. I think the stress applied to a 1 mm thick glass substrate by a 100 nm thick metal film will warp it enough to be practically measured by the kind of interferometric measurement you would use for lens figuring (I calculate 650 m radius of curvature with 200 MPa stress in film, which would cause 125 nm deflection in the middle of a 2.5 cm span) but there may be a more practical way to go about it if need be.
One idea I would try is pumping down with just air in the chamber first, and running the magnetron at low power to get a visible glow. Should work as a poor reactive ion etch to super clean and activate the glass surface. Then introduce the argon and proceed as usual. Those dangling oxygen (and maybe nitrogen) groups coupled with the complete absence of organics should help adhesion. As for surface quality, the fact that your windows got a super mirror surface but the slides didn’t just shows your deposition rate is too high - the flux of metal ions above the sputterer is way too dense which is causing agglomeration (blobbiness). Turn down the mA to something like 1/10th what you were using and it should improve a lot. Copper is the worst for this, although not as bad as silver or tin. Refractory metals cool faster, so titanium and tungsten can be coated at a higher current density and go much faster. With soft or low melting metals, you need to go low and slow.
The distancing will NOT solve his problem. It is ELECTRONS that are hitting the substrate and amorphise and crystal that is growing. You either use a facing magnetron topology (look it up) to try to trap those rogue electrons in between them, increase the magnetic field strength or decrease your power. You do NOT mind if positive ions hit your substrate, it is actually good for the final quality, what you want to avoid is those pesky electrons. This advise comes from someone who routinely can do ALO, ZNO, ITO with hugely good optical and electrical properties. Also, RF would be nice in some situations, but you can do ceramics also with DC, with reactive sputtering. For example for the Oxides you would just use Oxygen in your chamber, and for the Nitrides, you would just use Nitrogen! To have VERY good looking un-oxidised metal deposits, you can also use a small amount of Hydrogen in your Argon.
If you watch Ben Krasnow's channel he has a whole video on slide prep for sputter coating. It's suspicious that titanium coats well while copper doesn't though so I don't think slide prep is the issue. He does talk about plasma cleaning a lot which you didn't mention. I love this!
Titanium is used as an adhesive layer in many thin film structures. There are other thin film adhesive metals including: Cr, Al, Ta, Mo, Nb, V, & Hf. Copper won't stick too well by itself 😀.
Finally, you can settle the important question whether rice with a coating of 20 nm ITO tastes better than rice with 20 nm W. Great Magnetron, by the way, and thanks for all the great info on building the magnetron! If you want to get rid of the rusting issue on your magnetic yoke: ferritic stainless steels are magnetic. They also conduct heat better than their austenitic counterparts: 1.4003 (ANSI 410) is one.
Thanks for the advice! I would have loved to coat the puffed rice with gold and actually eat it. But getting a gold target was a little bit outside my budget.
In electron beam evaporation vacuum coating the cleanliness of a glass surface must be extremely high, like... atomically clean. When we do multi-layer dielectric coatings for coffee table sized optics to make them into mirrors for the National Ignition Facility fusion laser, the glass is cleaned by removing the top few layers of atoms. We scrub the glass with a toothpaste like thick suspension of nanosized alumina particles (0.05 um) called Baikalox and rinse with deionized 18 megaohm water. You can tell when the glass is clean because it will be perfectly hydrophyllic with absolutely no water beading anywhere. If you watch the DI water dry on the surface in a cleanroom you can see the Newton's fringes forming on the final few layers of water molecules as they sheet off and evaporate. When the glass is then loaded into the coater it is heated to a couple hundred C by quartz lamps to drive off the last layer of water molecules adhering to the dangling O and Si bonds of the glass surface. It sounds complicated and expensive but you could probably do it well enough with a simple setup and no cleanroom for very cheap with any common polishing slurry. The main metric of importance is the total lack of any hydrophobicity of the glass after cleaning.
@@ericlotze7724 Actual surface figure polishing was done at another facility before the glass was sent to us for coating. We're only removing a few atoms thickness with the final hand cleaning, so it doesn't change the optical properties.
The reason they're held to such stringent standards, is to avoid any anomalies in the reflection... which is a huge factor, when you're dealing with once-in-a-lifetime data. That level of precision isn't necessary for 99.9999% of the market. Sputtering is used in all kinds of industries for various purposes. Sometimes, such as with some aluminum applications, the layer is applied with the intent of near-immediate degradation/oxidation to protect the base material.
Your chamber is really coming along, love the custom magnetron. There's already a lot of good feedback on cleaning and process parameters already. Couple of things to add. Most common sputtering metals gold, silver, copper etc have very low adhesion to glass. Even with perfectly clean glass, as the coating thickness increases surface stresses build up and cause the coating to pull away. This causes the rough appearance on your slides. The viewports are farther away which slowed the deposition and reduced the surface stress. Increasing the slide distance to the magnetron will help but thick coatings will always be fragile unless the adhesion is improved. Typically an adhesion promoter layer is added between the glass and the thick conductor layer. A few nm of pure chromium is the usual recommendation since it sticks well to most things, but you could also try a high chrome content stainless, as an easy to source option.
No, the rough appearance on his glasses are due to very bad crystal quality. If you try to reactively sputter ATO, ZnO, ITO, AlO, or whatever transparent, you will see this effect even more pronounced. On metals you cannot really see how bad their crystal lattice is until you do some Xray on them. It has taken me years to perfect sputtering. In the literature you will find the ways to mitigate the low quality problem: either use facing magnetrons, limit the power, or increase the magnetic field. All these will help with much better quality but will reduce the throughput. The best way is to NOT let ANY electrons hit your surface. They either put a negative bias behind the substrate, or, like in my case, use a totally different geometry to planar sputtering...
So cool, I love magnetron sputtering! Tried it many times myself, with varying degrees of success. I really appreciate the helpful comments too, as there are a lot of potential mistakes you can make.
The reason for the rough coating was a combination of high deposition rate plus the high energy ions building up a static charge. On the viewport you are getting a lower depositing rate and the copper is mostly as neutral atoms by that point. ❤
Cool video, back in the mid 1990's I worked at a company in Sarasota Florida that produced pressure sensors for the automotive industry, I ran a CVC 601 Magnetron Sputtering System that we used to deposit SiO2 on pressure diaphragms then we Sputtered Nichrome on top of the SiO2 and after the diaphragms were removed from the Sputtering machine they went to a LASER where a Weatstone was etched into the Nichrome as the sensing element of the pressure sensors.
6:54 You revealed The Universal Truth about DIY. Nothing more. In my opinion this is important ingredient of being happy and mentally healthy, not to mention side effects of becoming good at what you do after several iterations of improvements. Kudos Sir 😸💪
Ok, perfect, this is what happened to me as well, after having finished my first sputtering machine. Despite sputtering being known for good adhesion, even the adhesion was bad on some cases. The worst part? You almost saw it with the copper deposition: bad crystal quality. Almost totally amorphous. You will see how bad the quality is if you try to do some ZnO, ITO, or anything transparent. So, it took me a whole MONTH to realise what was going on. The problem causer: ELECTRONS! High speed electrons BREAK your bonds! That is why your side glasses turned out ok, their crystal lattice was not bombarded with electrons. If you read the literature you will see they try to mitigate it that way (as you did by chance), or they use facing magnetrons. The problem there, is when you raise the magnetic strength, you get better quality as less electrons make it across, BUT you get also LESS atoms going across, reducing speed. I would suggest you either go to another geometry and break free from the standard magnetron setup, or go altogether to a molecular beam. Sputtering in that incarnation is pretty useless. I hope any of this helps. Regards from Greece.
Great to see that you managed to build such a nice setup. From past experience of getting glass coated, the companies that did it for me used a layer system with a titanium / titanium oxide layer for bonding to the glass, and then the actual desired coating on top of the titanium oxide. From some personal experience with magnetron sputtering: depending on the conditions like pressure you can produce clusters of the atoms of your target, which can behave very much different from atoms. Might affect the adhesion and later surface finish.
Having MOT-based power supply for such experiments is extremelly dangerous. It would be MUCH safier to have high frequency converter because even if you touch HV parts directly you would have a burn but the HF current will not fri your inside but 50 Hz 2000V definitely will. And it would be good to have some ground leak cut-off switch on HV side. __________ Drops of paraffin for insulation instead of Kapton tape would be much more reliable. __________ Regarding the solidity of the coating, I have two considerations. First: the surfaces to be coated must be very clean. Ben Krasnov spoke about this on the Applied Science channel. Second: probably, during direct propagation, a large number of low-energy particles reach the target and are not capable of forming a strong connection with the surface. And to the sight glass - on the contrary, the fastest. So perhaps the target object should not be placed on the main axis. This will lengthen the time, but only the best will hit the target.)
I am currently in the process of making my own sputtering system and this was just on time. And I would love any more videos on the sputtering system or PVD for that matter. Great video, great channel!
How did it go? I want to make this but I don’t know what I’m doing or where to start. Do you have any advice? My goal is to coat 3d printed items in metal.
it is like cooking. first try was too hot, so it couldn’t stick, but later on the window. also to have quicker results you can move the sample slowly and steadily. like roasting a marshmallow.
You should first process the slides with plasma cleaning to prepare the surface for sputtering. A cool experiment will be to sputter quartz over aluminum to make a durable first surface mirror.
@@AdvancedTinkering As a quick test, you might want to try flaming the microscope slides - pass them through a butane torch flame a few times. This apparently really cleans the surface, both by thermally evaporating any contaminants, and also reaction of the ions in the flame. Many years ago when I was playing with thermal deposition, I tried this and it really seemed to help.
@@BillDemosI had typed a long comment out on my phone, and a stupid telemarketing call ruined it. I just found this channel, watched the video, and read some comments, and you seem extremely knowledgeable on this subject, as well as the preparation involved. If someone were hoping to get a high level of cleanliness, using butane is inappropriate because...? The common sense part of my brain says that cleaning the target with acetone to get rid of any oils, doing a final clean with deionized and distilled water, and then drying the slide using a butane induced flame would be counter productive. This is because the possibility of unburned butane depositing more petroleum oils on the surface that aren't removed in the flaming passes would remain after the final pass. How correct is this thought process? Is there something else the flame cleaning would do to the target surface itself? Maybe some micro ablation from the forced air of the torch in a non-cleanroom/filtered air environment?
that magnetron looks more professional than the actual commercial stuff but keep in mind those magnets have a quite low curie point of like 85°C or something. there are H grades that withstand 150°C but at limited sizes.
Thank you very much! It has grown a bit larger than necessary because I lacked some experience in estimating the material thicknesses with which the milled components would still function. The magnets should never exceed 50°C through the water cooling. If I touch the target directly after sputtering, it is only slightly warm.
Man! You are so smart! You did an excellent job designing your parts. The whole video is very interesting and entertaining. I laughed aloud at some parts. I def noticed you being funny.😊
Love seeing the progress you're making and the process you're going through! It's just so interesting and I didn't know how much I needed more since watching Ben's videos on sputtering way back when.
I'm patientley waiting for that magnetron video! Hopefully with pricing I'm very curious how expensive PCBway's CNC is! Because my local CNC guy is very expensive
I think it is the deposition rate leading to a different grow on the substrate and on the view port glass. You view port glass is much further away. Another thing to consider is heat. Your plasma directly touches the substrate. I guess it will get hot during the sputtering. This can also reduce the quality of the sputtering process. And last but not least: You cleaned your glass substrate, right? At least with some isopropanol and then water in an ultrasonic bath for 10 min.
This is where having a very good knowledge of both physics and chemistry is most useful. Someone in an earlier comment mentioned glass being _atomically clean._ I'm not sure to what degree of cleanliness is needed and whether it can be achieved physically or chemically- possibly through a powerful oxidizer like Piranha solution. Other factors to consider would be obvious high voltage and heat dangers. Not sure if the charged particles have enough energy to generate X-rays of concern. If you have a copy of Strong's _Procedures in Experimental Physics,_ you could make some fascinating things, although one could make things of serious danger.
Now turn the whole thing upside down, put the plasma field on a CNC gantry over a stationary surface, and reduce the coating application area to sub-millimeter size. Boom, metal FDM printing, and you're rich.
Online machining services are great, but you're not a fully functioning human being without machine tools of your own! Great work though, it's only when you see the real practicalities of a technology that you actually understand it, Thanks!
13:08 I loved your reaction here at your fix working so well. I know the sound of joyous hand rubbing anywhere 🙏😁 Overall seems like a really nice setup and I look forward to seeing what more you do with it. I think it was a good idea to make the deeper dive on the magnetron a separate video and I'm looking forward to that, too.
FWIW, Ti is a common adhesion layer in semiconductor manufacturing because it sticks so well to Si / SiO2 / glass. If you put a second magnetron in there to do Ti before other metals you'll have far better results.
microwave magnetrons are usually powered by only half wave rectified DC, so using a **full wave bridge rectifier** (electroBOOM!) might be drawing too much current for the design of the transformer, making it hot.
That’s good to know! But with the new rectifier the MOT does not get hot anymore. But if I ever need to increase the current, I will have to add a fan.
Is this device capable of Plasma Assisted Chemical Vapor Deposition (PA-CVD) for thin films of Diamond Like Carbon (DLC) or heck even *proper diamond?*
16:32 The reason you get a perfectly coated side windows and not a perfect microscope slide is due to the temperature difference between the 2. The microscope slide is too near from the plasma and the material is not really condensing in slow atomic layer. Nest time, if you put methane gaz and a diamond sample, you will be able to grow diamond.cUS20110151226A1 - Synthetic CVD diamond. If you search in the world most favorite search engine this phrase : H BAR Patents for C30B 25 - Single-crystal growth by chemical reaction of reactive gases, e.g. chemical vapour deposition growth (13,302). You will have all the patent from 2010 about CVD (Control Vapour Deposition)
1. Increased heat on the substrate actually fixes quality problems. His problem is elsewhere, I have explained him in another comment. 2. To do diamond, you need Methane as you said, but only at around 1% in Hydrogen. If you don't use Hydrogen in the mix, you will get graphite instead of diamond.
@@BillDemos Thanks of the answer, your hypothesis is really good. Just need to repeat It with other material. Your anwser : @BillDemos il y a 1 heure Ok, perfect, this is what happened to me as well, after having finished my first sputtering machine. Despite sputtering being known for good adhesion, even the adhesion was bad on some cases. The worst part? You almost saw it with the copper deposition: bad crystal quality. Almost totally amorphous. You will see how bad the quality is if you try to do some ZnO, ITO, or anything transparent. So, it took me a whole MONTH to realise what was going on. The problem causer: ELECTRONS! High speed electrons BREAK your bonds! That is why your side glasses turned out ok, their crystal lattice was not bombarded with electrons. If you read the literature you will see they try to mitigate it that way (as you did by chance), or they use facing magnetrons. The problem there, is when you raise the magnetic strength, you get better quality as less electrons make it across, BUT you get also LESS atoms going across, reducing speed. I would suggest you either go to another geometry and break free from the standard magnetron setup, or go altogether to a molecular beam. Sputtering in that incarnation is pretty useless. I hope any of this helps. Regards from Greece.
You could nickel plate that iron plate pretty easily if you wanted to stop it rusting. Also I think PET film might be good for a sacrificial layer for your windows. It's easily available (florist shops have it for wrapping flowers) and should be fine in high vacuum. You could even rig it to be pulled off during runs if you needed to. Perhaps you could just cut a cylinder of it and sit it inside the whole chamber, it's cheap and stiff enough that it'll just sit there.
The copper coating the viewports is not colinear with the plasma. Perhaps the purest copper with the optimal deposition velocity can be found at the fringes. The viewports are far from your source and might also be colder than your intended deposition target.
hey i want to see you talk about the powersupply in details as well. You built it, i think you should make the content for it. your insights are very valuable.
To get a even coating the substrate often gets moved around. Also, the scratch test can be refined by pre scratching a grid in the thin Film, and then trying the tape test.
I wonder if you could lay down a thin film bimetallic strip on something water soluble like sugar? Would you get a thin sheet of metal foil or aren't the metal particles sort of welded to one another? If that would result in thin metal foil, it might be cool to try to make bimetallic strips or thermoplastic. In any case, thank you for your contribution to puffed rice cake technology.
13:20 Actually, we can see the green color of the copper ions in the plasma. Perhaps it didn't show clearly for you in your camera, but the video does show it clearly. Copper always has such a nice, pretty green glow.
The rough surface could be due to pitting from plasma etching. The slides were very close to the plume so even though the plasma was depositing metal ions, those same metal ions would cause etching and pitting because of their high kinetic energy. The viewports were further away and indirect, so they got all vapor and no etching. You can try moving the substrate further away from the plasma plume to minimize the etching.
You can 100% deposit ITO with thermal evaporation, ceramics give off an evaporation pressure suitable for coating well below even their melting point. Maybe the boats are unsuited for ceramics due to thermal conductivity or something, but I have with my own eyes seen it done with an e-beam.
@@AdvancedTinkering hi I have problem with my dc magnetron sputtering. Short circuit occurs every time I start increasing the power. Help me to solve this issue
Im excited for the prospects this device brings,cant wait to see more stuff sputtered on things. Can you try sputtering some metal oxides and other ceramics in the next videos? Or maybe you can even attempt making a beryllium mirror?
16:00 could it be something to do with how the target glass is directly above the sputterer? could the copper have too much energy when hitting the target and not stick properly?
Yes, I'm relatively sure the distance is at least one factor. As you can see, the edges of the microscope slide has a "mirror finish". Just like the vacuum chamber walls. I tried increasing the distance but it did not change much. The copper "flaked" in the middle. Maybe I have to preheat the substrate.
I would assume the issue with the glass is that the slides weren't sufficiently clean to get a firm connection. Maybe the copper is just more sensitive to impurities than the titanium? I don't really know, though, since ultimately I'm not at all trained in this. Another thing you could try could be putting a light layer of a metal that sputters easily, such as titanium, on first, just barely thick enough to coat it, then putting down a layer of copper on top of it, sort of like how chrome plating first deposits nickel or copper below the chromium.
To achieve perfect coating on glass - you need perfect clean glass surface first. IMHO, best way - cleaning by oxigen plasma. It burns out all kinds of organic residuals on suface. It may done in your vacuum chamber with minimum modifications. Other way - chemical cleaning and activation of surface.
I definitely have to improve my surface preparation. In this video I mainly wanted to test the magnetron. I plan on building a plasma cleaning setup inside the chamber.
I use plastic binder sleeves as viewport covers in my sputter chamber. You are using what looks like a type 2 unbalanced magnetron with a much stronger outer magnetic field. This is going to increase the energy of the species arriving at the deposition surface and will increase the compressive film stress. If the compressive film stress is higher than or very close to the maximum adhesion stress the deposited film will delaminate from the substrate. The viewports are seeing much less compressive stress and the film adhesion is higher. If you need any assistance I design industrial large are sputtering equipment for a living.
11:34 did you mean to say "violated microwave"? with that organspendeausweis gets a different meaning i do actually insanely relate to the "eating something and being annoyed that it isnt coated in metal" part also i really wonder with the viewport, theres various variations of ways to keep cameras clean, milling machines often use a spinning disk to spin away the liquid drops sprayed on it, maybe one could use some way of magnets or a charged mesh to just make a plasma shield for the windows? with the sight windows being coated and the glass slides not really quite as good, maybe different glass? the microscope ones are always greenish, the sight glasses would likely be boro or pure SiO2? even more likely tho, deposition rate id say.
Regarding the difference in deposition quality between the chamber sightglasses and the slides: your chamber glass is regularly exposed to high vacuum so it's probably extremely clean. How clean are the microscope slides though? Ben from Applied Science did a series on sputtering on his channel a while back and he did an entire video about cleaning...even the very fine airborne particulates and oils can be enough contamination to interfere with adhesion to the sputtering substrate.
How clean the surface is, is definitely an important factor. But I don't think it was the reason here. Because the surface created by thermal evaporation was that much better. The adhesion was still poor but it had a mirror finish. Nevertheless I want to build a plasma cleaning setup in the future.
The vacuum will cause the copper atoms to go in every direction. I would say a lower voltage and slower deposition rate will minimize deposition of unwanted elements.? I seem to recall, from another experiment, that the experimenters used mirros and placed the viewport behind a shield, though it was for telescope mirrors so the only element they used was aluminum. Obviously that wouldn't matter because their forty-five degree mirrors are already coated with aluminum. That would only deposit more aluminum and since the mirrors are so smooth it would never degrade the reflected image they see through their viewport. So in conclusion, you could use two flat mirrors, ( not regular mirrors like your bathroom mirrors, which have the reflective coating on the back), but two surface coated aluminized mirrors like they used to use in big projection screen televisions, to bounce the image around a shield, that protects your viewport. Coma splices everywhere! I know!
one of the differences i see between sputtering on the target and the interior of the chamer is distance and direction... maybe the coating's adhesion changes with distance and/or direction? material properties too perhaps?
I also considered this. Especially since the outer corners of the microscope slides have a better surface quality. I tested this and suspended the glass further away from the magnetron, but couldn't see a significant difference. Perhaps I was still too close to the plasma. Another comment also suggested that the glass should be kept as far away as possible.
There are good tutorials online. It's basically just a drag knive mounted to the extruder of the 3D printer. But I can try to include a short section about the conversion into the lithography video.
Interesting, I worked in the thin film market designing equipment for 30+ years. Lots to say, but IP tells me not too. 3 keys , Pumping Speed, Temperate, and deposition rate. Many others. It will get more interesting when you work on measuring rate.
Just watched how a blue led is made. I am guessing this is a similar process? Would love to see a 3d printed ball or cylinder coated in metal. Maybe a thermal detonator?
My idea is that the slides are too close to the magnetron, thinking it either get too much copper atoms or too fast or the path the copper takes trough the vaccum get directed by the magnet. Try using an m8 rod and nuts to place the sllides at random places inside the chamber and see what happens to the adhesion and surface flatness. i wonder how much would end up 8 meters away from the sputter if it was a big enough vaccum chamber.
15:51 Its well known that copper doesn't stick very well to glass, you need very high energies to make it do that so it embed it self some nm into the glass. Otherwise first use a chrome or nickel layer because especially the chrome stick very strong to glass and then copper onto that since that stick very hard to metals.
That custom magnetron is AWESOME. Wish I had the time and tools available to make something like that... How suitable is something like a microwave oven magnetron for sputtering?
Thank you! Unfortunately, I also do not have tools like a lathe or milling machine (yet). A microwave oven transformer is the cheapest way to build a power supply. But it has its disadvantages. Power supplies specifically designed for sputtering can detect a sudden increase in current (as in the case of an arc/short circuit) and shut off immediately. This function is not available with a microwave oven transformer, of course.
Could you use a wire mesh at some electrical potential in front of the glass to repel the copper ions? I know it sounds like a lot of effort but would basically lead to maintenance free operation.
At 7:26 you can see where I think you are losing power to plasma formation and perhaps even some minor internal (invisible) arcing, between the bolt heads and the ground plate, through those big holes in the PTFE spacer/insulator. I can't see what purpose those big outer holes serve, but if creating virtual leaks is the concern, I would strongly recommend making a replacement PTFE insulator where those axial through holes are replaced by small, blind radial slits (roughly 1-2mm wide, 0.5mm deep and long enough to connect to the headspace above the bolts with the vacuum) into the surface facing the bolts. This should reduce plasma formation and the associated losses to an acceptable amount, while also giving you a visual indication of any further problematic arcing.
Since my own native language is German, too, I would actually prefer the entire production to be in German - with English subtitles (since he himself seems to be a native German speaker). But then the audience will probably be too small, because the topics of this channel are already very "special". In my opinion, this is one of the best channels on TH-cam (I only discovered it by chance a few days ago...). 😍
@@donbastum4879, das stimmt, aber dafür habe ich eine Lösung: Unter „Einstellungen“ zunächst die automatischen Untertitel (englisch) einschalten, dann nochmal auf „Einstellungen“ gehen und die Untertitel automatisch auf deutsch (oder in jede andere Sprache) übersetzen lassen, das funktioniert erstaunlich gut. 🙂
I recently visited the IMSAS in Bremen, they used movable shields on there sputtering devices to prevent resedue from forming on the Glas. Maybe you could implement something similar.
arcing is a problem that has plagued magnetron sputtering since *checks google* as long as magnetron sputtering has been around for. I understand trying to be 100% home brew, or as close as possible, but I think you would be best served to get a commercial magnetron sputtering power supply. They can be had cheap off of ebay. Much work has been put into killing arcs as quickly as possible, and keeping as little stored energy int he supply as possible so that when arcs do happen a minimal amount of damage occurs.
I also wondered if a suitable coating exists. A thin film of oil would most likely prevent the metal from sticking but would be horrible for the vacuum. Maybe some non polar nano coating? If anyone has ideas, let me know.
@@AdvancedTinkeringNanoparticles was also my first thought but I am not sure if that would do anything. Another idea might be a thin layer of teflon mold release. Another thought: If you apply a bias voltage to the magentron the ejected particles should be charged. Then you could have a charged grid (or ITO coating) on the window to repel them.
@@MaxWithTheSax Repelling the ejected particles with an electric field would of course be the coolest and most convenient option. I will have a look at that. Mold release may work but I doubt it will stay on the glass surface for long when wiping it down. I think for now a piece of transparent vinyl is the easiest solution. Having to replace individual strips of packing tape gets tedious (because I must remove the view port each time). But having a stack of vinyl foil would allow me to just peal of a layer if it is too dirty. Only question is how much they outgas.
How did you clean the glass slides before sputtering? You probably want to clean them with a solvent immediately prior to coating to get rid of any remaining minute amounts of oil.
Bei Wasser würde ich nicht die Steckanschlüsse von Festo verwenden, sondern die zum Schrauben, die halten den Anschluss deutlich Dichter. So Lange kein wirklicher Druck da ist geht das zwar auch, aber würde ich bei der Elektrik lieber vermeiden. Klasse Video, Küsse auf deine Augen.
Very good! You are becoming more and more involved in the interesting part of vacuum technology! Sputtering is one of the most interesting ways of depositing materials, specially when reactive gasses are involved. Few tips:
1) Dont use such high pressures, try and stay in the 1x10-2mBar (get a baratron), if you go to high, the plasma will be super hot and the particles cant reach your substrate because of too much gas particles blocking them.
2) Keep the distance from target to substrate to a maximum, dont let the plasma touch whatever your trying to coat, it will destroy the layer(exactly what your problem is with adhesion).
3) Evaporation rate is usually very low with sputtering (5A/s is considered very high), so keep the energy down, let things be cool and take your time. If the layer gets too thick, and the substrate is cool, from my experience, anything above 2000A will start to flake off...
4) Use high pumping speeds and mass flow controller is a must that has a loop with baratron, i have used MKS systems that controls everything, from the gate valve position to the the mass flow controller, basically a closed servo loop that allows for steady pressure.
5) Consider getting a RF supply and a matching box, you can then do dielectric depositions, with DC you cant.
Thank you very much for the helpful tips!
Today I tested placing the substrate as far from the magnetron as my chamber allows. However, the layer was still very poor, especially where the layer thickness was too high. Interestingly, this problem did not happen when using thermal evaporation, even though the layer thickness was higher. Of course, adhesion to the glass was generally much worse, but at least it was a nice reflective surface.
Other comments suggested applying a layer of chrome first to improve the adhesion of the copper. I will test this as soon as I get hold of a chrome target.
Building a closed loop for the measuring device and the MFC is a very good idea! Sounds like a new project haha.
@@AdvancedTinkering Thin metal films are often under high stress, and in some cases this is the main limitation on achievable film thickness. The stress can be tensile or compressive, and depends on deposition rate, deposition process, substrate temperature and probably a few other things I don't remember.
I think the stress applied to a 1 mm thick glass substrate by a 100 nm thick metal film will warp it enough to be practically measured by the kind of interferometric measurement you would use for lens figuring (I calculate 650 m radius of curvature with 200 MPa stress in film, which would cause 125 nm deflection in the middle of a 2.5 cm span) but there may be a more practical way to go about it if need be.
One idea I would try is pumping down with just air in the chamber first, and running the magnetron at low power to get a visible glow. Should work as a poor reactive ion etch to super clean and activate the glass surface. Then introduce the argon and proceed as usual. Those dangling oxygen (and maybe nitrogen) groups coupled with the complete absence of organics should help adhesion.
As for surface quality, the fact that your windows got a super mirror surface but the slides didn’t just shows your deposition rate is too high - the flux of metal ions above the sputterer is way too dense which is causing agglomeration (blobbiness). Turn down the mA to something like 1/10th what you were using and it should improve a lot. Copper is the worst for this, although not as bad as silver or tin. Refractory metals cool faster, so titanium and tungsten can be coated at a higher current density and go much faster. With soft or low melting metals, you need to go low and slow.
The CVC 601that I ran used a 13.56 MHz RF source.
The distancing will NOT solve his problem. It is ELECTRONS that are hitting the substrate and amorphise and crystal that is growing. You either use a facing magnetron topology (look it up) to try to trap those rogue electrons in between them, increase the magnetic field strength or decrease your power. You do NOT mind if positive ions hit your substrate, it is actually good for the final quality, what you want to avoid is those pesky electrons. This advise comes from someone who routinely can do ALO, ZNO, ITO with hugely good optical and electrical properties.
Also, RF would be nice in some situations, but you can do ceramics also with DC, with reactive sputtering. For example for the Oxides you would just use Oxygen in your chamber, and for the Nitrides, you would just use Nitrogen! To have VERY good looking un-oxidised metal deposits, you can also use a small amount of Hydrogen in your Argon.
If you watch Ben Krasnow's channel he has a whole video on slide prep for sputter coating. It's suspicious that titanium coats well while copper doesn't though so I don't think slide prep is the issue. He does talk about plasma cleaning a lot which you didn't mention. I love this!
Yes, I know Ben's video. His channel is just brilliant overall. In the future, I want to install a plasma cleaning setup in my chamber.
Titanium is used as an adhesive layer in many thin film structures. There are other thin film adhesive metals including: Cr, Al, Ta, Mo, Nb, V, & Hf. Copper won't stick too well by itself 😀.
Finally, you can settle the important question whether rice with a coating of 20 nm ITO tastes better than rice with 20 nm W.
Great Magnetron, by the way, and thanks for all the great info on building the magnetron!
If you want to get rid of the rusting issue on your magnetic yoke: ferritic stainless steels are magnetic. They also conduct heat better than their austenitic counterparts: 1.4003 (ANSI 410) is one.
Thanks for the advice!
I would have loved to coat the puffed rice with gold and actually eat it. But getting a gold target was a little bit outside my budget.
Single gram bullion bars of gold are not too bad, and would be good for this application.
I wonder if there's any nutritional value if you eat puffed rice sputter coated with iron? It might make a great TH-cam scamercial product.
Could the Yoke be coated with Nickel for Example?
Very nice Magnetron.
I hope you are willing to share the Design Files.
In electron beam evaporation vacuum coating the cleanliness of a glass surface must be extremely high, like... atomically clean. When we do multi-layer dielectric coatings for coffee table sized optics to make them into mirrors for the National Ignition Facility fusion laser, the glass is cleaned by removing the top few layers of atoms. We scrub the glass with a toothpaste like thick suspension of nanosized alumina particles (0.05 um) called Baikalox and rinse with deionized 18 megaohm water. You can tell when the glass is clean because it will be perfectly hydrophyllic with absolutely no water beading anywhere. If you watch the DI water dry on the surface in a cleanroom you can see the Newton's fringes forming on the final few layers of water molecules as they sheet off and evaporate. When the glass is then loaded into the coater it is heated to a couple hundred C by quartz lamps to drive off the last layer of water molecules adhering to the dangling O and Si bonds of the glass surface. It sounds complicated and expensive but you could probably do it well enough with a simple setup and no cleanroom for very cheap with any common polishing slurry. The main metric of importance is the total lack of any hydrophobicity of the glass after cleaning.
That’s incredible. Cool stuff!
That must have been a really special (albeit maybe stressful because of that!) project to be a part of!
Did you use thise giant lapping machines for the Baikalox , or some other setup?
@@ericlotze7724 Actual surface figure polishing was done at another facility before the glass was sent to us for coating. We're only removing a few atoms thickness with the final hand cleaning, so it doesn't change the optical properties.
The reason they're held to such stringent standards, is to avoid any anomalies in the reflection... which is a huge factor, when you're dealing with once-in-a-lifetime data. That level of precision isn't necessary for 99.9999% of the market. Sputtering is used in all kinds of industries for various purposes. Sometimes, such as with some aluminum applications, the layer is applied with the intent of near-immediate degradation/oxidation to protect the base material.
For the sacrificial lens cover, my mind immediately thought of clear plastic wrap (e.g. "Saran wrap"). Your project is awesome!
This is exactly what we use in our sputter systems, and it works very well! We use a different sputter technique.
Perfect timing! I was about to eat my puffed rice for breakfast, but now I know how to prepare it properly. 👍
Your chamber is really coming along, love the custom magnetron. There's already a lot of good feedback on cleaning and process parameters already. Couple of things to add.
Most common sputtering metals gold, silver, copper etc have very low adhesion to glass. Even with perfectly clean glass, as the coating thickness increases surface stresses build up and cause the coating to pull away. This causes the rough appearance on your slides. The viewports are farther away which slowed the deposition and reduced the surface stress. Increasing the slide distance to the magnetron will help but thick coatings will always be fragile unless the adhesion is improved. Typically an adhesion promoter layer is added between the glass and the thick conductor layer. A few nm of pure chromium is the usual recommendation since it sticks well to most things, but you could also try a high chrome content stainless, as an easy to source option.
No, the rough appearance on his glasses are due to very bad crystal quality. If you try to reactively sputter ATO, ZnO, ITO, AlO, or whatever transparent, you will see this effect even more pronounced. On metals you cannot really see how bad their crystal lattice is until you do some Xray on them.
It has taken me years to perfect sputtering. In the literature you will find the ways to mitigate the low quality problem: either use facing magnetrons, limit the power, or increase the magnetic field. All these will help with much better quality but will reduce the throughput. The best way is to NOT let ANY electrons hit your surface. They either put a negative bias behind the substrate, or, like in my case, use a totally different geometry to planar sputtering...
So cool, I love magnetron sputtering! Tried it many times myself, with varying degrees of success. I really appreciate the helpful comments too, as there are a lot of potential mistakes you can make.
The reason for the rough coating was a combination of high deposition rate plus the high energy ions building up a static charge. On the viewport you are getting a lower depositing rate and the copper is mostly as neutral atoms by that point. ❤
Cool video, back in the mid 1990's I worked at a company in Sarasota Florida that produced pressure sensors for the automotive industry, I ran a CVC 601 Magnetron Sputtering System that we used to deposit SiO2 on pressure diaphragms then we Sputtered Nichrome on top of the SiO2 and after the diaphragms were removed from the Sputtering machine they went to a LASER where a Weatstone was etched into the Nichrome as the sensing element of the pressure sensors.
*Thanks for posting such a nice explanation!*
In terms of sound tech the voice over part was really neat and for my ears way better to listen to.
6:54 You revealed The Universal Truth about DIY. Nothing more. In my opinion this is important ingredient of being happy and mentally healthy, not to mention side effects of becoming good at what you do after several iterations of improvements. Kudos Sir 😸💪
Ok, perfect, this is what happened to me as well, after having finished my first sputtering machine. Despite sputtering being known for good adhesion, even the adhesion was bad on some cases. The worst part? You almost saw it with the copper deposition: bad crystal quality. Almost totally amorphous. You will see how bad the quality is if you try to do some ZnO, ITO, or anything transparent. So, it took me a whole MONTH to realise what was going on.
The problem causer: ELECTRONS! High speed electrons BREAK your bonds! That is why your side glasses turned out ok, their crystal lattice was not bombarded with electrons. If you read the literature you will see they try to mitigate it that way (as you did by chance), or they use facing magnetrons. The problem there, is when you raise the magnetic strength, you get better quality as less electrons make it across, BUT you get also LESS atoms going across, reducing speed.
I would suggest you either go to another geometry and break free from the standard magnetron setup, or go altogether to a molecular beam. Sputtering in that incarnation is pretty useless.
I hope any of this helps. Regards from Greece.
Great to see that you managed to build such a nice setup.
From past experience of getting glass coated, the companies that did it for me used a layer system with a titanium / titanium oxide layer for bonding to the glass, and then the actual desired coating on top of the titanium oxide.
From some personal experience with magnetron sputtering: depending on the conditions like pressure you can produce clusters of the atoms of your target, which can behave very much different from atoms. Might affect the adhesion and later surface finish.
Well, I do ion implantation first and then sputtering as normal.
Having MOT-based power supply for such experiments is extremelly dangerous. It would be MUCH safier to have high frequency converter because even if you touch HV parts directly you would have a burn but the HF current will not fri your inside but 50 Hz 2000V definitely will. And it would be good to have some ground leak cut-off switch on HV side.
__________
Drops of paraffin for insulation instead of Kapton tape would be much more reliable.
__________
Regarding the solidity of the coating, I have two considerations. First: the surfaces to be coated must be very clean. Ben Krasnov spoke about this on the Applied Science channel. Second: probably, during direct propagation, a large number of low-energy particles reach the target and are not capable of forming a strong connection with the surface. And to the sight glass - on the contrary, the fastest. So perhaps the target object should not be placed on the main axis. This will lengthen the time, but only the best will hit the target.)
I am currently in the process of making my own sputtering system and this was just on time. And I would love any more videos on the sputtering system or PVD for that matter. Great video, great channel!
How did it go? I want to make this but I don’t know what I’m doing or where to start. Do you have any advice? My goal is to coat 3d printed items in metal.
Its really impressive how you build a very professional setup in no time
The word “magnetron” said with a German accent is totally excellent.
it is like cooking. first try was too hot, so it couldn’t stick, but later on the window. also to have quicker results you can move the sample slowly and steadily. like roasting a marshmallow.
You should first process the slides with plasma cleaning to prepare the surface for sputtering. A cool experiment will be to sputter quartz over aluminum to make a durable first surface mirror.
Plasma cleaning is also a project I will work on in the future :)
@@AdvancedTinkering As a quick test, you might want to try flaming the microscope slides - pass them through a butane torch flame a few times. This apparently really cleans the surface, both by thermally evaporating any contaminants, and also reaction of the ions in the flame. Many years ago when I was playing with thermal deposition, I tried this and it really seemed to help.
@@imajeenyus42 I haven’t heard of that method before. Thanks a lot! I will try that.
He could increase the power and get to ion implantation. Surely there atoms will stick :)
@@BillDemosI had typed a long comment out on my phone, and a stupid telemarketing call ruined it.
I just found this channel, watched the video, and read some comments, and you seem extremely knowledgeable on this subject, as well as the preparation involved.
If someone were hoping to get a high level of cleanliness, using butane is inappropriate because...? The common sense part of my brain says that cleaning the target with acetone to get rid of any oils, doing a final clean with deionized and distilled water, and then drying the slide using a butane induced flame would be counter productive. This is because the possibility of unburned butane depositing more petroleum oils on the surface that aren't removed in the flaming passes would remain after the final pass. How correct is this thought process? Is there something else the flame cleaning would do to the target surface itself? Maybe some micro ablation from the forced air of the torch in a non-cleanroom/filtered air environment?
Titanium is easy, you just need a hot titanium wire in a vacuum. The Thought Emporium has some great spuddering videos.
that magnetron looks more professional than the actual commercial stuff
but keep in mind those magnets have a quite low curie point of like 85°C or something. there are H grades that withstand 150°C but at limited sizes.
Thank you very much! It has grown a bit larger than necessary because I lacked some experience in estimating the material thicknesses with which the milled components would still function. The magnets should never exceed 50°C through the water cooling. If I touch the target directly after sputtering, it is only slightly warm.
Man! You are so smart! You did an excellent job designing your parts. The whole video is very interesting and entertaining. I laughed aloud at some parts. I def noticed you being funny.😊
Love seeing the progress you're making and the process you're going through! It's just so interesting and I didn't know how much I needed more since watching Ben's videos on sputtering way back when.
Nice Video, I like how your production quality went way up
I'm patientley waiting for that magnetron video! Hopefully with pricing I'm very curious how expensive PCBway's CNC is! Because my local CNC guy is very expensive
Another great video, keep up the good work, and I am eagerly waiting for the Magnetron video.
I think it is the deposition rate leading to a different grow on the substrate and on the view port glass. You view port glass is much further away.
Another thing to consider is heat. Your plasma directly touches the substrate. I guess it will get hot during the sputtering. This can also reduce the quality of the sputtering process.
And last but not least: You cleaned your glass substrate, right? At least with some isopropanol and then water in an ultrasonic bath for 10 min.
I think this is your best video yet. I'm sure the editing and production took a lot of work, and it shows. Keep it up!
Thank you! I appreciate it! Yes the editing sometimes takes longer than filming...
This is where having a very good knowledge of both physics and chemistry is most useful. Someone in an earlier comment mentioned glass being _atomically clean._ I'm not sure to what degree of cleanliness is needed and whether it can be achieved physically or chemically- possibly through a powerful oxidizer like Piranha solution. Other factors to consider would be obvious high voltage and heat dangers. Not sure if the charged particles have enough energy to generate X-rays of concern. If you have a copy of Strong's _Procedures in Experimental Physics,_ you could make some fascinating things, although one could make things of serious danger.
Now turn the whole thing upside down, put the plasma field on a CNC gantry over a stationary surface, and reduce the coating application area to sub-millimeter size. Boom, metal FDM printing, and you're rich.
FDM full form please
Extremely informative. Thanks for the deep and thorough dive!
🥳
Wow such a cool video, very well done!
Thank you! I'm very excited for our next collaboration!
Online machining services are great, but you're not a fully functioning human being without machine tools of your own!
Great work though, it's only when you see the real practicalities of a technology that you actually understand it, Thanks!
Haha, that's true! As soon as my budget allows, I will definitely get a lathe and milling machine.
Also a corrosion-resistant first-surface mirror is relevant to my interests.
I hope you can make more frequent videos now with your new awesome setup. Love the videos, please keep them coming!
Thank you!
Time is still a limiting factor. But I will do my best to make videos more frequent.
Looks like the glass just got coated super duper slowly. Try putting a slide on the glass vieport after cleaning it the same way and see if it works
13:08 I loved your reaction here at your fix working so well. I know the sound of joyous hand rubbing anywhere 🙏😁
Overall seems like a really nice setup and I look forward to seeing what more you do with it. I think it was a good idea to make the deeper dive on the magnetron a separate video and I'm looking forward to that, too.
Here's a guy with the best toys.
FWIW, Ti is a common adhesion layer in semiconductor manufacturing because it sticks so well to Si / SiO2 / glass. If you put a second magnetron in there to do Ti before other metals you'll have far better results.
wonderful video absolutely amazing as always. keep it up 👍
Thank you a lot!
Bro thought he could casually flex his Organspendeausweis without us noticing
Shouldn't he have filled in his name and signed it?
@@pattheplanter Probably a spare one. They're free to order online, for example to give to friends and family
microwave magnetrons are usually powered by only half wave rectified DC, so using a **full wave bridge rectifier** (electroBOOM!) might be drawing too much current for the design of the transformer, making it hot.
That’s good to know! But with the new rectifier the MOT does not get hot anymore. But if I ever need to increase the current, I will have to add a fan.
Fascinating process, well explained too :)
Definitely interested in the power supply as well
That was fookin' fresh! 🌈🤗
This project is insanely interesting. I’d love a more in depth analysis of how you designed your parts.
Is this device capable of Plasma Assisted Chemical Vapor Deposition (PA-CVD) for thin films of Diamond Like Carbon (DLC) or heck even *proper diamond?*
should be possible, the magnetron is more or less the same. He might need a RF-source though...
I wonder if there some coating on the slides that may be interfering.
Cool thing! Never stop that man!!🤩🤩
You should make a pair of custom mirrored sunglasses. That would look awesome. The thought Emporium did something similar.
Thanks great project thousands of potential, no pun, experiments
16:32 The reason you get a perfectly coated side windows and not a perfect microscope slide is due to the temperature difference between the 2. The microscope slide is too near from the plasma and the material is not really condensing in slow atomic layer. Nest time, if you put methane gaz and a diamond sample, you will be able to grow diamond.cUS20110151226A1 - Synthetic CVD diamond. If you search in the world most favorite search engine this phrase : H BAR Patents for C30B 25 - Single-crystal growth by chemical reaction of reactive gases, e.g. chemical vapour deposition growth (13,302). You will have all the patent from 2010 about CVD (Control Vapour Deposition)
1. Increased heat on the substrate actually fixes quality problems. His problem is elsewhere, I have explained him in another comment.
2. To do diamond, you need Methane as you said, but only at around 1% in Hydrogen. If you don't use Hydrogen in the mix, you will get graphite instead of diamond.
@@BillDemos Thanks of the answer, your hypothesis is really good. Just need to repeat It with other material.
Your anwser :
@BillDemos
il y a 1 heure
Ok, perfect, this is what happened to me as well, after having finished my first sputtering machine. Despite sputtering being known for good adhesion, even the adhesion was bad on some cases. The worst part? You almost saw it with the copper deposition: bad crystal quality. Almost totally amorphous. You will see how bad the quality is if you try to do some ZnO, ITO, or anything transparent. So, it took me a whole MONTH to realise what was going on.
The problem causer: ELECTRONS! High speed electrons BREAK your bonds! That is why your side glasses turned out ok, their crystal lattice was not bombarded with electrons. If you read the literature you will see they try to mitigate it that way (as you did by chance), or they use facing magnetrons. The problem there, is when you raise the magnetic strength, you get better quality as less electrons make it across, BUT you get also LESS atoms going across, reducing speed.
I would suggest you either go to another geometry and break free from the standard magnetron setup, or go altogether to a molecular beam. Sputtering in that incarnation is pretty useless.
I hope any of this helps. Regards from Greece.
You could nickel plate that iron plate pretty easily if you wanted to stop it rusting. Also I think PET film might be good for a sacrificial layer for your windows. It's easily available (florist shops have it for wrapping flowers) and should be fine in high vacuum. You could even rig it to be pulled off during runs if you needed to.
Perhaps you could just cut a cylinder of it and sit it inside the whole chamber, it's cheap and stiff enough that it'll just sit there.
The copper coating the viewports is not colinear with the plasma. Perhaps the purest copper with the optimal deposition velocity can be found at the fringes. The viewports are far from your source and might also be colder than your intended deposition target.
hey i want to see you talk about the powersupply in details as well. You built it, i think you should make the content for it. your insights are very valuable.
In my second video about the different parts I also briefly covered the power supply: th-cam.com/video/MslftqA7LYw/w-d-xo.html
With the copper, you should try doing a base coat with titanium or something similar and then the copper will adhere a bit better.
To get a even coating the substrate often gets moved around. Also, the scratch test can be refined by pre scratching a grid in the thin Film, and then trying the tape test.
I wonder if you could lay down a thin film bimetallic strip on something water soluble like sugar? Would you get a thin sheet of metal foil or aren't the metal particles sort of welded to one another? If that would result in thin metal foil, it might be cool to try to make bimetallic strips or thermoplastic. In any case, thank you for your contribution to puffed rice cake technology.
13:20 Actually, we can see the green color of the copper ions in the plasma. Perhaps it didn't show clearly for you in your camera, but the video does show it clearly. Copper always has such a nice, pretty green glow.
The rough surface could be due to pitting from plasma etching. The slides were very close to the plume so even though the plasma was depositing metal ions, those same metal ions would cause etching and pitting because of their high kinetic energy. The viewports were further away and indirect, so they got all vapor and no etching. You can try moving the substrate further away from the plasma plume to minimize the etching.
You can 100% deposit ITO with thermal evaporation, ceramics give off an evaporation pressure suitable for coating well below even their melting point. Maybe the boats are unsuited for ceramics due to thermal conductivity or something, but I have with my own eyes seen it done with an e-beam.
I didn't know that. I always thought it was only possible using reactive thermal evaporation. Thanks for clarifying!
@@AdvancedTinkering SiO2 (bog standard glass) would be a good ceramic material to test this out with if you are ever at that point.
@@AdvancedTinkering hi I have problem with my dc magnetron sputtering. Short circuit occurs every time I start increasing the power. Help me to solve this issue
Im excited for the prospects this device brings,cant wait to see more stuff sputtered on things. Can you try sputtering some metal oxides and other ceramics in the next videos? Or maybe you can even attempt making a beryllium mirror?
16:00 could it be something to do with how the target glass is directly above the sputterer? could the copper have too much energy when hitting the target and not stick properly?
Yes, I'm relatively sure the distance is at least one factor. As you can see, the edges of the microscope slide has a "mirror finish". Just like the vacuum chamber walls.
I tried increasing the distance but it did not change much. The copper "flaked" in the middle. Maybe I have to preheat the substrate.
I would assume the issue with the glass is that the slides weren't sufficiently clean to get a firm connection. Maybe the copper is just more sensitive to impurities than the titanium? I don't really know, though, since ultimately I'm not at all trained in this.
Another thing you could try could be putting a light layer of a metal that sputters easily, such as titanium, on first, just barely thick enough to coat it, then putting down a layer of copper on top of it, sort of like how chrome plating first deposits nickel or copper below the chromium.
To achieve perfect coating on glass - you need perfect clean glass surface first. IMHO, best way - cleaning by oxigen plasma. It burns out all kinds of organic residuals on suface. It may done in your vacuum chamber with minimum modifications. Other way - chemical cleaning and activation of surface.
I definitely have to improve my surface preparation. In this video I mainly wanted to test the magnetron. I plan on building a plasma cleaning setup inside the chamber.
I use plastic binder sleeves as viewport covers in my sputter chamber. You are using what looks like a type 2 unbalanced magnetron with a much stronger outer magnetic field. This is going to increase the energy of the species arriving at the deposition surface and will increase the compressive film stress. If the compressive film stress is higher than or very close to the maximum adhesion stress the deposited film will delaminate from the substrate. The viewports are seeing much less compressive stress and the film adhesion is higher. If you need any assistance I design industrial large are sputtering equipment for a living.
11:34 did you mean to say "violated microwave"? with that organspendeausweis gets a different meaning
i do actually insanely relate to the "eating something and being annoyed that it isnt coated in metal" part
also i really wonder with the viewport, theres various variations of ways to keep cameras clean, milling machines often use a spinning disk to spin away the liquid drops sprayed on it,
maybe one could use some way of magnets or a charged mesh to just make a plasma shield for the windows?
with the sight windows being coated and the glass slides not really quite as good, maybe different glass? the microscope ones are always greenish, the sight glasses would likely be boro or pure SiO2? even more likely tho, deposition rate id say.
Regarding the difference in deposition quality between the chamber sightglasses and the slides: your chamber glass is regularly exposed to high vacuum so it's probably extremely clean. How clean are the microscope slides though? Ben from Applied Science did a series on sputtering on his channel a while back and he did an entire video about cleaning...even the very fine airborne particulates and oils can be enough contamination to interfere with adhesion to the sputtering substrate.
How clean the surface is, is definitely an important factor. But I don't think it was the reason here. Because the surface created by thermal evaporation was that much better. The adhesion was still poor but it had a mirror finish.
Nevertheless I want to build a plasma cleaning setup in the future.
Aerogel would also be neat. Would love to hold a block with a nice smooth metal finish
The vacuum will cause the copper atoms to go in every direction. I would say a lower voltage and slower deposition rate will minimize deposition of unwanted elements.? I seem to recall, from another experiment, that the experimenters used mirros and placed the viewport behind a shield, though it was for telescope mirrors so the only element they used was aluminum. Obviously that wouldn't matter because their forty-five degree mirrors are already coated with aluminum. That would only deposit more aluminum and since the mirrors are so smooth it would never degrade the reflected image they see through their viewport.
So in conclusion, you could use two flat mirrors, ( not regular mirrors like your bathroom mirrors, which have the reflective coating on the back), but two surface coated aluminized mirrors like they used to use in big projection screen televisions, to bounce the image around a shield, that protects your viewport. Coma splices everywhere! I know!
love this channel
Thank you! I appreciate hearing that!
That's pretty cool... I want one
one of the differences i see between sputtering on the target and the interior of the chamer is distance and direction... maybe the coating's adhesion changes with distance and/or direction? material properties too perhaps?
I also considered this. Especially since the outer corners of the microscope slides have a better surface quality. I tested this and suspended the glass further away from the magnetron, but couldn't see a significant difference. Perhaps I was still too close to the plasma. Another comment also suggested that the glass should be kept as far away as possible.
Id love to see the 3dprinter convert, vinyl cutters seems to be all the rage, but they're unreasonably priced at least here in aus.
There are good tutorials online. It's basically just a drag knive mounted to the extruder of the 3D printer. But I can try to include a short section about the conversion into the lithography video.
Interesting, I worked in the thin film market designing equipment for 30+ years. Lots to say, but IP tells me not too.
3 keys , Pumping Speed, Temperate, and deposition rate. Many others. It will get more interesting when you work on measuring rate.
Most glass has surface oils on them. If water doesn't make a film on a glass surface, it's dirty. Use solvent to clean.
You can use martensitic stainless steel for the magnetic yoke - it is at least magnetic.
Just watched how a blue led is made. I am guessing this is a similar process? Would love to see a 3d printed ball or cylinder coated in metal. Maybe a thermal detonator?
My idea is that the slides are too close to the magnetron, thinking it either get too much copper atoms or too fast or the path the copper takes trough the vaccum get directed by the magnet.
Try using an m8 rod and nuts to place the sllides at random places inside the chamber and see what happens to the adhesion and surface flatness.
i wonder how much would end up 8 meters away from the sputter if it was a big enough vaccum chamber.
Man that plasma looks so nice!! Way aren't you using the high voltage power supply you built few videos ago?
Thanks! The voltage produced by the power supply I built in an earlier video is way too high and it cannot supply enough current.
15:51 Its well known that copper doesn't stick very well to glass, you need very high energies to make it do that so it embed it self some nm into the glass. Otherwise first use a chrome or nickel layer because especially the chrome stick very strong to glass and then copper onto that since that stick very hard to metals.
That custom magnetron is AWESOME. Wish I had the time and tools available to make something like that...
How suitable is something like a microwave oven magnetron for sputtering?
Thank you! Unfortunately, I also do not have tools like a lathe or milling machine (yet). A microwave oven transformer is the cheapest way to build a power supply. But it has its disadvantages. Power supplies specifically designed for sputtering can detect a sudden increase in current (as in the case of an arc/short circuit) and shut off immediately. This function is not available with a microwave oven transformer, of course.
Some great thin film adhesion materials (3-10 nm): Cr, Al, Ta, Mo, Nb, V, & Hf.
Could you use a wire mesh at some electrical potential in front of the glass to repel the copper ions? I know it sounds like a lot of effort but would basically lead to maintenance free operation.
damn this is so cool
do you ever feel like a mad scientist? Sitting in a dark room watching a large contraption emit a massive blue light LOL
At 7:26 you can see where I think you are losing power to plasma formation and perhaps even some minor internal (invisible) arcing, between the bolt heads and the ground plate, through those big holes in the PTFE spacer/insulator.
I can't see what purpose those big outer holes serve, but if creating virtual leaks is the concern, I would strongly recommend making a replacement PTFE insulator where those axial through holes are replaced by small, blind radial slits (roughly 1-2mm wide, 0.5mm deep and long enough to connect to the headspace above the bolts with the vacuum) into the surface facing the bolts.
This should reduce plasma formation and the associated losses to an acceptable amount, while also giving you a visual indication of any further problematic arcing.
Very cool , great video . Please publish plans for the magnetron . Thank you
“Honey,what’s wrong, you haven’t touched your metalized puffed rice?”
*Subscribed!*
I appreciate it!
Very interesting video keep it up but what would be cool if it also has German subtitles.
Since my own native language is German, too, I would actually prefer the entire production to be in German - with English subtitles (since he himself seems to be a native German speaker).
But then the audience will probably be too small, because the topics of this channel are already very "special".
In my opinion, this is one of the best channels on TH-cam (I only discovered it by chance a few days ago...). 😍
@@tonga1968 das stimmt aber deutscher Untertitel wer doch cool oder ein zweit Kanal auf deutsch 😉
@@donbastum4879, das stimmt, aber dafür habe ich eine Lösung: Unter „Einstellungen“ zunächst die automatischen Untertitel (englisch) einschalten, dann nochmal auf „Einstellungen“ gehen und die Untertitel automatisch auf deutsch (oder in jede andere Sprache) übersetzen lassen, das funktioniert erstaunlich gut. 🙂
@@tonga1968 jaa cool das hat geklappt danke dir
I recently visited the IMSAS in Bremen, they used movable shields on there sputtering devices to prevent resedue from forming on the Glas. Maybe you could implement something similar.
Maybe with magnets.
Good job!!
arcing is a problem that has plagued magnetron sputtering since *checks google* as long as magnetron sputtering has been around for. I understand trying to be 100% home brew, or as close as possible, but I think you would be best served to get a commercial magnetron sputtering power supply. They can be had cheap off of ebay. Much work has been put into killing arcs as quickly as possible, and keeping as little stored energy int he supply as possible so that when arcs do happen a minimal amount of damage occurs.
I wonder if there are any coatings you could put on the window that would prevent the copper from sticking.
I also wondered if a suitable coating exists. A thin film of oil would most likely prevent the metal from sticking but would be horrible for the vacuum. Maybe some non polar nano coating? If anyone has ideas, let me know.
@@AdvancedTinkeringNanoparticles was also my first thought but I am not sure if that would do anything. Another idea might be a thin layer of teflon mold release.
Another thought: If you apply a bias voltage to the magentron the ejected particles should be charged. Then you could have a charged grid (or ITO coating) on the window to repel them.
@@MaxWithTheSax Repelling the ejected particles with an electric field would of course be the coolest and most convenient option. I will have a look at that. Mold release may work but I doubt it will stay on the glass surface for long when wiping it down. I think for now a piece of transparent vinyl is the easiest solution. Having to replace individual strips of packing tape gets tedious (because I must remove the view port each time). But having a stack of vinyl foil would allow me to just peal of a layer if it is too dirty. Only question is how much they outgas.
Perhaps you need to etch the glass slides. Maybe they are too smooth / not smooth enough?
How did you clean the glass slides before sputtering? You probably want to clean them with a solvent immediately prior to coating to get rid of any remaining minute amounts of oil.
Bei Wasser würde ich nicht die Steckanschlüsse von Festo verwenden, sondern die zum Schrauben, die halten den Anschluss deutlich Dichter. So Lange kein wirklicher Druck da ist geht das zwar auch, aber würde ich bei der Elektrik lieber vermeiden.
Klasse Video, Küsse auf deine Augen.