Ceramic pcb evaluation project brainstorming

[LaserSteve]

Leave the Brown flux out of the equation. It has nasty Flouro.borates  Boric Oxide is both a glass former and a metal flux. Boric Acid is readily available all over the US, It will usually lower the melting point of the glass. Baking it at 170 F removes I belive 5 water molecules [pentahydrate?] prior to use.  If I wanted to reduce my need for fluxes I would think about firing in wet Hydrogen, a reducing atmosphere, or under nitrogen or inert gas.

Hydrogen furnaces are wonderful tools and often forgotten.  Hydrogen however comes with its own safety issues. So the usual trick is to pack the furnace tube with Carbon, which forms Carbon Monoxide.  Adding carbon dust to glass usually just scavenges oxygen from the glass. That has its uses.

Steve

[coppercone2]

I had H2 before I got tired of paying a lease for it. It is weird especially how long the gas torch flames are, its a light saber. You need to setup a special work area to ensure that you won't fry something with your laser beam when you set the torch in the holder. But I thought the performance for welding aluminum was abysmal, I guess you need a whole different technique for using it.


I think it is relatively use able but you can't leave it unattended for any reason (the hydrogen furnace) in case the flame goes out. I always imagined though, if its out in the open, its not gonna be capable of doing squat, it seems that the roof is going to be the major culprit in any h2 problems.

[coppercone2]

the palladium silver trace looks better then the silver trace. its like fairly legitimate looking.

I think that white shit is less/comes off easier too.


Yeah this one catches my eye from a distance.

[coppercone2]

You could try Antimony doped Tin oxide - thats conductive. So is lead dioxide, but you said you don't want to play with lead. Theres also the titanium suboxides - Ti4O7 can be made from titania (TiO2) by reduction with carbon, hydrogen or titanium metal at 900C to 1000C.
Pretty sure magnetite is conductive too, and so is manganese dioxide.

The trick is picking something that will sinter readily at temperatures you can attain.

Ok I bought antimony tin oxide powder. Maybe I can learn a bit about how these things behave


I wonder how reasonable it is to make a "hybrid" with a LM317. Maybe you can even file it down to set the voltage divider ratio. Not with the wire bonding, but to solder a regulator to a ceramic PCB with ITO resistors, if they work at all.

[LaserSteve]

96% AR 4% H2 comes in standard tanks for lowering the  risks in active bonding/brazing.

We switched to the mixed gas version after a female graduate student let H2 bleed into the room through a 800 c' furnace tube for 40 minutes without igniting the burn-off pilot flame. . Cell phone call from arranged marriage fiance  back home was more important to her.  I didn''t have to yell at her. The instant  "shark attack" from three other graduate students was more than sufficient if not excessive.  Then her fiance called her back and told her in no uncertain terms that he would like his future bride alive and in one piece when she returned home.

So I understand your concerns, but tge flow rate required is miniscule in an electric furnace.

From what I have seen of arranged marriages, they work very well when the extended family is involved.

The CNTs grew better in the gas mixture then in the pure H2 anyways.

The amount of Hydrogen needed is small, the lack of Oxygen is the key.

Steve

[LaserSteve]

It looks like your headed toward thermal evaporation or sputter coating in vacuum. Especially with ITO.

With your mechanical skills both are within reach.  Thermal for aluminum for example can be done with just a roughing pump. I once visited a company that did Aluminizing for the auto industry.  Their main chamber pumps had six inch diameter pistons, not even a rotary vane pump. Roughing pump plus n2 flush, that was it.  They had man sized diffusion pumps, however it was cheaper to just flush the chamber with N2 from a Dewer of LN2.

STEVE

[coppercone2]

Will the ITO work with a frit though?

I got my ITO. I wanted to mix it up ... 3%? with powdered glass, just like with the silver, to try to make a resistor.


That stuff is all vacuum and way outside of my equipment level right now.


I saw the sputtering stuff on youtube, it all looks like "bell jar" vacuum type stuff. I think its more like a very high temperature solder pot and a coil, not a kiln


I am suspicious that it might be easy to make a evaporator out of A1 Kanthal wire, especially in a vacuum, if you have the right power supply. I think alot of the problem is lack of high current supplies, so things need to be matched for direct to outlet connection via heater impedance... but the same power level is possible in other structures so long you have a proper supply, I think...


If its in a vacuum you might not need a stupid amount of thermal insulator to make it work.

[coppercone2]

well I made a smoke blue blob of ITO and glass powder and I am baking it right now.

I will let it cool down slow in the oven and then test conductivity and adhesion.


This one is alot harder to mix, I probobly need to find some tiny ball bearings to shake with the ITO in a bottle. It likes making clumps, This test might be totally inconclusive because of poor mixing. I see what i need is a tiny quartz? mortar and pestle


Perhaps something called a "tissue grinder" will work. It looks just the right size.

Why does that sound so creepy? I am just making imitation resistors

[coppercone2]

should have done a bit more research, stayed 100% a powder at 800C.

Still, maybe a formulation with more glass will work

[T3sl4co1l]

I saw the sputtering stuff on youtube, it all looks like "bell jar" vacuum type stuff. I think its more like a very high temperature solder pot and a coil, not a kiln

A bit fancier than that, sputtering uses a magnetron (DC power into a vacuum diode with a magnet nearby).  Plasma bombards the surface, knocking loose molecules of the stuff.

Just plain old thermal evaporation (PVD) apparently works with ITO, load some in a boat and heat it up.

Tim

[coppercone2]

Lol, I think the ratio is between 95% to 30% glass to ITO/etc. With silver its 3% glass because silver is nice.


What would be nice is if you can get a stainless steel bowl with a vacuum dome on it, then hit the bottom with a torch. I think if its big enough it might not break a vacuum seal (say with that grease) but also spot heat the ITO in the middle enough to get it to evaporate. Doing this electrically seems like very difficult apparatus. I wonder if you can just torch a cooking pot with a lid on it and some stuff suspended in it. I think that might work. It says 600C, that means you might be able to use a propane blow torch. I just wonder if it would warp too much or something. Maybe wrap a copper tube around it to prevent the walls from getting hot. I just don't like the idea of trying to get thermocouple wire in there.

Every way I think about it, a probable cost over run occurs. Maybe one of those infrared thermometers can be used to hold a set point with the torch (I expect the absolute value is bullshit, but if you can put something in there that shows you what the correct reading is, you can try to maintain that reading). Then you don't have any feed throughs and shit like that. Like you can put a little ingot on the bottom of the pot that melts at the target temperature and see what the reading is on the IR thermometer from the bottom when it melts, then try to hold that number.


I wonder if one of those low quality piece of shit hot plates with the exposed coil can do it. They basically glow. it must be more then capable of over heating a pot on the bottom. I however don't have one.


I definately need to make some kind of zero cost experiment with what I have before going off on a vacuum tangent. I don't want to think aBOUT what kind of trouble ima get into with vacuum feed throughs.

[coppercone2]

Hmm even at 30% glass frit at 800C it did not stick.

I wonder if you need a higher temp or longer time, because I saw no evidence of any molten glass etc. It was slightly more durable then 3% though, but only slightly.




I wonder what would happen if you added some silver into it. It stuck great to the silver trace that was on the alumina already (this time I used a To-220 ceramic heat spreader as a substrate).


Maybe next time I can push it to like 950C. But I need my mixing tool because I am not convinced its mixing right.


However, before it broke off from the most gentle of probing, It did infact read as a resistor in the kOhms for a blob


And a random procesing tip, I noticed if you let it dry up with PVA , then you scrape it up, kinda chop powder it, and add more , it becomes more uniform as a paste



I tried it again, this time at 880C. I don't get it why it hardly sticks, I can rub it off with my finger. 


What is this magic of silver. The glaze is meant for cone 04 to 06. I could up it up to 1000 maybe. But the traditional thick film resistors with expensive iridum etc only get fired from 700 to 900.

I say it might not work lol 



I wonder actually if its the floor of the kiln. I think when I did my silver I had it on a little brick to elevate it from the floor. Perhaps its alot colder then I think! So maybe its a process error. I think I read something that said just putting stuff on the floor = problems.... but I don't think it could be that big a difference.

I wonder what those other oxides have that this one does not. I wonder if you can add small amounts of ruthenium etc to ITO to make it behave better.


I got some little porcelain mixing chemistry tray thingy so I can make a bunch of different mixtures and test them in parallel.

[LaserSteve]

GLAZES are often  pre-fired at the factory, then ground to powder.  Once fired the compounds added to aid in melting  / glass forming  are often evaporated or consumed, making it unlikely to wet some materials.

YOU may need to buy a low temperature  sealing frit.

Also, the COE of the glass has to nearly match the COE of the substrate in monst cases, otherwise the fused material may just fracture right off.

See attached

STEVE

[LaserSteve]

6.5 g of lead(II) oxide,
3.5 g of boric acid
0.5 g of zinc oxide

Fuses nicely to my Alumina crucible, but without adding potassium, calcium, or sodium oxides is otherwise a slightly  water soluable novelty. VERY TOXIC to make,  Adding trace amounts of Copper [II] oxide will color it blue and strengthen it a bit.

Requires a melt-fracture-remelt process, and you sure as hell do not want to breath the fumes when melting.

While that will form a glass at 850'-900'C, without trace additives like Bismuth Oxide, it's just not very useful.

Adding traces of other materials tends to drive it's Tg through the roof.

DID I mention it's brittle?

This mixture was a high school experiment at one time. NOT TODAY!  PbX compounds result in neurotoxicity, especially in young women.

Buy a commercial frit. The above mixture is a starting point for a series of glasses / glazes  now banned in modern countries.

I ran that in a fume hood. The oxides given off will condense on a kiln lid. It was part of an official project at a former employer.

Just an example of why you need to be careful.

The flare off in the gas furnace turned green from the Boron floating around and Yellow from the Lead..

I later learned that it was leaching the alumina from the crucible.

AGAIN, an example why you need to be careful.

Steve
 

[coppercone2]

well I do have another unanswered question, that is what silver oxide does if you sinter it. I am going to try that one.

Tin I regard as a troll metal like Al.

The thing with the frit kinda sounds like you need flux... because with the CTE, if that was the only problem it would pop off like a solid piece right? This stuff has 0 binding power. I too imagined there is a lack of some reaction going on.


Why  is flouride salt such a bad idea? It works for brazing. I think flouride might do whatever needs to be done with glass, because of HF reacting with glass. I am not terribly worried about a such a small amount (I do like as 1210 resistors worth) of flourine containing paste doing something... should i be?

My prototyping method is to smear it between parallel scotch tape and peel it off. It is a very small amount. All the brazing flux I use has flourides.. that has got to be way worse exposure. I don't braze that much, but that clearly makes some contamination getting into the flame.

Anyway, the silver oxide.. maybe it can make a low value resistor?


Regardless, for now I am happy that I can get a trace that is solderable. I can still do my filter and a PCB using regular resistors. I thought the iTO would be a free add on but I guess that is complicated like most things usually are.


And  for resistor development, I think that getting some of the traditional material (ruthenium, rhenium or iridium) and trying to sinter it with the glaze powder might get me some good direction about where to go. If it does not work, chances are its the glaze that is the problem, and maybe the ITO has merit. If it does work, chances are the other oxides might be further complicated.... its a modest price to answer so many questions.....



In that case, I would only have one unknown. if I just wanna work on the ITO directly being as cheap as possible to avoid expensive materials, I have two unknowns.. .. and it might not be worth the cost savings at all for my R&D effort! I have a whole body of literature to go through if I fail on a ruthenium resistor (mass produced).. unless I happen to get a PhD in inorganic chemistry. For all anyone here knows, getting ITO resistors might require a large number of special considerations, DESPITE it seeming like a logical choice for entry level research


I am imagining it kind of like a alien wanting to be a plumber. He saw the concept of a pipe. And then imagine he bases his initial designs and learning on aluminum pipe because its cheaper then copper. I mean for basic water stuff, any he never compared it to a copper pipe. It would be a giant school of hard knocks, or a alien electrician that thought he could setup infrastructure for his outpost with zinc wire ! . Had they spent a few dollars more on some copper, they would realize that their roadmap to success is totally freaking insane.. hopefully buying some precious oxide might prevent such a course of actions. Or that poor fool that decided to buy eagle cad (15 years ago I mean) without trying altium.

*****

The silver oxide glass attempt will be my last gamble for now, I will stick with traces and possibly experimenting with a traditional resistor if I decide to buy some reagents in the future, before I start trying to mess with frit glass composition and random oxides that sound good.

I have a whole bundle of work to figure out how to make a stencil that is good, paste spreading, recycling, etc. I did get my silver bearing soft solder (eutectic) so I can solder the traces with less disturbance.

I am thinking if silver oxide works, even if its low resistance, perhaps a ceramic "hybrid" low resistance PCB can be made for a low noise amplifier opamp chip, that uses low impedance resistors. Like a LT1028. My guess is based solely on the idea that I have good luck with silver, it has a good history of generally not irritating me. Tin is generally a dog (tin solder  ), indium is a T1000 and even copper oxide pisses me off because it means corroded copper, and lead (not eutectic in any form) makes me feel like I am a peon in the year 1400, or working on some posh dickheads bathtub liner in a manhattan pent house

[coppercone2]

a ha, I figured out why maybe they don't use silver oxide. It decomposes into silver at 300C. 

I figured it would be like copper oxide, without looking at the measured values


Well, maybe it makes some kind of silver glass foam material if it decomposes during sintering. very interesting

But it does melt at a low temperature which gave me a idea, infiltration resistors. If you put a frit on a substrate and soak it in liquid oxide. Maybe it would infuse and solidify to form a resistor. The easiest way to find out would be to get frit blocks and see if it wicks up in a pool of molten oxide. Maybe you can make a big resistor with that technique too. Its not exactly a shake and bake operation but it does make me wonder


It is at 570C, it looks grey still. It does not look like the silver trace near it. I wonder how this will play out when its done. I have also been reheating the initial silver/pall trace many times now to see what thermal cycling might do to the deposit, also to save on substrates. It does not look white like the silver did after sintering either.


Well the silver oxide got to 820. I turned it off. It looks different then the trace nearby, which looks rather white at this temperature (palladium-silver). I wonder what will turn out of this experiment.


The sample has a ability to burn human skin even when temperature is reduced to 300C after a short cooling period. The trace is a grey color, unsure if its because of the extra glass content or the silver oxide has this property even if it supposedly decomposed into silver.

When scraped, metalic silver is exposed. When measured, it has a similar conductivity to the silver palladium trace.


It appears if you for some reason have silver oxide, the result will be very similar *but a little crappy looking* compared to using fine silver powder. The resistance is approx 5% higher for a wider trace then the silver palladium one, I suspect mainly due to the increased glass content (15% or so).



So now you know, AgO is a usable feedstock if Ag is not available. It looks a bit bumpy and blistered. It is solderable but inferior. It would be interesting to know what a 3% glass Ag2O mix would do. Always nice to be able to use different reagents for the same process. 

If the process did work, I think it would have a similar conductivity to copper metal glass traces, which might be worth a experiment too.


Anyway, it was fun and I think it was worth a shot despite realizing that it would decompose right after I mixed it up.



Also further notes
1) porcelain mini bowls (like a sample plate) make for excellent mixing vessels for powder/PVA mixtures
2) small silicone spatula is great for mixing an applying it, but you need a stiff metal for the 'printing' part of screen printing, the silicone deforms and removes it from the cavity with any pressure applied


If anyone has more ideas, preferably non toxic, for resistors, I might try them out, but no lead, cadmium, or things that might otherwise foul my kiln. The price of the proper metals for resistors is very high that it seems that I could run a dozen or more cheap experiments and still be at a cost.

So far research results say
1) silver is good
2) silver oxide does not work because it decomposes, but you can seemingly make a usable PC trace with it if you desire with a similar end appearance to silver metal. Silver oxide is a dodgy looking material (looks like worthless slag dust)
3) ITO does not work with glaze, it might work if you mess around with frits



I am curious about powder made from various brazing rods used for sintering, I have a bunch of different rod types. Not sure how to powder them without getting too much grit. I was hoping maybe it floats on water or something for purifying the resulting powders. It should not be too hard to get 0.1 grams of whatever metal on sand paper.


I wonder how these traces react to current surges

[coppercone2]

I did a substantial amount of research and it seems that for anyone interested in resistors, tough luck. It seems that leaded BS glass was kinda like the standard and basically research efforts into lead free BS glass for binding oxides with good temperature coefficient and ETC is really specific and the information seems somewhat proprietary.

But I did find some interesting information that ceramic composition resistors are actually made of indium or tin oxide, similar to the idea someone here had... but none the less, getting the right frit is probobly a challenge.


Not saying its not possible but it looks like there is substantial compromises that would be made. Then again, the standards for people making components (they need to justify a assembly line with millions of units per year probably) is very high so they might be neurotic about certain details.


However ,researching ceramic comp resistors might be a wise bet, because their actually really god damn expensive, and have good RF performance, even if made in a through hole form. They are 10% resistors at 3-5$ each.... so messing around with this technology, instead of trying to replicate <10 cent thick film resistors... might actually be fruitful... especially since its a power product. That might be the dummy load no one was looking for.

[coppercone2]

I bought a large number of frits (they are actually cheap) with varying compositions for testing to see if performance is better. Some of them are called fluxes and it looks like some mixing is allowed for pottery.

The idea is to solder a wire to the trace and yank on it to see how hard it is to remove from the board, and if I have better performance then the glaze extract I am using.

I would hope the high magnesium one works, because it has the lowest temperature coefficient of expansion. Alumina is very low expansion.

This is in the interest of finding the one that can hold a edge mount coaxial connector the best for feeding in a signal, because then you don't need a large landing pad with lots of silver for a connector. then I will see if it can make a ground plane.

I found that cable stress reliefs can be made by using acrylic + silicone to make a variable pressure clamp that isolates test cables from sensitive joints (i.e. antennas, with threaded lexan and nylon bolts (yes it works))


I am also curious to see how alumina reacts to staked terminals, so I can have proper power connections to active circuits. Having wire to board is probobly obnoxious... I really hope I  can stake them on. Otherwise, I have to make due with a bolted connection, which should be OK with a small bolt (m2).

For vias I am concerned about getting repeatable ones with the recommeded process of stuffing the holes with silver and hoping it makes a via. Hopefully if I find a strong glaze I can drill out over-filled holes to make them dimensionally accurate, or I can experiment with small eyelets that are soldered in. I don't know how well RF rivets work out but at least its an option

For mounting the thing, the least work option would be to use plastic edge mount PCB standoffs, so I don't need to drill them, and so positioning is repeatable for comparative tests, and so it can survive being heated and cooled.

https://www.essentracomponents.com/en-au/p/pcb-support-pillars-locking-corner-edge-studded/mscehcbs-14-01?


I did a pull experiment on the silver oxide trace, its not well adhered compared to the silver metal. Oh well



I mixed up 5 samples of ITO / frit with 5 different frits in a 300mg ITO to 400mg Frit ratio. I am gonna grind it together and try them out. This experiment got gut wrenching when I started to think about using all that silver, I think I can mix up 200mg samples with the silver palladium for the trace tests. Even though ITO does not work its cheap lol 

[coppercone2]

I tried 5 different frits with ITO up to 780C.

The only one that kinda worked is 3134 frit. It made a lump that popped off the substrate kinda like a bad solder joint. This fairly thick lump read in the megaohms on a DMM.

Oddly enough, when mixed with the PVA glue, it did not form a nice dispersion like the other frits did, it was kind like clumping up in the PVA. Maybe that is a sign of compatibility.

The only frit that got close to the working point was 3134. The other ones were missing 100 degrees or so.

I wonder how it will do at a higher temperature. I don't have a good method to proceed, I thought to just make up the sample plate of various things and progressively try to heat it to a greater temperature


Should I be trying to melt them? Or right under? I think I will need to melt it to get it to adhere to the substrate.


I think that lump was promising, it looked like something that might belong in a electrical circuit.


I have an idea for a rapid fire furnace that I can use to heat a single resistors worth of material with less power using a quartz tube and some wire. The idea is just to push a sample into the tube and then push it out the other end when its cooked. I am not sure but I think i can run it off one of my lab supplies, then maybe I can get them done within 30 min.

[coppercone2]

does anyone know WTF is going on with the frit
https://digitalfire.com/material/ferro+frit+3134

3134

It seriously makes like a god damn playdough as soon as I add PVA. All the other frits make a kinda slurry. Reminds me of silly putty. is it the lack of alumina? it makes a god damn booger

[T3sl4co1l]

Hahaha, yeah too much soluble borate cross-links PVA. https://www.lsu.edu/science/chemistry/files/chemdemo/demos/silly-putty.pdf

Maybe a waxy sort of binder would do.

Tim

[coppercone2]

Well I got 5 samples up to 880C and turned it off to cool.

Some of them are stuck on now pretty good, but they have a very high resistance. I would say I can make a reliable at first glance 10-30 megaohm resistor. I made more like slurries then pastes though, so I expect them to have high impedance

The one that measured better (100k) is the low expansion magnesia, but it scratched off. I am short of the temperature it was in though, that one is 920C

Also, strontium flux was well ahead of its working temperatures, and it came off rather easily.


Calcium and barium seem to be sticking the best, but their well short of their melting temps.

The frit previously looks like it really melted into globs but they measure very high resistance (30+meg)



So it might be possible to make ITO resistors of somewhat decent robustness. But your gonna have a field day figuring it out exactly 


I think I need to run a test up to 1050 and actually try to get a paste of consistency and stencil it rather then dabbing slurries


I think it has a chance of working



and yeah, its hard to believe that powdered glass can actually do something like a chemical, that is cause a fluid to react in some water (not sure if its a chemical reaction, but it does 'react' in the strictest sense, like a physical reaction). Maybe its soluble, but you figure once its in a "glass" form nothing is happening. The only liquid I thought glass might ever do anything in is molten bases. They tell you its supposed to be the most inert thing. Because frit is ground glass. I am not like boiling it in hot stuff.



I wonder what kind of problems they have trying to make nuclear radioactive glass/cement for disposal lol. I bet alloys of metal are hard too.


And try making a decision about it, clearly you can squeegee in the 'blob' into a stencil if you try hard enough. is that appropriate? who knows

[T3sl4co1l]

I mean... "water glass" is produced as just that, a glass.  Solubility isn't unique to crystalline substances, nor is precipitation.  Depends what the substance, and solvent, both are.  Opals are amorphous silica, and chert/flint nearly so (micro (nano?) crystalline), both produced in sedimentary/altered environments (that is, altered by water).  Not that we traditionally call these materials "soluble", but for geological purposes, that's just a matter of time.

Tim

[coppercone2]

I think it means that most of what we consider to be glass behavior comes from silica and everything else is kinda some add on


https://digitalfire.com/material/fusion+frit+f-69 - works
https://digitalfire.com/material/ferro+frit+3134  - silly putty

So its not the boron, because the f69 made a fine slurry with the same liquid. They write something about alumina needed for a suspension on the 3134, its supposed to be mixed with clay . The F69 has 10% more boron and its OK

[T3sl4co1l]

Well yeah, that one's got no alkali (Li, Na, K). Very little (10^-6 maybe?) solubility in water.

With so much alkali, significant borate content will dissolve out from the other, probably silicate too.

Tim