My sputtering system

[Doug Coulter]

Joe, that's why we are here -- we are "writing the book" from a perspective of people who actually do this stuff, to share with others. In real life you just run into so many things that the usual oversimplified theorists miss, and we actually overcome all that and get 'er done. Usually all too much of this kind of info becomes someone's trade secret and is lost to the world.
We're fixing that problem.

Yeah, I'd have called that photo "a bad day at the office". But impressive none the less. John is the current expert here at any kind of depo or implantation, I'm just a beginner (who gets lucky some).

CD's are evaped, it's so much faster -- ridiculously so, and the process is far less critical. Fraction of a second a batch, vs an hour or so. Sputtering is like for things that Jerry is wanting to do. Ultra precise thickness optical coatings and things similar to that. Time is money, which is why a good optical AR coat is much more expensive than a CD.

Pick our brains all you want, that's the whole idea here. Gives us an excuse to create content that teaches the lurkers, and someday I'm sure you'll return the favor.
Because it's obvious you are a "maker", rather than a taker or faker or just talker. Our kinda guy.

I'll see if I can find some pix of my stuff when I was doing that, my evap thing is temporarily down -- that tank grew a fusion project in it. I got really decent results with some pretty simple setup.

[johnf]

Joe/ Doug

forgot to say that those films i'm referring to are accelerated by 15kV

I'm sure that even normal sputtering will include the Argon. Argon is difficult to ionise ie has a high work function (thats why it is used as the AR ions will impart their enegy to sputter the target material)
So in a short statement Joe your films will be a mix of Argon and Al. How much only RBS or similar technique can tell

the film pictured was 30% Ar just after implantation and 8% a week later.

Sputtered films tend to have more stresses than evap films due to the molecular bonding (my opinion)

Been doing a run of Ca,Ag,Ca,Ag on Si around 100nm for each layer to see what happens

came out looking alright -i'll see in a week if the film still is a good looker and if beam time is available an RBS spectrum

[Joe Jarski]

I was playing around with my TIG welder tonight in stick mode trying to see if it would work as an evap power supply. I was using it on a dummy filament made from 3 pieces of .032 SS safety wire twisted and wound like the ones on Lesker's site. It seems to work quite nicely. The current regulation is pretty decent, though not accurate (maybe because of power factor?). If I set it on 10A with a 12-15" section of "filament" there was an inrush spike, but then it would settle down with the meter reading about 20A and the filament would heat up to orange in about 5s and hold there with the voltage adjusting accordingly throughout the temperature rise. The only instability was when the current was set too high and the welder was trying to get over 50V to hit regulation, then the lights in my shop started to flicker, so I promptly concluded that test. So, I think I have a decent power supply for evaporation.

[johnf]

Joe
That pic is for Ar accelerated by 15kV after exiting the ion source so the argon aggregation is worse because of the higher ion energy.

I'm almost 100% sure that CD's are evaped for the Al film

[Joe Jarski]

John,
You are working with significantly higher voltages than I am at this point. I'd never be able to hit even 2kV until my vacuum system is upgraded. On the other hand, it looks like it may cause more problems anyway. My deposition rates at 500-600V seem to be good, at least to me anyway. After a 2 minute run I have a layer without any noticeable transparency. It may be less than that, but I've just been running it there.

I'm going to keep persueing sputtering along with setting up for evaporation, but now I know a few things to watch out for. It'll be interesting to see how much the argon ag is affected by the voltages.

[Doug Coulter]

If you can't get to 2kv, you probably also have gas purity issues, so that's a good thing to fix up no matter what you do. You don't want the stuff to be oxidizing on the way to the substrate!
(unless you are trying to do reactive sputtering, but purity is just as much an issue with that)
My rule here is to try and get to factor 100 or so better vacuum than the process requires for that reason. You'll be able to do both sputter and evap in the same rig with a little planning.
Sounds like your TIG welder is better at that than mine, mine goes kind of crazy trying that with it and isn't really meant to run that low a voltage.

What I did for evap was prep a chemistry ring-stand to go in the tank (Cu plated it so it wouldn't outgas so much). I added a shelf low on it to hold a cut off coffee can to catch dropped metal, and to keep from evaporating on to the whole tank -- I just cut holes in the side for the high current feedthrough rods to enter, and the cut off top of the can matches the underside of the ring-stand. I put my substrate on various jigs that sit on top of that, again, to keep from depositing the stuff all over places I don't want it to go. One time of ruining every feedthrough in the tank was enough to learn that one.

Here's a pic without the shields.

Simple evap

This is using a 3 wire tungsten heater from Lesker, one of the few things they sell that is cheap.

I then made some improvments and used a different type of heater that I have since added a shutter to, which helps quality -- you can preheat and outgas things before letting anything hit the target, then flip the shutter open and shut quicker than turning the heater on and off -- so as to get more control over coating thickness.

Here's a pic of the thing while I was building it.

Fancy source

I added a simple shutter, a small piece of sheet metal on a baling wire axle, and a couple small solenoids to flip it open and closed. This will let you deposit up to 3 different things in one run.
With the shutter, you find that anything reactive takes the vacuum down a lot better on heating it, then you flip the shutter open for only about 1 second to get your deposition done.
Of course, for best results, you need to get to e-6 millibar vacuum or better on most things, or you're going to get oxides in the mix.

There I am lucky -- after years of frustration, I finally just bought a couple turbo systems new, and nearly all my vacuum issues are now elsewhere. I've built one successful fusor with a diff pump, but it's so much more trouble to run that I won't bother doing that again -- that system will probably wind up doing depositions. Of course, I caught all kinds of crap from the scrounger crowd for just putting down the long bucks, instead of spending more years scrounging, but hey -- I ain't gonna live forever, and they don't count the costs of scrounging, time, gas money, buying defective stuff, you eat and pay rent while waiting, etc.

So the plan here is to scrounge when possible -- keep an eye out for any good deals and jump on them, but if something critical-path can't be scrounged, just buy the doggone thing new.

As always, click the pix for a larger version. This board is nice that way -- it does all that for you. I still use about 1024 pixel on a side, and use Gimp to scale and compress them better, it's just a good habit to get into.

[Jerry]

John,
You are working with significantly higher voltages than I am at this point. I'd never be able to hit even 2kV until my vacuum system is upgraded. On the other hand, it looks like it may cause more problems anyway. My deposition rates at 500-600V seem to be good, at least to me anyway. After a 2 minute run I have a layer without any noticeable transparency. It may be less than that, but I've just been running it there.

I'm going to keep persueing sputtering along with setting up for evaporation, but now I know a few things to watch out for. It'll be interesting to see how much the argon ag is affected by the voltages.

You souldnt need to get to 2kv. Even the nice Applied Energy MDX magnetron supttering supplies top out at 1200v and most DC sputtering papers I have read topped out at 800v or so.

[Joe Jarski]

Doug,
That's a nice looking stand. I like the multiple source idea too. Although I'll have to come up with a better power supply if I want to use the boat type. I don't think my welder would be too happy at ~1V & 300A. My welder works great TIG welding at the low end, it'll keep the arc lit down to 3A or so, not that you can weld there but it makes a nice light to find your way around at times before things start melting.

Sputtering voltage - that's what I kind of though when I built the power supply. My upper end is about about 750V. With the current increase at the pressure (about 10 mTorr) that I've been working at I'd need several kW at 2kV. That was all with a rotary pump doing a pump-purge-pump cycle to reduce contamination. I recently acquired a turbopump for the new system which will hopefully do wonders, although the drive is a bit flaky until it warms up - a subject for another post once I start digging into it.

Since aluminum wets tungsten, does it wick up into the stranded filaments when it melts? I thought maybe I could wrap the filament with 1100 Al MIG wire initially unless it'll just drip off. Does that sort of make sense or is it a silly idea?

[Doug Coulter]

Yes, the stand (with coffee can parts) and the boats work real well -- I have to unscrew the legs on the stand to get it through the door, but that's not too hard. I extended the FT's for the coil heater with similar copper rod, joined by 10-32 tapping and a threaded rod. The power supply for either is really simple -- chisel the HV winding off a microwave oven transformer and replace with a few turns of fat wire -- test with a few to see volts/turn. I'll have to get a pic of mine and put it up, but it was really more work for the big SSR/heatsink and snubbers and box holding the button and and the battery (to run the SSR) than anything else. I brazed brass bolts onto the wire for taps, so I could get the volts right for whatever and not need a huge variac for that. That boat isn't that bad, it's real thin stuff -- as you can see in the pic, one is broken because one of my kitties sneezed on it. If you've not yet, sign up for the Lesker catalog (link in vendors section), and look for the lower amp heaters, it's a big section there. As well, there are tips in the back for depositing almost anything you'd think of, what works best for each and so on, it's a very valuable resource, and the price is right.

Yes, the Al gets runny and gets over the whole heater, it's cool. You do a pre-run heat to make that happen before doing deposition, so it's uniform first (and outgassed well) I got a pound from Lesker in the form of little candy canes you crimp about 3 onto the heater turns. That's enough for a lot of runs here. So a pound is like a lifetime supply, and worth getting the really pure stuff. I tried alloys like MIG wire, not so good at all, and the Lesker catalog tech notes on "vapor pressure vs melting point" show why. There's almost no relation so things may not evaporate in the same ratios that existed in the source material, or even close. Yes, some does drop off, and the coffee can I put under that catches it. Not much if you raise the temperature slowly enough, but some is lost.

My source on sputtering volts etc is very old, pre magnetron, and that's how I did it -- no magnets, as at that point water cooling them was a little too daunting (and they only take a few seconds to get past curie temperature without it).

Jerry is very probably right about the volts. Since I was doing it the "old way" I used the volts for the old way, and it worked. John is mostly doing ion implantation, MBE kinds of things, so he's running in another world yet and very different conditions. And then for sputtering insulators, you need RF or pulsing stuff, and there's a whole 'nother world again.

I will still re-iterate that you have to want a much better vacuum system than what it takes to just get to process pressure. Even if you get 100x better (implying 99% purity) you don't get that good in a semi-sealed off system as it's still outgassing other junk unless the pump is going kind of full blast and you just overwhelm that with fresh gas input. For cheapo inert gases, maybe no sweat, but if you want to make TiN with ammonia -- well then -- nothing's going to like having that in it or in the air near you...or anything with one of the nastier semiconductor fab gases, or just do diamond film deposition with hydrocarbons -- you don't want tons flowing through the system for various reasons. So in general for best results, it's good to get it where you can get a lot better vacuum than you'd think you need, and have it not rise back up fast when you turn the pump off -- that's a key indicator that something is giving off gas in the system (you have to have a valve in the line to the pump to try that well -- turning off some pumps makes them leak or suck-back oil).

[Joe Jarski]

I've been plugging away at my new chamber, among other things, for the past week. I've got heaters, Al and a Lesker catalog on order. I did an initial pumpdown on the new chamber a few days ago just to check for any obvious leaks in the welds. I used tacky tape for seals because I don't have all of the O-rings or viewport window yet. Anyway, the first test seemed fairly successful considering that it hasn't been thoroughly cleaned and prepped - just wiped down with acetone. It took about 10 minutes to reach 15 mTorr with my rotary pump, which I think is decent for an 18" chamber and mediocre seals.