depositing thin films

i/Thin_film

I'm no expert, but it seems that the processes used would not be kind to the FR-4.

Hi,

I am no expert either, but I read that they can deposit the thin films at 70 degrees Celcius, ie onto plastic, so FR-4 should be possible I think. There are multiple different ways to deposit thin films, just looking for one that is compatible with a hobbyist budget.

cheers, Jamie

of

Well, that would be fun.

of

Silane gas...at your house? Seriously? Years ago the Exar fab in Sunnyvale had a silane leak. The fire department figured the best thing to do was to just let it burn.

Wafer fabs are kind of like nuclear reactors. Generally safe, but when something goes wrong, things go very wrong. But the semi biz has a decent safety record, well except for leaking all that TCE.

To deposit amorphous silicon on something gucky like FR4, you'd have to use low temperature evaporation or sputtering. PVD won't work because of the high temperatures required, just for a start.

Cheers

Phil Hobbs

Hi,

Thanks for the info. The potential application I was thinking of was a PCB based bolometer array. I saw your low cost bolometer previously, and thought up this one for fun :)

"

It is a simple XY grid of wires on a 2 layer PCB, with vias at the grid cross points, and the whole top surface is covered with an amorphous silicon thin film. The grid columns and rows (ie 16x16) are fed into a matrix decoder circuit and the resistance between adjacent vias is measured (corresponding to the amorphous silicon resistance which is proportional to the temperature of the silicon). Thermal images could be used to calibrate the bolometer, to add/subtract thermal impedance / heat capacitance etc for the thin film.

cheers, Jamie

hin

...

d of

I'm only familiar with metal sputtering, but I see mems have been done with sputtered polysilicon. Is 350 to 400 deg C compatible with FR4? Wiki says 140 deg C.

thin

=A0Apparently

Celsius.

method of

kind

just

Sputtering or evaporative PVD does not indicate that Silane would be involved. Both cases involve serious vacuum equipment though. Not clear if they can be scaled to garage or bench top size. Brings up an interesting collateral idea, can Kovar be sputtered? A micron to a mil of Kovar under the silicon could be very beneficial TCE wise.

"

That's a passive matrix design, i.e. no thin-film transistors. It's perfectly possible to sputter amorphous silicon at room temperature. Doing poly requires an annealing temperature high enough for recrystallization to proceed fast enough to be useful, usually around

There are a few things you might not have thought about:

1. Silicon-metal contacts form Schottky barriers unless you do something special. That means you're making back-to-back series pairs of diodes, which won't conduct well.

1. Surface cleanliness. It will be very difficult to make the silicon/copper interface clean enough to work.

2. Continuity. Making a thin continuous film of brittle material on top of FR4 will be hard. It generally requires a film thickness greater than the maximum surface relief in order to get good coverage. Even then, it is very likely to crack. (My infrared sensors started out using 400-angstrom Ni films on 9-um thick PVDF. I spent a _lot_ of time fixing broken connections with silver paint and a microscope, before switching to carbon ink.

1. Thermoelectric potentials. The thermoelectric potential between silicon and any metal is hundreds of microvolts per kelvin.

Cheers

Phil Hobbs

"

Hi,

thanks for all the information, for #1. I am not sure, but possibly gold plated PCB's would solve this. I don't know enough about this stuff but from this page it looks like that is what is done:

For #3, I can imagine that being very difficult, I am not sure how the flexible solar panels work but I have seen ones that can bend quite a bit without cracking. I've seen "roll up" thin film solar panels too, typically the thin film is put onto a thin stainless steel sheet I think, maybe that bond strength keeps it from cracking.

#4: Well maybe if gold plating is used on the contacts this could be overcome.. anyway that is what they appear to use on the microbolometers above.

I find this aspect of material science/electronics very interesting!

cheers, Jamie

For #4, I just realized that the thermoelectric potentials would cancel out between two vias, like two batteries put in series with one reversed. It wouldn't be exactly cancelled out due to differences in the two vias/silicon junctions though.

cheers, Jamie

us.

org...

ind

lms

en

to

of

nd

..."

be

If you seriously want to make a bolometer, you must consider the heat conduction out of the 'device' absorbing the heat. This includes the leads to measure the temperature.

There are all sorts of time constant trade offs involved.

The bolometers I've used were little cubes of semi-conductor (Ge Si or some compound sutff) held by thin gold wires in an intergrating sphere.

George H.

us.

org...

ind

lms

en

to

of

nd

..."

be

To avoid building Schottky diode junctions, the contact point is heavily doped.

silicon

Apparently

Celsius.

org...

be kind

films

possible I

just

when

have to

because of

was a

previously, and

"

grid

amorphous

into a

is

could be

around

something

diodes,

top

than

it

fixing

switching

gold

but

too,

microbolometers

Sounds pretty good for some purposes, not my choice for forming an image though. Of course optics at 7 to 14 micron is a bit strange as well.

?-)