Does more UV light result in less Photoresist exposure time?

[dapug]

I'm setting up a photoresist light box using a LED matrix.  I could lay the LED strips such that the LEDs are each evenly spaced on x,y axis...
.   .   .
.   .   .
.   .   .

but then got to thinking... why not just pack 'em in there?
. . . . .
. . . . .
. . . . .

I figure if I am using a Fresnel lens, I will get even coverage (collimation champion), and can probably even do boards that get pretty close to the edge.

But anyway, will more light reduce the needed exposure time?

[james_s]

Yes, it will. Commercial PCB production often uses high powered arc lamps for very fast exposure. The wavelength matters too, make sure the wavelength from the LEDs you use is close to what the resist is intended for.

[Berni]

Yep double the UV light should be roughly half the time. Tho its also the wavelength of the UV light that matters. Most UV LEDs produce the longer wavelength colors of UV while a lot of photoresist exposes faster with the sort of middle of the range wavelength UV.

Also if you plan to etch fine details then its smart to get a single bright UV source rather than many low power ones in a grid. A single point source of light casts sharper shadows since there is less light coming from an angle to creep around the edges of the mask making the edges of the shadow slightly blurry. I found that out the hard way when i got in to pcb etching and bought a filter-less blacklight tube (The regular filtered ones remove the harmful shorter wavelength UV that we want here). When i started doing 0.2mm track/space on my boards i noticed that it etched a lot more reliably in one direction. This is because the single long tube was more of a difused light source along its length but perpendicular to it it was closer to a point source since the tube is narrow. I reduced the problem by placing a box to the left and to the right of my PCB so that it blocked out the light coming from the ends of the tube and rotated the PCB so that the long side was running perpendicular to the tube, to compensate for the blocked light i had to increase the exposure time a bit too. After that the fine traces etched fine in both X and Y directions.

A particularly good lamp for exposing PCBs is a mercury vapor discharge bulb. The outer white light emitting phosphor coated glass can be broken off to expose the tiny mercury arc bulb inside. Not only is this a tiny point source of light that casts very sharp shadows but it is also very powerful and generates the shorter wavelength UV, so the exposure time can be seconds rather than minutes. But do be careful if you attempt something like this. The phosphor in the outer glass is very poisonus, mercury is aswell, they need high voltage to work, also they get very hot and can set things on fire. On top of it all it needs to be contained inside a box with a kill switch on the door, since looking in to one is very bad for your eyes. You are doing it at your own risk.

[james_s]

The phosphors used in color corrected mercury lamps is not particularly toxic, but there is a bigger issue here. These lamps rely on the sealed outer envelope filled with low pressure argon/nitrogen mix for two critical things. First it provides thermal insulation, a bare mercury arc tube run at the rated power will never fully warm up and reach full brightness. Second, exposure to atmospheric oxygen will quickly corrode the molybdenum ribbon used to pass through the pinch seals into the quartz arc tube.

You can get clear mercury vapor lamps that have no phosphor, though the glass outer envelope does attenuate the UV there is still a fair amount that gets through. A UV fluorescent tube works just fine though and runs a lot cooler than a mercury HID lamp, can be placed in a much more compact unit, comes on almost instantly without having to warm up.

[jpanhalt]

Yes, a point source gives a sharp shadow.  But no mercury lamp is a point source.  That is why in ordinary microscopy a collimated beam is used.  The same holds true when doing high-resolution X-rays for which a collimator is used.  By placing the image side of the transparency against the photoresist, one minimizes the effects of less dark shadows.

Besides the obvious fact that a mercury discharge lamp is not a point source, one needs to consider exposure intensity (see: blueskull @ Post #2).  If one had a point source at 6" from a 6"-wide plate, intensity at the edges is only 80% of what it is at the center. In my experience, I use roughly a 13 minute exposure.  Plus or minus 20% of that would be very easily noticed during development and etching. 

Another variable is the resist.  The OP has not come back to tell us which resist he/she is using.  I suspect he is a hobbyist and is using a positive resist.   For such resists, ordinary fluorescent lamps will work fine.  Even incandescent lamps will work, as will UV LED's.

John

[james_s]

Yeah that's something I forgot to mention earlier, the important part is to place the printed side of the pattern against the PCB and hold it down tight with a piece of glass. If you have the printed pattern tight against the board it doesn't matter if it's a point source because there is no leakage around or under the pattern.

I haven't done UV boards in a while but when I did make some, I used one of those U-shaped unfiltered UV tubes used in bug zappers.

[dapug]

Thanks everyone. Great tips!

I thought the answer (to the title question) was obvious, but still worth asking because I am not familiar with photoresist science.  Baking cookies, for example, does NOT happen faster by turning up the temperature/exposure - not with desired results anyway.  . But yes, I can see how 2x UV in this case would result in 2x speed.

I'm trying to strike a balance between quality and convenience.  UV LEDs in a nifty little box is more convenient than high powered lights and dedicating a larger space to the task. And my boards are for prototyping, not production (though I still want to mimic prod quality as much as possible).

I totally hear you about shadows and distorted light affecting the resist.  So here again, I am hoping to mitigate the issue by using a Fresnel lens.  For those not familiar with fresnel, the light can come in from any direction, but once it passes through the lens it is directed perfectly straight toward the object.  You can learn more via google or wikipedia.  In theory, the scattered light from an LED array will be directed straight down on the PCB.  That said... I could be completely wrong about how will this will help in the task of photoresist exposure.  It will be an experiment.

[james_s]

Well feel free to try it, or look around to see what others have done. I found that the small U shaped fluorescent tube in a simple reflector made out of aluminum sheet in a plywood housing gave perfectly good results in a compact box without any fancy optics. I think I got the tubes dirt cheap from Electronic Goldmine or something.

[richardlawson1489]

Consider this when calculating exposure time. Is your light source spectrum broadband (g-, h-, and i-line) or monochromatic? http://www.microchemicals.com/technical_information/exposure_photoresist.pdf