Weird PCB etching behavior

[ali6x944]

Heydo folks,
been trying to make PCBs for a while, and have successfully got very nice adherence after many failed trials using photoresist film when it is dry, the key was applying it to a wet clad. also made my own exposure box using a UVC tube...
lets just say photokeratitis isn't fun and eye protection is a must 
the exposures that i got seem rather good and relatively clean but for some reason the "sharpness" of the exposure is not consistent as shown in the images...
not sure why is it so, probably because the diagonal placement of the tube but i don't have concrete proof yet...
another weird thing was the exposure time, it seemed rather long about 15 mins compared to the literature online...
the development went very smoothly, but the etching was really problematic as follows:
a) the etching time took too long (about 26 hrs) for FeCl₃ with agitation and rubbing and heating to slightly above room temp, used a 3d printer heated bed and rubbed the board on occassion
b) the pattern seems to be under etched in some spots and over etched in others
c) there seems to be etchant bleeding under the mask
d) holes lift of completely

idk what i did wrong, i feel kinda confused not sure what should i go from here...
i think the problem is with the solution so i am going to test different formulations presented by nurdrage in his video

[ali6x944]

here are the exposure box and boards:

[ali6x944]

the boards

[tooki]

1. The waviness in the lines at the bottom of the exposed but unetched PCB make me think your stencil was not properly pressed onto the PCB. This allowed a variable air gap and ensuing blurring. Absolutely no gap is acceptable.

2. Why on earth did you use a UV-C lamp?!? They have a wavelength of 254nm, far below the 350-400nm that photoresist is designed for. This could very well explain the very long exposure time needed. You want a blacklight UV tube, not germicidal. And as you discovered the hard way, UV-C lamps are hazardous.

My guess is that the blurry exposure meant that the “unexposed” areas were actually very slightly exposed from the side, resulting in an invisible film of resist being left over many areas that should have been blank. Then you “compensated” for this by etching for an extremely long time, during which it etched away the free areas and then ate its way underneath the resist (including the invisibly thin areas that you were monitoring).

It shouldn’t be hard to find proper instructions for how to etch, including proper specs for the resist film you used.

[coppercone2]

26 hours is way too long how about like 26 minutes.

[janoc]

26 hours is way too long how about like 26 minutes.

Yeah that etchant is either really spent already or improperly prepared (too diluted). Or maybe it is not even Ferric Chloride at all ...


People also commented on the UV-C use which is completely wrong for this task. UV-A/B tubes or even those UV LEDs are a much better option. Seems that the OP didn't do their homework before attempting this.

[coppercone2]

Well clearly if you are using a 3d printer bed to heat a liquid there is a cost component involved in the decision making process 


I think people start exaggerating how wet chemistry is easy and then take liberties on top of that and you get fail... pop-sci seems to really focus on making bogus chemical processing stories. Its not that intuitive or easy, it took Bohr to start figuring out chemistry in the year 1600.... but people act like its smithing by neolithics in 20000 B.C.


And I still have yet to see someone that trusts a DIY plating job on any electromechanical component of significant force.

Folks, its technically complicated.



I think honestly they bullshitted so much about DIY chemistry that people don't feel a sense of pride in accomplishing something like a good etch or plating, because online you get like "lol thats easy" so no one wants to actually work on it. And since there is no pride in the work, all they do is compare costs and the process does not improve.


In my opinion, its way harder then soldering, which is seen as a high end skill.




Questions for the OP
1) do you have a immersion thermometer
2) how did you verify you etching solution
3) Are you using a way too small container because of hunches and figuring that you don't need too much etchant? I like when people start making chemical decisions based on hearsay about how much soap you need to wash the dishes. How many drops are in a tide pod?

[free_electron]

26 HOURS ?. with proper agitated (foam) and heated FeCl it shoud be like 10 minutes or even faster.
You got the mixture wrong.

[jpanhalt]

What etchant are you using?  Cu+2 (e.g., CuCl2) doesn't give the resolution that FeCl3 gives.  The chloride is important.  CuCl2 is faster, but the endpoint is harder to define.  My etch time at 60°C was on the order of just 10 minutes.

[tooki]

What etchant are you using?  Cu+2 (e.g., CuCl2) doesn't give the resolution that FeCl3 gives.  The chloride is important.

He states right in the original post what etchant he’s using.

[jpanhalt]

Sorry, I missed that.
John

[janoc]

Well clearly if you are using a 3d printer bed to heat a liquid there is a cost component involved in the decision making process 


I think people start exaggerating how wet chemistry is easy and then take liberties on top of that and you get fail... pop-sci seems to really focus on making bogus chemical processing stories. Its not that intuitive or easy, it took Bohr to start figuring out chemistry in the year 1600.... but people act like its smithing by neolithics in 20000 B.C.

...

In my opinion, its way harder then soldering, which is seen as a high end skill.

Come on, dude. Don't be ridiculous. BTW, plating & etching are two very different things, so that plating is possibly difficult has no bearing whatsoever on etching and vice versa.

We aren't talking Nobel level chemistry here. Moreover, poor Niels Bohr - he lived in 20th century, not 17th ... Basic chemistry processes, such as metal etching, were known for millennia, that we didn't understand why they work is not really relevant to the discussion in any way.

Moreover, if you buy a ready-made etchant, you don't have to do anything, just pour it into a tray. No need to even heat, just stir the solution every once in a while. Fresh ferric chloride will etch the board in 30 minutes even cold, no issues whatsoever.

If you buy dry ferric chloride crystals, all you need to do is to dissolve the crystals in water (doesn't even need to be distilled/deionized - if you are etching in a tray and not in a machine that could gum up it doesn't matter) to create a saturated solution i.e. until the crystals are not dissolving anymore. How are you making your tea or coffee if this is too complicated for you? Or, God forbid - cook a dinner?

That the OP used a 3D printer bed as a heater boggles my mind because all it would take is a drop of the etchant and the bed is ruined (never mind a brain fart causing the bed to move and splash the warm etchant all over the place). That's nothing to do with cost, more like outright stupidity (together with using that germicidal UV-C lamp). A water bath can be improvised in every kitchen and certainly costs less than a 3D printer.

Worse - you don't need any of it! Ferric chloride will etch cold just fine (unlike ammonium persulfate, that one must be heated), it only takes longer. I have never bothered to heat my ferric chloride, the risk of making a mess if I happen to spill it is not worth saving a few minutes for me.

Here the problem is not the supposed complexity of the process but that the OP has obviously "winged" it somehow, not bothering to read instructions or thinking they can just skip & adapt the steps of the process without understanding them. I could understand this if it was 30-40 years ago when we didn't have any resources or guides to learn from, only books and some magazines - or, if lucky, some friend who could mentor us. So trial and error was often the only way to figure things out. But today? With tons of good guides online, both in text and video format? 

[jmelson]

I use Riston dry film resist, applied with a heated roll laminator.  The board needs to be totally cleaned of any preservative, I rub with 1000 grit sandpaper until it appears evenly "frosted".  Then, rinse with acetone-soaked paper towel.  I expose with filtered black light fluorescent tubes.  The board and master artwork are held in a vacuum frame made of Plexiglas (acrylic sheet).  Exposure is one minute, any more causes light to leak through the black areas of the film and start to harden the resist that you DON'T want to harden.  I develop in a sodium carbonate solution and wipe the board with my fingers as the unexposed resist dissolves.  The developer solution is heated to 40 - 50 C.  Development of the resist takes about 2-3 minutes.  Then, wash the board in running water until it no longer feels slick.  Let dry for several hours.  Etch in heated FeCl3 and it should become completely etched in about 2-3 minutes with a spray etcher.  Sitting too long in the etchant bath will allow the etchant to undercut the resist.  If the part you want to etch away doesn't immediately turn salmon pink, then there is some contamination or preservative left on the board, or the resist did not completely wash away.
Jon

[ali6x944]

1. The waviness in the lines at the bottom of the exposed but unetched PCB make me think your stencil was not properly pressed onto the PCB. This allowed a variable air gap and ensuing blurring. Absolutely no gap is acceptable.

used a glass panel during the exposure that i compressed using some weights, so i am somewhat sure that it was compressed properly but that might be a possibility.

2. Why on earth did you use a UV-C lamp?!? They have a wavelength of 254nm, far below the 350-400nm that photoresist is designed for. This could very well explain the very long exposure time needed. You want a blacklight UV tube, not germicidal. And as you discovered the hard way, UV-C lamps are hazardous.

my bad i was following this https://sparks.gogo.co.nz/dry-film-tips.pdf , it is well explained and seemed very straightforward, however they didn't sell uv leds or uv led strips locally so i decided to improvise since the uvc tubes are more available, i tried searching locally for uvb and uva tubes first but they were either too long or too expensive for a single tube, so settled for uvc not thinking much of it, come to think about it that was really dumb and i should have been more diligent.

My guess is that the blurry exposure meant that the “unexposed” areas were actually very slightly exposed from the side, resulting in an invisible film of resist being left over many areas that should have been blank. Then you “compensated” for this by etching for an extremely long time, during which it etched away the free areas and then ate its way underneath the resist (including the invisibly thin areas that you were monitoring).
It shouldn’t be hard to find proper instructions for how to etch, including proper specs for the resist film you used.

actually i kinda see what you mean, i made sure i removed it during development, but that might be very much the case.

Well clearly if you are using a 3d printer bed to heat a liquid there is a cost component involved in the decision making process 

kinda, i really didn't use it in the beginning and used it after the 3 hr mark because it pretty much was not etching, from what i read it shouldn't take more that 45 mins to 1hr to etch a pcb of this size, i hate ferric chloride so much now like idk it should just work right? like i tried rubbing it with a soft sponge and bushing it with a paint brush for 3hrs with absolutely no effect but slight pitting, so i went online and saw people using their 3d printers as agitators so i did that and i checked on it periodically but no progress so i left it to do its thing for the entirety of Friday and checked on it Friday night and Saturday morning  and it looked almost done so i removed it and cleaned the resist with NaOH solution and noticed this weird etching pattern.

Questions for the OP
1) do you have a immersion thermometer
2) how did you verify you etching solution
3) Are you using a way too small container because of hunches and figuring that you don't need too much etchant? I like when people start making chemical decisions based on hearsay about how much soap you need to wash the dishes. How many drops are in a tide pod?

1) no but i can buy one since they are not very expensive here
2) i simply bought it from the store, i honestly don't know how i would verify it (maybe some sort of replacement reaction? or pour it directly into naked copper? )
3) i used a container that was slightly larger than the board, and filled it until the board was submersed completely by about a 1 cm or so of solution.

 

[Psi]

Some etchants will only work fast enough to be useful when very hot, while others will work fine at room temp.
I always used to warm up FeCl in an old microwave prior to using it.
But then I switched to peroxide and HCL and didn't have to any more.
I tried Ammonium Persulphate once but it had to be way to hot to be useful, it would cool down too quickly without active heating.

Whatever etchant you use it should be done in under 30min

[ali6x944]

That the OP used a 3D printer bed as a heater boggles my mind because all it would take is a drop of the etchant and the bed is ruined (never mind a brain fart causing the bed to move and splash the warm etchant all over the place). That's nothing to do with cost, more like outright stupidity (together with using that germicidal UV-C lamp). A water bath can be improvised in every kitchen and certainly costs less than a 3D printer.

Worse - you don't need any of it! Ferric chloride will etch cold just fine (unlike ammonium persulfate, that one must be heated), it only takes longer. I have never bothered to heat my ferric chloride, the risk of making a mess if I happen to spill it is not worth saving a few minutes for me.

Here the problem is not the supposed complexity of the process but that the OP has obviously "winged" it somehow, not bothering to read instructions or thinking they can just skip & adapt the steps of the process without understanding them. I could understand this if it was 30-40 years ago when we didn't have any resources or guides to learn from, only books and some magazines - or, if lucky, some friend who could mentor us. So trial and error was often the only way to figure things out. But today? With tons of good guides online, both in text and video format? 

honestly most of the process was very straightforward, no major issues, during film application i had slight bubbles because i did it dry as the laminator section didn't mention i should wet the clad but as soon as i figured that the film should be applied wet it worked without any issues as u could see from the developed exposure is, the development was also pretty straightforward, i pretty much followed the exact recipe and made stock solution 30g/L but the brand I used is Oxi Clean but its pretty much the same sodium carbonate and sodium percarbonate, same for the hardening step, but the etching is where something went wrong

[jpanhalt]

I have reviewed the thread in new light.  It appears you think everything is OK except the etching.

Like you , I used a piece of ordinary window glass to hold the mask against the photosensitized plate.  I used single pane window glass on both sides of the board and spring clips as one can get at stationary stores to hold them together.  You don't need really strong onee.  I used the weakest ones I could find.

For exposure, as others have said, do not use a germicidal lamp.  The resist has a photosensitizer in it and only requires UVA.  Phosphor coated fluorescent lamps work fine.  Mine puts out a lot of visible light too.  Note that window glass has very poor transmission of the 254nm  light emitted by germicidal lamps, so you really were not getting much exposure, which is why it took so long.

My developer and photoresist were different than the one you used and are not relevant.

A problem that I suspect is contributing to your very long etching time is incomplete removal of the photoresist.  When you are ready to etch, you can interrupt that process and restart it without problem.  I would dip my board in the ferric chloride etchant for a few seconds, remove and wash it.  The resist is hydrophobic, but the areas to be etched should wet well with water.  If not, then you need to develop more.  Another clue is the color of the copper.  My copper turned from shiny copper to more of a dull salmon color during that brief etch.  I used pre-coated boards that had not been abraded before coating.  Since your boards are abraded, the gloss may not be a clue, but wetting by water and color change should be clearly evident.

The attached photo shows the level of detail I usually got.

[tooki]

1. The waviness in the lines at the bottom of the exposed but unetched PCB make me think your stencil was not properly pressed onto the PCB. This allowed a variable air gap and ensuing blurring. Absolutely no gap is acceptable.

used a glass panel during the exposure that i compressed using some weights, so i am somewhat sure that it was compressed properly but that might be a possibility.

The waviness and selective blurriness demonstrate that it was not pressed down evenly — there clearly was a gap. The stencil could be uneven, or even the bare PCB itself could be uneven. The result is that they need to be compressed enough that whatever the source of unevenness, it gets pressed away. Look at your glass–stencil–PCB–support sandwich from the side, ideally with a light source behind it. It’s probably not as solid as you thought. What material was the support beneath the PCB?

Also, it dawned on me later that normal glass is a quite effective blocker of UV-C light (which is why UV-C tubes require a fused quartz tube, rather than glass). This would also have tremendously increased exposure time, just as jpanhalt says.

…from what i read it shouldn't take more that 45 mins to 1hr to etch a pcb of this size, i hate ferric chloride so much now like idk it should just work right?

The etch time should not even really depend on the board size, as long as there is sufficient etchant.

Ferric chloride is arguably the best etchant out there — fast, produces sharp, clean edges, and doesn’t go bad in storage — so as long as your bottle is real ferric chloride of sufficient concentration, it should etch quickly. Your bottle at least looks like the correct color, but you can easily test it by etching a piece of thoroughly cleaned scrap copper (or copper-clad PCB to which you have not applied any resist). Within seconds the shiny surface should go uniformly matte/satiny and look more pinkish than shiny copper.

[free_electron]

Itching solution.... if it itches you must scratch it.

[coppercone2]

If you are wondering, corrosion rate is a engineering unit. For stainless steel its specified in units like milimeters/year. For etchants its more like /hour. Its something you can measure and qualify.


It gets interesting when you study electrochemical machining. Truely interesting, because  chemical etching basically does not see hardness of materials... meaning it can get you some very 'impossible' geometry if you figure it out

[twospoons]

UVC wont go through plain window glass - the cutoff is around 320nm. That's why UVC lamps are made from quartz.
Lucky for you mercury also has emission lines at 365nm and 406nm.

[ali6x944]

…from what i read it shouldn't take more that 45 mins to 1hr to etch a pcb of this size, i hate ferric chloride so much now like idk it should just work right?

The etch time should not even really depend on the board size, as long as there is sufficient etchant.

Ferric chloride is arguably the best etchant out there — fast, produces sharp, clean edges, and doesn’t go bad in storage — so as long as your bottle is real ferric chloride of sufficient concentration, it should etch quickly. Your bottle at least looks like the correct color, but you can easily test it by etching a piece of thoroughly cleaned scrap copper (or copper-clad PCB to which you have not applied any resist). Within seconds the shiny surface should go uniformly matte/satiny and look more pinkish than shiny copper.

i used the rest of the failed test pcb and put that in the same plastic tub fully submersed in etchant with about 2cm of solution on top.
the solution was agitated by hand for 45mins 33secs to fully etch the board.

[tooki]

So your ferric chloride is fine. This proves the problem is with the photoresist. You have to get that right: proper exposure with the proper light source, with the stencil properly pressed against the sensitized PCB.

[ali6x944]

So your ferric chloride is fine. This proves the problem is with the photoresist. You have to get that right: proper exposure with the proper light source, with the stencil properly pressed against the sensitized PCB.

i am looking for these tubes, is there is a chance that they are the same tubes used for reptiles?

[tooki]

So your ferric chloride is fine. This proves the problem is with the photoresist. You have to get that right: proper exposure with the proper light source, with the stencil properly pressed against the sensitized PCB.

i am looking for these tubes, is there is a chance that they are the same tubes used for reptiles?

Not optimal, but would likely work decently. Reptile UV lamps are mostly UV-A (what you want), with 10-25% UV-B (which is largely wasted here).

All you need is simple UV-A, like used for disco blacklights, counterfeit money detection, bug zappers, and curing UV gel fingernail polish. For PCB exposure, it doesn't matter whether it's the kind that filters visible light (like disco blacklights and counterfeit money detection) or lets it out (like bug zappers).