I make my own pcb boards, and chemical of choice for etching is a very concentrated solution of ferric chloride, I use a pyrex glass dish and a tiny aquarium heater to heat it, I also use a air pump and air stone, to lift the ferric chloride via a pipe, and wash it over the board to be etched.
I have built a small extractor cupboard for fume extraction, It all works great! I am happy, but I got to thinking, how can you measure when the solution needs changing?? and what can you do to extend the life of it.
So far I have come up with the following, all of t is purely at the design stage, so comments welcome.

I figured best way was to measure the conductivity, my thinking goes like this, ferric chloride will start to conduct more as more copper Ions dissolve in the solution, at some point there will be a point that the reaction is too slow to be useful, so measure how much copper is dissolved by how much the solution conducts. First problem I can think of is what electrodes to use! I dont fancy gold, so decided on graphite from a pencil (very thin one), these shouldn't react at all, have a set distance between 2 electrodes, push a current through and measure, my main concern is gasses like chlorine, but if I use very low voltage and current (500mV 10mA), then that shouldnt be a problem, and everything is done in fume cupboard anyway. Pulse the current say every 10 mins for a second or so and measure the strength, then output to a DAC then to a analogue meter.
As too how to rejuvenate the solution I am working on that with my chemistry teacher .
So apart from it being a bit bonkers and pointless what do you think?

 

The product specifies how many sq. inches of material - copper - it will dissolve before it is saturated.
I googled, and got this link, of which there are more to choose from, if you google " "ferric chloride capacity"

http://books.google.com/books?id=m8sJBIMtETgC&pg=PA366&lpg=PA366&dq=ferric+chloride+capacity&source=bl&ots=2Z7oHNQaVg&sig=aFrQLK7cdyCcSw6DCK7NQ-N_wps&hl=en&sa=X&ei=OtvSUuDcCZOssQTCuIGoDA&ved=0CEMQ6AEwBA#v=onepage&q=ferric%20chloride%20capacity&f=false

Hope this helps

 

Interesting information, Looks like I am using it too strong to start with (42%), I have 20litres to use up, after that I might swap o something else, Regeneration while possible looks a utter waste of effort.

 

That's an excellent idea, if you can reliably test the ferric chloride by conductivity that would be of great help.

Have you done tests re dilution? Does the reading change if the FC is of a different dilution?

Because that will determine how useful the electrical test is to other people. Otherwise you might need two tests; one to measure dilution, then the other to measure electrically (and compare the two tests on a lookup table etc).

 

There may be wet chemical methods that could be used, such as precipitating the copper out of solution or such. Maybe your chemistry teacher can help you. It's just been too long since I took chemistry.

RB is correct that you'll need to correct for water evaporation in order to use conductivity. Total dissolved solids could be measured gravimetrically (with a balance, drying oven and such) but that would be a pain. There are quicker and easy methods to measure density, refractive index, freezing point depression and so on, and these might be useful.

Or you could design you own test. See how much weight is removed from a penny by overnight treatment with, say, 2 ml of your solution. Just an idea.

These ideas are my way of saying I'm not optimistic about measuring conductivity. I think the errors in the measurement will make the results not so useful to you. Conductivity is an intrinsic property, but measurements of it are extrinsic (and depend on scale, surface area and so on). Things like density, refractive index and freezing point are all intrinsic properties as well and can be measured more directly.

 

The solution I use is already made up, I get it very concentrated in large containers, I didnt know you were meant to dilute it .
Maybe conductivity isnt the right way to describe what I want to do. Copper is very conductive, so the more dissolved in a solution should have a greater effect on something???
What I am thinking of is maybe a two part test, one of specific gravity to take care of solution strength etc and a test to see how much better the solution is at conducting???
I am reading this back and I know I havnt explained it at all well, I cant find the right words :S, I will experiment.
I spoke to my chemistry teacher today, he is also convinced that the more copper that is dissolved thenI should see it affect something electrically, what that something is, I am not sure of yet.
RB thanks for the encouragement, I was about to give it up as a useless idea. I do agree I need to thoroughly test it out, somwhere there will be a relationship between temperature, specific gravity (concentration of ferric chloride) and the amount of copper dissolved. Once I have understood the relationship, then making a device to test the life left in a solution should be easy (ish). adding a temperature probe isnt a problem (metal case DS18B20 in heat shrink tubing to protect it from the chemical).
I dont know if there are specific gravity sensors??? I have only seen dad's home brew glass ones .
I need to design a proper experiment, any input and idea are welcome.
I am also working on a device to 'rub' the board against a sponge in the solution, doing it by hand reduces etch time significantly, making something to do it automatically isnt proving easy

 

LIGHT!!!!! LIGHT

I remember a machine that can spot fake wine and beer because of a light sensor! it detects even the tiniest amounts of added water to wine or beer, I will find out more about that, it would be great if i could shine a light and have a micro filter out say the green/blue of copper and other stuff, to give an accurate reading of feric chloride strength, then conductivity test to see how much life is left! as RB says then use a look up table to get a reading!
At least I now have a path to follow, if it dosnt work I will try something else

 

I dont know if there are specific gravity sensors??? I have only seen dad's home brew glass ones .

Those are excellent when used properly! In an industrial lab, density is measured by passing the solution under test through a vibrating tube. The frequency depends on the mass in the tube, it's internal volume is fixed, so the measured frequency is a direct measure of density. Temperature controlled of course.

At home or in a less equipped lab, you have to have a scale and calibrated volumetric device. Pipetting a known volume onto a scale is a tried and true approach. Or you use the spindles. Those are the preferred method in many cases.

Be careful to use the terms density and specific gravity accurately. They are often mistakenly used interchangeably. The latter is ratio of the density at a specified temperature to that of water at another specified temperature. A value of specific gravity quoted without those temperatures is a nearly useless value.

 

LIGHT!!!!! LIGHT

Wine and beer are both mostly water, so I'm not sure what this is about.

But yes, a spectrophotometer might help you. Copper compounds tend to be colored.

 

Or I could come at it from the other end, I could ignore what makes the up the solution be it water copper or ferric chloride, what I am actually after is a way of measuring how good or how effective the solution is at etching. That way I dont need to worry about the why's of it.
For example a test using a fixed current might tell me the solution is weak at etching, that may be due to a less concentrated solution or a nearly used up one. The reason its weak dosnt matter, what I am after is a way to tell how well it will etch.
So for example you mix a fresh solution and measure it, it will have a scale that goes from very strong to very weak, then as you use it you can put the electrodes back in and see when it gets weaker.
Not sure I have explained that well either :S.
I am building a test setup at the moment to experiment with

 

If you are going to measure electrical conductivity do not use DC current or voltage. Use AC. You can do this easily with two I/O pins on a microcontroller and a capacitor across the two pins.

Charge the capacitor and measure how long it takes to discharge.
Reverse the polarity of the charge and repeat the measurement.

 

AWWWWw now YOU are going to make me sound STUPID!

but I just have to ask. I assume I would use a electrolytic capacitor? say around 1uf?
But how can I change the direction of charge without damage to the cap?

Am I misunderstanding?? I think I am,

Take a pic

Just for the sake of argument, put a cap across pin 4 and pin 5 (random choice), pull pin 5 high and pin 4 low, charge and measure, then turn pin 5 low and pin four hi, charge and measure..
Why will this not damage a electrolytic cap? If you are not meant to reverse the polarity?

So sorry it sounds stupid, but I do have huge gaps in my knowledge as I tend to learn in a slightly random way, and I dont always start at the beginning, sometime I start in the middle and work back towards the ends.

 

I have a idea but first I will wait for the answer , then I will explain my idea and see what you think

 

Ok, dissolved copper is no longer metallic copper and doesn't behave like a metal. That said, you cannot use a simple conductivity test to determine the concentration of ferric vs copper chloride.

The reaction is quite nice from a chemistry teacher's point of view, ferric chloride (FeCl3) reacts with elemental copper to make ferrous chloride (FeCl2)plus cuprus chloride and eventually cupric chloride (CuCl2).

All free solvable ions conduct about the same - they are essentially electrolytes. Look up colligative properties in a physical chemistry book.

In any case, there are several methods to test - a UV- visible spectrometer will be the easiest. Plot the absorption vs Concentration at the peak absorption wavelength for each (ferric chloride and Coptic chloride). The. Use this when measuring the mixture. Ferrous chloride can also be measured. You will see some wavelength shift as concentrations become very dilute. Look for a "Spectronic 20" from
Perkin Elmer - the old standard in the lab.

Aside from that, atomic absorption or emission spectroscopy but out of reach technically for an afternoon project or price.

If you want to stick with an electric method, look into voltametry item to measure how many coulombs are needed to make the oxidation as you convert the available ferrous ion to ferric ion. Not a common method and accuracy can be difficult based on side reactions and dissolved oxygen changes.

In the end, a "Spec 20" is your best bet and can be found online for $40. Otherwise, do what I do - change out the ferric chloride when it seems too slow.

 

I mentioned earlier using a penny etching test. I should point out that a modern U.S. penny is relatively pure copper plated over almost pure zinc. For a test to be valid, I'd think you would want to avoid etching through into the zinc. Maybe spent copper wire (scrap leftover from home construction, for instance) would make a better test material.

 

AWWWWw now YOU are going to make me sound STUPID!

but I just have to ask. I assume I would use a electrolytic capacitor? say around 1uf?
But how can I change the direction of charge without damage to the cap?

Am I misunderstanding?? I think I am,

Take a pic

Just for the sake of argument, put a cap across pin 4 and pin 5 (random choice), pull pin 5 high and pin 4 low, charge and measure, then turn pin 5 low and pin four hi, charge and measure..
Why will this not damage a electrolytic cap? If you are not meant to reverse the polarity?

So sorry it sounds stupid, but I do have huge gaps in my knowledge as I tend to learn in a slightly random way, and I dont always start at the beginning, sometime I start in the middle and work back towards the ends.

Simple answer, don't use electrolytic capacitor. The value doesn't have to be so high. 1nF should do.

 

I am going to try out the pic CTMU, normaly used in cap touch applications, but might be useful in this. So job for tomorrow after school is to set up a test bench, have a play around and take it from there.
The solution is pre made, you meant to dilute it but I never have, so first experiment is to use same quantity in 2 containers, one i will measure as is, the other I will dissolve a known amount of copper wire, I will keep doing this and time how long it takes to dissolve completely, using the other solution as a standard.
Give it a few days and se if I have any useful data. Its a start point I guess.
Also I will do set dilutions and measure each. No idea how useful any of this will be, but it gives me a start point

 

If you use ground planes there is much less copper to dissolve.
I don't mean ground planes in between but a ground plane on the same side as the tracks to fill in what isn't used.

My CAD software uses ground planes and also copper fills.

 

...
The solution is pre made, you meant to dilute it but I never have, ...
...

You don't need to dilute. If you agitate and flush fresh solution over the PCB when etching the full strength solution etches very well, faster than a diluted solution. If you don't flush properly when etching the full strength solution can get patchy or eat under the track masks.

I just slosh by hand to keep the solution flushing over the PCB, and with full strength solution the etch only takes a few minutes. The best way to flush is to get a $10 fully submergible pump (they only have plastic parts exposed to the liquid) and with the PCB on an angle use the pump to deliver solution to the top of the PCB so it flushes over the surface.