Today is chemistry day...
I need an activator for through-hole plating. The chemicals I need for this are weeks away so I’m going to try some alternative processes.
One technique is to use graphite power in acrylic paint or gum Arabic but I’m going to start with an alternative chemical processes using silver ammonium acetate complex.
When working with strong acids and bases, silver salts and chemicals in general, wear gloves, eye protection and a lab coat. The hazards of acids and bases should be apparent, but many silver salts such as Silver Nitrate will burn and react with salts in skin resulting in black to purple staining that is tricky to remove (ammonium hydroxide and hydrogen peroxide might help here).
I’m missing some of the preferred starting chemicals so there’s a bit of work to do.
First some silver nitrate (I have some but only a few grammes). In a fume cupboard or outside, dissolve about 65 g of silver (mine was electrolytically recovered from photography developer) in a similar volume (65 ml) of hot fuming Nitric Acid. On standing the white silver nitrate crystals eventually form as the solution cools. Cool in a refrigerator and pour off the excess liquid (dilute Nitric Acid), and allow it to dry in the dark. While Silver Nitrate isn’t particularly light sensitive, it isn’t completely light stable either.
Figure 1. Silver Nitrate
Next make about 40 g of Sodium Acetate from 50 g of Sodium Bicarbonate and about 600 ml of white vinegar (which is about 8% acetic acid). Reduce with heat to about 100 ml. This will be off white due to impurities in the vinegar. If you want nice white sodium acetate then use an appropriate volume of Acetic Acid rather than vinegar.
Now make up a solution of 20 g Silver Nitrate in 10 ml of distilled water. Don’t use tap water because this is likely to contain chlorides that will precipitate out of solution. Add 50 ml of the Sodium Acetate solution, filter and wash the slight off white Silver Acetate and allow to dry in the dark (it is photosensitive).
We’re almost there. Dissolve about 10 g of the Silver Acetate in 10 ml of ~27% Ammonium Hydroxide and add 2 ml of Formic Acid drop by drop. Leave for 24 hours to react and filter out any reduced silver particles. Add a few drops of non-foaming detergent as a wetting agent and some Methylene Blue (or another azo dye) because the solution is clear making it difficult to see if it has coated the drill holes. As a late modification the solution should be diluted with distilled water to make up about 100 ml.
Figure 2. Silver Ammonium Acetate Complex with Formic Acid
To use the activating solution dip the drilled, de-burred and clean copper board in the activator and swish it gently from side to side to ensure that the holes are thoroughly wetted. Allow to drain and air dry. Now bake at about 180°C for about 30 minutes and allow to cool. All going well, the holes will be coated with silver ready for electroplating in an acidified Copper Sulphate bath.
What happens to the activator as it dries is that the excess ammonia evaporates and the silver complex is partially reduced to Silver Oxide and silver metal. During the bake the Formic Acid reduces the oxide to metallic silver particles which are then sintered to a conductive film.
My first test didn’t work well at all. Sure the solution laid down silver metal but it was thick and spongy and not conductive. I figure that this is due to my impatience trying to dry the activator too quickly, and also its concentration which I subsequently diluted to 100 ml total volume with distilled water (this modification is incorporated above).
I made a new test piece from double sided FRP and drilled an array of holes in it ranging from 0.5 to 3.0 mm in diameter followed by cleaning and de-burring. Interestingly some metallic silver was instantly laid down on the copper. After drying, baking and cleaning the edges of the board there was a 38 Ohm conductive path established between the two sides of the board. All of the holes appeared to be coated with a thin layer of conductive metallic silver. The coating is hard and ranges from bright silver to a black lustre. It will buff to bright silver but will rapidly wear through to expose the copper below because it is so thin. I suspect that I have over-buffed the board and significantly reduced the hole conduction in the process. In future I’ll just wash it with detergent and rinse with distilled water.
Figure 3. Test Piece 2 mm Hole Prior to Activation
Figure 4. Test Piece 2 mm Hole After Activation
(The activator has clearly formed metallic silver on the hole walls.)
The activator coating isn’t very thick; less than I can measure with a micrometer (I should have established the plating thickness by weight but the opportunity has been missed for now). The conductivity should have been significantly less than 38 Ohms, but this might be sufficient to allow for through-hole plating. I need to set up a plating bath and see what happens.
I’m going to use a Copper Sulphate Acid bath. Bath solutions seem to vary depending on what you choose to read and brighteners are proprietary (pretty much trade secrets). For the basic plating solution I’m going to try 200 g/litre of Copper Sulphate and 15 g/litre Sulphuric Acid. For a brightener I’m going to try 10 mg/litre of Thiourea and an extremely small quantity of Methylene Blue. These proportions are based on patented processes that you can look up on the interweb. There are many formulations (I like the term recipes but some folk seem to have a problem with this) with molasses, glucose, hydrochloric acid, hydrogen peroxide, surfactants and so forth. In my opinion less additives are generally better. The more stuff that you add to your bath chemistry the harder it will be maintain, and if you get it wrong then you have wasted a bunch of chemicals and reagents. I advise that you add stuff systematically on the basis of plating problems and literature advice.
A quick word on disposal. Copper Sulphate is a really good moss remover and it is sold for this specific use.. If you neutralise the acid with a Sodium Bicarbonate and dilute the solution then it can be used for this purpose.
I need some copper electrodes. Flattened copper plumbing pipe should do the trick and contains about 0.01 % phosphorus which I understand is required. I purchased a 3 Litre plastic container for the plating tank and an appropriately sized tray to contain drips and spills. Slots were made in the tank for the electrodes using a soldering iron against a metal straight edge. I purchased a ceramic frit stone from a pet shop and found an old aquarium air pump to agitate the solution. The stone was adhered to the base of the tank using marine silicone adhesive.
The fasteners must be stainless steel or copper. Steel will corrode very quickly in this application.
Figure 5. Acid Copper Plating Bath
The current density for plating will be between 100 and 215 A/sq m (10 to 20 A /sq ft) which I understand is relatively low. The cell voltage is sufficiently low that electrolysis (the disassociation of water into hydrogen and oxygen at the electrodes) should not occur.
For the first experiment I tried the through-hole activated test piece and plated for 30 minutes. The result wasn’t great. No copper was deposited in the holes whatsoever. I figure that this was due entirely to the vigorous bubble stream which prevented the electrolyte entering the holes and resulted in an upward ripped plating from the holes. As a consequence the plating was not even. I’ve added a cheap and nasty needle valve to the bubble line to reduce the bubbling, but a tube clamp would be a better option.
My second plating attempt was of a single sided board without holes and without bubbles. Straight from the bath the plating is salmon pink (indicative of pure copper) and after light buffing the plating was bright, shiny, even and hard. If Wikipeapdia is to believed then you can expect that all copper plating will have a dull salmon pink surface and need buffing - but I have seen videos that show bright finishes straight from the bath using proprietary brighteners. The average copper plating thickness was calculated on the basis of mass as 6 um (0.006 mm) so I expect that I have made a mistake in calculating the plating current.
We’re still not through-hole plating so I’m going to try a different activator chemistry. The following solution is described on the interweb and appears to produce good results.
Dissolve 30 g of Copper Sulphate in 140 ml of distilled water. Add 22 g of Calcium Hypophosphite and allow to stand for say 10 minutes with occasional stirring to fully react. A white precipitate of Calcium Sulphate will form. Filter and wash with a further 100 ml of distilled water, retaining the solute. Add 40 ml of aqueous Ammonium Hydroxide (27%) which will form a deep blue precipitate that will dissolve once all of the Ammonium Hydroxide has been added. Now add a further 10 g of Calcium Hypophosphite. Finally add about 5 ml of non-foaming liquid detergent which acts as a surfactant.
I tried using household dish washing detergent but this created a soap scum that I had to filter off, and I had to add an additional 4 g of Calcium Hyperphosphite which needs to be in excess at the bottom of the dip tray. Hopefully I haven’t ruined the activator. The purpose of the detergent is to act a surfactant (a wetting agent that lowers the surface tension). Rinse Aid as used in automatic dish washers will likely be more effective without forming a soap scum.
To use the activating solution, horizontally immerse the drilled, de-burred and clean board just below the surface ensuring that all holes are wetted. Allow the excess solution to drain off by holding the board vertically and rotating it from side to side and end to end. Now bake at 125 °C for at least 10 minutes followed by 175°C for a further 8 minutes and allow to cool. Clean the board with water and non-abrasive liquid soap. Rinse and plate.
The new activating solution worked first time with a plating current of about 600 mA (215 A /sq m) for 30 minutes. The only way to test the integrity of the vias was to cut the board around each plated hole, and trim off the plated board edges. They were all a dead short including the sections of board edge. I made some measurements of the 0.7 mm diameter holes. At a current of 1 A the voltage drop across the board was 2.6 mV. So the resistance of the via is about 2.6 milliOhms. Using the resistivity of copper (1.72E-8 Ohm m), the thickness of the board (1.5 mm) and the thickness of 1 oz copper (35 um) we can calculate an estimate of the average through hole plating thickness to be a bit under 4.5 um. This could easily be increased with more time in the plating bath or a higher plating current.
Figure 6. Plated Holes Cut for Testing
Figure 7. Via Resistance Measured at 1 Amp
Time for another plating test. This time I made up four similar test pieces each with fifteen holes ranging from 0.5 to 3 mm diameter. I applied my new copper based activator to Board A and my original silver based activator to Board B.
Figure 8. New Test Pieces
After baking, washing and drying the activated resistance of the silver was significantly lower than the copper (0.1 verses 3.4 Ohms) and the silver activated surface looked appreciably more even than the copper on the inside of the holes.
Figure 9. Activated Boards
(A is copper based activator. B is silver.)
After plating at 600 mA for 30 minutes Board B had one poorly plated 1 mm diameter hole, probably caused by a trapped air bubble. But the plating in the other holes was much more even than Board A, which had significant clumping, particularly in the smaller holes. Both boards buffed to a bright even surface shine with almost no effort. The plating thickness on both boards was calculated by change in mass as 11 um.
Figure 10. Board A. Uneven High-Build Plating
Figure 11. Board B. Clean Smooth Through-Hole Plating
The maximum measured resistance of the 0.5 mm vias at a current of 1 Amp for Board A was 7.0 milli-Ohms. Board B had substantially better conductivity at 4.9 milli-Ohms. From these measurements the effective plating thickness of the 0.5 mm holes on board A was approximately 2.2 u, and for Board B 3.2 um. This is well below the average board plating thickness of 11 um but the bubbler doesn’t actually ensure that the plating solution flows through the small holes. I’d do better to move the board from side to side every few minutes and do away with the bubbler or use a seal-less pump to agitate the plating bath.
The final test was soldering to Board B. The surface of the board took solder readily. With flux and sufficient heat (a challenge when both sides of the board are effectively copper planes) the solder flowed through the holes. It was impossible to remove the solder from the 0.5 diameter holes with solder wick. Larger dimensions were cleared easily and leaving a layer of solder over the plating.
On the basis of these tests the silver activator chemistry consistently produces a more even copper coating in the holes with a lower via resistance than the copper activator. However when it comes to etching with Ammonium Persulphate there is an unforeseen problem. The etchant stalls when it hits the silver layer and requires scrubbing and excessive etching time resulting in significant edge undercutting and risking damage to the resist. I’m going to remake the copper activator using a surfactant and not a detergent to see if this improves the performance. The new solution is made up using 5 mil of Rinse Aid and there were no issues with soap scum. Testing will wait for my next board.
In trying to modify the plating tank (Figure 5 above) to allow free side to side movement of the board I have managed to crack the plastic! That’s not good. I’ll be replacing the tank, maybe making it slightly deeper, and coming up with a motion unit for the boards to keep the plating solution flowing through the holes.
Here is the motion unit with a new tank. It is a simple Scottish Yoke design driven from a 12 Volt DC motor with an integral 36 RPM reduction gearbox, followed with a 3:1 helical spur reduction gear. It is mounted in a simple Perspex frame made to fit the top of the tank, complete with covers to protect the motor and the drive from the plating solution. The gears, yokes and cathode holders were 3D printed. The axle is 6 mm diameter stainless steel rod. The push/pull rods are 6 mm diameter aluminium extrusions. The yoke pins are 5 mm diameter stainless steel rod. The gears and cathode mounts and the yokes are fixed with M2 stainless steel screws making the whole assembly adjustable. The yoke drive wheels are press-fit to flats on the axle to ensure alignment and positive drive.
The unit provides exactly 20 mm of board movement with a cycle time of 5 seconds. This should be slow enough to prevent significant waves in the tank (obviously undesirable), while removing bubbles and replacing electrolyte in the through-holes. I expect that the bubble stone is now redundant as the board motion should ensure turbulent mixing of the plating solution. The motor drive is somewhat noisy even though the linear motion unit is smooth with low friction. I guess this is what you get for cheap.
Figure 12. Motion Unit
Figure 13. Scottish Yoke
I have mounted the anode copper plates in the motion unit. The whole assembly can be simply lifted off the tank, allowing the tank lid to be fitted for ready storage.
I’ve also fitted some removable radiation shields to my toaster oven using simple perforated steel angles. These significantly reduce direct radiation heating of the boards while allowing good convective air flow over the elements (I can feel the difference). But I fully expect that the temperature regulation (either with the thermostat or my electronic PID controller) will have deteriorated because the shield increases the oven’s thermal lag.
Figure 14. Radiation Shields Fitted in Toaster Oven
I’ll be measuring the performance in coming days, and probably need to modify the oven routine to the controller to account for the change in oven parameters. Having completed my experiments there is no easy way of turning this oven, which was specifically modified for IR reflow soldering, to a convection oven without undoing the reflow work. But these ovens are very low cost so I have purchased a new electronically controlled bench-top convection oven for just over $100. I’ve tested it extensively and while it has up to 20°C over-shoot it won’t significantly over-heat my boards, particularly if it is pre-heated for 10 minutes.
You can read more about the new oven tests and my experiments with reflow ovens here.
20 January 2019 and I have still have a problem with the through-hole plating process which appears to be poisoning the etchant. A new solution of my preferred ammonium persulphate etchant (200 g/l at 70°C) typically etches a plain double sided board in about 6 minutes. 30% HCL and 3% H202 etch even faster in less than 4 minutes. But when I etch an activated board everything slows down dramatically by a factor of ten (yes, a whole hour for ammonium persulphate). The problem is not with the electroplating process which consistently lays down salmon pink copper that buffs to a high shine with no effort, so it must be with the through-hole activation. Other folk have demonstrated excellent results with this method so I am doing something wrong.
Perhaps the problem is with my base chemicals? My ammonium hydroxide is good, I am reasonably confident that the Calcium Hypophosphite is true to label, but maybe there is a problem with my Copper Sulphate? I went cheap and cheerful here and purchased copper sulphate from a gardening shop as a moss killer. The MSDS lists this as 98% Copper Sulphate Pentahydrate but when I dissolve it in distilled water the solution turns quite cloudy and can only be cleared by filtration. Maybe there is some sort of flowing agent or other impurity here?
More experimentation is required...
The silver through-hole activation described above worked significantly better than the copper based activator, but both caused issues with the etchant. Silver, in particular, does not dissolve in Hydrochloric acid. There are some other electroless processes that I could try, but given the success of the silver through-hole plating perhaps I would do better to introduce another resist mask step and restrict the silver plating to the through-holes? This will also limit electroplating to the pads and activated holes and will save the etchant as there should be no exposed silver. Maybe this is the best way to proceed?
More to follow...