Back to Shoptalk Index:
Ferric Chloride - how diluting it with water affects the way it etches copper.

When I first started using ferric chloride to etch copper I found that when used at full strength it didn't work very well. There was lots of foul-biting, and weird things happened in the lines and areas that were supposed to be etching.  I diluted it with water, and have been using it for years without any trouble.  I use it in a flat tray with the plate lying in the tray face up, and enough ferric chloride solution to cover the plate.  I mostly agitate the solution as the plate etches by periodically lifting one corner or edge of the tray, to surge the solution back and forth across the plate.  If I want to ensure maximum etching in a particular area, I use an automotive battery-filler (available at automotive stores) to suck up some solution and baste the area of concern with it.  This washes deposits out of the lines, and ensures maximum exposure of the copper to the ferric chloride. (Battery-Filler)  (It's also good for removing bubbles if you're using zinc and nitric acid.)

Recently I decided to do some experiments to see how ferric chloride at different dilutions will etch copper.  I used as a "full strength" solution the solution suggested on the can of Iron Perchloride crystals (purchased from Graphic Chemical - I assume they're all pretty much the same) which was one pound of crystals dissolved in one pint (two cups) of water to make an approximate 40° Baume solution.  I also tried  6 other dilutions which were one part of the full strength solution diluted with 1, 2, 3, 4, 5, and 6 parts of water, for 7 dilutions in all.

I ran two sets of experiments. One was a thin copper wire with a weight on either end draped over the rim of a small jar containing the test solution. The idea was to see how long it took the solution to eat through the wire and allow the weights to drop.  The second was seven small test-plates uniformly aquatinted with a fairly coarse rosin hand-aquatint, each of the seven plates being etched for 15 minutes in one of the seven test solutions.  They were etched in a small plastic tray, flat, face up, and only agitated occasionally by tipping the tray to make waves of solution pass across the plate.  No battery-filler was used.  The test plates were cut from a 5x7 in. copper plate, one inch wide, for seven 1x5 in. plates.  After etching I inked and wiped them, placed them back together in the press and printed a 5x7 in. image of them. (see bottom of page)

Equipment & Procedures

Test Jar, empty.   Test Jar, in use.   Thin copper wire.

One weight is a scrap piece of zinc, which falls into the sink and makes a loud clatter. The other weight is a white glass knob.  The copper wire is a single strand from a piece of 22 AWG speaker wire made from 17 twisted strands per conductor.

These are the 7 test plates, aquatinted.

This is a test-plate in the ferric chloride.

To run the etch tests, I dissolved a 1-pound can of Iron Perchloride crystals in two cups (one pint) of water.  I let it sit overnight to cool off, and also a gallon plastic jug of water so both were at room temperature (80° F) (summer's here) by the next day. Then I put a cup of test solution in the jar and a cup of test solution in the small tray. Then I placed one end of the wire (the white glass weight) into the solution in the jar and started a stop-watch.  Then I placed a test plate in the solution in the tray and started a kitchen timer counting down, which was set for 15 minutes.  When the scrap of zinc clattered into the sink, I stopped the stop-watch and noted down the time elapsed.  When the kitchen timer started beeping I took the test plate out of the tray, rinsed and dried it and set it aside.  Then I put all the solution back into a gallon plastic jug and added two cups of water, mixed, and poured out two cups of the new dilution.  (I chose to ignore the fact that when you add the crystals to the water, you get about 2 2/3 cups of solution, not two cups, to avoid complication.) Since adding water to the ferric chloride solution caused a heat reaction which warmed it significantly, I put the two cups of test solution into a freezer long enough to cool it down to 80° F. before continuing with testing.  I repeated this cycle until I had tested all 7 of the solutions.

To check the accuracy of the "wire" test (not as part of this experiment) I repeated it 6 times when checking an old batch of 1:1 ferric chloride, which is the dilution I etch with, and compared the readings. I refreshed the test sample of solution for each test, and tried to do everything as uniformly as possible. The readings, in order of occurrence were: (1.) 7 min. 23 sec. (2.) 7 min. 4 sec. (3.) 7 min. 1 sec. (4.) 6 min. 37 sec. (5.) 6 min. 36 sec. (6.) 7 min. 47 sec.  Not very accurate, but accurate enough.  I suppose the main alternative to this would be taking readings with a hydrometer. I don't own a hydrometer, but I assume a series of similar repeated tests would yield very accurate and consistent results.  The question I have about hydrometers is not about accuracy, but about relevance.  A hydrometer accurately measures one very specific thing; that is the weight of whatever you're testing compared to the wieght of an equal volume of water at 60° F.  When you're mixing a fresh batch of ferric chloride, that reading is probably both accurate and relevant. Of course with a fresh batch of ferric chloride you probably already know what you've got before testing it with a hydrometer.  Throughout the life of that batch of ferric chloride as it gets used it becomes less and less ferric chloride, and more and more a completely different chemical mixture.  I don't know how the specific gravity of fresh ferric chloride compares to that of completely spent ferric chloride, which is no longer even ferric chloride.  So as your batch of ferric chloride gets older and information about how it's going to etch your plate becomes more and more critical, your hydrometer is increasingly giving you information about something that has nothing to do with etching.  Not to mention other issues like temperature, age, or whether or not some unknown party has strengthened, diluted, or contaminated the bath.  The wire test may only give you "ballpark" information, but at least it's the information you need.

Results
(Notice how almost no etching occurs when using full strength solution compared to radical increase in etching when diluted 1:1 with water, then slight decrease in etching as more water is added. When diluted with 6 parts water, the solution still etches about 7 times faster than when at full strength.)

Test Solution (each plate etched for 15 minutes)	Time required to etch through the wire
1 (full strength, 40° Baume)	2 hrs. 9 minutes
1:1 (1 pint full strength, 1 pint added water)	7 min. 44 sec.
1:2 (1 pint full strength, 2 pints added water)	8 min. 46 sec.
1:3	10 min. 22 sec.
1:4	13 min. 44 sec.
1:5	16 min. 53 sec.
1:6	18 min. 26 sec.

 
White deposits formed when etching with full strength solution.  

White deposits compared to normal etch when solution is diluted 1:1.  I didn't try etching a plate upside-down in the strong solution to see if there were less white deposits than when etching it face-up as I did with the plate on the left, but  I suspect that it wouldn't make much difference in the etching action because the wire showed the same dramatic difference in etching times between the full strength solution and the diluted solutions as the plates did, though I never noticed any white deposits on the wire.  If it was a gravity-related issue of something accumulating on the plate when face-up and blocking the etchant, it seems the wire should have etched dramatically faster than  the plate in the full strength solution.  I've also never used a vertical tank (or etched a plate upside-down), so if using one or positioning a plate face-down fixes the problem then it would appear to be gravity-related. I've always gotten clean etches using diluted ferric chloride and a face-up plate, so I've never worried about it.


This is the print made from the test plates. (flipped) There are 7 plates, hand aquatinted approximately equally, and each one etched face up in one of the 7 test dilutions of ferric chloride for 15 minutes.  The first one (1.) is from the 40° Baume solution, and is significantly lighter.  The other six don't show any significant difference.  (Slight differences might have been visible if I had used a box aquatint in stead of a hand aquatint.) It wouldn't be a bad idea to use a weak (1:6 or more) solution for etching fine aquatints to avoid dark areas from blobs of ferric chloride or dark edges where ferric chloride pools while draining off of the plate when removed from the tray.