evolve chlorine. The current varies with the cuprous-cupric concentration ratio; if a substantial amount of cuprous ion is present, a high current is required to maintain the anode potential. If this current is sufficiently high, cupric and cuprous ions will not diffuse to the cathode rapidly enough to keep the cathode polarization within the required range. This will result in hydrogen evolution. Attempts to use this technique in its simplest form-i.e., with the reference electrode in the position given as Reference I-on depleted solutions might result in precipitation of cuprous hydroxide because of removal of hydrogen ions. -4measure of self-biasing is available by use of the voltage drop through the solution. If the reference electrode, instead of being placed near the anode or shielded from potential gradients, is placed near the cathode (Reference 2): the voltage drop through the solution will appear as part of the anode potential. Now the larger flow of current in a depleted solution \vi11 cause a greater potential drop, causing the actual anode
potential to correspond to some higher cuprous-cupric ratio than the final control point. This, in turn, will cause the controller to appl) a lesser voltage to the cell than otherwise required. Hence, the cathode potential could be limited so that hydrogen ions are not discharged. As the solution is regenerated, the effective control point moves to\irard the final control point. As the solution approaches the final concentration, the curi-enr: decreases, decreasing cathode polarization, This lo\cer polarization permits progressively more reduction of cupric ions only to cuprous ions. At the control point there is no net reaction, because cuprous ion is oxidized to cupric a t the anode and cupric ions arc reduced to cuprous at the cathode. The current is very low.
cutting produces better printed wiring and photoengravings. Elimination of doum time and disposal and handling. of new and used etching solutions, coupled with the negligible cost of electricity and recovery of the etched copper. constitutes an extensive saving in the cost of etching. Continuous regeneration will keep etching rates constant. permitting assembly line etching. literature Cited
(1) Black, 0. D., Cutler: L. H., Isn. EXG. CHEM.50, 1539-40 (1958). (2j Photoengraoers Bull. 1947 (November) 19-24. (3) Photogrufhic J . 1915 (hprili 1 6 6 ~ 7 5 . (4) Swiggett, R. L., .Uodern Plnstrcs 31, NO. 8, 94-.5 (1954). (5) Vaaler, L. E., Photoengiarers Bull. 1948 (May) 41-4. RECEIVED for review hfay 8, 1958 ACCEPTED December 1, 1958
Conclusions
Cupric chloride solutions with excess chloride ion are superior to ferric chloride solutions for etching copper. The clean, smooth etch with less under-
Division of Industrial and Engineering Chemistry, Symposium on Chemical Aspects of Printed Wiring, 133rd hleeting, ACS, San Francisco, Calif., April 1958.
CORRESPONDENCE N e w Method for Etching Copper SIR: TVe disagree with several statements which appear in “New Method for Etching Copper” [Black and Cutler, IND.ESG. CHEM.50, 1539-40 (1958)l. 1. The authors state that hydrochloric acid was used “because it is inexpensive, easily handled, and readily obtainable.” O n the basis of these reasons it is strictly fortuitous that they chose the best acid. As we point out [IND. ENG. CHEM. 51, 293 (1959)], chloride ion is the essential ingredient in this system. We suggest that the authors try to devise a similarly regenerable system using other inexpensive, easily handled, and readily obtainable acids-e.g., sulfuric or nitric acid. 2. The authors conclude from their thermodynamic evaluation of the process that it is feasible and state that “thermodynamically the intermediate steps are unimportant.” it‘hile this may be true, kinetically the intermediate steps are of prime importance and one needs more than just cupric and hydrogen ions. This Ivhole process is based upon the ability of chloride ion to form stable soluble complexes with cuprous ion. The fact that the solution is a good etchant for copper is due to the formation of these chlorocuprous complexes. 3. The errors pointed out in 1 and 2 are a natural consequence of the Edisonian approach of these authors. They state, “therefore, the various ions present in the spent (ferric chloride) bath were
298
determined and simple solutions were prepared and carefully checked. . .” This further strengthens our feeling that the author‘s choice of hydrochloric acid was happily fortuitous and was made with no understanding of the s y s tem. 4. The oxidation of copper(1) to copper(I1) by oxygen is, we believe. well known and should not have been surprising to the authors. 5. The method of control uhich is used cannot be as simple as the authors state The addition of hvdrochloric acid must be based on the amount of copper etched rather than on the number of boards, since the amount of copFer etched mav vary Ttith the tIpe of board Ana1)sis of the solution on a routine basis is necessary.
LOUISH. SH.ARPE PACLD. GARN Eel1 Telephone Laboratories, Inc., Murray Hill, N. .J. SIR: Dr. Sharpe’s point that the chloride ion is necessary for the establishment of a commercially feasible regenerative system for etching copper is M ell taken and amply corroborated by the excellent experimental work done by him and Dr Garn. The formation of the complex ions of copper and chloride icas noted in our puhlication, although i t was not stressed. We were interested in the development of a commercial process to replace
INDUSTRIAL AND ENGINEERING CHEMISTRY
the ferric chloride sytem and \\.hen we found one \ve were surprised, not that the copper(1) is oxidized to copper(II), but that the reaction is so fast. The entire process is so simple that i t seemed odd that it had not been placed in use years before. The kinetics of the reaction had already established themselves as satisfactory, so \ye \yere interested in the thermodynamics only to make sure that kve were not dealing svith an impossible situation which had been caused to happen fortuitously but \vould later stop. The method of control which has been in operation in our commercial process for over 2 years is chat of checking the time required for etching a known thickness of copper. \\.hen this begins to increase appreciably. some spent solution is drained off and acid water added t u make up the required amount. In our setup this \could amount to perhaps 2 gallons‘ change at a time. Chemical checks are made only occasionally, perhaps once in 6 or 2 months, and no particular difficulty has been encountered. \$’e appreciate Dr. Sharpe’s bringing to everyone‘s attention more clearly the necessity of the chloride ion in making a rapid regenerative system.
OTISD. BLACK LEOS.+RDH. CUTLER Radio Corp. of America, Camden 2: N.J.