Electrochemical and Corrosion Behavior of Fe-Ni Alloys in Sulfuric Acid* R. R. SAYANO and KEN NOBE* Abstract “The anodic and cathodic polarization of pure iron and nickel, 20%, 40%, 60%, and 80% Fe-Ni alloys in de-serated 1N HSO4 has been investigated. The anodic Tafel slopes ranged from 41 10 50 mv. The cathodic Tafel slopes of the hydrogen evolution reaction were either 100 or 105 iy. The corrosion rates of the pure metals and alloys were determined by weight-loss, chemical ‘analysis, and polarization methods. The corrosion rates determined by the three methods were in reasonable agreement. Weight-loss and chemical analysis methods indicated @ minimum at 40% Fe. By the polarization method, a minimum was observed at 20% Fe. In addition, the differential capacitance attained a minimum at this same composition, It has been suggested that the minima at 20% Fe might be attributed to the possible existence of the ordered structure, FeNis This paper reports on the investigation of the electro- chemical behavior and corrosion of active iron-nickel alloys. Marsh’ has reviewed previous work on the corrosion rate of iron-nickel alloys. He reported that the corrosion rate of these alloys in sulfuric acid by weightloss measurements decreased substantially from about 2400 mdd for pure iron to 48 mdd for 60% iron. The corrosion rates were relatively constant for lower iron content alloys. Unfortunately, the oxygen content of the solutions was not noted. A similar decrease in the corrosion rate from 100% to 60% iron in aerated 5% sulfuric acid (800 to 150 mdd) with the weightloss remaining essentially constant for lower iron content alloys at 25 C (77 F) was reported by Landau and Oldach? The polarization characteristics of nickel-iron alloys in de-aerated sulfuric acid have been studied by Morioka et al® and by Economy et al.* Morioka? determined the poten- tiostatic polarization behavior of nickel-iron alloys in IM sulfuric acid. The increase in corrosion potential was nonlinear but monotonic from -470 to -270 mv vs SCE (saturated calomel electrode) for pure iron to pure nickel. The largest change in corrosion potential was observed between 100% to 50% iron, whereas, only a slight change was noted from 50% iron to pure nickel. Morioka? attributed the change in corrosion potential with alloying to the shift in the single electrode potential of iron in the alloy and the increase in polarization resistance for anodic dissolution, Economy ef al* investigated the anodic polarization of nickel-iton alloys in 0.001 to 10N sulfuric acid, In contrast ‘submitted for publication October, 1968, *School of Engineering and Applied Science, University of Cali fornia, Los Angeles, Calif. 260 to Morioka’s work,? in which the potential was increased from 0.02 to 0.05-volt increments every one to two minutes, Economy determined the polarization curves for steady-state conditions. The corrosion potential was found to display a regular dependence on composition in fair agreement with Morioka? (542 to -282 my vs SCE in IN H,SOq, pH = 0.56, from pure iron to pure nickel). The largest potential change occurred between 0 to 50% nickel in accord with Morioka’s observation. Experimental The nickeliron alloys were specially prepared by the Chicago Development Co., Riverdale, Md. The alloys were cast from 99.9% iron and 99.9% nickel by vacuum melting and casting followed by hot forging and swaging to nominal 3/8:inch diameter rods. ‘The chemical analyses of the alloys were determined (the nickel content by wet analysis and the impurities by spectrographic analysis) and are shown in Table 1. Pure iron (fom which the alloys were made) was not obtained from the supplier and Ferrovac E (99.9% Fe) from Crucible Steel was used. The electrodes were machined into cylinders 3/8inch in length and 3/8-inch in diameter. Electrodes were cut from the rod with an Al,0} abrasive cutoff wheel. Water was used at all times during cutting of electrodes to prevent heating of the electrode. After cutting to the proper length, the ends of the cylinders were polished with Al; 03 paper. The “mill scale” on the electrodes was removed by slicing off a small portion of the surface (~ .010-inch). On some electrodes, notably 100% and 80% Ni, the surface “mill scale” appeared to be quite deep and as much as 0.05-inch of the surface had to be removed before a smooth surface was achieved. The electrodes were then hand- CORROSION-NACETABLE 1 — Chemical Analysis of Fe-Ni Alloy Nom. % Fe Fe Ni al c cu Mo si Mg 100 Bal 0.02 0.010 0.006, 0.010 0.006 80 Bal 20.56 0.010 0.008 0.001 0.001 0,003 0.001 60 Bal 40.70 0.010 0.008 0.001 0.001 0.003 0.001 40 Bal 59.87 0.001 0.008 0.001 0.001 0.001 0.001 20 Bal 19.55 0.001 0.010 0.001, 0.001 0.001 0 0.100 Bal ois 0.001 0.001 Nom. % Fe Mn Sn s cr Co Ti PYWNO2H, 100 0.001 0.005 0.010 0.005 0.027 80 0.101 0.001 0.002 0.001 0.005 60 0.008 0.001 0.002 40 0.010 0.001 20 0.010 0.001 0 0.001 0.001 0.001 polished on the lathe using a waterproof Al, 03 paper. Grit 240 was used initially to remove the marks due to machining, followed by 360 grit paper. Polishing was carried out using an ample supply of water asa coolant and lubricant. After the final polishing, the electrodes were rinsed in water and wiped dry with laboratory tissue. Electrodes were degreased in a Soxhlet Column with hot benzene for four hours. The degreased electrodes were then annealed under vacuum (<10"$ mm Hg) at 650 C (1202 F) for one hour and slowly cooled (~ four hours) to room tempera- ture. Each alloy was annealed separately. The finished electrodes were then stored in the desiccator until used. ‘The electrochemical cell was similar to that described elsewhere.’ The electrolyte was prepared from reagent grade sulfuric acid and redistilled water (conductivity <10°* mho-cm™). The electrolyte used in all cases was IN H)SO,q (pH = 0.4). The cell and all accessories which were in direct contact with the electrolyte during an experiment were cleaned by soaking in cleaning solution which consisted of an equal volume mixture of concen- trated sulfurie-nitric acid. The cell was soaked for at least 12 hours with this cleaning solution followed by thorough rinsing in redistlled water. The cells and its accessories were then soaked in redistilled water for atleast four hours. The, cell was then given a final rinse with redistilled water and filled with the electrolyte. The electrolyte was de-aerated with prepurified njtrogen gas which was first passed through a gas washing bottle filled with IN H»SOq for at least 24 hours prior to an experiment. Stopcocks were made of Teflon. Special Teflon sleeves were used as sgreaseless seals for ground glass joints. Shielded electrical cables were used, except for the ground leads, with BNC connectors to minimize noise pickup. Electrode potentials were measured with a Keithley Electrometer, Model 603.1) The reference electrode was a saturated calomel electrode (SCE). The electrometer output keithley Electrometer, manufactured by Keithley Instruments, Inc., Cleveland, O} Vol. 25, No. 6, June, 1969 was monitored with 4 2-channel Servo-recorder.‘) Poten- tial transients of the current pulses were followed with the electrometer output connected to a Massa Meterite BSA-260 AR‘) oscillographic recorder. The d-< power supply was a Dressen-Barnes Model 32-101.4) Galvano- static polarization was achieved with the power supply connected in series with a precision resistance decade box, Model 1432-Q,*) and an ammeter, Model 269 VOM.) The pulser described previously was used for the capacitance measurements.® The pulse duration for capaci tance measurements was 5 msec. The pulser output was passed through a precision 100 ohm resistor (0.05% accuracy) and then to the auxiliary electrode. The voltage drop developed across this resistor was measured with one channel of the oscillographic recorder operated in the differential mode. The polarization transients were measured with the electrometer-oscillograph combination with potential being visually read off the electrometer just prior to interruption of polarizing current. The potential- time transients for determination of the double layer capacitance were measured with a Tektronix Type 564 ‘Storage Oscilloscope?) (with Type 3A3 vertical amplifier) operated in chopped-dual trace mode. The vertical amplifier of the scope was operated in the differential-a-c mode. This, climinated the use of a calibrated bucking circuit. One channel of the vertical amplifier was connected directly to the cell and the other channel across the precision 100 ohm resistor. The operation of the oscilloscope amplifier in the a-c mode did not affect the capacitance measurements and had the added advantage of eliminating the shift of the electrode potential when the oscilloscope was connected to ®servo-recorder, manufactured by Texas Instruments, Inc, Houston, Texas Meterite’ oscillographic recorder, Hamilton Corp., Waltham, Mass. instrument manufactured by Dresser-Barnes Electronies Corp., Pasadena, Calif. Resistance decade box, manufactured by General Radio Corp., Concord, Mass. (ammeter, manufactured by Simpson Electric Co., Chicago, I Fektronix oscilloscope, manufactured by Tektronix, Ine, Beaverton, Ore manufactured by Baldwin: 261the cell. The oscilloscope trace was recorded on Polaroid film for further analysis. The corrosion cell used for the weightloss and chemical analysis methods of determining the corrosion rate is shown in Figure 1. The stopper and all connections to the cell were made of Teflon. Electrodes were taken from the same batch as those used in the electrochemical experiments. These cells were cleaned in the same manner as the electrochemical cells. The electrodes were activated by immersing in SN H,SO, at room temperature for 15 minutes, thoroughly washed with redistilled water and then placed in the cell. Care was taken during this procedure to avoid touching the electrode surface. The electrodes were rotated at 100 rpm (vy = 5 m/sec). The corrosion potentials were read immediately. cy Teflon steppe —_ Thereafter, both corrosion potential and capacitance were measured every 15 minutes until an apparent steady-state potential was reached (~ 1 hour). ‘The steady-state galvanostatic polarization curves were determined, first cathodically, by increasing and then decreasing current density increments. Steady-state was obtained at each current value. Steady-state is defined here as when the change in the electrode potential was less than 1 mw during 5 minutes. The anodic polarization curve was then measured by increasing and then decreasing the current. Finally, the cathodic polarization was repeated. Between each polarization run, the corrosion potential was allowed to reach a stable value and then the differential capacitance was again measured. For the nonelectrochemical corrosion experiments, the electrodes were weighed on the analytical balance after the physical dimensions were taken with a caliper. The elec- trodes were then placed in the corrosion cell on the glass hooks. The content of nickel and iron in the cell was analyzed spectrophotometrically from time to time. The amount of iron was determined by the thiocyanate method and nickel by dimethylglyoxime.’ The iron was not extracted during chemical analysis of the alloy solutions for nickel since preliminary experiments indicated that there was no interference due to the presence of iron. Concen- tration of iron as high as 10 mg/liter did not interfere with the nickel analysis. In the chemical analysis of the alloy solutions the iron content was kept well below this value by dilution. The nickel ions did not interfere with the iron analysis. After about three weeks the electrodes were removed from the cell, and after drying, the final weight of each Glass Rod electrode was determined on the analytical balance. Nitrogen Gas Inlet Test Sample Results Corrosion Potentials ‘The steady-state corrosion potential as a function of alloy composition is given in Table 2. The corrosion potentials exhibited a large change from 100% to 40% iron, A similar change with alloy composition was reported by Morioka? and Economy.* The corrosion potentials of this work are compared with the results of Morioka? and Figure 1 ~ Corrosion cel. Economy* in Table 2 TABLE 2 — Corrosion Potential of Fe-Ni Alloys in De-aerated IN H2804 (mv vs SCE) Economy et al* Before Galvanostatic After Galvanostatic IN H,804 + SN K2S04 Morioka et al? Fe Polarization Polarization (pH= 0.56) IMH,S04 100 31735 “sun as 342 470 80 404 #3 414413, 42 370 60 “365 #10 -364 320 362 320 40 -329 410 334411 342 -290 20 33253 327 #10, 302 285 o 7323 +10 326 +4 282 270 262 CORROSION-NACEOpen-Circuit Differential Capacitance ‘The capacitances of the alloys were measured after the corrosion potential reached steady-state and before polar- ization. These values are plotted in Figure 2 as a function of alloy composition. The range of the differential capacitance of the unpolarized electrodes is indicated in Figure 2 by the vertical line through the average values. The largest variation of about + 4 uffem* was observed for pure nickel. Figure 2 shows that the capacitance decreased as the iron content decreased and reached a minimum of 13 yf/om? for 20% iron. The capacitance of pure nickel, 25 yf/em?, was substantially higher than the 20% iron alloy. The capacitance of pure iron was 33.0 + 1.1 pffem? and was found to be substantially lower than values reported previously by others.”** The average value of 25 Capacitance (uf em? 100 20 60 20 20 0 Alloy Composition (% Fe) Figure 2 — Differential capacitance of Fe-Ni alloys in de-serated 1N H2SOq. % Fe Cathodic Tafel Slope COverpotential (mw vs SCE) 8 uffem? for pure nickel was in reasonable agreement with the value of 20 pffcm? reported by Kronenberg? for electropolished nickel. Polarization Polarization runs were performed on each electrode in the following way. Cathodic and anodic pulse polarization runs were conducted before continuous polarization. How- ever, steady-state was not achieved for the pulse experi- ments. For the continuous runs, polarization was per- formed by increasing and decreasing current density increments. Cathodic polarization was conducted first, followed by anodic, and then a cathodic run was repeated. The continuous cathodic polarization results for the increasing current density increments are shown in Figure 3. These results indicate that the rate of the hydrogen evolution reaction (her) increased with increase in iron content with the exception of pure nickel. The rate of the her on pure nickel was greater than 20% and 40% iron alloys and slightly less than 60% iron. The cathodic Tafel slopes for the her are given in Figure 3. These values were approximately the same for all electrodes, either 105 my or 100 mi. Other investigators obtained cathodic Tafel slopes for the her on iron'®"" and nickel!?"? in reasonable agreement with this work. The cathodic polarization runs performed after the anodic runs yielded cathodic Tafel slopes of approximately 120 mv. For anodic polarization, the runs performed by decreasing current density increments were more consistent and reproducible than the runs by increasing current density increments. The Tafel plots of the former polariza ‘fel slopes which are given in Figure 4 ranged from 41 to 50 my. The pure iron and nickel exhibited smaller anodic Tafel slopes than the iron-nickel alloys. The value of 41 my for pure iron in H2SO4 was in agreement with the results of Bockris! 1 and Kelly.14 % Fe Anodic Tatel Slope 3 a 2 3 ‘Current Density (amp/cm“) 8 8 © 100 105 mw 200F 3 "go 105 . 500} eee 3 60 100 2 4 10s ool 40 105 09 + 70 100 so -700 we 0 10 107 7 0 "10 10? Courrent Density lamplem? Potential (nw va SCE) Figure 3 ~ Continous cathodic polarization of Fe alloys Figure 4 ~Continuous anodic polarization of Fe loys Vol. 25, No. 6, June, 1969 263Corrosion Current The corrosion rates which were determined by weight- Joss measurements and chemical analysis of the corrosion TABLE 4 — Percent Iron of Nickel-Iron Content in Solution During Corrosion 20% Iron Alloy cell solutions are given in Table 3. When the composition of Time 60% lon Alloy 40% ron Alloy on ‘Fe Fe ‘Fe each corrosion cell solution was plotted as a function of © ©? FF time, the resulting curve exhibited a nonlinear increase with 20, 604 449 203 time for approximately 50 hours followed by a linear 40, sna aos 23.4 increase. Therefore, approximately 5O hours were required 2 seo os 244 to reach a steady-state rate of dissolution for the alloys. 35, Sail ae ae Pure nickel was also observed to require about SO hours 16 393 406 21 before it reached a steady-state corrosion rate. Pure iron 200 393 403 22 was observed to have a constant rate of dissolution throughout the experiment. The slopes of the curves representing the total amount, of metal ions in solution as a function of time were the instantaneous corrosion rates. The steady-state values are the ones given in Table 3. The average corrosion rates determined from the total weight-loss of the electrodes which were exposed for approximately 500 hours are given in Table 3. The difference in weight of the electrodes before and after the experiment was used to determine the average corrosion rates by the weightloss method. The corrosion rates were then converted to an equivalent corrosion current density by application of Faraday’s law. It was assumed in this calculation that both nickel and iron went into the solution in the divalent state, ie, Ni= Ni"? +2€~ Fe=Fe? +27 Since the number of equivalence was the same for both, iron and nickel and from the chemical analysis of the solution, the alloys dissolved uniformly as shown in Table 4, the atomic weight of the alloys was taken as the sum of the pure component atomic weights multiplied by the atomic fraction. The atomic fractions were calculated from the alloy compositions. The corrosion current by the weightloss and the chemical analysis methods showed a large decrease of over two orders of magnitude from pure iron to 60% iron. A minimum in the corrosion rate was observed at 40% iron. The corrosion rate was about 1 majem? for pure iron and about 4 a/em? for pure nickel. The sharp decrease in the corrosion rate from 100% to 60% ion is in agreement with the results reported in Marsh's review! that for Fe-Ni alloys in the same range of iron content there was a decrease in weight-loss of NOTE: 80% fron alloy: at 147 hrs, 78.8% Fe and at 220 hes, 78.5% Fe, approximately 1/50 that of pure iron (100% Fe, 2400 mad 10" Ajem?; 60% Fe, 48 mdd = 2x 10° Ajem?). For alloys with iron content less than 60% Fe, the weight-loss ‘was essentially constant. The corrosion current obtained by electrochemical polarization was determined by extrapolating the cathodic Tafel curves to the corrosion potential. In Table 3 and Figure 5, the corrosion rates determined by continuous cathodic polarization are compared with the weight-loss and chemical analysis measurements. A minimum in the corrosion rate as determined by polarization was observed at 20% iron, ‘The corrosion current of about A/ont® for pure iron was in agreement with those obtained by Kelly*# (2.8 x 10°* AJem*) with the Type 1 behavior of Ferrovac E iron of Makrides!® (4.0.x 10~* A/em?). For pure nickel, the corrosion current was 6.8 x 107° A/cm? ‘The results of the corrosion rates which are given in Figure 5 show that there is excellent agreement between the corrosion rates determined by the weightloss and chemical analysis methods with the exception of the 80% iron alloy. It is indeed remarkable in view of the simple experimental procedure for the weight-loss method. Only ‘two measurements were required, the initial and final weights of the electrode, in addition to time-interval of the experiment. On the other hand, the chemical analysis method was very time consuming, requiring frequent analysis of the solution during the experiment. It should be noted that the geometry of the corroding specimens for the TABLE 3 ~ Corrosion Rate of Nie Alloys in De-sorated 1N H2S04 Weight Loss Weight Loss ‘Chemical Analysis Chemical Analysis Polarization 9% Fe (gmjem*-see) (gmjem?-sec) (Alem?) (Afom?) (fem?) 100 321x107 287x107 9.89 x 10 46 x10 80 273x107 1:23x 1077 420% 107* 0.66 x 107% 60 8.83 x 10722 7.01 1071? 2.99x 10° 237% 107° 18 x 107% 40 484x 10722 4.96 x 10720 1.62 107% 1.66 x 107% 45 x10 20 6.06 x 102° 6.00% 1072° 201x 10° 1.99 107% 25 x10" 0 115 x 107 115x107 3.78x 10° 377x107 68 x10 264 CORROSION-NACE10 a i © Polarietion 210 We weg los : Chemie alysis i 8 10° , 10° 10080. 60 40 20 ° Alloy Composition, % Fe Figure 5 — Corrosion rate of Fe-Ni alloys. weight-loss and chemical analysis methods was different than for the polarization method. For the former, the entire electrode surface, the cylindrical surface, the ends of the cylinder and the inner surface of the hole drilled through the center of the electrode was exposed to the corrosion environment, The entire surface area exposed was ‘measured to determine the corrosion rate per unit surface area, For the polarization experiments, only the cylindrical surface area was exposed to the corrosion environment. This difference in geometry should account, at least to some extent, for the discrepancy between the corrosion rates determined by the polarization method and the other two methods. It should also be noted that polarization was initiated several hours after immersion of the electrode in the solution. Since the chemical analysis method showed that approximately 50 hours were required to attain the steady-state corrosion rate, with the exception of pure iron, the corrosion rates determined by polarization were not the steady-state values. The corrosion rates before steady-state was achieved were larger than the steady-state values as indicated by the chemical analysis experiments. The results shown in Figure 5 are in accord with this observation for alloy compositions < 60% iron. ‘The results shown in Table 4 show that surface enrichment of the more noble component, nickel, did not ‘occur since throughout the experiment the solution compo- sition was essentially the same as the alloy composition. ‘The variation in the corrosion behavior of alloys ‘composition can often be explained in terms of Uhlig’s electron configuration theory." The critical composition Vol. 25, No. 6, June, 1969 of an alloy for which this theory is applicable is when the delectron vacancies of the electron acceptor metal become completely filled. For Fe-Ni alloys, the electrons are transferred from Fe to Ni.'® Since the critical composition of the Fe-Ni alloys has been determined to be 66% Fe,'® the alloy behaves like iron above this composition. For iton composition lower than 66% Fe, delectron vacancies of Ni remain and the alloy exhibits the behavior of nickel. Figure 5 shows that the corrosion rate of the Fe-Ni alloys exhibited a substantial decrease between 100% to 60% iron (almost three orders of magnitude by the weight-loss and chemical analysis methods). For subsequent, decrease in iron content, smaller changes in the corrosion ale were observed. These results are in accord with the electron configuration theory. ‘A minimum in the corrosion rate was observed for the 40% Fe alloy by both the weight-loss and chemical analysis methods as shown in Figure 5. The polarization method indicated a minimum in corrosion rate for the 20% Fe alloy. Utida and Saito’s'? results indicated a minimum in the corrosion rates of Fe-Ni alloys for 25% Fe in 10% HCL They observed a more shallow minimum for the 20% alloy in 10% HzSO4. The results of Landau and Oldach® indicated a minimum at 30% Fe. In view of the minima observed in the corrosion rate of Fe-Ni alloys between 40% to 20% Fe and the minimum in the open circuit differential capacitance at 20% Fe as shown in Figures 5 and 2, respectively, it is of interest to note that an order-disorder transformation is known to ‘occur in the composition range of 50% to 20% Fe for Fe-Ni alloys.'® The ordered state is a face-center cubic super- lattice, FeNis. Wakelin and Yates'® reported that the maximum changes in structure-sensitive properties for the Fe-Ni alloys occurred at the 25 atomic % Fe (atomic % is approximately weight % for the Fe-Ni alloys). The order- disorder transition temperatures ranged from about 494 to 500 C (924 to 932 F).!® Although in this study the Fe-N alloys were slowly cooled in the furnace after they were vacuum annealed at 650 C (1202 F) for 1 hour, it would seem that the alloys were not maintained for a sufficiently ong period in the range of the order-disorder transition temperatures to form the ordered structure. Nevertheless, the formation of the superlattice may have occurred, at least to some extent. In view of the observed minima at 40% and 20% Fe as mentioned above, it would be of fundamental interest and of practical importance to investi- gate in great detail the corrosion behavior of both the ordered and disordered structures. Acknowledgment This work has been supported by the University of California's sea water desalination program. One of the authors, R- R. Sayano, was the recipient of a NASA predoctoral traineeship during the course of this study. 265References 1. J. S. Marsh. The Alloys of Iron and Nickel-Vol. 1, Special Purpose Alloys, 1st ed., p. 491-S01, McGraw-Ilil, New York (1938), 2.R, Landau and C, S, Odach. Trans. Electrochem, Soc, 81, 521-571 (1942) 3. K, Shiobara, Y. Sawada, and S. Morioka, Trans. JIM, 5, 97-101 (1965). 4. G. Economy, R. Speiser, F. H. Beck, and M. G. Fontana. J Electrochem. Soc., 108, 337-343 (1961), 5. RR, Sayano and K. Nobe. Corrosion, 22, 81 (1966). 6. Methods for Chemical Analysis of Metals, ASTM, Philadelphia, Pa. (1963). 7, J. OM, Bockris and D. Drazic. Electrochim. Acta, 7, 293-312 (1962) 8, Y. V. Durdin, L. Kish, and V. L Kravtsov. Soviet Electrochem. ‘Vol. I, 123-134 (1961). 9. M, L. Kronenberg, J. C. Banter, E. Yeager, and F, Hovorka. J Electrochem. Soe:, 110, 1007-1013 (1963). 10. A.C. Makrides. J. Electrochem. Soc., 101, 869-877 (1960). 11. J. OM, Bockris, D. Drazic, and A. R. Despic. Electrochem, Acta, 4, 325-361 (1961). 12, E. 1. Mikhailova and Z. A. lofa, Electrokhim, 1965), 13, T. Hurlen. Dissolution of Nickel in Acid Chloride Solutions, Scientific Paper 64-186-390-Pi, Westinghouse Res. Lab., Pitts burgh, Pa., December, 1966. 14, E.J.Kelly.J. Electrochem. Soe., 112, 124-131 (1965). 15. H.H. Unlig. Z, Elektrochem., 62, 700-707 (1958). 16, J.S, Cho and J. H. Schulman. Sur. Sei, 245-251 (1964. 17. Y. Utida and M. Saito. Sei. Rept, Tohoku Imp. Univ., Sendai, Japan, 14, 295-313 (1925), 18, R. J. Wakelin and E, L, Yates. Proc. Phy. Soc,, B66, 221-240 (1953). 1, 107-110 DISCUSSION Reviewer Comments: The authors do not state how their approach differs greatly from past work. What specifically new does it add? What are they doing new or better than that which has been done before? If their approach does differ in a significant way, they should discuss the reasons for their results being more valid than others and why the results differ Reply by Authors: It seems to us that we have indicated that the approach of our work differed considerably from other work on iron-nickel alloys. According to Marsh's review, the corrosion studies of other workers on these alloys were by the weight loss method. Other more recent ‘work was mainly concerned with passivation characteristics. Our investigation, on the other hand, is on the active behavior of iron-nickel alloys. In addition, corrosion rates were determined by electrochemical and weight loss methods and by the periodic chemical analysis of the solution. The differential capacitances at the corrosion potential were also measured. To my knowledge, an investigation such as ours has not yet been reported by any other group In regard to the reviewer's last comment, it is often impossible to judge whether ones results are more valid than those of others in metallic corrosion studies. The complexities inherent in most corrosion systems make such comparisons difficult at best. Much of the work reviewed by Marsh was conducted with systems that were not very well defined, We have compared our results with those of others when it was appropriate. It is quite evident that considerable more work must be done to really understand the active behavior of iron-nickel alloys. Even after the enormous number of publications on iron corrosion in acid solutions, the mechanism of the corrosion of iron is not yet fully understood. Reviewer Comments: The authors have been thorough in their experiments, but I cannot understand their use of a calomel reference electrode. Even traces of chloride ion could be very significant. I feel that in order to properly define the system investigated the authors must show the level of chloride ion present in their system. Other impurities, such as mercury ion, are another problem also Dissolved H atoms could affect electron configuration also. Has this been taken into consideration? Reply by Authors: In respect to this concern of contami- nation of the solution due to mercury and chloride ions, 1 ‘can assure the reviewer that there was none. First of all, the reference electrode was in a separate compartment. Second, commercial calomel electrodes were used which were designed to minimize loss of the calomel. Third, there was a stopcock placed in the bridge between the test solution and the reference electrode compartment. This stopcock was closed. However, the conductivity between the test and reference electrode was large enough for precise potential ‘measurements. The input impedance of the electrometer was 10'* ohms. We have referred to our previous paper listed as Reference 5 for the description of the cell. Data of concentration of dissolved H atoms was not obtained. The interpretation of the change in corrosion potential with composition in terms of the electron configuration theory was only of secondary importance to use. It was merely to show that such an interpretation was a reasonable one. 266 CORROSION-NACE