Description:
BACKGROUND OF THE INVENTION
Magnetic thin film structures fabricated for computer memory applications are usually formed of Ni-Fe alloys which are prepared by vacuum evaporation techniques. Because of the inherent simplicity of electroplating as a manufacturing technique, attention has been directed to the application thereof to the fabrication of magnetic thin films. A severe problem in plating Ni-Fe magnetic films results when a plating current is initially applied to a Ni-Fe bath. The initial deposit is very rich in iron content and thereafter decreases in iron content until an equilibrium condition is reached and the alloy having the desired proportion of nickel and iron is plated. Since it is only in the initial layers plated that this variance in the proportions of nickel and iron is produced, usually the principal variance is produced within the first 500A. of film deposited. Therefore, this problem has not been too severe when the plated film is very thick. When the final film is to have a thickness of about 1,000A. or less, and the films are to be used in computer memories, which demand constant magnetic characteristics across the entire film, this initial iron rich deposit becomes a severe problem. This is especially so in terms of the magnetostriction of the deposited alloy, since zero magnetostriction is achieved with alloys including approximately 80 percent Ni and 20 percent Fe. When the alloy varies by any considerable degree from these proportions, it does not exhibit zero magnetostriction.
Electrodeposition of Ni-Fe alloys is accompanied by considerable hydrogen evolution which gives rise to alkalization in the vicinity of an electrode with subsequent formation of metallic hydroxides. Consequently, there is preferential deposition of Fe with the characteristics: (a) gradient in composition across film thickness up to approximately 1,000A.; (b) nonuniformity in composition in the plane of the film; and (c) inclusions in the films. In addition, the ratio of the metals in the deposit is not the same as the ratio of metal ions in the solution.
Ni-Fe films for memory application with thickness in the range of approximately 1,000A. to 1,200A. must satisfy stringent requirements in uniformity of both composition and physical properties. In the prior art, copending patent application Ser. No. 601,951 by J. M. Brownlow et al. filed Dec. 15, 1966, now abandoned, and commonly assigned, discloses use of specially shaped current pulses for satisfying these stringent requirements. In greater detail, the noted copending application by J. M. Brownlow et al. discloses that a shaped continuous current or a series of shaped current pulses are applied to effect the plating. The magnitude of the plating current, or of each of the plating current pulses, is initially significantly higher than that required to plate the desired alloy under equilibrium conditions in the bath. The current, or each current pulse, is thereafter decreased with time, preferably in inverse proportion to the square root of time, to provide films with uniform proportions of Ni and Fe throughout the film thickness.
Alternating current is known to have a significant influence on many electrode processes and it has been used in such electrochemical investigations as: (a) the study of electrical double layers as reported in the articles by Wien, Ann. Phys. Lpz., Vol. 58, page 815 (1896); D. C. Graham, J.Amer.Chem.Soc., Vol. 63, page 1207 (1941) and Vol. 68, page 301 (1946); and M. A. Proskurin et al., Trans. Faraday Soc., Vol. 31, page 110 (1935); (b) the kinetics of the formation and dissolution of oxide films as reported in the article by B. V. Ershler, Trans. 2nd Meeting on Metal Corrosion, Acad. Sci., U.R.S.S., Vol. 2, Page 52 (1943); (c) fast electrode reactions as reported in the articles by P. I. Dolin et al., Acta Physicochim., Vol. 13, page 747 (1940); and J. E. B. Randles, Disc. Faraday Soc., Vol. 1, page 11 (1947); and (d) in the electrodeposition and dissolution of metals as reported in the articles by A. T. Vagramyan et al., "Technology of Electrodeposition", Robert Draper Ltd. Teddington Page 95, (1961); and K. M. Gorbunova et al., J. Phys. Chem., 3, 542 (1955).
Further, A. T. Vagramyan et al. reported in Izv. A. N. SSSR, Otd. Khim Nauk, Vol. 3, Page 410 (1952) that alternating current can effect the grain size, brightness and porosity of electrodeposited metals; V. J. Marchese reported in the article J. Electrochem. Soc., Vol. 99, page (195239 (1952) that the superposition of a-c current on d-c current reduces internal stresses in electrodeposited nickel; and V. S. Pat. No. 2,619,454 issued Nov. 25, 1952 by P. P. Zapponi disclosed that the magnetic and mechanical properties of electroplated Ni-Co films could be improved by superimposing a-c current on d-c current during their codeposition. However, it did not disclose any relationship or critical dependence of any film properties on frequency of the alternating current.
OBJECTS OF THE INVENTION
It is an object of this invention to provide a method for the electrodeposition of alloy films which have uniform composition and uniform physical properties as a function of thickness.
It is another object of this invention to provide a method for obtaining alloy films of different composition from a plating bath of constant composition in a controlled manner.
It is another object of this invention to provide a method for matching the ratio of the metals in an electrodeposited alloy film to the corresponding ratio of the metal ions in the plating solution so that the ratio of the metal ions of the plating bath does not change with time.
It is another object of this invention to provide a method for the electrodeposition of alloy films by superimposing a-c current on d-c current with the peak amplitude and frequency of the a-c current being related to the pH of the electrodeposition solution.
It is another object of this invention to provide a method for plating alloy films from elements in a plating bath where the rate of disposition of a first one of the elements to be plated is limited by the diffusion rate of that element, and the rate of deposition of the second element is limited by the rate of the discharge of that element.
It is another object of this invention to provide a method for plating alloy films from elements in a complexing plating bath where the rate of disposition of a first one of the elements to be plated is limited by a chemical reaction of that element in the plating bath, and the rate of deposition of the second element is limited by the discharge rate of that element.
It is another object of the present invention to provide a method of electroplating Ni-Fe films which are uniform in their proportions of nickel and iron throughout the thickness of the films..
It is another object of this invention to provide a method of electroplating magnetic films using alternating current which may be successfully practiced with conventional plating baths to produce uniform binary alloys of nickel and iron.
It is another object of this invention to provide a method of plating Ni-Fe films for use in magnetic thin film memory applications in which the plating current is controlled to overcome the iron rich deposit which is usually produced when a direct current is first applied to a conventional Ni-Fe bath.
SUMMARY OF THE INVENTION
If the rate of deposition of one of the components of an electroplating solution having a given pH is under diffusion control or if it is controlled by a chemical reaction between the metal ion and its complexing agent in the electroplating solution, this invention provides a method of electrodepositing an alloy layer therefrom. There are in the solution a first concentration of a given metal and a second concentration of a given alloying agent and the layer is obtained by utilizing an applied alternating current superimposed on an applied direct current. The steps of the method of this invention for electrodeposition of Ni-Fe alloys comprise:
a. establishing said electroplating solution such that the concentration of the Fe metal is approximately in the range of 10-3 to 10-2 molar, the concentration of the Ni alloying agent is approximately in the range of 10-2 to 10-1 molar and that the concentration of both the metal and the alloying agent is approximately in the range of 10-1 to 10-2 molar, and that the pH thereof is given in relationship to the given concentrations of the metal and the alloying agent;
b. controlling the peak value excursions of the applied alternating current in relationship to the value of the applied direct current such that oxidation of the adsorbed hydrogen is the main anodic reaction of the electroplating solution; and
c. fixing the frequency of the applied alternating current in accordance with a plot of percentage of a component of the metallic alloy deposited from the electroplating solution versus frequency of the applied alternating current.
The plot of percentage of a component of the metallic alloy deposited from the electroplating solution versus frequency of the applied alternating current exhibits the following characteristics:
a. substantially a constant value over a low range of frequencies;
b. starting at a given point, an increasing percentage of deposit of the component over a range of higher frequencies; and
c. after a second point is reached, a substantially constant value over a high range of frequencies.
Generally, by superimposing a-c current on d-c current during the electrodeposition of the alloys, the following results are obtained by the practice of this invention:
1. The difference between the pH of solution at the surface of the cathode and pH in the bulk of the solution can be maintained approximately the same to limit hydroxide formation for iron group metals, and also in all cases where metal ions are used which readily form hydroxides, e.g., Zn, In, Cd.
2. The composition of an alloy electrodeposited from the same solution can be varied in the approximate range of 6 to 60 percent Fe by varying only the frequency.
3. The composition of an electrodeposited alloy film can be maintained constant over a thickness range of approximately 400A to 4,000A.
4. the ratio of the concentration of the metal to the concentration of the alloying constituent or agent in an electrodeposited alloy film can be made to reflect exactly the ratio of the concentrations of the respective ions in the solution.
Though the inventive method, as summarized above, is disclosed in this application as being applied principally to the fabrication of binary alloy films which include only nickel and iron, the inventive method can be employed to prepare ternary Ni-Fe alloys. Further, the Ni-Fe alloys, to which this method is principally directed, are only one example of a rather broad class of alloys which present similar problems when it is desired to plate a film which is uniform in composition throughout its thickness.
The following and other objects, features and advantages of the invention will be apparent from the following more particular description of preferred embodiments of the invention, as illustrated in the accompanying drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1A presents a schematic diagram illustrating an electrical arrangement for electrodeposition of an alloy film with combined d-c and a-c currents.
FIG. 1B illustrates the net current curve for the electrical arrangement of FIG. 1A.
FIG. 1C illustrates the net voltage curve for the electrical arrangement of FIG. 1A.
FIG. 2 illustrates the Fe content and the rate of alloy deposition as a function of log f in low Ni concentration solutions of pH = 3.0 and pH = 4.6 for (Fe/Ni)sol = 20/80, with I 2 mA/cm+2 and I p = .iota.13.7 mA/cm+2.
FIG. 3 illustrates the Fe content and the rate of alloy deposition as a function of log f in high Ni concentration solution of pH = 3.0 and 4.6 for (Fe/Ni)sol = 5/95, Id-c = 2 mA/cm+2 , I p = 13.75 mA/cm+2.
FIG. 4 illustrates the Fe content and the rate of alloy deposition in high Ni concentration solution of pH = 3.8 with Id-c of 2 mA/cm+2 and 5 mA/cm+2 and Ia-c peak = 13.75 mA/cm+2.
FIG. 5 illustrates the log of the direct current density versus potential in high citrate solution of pH = 9.25.
FIG. 6 illustrates the Fe content and the rate of alloy deposition as a function of log f with Id-c of 2 and 4 mA/cm+2, Ia-c peak = 15 maA/cm+2 and (Fe/Ni)sol = 20/80.
FIG. 7 illustrates the Fe content and the rate of alloy deposition in low citrate solution of pH = 9.25 (Fe/Ni)sol = 20/80 with Id-c of 2 and 4 mA/cm+2, and Ipeak = 15 mA/cm+2.
FIG. 8 illustrates the Fe content as a function of direct current density for f = 0, 30 and 400 Hz with Ipeak = 13.75 mA/cm+2, (Fe/Ni)sol = 5/95, and pH = 3.8.
FIG. 9 illustrates the Fe content as a function of direct current density for f = 0, 30 and 400 Hz in high citrate solution of pH = 9.25, and (Fe/Ni)sol = 20/80 and Ipeak = 16.7 mA/cm+2.
FIG. 10 illustrates the Fe content as a function of the amplitude of a-c current in low Ni concentration, (Fe/Ni)sol = 20/80 and high Ni solution, (Fe/Ni)sol = 5/95 for f = 20 and 100 Hz, and pH = 4.6.
FIG. 11 illustrates the Fe content and the rate of alloy deposition as a function of log f in high citrate solution, with Id-c = 2 mA/cm+2, Ipeak = 15 mA/cm+2 for T = 25° C and 40°C.
FIG. 12 illustrates the rate of alloy Ni-Fe deposition as a function of ω-1/2 in low nickel solution, (Fe/Ni)sol = 20/80 and pH = 3,00, Id-c = 2 (mA/cm+2, Ipeak = 13.75 mA/cm+2.
FIG. 13 illustrates the rate of Fe deposition as a function of ω-1/2 in high Ni solution, (fe/Ni)sol = 5/95 at pH = 3, 38 and 4.6, Id-c = 2 mA/cm+2, Ipeak = 13.75 mA/cm+2.
FIG. 14 illustrates the rate of Fe deposition as a function of ω-1/2 at Id-c of 2 and 4 mA/cm+2 in low and high citrate solution.
FIG. 15 illustrates the rate of Ni deposition as a function of ω-1/2 at 2 and 4 mA/cm+2 in low and high citrate solutions.
FIG. 16 illustrates the Fe content as a function of film thickness in acid and alkaline solutions.
APPARATUS FOR THE INVENTION
Apparatus for electrodepositing an alloy film for the practice of this invention is presented schematically in FIG. 1A and the net current and voltage curves therefor are shown in FIGS. 1B and 1C, respectively.
In FIG. 1A the electrolytic cell 10 consists of two compartments 12 and 14. The working compartment 12 includes a vessel 16 electrolyte 18, horizontal working electrode 20 masked on one surface with insulating material 22, and a platinum mesh auxiliary electrode 24. The working electrode 20 and auxiliary electrode 24 are connected to the external electrical circuit 23 by means of conductors 21 and 25, respectively. The working compartment 12 is connected to the reference compartment 14 by means of a Luggin capillary 26. The reference compartment 14 includes an electrolyte 30 contained in a vessel 28. The reference electrode 32 is saturated Calomel Electrode suspended in electrolyte 30. Reference electrode 32 is connected to the external electrical circuit 23 by conductor 33. The electrical circuit 23 includes a d-c power supply 36 having positive and negative terminals 38 and 40. A signal generator 42 is provided to produce an a-c current which is superimposed on the d-c current. A by-pass capacitor 43 connected between terminals 38 and 14 provides a path for the a-c current.
The negative terminal 40 of the d-c power supply 36 is connected to the working electrode 20 through conductor 41, variable resistor 44, conductor 49 ampere meter 66, and conductor 21. The current through the circuit as a function of time is monitored by dual-beam oscilloscope 50 via terminals 54 and 55 which is connected across the variable resistor 44 at connections 46 and 48. The potential on the working electrode 20 with respect to the saturated Calomel Electrode 32 is measured by volt-meter 62 which is monitored as a function of time by oscilloscope 50 at connections 56 and 57. Oscilloscope 50 presents trace 51 as function of time on tube face 52 of either the current measured by ampere-meter 66 or the voltage measured by volt-meter 62 as selected.
THEORY OF THE INVENTION
The effect of a-c current of variable frequency and amplitude on the composition and uniformity of electrodeposited Ni-Fe alloys will now be considered. Cases 1, 2 and 3 will be examined for the ways in which superimposed alternating current can affect the electrodeposition process.
CASE 1
In Case 1 for some portion of each cycle, the a-c component converts the electrode from cathode to anode as reported in the article by A. Brenner, "Electrodeposition of Alloys", Vol. 1, Academic Press, New York, Page 84 (1963 ).
During electrodeposition of most metals, discharge of H3 O++ or H2 O occurs concurrently with pH changes in the vicinity of the electrode surface. For metals with small hydrolysis constants, this alkalization will be reflected in the formation of metallic hydroxides, which subsequently can be incorporated into the deposit, thus causing non-uniformity. In the codeposition of two or more metals this phenomenon can cause preferential deposition of one metal. Further, a concentration gradient across the deposit thickness, which will be the most pronounced in the first 500A, is reported in the article by H. Dahms et al., J. Electrochem. Soc., Vol. 112, No. 8, 1965.
If a-c current is superimposed on d-c current during electrodeposition of such metals, during the time for which the electrode is the anode, oxidation of adsorbed hydrogen formed in the cathodic cycle will take place according to the reaction:
H+ + e- anod. H ads
In the ideal case of balancing the rate of cathodic discharge of H+ ions with its rates of oxidation and diffusion from solution, control of pH can be achieved on the surface such that pH (surface) ➝ pH(bulk). Hence, the above-mentioned difficulties should be minimized if not completely eliminated. Mathematically, this presents a complex problem. However, experimentally the condition can easily be found where there is no preferential deposition of one of the metals and where there is no composition gradient in the deposit; i.e., the condition of constant pH.
Case 2
In Case 2 the current is controlled by ionic diffusion in the electrolyte.
Passage of either direct or alternating current through an electrolytic cell will produce concentration changes, which are susceptible to mathematical treatment. Sand, as reported in the article, Phil. Mag., Vol. 1, Page 45 (1901), solved the diffusion equation for the case of electrolysis with a constant direct current. Further, E. Warburg and F. Kruger solved the diffusion equation for the case of sinusoidal alternating current as reported in the respective articles Wied. Ann., Vol. 67, Page 493 (1899); and J. Phys. Chem., Vol. 45, Page 1 (1903 ). Both treatments start from Fick's second law:
δc/δt = D (δ2 c)/(δX2)
where c is the concentration of one ionic species, D is its diffusion coefficient and x is the distance from the electrode into the solution.
Both d-c and a-c currents have the same boundary conditions; namely,
cx=0 = cx=∞ for t = 0
and
δc/δx = 0 for x ➝ ∞ and t > 0
Here, Cx=0 is the concentration at the electrode surface and cx=∞ is the bulk concentration.
The solution of Equation 1 for constant d-c current is:
c(o,t) = c∞ -2i/nF √ t/πD (2)
where i is the current density, t is time, n is the number of electrons involved in the electrode reaction and F is Faraday's constant. For steady state conditions
c0 - c∞ = - iδ/nfD (3)
or
ΔC = Kiδ
where δ is the thickness of the diffuse layer and K includes all constant terms.
For sinusoidal a-c current the solution of Equation 1 is:
where I is the amplitude of the current density, and ω = 2πf where f is the a-c current frequency.
At the electrode surface where x = 0, Equation 4 becomes
or
where A/√ω is the amplitude of concentration wave.
If both currents act simultaneously on the system, the net concentration changes can be obtained by adding together the concentration change that would be produced by each current taken separately (since the sum of a number of solutions of a linear differential equation is likewise a solution) as reported in the article by T. R. Roseburg et al., J. Phys. Chem., Vol. 14, Page 816 (1910). Thus:
δC = -[iδ + I √D/ω sin (ωt - π/4] /nFD 6
Consider electrodeposition of a binary alloy with one of the depositable metal ions under diffusion control and the other under charge transfer control. In such a case Equation 6 is applicable to only one constituent of the alloy and the other constituent will be deposited
as if the a-c current were not present, since a-c current does not effect charge transfer reactions.
The conditions will now be examined under which a-c current and d-c current have comparable affects on concentration change of Fe, which is deposited under diffusion control. For a 10-3 M Fe solution and a total direct current density of 2 mA/cm+2, the partial current for discharge of Fe is found to be 0.32 mA/cm+2, which from Equation 3 gives δc = 8.5 × 10-6 m cm+3. . Consider I is taken to be 15 mA/cm+2, the amplitude of the concentration wave from Equation 5 is 2.6 × 10-6 and 0.37 × 10-6 M cm-3 for a frequency of 20 and 1,000 Hz, respectively; i.e., a-c current of low frequency produces 30 percent and of high frequency produces 4 percent of the total concentration change. If the d-c current is increased, the affect of a-c current becomes even smaller (2 percent for 1,000 Hz and Id-c of 4 mA/cm+2). Clearly, the effect of diffusion becomes progressively smaller with increasing frequency. Theoretically, in accordance with Equation 5, the effect of a-c current can be increased by increasing its amplitude. Practically, it is not desirable to go too high into the anodic region, where dissolution of the alloy and oxide formation can take place.
CASE 3
In Case 3 deposition at the electrode is preceded by a chemical reaction in the solution.
If electrodeposition is carried out from a solution of complex ions, a reduction to the metallic state can take place either directly from the complex ion or this electrochemical step can be preceded by a chemical step or several steps in series.
If electrochemical reduction is preceded by a homogeneous chemical reaction of a type
(z1 +z2) kf z1 z2
Mm Cn mM + nC
then the rate of formation of the metallic ions is
v = kf CMC - kb (cM)m (cC)n (7)
where kf is the rate constant for dissociation of the complex and kb the rate constant for the recombination, and cMC, cM, cC are the concentrations of metallic complex, metal ion and complexing agent, respectively. Equation (7) can be written as
v = vo - kcMp (8)
where vo is the reaction exchange rate, k = kb cC is the reaction rate constant, and p is the reaction order.
As a result of diffusion and chemical reaction the change of concentration with time and distance at the surface of electrode can be represented by Fick's second law in extended form:
(δc)/(δt) = D(δ2 c)/(δx2) + v
Equation 9 applies to both direct and alternating currents. The direct current due to the deposition of metal with a slow chemical step and p = 1 is:
where cM is equilibrium concentration of metal ions determined by cM = KcMC /cC, K being the stability constant for a given complex, as reported by H. Gerescher et al., Z Physik Chem., Vol. 197, Page 92 (1951). When the concentration of metal ions at the surface, cs, becomes zero, a limiting reaction current ir is reached, given by
ir = - nF √vo cM D (11)
and from its value vo and k can be calculated (since at equilibrium v = 0 and vo becomes equal to k. cM). The reaction exchange rate is also related to the thickness of the reaction layer, δi, by the following equation:
δi = √DcM /vo (12)
Passage of a-c through a system where chemical reaction occurs prior to charge transfer will produce concentration changes which depend not only on ω-1/2 but also on k.
K. J. Vetter, as reported in the book "Electrochemical Kinetics", Academic Press, New York, Page 253 (1967), gives the concentration change as a difference of ohmic and capacitive components of the electrolyte, both of which are function of ω-1/2 and k/ω. The concentration wavelength as well as penetration depth are also dependent upon the same parameters. This derivation is valid only for very small differences between cM and cs. Further, the a-c and d-c solutions of the differential equation cannot be added in this case, since the differential equation is non-linear. Therefore, quantitative treatment has not been attempted. Qualitatively, it is expected that at a low frequency the concentration wave will be able to follow the slowly varying current, and that the penetration depth would be of the same length as d-c reaction layer thickness. At higher frequencies, the formation and decomposition of metal complexes will be increasingly less important, since they cannot follow fast changes of current. In addition the penetration depth of the concentration wave will become smaller. For both these reasons, it is to be expected that at high frequencies the d-c current behavior will dominate.
When two or more metallic complexes are present in the system, a-c current will affect them differently depending upon the value of k/ωfor each complex. Hence, in accordance with the principles of this invention, by superimposing a-c current on d-c current, the deposition kinetics of alloys can be affected in a practical way.
PRACTICE OF THE INVENTION
Measurements were performed with two compartment cells as shown in FIG. 1A. The cathode 20 was Cu-sheet or evaporated Ag on glass (2 × 2 cm), placed horizontally in one compartment 12 of the cell. The back of the electrode was masked by mask 22 so that electrodeposition was carried out on one side only. A Pt-mesh auxiliary electrode 24 was placed approximately 2 cm above the working electrode 20. The reference containing electrode 9 compartment 14 saturated Calomel electrode was connected with the main compartment 12 through a Luggin capillary 26 carefully bent to avoid any shielding effect.
The conventional electrical circuit is shown in FIG. 1A. Current time and potential time curves, FIGS. 1B and 1C, respectively, were simultaneously recorded on a dual-beam oscilloscope 50. It is important that the potential is recorded, since this provides a way of determining the conditions under which the oxidation of hydrogen takes place by an electrochemical mechanism which minimizes dissolution of alloy and avoids its oxidation.
Measurements were carried out in acid and alkaline solutions. The acid solutions had the following compositions: "Low Ni": 0.024 M NiSO4, 0.006 M FeSO4, 0.035 M NaKC4 H4 O6, pH = 3 or 4.6. The molar ratio of (Fe/Ni) in solution was 20/80. "High Ni": had composition as above for "Low Ni", but with 0.114 M NiSO4 and pH = 3, 3.8 or 4.6. The (Fe/Ni) ratio in solution was 5/95. The alkaline solutions were ammoniacal-citrate solutions, the compositions of which were: "High citrate": 0.125 M NiCO3, 0.032 M Fe dust, 0.301 M C6 H8 O7, 0.332 M (NH4)2 M C6 H5 O7 and NH4 OH for pH = 9.25. The "low citrate" solution had the same pH and concentration of Ni and Fe but it contained 0.127 M C6 H8 O7 and 0.137 M (NH4)3 H C6 H5 O7. The molar ratio of (Fe/Ni) in solution was 20/80.
The solutions were made of reagent grade chemicals and deionized water. The citrate solutions were prepared according to British Pat. No. 925,144.
After electroplating, the samples were cut into 1.5 × 1.5 cm squares and analyzed by the x-ray fluorescence technique for wt. % Fe (accuracy ± 1 wt. %) and thickness (accuracy ± 150A.).
The effect of frequency on the rate of deposition and on the composition of the deposited alloy were examined. In the FIGS. 2 and 3 the rate and percent Fe are shown as a function of log frequency in low and high nickel solutions, respectively, for conditions of constant pH, Id-C and Ipeak. On the left hand sides are given values for direct current plating only.
In accordance with the theory of this invention, a diminishing effect of a-c current with increasing d-c current in the system is expected. This prediction is clearly validated by FIG. 4. With Id-c of 2 mA/cm+2, the Fe content varies from 9.5 to 30 percent, but changes only from 15.2 to 18 percent with Id-c of 5 mA/cm+2 at constant Ipeak and pH = 3.8.
FIG. 5 shows a log current vs. voltage plot for the high citrate solution. It can be seen that for high values of total current, IFe reaches a limiting value, which is taken as the limiting reaction current according to Vetter's criteria as set forth hereinbefore in the Theory of the Invention section. In FIGS. 6 and 7 the deposition rates and percent Fe are given as a function of log frequency for two values of direct current density.
The variation in composition with the density of direct current at constant frequency and amplitude of alternating current is given in FIGS. 8 and 9 for acid and alkaline solutions, respectively. For the purpose of comparison, data for d-c current plating alone are also given and designated as f = 0.
From Equation 5 the amplitude of the diffusion concentration wave is expected to increase with increasing Ipeak, and that the iron content of both the surface electrolyte and the deposit should decrease. This is validated in FIG. 10.
Since temperature affects the equilibrium constant for the dissociation of metallic complexes, it can be expected to exert an influence on the deposition rate. In FIG. 11 deposition rates and percent Fe are given for the high citrate solution as a function of frequency for temperatures of 25° and 40°C. At higher temperatures the corrosion rate of the alloy becomes too large for meaningful study.
It is validated in FIGS. 2, 3, 4, 6, 7 and 11 that to a large extent the composition of the deposit is influenced by frequency. Since the percentage of one metal is a function both of its deposition rate and of the total rate of metal deposition, it is more meaningful to examine how the iron rate alone varies with frequency. The diffusion law predicts a linear dependence upon ω-1/2, e.g., Equation 6. From the plots given in FIGS. 12 to 15 it can be seen that the rate of Fe deposition is linearly dependent upon ω-1/2, approaching its d-c current value at high frequencies, where the contribution from the a-c component becomes negligible. However, there are two regions, one being that of low frequency, i.e., 20 to 100 Hz, and the other from 100 to 1,000 HZ for which the slope of the line has different values, being smaller at lower frequencies. The explanation of this behavior is discussed separately below for the two different types of solutions employed.
DEPOSITION FROM ACID SOLUTIONS
In the solution of pH = 3, the rate of alloy deposition is lower under a-c current plus d-c current, than under d-c current alone. This indicates that some dissolution of alloy is taking place. It might be argued that Fe dissolves faster than Ni, and that there is less Fe present in a deposit. However, there is no trend in the variation of alloy deposition rate with frequency. Further, in the solutions of pH = 4.6, the total rate is not affected by a-c current, but the Fe rate is lower and shows the same clearly defined two regions of different dependence on frequency. In the region of low frequency the contribution of a-c current is two-fold. Firstly, its affect on diffusion is the largest, and secondly there is an effect on the surface pH. When the potential is varied slowly, the electrode remains in the anodic region sufficiently long to allow oxidation of adsorbed hydrogen on its surface. Hence, pHsurface is brought back to its original value for the next cathodic cycle. If pHsurface does not increase, the formation of hydroxides does not occur, and there is no anomalous deposition of iron and the Ni deposition is not suppressed. This can be clearly seen from FIG. 12. With increasing frequency, the electrode spends less and less time in the anodic region, and the kinetic processes apparently cannot follow such rapid changes. As a result, pHsurface increases sufficiently to cause the formation of iron hydroxide, which prevents the discharge of Ni. At 100 Hz, Ni and Fe deposit with the same rate as shown in FIG. 12, even though the bulk concentration of Ni is four times higher than that of Fe. Above 100 Hz, Fe deposits with a higher rate than Ni. In FIG. 13 the rates of Fe deposition are shown for three values of bulk pH. Within the experimental error, Fe deposits from the solutions of pH = 3.8 and pH = 4.6 with the same rate, indicating that a-c current produced the same surface pH.
With increasing bulk pH, or by increasing the d-c current level, the a-c current component becomes less effective in controlling the pH of the surface as shown in FIGS. 2, 3, and 4. According to Bockris et al., as reported in the article in Electrochemistry Acta, vol. 4, page 325 (1961), the Fe rate is closely connected with pH through the relationship (δln iFe /δ log cOH -) = 1.
By examining FIG. 2, it can be seen that in the low Ni solution of pH = 4.6, the rate of alloy deposition is higher under a-c current plus d-c current than under d-c current alone. If adsorbed hydroxides block the surface, a hydrogen evolution reaction from the rather dilute bath might be kinetically the most favorable reaction. With a-c current present, adsorption of hydroxides doe not occur, and the rate is higher.
DEPOSITION FROM COMPLEX SOLUTIONS
Deposition of the alloys from complex solutions with superimposed a-c current on d-c current is interesting on account of the dependence on the k/ω ratio. Further, in such systems the two currents are more comparable since the reaction layer thickness for d-c current is approximately the same as the penetration depth for a-c current (approximately 7.5 . 10-5 cm). It can be seen from FIGS. 6 and 7 that by varying frequency alone the Fe content can be varied from 14 to 59 percent, or, by decreasing the concentration of complexing agent for Fe, from 8 to 63 percent.
In FIG. 14 rates of Fe deposition are given as a function of ω-1/2 for two values of d-c current and two concentrations of complexes of citrate ions. At Id-c = mA cm-2 a quite surprising effect is found, namely, Fe deposits with a higher rate from the solution containing more of its complexing agent. When Id-c is increased to 4 mA cm-2, Fe deposits with the same rate from both citrate solutions in low frequency region. However, at higher frequencies the situation becomes "normal", i.e., with more complexing agent less Fe ions are available for deposition. This "abnormality" can be explained if tee values of the rate constant are compared for low and high citrate solution.
The reaction exchange rate, vo, can be calculated from Equation 11 if the limiting reaction current, ir, is determined experimentally. For the high citrate solution ir = 1.54 mA cm-2, giving vo = 2.06 . 10-4 . For the low citrate solution, ir = 2.02 mA cm-2 and vo = 1.37 . 10-4 mole cm-3 sec-1. From these vo values, k is calculated to be 4.63 . 103 and 1.19 . 103 sec-1 for high and low citrate, respectively. The rate depends not only on ω-1/2 but also on the ratio of k to ω. This ratio varies from 37 to 0.74 in the high citrate solution, but only from 9.5 to 0.19 in the low citrate, when f is varied from 20 to 1,000 Hz. The rate constant is equal to ω at 740 Hz and 190 Hz for high and low citrate, respectively. Since k is an order of magnitude larger than ω at low frequencies in the high citrate solution, Fe deposits with a higher rate than from low citrate solution where k and ω are of the same order of magnitude. At 1,000 Hz the ratio of k/ω in both solutions are of same magnitude, i.e., 0.74 and 0.19, and there is very little difference in Fe rates as shown on the left side of FIG. 15.
By increasing Id-c, more material is required according to Faraday's law, and the effect of d-c current becomes more pronounced. When the frequency is increased, the effect of a-c current is still further diminished, and the transition to "normal" behavior is observed.
If the concentration of citrate ions is changed, changes are not expected in Ni rate, since Ni is present in solution as the [Ni (NH3)n ]++ complex. The data given in FIG. 15 supports the expectation.
By superimposing a-c current on d-c current it is expected, in accordance with the principles of this invention, that variation in Fe composition on the surface and consequently in the deposit will take place within the time of one cycle, i.e., approximately 10-2 sec. On a microscopic scale this means uniform composition, which is observed in practice of this invention, as shown in FIG. 16. The line at the bottom of the graph represents % Fe obtained from the solution with molar ratio of Fe/Ni = 5/95. The composition of the solution is reflected exactly in the deposit throughout its thickness.
EXAMPLES OF THE INVENTION
Alloys of 80-20 Ni-Fe are obtained by the practice of this invention from solutions having the parameters identified below:
(a)
NiCO3 = 16.2 g/l (45%Ni)
Fe dust = 1.78 g/l
Citric acid = 63.3 g/l
Nh4 -citrate = 75.5 g/l
pH = 9.25 at approximately 25°C
id-c = 2 ma/cm2
Ip = 15 ma/cm2
f = 60 Hz
Rate = 380A/min.
(b)
NiCO3 = 16.2 g/l (45%Ni)
Fe dust = 1.78 g/l
Citric acid = 63.3 g/l
Nh4 -citrate = 75.5 g/l
pH = 9.25 at approximately 25°C
id-c = 4 ma/cm2
Ip = 15 ma/cm2
f = 20 Hz
Rate = 380A/min.
(c)
NiCO3 = 16.2 g/l
Fe dust = 1.78 g/l
Citric acid = 26.6 g/l
Nh4 -citrate = 31.0 g/l
pH = 9.25 at approximately 25° C
id-c = 2 ma/cm2
Ip = 15 ma/cm2
f = 100 Hz
Rate = 100A/min.
(d)
NiSO4 . 6H2 O = 6.3 g/l
FeSO4 . 7H2 O = 1.7 g/l
NaK-tartrate = 10.0 g/l
pH = 3.0 at approximately 25°C
Id-c = 2 ma/cm2
Ip = 13.7 ma/cms
f ' 25 Hz
Rate = 22A/min.
(e)
NiSO4 . 6H2 O = 30.0 g/l
FeSO4 . 7H2 O = 1.7 g/l
NaK-tartrate = 10.0 g/l
pH = 3.0 at approximately 25°C
id-c = 2 ma/cm2
Ip = 13.7 ma/cm2
f = 100 Hz
Rate = 125A/min.