Copper deposition limiting current

Fig. 1. Experimental limiting-current curve for copper deposition from an acidified cupric solution. [From Hickman (H3).]...

Comparison of Interfacial Density Differences Resulting from the Deposition of Copper, the Cathodic Reduction of Ferricyanide, and the Anodic Oxidation of Ferrocyanide at the Limiting Current Density (7" = 25°C)... 

Therefore, criteria in the selection of an electrode reaction for mass-transfer studies are (1) sufficient difference between the standard electrode potential of the reaction that serves as a source or sink for mass transport and that of the succeeding reaction (e.g., hydrogen evolution following copper deposition in acidified solution), and (2) a sufficiently low surface overpotential and rate of increase of surface overpotential with current density, so that, as the current is increased, the potential will not reach the level required by the succeeding electrode process (e.g., H2 evolution) before the development of the limiting-current plateau is complete. 

Figure 3a is an illustration of the effect of surface overpotential on the limiting-current plateau, in the case of copper deposition from an acidified solution at a rotating-disk electrode. The solid curves are calculated limiting currents for various values of the exchange current density, expressed as ratios to the limiting-current density. Here the surface overpotential is related to the current density by the Erdey Gruz-Volmer-Butler equation (V4) ... 

Fig. 3. (a) Typical galvanostatic limiting-current curve for copper deposition at a copper disk in acidified CuS04 solution. The circles indicate the experimental curve. The solid curves were calculated using kinetic parameters as indicated, (b) Typical galvanostatic limiting current curve for ferricyanide reduction at a nickel electrode in equimolar ferri ferrocyanide solution with excess NaOH. [From Selman (S8).]... 

It is clear from the calculated limiting-current curves in Fig. 3a that the plateau of the copper deposition reaction at a moderate limiting-current level like 50 mA cm 2 is narrowed drastically by the surface overpotential. On the other hand, the surface overpotential is small for reduction of ferri-cyanide ion at a nickel or platinum electrode (Fig. 3b). At noble-metal electrodes in well-supported solutions, the exchange current density appears to be well above 0.5 A/cm2 (Tla, S20b, D6b, A3e). At various types of carbon, the exchange current density is appreciably smaller (Tla, S17a, S17b). 

From an analysis of the electrochemical mass-transfer process in well-supported solutions (N8a), it becomes evident that the use of the molecular diffusivity, for example, of CuS04, is not appropriate in investigations of mass transfer by the limiting-current method if use is made of the copper deposition reaction in acidified solution. To correlate the results in terms of the dimensionless numbers, Sc, Gr, and Sh, the diffusivity of the reacting ion must be used. 

As shown above, potentiodynamic generation of limiting currents is more rapid and, therefore, preferable in principle to the galvanodynamic technique. However, during a linear decrease of potential to the limiting-current condition, the current density initially rises more rapidly than when a current ramp is used. Therefore, in the case of copper deposition at the cathode, a linear potential ramp tends to yield a rougher deposit and a less well-defined plateau than a linear current ramp (see Section III,F). 

In many cases mass transfer is not the sole cause of unsteady-state limiting currents, observed when a fast current ramp is imposed on an elongated electrode. In copper deposition, in particular, as a result of the appreciable surface overpotential (see Section III,C) and the ohmic potential drop between electrodes, the current distribution below the limiting current is very different from that at the true steady-state limiting current. 

Limiting currents measured for a deposition reaction may be excessively high due to surface roughness formation near the limiting current. Rough deposits in the case of copper deposition have been mentioned several times in previous sections, since this reaction is one commonly used in limiting-current measurements. However, many other metals form dendritic or powdery deposits under limiting-current conditions, for example, zinc (N lb) and silver. Processes of electrolytic metal powder formation have been reviewed by Ibl (12). 

The models incorporate two microscopic parameters, the site density and the critical nucleus size. A fit of experimental current transients to the models allows conclusions, for example, concerning the effect of additives on nucleation rate. Fabricus et al. found by analysis of current transients that thiourea increases the nucleation density of copper deposited on glassy carbon at low concentration, but decreases it at higher concentration [112], Schmidt et al. found that Gold nucleation on pyrolytic graphite is limited by the availability of nucleation sites [113], Nucleation density and rate were found to depend on applied potential as was the critical nucleus size. Depending on concentration, critical nuclei as small as one atom have been estimated from current transient measurements. Michailova et al. found a critical nucleus of 11 atoms for copper nucleation on platinum [114], These numbers are typical, and they are comparable to the thermodynamic critical radii [86],... 

Chiba [43] has found that the electrodeposition of copper from a cupric-EDTA bath, in the presence of ultrasound, gave increased limiting current densities and increased cathodic efficiences while at the same time reducing the grain size. Walker [44,45] has shown that deposits obtained from sulphate baths in the presence of ultrasound show increased hardness. 

Drake [35] has measured the thickness of the diffusion layer during the electrodeposition of copper from an acidic-sulphate system. He obtained a value of approximately 200 pm for the silent system and values of 34 pm and 3.4 pm for ultrasonic frequencies of 1.2 MHz and 20 kHz respectively. The corresponding values of the limiting-current density were from 8 Am (silent) to 50 A m (1.2 MHz) to 500 A m (20 kHz), indicating a significant increase in the rate of deposition. 

As a result of that reductive process, a deposit of copper metal (denoted in Eq. 2.2 by s for solid ) is formed on the carbon electrode surface. The prominent anodic peak recorded in the reverse scan corresponds to the oxidative dissolution of the deposit of copper metal previously formed. The reason for the very intense anodic peak current is that the copper deposit is dissolved in a very small time range (i.e., potential range) because, in the dissolution of the thin copper layer, practically no diffusion limitations are involved, whereas in the deposition process (i.e., the cathodic peak), the copper ions have to diffuse through the expanding diffusion layer from the solution to the electrode surface. These processes, labeled as stripping processes, are typical of electrochemically deposited metals such as cadmium, copper, lead, mercury, zinc, etc., and are used for trace analysis in solution [84]. Remarkably, the peak profile is rather symmetrical because no solution-like diffusive behavior is observed. 

This assumption is based on three relevant indications. First, this wave results in a limiting-current. This means that steady-state transport phenomena control the rate of this reaction, which is not compatible with a possible oxidation of metallic copper to Cu(I) or Cu(II). If the latter were to be valid, a peak-shaped response should have been obtained because of the limited available amount of metallic copper (initially deposited by reduction of Cu(II) or Cu(I) in the reduction wave). In addition, the second voltammetric oxidation wave in the backward scan direction is actually compatible with such a dissolution reaction. 

Figure 6.7 Influence of pulse parameters on deposit morphology for copper deposition from a copper sulfete/sulfuric acid electrolyte [6.102]. p pulse current density ipj limiting pulse current density i average current density jj limiting current density under dc conditions.

Electroless metal deposition at trace levels in the solution is an important factor affecting silicon wafer cleaning. The deposition rate of most metals at trace levels depends mainly on the metal concentration and some may also depend on the interaction with other species as well. For copper the deposition rate at trace levels in HF solutions is different for n and p types. It depends on illumination for p-Si but not for n-Si. It is also different in HF and BHF solutions. In a HF solution the deposition process is controlled by both the supply of minority carriers and the kinetics of cathodic reactions. Thus, a high deposition rate occurs on p-Si only when both and illumination are present. In the BHF solution, the corrosion process is limited by the supply of electrons for p-Si whereas for n-Si it is limited by the dissolution of silicon because the reaction rate is indepaidmt of concentration and illumination. The amount of copper deposition does not correlate with the corrosion current density, which may be attributed to the chemical reactions associated with hydrogen reduction. More information on trace metal deposition can be found in Chapters 2 and 7. 

The polarization curves for copper deposition on the electrodes whose surfaces are shown in Fig. 24a and 26a are given in Fig. 27. It is obvious that the noticeable increase of the exchange current density attained by the application of the PO regime ( /o.eff = 3.3 mAcm-2 /r = 23.5) is followed by the minimal increase of limiting diffusion current density, relative to the one corresponding to the substrate from Fig. 24a. 

---