Controlled Wettability Graphite Electrodes for Selective Use in Electrolysis Cells

This invention relates to the production of metal such as aluminum from metal chloride dissolved in molten halide solvent bath by electrolyzing the bath in a monopolar or bipolar cell. More particularly, the invention relates to graphite electrodes used in such cells and to selective use thereof with respect to their wetting or non-wetting characteristics so as to prolong useful electrode life in such cells and to controlled methods of graphite electrode manufacture to achieve the desired wetting or non-wetting characteristics for such selective use.

One type of electrolytic cell used in the production of metal, such as aluminum, from metal chloride dissolved in a solvent salt bath includes a terminal anode, at least one intermediate bipolar electrode and a terminal cathode. These electrodes are typically situated in relatively closely spaced, generally parallel relationship wherein opposed anode-cathode faces provide interelectrode spaces through which the molten bath can move and be electrolyzed by passage of current from anode to cathode. Electrolysis of the metal chloride occurring within the interelectrode space results in molten metal depositing at the cathode and chlorine gas collecting at the anode. Cells of this type are described in U.S. Pat. Nos. 3,755,099 and 3,822,195, incorporated herein by reference. One of the important features of these cells is that the anode-to-cathode space or distance should be carefully maintained at a preselected level in order to achieve the high current efficiency and lower power consumption capabilities of the bipolar chloride electrolysis process. Obviously, any amount of wear occurring on either the anode or the cathode surface, as by erosion or other removal of electrode material, tends to increase the distance and, accordingly, increase the electrical resistance across the distance between anode and cathode. For the most part, the anode presents little problem since under most conditions chlorine is relatively non-corrosive to the carbonaceous materials employed for electrodes. However, experience has shown that some amount of electrode wear does occur on the cathode surface, and considerable effort has been expended to reducing or relieving this wear condition. Excessive cathode surface wear is a problem, not only in increasing power consumption as just explained, but can increase the resistance so much that the cell is considered uneconomical to operate, thus necessitating a costly shut-down, repair or replacement of the electrodes, and restarting the cell. In addition to the electrical resistance problems resulting from cathode wear, the carbonaceous material removed from the cathode surface can contaminate the bath. This alone can reach such an extreme as to necessitate shutting down the cell.

In accordance with the invention, it has been discovered that graphite electrode surfaces can exhibit either wetting or non-wetting behavior with respect to the metal deposited at the cathode, and that such behavior can be utilized in association with bath flow velocity and anode-cathode distance to minimize cathode surface wear. It has further been discovered that the wettability or the non-wettability of graphite electrodes can be established by carefully controlling the graphite manufacturing process.

Accordingly, it is an object of the present invention to provide for decreased cathode electrode wear in halide electrolytic cells used in producing metal such as aluminum from metal chlorides.

Another object is to provide a means for selectively positioning graphite cathode material based on its wetting characteristics so as to balance such with other cell operating conditions to minimize cathode wear.

Another object is to provide for selectively controlling the wetting characteristics of graphite electrode material by controlling the steps in manufacturing the graphite.

These and other objects will be apparent from the drawing, specification and claims appended hereto.

In accordance with the invention, it has been found that graphite cathode surface wear is reduced if the cathode surface is selected and controlled with respect to its wettability and with respect to bath flow rate over the cathode surface. Cathode graphite surfaces wetted by the metal deposited from the bath are used when the bath is moving over the cathode at a relatively low velocity. However, graphite cathode surfaces which are not wetted are used in regions of high velocity bath flow. It is to be appreciated that in electrolytic cells of the type here concerned, bath flow velocity can vary from cell to cell and within a single cell. Thus, in some electrolytic cells the bath flow velocity through the anode-cathode interelectrode space is relatively slow and in others it is more rapid. Moreover, there are cells which include regions wherein each effect occurs. In general, in the electrolytic cells of the type depicted herein and in the patents above referred to featuring one or more horizontal bipolar electrodes between an upper terminal anode and a lower terminal cathode providing more or less horizontal bath flow therebetween, it is difficult to avoid the occurrence of both fast and slow regions. In these cells a more rapid interelectrode bath flow velocity can occur in the upper interelectrode spaces and a lower flow velocity can occur in lower interelectrode spaces. Hence, one practice of the invention includes in single electrolytic cell the use of non-wettable cathode surfaces in regions of the cell where the higher flow rates occur, typically regions higher or further away from the terminal cathode and the use of wettable cathode surfaces in regions where low flow rates occur, typically regions lower or closer to the terminal cathode.

FIG. 1 is a sectional elevation illustrative of a cell for producing aluminum or other metal in accordance with the invention.

FIG. 2 is a schematic sectional elevation of an electrolytic cell useful in practicing the invention.

FIG. 3 is a schematic plan view of a cell of the type shown in FIG. 2.

A suitable cell structure for producing metal in accordance with the invention is illustrated in FIG. 1. The cell illustrated includes an outer steel shell 1, which is lined with refractory sidewall and end wall brick 3, made of thermally insulating, electrically non-conductive material which is resistant to molten alkali metal and metal chloride-containing halide bath and the decomposition products thereof. The cell cavity includes a sump 4 in the lower portion for collecting the metal produced. The sump bottom 5 and walls 6 are preferaly made of graphite. The cell cavity also accommodates a bath reservoir 7 in its upper zone. The cell is enclosed by a refractory roof 8, and a lid 9. A first port 10, extending through the lid 9 and roof 8, provides for insertion of a vacuum tapping tube down into sump 4, through an internal passage to be described later, for removing molten metal from the sump. A second port 11 provides inlet means for feeding the metal chloride into the bath. A third port 12 provides outlet means for venting chlorine.

Within the cell cavity are a plurality of plate-like electrodes which include an upper terminal anode 14, desirably an appreciable number of bipolar electrodes 15 (four being shown), and a lower terminal cathode 16, all of graphite. These electrodes are shown arranged in superimposed relation, with each electrode preferably being horizontally disposed within a vertical stack. Sloping or vertically disposed electrodes can also be employed, however, in either monopolar or bipolar electrode cell arrangements. In FIG. 1, the cathode 16 is supported at each end on sump walls 6. The remaining electrodes are stacked one above the other in a spaced relationship established by interposed refractory pillars 18. Such pillars 18 are sized to closely space the electrodes, as for example to space them with their opposed surfaces separated by 3/4 inch or less. In the illustrated embodiment, five interelectrode spaces 19 are provided between opposed electrodes, one between terminal cathode 16 and the lowest of the bipolar electrodes 15, three between successive pairs of intermediate bipolar electrodes 15, and one between the highest of the bipolar electrodes 15 and terminal anode 14. Each interelectrode space 19 is bounded by an upper surface 20 provided by the bottom of one electrode (which surface 20 functions as an anode surface) opposite a lower surface 21 provided by the top of another electrode (which surface 21 functions as a cathode surface). The spacing between anode and cathode surfaces is the anode-cathode distance in the absence of a metal layer of substantial thickness. When a layer of metal is present on the cathode surface, the effective anode-cathode distance is shorter than the distance between the graphite electrode surfaces 20 and 21. The bath level in the cell will vary in operation but normally will lie well above the anode 14, thus filling all otherwise unoccupied space therebelow within the cell.

Anode 14 has a plurality of electrode bars 24 inserted therein which serve as positive current leads, and cathode 16 has a plurality of collector bars 26 inserted therein which serve as negative current leads. The bars 24 and 26 extend through the cell wall and are suitably insulated from the steel shell 1. A suitable voltage is imposed across the terminal anode 14 and the terminal cathode 16, and this imparts the bipolar character to bipolar electrodes 15.

As indicated earlier, the sump 4 is adapted to contain bath and molten metal, and the latter may accumulate beneath the bath in the sump, during operation. Should it be desired to separately heat the bath and any metal in sump 4, an auxiliary heating circuit may be established therein.

A bath supply passage indicated by arrow 30 generally extends from the upper reservoir 7 down along the right-hand side (as viewed in FIG. 1) of the electrodes and into each interelectrode space 19. Thus, each of the interelectrode spaces 19 is supplied with a continual supply of the molten bath which travels across each interelectrode space 19 (moving right to left in FIG. 1) and exits the interelectrode space 19 turning upwardly as generally indicated by arrows 34 and 35.

The electrolyte employed for producing aluminum in accordance with the present invention typically comprises a molten salt bath composed essentially of aluminum chloride dissolved in one or more halides, particularly chlorides, of higher decomposition potential than aluminum chloride. By electrolysis of such a bath, chlorine is produced on the anode surfaces and aluminum on the cathode surfaces of the cell electrodes. The metal is conveniently separated by settling from the lighter bath, and the chlorine rises to be vented from the cell. In such practice of the present invention, the molten bath may be positively circulated through the cell by the buoyant gas lift effect of the internally produced chlorine gas, and aluminum chloride is periodically or continuously introduced into the bath to maintain the desired concentration thereof.

The bath composition, in addition to the dissolved aluminum chloride, will usually be made up of alkali metal chloride, although, other alkali metal halide and alkaline earth halide, may also be employed. A presently preferred aluminum chloride containing composition comprises an alkali metal chloride base composition made up of about 50-75 percent by weight sodium chloride and 25-50 percent lithium chloride. Aluminum chloride is dissolved in such halide composition to provide a bath from which aluminum may be produced by electrolysis, and an aluminum chloride content of about 1.kappa. to 10 percent by weight of the bath is generally desirable. As an example, a bath analysis as follows (in percent by weight) is satisfactory: 53 percent NaCl, 40 percent LiCl, 0.5 percent MgCl.sub.2, 0.5 percent KCl, 1 percent CaCl.sub.2, and 5 percent AlCl.sub.3. In such bath, the chlorides other than NaCl, LiCl and AlCl.sub.3 may be regarded as incidental components or impurities. The bath is employed in molten condition, usually at a temperature above that of molten aluminum and in the range between 660.degree. and 730.degree. C., typically at about 700.degree. C.

As described hereinabove, bath supplied from reservoir 7 through bath supply passage 30 is electrolyzed in each interelectrode space 19 to produce chlorine on each anode surface 20 and aluminum on each cathode surface 21. Electric current applied between the upper anode 14 and the bottom cathode 16 causes the interdisposed bipolar electrodes 15 to exhibit their bipolar behavior and effect electrolysis within each interelectrode space 19. The electrode current density can conveniently range from about 5 to 15 amperes per square inch, but preferred current density can vary from one particular cell to another and is readily determined by observation.

The chlorine produced at the anode is buoyant in the bath and its movement through the bath may be employed to effect bath circulation. That is, the chlorine rising up along the left side, when viewed in FIG. 1, of the cell creates a bath circulating effect including a sweeping of the bath through the interelectrode spaces 19. This sweeping action sweeps the aluminum produced on each cathode surface through an out of each interelectrode space 19 in the same direction as the bath, toward the left as viewed in FIG. 1, and permits the aluminum to then settle down into the sump 4.

As indicated hereinabove, the spacing between electrodes and the bath velocity through those spaces can vary from cell to cell and within a given cell. For the type of cell shown in U.S. Pat. No. 3,755,099, it will usually be found that the lower zones closer to the terminal cathodes 16 exhibit a lower bath velocity through the interelectrode spaces, whereas the higher zones closer to terminal anode 14 tend to exhibit higher bath flow rates through the interelectrode spaces 19.

In accordance with the invention, the wettability of a given graphite electrode material is readily determined by a test now described. Referring to FIGS. 2 and 3, there are schematically shown convenient arrangements for determining the wettability characteristics of electrode materials. In this type of arrangement, a small laboratory type electrolytic cell 200 has positioned therein an anode 314 together with two cathodes 316. The cathodes 316 may be identical or they may be different where it is desired to test two different electrode samples. Since the area of concern is the cathode surface, it is important that the surface 321 of the cathode 316 correspond to the cathode surface to be used in a production cell. That is, the cathode 316 should be taken from a larger electrode, or at least be representative of such material removed from a larger electrode, and be such that its surface 321 is representative of the cathode surface for the production electrode. It is also significant that the bath 213 contained within the cell 200 is preferably of substantially the same composition and temperature as anticipated in the production cell so as to minimize departures from production cell conditions.

A suitable size for the cathode blocks 316 is about 11/2 inches long by 5/8 inch thick by about 3/4 inch wide, and the cathodes are spaced from the anode 314 by a distance "d" which can suitably be 9/16 inch. The surface 321 should be aligned with the opposite surface 315 on the anode to be parallel and oppositely facing. The cell is operated at about 710.degree. C. at a current density of about 8 amperes per square inch. As is the case with production cells, a suitable bath contains 70% sodium chloride, 30% lithium chloride, to which is added about 7% aluminum chloride. The aluminum chloride content is maintained by periodic or continuous addition of aluminum chloride. The operating conditions are continuously maintained for a period of about 5 days during which aluminum is made continuously.

After about 5 days, the entire bath is drained from cell 200 and the cathodes are removed. The cathode surfaces 321, i.e. those closest to and oppositely facing the anode surfaces, are examined. The largest drop or droplet of aluminum found on the cathodic surface 321 is measured as an index of wettability. If this droplet is greater than one millimeter in its largest dimension in this test, the cathodic surface is considered to be wetted by the aluminum in the electrolyte bath. If, on the other hand, the largest droplet is one millimeter or less in its major dimension, the cathodic surface 321 is considered to be non-wetting.

As indicated hereinabove, the invention involves selection of cathode electrodes based on the wettability or non-wettability of the cathode surface in association with the electrolyte bath flow velocity over the cathode surface. The bath flow velocity is readily determined using a simulated water model of the cell, either full size or scaled down.

In accordance with the invention, cathode surfaces which exhibit wetting behavior are positioned to contact the bath where bath flow velocity over the cathode surface is relatively low, 1.5 feet per second or less, for instance, 0.3 or 0.5 to 1.4 or 1.5 feet per second. These will typically be found in the lower regions in cells of the type depicted in U.S. Pat. No. 3,755,099. One practice of the invention involves the use of relatively widely spaced electrodes in the cell regions which exhibit relatively low bath flow, especially where significant amounts of aluminum can accumulate on the cathode surfaces. In these regions the electrode gap, that is the distance between the anode surface and the opposed cathode surface, can be greater than 1/2 inch, for instance 5/8 to 3/4 inch, although distances of up to one inch can be useful, particularly where a significant collection of molten aluminum occurs on the cathode surface, such as sometimes can happen in the lower bath portions in a cell of the type depicted in FIG. 1 and in U.S. Pat. No. 3,755,099, that is lower regions of the cell closer to terminal cathode 16.

In those regions of electrolytic cells where the bath flow velocity at the cathode surface is relatively high, over 1.5 feet per second, for instance, 1.5 to 3 feet per second, the cathode surface should be non-wetted by the aluminum depositing there from the bath. Regions of high flow typically occur in the relatively higher regions of electrolytic cells of the type depicted in FIG. 1 and in U.S. Pat. No. 3,755,099, that is, regions closer to terminal anode 14. In regions of higher bath flow velocity, a preferred practice is to use relatively closely spaced electrodes, 1/2 inch or less, for instance 3/8 inch.

The practice of the invention includes the use in a single electrolytic cell of both high flow and low flow regions and the selective use of graphite electrodes in those respective regions based on the non-wettability or wettability of their cathode surfaces. Hence, one embodiment of the invention features the use of both high and low flow velocity regions in an electrolytic cell such that the bath flow between the anode and cathode in one or more interelectrode spaces 19 is relatively high, for instance greater than 1.5 feet per second. That same cell also includes a lower flow rate of about 1.5 feet per second or less in one or more other interelectrode spaces.The relatively high flow velocity can be 11/2 or 2 or more times the relatively low flow velocity. The practice of the invention places cathodes with non-wettable surfaces in the high flow regions and one or more cathodes with wettable surfaces in the lower flow regions, all in the same cell. The use of greater anode-cathode distances for the low flow regions and lesser anode-cathode distances for the high flow regions as just described can also be employed within a single cell.

The electrodes, including the bipolar electrodes 15, are comprised of graphite grade carbon, which can be produced from coke derived from coal or petroleum. In the case of petroleum coke, such is typically calcined at a temperature of about 800.degree. to 1600.degree. C. in order to drive off volatile impurities. In making an electrode, the calcined coke is blended with a pitch binder to provide a mixture having a pitch content of about 10 to 30%. This mixture is shaped such as by extrusion to provide a suitable size and configuration for use as an electrode or for cutting into electrodes. A shaped member can be cut to provide two or more electrode block pieces, after which the electrode is baked at about 700.degree. to 1600.degree. C. to drive off volatiles from the pitch binder. The next step usually involves immersing the baked block to impregnate it with liquid pitch to increase the density, after which it is again baked at about 700.degree. to 1600.degree. C. The baking and pitch treatment can be repeated one or more times to further increase the density. Finally, the carbonaceous material is graphitized at a typical temperature of about 2000.degree. to 3100.degree. C.

In the manufacture of graphitic carbonaceous electrode materials, non-wetting surface characteristics are generally favored by the use of higher graphitization temperatures, higher crystallinity of the graphite structure, higher graphite density and by the use of acicular or non-acicular coke as the starting material as distinct from isotropic coke. Onthe other hand, wetting characteristics are generally favored by lower graphitization temperatures and lower crystallinity and, to some extent, by lower density and by the use of isotropic coke as a starting material.

As just indicated above, the internal structure of the coke starting stock, the density and crystallinity of the graphite produced therefrom and expecially the graphitization temperature have a marked influence on the wettability or non-wettability of the graphite in contact with aluminum in a chloride reduction cell, and these aspects are now discussed in greater detail. In general, coke exhibits one of three internal structures, isotropic, acicular and non-acicular. The isotropic structure, as the name implies, is generally characterized by equiaxed grains or cells. Acicular, on the other hand as its name implies, is characterized by elongate, needle-like grains or cells. Non-acicular can be viewed as between the extremes represented by the isotropic and acicular structures. In the non-acicular structure, the grains or cells are non-equiaxed so as to be discernible from the isotropic, but are also clearly discernible from the needle-like character of the acicular structure. These characteristics are generally recognized in the art and the terms, as used herein, correspond to the general understanding in the art.

A significant consideration as to whether a particular specimen of graphite exhibits wettable or non-wettable behavior has been found to be the degree of crystallinity in the graphite structure. It is generally recognized that several useful measures of graphite crystallinity can be obtained from wide angle X-ray diffraction of the crystallite size and the interlayer spacing of graphite samples. The diameter, L.sub.a, and the height, L.sub.c, of the crystallite can be obtained from measurement of the broadening of the appropriate X-ray diffraction peaks. The interlayer spacing, d.sub.002, and d.sub.10, and the crystallite diameter, L.sub.a, remain more or less the same despite substantial changes in crystallinity. However, the degree of crystallinity correlates well with the crystallite height, L.sub.c, thus providing a simplified approach for the X-ray determination of the comparative crystallinity of graphite. This correlation is considered valid despite a simplified analysis to determine L.sub.c which is based principally on "size broadening" without allowing for strain effects or for distribution of layer spacings. That is to say that determination of L.sub.c can be made without accurate determination of the broadening parameters by a rigorous analysis of X-ray data which is complicated by a number of corrections as it generally recognized in the art of X-ray diffraction. It is suitable for purposes of the invention to evaluate the broadening parameters directly from experimental diffractometer traces and a smooth curve drawn through the profile of the trace. To determine L.sub.c, a base value of intensity is determined and a line parallel to the base line drawn at one-half of the peak height above the base line. Scherrer's equation can then be used to determine the value of L.sub.c.

In this equation, .lambda., B and .theta. are, respectively, X-ray wavelength, half width in radians, and peak angle in degrees. This method is described in a publication entitled "Measurement of Interlayer Spacings and Crystal Sizes inTurbostratic Carbons" by M. A. Short and P. L. Walker, Jr., Carbon, Vol. 1 (1963), pp. 3-9.

In general, a lower degree of crystallinity as reflected by a lower L.sub.c value correlates with a wetting characteristic, whereas a higher degree of crystallinity as reflected in a higher L.sub.c value correlates with a non-wetting characteristic. For instance, an L.sub.c of 350 angstrom units (A) or more correlates with non-wettable performance, whereas an L.sub.c value less than 350 angstrom units tends to characterize wettable performance.

Where isotropic coke serves as the starting material, the resulting graphite will, for all practical purposes, always exhibit a wettable characteistic with respect to aluminum in chloride reduction cells. The carbonaceous material can be graphitized at almost any temperature between 1800.degree. and 3000.degree. C. and still exhibit a wetting behavior which is more or less insensitive to density changes. Further, the L.sub.c value will practically always be less than 350 angstroms (A) and generally range from less than 100 to a maximum of about 300 angstroms.

Where acicular coke serves as the starting material, non-wetting behavior is favored where the graphitization temperature is equal to or greater than 2300.degree. C. This tends to produce an L.sub.c which exceeds 350 angstroms. Acicular coke can be produced to exhibit wetting behavior by graphitizing at a temperature of less than 2300.degree. C. which tends to result in a crystallinity characterized by an L.sub.c value of less than 350 angstroms. In the case of acicular coke as the starting material in producing the graphite, the density of the final graphite product can exert some influence on its wetting or non-wetting behavior. In general, a higher density tends to favor non-wetting behavior, whereas a lower density tends to favor wetting behavior. In general, the density can be controlled by the pitch impregnation employed in manufacturing the graphite. Repeating the pitch impregnation one or more times tends to increase the density.

In the case of non-acicular coke as the starting material, non-wettable behavior is favored by a graphitization temperature of 2500.degree. C. or higher which tends to result in a crystallinity characterized by an L.sub.c value of 350 angstroms or more. Graphite produced from non-acicular coke can be produced to exhibit wettable behavior by graphitizing at a temperature of less than 2500.degree. C. which tends to result in a crystallinity characterized by an L.sub.c value of less than 350 angstroms. Density is not as important as with acicular coke.

From the foregoing explanation, it is readily apparent that the graphitization temperature is of marked significance with respect to acicular and non-acicular coke in the production of graphite. In the case of acicular coke, density control becomes a factor but to a much lesser extent than graphitization temperature. Isotropic coke practically always results in wetting performance irrespective of graphitization temperature. The highest temperature to which the graphite has been heated is readily determined by subsequent X-ray diffraction analysis. As is known in the art, a standard curve relating X-ray parameter to highest temperature encountered can be developed for a given coke type and manufacturing sequence. Hence, this analysis is considered to reliably indicate the highest temperature employed in manufacturing graphite, that is, the graphitization temperature. Of significance in the use of wettable graphite is the fact that it can be less expensive to produce than non-wettable graphite, thus reducing costs, provided it is properly employed in accordance with the invention.

To this point, the invention has been described with an eye to starting with a single type of coke for a given graphite electrode production since this is the normal practice in commercial production. However, it is possible to use more than one grade of coke in producing a graphite electrode. In such case, the guidelines discussed above apply based on the dominant type of coke employed based principally on proportion and secondarily on comparative influence. With respect to comparative influence, isotropic coke is more influential than either acicular or non-acicular, and non-acicular is more influential than acicular. If a mixture of coke types includes 60 or 70% or more of any particular type, that type dominates. However, where different types are present in more or less equal amounts, then the above-stated order of influence applies. Obviously, as the degree of dominance diminishes, the certainty of the result can be likewise diminished. Hence, it is preferable in practicing the invention to use but a single type of coke as the starting material or at least, where a mixture is used, it is preferred to use a mixture characterized by a clear dominance such as a dominance of at least 80% proportion.

The invention and the improvements achieved thereby are illustrated in the following examples listed in table form. The data in Tables I and II show cathode wear rate as it varies with cathode graphite wettability and bath flow velocity in baths containing around 70% NaCl and 30% LiCl to which is added about 7% AlCl.sub.3. Wettability is determined in accordance with the herein-described test (FIG. 2 and particularly FIG. 3). The baths operating at about 710.degree. C. are electrolyzed to produce aluminum and the wear rate is determined for a measured time and converted to mm. of wear per year to provide a comparative wear estimate.

Table I illustrates the sensitivity of wettable graphite to a relatively high bath flow velocity of 2.5 feet/sec. (Example 2) but indicates a much lower wear rate for a low bath flow velocity of less than 0.1 feet per second (Example 1). A similar test at 1.4 feet per second bath veolcity resulted in a comparative wear rate estimate of only 3 mm. per year on a wettable graphite cathode surface. Non-wettable graphite (Examples 3 and 4) in this test had acceptable wear rates for either flow rate but not as good as the wettable graphite under low bath flow rate conditions.

Table II shows Examples 5 to 10 wherein starting with acicular coke (Examples 5 and 6) or with non-acicular coke (Examples 7 and 8), the graphite produced can be either wetting or non-wetting by molten aluminum in accordance with the invention. For instance, in Examples 5 and 6, increasing graphitization temperature from 2000.degree. C. to 2600.degree. C. changes the graphite from wettable to non-wettable. However, with isotropic coke (Examples 9 and 10) graphitization at either 1800.degree. or 2800.degree. still results in a wettable surface.

While the invention has been described with particular reference to electrolytic cells of the type shown in FIG. 1 featuring horizontal electrodes and horizontal interelectrode spaces therebetween for essentially horizontal bath flow through the interelectrode spaces, it is believed that the invention may also be useful in cells featuring non-horizontal electrodes such as vertical electrodes. In such case, the non-wettable cathode surfaces are to be used with higher velocity bath movement whereas wettable cathode surfaces are to be used in conjunction with lower bath velocity over the cathode surface.