US3320150A - Molded carbon materials - Google Patents

Description

United States Patent 3,320,150 MOLDED CARBON MATERIALS William Joseph Metrailer, Baton Rouge, La, assignor to Esso Research and Engineering Company, a corporation of Delaware N0 Drawing. Filed Sept. 6, 1963, Ser. No. 307,019 6 Claims. (Cl. 204294) This invention relates to carbon electrodes and the manner of their preparation from mixtures comprising relatively active and relatively inactive coke material. More particularly, it relates to the preparation of electrodes which can be utilized for obtaining of aluminum from its ores.

In the manufacture of aluminum by electrolytic reduction of alumina in a suitable fused bath, the necessary carbon electrodes have usually been manufactured from petroleum coke of relatively high purity. This application relates to a process for making improved coke electrode compositions which comprises formulating the composition of relatively high reactivity coarse coke and relatively low reactivity fine coke which, on molding and baking into an electrode, provides uniform oxidation of the electrode and substantially minimizes dusting of the electrode.

' Carbon electrodes heretofore have contained relatively coarse coke material, mixtures of coarse and fine coke materials, as well as mixtures of particles having a wide size particle distribution. The various particle sizes which have been used are normally obtained by grinding and screening coke from the same source.

One of the problems in using electrodes made from coke in the electrolytic reduction of aluminum has been the dusting of the electrodes during the reduction due to the different rates of consumption of the coke particles and the binder coke. The binder coke is the carbona-' ceous deposit from the pitch binder and is formed during the baking of the electrodes. The coke components that are consumed most slowly fall off the electrodes giving rise to particles of loose carbon which is called dust. This carbon dust short circuits the electrolysis bath and represents lost coke, that is, coke that does not reduce alumina.

Coke electrodes are generally formed by mixing the coke aggregate with a binder material, molding the mixture under pressure, and baking at elevated temperature for several hours. The baking step at elevated temperature removes substantially all of the volatile materials from the binder material leaving a binder coke. Normally, it is the relatively higher reactivity of this binder coke and its rapid oxidation which causes the dusting problem. This occurs when the binder material burns aWay-fromthe relatively slower burning coke aggregate, allowing the slower burning coke to fall out of the molded electrode since there is no binder left to hold it in place.

In an attempt to overcome the dusting problem, which is both an economic and process problem in the electrolytic reduction of alumina, several techniques have been tried. For example, mixtures of coarse aggregate coke and fine aggregate coke with various binder contents have been tried to obtain an optimum mixture and to avoid the dusting problem. To some extent these mixtures have helped to reduce the dusting. That is, it reduces the dusting over molded electrodes made with only coarse aggregates and binder. Another technique attempted was to mix different types of cokes, both types having wide particle size distributions. These mixtures, again, helped but did not effectively overcome the problem.

Petroleum cokes vary in their resistance to oxidation, that is, their reactivity. The reactivity, to some extent, is a function of the petroleum source and of the impurity 3,320,150 Patented May 16, 1967 concentration of the petroleum coke, that is, the ash content. Generally, the higher the impurity concentration, the higher the reactivity of the coke. However, the reactivity of a particular coke is also a function of the temperature and method by which the coke was produced, for example, batch destructive distillation as opposed to deposition on a heated surface. It is also dependent upon the temperature at which the coke was calcined.

In accordance with the present invention, a molded coke electrode with greatly minimized dusting characteristics is formulated from coarse, relatively large coke particles having a relatively high reactivity, in admixture with small fine coke particles having relatively low reactivity, and a binder. This mixture is molded under pressure and heated at high temperatures for a sufficient period of time to form a baked coke electrode. The baking or heating of the electrode volatilizes the volatile constituents of the binder material, leaving about 50 wt. percent of the binder as a binder coke of relatively high activity. It is the high reactivity of this binder coke which is the principal cause of the dusting of the coke electrodes.

The fine coke is selected to have a reactivity lower than the coarse coke and also preferably substantially lower than the binder coke. The fine coke thus presents a diffusion barrier to the consumption of the binder coke and thereby reduces the consumption of the filler material. It has been found that carbon bodies with improved resistance to oxidation and less tendency to form loose carbon dust can be prepared by selective use of less reactive carbon as the source of the finest portion of the coke aggregate. The low reactivity fine carbon and the carbon formed by baking out the binder form the filler between the large size coke aggregate particles in the baked out electrode. The carbon formed from the binder normally woud be consumed more rapidly than the coarse coke aggregate which makes up the bulk of the electrode. Since this binder bonds the electrode together, the rapid relative consumption results in loosening of the bond and the large coke aggregate will dust or flake off the electrode.

If fine aggregate is employed which has the same reactivity as the coarse aggregate, it will be consumed with the binder coke more rapidly and dusting will be a major problem. However, by using fines, in accordance with the present invention, of a lower reactivity relative to the larger coke aggregate, the rate of coke consumption on the entire surface of the electrode is more balanced and dusting greatly minimized. The low reactivity fines in the filler material acts as a diffusion barrier to reduce the rate of consumption of the high reactivity binder coke produced in baking the electrode.

The baking temperature and the binder content of the electrode varies depending on whether the electrode is the pre-baked or self-baking type. The self-baking electrode is generally baked at a lower temperature which is primarily controlled by the cell operating temperature.

The effectiveness of the relatively low reactivity fines in reducing dusting will depend on the particle size difference between the coarse aggregate and the fines aggregate and the relative reactivity difference between the coarse and the fines. The rate at which the coke of the same reactivity is consumed is generally inversely proportional to its size, that is, surface area/volume ratio; however, this is affected somewhat by the binder coke surrounding the particles. In order to effectively reduce the over-all consumption of the binder coke and fine coke between the larger coarse coke aggregate, the difference in reactivity of the fines desirably should be sufficient to offset smaller size effect. The effect of surface area/volume ratio in increasing the consumption of a coke particle can be seen by employing, in a pre-baked electrode, coke of the same relative reactivity, both as the coarse aggregate and as the fine aggregate, since the fine aggregate will form the filler material with the binder coke and will be consumed more rapidly than the larger coarse aggregate. However, the consumption of the filler will be slower than if there were no fines present in the filler.

Generally, the reactivity of a coke from a particular petroleum coke source which has been treated in the same manner will depend upon its relative size and impurity content. The various steps in which the coke is treated prior to forming the coke electrode, however, has an effect on its reactivity. For example, cokes deposited on a heated surface will frequently be less reactive than cokes formed by batch destructive distillation. However, this is not always so since the relative size of the particles being handled has an effect, as well as the calcining temperature, and the impurity content. The binder coke generally has the highest reactivity of the coke materials forming an electrode because this coke material is formed at the lowest temperature. The final baking temperature is lower than normal calcining temperatures. Also, binder coke is the least dense and generally coal tar binders are used which give a more reactive coke.

Generally, the following can be said to be true. A coke electrode made substantially of coarse coke aggregate and with little or no fines present and a binder coke will have relatively the largest dusting problem. An electrode made with the same coke but with a dumbbell particle size distribution of coarse aggregates and fine aggregates and a binder coke will also have a dusting problem although not as severe as the first case, since the fines will present a slight impediment to the rapid oxidation of the binder coke. Coke electrodes made from coke aggregate having a wide, even size particle distribution and a binder material will also have a dusting problem. Blends of different cokes of different reactivity but having wide size particle distribution will also have a dusting problem to the same extent that cokes of the same reactivity of wide size particle distribution will have a dusting problem.

There are three principal types of coke materials which can be used in formulating coke electrodes for reduction of aluminum. These are delayed cokes formed by batch destructive distillation, which are then calcined, fluid cokes formed by surface deposition at relatively low temperatures 900-1200" P.) which are then calcined, and cokes' formed by surface deposition at temperatures above about 1900" F. which do not need calcining. Within each of these three classes of coke, depending on the purity and the petroleum source, there is a wide range of relative coke activities.

In order to show the relative reactivity of different types of cokes, several calcined cokes from different sources were tested for reactivity and the results of the tests are reported below in Table I.

The relative reactivity is defined as the wt. percent of carbon which reacts with a flowing stream of CO at 1740 F. under identical time and test conditions. The carbon or coke samples are 14-35 mesh carbon particles.

In the manufacture of the electrode itself, the coke blend is admixed with and charged together with a carbonaceous binder to the fabrication system. The binders utilized are conventional and include materials such as aromatic coal tar pitch binders. Such binders generally have melting points lying between the range of 70-120" C. They contain small amounts of hydrogen (about 5% or less). The concentration of benzene and nitrobenzene in solubles represents preferably about 20-35% and 5-15%, respectively of the binder. The binder is utilized in an amount of 16-47 parts by weight per 100 parts of coke in the blend.

In general, two types of electrodes are employed by the industry: (a) pre-baked'and (b) Soderberg self-baking electrodes. In the former, a mixture comprising about -86% of coke aggregate and about 14-20% of coal tar pitch binder is molded at pressures of about 3000- 10,000 p.s.i. or extruded and then baked for periods up to 30 days at 1800-2200 F. These pre-formed electrodes are then used in electrolytic cells being slowly lowered into the molten alumina as they are consumed. Butts of the unconsumed electrodes are reground and used in subsequent electrode preparations.

The Soderberg process involves the continuous or intermittent addition of a coke and coal tar pitch binder introduced as a paste to the top of the cell as the electrode components in the lower part of the cell are consumed. In this operation, the paste represents a blend of about 68- 74% coke aggregate charge and 26-32% pitch binder. The cells usually operate at temperatures of 1700-1900 F. and the electrodes are consumed. The paste is baked into an electrode between the time it is added at the top and the time it is used by the heat dissipated from the cell. Both methods have in common the baking of the mixed coke aggregate charge and binder at a temperature in the range of 1700-2200 F. The Soderberg process, however, does not result in any unused butt materials.

In the electrolytic reduction of alumina using coke electrode formulations, carbon dioxide is formed in the reduction of the alumina to aluminum metal. The carbon dioxide formed is in contact with the coke electrode and has a tendency to oxidize some of the coke from the electrodes. This oxidation reaction results in the conversion of some of the carbon in the electrode to carbon monoxide and represents a loss of the carbon to the reduction process. The technique for measuring the reactivity of the coke electrodes is to contact coke electrode samples in a stream of carbon dioxide gas at the temperature at which the electrolytic reduction of alumina takes place, namely, about 1740 F. As previously set forth, various petroleum coke materials were tested to determine their reactivity by contacting coke aggregate samples of 14-35 mesh with carbon dioxide at 1742 F. The reactivity of the coke samples was determined by measuring the amount of coke that reacted with the carbon dioxide, that is, the amount of coke weight loss on contact with carbon dioxide under identical test conditions was a measure of its reactivity. It was found that this reactivity of the free coke aggregate could be related to the reactivity of molded coke electrodes formed therefrom.

In accordance with the present invention, it is preferred to use coarse aggregate particles having a wt. average size ratio 10/1 to 40/t1 times the wt. average particle size of the fines aggregate particles. For example, if the wt. average particle size of fines materials is 44 microns, it would be desirable to have the wt. average particle size of the coarse aggregate be above about 500 microns. The particle size distribution or the percentage of coarse particles relative to the percentage of fines particles present will generally be 1/1 to 10/1 and preferably 2/1 to 3/1, that is, the fines particles will comprise between about 10 and 50 wt. percent of the total aggregate. Depending on the particle size distribution and the relative size of the fines-to-coarse aggregates, it will be necessary to select a fines material of substantially less reactivity than the coarse material. Generally, a coke fines material will be selected having a ratio of relative reactivity to the relative reactivity of the coarse material of 1/ 2 to US, more generally 2/3 to 1/3.

In selecting the particle size distribution, it is preferable to have a dumbbell type distribution with below about wt. percent overlap of the particles. It is preferred that the fines material be sufliciently small so that they, in conjunction with the binder, will form a suitable filler material able to cement the large coarse aggregates together into a high density, high crush-resistant molded electrode. Suitable coarse particles that can be used in accordance with this invention will be particles having a particle size wherein 90% of the particles are between 2 and 48 mesh (about 300-10,000,u). The fines material which can be used in accordance with this invention to form electrode formulations will have about 90 wt. percent of the particles smaller than 100 mesh (150 microns), and preferably 40 wt. percent smaller than 325 mesh (44 microns). The amount of fines-to-coarse particles will depend on the particular use to which the formulated electrode is to be put but will generally be in the amount of 10-50 wt. percent of fines based on total electrode aggregate formulation. In accordance with this invention, calcined coke material or coke material not necessitating calcining is separated into two portions, depending on its relative reactivity to oxidation with carbon dioxide. The higher relative reactivity coke is used as is or ground to form a suitable coarse coke aggregate having a particle size of 248 mesh. The relatively low reactivity coke portion is ground to a particle size of 90% smaller than 100 mesh. Thirty-five to 50 wt. percent of fines based on combined weight of coarse and fines are mixed with the coarse aggregate and blended with 14-20 wt. percent coal tar pitch binder based on total electrode formulation weight and molded a pressure of about 5000 p.s.i. and subsequently baked at about 1800 2200 F. to form a molded coke electrode. Electrodes prepared in this manner have shown superior characteristics to crushing and dusting tendencies.

in a flowing carbon dioxide gas stream under a standard set of conditions. Upon completion of the test, the molded carbon body was cooled and weighed and then brushed with a stiif bristle brush. The loosely held carbon particles which could be brushed from the sample were weighed and reported as dust.

The molded carbon specimens used were prepared by mixing the coke aggregate and pitch binder for minutes at 300 F. The mixture was molded and baked slowly to a temperature between 1800 and 2000 F. for a period of 1050 hours. Two levels of binder content were used, depending on the molded sample, varying between 16 and 28 wt. percent binder. Each of the molded carbon electrode samples contained, as the coarse aggregate, particles having a particle size of 2 to 48 mesh and, as the fine aggregate, particle fines having the size ground milled to pass through a 100 mesh screen. The particle size of the coarse aggregate given in microns is 300l0,000 microns and of the fine particle size given in microns, 90% less than 150 microns and of which is less than 44 In the tests, each carbon electrode was molded, baked, and test carbon specimens cored from the electrode were contacted with carbon dioxide gas at 1742 F. for periods up to 16 hours. Depending on the particular sample, measurements were taken at 1.5, 4, 8, and 16 hours. The wt. percent loss at each of these times was taken. The loose carbon was then brushed off from separate samples and weighed.

Example I In accordance with the invention, a coke electrode was formulated wherein the fines aggregate was selected so that it had a relative reactivity about /2 that of the coarse aggregate, i.e., 2.2 to 4.7 (cokes 3 and 5, Table I). The electrode was made from equal portions of coarse particles 2-48 mesh and fines particles of less than 100 mesh. The results are shown below.

Wt. Percent Coke Burned in- Coke in Electrode Wt. Percent Relative in Electrode Reactivity 1.5 hrs. 4 hrs. 8 hrs. 16 hrs.

Coarse. 41.9 4. 7 0.74 1.98 3. 96 7. 92 Fine- 41.9 2. 2 0. 35 0.93 1. 86 3. 71 Binder. 16. 2 15.4 0.94 2. 5. 00 10. 00

Total cal. burned cnl'e 2.03 5. 41 10.82 21. Total measured burned coke 1. 8 5. 2 12. 7 33.8 Dust measured 0. 1 0. 5 7. 0 40. 9

This invention and its advantages will be better illustrated by the following examples of electrodes prepared in the manner taught.

Green fluid coke from several commercial fluid coking units was obtained and calcined at about 2400. Delayed fluid coke and a coal tar coke which had been calcined at about 2400 were also obtained. A commercial coal tar binder was used. The coal tar binder had a 215 F. softening point, a coking value of about 67.5 and a carbon to hydrogen ratio of 1.6. Tests were carried out using coke samples having different relative reactivities and different particle size distributions which were molded into electrodes to determine their dusting tendencies. Correlations based on these test data were used to calculate the dusting characteristics of other formulations. The reactivities of the molded carbon specimens were determined at 1742 F.

The above data clearly show the improved dusting characteristics of electrode formulations containing relatively reactive coarse aggregate and relatively less reactive fines aggregate. The dusting was substantially reduced with this mixture.

Example 11 Wt. Percent Coke Burned in- Coke in Electrode Wt. Percent Relative in Electrode Reactivity 1.5 hrs. 4 hrs. 8 hrs 16 hrs.

41.9 4. 7 0.74 1. 98 3.96 7. 92 41. 9 4. 7 0.74 1. 98 3.96 7. 92 Binderl6. 2 15. 4 0. 94 2. 50 5. 00 10. 00

Total Cal. Burned Coke 2. 42 6. 46 12. 92 25. 82 Total Measured Burned Coke 2.4 7. 5 15. 5 31. 92

Dust Measured- 0.3 5.9 27. 7 1 100.0

1 Sample disintegrated.

The above data clearly show that coke electrodes formulated from coarse and fine coke aggregates, having the same relative reactivity, experiences a considerable amount of dusting and after 16 hours disintegrate.

Example III sumed at about the same rate as the binder plus fines which would result in substantially no dust formation. The presence of the low reactivity fines in the binder would provide a diffusion barrier to the oxidation of the binder and reduce the consumption of the binder plus fines to about the same rate as that of the coarse aggregate.

Example V A coke electrode formulation containing a coarse coke aggregate having a relative reactivity of 6.5 and a fine aggregate having a relative reactivity of 2.2 (cokes 7 and 3) is made in accordance with this invention. The formulation contains equal portions of coarse aggregate of 2 Coke in Electrode Wt. Percent Coke Burned in Wt. Percent Relative in Electrode Reactivity 1.5 hrs. 4 hrs. 8 hrs. 16 hrs.

Coarse 54. 8 1. 21 2. 42 4. 84 Fine. 36. 5 0.80 1. 60 3. Binder 8.7 1. 34 2. 68 5. 36

Total cal. burned coke- 3. 6.70 13. 4

Total measured burned coke 3.0 1 6. 0 1 12. 0 Dust measurcd 0.3 2.7 13.0

1 Calculated based on burning rate and relative reactivity.

The above data and calculated results show that the amount of dust formed at a given carbon consumption to 48 mesh and fine aggregate of minus 100 mesh. The results are given below.

Wt. Percent Coke Burned in Coke in Electrode Wt. Percent Relative in Electrode Reactivity 1.5 hrs. 4 hrs. 8 hrs. 16 hrs.

Coarse 41. 9 6. 5 2 72 5. 44 10.88 Fine-.. 41. 9 2. 2 0.93 1. 86 3. 71 Binder 16. 2 15. 4 2. 50 5. O0 10. 00

Total cal. burned r'nlre 6.15 12. 30 24. 50 Cal. dusk 0.3 1.0 2.0

(burned) level is greater than that obtained in Example I where the ratio of the reactivity of the coarse-to-fine coke aggregate was about 2/ 1.

Example IV The calculations show that the coarse aggregate (again being about three times as reactive as the fines) would be consumed at about the same rate as the :binder plus fines which would result in substantially no dust formation.

A coke electrode formulation containing a coarse coke 45 E l Vl aggregate having a relative reactivity of 12.7 and a fine A coke e1 ctrode for lafo t k aggregate having a relative reactivity of 4.7 (cokes 8 and a0 re ate hsvin r 1 q i 3 e 5) is made in accordance with this invention. The fora re 5 .1 3 mulation contains equal portions of coarse aggregate of ade i g s. an f 2 to 48 mesh and fine aggregate of minus 100 mesh. The 50 e m 1 n con ams equa POI Ions 0 results are given below.

Wt. Percent Coke Burned in Coke in Electrode Wt. Percent Relative in Electrode Reactivity 1 5 hrs 4 hrs. 8 hrs. 16 hrs.

41. 9 12. 7 5. 3 10. 6 21. 2 41. 9 4. 7 1.98 3. 96 7. 92 Binder 16.2 15. 4 2.50 5. 00 10.00

Total cal. burned Mire 9. 78 19. 52 39.12 Cal. dust 0.5 2.0 4. 0

The calculations show that the coarse aggregate (being about three times as reactive as the fines) would be conminus 100 mesh. The results are given below.

Wt. Percent Coke Burned in- Coke 1n Electrode Wt. Percent Relative in Electrode Reactivity 1.5 hrs. 4 hrs. 8 hrs. 16 hrs.

Total cal. burned coke... 7.08 14.16 28. 32 Cal. dust- 9. 0 42 1 100 1 Sample would disintegrate.

The calculations show that coke electrodes formulated from a mixture containing coarse aggregate of low relative reactivity and fine aggregate of relative high reactivity would present severe dusting problems.

The above results clearly illustrate the novel and inventive features of the present invention. From the above results it is clearly seen that when fine particles having a relatively lower reactivity than the larger coarse aggregate particles are used to make a coke electrode, substantially less dusting of the coke electrode occurs at the same level of consumption (carbon loss). The minimization of the dusting of the coke electrodes is a function both of particle size distribution and relative reactivity of the particular coke particles utilized. The source of the coke aggregate in itself is not important in determining whether or not dusting will occur. It can be seen from the above that dusting can be minimized -by using coarse aggregates of fluid coke and fines of fluid coke as well as using coarse aggregates of delayed coke and fines of delayed coke if the relative reactivity of the fines is substantially less than the relative reactivity of the coarse aggregate. Other of the examples show that where coarse aggregates of delayed coke and fines of delayed coke, both having the same relative reactivity, or coarse aggregates of fluid coke and fines of fluid coke both having the same relative reactivity, are used in electrode formulations that dusting is a substantial problem.

The coke formulations of the present invention can be used as refractory lining, heating electrodes, or anywhere where reactivity or oxidation of the carbon in a formulated carbon body is a problem. The reactivity of the coke can be effected by treatment of the coke to obtain cokes of various degrees of relative reactivity. For example, a particular coke mate-rial having a specific reactivity can be separated into two portions and one treated to make it less reactive and the other treated to make it more reactive. The more reactive material can be used as the coarse aggregate and the less reactive material as the fine aggregate. A diflerence in reactivity can be produced :by other means, such as by treating particular coke material at different calcining temperatures or by adding specific chemicals to a particular coke material to either increase or decrease its relative reactivity.

It is to be understood that the invention is not to be limited to the specific examples which have been offered merely as illustrations and that modifications may be made without departing from the spirit of the invention.

What is claimed is:

1. A molded coke formulation comprising a coarse coke aggregate in which at least wt. percent of the particles have a particle size of 300-10,000 microns, a fines aggregate in which at least 90 wt. percent of the particles have a particle size of less than microns and a carbonaceous binder, wherein the ratio of reactivity of the fines aggregate to CO and the reactivity of the coarse aggregate to CO at the same reactivity test conditions is between about 1/2 respectively and 1/5 respectively.

2. The composition of claim 1 wherein the weight average particle size of the coarse aggregate is 10/1 to 40/1 times the weight average particle size of the fines aggregate.

3. The composition of claim 1 wherein the ratio of coarse aggregate to fines aggregate is 1/1 to 10/1.

4. The composition of claim 1 wherein the overlap of coarse and fines aggregate is less than 20 Weight percent of the total coke aggregate.

5. The composition of claim 1 wherein at least 40 wt. percent of the fines aggregate is less than about 44 microns.

6. The composition of claim 1 wherein the relative reactivity of the fines aggregate to the coarse aggregate is at least 1/2 to 1/3.

References Cited by the Examiner UNITED STATES PATENTS 2,653,878 9/1953 Sejersted et a1. 204-294 3,197,395 7/1965 Nelson 204-294 FOREIGN PATENTS 583,671 9/1959 Canada.

JOHN H. MACK, Primary Examiner. D. R. JORDAN, Assistant Examiner.

Claims (1)

1. A MOLDED COKE FORMULATION COMPRISING A COARSE COKE AGGREGATE IN WHICH AT LEAST 90 WT. PERCENT OF THE PARTICLES HAVE A PARTICLE SIZE OF 300-10,000 MICRONS, A FINES AGGREGATE IN WHICH AT LEAST 90 WT. PERCENT OF THE PARTICLES HAVE A PARTICLE SIZE OF LESS THAN 50 MICRONS AND A CARBONACEOUS BINDER, WHEREIN THE RATIO OF REACTIVITY OF THE FINES AGGRTEGATE TO CO2 AND THE REACTIVITY OF THE COARSE AGGREGATE TO CO2 AT THE SAME REACTIVITY TEST CONDITIONS IS BETWEEN ABOUT 1/2 RESPECTIVELY AND 1/5 RESPECTIVELY.