STATE OF ILLINOIS ^WiMiapiG. Stratton, Governor DEPARTMENT OF REGISTRATION AND EDUCATION Vera M. Binks, Director 1957 VANADIUM EFFLORESCENCE and its control by the USE OF FLUORSPAR D. L. Deadmore A. W. Men J. S. Machin REPORT OF INVESTIGATIONS 202 ILLINOIS STATE GEOLOGICAL SURVEY JOHN C. FRYE, Chief URBANA, ILLINOIS VANADIUM EFFLORESCENCE and its control by the USE OF FLUORSPAR D. L. Deadmore A. W. Allen J. S. Machin Illinois State Geological Survey Report of Investigations 202 Urbana, Illinois 1957 PRINTED BY AUTHORITY OF THE STATE OF ILLINOIS STATE OF ILLINOIS HON. WILLIAM G. STRATTON, Governor DEPARTMENT OF REGISTRATION AND EDUCATION HON. VERA M. BINKS, Director BOARD OF NATURAL RESOURCES AND CONSERVATION Hon. Vera M. Binks, Chairman W. H. Newhouse, Ph.D., Geology Roger Adams, Ph.D., D.Sc, Ll.D., Chemistry Robert H. Anderson, B.S., Engineering A. E. Emerson, Ph.D., Biology Lewis H. Tiffany, Ph.D., Pd.D., Forestry Dean W. L. Everitt, E.E., Ph.D., University of Illinois President Delyte W. Morris, Ph.D., Southern Illinois University GEOLOGICAL SURVEY DIVISION John C. Frye, Ph.D., D.Sc, Chief (60353—2800—7-57) CONTENTS Page Introduction 5 Clay samples and materials 6 Plan of experiment 7 Experimental procedures and techniques used 7 Vanadium analysis 7 Sample preparation 7 Extraction of vanadium 8 Water absorption 8 Particle size analysis of raw clays . 8 X-ray studies 8 Differential thermal analysis (D.T.A.) . 8 Base exchange 10 Wick tests 10 Experimental results , 10 Properties and composition of the unfired clays 10 Properties of the fired clays .12 Influence of fluorspar on the fired properties of pure clay minerals . . . . . . . . 21 Discussion of experimental results 24 Summary and conclusions 28 References . 29 ILLUSTRATIONS Figure Page 1 . Schematic sketch of differential thermal analysis apparatus g 2. Particle size distribution of unfired clays 10 3. Effect of fluorspar content and firing temperature on vanadium extracted and on water absorp- tion for clay S 12 4. Effect of fluorspar content and firing temperature on vanadium extracted and on water absorp- tion for clay M 12 5. Effect of fluorspar on the D.T.A. diagram of clay S. 19 6. Effect of fluorspar on the D.T.A. diagram of clay M 20 7. Effect of fluorspar content on the major exothermal peak temperature 21 8. Effect of fluorspar on the D.T.A. diagram of Georgia kaolin 23 9. Effect of fluorspar on the D.T.A. diagram of Anna, Illinois, kaolin 24 10. Effect of fluorspar on the D.T.A. diagram of illite 25 11. Relationship between water absorption and extracted vanadium for clay S 27 TABLES Table Page 1. Particle size analysis of the unfired clays 10 2. Estimates of quartz and clay mineral contents of the clays 11 3 . Vanadium content of unfired clays 1 1 4. Effect of firing temperature and fluorspar content on vanadium extractability and water ab- sorption of clays 13 5. Vanadium extracted by the Soxhlet and column methods 14 6. Total vanadium content of clays fired to various temperatures 15 7. Wick test results 16 8. Spectrochemical estimate of relative abundance of metals in colored efflorescence . . . . 17 9. Crystalline phases identified in clays by X-ray diffraction 18 10. Influence of fluorspar on the crystalline phases identified in Georgia and Anna kaolinite and illite 22 Digitized by the Internet Archive in 2012 with funding from University of Illinois Urbana-Champaign http://archive.org/details/vanadiumefflores202dead VANADIUM EFFLORESCENCE AND ITS CONTROL BY THE USE OF FLUORSPAR By D. L. DEADMORE,! A. W. ALLEN^ and J. S. MAGHINi ABSTRACT The effect of fluorspar on the water-soluble vanadium content of four fired buff -burning clays was studied. Fluorspar was found to inhibit the development of water-soluble vanadium compounds to a greater or lesser extent in all of the clays. The inhibitory effect was much more pronounced in the less refractory clays that contained considerable illite. The kaolinitic clays required more fluorspar and higher firing temperatures to prevent the development of water-soluble com- pounds of vanadium. INTRODUCTION The appearance of yellow, green, or brown stain on buff brick has been the subject of study and discussion by a number of authors, including Hammer (1953), Lovejoy (1934), Curry (1926), Palmer (1929), Bosazza (1940), Kal- launer and Hruda (1922), Seger (in Hecht and Cramer, 1902) and others. From their writings one may conclude that vanadium is the principal contrib- utor to colored stain on buff brick, and that molybdenum, iron, chromium and nickel make secondary contributions. Alkali and alkaline earth metals are also usually present in the colored stains, but are not believed to contrib- ute directly to the color. Study of the literature leads one to the tentative conclusion that the salts which cause colored efflorescence contain princi- pally vanadium (and a small amount of molybdenum) in the form of alkali and alkaline earth vanadates and that such salts are sensitive to pH. From alka- line solutions colorless salts crystallize out, but from acid solutions colored ^Illinois State Geological Survey, ^Department of Ceramic Engineering-, Univer- sity of Illinois. [ salts ranging from yellow through green to brown crystallize out (Jahr and Jander, 1934; Britton and Robin- son, 1930; Jander and Jahr, 1933). Very small amounts of vanadium cause colored stains on brick. Kallau- ner and Hruda (1922) found that the addition of as little as 0.001 percent of potassium vanadate to a pure kaolin produced yellow discoloration on bricks made therefrom. Bertrand (1950) states the amount of vanadium in clays to be usually small (within the limits of 10 to 1000 parts per million). Thus, even the smaller amount may cause ob- jectionable stain on light-colored brick. It has been observed (Bosazza, 1940 ; Hecht and Cramer, 1902; Kallauner and Kokta, 1941) that the vanadium in raw clay is not in a water-soluble form, but that heating the raw clay to tem- peratures of about 1000 °C converts at least a part of the vanadium to solu- ble vanadates. Palmer (1929) found that the addition of alkali carbonates to the clay increased the amount of vanadium efflorescence on the fired bricks. He and co-workers also found that if their test bricks were fired to 5] ILLINOIS STATE GEOLOGICAL SURVEY temperatures above 1000° C, vana- dium efflorescence did not appear. He believed that this effect was due to fu- sion of vanadium compounds with other materials in the brick. The exact mode of occurrence of va- nadium in clays is not known. Hatha- way (1955), on the basis of x-ray stud- ies of vanadium micas, recently con- cluded that the vanadium replaced some of the aluminum in octahedral positions in the clay-like lattice. In raw clays, such as are ordinarily used in brick manufacture, there are many impurities. Vanadium could conceiva- bly be present as water-insoluble vana- dium minerals (Fioletov, 1929), as re- placement for other ions in the clay lat- tice, as solid solution (Leutwein, 1941) with other impurities, as occu- pying base-exchange positions on the clay particles, or, as suggested by Ham- mer (1955), as incompletely weathered fragments of primary minerals, colloi- dal particles, or possibly organic com- plexes. The prevention of vanadium stain on buff brick has been the main object of most experimental studies that deal with such stain. The problem has usu- ally been approached in laboratory studies by investigating the effects of addition of small amounts of various materials to the clay before forming it into bricks. A number of materials have been reported to be effective. Lead and barium salts and various fluorides have all been reported to have inhibitory effects. Calcium fluoride in the form of fluorspar is the most prom- ising from the economic angle because it is a plentiful natural material and because it need not be of high purity. Fines that contain considerably less calcium fluoride than acid grade fluor- spar are relatively inexpensive. The small amounts of silica and other min- erals that such fines contain are incon- sequential in the brick. The object of the present investiga- tion was to obtain information of a fundamental nature on the formation of vanadium-type stain on bricks and on the effect of additions of fluorspar to the clay on the formation of such stains. Acknowledgments Thanks are due W. F. Bradley who gave help in the interpretation of x-ray data, to Juanita Witters who did the spectrographic work, and to W. A. White for the particle size determina- tions. All are staff members of the Illinois State Geological Survey. CLAY SAMPLES AND MATERIALS The four clay samples studied were from operating brick plant pits. Brick made from all four clays were known to have displayed colored efflorescence to some degree at times. The operators of these plants had all had trouble with ''green stain" to a greater or lesser de- gree. The plant methods used for con- trol of "green stain" included: a) fir- ing to higher temperatures; b) selec- tive mining of clay (one operator claimed to be able to recognize by its appearance clay which would give trouble; c) adding fluorspar. Clay S from an underground mine was of Pennsylvanian age and is de- scribed as an underclay associated with No. 2 coal in north central Illi- nois. Clay L, also of Pennsylvanian age, sometimes referred to as the Trade- water clay, comes from a large open pit also in north central Illinois. Clay M came from an open pit in western Indiana. The bed was less than 10 feet thick. It underlay 10 to 20 feet of overburden and a thin seam of coal. VANADIUM EFFLORESCENCE Clay W was taken from a newly opened pit a few miles from the clay M working and was also adjacent to a thin seam of coal. Several relatively pure clay minerals were collected so that the effect of fluor- spar on their fired properties might be studied. These included a pure, well crystallized Georgia kaolin, a rather impure and poorly crystallized kaolin from Anna, Illinois, and a sample of illite from Fithian, Illinois, which is re- putedly of high purity except for some free hydrous iron oxide (goethite). The fluorspar used was known as acid grade, that is, its calcium fluoride content was 98 percent. As received, it was light gray and like fine sand in texture. It was ball milled for two hours before use. PLAN OF EXPERIMENT The small amounts of vanadium compounds in clay and in brick make their direct study very difficult, if not impossible. This difficulty, of course, is common in studies involving trace ele- ments in mineral materials. The broad plan of experiment was to follow the vanadium content in certain buff -burn- ing clays that were subjected to heat treatment both with and without the addition of fluorspar. It was planned to do this mainly by determining the water-extractable vanadium in the fired samples. In addition, studies of the original test clays and of the fired products were made in the hope that such studies would throw light on any phenomena involving vanadium in even a secondary way. EXPERIMENTAL PROCEDURES AND TECHNIQUES USED Vanadium Analysis Vanadium analyses were carried out by the well known colorimetric phosphotungstate method described by Sandell (1950) and others. The an- alysis was supplemented in a few in- stances by spectrochemical analysis. Blanks were always run parallel with the analysis. The transmission meas- urements were made at 400 millimi- crons with a Coleman model 14 spectro- photometer. The results were recorded as parts vanadium per million parts of sample dried at 110° C. The reproduc- ibility of the analytical results was about ±10 percent of the amount of vanadium found in the range covered by most of the extraction experiments (30 p.p.m. and less). Most analyses were made in duplicate and the aver- age of the two reported. Sample Preparation Clays S and M were samples that had been ground at the plant. Clays W and L were ground in the laboratory in such manner as to give particle-size distributions as similar as possible to those of the plant-ground materials. When fluorspar was added to the clays, the composition was dry-mixed for 30 minutes in a covered laboratory model Simpson mixer with the rolls set to minimize grinding action. In all com- positions the tempering water was mixed in by hand until a uniform plas- tic mud was obtained. This mud was hand-formed in a steel mold into V X 1" X 6'' bars with 60° wedge tips at one end. The bars were dried thor- oughly before they were fired. The bars were fired in an electric furnace with natural circulation of air. The total firing time was 18 hours in all cases. The time required to reach maximum temperature was 8 to 10 hours, depending on the desired tem- perature. At the end of the firing pe- riod the power was shut off and the furnace allowed to cool for 48 hours ; at 8 ILLINOIS STATE GEOLOGICAL SURVEY the end of this period the fired bars were removed and used for all subse- quent tests. The samples on which differential thermal analyses were to be run were prepared separately in 10-gram batches. For compositions that con- tained fluorspar, the clay and fluorspar were weighed accurately on an analyt- ical balance and ground together in a mullite mortar. The material was mixed thoroughly with water to form a heavy paste, dried at 110° C, and re- ground to pass a 100-mesh sieve. The heating curve was run on this material. The clay (Georgia kaolin, Anna kao- lin, and illite) and fluorspar mixtures to be used for x-ray studies were pre- pared in small batches, as described above, for differential thermal analysis samples. Small amounts of each com- position were fired in platinum cruci- bles on the same schedule described for test bars. Extraction of Vanadium The water-soluble vanadium in the fired samples was determined by first crushing the fired bars and then grind- ing for two hours in a ball mill; 100 grams of ground, fired clay was then subjected to extraction for eight hours in a Soxhlet extractor using 300 milli- liters of water; at the end of the pe- riod the amount of vanadium in the ex- tract was determined. Water Absorption The water absorption of the fired test bars was determined by ASTM method C20-46 (1955). Particle Size Analysis of Raw Clays A particle size analysis in the range 0.5 micron to 30 microns was made on each raw clay. This was done by dis- persing the clay in water contained in a thermostated cylinder and withdraw- ing portions of suspension with a pi- pette at intervals in the usual manner. The portions withdrawn were evapo- rated to dryness and the weight of dry residue determined. X-ray diffraction patterns of each dry residue were made. Clay M contained more coarse mate- rial than the others. It was therefore wet-screened on 100-mesh and 200- mesh screens to determine the percent- age minus 149-microns and minus 74- microns. X-RAY Studies X-ray powder diffraction patterns were prepared for clays fired to various temperatures and containing various amounts of fluorspar. For unfired clays, oriented slides were prepared by allowing dispersed clay to settle out on clean microscope slides which lay at the bottom of the dish that contained the dispersion. The water was then evaporated on a steam bath leaving a layer of oriented clay particles on the slide. Diffraction pat- terns were then made from these slides. The diffraction patterns were the basis of estimates of the amounts and kinds of minerals present in all sam- ples, fired or raw. The estimates were made on the basis of accumulated data and experience of the X-ray Division of the Illinois State Geological Survey. The x-ray equipment used was a General Electric XRD-3 Recording Spectrometer. Differential Thermal Analysis (D.T.A.) A schematic sketch of the apparatus used to determine the heating curves is shown in figure 1. A two-hole nickel VANADIUM EFFLORESCENCE IIOV AC 60 c 0- lOA v-io I T 135 V Voriac Kanthal Furnace Winding, 32 -^ a. b. c. Vacuum Bottle Reference Junction at R.T IIOV. AC Leeds a Northrup c.r\r- — < 60 c X-Y Recorder Speedomox Type 6 Pen X-axis (0^5mV) ( I mV/in.) Compensating Leads Chart Y-axis (IOO°C/in.) (IOO*'-IIOO''C) Thermocouples : a. Crucible temp. 1- Pf 2- pt + 10% Rh b. Standard temp c. Sample temp. 3- Alumel 4- Chromel 5- Alumel Copper Leads Leeds a Northrup D.C. Pre-Amplifier Variable Gain IIOV AC 60 c Fig. 1. — Schematic sketch of differential thermal analysis apparatus. block crucible was used. The sample was uniformly packed in one hole and calcined AI2O3 in the reference hole. The differential thermocouple was made of chromel-alumel wire and the crucible or furnace temperature ther- mocouple was platinum-platinum +10 percent rhodium. The furnace temper- ature thermocouple was inserted in a third hole in the nickel crucible mid- way between the sample and reference material holes. The heating rate was controlled by hand and it was easy, after some practice, to maintain the 10°C/minute temperature rise, used here, within the limits of 9.5° to 10.5° C/min. In some runs a block of nickel, lying loosely on top of the crucible, cov- ered both holes of the crucible. The pre-amplifier in the differential thermocouple circuit made it possible to easily change the sensitivity of the X-axis within wide limits. By chang- ing the multiplier switch one could ob- tain sensitivities of 5.0/x volts/inch to 200|U, volts/inch in six steps. In sev- eral runs the sensitivity was changed in the middle of the run to bring out details of thermal effects. The original diagrams were recorded on chart paper scale 100° C per inch with lines at every 10 °C, making it possible to esti- mate the furnace temperature to a few degrees centigrade. The position of the 10 ILLINOIS STATE GEOLOGICAL SURVEY major peaks was estimated from the original diagrams. It will be noticed in later sections that all thermodiagrams start at 100° C. This is due to the fact that the recorder does not start to re- cord until the furnace has reached this threshold temperature. Base Exchange A few very brief exploratory experi- ments were carried out to see if vana- dium could be exchanged. Three 100- gram samples of each of the unfired clays were placed in 1000 ml. beakers and 500 ml. of 2N solutions of AICI3, CaCl2 and NH4C2H3O2 were added. These were each stirred for 15 minutes with a mechanical stirrer and then placed on the steam bath for 24 hours. After digestion the clay was filtered off, using a Buchner vacuum funnel, and the filtrate concentrated to 100 ml. This concentrated filtrate was analyzed for vanadium by the spectrochemical method. Wick Tests The wick test consisted of standing the fired bars on end in distilled water and recording the appearance of the material on the tapered tip after seven days. The 60° tapered tip served to con- centrate the efflorescent salts and made the method quite sensitive. Some sam- Table 1.- -Particle Size Analysis of the Unfired Clays Particle size Weight percent finer than (Microns) Clay W Clay L Clay S Clay M 149 _ _ 93.2 74 — — — 64.5 30 97.5 47.8 91.3 55.0 20 93.4 44.3 86.5 51.7 10 82.2 38.8 76.3 47.8 5 59.8 32.9 58.4 41.4 2 30.7 23.5 32.5 29.7 1 19.4 16.9 22.1 21.2 0.5 12.3 12.0 14.5 16.4 pies of the efflorescent material on the tips were studied to determine metals and ions present. EXPERIMENTAL RESULTS Properties and Composition of Unfired Clays The results of the particle size an- alysis are given in table 1 and shown graphically in figure 2. 20 30 40 50 60 70 80 90 Weight Percent Finer Than Fig. 2. — Particle size distribution of unfired clays. From these results it appears that clays W and S are very fine, nearly 100 percent is less than 30 microns, where- as clays M and L are comparatively coarse, only 50 percent of the clay be- ing less than 30 microns. However, the amount of very fine material (less than 1/2 micron) is approximately the same for all clays. The relative amounts of free quartz, illite, kaolinite, and montmorillonite+ chlorite were estimated from the ori- ented slide type of x-ray pattern. It must be emphasized that these are only VANADIUM EFFLORESCENCE 11 Table 2. — Estimates of Quartz and Clay Mineral Contents of the Clays Particle Clay S Clay W Clay M (Microns) Q* I* K* Q I K Q I K M + C* 149 52 15 30 3 30 22 50 27 25 60 18 23 23 53 — 20 18 52 27 28 55 20 20 20 53 — 10 18 55 30 25 57 19 15 16 60 — 5 10 57 33 20 60 20 10 15 63 10 2 5 57 38 10 66 23 15 65 20 1 57 43 72 28 15 65 20 quartz; I=:illite; K:=kaolinite; M+C = montmorillonite and chlorite. estimates and do not include other minor impurities such as pyrite, cal- cite, organic matter, etc. However, they are accurate enough to show up major differences between the clays. The re- sults are shown in table 2. X-ray pat- terns of clay L were not made ; there- fore, its composition was not esti- mated. Clays S and W contained no mont- morillonite+ chlorite and clay M con- tained only a small amount. The esti- mates for clays S and W, at 30-micron particle size, are taken to represent ap- proximately the whole clay. However, for clay M the whole clay composition is probably represented fairly closely only at 149 microns. Thus the total clay mineral content of clays S and W is nearly 80 percent of the clay, but for clay M it is approximately 45 percent. Clay M has twice the free quartz con- tent of either S or W. A large portion of the quartz occurs as coarse material in whole clay M. In clays S and W the quartz content is much lower and its particle size is distributed over a lower range. So the difference between the particle size distribution curves of clays S and W compared to that of M is due predominately to differences in quartz grain size. The ratio of illite to kaolinite in the whole clay S is about two to one, in whole clay W it is three to one, but in whole clay M it is 0.5 to one. This information is believed to be vital, as is shown below. The vanadium content of the unfired whole clays and two fractions, one less than 2 microns and one more than 2 microns, were determined. Results of the analysis are shown in table 3. Whole clays S and W have higher va- nadium contents than M and L. This may be due to the fact that S and W have higher clay mineral contents. These data (table 3) indicate that the vanadium is more concentrated in the finer fractions of the whole clays. Table 3. — Vanadium Content of Unfired Clays In parts per million Clay Fraction Clay S Clay W Clay M Clay L Whole clay 138 140 56 70 >2 microns 83 88 13 15 < 2 microns 144 142 84 143 In order to determine the water ex- tractability of vanadium in the unfired whole clays, samples of each dried clay were extracted for 8 hours in a Soxhlet continuous extractor. The results showed for clay L, 0.30 p.p.m. of vana- dium in the extract ; clay W, 0.80 ; clay M, 0.40; and clay S, 0.40. 12 ILLINOIS STATE GEOLOGICAL SURVEY 900 1000 Temperature (°C) Fig. 3. — ^Effect of fluorspar content and firing temperature on vanadium extracted and on water absorption for clay S. If the vanadium were associated with the clay minerals, it was thought that it might be in an exchangeable posi- tion. An attempt was made to ex- change AP^ Ca2% and NH^* cations. The results showed that no detectable amount of vanadium could be ex- changed for the above ions in any of the clays. Properties of Fired Clays This section centers around the ef- fect of temperature and fluorspar con- tent on the extractability of vanadium from clays and an investigation of other properties which might directly or indirectly shed light on the mech- anism of these two effects. The effect of firing temperature and fluorspar content on the amount of va- nadium extractable by the Soxhlet 17 16 15 14 y^' ;;^:^^^~Cr^' -\4 k^^^^-:^^: "^ Water Absorption ] ^>0^ N N s^% Fluorspar 1? 1 \ \ A ^ \ / \ 1 \ e \ IN / Vonodium Extracted \ / \0% Fluorspar -\ / ,/:^ — ^ ' \ [^ r °~^~ __4 ^ K \ -^. ^ ' 800 900 1000 1100 1200 1300 Temperolure (°C) Fig. 4. — Effect of fluorspar content and firing temperature on vanadium extracted and on water absorption for clay M. method and the water absorption of the fired clay bars are given in table 4. A portion of the data in table 4 for clays M and S are shown in figures 3 and 4. Several things stand out in these figures. First, the amount of va- nadium extracted remains constant and small until the temperature reaches 650° to 700°C, then the amount of vanadium extracted rises rapidly with temperature to about 950°C, after which it drops. It will be noted that the drop in the amount of extracted vanadium with respect to rising temperature is much more rapid for clay S than for clay M. Second, the water absorption of both clays S and M reaches a maximum at about 950 °C, but for clay S it drops much more rapidly with rising temperature than for clay M. It was noticed that clay S began to overfire between 1200° and 1300° C, whereas clay M showed no signs of overfiring at 1300 °C. Clay W also showed signs of overfiring at 1200° to 1300°C. VANADIUM EFFLORESCENCE 13 Table 4. — Effect of Firing Temperature and Fluorspar Content on Vanadium Extractability (by Soxhlet Method) and Water Absorption of Clays S, M, L and W Top figure — vanadium p. p.m. Bottom figure — water absorption, percent Clay Fluorspar content - (percent) Degrees C 100 650 725 800 850 900 S 0.40 0.12 16.1 . 1.5 16.4 — 6.1 16.7 S y2 — — 0.70 17.0 1.6 16.5 2.5 16.8 4.5 16.4 S 1 — — 0.92 16.5 1.9 15.9 3.1 16.6 3.2 16.7 s 2 — — 0.65 16.2 1.8 15.6 2.9 16.1 2.5 16.6 s 4 — — 1.4 15.8 4.0 15.5 3.6 16.4 15.1 M 0.40 0.55 13.6 — 2.3 — 7.4 14.9 M Vi — — — 1.6 15.6 — 2.1 15.6 M 1 — — 1.8 14.8 — 1.7 14.9 M 2 — — — 1.9 15.2 — 1.6 16.1 M 4 — — — 1.6 14.5 — 2.0 15.9 L 0.30 0.45 14.1 — 2.2 15.0 — 5.2 14.7 W 0.80 0.12 19.3 — 3.1 18.2 — 8.5 18.1 Fluorspar, in general, reduces the amount of vanadium extracted from both clays S and M. The shape of the curves for clay S at one-half and two percent fluorspar shows that the ex- traction peak has been shifted back to lower temperatures, but for clay M the peak at one and two percent fluorspar has been slightly shifted to higher temperatures. The amount of vanadium extracted compared to the amount of vanadium in the whole clay is rather small. It was suspected that possibly the Soxh- let method of extraction was not re- moving all the soluble vanadium. So to check the extraction efficiency of the Soxhlet method, clay S, fired at 900 °C and containing zero to four percent fluorspar, was extracted in a continu- ous flow column extractor, which was thought to be more efficient. The pro- cedure was to first extract by the usual Soxhlet method, then extract this same clay with several 500 cc charges of water in the column extractor. Each 500 cc through the column was ana- lyzed for vanadium and the results are shown in table 5. 14 ILLINOIS STATE GEOLOGICAL SURVEY Table 4. — Continued Clay Fluorspar content — (percent) Degrees C 950 1000 1100 1200 1300 S 8.5 16.0 6.1 12.6 1.3 7.6 0.67 5.3 0.70 10.6 S y2 3.3 16.0 1.6 11.5 0.42 5.3 — — S 1 1.8 14.7 1.1 11.0 0.40 5.0 — — S 2 0.90 13.0 0.93 10.3 0.20 4.7 — — S 4 1.2 12.6 0.86 10.3 0.22 6.7 — — M 8.6 15.5 6.7 15.3 4.5 15.0 2.7 12.7 0.92 12.1 M Vi 3.3 15.8 2.7 15.9 2.2 15.7 1.1 14.1 — M 1 2.7 16.0 2.9 16.2 2.2 15.7 0.85 14.6 — M 2 2.1 16.6 2.7 15.2 1.4 14.6 0.55 14.7 — M 4 1.6 16.9 1.3 17.0 1.3 16.5 0.50 15.9 — L 7.3 15.7 6.4 15.3 2.3 15.2 1.6 12.2 0.77 11.7 L 9.2 16.7 4.6 12.8 0.80 2.5 0.67 4.2 0.80 13.2 Table 5. — Vanadium Extracted (p.p.m.) BY THE Soxhlet Method AND THE COLl UMN Method Firing temperature 900°C Column method Total Fluorspar (percent) Soxhlet After 8 hrs. Clay After 1st 500 cc After 2nd 500 cc After 3rd 500 cc (both methods) S S S S S 1 2 4 6.1 4.5 3.2 2.5 4.7 2.6 2.2 2.2 1.2 0.7 0.6 0.7 0.4 0.4 0.2 0.3 12.4 8.7 6.7 5.4 5.9 It looks as though eight hours of Soxhlet extraction removes about one- half the total amount of extractable vanadium removable by the combined methods. It also appears that the Soxhlet method is a relative test pro- cedure and not an absolute method; however, it should be noted that the general trend of the vanadium ex- tracted vs. the fluorspar content for the Soxhlet method is the same as the totals column in table 5. VANADIUM EFFLORESCENCE 15 Table 6. — Total Vanadium Content of Clays S, M, L, and W Fired to Various Temperatures Vanadium in p. p.m. Fluorspar (percent) Degrees C Clay 100 900 1000 1200 1300 S 138 _ 144 122 100 S ¥2 — — ■ 146 — S 1 — — 129 — S 2 — 146 135 — S 4 — 132 142 — — M 56 56 56 54 44 M Vi — 52 43 — . M 1 — 43 50 . M 2 — 35 50 . M 4 — 45 — — — L 70 — 72 66 38 W 140 — 141 143 108 The amount of vanadium extracted under the best conditions is small com- pared to the total amount present in the raw clay, indicating that the con- version of vanadium from the insolu- ble to the soluble form is incomplete. Firing to high temperatures or the use of fluorspar appear to reduce the amount of vanadium extracted. It has been reported that vanadium is lost when mixtures of calcium fluoride and vanadium pentoxide are heated to- gether (Prandtl and Mautz, 1911), a fact which suggests that fluorspar may control efflorescence by promoting elim- ination of vanadium by volatilization. To check this, clays S and M, fired to various temperatures and containing various amounts of fluorspar were an- alyzed for their vanadium content. The results are shown in table 6. It appears that fluorspar does not cause any significant loss of vanadium, at least within the reliability of the data and at temperatures up to 1000 °C in a clean oxidizing atmosphere. The data for all clays which contain no fluorspar indicate some loss of vana- dium at temperatures of 1200° to 1300°C. The bars of fired clay were allowed to stand for seven days in distilled water. At the end of this period the color of the tips of the bars was re- corded as shown in table 7. No attempt was made to estimate the amounts of efflorescence except for clay M. The dotted lines in table 7 show the approximate concentrations of fluor- spar and firing temperatures required to eliminate the appearance of colored efflorescence on the wick test bars. From the results of table 7 the bars containing no fluorspar show colored efflorescence in the vicinity of the ex- tractability peak shown in table 4 and figures 3 and 4. On the low-tempera- ture side of the peak, that is up to about 800° C, white efflorescence is formed, then colored efflorescence up to 950° C. On the high-temperature side of the extractability peak, colored efflores- cence is present until the temperature is high enough to eliminate all efflor- escence. In clay S fluorspar effectively eliminates the colored efflorescence at lower temperatures ; however, it is not nearly as effective in clay M. The efflorescence formed on the tip of the bars of clays S, L, and W was a 16 ILLINOIS STATE GEOLOGICAL SURVEY Table 7. — ^Wick Test Results After Seven Days N = No efflorescence W = White efflorescence only C = Colored efflorescence easily detected # = Colored efflorescence detected only by close examination Lines indicate approximate limits of colored efflorescence Fluorspar (percent) Degrees C Clay 650 725 800 850 900 950 1000 1100 1200 1300 S W — W — C c c N N — S 1 2 — W w w W w w w w w C C c c c N N N — ^ S N N — S N N — S 4 — w w w N N N N — — M W — w — C C c c c N M Vi — — w — C C c C C — M 1 2 : : w w : c c c c c c c# c# C# — M N — M 4 — — w — c c c% c# N — L w — C — c c c c N N W w — w — c c c N N N yellow-orange color but that on bars of clay M showed much more green. The efflorescent material on the tip of bars made from clays S, L, M and W con- taining no fluorspar and fired to 950 °C and 1000° C was examined spectro- chemically. The results are given in table 8. The numbers indicate only relative abundances and are not to be taken as representing composition in any units. Microchemical tests showed that a considerable amount of sulfate was present in this efflorescent mate- rial. These results show that vanadium is rather abundant in the efflorescence formed on clays S and W and is fairly abundant in that formed on clays L and M. Molybdenum occurs in the ef- florescence on all the clays, but is not as abundant as vanadium. It is probably a fair guess to say that vanadium com- pounds with Na, K, and Ca, are the main sources of the color in the efflor- escent material. Colored molybdates are possibly contributing some color, as indicated by Palmer (1929). Fluorspar in amounts larger than 2 percent darkens the fired color of both clays S and M at temperatures above 900° to 1000°C. An x-ray examination of clays S and M was carried out. Table 9 summarizes the crystalline phases identified in each sample. The quartz and anatase are impuri- ties in the raw clay, but the rutile and hematite are formed from impurities when the clay is fired. The illite (as the anhydrous modification) (Grim and VANADIUM EFFLORESCENCE 17 Table 8.- -Spectrochemical Estimate OF THE Relative Abundance of Metals in Colored Efflorescence No fluorspar Clay Metal L M S W Fired to 950° V 14 18 38 47 Na 7 4 39 24 K 15 18 19 24 Ca 59 56 Tr.* Tr. Mg 4 4 4 5 Mo 2 6 16 14 Cr Tr. Tr. Tr. Tr. Sr Tr. Tr. Tr. Tr. Fe Tr. N. D.* ] M. D. N. D. Li N. D. Tr. Fired to 1000° G Tr. Tr. V 10 15 27 44 Na 6 5 8 13 K 32 37 38 37 Ca 47 37 25 5 Mg 4 6 2 1 Mo 0.5 6 8 5 *Tr.= trace; N.D. =:not detected. Bradley, 1940) in clays S and M de- creases in amount until it disappears between 850° and 950°C. In clay S, which contains more illite than clay M, some spinel is formed at 800 °C and is still detectable at 1000°C when no fluorspar is present, but when four per- cent fluorspar is present no spinel is detected above 900° C. Up to about 900 °C the fluorspar remains in the clay and is identifiable as such in both clays S and M. Fluorspar is detected only at four percent concentration because less than this cannot be detected by x-rays. By 950 °C the fluorspar is no longer detectable but it has produced changes in the detectable crystalline phases that are not observed in the ab- sence of fluorspar until much higher temperatures. Clay S that contains two percent fluorspar forms some detectable mullite at 950 °C, but at four percent fluor- spar, mullite is clearly detectable along with some feldspar. It was not possi- ble to determine which feldspar is formed, but it is believed to be anor- thite. At 1000 °C mullite can be de- tected at one-half percent fluorspar, and at two percent fluorspar the feld- spar phase is also present. At 1100° C mullite is detected for the first time in the untreated clay, but feldspar is not detected until at least two percent fluorspar is present. In general, the free quartz content decreases gradu- ally from 900 °C on in clays with and without fluorspar; however, in the clay that contains fluorspar the de- crease is much more rapid. Four per- cent fluorspar at 1100° C causes almost a complete disappearance of the free quartz. In review, the addition of fluorspar to clay S produces mullite, in amounts detectable by x-ray examination, at temperatures 100° to 150°C lower than that required to produce mullite in the absence of fluorspar. Fluorspar in clay S also produces a feldspar which is not detected at any temperature in the ab- sence of fluorspar. It will also be ob- served that no cristobalite is observed at any temperature or fluorspar con- centration. This is probably due to free quartz being used up in forming glass before it can invert to cristoba- lite. The very fine particle size of the free quartz would allow it to form glasses readily with the high flux con- tent of the illite and the added fluor- spar. Clay M, containing two percent and four percent fluorspar, shows mullite and cristobalite at 950°C. At 1000°C only one-half percent of fluorspar pro- duces some mullite, and with two per- cent, cristobalite and feldspar are forming in addition to mullite. Near 1100°C, mullite and cristobalite are de- 18 ILLINOIS STATE GEOLOGICAL SURVEY Table 9. — The Crystalline Phases Identified in Clays S and M by X-Ray Diffraction Clay Fluorspar (percent) Degrees C 800 850 900 950 1000 1100 1200 1300 S I, Q SI, A I. Q A, SI I, Q A, SI I, Q R, A Q, SI R, A Mu, Q — Mu, Q S 'A a a a « Mu, Q Mu, Q - — S 1 u (( a Q Mu, Q Mu, Q - - — S 2 u u u Mu, Q Mu, Q F Mu, Q F - - — S-- 4 I, Q SI, Sp. A I- Q SI, Sp. A si, Sp. A Mu, Q F Mu, Q F Mu, Q F — — M I. Q A — A Q, A R, H Q, H Q, Mu C, H 8: Mu H Q, Mu M Vi a — u u Q, Mu Q, Mu C, R H c Mu — M 1 u u Q, Mu A, R H Q, Mu Q, Mu C, F R, A H 8; Mu F M 2 u — u Q. Mu C, F H Q, Mu C,F R Q, Mu C, F R, A Q, c, Mu F — M 4 I, Q A, Sp. A, Sp. Q, Mu C, F H Q, Mu C, F R Q, Mu C, F R, A c. Mu F — Mu = mullite; Q=: quartz; C = cristobalite; F = feldspar; R = rutile; Azzanatase; H = hematite; I = illite; SI = spinel; Sp. = fluorspar. tected for the first time in the absence of fluorspar, and with one percent fluorspar content the feldspar phase begins to appear. The effect of fluor- spar on clay M is similar to that on clay S. The main difference is that cristo- balite is produced from clay M and none is produced by firing clay S under any condition used here. The free quartz content of clay M decreases only a very little up to 1200° C, even with the addition of fluorspar. What de- crease takes place is probably due to conversion to cristobalite, since the amount of flux constituent (illite) is low and therefore little glass is formed. Also a hindering factor in the solution of the free quartz in the small amount of glass is the large particle size of most of the free quartz. In review, it looks as though fluor- spar aids the formation and growth of mullite crystals at much lower temper- atures than they otherwise would de- velop, thus permitting mullite to be de- tected by x-ray diffraction techniques. Without fluorspar, mullite of sufficient crystallinity to be detected by x-rays is produced, but only at a 150 °C higher temperature. Feldspar is produced in clays S and M only when heated with fluorspar. It is not known whether the feldspar is formed on heating or crys- tallizes from a glassy phase on slow VANADIUM EFFLORESCENCE 19 lOO^C 200* 300* 400* 500* 600* 700* 800* Fig. 5. — -Effect of fluorspar on the D.T.A. diagram of clay S. 900" 1000' cooling. When the free quartz is low and very fine grained, when the illite, which is a flux-carrying material, is high, and the kaolinite content is low, cristobalite is not detected. When the quartz is very coarse, little flux ma- terial is present, and the kaolinite con- tent is high, cristobalite is detected for the first time at 1100° C without fluor- spar and at 950 °C with four percent fluorspar. This cristobalite could be from decomposition of the kaolinite or from inversion of the quartz. The differential heating curves for clays S and M containing zero, one, two, and four percent fluorspar are shown in figures 5 and 6. The heating rate for both clays was 10°C/min. and in both cases the nickel sample holder was covered with a nickel block. In figure 5, the curve for zero fluorspar was determined at sensitivity 10 (50 microvolts/in.) up to 500 °C then switched to scale 20 (100 micro- volts/in.) to 600 °C then back to scale 10 for the remainder of the diagram. The rest of the curves in figure 5 and all the curves in figure 6 were run on scale 20 to 800°C then on scale 10 to 1000°C. In figure 5, the presence of the endo- therms at 150° and 580°C and the exo- therm at 950° C would indicate the presence of only poorly crystalline kao- lin and the illite content (known to be present from x-ray analysis) would be missed entirely, as shown by Grim (1947) in his work on mixtures of clay minerals. The small endothermal peak at about 450 °C is due to pyrite (Grim 20 ILLINOIS STATE GEOLOGICAL SURVEY lOCC zoo** 300* 400** 500" 600* 700* 800* Fig. 6. — Effect of fluorspar on the D.T.A. diagram of clay M. 900* 1000* and Rowland, 1944) . This peak is exo- thermic when the crucible is not cov- ered and there is sufficient oxygen to burn the sulfur as it is released and thus produce heat. It appears that the oxidation of the sulfur (from the de- composition of the pyrite) and the or- ganic matter produces the broad exo- thermic hump between 600° and 700 °C. In short, the restricted oxygen supply pushes the elimination of the sulfur and organic matter to higher temperatures. Figure 5 shows that the addition of fluorspar has little effect on the shape of the curves up to about 900 °C. How- ever the exothermic peak between 900° and 1000 °C is sharpened and shifted to lower temperatures, and at four per- cent fluorspar a new exotherm is noted at 950 °C. The alteration of the posi- tion and shape of the exothermal peak between 900° and 950 °C coincides with the temperature at which fluorspar disappears and mullite and feldspar appear in the x-ray analysis (see table 9). This shift of the exothermal peak is linear for clay S, as shown in figure 7. This indicates that the fluorspar lowers the temperature at which the mullite nuclei first appear. The fact that it can be detected by x-ray an- alysis as soon as it is formed indicates that fluorspar also influences the growth and perfection of the mullite nuclei. The second faint exothermal peak at 950 °C may be associated with perfection of the mullite nuclei or pos- sibly with the formation of the feld- spar phase. These observations agree essentially with the current literature except that VANADIUM EFFLORESCENCE 21 940 910 K ^ \ \ L ^ "^ Clay M Clay S 2 3 Percentage of fluorspor Fig, 7. — Effect of fluorspar content on the major exothermal peak temperature. the feldspar phase is not reported elsewhere. Palmeri (1951) studied with x-rays and the electron micro- scope the effect of mineralizers on the mullitization of kaolinite. He found that in the absence of a mineralizer the kaolinite retained its characteristic shape up to 1190°C, but when three percent CaF2 is added, the particle shape disappears at less than 1000 °C and mullite is detected in the x-ray pat- tern. Nazarenko and Sviridenko (1953) studied the influence of min- eralizers on mullite formation from AI2O3 and Si02 (quartz). They found LiF the most effective mineralizer and CaF2 only slightly less effective. They showed that the fluorides were more effective than the corresponding car- bonates. The curves in figure 6 are very simi- lar to those in figure 5 except that the exothermal peak between 600° and 700 °C has been ehminated. This prob- ably indicates less organic matter in clay M. The peak between 900° and 1000 °C is sharpened up slightly and shifted to lower temperatures as for clay S. However, figure 7 shows that the shift is not linear as it was for clay S. A second small exothermal peak in clay M appears at about 950 °C when the fluorspar content is two percent or more. The reason for the appearance of this second small exotherm at 950 °C may be the same as given for clay S, that is, it could be associated with the mullite and/or feldspar development. Influence of Fluorspar on the Fired Properties of "Pure'' Clay Minerals Clays S and M contain mixtures of the clay minerals illite and kaolinite, so in an attempt to get a clearer picture of the influence of heat and fluorspar on these minerals, more or less pure samples of each were investigated with two and four percent fluorspar addi- tions. The x-ray patterns were ana- lyzed after firing to various tempera- tures and the differential heating curves were run. A general discussion of the influence of heat and fluorspar on the clay minerals follows. The Georgia kaolin was a highly crystalline mineral with about five to ten percent of impurities, which were mainly free quartz, anatase, and illite. The illite sample from Illinois contained no free quartz, but contained some hy- drous iron oxides (as goethite) as shown in the D.T.A. diagram (fig. 10) by the endothermic peak between 350° and 400°C (Grim and Rowland, 1942). The kaolin from Anna, Illinois, was of very low crystallinity and contained a large amount (20 to 25 percent) of free quartz. The crystalline phases identi- fied in each fired sample are summa- rized in table 10. On heating the Georgia kaolin that contained no fluorspar, the only new crystalline phases up to 950 °C are due to impurities. At 1000 °C some y-Al203 22 ILLINOIS STATE GEOLOGICAL SURVEY Table 10. -Influence of Fluorspar on the Crystalline Phases Identified in Georgia Kaolin, Anna Kaolin, and Illite Clay Fluorspar (percent) - Degrees C 900 950 1000 1100 1200 Georgia kaolin 2 I, A u I, A 7-Al,03 A, Mu Mu Mu, F, C Mu Mu, F, C u 4 A, Sp. Mu, F Mu, F Mu, F, C Mu, F, C Anna kaolin Q Q Q Q, Mu Mu, Q, C u 4 Q Mu, Q Mu, Q, F Mu, Q. F C Mu, Q, F C Illite I, H SI, H SI, H Mu, SI, H Mu, SI H a 2 u Mu, SI, H u SI, H a u 4 I, H, SI Sp. Mu, SI, H F SI, H, F u SI, H Mu=:mullite; Q= quartz; C = cristobalite; F = feldspar; R = rutile; Arranatase; H=: hematite; I = illite; 81 = spinel; Sp. = fluorspar. and mullite are present and at 1100° C only mullite is detected. When two per- cent fluorspar is added the same phases are present up to 1000°C, but at 1100° feldspar and cristobalite are detected, in addition to mullite. At four percent fluorspar the unreacted CaF2 was de- tected at 900 °C, but at 950° C only mul- lite and feldspar were detected and by 1100 °C cristobalite has also appeared. It should also be noted that no visual evidence of vitrification was apparent in any of these samples. The Anna kaolin shows no detecta- ble mullite until 1100°C and at 1200°C cristobalite also is detected. At four percent fluorspar mullite is detected at 950°C; at 1000°C feldspar and mullite are present and at 1200 °C, mullite, feldspar, and cristobalite are detected. The free quartz content remains the same to 1100 °C when no fluorspar is present, then decreases slightly at 1200 °C. However, when fluorspar is present the free quartz decreases grad- ually from 1000°C on, and at 1200°C it is less than half the original content. It should be noted that Anna kaolin itself became progressively harder to grind as the temperature increased, but when four percent fluorspar was added the difficulty in grinding in- creased markedly with temperature. This may indicate some vitrification. X-ray patterns of the illite speci- mens showed a general lack of crystal- line material and in all cases where crystalline phases did develop the amounts were very small. Consequent- ly, the interpretation of the illite pat- terns given in table 10 is very poor and not considered very reliable. The pres- ence of fluorspar seems to promote mullite and feldspar formation at 950°C to 1000°C; however, by 1100°C to 1200° C neither can be detected. It is possible that the fluorspar has pro- moted the solution of the feldspar and mullite in the glass at high tempera- tures. Figure 8 shows that the only influ- ence of fluorspar on Georgia kaolin is the shifting of the 960 °C exothermic peak to slightly lower temperatures. VANADIUM EFFLORESCENCE 23 IOO*'C 200** 300*^ 400" 500** 600* 700* 800* 900* 1000' Fig. 8. — Effect of fluorspar on the D.T.A. diagram of Georgia kaolin. At four percent fluorspar the shift amounts to about 15° C. The thermo- gram for Anna kaolin (fig. 9) shows a shift of the 940° exothermic peak to about 920 °C. A second small exother- mic peak between 900° and 1000°C does not appear in either of these clays. So the effect of fluorspar on well crystallized and poorly crystallized kaolins appears to be very similar up to 1000 °C in spite of the large amount of free quartz impurity in the Anna kaolin. In figure 10, at zero percent fluor- spar, the only peak not due to illite is the slight endotherm between 350° and 400 °C. This has been interpreted as coming from geothite (FeO(OH)) (Grim and Rowland, 1942), an impur- ity in the clay. On the addition of two percent fluorspar the only difference is a sharpening up of the exotherm at about 920 °C. On the addition of four percent fluorspar the intensity of the endotherm at about 900 °C and the exo- therm at 930 °C is much diminished and at 950 °C a second small exotherm appears. According to Grim and Row- land (1942), (for illite containing no fluorspar) the first endotherm at about 160 °C is due to the loss of adsorbed moisture. The next endotherm, which is associated with the illite, is at about 580 °C ; this is where most of the lattice water is lost forming the anhydrous modification. From 580 °C on, the an- 24 ILLINOIS STATE GEOLOGICAL SURVEY IOO°C 200*> 300** 400^* 500** 600* 700* 800* 900* Fig. 9. — Effect of fluorspar on the D.T.A. diagram of Anna, Illinois, kaolin. 1000* hydrous modification of the illite slow- ly disappears and the endotherm at about 880 °C is the point at which the anhydrous modification is destroyed. The exotherm at about 920 °C is asso- ciated with the formation of the spinel. From this, one may guess that the slight exotherm at 950 °C (in the sam- ple containing four percent fluorspar) may well be due to crystallization of mullite or feldspar. Discussion of the Experimental Results The clays studied here fall into two general categories. Clays S and W are fine-grained and contain about 80 per- cent clay minerals, which are com- posed largely of illite. These clays are not highly refractory, as is shown by the fact that water absorption values fall rapidly above 950 °C and that they overfire between 1200° and 1300° C. When fluorspar is present they overfire at lower temperatures. The high illite content is probably responsible for the low refractoriness of these clays. Clays M and L are coarse-grained and clay M contains only 50 percent clay minerals which are made up large- ly of kaolinite. The coarse fraction of clay M is mainly quartz. Both of these clays are refractory, as indicated by water absorption values. In clay M this is attributed to the high kaolinite and quartz and low illite content. In fact, clay M is currently being produced and sold as a refractory for low- to moder- ate-temperature use. The clays of both categories show very little extractable vanadium after heating up to 650 °C. However, after VANADIUM EFFLORESCENCE 25 IOO**C 200*' 300* 400** SOO** 600* 700* 800* Fig. 10. — Effect of fluorspar on the D.T.A. diagram of illite. 900' 1000" heating above this temperature, the amount of vanadium extracted rises rapidly until a maximum is reached at around 950 °C. Heating above this tem- perature causes the amount of extract- able vanadium to fall. However, clays S and W show a more rapid decline than clays M and L. The curves show- ing water absorption versus firing tem- perature show this same trend. For clays S and W the water absorption falls very rapidly above 950 °C whereas clays M and L show only a gradual de- crease in water absorption and even when heated to 1300 °C never reach the low values that clays S and W attain. Figures 3 and 4 show these relation- ships for clays S and M, which are rep- resentative members of the two major categories. From this it appears that the amount of vanadium extractable follows the water absorption curves when no fluorspar is preisent. When the water absorption curve falls rap- idly with rising temperature, the ex- tractability curve also falls rapidly. The general shape of the curves in figures 3 and 4, when no fluorspar is present is consistent with two succes- sive mechanisms. From 650° to 950 °C a reaction is taking place that pro- duces the highly colored, water-soluble vanadium compounds from the insolu- ble compounds in the raw clay. The water-soluble compounds are believed to be mainly vanadium compounds with Na, K, and Ca. At 950 °C this reaction appears to be arrested by the onset of vitrification, and above 950 °C the vit- reous phase formed from the clay seems to take the water-soluble vana- dium compounds into solution, or in 26 ILLINOIS STATE GEOLOGICAL SURVEY essence make them water-insoluble. If a greater amount of vitreous phase is formed at a given temperature above 950°C, more of the soluble vanadium compounds will be dissolved in this vit- reous phase and hence the extractabil- ity curve will fall more rapidly. If this is correlated with the clay mineral compositions of the clays, one sees that clay S, with the highest illite content, shows the most rapid decrease in the amount of extractable vana- dium. It is known that illite begins to vitrify (Grim and Rowland, 1942) near 950 °C, the peak extractability temper- ature, and with rising temperatures produces more vitreous phase with which to take water-soluble vanadates into solution. Clay M, which contains a much smaller proportion of illite, produces much less vitreous phase, hence the decline in the extractability curve is not as rapid. It is of course possible that this may be just fortui- tous and that other unsuspected fac- tors may be more important than the type of clay minerals present. The wick test results correspond very well, visually, with the extraction curves shown in figures 3 and 4. It was noted for clay S (containing no fluor- spar) that when a sufficiently high tem- perature has been reached, where the extractable vanadium has fallen to a level of one to two p.p.m., no colored efflorescence is noted on the bar tips after a seven-day wick test, but for clay M (containing no fluorspar) the extractable vanadium must be less than one p.p.m. to be assured of no col- ored efflorescence. There are, of course, some exceptions, but these appear to be good figures for the major part of the experimental data. Clay M, con- taining no fluorspar, must be fired to 1200° to 1300 °C to produce a bar with no seven-day efflorescence. For clay L the corresponding temperature is 1100° to 1200°C and for clays S and W it is 1000° to 1100°C. When four per- cent fluorspar is present clay S shows no seven-day efflorescence after firing to 900° to 950° C but clay M must be fired at 1100° to 1200°C to effectively eliminate colored efflorescence. It is felt that the rate of decline of the extractabiHty curve, above 950 °C, is an important factor in controlling vanadium staining without the addi- tion of fluorspar. The ideal situation would be a sudden drop from the peak to less than one p.p.m. in the tempera- ture range of 950° to 1050°C, the usual firing temperature employed in indus- trial manufacture. Since this is not the usual case, fluorspar can be added to help control the vanadium efflores- cence. Figures 3 and 4 show the inffuence of fluorspar on clays S and M. The fluorspar greatly reduces the amount of extractable vanadium in both clays S and M. It will be noted for clay S, in figure 3, that with increased fluorspar content the extractability peak was shifted to lower temperatures. This indicates that the arresting reaction has started at lower temperatures. However, the slope of the extraction curve at temperatures above the ex- tractability peak is about the same as it was at zero percent fluorspar. Clay M shows no shift in the extract- ability peak to lower temperatures until the concentration of fluorspar reaches four percent. The slope above the peak, that is, on the high tempera- ture side, at all fluorspar concentra- tions is about the same. Evidently the fluorspar does not change the slope of the curve above the extractability peak, but does cause changes in the properties of the glass phase and vari- VANADIUM EFFLORESCENCE 27 ous other changes which tie up the soluble vanadates at lower tempera- tures. The shift in the D.T.A. curves and the x-ray analysis show that adding fluorspar apparently promotes the growth of crystalline phases at lower temperatures. The fact that new phases are developed at lower tempera- tures probably has no direct connec- tion with vanadium efflorescence. How- ever the breakup of the clay lattice, the variation in the extractable vana- dium, and the appearance of new phases all within the same limited tem- perature range, suggest strongly that these phenomena are in some way in- terrelated. It is a well established fact that new reactions take place more readily immediately following the break up of existing phases. Figure 11 shows water absorption versus the vanadium extracted, for all temperatures above the extractability peak shown in figure 3. This shows that at zero percent fluorspar the amount of extractable vanadium fol- lows the water absorption. However, when fluorspar is introduced the amount of extractable vanadium falls rapidly, although the water absorption remains constant for a considerable pe- riod at temperatures of 800° to 950 °C. This would seem to indicate that the amount of extractable vanadium is greatly decreased without increasing the amount of glassy phase (lowering of the water absorption) . So from this it looks as though the fluorspar must be influencing the properties of the vit- reous phase already present. Recent work by Yee, Machin, and Andrews (1955) has shown that fluorspar de- creases the viscosity of enamel glasses. Three percent of fluorspar reduced the viscosity 50 percent at 900 °C. This suggests that the fluorspar may be re- t 5 o Contoining^ Fl^iorspor » ^^,A-''t)% Fluorspar ^^ 0.6 08 1.0 2.0 Vonodium Exfrocted (p.p. m.) Fig. 11. — Relationship between water absorption and extracted vanadium for clay S. ducing the viscosity of the vitreous phase formed in the clays so that it can diffuse more readily through the struc- ture and dissolve more of the vana- dium-containing compounds, thus ty- ing up the vanadium so that it cannot be water-extracted. At the same time, this more fluid liquid may be catalyz- ing the formation and growth of new phases at lower temperatures. The water absorption, after the constant in- terval, drops rather rapidly, beginning at temperatures of 950 °C or higher, indicating that additional glass is forming, which lowers the water ab- sorption of the bar and further de- creases the amount of extractable va- nadium. Whatever the true mechanism may be, it is important to notice that low concentrations of fluorspar are effec- tive in preventing vanadium efflores- cence at low firing temperatures only when certain types of clay minerals are present. In short, the clay minerals must in themselves be conducive to vit- rification if small concentrations of 28 ILLINOIS STATE GEOLOGICAL SURVEY fluorspar are to be effective in control- ling vanadium efflorescence in the usual commercial firing range. Another important fact to be noted is that all the firing operations in the present study were carried out in an electric furnace with a clean, oxidizing atmosphere. The influence of combus- tion gases in commercial coal or gas- fired kilns might substantially change the results reported here. SUMMARY AND CONCLUSIONS Fluorspar has an inhibiting effect on the tendency of brick made from buff- firing clays to develop vanadium stain. This effect is much more pronounced in cases where the clays are of low re- fractory character, that is, where a considerable proportion of the clay minerals present are illitic. For clays in which the clay minerals are predom- inantly kaolinitic and the quartz con- tent is high, the effect of fluorspar is small and requires higher tempera- tures and more fluorspar for its devel- opment. The most probable mechan- ism for this inhibiting effect is be- lieved to be solution of the vanadium compounds in glass that begins to form as vitrification is initiated in the brick. The vanadium is essentially in a water-insoluble form in all of the four raw clays tested. It was more concen- trated in the fine fractions, which in- clude most of the clay minerals. This of course does not preclude the possi- bility that vanadium may occur in some finely divided state independent of the clay minerals. The evidence fa- voring the theory that vanadium is in- timately associated with the clay min- eral may be summarized as follows: 1) The vanadium is more concen- trated in the fine fractions. 2) The greatest amount of water- soluble vanadium appears in the tem- perature region where breakup of the clay lattice (indicated by x-ray stud- ies and D.T.A. curves) occurs. 3) Hathaway (1955) has presented evidence that vanadium occurs as a substituent for aluminum in the lattice of clay-like silicates containing vana- dium. Vanadium remains relatively water- insoluble in clays which have been heated to temperatures below 700° C. As the heating temperature is carried higher, progressively more of the vana- dium becomes water-soluble until about 950 °C is reached when the amount of water-soluble vanadium begins to fall off rather rapidly (more rapidly for illite-bearing clays). Analyses of the colored materials on the tips of wick test bars indicate that colored efflorescence contains princi- pally vanadium as the colorant with smaller amounts of molybdenum. So- dium, potassium, calcium and magne- sium are present in amounts propor- tionately greater than vanadium. Sul- fate is present in considerable quan- tity. The relative abundance of cal- cium and the alkalies as compared to vanadium leads one to suspect that in many cases where there appear to be large amounts of colored efflorescence on a brick wall the effect may really be largely due to ordinary white efflores- cence which is masked by a strong dye- ing action of highly colored vanadium compounds. No evidence was found to indicate that significant amounts of vanadium are eliminated by volatilization at tem- peratures likely to be used in firing building brick. VANADIUM EFFLORESCENCE 29 REFERENCES ASTM, 1955, Standard Method of Test for Water Absorption: A.S.T.M. Stand- ards. Part 3, p. 736-738 (1955). Bertrand, D., 1950, The Biogeochemis- try of vanadium: Bull. Am. Museum of Nat. Hist., V. 94, art. 7, p. 412. BosAZZA, V. L., 1940, Occurrence of va- nadium and molybdenum in clays: Na- ture, V. 146, no. 370, p. 746. Britton, H. T. S., and Robinson, R. A., 1930, Physicochemical studies of com- plex acids, II, vanadic acid: Jour. Chem. Soc, 1930, p. 1261. Curry, E. R., 1926, Notes on green scum- ming: Jour. Am. Ceram. Soc, v. 9, p. 392. FiOLETOV, A., 1929, Investigation of some rare constituents of clays : Keram. Rund- schau, V. 37, p. 659. Grim, R. E., 1947, 1948, Differential thermal curves of prepared mixtures of clay minerals: Am. Mineralogist, v. 32, no. 9 and 10, p. 493; Illinois Geol. Sur- vey Rept. Inv. 134, p. 11. Grim, R. E., and Bradley, W. F., 1940, Investigations of the effect of heat on the clay minerals illite and montmorillon- ite: Illinois Geol. Survey Rept. Inv. 66; Jour. Am. Ceram. Soc, v. 23, p. 242. Grim, R. E., and Rowland, R. A., 1942, Differential thermal analysis of clay minerals and other hydrous materials: Am. Mineralogist, v. 27, no. 11, p. 750; Illinois Geol. Survey Rept. Inv. 85. Grim, R. E., and Rowland, R. A., 1944, Differential thermal analysis of clays and shales, a control and prospecting method: Illinois Geol. Survey Rept. Inv. 96, p. 10. Hammer, A. J., 1953, You can eliminate green efflorescence: Brick and Clay Rec- ord, V. 123, no. 3, p. 37. Hammer, A. J., 1955, Occurrence of va- nadium in clays, I: Ceram. Age, v. 65, no. 4, p. 47. Hammer, A. J., 1955, Occurrence of va- nadium in clays, VI: Ceram. Age, v. 66, no. 4, p. 28. Hathaway, J. C, 1955, X-ray diffracto- meter studies of some vanadium sili- cates, trace elements investigation : U. S. Geol. Survey Rept. TEI 542. Hecht, H., and Cramer, E., 1902, The collected writings of Herman August Seger, v. I, p. 381: The Chemical Pub- lishing Co., Easton, Pa. Jahr, K. F., and Jander, G., 1934, Poly- vanadates which crystallize from potas- sium vanadate solutions of different hy- drogen-ion concentration: Z. anorg. all- gem. Chem., V. 220, p. 201. Jander, G., and Jahr, K. F., 1933, Build- mg-up and breaking-down of higher molecular inorganic compounds in solu- tion, for example, vanadium acids, poly- vanadates and vanadium salts: Z. anorg allgem. Chem., v. 212, p. 1. Kallauner, O., and Hruda, O., 1922, The occurrence of vanadium ceramic raw materials and products and its effect on the fusibility as well as the color and formation of efflorescence in pure kaolin and a typical brick clay: Sprechsaal, v. 55, no. 30, p. 333; no. 31, p. 345. Kallauner, O., and Kokta, J., 1941, Vanadium efflorescence: Stavivo. v 22 p. 189. Leutwein, F., 1941, Geochemie und Vor- kommen des Vanadiums: Ber. freiber- ger geol. Ges., v. 18, p. 73. LovEJOY, E., 1934, Classification of fluor- escence: Bull. Am. Ceram. Soc, v. 13, no. 1, p. 22. Nazarenko, M. F., and Sviridenko, V. A., 1953, Mineralizing effect of some ad- ditions on the mullitization process: Vestnik Akad. Nauk. Kazakh., S.S.R., v. 10, no. 11, p. 85. Palmer, L. A., 1929, Vanadium and molybdenum compounds in clays: Jour. Am. Ceram. Soc, v. 12, p. 37. Palmeri, V. R., 1951, Mullite formation by decomposition of kaolinite: Anales Soc. Cient. Argentina, v. 152, 127. Prandtl, W., and Mautz, H., 1911, Ac- tion of calcium fluoride on vanadium pentaoxide: Ber., v. 44, p. 2582. Sandell, E. B., 1950, Colorimetric de- termination of traces of metals, second ed., p. 613: Interscience Publishers, Inc., N. Y. Yee, T. B., Machin, J. S., and Andrews, A. I., 1955, Relation of composition to viscosity of enamel glasses: Jour. Am. Ceram. Soc, v. 38, p. 378. 30 ILLINOIS STATE GEOLOGICAL SURVEY Illinois State Geological Survey Report of Investigations 202 29 p., 11 figs., 10 tables, 1957