Table of Contents
A fluorometric method using morin in an alkaline-buffered chelating solution has been developed by the Bureau of Mines for determining beryllium in ores and mill products. Apparatus for measuring fluorescence intensities of the beryllium-morin reaction is described.
Preparation of the sample for analysis consists essentially of digestion with sulfuric, nitric, and hydrofluoric acids, fusion of the residue with sodium carbonate and sodium borate, solution of the melt with sulfuric acid, and proper dilution. Chemical separation of other ions is usually unnecessary. The method gave reliable beryllium determinations on samples in the 0.01- to 0.37-percent BeO range. The method is particularly applicable to low-grade beryllium ores; however, semiquantitative beryllium determinations have been made on beryl containing 12.0 percent BeO.
Assay Determination of Beryllium
The accurate determination of beryllium is rather difficult, and most analytical procedures in use are lengthy, tedious processes requiring a great deal of patience and skill. Four general types of methods are in use at present, namely, gravimetric, colorimetric, fluorometric and spectrochemical. Selection of one of these methods is largely a matter of personal preference; however, the number of determinations required, the grade of material to be analyzed, and funds available for purchasing the necessary equipment are all important considerations.
The gravimetric method is very satisfactory for determining beryllium in medium- to high-grade ores ranging from 8 to 12 percent beryllia, but it is not dependable for low-grade materials because it is very difficult to separate completely the beryllium from all associated elements. These are precipitated with the beryllium and reported as beryllium. The colorimetric and fluorometric methods are reasonably fast if set up on a production-line basis. Their accuracy depends upon the sensitivity and specificity of the dyes employed, particularly for the traces of various elements, some of which are rather difficult to separate completely from the beryllium. In general, these methods are most satisfactory for low- to medium- grade materials. The spectrochemical method appears to be the most satisfactory of the group; but, owing to the expensive equipment required, it is generally available only to the larger laboratories, where it can be fully utilized in making the wide variety of analyses to which it is so ideally suited.
During the course of pegmatite-beneficiation studies for recovering beryl at the Rapid City Experiment Station of the Bureau of Mines, a need arose for a rapid and reliable beryllium assay of ores ranging in beryllia content from 0.01 to 0.30 percent. This range covers the various grades of material that are being studied for the development of a suitable process of beryl recovery end closely covers the grade range of the beryl reserves of the United States as estimated by the Federal Geological Survey.
The fluorometric method described in this report is a modification of the, method reported by Fletcher; but employs morin, which was investigated by Sandel and is also described in various books on chemical analysis. It measures the amount of beryllium present by comparing fluorescence intensities between unknowns and similarly prepared standards. The fluorescence is produced in an alkaline-buffered chelating solution by exposure to ultraviolet light, and the fluorescence intensities are measured with a photoelectric fluorometer. It may be used as a qualitative test for beryllium by visual observation of the yellow-green fluorescence.
This method not only gives very reliable beryllium determinations on ores and beneficiation products of the indicated grade range, but may also be used to determine the beryllia content of materials ranging up to 12.0 percent or more beryllia.
Acknowledgment is made to J. B. Cunningham, professor of metallurgy, University of Arizona, for preparation of standard mixtures and consultation, and to A. L. Lingard, South Dakota Engineering and Mining Experiment Station, for spectrographic analyses.
Reagents & Equipment
All reagents used are C.P. chemicals and are tested for fluorescence by running blank samples in the procedure subsequently described.
Sulfuric acid – 1:1 and concentrate.
Hydrofluoric acid – 50 percent.
Nitric acid – 70 percent.
Sodium hydroxide – 2-molar solution,
Fusion mixture – 3 parts sodium carbonate and 1 part sodium borate, anhydrous.
Alkaline-buffered chelating solution – 86 grams boric acid, 154 grams citric acid, 288 grams of sodium hydroxide, and 150 milliliters of Versene T (derivative of Ethylene Diamine Tetra Acetic acid) diluted to 3 liters with distilled water. No extraordinary precautions are needed in storing this solution, and it may be kept in an ordinary glass-stoppered bottle.
Morin – 0.01 percent in reagent-grade methyl alcohol. Technical-grade morin is used, and the solution is stored in a brown bottle.
Standard beryllium solution – A stock solution of beryllium sulfate is prepared from beryllium metal of known assay. 1.8000 grams of beryllium is dissolved in 35 milliliters of 1:1 sulfuric acid and diluted to 500 milliliters. One milliliter of this stock solution in equal to 10 milligrams of beryllium oxide. A working solution containing 1 microgram of beryllium oxide per milliliter is prepared from the stock solution and stored in a polyethylene bottle. Working solutions are prepared monthly, using distilled water acidified to a pH of 8 with sulfuric acid.
Special equipment necessary for determining beryllium by this method consists of a Beckman spectrophotometer and a fluorometric attachment. This fluorometric attachment is similar to that described by Fletcher. Briefly, it consists of a light-tight cell compartment equipped with suitable filters and a lamp housing. The light source is a General Electric B-H-4 mercury lamp, and the lamp emission is controlled by a constant voltage transformer and by proper ventilation of the lamp housing. Two filters are employed with a blue-sensitive phototube. These filters are Corning Nos. 5874 and 3486. The primary filter, Corning No. 5874, eliminates all light lying between 4,100 and 7,000 A., whereas the secondary filter, No. 3486, absorbs all light of a wave length leas than 5,000 A. The B-H-4 lamp has a strong emission of light at 3,650 A. and, when used in combination with the primary filter, provides light of the proper wave length for excitation of fluorescence in the morin- beryllium solutions. The fluorescent radiation from the solution, which is in the range of 5,600 A., passes through the secondary filter to the phototube and is measured. The phototube is insensitive to light above, 7,000 A. The receiving and amplifying systems of the Beckman spectrophotometer are used for measuring the intensities of the fluorescent light.
Optical glass cells, 31-millimeter-cube inside measurement, with wall thickness of 1 millimeter, are used to hold the solutions while measuring the fluorescence.
Beryllium Assay Procedure
Pulverize the sample to pass through a 200-mesh screen, mix well, and then weigh 20 milligrams of high-grade ore or 500 milligrams of low-grade rock into a 100-milliliter platinum dish. Digest the sample with 1 milliliter of concentrated sulfuric acid, 1 milliliter of nitric acid, and 10 milliliters of 50-percent hydrofluoric acid. Evaporate to dryness on a hot plate, and ignite over a burner until sulfur trioxide is expelled so as to remove all free acid. Fuse the residue with 5 grams of the fusion mixture for 5 minutes at 1,000° C. in a muffle furnace. Digest the fusion cake with 25 milliliters of 1:1 sulfuric acid, and filter into a 1-liter volumetric flask. Aliquot up to 10 micrograms of beryllium oxide from the unknown solutions into 100-milliliter beakers. This is a trial-and-error operation which must be arrived, at by familiarization with the work. Adjust the pH to 2 with 2- molar sodium hydroxide and add 10 milliliters of the chelating solution. Adjust the volume to approximately 20 millilitere with, distilled water, mix, and bring to a boil on a hot plate. Cool the solution and add 10 milliliters of the morin solution, transfer to a 100-milliliter volumetric flask, add distilled water to make volume to 100 milliliters, and mix well. The pH of this final solution is now approximately 13.1. Till the optical cells, place in the fluorometer, and read the percent transmission, which is a measure of fluorescence. Prepare standard beryllium solutions in the same manner as the unknowns, and prepare a standard curve ranging in beryllium oxide content from a blank to 10 micrograms. These curves should be prepared daily end checked at frequent intervals while running a series of unknowns. Determine the micrograms of beryllium in the unknown solutions from the standard calibration curve, and calculate the percentage of beryllium oxide. A standard curve is shown in figure 1.
Borates in Glass
In the early 1970’s, the standard purified product became 5-mol borax, and Rasorite was removed from the market while Rasorite is still sold outside the United States. Fiberglass producers began serious evaluations of borate ores, principally colemanite and ulexite from California. The Death Valley colemanite from the open pit operations had between 34 and 48% boron oxide, depending on whether or not it was calcined. This product was accepted for economic reasons, and because it melted better in existing glass furnaces, and it replaced the more expensive boric acid (Table 2). As the demand grew, and the mining and blending problems preventing a uniform product were solved, colemanite became a major raw material. Underground mining was begun to develop better and larger orebodies, flotation operations were installed, and the level of impurities and stability of composition desired by the glass industry was achieved with about 38% boron oxide content in uncalcined California colemanite.
Turkish colemanite has been imported for E glass, but only under customer-supervised mining and production in Turkey. The ore is generally shipped to the east coast of the United States, where it is blended and ground, yielding a product of about 40 to 45% boron oxide. Because of freight costs, the Turkish colemanite, although higher in boron content than that of California, is used generally in the United States east of the Mississippi River, while California products primarily service the western United States.
Ulexite from California was used in insulation glasses, but mining and processing considerations have brought about its withdrawal except for limited uses. Instead, Turkish ulexite is brought to the east coast of the United States, blended and ground, and sold at 38% boron oxide content. Ulexite has been used in place of domestic borax for cost reasons in products in which sodium can be tolerated in the glasses.
European users find Rasorite, which is not offered in the United States, suitable for use in insulation glasses. Ulexite and colemanite from Turkey also supply the needs of Europe. South American glass makers use either Rasorite, 5-mol borax, or boric acid.
The choice of boron raw materials is also influenced by the size and type of glass-melting facilities. E glass is produced mainly in fairly large gas-fired furnaces, with recuperators for preheating combustion air. With operating temperatures above 2700°F and outputs of 75 st of glass per day or more, fuel efficiency is a major cost consideration. The use of colemanite in place of limestone and boric acid allows faster batch melting, sometimes increases furnace output, and yields better fuel efficiency. Ulexite is usually a replacement for borax, but the effect on batch melting is not as pronounced as for colemanite in E glass, since the batches already contain soda ash, a very active flux. In glasses where purity or color is critical, the purer boric acid or borax is used, because both colemanite and ulexite have enough iron oxide, sulfate, and other deleterious components to preclude their use.
Boron is beneficial in glass because it reduces thermal expansion, improves chemical durability, modifies optical and electrical properties, increases heat resistance, and improves resistance to thermal shock. As boron content increases in many compositions, phase separation occurs upon heating, and this separation is the basis for the high-silica Vycor products. Regular melting and forming techniques are used, but by heat treatment of the finished article, a continuous borate phase is formed in the high-silica glass, which is then leached out with a hot acid solution. The porous structure is then consolidated to a transparent 96% silica glass by heating above 1650°F, giving products very resistant to acids and thermal shock. The porous glass can also be used as a desiccant or as a catalyst support.
Boron is used as a flux in place of alkali oxides in the higher-silica glasses. Boron retains its trigonal planar coordination which reduces the cohesiveness of the silica structure in glass. As a network-forming ion, boron oxide does not increase the thermal expansion of vitreous silica as much as network-modifying ions do, making it an effective flux and additive to increase thermal shock resistance in commercial products.
Resistance to water attack is achieved when boron is used to enhance chemical properties, such as in fiberglass insulation. Water attack involves the removal of alkali ions from the glass surface. This process can destroy the silica network if the alkalinity and temperature are adequate. This causes frosting of the glass surface which is detrimental in many products. In the case of fiberglass, water attack tends to destroy the bond with the organic materials in the product and the resulting deterioration can cause failure. Most acid-resistant glasses are a combination of fairly-high silica and boron; these glasses generally have lower alkali content which makes them very good for most chemical uses, such as glass no. 2 in Table 1. A summary of chemical durability is given by Adams and Walters (1984).
In the glass industry, the use and demand for boron is very firmly established. The electronic and technical uses are varied and are expected to provide growth of demand well into the future. With energy costs high, new building construction requires adequate insulation, so the fiberglass insulation industry appears to offer a steady demand for boron products with about 30 large plants producing a wide variety of products in the United States and Canada. The glass textile producers have been expanding output to meet plastic reinforcing demand for structural and electronic uses, as well as for reinforcing mats for plastics, roofing mats, shingles, underground storage tanks, and the many products that use the high tensile strength, high modulus of elasticity, and chemical stability of E glass. About 16 plants now produce continuous textile products in the United States and Canada. Large producers of fiberglass products are also located in Europe, South America, Australia, and the Far East, while boron for these demands is met essentially from California or Turkey.
Borates In Glass by Gagin, Lawrence V.
Organization: The American Institute of Mining, Metallurgical, and Petroleum Engineers
Pages: 4, Publication Date: Jan 1, 1985
Experimental Results And Discussion
During the past 12 months several hundred beryllium determinations have been made by this method on a wide verity of products, including all grades of ore and beneficiation products, as well as leach liquors and residues resulting from beryllium extractive studies on beryl ores and flotation concentrates. For the purpose of this report, however, seven standard mixtures of known beryllium oxide content were prepared, using high-grade beryl and either beryllium-free quartz or beryllium-free soda feldspar. These samples ranged in beryllia content from 0.0124 to 0.372 percent. Each sample was analyzed for its beryllia content 3 or 4 times. The average beryllia content of each sample was calculated and the maximum deviation from the average figured. These results are given in table 1. The maximum deviation from the average analysis was 7.7 percent and occurs, as is to be expected, in the lowest grade sample. Here the maximum deviation was only 0.001 percent or 10 p.p.m. of beryllia.
Beryllium reacts with morin in an alkaline solution to produce a strong yellow-green fluorescence, and Sandel indicates that as little as 1 part per billion of beryllium can be detected. However, one of the greatest hazards in fluorescence measurements here is the variation of fluorescence intensity with pH. Figure 2 shows the variation of relative fluorescence intensity with pH for the beryllium- morin reaction. From the standpoint of the fluorescence Intensity-pH relationship, a pH of about 12.5 appears optimum; however, specificity for beryllium favors a more alkaline medium. Solutions well-buffered at pH 13.1 were found to be most satisfactory. Fluorescence intensity decreases rapidly as the pH falls below 12.4.
The concentration of morin is a critical factor in measuring beryllium-morin fluorescence. Figure 3 shows the variation of relative fluorescence intensity with the amount of morin. Blank samples fluoresced slightly owing to morin and were zeroed with the dark current adjustment of the spectrophotometer.
Analyses of the normal run of beryl-bearing pegmatite rocks offers little complications with respect to interfering ions. The Versene T is a special complexing agent for iron in caustic solutions. Other ions, such as aluminum, lithium, calcium, and silica, can be controlled by proper dilution of the sample solution to a point where they have a minimum effect upon the beryllium determination. Solutions containing excessive amounts of aluminum increase the morin-beryllium fluorescence, but this is due more to a fall in pH caused by caustic consumption through aluminate formation than to the presence of aluminum. This can be corrected by adding enough sodium hydroxide to convert the aluminum to the aluminate. Ferric iron is found to lower the fluorescence; however, excessive amounts of iron can be removed by heating the sample solution with sodium hydroxide, being careful not to boil. Beryllium, aluminum, and iron precipitate; however, the aluminum and beryllium will redissolve, leaving a ferric hydroxide precipitate. Calcium causes the beryllium analyses to be low, and lithium will cause it to be high, whereas soluble silica has no appreciable effect on the analysis.
To study the effect of the more common elements found in pegmatite rock on the morin fluorometric method of beryllium determination, a series of standard 5-microgram beryllia solutions is contaminated with varying amounts of aluminum, lithia, calcium, and iron. Figures 4, 5, 6, and show the effect of varying amounts of aluminum, lithia, calcium, and iron, respectively, on the observed beryllia content of the standard solutions.
Aluminum causes the observed beryllium content to be 15 percent high when present in amounts ranging from 2 to 10 milligrams. Two milligrams of lithia increases the fluorescence as much as 10 percent. Ten milligrams of calcium causes the beryllia to be approximately 10 percent low; however, smaller quantities in the range of 7 milligrams cause it to be approximately 10 percent higher than 5 micrograms. One milligram of ferric iron is found to lower the fluorescence 15 percent.
Although these elements appear to cause large variations in the analysis of beryllium, they are well within the limits of the accuracy of the method. For example, in a sample of rock containing 0.01 percent beryllia, an aliquot protion that contained 5 micrograms of beryllia would equal 50 milligrams of rock. Normal pegmatite rock seldom contains more than 20 percent alumina; therefore, the aliquot would contain 10 milligrams of alumina or slightly more than 5 milligrams of aluminum. While this is enough aluminum to cause a 15-percent increase in the observed beryllia content of the sample, it can be tolerated, since the error would occur in the third decimal place, and, on the basis of the 0.01 percent beryllia, this would make very little difference in the final answer. Now, assume that the rock sample contains 0.1 percent beryllia; a 5-microgram beryllia aliquot would then be equal to 5 milligrams of rock. On the basis that the rock contains 20 percent alumina, only 0.5 milligram of aluminum would be present, and from figure 4 we see the observed beryllia content to be 5.3 micrograms at this concentration of aluminum. The error is now only 6 percent. As the beryllia content of the sample is increased further, the error caused by the presence of alumina decreases, since smaller aliquots would be used.
On the same basis as the previous example, up to 4 percent lithia and 20 percent calcium could be tolerated without appreciably affecting the beryllium analysis of rock containing 0.01 percent beryllia.
Figure 7 shows that 1 milligram of iron causes a 15-percent decrease in the observed beryllia-morin fluorescence of a standard 5-microgram beryllia solution. A 0.01-percent-beryllia sample therefore could contain up to 2 percent iron without appreciably affecting the analysis. Normal pegmatite rock seldom contains over 7 percent iron; however, rock that has a high content of tourmaline and other dark or iron-stained minerals probably would exceed the 8-percent iron tolerance, end iron separation would be required.
As in the case of aluminum, the interference caused by lithium, calcium, and iron with the morin-beryllium fluorescence decreases as the beryllium content of the sample increases. It is therefore possible to control the interfering ions by proper dilution of the sample solutions.
Factors Affecting Morin-Beryllium Fluorescence
Such factors as temperature, time, and light all affect the morin-beryllium fluorescence. However, they may all be offset by preparation and comparison of standards and unknowns at the same time. The small temperature changes in normal room conditions have no noticeable effect on the fluorescence. Fluorescence Intensities tend to more more constant with higher morin concentrations. Large temperature changes, however, should be avoided. With respect to time, the fluorescence increases slightly shortly after preparation of the solutions, but then a gradual fading begins. The known and unknowns should be compared at reasonably uniform time intervals. This can be accomplished by scheduling the introduction of the morin into the sample solutions at proper time intervals to allow for the necessary time to measure and record the fluorescence. The rate of fading of the fluorescence, however, is relatively slow, and timing is not too critical a factor. Light, like time, causes the morin-beryllium fluorescence to fade. The rate of fading by light is relatively slow, and analyser; may be made in a well-lighted laboratory without appreciable effect upon the accuracy of the results. It is best, however, to avoid exceptionally strong light, such as direct sunlight. Again, like time, a time schedule will eliminate the effects of light.
Although the technical-grade morin used in this procedure is a relatively impure product, which may vary considerably from shipment to shipment, it has so far given very satisfactory results. The ability to obtain such satisfaction is attributed to the fact that standard beryllium curves are prepared daily or for each group of unknowns run.