This invention relates to a process for producing indus¬ trially usable gypsum from a commercial by-product, in parti¬ cular fluoroanhydrite. More particularly, this invention re¬ lates to an improvement in processes for the production of hydrogen fluoride and waste product fluoroanhydrite.

In the industrial production of hydrogen fluoride, fluorspar is reacted with concentrated sulfuric acid in an externally heated reaction vessel to co-produce hydrogen fluoride and fluoroanhydrite. For every ton of fluorspar consumed, approximately 1.75 tons of anhydrite is produced. After reaction, the hydrogen fluoride is drawn off and con¬ densed while the fluoroanhydrite is generally slurried with water, neutralized with lime or limestone and pumped to dis¬ posal ponds.

The fluoroanhydrite is contaminated with excess sulfuric acid and fluoroaluminate impurities. Historically, because of the contaminants, the fluoroanhydrite has not been com¬ mercially usable and has been allowed to hydrate naturally over a several year period of time. Thus, large tonnages of mixed fluoroanhydrite-fluorogypsum materials have been accumu¬ lated in these ponds over the years. Descri tion of the Prior Art

U.S. patent 3,825,655 discloses an improvement in hydro¬ gen fluoride processes said to produce a coarse-grained sulfate

SUBSTITUTE SHEET

co-product in either anhydrite or gypsum form said to be industrially useful.

However, it has been shown that certain fluoroaluminum

-2 impurities still present, e.g. IA1F..H O) ] ^*" cause deleterious effects in attempts to manufacture gypsum products from the fluoroanhydrite. These impurities inhibit the hydration of the fluoroanhydrite to gypsum. These impurities inhibit the hydration of the fluoroanhydrite to gypsum. These impurities raise the calcining temperature during conversion of fluoro- gypsu to calcium sulfate hemihydrate; and finally, they result in producing gypsum products such as industrial plasters, building plasters and gypsum wallboard of poor quality. From the above, it is apparent that there is a need in the art for effective and economical means for removing a substantial portion of the impurities that originate in the hydrogen fluor¬ ide process. Furthermore, there is a need in the art for making calcium sulfate products that may be used in wallboard and other industrial and construction materials. The process of the present invention offers a solution to these needs. Summary of the Invention

These and other objects and advantages are realized in accordance with the present invention wherein fluoroanhydrite obtained directly from a conventional hydrogen fluoride pro¬ duction process is rapidly hydrated to gypsum in a dilute acidic water slurry to produce a gypsum free of harmful sol¬ uble and syncrystallized impurities and which is acceptable for gypsum board manufacture. The process comprises adding small amounts of fluoroanhydrite to a fluid reaction medium containing large quantities of small gypsum seed crystals (about 40-90% by weight of the solids of the medium) and about 0.5-20% by weight of a soluble sulfate accelerator. It has been found that the hydration reaction of fluoroanhydrite is very rapid in the presence of large quantities of .-mall sized gypsum seed crystals and these seeds magnify the effect of soluble sulfate hydration accelerators such as sulfuric acid or sodium sulfate, especially in the earlier stage.-- of reaction. As hydration proceeds to build up large sized gypsum particles, the medium is filtered to separate about 15 ^* .. of the

SUBSTITUTE SHEET

solids of the medium as gypsum product; and it is preferred to recycle the remaining 85% solids for further proce. ^« -.. ^; i ng . Brief Description of the, Drawings

Figure 1 is a diagrammatic flow chart illustrating one preferred method of the present invention and Figures 2-5 are plots of percent conversion of fluoroanhydrite to gypsum ver¬ sus time under various conditions of the present invention. Description of the Preferred Embodiments

The process comprises a continuous, rapid conversion of fluoroanhydrite, produced by a hydrogen fluoride plant, to gypsum by adding small amounts of the fluoroanhydrite to a fluid reaction medium containing a soluble sulfate (preferably sulfuric acid and sodium sulfate) and large amounts of small gypsum seed crystals. The fluid reaction mixture has a total solids content of about 20-60% by weight, with preferably about 50-90% of the total solids being small sized gypsum seed cry¬ stals. The reaction mixture also contains about 0.5-20% soluble sulfate accelerator and fluoroanhydrite is added to the mixture in amounts less than the amount of gypsum seed crystals. The fluoroanhydrite is hydrated to large sized gypsum particles on the gypsum seed crystals. About 10-30% of the gypsum particles in any one cycle is recovered as large size gypsum crystal product. Thereafter, the separated, large crystal gypsum product is passed on to separate processing for conversion to building plaster products such as wallboard. The remaining reaction mixture suspension contains a maior propor¬ tion of small size gypsum crystals and is recycled in the pro¬ cess to act as a supply of seed crystals for processing subse¬ quent additions of fluoroanhydrite.

The fluoroanhydrite feed may be produced by .my of the conventional hydrogen fluoride production processes, preferably the cooled, dry product directly from the hydrogen fluoride reactor. This material has a particle si__e of less than 10 micrometers, generally less than 5 micrometers, as obtained from the reactor. Stockpiled fluoroanhydrite cake m.iy be easily broken up into suitable particle sices for pr eessing in accordance with the present invention.

SUBSTITUTE SHEET

OMPI

The sulfuric acid is conveniently the same as that used as feed in the hydrogen fluoride reactor, diluted to a weight concentration of about 1% to about 20%, preferably about 5-15 % and most preferably about 7.5%. The necessary soluble sulfate moiety may be provided in whole or in part by residual acid in the fluoroanhydrite from the hydrofluoric acid reactor. Any sulfuric acid added beyond that co ing in with unwashed fluoro¬ anhydrite should be reaction grade. The accelerator, prefer¬ ably soluble sulfate may further 'be provided in whole or in part by soluble sulfate salts, such as sodium, potassium, magnesium, aluminum, iron, zinc and other soluble sulfates. Combinations of accelerators are preferred such as mixtures of ammonium and potassium sulfate combined with sulfuric acid. A. particularly preferred combination is about 7% sulfuric acid and 5-15% sodium sulfate in the liquid solution of the reacting slurry. Above about 20% or below about 1% sulfuric acid are generally unsatisfactory; since in the former solubility of calcium sulfate decreases and the resultant gypsum product is unstable and requires too much neutralization for use in gypsum construction products, and the latter requires uneconomical conversion times.

The gypsum seed material may be any natural ground gypsum or high purity synthesized calcium sulfate dihydrate of about 5-60 micrometers size and preferably 15 to 30 micro¬ meters particle size. For the start-up of the process of this invention, the gypsum seed may be either ground, naturally occuring gypsum or prepared from industrial processes yielding a high purity gypsum of the appropriate particle size, such as from titanium dioxide preparation, citric acid preparation, flue gas desulfurization processes and the like. Once steady state is achieved, conveniently the gypsum seed is obtained as recycle of the fluid reaction medium after havinα separated larger size dihydrate particles as gypsum product. Such larger particles of gypsum are of convenient size for gypsum board or plaster production.

In implementing the process, it is important that there be a sufficient quantity of small sized, pure gypsum seed

SUBSTITUTE SHEET

crystals and sufficient soluble sulfate present in the re¬ action medium to achieve a rapid conversion of the fluoroan¬ hydrite to large gypsum particles, to keep impurities in the fluoroanhydrite solubilized, and to maintain easy separation of the highly pure gypsum product from the recycle seed cry¬ stals. Thus, generally 50-90% by weight, and more preferably 60-75%, of the solids of the reaction slurry should be small sized, pure gypsum seed crystals. About 5-20% concentration, and more preferably 5-10%, of sulfuric acid and water soluble sulfate salt in the reacting slurry will provide optimum rapid conversion.

The process is carried out, as illustrated in Figure 1 in one preferred embodiment by metering cooled and dried fluoro¬ anhydrite powder (1) into a recycling reaction medium in a first mixing zone (2). All of the vessels used in the process are conventional. At start-up all of the materials are fed to the first mixing zone from an outside source, but when steady state operation has been achieved, the reaction medium is supplied with recycled materials from the separator (3) plus any necessary makeup of water, sulfuric acid and soluble sul¬ fate accelerator. Upon steady state operation, very little makeup of soluble sulfate and sulfuric acid is necessary except to compensate for losses and purged bleed off (6) to avoid buildup of soluble fluorides in the system. While four hydrating tanks are illustrated in Figure 1, conventional mixing vessels or compartmentalized cells, at ambient temperatures, from a single through a half dozen units may be utilized depending upon the scale of operations. The hydrating tanks o o may be operated at from about 0 C to about 42 C, although o ambient temperatures of about 20-22 C are preferred.

The hydration of the fluoroanhy rite proceeds progres¬ sively as the material advances from on e compartment to the next. Numerous tests have shown that if fluoroanhydrite con¬ taminated with sulfuric acid is hydrated directly wi h a major proportion of recycle gypsum seed crystals, the con¬ version takes place in a matter of minutes. The hydrating mixture seems to be acidic enough to keep complex fluoro-

SUBSTITUTE SHEET

- - aluminum species from precipitating out on the surface of the growing gypsum crystals. Otherwise, the fluoroaluminum im¬ purity would inhibit the crystal growth and co-crystallize with the growing gypsum particles thereby rendering them un¬ suitable for production of building construction material products. If there is sufficient proportion of small size seed crystals in the reaction mixture, hydration is very rapid as shown in Figures 2 and 3. Furthermore, the degree of con¬ version can be controlled so as to aid in maintaining any desired recycle proportion as shown in Figures 4 and 5.

Under usual operating conditions, the fluoroanhydrite is very rapidly converted (generally 10 minutes to 2 hours) to sufficiently large sized gypsum particles for convenient separation by conventional separation equipment e.g. , one or more hydrocycloneε, centrifuges, fixed or moving bed filters and the like. In the preferred operation, a first separation is made to classify and separate very small sized gypsum cry¬ stals for recycle in the reaction by a hydrocyclone or the hydraulic separator (3) illustrated in Figure 1. The larger gypsum particles are than passed to a further separator such as bed filter (4) for separation of very large size product gypsum particles. The acidic filtrate from the filter (4) can be partly neutralized by lime or limestone addition (7) to immobilize the fluoroaluminum impurities for purging from the system (6) as their concentration builds up to an interfering level. The product gypsum underflow from filter (4) may be washed (not shown) to remove residual filtrate. The filtrate and wash waters may then be combined to reconstitute the fluid reaction medium for recycle. Optionally, a certain amount of lime treated fluoroanhydrite from waste piles may be used to neutralize the acidic filtrate. Alternatively, the acidic filtrate can be used for washing hydratod, lime treated fluoro- gypsum from waste piles to remove some of the 1uoroaluminum impurities from that gypsum making if suitable for use as seed crystals.

Example 1 Fluoroanhydrite from a manufacturer of hydrofluoric acid is treated as set forth in Figure 1.

SUBSTITUTE SHEET OMFI WIFO

Ta!ble la

Process Point A B C D E F G

Gypsum 30.0 37.5 34.8 54.1 81.1 0 32.2

Anhydrite 10.0 4.1 3.8 5.9 8.9 0 3.5

Sodium Sulfate 3.0 3.0 3.2 2.1 0 2.8 3.2

Sulfuric Acid 6.0 6.0 6.3 4.1 0 8.5 6.4

^" Water 51.0 49.4 51.9 33.8 10.0 88.7 54.7

Total wt. % 100.0 100.0 100.0 100.0 100.0 100.0 100.0

After obtaining steady state conditions, 25.7 short tons (51400 lbs., 23315 kg) per hour of fluoroanhydrite from the hydrogen fluoride reactor are introduced into a mixing appara¬ tus such as that shown schematically in Figure 1 which comprises 4 hydration compartments. 11.25 short tons of sodium sulfate and 11.25 short tons of 100% sulfuric acid are introduced into the first compartment before reaching steady state to provide on reaching steady state conditions, a reaction medium as set forth in Table la, the slurry having a specific gravity of 1.37 grams per milliliter and flowing at a rate of 1143.8 gallons (4329.754 liters) per minute. At various points in the pro¬ cess data are taken of the materials at the points identified in Figure 1 and set forth in Table la, the points corresponding to letters in the figure.

Table lb

^• Process Point B D G

Gypsum 32.6 41.3 32.6 8.7 8.7 3 2 . 6

Anhydrite 9.3 2.5 1.9 0.5 0.5 1 . 9

Sodium Sulfate 4.3 4.3 3.7 0.6 4 . 3

Sulfuric Acid 7.8 7.8 6.8 1.0 7 . 8

Water 61.0 59.2 51.0 8.2 1.0 61 . 0

Total gms 115.0 115.0 96.0 19.0 10.2 1 17 . 6

Relative Weight Flows, g/min. Solids 41.0 43.7 34 9.2 9.2 34 . 5

Liquids 73.1 71.3 61.5 9.8 1.0 7 3 . 1

SUBSTITUTE SHEET

( OMPI

- 6 -

Product Properties Vicat Set

Mortar Cube Strengths of ,47.7 p.c.f. cubes: 863 psi dry

2 compressive strength; 517 humid

2

Wallboard Bond Load Capability at humid conditions: 9.8 pounds compressive strength

The run was repeated with fluoroanhydrite from a dif¬ ferent manufacturer. After achieving steady state conditions, 7.4 grams per minute of the fluoroanhydrite were metered into the process as set forth in Figure 1 with a flow rate through the reactor vessels (2) of 115 grams per minute. Data taken at various process points is set forth in Table lb. The pro¬ duct gypsum was calcined to hemihydrate and properties of gypsum products made from it are also set forth in Table lb. As set forth in Table lb, acceptable gypsum board and plaster product properties were obtained.

Example 2 A fluoroanhydrite from hydrogen fluoride production was analyzed as follows:

Wet Chemical Analysis (in weight %) o o Loss on Ignition, 40 -230 C 2.79% CaO 38.35

S0 ₃ 57.29

MgO 0.03

SiO 0.14

Fe ₂ 0 0.12

Excess Loss on Ignition,

230°C-950°C 1.39 pH 1.70

Water Soluble Salts (in ppm) Potassium 118

Sodium 8

Magnesium 25

Chloride 40

Fluoride 5

1 Elapsed time from mixing 50 q standard accelerated plaster water mix to when a 300 g Vicat needle will not pene¬ trate more than half way into the setting slurry. 2 Humidified for 16 hours at 90°F and 90% relative ^humidity . ^^gS Xf

^■ OMPI _

SUBSTITUTE SHEET i__.™«_f_°wo _< ^* _i.

The rate of hydration to gypsum with varying proportions of dihydrate seeding and accelerators was evaluated. For this, 300 gram aliquots containing the appropriate amount of fluoro¬ anhydrite and gypsum were combined with 600 ml aliquots of water containing the accelerator, stirred for 24 hours, and samples taken for hydration reaction analysis at 2 , 4, 6 and 24 hours.

Increasing the amount of gypsum seeding has a direct influence on the rate of conversion of fluoroanhydrite to gypsum as shown in Figure 2. For example, at 3 hours the per¬ cent conversion for a mixture of fluoroanhydrite with 10% α gypsum seed and 1% sodium sulfate at 22. C increased about 5 times when using a 50% gypsum seed in the system. It is dif¬ ficult to reach a conversion much above 95% since the surface area of the anhydrite becomes a limiting factor. A way around this is to hydraulically classify, as at separator (3) , the mixture after a suitable residence time and recycle the small anhydrite crystals for further conversion. In this manner a gypsum product having a purity greater than 95% is produced.

Figure 3 shows the effect of seed crystal surface area on the rate of conversion. In this evaluation, although the seed crystals were provided as 50% of the total solids, the surface area of the seed was varied by using crystals of dif¬ ferent sizes. Again, there appeared to be a direct dependence of conversion rate as a function of surface area. For example, at 2 hours a 50% seeding went from a conversion of 16% to about 78% by using 7 micrometer seeds having a surface area of 6100 square centimeters per gram in contrast to 32 micrometer seed having a surface area of 900 square centimeters per gram. Thus, a 6.7 fold increase in surface area provided an increase of about 5 times in conversion rate. In this case, the sodium sulfate concentration becomes a limiting factor because of greater ion pair formation and tying up of reactant water through hydration of the salt ions.

The soluble sulfate salt concentration was varied as shown in Figure 4, for a 50% seed system. Again, there seemed to be a direct correlation between the salt concen-

SUBSTITUTE SHEET

tration at the lower levels and the conversion rate. For example, at 3 hours the rate was approximately doubled by increasing the sodium sulfate concentration by a factor of 2. However, this relationship was complicated by the 1-2% sulfuric acid present as an impurity in the fluoroanhydrite.

The effect of sulfuric acid concentration is shown in Figure 5. Sulfuric acid is not as effective an accelerator as the soluble cationic sulfate salts. This is because its sul¬ fate ion concentration is probably only 1/lOth mole percent in comparison to that of sodium sulfate or potassium sulfate. Also, the rate of conversion becomes much slower above about 60%; however, this may be augmented by mixtures of sulfuric acid and the sulfate salts. For example, it is believed that combinations of sulfuric acid having a concentration between about 5-10% with sodium sulfate having a concentration between about 0.5-3% are at least additive in hydration; and thus, mix¬ tures are preferred.

Example 3

An extended run for 16 hours continuous using 75% seeding with 10% sulfuric acid and 5% sodium sulfate acceler¬ ators in the liquid phase was performed to evaluate conversion and fluoride buildup on recycle as follows:

A fluoroanhydrite from a hydrogen fluoride reactor oper¬ ation was obtained. The material had the following analysis: 38.98% CaO, 54.79% SO , 0.03% MgO, 0.08% SiO , 0.10% Fe 0 , 0.10% Al 0 , 0.13% P and 0.21%F. Loss on ignition between 40°C and 230°C was 3%; and X-ray diffraction analysis showed only anhydrite as the calcium sulfate form present.

Fluoroanhydrite (150 parts by weight) and 99% purity gypsum seed crystals (450 parts by weight and 7 micrometers particle size) were added with stirring to 1200 parts by weight of a solution containing 10% by weight of sulfuric acid and 5% by weight of sodium sulfate. This mixture formed a 33% solids hydrating media with the seed crystals constituting 50% based on the total solids of the media. The slurry was con¬ tinuously stirred for a 16 hour interval. Every 3-4 hours a 370 parts by weight portion of the slurry was diverted, fil-

SUBSTITUTE SHEET

OMPI

tered and washed, and anlayzed for gypsum and fluoride content; and a fresh portion of 100 parts by weight of the fluoroanhy¬ drite plus 270 parts by weight of the 10% sulfuric acid and 5 ^C . sodium sulfate solution was ball milled to reduce gypsum crystal size then returned to the mixing slurry. Analysis of the sampled portions was as follows:

Time Interval % Gypsum Content PPM Fluoride Particle Size

4 97.8% 18 ppm 13.32

8 97.5% 22 ppm 17.21

12 96.9% 35 ppm 18.58

16 96.8% 119 ppm 20.82

From the above, it can be seen that periodic purging of fluoride from the recycle is desirable. From this and earlier examples it is seen that conversion rate is greatest in about the first hour or so. Most practical operations will require only about 60-90% conversion in order to recycle some anhydrite. It should be noted that in this and the foregoing examples, the results are expressed in weight of fluoroanhydrite converted. Since, for example, most of the slurries utilized 50% gypsum solids present by the seeding, at a 50% total suspended solids slurry, a 78% conversion of the anhydrite present in the first hour or less provides a 90% gypsum content.

Example 4 A composite of the samples from Example 2 which contained over 70% gypsum on hydration in 24 hours, was calcined to cal¬ cium sulfate hemihydrate plaster and compared with commercial plaster made from natural gypsum rock.

Standard plaster slurries were made with water and set accelerators, with results of setting properties as follows:

Table 2

Setting

Plaster Agent Time of Temperature Ri s e S e t Source Addition Vicat Set M inut e s lyiii i n .

Natural Rock plaster — 32 minutes 37.-5 110 A

0. lg 5.75 15 11 3 (' . 7

Fluoroanhydrite source plaster -- 8 15 1 18

0.1 3 7 1 19 . 8 1 . ι>

SUBSTITUTE SHEET

Additional fluoroanhydrite source plaster samples were blended in about equal weight proportion with natural rock plaster and then blended' with conventional additives for gypsum board manufacture with results as follows: stucco color: gray;

PH 7.2;

2 reground Blaine surface area: 6600cm~/g mortar cube strength of 43.81 lb/ft. density cubes: 720 psi dry compressive wallboard bond load capability, on 16 hour conditioning at

90°F/90% relative humidity: 13.53 pounds

Satisfactory plaster, stucco and wallboard properties were obtained from calcination of the fluoroanhydrite source plaster.

SUBSTITUTE SHEET

OMPI ^ V/IPO ^