Process for preparing crystalline ammonium sulfate product

Abstract

The invention relates to a process for preparing a crystalline ammonium sulfate product, which process comprises: a) subjecting in a crystallizer a feed solution of ammonium sulfate to crystallization to form a first slurry of ammonium sulfate crystals; b) subjecting the first slurry of ammonium sulfate crystals to a first size classification to yield a first coarse ammonium sulfate crystal fraction and a first fine ammonium sulfate crystal fraction; c) recycling at least part of the first fine ammonium sulfate crystal fraction to the feed solution of ammonium sulfate; and d) recovering a crystalline ammonium sulfate product from the first coarse ammonium sulfate crystal fraction, characterized in that: e) a second size classification is carried out on a second slurry of ammonium sulfate crystals to yield a second coarse ammonium sulfate crystal fraction and a second fine ammonium sulfate crystal fraction.

Claims

1. A process for preparing a crystalline ammonium sulfate product, which process comprises: a) subjecting a feed solution of ammonium sulfate to crystallization in a crystallizer to form a first slurry of ammonium sulfate crystals; b) subjecting the first slurry of ammonium sulfate crystals to a first size classification to yield a first coarse ammonium sulfate crystal fraction and a first fine ammonium sulfate crystal fraction; c) recycling at least part of the first fine ammonium sulfate crystal fraction to the feed solution of ammonium sulfate; d) recovering a crystalline ammonium sulfate product from the first coarse ammonium sulfate crystal fraction, and e) subjecting a second slurry of ammonium sulfate crystals to a second size classification to yield a second coarse ammonium sulfate crystal fraction and a second fine ammonium sulfate crystal fraction, wherein (i) the second slurry of ammonium sulfate crystals is the first fine ammonium sulfate crystal fraction or (ii) the second slurry of ammonium sulfate crystals is withdrawn directly from the crystallizer.

2. The process according to claim 1, wherein option (ii) of step e) comprises determining a threshold crystal size of the second size classification by adjusting density of the first slurry of ammonium sulfate crystals in the crystallizer.

3. The process according to claim 1, which further comprises recycling at least part of the first fine ammonium sulfate crystal fraction to the feed solution of ammonium sulfate.

4. The process according to claim 3, which comprises dissolving at least some of the crystals in the at least part of the first fine ammonium sulfate crystal fraction to be recycled before the at least part of the first fine ammonium sulfate crystal fraction is recycled to the feed solution of ammonium sulfate.

5. The process according to claim 1, which comprises recycling at least part of the second fine crystal fraction to the feed solution of ammonium sulfate.

6. The process according to claim 1, which comprises dissolving at least some of the crystals in the at least part of the second fine ammonium sulfate crystal fraction to be recycled before the at least part of the second fine ammonium sulfate crystal fraction is recycled to the feed solution of ammonium sulfate.

7. The process according to claim 1, wherein the first slurry of ammonium sulfate crystals comprises 0.1 to 50 vol. % ammonium sulfate crystals relative to the volume of the slurry.

8. The process according to claim 1, wherein the first size classification has a threshold of from 0.5 to 2.0 mm.

9. The process according to claim 1, wherein the second size classification has a threshold smaller than the threshold of the first size classification and is from 0.1 to 1.0 mm.

10. The process according to claim 1, wherein the first size classification utilizes a filter or a hydrocyclone.

Description

(1) FIG. 1 illustrates the prior art, as exemplified in WO 02/081374.

(2) FIG. 2 illustrates an embodiment derivable form prior art processes.

(3) FIGS. 3, 4 and 5 each illustrate embodiments of the process of the present invention.

(4) FIG. 1 depicts a typical set-up according to the prior art. A fresh solution of ammonium sulfate enters, through line (a), a mixing unit (1), where it is mixed with input from line (g) to form a feed solution of ammonium sulfate. The feed solution of ammonium sulfate is then passed through line (b) into crystallizer (2), where it is subjected to crystallization, such that a slurry of ammonium sulfate crystals is produced. The slurry of ammonium sulfate crystals passes through line (c) to recovery unit (4), where crystalline ammonium sulfate is separated and removed through line (e), and ammonium sulfate solution is recycled through line (g) to mixing unit (1).

(5) FIG. 2 depicts a second embodiment of the prior art. The set-up is similar to that of FIG. 1, with the exception that a first size classifier (3) is inserted. The operation is substantially as described for FIG. 1, with the following additions. The slurry of ammonium sulfate leaving the crystallizer (2) passes via line (c) into first size classifier (3), where it is separated into a fine ammonium sulfate crystal fraction and a coarse ammonium sulfate crystal fraction. The resulting fine ammonium sulfate crystal fraction is passed via line (f) to mixing unit (1). The coarse ammonium sulfate crystal fraction is passed via line (d) to separation unit (4).

(6) FIG. 3 depicts an embodiment of the present invention. In addition to the set-up of FIG. 2, a second size classifier (5) is added, as is a dissolution unit (6). The operation is substantially as described for FIG. 2, with the following additions. The fine ammonium sulfate crystal fraction leaving the first size classifier (3) via line (f) enters second size classifier (5), where it is separated into a second fine ammonium sulfate crystal fraction and a second coarse ammonium sulfate crystal fraction. The threshold of the second classifier (5) is smaller than that of the first size classifier (3). The second fine ammonium sulfate crystal fraction passes via line (h) into dissolution unit (6), wherein at least some of the crystals therein are dissolved by solvent, which enters via line (i). The resulting stream of ammonium sulfate is passed via line (j) to mixing unit (1). The second coarse ammonium sulfate crystal fraction is recycled via line (k) to mixing unit (1).

(7) FIG. 4 depicts a further embodiment of the present invention. In addition to the set-up of FIG. 2, a second size classifying step is introduced. A slurry of fine ammonium sulfate crystals is removed from the upper part of the crystal bed in the crystallizer via line (m) and passed to a dissolution unit (7). Due to gravitational and drag effects, lighter (therefore smaller) crystals rise to the upper part of the crystal bed in the crystallizer. By controlling the slurry density in the crystallizer, the threshold size of the second size classifying step is controlled. A coarse ammonium sulfate crystal fraction therefore remains in the crystallizer. At least some of the removed fine ammonium sulfate crystals are dissolved in the dissolution unit (7), by solvent which enters via line (n). The resulting ammonium sulfate stream passes via line (p) into mixing unit (1).

(8) FIG. 5 depicts a further embodiment of the present invention. It combines the set-up and operation of each of FIG. 3 and FIG. 4. In addition to the set-up of FIG. 2, a second size classifier (5) is added, as are dissolution units (6) and (7). The fine ammonium sulfate crystal fraction leaving the first size classifier (3) via line (f) enters second size classifier (5), where it is separated into a second fine ammonium sulfate crystal fraction and a second coarse ammonium sulfate crystal fraction. The threshold of the second classifier (5) is smaller than that of the first size classifier (3). The second fine ammonium sulfate crystal fraction passes via line (h) into dissolution unit (6), wherein at least some of the crystals therein are dissolved by solvent, which enters via line (i). The resulting stream of ammonium sulfate is passed via line (j) to mixing unit (1). The second coarse ammonium sulfate crystal fraction is recycled via line (k) to mixing unit (1).

(9) Further, a fine ammonium sulfate crystal fraction is removed from the crystallizer via line (m) and passed to a dissolution unit (7). A coarse ammonium sulfate crystal fraction therefore remains in the crystallizer. At least some of the removed fine ammonium sulfate crystals are dissolved in the dissolution unit (7), by solvent which enters via line (n). The resulting ammonium sulfate stream passes via line (p) into mixing unit (1).

(10) The present invention is illustrated by, but not limited to, the following examples.

EXAMPLES

Comparative Example 1

(11) The set-up was substantially as shown in FIG. 1. A 300 m.sup.3 fluid bed Oslo-type crystallizer (2) was used with an external circulation circuit having a Begemann impeller pump (capacity 5000 m.sup.3.Math.hour.sup.1) and a heat exchanger. The crystallizer was operated by evaporation at a temperature of 90 C. 180 m.sup.3 of aqueous crystal slurry was present comprising ammonium sulfate crystals in a saturated ammonium sulfate solution in water. In the lower part of the crystallizer (2) was a crystal bed (effectively a dense slurry) with an ammonium sulfate crystal concentration in the range of 40 to 50 wt. %.

(12) Ammonium sulfate feed solution, which was obtained as a by-product in a production process for producing caprolactam, and which comprised ammonium sulfate dissolved in water (40 wt. % ammonium sulfate with respect to the solution) was introduced via line (a) into a mixing unit (1). Ammonium sulfate passed through line (b) into crystallizer (2). A slurry of ammonium sulfate crystals exited the crystallizer via line (c), through which it passed into recovery unit (4).

(13) In mixing unit (1) fresh ammonium sulfate feed solution (delivered via line (a)) was mixed with recycled feed (delivered via line (g)) from the recovery unit (4). Typically 79 wt. % ammonium sulfate crystals discharged as product via line (e) was retained by a sieve with a sieve size of 1.4 mm. The performance of the system was translated into a model which described the system using a combined mass, heat and population balance using the Borland Delphi 5.0 programming language. The population balance describing the crystal size distribution in the system was implemented according to a first order discretization scheme similar to the description given by M. J. Hounslow, R. L. Ryall, V. R. Marshall in A discretized population balance for nucleation, growth, and aggregation; AlChE J., 34 (1988) 1821-1832. The description of primary crystal nucleation and crystal growth were obtained from lab-scale experiments, whereas the description of secondary crystal nucleation was calibrated on the basis of production data. The model was used to simulate the performance of the described crystallizer system, consisting of a start-up period of approximately 30 hours and a steady-state operation period of 90 hours. The performance of the crystallizer was characterized by the weight % of crystals produced during the steady-state operation retained by a sieve of 1.4 mm as compared to the total weight of crystals produced in that period. The relative weight of crystals with a crystal size above 1.4 mm is called the granular efficiency. For Comparative Example 1, a granular efficiency of 79% was calculated.

(14) The simulation model was used to investigate the examples according to the invention by adding appropriate units to the simulation model of Comparative Example 1.

Comparative Example 2

(15) The set-up was substantially as shown in FIG. 2. In this embodiment the simulation model developed as used in Comparative Example 1 was transformed into the equipment configuration as depicted in FIG. 2, by installing a first size classifier (3) between the crystallizer (2) and the recovery unit (4). The first size classifier (3) separated the particles according to size. The separation of the particles is described by Equation (1):

(16) $\begin{array}{cc}\frac{\mathrm{separated}\mathrm{crystals}}{\mathrm{feed}\mathrm{crystals}}=\mathrm{sep}\mathrm{fac}.Math.\exp \left[-{\left(\frac{d_p}{d_{\mathrm{sep}}}\right)}^n\right]& \left(\mathrm{Equation}1\right)\end{array}$

(17) Here, sep fac is a proportionality coefficient representing the fraction of crystals separated from the stream (expressed as a percentage), d.sub.p is the size of the crystal based on sieve analysis, d.sub.sep is the characteristic crystal diameter for separation and n is the separation sharpness. Both d.sub.sep and n are parameters of the equipment. In Comparative Example 2 different threshold sizes of the first size classifier (3) were tested in the range of from 0.3 to 1.2 mm while setting the separation sharpness (n) to a value of 20, indicating a sharp separation by the first size classifier (3).

(18) TABLE-US-00001 TABLE 1 Threshold size of first size Granular Example no. classifier (3) efficiency Type of operation Comp. Ex. 1 Not applicable 79.1% Stable operation Comp. Ex. 2-I 0.3 mm 71.2% Stable operation Comp. Ex. 2-II 0.6 mm 60.0% Stable operation Comp. Ex. 2-III 0.8 mm 52.7% Stable operation Comp. Ex. 2-IV 1.0 mm 44.7% Stable operation Comp. Ex. 2-V 1.2 mm Not applicable Unstable operation

(19) Table 1 gives the results of Comparative Example 1 together with the results for different threshold sizes of the first size classifier (3) of Comparative Example 2. The granular efficiency dropped immediately when fines were returned to the crystallizer. The granular efficiency gradually dropped from approximately 80% to 45% as the threshold size of the first size classifier was increased from zero (no first size classifier as in Comparative Example 1) to 1 mm. When a threshold size of 1.2 mm was used an unstable operation of the system was seen because the system contained too large a quantity of small crystals.

Comparative Example 3

(20) The set-up was substantially as shown in FIG. 2, but with a dissolution unit inserted into line (f) (as described in relation to FIG. 3). To this dissolution unit water is added. In this embodiment, the simulation model developed as used in Comparative Example 1 was transformed by inserting a dissolution unit with a water feed into line (f). The fine ammonium sulfate crystal fraction from first size classifier (3) passed via line (f) into the dissolution unit, where the crystals therein were dissolved by addition of water. The resulting stream was returned to mixing unit (1). Different threshold sizes for the first size classifier (3) were tested in the range of from 0.3 to 1.4 mm, while setting the separation sharpness to a value of 20, indicating a sharp separation by the first size classifier.

(21) TABLE-US-00002 TABLE 2 Threshold size of first size Granular Example no. classifier (3) efficiency Type of operation Comp. Ex. 1 Not applicable 79.1% Stable operation Comp. Ex. 3-I 0.3 mm 79.1% Stable operation Comp. Ex. 3-II 0.6 mm 80.1% Stable operation Comp. Ex. 3-III 1.0 mm 84.4% Stable operation Comp. Ex. 3-IV 1.2 mm 80.4% Borderline stability Comp. Ex. 3-V 1.4 mm Not applicable Unstable operation

(22) Table 2 gives the results of Comparative Example 1 together with the results for different threshold sizes of the first size classifier (3) of Comparative Example 3. When a small threshold size was used (0.3 mm) no significant increase in granular efficiency was seen. This is because the crystallizer contained a limited amount of fines. However, an increase in granular efficiency was seen as the threshold size was increased up to 1.0 mm. Increasing the threshold size to a value of 1.2 mm decreased the granular efficiency again. This is because the system tended to become unstable, with a large variation in granular efficiency. The situation became worse when the threshold size was set to 1.4 mm, resulting in an unstable system that was not controllable and therefore would not result in a steady-state operation.

Example 1

(23) The set-up was substantially as shown in FIG. 3. In this embodiment according to the present invention, the simulation model developed as used in Comparative Example 1 was transformed into the equipment configuration as depicted in FIG. 3 by installing a first size classifier (3) between the crystallizer (2) and the recovery unit (4); a second size classifier (5) and a dissolution unit (6), connected by line (h), in line (f); a line (k) which returned a second coarse ammonium sulfate fraction back to mixing unit (1); and a line (i) through which water was fed to the dissolution unit (6). The resulting stream of ammonium sulfate is passed via line (j) to mixing unit (1). This is as described above in relation to FIG. 3.

(24) The threshold size of the first size classifier (3), was tested within a range of from 1.4 to 1.6 mm. The second size classifier (5) was maintained at 0.6 mm. The separation sharpness of both classification units were set to a value of 20, indicating a sharp separation by the classification unit. The separation factor (Equation 1) has been varied at values of 0%, 50% and 100%. Separation factor is the proportion of slurry subjected to size classification. A separation factor of 50%, indicates that half of the slurry is subjected to the size classification; the other half bypasses the size classifier. A separation factor of 0% means that there is no size classification.

(25) TABLE-US-00003 TABLE 3 Threshold size Separation factor of first/second size first/second size Granular Example no. classifier classifier efficiency Comp. Ex. 1 Not applicable Not applicable 79.1% Ex. 1-I 1.4 mm/0.6 mm 50%/0% 75.9% Ex. 1-II 1.4 mm/0.6 mm 50%/100% 83.4% Ex. 1-III 1.4 mm/0.6 mm 100%/100% 89.5% Ex. 1-IV 1.6 mm/0.6 mm 100%/100% 94.9%

(26) Table 3 gives the results of Comparative Example 1 together with the results for different threshold sizes of the first size classifier (3) tested in Example 1. Ex. 1-I effectively used the same system as that of Comp. Ex. 3-V, with the difference that the separation factor of the first size classifier (3) was set to 50%, rather than 100%. This stabilized the operation of the crystallizer, but the resulting granular efficiency of the system was lower than in Comparative Example 1 (75.9% rather than 79.1%). In general the granular efficiency was shown to increase by increasing the separation factor of the second size classifier. From Ex 1-I to Ex1-II, the granular efficiency increased from 75.9 to 83.4%, corresponding to an increase in separation factor of the second size classifier of from 0 to 100%. This increase in granular efficiency of 7.% demonstrates the effect of introducing a second size classification. A further increase of the granular efficiency from 83.4 to 89.% was shown by increasing the separation factor of the first size classifier from 50% to 100% (Ex 1-II to Ex 1-III). An increase of the granular efficiency to 94.9% was shown to be possible by increasing the threshold of the first size classifier from 1.4 to 1.6 mm. This demonstrates that optimizing size classification threshold size significantly increases the granular efficiency. Overall an increase in granular efficiency of 15.8% is demonstrated.

Example 2

(27) The set-up was substantially as shown in FIG. 4, but with a dissolution unit inserted into line (f) (as described in relation to FIG. 3). In this embodiment according to the invention double classification was achieved by applying a special feature of the Oslo-type crystallizer, namely internal size classification. Internal size classification takes place in the upper part of the fluid bed where the average crystal size has the lowest value. A fraction of the slurry was removed from the upper part of the fluid bed by drawing off 15 m.sup.3 hour.sup.1 of ammonium sulfate slurry at a liquid velocity of 21 and 42 mm.Math.s.sup.1. Thereby the second fine ammonium sulfate crystal fraction was formed; the second coarse ammonium sulfate crystal fraction being the slurry remaining in the crystallizer. The first size classifier had a threshold of 1.6 mm and a separation sharpness of 20.

(28) TABLE-US-00004 TABLE 4 Example no. Liquid velocity Granular efficiency Comp. Ex. 1 Not applicable 79.1% Ex. 2-I 21 mm .Math. s.sup.1 91.1% Ex. 2-II 42 mm .Math. s.sup.1 91.3%

(29) Table 4 gives the results of Comparative Example 1 together with the results of the two experiments of Example 2. Granular efficiency increased from 79.1% to 91.1%, by introducing a second size classifier with a liquid velocity of withdrawal of 21 mm.Math.s.sup.1. This shows that an increase in granular efficiency can be achieved by a second size classification, even when that is not directly linked to the first size classification. Doubling the liquid velocity of withdrawal from 21 to 42 mm.Math.s.sup.1 (Ex. 2-II) only increased the granular efficiency from 91.1 to 91.3%.