REACTION CRYSTALLIZATION OF SULFAMIC ACID FROM UREA AND FUMING SULFURIC ACID

Transcription

1 REACTION CRYSTALLIZATION OF SULFAMIC ACID FROM UREA AND FUMING SULFURIC ACID Ken TOYOKURA,Kenji TAWAand Junji UENO Department of Applied Chemistry, Waseda University, Tokyo 160 Reaction crystallization of sulfamic acid from urea and fuming sulfuric acid was studied in a continuous well-mixed crystallizer. In this work, growth rate and nucleation rate were correlated with retention time. Secondary nucleation obtained from two kinds of agitators was studied briefly on the basis of the collision model. On the other hand, reaction rate was represented by the pseudo-first order of urea concentration in a vessel and the correlation between reaction rate and retention time was also obtained. Because both reaction rate and crystallization rate were correlated with retention time, a new design method without use of correlative equations of supersaturation is proposed and a chart based on retention time is submitted. Intr oduction A numberof studies of crystallizer design have been carried out in recent years. Most concerned physical crystallization and have been satisfactorily applied for some industrial purposes. On the other hand, reaction crystallization is considered an important field in industrial operations and its design requires study. But the phenomena of reaction crystallization are complicated, and design of reaction crystallizers is not easy at present. Studies of reaction crystallization considering both reaction rate and crystallization rate have scarcely been reported. Design for reaction crystallization has been generally attempted by applying the theory obtained from physical crystallization. In reaction crystallization, reaction rate is effective in creating supersaturation and the idea of classification of reaction crystallization by rate of reaction is devised for development of design theories by the authors. The criterion of the classification is whether most of the chemical reaction occurs before reactants are mixed well in a vessel, or after. In the latter case, the slow reaction group, supersaturation is supposed to be homogeneous throughout a well-stirred vessel, but in the former case distribution of supersaturation in a vessel might markedly occur. Reaction crystallization of sulfamic acid is thought to occur after urea and fuming sulfuric acid are mixed well in a vessel. Its reaction rate is easily estimated by observing the generation rate of carbon dioxide gas. Therefore, this system was adopted as an appro- Received March 2, Correspondence concerning this article should be addressed to K. Toyokura. 24 priate one for study of reaction crystallization with slow rate of reaction. Growth rate and nucleation rate of crystals necessary for design are correlated with supersaturation in general. But the operational supersaturation is often difficult to observe. Particle size of product crystal is generally important and the correlation between particle size of product crystal and retention time in a continuous well-mixed vessel was proposed3} for slightly soluble system, and a new design theory based on retention time has been proposed in conformity with crystallization theory without using the correlative equation with supersaturation. Based on this theory, the above relation was extended to reaction crystallization of sulfamic acid and a correlation was obtained also for this system. On the other hand, the correlation between reaction rate and retention time was derived from the material balance at steady state. Then both reaction rate and crystallization correlated with the retention time. rate were 1. Experimental Apparatus and Procedure 1. 1 Experimental apparatus A diagram of this experimental apparatus is shown in Fig. 1. The crystallizer was a stainless steel beaker, 150 mmin diameter and 170 mmindepth. Thisvessel was dipped in a constant-temperature bath. Agitators used were of two kinds; one was the paddle type with two blades, 25mm in width and 100mm in length, and the other was the paddle type with four blades each of which had an inclination of 45 degrees, 25mm in width and 130mm in length. Carbon dioxide generated by reaction was stored in a gas tank by reduction of the inside pressure to 10 mmhg.removal of reactants was performed with an aspirator. JOURNAL OF CHEMICAL ENGINEERING OF JAPAN

2 1. 2 Experimental procedure The reaction of sulfamic acid is written as Eq. (1). NH2CONH2+SO3+H2SO4-»2NH2SO3H+CO2 (1) Urea dissolved in 100% sulfuric acid and 30% fuming sulfuric acid were dropped (25 ml each) into the vessel atthe same time from the feed tanks (3) in Fig. 1. The interval of addition was 1/10 of the retention time (The volume of solution in the vessel was 500ml). Aportion equivalent to volumeincrease as a result of adding the feed solution was drawn out from the outlet (4) as slurry and the vessel volume was thus kept constant. Carbon dioxide generated by reaction was stored in the gas tank (6) after being passed through the washer (5) filled with dilute sulfuric acid. Its volume was plotted against the elapsed time. When its as slope becameconstant, reaction rate was regarded steady state. And crystallization rate was considered to reach steady state after more than five times the retention operation continued time from attaining the steady state of reaction rate. was stopped. The weight of crystals Then the operation in the vessel and the crystal size Experimental distribution were measured. conditions were as follows : reaction temperature 90 C, molar ratio SO3/urea 1.8, the revolution rate per minute rpm for the twoblade agitator, rpm for the four-leaning blade agitator. These conditions were chosen from a preliminary test of batch operation using the same vessel in order to make the progress of reaction smooth by dropping 30% fuming sulfuric acid into the vessel containing urea dissolved in 100% sulfuric acid under several temperature ranges. Conditions chosen here may not always be the best, but this is an easy example to obtain data for discussion of reaction crystallization based on reaction rate and crystallization rate. 2. Results 2. 1 Reactionrate Production rate of sulfamic acid is given as twice the generation rate of carbon dioxide from Eq. (1). Reaction rate of urea is equal to the generation rate of carbon dioxide. The relation between urea concentration in a vessel CA[mol//] calculated from the marerial balance and reaction rate ra [mol//*min] is shown in Fig. 2. Here, circles mean the data of the preliminary batch operation and triangles those of the continuous one. From this, it was concluded that reaction rate was correlated with the pseudo-first order of urea concentration in a vessel, and Eq. (2) was obtained. r^=0.096 C^ (2) In continuous operation of this reaction, the material balance is expressed as Eq. (3). VOL. 12 NO Fig. 1 Experimental apparatus Fig. 2 Correlation between reaction rate and urea concentration in a vessel CAoG-CAG-rA V=O (3) where C^0[mol//] is the initial concentration of urea, G[l/min] is the volumetric flow rate and V[l] is the volume of crystallizer. Substituting Eq. (2) for ra in Eq. (3), CA becomes Eq. (4). CA={l/(l &)}CAo (4) where (9[min] is the retention time and is equal to V/G. The term in the parenthesis of Eq. (4) is approximated as Eq. (5) in the range of retention time used, from 30 to 250 minutes. l/(l+0.096<9)=5.8 <9- -9 (5) Consequently, Eq. (4) becomes Eq. (6) for the case of CA =2.0. CA=U.6&-0-9 (6) Combination ofeq. (2) and (6) gives Eq. (7). ra=l.l (7) To make sure of this relation, the reaction rate given in a continuous operation was plotted against the retention time. In Fig. 3, circles mean the data obtained from the two-blade agitator and triangles those from the four-leaning blade agitator. The line 25

3 Fig. 3 Correlation between reaction rate and retention time Fig. 4 Correlation between production rate and reaction rate Fig. 5 Size distribution of this figure is the relation of Eq. (7), which is regarded as holding good nearly throughout this system. Andthis figure makes it clear that reaction rate was not affected by the difference of agitators for the range of these experimental conditions. On the other hand, production rate of sulfamic acid P[mol//-min] is given as twice the reaction rate from Eq. (1), but the correlation was given by Eq. (8), as shown in Fig. 4. The distinction of data points in Fig. 4 is the same as in Fig. 3. P=arA (8) In these tests, the value ofa was This difference between test data and the value estimated from Eq. (1) is supposed to be caused by the double decomposition of product crystals and others, but details were not studied here Growthrate and nucleation rate Growthrate was measuredby a commonmethod using the results of the size distribution as follows. Figure 5 is an example of the size distribution obtained from this experiment. The vertical axis is the logarithm of the dimensionless numberof particles and the horizontal axis is the particle size. Growth rate was calculated from the slope of this straight line expressed by the tangent using the next equation. tangent= -2. 3O3(dl/d0)O (9) In a continuous well-mixed crystallizer, the straight line in this method shows that AL law is applicable. When AL law is applied to a continuous well-mixed crystallizer, following equations have already been reported2k P=(l-e)PcV'/6 (10) lm= 3(dl/dd)0 (1 1) Fi= (9/2)P/Pc V' lm3 (12) By these equations, dominant size Im and nucleation rate Fvr- in these tests were calculated from test data of P, dl/dd and 0. On the other hand, the correlation between growth rate and retention time has already been reported3) in the latest paper for the precipitation of slightly soluble salts by ionic reaction. Since in this reaction crystallization of sulfamic acid the solubility of sulfamic acid is comparatively small and this relation is convenient to apply to discussion of industrial operations, growth rate and nucleation rate calculated by Eq. (12) were plotted against retention time in Figs. 6 and 7, respectively. From these figures, the correlations were given as follows: Whenusing the two-blade agitator, dl/d6=2l.60- '87 (13) F^H^x lo (14) Whenusing the four-leaning blade agitator, dl/d6=2sj (15) 3. Discussion Frv=635x Wl0-1 2" (16) 3. 1 Crystallization rate In a continuous well-mixed crystallizer, the relation between growth rate and retention time is derived from Eqs. (ll) and (12) as Eq. (17) dl / P \1/3 "'-=( ) /7/-l/3@-l (IJ\ dd \69cv 26 JOURNAL OF CHEMICAL ENGINEERING OF JAPAN

4 The power of the retention time for growth rate in Eq. (17) is different from that in Eqs. (13) and (15). This supposes that production rate and nucleation rate are functions of retention time, respectively. On the other hand, growth rate obtained from the four-leaning blade agitator was about 1.3 times that from the two-blade agitator. When growth rate is postulated to be represented by Eq. (18) as the function of supersaturation AC, the difference of growth rate caused by the two kinds of agitators in Fig. 6 is considered to result from the difference of supersaturation because the effect offluidity is thought to be not so great under good mixing conditions in each case. di/do=k(ac) (18) In Fig. 7, though scattered plots can't show the situation, secondary nucleation rate obtained from the two-blade agitator was about 2.3 times that from the four-leaning blade agitator. Mechanism of secondary nucleation which is considered to take place in a vessel is complicated, but an explanation was given by applying the collision model reported by de Jong et al}\ In these tests, the difference of experimental conditions for nucleation rate is the revolution rate per minute and the size and shape of agitator, and the factor affecting secondary nucleation is considered to be the revolution rate and impeller area for the direction of revolution. The power of revolution rate for secondary nucleation rate has been reported as 3 by de Jong and as 2 to 3 by the experimental results of K-alum system carried out by Toyokura4). Considering that the contribution of supersaturation to secondary nucleation rate is the same as that to growth rate in the collision model4}, and assuming that secondary nucleation rate is proportional to impeller area for the direction of revolution, the power of revolution rate for secondary nucleation rate was calculated as 2.8 in this case. This experimental value was in the range described above, but details are subjects for the future Application to design method The correlation between crystallization rate and retention time make the following method possible. Combination ofeqs. (10) and (12) gives Eq. (19). i^=(9/2)(l -s)/6> /m3 (19) Since the production rate is a function of reaction rate, 1- was shown to be a function of the retention time like Eq. (20), from Eq. (10). l- =0.096>0 1 (20) In Eq. (19), when 1-s is constant or a function of retention time, this equation shows the relation of Fv\ & and Im. Figure 8 shows the relation between Fv' and 0 with Im parameters when 1-s is represented by a function of retention time like Eq. (20). Two Fv'-0 lines in Fig. 8 are test data shown in Fig. 7. VOL. 12 NO Fig. 6 Correlation between growth rate and retention time Fig. 7 Correlation between nucleation rate and retention time Fig. 8 Correlation between nucleation rates and retention time for determination of design conditions If the relation between Fv' and 6 is obtained in a crystallization experiment, a series of crosspoints are given from two kinds of Fvr-0 lines, like this figure 27

5 and Fv' and 6 are decided for the desired particle size Im. The crystallizer volume is decided from Eq. (12) for the production rate set up. According to this chart, a long retention time is necessary to obtain large particle size, and to get large particle size with good efficiencies, Fv'-& line must be lowered. These two Fv'-& lines show the possibility of lowering this line by studying secondaly nucleation. This chart is considered to indicate the direction of development of processes like this, but detailed study of secondary nucleation is necessary for this purpose. Nomenclature ca AC Fv' G K I Im urea concentration in a vessel [mol//j initial concentration of urea [mol//] supersaturation [mol// or Kg-mol/m3] nucleation rate per unit volume [number/m3 à"hr] volumetric flow rate overall growth rate coefficient crystal size dominant size of crystals numberof crystals pi production rate reaction rate of urea volume of crystallizer coefficien t void fraction retention time time density of crystal [Kg/hr, mol//- min] [mol// - min] [m3] [-] [-] [min or hr] [min or hr] [Kg/m3] Literature Cited 1) E, P. K. Ottense and E. J. de Jong: /. Cryst. Growth, 13, 114, 500 (1974). 2) Shirotsuka, T. and K. Toyokura: AIChE Symposium Series, 67, No. 110, 145 (1971). 3) Toyokura, K. and K. Ono: Kagaku Kogaku Ronbunshu, 4, 38 (1978). 4) Toyokura, K. : to be published in "Industrial Clystallization '78", E. J. de Jong-editor, North Holland Publishing Company, Amsterdam. (Presented at the 42 nd Annual Meeting of The Soc. of Chem. Engrs., Japan, at Hiroshima, April 1977.) 28 JOURNAL OF CHEMICAL ENGINEERING OF JAPAN