EFFECT OF SOLIDS CONCENTRATION ON SOLID-LIQUID MASS TRANSFER IN AN AGITATED DISSOLUTION SYSTEM Chongguang Yu¹, Rajarathinam Parthasarathy 1 *, Jie Wu 2 and Dr. Nicky Eshtiaghi 1 1 School of Civil, Environmental and Chemical Engineering, RMIT University, Melbourne, 3001, Australia 2 CSIRO, Division of Process Science and Engineering, CSIRO, Highett, 3190, Australia *Corresponding author, E-mail: rchrp@rmit.edu.au ABSTRACT Suspension of solids in mechanically agitated vessels is important for many industrial solid-liquid processes such as dissolution and leaching. Although solid-liquid mixing in agitated vessels has been studied extensively, mass transfer in such systems, especially under high solid loadings, has received insufficient attention. Also, the current knowledge on the effect of active particle mass fraction on mass transfer is limited. This paper investigates the effects of solid concentration and active particle mass fraction on solid-liquid mass transfer in an agitated dissolution system under baffled conditions. Experiments were carried out in a 0.2 m diameter vessel using Rushton turbine as the impeller. Water and glass particles (inert particles) were used as the liquid and solid phases, respectively. Glass particles coated with benzoic acid were used as active particles. Total solids concentration C V was varied from 3 to 30% (v/v) and the mass fraction of active particles in the total solids was varied from 0.03 to 0.10. Volumetric solid-liquid mass transfer coefficient k sl a p was determined by measuring the changes in the conductivity values of water due to the dissolution of benzoic acid from solid surface. Experimental results show that k sl a p at just off-bottom solid suspension condition increases with increase in C v. Higher the mass fraction of active particle in total solids, greater the k sl a p for all C v. These results imply that it is possible to intensify the dissolution process by increasing the solids concentration in the system and still achieve significant improvement in mass transfer. INTRODUCTION Solid-liquid mixing in a mechanically agitated vessel is one of the most common unit operations in chemical industry. Examples of industrial operations which are designed on the knowledge of solid-liquid mass transfer between suspended solids and fluids are bioreactor, slurry reactor, crystallisation and growth processes, dissolution processes, ionexchange and polymeric processes (Paul, 2003). In all these operations, the process efficiency is influenced significantly by the rate of mass transfer between the solid and liquid phases and the degree of solid suspension.
Solids concentration is usually defined as the concentration of total solids, including both active and inert particles, in a solid-liquid system. Most of the papers in the literature focussing on the effects of operating and system parameters on solid-liquid mass transfer involve low total solids concentrations. However, according to a recent paper by Wu et al (2010), a higher return on capital investment in mineral industry can be obtained by increasing the system throughput and using high total solids concentration in the slurry feed. This method is known as process intensification in which production yield per unit volume per unit time can be increased significantly (Wu et al., 2007). Recently, Wang (2010) and Wu et al (2010) have reported that energy efficiency in a solid-liquid agitated vessel could be increased significantly by operating the system under high solids concentration. They have also reported that impeller energy efficiency will be maximised by operating the system at an optimum solids concentration which is higher than that is normally used in industrial slurry tanks. However, the knowledge on the effects of high solids concentration on solid-liquid mass transfer is lacking significantly in the literature. This study aims to investigate the effect of solids concentration on mass transfer by determining the solid-liquid mass transfer coefficient at various solids concentrations. Effects of solids concentration on solid-liquid mass transfer coefficient have been investigated by many researchers in the past but these studies are limited to low solids concentrations. Also there is no consensus in their findings on solids concentration effect. Barker and Treybal (1960) reported there is no significant effect of solids concentration on solid-liquid mass transfer coefficient from their study involving 10% (volume) active particles. In contrary, Harriott (1962) and Brain et al (1969) have found that solid-liquid mass transfer coefficient increases with increase in active particle concentration. In agitated mass transfer operations carried out in mineral processing industry, not all ore particles contribute to the mass transfer. Significant amount of ore particles is inert and does not participate in mass transfer. However, their presence influences various operating parameters such as just off-bottom suspension impeller speed and power input, and thus the mass transfer rate. The information available on the effect of inert particle concentration on solid-liquid mass transfer coefficient is very limited. In one of the few studies made on inert particle concentration effect on mass transfer, Iwanaka et al (1985) found that solid-liquid mass transfer coefficient would decrease with an increase in inert particle concentration. Yang and Renken (1998), on the other hand, reported that mass transfer rate would increase with increase in inert particle concentration but for a fluidised bed. It was clear from our literature review that most of the previous solidliquid mass transfer studies involved solids that were all active in mass transfer. But, the ore particles processed in mineral industry are usually a mixture of active and inert particles with either a constant or varying volume ratio. Very few attempts have been made so far to investigate the effects of varying mass fraction of active particles on solidliquid mass transfer coefficient. This study aims to do that. 2
EXPERIMENTAL Experimental set-up All experiments were carried out in a 0.2 m diameter (T) cylindrical, flat-bottom perspex tank placed inside a rectangular outer perspex tank (Figure 1). The cylindrical tank was equipped with four equally spaced baffles with width (B) to tank diameter ratio (B/T) of 1/12. The space between the inner and outer tanks was filled with tap water to minimize the optical distortion caused by curvature of the inner tank during flow visualisation. A 6-bladed Rushton turbine with a diameter of ⅓T was used as the impeller. The impeller was located at the axis of the tank. Impeller off-bottom clearance was maintained at ¼T to minimise the impeller power input required for just off-bottom suspension (Levins and Gladstone, 1972; Nienow and Miles 1978; Nienow and Bujalski, 2002). Impeller Speed and Power Input Impeller speed was read from motor (Heidolph RzR-2102) display. Impeller power input (W) was determined by measuring the torque experienced by the impeller shaft and using the following equation: P = 2πNτ (1) where N is the impeller speed (rps) and τ is the impeller torque (N.m). A torque transducer (Burster Series-NR) attached to the impeller shaft was used to measure the torque. Voltage signals from the torque transducer was converted into digital signal by an A/D converter and sent to a computer for data logging. Impeller Speed for Just Off-bottom Suspension (N js ) In most of the studies on solids suspension in agitated vessels, impeller speed required to achieve just off-bottom suspension is determined using the criterion proposed by Zwietering (Zwietering, 1958). According to this criterion, the just-off-bottom suspension is the condition at which no solids remain stationary at the tank bottom for more than 1 2 s. This criterion is not useful for suspensions with high solids concentration because it is possible for a small proportion of particle to remain stagnant in bottom corners and not being suspended. To achieve the suspension these particles, impeller speed needs to be increased significantly without any consequential benefit in terms of mass transfer or reaction conversion (Wu et al, 2006). Therefore an alternative approach proposed by Wu et al was used in this work. This approach involves using sedimentation bed height (H B ) at the tank bottom for N js determination. According to this method, N js is the impeller speed at which the sedimentation bed disappears (H B = 0). To determine N js in our experiments, the impeller speed was increased initially to a sufficiently high value until a uniform suspension was achieved. Impeller speed was then decreased gradually until a thin settled solid bed appears at the tank bottom. The impeller speed was then increased slightly until the solid bed disappears. The speed at which the 3
solid bed disappears was designated as N js. This method was tested many times for reproducibility and N js was found to vary by ± 2 to 3 rpm. Specific Impeller Power Input at N js (ε js ) Specific impeller power input (ε js ) in this work was determined according to the following equation: ε P 2πNτ js js = = (2) M s M s where P js is the impeller power input at N js and M s is the mass of solids suspended. This definition of ε js was proposed by Drewer et al (2001) on the consideration that the rate of mass transfer in solid-liquid systems is independent of vessel volume and dependent only on the surface area and mass of particles at N js. Further increase in specific power input or impeller speed will not increase the solid-liquid interfacial area any further and is therefore unnecessary. Based on this consideration, specific impeller power input was determined according to equation (2) and used throughout in this work. However it should be noted that this measure is not applicable for systems requiring homogenous suspension. Active and Inert Particles Tap water and glass particles were used as the liquid and solid phases, respectively in this work. Benzoic acid was chosen as the solute that is transferred from the solid to liquid phase in mass transfer experiments. Benzoic acid is frequently used in mass transfer studies because it readily dissolves in water and its concentration can easily be measured using conductivity of the solution. To simulate the dissolution process that occurs in mineral processing operations where mass is transferred from ore particles to the liquid phase, a model system involving glass particles coated with benzoic acid and water was chosen in this work. Spherical glass particles used in this work were supplied by Potters Industries (Grade B). Laboratory grade benzoic was used in the coating process. Details of benzoic acid and glass particles are shown in Table 1. Coating of glass particles with benzoic acid involved melting of the benzoic acid particles at 120ºC and soaking the glass particles in the molten acid. The slurry containing glass particles and benzoic acid solution was then cooled to room temperature under constant stirring. This process was repeated with the same batch of glass particles until about 90% of glass particles were coated. Mass of the glass particles was measured before and after the coating to determine the mass of benzoic acid present in the coating. Similarly, Sauter mean diameter (d 32 ) of the particles before and after the coating was determined from particle size distribution data obtained using Malvern particle size analyser (2400E) to find out the change in d 32 due to coating. In mass transfer experiments, coated (active) glass particles were mixed with inert (uncoated) glass particles and used. The mass fraction of active particles (M a ) was varied over a range of 0.03 to 0.1. 4
Tab.1: Particle properties Substance d32(µm) Standard deviation (µm) 3 Density ( kgm ) Benzoic acid 320 69 1270 Glass particles 350 76 2500 Coated particles 355 77 2500 Fig.1: Schematic of the experimental setup, T = 200 mm. Mass Transfer Measurement Diffusional mass transfer from suspended solids was measured using the dissolution of benzoic acid from the coated glass particle in water. Noyes and Whitney (1987) proposed the following equation which relates the rate of solute concentration change in the liquid phase to the mass transfer coefficient k sl. dc dt = k a ( C C) (3) sl p s where k sl (ms -1 ) is the solid-liquid mass transfer coefficient and a p (m 2 /m 3 ) is the solidliquid interfacial area, C s (kgm -3 ) is the saturation concentration of solute in the liquid 5
phase and C (kgm -3 ) is the concentration of solute in the liquid phase. The product of k sl and a p is the volumetric solid-liquid mass transfer coefficient k sl a p (s -1 ). On integrating equation (3) with respect to time, the following equation is obtained: Cs Ct ksla pt = ln (4) C C s C t (kgm -3 ) represents the concentration of benzoic acid in water at a given time t (s) and C 0 (kgm -3 ) is the initial concentration of benzoic acid in water, in this case it is close to Cs Ct zero. When the benzoic acid concentration ( ) is plotted against time on a Cs C 0 logarithmic graph, a straight line is obtained. Mass transfer rate constant k sl a p is then found from the slope of the straight line. Concentration of benzoic acid in water was determined by measuring the conductivity of water. Conductivity of benzoic acid solution varies linearly with benzoic acid concentration in water. A calibrated conductivity meter was used in our experiments to measure the changes in benzoic acid concentration. The temperature of the solid-liquid system was maintained at 25ºC in all experiments. To determine the mass transfer coefficient k sl from k sl a p, interfacial area a p is required. It was determined using the following equation: a 6C ( d32 ) a 0 a p = (5) where (d 32 ) a and C a are the Sauter-mean diameter and volume fraction of active particles (coated glass particles) in the slurry. The volume fraction of active particles C a in total solids was determined by substituting C a = C v.r ai in equation (5) as follows: a 6C R v ai p = (6) ( d 32 ) a where C v is the volume fraction of total solids in the slurry (v/v) and R ai is the volume fraction of active particles in total solids. RESULTS AND DISCUSSION Optimum Solids Concentration in Terms of Power Consumption The ε js values obtained at N js at various solids concentrations C v % (v/v) are shown in Figure 2. The solid-liquid system used in this experiment was water and uncoated (inert) particles. It can be seen clearly from this Figure that ε js decreases with increase in C v, reaches a minimum and starts increasing thereafter. The C v value at which ε js is minimum is designated as the optimum solids concentration (C v ) opt. This is purely based on the impeller energy efficiency. In other words, at (C v ) opt, more solids per Watt of power input are suspended off the tank bottom compared to those at other C v. For example, 6
0.036 kg of solids can be suspended per Watt for C v = 5% (v/v) while 0.063 kg of solids can be suspended per Watt for C v = 25% (v/v). These results show that energy efficiency in solid-liquid mixing can be increased significantly by operating the system at (C v ) opt. These results are consistent with the findings reported in previous studies (Wu et al. 2007; Wang 2010; Wu et al. 2010). These studies reported (C v ) opt values of 25 to 35% (v/v) depending on the impeller type. 60 50 40 Minimum ε js ε(w/kg) 30 20 10 0 0% 5% 10% 15% 20% 25% 30% 35% 40% Soilds concentration Cv (v/v%) Fig.2: Specific impeller power input ε js at N js vs. total solids concentration C v, System: uncoated (inert) glass particles-water. Effect of Solids Concentration on Solid-Liquid Mass Transfer Rate The effect of solids concentration on volumetric solid-liquid mass transfer coefficient k sl a p is shown in Figure 3 for a range of active particle mass fraction values M a. It is clear that for all M a values, k sl a p increases initially with increasing C v and reaches a maximum value and decreases thereafter. For all M a values, k sl a p is maximum for a C v of 15% (v/v). These results indicate that this system will yield optimum results in terms of mass transfer rate when it is operated at C v = 15% (v/v). The decrease in k sl a p for C v > 20% (v/v) could be due to many reasons. The N js required for suspending solids at C v > 20% (v/v) was fairly high and it led to surface aeration of bubbles which could be interfering with mass transfer. Also the decrease in k sl a p at C v > 20% (v/v) is certainly due to a decrease in k sl because a p values at these C v are higher than that at C v = 15% (v/v). In order to understand the influence of C v on k sl a p clearly, therefore, it is essential to study the trend in k sl as a function of C v. Figure 3 also illustrates that higher M a value leads to higher value of k sl a p indicating that it is more efficient to operate this system at a higher mass fraction of active particles for achieving higher k sl a p. It is interesting to note that at low solid concentrations (1-5% (v/v)), M a has greater effect on k sl a p. But as C v increases, the effect of M a becomes less on k sl a p. 7
Fig.3: Effect of solids concentration C v on volumetric solid-liquid mass transfer coefficient k sl a p for various mass fraction of active particles M a. Effect of Solids Concentration on Solid-Liquid Mass Transfer Coefficient Figure 4 shows k sl values as a function of C v for various M a values used in this work. It is interesting to note that as C v increases, k sl decreases for all M a. It is also clear that k sl value for a given C v decreases with increase in M a. For C v = 5% (v/v), k sl value is maximum for all M a values. When these results are considered along with k sl a p results shown in Figure 3, it is clear that the increase in k sl a p with increase in C v is mainly due to the increase in a p for C v < 15% (v/v). For C v > 15% (v/v), the decrease in k sl a p with increase in C v is certainly due to the decrease in k sl values. In this case, the extent of decrease in k sl with increase C v is greater than the increase in a p with increase in C v. Effect of Mass Fraction of Active Particles on Solid-Liquid Mass Transfer Coefficient It is clear from Figure 4, k sl decreases as M a increases. However the extent of decrease is smaller for C v > 20% (v/v). In other words, there is no significant difference in k sl values with increase in M a values for C v = 20 and 25% (v/v). Considering k sl a p data in Figure 3, for C v > 20% (v/v), it seems that a p for active particles increases by a much greater extent than the decrease in k sl. These results indicate that there might be an optimum operating C v value for this system; however, this needs to be determined by taking both k sl and ε js values together. Further studies are conducted in this work to determine an optimum C v. 8
Fig.4: Effect of solids concentration C v on mass transfer coefficient k sl at various mass fraction of active particles M a. CONCLUSION Effect of solids concentration on ε js (specific power input at N js ) and k sl (solid-liquid mass transfer coefficient) was studied using coated benzoic acid particles/water system. A minimum value for ε js was found at an optimum solid concentration of 25 % (v/v). The k sl a p for this system was found to be the highest for total solids concentration of 15% (v/v). In addition, k sl a p value for a given total solids concentration was found to increase with an increase in mass fraction of active particle M a. However, k sl was found to decreases as solids concentration increases. Furthermore, k sl for a given solids concentration was found to decrease as mass fraction of active particles increases. Further studies are required to determine an optimum operating solids concentration which will yield better mass transfer results with higher impeller energy efficiency. In addition, the effect of impeller type on mass transfer under high solids concentration condition will be studied. REFERENCES Barker, J. J. and R. E. Treybal (1960). "Mass Transfer Coefficients for Solids Suspended in Agitated Liquids." Aiche Journal 6(2): 289-295. 9
Brian, P. L. T., H. B. Hales, et al. (1969). "Transport of heat and mass between liquids and spherical particles in an agitated tank." Aiche Journal 15(5): 727-733. Drew, G. R., N. Ahmad, et al. (2001). A optimum concentration for suspension of solids in stirred vessel. Mixing and Crystallization. Netherlands, Kluwer Academic Publishers Edward L. Paul, V. A.-O., Suzanne M. Kresta, Ed. (2003). Handbook of Industrial Mixing: Science and Practice. NJ, John Wiley & Sons. Harriott, P. (1962). "Mass Transfer to Particles.1. Suspended in Agitated Tanks." Aiche Journal 8(1): 93-101. Iwanaka, H., K. Tojo, et al. (1985). "Mass Transfer Characteristics in a slurry reactor" Chemical Engineering Communications 34(1-6): 205-212. Levins, D. M. and Glastonb.Jr (1972). "Particle-Liquid Hydrodynamics and Mass- Transfer in a Stirred Vessel.1. Particle-Liquid Motion." Transactions of the Institution of Chemical Engineers and the Chemical Engineer 50(1): 32-&. Nienow, A. W. and W. Bujalski (2002). "Recent Studies on Agitated Three-Phase (Gas- Solid-Liquid) Systems in the Turbulent Regime." Chemical Engineering Research and Design 80(8): 832-838. Nienow, A. W. and D. Miles (1978). "The effect of impeller/tank, configurations on fluid-particle mass transfer." The Chemical Engineering Journal 15(1): 13-24. Noyes, A. A. and W. R. Whitney (1897). "The Rate of Solution of Solid Substances in Their Own Solutions." J. Am. Chem. Soc. 19: 930. Wang, J. (2010). Suspension of High Concentration Slurry in Agitated Vessels. School of Civil, Environmental and Chemical Engineering. Melbourne, RMIT university. Degree of Master of Engineering: 125. Wu, J., L. Graham, et al. (2010). "Mixing Intensification for the Mineral Industry." Canadian Journal of Chemical Engineering 88(3): 447-454. Wu, J., L. J. Graham, et al. (2006). "Energy efficiency study on axial flow impellers." Chemical Engineering and Processing: Process Intensification 45(8): 625-632. Wu, J., L. J. Graham, et al. (2007). "Intensification of mixing." Journal of Chemical Engineering of Japan 40(11): 890-895. Yang, J. and A. Renken (1998). "Intensification of mass transfer in liquid fluidized beds with inert particles." Chemical Engineering and Processing: Process Intensification 37(6): 537-544. Zwietering, T. N. (1958). "Suspending of Solid Particles in Liquid by Agitators." Chemical Engineering Science 8(3-4): 244-253. 2