Separation and Purification Technology 38 (2004) 139–147 Development of a novel moving bed with liquid-pulse and experimental analysis of nickel removal from acidic solution Hideaki Tokuyama a,∗, Shougo Maeda b, Katsuroku Takahashi b a Department of Chemical Engineering, Hiroshima University, 1-4-1 Kagamiyama, Higashi-Hiroshima 739-8527, Japan b Department of Chemical Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan Received 21 July 2003; received in revised form 27 October 2003; accepted 3 November 2003 Abstract A prototype of moving bed has been developed and applied to remove Ni from model solutions of industrial wastewater for the recycling of the solution and the metal. The apparatus has been simply designed by installing a plate with a hole to the bottom of resin bed and by giving a pulse to liquid. This apparatus allows us to control the resin flow rate by changing the liquid flow rate and the cross-sectional area of the hole. The liquid-pulse gives dual advantages; the prevention of clogging of resins and stabilization of the resin flow rate. Removal of Ni from the acidic solution has been carried out by using a strong cation resin, DIAION SK1B and the column performance has been evaluated in terms of removal fraction of Ni and ion-loading fraction of the resin under the various conditions. The column showed the high performance; the removal fraction was more than 0.9 in every system investigated and the ion-loading fraction was over 0.8 in some cases. Good separation of high removal and ion-loading fractions is achieved by the operation with the large bed height and with the large hole diameter. The continuous removal of Ni in the various concentrations of HCl has been discussed with the ion exchange equilibrium and kinetics. The values of equilibrium coefficient and ion exchange capacity have been determined in relation to the concentration of hydrogen ion. The effective diffusivity of Ni within the resin particle increases with the increase in the concentration of hydrogen ion. © 2003 Elsevier B.V. All rights reserved. Keywords: Ion exchange; Moving bed; Nickel; Equilibrium; Kinetics 1. Introduction In recent years, the recovery of heavy metals from industrial process solutions has attracted a great deal of attention. The necessity to recover metals is mainly due to the rise in the environment awareness and the consequent severity of legislation regarding the dis- ∗ Corresponding author. Tel.: +81-824-24-7720; fax: +81-824-24-7720. E-mail address:
[email protected] (H. Tokuyama). posal of toxic substances; for instance, the effluent limits for Ni ions in Germany, Switzerland and the US are 0.5 ppm (1991), 2 ppm (1991) and 3.98 ppm (2001), respectively [1]. No side-contamination of solution during removal of metals is preferable to the recycling all solution and the metals, which is the ideal solution for wastewater problems. In this regard, ion exchange method must be advantageous because target ions are removed without adding chemicals. In the methods of extraction and precipitation, additional facilities will be indispensable to remove residual 1383-5866/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.seppur.2003.11.001 140 H. Tokuyama et al. / Separation and Purification Technology 38 (2004) 139–147 Nomenclature a surface area of resin particles per unit volume (m2/m3) A cross-sectional area of column (m2) Ah cross-sectional area of a hole in Teflon plate (m2) C concentration in liquid phase (kmol/m3) Ce equilibrium concentration in liquid phase (kmol/m3) Cin feed concentration of Ni in liquid phase (kmol/m3) Cout concentration of Ni in effluent (kmol/m3) C0 initial concentration in liquid phase (kmol/m3) C∗ concentration of Ni in liquid phase in equilibrium with q (kmol/m3) dp particle diameter (m) De effective diffusivity within resin particle (m2/s) DF diffusivity of Ni in liquid phase (m2/s) Dh diameter of a hole in Teflon plate (m) h moving bed height or fixed bed height (m) kF mass transfer coefficient in liquid phase (m/s) kS mass transfer coefficient in resin phase (m/s) K equilibrium constant (m3-dry resin/m3) KF overall mass transfer coefficient (m/s) m distribution ratio of Ni at liquid–resin film interface (m3/m3-dry resin) q average concentration within resin particle (kmol/m3-dry resin) qe equilibrium concentration in resin phase (kmol/m3-dry resin) qmax concentration of Ni in resin phase in equilibrium with Cin (kmol/m3-dry resin) qout outlet concentration of Ni in resin phase (kmol/m3-dry resin) Q ion exchange capacity (kmol-Ni/m3-dry resin) QL liquid flow rate (m3/s) QR resin flow rate (m3-dry resin/s) Re Reynolds number, udpρ/µ Sc Schmit number, µ/ρDF t time (s) u superficial velocity of aqueous solution (m/s) VL volume of solution (m3) VR volume of resins (m3-dry resin) Greek letters ε void fraction of bed µ viscosity of liquid (Pa s) ρ density of liquid (kg/m3) chemicals. Studies have been carried out with regard to the application of ion exchange resins for the re- covery of heavy metals from various wastewaters of metal plating, alloy, pigments and so on [2–5]. Ion exchange resin is widely used for separation processes in industries. Most of them are operated with the packed bed. The narrow adsorption band in the packed bed results in the use of a large amount of resins and the non-working resins give large pres- sure drop, resulting in a low throughput. A counter- current operation in a continuous process increases the viability of process to reuse solution. A number of solid–liquid contactors, such as moving and multi- stage fluidized bed, have been developed [6–10] and those contactors show high performance; a high level in ion removal and in ion-loading of resin [9], and a continuous mutual separation of ions [10]. However the control of resins, a steady flow through the column and a constant supply and withdrawal, is rigorous [11]. The difficulty is one of the main factors that disturb the application of the countercurrent operations. In this study, a prototype of moving bed has been developed to obtain the stable flow of resins easily. The apparatus has been simply designed by installing a plate with a hole to the bottom of resin bed and by giving a pulse to liquid. Our unique technique to con- trol resin flow is an excellent way to prevent resin pul- verization compared with mechanical control by valve [6,7]. The column with a cation resin has been ap- plied for the removal of Ni from the model solution of wastewater, such as a rinsewater for metal finishing [4] and a solution leached from oil fly ash [12]. Its H. Tokuyama et al. / Separation and Purification Technology 38 (2004) 139–147 141 feasibility for metal recovery has been examined from the fractions of the removal of Ni and the ion-loading of resins under the various conditions of bed height, hole size and concentration of HCl. In addition, the ion exchange equilibrium and kinetics have been in- vestigated and related with concentration of HCl. 2. Experimental 2.1. Measurement of ion exchange equilibrium A commercial strong cation resin, DIAION SK1B (Mitsubishi Chemical Co.), was used in this study and the characteristics are listed in Table 1. Prior to all measurements, the resins were converted from Na-form to H-form with HCl solution. The resin par- ticles were sieved to obtain narrow profile in diameter distribution. Model solution of wastewater is the aque- ous solution of NiCl2 dissolved in HCl solution. The solution were prepared as follows; the initial concen- tration of Ni, CNi,0, was 0.0170 kmol/m3 and that of HCl, CH,0, was 1.0, 0.5, 0.1, 0.05 or 0.01 kmol/m3. Ion exchange equilibrium between the aqueous so- lution of Ni and the resin was measured by a batch method, shaking for 2 h in a test tube at 298 K. The equilibrium concentration of Ni in the liquid phase, CNi,e, was measured with ICP-AES. The equilibrium concentration of Ni in the resin phase, qNi,e, was de- termined from the mass balance as follows. qNi,e = ( VL VR ) (CNi,0 − CNi,e) (1) where VL and VR are the volumes of the solution (20 cm3) and the resins (0.01–4.0 cm3-dry resin), re- spectively. In addition, the concentrations of H+ were determined from the pH measurement or the titration with NaOH solution. Table 1 Characteristics of DIAION SK1B Polymer matrix Gel, styrene-DVB Functional group –SO3− Density (kg/m3) 1276 Particle diameter (mm) 0.71–0.85 Moisture retentiona (%) 43–50 a Data based on the supplied technical catalogue. 2.2. Kinetics measurement Effective diffusivity of Ni within the resin parti- cle was determined by a shallow fixed bed column method. The bed with SK1B was 10.0 mm in height and 12.5 mm in inside diameter. The feed concentra- tions of Ni and HCl were the same as those in the equi- librium experiment. The solution was fed to the bottom of the bed at a constant velocity (u = 0.136 mm/s) by a pump at 298 K. The concentrations of Ni, Cout, in the effluent with time were measured by ICP-AES. 2.3. Moving bed operation The schematic diagram of the moving bed apparatus is shown in Fig. 1. The bed column made of glass pipe has the inside diameter of 13.0 mm and the bed height, h, is variable (30, 51 and 97 mm). A Teflon plate with a hole in the center (No. 2 in Fig. 1) is installed at the bottom of the resin bed. The plate works to interrupt the liquid and resin flows in the column. The four kinds of the plate were used and their dimensions were as follows; the thickness of 10.0 mm and the diameter of the hole, Dh, of 3.11, 3.28, 3.57 and 4.06 mm. The resin stocked in the feed tank falls down through the Fig. 1. Schematic diagram of moving bed: (1) moving bed, (2) Teflon plate for QR-controlling, (3) pump for giving liquid pulse, (4) feed tank, (5) pump to feed, (6) leveler, (7) manometer, (8) resin feed tank, (9) orifice, (10) resin storage tank. 142 H. Tokuyama et al. / Separation and Purification Technology 38 (2004) 139–147 bed and the hole by the gravity and is finally exhausted into the storage tank. Liquid is fed to the bottom of the column by a pump through a leveler and overflows from the top of the column through another leveler. Another pump (No. 3 in Fig. 1) gives a pulse to the liquid, which prevents the cogging of the resin at the hole. The frequency and the amplitude of the pulse used were 1.2 s−1 and 0.225 cm3, respectively. The liq- uid and resin flow rates, QL and QR, were determined by an orifice-manometer and by the resin volume dis- charged during a fixed time, respectively. The resin flow rate was examined for the variables of QL, Dh and h. During countercurrent operation, the instantaneous QR was measured with/without liquid-pulse to reveal that effect. In addition, the time from stopping the pulse to clogging of resins was measured for various QL’s. Nickel removal from the model solution of wastew- ater was continuously operated with SK1B resin bed. The solutions were prepared as follows; the feed con- centration of Ni, Cin, was 0.00852 kmol/m3 and that of HCl, CH,0, was 0.01, 0.05 or 0.1 kmol/m3. The concentrations of Ni in the effluent were measured with ICP-AES at a steady state after flowing resin of twice bed volume. The removal fractions were examined for the variables of QL, Dh, h and CH,0 at 298± 1 K. 3. Results and discussion 3.1. Ion exchange equilibrium The ion exchange equilibrium between the aqueous solution of Ni and SK1B resin is shown in Fig. 2. The values of qNi for a given CNi decrease with increasing the concentration of HCl, which indicates the higher instability of Ni on the resin in a higher HCl solution. This implies that the resins loaded with Ni can be re- generated by using HCl solution of higher concentra- tion. The solid curves refer to the isotherms calculated on the basis of equilibrium for an ion exchange re- action. To establish the reaction, the exchanging ratio of Ni to the functional group of the resin was deter- mined. The ratio was obtained as 1/2 from the relation between amount of Ni adsorbed and that of H+ des- orbed, which was examined from the concentrations of each ion in the liquid phase before and after equi- 10-4 10-3 10-2 0.1 1 C H, 0 [kmol/m3] 1.0 0.5 0.1 0.05 0.01 q N i, e [ km ol /m 3 - dr y re si n] C Ni, e [kmol/m3] Fig. 2. Equilibrium isotherms between aqueous solution of Ni and SK1B resin for various initial concentrations of HCl at 298 K. librium. The ion exchange reaction is as follows: 2R–H+ Ni2+ ↔ R2–Ni+ 2H+ (2) where R represents the functional group and R–H de- notes the strong cation resin. According to Eq. (2), the equilibrium constant, K, and the ion exchange ca- pacity, Q, were determined by the following equations to be 1.62 m3-dry resin/m3 and 1.58 kmol-Ni/m3-dry resin, respectively. K = qNi,eC 2 H,e q2H,eCNi,e (3) Q = qNi,e + qH,e2 (4) The isotherms calculated with these values are in good agreement with the observed values in Fig. 2. 3.2. Effective diffusivity of Ni within a resin particle The effective diffusivity, De, of Ni within the resin particle was determined by the fitting the simulated breakthrough curve to the observed one in a fixed bed experiment. In the present study, the simulation was conducted by the method in our previous study [9], which is more simplified and free from the complexity of the numerical integration although the several mod- els have been presented to determine De from fixed bed experiment [13–15]. The outlet concentrations in the liquid and resin phases are determined by solving H. Tokuyama et al. / Separation and Purification Technology 38 (2004) 139–147 143 simultaneously the following mass balance equations for the differential time and height, dt and dh, with the Runge–Kutta method. u ( dCNi dh ) t = −KFa(CNi − C∗Ni) (5) (1− ε) ( dqNi dt ) h = −KFa(CNi − C∗Ni) (6) where qNi is the average concentration of Ni within the resin particle. CNi and C∗Ni are the concentration of Ni in the liquid phase and that in equilibrium with qNi, respectively. The value of C∗Ni is determined from Eqs. (3) and (4). a is the surface area of resin particles per unit volume and is estimated from the void fraction of the bed, ε, (0.26 for spherical particles in stationary bed) and a particle diameter, dp, through the relation of a = 6(1−ε)/dp. The overall mass transfer coefficient, KF, is expressed as follows. 1 KF = 1 kF + 1 mkS (7) where m is the distribution ratio of Ni at liquid–resin film interface. An approximation of the value of m was made on equilibrium relation, m ≈ qNi/C∗Ni. Since the value of KF is determined for each differential volume and time, m is locally taken as constant value within the volume and time. The mass transfer coefficient, kF, in the liquid phase is determined from the correlation developed by Wilson and Geankoplis [16].{ kF (u/ε) } Sc2/3 = 1.09Re−2/3 (8) The mass transfer coefficient, kS, in the resin phase is estimated from the mass balance with the Crank’s equation [17]. kS = − dp 6t ln { 6 π2 ∞∑ n=1 1 n2 exp ( −4Den 2π2t d2p )} (9) The observed breakthrough curves and simulated ones for the various initial concentrations of HCl are shown in Fig. 3, where Cout/Cin is the concentration ratio of the effluent to the feed. The values of De rep- resented in the figure were obtained by the best fit, in which the shape of the curve calculated was sensitive to the variation of De. The value of De increases with the increase in the concentration of HCl. This can be 0 2000 4000 6000 8000 0.0 0.2 0.4 0.6 0.8 1.0 C H,0 D e X 1012 [kmol/m3] [m2/s] 1.0 10 0.5 5.0 0.1 1.4 0.05 1.1 0.01 0.7 C ou t/C in [- ] time [sec] Fig. 3. Breakthrough curves at various initial concentrations of HCl at 298 K. Solid curves are calculated with De for each CH,0. explained by considering that surface diffusion resis- tance is dominant compared with pore diffusion resis- tance for the transfer of Ni within the resin. To prove that theory, the diffusivity of Ni, DF, in the pore-liquid was estimated by the Vinograd–McBain equation [18]. As the results, DF is about 2 orders of magnitude larger than De obtained and oppositely decreases with increasing the concentration of HCl. These facts sup- port the theory of which surface diffusion resistance is dominant. Surface diffusion is known to be highly dependent on the amount adsorbed as has been discussed for physical adsorption system [19,20], and the surface diffusion coefficient increases with the increase in the amount adsorbed. In the present study, De was ob- tained for the mean amount adsorbed from zero to the value in equilibrium with the feed solution. In our ex- perimental system, larger value of De was obtained at the higher concentration of HCl despite the smaller adsorption amount (see Fig. 2). The surface diffusion by the ion exchange, especially between divalent and monovalent ions, can be explained as follows. In the process of surface diffusion, Ni diffuses to the center of intraparticle by repeating the adsorption/desorption at local adsorption site. Divalent Ni ion bonded on the double points might be hard to be exchanged compared with monovalent ion bonded on the single point. When there are the large amounts of free hydrogen ion in the pore-liquid close to Ni adsorbed, the ion exchange of Ni can be promoted, resulting in the larger De. 144 H. Tokuyama et al. / Separation and Purification Technology 38 (2004) 139–147 0 20 40 60 80 100 120 0 1 2 3 4 5 6 7 8 Q L /A h [cm/min] Q R /A h [ cm /m in ] D h [mm] h [mm] 3.11 30 3.28 30 3.57 30 4.06 30 3.57 51 3.57 97 Fig. 4. Relation between resin and liquid velocities through a hole in plate. Conditions of the pulse; the frequency and the amplitude are 1.2 s−1 and 0.225 cm3. 3.3. Control of resin flow rate in a moving bed The relation between the liquid and resin flow rates in the moving bed is shown in Fig. 4, where the volumetric flow rates of QL and QR are divided by the cross-sectional area, Ah, of each hole. These variables correspond to the superficial velocities of the liquid and resin through the hole. The relation of these velocities is shown in a dot curve and is un- changeable for various operating conditions of hole diameters and bed heights. Changing QL and Ah (i.e. Dh) successfully control QR in our apparatus. For ex- ample, small QR is obtained by increasing QL and/or decreasing Dh. This is because the superficial veloc- ities of QR/Ah is almost equal to the relative velocity between the terminal settling velocity of resins and QL/Ah. The instantaneous QR with/without liquid-pulse during countercurrent operation was measured to re- veal the effect of the liquid-pulse on QR. The results are shown in Fig. 5 as time courses of QR. The ap- plying liquid-pulse (1.2 s−1, 0.225 cm3) keeps QR constant without the clogging of resins at the hole. On the other hand, QR without the pulse is varied largely with time and is smaller than that with the pulse. One example of the results obtained under no pulse condition is drawn in Fig. 5. The instantaneous QR’s without the pulse were different from one experiment to another and the clogging of resins occurred in 0 10 20 30 40 50 60 70 0.0 0.1 0.2 0.3 0.4 0.5 with pulse (1.2 s-1, 0.225 cm3) without pulse Q R [ cm 3 - dr y re si n/ m in ] time [min] Fig. 5. Instantaneous QR with/without liquid-pulse against time during countercurrent operation. Conditions; QL = 5.05 cm3/min, Dh = 3.57 mm. some fractions. The increase in the frequency and/or the amplitude of the pulse allowed QR to be large and the clogging of resins to be avoided. However, too strong liquid-pulse made QR small oppositely. These conditions of the pulse were suitable to obtain steady and large QR. The time to the clogging of resins after stopping the pulse was investigated in another aspect to reveal the function of the liquid-pulse. The average values and the standard deviations are plotted against QL in Fig. 6. It must be noted that there is no clogging while the liquid-pulse are given. Although the standard de- viations are far from the average values due to the 8 9 10 11 10 100 1000 ti m e to c lo gg in g of r es in s [s ec ] Q L [cm3/min] Fig. 6. Time to clogging of resins against various QL’s after the pulse was stopped. The average values (circle key) and the standard deviations are represented. Condition: Dh = 4.06 mm. H. Tokuyama et al. / Separation and Purification Technology 38 (2004) 139–147 145 4 6 8 10 12 14 10-3 10-2 10-1 Q L [cm 3 /min] 0.0 0.2 0.4 0.6 0.8 1.0 D h [mm] 4.06 3.57 C ou t/C in [ -] q o ut /q m ax [ -] Fig. 7. Cout/Cin (closed key) and qout/qmax (open key) against liquid flow rate for various diameters of hole in moving bed operation. Conditions: h = 30 mm, Cin,H = 0.01 kmol/m3. uncertainty of clogging, the time tends to decrease with the increase in QL. This means that the clogging can easily occur at large QL, in which QR become contrarily small (see Fig. 4). In such a crucial condi- tion, giving the liquid-pulse is quite advantageous to obtain the stable QR. 3.4. Recovery of Ni with an ion exchange moving bed The column performance for the continuous re- moval of Ni was evaluated in terms of Cout/Cin and qout/qmax, which represent the removal fraction of Ni and the ion-loading fraction of resins. The concen- tration of Ni in the resin phase in equilibrium with the feed solution, qmax, is determined by Eqs. (3) and (4) with the feed concentrations of Ni and HCl. The outlet concentration of Ni, qout, in the resin phase is determined from the mass balance as follows. qout = ( QL QR ) (Cin − Cout) (10) The fractions of Cout/Cin and qout/qmax are plotted against QL for variables of Dh, h and CH,0 in Figs. 7–9. These figures show the moving bed high performance; more than 90% of Ni was removed in every system in- vestigated and qout/qmax over 0.8 were simultaneously obtained in some cases. Large qout/qmax is preferable to save energy in the regeneration of resins. Both frac- 4 5 6 7 8 9 10 10-4 10-3 10-2 10-1 Q L [cm 3 /min] C ou t/C in [ -] q o ut /q m ax [ -] 0.0 0.2 0.4 0.6 0.8 1.0 h [mm] 30 51 97 Fig. 8. Cout/Cin (closed key) and qout/qmax (open key) against liquid flow rate for various bed heights in moving bed operation. Conditions: Dh = 3.57 mm, Cin,H = 0.01 kmol/m3. tions increase with the increase in QL as shown in Figs. 7–9. This is attributable to the increase in the amount of the solution to be treated by a unit amount of the resin by the synergy effect of QL and QR (see Fig. 4). The effect of Dh on the column performance is shown in Fig. 7. The fraction Cout/Cin was unaffected by Dh. On the other hand, enlarging Dh for a given QL resulted in the undesirable decrease of qout/qmax 4 5 6 7 8 9 10 10-4 10-3 10-2 Q L [cm 3 /min] 0.0 0.2 0.4 0.6 0.8 1.0 CH, 0 [kmol/m 3] 0.1 0.05 0.01 C ou t/C in [ -] q o ut /q m ax [ -] Fig. 9. Cout/Cin (closed key) and qout/qmax (open key) against liquid flow rate for various initial concentrations of HCl in con- tinuous operation with moving bed. Conditions: Dh = 3.57 mm, h = 51 mm. 146 H. Tokuyama et al. / Separation and Purification Technology 38 (2004) 139–147 because of the decrease in the residence time of resins with the increase in QR. However, most important point is that the large throughput is attained at large Dh, in which both of QR and QL become large (see Fig. 4). The fraction qout/qmax at Dh of 4.06 mm is larger than that at Dh of 3.57 mm in the operation at each maximum QL. As long as the removal fraction is still kept at high value, better separation is attained by the operation with large Dh. The effect of the bed height on the performance is shown in Fig. 8. The remarkable decrease of Cout/Cin is observed with increasing h. In particular, Cout with 97 mm bed is 1000 times less than the feed concen- tration. This decrease of Cout/Cin can be explained by considering the equation to determine the bed height. The bed height in mass transfer operation with countercurrent flow is given by the product of height of a transfer unit, HTU and number of a transfer unit, NTU. h = ( QL/A KFa )∫ Cin Cout dC C − C∗ (11) where HTU is represented as (QL/A)/(KFa) and NTU is the integral part. Eq. (11) shows that NTU increases and Cout decreases with the increase in h and/or KF. The values of KF calculated on the basis of experi- mental data were, for example, 1.82 × 10−5 m/s for h = 30 mm and 1.57 × 10−5 m/s for h = 51 mm at QL = 7.2 cm3/min. Although the KF value somewhat decreased with the increase in h, Cout largely depends on h. The change of KF for h might be due to the residence time of resins through the column. On the other hand, there is apparently no dependence of h on qout/qmax in Fig. 8 due to the quite high removal fraction. As a result, the operation with the large bed height achieves very high removal fraction without any decrease of qout/qmax. The removal of Ni in the various initial concen- trations of HCl is illustrated in Fig. 9. The fraction Cout/Cin decreases with the decrease in CH,0. It must be noted that the values of qmax are different in each feed concentration of HCl; qmax calculated by Eqs. (3) and (4) are 1.53, 1.33 and 1.13 kmol/m3-dry resin for 0.01, 0.05 and 0.1 kmol/m3 of CH,0, respectively. Thus, qout, corresponding to the net amounts ad- sorbed, for CH,0 = 0.1 kmol/m3 was relatively small at smaller QL (about 4.8 cm3/min) and was almost same at larger QL (about 9.1 cm3/min) compared with that for CH,0 = 0.01 kmol/m3. Cout decreases with the decrease in CH,0. This behavior can be explained on the basis of Eq. (11) by considering equilibrium and kinetics as follows. HTU has one variable of KF under the constant QL, where KF is expressed as Eq. (7). Since kF determined from Eq. (8) is constant under the constant QL, KF is affected by only kS, assuming no change of m with CH,0. As a result, the decrease in CH,0 gives the increase of HTU through the decreases of De (see Fig. 3), kS (see Eq. (9)) and KF. From the relation between HTU and NTU, NTU must decrease with the decrease in CH,0, which means the increase of Cout for an isotherm. However, Cout decreased with the decrease in CH,0 in Fig. 9. By considering that the ion exchange equilibrium is affected by the concentration of H+ (see Fig. 2), C∗ is smaller at lower concentration of H+ accord- ing to the relation between isotherm and operating line and NTU must be small despite the small value of Cout. 4. Conclusions A prototype of moving bed has been constructed and the feasibility has been examined for Ni removal from model solutions of wastewater. The resin flow rate is successfully controlled by the liquid flow rate and the cross-sectional area of a hole in the plate in- stalled at the bottom of resin bed. Applying the ap- propriate liquid-pulse prevents resins from clogging at the hole and to stabilize the resin flow rate. The mov- ing bed with SK1B resin shows the high performance; Cout/Cin of more than 0.9 and qout/qmax of more than 0.8 were simultaneously obtained. Better separation is attained by the operation with the large diameter of a hole and the large bed height; the former gives high ion-loading fraction of resins with the large through- put and the latter high removal fraction without any reduction of the ion-loading fraction. The removal of Ni in the various initial concentrations of HCl has been discussed with the ion exchange equilibrium and ki- netics. The values of equilibrium coefficient and ion exchange capacity were determined, taking account of the exchanging ratio, 1/2, of Ni to the functional group of the resin. The effective diffusivity of Ni within the resin particle increases with the increase in the con- centration of HCl. H. Tokuyama et al. / Separation and Purification Technology 38 (2004) 139–147 147 References [1] P.B. Spoor, L. Grabovska, L. Koene, L.J.J. Janssen, W.R. ter Veen, Chem. Eng. J. 89 (2002) 193. [2] K. Dorfner, Ion Exchangers, Walter de Gruyter, New York, 1991, pp. 677–683. [3] C. Simpson, S.H. Laurie, Hydrometallurgy 51 (1999) 335. [4] T.H. Karppinen, A. Yli-Pentti, Sep. Sci. Technol. 35 (2000) 1619. [5] R. Koivula, J. Lehto, L. Pajo, T. Gale, H. Leinonen, Hydrometallurgy 56 (2000) 93. [6] M.J. Slater, Br. Chem. Eng. 14 (1969) 41. [7] N.K. Hiester, R.C. Phillips, Chem. Eng. 61 (1954) 161. [8] B. Biscans, J.P. Riba, J.P. Couderc, Int. Chem. Eng. 28 (1988) 248. [9] H. Tokuyama, S. Nii, F. Kawaizumi, K. Takahashi, Ind. Eng. Chem. Res. 41 (2002) 3447. [10] H. Tokuyama, S. Nii, F. Kawaizumi, K. Takahashi, J. Chem. Eng. Jpn. 36 (2003) 486. [11] H. Tokuyama, M. Natsume, S. Nii, F. Kawaizumi, K. Takahashi, J. Chem. Eng. Jpn. 36 (2003) 98. [12] H. Tokuyama, S. Nii, F. Kawaizumi, K. Takahashi, Sep. Sci. Technol. 38 (2003) 1329. [13] K. Mimura, K. Hashimoto, J. Chem. Eng. Jpn. 10 (1977) 490. [14] A. Hasnat, V.A. Juvekar, AIChE J. 42 (1996) 161. [15] J.-H. Koh, P.C. Wankat, N.-H.L. Wang, Ind. Eng. Chem. Res. 37 (1998) 228. [16] E.J. Wilson, C.J. Geankoplis, Ind. Eng. Chem. Fund. 5 (1966) 9. [17] J. Crank, The Mathematics of Diffusion, Oxford, UK, 1975, pp. 89–103. [18] J.R. Vinograd, J.W. McBain, J. Am. Chem. Soc. 63 (1941) 2008. [19] M. Suzuki, T. Fujii, AIChE J. 28 (1982) 380. [20] E.R. Gilliland, R.F. Baddour, G.P. Perkinson, K.J. Sladek, Ind. Eng. Chem. Fund. 13 (1974) 95. Development of a novel moving bed with liquid-pulse and experimental analysis of nickel removal from acidic solution Introduction Experimental Measurement of ion exchange equilibrium Kinetics measurement Moving bed operation Results and discussion Ion exchange equilibrium Effective diffusivity of Ni within a resin particle Control of resin flow rate in a moving bed Recovery of Ni with an ion exchange moving bed Conclusions References