DEFLUORIDATION STUDIES WITH FISHBONE CHARCOAL IN DIFFERENT ADSORPTION SYSTEMS

Abstract

Fluoride in drinking water can cause beneficial or 1e..riinental effects depending on its concentration and the total amount ingested. The beneficial effect of fluoride is its 3apability to prevent dental caries (upto an optimum amount of 1.0 (Jg'ii). An excessive fluoride causes dental or skeletal fluorosis. Fhe Indian standards for drinking water supply recommend an acceptable fluoride concentration of 1.0 mg/L and an allowable Quor.de concentration of 1.5 mg/L in potable waters. Ground *alara in India contain fluoride in varying concentrations. In the B&jor.ity of cases, it ranges from 1.5 to 6.3 rag/L. In some si ,jftoions, values as high as 16 to 18 mg/L (in one case even 36 i"jg/L) have been reported. In certain parts of about thirteen states of India, fluorosis is endemic, as a result of which large sections of rural, population are subjected to permanent bone ii..orders such as bent knees, stiffened joints etc. Hence, the ie:f fluoridation of water is necessiated. >e luoridation methods The different methods tried for defluorLdation are of three basic types via. adsorption methods, ion exchange methods and ?r«oipitation methods. Extensive research has been done for lefluoridation of water by many researchers including Boruff, la w, Culp and Stoltenberg, Savinelli and Black, Harmon and Caiichman, Nawlakhe et al., Wu and Nitya, Choi and Chen, Rubel and "foosley, Bulusu et al. and Hao and Huang. A comprehensive review III IV on the subject has been presented elsewhere (Killedar and Bhargava, 1988a, 1988b). The various materials mainly used for defluoridation by the above three methods include activated carbon, animal bone and bone charcoal, tricalcium phosphate, synthetic ion exchangers, activated alumina, lime and alum. Commercial activated carbon showed a good fluoride removal capacity but since the process was found pH dependent (optimal pH value = 3.0), the finished water would need a pH adjustment which prohibits its field applicability. The synthetic ion exchangers exhibited poor fluoride removal efficiency due to its lower affinity for fluoride. The animal bone charcoal and activated alumina have been used widely in field scale but it requires efficient control and expensive regeneration. Lime precipitation method has a limited application in the sense that it is suitable only for waters having high magnesium content. The requirement of high doses of alum with lime restricts its use for treatment of water for suitabale situations only. Thus, the different existing methods have certain limitations and their own merits and demerits. The selection of an appropriate method would depend on many factors including material cost, capital cost, operation cost, durability and efficiency of the system and the maintenance. The defluoridating materials being used at present are relatively costlier. The investigations of feasible low cost materials for fluoride removal are needed. Preliminary investigations were carried out to determine the feasibility of some pretreated low cost/waste materials viz. blast furnace slag, natural dolomite, tammrind seed shell carbon and fishbone charcoal for fluoride removal from water (Killedar and Bhargava, 1991b) The batch adsoprtion studies with these materials Lead to conclude that: (i) Blast furnace slag removed about 50% of fluoride within 15 min of contact but afterwards due to desorption the fluoride removal reduced to 24% in 180 min of contact, (ii) Dolomite removed about 18% of fluoride from fluoride solution prepared in distilled water adjusted to pH 10.0. In the fluoride solution prepared in tap wter (pH = 10.0), the removal was observed to be 50%. This shows that the dolomite can be recommended for the treatments of waters containing natural magnesium, (iii) Tammrind seed shell carbon showed a lower efficiency (about 15%) for fluoride removal, and (iv) Fishbone charcoal showed a good potential for fluoride removal. It instantaneously adsorbed fluoride and about 50% removal was obtained w.thin 180 min of contact. These preliminary investigations provided an encouragement to use the fishbone charcoal for further detailed investigations. Fishbone la comparatively cheap material available abundantly in coastal areas or nearby areas of perrenial rivers. Earlier, the bone charcoal prepared out of animal bones has been successfully used in many full scale installations for defluoridation of the drinking water in Southern California. Smith and Smith have suggested the removal mechanism of fluoride by bone is an ion-e> change phenomenon. The use of fishbone charcoal <and the results of its feasibility studies in batch operation, column operation or moving VI media reactor operation have not been widely reported. The objective of present studies, was to carry out detailed investigastions for fishbone charcoal as a defluoridating medium and to evaluate its performance in the three adsorption systems viz., the batch, the column and the moving media reactor operations. BATCH ADSORPTION STUDIES In a batch adsorption system some significant factors that affect the adsorption process, include the contact time (t), the adsorbent size, the initial solute concentration (C ), the adsorbent dose (Ws); pH, stirring rate and temperature. The effects of these variables on fluoide removal by fishbone charcoal have been investigted in a batch process. Fishbone charcoal was prepared by carbonising the cleaned and pulverised fishbone in an electric furnace in a closed retort at 1000°C for 2.0 hours. The cooled material was sieved to obtain six adsorbent sizes (size I to 3ize VI), having geometrical mean diameters (d) of 1:0.714 mm (600 pm - 850 /jib size range), 11:0.461 mra(355 (Jm-600 Mm size range), 111:0.253 mm (180 <um - 355 turn size range), IV: 0.150 mm (125 fum - 180 <-im size range), V: 0.106 mm (90 t-im - 125 L>m size range) and VI: 0.549 mm (355 ^m - 850 /Jm size range). The specific surface area of this material (for size VI) was determined to be 85 m2/g. The density was determined to be 1.8 / 3 g/cm . The test solutions of different initial fluoride concentrations (Cq = 5 mg/L, 10 mg/L, 20 mg/L and 50 mg/L) used in some batch studies were prepared by adding the required amount of Vll sodium fluoride in the distilled water (D.W.) as per the standard methods. These test solutions of above mentioned concentrations were used an batch experiments aimed at studying the effects of contact time, particle size, initial solute concentration and pH. The batch adsorption runs aimed at studying the effects of pH, initial solute concentration and adsorbent dose, under simulated field conditions, the test fluoride solutions having different fluoride concentations of 3.0 mg/L, 6.5 mg/L, 10.0 mg/L, 20.0 mg/L and 30.0 mg/L were prepared by adding the appropriate quantities of sodium fluoride in the tap water (T.W.). The tap water used had pH of 7.0 - 8.1; total dissolved solids of 170 mg/L; total hardness (as CaCo3) of 112 mg/L; calcium (as Ca++) of 44 mg/L; chlorideMas CI ) of 10 mg/L; sulphates (as SO") of 1.5 mg/L; alkalinity (as CaCo3) of 75 mg/L and fluorides (as F~)of less than 0.1 mg/L. The interference of competing ions and presence of calcium in tap water is determined to be negligible on the basis of trial test run conducted to evaluate the fluoride removal though adsorption on the fishbone charcoal. No precipitates of CaF2were virtually observed in the test solutions when prepared in the tap water. The Initial feasibility studies conducted to evaluate the effects of particle size (Size III and Size VI) and adsorbent dose (Ws = 2,4,8,12 and 16 g/L) and contact time (0-180 min) revealed that the fishbone charcoal instantaneously adsorbed fluoride and the adsorption increased with time till equilibrium was attained in about 180 min. The plot of amount of fluoride adsorbed per unit weight of adsorbent (rng/g) versus time (min1/2) V11X did not pass though origin indicating that pore diffusion is not the only rate limiting step in the removal of fluoride. The overall percentage removal was more for finer adsorbent for any adsorbent dose (Bharagava and Killedar, 1989a). The further elaborated studies to determine effect of adsorbent size (size I to size V) on fluoride removal for a contact time of 0-480 min (C0= 10.0 mg/L , D.W.base, W^ 4 g/LJ have showed that for any 3i2e Qf adsorbent) ^ ^ of fluoride uptake was high for a contact time ranging from 0 to 180 min while the rate had significantly decreased beyond 180 min till the equilibrium stage is attained in about 480 min. To explore the dependency of the adsorption rate on adsorbent size, the plots were prepared between the residual fluoride concentration versus contact time for different adsorbent sizes. The rate constants (*,) and the particle size were found to follow the relationship of the type shown in Eq. 1 (Bhargava and Killedar, 1993 ). k _ 10[0.058(l/d)-3.90J s (1) In Eq. 1. ks represents the rate constant, in mg/L.mm-' and di, the geometrical mean diameter of the given adsorbent size, i„ mm. Eq. 1can be used to compute the rate constant for a given Particle size of adsorbent under the observed test conditions. Studies were also conducted to determine the effects of contact time and initial fluoride concentration on fluoride removal bv fishbone charcoal. The Plots between the percentage fluoride removal and contact time for the various initial fiuoride IX concentrations (CQ = 5.0,20.0, 50.0 and 100 mg/L, D.W.base) were also prepared. The fluoride removal increased with time for any C o value. FOJ a given size and dose of adsorbent the percentage fluoride removal at contact time of 180 min decreassed with increase in initial fluoride conentration (C = 5, 20, 50 and 100 mg/L of D.W.base). However, the amount of fluoride adsorbed per unit weight of adsorbent (mg/g) increased with initial fluoride concentration. (Bhargava and Killedar, 1993 ). Comprehensive batch experiments were conducted to determine the effects of some parameters such as contact time (0-540 min), initial solute concentration (CQ = 3.0, 6.5, 10.0. 20.0 and 30.0 mg/L) and dose of adsorbent (Wg = 2,4,8,16 and 32 g/L) on fluoride removal by fishbone charcoal under simulated field conditions. The Plots of percentage fluoride removal versus contact time for various Ws and Co values were obtained. The fluoride removal increased with time for any Ws and Cq value. The fluoride removal increased with Wg at any time for agiven Rvalue. Using these Plots,an integrated empirical relationship has been developed to correlate various parameters such as fluoride removal (R), contact -tin..: (t), initial fluoride concentration CO and dose of adsorbent (y under simulated field conditions as shown in Eq. 2 (Bhargava and Killedar, 1991c). "W~ z {~~ : )t {C /[( _—!ls + 0'33 W3 ° 5.44 x 10-3+ 6.758 x 10-3W s W + / S V 0.0167 + 0.014 W }Co3}t (2) In Eq.2, R is flouride removal,in % , t is the contact time, in min, Cq is the initial fluoride concentration, in mg/L and W is s the dose of adsorbent, in g/L. The empirical relationship shown in Eq. 2 can be exploited to predict the percentage fluoride removal at any time t for the chosen values of Wg under the observed test conditions (size VI; PH = 8.0 and T.W. base). Such empirical equation can facilitate the operation of batch reactors for obtaining the required removal efficiency by controlling the variables such as contact time and dose of adsorbent. Batch adsorption studies were conducted through jar tests, to study the effects of CQ (3.0, 6.5, 10.0, 20.0, 30.0 mg/L) and Wg (2,4,8,16 and 32 g/L) on the designated equilibrium concentration corresponding to the selected and designated equilibrium time of 240 min. The data for these tests were used to prepare plots between the equlibrium concentration (C ) and W for all C values. The equilibrium concentration was found to decrease with increasing Ws values at a. given Cq value. The plots of Ce versus Ws for various Cq values followed by the relationship (Bhargava and Killedar, 1991c) of type as shown in Eq. 3. 1 _ 1 ce - "IT *+ ee,l "Ws (3) In Eq. 3, ex represents a coefficient. The values of coefficient e, for different Cq values were obtained from the regression of data of plots between Ce versus ^. The plot of e versus Cq appeared to follow the relationship indicated in Eq. 4! XI (eQ+eQlog C ) ej - 10 (4) In Eq. 4, e2 and eg are coefficients and their values were determined as 0.2316 and 1.8775 respectively. Substitution of Eq. 4 in Eq. 3 yielded an empirical equation shown in Eq.5. ii. -ji tri0(*.213e-1.e77io«conMs (6) e o In Eq.5, Ce and CQ are respectively the equilibrium and initial fluoride concentrations in mg/L and W is the dose of adsorbent in g/L. Equation 5 can be used to predict the equilibrium fluoride concentration (C ) for any values of C (in e o the range of 3.0 to 30.0 mg/L) and Ws (in the range of 2.0 to 32.0 g/L for the observed test conditions (PH = 8.0, T.W. base and for adsorbent Bise VI). The observed equilibrium concentration, C^ and the fluoride adsorption capacity, ge (expressed as the amount of fluoride adsorbed per unit weight of adsorbent) poorly correlated in the Langmuir or Freundlich isotherms, and a linear relationship (Bhargava and Killedar, 1991c) as shown in Eq. 6 has been evolved. q_ = 0.2564 + 0.0997 C ,c, e e . (o; In Eq 6, qe repesents the fluoride adsorption capacity of fishbone charcoal, in mg/g and Ce represents the equilibrium fluoride concentration, in mg/L. Eq. 6 is applicable for the same observed test conditions mentioned for Eq.5. The solution PH generally shows a significant impact on the xii solute removal efficiency during the adsorption process. Maier reported the fluoride removal by bone charcoal to be pH dependent such that at lower pH values of raw water, the removal efficiency increased significantly. To explore this aspect, for fluoride removal by fishbone charcoal, batch adsorption studies were conducted at different pH values of 2,4,6,8 and 10 for an initial fluoride concentration of 10 mg/L and 11 mg/L of the test solutions prepared in distilled water (D.W.) and tap water (T.W.) respectively (for Wg = 8 g/L and t - 0-480 min). The prepared variation curves of the fraction of fluoride remaining in solution, C (the ratio of C , the residual concentration at any time t, to C , the initial fluoride concentration) at different times and pH values (for the D.W. and T.W. base solutions) showed that the C value decreases with increase in contact time at any pH value and increases with pH at any time. The plots of C versus t for various pH values are seen to follow a first order reaction represented by Eq. 7. In (C ) = - kpR t (7) In Eq. 7, kpR represents the rate constant. The rate constants vary from 235 x 10~3 hr_1 to 32 x 10~3 hr_1 for a PH range of 2 to 10 (D.W. base) respectively and from 211 x 10~3hr_1 to 31 x 10 ' hr for a pH range of 2 to 10(T.W. base ) respectively. i Since the rate constants kpR are varying with the pH, the values of kpH were correlated with pH as shown in Eq. 8. Xlll kpH " e3 + e4 ( 1/pH) (8) In Eq. e3 and e4 are coefficients.The values of e„ and e.were o 4 determined as 0.015 and 0.462 (for distilled water base), and 0.006 and 0.44 (for tap water base). These values of coefficients are representative of the observed test conditions. The predictive relationships were obtained (from Eqs. 7 and 8) and are presented * in Eqs. 9 and 10 (Bhargava and Killedar, 199lc). C = exp { -[0.015 + 0.462 (1/pH)] t } (9) C = exp { -[0.006 + 0.440 (1/pH)] t } (10) Eqs. 9 and 10 can be used to predict the C value at any time t in the PH rnge of 2 to 10 for distilled water base and tap water base solutions respectively under the observed test conditions. The batch experiments aimed to investigate the effects of stirring rate and temperature revealed that the fluoride removal at any time increased with stirring rate (in the range of 20 to 100 rpm ;for Cq =10.0 rag/L (T.W.base) and Ws = 8 g/L). The increase in stirring rates showed a decrease in boundary layer thickness as interpreted from the prepared plots of q^the amount of fluoride adsorbed per unit weight of adsorbent) in mg/g versus the time^from the data of this study(as also interpreted by Mckay et al. in their studies). The fluoride removal at a given temperature is found to increase with time. The initial rates of adsorption increased with increase in temperature(in the range of i0°C to 60°C. For C = 10.0 mg/L (T.W. base) and Ws = 8.0 g/L) whereas ^ ^.^ removal capacity at equilibrium decreased with increase in temperature This may be happening because the rise in temperature XIV increases the escaping tendency of molecules from an interface, and therefore diminishes the extent of adsorption, resulting in a decrease in equilibrium capacity as interpreted also by McKay et al . CKilledar and Bhargava, 1993aD. The nature of the rate controlling step can be obtained by determining the activation energy of the process. This can be achieved by studying the effect of temperature on rates of adsorption that are expected to increase with an increase in temperature. A plot of pseudo rate constants versus temperature yields a straight line slope of which gives the activation energy of adsorption. The activation energy can be computed from Van* tHoff - Arhenius equation (20) The magnitute of activation energy is useful in asserting the mechanism of adsorption, a diffusional controlled adsorption reaction will have a small activation energy, i.e. < 5-10 Kcal/mole. A surface reaction controlled adsorption reaction will have a greater activation energy, i.e. > 10 Kcal/mole. The plots of amount of fluoride adsorbed per unit weight of adsorbent versus time0-5 at different temperatures showed that the linear portions of these plots if extended back do not pass through origin which indicate that the pore diffusion is not only rate controlling step during entire time of adsorption, while film diffusion appears to be rate limiting during initial time of adsorption. The activation energy was found to be 1.12 Kcal/mole which is within the range of values for diffusion controlled processes i.e. <5-10 Kcal/mole. Weber CI 9725 and Humenick CI9773 have described an approach for determination of the overall mass transfer coefficient CKaZ), Film diffusion coefficient CK ) and pore surface diffusion coefficient CK^ using isotherm and column data for any solute sorbent system. The similar approach has been attempted to compute the values of Ka, L & K in case of f s fluoride fishbone charcoal adsorption system. It has been found that the pore surface diffusion coefficient is negligible while the overall mass transfer coefficient and film diffusion coefficient worked out to be 2.86 g/min. cm3 _3 and 1.ai x lO cm/sec respectively. A comparative study was carried out to access the fluoride removal efficiencies of fishbone and animalbone charcoals. The animalbone charcoal was prepared by the same methodology as described earlier for fishbone charcoal. The fluoride removal was found to be a function of time and dose of adsorbent CWg = 2.4.8. 16.32 g/L3 at a given initial fluoride concentration CCq = 10.0 mg/L, T.W. base). The fishbone charcoal showed a comparatively higher fluoride removal than animalbone charcoal at any time and XV XVI dose of adsorbent. This indicates that the use of fishbone charcoal shall be economical over animal bone charcoal for fluoride removal due to its higher efficiency and lesser cost of production. (Bhargava and Killedar, 1991a). COLUMN STUDIES Fixed bed adsorption studies were conducted to determine the effects of some of the variables such as the flow rate (Q), initial fluoride concentration (CQ) and bed depth (D) on the effluent concentration (Ce) at different times. The breakthrough curves of effluent concentration versus time at different flow rates (Q = 15,25,50 and 75 ml/min) were analysed. The volume of treated effluent (V) at different flow rates at the desired breakthrough concentration of C0 - l.fl mg/L of F~ were found to correlate by the relationship of the type shown in Eq. 11. ff a2 + b2 ( 1/Qw> (11) In Eq. 11, v represents the volume of treated effluent volume, in L at a flow rate of Qw, in ml/min per gram of the adsorbent which yields an effluent concentration of fluoride, Ce of 1.0 mg/L. a and bj are the coefficients. The values of coefficients a{ and hj were determined as 1.748 and 2.80 respectively, through a regression analysis of the data obtained from breakthrough curves mentioned earlier. The relationship evolved (Eq.ll) can be used to determine the total weight of the adsorbent required to treat a known influent volume at a given flow rate at an initial fluoride concentration of 5.0 XV11 rag/L. The similar analysis of breakthrough curves of the column experiments conducted to study the effect of initial fluoride concentration (C = 2.5, 6.0, 10.0 and 20.0 mg/L) on effluent o fluoride concentration revealed that the treated effluent volume (corresponding to breakpoint of C = 1.0 mg/L) decreassed with increase in C value and followed a relationship of the type shown in Eq. 12. V = a2 *- b2 (1/CQ) (12) In Eq. 12, V represents the volume of treated effluent in L at an initial solute concentration, C corresponding to C = 1.0 mg/L, a„ and b^ are coefficients. The values of the coefficients a„ and b„ were determined as 0.897 and 44.58 respectively. The evolved relationship (Eq.12) can be exploited to predict the volume of treated effluent at any initial solute concentration (in the range of 2.5 mg/L - 20.0 mg/L at a flow rate of 15 ml/min and at an effluent fluoride concentration of 1.0 mg/L). The values of treated effluent volume (corresponding to Ce = 1.0 mg/L) for the different bed depths are found to be 5.6L, 7.65L and 11.92L for 15 cm, 30 cm and 50 cm depths respectively. The value of volume of treated effluent per unit depth i3 determined to be 0.228 L/cm on the average from the slope of the plot between the treated effluent volume versus bed depth. Further, detailed column runs were conducted to determine the integrated effects of flow rate (Q = 15,30,50 and 75 ml/min), XV111 initial fluoride concentration (C = 2.5, 5.0, 10.0 and 20.0 mg/L) and bed depths (D = 20, 42 and 65 cm) on fluoride removal by fishbone charcoal. The adsorbent size (355 ym - 850 Mm size rang) and pH (8.0, T.W. base) were kept constant. The values of the treated effluent volume (V) (corresponding to the breakthrough concentration of 1.0 mg/L) for various flow rates, initial fluoride concentations and bed depths were calculated from the column runs data. Plots were prepared between the treated effluent volume and flow rate for the various C o values and column bed depths. Using these plots, empirical relationship (Eq.13) correlating the volume of treated effluent with the variables viz. flow rate, initial fluoride concentration and bed depth was developed (Killedar & Bhargava 1993 b) log V={(( 0.640+0162 Cq )+ 00112 D] _|" (-1.836+0.535 -CQ)-(4.013x10 4+5.647xl0~4 C)D] 1 (13) In Eq. 13, V is the volume of treated effluent at breakthrough concentration of 1.0 mg/L of F_ in L, C is the o initial fluoride concentration in mg/L, D is the column bed depth in cm and Q is the flow rate in ml/min/crn2. The polynomial based fitting concept has been used to develop an integrated expression for the prediction of the treated effluent volume (V) in terms of the variables such as flow rate (Q), column bed depth (D), and initial fluoride concentration (CQ). The integrated expression developed is shown in Eq.14. XIX V = [bQ + bj log Q + b2 D + b3 log Co] 3 (14) In Eq.14, b0, bj, b2 and b3 are the function coefficients whose values were determined as 3.8277, -0.8869, 0.0263 and -1.949 respectively with Q, D and Cq expressed in ml/rain/cm2, cm and mg/L respectively. The relationships obtained in Eqs.13 and 14 can be exploited to estimate the amount of volume of treated effluent for the desired CQ, Q and D values under the observed test conditions. MOVING MEDIA REACTOR STUDIES Adsorption from aqueous solution involves the concentration of sorbate on the sorbent surface. The efficiency of an adsorption reactor depends upon the mode of contact between the sorbent and sorbate.According to Weber,the fixed bed continuously operating column reactors are found to be more efficient than the batch reactors as it provides better utilisation of sorbent capacity. Hassler and Weber had suggested the use of continuously operating expanded bed reactor to increase the sorbent effectiveness. In a moving media reactor, the solid sorbent is added in the reactor top and the spent sorbent is withdrawn from the bottom. The solution from which adsorption takes place flows upwards through the regions of partially utilized sorbent to regions of freshly added sorbent material. Bhargava and Bhatt in their studies developed a continuously operated moving media reactor and observed that this operation facilitates a fuller utilization of the sorbent capacity. Experiments were carried out to investigate the effects of XX some variables such as sorbent mass input rate «„), sorbate flow rate (Q , and initial solute oonoentration i%) on fluoride removal by fishbone oharooal in a moving media reactor „.*«, During experimentation, the sorbent .i*. ,366 ,„„ -eM „„, ,lM ra"8el 0"d "" (8'0; •''•"• •«•> <« U.i Glutton, were kBpl constant while the sorbent mass Input rate (W =0 5 ,a ., sm .*'• "».*•»• ,i.o, 2.0 and 3.0 g/min,, the aorbate flow ^ (Q .& ^ ^ ^ ^ and 60 ml/min, and lnltlal £luor.de conoentrat.on (c _z g 0 «... ».. and 30.0 M/Il) „ere varied Th/ ^^ concentrations attained in effluent were determined for the various test conditions. Using these data, plots of c <the ratio of the attained e,uiliblr». effluent concentration, "c to the initial sorbate concentration, „„> and % (the fluoride removal capacity of the sorbent, versus the variables W , Q and C0 were prepared. The 5. and %values varied iaverael" ith W -d varied directly with 9 and „„. Emplrical relatlonslu^ . <Bhargava and KiHedar, 1992a,have been developed ^ ^ Plot, to predict Cm and %values with respect to the sorbent mass input rate, the sorbate flow rate and the initial sorbate concentration respectively for tK« „k the observed test conditions (Bhargava and Killedar,1992b). The data of these investigations poorly confirmed the conventional adsorption isotherms. An isotherm for moving media adsoprtion system has been evolved. „hich follous the r.1Mlonabip (Bhargava and Killedar, 1992b, of type shown i„ Eq.15. qem = H + b6 Cem ^•^ 9en represents the fluoride removal Per'unit"3urfaoe XXI area of sorbent mass input per unit time in /Jmol/min/m2 ,C represents the equilibirum effluent concentration in moles/min and £„, bfi are coefficients. The values of coefficients a„ and b„ b 6 were determined as 3.528 x 10~8 and 4.864 x 10~5 respectively. The Cm value is found to be a function of sorbent-sorbate input rates ratio (Wsm / Cq Q ). A model is obtained to predict the Cm value for given sorbent sorbate input rate ratio for the observed test conditions as shown in Eq. 16 (Bhargava and Killedar, 199.<3b) . -jr- l« +m-c% (is) m In Eq. 16, m is a constant. The value of m is to be determined using the experimental data. For present studies, the value of m was calculated to be 7.005 x 10~5. The model presented in Eq.16 can be exploited to determine the Cm value attainable for a given ^3m/CQQ value under the observed test conditions. A methodology is presented to detemine the optimal value of the sorbent mass input rate which can be used in actual operation of reactors to strike a balance between the fluoide removal and the fluoride removal capacity of the sorbent, under the experimental test conditions (Bhargava and Killedar,1991b ) Comparative Performance in Batch,Fixed Bed and Moving Media Reactor Operations 10 assess the fluoride removal efficiency of the fishbone charcoal in the three adsorption systems viz. batch,fixed bed and moving bed, the fluoride removal capacity (^) „as determined for the desired equilibrium(and residual) concentration of 1.0 mg/L. em XX11 The comparison was done at a flow rate of 15 ml/rnin for fixed bed and moving bed systems and at initial flouride concentrations(C ) o of 2.5,5.0 and 10.0 mg/L. It is observed that the lowest flouride removal capacity was obtained for all CQ values in case of the moving media reactor while the fixed bed reactor has shown the highest fluoride removal capacity for Cq = 5.0 and 10.0 mg/L. The batch reactor has shown the highest capacity for C =2.5 mg/L. This indicates that the fixed bed system shall be most suitable and also convenient way for defluoridation of drinking water.The use of fixed bed can also be recommended in the form of home filters at domestic level due to its highest efficiency and convenience.The lowest fluoride removal capacity of fishbone charcoal in moving media reactor system indicates that this system is not economical and appropriate for drinking water defluoridation. However,this system may be reinvestigated for treatment of industrial fluoride wastes of very high initial concentrations. With the estimated cost of fishbone charcoal as Rs*4.0/kg, the comparative cost of per gm fluoride removal by batch,fixed bed and moving media systems works out to respectively Rs.10.0,4.0 and 160.0 on an average basis.The fixed bed column system therefore is more economical than the other systems. Rs.100.00 = US $ 6.00 XX111 SUMMARY The feasibility and the fluoride removal efficiency of fishbone charcoal was investigated in batch operation, fixed bed column operation and moving media reactor operation. The effects of various variables studied include size and dose of adsorbent, initial solute concentration, contact time, pH, stirring rate and temperature for batch studies; flow rate, initial solute concentration and bed depth for fixed bed column studies and sorbent mass input rate, sorbate flow rate and initial sorbate concentration for moving media reactor studies.Various aspects of the solute removal and related effects of the various environmental factors have been discussed in context of the removal efficiencies. Predictive empirical relationships have been developed in respect of various system variables for the observed test conditions which are useful to assess the fluoride removal efficiency of fishbone charcoal for the adsorption systems studied viz. the batch system, the fixed bed column system and moving media reactor system. Defluoridation of drinking water with fishbone charcoal is feasible for isolated communities of small sizes as well as at domestic level in the form of home filters. The higher fluoride removal effiiency and low unit cost of production of fishbone charcoal over animalbone charcoal claims it an economically feasible defluoridating material for small installations.