Environ. Sci. Technol. 2005, 39, 885-888
Adsorption onto Fluidized Powdered Activated Carbon Flocs-PACF A N A L IÄ D I A S E R P A , I V O A N D R EÄ H . S C H N E I D E R , A N D JORGE RUBIO* UFRGS, PPGEM, Avenue Osvaldo Aranha 99/512, CEP: 90035-190, Porto Alegre RS, Brazil
This work presents a new adsorption technique where the adsorbent (powdered activated carbon-PAC) is in the form of suspended flocs formed with water-soluble polymer flocculants. Thus, the adsorption of a typical dye, methylene blue (MB), was studied onto polyacrylamide flocs of PAC (PACF) in a fluidized bed reactor. The technique is based on the fact that the adsorption capacity of PAC does not decrease after flocculation because the adsorbed polymer occupies only a few surface sites, in the form of trains, loops, and tails. Moreover, the adsorption was found to proceed through a rapid mass transfer of MB to the adsorbing PAC flocs, in the same extent as onto PAC. Because of the rapid settling characteristics of the aggregates formed, the two phase separations, loaded PAC and solution, become easier. Thus, the technique offers the advantages of conducting simultaneously both adsorption and solid/liquid separation all in one single stage. Results obtained showed that high MB removal values can be attained in a fluidized bed reactor (>90%) and that PACF presents a much higher adsorption capacity (breakthrough points) than granulated activated carbon (GAC) in the same adsorbing bed. It is believed that this technique highly broadens the potential of the use of powdered activated carbon or other similar ultrafine adsorbents.
Introduction Liquid-phase adsorption has been shown to be an effective method for color removal from aqueous streams, and activated carbon is the most widely adsorbent used for this process (1). A number of studies concerning the use of activated carbons in the removal of dyestuffs has been reported with granular activated carbon (GAC) (2-7), powdered activated carbon (PAC) (8-12), and the combined use of powdered activated carbon with the activated-sludge process (PACT) (13). Most of the activated carbon systems use granular activated carbon (GAC) in fixed or expanded bed reactors (1, 4, 6). However, dyes and other high molecular weight pollutants do not penetrate in the inner pores of the particles, leading to a rapid adsorbent saturation (2, 3, 12, 14). Powdered activated carbon (PAC) presents a larger surface area available and adsorption capacity, but operational problems such as pore clogging and pressure drop makes its use in packed bed systems almost impossible (5). The use of PAC in fluidized reactors is also unlikely since the particles settling rate is very low (well stable dispersions are formed) and escape readily from the adsorption column. This problem limits the * Corresponding author phone:
[email protected]. 10.1021/es035276w CCC: $30.25 Published on Web 12/10/2004
+55-51-33163540; e-mail:
2005 American Chemical Society
application of most finely divided adsorbents to batch or continuous stirring tank reactors where the cost of the solidliquid separation of the loaded PAC from water and its regeneration are the main drawbacks (1). An alternative to these problems is to flocculate the powder activated carbon to increase the particles size and density. For instance, Snoeying et al. suggested the aggregation with alum (coagulation), clay (aggregation or heteroagregation), and if necessary, with polyelectrolyte (flocculation). These authors used a floc blanket reactor to conduct pilot-scale studies removing 2,4,6-trichlorophenol (10) and atrazine (11) by adsorption onto a flocculated mixed pulp of alum, clay, and activated coal. This work presents an alternative basis to solve the problem of using powdered adsorbents at higher flow-rate (high loading rates), by employing well-structured, strong, big, and fluidized polymeric flocs. It is believed that by fluidizing the flocs, the loading rates achievable are higher than in static (small expansion) blankets. Thus, it was discovered that after flocculation (big flocs) with commercial polymer flocculants, the adsorption capacity of PAC did not decrease, and because of the rapid settling characteristics of the flocs, the solid/liquid separation is fairly easy. This allowed the use of expanded or fluidized reactors for adsorption/separation of dyes, such as methylene blue (used as a model), onto PAC flocs (PACF). Accordingly, the aim of this work is to show comparative results in the adsorption of methylene blue on GAC, PAC, and PACF in fluidized bed reactors.
Experimental Procedures Methylene Blue Solutions. Commercial grade methylene blue (MB) was supplied by Importadora Quı´mica Delaware Ltda (Porto Alegre, Brazil). The chemical structure is shown in Figure 1, and the dye solutions were prepared using deionized water for the isotherm and kinetics studies. But, because of the high water volumes required and practical implications (closer to real cases), the municipal tap water was employed for the continuous semi-pilot studies in the fluidized reactor. Activated Carbon. GAC, Indu ´strias Quı´micas Carbomafra (Curitiba, Brazil), supplied the granular activated carbon. According to the manufacturers, the particle size was 0.295-0.833 mm, bulk density was 0.45-0.55 g cm-3, and iodine number was 900 mg of I2 g-1. PAC corresponds to ground GAC material in a laboratory mill and screened to 100% less than 147 µm. Surface area measurements by BET nitrogen adsorption method (16) provided a surface area of 565 m2 g-1 for the GAC and 599 m2 g-1 for the PAC. Flocculants. Polymeric flocculants were all supplied by Nalco do Brasil S. A. The reagent studied were Nalco 8589 (high molecular weight cationic polyacrylamide), Nalcolyte 8105 (cationic coagulant), Nalco 9915 (high molecular weight anionic polyacrylamide), and Nalcolyte 4032 (nonionic polyacrylamide). The polymers were prepared using deionized water in a stock solution concentration of 1 g/L. All polymer solutions were used within a 1 week period after preparation. Flocculation and Adsorption Studies. Flocculation studies were carried out using 1.0 L glass Beckers. Activated carbon-water suspensions in a concentration of 2.0 g L-1 were stirred during 1 min (with the polymer flocculant) at 70 rpm to promote a uniform distribution of the flocculant. The stirring speed was reduced to 20 rpm for 1 min to create low shear conditions allowing floc formation. The suspensions were transferred to a 35 cm column, and the flocculation efficiency was evaluated by the settling rate, visual and VOL. 39, NO. 3, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
885
TABLE 1. Evaluation of Activated Carbon Flocculation Using Commercial Polymers polymer
FIGURE 1. Structure formula of methylene blue. Main characteristics: molecular weight, 319.85 g mol-1; solubility, 40 g L-1 (15).
none Nalcolyte 8105 Nalco 8589 Nalco 9915 Nalcolyte 4032
best concn (mg g-1)
0.5 2.0 1.0
settling rate (cm s-1)
flocculation performance
1.5 1.5 1.1
nulla null excellentb excellent goodc
a Null: no flocs were formed. b Excellent: flocs size larger than 3 mm. c Good: flocs size of about 1-3 mm.
FIGURE 2. Semi-pilot scale fluidized bed used for methylene blue adsorption: (1) methylene blue tanks, (2) centrifugal pump (0.5 hp), (3) valve, (4) flow meter, (5) distribution grid, (6) fluidized bed (i.d. ) 0.04 m, h ) 1.06 m), (7) treated effluent collection zone, (8) sampling valve, and (9) manometer. microscope observations of flocs quality. The best concentration for each flocculant was defined as the minimum dosage necessary to get homogeneous, big, and resistant (to shearing) flocs. Adsorption isotherms were measured at 25 °C by varying the initial dye concentration from 0.06 to 0.78 mmol L-1 and keeping the concentration of activated carbon constant in 2 g L-1 (GAC, PAC, or PACF). The MB solutions (0.1 L) were stirred at 20 rpm for 24 h in a dark environment to reach equilibrium. Isotherms were analyzed by the Langmuir adsorption model, which describes the adsorption reaction as (17)
m)
MbC (1 + bC)
or by its linearized form
C 1 C ) + m bM m where m is the mass of solute adsorbed per unit weight of adsorbent, M is the mass of solute adsorbed per unit weight of adsorbent at saturation, b is the constant accounting for the energy of interaction, and C is the measured concentration in solution at equilibrium. Adsorption rate was determined at 25 °C varying the reaction time within 1 and 300 min using an initial methylene blue concentration of 0.156 mmol L-1 and activated carbon concentration of 0.43 g L-1. The MB solutions (1 L) were stirred at 20 rpm for the three forms of activated carbon studied: GAC, PAC, and PACF. Fluidized Bed Semi-Pilot Adsorption Studies. In the semi-pilot studies, the fluidized bed reactor shown in Figure 2 was used in experiments with GAC and PACF. The dye solutions were prepared using both tanks (Figure 2) through pump circulation. The PAC flocs (100 g) were prepared separately in a 1 L pulp volume with the help of a mechanical stirrer employing high speed stirring (70 rpm during 1 min), addition of the flocculant, and mild agitation for the floc 886
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 3, 2005
formation (20 rpm during 1 min). Then, the flocs were transferred to the column, and the MB solution was fed into the column in a flow rate of 0.35 L min-1. Evaluation of the PACF and GAC losses in the column was carried out by the mass weight (dry basis) differences between the activated carbon added (feed) and that after a complete adsorption cycle. This follows the fact that the mass of the adsorbed dye and flocculant are negligible as compared to the mass of the activated carbon in the reactor. Methylene Blue Analysis. The concentration of MB in solution was determined spectrophotometrically at 669 nm, using visible-ultraviolet spectrophotometer CG 8000, both before and after the introduction of activated carbon in the system. The amount of MB abstracted from the solution per unit weight of mass was determined by the difference between the two analyses. Surface Area Measurements. Because adsorption occurs in aqueous solution, specific surface area was determined by the dye adsorption method (18). The surface area was calculated from the saturation adsorption assuming a crosssectional area of 108 Å2 for the methylene blue molecule. Photographs. The PACF photographs were taken with a digital camera Sony model DSC, at the low stirring speed flocculation stage, by the top view of the glass beakers.
Results and Discussion Flocculation tests of the activated carbon were conducted using different commercial synthetic flocculants. Table 1 presents the results obtained including the best dosages required. Excellent results were attained with both cationic and anionic flocculants, but Nalco 8589 was selected since flocculation was achieved at lower polymer dosages. The larger flocs, which presented the higher settling rate, were attained with a polymer concentration of 0.5 mg g-1. The flocs showed a filamentous shape with typical length between 3 and 8 mm and were very resistant to the hydraulic conditions of the system (Figure 3). Figure 4 presents adsorption isotherms of MB by GAC, PAC, and PACF. The linear correlation coefficients and the Langmuir model constants obtained are summarized in Table 2. According to these constants, best results were found with PAC, either flocculated or not. The maximum capacity attained was 0.413 mmol g-1 with PAC and 0.398 mmol g-1 with PACF, while the adsorption capacity of the GAC was significantly lower (0.195 mmol g-1). Figure 5 presents the effect of reaction time on MB adsorption onto the three forms of activated carbon (PACF were formed with 0.5 mg g-1 of Nalco 8589). Results show the effect of particle size and surface area available on the adsorption of dye molecules by activated carbon. Thus, the PACF had a similar behavior as compared to PAC, both having higher adsorption rates and accumulation capacities when compared with GAC. Figure 6 illustrates typical breakthrough curves obtained in tests using the fluidized bed reactor to remove MB from solution. The two experiments were conducted with the same
FIGURE 5. Adsorption of methylene blue as a function of time by GAC, PAC, and PACF at 25 °C. Initial methylene blue concentration of 0.156 mmol L-1 and activated carbon concentration of 0.43 g L-1. FIGURE 3. Powdered activated carbon flocs obtained with 0.5 mg g-1 of Nalco 8589, a high molecular weight cationic polyacryamide.
FIGURE 4. Adsorption isotherms of methylene blue by GAC, PAC, and PACF at 25 °C.
TABLE 2. Langmuir Adsorption Model Constants Obtained from Isotherms adsorbent powdered activated carbon flocs, PACF powdered activated carbon, PAC granulated activated carbon, GAC
FIGURE 6. Methylene blue adsorption breakthrough curves using GAC and PACF in fluidized beds. Conditions: flow rate, 0.35 ( 0.01 L min-1; methylene blue initial concentration, 0.156 mmol L-1; mass of activated carbon, 100 g; and fluidized bed dimensions (i.d. ) 0.04 m, h ) 1.06 m).
TABLE 3. Surface Areas of PACF, PAC, and GAC Available for MB Adsorption
M (mmol g-1)
b (L mmol-1)
R
0.398
278.0
0.997
adsorbent
specific surface area (m2 g-1)
0.413
404.0
0.997
0.195
47.6
0.978
powdered activated carbon flocs, PACF powdered activated carbon, PAC granulated activated carbon, GAC
258 268 127
mass of activated carbon (100 g) and flow rate (0.35 L min-1) with less than 1% of activated carbon mass loss in both experiments. Results show that the PACF in the fluidized bed reactor removed efficiently the dye from the water solution (>90% adsorption). The high surface area of PACF and the floc configuration in the bed promoted the necessary conditions for the adsorption of MB molecules onto the solid phase. Using the same experimental conditions, incomplete removal of MB was observed with GAC, showing lower adsorption capacity and earlier breakpoints. The fluidized reactor functions in a semicontinuous mode, until the saturation of the powdered activated carbon flocs, PACF, were attained (in about 3 h operations) with the flocs remaining unaltered. After saturation, the column has to be reloaded with PACF and the effluent bypassed to another column. Or, a system of solids addition and removal like in the floc blanket reactor may be employed (10-11). Yet, the
operating conditions used in this work allowed an hydraulic loading capacity of 19-24 m3 m2 h-1, higher than the values for the floc blanket reactor (1.3 m3 m2 h-1) reported by Campos et al. (11) and other solid liquid settling equipments (1). This is explained by the big size of the flocs, which are possible to generate. Results obtained with PACF can be explained by the fact that the adsorbed flocculant occupy only a few sites at surfaces of the solid because of the well-known adsorption configuration in the form of tails, trains, and loops (19). Moreover, the loss of surface area of PACF (measured by the MB method and shown in Table 3) due to the flocculant adsorption and the flocs formation (particle aggregation) was negligible, revealing the importance of the internal area. Results also demonstrated the higher adsorption efficiency (process kinetics and MB surface accumulation) of PAC (flocculated or not) over GAC revealing the effect of particle size and interfacial area. The importance of adsorbent particle VOL. 39, NO. 3, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
887
size for adsorption of dyes (i.e., Basic Blue 26, Disperse Blue 7) and other organic compounds is very well-reported in the literature (3-5). However, this effect is not so pronounced when considering organic molecules of low molecular weight (i.e., phenol) (20, 21). It is believed that adsorption onto polyacrylamide flocs of PAC suspended in expanded or fluidized reactors technique can be successfully applied for the removal of the former dye compounds from wastewaters. Finally, the adsorbing polymeric flocs technique, developed in this work, showed results that appear to broaden the potential use of powdered activated carbon (PAC) or other similar powdered adsorbents. The adsorption of MB onto powdered activated carbon flocs, PACF, was efficient in a fluidized reactor (>90% adsorption) showing advantages (kinetics and adsorption capacity) when compared to granulated carbon (GAC). The PACF process offers the same rapid mass transfer of the adsorbate to the PAC with the advantage of allowing solid-liquid separation, all in one single stage. This technique avoids the operational problems encountered in packed bed systems, like pore clogging and pressure drop. This process appears to be an excellent alternative in the treatment of effluents bearing (among others) a variety of dye and surfactants, namely, wastewaters from the textile industry, laundry houses, and stone inking plants.
(3) (4) (5) (6)
(7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19)
Acknowledgments The authors thank all institutions supporting research in Brazil and to the colleagues and students for their friendship.
Literature Cited (1) Metcalf & Eddy, Inc. Wastewater Engineering: Treatment and Reuse; McGraw-Hill: Boston, 2003. (2) Al-Degs, Y.; Khraisheh, M. A. M.; Allen, S. J. Water Res. 2000, 34(3), 927-935.
888
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 3, 2005
(20) (21)
McKay, G. J. Chem. Technol. Biotechnol. 1982, 32, 759-772. McKay, G. J. Chem. Technol. Biotechnol. 1982, 32, 773-780. McKay, G. Chem. Eng. Res. Des. 1983, 61, 29-36. O’Brien, R. P.; Rizzo, J. L.; Schuliger W. G. In Control of Organic Substances in Water and Wastewater Pollution Technology Review; Berger, B. B., Ed.; Noyes Data Corporation: Park Ridge, NJ, 1987; pp 337-362. Shi, J.; Zhao, X.; Hickey, R. F. Water Environ. Res. 1995, 67(3), 302-309. Yeh, R. Y. L.; Thomas, A. J. Chem. Technol. Biotechnol. 1995, 63, 55-59. Juang, R.-S.; Swei, S.-L. Sep. Sci. Technol. 1996, 31(15), 21432158. Najm, I. N.; Snoeyink, V. L.; Richard, Y. Water Res. 1993, 27 (4), 551-560. Campos, C.; Snoeyink, V. L.; Marin ˜ as, B.; Baudin, I.; Laine´, J. M. Water Res. 2000, 34 (16), 4070-4080. Li, Q.; Snoeyink, V. L.; Marin ˜ as, B.; Campos, C. Water Res. 2003, 4863-4872. Costa, C.; Marquez, M. C. Water Res. 1998, 32 (1), 107-114. Pelekani, C.; Snoeyink, V. L. Carbon 2000, 38, 1423-1436 Merck. The Merck Index, 12 ed.; Merck and Co., Inc.: Rahway, NJ, 1996. Allen, T. Sizing Techniques, Surface Area, and Porosimetry; Chapman and Hall: New York, 1990. Adamson, A. W. Physical Chemistry of Surfaces; Wiley: New York, 1997. van den Hul, H. J.; Lyklema, J. J. Am. Chem. Soc. 1968, 9, 30103015. Dobias, B. Coagulation and Flocculation: Theory and Applications; Marcel Dekker Inc.: New York, 1993. Abdo, M. S. E.; Nosier, S. A.; El-Tawil, Y. A. 1997 J. Environ. Sci. Technol. 1997, A32 (4), 1159-1169. Wang, R. C.; Kuo, C. C.; Shyu, C. C. J. Chem. Technol. Biotechnol. 1997, 68, 187-194.
Received for review November 16, 2003. Revised manuscript received October 25, 2004. Accepted October 26, 2004. ES035276W
