Decolorization and Sedimentation of a Textile Dye by Dissolution of Metallic Aluminum: A Statistical Investigation

Decolorization and Sedimentation of a Textile Dye by Dissolution of Metallic Aluminum: A Statistical Investigation
of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  ENVIRONMENTAL ENGINEERING SCIENCEVolume 20, Number 6, 2003©Mary Ann Liebert, Inc. Decolorization and Sedimentation of a Textile Dye byDissolution of Metallic Aluminum: A Statistical Investigation Ahmet Gürses, 1,* Mehmet Yalcin, 1 Çetin Dogar, 2 and Kemal Doymus 1 1 K.K. Egitim Fakültesi Department of Chemistry Atatürk University25240 Erzurum, Turkey 2  Erzincan Egitim Fakültesi Atatürk University25240 Erzincan, Turkey ABSTRACT In this study, the decolorization and sedimentation of strong colored solutions containing a reactive tex-tile dye,  Remazol Red RB , were investigated. Aluminum hydroxides from dissolution of metallic aluminumwere used as flocculant. Experiments were performed according to “2 4 full factorial experimental design.”By using MINITAB software, analysis of variance was carried out. The parameters investigated were alu-minum dissolution temperature, initial pH, aluminum dissolution time, and effect of surfactant. Aluminumdissolution temperature and dissolution pH were found to be statistically significant parameters in the de-colorization process. From the variance analysis of settling rate, the amount of dissolved aluminum andthe mean distribution of floc size, it was found that pH was a significant parameter common in all of theseprocesses. These results suggest that decolorization by flocculation with direct dissolution of metallic alu-minum may be practically used for the removal of reactive dyes from aqueous solutions. Key words: decolorization; textile dye; sedimentation; metallic aluminum 667 * Corresponding author: Atatürk University, K.K. Egitim Fakültesi, Department of Chemistry, 25240 Erzurum, Turkey. Phone :90 442-23-4004; Fax : 90 442-236-0955;  E-mail : INTRODUCTION T HETEXTILEINDUSTRY consumes considerableamounts of water in the manufacturing process. Thewater employed in the dyeing and finishing processeseventually ends up as wastewater requiring treatment be-fore final discharge. Frequent changes of dyestuff em-ployed in the dyeing process cause considerable varia-tion in the wastewater characteristics, particularly pH,color, and COD (Lin, 1993). Strong color is a significantcomponent of the textile wastewater. The traditionalwastewater treatment methods consist of biological,physical, and chemical treatments, and also their variouscombinations (Peters et al. , 1985; Rebhun and Lurie,1993; Do and Chen, 1994; Lin and Peng, 1994, 1996;Pols and Harmsen, 1994; Szpyrkowicz et al., 1995; Jiangand Graham, 1996; Morais et al. , 1999). Although thecost of biological treatment is less than chemical and  physical treatments, biological treatment’s poor perfor-mance is due to the toxicity of the dissolved or suspendedmatters.The coloring components of wastewater can be effec-tively removed by physical and chemical methods in-cluding chemical and electrochemical coagulation, ad-sorption, ultrafiltration, and ozonation (Thebault et al. ,1981; Van Benschoten and Edzwald, 1990; Churchley,1994; Gahr et al. , 1994). Flocculation Flocculation plays an important role in wastewatertreatment. Its objective is to bring colloidal and particu-late matter into an aggregated form so that they may beseparated from the process stream with further treatment(Odegaard, 1995). In addition, flocculation also enablesdissolved organic substances with high molecular weightand suspended particles to become entrapped in growingfloc. Hence, these substances can be removed during theflocculation and settling processes. In this respect, a par-ticularly important aspect of flocculation is to adapt thecharacteristics of the flocs to the subsequent separationstep (direct filtration, sedimentation, flotation) (Bern-hardt and Schell, 1993; Bhole, 1993).Although there are some examples of works in the lit-erature which describes wastewater treatment by coagu-lation-flocculation by using Al 3 1 and Fe 3 1 salts as co-agulant-flocculant (Thebault et al. , 1981; Peters et al .,1985; Jiang and Graham, 1996), a review of this litera-ture shows no description of the efficiency of flocculationand decolorization using direct dissolution of metallicaluminum. Temperature and pH tolerance of conventionalbiological and chemical methods is very limited (Zoltekand Melear, 1981; Peters et al ., 1985). Under normal con-ditions, because textile wastewaters are characterized bya high factory discharge temperature (60–70°C) and pH(9–11), decolorization using aluminum dissolution pro-cess, as conducted in this study, can be carried out in theconventional textile wastewater treatment process with-out pH and temperature adjustment. The flocs were pro-duced by dissolving metallic aluminum in high temper-ature and pH.This study aimed to achieve the following goals: first,it was aimed to investigate the effects of dissolution tem-perature, dissolution pH, dissolution time and surfactant( Cethyltrimethylammonium Bromide ) on both settling ve-locity of flocs and decolorization of simulated waste-water. Second, it was aimed to statistically investigate theeffect of the amount of dissolved aluminum and the con-sequent mean distribution of floc sizes on the decol-orization process. In addition, it was also aimed to de-velop a mathematical model showing the effect of theamount of dissolved aluminum and the consequent meandistribution of floc size on the percentage of decoloriza-tion. This study provides information about direct use of metallic aluminum, instead of aluminum and iron salts indye removal from textile wastewater with high pH andtemperature characteristics. So, this study focus in the po-tential of use in treatment of textile wastewaters of wastealuminum materials such as aluminum foils. MATERIAL AND METHODS  Experimental Decolorization was conducted using a pilot-scale floc-culation process under various pH and temperature con-ditions. From experiments, it was determined that the pro-duction of the flocs occurs only in a narrow pH rangebetween 9 and 10. This suggests that the textile dye af-fects the formation of aluminum hydroxide. Similar re-sults were reported previously about the effect of dis-solved species on aluminum hydroxide precipitation(Duan and Gregory, 1998). Square aluminum plates, 4 3 4 cm in dimension with 0.10 mm thickness, were usedto produce the aluminum ions. Aluminum plates werecleaned by 15% (wt) HCl solution prior to use within afew seconds. Solutions of 25.0 mgL 2 1 of  Remazol Red  RB textile dye were prepared for use in these experiments.  Remazol Red RB (color index name:  Reactive Red 198  )is a textile dye with a negative charge in aqueous solu-tion because of vinyl sulfone groups (Ganesh, 1994).After centrifuging the supernatant samples obtainedfrom experiments, the residual dye concentrations werespectrophotometrically measured at the appropriatewavelength (  l 5 506 nm) against blank solutions(Bosch-Lamb UV Spectrophotometer). Decolorizationpercentage was calculated by comparing the differencebetween dye concentrations before and after treatments.After the experiments, the consequent amount of sludgeand dissolved aluminum, using different weight of platesafter and before treatments, were determined.The pH of the prepared solutions was adjusted by aque-668GÜRSES  ET AL.  ous NaOH and HCl solutions. The metallic aluminumplate was then contacted with dye solutions for periodsof 5 to 10 min at high pHs (12, 12.5, and 13) and tem-peratures (from 55 to 65°C). Dissolution pH was contin-uously controlled using a pH electrode immersed in thesolution and continuous adjustment. Temperature wascontrolled using a thermostatic water bath. A cationic sur-factant, CTAB, was used to enhance flocculation. Finally,pH values of the mixtures contained in the glass reactorswere lowered to 9.5. Flocs appeared in seconds after pHwas lowered. A glass column (60 cm height, 7 cm di-ameter) immersed in a water bath with thermostat wasused to measure the height of the solid–liquid interface.To determine the sedimentation rate, the swept distanceof the solid–liquid interface was measured at certain timeperiods. Initial rates of solid–liquid interface were usedin statistical analysis. The amount of dissolved aluminumwas calculated as the difference between weights of alu-minum plates before and after the dissolution reaction.The floc sizes were microscopically determined for everyexperiment. In addition, the floc sizes were verified us-ing sieves suitable to ASTM standards. These resultsshowed that significant floc breakage did not occur in thesieving procedure. In some experiments, because an alu-minum hydroxide network was formed, floc size mea-surements could not be carried out. Therefore, the meanfloc size was taken as zero. In addition, despite waitinglong enough, sedimentation did not occur in the experi-ments. Experiments in this study were replicated twice.In plots, mean values of the measurements were used. Statistical analysis and modeling In this study, general statistical model was used. Thegeneral statistical model used in this study follows: Y  5 m 1  b  X  1 1  b 1  X  2 1  b 2  X  3 1  b 3  X  1 *  X  2 1  b 4  X  1 *  X  3 1  b 5  X  2 *  X  3 (1)where Y  is the dependent variable and m is overall meanof all treatments.  b s,  X  s, and  X  3  X  s represent the coef-ficients of independent variables, independent variables,and interactions on the Y  parameter, respectively.It is often good practice to code the quantitative inde-pendent variables, prior to introduction into a statisticalmodel. For example, supposes a variable T  , takes on threelevels: 50, 100, and 150. We can code (or transform) thevariable measurements using Equation (2)  x 5 (2)Then, the coded levels  x 52 1, 0, and 1 correspond tosrcinal levels 50, 100, and 150. There are two relatedreasons for coding variables. First, considerable compu-tational error may occur during the inversion process if the numbers in matrix vary greatly in absolute value. Sec-ond, the coding variable, which pertains to the problemof multicolinearity, is beyond this discussion (Menden-hall and Sincich, 1995). To allow selection of the statis-tically significant variables, main variables were modi-fied to their variables coded in the following manner(Hines and Montgomery, 1990):  X  1 ,  X  2 ,  X  3 , and  X  4 are srcinal variables; where theseare dissolution time, dissolution temperature, initial pH,and cationic surfactant concentration, respectively. Ac-cordingly,  x 1 ,  x 2 ,  x 3 , and  x 4 are the coded variables, where  x 1 5 (  X  1 2 7.5)/2.5;  x 2 5 (  X  2 2 60)/5;  x 3 5 (  X  3 2 12.5)/0.5;  x 4 5 (  X  4 2 200)/100The levels of the parameters are given in Table 1. Inter-actions among independent variables were accounted inthe following way: suppose the  y -dependent variable andindependent variables  x 1 and  x 2 are said to interact if thechange in  y for a one-unit change in  x 1 (when  x 2 is heldfixed) is dependent on the value of  x 2 (Mendenhall andSincich, 1995). Higher order interactions generally havesmall effects relative to low order interactions, becausea system with several parameters is controlled primarilyby the independent variables themselves and low-orderinteractions among these variables. Therefore, presentwork considers only two-way interactions (e.g.,  x 1 3  x 2 ).Higher order interactions are relegated within the modelto residual error (Hines and Montgomery, 1990). Exper-iments in the center points (experiments with zero valueof code in Table 2) could be used to construct the check( T  2 100) }} 50DECOLORIZATION AND SEDIMENTATION OF A TEXTILE DYE669ENVIRON ENG SCI, VOL. 20, NO. 6, 2003 Table 1. Original and coded variables. Coded variable ( x ) Main variableOriginal variable ( x )  2 101 Dissolution time, min  X  1 005007.5010 Dissolution temperature, °C  X  2 0550600.065 Initial pH  X  3 012012.5013 Surfactant conc., mg/L  X  4 1002000.300  670GÜRSES  ET AL. Table 2. 2 4 Full factorial experiment design with one replicate and four center points.  Deco. %Al, gSed. rate a cms 2 1 Floc size, m mSludge (kg/m 3 )RunTime, x 1 Time, x 2  pH, x 3 Surf, x 4 0.00.0090.000.000.101  2 1  2 1  2 1  2 11000.0180.9691.60.202  2 1  2 1  2 1  2 194.50.0100.000.000.103  2 1  2 1  2 1  2 182.50.0140.000.000.204  2 1  2 1  2 1  2 11000.1641.31108.51.905  2 1  2 1  2 1  2 11000.4020.61155.44.606  2 1  2 1  2 1  2 11000.2711.14163.33.107  2 1  2 1  2 1  2 11000.2610.58161.83.008  2 1  2 1  2 1  2 10.00.0050.000.000.069  2 1  2 1  2 1  2 10.00.0040.000.000.0510  2 1  2 1  2 1  2 175.30.0230.000.000.2711  2 1  2 1  2 1  2 11000.2140.000.002.5012  2 1  2 1  2 1  2 180.10.1250.000.001.4013  2 1  2 1  2 1  2 11000.2031.170.002.3014  2 1  2 1  2 1  2 11000.2350.61162.92.7015  2 1  2 1  2 1  2 11000.2070.65165.62.4016  2 1  2 1  2 1  2 11000.0740.97167.90.8617  2 0  2 0  2 0  2 01000.0751.01171.20.8718  2 0  2 0  2 0  2 01000.0800.96167.00.9219  2 0  2 0  2 0  2 01000.0721.07169.30.8320  2 0  2 0  2 0  2 0 a The linear rate values calculated for first 9 min. Table 3. Analysis of variance for decolorization percentage.  Estimated effects and coefficients for decolorizationTermEffectCoefStd Coef  t -value pConstant81.625.36215.220.000Time  2 16.58  2 8.295.995  2 1.380.200Temperature34.0217.015.9952.840.019pH40.9820.495.9953.420.008Surfactant conc.  2 15.20  2 7.605.995  2 1.270.237Time 3 Temperature13.406.705.9951.120.293Time 3 pH11.605.805.9950.970.359Time 3 Surfactant conc.5.422.715.9950.450.662Temperature 3 pH  2 29.0414.525.995  2 2.420.038Tempt. 3 Surfactant conc.14.787.395.9951.230.249pH 3 Surfactant conc. P , 0.05 accepted as statistically significant.  Analysis of variance for decolorization percentageSourceDFSeq SSAdj SSAdj MS  F a pMain effects413,37013,3703342.65.810.0142-Way interactions66,0396,0391,006.51.750.216Residual error95,1765,176575.1Curvature11,6901,6901689.93.880.084Lack of fit53,4863,486697.2  2 4E 1 061.000Pure error3  2 0  2 0  2 0.0Total1924,585 a F  -values calculated using residual error.  for curvature in the model to be fitted. That is, the checkfor curvature is used to determine whether our model re-quires square or cubic forms of variables (Bennett andFranklin, 1967; Hines and Montgomery, 1990; Strange,1990). The experiments were performed according to “2 4 full factorial design” (Table 2). To determine the effectsof four parameters, analysis of variance was performedby MINITAB software. RESULTS AND DISCUSSION The outputs of ANOVA for the decolorization per-centage, the sedimentation rate, the amount of dissolvedaluminum, and mean floc size are shown in Tables 3–6.From Table 3, it can be seen that at least one parameteris statistically significant, because the  p -value of com-bined effects (0.014) is less than 0.05. In this sense, Table3 shows that dissolution temperature (0.019), initial pH(0.08), and the interaction between dissolution tempera-ture and initial pH (0.038) are statistically significant. Inaddition, Table 3 suggest that linear statistical model canfit to the data since the value of “lack of fit” is biggerthan 0.05. Moreover, it is also evident in Table 4 that pH(0.027) is a statistically significant factor for sedimenta-tion rate. Converse to the decolorization percentage, thelinear model cannot fit to the data since the value of “lackof fit” are smaller than 0.05. Since the linear model couldnot be constructed between sedimentation rate and pH,quadratic and cubic forms of these parameters should beincluded to the model. Tables 5 and 6 indicates that pHis a significant parameter for “the amount of dissolvedaluminum” and “mean floc size”. The values of “lack of fit” for both “the amount of dissolved aluminum” and“mean floc size” suggest that linear model cannot fit tothese data. Similar to the sedimentation rate, quadraticand cubic forms should be included to the model for “theamount of dissolved aluminum” and “mean floc size.”In this study, the goal was to remove the reactive dyefrom aqueous solutions. It was observed that just afterpH adjustment and surfactant addition, flocs were formedrapidly. Due to the high dye concentration and dissolvedaluminum concentration, mainly Al(OH) 4 2 , and also theexistence of cationic surfactant, the adsorption and/or en-trapment of dye molecules by the formed aluminum hy-droxide occurred just in few seconds following the low-ering pH. At the conditions investigated, any change incolor intensity of solution was do not observed after dis-solution of the metallic aluminum prior to the pH ad- justment and precipitation step. This suggests that theDECOLORIZATION AND SEDIMENTATION OF A TEXTILE DYE671ENVIRON ENG SCI, VOL. 20, NO. 6, 2003 Table 4. Analysis of variance for sedimentation rate.  Estimated effects and coefficients for sedimentation rateTermEffectCoefStd Coef  t -value pConstant0.55170.10815.100.000Time0.11430.05720.12090.470.648Temperature  2 0.1327  2 0.06630.1209  2 0.550.597pH0.63730.31870.12092.640.027Surfactant conc.  2 0.2710  2 0.13550.1209  2 1.120.291Time 3 Temperature  2 0.2427  2 0.12130.1209  2 1.000.342Time 3 pH  2 0.1260  2 0.06300.1209  2 0.520.615Time 3 Surfactant conc.0.18900.09450.12090.780.454Temperature 3 pH0.10760.05380.12090.450.667Tempt. 3 Surfactant conc.0.15600.07800.12090.650.535pH 3 Surfactant conc.  2 0.0307  2 0.01530.1209  2 0.130.902  Analysis of variance for sedimentation rateSourceDFSeq SSAdj SSAdj MS  FpMain effects42.041322.041320.510332.180.1522-Way interactions60.589410.589410.098240.420.848Residual error92.104692.104690.23385Curvature11.018491.018491.018497.500.025Lack of fit51.079741.079740.21595100.350.002Pure error30.006460.006460.00215Total194.73542
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks