Boron removal from aqueous solutions by batch adsorption onto cerium oxide using full factorial design

Boron removal from aqueous solutions by batch adsorption onto cerium oxide using full factorial design

Desalination 223 (2008) 106–112 Boron removal from aqueous solutions by batch adsorption onto cerium oxide using full factorial design Nes¸ e Öztürk*...

162KB Sizes 0 Downloads 5 Views

Desalination 223 (2008) 106–112

Boron removal from aqueous solutions by batch adsorption onto cerium oxide using full factorial design Nes¸ e Öztürk*, Duygu Kavak Department of Chemical Engineering, Faculty of Engineering and Architecture, Eskis¸ehir Osmangazi University, 26480 Mes¸elik, Eskis¸ehir, Turkey Tel. +90-222-2393750/3280; Fax +90-222-2393613; email: [email protected] Received 19 December 2006; accepted 3 January 2007

Abstract In this study removal of boron from aqueous solutions by adsorption was investigated and experimental design was applied. Cerium oxide was used as adsorbent. In the first part of study 3 parameters effect performance and two levels of these parameters were investigated. Parameters that were chosen temperature (20 and 40°C), particle type (granule and powdered) and initial pH of boron solution (6.18 and 10). In the second part of the study linear law Langmuir and Freundlich adsorption models were applied to describe the equilibrium isotherms. Maximum boron removal was obtained at original pH value of boron solution and 40°C using powdered cerium oxide. It was verified that temperature, temperature-particle type, pH-particle type-temperature variables had influence on the boron adsorption at the level of 5% probability. The influence order was temperature > temperature-particle type > pH-particle type-temperature. Keywords: Boron removal; Batch adsorption; Cerium oxide; Full factorial design

1. Introduction Turkey has the largest boron reserve which is approximately 90 million tons in the world. It was estimated that Turkey has about 70% of the known reserves of the world [1]. Boron compounds are used in a wide range of industrial applications. Due to use and production of boron *Corresponding author.

compounds are introduced into the environment in the form of waste. A very low boron content is required in irrigation water for certain metabolic activities. However if it is present in amounts larger than required it becomes toxic. Referring to Nable et al., [2] safe concentrations of boron in irrigation water are 0.3 mg/L for sensitive plants, 1–2 mg/L for semi tolerant plants, and 2–4 mg/L for tolerant plants [2].

Presented at the conference on Desalination and the Environment. Sponsored by the European Desalination Society and Center for Research and Technology Hellas (CERTH), Sani Resort, Halkidiki, Greece, April 22–25, 2007. 0011-9164/06/$– See front matter © 2008 Published by Elsevier B.V. doi:10.1016/j.desal.2007.01.176

N. Öztürk, D. Kavak / Desalination 223 (2008) 106–112

There are several physicochemical treatment processes typically used to remove boron from water and wastewater. These are adsorption, ion exchange, chemical precipitation, solvent extraction, and a combination of adsorption and solvent extraction, membrane processes and ultrafiltration [3–15]. Two or more methods must be applied to high boron content wastewater. Adsorption can be used for advanced treatment of boron. A technique known as statistical design of experiments is a powerful technique for process characterisation, optimisation and modelling. It basically involves the process of planning and designing an experiment so that the appropriate data may be collected which then can be analysed and interpreted, resulting in valid and objective conclusions. Experiments in which the effects of more than one factor on response are investigated are known as full factorial experiments. In a full factorial experiment, both of the (−1) and (+1) levels of every factor are compared with each other and the effects of each of the factor levels on the response are investigated according to the levels of other factors. By factorial planning of the experiments it was possible to investigate the effects of all the variables simultaneously [16]. In this study boron removal from aqueous solutions by adsorption method was investigated using cerium oxide. Two level factorial design was applied to investigate the effects of the parameters and their interactions on boron removal by batch adsorption. The regression equation was obtained. The linear law, Langmuir and Freundlich isotherm models were tested for their applicability. 2. Experimental 2.1. Materials and methods Cerium oxide was used as granule and powdered (200–275 mesh) for boron removal from aqueous solutions by adsorption. Adsorbent was dried at 105°C for 2 h before being used. The aqueous solution of H3BO3 was prepared by

107

using the analytical grade Merck product. The solution was prepared in such a manner that the initial boron concentration in adsorption experiments was held at 100 mg/L (pH = 6.18). pH was measured with pH meter (Consort P903). All the tests were done in capped volumetric flask (100 mL) by adding adsorbent (1 g) and 50 mL H3BO3 solution. Temperature was kept constant in a water bath with shaker (memmert). Shaking speed was 150 rpm. Adsorbent and solution mixtures were shaken 24 h. After adsorption samples were centrifuged and boron in supernatants was analyzed. All of the tests were duplicated. Boron analyzed using HACH DR 2000 Spectrophotometer by carmine method. For adsorption isotherm study 1 g of adsorbent was contacted with 50 mL of H3BO3 solution at concentrations of 5, 10, 25, 50, 100 and 600 mgB/L at 20°C for 24 h with continuous shaking. 2.2. Statistical design of experiments The use of statistical design of experiments is advantageous as it allows one to obtain conditions through a relatively smaller number of systematic experiments. Using a proper design matrix one can obtain a regression equation, which highlights the effect of individual parameters and their relative importance in given operation/ process. The interactional effects of two or more variables can also be known, which is not possible in a classical experiment [17]. The principal steps of statistically designed experiments are: determination of response variables, factors and factor levels, choice of the experimental design and statistical analysis of the data. Today, the most widely used experimental design to estimate main effects, as well as interaction effects, is the 2n factorial design, where each variable is investigated at two levels. Research can be designed for multiple factors and treatments, but data analysis and treatment establishment becomes more complex and time

108

N. Öztürk, D. Kavak / Desalination 223 (2008) 106–112

Table 1 Actual and vis a vis coded values of parameters in 23 full factorial design for boron removal by adsorption Level of variables

First level Second level

Temperature

Initial pH of solution

Particle type

Actual (x1)

Coded (X1)

Actual (x2)

Coded (X2)

Actual (x3)

Coded (X3)

20 40

− +

Original pH (6.18) 10

− +

Granule Powdered

− +

consuming as the number of factors and treatments increase [16]. So, 23 factorial design was selected in this study. The number of experiments (N) required for understanding all the effects is given by ak = 23 = 8 where a is the number of levels and k is the number of factors. The regression equation developed from different sets of experiments show the dependence of yield on individual parameters as well as interactions for simultaneous variations of parameters.

done two times. Variables and levels for the study are given in Table 1. The higher level variable was designated as (+1), and lower level as (−1). In Table 1, X1, X2 and X3 represent the levels of temperature, initial pH of solution and particle type, respectively and x1, x2 and x3 are the corresponding values in coded forms. The experimental matrix along with natural and coded scales is shown in Table 2. The regression equation for the matrix is represented by the following expression

3. Result and discussion

Yi = b0 + b1X1i + b2 X2i + b3 X3i + b12 X1i X2i + b13 X1i X3i + b23 X2i X3i + b123 X1i X2i X3i

3.1. Statistical analysis In order to examine the main factors and their interactions for the boron removal by adsorption 23 full factorial design was applied. In 23 full factorial design total number of experiments becomes 23 = 8, but totally 16 experiments were performed, because each experiment was

(1)

The main and interaction coefficients have been calculated by following relations [18]:

b0 =

Yi

∑N

(2)

Table 2 Experimental matrix Serial number (i)

Temperature Actual

1 2 3 4 5 6 7 8

20 40 20 40 20 40 20 40

Initial pH of solution

Particle type

Response

Coded X1i

Actual

Coded X2i

Actual

Coded X3i

(Yi)

− + − + − + − +

6.18 6.18 10 10 6.18 6.18 10 10

− − + + − − + +

Granule Granule Granule Granule Powdered Powdered Powdered Powdered

− − − − + + + +

Y1 Y2 Y3 Y4 Y5 Y6 Y7 Y8

Y9 Y10 Y11 Y12 Y13 Y14 Y15 Y16

N. Öztürk, D. Kavak / Desalination 223 (2008) 106–112

109

Table 3 Design of trial runs (in coded form) for boron removal by adsorption in two replicate experiments Trial no.

X1

X2

X3

X1X2

X1X3

X2 X3

X1X2 X3

Y adsorbed boron amount (g/L)

Y adsorbed boron amount (g/L)

Y average adsorbed boron amount (g/L)

1 2 3 4 5 6 7 8

− + − + − + − +

− − + + − − + +

− − − − + + + +

+ − − + + − − +

+ − + − − + − +

+ + − − − − + +

− + + − + − − +

10 12 16 14 9 21 14 16

12 10 12 17 7 17 11 14

11 11 14 15.5 8 19 12.5 15



X jiYi

bj =

bnj =



(3)

N ( X ni X ji ) Yi N

(4)

where Yi is the response (adsorbed boron amount), and Xji values (j = 1,2,3; i = 1,2,3...16) represent the corresponding parameters in their coded forms (Table 2), b0 gives the average value of the results obtained for the adsorbed boron amount; b1, b2 and b3 are the linear coefficients (independent parameters); b12, b13, b23 and b123 are the interaction coefficients and N is the number of total experiments. Coefficient b12, b13, and b23 show the interacting effects of two variables at a time and b123 shows the interacting effect of all three variables taken at a time. The design matrix and the results showing adsorbed boron amount are shown in Table 3. The values of regression coefficient determined are given in Table 4. The values of these coefficients were incorporated in Eq. (1), which takes the following form: Y = 13.25 + 1.875 X1 + X2 + 0.375 X3 − 0.875 X1X2 + 1.5 X1X3 − 0.875 X2 X3 − 1.25 X1X2 X3

(5)

The effect of individual variables and interactional effects can be estimated Eq. (5). According to Eq. (5), temperature, initial pH of solution

and particle type have a positive effect on boron removal by adsorption in the range of variation of each variable selected for the present study. On the other hand, the greatest effect on boron removal was supplied by temperature. The interactions between temperature and initial pH of solution, and particle type, initial pH of solution and temperature (triple effect), initial pH of solution and particle type effect have a negative effect on boron removal. On the other hand, the interaction between temperature and particle size have positive effect on boron removal at studied conditions. Variance of every factor and the sequence of importance of the factors were determined by the F-test (Fisher test) method [16]. Using Table 4 Values of model coefficients Coefficients

Values

b0 b1 b2 b3 b12 b13 b23 b123

13.25 1.875 1 0.375 −0.875 1.5 −0.875 −1.25

110

N. Öztürk, D. Kavak / Desalination 223 (2008) 106–112

Table 5 F ratios and decisions according to analysis of variance Source of variation

F ratio

Decision a = 0.1

Decision a = 0.05

Decision a = 0.01

X1 X2 X3 X1X2 X1X3 X2X3 X1X2 X3

13.636 3.879 2.97 0.545 8.727 2.969 6.061

Effective Effective Ineffective Ineffective Effective Ineffective Effective

Effective Ineffective Ineffective Ineffective Effective Ineffective Effective

Effective Ineffective Ineffective Ineffective Ineffective Ineffective Ineffective

Y 0.1 = 13.25 + 1.875 X1 + X2 + 1.5 X1X3 − 1.25 X1X2 X3

(6)

Y 0.05 = 13.25 + 1.875 X1 + 1.5 X1X3 − 1.25 X1X2 X3

(7)

Y 0.01 = 13.25 + 1.875 X1

(8)

The most important parameter which is efficient in boron adsorption from aqueous solution is temperature which is followed by interactions temperature-particle type, temperature-initial pH of solution-particle type and, initial pH of

solution at 90% confidence level. From the statistical analysis, it can be concluded that adsorption was favoured by increase in temperature, initial pH of solution and powdered cerium oxide more effective than granule cerium oxide. 3.2. Adsorption isotherms The adsorption isotherm of boron adsorption is given in Fig. 1. The concentration of the solid phase is expressed as qe mg adsorbate (solute)/g adsorbent (solid) and in the solution phase as Ce mg adsorbate/L solution. Data that follow linear law can be expressed: qe = K Ce

(9)

where K is a constant determined experimentally (L/g adsorbent). This linear isotherm is not common, but in the dilute region it can be used to approximate data of many systems [19].

qe (mg/g)

Fisher’s test, not only effects and interactions without meaning can be eliminated but the ones that have more influence on the boron removal by adsorption process also be verified. According to analysis of variance, calculated F ratios and decisions at 0.1, 0.05 and 0.01 probability levels are given in Table 5. Estimated F values were compared with Fisher’s value (F0.1(1.8)) = 3.46; (F0.05(1.8)) = 5.32; (F0.01(1.8)) = 11.26. At the 90% confidence level, X1, X2 variables and X1X3 and X1X2 X3 interactions was effective on boron removal from aqueous solutions. At the 95% confidence level X1, X1X3, X1X2 X3 effective and at the 99% confidence level only X1 effective on boron removal by adsorption. It can be assumed that the following equations adequate at 90%, 95%, 99% confidence levels respectively.

12 10 8 6 4 2 0

y = 0.0259x – 0.2857 R 2 = 0.9656

0

100

200 300 Ce (mg/L)

Fig. 1. Adsorption isotherm.

400

500

N. Öztürk, D. Kavak / Desalination 223 (2008) 106–112

The Langmuir isotherm is represented by the following equation Ce /qe = 1/qob + Ce /qo

(10)

where qo signifies the adsorption capacity (mg/g) and b is related to the energy of adsorption (L/mg). The linear plot of Ce/qe versus Ce shows that adsorption follows a Langmuir isotherm. Experimental data were not fitted to Langmuir isotherm and not shown. The Freundlich isotherm was also applied for the boron removal by adsorption Freundlich isotherm model is given by the following equation:

111

4. Conclusions The major conclusions derived from the present work are • The parameters affecting adsorbed boron amount are given as temperature, temperatureparticle type, pH-particle type-temperature at the level of 5% probability. • Maximum boron removal was obtained at original pH value of boron solution and 40°C by powdered cerium oxide. • Linear law is obeyed better than Freundlich and Langmuir isotherms. Acknowledgement

log qe = log Kf + (1/n) log Ce

(11)

where Kf and n Freundlich adsorption isotherm constants. Values of Kf and n were calculated from the intercept and slope of the plots of log qe versus log Ce. Isotherm constants and R2 values were given in Table 6. In general, as the Kf value increases, the adsorption capacity of adsorbent increases. If n is equal to unity the adsorption is linear and adsorption sites are homogenous in energy and no interaction takes place between the adsorbed compounds. If the n value smaller than 1 adsorption is favourable. If the value of n is greater than 1 adsorption bond becomes weak; unfavourable adsorption occurs as a result of the adsorption capacity decreases [7]. As can be seen from Table 6 (n > 1) boron adsorption onto cerium oxide is unfavourable.

This study was financially supported by Osmangazi University Research Fund (Project No: 200315038). References [1]

[2]

[3]

[4]

[5]

Table 6 Isotherm constants Linear isotherm constant

Freundlich constants [6]

2

K (L/g)

R

0.0264

0.97

Kf (mg/g)

n

R2

0.048

1.31

0.75

B. Karasu, G. Kaya, H. Yurdakul and H. Topkaya, Boraks kati atiklarinin duvar karosu bünyelerinde kullanimi ve mikro yapiya olan etkilerinin incelenmesi, Bor Sempozyumu, Balikesir Üniversitesi, 20–22 Haziran (2002), sayfa 18. R.O. Nable, G.S. Banuelos and J.G. Paull, Boron toxicity, Boron in Soils and Plants: Reviews, B. Dell, P.H. Brown and R.W. Bell, (Eds.), Kluwer Academic Publishers, Netherlands, 1997, pp. 182–198. O. Okay, H. Güçlü, E. Soner and T. Balkas¸, Boron pollution in the Simav River, Turkey and various methods of boron removal, Water Res., 19 (7) (1985) 857–862. N. Öztürk and D. Kavak, Boron removal from aqueous solutions by adsorption using full factorial design, Fresenius Environ. Bull., 12 (12) (2003) 1450–1456. N. Öztürk and D. Kavak, Boron removal from aqueous solutions by adsorption on waste sepiolite and activated waste sepiolite using full factorial design, Adsorption, 10 (2004) 245–257. N. Öztürk and D. Kavak, Adsorption of boron from aqueous solutions using fly ash: batch and column studies, J. Hazard. Mater., B127 (2005) 81–88.

112

N. Öztürk, D. Kavak / Desalination 223 (2008) 106–112

[7] Y. Seki, S. Seyhan and M. Yurdakoç, Removal of boron from aqueous solution by adsorption on Al2O3 based materials using full factorial design, J. Hazard. Mater., B138 (2006) 60–66. [8] O. Recepogalu and Ü. Beker, A preliminary study on boron removal from Kizildere/Turkey geothermal waste water, Geothermics, 20 (1/2) (1991) 83–89. [9] M.O. Simonnot, C. Castel, M. Nicolai, C. Rosin, M. Sardin and H. Jauffret, Boron removal from drinking water with a boron selective resin. Is the treatment really selective? Water Res., 34 (1) (1999) 109–116. [10] N. Nadav, Boron removal from seawater reverse osmosis permeate utilizing selective ion exchange resin, Desalination, 124 (1999) 131–135. [11] S. S¸ahin, A mathematical relationship for the explanation of ion exchange for boron adsorption, Desalination, 143 (2002) 35–43. [12] E. Kalafatogalu, M. Örs, S.S. Özdemir, M. Koral and F. I·s¸bilir, Treatment of boron Compound Production Plant’s Wastewater, UKMK3, Erzurum, 1998, pp. 1317–1322. [13] H. Hos¸gören, S. Tural, F. Kahraman, M. Togrul and M. Karakaplan, Solvent extraction of boron with 1,2-dihydroxy-4-oxadodecane (DHD) n amyl

[14]

[15]

[16]

[17]

[18]

[19]

alcohol, Sol. Extract. Ion Exch., 15 (2) (1997) 249–257. M. Badruk, N. Kabay, M. Demircioglu, H. Mordogan and U. Ipekoglu, Removal of boron from wastewater of geothermal power plant by selective ionexchange resins. I. Batch sorption-elution studies, Separ. Sci. Technol., 34 (13) (1999) 2553–2569. M.R. Pastor, A.F. Ruiz, M.N. Chillon and D.P. Rico, Influence of pH in the elimination of boron by means of reverse osmosis, Desalination, 140 (2001) 145–152. D.C. Montgomery, Design and analysis of experiments, 4th edn., John Wiley & Sons, USA, 1997, pp. 7–8. B.P. Singh, L. Besra and S. Bhattacharjee, Factorial design of experiments on the effect of surface charges on stability of aqueous colloidal ceramic suspension, Coll. Surface. Physicochem. Eng. Aspect., 204 (2002) 175–181. R.N. Sahoo, P.K. Naik and S.C. Das, Leaching of manganese from low-grade manganese ore using oxalic acid as reductant in sulphuric acid solution, Hydrometallury, 62 (2001) 157–163. W.J. Weber, Physicochemical Processes for Water Quality Control, John Wiley and Sons Inc., New York, USA, 1972.