So Young Lee*, Jong-Un Lee**, Heechul Choi* and Kyoung-Woong Kim* * Kwangju Institute of Science and Technology, Kwangju 500-712, Republic of Korea (E-mail:
[email protected]) ** Chonnam National University, Kwangju 500-757, Republic of Korea Abstract Degradation or detoxification of pollutants by SAT system was generally focused on organic contaminants although the regulation of water reuse has provisions of heavy metals. This study is to evaluate the feasibility of SAT for metals such as Cd, Cr and Pb with the pilot scaled column experiment. The desorption possibility of sorbed metals was also examined in the condition of continuous water and even acidic water (pH 4.3) injection. Fractional factorial analysis is a tool frequently used to identify factors or variables that have an effect on a certain response. A two-level fractional factorial analysis was planned to study the effect of four factors on Pb sorption from the experiments; soil particle size, TOC in solution, Pb concentration in solution and flow rate. The main effects among the factors were obtained by ANOVA based MINITAB software. The effects of four factors were also converted into coefficients; those values may build an empirical model to predict the metal sorption of soils. Keywords Fractional factorial analysis; heavy metals; SAT (Soil Aquifer Treatment); sorption
Introduction General idea of SAT
Soil aquifer treatment is an economical and aesthetic wastewater reclamation system. Since the soil and aquifer can act as a natural filter, SAT system can remove suspended solids, biodegradable materials, bacteria, viruses and other microorganisms (Bouwer, 1987). Significant reductions of nitrogen and phosphorus were reported, however, removal of heavy metals from wastewater is also possible with the sorption and physico-chemical stabilization (NCSWS, 2001). Confronted with water shortage, re-use of discharged water from the sewage treatment plant can be an alternative water resource. Considering the huge volume of discharged water from the plant, it is directly abandoned to the river and the sea without reuse, and the cost effective and easy operation of SAT system, application of SAT to reuse wastewater can be a good way to secure water resource.
Water Science and Technology Vol 50 No 2 pp 263–268 © IWA Publishing 2004
Sorption behaviors of heavy metals in SAT (soil aquifer treatment) system
Fractional factorial analysis
Factorial analysis is used to identify the key factors that dominantly influence the process (Normand and David, 1997). Using a full factorial experiment permits direct evaluation of all the main and interaction effects. However, when a large number of independent factors is being investigated simultaneously, such a full factorial experiment is very costly in terms of the number of runs undertaken. It is generally found that interactions among the large number of factors are less likely to be important, high-order interactions can be safely ignored (Tippett, 1934). Consequently, fractional factorial experiments are often used, and fractional factorial analysis is enough to screen factors and to search main effect with confounding and aliasing (Montgomery, 1991). The objectives of this study are to evaluate the possibility of metal SAT system and to build a sorption model which can estimate sorption capacity of metal onto soil with different conditions. This model can suggest guidelines to the operation of metal SAT in field
263
application by forecasting how many metals can be sorbed or how long those fields can be endured in the system. Materials and methods Column experiment
So Young Lee et al.
The soils collected from the riverbank of the Youngsan River in Republic of Korea were packed into a pilot-scale column (150 cm height and 10 cm diameter). The inlet water was collected from discharged water of a sewage treatment plant and spiked as 100 ppm of Cd, Cr and Pb to observe the result clearly. Figure 1 shows the procedure of column experiment and Table 1 presents the properties of soils used in the experiment. After 6 pore volume injections of discharged water, distilled waterand acidic water (pH 4.3) were also infused each for 2 pore volumes to check the feasibility of metal desorption which already sorbed onto soils. Fractional factorial experiment
Four factors including particle size (A), TOC in solution (B), Pb concentration in solution (C) and flow rate (D), were investigated, and each of which was tested at two levels. Eight runs were selected by using the following defining contrast: I = ABCD. The related experimental conditions are shown in Table 2 and the combinations of factors for runs are presented in Table 3. The experiments were carried out with a small column (12 cm height and 3 cm diameter) test (Figure 2). The packed soils were the same with a pilot-scale column experiment and TOC in solution was controlled with distilled water (– level) and discharged water from plant (+ level). Lead concentration in solution was adjusted with PbCl2. Samples were collected at different times (6, 12, 24, 36, 48 min and 1, 2, 3, 4, 5, 6 h) and analyzed by ICP-AES. The experimental results at each treatment were statistically analyzed by MINITAB software.
Table 1 Soil properties Size fraction (µm) Bulk density (kg/L) Pore volume (mL) pH Soil weight (kg) Organic matter content (%)
420~2,000 1.616 4500 6.68 15.7 n.d
Pump
Water level control
150 cm
Water reservoir 10cm 264
Figure 1 Column experimental design
Table 2 Factors and levels investigated for the fractional factorial design Code
Levels
Particle size (µm) TOC in solution (mg/L) Pb concentration in solution (mg/L) Flow rate (L/m2•h)
– 420–600 0 10 339.5
+ 850–1,800 6.3 50 679
Table 3 Experimental conditions investigated in the 24–1 fractional factorial experiment (Defining contrast: I = ABCD) Run
CD AB AC (1) BC BD AD ABCD
A (µm)
B (mg/L)
C (mg/L)
D (L/m2•h)
420–600 850–1,800 850–1,800 420–600 420–600 420–600 850–1,800 850–1,800
0 6.3 0 0 6.3 6.3 0 6.3
50 10 50 10 50 10 10 50
679 339.5 339.5 339.5 339.5 679 679 679
So Young Lee et al.
` A B C D
Factors
3 cm Flow control
12 cm
Figure 2 Fractional factorial experiments
Results and discussion
During the pilot-scale column experiment, spiked heavy metals (each 100 ppm of Cd, Cr, and Pb) were all sorbed onto soil particles in one pore volume (Figure 3). Desorption was not observed in the injection of distilled water and even in the case of acidic water infusion. According to Alloway (1995), heavy metal cations are most mobile under acid condition. Concentration in solution (ppm)
120 Cd 100
Cr Pb
80 60
Inlet of distilled water (pH 4.3)
Inlet of distilled water
40 20 0 0
1
2
3
4
5
6
7
8
9
Pore Volume
Figure 3 Pilot-scale column device
265
Cumulative amounts of Pb (mg)
So Young Lee et al.
As the SAT system depends on the sorption ability of soils, desorption may be apprehended. However, sorption/desorption capacities are different in the comprehensive conditions of soil (Christensen, 1985). In this experiment, it is supposed that heavy metals are suitable for the removal by SAT system and metal SAT system is safe in the condition of acid rain despite increasing metal mobility in the acidic circumstances. The results of fractional factorial experiment show cumulative sorption amounts of Pb onto soils with different runs (Figure 4). Generally, metal sorption reaches equilibrium within 6 hours and in the sufficient metal concentrations in solution, the metal sorption onto soil particles can be completed in one hour. In this treatment, some Pb sorptions were quickly progressed in one hour. On the other hand, the other sorption lasted during 6 hours. Therefore, operating time is an important factor for SAT system, especially in the starting time. As the figure of fractional factorial runs may not show each effect of four factors, results were statistically analyzed with MINITAB software and the fitness of model was also investigated. Figure 5 shows the main effects of four factors that effect on Pb sorption onto soils. The Pb concentration in inlet water is the most important factor that dominates Pb sorption capacity of soils. During one hour, more than 90% of the sorption procedure can be explained with four factors; Significance shows how many portions of results can be described with given factors. After one hour, the effect of Pb concentration decreased and contribution of TOC in solution increased. The effects of particle size and flow rate were negligible during 6 h of leaching compared with other two factors. Because the factorial analysis is effective within the strait trend of each factor’s response, the level of particle size between high and low may be insufficient to the greatest results. The previous results of sorption capacity according to four particle size fractions ( 60 min) Y (t) = {–0.0496Ln(t) + 0.1745} + {0.0001Ln(t) – 0.0007} × {A} +{0.025Ln(t) – 0.15} × {B} + {0.0394t–0.4344} × {C} +{0.0037t–0.5411} × {D} + {0.0006t–0.4996} × {A} × {B} The simulation results obtained by estimated sorption model can predict the sorption behavior and capacity with different levels of factors. However, expended scale work is needed to evaluate the main scale-up parameter of the process and these results constitute useful and basic information on the precise prediction modeling studies. y = 1E-04Ln(x) + 0.0055
0.01 0
Coefficient
0
20
40
-0.01 -0.02 -0.03
60
A: y = -7E-07Ln(x) - 0.0002 D: y = -5E-06Ln(x) + 0.0003 AB: y = 5E-05e0.0006x y = -0.0013Ln(x) - 0.0336
-0.04 y = 8E-06x - 0.0421
constant
-0.05
A
B
C
D
AB
Time (min)
Figure 6 Estimated coefficients for the regression model (during 1 hour) 0.02
Coefficient
-0.04
A: y = 0.0001Ln(x) - 0.0007
C: y = 0.0394x-0.4344
0 100 -0.02
200
300 D: y = 0.0037x-0.5411
y = 0.025Ln(x) - 0.15
400
AB: y = 0.0006x-0.4996
-0.06 -0.08 -0.1 -0.12
y = -0.0496Ln(x) +0.1745
-0.14
co nsta nt
A
B
C
D
AB
Time (min)
Figure 6 Continued (from 1 to 6 hour)
267
Conclusion
So Young Lee et al.
Metal SAT system is effective to remove heavy metals from wastewater and the system is safe without metal desorption even in the acid rain conditions. To the application of metal SAT system in the field, predicting and monitoring the possible amounts of metal sorption is needed. Starting from the theory of two-level fractional factorial analysis, an empirical model has been developed to describe the metal sorption, relating metal concentrations in leaching solution at different process conditions. Among four factors (particle size, TOC in solution, Pb concentration in solution and flow rate), the most significant factor on the metal sorption was found as metal concentration in solution and the sorption procedures can be well explained with those factors during an hour procedure. The regression analysis of these effects was performed to obtain useful empirical equations. The predictive equations can examine the safety of soil and water and show the time when soils are changed for the effective metal SAT system. Acknowledgements
This research was supported by grants from the Kwangju Institute of Science and Technology (K-JIST) and Water Reuse Technology Center (WRTC) at Kwangju Institute of Science and Technology (K-JIST). References Alloway (1995). Heavy Metals in Soils, 2nd edn., Chapman & Hall, pp. 12–13. Bouwer, H. (1987). Soil-aquifer treatment of sewage. Paper prepared for the Land and Water Development Division, FAO, Rome. Christensen, T.H (1985). Cadmium soil sorption at low concentrations. Water, Air and Soil Pollution, 26(1), 265–274. Montgomery D.C. (1991). Design and Analysis of Experiments, 3rd edn., Wiley. National Center for Sustainable Water Supply (NCSWS) (2001). Investigation on soil-aquifer treatment for sustainable water reuse (Research project summary). Frigon N.L. and Mathews D. (1997). Practical Guide to Experimental Design, Wiley, 163–164. Tippett, L.H.C. (1934). Application of Statistical Methods to the Control of Quality in Industrial Production, Manchester Statistical Society.
268
