Scale Up Of Adsorption Column Packed With Conditioned

Waste Lignocellulosics
O. Kopsidas, I. G. Konstantinou, D. K. Sidiras and F. A. Batzias.

Laboratory of Simulation of Industrial Processes, Department of Industrial Management and Technology, University
of Piraeus, 80 Karaoli and Dimitriou, GR 18534 Piraeus, Greece

Abstract. The simulation and the scale up of the column kinetics of methylene blue adsorption on conditioned waste
lignocellulosics were investigated, using untreated barley straw as control, in order to explore its basic use as a low cost
absorbent for wastewater dye removal and simulating methylene blue adsorption. The BET surface area, the adsorption
capacities (estimated according to Freundlich model), indicated the significant adsorption properties of the original material
that can be attributed to the removal of the hemicelluloses during sulfuric acid treatment, resulting in the ‘opening’ of the
pores of lignocellulosic matrix’s structure and the increasing of the BET surface area.
Keywords: adsorption; barley straw; column; methylene blue; scale up.
PACS: Replace this text with PACS numbers; choose from this list: http://www.aip.org/pacs/index.html

1. INTRODUCTION

The utilization of inexpensive, locally available and effective materials as potential substitutes of commercial
activated carbon for the removal of dyes in aqueous solutions is commonly discussed in literature. The
environmental impact of the dyes present in the wastewater streams of many industrial sectors, such as dyeing,
textile, paper, tannery and the paint industry, has drawn a lot of attention, emphasizing the necessity for their
removal. Undoubtedly, low-cost adsorbents offer a lot of promising benefits for commercial purposes in the future.
Feasibility of using low cost materials, such as cotton waste, rice husk (1), betonite clay (2), neem leaf powder,
(3) powdered activated sludge, perlite (4), bamboo dust, coconut shell, groundnut shell, rice husk and straw, duck
weed (5), sewage sludge (6), gram husk (7), coal bottom ash (8-9), bagasse fly ash (10), blast furnace slag (11) and
deoiled soya (12-13), as adsorbents for removal of various dyes from wastewaters. The suitability of a range of
materials as dye adsorbents (e.g. activated carbons from bamboo dust, coconut shell, groundnut shell, rice husk and
straw) is often established using (i) the kinetics of methylene blue adsorption during batch and continuous (column)
processes (14) and (ii) the BET surface area.
In the present study, the scale up of adsorption column, the BET surface area and the removal of chemical grade
methylene blue by barley straw were studied using untreated barley straw as control. The results indicate that the
lignocellulosic material exhibits advanced adsorption properties and high BET surface area values. Considering the
abundance and low cost of the original material the untreated barley straw could be an extremely cost-effective
adsorbent for basic dyes.

2. METHODS

2.1 Material Development

The barley straw used was obtained from Thessaly, a Greek region, as a suitable source for full-scale/industrial
applications. The moisture of the material was 8.5% w/w. After grinding by a hummer mill and screening, the
fraction with particle sizes between 1.5 and 2.5 cm was isolated as ‘coarse grinded barley straw’. It was saturated for
24 hours before the experimental procedure. The chemical composition and the BET surface area of the raw material
are presented in Table 1.
The composition of the untreated barley straw (cellulose, hemicelluloses, lignin etc as dry weight of product %
w/w of the untreated dry material) and the BET surface area values of the material are presented in Table 1.

2.2 Experimental Setup

Experimental investigations before designing a full-scale column are usually made in a small diameter fixed bed.
The procedure is time-consuming and expensive. In this paper, fixed-bed up-flow absorber studies were conducted
in a 137 cm x 19 cm plexiglass column and in a 15 cm x 3 cm stainless steel column filled with 2730 g and 14 g of
coarse grinded barley straw, respectively.
The experimental set-up consisted of three parallel columns, fed by a multi-channel peristaltic pump at a
constant flow rate, ranging from 5 - 15 mL.min -1. The methylene blue influent concentration was 14 mg.L -1.
Interconnective tubing and fitting were made of polytetrafluoroethylene (PTFE). Effluent samples were analyzed to
yield output concentration breakthrough curves.

2.3 Analytical Techniques

The degree of crystallinity of wood cellulose was measured with the X-ray diffraction method proposed by Segal
et al. (15). Following the technique proposed by Saeman et al. (16), the lignocellulosic materials were hydrolyzed to
glucose and reducing sugars in nearly quantitative yields; the filtrates were analyzed for glucose using the Glu-cinet
test and for reducing sugars using the Somogyi technique (17). Based on these results the cellulose and
hemicelluloses contents of the adsorbents were estimated. Finally, the acid-insoluble lignin (Klason lignin) was
determined according to the Tappi T222 om-88 method.
The BET (Brunauer, Emmet and Teller) surface area of the original straw was measured from the N 2 adsorption
isotherm with an ‘Areameter II’ (Juwe - Laborgeräte GmbH) sorptiometer in accordance with DIN 66132 (18). Prior
to this measurement, the samples were dried under vacuum at 150oC overnight to clean the surface.
The concentrations of methylene blue in the solution was obtained by measuring O.D. at 663 nm and ( max ),
using a HACH DR4000U UV–visible spectrophotometer. The effect of the pH at the max value for specifically
severe conditions (pH > 12 and adsorption time t = 7 days) was taken into account.

3. RESULTS AND DISCUSSION

3.1 Column Studies

Oulman (19) proposed the use of a bed depth service model for simulating GAC (granular activated carbon)
adsorption beds. The model, first developed by Bohart and Adams (20), was based on surface reaction theory and is
equivalent to the logistic curve (21, 22, 23 – 29). The Bohart-Adams equation is as follows
C  KN x
ln i  1   K  Ci  t (1)
C  u
in which C = effluent concentration (mg.L-1); Ci = influent concentration (mg.L-1); K = an adsorption rate coefficient
(L.mg-1.min-1); N = an adsorption capacity coefficient (mg.L -1); x = bed depth (cm); u = linear velocity (cm.min-1);
and t = time (min). The above equation can be rewritten as
C 1
 (2)
Ci 1  e a  bt
where a = K.N.x/u and b = K.Ci.
Eq. (2) corresponds to the ‘logistic curve’, an S-shaped curve, which is symmetrical around its midpoint at t =
a/b = (Nx)/(uCi), C = Ci/2. One of the limitations of the ‘simple logistic function’ is that it requires symmetry. Clark
(25) has developed an alternative to the ‘simple logistic function’, called the ‘generalized logistic function’, which
incorporates the parameter n of the Freundlich (30) adsorption isotherm:
 1
 Ci n 1  n 1
C (3)
 1  Ae  
 rt 

where n = inverse of the slope of the Freundlich isotherm; A = ea; r=b.

Eq. (3) was applied to the effluent data from the column adsorber (Fig. 1). Transforming eq. (3):
n 1
 Ci 
   1  A  e r t (4)
C
Linearizing eq. (4):
 Ci  n1 
ln    1  ln A  r  t (5)
 C  
The values of A and r can be thus estimated from the column effluent data assuming Ci = column influent and C
= the column effluent at time t. As can be observed in Fig. 2, the theoretical estimations - curves according to the
theoretical model expressed by the eq. (3) - sufficiently simulate the experimental data. In addition, the adsorption
rate coefficient (K) and the adsorption capacity coefficient (N), shown in Table 2, were estimated from the A and r-
values. The adsorption capacity coefficient values were found to be significant high for the tested material (Figs 3a,
b, c and d).
Introducing the N parameter (in mg.L-1), as a variable, possibly depended on Cp, we observe an almost linear
relation between them:
N  a N  bN .C p (6)
By combining relations (2) and (6) we obtain eq. (7), which simulates qm parameter (in mg.g-1), growth as a
function of the mild acid hydrolysis processing time, including preheating period > (40+30)/60 = 1.17 h:
N  N e  ( N e  N 0 )e  k N .t (7)
where Ne is the value of N for t   , N0 is the value of N for t = 0 (at the beginning of the preheating period) and
kN is the kinetic constant in h -1. The parameters of eq. (6) were found by non-linear regression analysis (Table 3). The
solid line is for the methylene blue absorption on the coarse grinded barley straw fraction in Fig. 3c. By the
assumption that kN = kp then N e  a N  bN C pe , and N 0  N e  bN .C pe .
Alternatively, the above eq. (7) comes from the combination of eq. (3) with the following equation
N  a'N  b'N .B (8)
by the assumption that kN = kB. In this case N e  a' N b' N Be , and N 0  N e  b' N  Be  B0  .
The results from the fitting of eq. (8) to the experimental data are given in Fig. 3b and d. The solid line in Fig. 3d
is for the methylene blue absorption on the coarse grinded barley straw fraction. The eq. (8) parameters’ values
estimated with linear regression analysis are given in Table 3. The N values are increasing linearly vs. BET surface
area increasing for coarse grinded barley straw fractions.
Applying the bed depth service model for the simulation of straw adsorption beds, the adsorption capacity
coefficient (N) values were found to be significant high. In addition, the linear correlations of the KF values, the
adsorption capacities (qm) and the adsorption capacity coefficient values (N) vs BET surface area were found to be
statistically significant, as is evident from the correlation coefficient R-values presented in Table 3.
The 15 cm column, filled with 14 g of coarse grinded barley straw, gives us the values: R = -0.9413, N = 786
mg/L, K = 0.00079, Q = 22.52 mL/min, E = 4.91 cm2 , u = 4.591 cm/min and g/L = 0.1902.
The 137 cm column (1st day), filled with 2730 g of coarse grinded barley straw, gives us the values: E = 283.39
cm2 and g/L = 0.0703.

3.2 Discussion
The column kinetics of methylene blue adsorption on barley straw was simulated herein, using untreated barley
straw as control, in order to facilitate its potential use as a low cost adsorbent for wastewater dye removal.
Economies arise when the facility that can use such adsorption materials is near a source of a lignocellulosic waste
as agricultural residues, thus saving transportation cost and contributing to industrial ecology at local level.
 The simulation results indicate the significant adsorption properties of the barley straw. This can be attributed to
the ‘opening’ of the pores of the structure of the lignocellulosic substrate and the increasing of the BET surface area
of the absorbent material by the removal of the hemicelluloses and the amorphous part of cellulose. The linear
correlation of the BET surface area vs. the removed by the acid hydrolysis polysaccharides (cellulose and
hemicelluloses) fraction was found to be statistically significant. Moreover, the linear correlations of the KF values,
the adsorption capacities (qm) and the adsorption capacity coefficient (N) values vs. the BET surface area, were also
found to be statistically significant, allowing good predictability for industrial applications.
Below is an example equation created with Word 97’s Equation Editor. To move this equation, highlight the
entire line, then use cut and paste to the new location. To use this as a template, select the entire line, then use copy
and paste to place the equation in the new location.
d  F1  d  F1 
 SAm2 cos  ,  SAm3 cos  .
d 2 d 3

ACKNOWLEDGMENTS
The authors thank the Research Center of the University of Piraeus for the financial support.

REFERENCES
1. G. McKay, G. Ramprasad and P.P. Mowli, Water Air Soil Pollut. 29, 273-283 (1986).
2. K. R. Ramkrishna and T. Viaraghavan, Water Sci. Technol. 36, 189-196 (1997).
3. K. G. Bhattacharya and A. Sharma, Dyes Pigm. 65, 51-59 (2005).
4. M. Dogan, M. Alkan, A. Turkyilmaz and Y. Ozdemir, J. Hazard. Mater. 109, 141-148 (2004).
5. P. Waranusantigul, P. Pokethitiyook, M. Kruatrachue and E. S. Upatham, Environ. Pollut. 125, 385-392 (2003).
6. M. Otero, F. Rozada, L. F. Calvo, A. I. Garcia and A. Moran, Biochem. Eng. J. 15, 59-68 (2003).
7. R. Jain, and S. Sikarwar, Int. J. Environ. Pollut. 27, 158-178 (2006).
8. V. K. Gupta, A. Mittal, L. Krishnan and V. Gajbe, Separat. Purificat. Technol. 40, 87-96 (2004).
9. A. R. Dincer, Y. Gunes and N. Karakaya, J. Hazard. Mater. 141, 529-535 (2007).
10. A. Mittal, L. Kurup (Krishnan) and V. K. Gupta, J. Hazard. Mater. 117, 171-178 (2005).
11. V. K. Gupta, D. Mohan, S. Sharma and M. Sharma, Separat. Sci. Technol. 35, 2097-2113 (2000).
12. V. K. Gupta, I. Ali, Suhas and D. Mohan, J. Colloid. Interface Sci. 265, 257-264 (2003).
13. A. Mittal, L. Krishnan and V. K. Gupta, Separat. Sci. Technol. 43, 125-133 (2005).
14. N. Kannan and M. M. Sundaram, Dyes Pigments 51, 25-40 (2001).
15. L. Segal, J. J. Greely, A. E. Martin and C. M. Conrad, Textile Res. J. 29, 786-795 (1959).
16. J. F. Saeman, J. F. Bubl and E. E. Harris, Ind. Eng. Chem. Anal. Ed. 17, 35-37 (1945).
17. M. Somogyi, J. Biol. Chem. 195, 19-23 (1952).
18. DIN 66132, Determination of specific surface area of solids by adsorption of nitrogen; single-point differential method
according to Haul and Dümbgen (1975).
19. C. S. Oulman, J. Am. Wat. Works Assoc. 72, 50-53 (1980).
20. G. S. Bohart and E. Q. Adams, J. Chem. Soc. 42, 523–544 (1920).
21. S. J. Allen, Q. Gan, R. Matthews and P. A. Johnson, Bioresource Technol. 88, 143-152 (2003).
22. F. A. Batzias and D. K. Sidiras, Biores. Technoldoi :10.1016/j.biortech.2006.05.020 (2006).
23. R.A. Hutchins, Chem. Eng. 80, 133-138 (1973).
24. B. M. Van Vliet and W. J. Weber, J. Wat. Poll. Contr. Fed. 53, 1585-1598 (1981).
25. R. M. Clark, J. Am. Wat. Works Assoc. 79, (1) 33-131 (1987).
26. K. S. Low, C. K. Lee and A. Y. Ng, Bioresource Technol. 68, 205-208 (1999).
27. G. M. Walker and L. R. Weatherley, Chem. Eng. J. 84, 125-131 (2001).
28. U. Kumar and M. Bandyopadhyay, J. Hazard. Mater. B129(1-3), 253-259 (2006).
29. A. Cambiella, E. Ortea, G. Ríos, J. M. Benito, C. Pazos and J. Coca, J. Hazard. Mater. B131(1-3) 195-199 (2006).
30. H. M. F. Freundlich, Zeitschrift für Physikalische Chemie 57, 385–471 (1906).

TABLE 1. Barley Straw Plexiglass Column 137 cm

x (cm) N(mg/L) K Q (L/min) R
27 50 0.00063 1.32 -0.9599
67 188 0.00076 1.33 -0.9502
 107 170 0.00083 1.22 -0.9333
137 161 0.00083 1.58 -0.9568

14 14
12 12
10 10
C (mg/L)

C (mg/L)
8 8
15 cm
6 6 137 cm
4 27 cm
4
2 67 cm
2
107 cm
0
0
0 2 4 6 8
0 100 200 300 400 500 600 700
V (L)
V (L)
FIGURE 1. This is the Style for Figure Captions. Center this text if it doesn’t run for more than one line.