w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 7 8 3 e2 7 9 4

Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

Characterisation of adsorbents prepared by pyrolysis of

sludge and sludge/disposal filter cake mix

I. Velghe a,b, R. Carleer b, J. Yperman b,*, S. Schreurs a, Jan D’Haen c

a
NuTeC, Department TIW, XIOS, Agoralaan Gebouw H, 3590 Diepenbeek, Belgium
b
Laboratory of Applied and Analytical Chemistry, CMK, Hasselt University, Agoralaan Gebouw D, 3590 Diepenbeek, Belgium
c
Institute for Materials Research, Wetenschapspark 1, 3590 Diepenbeek, Belgium

article info abstract

Article history: Copper and zinc removal from water (pH ¼ 5.0) using adsorbents produced from slow and
Received 18 November 2011 fast pyrolysis of industrial sludge and industrial sludge mixed with a disposal filter cake
Received in revised form (FC), post treated with HCl, is investigated in comparison with a commercial adsorbent
13 February 2012 F400. The results show that a pseudo-second order kinetics model is followed. The Lang-
Accepted 17 February 2012 muireFreundlich isotherm model is found to fit the data best. The capacity for heavy metal
Available online 24 February 2012 removal of studied adsorbents is generally better than that of commercial F400. The
dominant heavy metal removal mechanism is cation exchange. Higher heavy metal
Keywords: removal capacity is associated with fast pyrolysis adsorbents and sludge/FC derived
Adsorbent adsorbents, due to enhanced cation exchange. Improvement of Zn2þ removal via 1 N HCl
Pyrolysis post-treatment is only effective when exchangeable cations of the adsorbent are
Metal substituted with Hþ ions, which boost the cation exchange capacity. Increase of temper-
Sludge ature also enhances metal removal capacity. Fast pyrolysis sludge-based adsorbents can be
Disposal filter cake reused after several adsorptionedesorption cycles.
ª 2012 Elsevier Ltd. All rights reserved.

1. Introduction quality level of copper and zinc in surface water used for
drinking water production is 1 mg/l and 5 mg/l respectively
Elements like chromium, copper and zinc are indispensable in (Vlarem II).
small concentrations for human and other living species. But Removal of heavy metals from water is thus an important
at high concentrations and depending on their oxidation state environmental issue. Many technologies have been developed
(e.g. chromium), these metals are toxic. Especially copper and to remove heavy metals from waters, such as chemical
zinc, besides cadmium, mercury and lead, causes environ- precipitation, ion exchange, membrane separation, biological
mental problems in Flanders (Belgium). The presence of these treatment and adsorption. But most commercial carbons have
metals are due to historical industrial contamination (MIRA, limited ability to remove metals (Lima and Marschall, 2005).
http://www.milieurapport.be). Elevated levels of copper result Besides this, the high cost of activated carbons stimulates
in health problems such as liver and kidney failure or Wilson’s researchers to develop low cost adsorbents effective to
disease (http://www.ninds.nih.gov), apart from damage to remove metals.
plants and decrease in the yield of farmlands. Symptoms of The last decade, sewage sludge and industrial sludge are
zinc toxicity include irritability, muscular stiffness, loss of being generated in increasing amounts due to the rapid
appetite and nausea. In Flanders (Belgium), the environmental urbanization and industrialization. The conventional sludge

* Corresponding author. Tel.: þ32 11 268 320; fax: þ32 11 268 301.
E-mail addresses: [email protected] (I. Velghe), [email protected] (R. Carleer), [email protected] (J. Yperman),
[email protected] (S. Schreurs).
0043-1354/$ e see front matter ª 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.watres.2012.02.034
2784 w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 7 8 3 e2 7 9 4

disposal options are landfill, application to farmland and aerobic waste water treatment plant (leftovers from basic
forestry, and incineration. However, cost effective and chemistry, plastic and fertilizer production processes) and
renewable disposal possibilities are of concern. A potential a filter cake (FC) produced during polyetherol production
alternative is e.g. the manufacture of adsorbents. It is process. FC mainly exists out of phosphate and carbonate
demonstrated that sludge can be converted into adsorbents crystals, and magnesium silicates besides small amounts of
(Mendez et al., 2005; Seredych and Bandosz, 2006; Rio et al., organics (data obtained by BASF). The sludge and FC are dried,
2006a; Bhatnagar et al., 2008). Moreover, it is showed that crushed and sieved through a 2 mm mesh. Properties of oven
carbonaceous adsorbents obtained from waste metal dried sludge, FC and sludge/FC mix (85:15 w/w) are found in
sludge cost 0.1e0.2 US $ per kg (Bhatnagar et al., 2008), making Table 1. As expected, FC contains high amounts of silicon and
them useful materials in terms of cost as compared to phosphor, but also high amounts of potassium.
commercial carbons which cost more than 3.0 US $ per kg 100 g oven dried (100  C) sludge, as well as 100 g of homo-
(Ahmaruzzaman, 2011). geneous mix of oven dried (100  C) sludge/FC (85 g/15 g) are
The method of preparation, including the characteristics of used separately for slow and fast pyrolysis. Slow and fast
starting material, pyrolysis process, activation agents, acti- pyrolysis experiments are performed in a home-built semi-
vating process, acid washing step, etc., can strongly influence continuous lab-scale reactor. An extensive and updated
the adsorption characteristics of the activated carbon. The description of the reactor setup is given by Smets et al. (2011).
selectivity and sensitivity of metal removal process are also Pyrolysis atmosphere is ensured by a 2  70 ml/min nitrogen
influenced by aqueous solution pH, contact time, initial metal flow. Sand (700 g) is used as a heat transfer medium. Slow
concentration, type of metal, adsorbent nature (e.g. surface pyrolysis is performed at a slow heating rate of 5  C/min up to
chemistry, specific porosity, high volume of mesopores) and 450  C during 1.5 h on the 100 g sludge or sludge/FC mixture
adsorbent concentration. The production of slow pyrolysis and 700 g sand, in constant rotation using an Archimedes’
sludge-based adsorbents for copper and zinc removal is screw. Fast pyrolysis is performed at 450  C: 100 g sludge or
reported by several authors (Mendez et al., 2005; Seredych and sludge/FC mixture is directly injected into 700 g heated sand at
Bandosz, 2006; Rio et al., 2006a; Kang et al., 2006; Hammaini a rate of 62 g/min. The yield of adsorbents prepared by
et al., 2007; Bouzid et al., 2008; Wang et al., 2011; pyrolysis ranges between 44 and 50 wt%. After cooling down,
Ahmaruzzaman, 2011), but limited information over fast samples are washed with 0.01 N or 1 N HCl to remove inor-
pyrolysis produced adsorbents is available. ganic matter. They are further washed several times with
In this work, slow and fast pyrolysis produced adsorbents
of industrial sludge and sludge mixed with a disposal filter
cake (FC), are investigated. Slow pyrolysis is characterized by
a slow heating rate and a long residence time, while fast Table 1 e Properties of oven dried sludge, FC and sludge/
pyrolysis is characterized by a fast heating rate and a short FC mix.
residence time. Different parameters of the pyrolysis process Sludge FC Sludge/FC mix
can influence the composition (Williams and Slaney, 2007)
TGA analyses (wt%, dry base)
and thus the adsorption characteristics of the pyrolysis solid
Moisture 7.0 2.0 3.0
adsorbent. The solid adsorbent can be easily made at low cost, Volatiles 52.0 38.2 54.0
since pyrolysis of waste materials to obtain valuable liquid Fixed carbon 18.0 0.0 14.0
and gas products is of growing interest, and thus large Ash 23.0 59.8 29.0
amounts of the solid byproduct will be available. The obtained Ultimate analyses (wt%, dry base)
adsorbents are post-treated with HCl, to partially dissolve the C 34.6 19.0 35.2
H 5.5 3.8 5.0
inorganic matter and to evaluate the influence of HCl washing
N 5.6 0.3 5.4
on heavy metal adsorption. Smith et al. (2009) showed that
O 28.5 15.3 24.1
washing of sewage sludge adsorbents with HCl can yield S <0.1 <0.1 <0.1
improvements in their ‘Brunauer, Emmet and Teller’ (BET) ICP/AES (g/kg)
surface areas. This might positive influence heavy metal Pb 0.008  0.004 e e
removal characteristics. Co 0.040  0.001 e 0.040  0.001
The aim of present study is to compare the obtained Cu 0.207  0.006 e 0.151  0.002
Cr 0.132  0.001 e 0.130  0.002
adsorbents together with a commercial F400, based on prop-
Mn 0.068  0.001 e 0.065  0.001
erties and ability for copper and zinc removal from water. The
Ni 0.089  0.001 e 0.082  0.001
influence of the starting material, pyrolysis process, HCl Ba 0.077  0.020 e 0.057  0.003
treatment and temperature on adsorption capacity and Zn 1.2  0.02 e 0.95  0.011
mechanism of adsorbents is examined. Al 4.0  0.30 0.014  0.003 3.5  0.10
Fe 6.8  0.74 0.018  0.003 6.5  0.01
Na 13.3  4.30 2.5  0.12 14.1  0.31
Mg 6.5  0.60 0.160  0.005 6.7  0.01
2. Experimental
K 4.2  0.37 156  3.40 37.5  0.22
P 8.5  0.40 135  4.00 26.5  0.47
2.1. Adsorbent preparation Ca 87.7  4.80 0.350  0.016 89.4  2.42
Si 63.2  3.12 104  4.20 82.0  1.40
Two different waste streams are supplied by the chemical
‘e‘: below detection limit.
company BASF (Antwerp, Belgium): i.e. sludge from the
 w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 7 8 3 e2 7 9 4 2785

water to remove traces of acid. The obtained adsorbents are obtain acidic surface groups. 200 mg of adsorbent is added to
dried and put in an air-free container until analyses. The 25 ml 0.01 M NaOH and incubated during 48 h in a closed
commercial activated carbon/adsorbent used for comparison vessel. Titration is done with 0.01 M HCl in a nitrogen
is a Chemviron product: F400. atmosphere.
SS, FS, SSFC, FSFC are notations used for adsorbents Scanning Electron Microscopy-Energy Dispersive X-ray
obtained by slow pyrolysis of sludge, adsorbents obtained by (SEM-EDX) is performed with a FEI Quanta 200 FEG-SEM
fast pyrolysis of sludge, adsorbents obtained by slow pyrol- equipped with a Bruker Si(Li) EDX detector. For the surface
ysis of mix sludge/filter cake, and adsorbents obtained by study, secondary electron (SE) images are used. For the cross-
fast pyrolysis of mix sludge/filter cake respectively. The sectional study, backscattered electron (BSE) images are made
extension 0.01 or 1 reflects washing with 0.01 N or 1 N HCl because the BSE yield is related with the chemical composi-
respectively. tion. Porosity characteristics (BET; VT) are analysed by
A schematic overview of the experimental setup is pre- adsorption of nitrogen in 77 K (Gas Sorption Analyser NOVA
sented in Fig. 1: preparation of adsorbents and different 2200 (QUANTACHROME)).
analyses performed on each adsorbent. A detailed description
of the analyses is given in 2.2 and 2.3. 2.3. Aqueous adsorption studies

2.2. Properties/composition of the adsorbents Stock solutions of 0.1 g/l Cu2þ and Zn2þ are prepared by dis-
solving CuCl2.2H2O and ZnCl2 (AR-grade) in de-ionized water
Ultimate analyses are used to determine the CHNS- and O- from a Millipore-Q water system. These solutions are diluted
content, using a Flash EA 1112 Elemental Analyser of Thermo to 2e50 mg/l for adsorption experiments.
Electron Corporation. Ash content is determined by ther-
mogravimetric analyses (TGA) with a DuPont Instruments 951 2.3.1. Adsorption kinetics
Thermogravimetric Analyzer and by ASTM standard D5142- 30 mg of adsorbent is weighted and placed in a set of 12
02a. The metallic elements are determined with an OPTIMA vessels of 50 ml of Cu2þ or Zn2þ aqueous solution (38 or 25 mg/l
3000 DV ICP/AES. ICP/AES analyses are done in triplicate respectively). The pH is adjusted to 5.0 using diluted HCl or
(250 mg sample). pH value is measured by stirring 4 g of NaOH solutions. Adsorption study is performed in a thermo-
adsorbent with one litre of water during 24 h in a closed vessel static bath agitated at 200 rpm at 20  C. Samples are taken at
at 20  C. 12 different time points (5 mine72 h). The concentration of
FTIR analyses are carried out on a Bruker IFS 48 and heavy metal in each solution is measured by using ICP/AES.
a Bruker Vertex 70 FTIR spectrometer. Potentiometric titra- Adsorbent-free controls are run in parallel in all experiments.
tion is done according to Boehm method (Boehm, 2002) to The data are analysed using three kinetic models: Pseudo-first

Fig. 1 e Experimental setup.

2786 w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 7 8 3 e2 7 9 4

order kinetics, Pseudo-second order kinetics and Intraparticle 2.3.3. Dynamic adsorption and desorption
diffusion model (Ho and McKay, 1998). 150 mg of adsorbent is packed in a glass column (0.5 cm
Pseudo-first order kinetics model: diameter and 2 cm length), with a glass wool plug at both ends

 to guarantee constant packing. 15 mg/l Zn2þ solution is passed
q ¼ qe 1  ek 1t (1) through the column with a constant flow rate (2 ml/min)
with qe and q the amount of heavy metal adsorbed per unit during adsorption. 1 N HCl solution is used during desorption.
weight of adsorbent at equilibrium and time t, respectively, The effluent fraction is collected every 3 min, the pH of each
and k1 the constant of pseudo-first order rate. fraction is measured and its heavy metal concentration is
Pseudo-second order kinetics model: determined by ICP/AES after adding HNO3 to the different
effluent fractions. The amounts of Zn2þ adsorbed and des-




t = q ¼ 1= k2 q2e þ t=qe (2) orbed Q (mg/g) are calculated according to Yin et al. (2010):

with k2 the constant of the relevant adsorption rate. X

n

Intraparticle diffusion model: Q¼ ðððC0  Ci Þvti Þ=mÞ (10)

i¼1

q ¼ kint t1=2 (3)

X
n

with kint the constant of the relevant adsorption rate. Q¼ ððCi vti Þ=mÞ (11)
i¼1

2.3.2. Adsorption isotherm where C0 is the initial concentration, Ci is the concentration of

Adsorption isotherm experiments are carried out with 30 mg heavy metal ion in the effluent at time ti; v is the flow rate of
of adsorbent, weighted and placed in a 50 ml of Cu2þ or Zn2þ solution passing through the column; ti is the time of col-
aqueous solution of desired concentrations (2e50 mg/l). The lecting the effluent; m is the mass of adsorbent.
pH is adjusted to 5.0 using diluted HCl or NaOH solutions. The
temperature is 20  C and the contact time is 48 h (experi-
mentally determined) to assure equilibration of adsorption. 3. Results and discussion
The concentration of heavy metal in each solution is
measured using ICP/AES. Adsorbent-free controls are run in 3.1. Chemical composition
parallel and all experiments are done in duplicate. Three
kinds of isotherm models are studied: Langmuir, Freundlich Table 2 presents the contents of carbon, hydrogen, nitrogen,
and LangmuireFreundlich isotherm models. oxygen, ash and metals of the different adsorbents. As expec-
The Langmuir equations are (Langmuir, 1918): ted, an enrichment of inorganic species is found into the

 adsorbents due to the release of volatiles during pyrolysis
qe ¼ qm KL Ce ð1 þ ðKL Ce ÞÞ (4)
process (Tables 1 and 2). For the 0.01 N HCl post treated
or adsorbents, K, Na and P are partially leached. Mainly Ca and Si




  are present in significant quantities in these adsorbents. Add-
Ce =qe ¼ 1= qm KL þ 1=qm Ce (5) ing FC to the sludge before pyrolysis results in much higher K
with Ce the equilibrium concentration of heavy metal in and P contents (although a lot of K and P is leached through acid
solution, qm the monolayer adsorption capacity and KL the washing) and a small increase of the Si content in SSFC_0.01
constant related to adsorption affinity. and FSFC_0.01 adsorbents. For the 1 N HCl post-treated adsor-
The Freundlich equations are (Freundlich and Helle, 1939): bents, all metals e except Pb, Co, Cu, Cr and Ni e are partially
leached, and especially Ca and Si are removed in high amounts.
qe ¼ KF Ce1=n (6) Except for Zn, minor concentrations of heavy metals Pb,
Co, Cu, Cr, Mn, Ni and Ba are present in all prepared adsor-
or
bents. A major concern with the use of these derived adsor-
log qe ¼ log KF þ 1=nlog Ce : (7) bents is the possible release of these hazardous heavy metals
during effluent treatment applications. Therefore, leaching of
with KF the relative adsorption capacity constant of the
heavy metals is evaluated during different heavy metal
adsorbent. n is related to the uptake intensity of the adsorp-
removal experiments, but none of these heavy metals are
tion constant KF.
found in detectable amounts in the leaching solutions.
The LangmuireFreundlich equations are (Marczewski
et al., 1988):
3.2. Surface functionality features



qe ¼ qm ðKLF Ce Þn = 1 þ ðKLF Ce Þn (8)
FTIR spectroscopy is performed to provide information of
or surface functional groups. Similar results are obtained for both




  attenuated total reflection (ATR) FTIR and transmission FTIR
Cne =qe ¼ 1= qm KnLF þ 1=qm Cne (9)
using KBr pellets. The characteristic bands of several func-
with KLF the Langmuir-type equilibrium constant defined by tional groups are identified and presented in Table 3. SS_0.01,
the Van ‘t Hoff equation for heterogeneous solid, and the FS_0.01 and FS_1 adsorbents have similar FTIR spectra and the
exponential term n the heterogeneity parameter of the site absorption peaks correspond to the presence of carbonate and
energies (0 < n < 1). silicate functional groups (alumino-silicates). For SS_0.01 and
 w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 7 8 3 e2 7 9 4 2787

Table 2 e Ultimate, proximate and ICP/AES analyses of adsorbents.

SS_0.01 FS_0.01 SSFC_0.01 FSFC_0.01 F400 FS_1 FSFC_1

Ultimate & proximate analyses (wt%, dry base)

C 38.1 26.5 26.4 26.2 83.4
H 1.6 1.2 1.0 1.0 0.4
N 5.3 3.7 3.7 3.0 0.7
O 18.4 21.3 15.2 16.4 1.4
Ash 40.0 49.0 55.8 55.8 6.2
ICP/AES (g/kg)
Pb 0.033  0.002 0.029  0.003 0.028  0.001 0.030  0.001 e 0.030  0.001 0.033  0.001
Co 0.080  0.003 0.083  0.001 0.070  0.004 0.092  0.001 e 0.082  0.001 0.101  0.001
Cu 0.440  0.006 0.410  0.015 0.400  0.056 0.414  0.001 e 0.412  0.003 0.400  0.008
Cr 0.280  0.005 0.293  0.001 0.383  0.015 0.340  0.007 e 0.330  0.002 0.411  0.003
Mn 0.150  0.005 0.142  0.001 0.121  0.005 0.153  0.002 e 0.030  0.001 0.040  0.001
Ni 0.180  0.005 0.180  0.001 0.240  0.008 0.221  0.001 e 0.182  0.001 0.220  0.001
Ba 0.230  0.001 0.200  0.002 0.212  0.004 0.230  0.001 e 0.100  0.003 0.060  0.001
Zn 3.4  0.10 2.7  0.16 2.1  0.02 2.3  0.04 e 0.531  0.015 1.5  0.05
Al 15.3  0.16 12.0  0.09 9.3  0.16 9.9  0.89 7.6  0.32 5.4  0.01 6.1  0.76
Fe 16.3  0.54 14.7  0.16 14.0  0.14 15.5  0.02 3.5  0.31 9.0  0.14 8.7  0.58
Na 9.6  0.12 6.6  0.31 10.0  0.33 9.7  0.51 0.660  0.027 0.331  0.050 0.530  0.026
Mg 15.3  0.21 11.5  0.14 9.9  0.36 9.6  0.32 0.322  0.007 0.450  0.026 1.0  0.01
K 3.7  0.11 3.37  0.08 18.5  0.34 18.8  0.51 0.904  0.003 0.900  0.001 2.1  0.08
P 14.3  0.36 14.5  0.02 54.5  0.93 49.6  2.42 e e 0.790  0.047
Ca 190  2.18 196  4.75 192  0.49 197  4.30 0.400  0.017 4.1  0.06 2.8  0.62
Si 211  2.97 226  1.14 230  4.25 239  2.94 127  1.22 21.1  0.97 42.6  1.53

‘e‘: below detection limit.

FS_0.01, high amounts of Ca and Si are found (Table 2), which presence of more silicate functional groups. These hypotheses
supports the FTIR results. For FS_1 adsorbents, high amounts are in agreement with the ICP-AES results e i.e. higher P, K and
of Ca and Si are leached as a result of high HCl treatment (Table Si contents in SSFC_0.01 and FSFC_0.01 adsorbents (Table 2) e
2), but carbonate and silicate groups still remain. and with the adsorption isotherm results which will be dis-
Compared with the sludge derived adsorbents, SSFC_0.01 cussed in 3.4.2 (i.e. the significant release of e.g. phosphate
and FSFC_0.01 adsorbents have less pronounced absorption groups from sludge/FC adsorbents). FTIR spectra of FSFC_1
peaks at 1420 cm1 and stronger absorption peaks at adsorbents show less pronounced absorption peaks than
1040 cm1 (w1034 cm1) and 564 cm1, which can be those of FSFC_0.01 and SSFC_0.01 adsorbents, indicating that
explained in terms of less carbonate groups on their surface, high amounts of functional groups like carbonate, silicate and
but more phosphate groups associated with e.g. calcium or phosphate groups, are removed through HCl treatment. Only
potassium. The stronger absorption peak at 1040 cm1, also weak absorption peaks associated with alumino-silicates are
associated with SieOeSi structures, also indicates the found for the commercial F400 adsorbents. Both FTIR results
of FSFC_1 and F400 adsorbents are again in agreement with
the ICP-AES results (Table 2).
Boehm titration identifies the surface charge for the
Table 3 e FTIR assignment. adsorbents in an effort to quantify the amount of various
Functionality Assignment Wavenumber oxygen containing groups on the adsorbent surface. The
(cm1) results demonstrate the negligible presence of acidic func-
tional groups on these surfaces: 0.262 (SS_0.01), 0.348 (FS_0.01),
Carbonates CO2
3 out of plane 875
0.272 (SSFC_0.01), 0.367 (FSFC_0.01), 0.603 (FSFC_1) and 1.018
(CaCO3) deformation (n2)
Asymmetric CeO 1420 (FS_1) mmol Hþ/g. The acidic surface charges are slightly
stretching (n3) higher than those of F400 (0.186 mmol Hþ/g), but are still
OeCeO bending (in plane 710 negligible. The results confirm the FTIR data, since no
deformation (n4)) absorption peaks associated with acidic functional groups are
Combination bands 2500e1790 detected. 0.01 N HCl post-treated adsorbents are basic (pH of
(overtones)
8.1 (SS_0.01), 7.2 (FS_0.01), 7.8 (SSFC_0.01) and 7.6 (FSFC_0.01)),
Alumino- OeH stretching 3300e3500
while FS_1 and FSFC_1 adsorbents are rather acidic (pH of 4.7
silicates OeH bending 1600e1650
SieOeSi bending 1040 for both adsorbents).
SieOeSi asymmetric 460
stretching
3.3. Porosity
Phosphates PO3
4 deformation 1034
vibrations (n1 and n3)
PeO and OePeO stretching 564 SSFC_0.01 and FSFC_0.01 adsorbents obtain higher SBET and VT
values (60 m2/g and 0.10 cm3/g respectively) than SS_0.01 and
2788 w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 7 8 3 e2 7 9 4

FS_0.01 adsorbents (15 m2/g and 0.02 cm3/g respectively). and isotherm adsorption experiments are performed at pH
Additionally, treatment with 1 N HCl further improves the BET of 5.0.
surface area and porosity only for FSFC_1 adsorbents (257 m2/ Kinetic investigation of Cu2þ and Zn2þ shows that after
g and 0.30 cm3/g respectively); no improvement is found for 48 h, a steady state is reached with 0.01 N HCl post-treated
FS_1 adsorbents. Although adding FC to sludge and HCl post- adsorbents and commercial F400 adsorbents. With SS_0.01
treatment of the sludge/FC derived adsorbents improve BET and F400 adsorbents, a fast increase is found within the first
surface area and porosity of the obtained adsorbents, they are 4 h and equilibrium is slowly reached for both Cu2þ and Zn2þ
still negligible compared with those of F400 adsorbents adsorption. A fast increase is also observed with FS_0.01,
(1137 m2/g and 0.57 cm3/g respectively). SSFC_0.01 and FSFC_0.01 adsorbents for Zn2þ adsorption. For
SEM surface data show that all obtained 0.01 N HCl post- Cu2þ adsorption, equilibrium is slowly reached with these
treated adsorbents have an irregular surface with large adsorbents. Looking at heavy metal adsorption capacity, these
diameter pores, craters and inorganic matter. Craters are adsorbents have higher heavy metal adsorption capacity than
defined as not fully open cavities which can open and form SS_0.01 and F400 adsorbents. All examined adsorbents,
pores when the adsorbents are post-treated with 1 N HCl. The including F400 adsorbents, reveal a slightly higher Zn2þ
surface of F400 adsorbents is more intact, without a lot of adsorption capacity than for Cu2þ. During kinetic studies,
macropores. SEM cross-section data confirm these conclu- leaching of heavy metals present in adsorbents is not found in
sions i.e. the derived adsorbents are highly macroporous and detectable amounts in the aqueous solutions.
significant differences are not observed between the different Based on the correlation coefficient between the different
adsorbents except for inorganic matter content. Examples of kinetic models and the obtained results, only the pseudo-
SEM images are given in Fig. 2aec i.e. SEM surface image of second order kinetics model is suitable to fit the results
FS_0.01 adsorbent; and SEM cross-sectional images of FS_0.01 (R2 > 0.99, Table 4). The pseudo-second order model assumes
and FS_1 adsorbent. that the rate-limiting step is a chemical sorption between the
adsorbate and adsorbent (chemisorptions) (Ho and McKay,
3.4. Aqueous adsorption studies 1998; Mohan et al., 2007; Li et al., 2010). The reaction rate of
the pseudo-second order kinetics (k2) is small, indicating that
3.4.1. Adsorption kinetics the adsorption is rather a slow surface chemisorption process.
Bulk precipitation of copper or zinc (as copper or zinc oxides)
occurs when the equilibrium pH of water is above 6 or 3.4.2. Adsorption isotherm
between 7 and 11 respectively. To avoid bulk precipitation Adsorption isotherm experiments are performed at pH of 5.0
and to exclude pH influences on heavy metal removal, kinetic for a period of 48 h, reaching equilibrium according to general

Fig. 2 e SEM surface images of (a) FS_0.01 adsorbent; and SEM cross-sectional images of (b) FS_0.01 adsorbent and (c) FS_1
adsorbent. Bright spots indicate inorganic matter.
 w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 7 8 3 e2 7 9 4 2789

SSFC_0.01 > FS_0.01 > FSFC_0.01 >> SS_0.01 > F400

Table 4 e Results of adsorption kinetics.
The monolayer adsorption capacity qm for Zn2þ follows the
Metal Conditie C0 qe,exp Pseudo-second order order:
(mg/l) (mg/g) kinetics
FSFC_0.01 > FS_0.01 > SSFC_0.01 >> F400 > SS_0.01
k2 qe,cal R2 Based on Freundlich and/or Langmuir isotherm constants
(g/(mg min)) (mg/g) for Cu2þ and Zn2þ adsorption, similar conclusions can be
Cu2þ SS_0.01 38 2.3 8.00E-03 2.3 0.9991 made: FSFC_0.01, SSFC_0.01 and FS_0.01 adsorbents show
FS_0.01 38 20.1 1.60E-04 21.7 0.9853 higher values than SS_0.01 and F400 adsorbents.
SSFC_0.01 38 17.9 4.80E-04 18.6 0.9963 Generally, it can be concluded that the pyrolysis process
FSFC_0.01 38 19.9 5.20E-04 20.3 0.9982 itself is an important factor for the efficiency of heavy metal
F400 38 2.2 5.40E-03 2.3 0.9987 removal. 0.01 N HCl post-treated fast pyrolysis adsorbents
Zn2þ SS_0.01 25 5.5 7.10E-03 5.5 0.9985
show higher Cu2þ and Zn2þ removal than 0.01 N HCl post-
FS_0.01 25 17.6 4.10E-04 17.7 0.9939
treated slow pyrolysis adsorbents. Moreover, adding FC to the
SSFC_0.01 25 12.8 2.60E-03 12.8 0.9995
FSFC_0.01 25 15.8 1.00E-03 16 0.9977 sludge before carbonization/pyrolysis will improve aqueous
F400 25 6.1 1.60E-03 6.9 0.9979 Cu2þ and Zn2þ removal, especially in the case of slow pyrolysis
derived adsorbents. In the case of fast pyrolysis, the difference
is less pronounced. Except for SS_0.01 adsorbents, the
obtained adsorbents show much better Cu2þ and Zn2þ
kinetic studies (see 3.4.1). The experimental data are analyzed removal than the commercial F400.
with Langmuir, Freundlich and LangmuireFreundlich equa- The mechanism of metal removal can give an explanation
tions and presented in Table 5. The data of 1N HCl post- of above obtained results. As the experiments are performed
treated adsorbents will be discussed in Section 3.4.3. The under pH controlled conditions (pH ¼ 5.0), bulk precipitation
highest regression coefficient R2 is mostly obtained from can be excluded as metal removal mechanism. According to
LangmuireFreundlich model. However, the pure Langmuir or kinetic studies, metal removal is based on chemisorption
Freundlich models are also found to fit the data well, indicating processes. In this study, the important removal mechanism is
a complex adsorption mechanism ruled by different attraction in favor of a cation exchange mechanism. A minor release of
circumstances as e.g. exchanges and surface precipitations. Ca2þ and a significant release of Kþ from F400 and SS_0.01
The parameter n (LangmuireFreundlich model) indicates adsorbents occur during uptake of Cu2þ. Significant amounts
that SSFC_0.01 adsorbents have a more heterogeneous surface of both Ca2þ and Kþ and a minor amount of Mg2þ are released
(n < 0.5) in comparison with the other adsorbents within Cu2þ from FS_0.01, SSFC_0.01 and FSFC_0.01 adsorbents, which
adsorption experiments. The same is observed for FSFC_0.01 results in a higher Cu2þ uptake than SS_0.01 and F400 adsor-
and SSFC_0.01 adsorbents within Zn2þ adsorption experi- bents. The cation exchange during the adsorption of Cu2þ
ments. The n values higher than 0.5 obtained for SS_0.01, (equimolar concentration) on FSFC_0.01 adsorbents is pre-
FS_0.01 and F400 adsorbents within both Cu2þ and Zn2þ sented in Fig. 3 as an example, other figures are omitted for
adsorption experiments, and for FSFC_0.01 adsorbents within brevity. Besides the above described ions, no release of other
Cu2þ adsorption experiments, suggest that the adsorption metals, i.e. hazardous metals, is observed.
sites of the homogeneous surface can be considered similar in The low cation exchange capacity and thus the low Cu2þ
their heavy metal adsorption capability (wLangmuir model). uptake of F400 adsorbents can be explained by the low content
According to LangmuireFreundlich model, the monolayer of inorganic material i.e. exchangeable metals (Table 2, SEM
adsorption capacity qm for Cu2þ follows the order: data). Slow pyrolysis and fast pyrolysis adsorbents have

Table 5 e Langmuir, Freundlich and LangmuireFreundlich isotherm constants for Cu2D and Zn2D adsorption on different
adsorbents.
Metal Langmuir Freundlich LangmuireFreunlich
2 2
qm (mg/g) KL (l/mg) R KF (mg/g) n R qm (mg/g) KLF (l/mg) n R2

SS_0.01 Cu2þ 3.95 0.12 0.939 0.62 1.97 0.947 3.95 0.12 1.00 0.939
FS_0.01 25.64 0.19 0.949 6.08 2.69 0.981 25.70 0.19 0.99 0.989
SSFC_0.01 20 0.38 0.988 7.38 3.64 0.967 41.66 0.02 0.37 0.989
FSFC_0.01 21.28 1.15 0.998 8.93 3.44 0.905 21.78 1.10 0.82 0.998
F400 2.95 0.08 0.960 0.29 1.56 0.963 2.95 0.08 1.00 0.960
SS_0.01 Zn2þ 6.71 0.15 0.977 1.39 2.27 0.978 9.1 0.08 0.74 0.980
FS_0.01 28.57 0.15 0.995 4.41 1.84 0.976 44.24 0.05 0.75 0.998
SSFC_0.01 21.74 0.28 0.925 8.17 4.37 0.953 38.17 0.22 0.34 0.971
FSFC_0.01 26.32 0.23 0.965 6.59 2.68 0.993 68.96 0.1 0.47 0.919
F400 11.24 0.07 0.968 1.06 1.68 0.987 18.62 0.02 0.76 0.991
FS_1 Zn2þ 24.39 1.24 0.991 15.10 6.99 0.931 24.39 1.54 0.89 0.997
FSFC_1 5.81 0.78 0.933 2.22 3.24 0.840 6.45 0.46 0.85 0.933
2790 w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 7 8 3 e2 7 9 4

groups (XSi-O þ 2Hþ XSi-OH þ Hþ XSi-OHþ2 ) (Seredych

and Bandosz, 2006). After addition of diluted HCl, hydroxides
generated will be neutralised and more metal oxides could
dissolve into the water. Secondly, it is possible that cation
exchange occurs with Hþ ions instead of Cu2þ or Zn2þ ions.
The total amount of cations released is thus higher than the
amount of Cu2þ or Zn2þ adsorbed, indicating that the cation
exchange mechanism is complex and that there could be
a competition between the different cations for adsorption.
Besides cation exchange mechanism, metal precipitation at
the surface and physical adsorption in pores could be other
metal removal mechanisms. Since all derived washed adsor-
bents are weak alkaline and porosity is low, the contribution
Fig. 3 e Cation exchange during the adsorption of Cu2D on of these mechanism is rather limited. Metal complex forma-
FSFC_0.01 adsorbents. tion with e.g. silicon oxide groups at the carbon surface could
also play a role. But due to addition of diluted HCl, Hþ ions
could compete with Cu2þ or Zn2þ ions for binding with the
silicon oxide sites, reducing the Cu2þ or Zn2þ adsorption. In
similar chemical compositions (Table 2) and surface features conclusion, it could be stated that cation exchange mecha-
(silicate and carbonate functional groups), but it seems that nism dominates for these adsorbents. This is in agreement
during slow pyrolysis, metals and surface functional groups with the findings of other authors as reviewed by Smith et al.
are probably more encapsulated within the slow pyrolysis (2009).
adsorbents. This will prevent Mg2þ release and decreases Ca2þ
release (e.g. silicate and carbonate groups) from SS_0.01 3.4.3. The influence of HCl treatment
adsorbents, resulting in a lower Cu2þ uptake compared with The fast pyrolysis adsorbents are post-treated with 1 N HCl
FS_0.01 adsorbents. Adding FC to the sludge before pyrolysis and the obtained adsorption isotherm data (pH ¼ 5.0) are
enhances e.g. Ca, K and P amounts in the obtained carbons compared with those of FS_0.01 and FSFC_0.01 adsorbents
(Table 2) and phosphate functional groups on the surface are (Table 5). During isotherm experiments with these 1 N HCl
formed. This results in a significant higher Ca2þ release, post-treated adsorbents, no release of hazardous heavy
a slightly higher Kþ release and a release of phosphates from metals is observed.
the sludge/FC derived adsorbents compared with the sludge Again, the LangmuireFreundlich model fits the experi-
derived adsorbents. Higher Cu2þ uptake is thus obtained on mental data of 1 N HCl post-treated adsorbents best. The
SSFC_0.01 adsorbents compared with SS_0.01 adsorbents. surface heterogeneity of these adsorbents, determined by
However for FSFC_0.01 adsorbents, the Cu2þ uptake is lower parameter n, is lower than those of 0.01 N HCl post-treated
than for FS_0.01 adsorbents, although the total amount of adsorbents, which is probably due to higher removal of inor-
cations released from FSFC_0.01 adsorbents is higher. A ganic matter (Table 2). This suggests that the adsorption sites
possible explanation could be the competition between the of the homogeneous surface can be considered similar in their
released and adsorbed ions, e.g. the effect of the ionic Zn2þ adsorption capability.
strength, resulting in a lower Cu2þ uptake on the FSFC_0.01 For FS_0.01 and FSFC_0.01 adsorbents, the qm values of
adsorbents. The effect of the ionic strength could also give an LangmuireFreundlich model are quite high, and not compa-
explanation for subsequent observation: the amounts of Ca2þ, rable with those of Langmuir model. Therefore, to compare
Kþ and phosphates released from FSFC_0.01 adsorbents are 0.01 N HCl with 1 N HCl post-treated adsorbents, not only the
higher than SSFC_0.01 adsorbents (wencapsulation effect), isotherm constants of the Langmuir and LangmuireFreund-
but the Cu2þ uptake on FSFC_0.01 adsorbents is lower. Similar lich model should be taken into consideration, but also those
trends are observed for Zn2þ uptake, but the effect of ionic of Freundlich model. Moreover, all three models are found to
strength on Zn2þ adsorption is less pronounced: higher Zn2þ fit the experimental data well. Based on the three models, it
adsorption is thus related with higher amounts of cations can be concluded that treatment with higher concentration of
released or exchanged. The consistent higher Zn2þ HCl enhances Zn2þ adsorption only for FS adsorbents. For
uptake compared with Cu2þ uptake on the same adsorbents is FSFC adsorbents, Zn2þ adsorption is decreased when adsor-
due to the consistent higher Ca2þ exchange or release. bents are treated with higher concentrations of HCl.
For all adsorbents studied, the total amount of cations During treatment with 1 N HCl, more inorganic material is
released (Fig. 3, sum cations released) is higher than the removed from these adsorbents (Table 2, SEM data), resulting
amount of Cu2þ or Zn2þ adsorbed. This phenomenon can be in a lower cation exchange (Ca2þ and ½ Kþ) compared with
explained by the addition of diluted HCl during isotherm 0.01 N HCl post-treated adsorbents (Ca2þ, ½ Kþ and Mg2þ).
studies to obtain a solution pH of 5.0. An increase of solution Secondly, the total amount of cations released from FSFC_1
pH can be caused by the dissolution of earth alkaline or alkali adsorbents is much lower than those of FS_1 adsorbents
metal oxides like calcium, sodium or magnesium oxides (Fig. 4), indicating that cation exchange of Ca2þ and Kþ is more
present on adsorbents into water generating hydroxides pronounced for FS_1 adsorbents. FTIR data confirms a reduc-
(Mnþ
m O þ H2O / mM
nþ
þ 2OH), or by the adsorption of Hþ tion of surface functional groups (carbonate and phosphate
ions present in water on the silicon or aluminium oxide groups) for FSFC_1 adsorbents, suggesting that the possible
 w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 7 8 3 e2 7 9 4 2791

capacity increased from 20.41 (5  C) to 34.48 (45  C) mg/g and

from 16.13 (5  C) to 29.42 (45  C) mg/g for FSFC_0.01 and FS_1
adsorbents, respectively. Zn2þ adsorption capacity of the
adsorbents increases thus with enhanced temperature.
In order to describe thermodynamic behavior of the
adsorption of metal ions on adsorbents, the thermodynamic
parameters including the Gibbs free energy (ΔG ), enthalpy
(ΔH ) and entropy (ΔS ) parameters are calculated from equa-
tions (Sari et al., 2008):

DG+ ¼ RT ln Kd

DG+ ¼ DH+  TDS+

2D
Fig. 4 e Cation exchange during the adsorption of Zn on
where R is the universal gas constant, T the temperature and
FS_1 and FSFC_1 adsorbents. The sum of cations released
Kd (qe/ce) is the distribution coefficient.
are: Ca2D D ½ KD.
The Gibbs free energies are all negative within the different
conditions tested (the adsorption of Cu2þ and Zn2þ on
different adsorbents at 20  C, the adsorption of Zn2þ on
involvement in cation exchange is rather limited. Thirdly, the FSFC_0.01 and FS_1 adsorbents at different temperatures),
total amount of cations released (Ca2þ þ ½ Kþ) is much lower indicating that the adsorption is thermodynamically feasible
than the Zn2þ adsorption for 1 N HCl post-treated adsorbents, and spontaneous. The initial Zn2þ and Cu2þ concentration has
especially for FS_1 adsorbents (Fig. 4). For FS_1 adsorbents, an impact on the thermodynamic behavior of the adsorption
another adsorption mechanism than the above described process: a minimum ΔG is obtained for the lowest initial
cation exchange dominates, which is less pronounced for concentration, while a maximum ΔG is obtained for the
FSFC_1 adsorbents. This results in a higher Zn2þ adsorption highest initial concentrations. Fig. 5 shows the ΔG within the
for FS_1 adsorbents compared with FSFC_1 adsorbent. Zn2þ adsorption conditions. The adsorption of Zn2þ on FS_1
Compared with FSFC_0.01 adsorbents, a lower Zn2þ adsorp- adsorbents is thermodynamically the most feasible. The same
tion for FSFC_1 adsorbents is found. trends are observed within the temperature conditions: at
As described in literature, HCl treatment can improve BET each temperature, the minimum ΔG is obtained for the
surface areas and porosity (Smith et al., 2009), which might lowest initial concentration and vice versa.
result in higher metal adsorption. Improved BET surface area ΔG decreases with increasing temperature at each
and porosity is found for FSFC_1 adsorbents compared with concentration, as presented in Fig. 6 for Zn2þ adsorption on
FSFC_0.01 adsorbents. No improvement is found for FS_1 FSFC_0.01 adsorbents. The same is found for FS_1 adsorbents.
adsorbents (see Porosity data). It can be concluded that BET Zn2þ adsorption is thus more feasible at higher temperature,
surface and porosity are not directly correlated with Zn2þ which is in agreement with the isotherm experiments. The
uptake. enthalpy change decreases from 28.2 to 14.4 kJ/mol and from
Mendez and Gasco (2005) showed enhanced Naþ, Ca2þ and 52.8 to 15.0 kJ/mol, while the entropy change decreases from
2þ
Mg uptake of HNO3 washed carbonized sludge due to 0.20 to 0.09 kJ/(mol K) and from 0.25 to 0.10 kJ/(mol K) with
substitution of exchangeable cations in char with Hþ ions, increasing initial Zn2þ concentration from 2 to 38 mg/l Zn2þ
which boost the char’s cation capacity. During adsorption for FS_1 and FSFC_0.01 adsorbents respectively. The positive
experiment with FS_1 adsorbents, NaOH has to be added enthalpy values indicate the endothermic nature of the
constantly to the Zn2þ solution to maintain the desired pH of process. The decrease of the positive entropy values suggests
5.0, probably due to the release of Hþ ions. This is not observed
with FSFC_1 adsorbents, suggesting that the exchangeable
cations are only significantly substituted with Hþ ions in the
case of FS_1 adsorbents. It can be concluded that the addi-
tional adsorption mechanism, especially for FS_1 adsorbents,
is probably a cation exchange with Hþ ions present on the
surface. Secondly, besides the dominant cation exchange,
surface complex formation could be enhanced for FS_1
adsorbents: the silicon oxide sites could be more accessible
after acid treatment. The involvement of this mechanism will
be minimal for FSFC_1 adsorbents (FTIR data).

3.4.4. Thermodynamics
Zn2þ isotherm experiments are performed at different
temperatures (5, 20, 45  C) with FSFC_0.01 and FS_1 adsor-
bents. The highest regression coefficient is obtained from
Freundlich model and LangmuireFreundlich model for Fig. 5 e Zn2D adsorption thermodynamics of different
FSFC_0.01 and FS_1 adsorbents respectively. The adsorption adsorbents.
2792 w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 7 8 3 e2 7 9 4

adsorbents, a release of Ca2þ, Kþ and phosphates is observed.

Leaching of other metals, i.e. hazardous metals, is not detec-
ted. As already formulated in the isotherm studies, there is
a correlation between Zn2þ adsorption and the release of the
considered cations: e.g. FSFC_0.01 adsorbents obtain the
highest Zn2þ adsorption, because the highest amount of
cations is released from these adsorbents. For F400 adsor-
bents, the total amount of cations released equals the amount
of Zn2þ adsorbed (adsorption curve and total cations released
curve coincide completely). Cation exchange will thus be the
only metal removal mechanism for F400 adsorbents. For
SS_0.01 adsorbents, the total amount of cations released
equals the amount of Zn2þ adsorbed during the first 20 min.
Fig. 6 e Zn2D adsorption thermodynamics of FSFC_0.01
From 20 min until 47 min, there is still a low release of Ca2þ
adsorbents.
and Kþ, while no increase of adsorbed Zn2þ amounts is
observed. Except for the first effluent fraction of SSFC_0.01 and
FSFC_0.01 adsorbents, which contains high Kþ amounts, the
a reduction in the randomness at the solid/solution interface. total amount of cations released equals the amount of Zn2þ
In general, increasing the temperature could enhance Zn2þ adsorbed at the start of the experiment for FS_0.01, SSFC_0.01
removal, especially at low initial concentrations (2 mg/l), but and FSFC_0.01 adsorbents. This is followed by a period
also at higher initial concentrations (38 mg/l). whereas the total amount of cations released is lower than the
amount of Zn2þ adsorbed, indicating that another mechanism
3.4.5. Dynamic adsorption and desorption than cation exchange is responsible for Zn2þ adsorption. Then
Dynamic adsorption experiments are performed with all 0.01 the total amount of cations released will be higher than the
N HCl post-treated adsorbents, while desorption experiments amount of Zn2þ adsorbed, finally resulting in a short period
are only performed with FSFC_0.01 and FS_0.01 adsorbents. where a low release of Ca2þ and Kþ is observed while Zn2þ
During adsorption studies, SS_0.01, FS_0.01, F400 adsor- adsorption is terminated. It seems that after a certain time,
bents differ from SSFC_0.01 & FSFC_0.01 adsorbents. For there is a competition between the release and the adsorption
SS_0.01, F400 and FS_0.01 adsorbents, the maximum adsorp- of different cations, finally resulting in a stabilisation of the
tion capacity is relatively quickly reached (21 min, 60 min and adsorbed Zn2þ amount. For all studied adsorbents, surface
89 min, respectively) with the adsorption capacity being 0.35, precipitation could be the additional metal removal mecha-
1.31 and 2.65 mg Zn2þ/g, respectively. For SSFC_0.01 and nism as all adsorbents are basic and zinc hydroxide could
FSFC_0.01 adsorbents, the adsorption capacity increases precipitate at the alkaline surface. Moreover, Zn2þ could form
quickly in the first 50 min and then slows down, until equi- a complex with e.g. silicate functional groups present on the
librium is reached (189 min and 207 min, respectively) with surface. This is probably true for FS_0.01, SSFC_0.01 and
the adsorption capacity being 5.79 and 10.5 mg Zn2þ/g, FSFC_0.01 adsorbents during a certain time, as discussed
respectively. Fig. 7a and b presents the adsorption of Zn2þ, the above.
total amount of cations released (Ca2þ þ ½ Kþ) and the pH For SSFC_0.01 and FSFC_0.01 adsorbents, bulk precipitation
measurements of the effluent fractions of FS_0.01 and of Zn2þ could be an additional metal removal mechanism as
FSFC_0.01 adsorbents. Other figures are omitted for brevity. the pH of the effluent fractions of these adsorbents are more
From adsorption isotherm experiments, it can be stated basic than those of SS_0.01, FS_0.01 and F400 adsorbents. The
that Zn2þ removal can occur by cation exchange. A release of initial pH of the Zn2þ solution is 5.4, without adsorbents. For
Ca2þ and Kþ from SS_0.01, FS_0.01 and F400 adsorbents is SSFC_0.01 and FSFC_0.01 adsorbents, the pH of the effluent is
indeed observed. While for SSFC_0.01 and FSFC_0.01 changed to 7.5 at the beginning of dynamic adsorption

Fig. 7 e The adsorption of zinc, the total amount of cations released (Ca2D D ½ KD) and pH measurements of the effluent
fractions of the dynamic adsorption experiments with (a) FS_0.01 adsorbents and (b) FSFC_0.01 adsorbents.
 w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 7 8 3 e2 7 9 4 2793

experiments probably caused by the dissolution of K2O first cycle, and 96.4% and 92% for the second and third cycle
generating OH and Kþ, and decreased to 5.95 at the end of the respectively.
experiment. During the first 20 min, these effluents are light In conclusion, the reuse of the adsorbents in adsorp-
alkaline (pH > 7) and zinc hydroxide can be formed and tionedesorption cycles is only possible for FS_0.01 adsorbents.
precipitated. This metal removal effect could not be measured
as the effluent fractions are made acidic before ICP/AES
measurement. 4. Conclusions
Adsorptionedesorption cycles of Zn2þ are presented in
Fig. 8 for FS_0.01 and FSFC_0.01 adsorbents.  This study focuses on the adsorption of Cu2þ and Zn2þ from
For FS_0.01 adsorbents, 2.65 mg/g is adsorbed within aqueous solution (at pH ¼ 5.0) on low cost adsorbents
89 min during the first cycle, as described above. 2.53 mg/g of produced by slow and fast pyrolysis of industrial waste
the 2.65 mg/g could be removed within 30 min through 1 N HCl sludge or sludge/FC mix, post-treated with 0.01 or 1 N HCl.
and thus the desorption efficiency is 95.5%. Other inorganic  It is found that the adsorption follows the pseudo-second
material is also removed during this step. During the second order kinetics model well and steady state is reached within
adsorption cycle, 7.23 mg/g can be adsorbed within 130 min, 48 h. LangmuireFreundlich isotherm model fits the data
as a result of cation exchange with Kþ and especially Hþ ions. best. All obtained adsorbents, except SS_0.01 and FSFC_1
The effluent pH decreases significantly, which proves the adsorbents, show better metal removal capacity than the
exchange of Hþ ions. Surface complex formation also occurs, commercial F400.
as the total amount of cations released (½ Kþ and ½ Hþ) is  Cation exchange mechanism is found to be the dominant
lower than the Zn2þ amount adsorbed. The desorption effi- heavy metal removal mechanism. Adding FC to the sludge
ciency during the second desorption cycle of 20 min is only enhances adsorption capacity, because higher amounts of
49%: 3.25 mg/g of the 7.23 mg/g could be desorbed and thus exchangeable ions e.g. Ca2þ ions, Kþ ions and phosphates
3.98 mg/g is left on the adsorbent. 3.47 mg/g is adsorbed within are obtained. During slow pyrolysis process, exchangeable
45 min during the third adsorption cycle, as a result of cation ions are more encapsulated within the obtained adsorbents,
exchange (½ Kþ and ½ Hþ) and as a result of surface complex reducing the cation exchange capacity.
formation. The same amount of Kþ and reduced amounts of  Improvement of Zn2þ removal via 1 N HCl post-treatment is
Hþ are exchanged in comparison with the second adsorption only effective for FS adsorbents due to substitution of
cycle, resulting in reduced Zn2þ adsorption. Moreover, surface exchangeable cations with Hþ ions, which boost the cation
complex formation is also reduced due to blocking of the sites exchange capacity.
in an irreversible way during the second cycle. The desorption  Zn2þ adsorption on adsorbents is an endothermic and
efficiency during the third desorption cycle of 20 min is again spontaneous process at 5e45  C. Zn2þ adsorption capacity
94.5%: 3.28 mg/g of the 3.47 mg/g is desorbed and 4.17 mg/g is increases with enhanced temperature.
left on the adsorbent.  The reuse of the adsorbents in adsorptionedesorption
For FSFC_0.01 adsorbents, the adsorption capacity is cycles is only possible for FS_0.01 adsorbents.
reduced significantly and systematically during second and  No release of hazardous heavy metal present on adsorbents
third adsorption cycle. During the second adsorption cycle, is observed during different adsorption experiments.
a release of Ca2þ, Kþ and phosphates and a minimal release of  In conclusion, it can be stated that the production of
Hþ are observed. The involvement of surface complex adsorbent of sludge and especially of sludge/FC mix can be
formation mechanism is minimal (the reduction of surface useful in wastewater treatment for heavy metal removal.
functional groups, FTIR data), resulting in a lower Zn2þ
adsorption compared with FS_0.01 adsorbents. During the first
desorption cycle, high amounts of inorganic material Acknowledgements
including Zn2þ present on the adsorbents, are removed.
Nevertheless, the desorption efficiency is still 98.5% for the The authors wish to thank M. Vanhamel for her help with
FTIR, the ‘Department of polymer and carbonaceous matrial,
Wroclaw University of Technology, Poland’ for the porosity
analyses and BASF (Antwerp, Belgium) for the supply of sludge
and disposal filter cake samples.

references

Ahmaruzzaman, M., 2011. Industrial wastes as low-cost potential

adsorbents for the treatment of wastewater laden with heavy
metals. Advances in Colloid and Interface Science 166, 39e59.
Bhatnagar, A., Minocha, A.K., Pudasainee, D., Chung, H.K.,
Kim, S.H., Kim, H.S., Lee, G., Min, B., Jeon, B.H., 2008.
Vanadium removal from water by waste metal sludge and
cement immobilization. Chemical Engineering Journal 144 (2),
Fig. 8 e Adsorptionedesorption cycles of Zn2D. 197e204.
2794 w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 7 8 3 e2 7 9 4

Boehm, H.P., 2002. Surface oxides on carbon and their analysis: during bio-oil production. Journal of Colloid and Interface
a critical assessment. Carbon 40, 145e149. Science 310, 57e73.
Bouzid, J., Elouear, Z., Ksibi, M., Feki, A., Montiel, A., 2008. A study Rio, S., Le Coq, L., Faur, C., Le Cloirec, P., 2006a. Production of
on removal characteristics of copper from aqueous solution by porous carbonaceous adsorbent from physical activation of
sewage sludge and pomace ashes. Journal of Hazardous sewage sludge: application to wastewater treatment. Water
Materials 152 (2), 838e845. Science and Technology 53 (3), 237e244.
Freundlich, H., Helle, W.J., 1939. Ubber die adsorption in Sari, A., Mendil, D., Tuzen, M., Soylak, M., 2008. Biosorption of
lusungen. Journal of the American Chemical Society 61, 2e28. Cd(II) and Cr(III) from aqueous solution by moss (Hylocomium
Hammaini, A., Gonzales, F., Ballester, A., Blazquez, M.L., splendens) biomass: equilibrium, kinetic and thermodynamic
Munoz, J.A., 2007. Biosorption of heavy metals by activated studies. Chemical Engineering Journal 144, 1e9.
sludge and their desorption characteristics. Journal of Seredych, M., Bandosz, T.J., 2006. Removal of copper on
Environmental Management 84, 419e426. composite sewage sludge/industrial sludge-based adsorbents:
Ho, Y.S., McKay, G., 1998. A comparison of chemisorptions kinetic the role of surface chemistry. Journal of Colloid and Interface
models applied to pollutant removal on various sorbents. Science 302 (2), 379e388.
Transactions of the Institutions of Chemical Engineers 76, Smets, K., Adriaensens, P., Reggers, G., Schreurs, S., Carleer, R.,
332e340. Yperman, J., 2011. Flash pyrolysis of rapeseed cake: influence
Kang, H.Y., Park, S.S., Rim, Y.S., 2006. Preparation of activated of temperature on the yield and the characteristics of the
carbon from paper mill sludge by KOH-activation. Korean pyrolysis liquid. Journal of Analytical and Applied Pyrolysis 90
Journal of Chemical Engineering 23 (6), 948e953. (2), 118e125.
Langmuir, I., 1918. The adsorption of gases in plain surfaces of Smith, K.M., Fowler, G.D., Pullket, S., Graham, N.J.D., 2009. Sewage
glass, mica, and platinum. Journal of the American Chemical sludge-based adsorbents: a review of their production,
Society 40, 1361e1403. properties and use in water treatment applications. Water
Li, Z., Tang, Q., Katsumi, T., Tang, X., Inui, T., Imaizumi, S., 2010. Research 43, 2569e2594.
Leaf char: an alternative adsorbent for Cr(III). Desalination Vlarem II, annex 2.3.1.
264, 70e77. Wang, X.J., Xu, X.M., Liang, X., Wang, Y., Liu, M., Wang, X.,
Lima, I.M., Marschall, W.E., 2005. Granular activated carbons from Xia, S.Q., Zhao, J.F., Yin, D.Q., Zhang, Y.L., 2011. Adsorption of
broiler manure: physical, chemical and adsorptive properties. copper(II) onto sewage sludge-derived materials via
Bioresource Technology 96, 699e706. microwave irradiation. Journal of Hazardous Materials 192,
Marczewski, A.W., Derylo-Marczewska, A., Jaroniec, M., 1988. 1226e1233.
Correlations of heterogeneity parameters for single-solute and Williams, P.T., Slaney, E.J., 2007. Analysis of products from the
multi-solute adsorption from dilute solutions. Journal of the pyrolysis and liquefaction of single plastics and waste plastic
Chemical Society Faraday Transactions 84, 2951e2957. mixtures. Resources, Conservation and Recycling 51,
Mendez, A., Gasco, G., 2005. Optimization of water desalination 754e769.
using carbon-based adsorbents. Desalination 183 (1e3), Yin, P., Xu, Q., Qu, F., Zhao, G., Sun, Y., 2010. Adsorption of
249e255. transition metal ions from aqueous solutions onto a novel
Mendez, A., Gasco, G., Freita, M.M.A., Siebielec, G., Stuczynski, T., silica gel matrix inorganic-organic composite material. Journal
Figueiredo, J.L., 2005. Preparation of carbon-based adsorbents of Hazardous Materials 173, 710e716.
from pyrolysis and air activation of sewage sludges. Chemical Environmental Report Flanders, 2010. MIRA. http://www.
Engineering Journal 108 (1e2), 169e177. milieurapport.be/Upload/Main/MiraData/MIRA-T/02_
Mohan, D., Pittman Jr., C.U., Bricka, M., Smith, F., Yancey, B., THEMAS/02_03/AG_ZWARE_METALEN.PDF (accessed
Mohammad, J., Steele, P.H., Alexandre-Franco, M.F., Gomez- 28.10.11.).
Serrano, V., Gong, H., 2007. Sorption of arsenic, cadmium, and http://www.ninds.nih.gov/disorders/wilsons/wilsons.
lead by chars produced from fast pyrolysis of wood and bark htm(accessed 01.02.12.).