Water Qual. Res. J. Canada, 2006
•
Volume 41, No. 2, 185–189
Copyright © 2006, CAWQ
Comparison of Analytical Techniques for Analysis
of Arsenic Adsorbed on Carbon
Nirbhay Narayan Yadav,1 Saravanamuthu Maheswaran,1† Vaithiyalingam Shutthanandan,2
Suntharampillai Thevuthasan,2 Todd R. Hart,2 Huu Hao Ngo3
and Saravanamuthu Vigneswaran3*
1
Nanoscale Organisation and Dynamics Group, University of Western Sydney,
Penrith South DC, NSW 1797, Australia
2
Environmental Molecular Sciences Laboratory, Pacific Northwest National Laboratory,
Richland, Washington, 99352, United States
3
Faculty of Engineering, University of Technology - Sydney, Sydney, NSW 2007, Australia
Activated carbon (AC) has been used extensively to treat arsenic-contaminated groundwater for a number of years. To date,
attempts to quantify directly the amount of arsenic removed by the activated carbon using nondestructive methods has been
limited. High-energy ion beam based proton induced x-ray emission (PIXE) is ideally suited to investigate the issues regarding the quantification of trace metals in solids. In this study, after the adsorption of arsenic on activated carbon, arsenic concentration in granular activated carbon (GAC) and powder activated carbon (PAC) were quantified using PIXE. The PIXE
results were compared with atomic absorption spectrometry (AAS) and inductively coupled plasma optical emission spectrometry (ICP-OES) measurements. Some differences are observed between these measurements. The differences are greater
in the case of GAC compared to PAC. These differences are mainly due to the inhomogeneous structure of GAC and PAC,
which includes the variable surface properties such as surface area and pore sizes in each granule or particle. The larger differences are mainly due to the increased particle dimensions of GAC compared to PAC and the nature of the internal pore
structure of GAC, which results in different amounts of arsenic adsorbed on different granules of GAC or even in different
regions of one granule. This inhomogeneity of arsenic concentration is clearly visible in the arsenic concentration map generated for a single GAC particle using microbeam PIXE.
Key words: activated carbon, PIXE, inhomogeneous structures, adsorption
Introduction
atoms are analyzed with a suitable spectrometer. Measurements of these characteristic x-rays with sufficient resolution to distinguish x-ray energies from adjacent elements in the periodic table provide a fingerprinting of the
elemental constituents of the sample. The quantification
with high accuracy and high sensitivity makes PIXE superior to all other x-ray based techniques.
To date, bulk analysis techniques used to quantify
the adsorbed trace metals in AC have been restricted to
wet chemical techniques and neutron activation analysis
(NAA). These techniques have many inherent problems
such as the extremely laborious and time-consuming
sample preparation required for wet chemical techniques
such as AAS and ICP-OES. A relatively large amount of
sample is required for NAA and it can only offer its
highest sensitivities with a very high neutron flux achievable in a nuclear reactor (Krachler et al. 2002).
In this work, granular activated carbon (GAC) and
powder activated carbon (PAC) were used as adsorbents
to remove arsenic from water. The amount of arsenic
adsorbed onto GAC and PAC was measured directly by
means of PIXE. The results were then compared with the
The presence of toxic elements including arsenic in trace
amounts in groundwater is a major issue in several locations, and in developing countries in particular (Tseng et
al. 1968; Cebrian et al. 1983; Meranger and Subramanian
1984; Dhar et al. 1997; Burkel and Stoll 1999; Koch et al.
1999; Karim 2000). Detecting the amount of trace elements removed by activated carbon directly in the adsorbent with high accuracy and sensitivity is a challenging
task. A rapid, highly sensitive, nondestructive, quantitative, multi-elemental analytical method is required to
investigate the issues associated with trace elements. Proton induced x-ray emission (PIXE) is such an analytical
tool and it is suitable for simultaneously quantifying trace
elements with sensitivity of at least parts per million
(ppm) (Grime 1998). In PIXE, the sample is irradiated by
accelerated charge particles (usually protons), and the
emitted x-rays during the de-excitation process within the
* Corresponding author; [email protected]
†
Deceased
185
186
Yadav et al.
adsorbed amounts determined from ICP-OES and AAS
measurements.
Experimental
Equilibrium and kinetic adsorption experiments were
carried out to evaluate different aspects of adsorption
mechanisms in GAC and PAC. In the equilibrium experiments, different amounts of GAC (between 0.1 and
10.0 g/L) were added to a solution with initial arsenic
concentration of 4.9 mg/L and then agitated on a platform mixer for 72 h. The ambient temperature was
22ºC. In the case of PAC, between 0.1 and 5.0 g/L PAC
was added to a solution with initial arsenic concentration of 1.0 mg/L. In the kinetic experiments, 5 g/L of
GAC was added to a solution with initial arsenic concentration of 5.3 mg/L and then agitated for between
5 min to 9 h using a mechanical stirrer at 100 rpm. With
PAC, 5 g/L was added to a solution with initial arsenic
concentration of 1.0 mg/L and then agitated for between
5 min and 8 h. The GAC and PAC used were taken from
single batches for which the characteristics are given in
Tables 1 and 2, respectively. The most important characteristic of the GAC was its extremely large surface area
(more than 1000 m2/g GAC) and high porosity that give
it an advantage in adsorbing substances from water.
The PAC was removed from solution by filtration
through No. 42 Whatman filter paper (which was tested
for arsenic adsorption). This was not necessary in the
case of GAC because of its rapid natural settling and
mechanical resistance to the shaking process. After the
GAC and PAC were separated from the solution, they
were dried at room temperature in a desiccator, then
measured directly for arsenic concentration using PIXE.
It should be noted that carbon was not washed before
the drying process. The arsenic concentration in solution
was measured using flame AAS and ICP-OES.
PIXE measurements on these samples were carried
out at the Environmental Molecular Sciences Laboratory
(EMSL) accelerator facility, Pacific Northwest National
Laboratory, U.S.A. The details of the accelerator facility
and the end stations are described elsewhere (Thevuthasan et al. 1999). The experiments were performed
in a high vacuum chamber which is equipped with many
capabilities including PIXE. The vacuum system consists
of a 17-inch (43.2 cm) diameter cylindrical aluminium
chamber with load lock capabilities. A CCD zoom camera provides visual indication of the size and location of
the beam spot on target. Since the whole chamber is electrically isolated from the surroundings, it acts as a Faraday cup and the charge collection can be accurately performed using the beam current measured in the chamber.
PIXE measurements were carried out using a
2.5 MeV proton beam at an incidence angle of 45º. A
typical beam current of 10 nA with a beam diameter of
0.5 mm was used. A translatable 80-mm2 Si(Li) detector
(GRESHAM) was used for PIXE measurements at an exit
angle of 39º with respect to the surface normal. A filter is
used to balance the transmission of low- and high-energy
x-rays. Known as a funny filter, it is made out of a
graphite disc with a small hole in the middle. The filter is
attached to the front of the x-ray detector. A total dose of
0.5 µC (approximately 1.6 × 1015 H+/cm2) was used to
collect each PIXE spectrum. Theoretical simulations of
PIXE experimental data were performed to determine
elemental composition using the GUPIX program
(Maxwell et al. 1989). PIXE results were compared to
the adsorbed amount of arsenic determined using AAS in
the case of GAC and ICP-OES for PAC.
Fly ash standard reference materials (SRM 2689)
from the National Institute of Science and Technology
(NIST) were used for calibration and validification of
the PIXE method. The measurements from each technique (PIXE, AAS and ICP-OES) were repeated three
times for each sample.
Results and Discussion
A typical PIXE experimental spectrum from one of the
PAC samples (after undergoing adsorption) is shown in
Fig. 1. The elements present were identified by their characteristic energies. The area under each peak directly corresponds to the concentration of that element in the sample.
TABLE 2. Characteristics of the powder activated carbon
(PAC) used (supplier: James Cumming & Sons Pty Ltd.,
Australia)
Specification
TABLE 1. Characteristics of the granular activated carbon
(GAC) used (supplier: Calgon Carbon Corp., U.S.A.)
Specification
Iodine number (mg/g min)
Maximum ash content
Maximum moisture content
Bulk density (kg/m3)
BET surface area (m2/g)
Nominal size (m)
Mean pore diameter (Å)
GAC
800
5%
5%
748
1112
3 × 10-4
26.14
Iodine number (mg/g min)
Maximum ash content
Maximum moisture content
Bulk density (kg/m3)
Surface area (m2/g)
Nominal size
Type
Mean pore diameter (Å)
Micropore volume (cc/g)
Mean diameter (µm)
Product code
PAC-WB
900
6%
5%
290–390
882
80% min finer than 75 µm
Wood based
30.61
0.34
19.71
MD3545WB powder
A New Analytical Method of Arsenic Adsorbed on Carbon
Fig. 1. A typical PIXE spectrum from one of the PAC samples after an adsorption experiment. The 2.5 MeV proton
beam was used for the PIXE measurements.
In the present work, the amount of arsenic was determined
from GUPIX simulations. In the experimental spectrum,
peaks at 10.5 ± 0.2 keV and 11.7 ± 0.2 keV correspond to
arsenic Kα and Kβ, respectively. Arsenic concentration was
determined using the Kα peak because the area underneath
the Kα peak is much larger than that of the Kβ, thus the use
of Kα peak leads to greater accuracy and sensitivity.
Determination of Arsenic Concentration in GAC
The results of adsorption equilibrium experiments on GAC
samples are shown in Fig. 2. GAC was removed from the
solution and the concentration of arsenic adsorbed in the
GAC was directly measured using PIXE. When the GAC
particles were mounted on the PIXE sample holder, the
inhomogeneous GAC particle geometries generated voids
Fig. 2. PIXE and AAS measurements of arsenic concentration
in GAC after the adsorption equilibrium experiments.
187
between the particles. The concentration of the remaining
arsenic in the treated solution was measured using AAS.
The amount of arsenic adsorbed by GAC was calculated
by subtracting the amount measured by AAS from the
amount of arsenic present in the initial solution. These
results were compared with the arsenic concentrations
measured by PIXE. The discrepancies exist in the arsenic
concentrations measured by these two methods due to the
inhomogeneous geometries associated with GAC particles
and they are clearly visible in the results (Fig. 2). These discrepancies are especially present in samples below 5 g/L
but are significantly reduced for samples above 5 g/L. This
could have been a result of a more even distribution of
arsenic uptake by GAC when a higher dose was used.
The results of the kinetic experiments are shown in
Fig. 3. Here again, significant differences are noted
between both PIXE and AAS measurements. As in the
case of equilibrium experiments, the adsorbed amounts
of arsenic measured using PIXE are highly irregular
when compared to those determined from AAS due to
the inhomogeneous GAC particles.
To understand the inhomogeneous adsorption of
arsenic in activated carbon (especially the GAC),
microbeam (with 20-micron beam size) PIXE measurements were carried on two GAC particles chosen arbitrarily. Arsenic concentration maps generated from the
micro PIXE measurement are shown in Fig. 4. The
results from these measurements clearly show that different regions within each GAC particle adsorb different
amounts of arsenic. In addition, arsenic adsorption
between the two granules is also different.
Determination of Arsenic Concentration in PAC
The results of adsorption equilibrium experiments are
shown in Fig. 5. Similar experimental procedures
Fig. 3. PIXE and AAS measurements of arsenic concentration
in GAC after the adsorption kinetic experiments.
188
Yadav et al.
Fig. 4. Mapping of arsenic concentration in two arbitrarily chosen GAC particles
using micro-PIXE measurement. The measurements were taken using 20-µm diameter
incident proton beam (Maheswaran et al. 2003).
described above for GAC were followed in determination of arsenic concentration in PAC. PAC was taken
out from the solution after treatment and arsenic concentration in PAC was directly measured using PIXE.
Arsenic concentration in PAC was also measured using
ICP-OES measurements after stripping all the arsenic
from the PAC (ICP-OES direct). The concentration of
remaining arsenic in treated solution was measured by
ICP-OES. The amount of arsenic adsorbed by GAC
(for ICP-OES measurements) was calculated by subtracting arsenic concentrations determined in treated
solution from the amount of arsenic present in the initial solution. The amount of arsenic adsorbed onto
PAC determined by PIXE was similar to the values
determined by ICP-OES direct and ICP-OES. Since the
density of activated carbon is significantly higher in the
PIXE measurement area for PAC compared to GAC,
the voids during the sample mounting were significantly reduced in the case of PAC. As a result, both
PIXE and ICP-OES measurements regularly gave consistent results in the case of PAC.
The results of adsorption kinetic experiments are
presented in Fig. 6. As observed in the equilibrium
experiments, the discrepancy between the values determined from PIXE and ICP-OES was less compared to
that of GAC experiments.
The differences between the arsenic concentrations
measured using PIXE and the other techniques are significant in GAC compared to PAC. These differences are
mainly due to the inhomogeneous structure of GAC,
which includes the variable surface properties such as
surface area and pore sizes in each GAC granule. Quantitative analysis of PIXE data from specimens with inhomogeneous structure is still an emerging field. Despite its
many advantages described previously, the analysis of
inhomogeneous structures creates uncertainties due to
unpredictable cross-sections, attenuation coefficients and
stopping powers throughout the range of impinging pro-
Fig. 5. PIXE and ICP-OES measurements of arsenic concentration in PAC after the adsorption equilibrium experiments.
Fig. 6. PIXE and ICP-OES measurements of arsenic concentration in PAC after the adsorption kinetic experiment.
A New Analytical Method of Arsenic Adsorbed on Carbon
tons (Sjöland et al. 2000). Further difficulty arises due to
the use of a relatively small volume of samples (one or
two particles of GAC) in PIXE measurements to cover
the entire proton beam of 0.5 mm in diameter and the
existing “voids” between the GAC particles. The area
examined by the PIXE beam might not have been representative for the whole sample. A sampling technique,
such as palletization, can be employed to get better representation of arsenic concentration throughout inhomogeneous samples or appropriate corrections need to be
explored and carried out.
Conclusions
We have demonstrated that the PIXE technique can be
used effectively to determine trace arsenic concentration directly in PAC. The inhomogeneous structure of
GAC makes it difficult to obtain consistent values for
the arsenic concentration from the PIXE measurements.
Different sampling techniques should be employed to
get better representation of inhomogeneous samples or
appropriate corrections need to be explored. The
micro-PIXE images provide valuable information on
the distribution of adsorbate (in this case, arsenic) onto
pores of adsorbent.
Acknowledgements and Dedication
This paper is dedicated to the memory of Saravanamuthu
Maheswaran (1959–2005) who tragically passed away
during the process of finalizing it. A portion of this work
was conducted at the Environmental Molecular Sciences
Laboratory and the financial support of the Office of Biological and Environmental Research (OBER), U.S.
Department of Energy (DOE) is greatly appreciated.
References
Burkel RS, Stoll RC. 1999. Naturally occurring arsenic in
sandstone aquifer water supply wells of Northeastern
Wisconsin. Groundwater Monit. Remediat. 19(2):
114–121.
Cebrian ME, Albores A, Aguilar M, Blakely E. 1983.
Chronic arsenic poisoning in the north of Mexico.
Human Toxicol. 2:121–133.
Dhar RK, Biswas BK, Samanta G, Mandal BK,
Chakraborti D, Roy S, Jafar A, Islam A, Ara G, Kabir
189
S, Khan AW, Ahmed SA, Hadi SA. 1997. Groundwater
arsenic calamity in Bangladesh. Current Science
73(1):48–59.
Grime GW. 1998. Analysis of individual environmental
particles using MicroPIXE and nuclear microscopy.
X-Ray Spectrometry 27:221–231.
Karim M. 2000. Arsenic in groundwater and health problems in Bangladesh. Water Resour. 34(1):304–310.
Koch I, Feldman J, Wang L, Andrewes P, Reimer KJ, Cullen
WR. 1999. Arsenic in the Meager Creek hot springs
environment, British Columbia, Canada. Sci. Total
Environ. 236:101–117.
Krachler M, Emons H, Barbante C, Cozzi G, Shotyk W.
2002. Inter-method comparison for the determination
of antimony and arsenic in peat samples. Anal. Chim.
Acta 458:387–396.
Maheswaran S, Yadav NN, Shutthanandan V, Thevuthasan
S, Ngo HH, Vigneswaran S. 2003. Evaluating the use
of proton induced x-ray emission (PIXE) technique in
quantifying the arsenic in an adsorbent. In Proceedings
of Asian Waterqual 2003, IWA Asia-Pacific Regional
Conference, Bangkok, Thailand.
Maxwell JA, Campbell JL, Teesdale WJ. 1989. The Guelph
PIXE software package. Nuclear Instruments and
Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 43(2):218–230.
Meranger JC, Subramanian KS. 1984. Arsenic in Nova
Scotian groundwater. Sci. Total Environ. 39:49–55.
Sjöland KA, Munnik F, Wätjen U. 2000. Thick-target correction in PIXE for randomly inhomogeneous samples.
Nuclear Instruments and Methods in Physics Research
Section B: Beam Interactions with Materials and Atoms
161–163:264–268.
Thevuthasan S, Peden CHF, Engelhard MH, Baer DR, Herman GS, Jiang W, Liang Y, Weber WJ. 1999. The ion
beam materials analysis laboratory at the environmental molecular sciences laboratory. Nuclear Instruments
and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 420(1–2):81–89.
Tseng W-P, Chu HM, How SW, Fong JM, Lin CS, Yeh S.
1968. Prevalence of skin cancer in an endemic area of
chronic arsenicism in Taiwan. J. Nat. Cancer Institute
40(3):453–463.
Received: August 3, 2005; accepted: January 30, 2006.