Elemental Mercury Adsorption by Activated Carbon
Treated with Sulfuric Acid
Paper # (To be supplied)
Yong Hua Li, Shannon D. Serre, Chun Wai Lee, and Brian K. Gullett
U.S. Environmental Protection Agency
Office of Research and Development
National Risk Management Research Laboratory (MD-65)
Research Triangle Park, NC 27711
ABSTRACT
Adsorption of elemental mercury (Hg°) at 125 °C by a sulfuric acid [H2S04, 50%(w/w) solution]
treated carbon has been studied for the removal of Hg from flue gas. The pore structure of the
sample was characterized by nitrogen (N?.) adsorption at -196 °€ and the t-plot method analysis.
Thermogravimetric analysis (TGA) was used to examine the thermal behaviors of the treated
sample. Results from this study show that the tkSCVtreated carbon has a very high Hg°capture
capacity (27 mg/g) via a physisorption mechanism. It is shown that the pore volume and surface
area of the treated carbon were drastically reduced due to adsorption of the H2SO4 in the
micropores of the carbon, resulting in a narrower microporosity. A significant amount of water
(H2O) was found to remain in the carbon through hydrogen bonding with the H2SO4 molecules.
It is suggested that the Hg° adsorption capacity of the I^SCU-treated carbon can be explained by
the enhanced adsorption potential and enthalpy of adsorption. The enhancements may be
resulting either from the narrow microporosity due to overlapping of the adsorption field from
the neighboring pore walls, or from the increased surface polarity with which coulombic
contributions from polarization energy to the overall adsorption potential could be significant.
INTRODUCTION
Mercury (Hg) has been identified by the U.S. Environmental Protection Agency (EPA) as a toxic
pollutant of environmental concern. Regulations governing Hg emissions from coal-fired power
plants are to be issued by EPA by December 15, 2004'. Development of Hg control technology
requires maximizing sorbent effectiveness for optimal Hg removal. Activated carbons have been
shown to be an effective sorbent for removal of Hg from combustion flue gases. The control of
Hg emissions is strongly dependent on the form of Hg emitted. A major portion of the Hg
emitted from many coal-fired power plants is in the elemental form (Hg°). Hg° is more difficult
to capture than its oxidized form, due to its higher volatility and insolubility in water. Studies
evaluating the effectiveness of Hg° adsorption by activated carbons have shown that several
factors influence the efficiency of Hg° removal by activated carbons, including carbon
characteristics, flue gas composition, and Hg speciation2"6. Different activated carbons have
been examined in bench-, pilot-, and full-scale tests7. However, the mechanism involved in Hg°
adsorption is not well understood, the important parameters of selecting activated carbons for
-------
Hg° removal in the flue gas, which could determine the adsorption capacity and reaction kinetics,
have not been established.
Activated carbons treated with H2S04 exhibit high Hg° sorption capacity8. By treating granular
activated carbon in an aqueous 70% H2SO4 solution, and then drying it at 100-200 °C for several
hours, the treated activated carbon at 110-120 °C has a very high Hg° capture capacity, while the
presence of sulfur dioxide (SO2) has no impact on Hg° removal. Since activated carbon is used
to remove H2SO4 in industrial applications , it would be mutually beneficial if the 'H2SO4-
exhausted activated carbon' could be used lor the control of Hg° emissions. Furthermore, studies
of Hg° adsorption by the H2S04-treatcd carbon could lead to a better understanding of the Hg°
adsoiption mechanism and, hence, a more effective carbon sorbent. The work reported in this
paper investigates the adsorption of Hg° by an H2S04-treated carbon.
EXPERIMENTAL
A commercially available activated carbon. BPL (Calgon Carbon Corporation, Pittsburgh, PA)
was treated with a 50% (weight percentage, w/w) H2SO4 solution at room temperature (25 °C).
About 20 g of sample was placed in a glass beaker containing 300 mL of H2S04 solution, and
stirred for 20 hr. The supernatant liquid was then decanted, and the sample was oven-dried at
150 °C for 20 hr. The samples were stored in a sealed container for later use.
The thermal behaviors of both the as-received (designated AR) and the H2SC>4-treated
(designated H2S04) samples were evaluated, using a thermogravimetric analyzer (TGA-7, Perkin
Elmer). About 20 mg of sample was heated from room temperature to 110 °C and held for 2 hr
in a helium (He) flow. The sample was then heated at a constant rate of 10 °C/min to 900 °C and
held for 30 min. The weight loss of the sample as a function of time (hence, temperature) was
obtained.
The pore structures of the samples were characterized by N2 adsorption at -196 °C using a
volumetric adsorption apparatus (ASAP 2400, Micromeritics). The sample was degassed at
150°C for 24 hr at a pressure of 1.33x103 Pa prior to the adsorption measurement. The total
surface area was evaluated by the BET (Brunauer-Emmett-Teller) method10. The total pore
volume was obtained from the N2 adsorption isotherm at P/Pq = 0.99. The t-plot method" was
used to analyze the micropore structure of the samples.
Hg° adsorption at 125 °C was performed at a I Ig° concentration of 55 ppb [450 ug/m\ standard
temperature and pressure, (8TP)] in a N2 atmosphere, and a total flow rate of 370 mlVmin (STP).
A quartz fixed-bed reactor (1.27 cm inside diameter), surrounded by a temperature-controlled
electrical furnace, was used for the Hg° adsorption experiments. The Hg° -laden gas mixture was
generated using a Hg° permeation tube contained in a constant temperature system
(Dynacalibrator Model 190, VICI Metronic). An ultraviolet (UV) Hg° analyzer (Model 400A,
BUCK Scientific) was used to continuously measure the concentration of Hg° in the reactor
outlet stream. In a typical experiment, about 10-20 mg of carbon sample was mixed with 2 g of
sand and loaded into the reactor, and then heated to 125 °C for 30 min until isothermal condition
was achieved. The Hg°-laden N2 flow was switched to pass through the reactor until the exit Hg°
2
-------
concentration reached the inlet Hg° concentration (baseline). The Hg° adsorption capacity was
calculated from the area between the baseline Hg° concentration and the breakthrough curve.
The experimental setup is described in detail elsewhere12.
RESULTS AND DISCUSSIONS
Thermal Behavior
Figure 1 shows the weight loss profiles of the as-received and the H2SCVtreated carbon samples
obtained from the 'IGA analysis. The weight loss of the BPL-AR sample accounts for about 3%
Figure 1: Weight loss profiles of the as-received
and H2S()4-treated samples
¦ — -• - ' i 1000
I
800
o
o
400 I
sD
f—
!
200
0 50 100 150 200 250 300
Time (min)
from 110 to 900 °C. Most of the weight loss occurred at above 400 °C, which is due to the
decomposition of oxygen surface complexes from the carbon surface13. The weight loss profile
for the H2S04-treated sample corresponds to three different temperature regimes: 110 °C, 200-
400 °C, and > 400 °C. As shown in Figure 1, a weight loss of about 15% occurred when the
H2SC>4-treated sample was heated at 110 °C for 2 hr under a He atmosphere. This weight loss
can be attributed to removal of water from the sample. Although the sample was dried at 150 °C
in an oven for 20 hr after the H2SO4 treatment, H20 has a strong affinity through hydrogen
bonding to the adsorbed H2SO4 on the carbon surface 14,15. Readsorption of H20 from the
atmosphere during sample handing is unavoidable and is effected by extensive hydrogen bonding
with the H2SO4 molecules on the carbon surface. A high weight loss of about 40% of the total
sample weight was found occurring in the temperature range of 200-400 °C. This can be
100
/ BPL-AR
80
60
40
20
/
I
3
-------
explained by the loss of H2S04 from the carbon surface around the boiling points (~ 300 °C) of
hydrates of H2SO4 l5. This observation also agrees well with that reported by Gomez-Serrano et
al. 16 who showed that a high weight loss occurred below 400 °C in the heating of a H2SO4-
trcated carbon from 30 to 800 °C. A weight loss of 34% was reported for an activated carbon
treated with a 62% (w/w) H2S04 solution for 24 hr. The presence of HSO4" and S042 ions on the
carbon surface is evident by their FT1R (Fourier transform infrared) spectra, which account for
the mass uptake of the activated carbons after treatment with the H2SO4 solution. As shown in
Figure 1, the weight loss of the H2S04-treated sample at above 400 °C is small, about 5% of the
total mass of the sample. This suggests that the thermal response of the sample at higher
temperatures was similar to that of the as-rcceived sample.
Pore Structures
Figure 2 shows the isotherms of N2 adsorption at -196 °C for the BPL-AR and the BPL-H2SO4
samples. The BET surface areas (Sbet) and pore volumes (Vt), derived from the isotherms, are
presented in Table 1. The adsorption of H2SO4 into the pores of the activated carbon decreased
the pore volume and surface area of the BPL-H2SO4 sample drastically; to only 36 and 22%,
respectively, compared to those of the as-received carbon. Both isotherms are type I, and
adsoiption occurs at a low relative pressure, which are indicative of the microporous
characteristics of the two samples. There is a widening of the microporosity (opening of the
knee) of the BPL-AR sample, indicating the presence of supermicroporosity (pore size between 7
Figure 2: N2 adsorption isotherms of samples
400 -
CL
go
eo
BPL-AR
300
E
o
o 200 -
¦d
C3
100 -
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Relative pressure (P/P0)
4
-------
and 20 A)!l'17,18. The very closed knee, shown in the isotherm of the BPL-H2SO4 sample,
indicates an absence of supermicroporosity and the presence of very narrow microporosity (< 7
A)'1,
Table 1. Pore structure characteristics of samples
Sample
Sbet
(m2/g)
v,
(em3/g)
S£*
(m2/g)
C *
(m2/g)
V *
*5
(cm7g)
BPL-AR
1136
0.58
52
1080
0.48
BPL-H2SO4
253
0.21
46
207
0.10
* Derived from the t-plot method
Figure 3 shows the t-plots of the two samples, the volume adsorbed versus the statistical
thickness. The thickness of the adsorbed molecules was calculated using the de Boer equation".
Two distinctive linear portions of the curves are clearly shown in the figure, indicative of
micropores in the presence of mesopores". The broad curvature between the two linear portions
of BPL-AR is an indication of a wide distribution of micropores. The micropore volume of the
sample is obtained by extrapolating the upper linear part of the curve to the volume axis, and the
external surface area can be obtained from the slope. The total surface area can be obtained from
the initial slope, passing through the origin. The micropore surface area is then obtained by the
difference between the total surface area and the external surface area. In the absence of
sufficient N2 adsorption data at low relative pressures, the total surface area from the lower linear
portion of the t-curve cannot be easily calculated; thus, the BET surface area was used to
Figure 3: t-plots of the samples
400
BPL-AR
E
u
200
C3
0
2
4
8
6
10
Statistical thickness (A)
5
-------
calculate the micropore surface area. The t-plot analysis results are shown in Table 1. The
mesopore surface area (Se) of the H2S04-treated sample does not change significantly from that
of the as-received sample, 46 versus 52 m2/g, respectively. This is consistent with their similar
N2 adsorption isotherms at high relative pressures as shown in Figure 2. However, the micropore
surface area (Sm) has been reduced from 1080 to 207 m2/g after H2SO4 treatment. Therefore, it is
quite likely that the decreases of the pore volume and surface area of the I^SCVtreated sample
are due to the reduction of the micropores. It appears that H2SO4 is strongly adsorbed in the
micropores of the carbon, resulting in very narrow micropores and also a homogeneous
microporosity distribution. Considering that the size of S042 is about 5 A19, it can be suggested
that the narrowest micropores (< 5 A) which existed in the untreated sample could be blocked
after H2S04 adsorption. Our previous results12 on carbon dioxide (C()2) adsorption at 0 °C of the
BPL-AR sample have shown that the average micropore width evaluated by using the DR
(Dubinin-Radushkevich) method11 is about 15 A. The presence of newly formed
ultramicropores (< 5 A) after H7SO4 treatment, in which N2 at -196 °C has severe diffusion
limitation in such narrow micropores11, l8, could also account for the observed low adsorption of
N2 at such a temperature.
Mercury Adsorption and Dcsorption
Figures 4a and 4b show the Hg° adsorption breakthrough curves of the HjSCVtreated (-H2SO4)
and as-received (-AR) samples, respectively. Another experiment was also performed with
H?S04-treated carbon heated at 420 °C in a N2 atmosphere to remove the adsorbed H2SO4. The
breakthrough curve of the heat-treated sample (-H2SO4-420) is also shown in Figure 4b. More
than 70 hr was needed for the BPL-H2SO4 sample to reach adsorption equilibrium. The total Hg°
capture capacity of the H2S04 treated sample estimated from the breakthrough curve is 27 mg/g
(sorbent), which is an extremely high Hg° adsorption capacity compared to those of other
sorbents. The Hg° adsorption capacity of the BPL-AR and BPL-H2S(V42Q samples are very
low, 40 and 16 j.tg/g. respectively. Removal of the adsorbed H2S04 from the carbon surface
through heat treatment at 420 °C reduced the Hg° adsorption capacity of the sample to a level
similar to that of the as-received sample.
Figure 4a shows an initial increase of Hg° concentration followed by a slight decrease in the first
2 hr. This is attributed to the pores that were initially blocked by the adsorbed H20 associated
with the H2S04 molecules. Following H20 evaporation from the surface, the pores opened up
and became available for Hg° adsorption, which is evident in the TGA analysis shown in Figure
1. Consequently, the Hg° breakthrough concentration decreased, and more Hg° was adsorbed on
the surface.
To examine the thermal stability of the Hg° adsorbed by the FLSCVtreated sample, the Hg°-
laden gas was changed to N2 at the end of the Hg° adsorption experiment, and the reactor was
purged with N2 at the same temperature (125 °C) as during its exposure to Hg°. The Hg°
concentration measured in the reactor outlet stream as a function of N2 purge time is shown in
Figure 5. Hg° immediately desorbed from the sample at a high desorption rate for about 10 hr
and then decreased to a much lower rate. The desorbed Hg° calculated from the sample after 93
hr of N2 purge was 19.4 mg/g, 72 % of the total Hg°adsorbed. In order to further facilitate the
6
-------
Figure 4a: Mercury adsorption breakthrough
curve of BPL-H2SO4
60
50
i
30
20
I
1
j
i
10
0 I
0
70
30 40
50
60
10
20
Time (hr)
Figure 4b: Mercury breakthrough curves of BPL-AR
and BPL-H2SO4-42O samples
60 -
50 '
30
BPL-AR
* " BPL-H9SO4-42O
'oo 20 1
10
I
«
I
0 — -- ; - -"1 -¦ i ¦ 1 1 !
0 20 40 60 80 100 120 140
Time (min)
7
-------
desorption process, the sample was then heated under a constant heating rate (10 °C/min) to 900
°C. A total of 2.5 mg/g of Hg° was desorbed at the higher temperatures (125-900 °C). Based on
the above results, it can be suggested that most of the Hg° (about 70 %) captured by the H2SO4-
treated sample was physically adsorbed in the micropores of the sample, since it can be desorbed
at the adsorption temperature (125 °C) in a N2 atmosphere.
Figure 5: Mercury concentration profile of BPL-H2S04
sample during N2 purge
60 - - —- ¦ ¦ -
50
XI
& 40
30
20
60
10
0
80
100
0
20
40
60
Time (hr)
The forces of attraction involved in physical adsorption include both van der Waals forces and
electrostatic (or coulombic) forces, van der Waals forces are always present, while the
electrostatic attractions are significant only in the cases of adsorbents that exhibit ionic
characteristics. H2SO4 is a polar molecule, which could form hydrogen bonds and be ionized in
an aqueous solution . The effect of polarity in enhancing the energy of interaction is discussed
in the literature11. The presence of HSO4" and SO42" ions, and the formation of extensive
hydrogen bonding resulting from H2SO4 adsorption could create a strong electric field on the
carbon surface. The coulombic contributions to the overall adsorption potential could be
significant due to the specific interaction between the ionic surface and Hg°. It was observed
that the surface moisture of activated carbons has a strong enhancement effect on Hg°
adsorption2'12. The effect has been interpreted in terms of chemisorption, based on the results12
of temperature-programmed desorption and an X-ray absorption fine structure (XAFS)
spectroscopy analyses of adsorbed Hg°. It has been suggested that interactions between H2O and
carbon-oxygen complexes may create certain active sites, or affect surface conditions, which
favor Hg° adsorption12. It is useful to distinguish between physisorption and chemisorption in
8
-------
discussing the fundamentals of adsorption. However, it is not always possible to categorize a
particular system unequivocally. In fact, hydrogen bonding is formed when I hO is adsorbed on
the carbon surfaces, thus resulting in an increase of the surface polarity. Therefore, the
coulombic forces due to the surface polarity might play a significant role in Hg° adsorption.
With the presence of very narrow micropores in the H2S04-treated carbon, the enhanced overall
interaction potential and, therefore, the enthalpy of adsorption could also lead to a high Hg°
adsorption capacity of this sample. The interaction energy between the free surface of a solid
and an adsorbate molecule is rather different from that of the narrow micropore as a consequence
of the overlapping of the adsorption field from the neighboring pore walls. This overlap leads to
a strong adsorption of the gas by the micropore, and to an enhancement of the enthalpy of
adsorption11. It has been shown that the critical parameter for the enhancement of interaction
energy is not the pore size itself, but rather the ratio of the size of the pore to that of the adsorbate
molecule". The upper limit of size at which the adsorption process with an enhancement of
interaction potential was suggested to be < 2
-------
be significant, or from the narrow microporosity as a consequence of the overlapping of the
adsorption field from the neighboring pore walls.
ACKNOWLEDGEMENT
This research was supported in part by an appointment (Y. H. Li) to the Postdoctoral Research
Program at EPA's National Risk Management Research Laboratory administered by the Oak
Ridge Institute for Science and Education through interagency agreement DW89938167 between
the U.S. Department of Energy and the EPA.
REFERENCE
1. U.S. EPA, Mercury, www.epa.gov/mercury, Washington, DC, January 2001.
2. Krishnan, S. V., Gullett, B. K., Jozewicz, W. Environ Sci Technol 1994; 28: 1506-1512.
3. Huang, H. S., Wu, J. M., Livengood, C. D. Hazardous Waste & Hazardous Materials 1996;
13: 107-119.
4. Korpiel, J. A., Vidic, R. D. Environ Sci Technol 1997; 31: 2319-2325.
5. Carey T. R., Hargrove, O. W. Jr., Richardson, C. F., Chang, R., Meserole, F. B. J. Air &
Waste Manage. Assoc. 1998; 48, 1166-1174.
6. Rostam-Abadi, M,, Chen, S. G., Hsi, H. C., Rood, M., Chang, R., Carey, T., Hargrove, B.,
Richardson, C., Rosenhoover, W., Meserole, F. Presented at the EPRI-DOE-EPA Combined
Utility Air Pollutant Control Symposium, Washington, DC, 1997.
7. Carey, T. R., Richardson, C. F., Chang, R., Meserole, F. B., Rostam-Abadi, M., Chen, S.
Environmental Progress 2000, 19, 167-174.
8. U.S. Patent No. 3,876,393, 1975.
9. Schuliger, W. G., In Carbon Adsorption Handbook, eds P. N. Cheremisinoff & F. Ellerbusch.
Ann Arbor Sci., Ann Arbor, MI, 1980, p. 82.
10. Brunauer, S., Emmett, P. H., Teller, E. J. Amer. Chem. Soc. 1938, 60, 309.
11. Gregg, S. J., Sing, K. S. W. In Adsorption, Surface Area and Porosity, Second Edition, 1982,
Academic Press, London, UK.
12. Li, Y. H„ Lee, C. W„ Gullett, B. K. Carbon, 2001, In press.
13. Bansal, R. C., Donnet, J. B., Stoecki, F. Active Carbon. New York, NY, and Basel,
Switzerland: Marcel Dekker, 1988.
14. Kuczkowski, R. L„ Suenram, R. D„ Lovas, F. J. J. Am. Chem. Soc. 1981, 103, 2561-2566.
15. Sander, U. H. F., Rother, U., Kola, R. In Sulphur, Sulphur Dioxide and Sulphuric Acid. 1984,
The British Sulphur Corporation Ltd. p. 266, London, UK.
16. Gomez-Serrano, V., Acedo-Ramos, M. J., Lopez-Peinado, A. J. Chem. Tech. Bio technol.
1997,68, 82-88.
17. Raymundo-Pinero, E., Cazorla-Amoros, D., Salinas-Martinez de Lecea, C., Linares-Solano,
A. Carbon 2000, 38, 335-344.
18. Marsh, H. Carbon 1987, 25, 49-58.
19. Jenkeis, H. D. B. J. Chem. Educ 1979, 58, 576.
20. Cotton, F. A., Wilkinson, G. In Advanced Inorganic Chemistry, Fifth Edition, John Wiley &
Sons, Inc., New York, NY, 1988, pp.114-115.
10
-------
TECHNICAL REPORT DATA
NRMRL-RTP-P-613 (Please read Instructions on the reverse before complet.
1 REPORT NO.
EPA/600/A-01/112
2
3. RECIi
4. TITLE AND SUBTITLE
Elemental Mercury Adsorption by Activated Carbon
Treated with Sulfuric Acid
5. REPORT DATE
6. PERFORMING ORGANIZATION CODE
7. AUTHORS
8. PERFORMING ORGANIZATION REPORT NO
Y.H. Li, S.D. Serre, C.W. Lee, and B.K. Gullett
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
See Block 12
11. CONTRACT/GRANT NO.
U.S.DoE IAG DW89938167 (Li)
12, SPONSORING AGENCY NAME AND ADDRESS
U. S. EPA, Office of Resea
rch and Development
13. TYPE OF REPORT AND PERIOD COVERED
Published paper; 2/00 - 1/01
Air Pollution Prevention and Control Division
Research Triangle Park, North Carolina 27711
14. SPONSORING AGENCY CODE
EPA/600/13
15 supplementary notes APPCD project officer is Chun Wai Lee, Mail Drop 65, 919/541-7663.
Y.H. Li is an ORISE postdoctoral fellow. For presentation at EPA/DOE/EPRI Megasympo-
sium, Chicago, IL, 8/20-24/01.
16 abstract The paper gives results of a study of the adsorption of elemental mercury at
125 C by a sulfuric-acid (H2S04, 50% w/w solution)-treated carbon for the removal of
mercury from flue gas. The pore structure of the sample was characterized by nitrogen
(N2) at -196 C and the t-plot method analysis. Thermogravimetric analysis (TGA) was
used to examine the thermal behaviors of the treated sample. Study results show that
the H2S04-treated carbon has a very high elemental mercury capture capacity (27 mg/g)
via a physisorption mechanism. The pore volume and surface area of the treated carbon
were drastically reduced due to adsorption of the H2S04 in the micropores of the carbon,
resulting in a narrower microporisity. A significant amount of water (H20) was found to
remain in the carbon through hydrogen bonding with the H2S04 molecules. It is suggested
that the elemental mercury adsorption capacity of the H2S04-treated carbon can be ex-
plained by the enhanced adsorption potential and enthalpy of adsorption. The enhance-
ments may result either from the narrow microporisity due to overlapping of the adsorp-
tion field from the neighboring pore walls, or from the increased surface polarity with
which coulombic contributions from polarization energy to the overall adsorption poten-
tial could be significant.
17
KEY WORDS AND DOCUMENT ANALYSIS
a DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Mercury (Metal)
Adsorption
Activated Carbon
Sulfuric Acid
Flue Gases
Pollution Control
Stationary Sources
13B
07B
14G
11G
21B
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
21. NO. OF PAGES
20. SECURITY CLASS (This Page)
22. PRICE
EPA Form 2220-1 (Rev. 4-77 ) PREVIOUS EDITION IS OBSOLETE forms/admin/techrpt fmn 7/8/99 pad
------- |