das2016
das2016
To cite this article: Dipa Das, Debi Prasad Samal & Bhim C. Meikap (2016) Removal of
CO2 in a multistage fluidized bed reactor by diethanol amine impregnated activated
carbon, Journal of Environmental Science and Health, Part A, 51:9, 769-775, DOI:
10.1080/10934529.2016.1170462
Article views: 15
removal efficiency of CO2 have been investigated and presented. The percentage removal of CO2 has
been found close to 80% under low gas flow rate (0.188 m/s), high solid flow rate (4.12 kg/h) and weir
height of 50 mm. From this result it has been found out that multistage fluidized bed reactor may be a
suitable equipment for removal of CO2 from flue gas.
Introduction
activated carbon,[3–5] coal,[6–8] calcium based adsorbents,[9–12]
The onset of the industrial revolution, burning of fossil fuels sodium or potassium based adsorbent [7,13] etc. Among which
and deforestation has led to increase in atmospheric CO2, thus activated carbons (ACs) are highly microporous materials hav-
resulting in climate change and global warming. As the concen- ing large specific surface area and considered to be the most
tration of the carbon dioxide in air increases, it traps more heat efficient adsorbents.[14] As CO2 comes from acidic pollutants,
and raises the temperature of the earth’s surface due to which various efforts have been made to increase the alkalinity of
various natural calamities like change in rainfall patterns, ACs. Different methods are used to increase the alkalinity of
increases of sea level, and lots of bad impacts on vegetation, the surface. The impregnation of ACs with a basic solution [15–
18]
human beings and wildlife occur. Fossil fuels based plants is the most effective one. The main advantage is that it is
account for 52% of the world current CO2 emission. Petroleum easy to handle the solids with no corrosion problems.[19] Amine
oil, natural gas, and coals are the main fossil fuels burned by containing solvents are generally environmental friendly which
humans. The atmospheric concentration has increased to are generally non-toxic, cheaper and easily available. Hence sol-
384 ppm in 2007 from its pre-industrial level of 280 ppm and is vent containing ammonia group like diethanol amine was taken
expected to reach 550 ppm by 2050 even if CO2 emission is sta- here as impregnating solvent for activated carbon for amine
ble for the next four decades.[1] The present concentration has group. Till now this activated carbon adsorbent has been tested
reached down to 401.52 ppm in March, 2015 based on Scraps in common reactors and single-stage fluidized reactor. To
CO2 Program at the Mauna Loa Observatory in Hawaii. Unfor- improve the efficiency of CO2 adsorption these amine impreg-
tunately there will be no big change in the next few years as far nated adsorbents have been used in our newly designed four-
as energy consumption is concerned. Biomass-based fuels, solar stage fluidized bed reactor. The major advantage of multistage
energy and nuclear energy are the other forms of energy sour- fluidized bed reactor is that the residence time distribution of
ces but they still cannot replace fossil fuels on a larger scale. By solid becomes narrow so flow pattern is plug flow. Also gas
2030 the energy demand will rise by 53%.[2] For reduction of bypassing is reduced. Multistage fluidized bed reactor with a
the atmospheric CO2 concentration attention has been given to down comer using a suitable adsorbent is a very recent concept
develop process equipment for carbon dioxide removal from for removal of carbon dioxide from flue gas.[20–23] But till now
the flue gas in a very cost effective manner to meet the demands no work has been done to capture CO2 from flue gas by amine
of technology and with stringent environmental laws and regu- impregnated activated carbon. The aim of this work was to
lations. The important thing is the selection of a suitable adsor- evaluate the influence of the operation of four-stage fluidized
bent which can adsorb CO2 from the flue gas in a cost effective bed reactor for sorption of CO2 gas on diethanol amine impreg-
manner. Different adsorbents used for CO2 capture are nated activated carbon particle for a wide range of operating
CONTACT Bhim C. Meikap [email protected] Department of Chemical Engineering, School of Engineering, Howard College, University of
Kwazulu-Natal, Durban, South Africa.
Color versions of one or more of the figures in this article can be found online at www.tandfonline.com/lesa.
© 2016 Taylor & Francis Group, LLC
770 D. DAS ET AL.
conditions. DEA impregnated activated carbon is an efficient Experimental setup and procedure
adsorbent for carbon dioxide (CO2) due to the formation of
Four-stage fluidized bed reactor was developed and used in this
more stable carbamate ion.[24,25] So DEA based adsorbed has
study, and the schematic diagram is shown in Figure 1. The fluid-
the ability to adsorb more CO2 compared to non-impregnated
ized bed column consisted of four stages (0.21 m height per stage
activated carbon.
and 0.095 m internal diameter). The stages were assembled
together with a flanged joint. Four stainless steel plates (S1, S2, S3,
Materials and methods S4) of 0.002 m thickness were used as internal baffles between two
stages. Hole of diameter 0.002 m on a triangular pitch arrangement
Preparation of sorbent
was present in each plate. To avoid solids from falling down
Chemicals were supplied by MERCK Specialities Private Limited, through the plate, grid plates were covered with fine weir mesh
Mumbai, India. Green coconut shells were collected from nearby (100 mesh size) with openings smaller than particle size. Down
local market of IIT Kharagpur, then cut into small pieces, and comers (D1, D2, D3, D4) were made up of Perspex cylinder of
washed with tap water to remove dirt followed by drying in the 0.024 m internal diameter and height of 0.265 m. Each section was
sunlight for 15–20 days till it becomes completely dry. Dried mate- provided with down-comers and further fitted with a cone of diam-
rials were kept inside the furnace at 150 C for 24 h to get rid of eter 0.007 m and 0.024 m height at the exit end so that up-flow of
moisture and other volatile impurities. Dried samples were crushed the gas through the down-comer is reduced as a result of which sta-
with a locally made crusher and sieved to a size of 512 mm. Then ble operation is maintained. On the gas distributor, the down-
chemical activation was done with ZnCl2. According to the litera- comers were further fitted in special threading arrangement and
ture chemical activation is better than physical activation in order there was a provision for adjustment of weir height as required.
Downloaded by [University of York] at 23:51 08 June 2016
to avoid high temperature and prolonged time activation inside The weir height was considered to be bed height. The material
furnace. flows from stage to stage through the down-comer. There were
500 gm of dried precursor was mixed well with 2000 mL of provisions for measuring pressure drop. For uniform distribution
concentrated solution of ZnCl2 that contains 500 gm of ZnCl2. of the gas to the fluidization column, gas distributor was present at
The impregnation ratio (activating agent/precursor) was 100%. the bottom of the column. Calibrated rotameter was fitted to mea-
The powdered material in slurry form was properly mixed and sure the air flow rate. A conical hopper was attached at the bottom
kept for 24 h for proper soaking of ZnCl2 on the surface of of the column for storage of the solid. A feeding funnel was present
powdered precursor. The slurry was kept inside the oven at at the top of the column to hold the activated carbon particle and it
100 C for 24 h. The resulting chemical impregnated samples was attached to the screw feeder. Screw feeder was fitted to a motor
were placed inside a galvanized iron pipe of dimensions of of 0.25 HP, and the speed of the motor was controlled by a variable
length 8 cm and inner diameter of 1.5 cm and kept inside the rheostat. 5 HP compressor was used to supply the air as fluidizing
furnace. The material inside the furnace was heated (10 C gas. The solid is fed to the first stage of the down-comer from the
min¡1) to the final carbonization temperature of 600 C under top of the funnel connected to screw feeder and then through Per-
the nitrogen flow rate of 100 cm3 min¡1STP. The material was spex tube (0.011 m internal diameter). The gas leaving the top stage
kept inside the furnace for 1 h at 600 C. Then it was cooled is passed through 0.14 m diameter cyclone. To collect the fines
under the constant flowing of nitrogen gas till the temperature coming out from the fluidized bed, a bag is attached at the bottom
reaches the room temperature. The dried material was washed of the cyclone. The solids were fed from the top through the screw
with 0.5 N HCl for 2–3 times and then washed with warm dis- feeder to the first stage of the down comer of the reactor. Com-
tilled water to remove any kind of residual organic and mineral pressed air at 2 kg/cm2 pressure and 80–150 L/min flow rate was
matter. Then it was finally mixed with cold water till the solu- passed through the pipeline and regulated by valve. At that time
tion becomes neutral. Finally, the sample was dried for 24 h at CO2 from the cylinder was passed through a gas regulator at a cer-
100 C inside an oven till it becomes completely dry. tain flow rate. CO2 and air mixture was passed through the throat
of the ejector and the mixture was fed into the gas chamber at the
bottom stage of the fluidized bed reactor. Counter-current flow of
Preparation of diethanol amine impregnated activated
the solids and the gas flow occurred at each stage. The experiments
carbon
were conducted by setting the gas velocity to 0.188–0.353 m/s cor-
The dried activated carbon was impregnated with diethanol responding to solid flow rate of 2.15–4.12 kg/h. The weir height of
amine solution (HO-CH2-CH2-NH-CH2-CH2-OH) (in the the down comer was kept at 0.03 m and 0.05 m. From the literature
weight ratio 0.4). Then this amine-impregnated activated car- we have studied that the flue gas containing CO2 coming out from
bon was dried in the oven at temperature of 100 C for 48 h till stack of industries generally has very high concentration. Hence we
it becomes dry and kept inside an air tight container for our have taken an approximate range (low concentration to high con-
experiment. centration) in ppm scale of 3000–20000 ppm. For each gas flow
rate, the inlet CO2 loadings were varied from 3000 to 20000 ppm in
Characterization of adsorbents four stages.
The adsorbent used for removal of CO2 in our experimental
setup was diethanol amine impregnated activated carbon
Sampling and analysis
(DEA) of impregnation ratio (IR) 0.4. Physical characteristics
and pore structure parameters of DEA impregnated AC are When all stages of the reactor were identical in their opera-
given in Table 1. tion and the pressure drops across each stage were almost
JOURNAL OF ENVIRONMENTAL SCIENCE AND HEALTH, PART A 771
Table 1. Physical characteristics and pore structure parameter of DEA impregnated AC.
Average particle Density Sphericity Porosity BET SSA VTot VMicro Amicro Avg. pore
Sample diameter (mm) (kg/m3) (Fs) (emf) (m2/g) (cm3/g) (cm3/g) (m2/g) radius (A0)
DEA-AC 512 2870 0.62 0.77 572.776 0.2621 0.198 9.79635 491.324
equal, then it indicated steady and stable operation reactor. CO2 mixture from the cylinder was taken inside Orsat
At that time samples at the inlet and outlet of the column apparatus. By noting the level difference of initial and final
were drawn with the help of aspirator bottles and the marking of measuring burette the volumetric percentage of
obtained CO2 gas samples were analyzed by Orsat analysis. concentration of CO2 has been found out and then the vol-
The gas samples i.e. concentration of CO2 in CO2 C air umetric concentration of CO2 was converted to ppm.
mixture were analyzed for carbon dioxide by the “Orsat
Analysis” method. Simply we took an aspirator bottle filled Orsat analysis method
with kerosene; one end of the bottle was fitted below the Figure 2 shows the Orsat apparatus. It consists of measuring
inlet section of the first stage fluidized bed reactor through burette and absorption pipette. Measuring burette is con-
a pipe and the other end is connected to the collecting jar. nected to leveling bottle which contains a mixture of potas-
By downward displacement of kerosene we collected CO2 C sium dichromate, water, and sodium chloride. Absorption
air mixture into the aspirating bottles. Then this air and pipette contains potassium hydroxide which absorbs CO2. The
Downloaded by [University of York] at 23:51 08 June 2016
aspirating bottle is used to collect and analyze CO2 sample. The % Removal of CO2 in stage i can be calculated as
First the mixture of solution (potassium dichromate, sodium
chloride, water) in measuring burette is adjusted as 100 mL Ci C 1 ¡ Ci
using a leveling bottle by opening and closing the inlet valve. hCO2 D £100 (2)
Ci C 1
The potassium hydroxide level in adsorption pipette is noted,
and then one end of aspirator bottle is connected to the capil-
lary tube and the other end is connected to the leveling bottle where Ci and C iC1 are outlet and inlet carbon dioxide concen-
that contains water. Then the valves of aspirator bottle and trations in gas.
inlet valve are opened. After lifting the water bottle, the read-
ing in burette decreases as CO2 enters into the burette simul-
Results and discussion
taneously. When the reading reaches zero, the inlet valve is
closed. The absorption pipette valve is opened, and then the Effect of gas flow rate on the percentage removal
leveling bottle is adjusted for 30–50 times till all the CO2 gas efficiency of CO2
samples are absorbed in the KOH. After that the valve was
Figure 3 represents the percentage removal efficiency of CO2 at
closed and the final reading is noted. The change in reading
different inlet CO2 loadings and at different gas flow rates. It
gives the volume % of CO2 absorbed by the KOH solution.
may be seen from the figure that increasing the inlet CO2 load-
Then it was converted to ppm. The percentage removal of
ing decreases the removal efficiency of CO2 at a particular solid
CO2 has been calculated for each experimental run by Equa-
velocity and weir height. The maximum removal efficiency of
tion 1:
carbon dioxide for inlet concentration 3000 ppm was 75.8% at
50 mm weir height with solid flow rate 2.15 kg/h. The mini-
CO2 ¡ CO2
inlet outlet
% Removal of CO2 D £ 100 (1) mum removal efficiency of carbon dioxide for inlet concentra-
CO2 inlet tion 20000 ppm was 70.9% at 50 mm weir height under the
JOURNAL OF ENVIRONMENTAL SCIENCE AND HEALTH, PART A 773
Figure 4. Effect of gas velocity on CO2 removal efficiency at hw D 50 mm and Figure 5. Effect of solid flow rate on CO2 removal efficiency at hw D 50 mm and
Sa D 4.12 kg/h at Ug D 0.188 and 0.353 m/s. Ug D 0.188 m/s at Sa D 2.15 and 4.12 m/s.
774 D. DAS ET AL.
Downloaded by [University of York] at 23:51 08 June 2016
the solid hold in each stage increases. These increases in hold- Effect of weir height on the percentage removal efficiency
up of solids augment to adsorb more quantity of CO2 on DEA of CO2
impregnated activated carbon particle. Figures 7 and 8 represent the effect of inlet CO2 concentration on
the percentage removal efficiency of CO2 at different weir heights.
It can be seen from these figures that the percentage removal of
CO2 at the higher weir height was maximum as the solid reactants
available was maximum. At 50 mm weir height, solid flow rate of
2.15 kg/h, and the gas velocity 0.353 m/s the highest removal effi-
ciency was 74.2% for 3000 ppm inlet CO2 concentration and
66.5% for 20000 ppm inlet concentration. At solid flow rate of
4.12 kg/h and the gas velocity 0.188 m/s the range of removal effi-
ciencies was 79.6% for 3000 ppm inlet CO2 concentration and
72.3% for 20000 ppm inlet concentration. As observed, increasing
the weir height increased the removal efficiency due to increase in
bed volume resulting in more gas-solid interaction. However, the
effect of weir height at lower concentration was not as much as
observed at higher concentration indicating the presence of less
quantity of reactive solids at lower height.
The actual reaction is shown for DEA in presence of solvent [7] Mazumder S.; Van Hemert P.; Busch A.; Wolf K.A.; Tejera-Cuesta P.
water: Flue gas and pure CO2 sorption properties of coal: a comparative
study. Int. J. Coal Geology. 2006, 67, 267–279.
OH ¡ CH2 ¡ CH2 ¡ NH ¡ CH2 ¡ CH2 ¡ OH C CO2 [8] Saghafi A.; Faiz M.; Roberts D. CO2 storage and gas diffusivity prop-
erties of coals from Sydney Basin, Australia. Int. J. Coal Geology.
$ OH ¡ CH2 ¡ CH2 ¡ NH C ¡ COO ¡ ¡ CH2 ¡ CH2 ¡ OH 2007, 70, 240–254.
[9] Manovic V.; Anthony E. Sequential SO2/CO2 capture enhanced by
(2 carbamate i.e. zwitterion) steam reactivation of a CaO-based sorbent. Fuel. 2008, 87, 1564–
1573.
OH ¡ CH2 ¡ CH2 ¡ NH C ¡ COO ¡ ¡ CH2 ¡ CH2 ¡ OH [10] Nikulshina V.; Gebald C.; Steinfeld A. CO2 capture from atmo-
spheric air via consecutive CaO-carbonation and CaCO3-calcina-
C OH ¡ $ OH ¡ CH2 ¡ CH2 ¡ N ¡ COO ¡ tion cycles in a fluidized-bed solar reactor. Chem. Eng. J. 2009,
¡ CH2 ¡ CH2 ¡ OH C H2 O 146, 244–248.
[11] Salvador C.; Lu D.; Anthony E.J.; Abandes JC. Enhancement of CaO
The rate determining step is the 1st which results in the for- for CO2 capture in an FBC environment. Chem. Eng. J. 2003, 96,
mation of a secondary carbamate ion for diethanol amine solu- 187–195.
tion. The second step occurs rapidly in the presence of base. So [12] Seo Y.; Jo S.H.; Ryu C.K.; Yi CK. Effects of water vapor pretreatment
the tendency to adsorb more CO2 depends critically on the first time and reaction temperature on CO2 capture characteristics of a
sodium-based solid sorbent in a bubbling fluidized-bed reactor. Che-
step. Thus more is the stability of carbamate ion formed more mosphere. 2007, 69, 712–718.
is the forward side reaction and hence adsorption. Due to this [13] Yi C.K.; Jo S.H.; Seo Y.; Lee J.B.; Ryu CK. Continuous operation of
chemistry, DEA based adsorption has the ability to adsorb the potassium-based dry sorbent CO2 capture process with two fluid-
more CO2 compared to non-impregnated activated carbon. ized-bed reactors. Int. J. Greenhouse Gas Control. 2007, 1, 31–36.
Downloaded by [University of York] at 23:51 08 June 2016