introduction - university of edinburgh · web viewcoal-fired power plants produce about one third...
TRANSCRIPT
Testing the Stability of Novel Adsorbents for Carbon Capture
Applications using the Zero Length Column Technique
Xiayi Hua,c, Stefano Brandania*, Annabelle I. Beninb and Richard R. Willisb
a. Institute for Materials and Processes, School of Engineering, the University of Edinburgh, UK
b. New Materials Research, UOP LLC, a Honeywell Company, Des Plaines, IL, USA
c. College of Chemical Engineering, Xiangtan University, Hunan Province, P. R. China
ABSTRACT
In this paper, a semi-automated ZLC technique was used to study the stability of
novel adsorbents in the presence of water, SOx and NOx impurities present in coal-
fired power plant flue gas. The tests were carried out at 38°C and 0.1bar pressure on
the most promising materials of the M/DOBDC ((M = Co, Ni, Mg)) MOF series, as
well as on commercial 13X zeolite pellets. The experimental results indicated that
even if the DOBDC family shows high CO2 adsorption capacity at low partial
pressure at ambient conditions, impurities have a strong effect on their stability. The
ZLC system provides quantitative information on the deactivation of samples due to
SOx and NOx in a relatively short time and using less than 15 mg of sample. The fact
that the treatment can be repeated in situ in the apparatus used also for measuring the
CO2 capacity of the sample has shown that the ZLC can be a valuable tool in
screening novel adsorbents for carbon capture applications.
Keywords: Zero length column (ZLC), adsorption, MOFs, CO2 capture, stability, flue
gas, SOx, NOx.
*Corresponding author.
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Email address: [email protected]
1. Introduction
Coal-fired power plants produce about one third of CO2 emissions; therefore, the
capture of CO2 from the flue exhaust of power plants may play a significant role in
mitigating the effects of CO2 on climate (Willis et al., 2010), since fossil fuels will
still be a dominant source of energy supply in the foreseeable future in China and the
USA. Typically the flue gases produced by coal fired power plants contain 12 to 16%
CO2 by volume at ambient conditions, and impurities including water, O2, HCl,
hydrocarbons, CO, SOx and NOx (Granite and Pennline, 2002).
Adsorption in porous materials is known to be a potential technology for post-
combustion capture because of the lower energy requirement of physisorption, but
traditional zeolites may be limited by their high regeneration temperature (Harlick and
Tezel, 2004), and the presence of water in flue gas can significantly affect their CO2
capacities (Brandani et al., 2004). Metal organic frameworks (MOFs) are known for
their large surface areas, controllable pore structures, and versatile chemical
compositions (Férey, 2008), some of them show extremely high CO2 capacity and
very desirable isotherm shapes, and have therefore drawn considerable attention as
potential materials for CO2 capture (Dybtsev et al., 2003; Llewellyn et al., 2006;
Natesakhawat et al., 2006). Given the large number of possible novel MOF materials
2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
which can be investigated, it is important to develop rapid screening techniques which
can assess experimentally the potential of these novel materials. In previous
contributions the use of a purposely designed Zero Length Column (ZLC) system has
been shown to provide a valuable tool to determine both adsorption capacities (Hu et
al., 2015a) and mass transfer kinetics (Hu et al., 2015b) using less than 15 mg of
sample. The methodology was applied to the recently developed M/DOBDC (M =
Zn, Ni, Co, Mg) MOFs, which are considered to be potential candidates for carbon
dioxide capture because of their extra high CO2 uptake, especially at low CO2 partial
pressures (Millward and Yaghi, 2005; Mofarahi et al., 2008; Yazaydın et al., 2009).
For MOFs to be useful in an industrial application, it is important to understand how
stable these materials will be when in contact with impurities present in the flue gas.
Liu et al. (2010) investigated the effect of water in the adsorption of CO2 on Ni-MOF-
74 and HKUST-1 at 0.1 atm and 25 ºC. Ni-MOF-74 exhibited a higher degradation
resistance, maintaining most of its CO2 capacity after several cycles of water exposure
and thermal regenerations. Similarly, Kizzie et al. (2011) found that despite the higher
CO2 capacity, the Mg-MOF-74 sample exhibited the highest loss of capacity after
exposure to water, while Co- and Ni-MOF-74 resulted the most stable samples
maintaining 85% and 60% of the original capacity, respectively, after exposure to a
mixture with a relative humidity of 70%. More recently experiments on MOFs
stability under flue gas exposure have been presented by (Han et al. (2012); Sang et
al. (2010)). The experiments were carried in a high throughput apparatus consisting of
3
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
36 sample cells each of them provided with a pressure sensor. Small changes in the
capacity were observed for most of the samples after the short term exposure to humid
air and acid gases, while a significant loss in the CO2 adsorption was registered after
the long term exposure. Mangano et al. (2016) used the ZLC to clarify the effect of
impurities and humid flue gas, on the CO2 adsorption of Mg- and Ni-CPO-27 MOF
crystals by exposing the samples to a synthetic flue gas mixture containing water. The
experimental results showed that Mg-MOFs exhibited the rapid deactivation simply
with water present, while Ni-CPO-27 exhibited an initial decrease followed by a more
stable performance through several cycles.
In this contribution, the ZLC is used to study the effect of water and SOX and NOX
impurities on the stability of M/DOBDC and commercial zeolites. The key
advantages of this approach are the use of small samples and the ability to determine
the amount of impurities that are adsorbed by the solid.
2. Materials and Methods
2.1. Adsorbent Synthesis
Most MOF materials are synthesized through solvothermal reactions. The synthesis
procedures for the seven MOF samples used in our research are modified from the
literature (Hu et al., 2015a).
Mg/DOBDC (DOBDC = dioxybenzenedicarboxylate) was synthesized by the
Matzger group at the University of Michigan. Linker precursor 2, 5-
4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
dihydroxyterephthalic acid (2.98 g) was dissolved in 180 mL 1-methyl-2-
pyrrolidinone (NMP) + 20 mL deionized water using sonication. Mg (OAc) 2. 4H2O
(6.44 g) was added to a 0.5 L jar. The 2, 5-dihydroxyterephthalic acid solution was
filtered into the jar. The solution was sonicated to dissolve completely the magnesium
acetate and the jar was placed in a 120 °C oven for 20 hours. The mother liquor was
decanted while still hot and the yellow solid was washed with methanol (3 × 200 mL).
The methanol was decanted and replaced with fresh methanol (200 mL) three times
over the course of three days. The yellow solid was evacuated at 250 °C for 16 hours
and then transferred to a glove box for storage. One batch produced ~3 g of
Mg/DOBDC.
Ni/DOBDC, nickel(II) acetate (18.7 g, 94.0 mmol, Aldrich) and 2,5-
dihydroxyterephthalic acid (DOBDC, 37.3 g, 150 mmol, Aldrich) were placed in 1 L
of mixed solvent consisting of equal parts tetrahydrofuran (THF) and deionized water.
The mixture was then put into a 2 L static Parr reactor and heated at 110 °C for three
days. The as-synthesized sample was filtered and washed with water. Then the sample
was dried in air, and the solvent remaining inside the sample was exchanged with
ethanol six times over eight days. Finally, the sample was activated at 150°C under
vacuum with nitrogen flow.
Co/DOBDC, the reaction of cobalt (II) acetate and 2, 5-dihydroxyterephthalic acid
(C8H6O6) in a mixture of water and tetrahydrofuran (molar ratio 2:1:556:165) under
autogenous pressure at 110°C in a Teflon-lined autoclave (50 percent filling level).
5
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
This yielded a pink-red, needle-shaped crystalline substance [Co2(C8H2O6)
(H2O)2]8H2O. After the hypothetical removal of the noncoordinating solvent
molecules in the channels, the empty channels occupy 49 percent of the total volume
of the unit cell, and the average cross-sectional channel dimensions are 11.08×11.08
Å2. The empty volume increases to 60 percent if the coordinating water is also
removed. The water molecules excluding and including the coordinating water
account for 29.2% and 36.6% of the mass, respectively.
The 13X pellets are commercial materials. 13X-APG adsorbent is the sodium form of
X zeolite possessing an aluminosilicate binder.
2.2. ZLC Experimental Setup
The ZLC apparatus is described in detail in our previous contribution (Hu et al.,
2015a). Figure 1 shows a schematic diagram of the new semi-automated ZLC
apparatus used in our experiments to test the stability of novel adsorbents.
6
1
2
3
4
5
6
7
8
9
10
11
12
13
14
TCD
ZLC column
Capsule
Vent
ZLC oven
Mass spectometer
PC
Preheat oven
On-Off Valves
Oven (Dosing volume)
Mass flow controllers
High flow rate
Low flow rate
Digital pressure transducer
Cylinders (4L)
Hel
ium
Car
bon
Dio
xide
Nitr
ogen
Drying columns
Vacuum pump
Differential pressure
transducer
Flue gas
Vent
Oth
er
Figure 1. Schematic diagram of the experimental system showing details of the
developed semi-automated zero length column system.
For the stability experiments the key feature is the gas dosing system which is
designed to control the amount of water in the mixture. The dosing system includes
four stainless steel cylinders (1L each) inside an oven (Sanyo, Japan) which allows
keeping the gas mixtures at a temperature sufficiently high to avoid any condensation
of water. Water is added to the dosing volumes using a detachable capsule similar to
that described by Zielinski et al. (2007). Initially the dosing volume is evacuated and
filled to approximately 0.5 bar with either the carrier gas or the gas mixture and the
valve that connects the capsule to the dosing cylinders is opened slowly. The amount
of water dosed is checked by a mass balance with the equilibrium pressure in the
system and by detaching the capsule and recording the mass change of the capsule
using a Mettler-Toledo balance (Model No.XS205). Mixtures with known
7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
concentration of water vapour can be prepared accurately up to 1% v/v. This limit is
imposed in order to avoid condensation of water in the mass flow controllers. Given
the high affinity to water of all the materials studied, this concentration allows to
reach near saturation in the adsorbed phase, ie a higher partial pressure of water
would not change the adsorbed phase water concentration.
In the flow control part, a low flow rate of 1-5 cc/min was applied to the
measurement, and all the lines in the system were heated to 90 C to avoid water
condensation.
Four different mixtures were prepared in the dosing volume (i) 16% v/v CO2 in
nitrogen, which is for the CO2 capacity measurement on the adsorbent; (ii) 1% v/v
water in nitrogen, which is for the initial water stability measurements, (iii) 1% v/v
water and 16% v/v CO2 in nitrogen, for a further stability measurement; (iv) 16% v/v
CO2, 100 ppm SO2 and 10 ppm NO, with balanced of N2, supplied by BOC, being
utilized. Water could not be included in the mixture supplied by BOC, since at high
pressure (200 bar) it would have condensed in the cylinder leading to safety and
corrosion issues. Since the ZLC system contains an independent dosing volume, a
known concentration of water was added to the flue gas. Oxygen was not included
since it is incompatible with SO2, but this would likely not have a major effect, since
adsorption of O2 and N2 on MOFs are similar. The characteristics of the synthetic wet
flue gas are shown in Table 1.
8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Table 1. Key wet flue gas characteristics for ZLC tests
Flue gas CO2 H2O NO SO2 N2
Concentratio
n16% 1% 10ppm 100ppm Balance
The ZLC cell is made of a Swagelok 1/8 in. fitting and the 10-15mg of sample is
sandwich packed by two porous sinter discs in the fitting and weighted in a Mettler-
Toledo XS205 dual-range balance. Prior to the experiment, the ZLC column was
thermally regenerated with a ramping rate of 1°C/min to the required temperature, as
described by Hu et al. (2015a), and held at this temperature for 12h at a very low flow
rate of the helium purge gas, 1 cc/min. After regeneration, the oven temperature was
reduced to the desired temperature (38 °C) automatically. A typical ZLC experiment
was then carried out on a fresh sample, which was first equilibrated with nitrogen
stream containing 16% of CO2 that was prepared in the dosing volume. At time zero,
the flow was switched to a pure nitrogen purge stream at the same volumetric flow
rate. For each sample, the ZLC experiment runs at 2 different flow rates performed in
the range of 1-3 cc/min and from the desorption curve the CO2 adsorption capacity
was determined (Brandani et al., 2003; Hu et al., 2015a; Wang et al., 2011).
Thereafter, the sample was exposed to the relevant gas mixture for 24 hours, followed
by intermediate regeneration (same procedure as fresh sample). The ZLC runs were
carried out on the adsorbent to evaluate the residual CO2 capacity. If the sample was
9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
still active, a further 24 hours exposure was applied until the sample became inactive
or until it maintained a constant capacity.
3. Theory
3.1. ZLC measurement
In a ZLC experiment, the column is short enough to be treated as a well-mixed cell
with negligible external mass and heat transfer resistance. In a desorption experiment
the differential mass balance is therefore given by:
V sd q̄dt
+V gdcdt
+Fc=0 (1)
where Vg is the volume of gas; Vs represents the volume of solid; F is the volumetric
flowrate; q̄ is the average adsorbed phase concentration and c is the sorbate
concentration in the gas phase. During the experiment the gas phase concentration is
measured and the carrier flowrate is assumed constant.
With the assumption that equilibrium is maintained between the gas and the solid
phase; ideal gas behaviour; isothermal system; and a constant carrier flowrate, ie
calculating the total flowrate as F = Fc/(1y), Eq. (1) is integrated to obtain:
q¿=(∫0∞ y
1− yd( Fc t
V S )−∫0
ty
1− yd ( Fc t
V S )−V g
V S⋅y ) C
(2)
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Where is the flowrate of the carrier gas, y is the mole fraction and C is the total
concentration which can be calculated from the ideal gas law, C= P
RT . While the
flowrate of the carrier gas is not strictly constant, Eq. 2 is normally valid up to mole
fractions of the adsorbate of 0.4-0.5. Integration of Eq. 2 allows to calculate either the
total capacity or the full adsorption isotherm.
3.2. Analysis as breakthrough curves
The ZLC can also be seen as a breakthrough experiment, especially for strong
adsorbed components, i.e. SO2, based on the mass balance, the total accumulation in
the column can be written as:
[ε dcdt
+(1−ε ) d qdt ]V =( Fc )IN −( Fc )OUT
(3)
With the assumption that gas and solid phase are at equilibrium; ideal gas behaviour;
and an isothermal system, the eq. (3) can be rewritten in terms of the carrier gas
flowrate and the adsorbate mole fraction as
[ ε+(1−ε )K ]V d y
dt=(FCarr
y1− y )
IN−(FCarr
y1− y )
OUT (4)
Where K=
(q1−q0)(c1−c0) represents the slope of the secant of the adsorption isotherm
between the two end points.
11
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
The retention volume for a packed ZLC column, which is easily determined from the
first moment of the response at a known flow rate is directly related to the
dimensionless Henry constant, Assuming a constant carrier flowrate, eq. (4) is
integrated to obtain:
−[ε+(1−ε ) K ]=FCarr
( y1− y0 )V ∫0
∞ ( y1
1− y1− y
1− y )dt (5)
Eq. (5) is only suitable for small step changes and the carrier flowrate can be assumed
to be constant.
4. Results and Discussion
4.1 Water stability
Figure 2 shows the ZLC response curves at 38 °C for Co/DOBDC after exposure to
mixtures of 1% v/v water in nitrogen. Since the area between the desorption curve and
the blank response is proportional to the CO2 capacity, it is possible to see clearly the
effect of water on the stability of this material. Figure 3 shows the results for
Ni/DOBDC which is very stable in the presence of water. Figure 4 shows the results
for Mg/DOBDC which is better than the cobalt sample, but loses half of its CO2
capacity after the first 24 hours of exposure. Thereafter, the rate of loss in capacity
decreases.
Figure 5 shows the relative CO2 capacity obtained from the mass balance of the ZLC
(Brandani et al., 2003; Brandani, 1998; Hu et al., 2015a) versus specific gas volume
12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
passed over the sample, i.e. the gas flowrate times the exposure time divided by the
sample mass. The figure shows the quantitative comparison between the three MOFs.
Although these three MOFs are synthesized via solvothermal reactions, Co/DOBDC
and Mg/DOBDC clearly lack stability to moisture. These two materials are therefore
very unlikely candidates for carbon capture applications.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 5 10 15 20 25 30 35 40Ft, cc
c/c 0
Fresh Sample
Exposure for 24 hours
Blanks
Figure 2. Co/DOBDC stability results after exposure to humidity (1% v/v water
in nitrogen) gas. Ft plot gives a direct CO2 capacity comparison with different
exposure work and also blanks.
13
1
2
3
4
5
6
7
8
9
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 2 4 6 8 10 12 14Ft, cc
c/c 0
Fresh Sample
Exposure for 24 hours
Exposure for 48 hours
Blanks
Figure 3. Ni/DOBDC stability results after exposure to humidity (1% v/v water
in nitrogen) gas.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 5 10 15 20Ft, cc
c/c 0
Fresh Sample
Exposure for 24 hours
Exposure for 48 hours
Exposure for 72 hours
Blanks
Figure 4. Mg/DOBDC stability results after exposure to humidity (1% v/v water
in nitrogen) gas.
14
1
2
3
4
5
6
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 500 1000 1500 2000 2500
Rel
ativ
e C
apac
ity
Specific Gas Volume, m3/kg
1% Wet Gas Exposure withCo/DOBDC Powder
1% Wet Gas Exposure with Ni/DOBDCPowder
1% Wet Gas Exposure withMg/DOBDC Powder
Figure 5. Comparison of the stability tests with humidity gas (1% v/v water in
nitrogen).
4.2 Water and CO2 stability
The three MOFs were then exposed to 1% water, and 16% CO2 in nitrogen. The
results which can be seen in Figures 6-8 show a similar trend observed for the water
stability, ie the Ni/DOBDC sample is the most stable. What is important to note is that
in the presence of water the resulting acidic environment does have an impact and
accelerates the deactivation for all the samples, including the Ni/DOBDC. This effect
was not observed in previous studies.
A quantitative comparison of the three MOFs is shown in Figure 9.
15
1
2
3
4
5
6
7
8
9
10
11
12
13
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 5 10 15 20 25 30 35 40Ft, cc
c/c 0
Fresh Sample
Exposure for 24 hours
Blanks
Figure 6. Co/DOBDC stability results after exposure to humidity (1% v/v water,
16% v/v CO2 in nitrogen) gas.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 2 4 6 8 10 12 14Ft, cc
c/c 0
Fresh Sample
Exposure for 24 hours
Exposure for 48 hours
Exposure for 72 hours
Exposure for 96 hours
Blanks
Figure 7. Ni/DOBDC stability results after exposure to humidity (1% v/v water,
16% v/v CO2 in nitrogen) gas.
16
1
2
3
4
5
6
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 5 10 15 20Ft, cc
c/c 0
Fresh Sample
Exposure for 24 hours
Exposure for 48 hours
Blank
Figure 8. Mg/DOBDC stability results after exposure to humidity (1% v/v water,
16% v/v CO2 in nitrogen) gas.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 500 1000 1500 2000 2500
Rel
ativ
e C
apac
ity
Specific Gas Volume, m3/kg
1% Wet Gas(16%CO2) Exposure withCo/DOBDC Powder
1% Wet Gas(16%CO2) Exposure withNi/DOBDC Powder
1% Wet Gas(16%CO2) Exposure withMg/DOBDC Powder
Figure 9. Comparison of the stability tests with humidity gas (1% v/v water, 16%
v/v CO2 in nitrogen).
17
1
2
3
4
5
6
4.3 Wet flue gas stability
An important feature of the ZLC system is the fact that the gas at the outlet can be
monitored continuously. This allows to ensure that during the exposure the sample
reaches equilibrium, i.e. that enough impurities are adsorbed. When using very dilute
impurities one must ensure that the sample saturates over the time of exposure.
Figure 10 shows a typical MS signal (CO2, H2O, SO2 and NO) for the Co/DOBDC
sample during the first 24 hours exposure. It is clear that CO2 and NO reach
equilibrium in a few minutes. For the more strongly adsorbed species the ZLC is
behaving like a normal breakthrough apparatus and more than 8 hours are needed to
reach equilibrium for H2O and SO2. Similar curves were obtained for the different
samples.
1.00E-15
1.00E-14
1.00E-13
1.00E-12
1.00E-11
1.00E-10
1.00E-09
1.00E-08
1.00E-07
0 4 8 12 16 20 24
MS
sign
al, a
mps
Exposure Time, hour
Mass 44 CO2
Mass 18 H2O
Mass 30 NO
Mass 64 SO2
18
1
2
3
4
5
6
7
8
9
10
11
12
Figure 10. Monitored MS signal (CO2, H2O, NO, SO2) of first 24 hours wet (1%
v/v water) flue gas exposure for Co/DOBDC.
The results of wet flue gas stability on M/DOBDC MOFs and zeolite 13X are shown
in Figures11-14. From Figure 11, one can see that the behaviour of the Co/DOBDC
sample is significantly different from the other two stability tests. What is interesting
in this case is the fact that some residual capacity is observed and the shape of the
isotherm has changed. The effect of the wet flue gas seems to result in stronger
adsorption sites, so the isotherm, after the first treatment, shifts to a non-linear one.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 5 10 15 20 25 30Ft, cc
c/c 0
Fresh Sample
Exposure for 24 hours
Exposure for 48 hours
Exposure for 72 hours
Exposure for 96 hours
Blanks
Figure 11. Co/DOBDC stability results after exposure to wet (1% v/v water) flue
gas.
The Ni/DOBDC sample showed excellent stability in water, with full regeneration at
150 °C (Figure 11). When the samples were tested with wet flue gas there was a
progressive loss in capacity. What is interesting to note from the stability results is
19
1
2
3
4
5
6
7
8
9
10
11
12
13
14
that the capacity loss seems to be uniform, i.e. at each extended exposure more
material is deactivated but the general shape of the isotherm and of the ZLC response
remains the same.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 5 10 15 20Ft, cc
c/c 0
Fresh Sample
Exposure for 24 hours
Exposure for 48 hours
Exposure for 72 hours
Exposure for 96 hours
Blanks
Figure 12. Ni/DOBDC stability results after exposure to wet (1% v/v water) flue
gas.
Figure 13 shows the results for the Mg/DOBDC sample. There is an additional effect
due to the flue gas, but it is clear that in this case most of the deactivation is due to the
effect of water. For this sample, the shape of the ZLC curve does not change
qualitatively.
Figure 14 shows the results for zeolite 13X plotted in semi-logarithmic form because
the effect of purities in wet flue gas on this zeolite was very small. There is some loss
in the first 24 hours of exposure, followed by a slower decay.
20
1
2
3
4
5
6
7
8
9
10
11
12
13
14
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 5 10 15 20Ft, cc
c/c 0
Fresh Sample
Exposure for 24 hours
Exposure for 48 hours
Blanks
Figure 13. Mg/DOBDC stability results after exposure to wet (1% v/v water) flue
gas.
0.01
0.1
1
0 10 20 30 40 50 60Ft, cc
c/c 0
Fresh Sample
Exposure for 24 hours
Exposure for 48 hours
Exposure for 72 hours
Blanks
Figure 14. semi-logarithmic plot of 13X stability results after exposure to wet
(1% v/v water) flue gas.
21
1
2
3
4
5
6
Since the ZLC is behaving like a normal breakthrough apparatus in this case, the first
moments (mean residence time) analysis was applied to the SO2 breakthrough curve
(i.e. Figure 10) for the three M/DOBDC MOF and 13X samples, which is related the
adsorption capacity. For Co/DOBDC, it is clear in Figure 15 that there is a very sharp
reduction after the first exposure. However, in the case of Ni/DOBDC (Figure 16), the
SO2 capacity follows a similar trend as the CO2 capacity. In Mg/DOBDC (Figure 17),
a clear reduction of SO2 capacity is shown as well. For zeolite 13X (Figure 18), the
SO2 capacity also shows a slower decay as CO2 capacity measured after exposure.
SO2 Adsorption Capacity (Co/DOBDC)
0
1
2
3
4
5
6
24 48 72 96
Time, h
t, s
(10*
4)
AdsorptionCapacity
Figure 15. Moment analysis on SO2 breakthrough curves for Co/DOBDC
during long time wet flue gas exposure.
22
1
2
3
4
5
6
7
8
9
10
11
12
13
SO2 Adsorption Capacity (Ni/DOBDC)
0
1
2
3
4
5
6
7
24 48 72 96
Time, h
t, s
(10*
4)
AdsorptionCapacity
Figure 16. Moment analysis on SO2 breakthrough curves for Ni/DOBDC during
long time wet flue gas exposure.
SO2 Adsorption Capacity (Mg/DOBDC)
0
1
2
3
4
5
6
7
24 48 72 96
Time, h
t, s
(10*
4)
AdsorptionCapacity
Figure 17. Moment analysis on SO2 breakthrough curves for Mg/DOBDC during
long time wet flue gas exposure.
23
1
2
3
4
5
6
7
SO2 Adsorption Capacity (13X)
0123456789
10
24 48 72 96
Time, h
t, s
(10*
4)
AdsorptionCapacity
Figure 18. Moment analysis on SO2 breakthrough curves for 13X during long
time wet flue gas exposure.
To confirm that the samples had degraded, the ZLC columns treated with the wet flue
gas were sealed and sent for further characterisation X-Ray Diffraction (XRD) to
compare the crystallinity of the freshly activated sample as well as the effect of flue
gas.
The typical XRD patterns shown in Figure 19 give the comparison of fresh activated
Mg/DOBDC and the sample after being exposed to wet flue gas for a long time. The
significant reduction in peak intensity for samples exposed to the flue gas compared
to the fresh activated sample indicates the transition to an amorphous structure. This
24
1
2
3
4
5
6
7
8
9
10
11
12
confirms the reason of the CO2 capacity reduction after the exposure experiment
compared to the original sample.
Figure 19. XRD spectra for Mg/DOBDC original sample (up), exposed to 1% wet
flue gas at 38ºC, 3cc/min for 72h.
Figure 20 shows the quantitative comparison of effect of the wet flue gas on CO2
adsorption for M/DOBDC powders and 13X zeolite pellets. It is obvious that the flue
gas does not affect CO2 adsorption in 13X as much as for all the M/DOBDC MOFs.
Zeolite 13X only showed some loss of capacity after the first exposure, followed by a
slow secondary decay.
25
1
2
3
4
5
6
7
8
9
10
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 500 1000 1500 2000
Rel
ativ
e C
apac
ity
Specific Gas Volume, m3/kg
1% Wet Flue Gas Exposure with Co/DOBDC Powder1% Wet Flue Gas Exposure with Ni/DOBDC Powder1% Wet Flue Gas Exposure with Mg/DOBDC Powder1% Wet Flue Gas Exposure with 13X Powder
Figure 20. Comparison of the stability tests with wet (1% v/v water) flue gas.
5. Conclusions
The ZLC apparatus has been shown to provide an effective system to test the stability
of novel M/DOBDC MOF materials subject to various mixtures related to flue gases
of coal-fired power plants. The use of very small samples results also in a much
reduced consumption of gases for the testing procedures in comparison to traditional
breakthrough experiments. The fact that the exposure to impurities and water can be
repeated in situ in a relatively simple way has shown that the technique is an
important tool in the screening of novel adsorbents for carbon capture applications.
By monitoring directly also the impurities, the ZLC allows to determine the time
26
1
2
3
4
5
6
7
8
9
10
11
needed to saturate the samples with impurities, pointing to an important factor to
consider when comparing different studies on the stability of novel materials.
All the DOBDC MOFs were found to be strongly affected by either water (Mg and
Co forms) or wet flue gas conditions (Ni/DOBDC). The lack of water stability tends
to exclude Mg/DOBDC and Co/DOBDC as potential materials in practical
applications. The progressive deactivation of Ni/DOBDC under acidic conditions
emphasises the need to include guard beds in the design of a carbon capture system
based on this material.
Acknowledgements
This project was supported by the U.S. Department of Energy through the National
Energy Technology Laboratory under Award No. DE-FC26-07NT43092, National
Natural Science Foundation of China (21506179) and Research Foundation of
Education Bureau of Hunan Province, China(17B255). However, any opinions,
findings, conclusions, or recommendations expressed herein are those of the authors
and do not necessarily reflect the views of the DOE.
References
Brandani, F., Rouse, A., Brandani, S., Ruthven, D.M., 2004. Adsorption Kinetics and
Dynamic Behavior of a Carbon Monolith. Adsorption 10, 99-109.
27
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Brandani, F., Ruthven, D., Coe., C.G., 2003. Measurement of Adsorption Equilibrium
by the Zero Length Column (ZLC) Technique Part 1: Single-Component
Systems. Ind. Eng. Chem. Res 42, 1451-1461.
Brandani, S., 1998. Effects of nonlinear equilibrium on zero length column
experiments. Chem. Eng. Sci 53, 2791-2798.
Dybtsev, D.N., Chun, H., Yoon, S.H., Kim, D., Kim, K., 2003. Microporous
manganese formate: a simple metal-organic porous material with high framework
stability and highly selective gas sorption properties. J. Am. Chem. Soc 126, 32-
33.
Férey, G., 2008. Hybrid porous solids: past, present, future. Chem. Soc. Rev 37, 191-
214.
Granite, E.J., Pennline, H.W., 2002. Photochemical Removal of Mercury from Flue
Gas. Ind. Eng. Chem. Res 41, 5470-5476.
Han, S., Huang, Y., Watanabe, T., Dai, Y., Walton, K.S., Nair, S., Sholl, D.S.,
Meredith, J.C., 2012. High-Throughput Screening of Metal-Organic Frameworks
for CO2 Separation. ACS Comb. Sci 14, 263-267.
Harlick, P.J.E., Tezel, F.H., 2004. An experimental adsorbent screening study for CO2
removal from N2. Micropor. Mesopor. Mat 76, 71-79.
Hu, X., Brandani, S., Benin, A.I., Willis, R.R., 2015a. Development of a Semi-
Automated Zero Length Column Technique for Carbon Capture Applications:
Rapid Capacity Ranking of Novel Adsorbents. Ind. Eng. Che. Res 54, 6772-6780.
Hu, X., Brandani, S., Benin, A.I., Willis, R.R., 2015b. Development of a
Semiautomated Zero Length Column Technique for Carbon Capture
Applications: Study of Diffusion Behavior of CO2 in MOFs. Ind. Eng. Che. Res
54, 5777-5783.
Kizzie, A.C., Wong-Foy, A.G., Matzger, A.J., 2011. Effect of Humidity on the
Performance of Microporous Coordination Polymers as Adsorbents for CO2
Capture. Langmuir : the ACS journal of surfaces and colloids 27, 6368-6373.
28
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Liu, J., Wang, Y., Benin, A.I., Jakubczak, P., Willis, R.R., Levan, M.D., 2010.
CO2/H2O adsorption equilibrium and rates on metal-organic frameworks:
HKUST-1 and Ni/DOBDC. Langmuir : the ACS journal of surfaces and colloids
26, 14301-14307.
Llewellyn, P.L., Bourrelly, S., Serre, C., Filinchuk, Y., Férey, G., 2006. How
Hydration Drastically Improves Adsorption Selectivity for CO2 over CH4 in the
Flexible Chromium Terephthalate MIL-53 †. Angewandte Chemie 45, 7751-
7754.
Mangano, E., Kahr, J., Wright, P.A., Brandani, S., 2016. Accelerated degradation of
MOFs under flue gas conditions. Faraday Discussions 192, 181-195.
Millward, A.R., Yaghi, O.M., 2005. Metal-organic frameworks with exceptionally high
capacity for storage of carbon dioxide at room temperature. J. Am. Chem. Soc
127, 17998-17999.
Mofarahi, M., Khojasteh, Y., Khaledi, H., Farahnak, A., 2008. Design of CO2
absorption plant for recovery of CO2 from flue gases of gas turbine. Energy 33,
1311-1319.
Natesakhawat, S., Culp, J.T., Christopher Matranga, A., Bockrath, B., 2006.
Adsorption Properties of Hydrogen and Carbon Dioxide in Prussian Blue
Analogues M3[Co(CN)6]2, M = Co, Zn. J. Phys. Chem. C 111, 1055-1060.
Sang, S.H., Choi, S.H., Duin, A.C.T.V., 2010. Molecular Dynamics Simulations of
Stability of Metal—Organic Frameworks Against H2O Using the ReaxFF
Reactive Force Field. Chemical Communications 46, 5713-5715.
Wang, H., Brandani, S., Lin, G., Hu, X., 2011. Flowrate correction for the
determination of isotherms and Darken thermodynamic factors from Zero Length
Column (ZLC) experiments. Adsorption 17, 687-694.
Willis, R.R., Benin, A., Snurr, R.Q., Yazaydın, Ö., 2010. Nanotechnology for Carbon
Dioxide Capture, in: Garcia-Martinez, J. (Ed.), Nanotechnology for the Energy
Challenge. Chemistry International, Weinheim, pp. 359-401.
29
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Yazaydın, A.Ö., Benin, A.I., Faheem, S.A., Jakubczak, P., Low, J.J., Willis, R.R.,
Snurr, R.Q., 2009. Enhanced CO2 Adsorption in Metal-Organic Frameworks via
Occupation of Open-Metal Sites by Coordinated Water Molecules. Chem. Mater
21, 1425-1430.
Zielinski, J.M., Peter Mckeon, A., Kimak, M.F., 2007. A Simple Technique for the
Measurement of H2 Sorption Capacities. Ind. Eng. Chem. Res 46, 329-335.
30
1
2
3
4
5
6