introduction - university of edinburgh · web viewcoal-fired power plants produce about one third...

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

mailto:[email protected]

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



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




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


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





Blanks

Figure 7. Ni/DOBDC stability results after exposure to humidity (1% v/v water,


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



Blank

Figure 8. Mg/DOBDC stability results after exposure to humidity (1% v/v water,


0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 500 1000 1500 2000 2500

Rel

ativ

e C

apac

ity


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





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





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



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




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


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

introduction - university of edinburgh · web viewcoal-fired power plants produce about one third...

Documents