F. Akhtar*ac,
N. Keshavarzia,
D. Shakarovaa,
O. Cheungab,
N. Hedinab and
L. Bergströma
aDepartment of Materials and Environmental Chemistry, Stockholm University, Stockholm 10691, Sweden. E-mail: farid.akhtar@ltu.se; Tel: +46 920 491793
bBerzelii Center EXSELENT on Porous Materials, Stockholm University, Stockholm 10691, Sweden
cDivision of Materials Science, Luleå University of Technology, Luleå 97187, Sweden
First published on 20th October 2014
Monoliths of microporous aluminophosphates (AlPO4-17 and AlPO4-53) were structured by binder-free pulsed current processing. Such monoliths could be important for carbon capture from flue gas. The AlPO4-17 and AlPO4-53 monoliths exhibited a tensile strength of 1.0 MPa and a CO2 adsorption capacity of 2.5 mmol g−1 and 1.6 mmol g−1, respectively at 101 kPa and 0 °C. Analyses of single component CO2 and N2 adsorption data indicated that the AlPO4-53 monoliths had an extraordinarily high CO2-over-N2 selectivity from a binary gas mixture of 15 mol% CO2 and 85 mol% N2. The estimated CO2 capture capacity of AlPO4-17 and AlPO4-53 monoliths in a typical pressure swing adsorption (PSA) process at 20 °C was higher than that of the commonly used zeolite 13X granules. Under cyclic sorption conditions, AlPO4-17 and AlPO4-53 monoliths were regenerated by lowering the pressure of CO2. Regeneration was done without application of heat, which would regenerate them to their full capacity for CO2 adsorption.
Microporous powders are usually structured into granules and beads because beds of micron-sized powders exhibit very large pressure drops. A large pressure drop can result in clogs or blockages in some gas separation processes.8,9 Typically, structured adsorbents are produced by shaping a mixture of the porous powder with an inorganic and organic binder into a body of the desired geometry.10–12 The powder body is thermally treated to increase the mechanical strength. While aluminophosphates, silicoaluminophosphates and other ortho and pyro-phosphates powders have been processed to produce hierarchically porous catalytic supports with catalytic components, (e.g. Pt, Ag on zeolite Y, and zeolite ZK-5),13–17 reports on structuring of AlPO4-n powders to produce structured adsorbents are sparse.18
The efficiency of an adsorbent is decreased when the active component – the microporous powder – is diluted by a large proportion of an inert binder. Moreover, the inert binder is selectively removed in a chemically corrosive environment.19 The selective removal of the binder can result in lowered mechanical stability of the structured adsorbent. The presence of the inert binder may also alter the adsorptive properties of the structured adsorbent.8 Hence, there is a great interest in developing binder-less processing routes that can produce mechanically strong structured adsorbents with maximized volume efficiency. Previously, it was demonstrated in an elaborate, multi-step process, that clay and silica binders can be converted into active porous materials by hydrothermal treatment.20–24 We developed a versatile pulsed current processing (PCP) route to directly produce mechanically strong, yet binder-free, hierarchically porous monoliths from e.g. microporous zeolite, mesoporous silica and macroporous diatomite powders.25–29
In this work, we demonstrate that mechanically stable and binder-less structured adsorbents of AlPO4-17 and AlPO4-53 can be produced by PCP. With the optimized PCP temperature and pressure, the PCP-produced monoliths displayed a high CO2 capture capacity and outstanding CO2-over-N2 selectivity. The CO2 working capacity in a typical PSA process was evaluated and compared with commercial granules of zeolite 13X. Cyclic adsorption capacity and regeneration conditions to a full CO2 adsorption capacity were determined for the AlPO4-17 and AlPO4-53 monoliths as well.
Nitrogen adsorption–desorption experiments were performed at −196 °C on a Micrometrics ASAP2020 surface area analyzer (Micromeritics, Norcross GA, USA). The specimens were degassed at high vacuum (1 × 10−4 Pa) at 300 °C for 6 hours. The Brunauer–Emmet–Teller (BET) surface area was calculated using the nitrogen uptake of the specimen in the relative pressure range of 0.05–0.15 p/po. The CO2 and N2 adsorption measurements were performed on a Micrometrics Gemini V 2390 apparatus (Micromeritics, Norcross GA, USA) equipped with the room temperature add-on. CO2 and N2 adsorption measurements were recorded at 0 °C and 20 °C within a pressure range from 0 to 101 kPa. Isothermal conditions (±0.1 °C) were maintained by a circulating bath (Huber Ministat 230) which contains a low molecular weight siloxane polymer. The temperature in the Dewar flask was measured by an external thermocouple and cross-calibrated to that of the circulating bath. Prior to adsorption measurements, the calcined AlPO4-n powders and the consolidated monoliths were pre-treated under a flow of dry N2 gas at a temperature of 300 °C for 8–10 h. The cyclic performance of the monoliths was tested by recording the CO2 uptake of the samples after regeneration by only vacuum at room temperature.
The traditional Langmuir isotherm model with two parameters was used to describe the adsorption isotherms of CO2 and N2. The traditional Langmuir isotherm model can be written as:3
(1) |
Monoliths | AlPO4-17mPCP650 | AlPO4-53mPCP400 |
---|---|---|
a BET surface area is calculated from N2 adsorption data recorded at −196 °C.b Langmuir surface area is calculated from CO2 adsorption data at 0 °C.c Macropore volume, macroporosity, and median pore diameter are determined by mercury intrusion porosimetry. | ||
aSBET (m2 g−1) | 464 | 223 |
bSLangmuir (m2 g−1) | 548 | 256 |
cVMacropore (cm3 g−1) | 0.26 | 0.32 |
cMacro-porosity (vol.%) | 31.0 | 41.0 |
cMedian pore diameter (μm) | 3.50 | 0.70 |
Mechanical strength (MPa) | 1.05 ± 0.10 | 0.85 ± 0.10 |
The SEM micrographs of the monoliths (Fig. 1) show that the rod-like AlPO4-17 and polyhedral AlPO4-53 crystals have retained their well-defined and faceted morphologies after PCP. Fig. 1 display pores in-between the crystals in the PCP AlPO4s. The median diameters of these macropores have been quantified by mercury intrusion porosimetry (Table 1), which show that the macropores are significantly larger in the AlPO4-17 monolith when compared to the AlPO4-53 monolith. Large macropores are advantageous as they limit the pressure drop over an adsorption column and enhance the mass transport of gas molecules.30,32–34
Fig. 1 Scanning electron micrographs from fractured surfaces of: (a) AlPO4-17mPCP650; (b) AlPO4-53mPCP400. The insets show photographs of the monoliths. |
An optimum balance between adsorption activity, mass transfer and mechanical stability is required for gas separation by swing adsorption processes.8,33,35 The combination of an optimized temperature and pressure during PCP has resulted in structured monoliths of AlPO4-17 and AlPO4-53 with a high CO2 capture capacity and a relatively high mechanical strength (Table 1). The CO2 and N2 uptake on these monoliths of AlPO4-17 and AlPO4-53 (Fig. 2) show that the PCP temperature has significantly influenced the capacity for adsorption of CO2. The AlPO4-17 monoliths that have been treated by PCP at 650 °C have only a 12% reduced capacity as compared to the powder, Fig. 2a and b. We ascribe this minor reduction in the CO2 uptake to the bonding of AlPO4-17 crystals at contact points. These contact points can alter the local microporous structure27,28 during the PCP treatment by local amophization or phase transformation to a non-adsorbing phase. If the temperature during the PCP is higher than 650 °C, the capacities of the AlPO4-17 monoliths are further reduced. AlPO4-17 based monoliths that have been consolidated at 750 °C can only adsorb 1.2 mmol g−1 of CO2. Similarly, the AlPO4-53 monoliths that have been subjected to PCP at 400 °C and 50 MPa pressure also have only a slight decrease in the capacities to adsorb CO2 and N2, Fig. 2c and d. The N2 uptake on AlPO4-17 and AlPO4-53 powders and monoliths is small, as expected from the low electric quadrupole moment of N2 (−4.6 × 10−40 C m−2) compared to CO2 (−14 × 10−40 C m−2).36
When the AlPO4-17(p) and AlPO4-53(p) have been processed above the optimal temperature for PCP, the microporous powders transformed into a new dense AlPO4 phase and a crystalline AlPO4 phase, respectively (Fig. 3). These monoliths then have negligible capacities to adsorb CO2 and N2 (Fig. 2). Analyses of the XRD data (Fig. 3) show that AlPO4-17 transforms into a sodalite AlPO4 phase and that AlPO4-53 transforms into tridymite AlPO4 phase at 950 °C. It should be mentioned that although the sodalite structure is porous, the pore windows are too small for diffusion of CO2 and N2 molecules.
The CO2 adsorption capacity at 0 °C of the AlPO4-17mPCP650 is 2.5 mmol g−1 and 1.65 mmol g−1 for the AlPO4-53mPCP400 at a partial pressure of 100 kPa (see Fig. 2). The N2 adsorption is low on both monoliths, but is significantly lower on the AlPO4-53mPCP400, 0.05 mmol g−1 (100 kPa) than that of the AlPO4-17mPCP650, 0.24 mmol g−1 (100 kPa). The difference in CO2 and N2 adsorption capacity of AlPO4-17mPCP650 and AlPO4-53mPCP400 suggest that AlPO4-17mPCP650 have high capacity for CO2 adsorption and AlPO4-53mPCP400 display an extraordinary high CO2-over-N2 selectivity.
CO2 selectivity is an important requirement for separation of CO2 from flue gas. We can obtain a simple estimate of the CO2-over-N2 selectivity, SCO2/N2, for a typical flue gas mixture that contains 15 mol% CO2 and 85 mol% N2 as the ratio of equilibrium mole fraction of CO2 adsorbed at 15 kPa over the equilibrium mole fraction of N2 adsorbed at 85 kPa , as follows.
(2) |
The SCO2/N2 at 20 °C is 17 for AlPO4-17mPCP650 and 102 for AlPO4-53mPCP400. The high CO2-over-N2 selectivity of AlPO4 has been ascribed to kinetic or molecular sieving effects.3,37 Molecular sieving or kinetic effects are related to the kinetic diameter of CO2 (3.3 Å) and N2 (3.64 Å) and the size of the 8 ring window of AlPO4-17 and AlPO4-53. AlPO4-17 is a 8-ring aluminophosphate with a window size of 3.6 × 5.1 Å2.38 AlPO4-53 is also a 8-ring aluminophosphate with a window size of 4.3 × 3.1 Å2 along [100] direction,38 which is close to the kinetic diameter of CO2 molecule. More physically correct estimates of the thermodynamic selectivity make use of ideal adsorbed solution (IAS) theory developed by Myers and Prausnitz.39–41 It allows estimation of the co-adsorption equilibriums for CO2 and N2 mixtures from the single component isotherms of CO2 and N2. In IAST, Myers and Prausnitz39 defined selectivity within a two phase model as the ratio of mole fraction of CO2 in the adsorbed state (xCO2) over the mole fraction of CO2 (yCO2) in the gas phase divided by the same relative fractions for N2 (xN2,yN2).
(3) |
Table 2 shows that the binary CO2-over-N2 selectivity is very high for AlPO4-53mPCP400. Hence, the favourable pore window dimensions of AlPO4-53 are preserved after PCP and the monolith could selectively retard the diffusivity of N2 and reduce its uptake kinetically or by molecular sieving.3,26,40,42 The AlPO4-17mPCP650 has high CO2 uptake but the CO2-over-N2 selectivity is significantly smaller compared to AlPO4-53mPCP400. CO2 separation in industrial practice, e.g. flue gas scrubbing and natural or biogas upgrading, is considered economically feasible by adopting pressure swing adsorption (PSA) or vacuum swing adsorption process (VSA) at moderate temperatures, providing that the adsorbent has a low pressure drop, high working capacity and also small sensitivity to water adsorption.43–46 In the interest of potentially using AlPO4 monoliths in swing adsorption carbon capture processes (PSA or VSA), the amount of CO2 that can be removed per kilogram of structured AlPO4 monoliths needs to be determined. We have estimated the CO2 capturing capacities in an idealized PSA process using the CO2 adsorption isotherms in Fig. 2. The estimates in Fig. 4 assume that the flue gas from a small scale combustion plant contains 15 mol% CO2 and 85 mol% N2. The total pressure of the process is assumed to swing from 1 bar to 6 bar in a simple and hypothetical PSA process. The temperature of the flue gas is assumed to be 20 °C, this is mainly because CO2 adsorption data are quite commonly reported at this temperature. We consider a small scale combustion plant only, as we doubt that it will be economic or technically possible to compress the full flue gas stack. In this gas mixture, the partial pressure of CO2 at 1 bar and 6 bar in the flue gas corresponds to 0.15 bar and 0.90 bar, respectively. For these conditions, the CO2 capture capacity can be defined as the difference between CO2 uptake at 0.9 bar and 0.15 bar and represent the moles of CO2 gas that can be removed in a PSA cycle per kilogram of the adsorbent.
Monoliths | Adsorbate | qma (mmol g−1) | bb (1 kPa−1) | KH (qm × b) CO2 | KH (qm × b) N2 | KH CO2/KH N2 | Binary selectivity (αCO2/N2)c |
---|---|---|---|---|---|---|---|
a Obtained from fitting adsorption isotherm at 293 K by Langmuir model.b Obtained from fitting adsorption isotherm at 293 K by Langmuir model.c Calculated by ideal adsorption solution theory at 100 kPa in CO2 and N2 binary mixture of composition 15 mol% CO2 and 85 mol% N2. | |||||||
AlPO4-17mPCP650 | CO2 | 4.389 | 0.0067 | 0.029 | — | 12.78 | 15 |
AlPO4-17mPCP650 | N2 | 1.095 | 0.0021 | — | 0.0023 | ||
AlPO4-53mPCP400 | CO2 | 1.878 | 0.0157 | 0.0295 | — | 99.33 | 2800 |
AlPO4-53mPCP400 | N2 | 0.0799 | 0.0039 | — | 0.0003 |
Fig. 4a and b show the CO2 capture capacity of AlPO4-17mPCP650 and AlPO4-53mPCP400 in the highlighted region that corresponds to these PSA conditions, i.e. varying pressure from 1 bar to 6 bar. The CO2 capture capacity of monoliths of AlPO4-17mPCP650 is 1.4 mmol g−1. The CO2 capture capacity is comparable to that of several MOFs with large uptakes of CO2, e.g. MOF 5,46 Mg-MOF-74,46 MOF-50847 and higher compared to zeolite adsorbents48,49 (Table 3).
Adsorbent (monoliths/powder) | CO2 capture capacity (mmol g−1) |
---|---|
AlPO4-17mPCP650 | 1.4 |
MOF-5 | 0.7 |
Mg-MOF-74 | 2.1 |
MOF-508 | 1.6 |
Silicalite | 0.6 |
HZSM-5 | 1.0 |
The CO2 capture capacity of monoliths of AlPO4-53mPCP400 at 20 °C is 0.8 mmol g−1. The CO2 capture capacity is lower than AlPO4-17mPCP650, however the CO2 selectivity is higher on AlPO4-53mPCP400. When compared with the CO2 capture capacity of commercial granules of zeolite 13X, AlPO4-17mPCP650 and AlPO4-53mPCP400 display higher CO2 capacities. Zeolite 13X is widely researched and accepted as a standard material with a potential use in CO2 capture.30,31,50,51 The CO2 capture capacity of 13X granules is 0.7 mmol g−1 and 0.67 mmol g−1 in the first and second adsorption cycle, respectively. The total CO2 adsorption capacity of 13X granules is reduced irreversibly after the first cycle from 2.9 to 2.5 mmol g−1 (Fig. 4c). This irreversible reduction is probably related to chemisorption of CO2 on 13X granules as the reduced CO2 capacity between the two cycles cannot be overcome without high temperature regeneration.50,51 It should be noted that the flue gas contains water vapors which could reduce the CO2 capture capacity of AlPO4 monoliths and 13X granules further. However, it has been reported that aluminophosphates are less hydrophilic at lower partial pressure of water vapors.2,52 Liu et al.3 reported that the water adsorption capacity of AlPO4-17, AlPO4-53 and 13X powders was 0.17, 0.15 and 0.37 g g−1, respectively. Zeolite 13X has a so-called type-1 water adsorption isotherm53 and the AlPO4-ns have so-called type-V water adsorption isotherms.52 These shape differences implied that AlPO4-17 and AlPO4-53 not only adsorbed less water in the low pressure region compared to 13X zeolite, but that they are significantly less hydrophilic. Therefore, there will be no or only a slight reduction in the CO2 capture capacity from (wet) flue gas. 13X granules show limited CO2 capture capacity under PSA conditions (Fig. 4c) however they may be more useful materials in a VSA process where the pressure varies between 0.01 to 0.3 bar (ref. 48) for CO2 capture from dry flue gas.
The cyclic adsorption performance of an adsorbent is an important property for its long term usage. Fig. 5 shows that the CO2 capture capacity of AlPO4-17mPCP650 (Fig. 5a) and AlPO4-53mPCP400 (Fig. 5b) monoliths do not show any significant changes over five adsorption cycles. The monoliths have been regenerated by lowering the pressure (near-vacuum conditions) and without applying heat. The unchanged CO2 capture capacities are attributed to the absence of chemisorption on the less hydrophilic framework of AlPO4-17 and AlPO4-53 materials.3,52 The cyclic adsorption performance of AlPO4 monoliths is superior to zeolite 13X, which loses a fraction of its CO2 adsorption capacity after first adsorption cycle (Fig. 4c). Typically, zeolites require heating to a high temperature for regeneration to their full CO2 adsorption capacity.26,42,51 Overall, the AlPO4-17mPCP650 and AlPO4-53mPCP400 show high mechanical stability, high CO2 capture capacity in PSA, low hydrophilicity, cyclic performance and easy regeneration. These render them as potential materials with good CO2 capture capacities, long life time and low cost for CO2 capture from flue gas. They are in particular interesting for CO2 removal processes that can tolerate a pressurization step.
This journal is © The Royal Society of Chemistry 2014 |