Rotary wheel adsorber for carbon capture

Transcription

1 Rotary wheel adsorber for carbon capture E. Mangano 1, E. Shiko 1,A. Greenaway 3, A. Gibson 2, A. Gromov 2, M. M. Lozinska 3, E. Campbell 2, P. A. Wright 3, S. Brandani 1 1 University of Edinburgh, School of Engineering; University of Edinburgh, School of Engineering, Edinburgh SCCS Scottish Carbon Capture and Storage Centre 2 University of Edinburgh, School of Chemistry; 3 University of St. Andrews; e.mangano@ed.ac.uk s.brandani@ed.ac.uk Best paper of the Carbon Management: Recent Advances in Carbon Capture, Conversion, Utilization and Storage session at ACS Fall meeting 2015 EPSRC: EP/J02077X/1

2 Partners: Adsorption Materials and Processes for Gas fired power plants AMPGas The University of Edinburgh (Coordinator) Industrial Partner: University of St. Andrews Heriot Watt University Other industrial contributions: Howden Group Ltd Chemviron Carbon; Purolite; Thomas Swan and UOP 2

3 AMPGas Project Aims: Apply a range of experimental techniques to determine equilibrium and kinetic properties of nanoporous materials, which are being developed for CO 2 capture from dilute streams; Predict the performance of an integrated adsorption process based on rapid thermal swing; Demonstrate the proposed process using a bench scale rotary wheel adsorber. Materials: Thanks to the expertise of the partners different materials can be tested: Zeolites (St. Andrews University) Amine containing MOFs (St. Andrews University) Amine based Silicas (Heriot Watt University & St. Andrews University) Amine containing Carbon and Carbon Nanotubes (University of Edinburgh) 3

4 Adsorbents: the challenges Tailoring novel adsorbents for CO 2 separation from dilute streams (4 9% CO 2 ): Structures Physisorption (optimised zeolites) Cation types Cation distribution Hydrophilicity Supporting material Pore size/volume Chemisorption (amine based adsorbents ) Amine groups Synthesis process Chemical and thermal stability Behaviour in presence of water 4

5 Porous Materials Cation Gating Zeolites Small Pore MOFs Aluminosilicates which contain 8 membered windows, such as RHO, ECR 18 Extra framework cations adopt positions in window sites Co operative interaction between CO 2 and cation causes cation to move allowing CO 2 to pass through window Metal organic frameworks consist of organic linker groups coordinated to metal clusters (nodes) Ability to modify both metal source and organic linker makes MOFs highly versatile Functionalisation of organic linkers can have a profound effect on the adsorption properties of the material. Highly selective for CO 2 / N 2 and CO 2 / CH 4 5

6 Cation Gating Zeolites Recently discovered which can be used to separate adsorbates with differing physical properties, for example the separation of CO 2 from N 2. This is a separate mechanism from the more common molecular sieving separations that zeolites are commonly used for. Has potential to be used for size inverse separations. As molecules with a larger polarizability, dipole and quadrupole moments interact more strongly with the cation. M. M. Lozinska, E. Mangano, et al. J. Am. Chem. Soc. 2012, 134,

7 Cation Gating Zeolites: Na RHO Simple zeolite structure, with lta cages connected via d8r cages. Exhibits promising gas adsorption / separation properties but slow Na M. M. Lozinska, E. Mangano, et al. J. Am. Chem. Soc. 2012, 134, M. M. Lozinska, P. A. Wright, et al. Chem. Mater., 2014, 26,

8 Cation Gating Zeolites: Expanding the Rho series ECR 18 is the synthetic form of zeolite Paulingite All openings 8 rings 7 different cage types. Similar interpenetrated backbone to Rho 8

9 Cation Gating Zeolites: ECR 18 Na,H ECR Relative intensity (a.u.) Time (s) g CO 2 /N 2 (50:50) 20 cc/min 303 K Structural changes of synthetic paulingite (Na,H ECR 18) upon dehydration and CO 2 adsorption A. G. Greenaway, J. Shin, P. A. Cox, E. Shiko, S. P. Thompson, S. Brandani, S. Bong Hong, P. A. Wright*, Zeit. Krist. Crystalline Materials, 2015, 230,

10 ZSM 25 Solved structure of ZSM 25 : related to ECR 18 and Rho ECR 18 ZSM 25 A zeolite family with expanding structural complexity and embedded isoreticular structures P. Guo, J. Shin, A. G. Greenaway, J. Gi Min, J. Su, H. J. Choi, L. Liu, P. A. Cox, S. B. Hong, P. A. Wright, X. Zou Nature, 2015, 524,

11 ZSM 25 Adsorption Isotherms on ZSM 25 Adsorption isotherms at 298 K of CO 2, CH 4 and N 2 Good regenerability. Inset shows capacities over 100 cycles Kinetics of CO 2 uptake NaTEA ZSM 25, NaTEA ECR 18, Na Rho and K chabazite Presence of secondary cage structure speeds up adsorption in Na form of the zeolites ECR 18 and ZSM 25 11

12 Small Pore MOFs Model Sc 2 BDC 3 series is hydrophobic Isostructural series with functional groups Resulting MOFs exhibit different adsorption properties Amine functionalised most selective CO 2 at 273 K N 2 at 77 K 12

13 CO 2 adsorption on Sc 2 (NH 2 BDC) 3 Thermodynamics from isotherms C/C Kinetics (Zero Length Column) 11 ml/min 21 ml/min 32 ml/min blank 11 ml/min blank 21 ml/min blank 32 ml/min C/C ml/min 21 ml/min 32 ml/min blank 11 ml/min blank 21 ml/min blank 32 ml/min t(min) Ft(ml) Desorption curves of CO 2 from Sc 2 (BDC NH 2 ) 3 and an empty ZLC column at different flowrates. The normalized decrease in concentration (C/C 0 ) is plotted against a) time (t) and b) Ft scales. Heat of adsorption 31(±3) kj mol 1 Desorption is under equilibrium conditions even at fastest flow rate In situ synchrotron IR microspectroscopy of CO 2 adsorption on the functionalised MOF Sc 2 (BDC NH 2 ) 3 A. Greenaway, B. Gonzalez Santiago, P. M. Donaldson, M. D. Frogley, G. Cinque, J. Sotelo, S. Moggach, E. Shiko, S. Brandani, R. F. Howe and P. A. Wright Angew. Chem. Int. Ed. 2014, 53,

14 Amine functionalised carbons: preparation 3 Main types of material prepared: 1. CNTs grafted with a basic amino functionalities 2. Activated carbon grafted with amino functionalities 3. A physical impregnation of amino groups to the surface of two different types of porous carbon Amine Grafted carbon nanotubes (CNT CO NHR) Amine grafted porous carbon Amine impregnated porous carbon (various loadings) EDA DETA TETA PEI (MW600) PEI (MW10000) PEI (MW750000) Advantage: Carbon materials can be heated ohmically for cyclic regeneration of the adsorbent 14

15 Multi walled carbon nanotubes (MWCNT) Carbon nanotubes are cylindrical allotropes of carbon Two varieties: single walled and multi walled Large specific surface area ranging from m 2 g 1 Surface can be readily functionalised to modify the material s properties Functionalization methods can be applied to cheaper activated carbon for large scale gas separation Thermal cyclic regeneration through ohmic heating The highly ordered structure makes particularly interesting the study of the kinetics 15

16 Functionalization of Carbon Materials Scheme: Functionalization of carbon nanotubes with basic amine moieties Image: MWCNT/agarose aerogel produced by lyophilisation Scheme LHS: Carbon nanotubes can be functionalized for selective carbon capture Image RHS: Functionalized CNTs can be utilised to create 3D structures with high specific surface area 16

17 Physical impregnation of amine onto porous carbons N 2 isotherms, 77K STP (cc/g) BET (m 2 g -1 ) micro-ac 1336 micro-ac-teta(10) 953 micro-ac-teta(30) 219 micro-ac-teta(50) 24 micro-ac-teta(70) 6.7 STP (cc/g) BET (m 2 g -1 ) meso-ac 816 meso-ac-teta(10) 510 meso-ac-teta(30) 274 meso-ac-teta(50) 162 meso-ac-teta(70) Relative Pressure (P/P 0 ) DFT: Pore size distribution Pore Volume (cc/g) meso-ac meso-ac-teta(10) meso-ac-teta(30) meso-ac-teta(50) meso-ac-teta(70) Relative Pressure (P/P 0 ) Pore Volume/ cc g 1 micro AC meso AC Raw TETA TETA TETA TETA Pore width (nm) 17

18 CO 2 Uptake Temperature Dependence Thermogravimetric analysis (TGA): micro AC Pore width 2 nm 35 C 50 C 75 C micro AC TETA 50 Pore width 2 nm 35 C 50 C 75 C 90 C q/ mmol g q (mmol g -1 ) Time/ s Raw AC material CO 2 capacity as temperature Time (Hours) Impregnated AC material CO 2 capacity as temperature ΔH ADS = 27 kj mol 1 Physisorption Heat of Adsorption > 50 kj mol 1 chemisorption Heat of Adsorption < 50 kj mol 1 physisorption ΔH ADS = 90 kj mol 1 Chemisorption 18

19 CO 2 uptake of carbon supports loaded with various wt % of TETA (a) q (mmol g -1 ) micro AC TETA Pore width 2 nm micro-ac-teta(10) micro-ac-teta(30) micro-ac-teta(50) micro-ac-teta(70) (b) q (mmol g -1 ) meso AC TETA Pore width 2 10 nm meso-ac-teta(10) meso-ac-teta(30) meso-ac-teta(50)-run1 meso-ac-teta(50)-run2 meso-ac-teta(75) Time (hours) Time (hours) J.A.A. Gibson, A.V. Gromov, S. Brandani, E.E.B. Campbell, Microporous and Mesoporous Materials, 208 (2015)

20 CO 2 Cyclic Experiments q (mmol g -1 ) meso AC PEI600, 4 cycles 0.1 bar CO 2, 75 C time (hours) q Heat Heat flow (mw) PEI 9% drop in capacity over 4 cycles N H 14 meso AC TETA 30, 4 cycles 0.1 bar CO 2, 75 C q Heat flow q (mmol g -1 ) Heat (mw) time (hours) 20

21 Experimental approach for novel adsorbents Zero Length Column Extended Zero Length Column Dual Piston PSA Rotary Wheel Adsorber mg ~ 50 mg ~ 10 g ~ 1 Kg Equilibrium Kinetics Stability Clear separation Useful for TSA evaluation Heat transfer Mass transfer Pressure drop Full cycle performance Purity Recovery Productivity 21

22 Real rotary system 22

23 Testing Novel Adsorbents: the Zero Length Column An experimental technique should allow us to: Rank CO 2 capacity of materials rapidly Require only small samples Interpret the results easily Allow to determine kinetics Allow to test the materials with water Allow to test the materials with SO x and NO x Signal A properly designed ZLC system can deliver 0 on all of these requirements Full sat. Time, s Partial sat. 23

24 Experimental approach for novel adsorbents Ranking Good sample Bad sample Equilibrium Kinetics Stability 24

25 Ranking of CO 2 capacity for Amine based Carbons (UoE) micro AC meso AC micro AC TETA(10) micro AC TETA(30) micro AC TETA(50) micro AC TETA(65) meso AC TETA(10) meso AC TETA(30) meso AC TETA(50) meso AC TETA(75) meso AC PEI600(20) meso AC PEI600(40) meso AC PEI600(60) meso AC PEI600(75) meso AC PEI600(100) meso AC 2 PEI600(290) q (mmol g 1 ) Conditions: 75 C, 10% CO 2, 90% helium 25

26 Ranking of CO 2 capacity for Zeolites (UoStA) T = 35 C; P CO2 = 0.1 atm

27 Equilibrium controlled experiment 1 1 C/C ml/min 21 ml/min 32 ml/min blank 11 ml/min blank 21 ml/min blank 32 ml/min C/C ml/min 21 ml/min 32 ml/min blank 11 ml/min blank 21 ml/min blank 32 ml/min t(min) Ft(ml) 27

28 ZLC Partial loading experiment on Na Rho (1μm) T = 35 C; P CO2 = 0.1 atm R 2 /D = 167 min 28

29 Experimental approach for novel adsorbents Ranking Good sample Bad sample Equilibrium Kinetics Stability 29

30 ZLC scale up: Extended ZLC for breakthrough experiments Sample: meso AC PEI1200(200) Conditions: 75 C, 10% CO 2 90%N 2 Model under development to fit E ZLC data Capacity: 1.8 mmol g 1 Evidence of chemical reaction Allows to run breakthrough experiments with ~ 50 mg of sample 30

31 Breakthrough experiment on Li Rho CO 2 /CH 4 selectivity P tot = 1 atm T = 35 C Y CO2 = 0.05 Y CH4 = 0.4 Carrier gas: He 31

32 Breakthrough experiments on Paulingite Adsorption CO 2 /N 2 selectivity P tot = 1 atm Y CO2 = Y N2 = 0.3 in He Evidence of structural change 32

33 Experimental approach for novel adsorbents Good sample Ranking Good sample Bad sample Equilibrium Kinetics Stability 33

34 Process scale up: Dual Piston PSA Benefits of DP PSA Direct test of the separation performance Single column required Closed system with total reflux; only small amount of gas needed Rapid testing of adsorbent materials Many different experiments are possible Particularly suitable to measure kinetic and equilibrium properties of novel adsorbent materials Dosing system Oven Pistons Aim Testing of novel materials for the separation of CO 2 from flue gas Column Pressure readings 34

35 Experimental approach for novel adsorbents Good sample Equilibrium Ranking Good sample Kinetics Stability All information for the model acquired! Bad sample 35

36 Adsorption model hierarchy Micropore Equilibrium Complete model Isothermal Non- Isothermal: 1T Non- Isothermal: 2T Non- Isothermal: 3T Pressure drop No Pressure drop No Film resistance Film resistance No Macropore Macropore LDF Macropore Diffusion Dusty Gas Model Micropore LDF Micropore Diffusion MS-Surface diffusion Now including also Ideal Adsorption Solution Theory methods for multicomponent adsorption 36

37 General adsorption cycle simulator Column 1 Adsorption Feed Pressurisation Column 2 PE Evacuation Purge PE Adsorption systems Multiple adsorption columns Connected by splitters, mixers, valves and tanks Series of cycle steps: pressurisation, feed, purge, Extend column simulation to general adsorption cycles Modular system with different units: adsorption columns, valves, splitters, tanks,... Arbitrary number and connection of the units Simulate different cycle configurations by time events, e.g. switching of valves 37

38 Buffer unit for unibed approach All columns cycle through the same steps Steps with interaction between two columns Output of one column is input of the other column Add a buffer unit for each interaction pair Data in buffer unit is half a cycle out of date Same result at Cyclic Steady State Order of magnitude faster 38

39 Experimental approach for novel adsorbents Good sample Equilibrium Ranking Good sample Kinetics Stability All information for the model acquired! Bad sample 39

40 Rotary Wheel Adsorber for carbon capture Advantages Due to very low concentration of CO 2 thermal swing adsorption is required for rapid regeneration of the adsorbent. A properly designed rotary wheel adsorber: Can treat large volumes of gas Lower capital cost (no multiple columns, piping, valves, etc ) Efficient heat integration Low pressure drop Can perform rapid temperature swings Thermal cycles of few minutes: 10 times faster than traditional TSA in fixed bed Significant reduction of the size of the capture plant 40

41 Bench scale Rotary Wheel Adsorber for carbon capture Rotating part Stationary part Stationary part 12 columns rotary system Each column is detachable and can be independently tested Up to 24 thermocouples (2 per column) Large amount of data to be sent in real time Max. rotational speed 1 rpm Regeneration using electrical heating elements One of the first LiFi communication on moving elements Real time computer for data acquisition and control of the system 41

42 Bench scale Rotary Wheel Adsorber for carbon capture 42

43 Experimental setup being built at UoE Rotating can and adsorption columns Sector plate PTFE sector plate sealing 43

44 RWA concept system control LED ring TC data + position LiFi receiver Gas MFC NI CRIO real time computer D P transducers Slip rings for H E 60 W AC motor 0 1 rpm 44

45 Preliminary simulations Adapting Cysim cycle simulator for the simulation of some base scenarios: Column: L = 20 cm, ID = 2 cm. Adsorbents: 0.05 bar TRI PE MCM 41(Y. Belmabkhout, et al., 2010) 13X (isotherms measured in our lab) P tot = 1 bar Feed concentration: 5% CO 2 in N 2 Adsorption:35 C Regeneration methods: Electrical heating; Steam 45

46 Case 1 : Electrical heating TRI PE MCM 41 13X 46

47 Case 1 : temperature distribution TRI PE MCM 41 13X 47

48 Case 1 : concentration distribution TRI PE MCM 41 13X 48

49 Case 2 : Purge with steam TRI PE MCM 41 13X 49

50 Case 2 : temperature distribution TRI PE MCM 41 13X 50

51 Case 2 : concentration distribution TRI PE MCM 41 13X 51

52 LiFi how does it work? On Off Intensity Spectrum: Unregulated (free) Huge Safe Existing Infrastructure Inexpensive devices Time Prof. Harald Haas, Dr. Stefan Videv 52

53 Recent hero demonstrations 5 mw 3.5 Gbps from single color LED at 5mW 1.1 Gbps at 10 m at 5mW Prof. Harald Haas, Dr. Stefan Videv 53

54 Conclusions Several materials have been developed and tested using different techniques (ZLC, TGA, Breakthrough) Some of the amine functionalised carbons show a clear chemisorption process Some of the zeolitic frameworks show evidence of structural modification associated to the presence of CO 2 A novel bench scale rotary wheel adsorber has bee designed and is being built at the UoE CySim is being modified to predict the performance of the bench scale prototype A novel LiFi communication system (one of the first on moving elements) is being developed for the data acquisition in the RWA 54

55 Acknowledgments We gratefully acknowledge EPSRC for funding the AMPGas project (EP/J02077X/1) 55