Aqueous-phase reforming a pathway to chemicals and fuels

Alexey Kirilin Aqueous-phase reforming a pathway to chemicals and fuels Laboratory of Industrial Chemistry and Reaction Engineering Process Chemistry Centre Åbo Akademi

Agenda 2 Short Introduction, biomass, what is APR? Main tasks of the doctoral research Experimental methods Choice of catalytic systems Catalytic results Summary

Introduction - Biomass 3

Catalytic transformation of biomass 4 Hydrolysis Cellulose Hemicellulose Hydrolysis Oxidation Oligomers Aldoses Hydrogenation Sugar acids Chemicals Sugar alcohols Esterification Aqueous reforming Fuels Lubricants

Aqueous phase reforming (APR) 5 Biomass RR=11/5

The main tasks of the research 6 Task 1. Investigation of reaction products and intermediates Task 2. Synthesis and characterization of catalysts Task 3. Catalyst stability studies Task 4. Investigation of reactant structure on process properties Task 5. Influence of a second metal (Re) Task 6. Modelling of reaction kinetics

Experimental setup 7 Reaction conditions: 0.5 g of a catalyst 225 C 30 bar Carrier gas N 2 (30 ml/min) Continuous fixed-bed reactor 10 wt.% of xylitol used as a feed Reagents in the liquid phase

Reactor s scheme 8 H 2 (for reduction) MFC MFC HPLC pump N 2 (1% He) mass flow controller catalyst furnace MFC GC sampling collector P safety valve pressure controller effluent

Product analysis 9 APR of Polyols Gas phase Liquid phase Micro-GC HPLC TOC On-line sampling Volatile compounds were identified by means of Solid-phase microextraction (SPME) + GC-MS TOC total organic carbon analysis Mass balance 95-100% (by carbon analysis)

Important parameters in APR 10 G L H 2, CO 2, alkanes Oxygenates 100% Catalyst activity, stability and selectivity to H 2 Product distribution between phases Ratio H 2 /CO 2 v ( H ) 2 v( C ) Selectivity to H 2 (%) = 1/ RR 100% in gas RR reforming H 2 /CO 2 ratio (11/5 fo xylitol)

Choice of active metal and support 11 Based on catalytic activity in the APR of ethylene glycol Ni Pt Al 2 O 3 TiO 2 Ru Pd C MgO

Choice of active metal and support 11 Based on catalytic activity in the APR of ethylene glycol Ni Pt Al 2 O 3 TiO 2 Ru Pd C MgO C-C cleavage WGS activity Pt/Al 2 O 3, Pt/TiO 2, Pt/C and bimetallic catalyst PtRe/TiO 2

Catalyst stability Bildningshastighet Formation rate [x10 4 4 ], mol/min 12 1.8 1.6 1.4 1.2 1.0 0 30 60 90 120 Time on stream, h Catalyst showed stable performance within more than 120 h time on stream Time. h CO 2 [ 10 4 ]. mol (X C x H y )[ 10 5 ]. mol C x H y /CO 2 7 1.7 6.9 0.4 123 1.6 6.4 0.4

Distribution of carbon Carbon content, % 13 benchmark catalyst Pt/Al 2 O 3 G H 2, CO 2, alkanes 100% L Oxygenates 100 Liquid 80 60 40 20 Gas 0 0.5 1.0 1.5 2.0 2.5 3.0 WHSV, MVKV, h -1 t -1 WHSV weight hourly space velocity g subs /g cat /hour [h 1 ]

Distribution of carbon Carbon content, % 13 benchmark catalyst Pt/Al 2 O 3 G H 2, CO 2, alkanes 100% L Oxygenates 100 Liquid 80 60 40 20 Gas 0 0.5 1.0 1.5 2.0 2.5 3.0 WHSV, WHSV, h -1 h -1 WHSV weight hourly space velocity g subs /g cat /hour [h 1 ]

Gas phase composition Yield, % Yield, % Yield, % 14 Studied by micro-gc and GC-MS, benchmark catalyst Pt/Al 2 O 3 60 45 CO 2 3.0 2.5 2.0 propane etane 30 1.5 1.0 n-butane 15 H 2 0.5 0.0 СО 0 0 1 2 3 WHSV, h -1-1 0.5 1.0 1.5 2.0 2.5 3.0 WHSV, MKMV, ht -1-1 0.25 0.20 0.15 0.10 0.05 n-hexane iso-butane n-pentane 0.00 0.5 1.0 1.5 2.0 2.5 3.0 WHSV, h -1-1

Liquid phase 15 Studied by HPLC and SPME benchmark catalyst Pt/Al 2 O 3 изосорбид isosorbid Carbon in liquid. % Product distribution. % Alcohols Diols Triols Ketons Acids Other Total amount detected. % 24.4 17.3 49.9 4.2 5.0 4.9 18.7 90.6

Influence of reactant structure Yield of H 2, % Alkane selectivity, % 16 30 30 25 25 20 20 Reforming of xyltiol 15 15 10 Reforming of sorbitol 10 5 Reforming of xyltiol Reforming of sorbitol 1.0 1.5 2.0 2.5 3.0 3.5 4.0 WHSV, h -1 Higher yields of hydrogen for xylitol substrate with shorter carbon chain 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 WHSV, h -1 Higher selectivity to alkanes at lower space velocities Conditions: 225 С, 30 atm., N 2 30 ml/min, 10 wt.% solution, Pt/Al 2 O 3

Influence of reactant structure Yield of H 2, % Alkane selectivity, % 16 30 30 25 25 20 20 Reforming of xyltiol xylitol 15 15 10 Reforming of sorbitol 10 5 Reforming of xyltiol xylitol Reforming of sorbitol 1.0 1.5 2.0 2.5 3.0 3.5 4.0 WHSV, h -1 Higher yields of hydrogen for xylitol substrate with shorter carbon chain 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 WHSV, h -1 Higher selectivity to alkanes at lower space velocities Conditions: 225 С, 30 atm., N 2 30 ml/min, 10 wt.% solution, Pt/Al 2 O 3

Catalysts 17 1. Effect of a second metal: series Pt/TiO 2, Pt-Re/TiO 2 Re/TiO 2 Prepared by incipient wetness impregnation, from HReO 4 and or (NH 3 ) 4 Pt(NO 3 ) 2 2. Effect of a support material and support structure one commercial Pt/C (Degussa) four prepared and characterized materials 5% Pt/C (Degussa) 2.5% Pt/TiC-CDC (CDC carbide-derived carbon) 5% Pt/Sibunit from (NH 3 ) 4 Pt (HCO 3 ) 2 5% Pt/Sibunit H 2 PtCl 6 5% Pt/BAC (Birch active carbon)

Catalysts 17 1. Effect of a second metal: series Pt/TiO 2, Pt-Re/TiO 2 Re/TiO 2 To enhance selectivity to hydrocarbons 2. Effect of a support material and support structure one commercial Pt/C (Degussa) four prepared and characterized materials 5% Pt/C (Degussa) 2.5% Pt/TiC-CDC (CDC carbide-derived carbon) 5% Pt/Sibunit from (NH 3 ) 4 Pt (HCO 3 ) 2 5% Pt/Sibunit H 2 PtCl 6 5% Pt/BAC (Birch active carbon)

Example: APR of xylitol over Pt-Re/TiO 2 18 Yield of H 2, % 35 30 25 20 15 Pt-Re/TiO 2 Pt/Al 2 O 3 Higher yields of hydrogen for monometallic Pt catalyst Higher selectivity to alkanes in the presence of bimetallic Pt-Re catalyst 10 Pt/TiO 2 5 1 2 3 4 5 6 WHSV, h -1 Conditions: 225 С, 30 atm., N 2 30 ml/min, 10 wt.% solution In cooperation with prof. C. Hardacre from Queen s University, Belfast

Yield of H 2, % Selectivity to alkanes, % Example: APR of xylitol over Pt-Re/TiO 2 19 35 30 35 Pt-Re/TiO 2 30 25 25 Pt/TiO 2 20 20 15 Pt-Re/TiO 2 Pt/Al 2 O 3 15 10 Pt/TiO 2 10 Pt/Al 2 O 3 5 5 1 2 3 4 5 6 WHSV, h -1 Higher yields of hydrogen for monometallic Pt catalyst 1 2 3 4 5 6 WHSV, h -1 Slightly higher selectivity to alkanes in the presence of bimetallic Pt-Re catalyst Conditions: 225 С, 30 atm., N 2 30 ml/min, 10 wt.% solution In cooperation with prof. C. Hardacre from Queen s University, Belfast

Characterization Uptake of H 2, a.u. Frequency, % Concentration of NH 3, % 20 Transmission electron microscopy Particle size distribution Pt/BAC Pt/BAC 20 15 10 ~2 nm 5 20 nm 0 2 4 6 8 10 12 14 16 d, nm Temperature programmed reduction (H 2 ) Temperature programmed desorption (NH 3 ) Pt/Al 2 O 3 Pt/Al 2 O 3 350 C 0.6 386 K 75 C 180 C 0.5 0.4 0.3 0.2 0.1 379 K 20 K min -1 370 K 15 K min -1 10 K min -1 358 K 5 K min 349 K -1 3 K min -1 0.0 0 50 100 150 200 250 300 350 400 450 Temperature, C 300 350 400 450 500 Temperature, K

APR of xylitol over Pt/C Rate of alkanes formation [x10 5 ], mol min -1 21 Effect of acidity 1.6 1.4 1.2 1.0 0.8 0.6 3.0 2.9 2.8 2.7 2.6 2.5 2.4 2.3 2.2 2.1 H 2 /CO 2 0.4 0 20 40 60 80 100 120 140 160 NH 3 desorbed, mol g 2.0 1.9 Supervised visiting PhD student Benjamine Hasse for 2 months In cooperation with prof. B. Etzold from University of Erlangen-Nürnberg

Reaction pathways Based on experimental data 22

Reaction pathways Based on experimental data Pathway to hydrogen 22

Reaction pathways Based on experimental data 22 Pathway to alkanes

Reaction pathways Based on experimental data Pathway to hydrogen 22 Pathway to alkanes Modelling of kinetics was performed at ÅA A.V. Kirilin, J. Wärnå, T. Salmi, D. Yu. Murzin, Kinetic Modeling of Sorbitol Aqueous-Phase Reforming over Pt/Al 2 O 3, Ind. Eng. Chem. Res., 2014, 53, 4580 4588

Modeling Two main pathways are selected 23

Modeling 24

Modeling Weisz-Prater modulus robsr cd eff 2 for 1-st order reaction Φ < 1, for 0 order reaction Φ < 6 for 2 order reaction Φ < 0.3 r obs maximal initial reaction rate R mean radius of the catalyst particle c substrate concentration. In our case R = 1.25 10-4 m (125 µm) 25 effective diffusion coefficient (D eff ) of substrate (sorbitol) in water Wilke-Chang: 8 o 7.4 10 ( M porosity (0.3-0.6) D ef D D AB tortuosity (2-5) 0. 6 B Vb ( A) = 2.6 (water), M B is molecular weight of solvent, B = 0.11888 cp is solvent dynamic viscosity, T (K) = 498K, P=30 bar), V b(a) = 122.15 cm 3 mol -1 is liquid molar volume at solute s normal boiling point. B ) 1 2 T [cm 2 /s] Assuming / = 1/10 the diffusion coefficient of sorbitol is calculated to be D ef =1.18 10-9 m 2 /s (498 K and 30 bar). The concentration of the substrate in the solvent is equal to 0.515 mol l -1. r max = 1.93 10-4 mol l -1 s -1 at 2.7 h -1 Φ=0.005 substrate diffusion inside the catalyst pores does not affect the reaction rate

P=S+W-I, where S is the number of steps, W is the number of balance (link) equations, and I is the number of intermediates 26 P=17+2-11=8 basic routes (no intermediates are there) N (1) : C N (2) : C N (3) : C 6 5 4 O H 6 O H 5 O H 4 14 12 10 Reaction pathways C5O5 H12 H2 C4O4H10 H2 C3O3 H8 H2 CO CO CO N (4) : C 3 O H 3 8 C2O2H 6 H2 CO N (5) : CO H2O CO2 H2 N (6) : N (7) : N (8) : C4O4H10 4H2 C4H10 4H2O C3O3 H8 3H2 C3H8 3H2O C2O2H 6 2H2 C2H6 2H2O D. Murzin & J. Wärnå

on metal sites: C O H K C 6 6 14 1 C6O6H 14 on acid sites: V K C ' C ' 4O4H10 12 C4O4H10 V C O H K C 5 5 12 3 C5O5H 12 V 27 and so on for each component Total coverage: V 1 V ' 1 K 1 K 1CC O H K CC O H K CC O H K CC O H K CC O H K C 6 6 14 3 5 5 12 5 4 4 10 7 3 3 8 9 2 2 6 10 H 2 O 12 C C Rate for basic route 1: r (1) k 2 C O H C O H K 1 C K 4O4H10 14 C3O3H 8 16 k K C C O H C O H V k2k1cc 6O6H14 1 K C K C K C K C K C K 1 6 6 6 6 14 14 2 3 1 5 6 5 6 12 14 5 C O H Generation rate for compound: 4 4 10 C 7 dc C 2 O 2 C O H 3 3 H 8 C O H 6 9 6 6 14 r d (1) C O H 2 2 6 10 C H O 2 D. Murzin & J. Wärnå

Modeling C6OH 28 r (1) k 2 C O H 6 C O H k K C C O H C O H V k2k1cc 6O6H14 1 K C K C K C K C K C 1 6 6 6 14 14 2 3 1 5 6 5 6 12 14 5 C O H 4 4 10 7 C O H 3 3 8 9 C O H 2 2 6 K 10 C H O 2 Simplification of the model because of overparametrization 3.5 x 10-4 (1) ' r k2c C O H... 6 6 14 3 2.5 dc (9) r k ' 18C C O H C 6 O 6 H 14 (1) (9) r r d 3 3 8 2 1.5 1 97% 0.5 0 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 WHSVh D. Murzin & J. Wärnå

Summary 29 1. Aqueous-phase reforming of two most abundant polyols has been studied over stable catalysts 2. Effect of reactants structure on selectivity to main product formation reveals that higher selectivities to hydrogen can be achieved in APR of xyltiol compared to sorbitol (up to 85%) 3. Addition of Re as a second metal component leads to enhanced alkane formation 4. Advanced reaction scheme was proposed on the basis of experimental data 5. For the first time reaction sensitivity in APR of xylitol was studied over Pt/C 6. The reaction kinetics was modeled based on kinetic data and mechanistic considerations for sorbitol transformation during APR. APR is a powerfull method for hydrogen production from renewables Kinetic model proposed can be further developed and applied to description of APR data for other substrates as well as for APD/H process

Financial support 30 Graduate School in Chemical Engineering (GSCE, 2010-2013) Fortum Foundation (2009-2011) Academy of Finland in cooperation with North European Innovative Energy Research Programme (N-Inner, 2010-2013) Haldor Topsøe Scholarship Programme (2010-2012) COST Action CM0903 (UbioChem, 2010-2013)

Alexey Kirilin Aqueous-phase reforming a pathway to chemicals and fuels Laboratory of Industrial Chemistry and Reaction Engineering Process Chemistry Centre Åbo Akademi