Micro-Structured Monolithic-Catalyst Bed as C d Hi h Th h RCompact and High Throughput Reactor

System for Conversion of Syngas to Liquid F lFuels

Wei LiuWei LiuPacific Northwest National Lab

Richland, WA

Presentation at Technical Session of Syngas Production and Gas-to-LiquidsAIChE Annual MeetingAIChE Annual Meeting

Nov. 11th, 2010Salt Lake City, UT

Acknowledgment

Colleague at Pacific Northwest National Lab 

Yong Wang, John Hu, Wayne Wilcox, Shari Li, David King

Thi k t d d l b t di t d h dThis work was supported  under laboratory‐directed research and development (LDRD) program of PNNL by Energy Conversion Initiative and Energy & Environmental Directorate. 

Pacific Northwest National Laboratory (PNNL) is operated by Battelle for the Department of Energy. 

2

Syngas Conversion Critical to Production of Chemicals or Liquid Fuels from Various Carbon Sources

Coal

Heavy oil, tar sand

Biomass

GasificationMethanol

Suitable H2/CO ratio

Wastes

N t l Steam-

Gas cleanup & conditioning

Catalytic conversion of

syngas (H2/CO)

Oxygenates

HydrocarbonsNatural gas

CO2 + H2O

Steamreforming

Hydrocarbons

Solar photochemical

FT reactionCO2 + H2O photochemical

electrochemicalthermochemical

FischerFischer--Tropsch type of Tropsch type of SynthesisSynthesis

3

Catalysis chemistries studied extensively

Main reaction:

Highly exothermic and complex reaction paths

*222 )(2)( CHOHgHgCO

)(22 paraffinHC nn

+H2Chain growth

Main reaction: (H ~ -10 MJ/kg-product)

222 )()( gg)(2 olefinHC nn

CokeSide reaction: Coke

422 )(3)( CHOHgHgCO

222 )()( COHgOHgCO

Side reaction:

• Co-based catalysts for fixed beds

F b d t l t f l b d

4

• Fe-based catalysts for slurry beds

Step‐out reactor needed to overcome  “scale‐economy”  barrier for distributed gas‐to‐liquid conversion

• Current FT reactor commercialized for large-scale, integrated plants. Cost-prohibitive when processing capacities are one or two orders of magnitude smaller than refineries or petrochemical complexes.

• Most today’s needs of gas-to-liquid conversion are like distributed production at small capacity scales, coal gasification, biomass gasification, remote natural gases, solid wastes, other renewable energy sources.

Furnace Gas compressor Gas cleanup

P k d ll t b d

Conversion Cleanup & Catalytic Gas/liquid Liquid F d t k

Recycle of un-converted syngas gas and methane

Packed pellet beds or slurry reactor of catalyst fines

Conversion to syngas

Cleanup & conditioning

Catalytic reactor separator hydrocarbonFeedstock

5

Major portion of capital and operation costs is associated with the gas recycle system, beyond the reactor itself

Present objective: compact, high‐throughput, modular reactor unit 

Recovery of residual gas as fuel gas

Planar heat exchanger-type reactor configuration:-- Hot zone is loaded with monolithic-structured catalyst bed

Cooling

Reaction

Conversion to syngas

Cleanup & conditioning

Novel catalytic

Gas/liquid separator

Liquid hydrocarbonFeedstock

g g

y g g yreactor

• Achieve nearly complete syngas conversion at one pass to dramatically simplify process flow diagram, and thus, capital and operation costs

Limitations of conventional reactors (packed or slurry beds):• CH4 selectivity is too high at deep CO conversion• Hydrodynamics, such as back mixing• Catalyst deactivation

6

A holistic approach  toward catalyst & reactor design 

Li & Kwauk “Exploring complex systems in chemical engineering – the multi-scale methodology”. Chem. Eng. Sci. 58 (2003) 521-535, a powerful approach to all catalytic reactor systems.

Address mass and heat transfer at

catalytic reactor systems.

• Reactor scale• Scale of individual reaction unit• Catalyst structure scale

7

Reactor‐scale design to address heat management at macro‐scale

Catalyst structure insert1cm x 1m x 2m

Planar, heat-exchanger-type design: • Take away reaction heat by adjacent cooling 1cm x 1m x 2my y j g

streams

• Control temperature profiles by feed gas and cooling fluid distribution, flow rates, and heat gexchanging area

• Utilize existing reactor fabrication technologiesg

• Scale up by numbering up reactor modules

8 Cooling channel

Micro‐structured catalyst bed to address heat and mass transport at scale of individual reaction zone

Ceramic honeycomb as catalyst support structures• Structured catalyst bed comprises catalyst 

coatings on a continuous inert support.S t t i id 3 D th l• Support matrix provides 3‐D thermal conduction network for rapid heat transfer from catalyst surfaces to the heat exchanging interface.g g

Cooling plateCooling plate

SyngasSyngas

9

Catalyst channel‐scale design to addresses gas/liquid/catalyst mass transport at reaction pointsBulk fluid

Channel surface

External surface of catalyst coating

Catalyst pore

H2OOilCH4

H2 (g)CO (g)

I IVIIIIICatalyst surface

St III &IV b ti i d b

surface catalyst coating p CH4CO (g)

H2OOilCH4

SyngasT l flDi d b bbl fl

Steps I & II can be controlled by channel geometries (size, shape) and flow conditions (Ul, Ug)

Steps III &IV can be optimized by thickness and pore structures of the catalyst coating layer

CH4

Inert support

Catalyst layerLiquid film

Sy gasTaylor flowMist flow

Dispersed bubble flow

10

Experimental testing of key design principlesExperimental testing of key design principles

FT catalyst coating on ceramic monolith substrate of 1mm channel size 

Coated catalyst channel

Coating texture of channel surface

Cross-sectional view of coating

channel

Al Si Co La Re

76.7% 2.5% 19.0% 0.8% 1.0% Al Si Co La Re

CompositionComposition

67.7% 1.4% 28.6% 1.1% 1.3%

12

Reaction testing in a ½” OD stainless steel tube sheathed with silicone oil heat‐exchanging fluid

Thermocouple

Feed gas

All channels are open after 45-

day testing

Monolithic-

Feed gasday testing

Monolithic-structured bed

Af iSilicone oil

• After reaction was quenched, existence of wax was found in all channels on bottom of thechannels on bottom of thebed

Suggest uniform reaction among all channels

13

g

Steady‐state performance of monolithic‐structured catalyst bed (25 bar, 210oC)

Nearly complete H2 or CO conversion at base WHSV at CH4 selectivity <10 wt%.  Stable temperature control

WHSV =base

80

90

100

ty, % 230

240

250WHSV =baseH2/CO=1.5

WHSV =baseH2/CO=2.0

WHSV =2xbaseH2/CO=2.0

50

60

70

CH

4 se

lect

ivi

200

210

220

ratu

re, o

C

30

40

50

conv

ersi

on o

r

180

190

200

Bed

tem

per

CO conversionCH4 selectivityT topT bottom

0

10

20CO

c

150

160

170T bottom

14

00 24 48 72 96 120 144 168 192 216 240 264 288

Time on stream, h

150

Impact of reaction temperature on performance of structured catalyst bed• At a given WHSV, CO conversion and CH4 selectivity increase with

temperature, as expected110 CO at WHSV=38.7 Feed H2/CO ratio =2

8090

100110

ty

CH4 at WHSV=38.7CO at whsv=77.4CH4 at WHSV=77.4

25 bar

50607080

or s

elec

tivit

20304050

Con

vers

ion

o

010

20

205 210 215 220 225 230

C

15

205 210 215 220 225 230

Temp, oC

Impact of feed gas superficial linear velocity on performance of structured catalyst bed

• CO conversion decreases with increasing Vg (= increasing WHSV), as expected.

• CH4 selectivity increases with Vg higher CH4 selectivity at lower CO conversion opposite to results of conventional slurry or packed particle beds opposite to results of conventional slurry or packed particle beds

100110

%

Feed H2/CO ratio =225 bar

60708090

sele

ctiv

ity, %

CO at 210oC

30405060

vers

ion

or s CO at 210oC

CH4 at 210oC

CO at 225oCCH4 at 225oC

01020

Con

v

16

0.00 0.05 0.10 0.15 0.20

Vg, cm/s

Steady‐state performance of crushed monolith catalyst particle• 60-200mesh (~170um), H2/CO = 2, 25 bar, 210oC• Very difficult to control temperature. CO conversion was unstable under const

conditions. Very high CH4 selectivity

80

90

100

ity, %

230

240

250CO conversion

CH4 selectivity

T top

WHSV = 0.5 x base

60

70

80

r CH

4 sel

ectiv

i

210

220

230

ture

, ºC

p

T bottom

30

40

50

O c

onve

rsio

n o

180

190

200

Tem

pera

t

0

10

20CO

150

160

170

17

0 24 48 72 96 120

Time-on-Stream, h

Comparison of structured bed to crushed monolith particle bed  under same temperature• The structured bed shows CO conversion activity a few times higher than the

crushed particle bed.

• The structured bed enables <10% CH4 selectivity at nearly complete one-pass CO 4 y y p pconversion.

100 210oC, 25 bar1820 Typical particle bed

60

80

sion

% Structured bed

1012141618

ctiv

ity, % Paradigm shift

by structured bed

20

40

O c

onve

rs

Particle bed 468

10

CH

4 se

lec

210oC25 b

0

20

0 20 40 60 80 100

CO Particle bed

02

0 20 40 60 80 100

C 25 bar

18

WHSV, g/g/h CO conversion, %

Comparison of structured bed to crushed monolith particle bed at different temperature • Structured bed consistently provides CO conversion a few times higher

than the particle bed under the same reaction conditions

110 Feed H2/CO ratio =2

8090

100110 Feed H2/CO ratio =2

25 bar

607080

r sel

ectiv

ity CO monolith WHSV=38.7CH4 monolith WHSV=38.7CO particle whsv=35CH4 particle WHSV=35

304050

onve

rsio

n or CH4 particle WHSV 35

01020

20 210 21 220 22 230

Co

19

205 210 215 220 225 230Temp, oC

Proposed physical models for multiphase reaction in different catalyst beds 

Packed particle bed or slurry bed may be uniform at macro-scale. Segregation of gas/liquid/solid (catalyst) is inevitable at individual particle scale

Micro-structured catalyst channel provides uniform contacting of gas/liquid/catalysty p g g q y

Catalyst surface coverage by a dynamic, thin liquid film favors for low CH4selectivity and high reaction rate.

SSyngas gas

Syngas gasCatalyst

layerLiquid film

Syngas Catalyst layer

Liquid film

Syngas

Inert supportInert support

20

Structured catalyst channel Packed catalyst pellets Slurry bed

Concluding remark

For gas-to-liquid (FT) reaction, micro-structured catalyst bed designs lead to some unexpected, exceptionally good performance attributes with a common catalyst composition:

• >90% CO conversion at <10% CH4 selectivity• A few times of higher activity than the same catalyst material in crushed particle

form• Good temperature control in conventional sizes of reaction tubes (micro-channel

reactor is not necessary)• Catalyst bed free of plugging by wax• A compact, high throughput gas-to-liquid catalytic reactor is possible

Note: poor performances obtained with conventional catalyst particle or slurry beds b d t th t d i i th th d t t l t t i l it lfmay be due to the reactor design issues rather than due to catalyst material itself.

21