G-L Taylor Flow reactor system Oxidation of ethylbenzene

R. Sumbharaju

L.A. Correia D.F. Meyer

Y.C. van Delft A. de Groot

This presentation was presented at the CAMURE-8 & ISMR-7 conference,

Naantali, Finland, 22-25 May, 2011 ECN-M--11-084 AUGUST 2011

www.ecn.nl

G-L Taylor Flow reactor system Oxidation of ethylbenzene

R. Sumbharaju, L.A. Correia, D.F. Meyer, Y.C. van Delft and A. de Groot

2 23-8-2011

Contents• ECN – Research E&I • Research PI

- Process and technology selection• Structured reactor• Realization prototype reactor

- Selection of G/L oxidation process- Prototype

• Experimental- Process window

• Kinetics• Hydrodynamics

- Results- Summary- Future work

3 23-8-2011

• Largest Dutch R&D institute on energy • Independent and committed• Link between fundamental academic research and market application

Mission statementECN develops and brings to market high-level knowledge and technology for a sustainable energy society

ECN Mission and Units

Solar energy Wind energyBiomass Policy Studies

ECN Business units

Efficiency & Infrastructure

Research areas ECN - E&I

4 23-8-2011

by products

rawmaterials

clean air

clean water

Scheiding

Separation

Separation

Separation

SeparationREACTOR

Waste heat

1. Industrial waste heat

3. Process intensification/ Multifunctional reactors

2. Molecular Separation Technology

Process and technology selection

5 23-8-2011

NH3+MeOH21%

Others (exo)11%

Others (endo)5%

Oxidation30%

Cracking19%

Oxi

datio

n

Cra

ckin

g

Am

mon

ia

Hyd

roge

natio

n

Deh

ydro

gena

tion

Hyd

ratio

n

Deh

ydra

tion

Con

dens

atio

n

Alk

ylat

ion

Isom

eriz

atio

n

MR / pervaporation membrane 0 0 0 0 0 7 2 26 0 0MR / hydrogen selective 0 0 21 0 19 0 0 0 0 0Catalytic membrane reactor 5 3 1 0 1 0 1 0 1 1HEX-reactor 48 29 0 2 6 19 3 0 15 0Structured (monolith) reactor 22 0 0 1 0 15 2 1 2 1Micro-reactor 54 28 3 5 8 27 5 4 8 3Short residence time reactor 11 9 0 0 2 0 2 0 2 0Reactive distillation 4 20 14 1 11 2 2 20 3 11Sorption enhanced reactor 7 21 21 1 19 4 1 26 6 18Supercritical reactor 8 0 0 10 0 10 4 11 0 4Advanced distillation 7 14 1 19 4 0 6 18

Relative energy usage NL

J. Hugill Monolith reactors for gas-liquid reactions: Internal ECN Report

Technologies for process intensification

6 23-8-2011

Structured reactor and Taylor flow

Close to plug flow

Thin liquid layerseparates gas/solid Intensive mixing in the slugs

• Good mass transfer• Good mixing in the liquid slugs• Overall plug flow• Good heat transfer

Structured reactor - 2-phase flow pattern

7 23-8-2011

Flow patterns depend on VG ,VL[1]Best operating condition for the reaction in the channels is Taylor flow (PF).

[1] Kreutzer et al (2001), Chem. Engg. Sci. 56 (2001) p.p.:6015-6023

Bubble flow Taylor flow Unstable slug flow

Annular flow

8 23-8-2011

Process selection for energy savings in NL [1]

• Inventory of oxidation processes in the NL• Selection of top 5 oxidation processes in NL based on energy usage• Selection of one process based on

-Total production capacity in the NL-Reaction heat-Temperature-Pressure-Catalyst

• Result : Styrene - Propylene oxide process

Styrene process

9 23-8-2011

Focus area

EB oxidation to EBHPProcess conditions:P = 2 barT = 140 °CConversion = 10 wt%Selectivity = 90 %Residence time = 1 hr

Realization prototype reactor• G/L oxidation reactions are highy exothermic in nature • Formation of explosive gas mixtures • Demanding high safely operating equipment

Procedure for realization prototype reactor:- Chemical and physical hazards- Conceptual design of installation - Critical operation hazards and required safety measures- Detailed design of installation- Prototype - computer controlled

10 23-8-2011

11 23-8-2011

Prototype Taylor flow reactor

Prototype Taylor flow reactor - control unit

12 23-8-2011

13 23-8-2011

Prototype Taylor flow reactor

Reactor Section G/L separator and product Section

Feed Section

Preliminary oxidation study

14 23-8-2011

Demonstrating that in TF mode oxidation can be performed with improved yield.

• Verification of flow pattern with water/N2 and different P and T.

• Development of a kinetic model for EB oxidation

Defining process window

Operating window - fluid flow

15 23-8-2011

◊ Unstable region□ Border region○Taylor/Plug flow

Prediction of Temp. window by Kinetic model

16 23-8-2011

[1] A.Yurdakul “Reactor Modelling of GL Reactions in Small Channels Intermediate Report” ECN-ei-2009-265

17 23-8-2011

Experimental Program (1)Reaction steps in oxidation of Ethylbenzene

Autocatalytic reaction• Initiation (Low residence time improve initiation rate)• Propagation• Termination

Competing reactions for byproducts• Decomposition of Ethylbenzene hydroperoxide

L.A. Tavadyan et al., Kinetics and Catalysis, 44, 91-100, 2003

Preliminary results Different commercial initiators: A, B, C (4 mol%)

18 23-8-2011

Initiator A shows highest Yield

19 23-8-2011

Experiment vs Kinetic model

• Industrial yield is in agreement with the model at 140 °C• Experimental yield is lower than the model at 180 °C• By-product formation in our reactor at 180 °C is higher than predicted

Summary

• Prototype CRS can handle dangerous reactions without reaching explosive situations.

• Oxidation in TF mode is demonstrated.• In the prototype CRS EBHP yield is lower than expected. • Hydrodynamic results are in agreement with the literature.• Yield highly depends on initiator type.

20 23-8-2011

Plans for future

• Experimental kinetic studies (increasing yield) - Reactor geometry- Process conditions

• Other oxidation reactions

21 23-8-2011

22 23-8-2011

Questions

Commercial Initiators:A

23 23-8-2011

• Increase of temperature increases conversion and selectivity• wt% increase in concentration has negative effect on selectivity• 180 °C is promising for selection

24 23-8-2011

Small channels- Small fluid mixture hold-up: safe operation of explosive mixtures

Increased mass and heat transfer - Heat transfer: Large surface-area/volume ratio - Improved temperature control- Mass transfer: Large contact area of the phases.- Improved mixing in Taylor flow (TF) operation

Structured reactor in 2-phase flow (1)

[1] F.Kaptein et al ,CATTECH, Vol 3(1999), No.1,24-41

Hydrodynamics – Flow windowExperimental vs literature

25 23-8-2011

Taylor flow boundary [1]• Dimensionless number, We, flow pattern map for all

experimental conditions, p and t shows a sharp border between Taylor and bubble flow:

6.05.5 GSLS WeWe =

Gas hold-up [1]• Armand correlations relates gas hold-up to the gas volume

fraction

• Found Armand correlation is: 0.88.

GG βε 833.0=

LG

GG VV

V+

=βSSSS

SG LG

WLG

G+

−++

= )16

(πε

Gas slug length (GS) [1]• The gas slug length scales linearly with the ratio VG/VL .

• For our system, α1=1.24 and α2=0.59 (α1=1.0 and α2=1.0 [2])L

G

VV

WGs

21 αα +=

[1] Jun, Y., et al, Chem. Engineering Science, 63, pp. 4189-4202, 2008[2] van Steijn, V. (2010): Formation and Transport of Bubbles in Micro-Fluidic Systems. pp. 136, Delft, 2010.