CARBON MATERIALS FOR ENVIRONMENTAL PROTECTION

Professor Barry Crittenden

Department of Chemical Engineering

Faculty of Engineering and Design

University of Bath, Bath, UK

CESEP’09, Torremolinos, Spain, 25-29 October 2009

AN ENGINEER’S PERSPECTIVE

OF

THE ROLE OF ACTIVATED CARBON MONOLITHS

Granular activated carbons (GAC) have enjoyed a long and successful record in removing pollutants from aqueous and gaseous environments.

A VERY GOOD MATERIAL…

Granular activated carbons (GAC) have enjoyed a long and successful record in removing pollutants from aqueous and gaseous environments.

With pressures now to reduce energy demands and CO2 emissions in all forms of processing, focus is shifting towards ways of operating AC systems at very much reduced pressure drops.

NOW FOCUS ON ENERGY…

Granular activated carbons (GAC) have enjoyed a long and successful record in removing pollutants from aqueous and gaseous environments.

With pressures now to reduce energy demands and CO2 emissions in all forms of processing, focus is shifting towards ways of operating AC systems at very much reduced pressure drops.

As an example, activated carbon monoliths (ACMs) can be used to remove volatile organic compounds (VOCs) from air streams with substantially reduced energy demands because (i) they have intrinsically low pressure drops, and (ii) they can be thermally cycled much faster than GAC and hence their carbon inventories can be reduced considerably.

TO REDUCE ENERGY DEMANDS…

Granular activated carbons (GAC) have enjoyed a long and successful record in removing pollutants from aqueous and gaseous environments.

With pressures now to reduce energy demands and CO2 emissions in all forms of processing, focus is shifting towards ways of operating AC systems at very much reduced pressure drops.

As an example, activated carbon monoliths (ACMs) can be used to remove volatile organic compounds (VOCs) from air streams with substantially reduced energy demands because (i) they have intrinsically low pressure drops, and (ii) they can be thermally cycled much faster than GAC and hence their carbon inventories can be reduced considerably.

Replace the large GAC bed with a much smaller volume of structured ACM

BY REDUCING THE INVENTORY OF CARBON…

Granular activated carbon needs to be heated with steam or hot gas for its regeneration. This is a slow process, causing thermal swing adsorption (TSA) cycle times to be high, eg 8 hours. Recovery of the solvents is more difficult if steam is used although this is a good heating medium.

COMPARE AND CONTRAST

Granular activated carbon needs to be heated with steam or hot gas for its regeneration. This is a slow process, causing thermal swing adsorption (TSA) cycle times to be high, eg 8 hours. Recovery of the solvents is more difficult if steam is used although this is a good heating medium.

Activated carbon monoliths can be heated electrically at low potential difference. This is a fast process, thereby allowing operation at much shorter cycle times, eg 60 minutes. The consequence is that much less adsorbent is required and so the equipment is much smaller.

COMPARE AND CONTRAST

COMPARE THE PRESSURE DROPS…

4

4.28

a

Q

L

P

pp d

u

e

e

d

u

e

e

L

P 2

323

2 175.1

1150

Skin friction only (laminar flow)

Patton et al (2004)

Skin friction and form drag with the latter as a function of u2 predominating

Ergun (1952)

A WORLD-WIDE PROBLEM

UK GDP = 4.6% of global GDP (Scaruffi, 2004)

UK VOC emissions reduced to 1.5 million tonnes in 2002 (Environment Agency)

Worldwide VOC emissions therefore are approximately 32.6 million tonnes/year

Adsorption is used for 25% of total VOC control market (Frost & Sullivan, 2000)

Therefore about 8 million tonnes/year might be controlled by adsorption worldwide

A typical VOC concentration might be 2 g/m3

Volume of air to be cleaned therefore could be 130,000 m3/s

Assume an average 20% loading (w/w) on both the granular and monolithic adsorbents (it has been shown that the kinetic performances of the two systems are quite similar).

Assume an 8.0 hour cycle time for granular (spherical) adsorbent.

Assume that the Ergun equation is applicable for the pressure drop through a bed of granular activated carbon (GAC).

Assume a 1.0 hour cycle time for monolith adsorbent.

Assume that the Poiseuille equation is applicable for the pressure drop for laminar flow through the monolith channels.

What is the potential benefit from switching from GAC to ACM for VOC control?

BASIS OF PRESSURE DROP COMPARISON

GLOBAL POTENTIAL

Pressure Drops for Worldwide Control of VOCs Packed bed Monolith Pellet diameter, mm 1.0 - Square channel monolith dimension, mm - 0.63 Square channel monolith wall thickness, mm - 0.43 Air flow rate, m3 s-1 130,000 130,000 Bed volume, m3 50,000 6250 Reynolds number 100 157 Superficial velocity, m s-1 1.55 1.55 Bed voidage or monolith fractional free volume 0.4 0.4 fp or fm 49.7 0.41 Bed cross sectional area, m2 83,870 83,870 Bed length, m 0.60 0.075 Pressure gradient, N m-3 70,747 5790 Pressure drop, N m-2 42,450 434 Power requirement, MW 5500 56

Potential power saving is substantial.

Potential for CO2 reduction depends on energy sources.

EFFECT OF SUPERFICIAL VELOCITY

Effect of Superficial Velocity on Pressure and Energy Losses Packed Bed Monolith Reynolds number 1000 100 10 1570 157 15.7 Bed volume, m3 50,000 50,000 50,000 6250 6250 6250 Sup. velocity ms-1 15.5 1.55 0.155 15.5 1.55 0.155 ε 0.4 0.4 0.4 0.4 0.4 0.4 fp or fm 34.5 49.7 201.6 0.041 0.41 4.1 Bed XSA, m2 8387 83,870 838,700 8387 83,870 838,700 Bed length, m 6.0 0.6 0.06 0.75 0.075 0.0075 ΔP/L, N m-3 4.91x106 70,747 2870 57,900 5790 579 ΔP, N m-2 29.5x106 42,450 172 43,430 434 4.34 Power, MW 3.84x106 5500 22 5600 56 0.56 Power ratio 685 98.2 39.3 1 1 1

The effect of velocity squared in the packed bed is very significant.

MANUFACTURE OF ACMs

Gadkaree Carbon from high carbon yield phenolic resin impregnated on ceramic honeycomb support

1998

Yates et al. Activated carbon mixed with silicate clay before extrusion

2000

Tennison et al. Binder-less activated carbon made from extruded phenolic

Novolak resin

2001

Fuertes et al. Carbon from phenolic Novolak resin mixed with Nomex fibres

2003

Valdés-Solis et al. Carbon from phenolic Novolak resin dip-coated on ceramic

support

2003 30% linear & 50% volumetric shrinkage

on carbonisation

1. Phenolic resin

2. Controlled cure

3. Mill & classify

4. Extrude

5. Carbonise

6. Activate

MAST CARBON MANUFACTURING STEPS

EXTRUDED ACMs IN VOC RECOVERY UNIT

Place R N, Blackburn A J, Tennison S R, Rawlinson A P and Crittenden B D

“Method and equipment for removing volatile compounds from air”US Patent 6964695 (2005), European Patent EP 1372917 (2005)

Winner – IChemE Severn Trent Water Safety Award 2002

ISOTHERMS FROM HIDEN ANALYTICAL INTELLIGENT GRAVIMETRIC ANALYSER

Tóth model:

The adsorption capacity is deemed to be fixed but the

affinity parameter b is allowed to vary with temperature.

DCM

nnmax

bp

bpqq

1

1

EFFECTIVE DIFFUSION COEFFICIENT FROM THE IGA

DCM on ACM Sqr-21

FITTING LDF EQUATION TO IGA KINETIC DATA

LDF is described by the following (Chagger et al, 1995):

kteqtq 1

q(t) is the uptake at time t, Δq is the total change in uptake for a given pressure increment and k is the adsorption rate constant.

ottkoo eqqqtq 11

qo is the equilibrium uptake at to (the initial time of the pressure increment) and q1 is the equilibrium uptake at the

equilibrium time.

ORIGINLAB software was used to fit the experimental data to the LDF model equation.

FUNDAMENTAL DESIGN

Gas phase

Solid phase

Gas flow

0

x y

z

tw

a

Decisions:

3D diffusion & convection in the channel gas phase

Various flow regimes in the channels (eg plug, axially dispersed plug, fully developed, developing)

3D diffusion in the solid phase

Adsorption at the gas-solid interface (eg Langmuir, Tóth)

Isotropic, anisotropic solid phase

Isothermal, non-isothermal

Uniform, non-uniform channels

FLOW AND CONVECTION

02

2

2

2

2

2

x

cu

z

c

y

c

x

cD

t

caveM

2

2

2

2

2

2

z

q

y

q

x

qD

t

qeff

Mass balance gas phase:

Mass balance solid phase:

Tóth model:

0,2

2

2

2

2

2

z

T

y

T

x

Tzyuc

z

T

y

T

x

T

t

Tc ggg

pggggg

gg

pgg

02

2

2

2

2

2

z

T

y

T

x

T

t

Tc sss

ss

pss

0TTAdt

qdH

z

Tsg

az

ss

45.0

095.01977.2

PrRe

L

dhdNu ch

g

ch

Energy balance gas phase:

Energy balance solid phase:

Energy balance interface:

Gas channel heat transfer coefficient:

nnmax

bp

bpqq

1

1

BREAKTHROUGH OF DICHLOROMETHANE

0 20 40 60 80 100 120 140 160 180 2000.0

0.2

0.4

0.6

0.8

1.0

c/c 0

Time [min]

7 litres/min (STP) air flow (Re = 43; uave = 0.98 m/s) with 2000 ppmv DCM 103 mm long by 18.6 mm diameter ACM with

nominal channel size of 0.7 mm, nominal wall thickness of 0.35 mm fractional free cross sectional area of 0.44.

0 20 40 60 80 100 120 140 160 1800.0

0.2

0.4

0.6

0.8

1.0

c/c 0

Time [min]

0

5

10

15

20

25

30

35

40

45

0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4

d/dave

Nu

mb

er o

f C

han

nel

s

Actual distribution Normal distribution

smaller non-square channels

MODEL TAKES CHANNEL NON-UNIFORMITY INTO ACCOUNT

FINALLY – A COMPARISON WITH GRANULES

CONCLUSION

Once the channel shape has been selected, two independent design variables (the channel dimension and the wall thickness) can be altered to provide internal and external mass transfer coefficients similar to those for an equivalent bed of granules. GAC has one less independent design variable.

Reducing the channel dimension improves the monolith external mass transfer performance but increases its pressure gradient. Reducing the wall thickness improves its internal mass transfer performance.

Channels just less than 1 mm and walls just less than 0.5 mm provide the same free space and mass transfer performance equivalent to a bed of 1-2 mm granules but with a considerably reduced pressure drop. This does not take into account the even greater reduction in pressure drop achievable because the ACM is electrically regenerable.

Form drag and roughness are the principal weaknesses of granular materials. In laminar flow, the roughness of monoliths does not affect the pressure drop, and hence energy requirements and carbon dioxide emissions.

QUESTIONS?