Modelling of Eucalyptus Globulus

Kraft Pulping

H. Sixta, E. Rutkowska, P. Schröder, P. Wollboldt

3rd International Colloquium on Eucalyptus Pulp

March 4-7, 2007 - Belo Horizonte, MG Brazil

06-03-2007 | 2

Outline

Objective and review of existing cooking models

Structure of selected kinetic model for Delignification / cellulose degradation

Depolymerization

HexA formation

Validation of the model

Structural characterization of lignin during kraft pulping using

a new method for lignin isolation

Summary

06-03-2007 | 3

Objective, Background

Eucalyptus globulus Labill, one of the most interesting species

among the > 600 species comprising the genus Eucalyptus

represents the fiber source with the highest pulp yield and quality in

both PP and DP.

Major wood source for the production of market hardwood bleached

kraft pulp in Europe (Spain, Portugal). Growing interest to grow E.

globulus in plantations in Australia and South America.

However, little information is available about delignification kinetics

of the Eucalyptus species in general and of E. globulus in particular

It is the objective to develop a comprehensive kinetic model to

predict delignification, carbohydrate degradation, depolymerization

(viscosity loss) and HexA formation and degradation.

Literature Review on Modeling of E.

globulus Kraft pulping

06-03-2007 | 5

Three Phase Model1

Concept of Consecutive Reactions

Initial delignification: Lignin > 0.82*L0

1A. Santos et al. Ind. Eng. Chem. Res. 1997, 36, 4114-4125; M.A. Gillarranz et al. Ind. Eng. Chem. Res. 2002, 41, 1955-1959

LTTR

Expdt

dL

37700252

LSHOHTR

Expdt

dL

16.015.011 000.106

1013.3

LOHTR

Expdt

dL

59.08 000.86

1013.1

7.2

Ld

OHd

92.0

Ld

OHd

55.0

Ld

OHd

Transition from the bulk to the residual phase is dependend on [OH-] and T:

DL/L0=0.022*[OH-]+6.9*10-4*T+0.82

Min = 0.93 (0.7 OH-/150°C), Max =0.97 (1.6M OH-/180°C)

Residual delignification: Lignin < 0.05 (0.03 – 0.065) *L0

Bulk delignification: 0.05*L0 < Lignin < 0.82*L0

06-03-2007 | 6

0 50 100 150 200 250 300 350

1

10

100

exper. results: [OH-] = 0.1 M

Ka

pp

a n

um

be

r, o

dw

time [min]

Prediction of Step Changes: [OH-]=0.1 M 1.0 M

0 50 100 150 200 250 300 350

1

10

100 step change

exper. results: [OH-] = 0.1 M [OH

-] = 1.0 M

Kap

pa

nu

mb

er,

od

w

time [min]

0 50 100 150 200 250 300 350

1

10

100

exper. results: [OH-] = 0.1 M [OH

-] = 1.0 M

simulation: [OH-] = 0.1 M

Ka

pp

a n

um

be

r, o

dw

time [min]

0 50 100 150 200 250 300 350

1

10

100

exper. results: [OH-] = 0.1 M [OH

-] = 1.0 M

simulation: [OH-] = 0.1 M [OH

-] = 1.0 M

Ka

pp

a n

um

be

r, o

dw

time [min]

T = 160 °C

Experimental

06-03-2007 | 8

Wood Source E. globulus, M‘Bopicua, Uruguay (ENCE)

Parameter, % od Value

Klason Lignin KL 22,1

Acid Soluble Lignin AL 4,6

Cellulose C 47,3

4-O-methylglucuronoxylan X 21,8

Glucomannan GM 2,2

Resins, Ash 2,0

Total 100,0

Parameters

L1 L2 L3 L4 L5

[OH-] mol/L 0,37 0,10 0,27 0,52 0,94 1,23

[SH-] mol/L 0,16 0,17 0,29 0,64

[Na+] mol/L 1,5 0,80 1,50 2,50

L/S kg/kg 10 40 40 40 40 40

Temperature (T) °C 100 140 150 160 170

Time (max at T) min 60 1448 565 227 114

Reaction KineticsImpreg

nation

Reaction conditions

Structure of Selected Kinetic Model

06-03-2007 | 10

Model Structure2: Delignification, Carbohydrate Degradation

2 N. Anderson, D.L. Wilson, U. Germgard: Nordic Pulp and Paper Research J.,2003, 18 (2), 200-209

jijWji

Wkdt

dWi ,,

,

r

cbaW

WW kNaHSOHTR

EAExpAk ji

jiji

1

443

1,

,,

General rate expression / general solution of first-order equations:

tkExpWW jWj

jitoti i

,

3

2

0,,

Dependency of rate expressions on the reaction conditions:

06-03-2007 | 11

Concept of Interchanging Species2

2 N. Anderson, D.L. Wilson, U. Germgard: Nordic Pulp and Paper Research J.,2003, 18 (2), 200-209

0,1 0,5 0,9 1,3

4

7

10

13

16

0,8 1,3 1,8 2,3

4

7

10

13

16

0,1 0,3 0,5 0,7

4

7

10

13

16

140 150 160 170

4

7

10

13

16

L* ,

kappa o

dw

[OH-], mol/L

[Na+], mol/L

[HS

-], mol/L

T, [°C]

0,0 0,3 0,6 0,9 1,245

48

51

54

57

140 150 160 17045

48

51

54

57

CH

* , %

od

w

[OH-], Mol/L

T, °C

W2 and W3 interchange reversibly: W2 = W3 = W* = f(conditions)

106.4308.0625.00.5045.0025.0001487.0 NaHSOHL

OHTOHCH 48.045.6844.55

06-03-2007 | 12

Kinetics of Cellulose Chain Scissions

Kinetic Constants

tNaOHTR

EAExpA

DPDP

edC

ntn

0,,

111

Units

DPn,0 4900

ln A (1/M*min) 36,48

EAc kJ/mol 180,30

d 1,08

e 0,74

Model

parameter

An increase in ionic strength relates to an increase in cellulose degradation

0 5 10 15 20 25600

800

1000

1200

1400

[Na+], mol/L: 0.8 1.5 2.5

Vis

co

sity [

mL

/g]

Kappa number

06-03-2007 | 13

Diffusivity of Alkali, Chemical Consumption

3R.R. Gustafson et al.: Ind. Eng. Chem. Process Des. Dev, 1983,22,87-96. 4S. McKibbins: Tappi J., 1960, 43

(10), 801-805; 5T. Christensen et al.: Tappi Proceedings – Annual Meeting, 193, 239-246

Gustafson3 corrected diffusivity based on McKibbins4 with respect to

pH and lignin content with the ECCSA data published by Hartler;

58.0 13.0#00364.04.2452

057.0 55.0

OH

TExpTD

Componentkg/kg removed

Lignin 0,20 0,04

Carbohydrates 0,45 0,00

EA-consum-

ption

HS-consum-

ption

Consumption of active cooking chemicals (in the range of the data

reported by Christensen5 et al.):

06-03-2007 | 14

Course of [OH-] during Batch Cooking

E. globulus Kraft

Chip thickness 4.0 mm

Moisture content 35%

Chip density 0.50 kg/L

L/S 4:1

[OH-]t=t0 1.6 M

[HS-]t=t0 0.245

Temperature (at) 160 °C

time_to 90 min

time_at 112 min

H-Factor 800

0 50 100 150 2000,0

0,4

0,8

1,2

1,6

Calculated: Free-[OH-] Bound-[OH

-] Chip centre-[OH

-]

Experimental:

[OH

- ], m

ol/L

time [min]

06-03-2007 | 15

Change of HexA is controlled by MeGlcA and HexA concentration (no dissolution considered)

Model Structure6: Formation and Degradation of HexA

MeGlcA-(1->2)--D-Xylp HexA-(1->2)--D-Xylp

-D-Xylp

k1

k2k3

-CH3OH

MeGlcA HexA

HexAkMeGlcAkdt

HexAd 21

tkExptkExpkk

MeGlcAkHexA 21

12

0 1

Solution of the differential equation describing a consecutive reaction

[MeGlcA]0 = 166 mmoL/kg wood

[HexA]0 = 0 mmoL/kg wood

4S. Danielsson, K. Kisara, M.E. Lindström: Ind. Eng. Chem. Res. 2006, 45, 2174-2178

06-03-2007 | 16

Effect of Temperature on HexA profile

Model describes suffenciently well the pattern of a simple successive first-order reaction. Activation energies quite comparable to those reported by Danielsson et al.

0 200 400 1000 1500

0

10

20

30[OH

-] = 0.52 M

[HS-] = 0.28 M

[Na+] = 1.5 M

140 °C 150 °C 160 °C 170 °C

He

xA

[m

mo

l/kg

od

wo

od

]

time at temperature [min]

temperature k1 k2

[°C] [min-1

] [min-1

]

140 0,0019 0,0058

150 0,0040 0,0135

160 0,0078 0,0329

170 0,0192 0,0927

EA [kJ/mol] 116,5 140,3

ln A [min-1

] 27,6 35,6

Validation of the Kinetic Model

06-03-2007 | 18

Prediction of Step Changes at L/S = 40:1

Immediate change from 0.1 M [OH-] to 1.0 M [OH-]

Agreement between

calculated and

measured kappa

numbers

satisfactory

0 50 100 150 200

10

100

[OH

-], mo

l/L

kappa number: calculated measured

Ka

pp

a n

um

be

r

reaction time at 160 °C, min

0,0

0,4

0,8

1,2

[OH

-]: free bound chip centre

0 50 100 150 200

10

100

[OH

-], mo

l/L

kappa number: calculated measured

Ka

pp

a n

um

be

r

reaction time at 160 °C, min

0,0

0,4

0,8

1,2

[OH

-]: free bound chip centre

0 50 100 150 200

10

100

[OH

-], mo

l/L

kappa number: calculated measured

Ka

pp

a n

um

be

r

reaction time at 160 °C, min

0,0

0,4

0,8

1,2

[OH

-]: free bound chip centre

06-03-2007 | 19

Prediction of Kappa and Viscosity at L/S = 3:1 Pilot Plant trials with E. globulus

H-Factor: 300, 450, 600, 800 [OH-]0 1.60 M

[HS-]0 0.25 M

[Na+] 2.00 M

T 160 °C

0 50 100 150 2000

10

20

50

100

150

Vis

co

sity

[ml/g

]

Kappa#: calculated experimental; temperature

Ka

pp

a #

[o

n p

ulp

] /

T,

°C

time [min]

600

900

1200

1500

( )

Viscosity: calculated experimental

Good fit between

measured and calculated

pulp properties.

Surprisingly, prediction of

viscosity more precisely

as compared to the

kappa number

Structural Characterization of Lignin

during Kraft Pulping using a New

Method for Lignin Isolation4

4Schröder, P., Wollboldt, R.P., Weber, H.K., Fasching, M., Sixta, H.: Lignin Isolation from completely dissolved wood, submitted to Holzforschung

06-03-2007 | 21

Lignocellulose

Preparation

Lignin/Carbohydrate

Separation

Removal of

Metal Ions

Lignin

Purification

Dissolution

Removal of

Extractives

Lignin

Acetylation

Isolation of Dissolved Wood Lignin (DWL)4

◆ drying ◆ milling in cutting mill (40 mesh) ◆ extraction in

aceton:water=9:1 ◆ drying at 0.003 mbar

◆ ball milling under N2, 48 h ◆ dissolution in 30 mL

2 vol DMSO + 1 vol NMI per g substrate

◆ precipitation in 600 mL 9 vol dioxane + 1 vol water ◆

centrifugation at 5000 rpm to separate cellulose ◆

destillation of solvent, recovery of raw lignin

◆ dissolution of raw lignin in 75%HOAc ◆ centrifugation,

filtration, precipitation in water ◆ centrifugation at

13000 rpm ◆ drying of purified lignin

◆ DCM- and acetone extractions in ASE

◆ lignin dissolution in 5 vol acetone + 1 vol water ◆

addition of washed ion exchanger Amberlite IR748C,

shaking overnight ◆ filtration ◆ washing, freeze-drying

◆ acetylation according to S.Y. Lin and C.W. Dence in

acetic anhydride/pyridine 4Schröder, P., Wollboldt, R.P., Weber, H.K., Fasching, M., Sixta, H.: Lignin Isolation from completely dissolved wood, submitted to Holzforschung

06-03-2007 | 22

Characterization of Lignin Samples

Origin DWL Yield K Lignin AS Lignin Xylan% odw [ ] odw %odw % odw % odw

wood 1 100 141 141 22,4 4,0 15,6

precook11 2 90 101 91 19,5 4,3 14,6

CBC266 3 65 73 48 5,9 3,4 10,4

CBC256 4 54 34 19 1,5 0,8 9,7

CBC211 5 51 14 7 1,1 0,6 9,2

kappa number

0 50 100 150 200 250

10

100

DWL isolated DWL to be isolated

ka

pp

a#

, o

dw

time at 160 °C, min

0 50 100 150 200 250

10

100

= 7

= 19

= 48

= 91

= 141

DWL isolated DWL to be isolated

ka

pp

a#

, o

dw

time at 160 °C, min

DWL, # Yield Xylan Total Sugars Methoxyl% od % od % od % od

141 31 4,4 7,1 18,5

91 52 0,4 0,9 21,8

48 48 0,4 0,7 16,0

DWL appears to be a

facile new method for

lignin isolation,

applicable also for the

isolation of residual

lignin of low-kappa

number pulps

06-03-2007 | 23

HSQC spectrum of E. globulus kraft pulp, # = 91

CDCl3

OMe-5:

-:

g5

g6

g2

S 2,6

a-carbonylS 2,6

-:g

-5:a

O:a:a

O:a

-:a

O:

ppm

ppm

E. glob

wood kappa 91 kappa 48

ß-O-4: a 73,9 6,1 0,4 12% -62%

ß-O-4: a S2 + ß-1? 75,0 6,1 0,4 46% -50%

ß-O-4: a S3 74,8 5,8

Sß-O-4: a -- -- 0,8 40% -53%

ß-O-4: 80,7 4,4 12,4 47% -12%

s-2,6 103,9 6,6 15,4 58% -3%

s-2,6-a carbonyl 105,8 7,1 2,4 1% -42%

g-2 111,0 6,9 2,5 0% -26%

g-6 119,3 6,9 2,4 17% -24%

g-5 122,3 6,9 1,4 -10% -48%

ß-ß: 54,1 3,0 0,4 14% -24%

ß-ß: a 85,5 4,7 1,6 33% -27%

ß-ß: g 71,7 3,9 1,2 12% -32%

ß-ß: g 71,7 4,3 1,0 24% -27%

ß-5: a 88,0 5,4 0,5 -25% -48%

ß-5: 50,2 3,8 0,5 -13% -19%

HSQC acetylated DWL

E. globulus Kraft Pulp

change rel. to nativeAssignment 13C

1H Int (OMe)

06-03-2007 | 24

Quantitative 13C NMR

aliph.OAc

native

kappa 91

kappa 48

phen.OAc

ArC-O ArC-C ArC-H

06-03-2007 | 25

Structural Changes of E. globulus Lignin

during Kraft Pulping

After initial delignification, kappa number 91 odw preferential dissolution of G type units, DWL enriched in S type

units

increase in -O-4-units

moderate increase in -structures

no change in the amount of quarternary carbons in aromatic

nuclei, ArC-C

decrease in -5: a-O-4

During bulk delignification, kappa number 48 odw decrease in S type units, predominantly s-2,6.

decrease in -O-4-units

moderate decrease in -structures

slight increase in the amount of ArC-C

no signicant changes in the amount of aliphat. and aromatic OH

06-03-2007 | 26

Conclusions

Comprehensive kinetic model, based on the concept of Andersson

et al., adequately describes delignification, carbohydrate

degradation, HexA formation and degradation and cellulose chain

scissions.

Distribution model introduced by Andersson et al. suitable for offline

studies and industrial kraft cooks.

New Method for Lignin Isolation: The combination of total

dissolution of lignocellulosic substrates with elements from the

classical MWL preparation appears to be a facile new method for

lignin isolation and presumably also for carbohydrate isolation.

06-03-2007 | 27

Acknowledgment

Wood Kplus and Lenzing AG provided financial

support

Co-workers of Pulp R&D department of Lenzing

AG

06-03-2007 | 28

HexA vs. Kappa number

0 10 20 30 60 80 100

0

10

20

30

[OH-], Mol/L: 0.10 0.52 1.23

He

xA

[m

mo

l/kg

od

w]

Kappa number on odw0 10 20 30 40 60 80 100

0

10

20

30

Temperature, °C: 140 150 170

He

xA

[m

mo

l/kg

od

w]

Kappa number on odw

0 10 20 30 60 80 100

0

10

20

30

[HS-], Mol/L: 0.17 0.28 0.64

He

xA

[m

mo

l/kg

od

w]

Kappa number on odw0 10 20 30 60 80 100

0

10

20

30

[Na+], mol/L: 0.8 1.5 2.5

He

xA

[m

mol/kg o

dw

]

Kappa number on odw

Influence of [OH-] Influence of T

Influence of [SH-]

Influence of [Na+]