TRITA-PMT REPORT 2000:8
ISSN 1104-7003
ISRN KTH/PMT/R--00/8--SE
Precipitation of Kraft Lignin under
Alkaline Conditions
Jonas Sundin
Royal Institute of Technology
Department of Pulp and Paper Chemistry and Technology
Stockholm 2000
Precipitation of Kraft Lignin under
Alkaline Conditions
Jonas Sundin
Doctoral Thesis
Royal Institute of Technology
Department of Pulp and Paper Chemistry and Technology
Stockholm 2000
Division of Pulp Technology
Department of Pulp and Paper Chemistry and Technology
Royal Institute of Technology
SE-100 44 Stockholm
Sweden
Akademisk avhandling som med tillstånd av Kungliga Tekniska Högskolan i
Stockholm framläggs till offentlig granskning för avläggande av teknologie
doktorsexamen, fredagen den 1 december 2000. Avhandlingen försvaras på svenska.
TRITA-PMT REPORT 2000:8
ISSN 1104-7003
ISRN KTH/PMT/R--00/8--SE
Stockholm 2000 KTH Högskoletryckeriet
To my dear wife Maria,
and to my wonderful children
Richard, Jonathan and Linnea
Abstract
In this work, the influence of metal cations on the removal of lignin during kraft pulp
washing has been studied. The interaction of metal cations with kraft lignin in alkaline
solutions has been elucidated by, among other things, the application of ion-selective
electrodes. The effect of the interaction of metal cations with kraft lignin in pulp
washing, as well as its consequences in oxygen delignification and hydrogen peroxide
bleaching stages are reported. An attempt is made to relate the laboratory findings to
industrial process conditions.
The high molecular mass fraction (>1000 g/mol) of kraft lignin was precipitated by
metal cations (Na+ , Ca2+, Mg2+ , Al3+ ) at pH 9, but only calcium and magnesium ions
caused any noticeable precipitation at pH 11-13. Since only the high molecular mass
fraction of the lignin was precipitated, the precipitation can be regarded as the
coagulation of a colloid. The critical coagulation concentration (ccc), that is, the
concentration of the metal cation when the coagulation occurs, was about 3 mmol/L
for Ca2+ at pH 11 and ambient temperature. The ccc(Ca 2+ ) was the same for birch and
spruce kraft lignin but higher for dissolved lignin after oxygen delignification.
A method was developed to estimate the free calcium ion concentration in lignin
solutions during titration with calcium ions by the use of ion-selective electrodes. An
end point was detected for the reaction between calcium ions and lignin. At calcium
ion concentrations below the end point, almost all the added Ca2+ was bound to the
lignin. The maximum content of calcium ions bound to the lignin was found to be
about 30 Ca2+ /100 C-9 units in the kraft lignin. This was confirmed by two other
methods.
The precipitation of kraft lignin by calcium ions during washing of a laboratorycooked kraft pulp resulted in a pulp with a kappa number 15 units higher than a pulp
washed in the absence of calcium ions. The degree of swelling of the fibres in these
pulps was constant. The corresponding difference in kappa number for laboratory
washing of an industrial unbleached softwood kraft pulp was 5 kappa number units
for a pulp produced for bleached products and 10 kappa number units for a pulp
i
produced for unbleached products (sackpaper grade).
The precipitated lignin was darker than the other residual lignin in the pulp. Pulps
containing calcium-precipitated lignin yielded results on oxygen delignification
similar to those of pulps without precipitated lignin. In an alkaline hydrogen peroxide
stage, the precipitated lignin seemed to be more reactive than the other part of the
residual lignin. This is probably due to a higher content of phenolic groups in the
precipitated lignin.
An attempt to estimate the effect of precipitation of lignin by calcium ions during
industrial brownstock washing in a mill, where the calcium and magnesium
concentration in a number of filtered liquors was determined, resulted in 1-2 units
increase in kappa number for a pulp produced for bleached products in a fibre line
with an oxygen stage. The estimate of the effect of magnesium ions is uncertain, but it
may be considerably higher due to the higher concentrations of magnesium in process
liquors.
Keywords: Alkali lignin, Black liquors, Bleaching, Calcium, Coagulation, Colloids,
Kraft pulp, Magnesium, Metals, Mills, Precipitation, Washing
ii
Contents
INTRODUCTION ......................................................................................................... 1
Washing of kraft pulps ............................................................................................ 2
Chemical processes and operations in pulp washing.................................. 3
Technical aspects of pulp washing.............................................................. 6
Lignin ...................................................................................................................... 6
Determination of metal ions in pulp process liquors............................................... 8
Objectives.............................................................................................................. 10
RESULTS AND DISCUSSION .................................................................................. 11
Precipitation of kraft lignin by metal cations in alkaline solutions....................... 11
Influence of different metal cations .......................................................... 11
Influence of molecular mass and functional groups ................................. 16
Influence of different lignins..................................................................... 19
Precipitation of kraft lignin by metal cations during pulp washing ...................... 22
Influence of pH, [Ca2+] and temperature .................................................. 22
Fibre swelling............................................................................................ 24
Influence of wood species......................................................................... 27
Influence of lignin concentration .............................................................. 28
Influence of precipitation of lignin during pulp washing on oxygen delignification
and pulp bleaching................................................................................................. 31
Oxygen delignification.............................................................................. 31
Chromofore content in kraft and oxygen-delignified pulps...................... 34
Final bleaching.......................................................................................... 39
Studies on the interaction between kraft lignin and calcium ions using an ionselective electrode ................................................................................................. 42
Determination of end points...................................................................... 42
Determination of free calcium ion concentration during titrations........... 45
Concentration of metal ions in industrial brownstock washing and its influence on
lignin precipitation ................................................................................................ 51
CONCLUSIONS.......................................................................................................... 57
ACKNOWLEDGMENT.............................................................................................. 60
LITERATURE............................................................................................................. 62
APPENDICES ............................................................................................................. 70
Definition of technical terms and abbreviations.................................................... 70
Nomenclature in bleaching stages......................................................................... 72
iii
List of papers
This thesis is a summary of the following publications, referred to in the text by the
roman numerals I-VI.
I
Sundin, J. and Hartler, N., “Precipitation of kraft lignin by metal cations in
alkaline solutions”, Nord. Pulp Pap. Res. J. 15(4):312-318 (2000).
II
Sundin, J. and Hartler, N., “Precipitation of kraft lignin by metal cations during
pulp washing”, Nord. Pulp Pap. Res. J. 15(4):320-325 (2000).
III
Sundin, J., “Influence of precipitation of lignin during pulp washing on oxygen
delignification and pulp bleaching”, Submitted to Nord. Pulp Pap. Res. J.
(2000).
IV
Sundin, J. and Svedberg, M., “Studies on the interaction between kraft lignin
and calcium ions using an ion-selective electrode. Determination of end points”,
Submitted to J. Pulp Pap. Sci. (2000).
V
Sundin, J. and Svedberg, M., “Studies on the interaction between kraft lignin
and calcium ions using an ion-selective electrode. Determination of free calcium
ion concentration during titrations”, Submitted to J. Pulp Pap. Sci. (2000).
VI
Sundin, J. and Hartler, N., “Concentration of metal ions in brownstock washing
and its influence on lignin precipitation”, Submitted to Nord. Pulp Pap. Res. J.
(2000).
These publications are appended at the end of this thesis.
iv
Introduction
Paper products, which can be defined as layered sheets usually made of cellulosecontaining fibres formed on a fine screen from water suspensions, have been made
since about the year 105 AD in China ("paper" Encyclopædia Britannica Online). The
invention of the printing process in the mid 15th century greatly increased the demand
for paper, and at the beginning of the 19th century wood pulps began to replace rags
as the principal source of fibres for papermaking. Today almost 14 million tons of
pulp and paper products are produced annually in Sweden alone, and about 300
million tons of paper products world-wide. The currently dominating processes for
the production of pulp for production of paper are the kraft process and mechanical
processes, of which the first contributes more than 60% of the total amount of pulp
produced in Sweden. About 30 % of the amount of kraft pulp produced in Sweden is
unbleached (Skogsindustrierna 2000).
Wood consists mainly of three polymers; cellulose, hemicellulose and lignin. In the
production of bleached chemical pulps, lignin is removed almost completely with as
little degradation of the cellulose as possible. In the fibre line of a kraft mill producing
bleached pulp, wood chips are fed to the digester, where most of the lignin is
removed, and the fibres are liberated from each other. The pulp is screened to remove
particles that are not single fibres. Often the pulp is then further delignified in an
oxygen stage before it enters the bleach plant. Earlier the bleaching was mainly
carried out with chlorine-containing chemicals, as for example elemental chlorine.
This resulted, however, in the discharge of toxic and persistent chlorinated
substances. Therefore, environmental pressures initiated the development of today’s
bleaching technology, which is dominated by the totally chlorine free (TCF) and the
elemental chlorine free (ECF) concepts.
To remove products formed during the reactions aiming in degrading the lignin, the
pulp is washed. Along the fibre line, washing of the fibres is carried out a number of
times. In some mills, the first washing stage is already in the final part of the digester
(high-heat washing). After the digester, the pulp is washed, often in several stages,
and this is commonly referred to as brownstock washing. The post-oxygen washing is
1
especially important since this is the final stage of the closed part of the pulp mill
before the pulp enters the not completely closed part of the mill, that is, the bleach
plant. In the bleach plant, the washing operations are also extensive, since the pulp is
washed after each bleaching stage in the bleaching sequence, which normally consists
of about five stages.
Washing of kraft pulps
The primary reason for washing the pulp is to remove soluble substances from the
pulp. Furthermore, pulp is washed to recover cooking chemicals and dissolved
organic material, the latter being recovered for its heating value. However, the
washing must be carried out with a minimum of liquor since this has to be removed
by evaporation before incineration in the recovery boiler.
Due to environmental concerns, kraft pulp mills have increased the degree of system
closure during recent years, since this has been a way of decreasing liquor discharges.
In the mid-1970s the amount of liquor sewered to recipient was normally above 200
m3/ton dry pulp (SSVL 1974) whereas today, in some mills, it has reached a level as
low as 5 m3/ton dry pulp (Sandström et al. 1998). One important factor permitting
lower discharges has been improved washing, both before and after oxygen
delignification and in the bleach plant (Annergren, Sandström 1996). Improved
washing before oxygen delignification reduces the carry-over of organic material
from the digester. This reduction improves the efficiency of the oxygen stage as well
as the strength properties of the pulp (Miller et al. 1991; Andtbacka 1998). Improved
washing before the bleach plant reduces the consumption of bleaching chemicals and
reduces the amount of AOX formed in an ECF sequence, and it improves the
selectivity in a TCF sequence (Dence, Reeve 1996). It also facilitates increased
closure of the bleach plant, since the carry-over of both inorganic and organic
substances to the bleach plant is reduced. Another important aspect of brownstock
washing is the removal of extractives, since pulp washing is one of the more
important process stages in deresination and pitch control in a kraft pulp mill (Allen
1997).
2
One drawback of increased system closure is that it leads to higher levels of organic
and inorganic substances in the liquors, and this may cause process problems, such as
scaling, the formation of deposits and corrosion (Ulmgren 1997). It may also give rise
to the precipitation of dissolved lignin (paper 2) and extractives (Björklund-Jansson et
al. 1985). This may cause problems not only in washing stages but also in the
subsequent process stages, since these substances interfere with bleaching and
papermaking.
These problems may, to some extent, be overcome by purging the non-process
elements (NPEs) in process streams exiting the process, as for example green liquor
dregs (Ulmgren 1997). Another approach, which has not yet been applied industrially,
is the leaching of wood chips before the digester (The Ecocyclic Pulp Mill 2000). The
basis of this approach is that most of the NPEs enter the mill with the wood (Ulmgren
1997). Softwood contains about 600-1200 ppm calcium (Fossum et al. 1972; Bailey,
Reeve 1994; Bailey, Reeve 1996; Allison et al. 2000) and hardwood up to 2600 ppm
(Allison et al. 2000). In the case of calcium, Douek and Allen (1978) found that about
80% of the calcium to the digester and brownstock washers came from the wood and
about 20% from the white liquor.
Chemical processes and operations in pulp washing
In the washing of pulp a number of chemical processes are taking place, which may
be divided into chemical reactions and material transport. Since the conditions in the
stage preceding the washing and in the washing stage conditions differ considerably,
this results in altered driving forces for the chemical reactions. Some examples of
these are a change in pH, a change in temperature and a change in concentration of
substances. In the latter case, this may be not only substances directly involved in the
reaction but also substances contributing to the ionic strength and thus changing, for
example, the equilibrium constants.
In literature there are several papers reporting on the interaction of metal cations with
pulp fibres. One approach of relating the amounts of different metal cations in pulp
fibres under different conditions is based on ion-exchange theories (Ohlsson, Rydin
1975; Bryant, Edwards 1996). Another common approach is to describe these
3
processes as sorption phenomena. Thrinh and Crotogino (1987b) reported the strong
influence of pH on the amount of sodium ions bound to the fibres, and described the
sorption of sodium ions by a Langmuir type equation. A third approach is the
application of the Donnan equilibrium theory (Towers, Scallan 1996). However, the
different approaches all relate the origin of the phenomena to a common issue, the
charged groups in the fibre. Rosen (1975) showed that both the carboxylic groups in
the lignin and carbohydrates and the phenolic groups in the lignin take part in the
reactions.
Three processes are important for material transport in the washing of pulp: dilutiondewatering, displacement and diffusion.
Displacement is the process whereby the original liquor surrounding a fibre bed is
ideally replaced by moving the wash liquor in a piston-like movement through the
fibre bed. In reality, the displacement is disturbed by the mixing of the two liquors in
the wash front and by channeling effects in the bed (Rydholm 1965, p 719). Sodium
ions and lignin have been found to be transferred from unwashed pulp to the wash
liquor at similar rates (Edwards, Rydin 1975; Trinh, Crotogino 1987a). On the other
hand, Grähs (1976) found that lignin was displaced faster than sodium ions in
displacement washing but Grén and Ström (1985) claimed that this difference was
due to sorption effects on sodium ions.
McKibbins (1960) used Fick’s second law of diffusion in the area of washing of kraft
cooked wood chips that is, for instance, applicable to high-heat washing. He found
that experimental results on the extraction of sodium ions correlated well with
diffusion theory. He also reported higher diffusion rates at higher temperatures, which
is expected from diffusion theory. In free diffusion, the relationship between
temperature and diffusion coefficient is given by the Stoke-Einstein equation (Bird et
al. 1960, p 514). This equation predicts that the diffusion coefficient will increase
somewhat more than linearly with temperature, since the solute viscosity will
decrease with temperature:
4
DAB =
kT
6πµ B RA
(1)
where DAB is the diffusivity or diffusion coefficient of particle A in solute B, µB the
solute viscosity and RA the radius of the particle, where the hydrodynamic radius (Rh)
is often used. However, Favis et al. (1983b) found a much higher increase in
diffusivity of lignin with increasing temperature at temperatures above 70 °C, when
leaching an unbleached pulp at 1% pulp consistency in distilled water, than is
predicted by Eq. 1. Eriksson and Grén (1997) observed a similar dependence of the
diffusivity of lignin in pulp on temperature. Favis explained this deviation as being
due to physical changes in the fibre wall structure associated with the thermal
softening of the hemicellulose in the fibre wall, and he referred to an earlier
investigation where the thermal softening temperature, Tg, for hemicelluloses in
samples isolated from wood had been determined to be about 55 °C at about 20%
water content (Goring 1963). Later studies have shown that Tg for hemicelluloses in
wet conditions is definitely lower, and approaches the freezing point of water (Back,
Salmén 1982; Salmén 1990). Moreover, Favis et al. assumed in their calculations of
diffusivity that the lignin’s hydrodynamic radius was unchanged with temperature.
This may be questionable since lignin is known to associate/disassociate when
conditions change. Li and MacLeod (1993) also found a positive effect of increasing
temperature during leaching, but the increase in leaching with temperature was
accelerated to a much less extent above 70 °C than Favis had reported. Li carried out
the experiment at 0.1 M NaOH, while Favis et al. (1983a) worked at pH 6-8 and
Eriksson at pH 7-8.
The transport of lignin out of the fibre wall by diffusion has been modelled. Favis et
al. (1981) approximated the fibre as an infinite flat plate, while Li and MacLeod used
a hollow cylinder. The predictions of these two models are similar when the ratio of
the outer to the inner radius of the cylinder is less than 10 (Crank 1975, p. 83). Li and
MacLeod used a ratio of 1.5 in their calculations. Jacobson and Nguyen (1998b;
1998a) presented results from calculations using the same model as Li and MacLeod
but with several different boundary conditions relevant for washing on an industrial
5
scale.
Technical aspects of pulp washing
The washing of kraft pulp in pulp mills is based on dilution/extraction and
displacement, and usually on combinations of these (Crotogino et al. 1987). Parallel
with these processes, diffusion of soluble substances from the interior of the fibre wall
to the liquor surrounding the fibre takes place. There are a number of different types
of washing equipment available. For an overview, see for example Gullichsen and
Fogelholm (2000). A combination of a number of apparatus using different washing
principles is often used in a wash line. Several parameters and models have been
developed to describe the properties and efficiency of washing equipment, for a
review of these see, for example, Pekkanen and Nordén (1985) and Crotogino et al.
(1987).
In industrial pulp washing systems, there are a number of problems associated with
the apparatus. Some examples are the formation of scale on dewatering equipment,
which reduces the dewatering capacity, and the inclusion of air which gives rise to the
formation of foam. These runnability problems lower the efficiency of the washing
and may increase the carry-over of dissolved substances to the subsequent stages,
increase the wash liquor demand, which increases the evaporator load, and reduce the
production capacity.
Lignin
Native wood lignin is an amorphous, branched polydisperse macromolecular
substance. The building blocks in these polymers are phenylpropane units. Normal
softwood contains 26-32 % lignin, while normal hardwood contains 20-28% lignin
(Sjöström 1981). The lignin contains several functional groups, of which phenolic
hydroxyl, benzylic hydroxyl and carbonyl groups are among the more important for
the reactivity of the lignin. Kraft lignin contains acidic groups that bind cations.
Catechol compounds, that is, aromatic compounds containing ortho-substituted
dihydroxyl groups, are known to form stable complexes with multivalent metal
cations (Sillén, Martell 1971). These compounds have been found in kraft black
6
liquor (Niemelä 1989). Besides the aromatic hydroxylic groups, aliphatic carboxylic
groups in the lignin may also form complexes with metal cations (Mathuthu, Ephraim
1993).
Kraft lignin has a broad molecular mass distribution (Sarkanen et al. 1982). During
the cook, the lignin is partly degraded and hydrophilic groups are formed in the lignin
(Gellerstedt, Lindfors 1984a; Robert et al. 1984). The ionization of these groups is
essential in order to make the lignin water-soluble (Gellerstedt, Lindfors 1987), which
is a prerequisite for removing lignin during cooking and washing.
Different conformations or configurations of lignin macromolecules in solution have
been reported in the literature. Goring summarized his and others work (1971) and
concluded that alkali lignins behave like Einstein spheres, that is, as a compact,
incompressible sphere. In a later work, Goring et al. (1979) studied lignosulfonates
and found that they adopted the conformation of a flat disk when adsorbed on a
carbon film. When presenting their model for the leaching of lignin, Favis and Goring
(1984) concluded that this conformation was compatible with the proposed
interrupted lamella model for the ultrastructural arrangement of lignin and
carbohydrates in the wood cell wall (Kerr, Goring 1975). However, Goring stated
earlier (1971) that alkali lignins behave more like Einstein spheres than
lignosulfonates. Furthermore, in the paper where Goring and co-workers reported that
the lignosulfonates form flat disks on carbon film, they stated that these molecules
would probably assume a spherical conformation in solution. Both Dong and Fricke
(1995) and Garver and Callaghan(1991) found that kraft lignin in alkaline solution
adopted the conformation of a compact sphere. In a paper regarding the modelling of
the three-dimensional structure of lignin, Jurasek (1995) discusses the influence of
spatial restrictions on the shape of the formed lignin in different parts of the cell wall,
and relates these differences in shape to chemical differences in the lignin. This may
indicate that lignin probably exists in different conformations depending on the spatial
origin of the lignin macromolecule in the cell wall, and this may perhaps be an
7
explanation of the different conformations obtained in the experimental studies.
Colloids are generally defined as systems with a dispersed phase having a size in, at
least one dimension, between 1 and 1000 nm (Everett 1988). Since lignin is a
macromolecule with colloidal dimensions (Garver, Callaghan 1991; Jurasek 1995), it
may also take part in reactions characteristic to colloids. Marton (1964) stated that
lignin behaves like a hydrocolloid; the lignin precipitates as the pH is lowered with
the simultaneously protonation of the acidic groups in the lignin. He also found that
the pH at which the lignin starts to precipitate increases with increasing concentration
of sodium ions. Numerous studies have revealed association phenomena between
lignin molecules in solution. Increased lignin concentration (Sarkanen et al. 1984;
Rudatin et al. 1989), increased ionic strength (Rudatin et al. 1989) and decreased pH
(Benko 1964; Lindström 1979; Woerner, McCarthy 1987; Rudatin et al. 1989;
Garver, Callaghan 1991) have been found to promote the association of lignin
molecules.
Lindström (1980) showed that lignin is coagulated in the presence of multivalent
metal cations, but he only studied the pH 4-9 range. Moreover, in most studies of the
interaction between lignin and metal cations, only the high molecular mass fraction of
isolated lignin has been used (Lindström 1980; Nyman et al. 1986).
Determination of metal ions in pulp process liquors
The total concentration of calcium has been measured in green, white and black
liquors, and in pulp and bleach-plant filtrates using digestion under highly oxidizing
conditions followed by analysis with flame atomic absorption (AAS) or inductively
coupled plasma (ICP) (Järvinen, Välttilä 1998; Sirén 1998). These methods are useful
when, for example, material balances are to be studied.
When equilibrium reactions are to be studied, the use of ICP or AAS techniques is
limited, since they only measure the total concentration, while the reactions and
equilibrium are dependent on the free (unbound) concentration. However, if proper
sample pre-treatment is used, the ICP/AAS techniques may be useful for determining
other concentrations than the total, for example the total concentration of soluble
8
compounds. Careful filtration is an example of such pretreatment.
One way to determine the concentration of free ions is by using an ion-selective
electrode, the voltage of which is dependent on the concentration of the free ions
(Morf 1981):
E = E ′ + S ⋅ log aion
(2)
where E is the measured voltage; E’ is the voltage resulting from combining factors
that are assumed to be constant, for example the voltage of the reference electrode; S
is the slope, which in ideal cases (nernstian response) is equal to 29.6 mV per
log[Ca2+ ] for calcium ions at 25 °C, and aion is the activity of the measured ion and is
related to the concentration by the Debye-Hückel formula:
a = γ ⋅c
(3)
where γ is the activity coefficient, which is dependent on the ionic strength.
Some examples of more frequently used ion-selective electrodes employed for the
determination of ions in liquors from the pulp industry are the sulfide (Swartz, Light
1970; Papp 1971a; Papp 1971b; Korhonen, Lumme 1978), sodium (Lenz, Mold 1971;
Korhonen et al. 1973b; Korhonen et al. 1973a) and pH electrodes (Milanovo, Dorris
1994; Ulmgren et al. 1994). The membranes of these electrodes are made of solid
state material, Ag2S in the sulfide electrode and glass in the other two. The calcium
electrodes are of a type called the liquid membrane electrode. The membrane is made
of PVC containing a calcium-ion-selective molecule, an ionophore, which reacts with
free calcium ions in the sample. This type of electrode is somewhat more sensitive to
aggressive samples than the solid state electrode, since there is a risk of leakage of
ionophores and other additives from the membrane.
Few attempts have been made to use calcium ion-selective electrodes to determine the
free concentration of calcium ions in pulp mill liquors. Lindström et al. (1988) studied
the interaction between calcium ions and lignin (Indulin AT) with a calcium
electrode, but the lignin was removed from the solution before the measurements
were made.
9
Objectives
The aim of this work was to study the influence of metal cations on the removal of
lignin from kraft pulps during washing. The cause of the interest in this topic was the
observation made by Mao and Hartler (1994) of a considerable difference in lignin
content of pulps washed in tap water and in deionized water. Since Goring and coworkers had reported a strong influence of divalent metal cations on the leaching of
lignin from a kraft pulp, and since the calcium concentration in the tap water used by
Mao was of the same order as the calcium concentration in the process liquors in a
kraft pulp mill (Douek, Allen 1978), the focus of the work was on calcium ions.
10
Results and discussion
Precipitation of kraft lignin by metal cations in alkaline
solutions
(Papers I and III)
In this section, the interaction between some metal cations and kraft lignin under
different conditions in alkaline solutions is described. The next section deals with the
same issues, but with pulp fibres present.
Influence of different metal cations
Fig. 1 shows the typical dependence of the lignin concentration on metal cation
concentration. When the cation concentration, in this case the sodium ion
concentration, exceeded a certain limit, the lignin started to precipitate and the lignin
concentration in the solution decreased. If the lignin is considered to be a colloid, the
precipitation is termed coagulation, and the cation concentration at which the
coagulation occurs is called the critical coagulation concentration (ccc).
Lignin concentration, g/L
0.5
¡‘
¦
¦
²
¡‘
²
¡ ‘
²
¦
¡
¦
‘
²
¦
¡
²
‘
0.4
‘
¡
²
¦
¡
²
²
¡
‘
0.3
‘
0.2
²
¡
‘
¦
0.1
pH 13, 25 °C
pH 11, 25 °C
pH 11, 80 °C
pH 9, 25 °C
¦
¦
0
0
500
1000
1500
2000
[Natot], mmol/L
Fig. 1. The effect of sodium ion concentration on lignin in solution at different pH
values at 25 and 80 °C. The lignin concentration was determined from UV
absorbance spectra at 280 nm after filtration with a 0.2 µm filter. The lignin solution
was prepared by dilution of a spruce black liquor obtained by laboratory cooking of
spruce wood chips at 22% E.A. and 30% sulfidity, resulting in a pulp with a kappa
number of about 25.
11
Lignin concentration, g/L
0.5
² pH 13
¡ pH 11
¦ pH 9
¦
¦
²
¦
¡²
¡¦ ²
¡²²
0.4
¡
0.3
¦
0.2
²
¡
¦
¡
0.1
²
¡
¦
²
¡
¦
0
0
5
10
15
20
25
30
35
[Catot], mmol/L
Fig. 2. The effect of calcium ion concentration on spruce kraft lignin in solution at
different pH values at 25°C. See Fig. 1 for a more extensive description of the
experimental conditions.
At 25 °C, precipitation occurred only at pH 9. This indicates that the lignin is not as
easily precipitated by sodium ions when the pH is higher than this level. The increase
in the stabilization against coagulation with increasing pH is probably caused by
ionization of phenol groups in the lignin, which leads to both an increased
electrostatic repulsion and an increased steric repulsion between lignin molecules.
An increase in the temperature seems to have a destabilizing effect on lignin.
However, this may also be explained by an effect of the decrease in the autoprotolysis
constant of water (Kw) with increasing temperature.
Calcium ions had a much more pronounced effect on lignin precipitation than sodium
ions, as shown in Fig. 2. The precipitation of lignin occurred at a considerably lower
concentration of calcium than of sodium ions. Moreover, lignin precipitation by
calcium ions occurred also at pH 11 and 13, in contrast to the precipitation with Na+,
which occurred only at pH 9 at 25 °C, as seen in Fig. 1. The increase in ccc from pH
11 to pH 13 may, in addition to the effects discussed above for sodium ions, be due to
the formation of CaOH+ and Ca(OH)2(s), which lowers the concentration of free Ca2+
ions in solution (Sillén, Martell 1964). The concentration of carbonate in the solutions
was probably low since deaerated deionized water was used in the preparation of most
liquors used in this work and the liquors were kept under a nitrogen atmosphere.
12
Lignin concentration, g/L
0.5
‘
¡
‘
¡¡
0.4
‘¡
0.3
0.2
‘
25 °C
¡
‘
¡
‘
¡
‘
¡
‘
80 °C
0.1
0
0
5
10
15
20
25
30
35
[Catot], mmol/L
Fig. 3. The effect of calcium ion concentration on spruce kraft lignin in solution at pH
11, at 25 and 80 °C. See Fig. 1 for a more extensive description of the experimental
conditions.
As was the case with sodium ions, an increase in temperature lowered the ccc as well
as the lignin concentration at a calcium ion concentration above the ccc, as shown in
Fig. 3.
The large difference in ccc between monovalent sodium ions and divalent calcium
ions for the coagulation of kraft lignin has been observed earlier (Lindström 1980;
Nyman et al. 1986; Månsson, Öster 1988), and is in accordance with the theory of
Derjaguin, Landau, Verwey and Overbeek (DLVO) (Evans, Wennerström 1994). The
DLVO theory relates the interaction free energy to the particle separation distance,
considering the attractive van der Waals and the repulsive electrostatic forces. This
theory predicts that the concentration of an electrolyte at which a colloid is
precipitated is related to 1/z6 for highly charged surfaces and 1/z2 for lowly charged
surfaces, where z is the valency of the ion. The effect of the electrolyte is a
compression of the electrical double layer; this reduces the repulsive electrostatic
force, and coagulation may occur. Whether the surface is regarded as highly or lowly
charged depends on the value of the expression: zeΦ0/kT, where z is the valency of the
electrolyte, e is the charge of an electron, Φ0 is the surface potential and k is the
Boltzmann’s constant. The expression is >>1 for a highly charged surface and <1 for
a lowly charged surface. Considering the zeta potentials of lignin determined by Dong
13
et al. (1996), the surface charge of lignin above pH 8 should probably be regarded as
being intermediate between a lowly and highly charged surface.
The DLVO theory predicts that the ccc will be higher at 80 °C than at 25 °C. The
results in Fig. 1 and 3 show instead a decrease with temperature. When studying
effects of cations on the high molecular mass fraction of Indulin at pH 6, Lindström
(1980) also found a decrease in ccc with increasing temperature for Ca 2+ but not for
Na+ . He explained these results as being due to a reduced steric stabilization at higher
temperatures.
One of our observations is that the precipitate was denser at 80 °C than at 25 °C. This
is in accordance with results from the precipitation of lignin with carbon dioxide
(Alén et al. 1979). On the other hand, Lindström (1982) found that syneresis, that is,
the densification of a gel, upon heating of a lignin gel occurred only below pH 9.
The results of the Mg2+ addition are given in Fig. 4. The ccc decreased substantially
when the pH was increased from 9 to 11. At pH 13, the ccc was intermediate between
the ccc at pH 9 and 11, but the remaining lignin concentration in solution at
magnesium ion concentrations above the ccc was higher at pH 13 than at pH 9 and 11.
Lignin concentration, g/L
0.5
‘¡
¦¦ ¦
²
¦
¡
‘
¦²
²
¡
²
0.4
²
² pH 13, 25 °C
¦ pH 9, 25 °C
‘ pH 11, 80 °C
¡ pH 11, 25 °C
¦
²
¡
0.3
²
²
‘
0.2
¦
¡
‘
¦
‘
¡
0.1
‘
¡
‘
¡
0
0
5
10
15
20
25
30
35
[Mgtot], mmol/L
Fig. 4. The effect of magnesium ion concentration on spruce kraft lignin in solution at
different pH values at 25 and 80 °C. See Fig. 1 for a more extensive description of the
experimental conditions.
14
Equilibrium calculations were performed with the program Mineql+ (2000) (using
USEPA’s thermodynamic database MINTEQA2). From these calculations at the
actual total magnesium ion concentration, the dominating magnesium species above
pH 10 is Mg(OH)2. In comparison, Ca2+ is still the dominating calcium species at pH
12. In spite of the calculated low concentration of free Mg 2+ ions at pH 11, the
coagulation by magnesium was more effective at pH 11 than at pH 9, both in terms of
ccc and in remaining lignin concentration in solution at magnesium concentrations
above ccc. Heterocoagulation might be the explanation of this behaviour.
Heterocoagulation occurs between two different types of particles. When the two
particles have different charges, both the van der Walls and the electrostatic forces are
attractive, and this provides the conditions for an effective coagulation. Lignin
molecules have a negative zeta potential above pH 1 (Dong et al. 1996). Isoelectric
point (IEP) values of Mg(OH)2 are reported to be above pH 12 (Parks 1965;
McLaughlin et al. 1993), and the Mg(OH)2 particles are positively charged below the
IEP (McLaughlin et al. 1994). Thus, heterocoagulation between positively charged
Mg(OH)2 particles and negatively charged lignin molecules may explain the effective
coagulation at pH 11.
Lignin concentration, g/L
0.5
²
¦
0.4
Al
²
¦
¡
²
3+
Ca
0.3
¡
¡
2+
¡
Na+
¦
²
0.2
¦
¡
² ² ²
¦ ¦ ¦
0.1
0
0.1
¡
1
10
¡
100
1000
10000
Metal ion concentration mmol/L
Fig. 5. The influence of Al3+, Ca2+ and Na+ on spruce kraft lignin in solution at pH 9
and 25 °C. See Fig. 1 for a more extensive description of the experimental conditions.
15
Results for three cations of different valences are given in Fig. 5. The ccc decreased
with increasing valency, which is in accordance with the DLVO-theory. Compared
with the ccc for Na+ , the ccc for Ca2+ was about three times lower than predicted by
the DLVO theory, assuming that the lignin has a highly charged surface, as discussed
above. This difference between theory and experimental results indicates that there
are specific interactions between Ca2+ and functional groups in the lignin. One
example of such groups is the catechol groups, which are known to form complexes
with multivalent metal cations (Sillén, Martell 1971).
Influence of molecular mass and functional groups
Since the precipitation of lignin is a colloidal phenomenon, the molecular mass of the
lignin can be expected to be important.
A solution of Indulin AT was separated into different molecular mass fractions by
dialysis. The molecular mass cut-off of the dialysis membrane was 1000 g/mol. Fig. 6
shows that the fraction of Indulin below 1000 g/mol was not affected by Ca2+ at all.
The lignin in the fraction above 1000 g/mol is on the other hand precipitated to a
Lignin concentration, g/L
0.06
Mw<1000 g/mol
0.05 ¡¡¡ ¡
0.04
0.03
¡
¡
¡
²
¦²
¦
²
¦
²
0.02
Undialysed
²
²
¦
¦
0.01
²
²
¦
Mw>1000 g/mol
¦
0
0
5
10
15
20
25
30
35
[Catot], mmol/L
Fig. 6. The effect of calcium ion concentration on lignin in solution at pH 11 and
25°C for different molecular mass fractions separated by dialysis. The lignin solution
was prepared from Indulin AT, which is a commercially available acidified pine kraft
lignin. The dialysed samples were prepared in dialysis tubes with a molecular mass
cut-off of 1000 g/mol immersed in water at pH 11 at room temperature.
16
0.03
1000 g/mol
Absorbance, a.u.
5400 g/mol
152 g/mol
0.02
Without calcium
With calcium
0.01
0.00
0
10
20
30
40
50
60
70
80
Time, min.
Fig 7. Molecular mass distributions determined by GPC for lignin at pH 11 and 25
°C in a dilute spruce black liquor. One sample was treated with 32 mmol/L Ca2+
before GPC; both samples were filtered with a 0.2 µm filter before GPC.
greater extent, giving a lower lignin concentration in solution at Ca2+ concentrations
above the ccc, while the ccc is unchanged compared to the undialysed sample. Results
from gel permeation chromatography (GPC) verified that it was mainly the high
molecular mass lignin that was precipitated, Fig. 7.
An explanation of why only larger molecules coagulate is that the van der Waals
attractive forces are much more long range for larger molecules/colloid particles than
for small molecules (Everett 1988, p. 35).
Lignin is known to associate and form larger aggregates when the pH is lowered
(Rudatin et al. 1989). This increase in size probably increases the fraction of lignin
that is large enough to be precipitated by Ca2+ . This may explain the decrease in
lignin concentration after coagulation (for example at a calcium ion concentration of
15 mmol/L) with decreasing pH, as shown in Fig. 2.
The contents of phenolic hydroxyls and carboxylic acids are given in Table 1. In order
to avoid the contribution of carboxylic acids in carbohydrates, the content of
carboxylic acids was determined on lignin purified from the black liquor. The
17
Table 1. Content of functional groups in spruce kraft lignin.
Sample
Phenolic hydroxyls
mmol/g
Black liquor
Black liquor + Ca
3.1
2+ d)
Lignin
Lignin + Ca
2+ d)
a)
per 100 C-9 units
Carboxylic acids
b)
mmol/g
c)
per 100 C-9 units
56
4.3
78
2.7
49
1.1
20
3.4
62
2.6
47
a)
Confidence limits of analysis: ±0.15
b)
Calculated using a molecular mass of 182 g/mol for C-9 units.
c)
Confidence limits of analysis: ±0.1
d)
Remaining solution after filtration of a solution containing 10 mmol/L Ca2+ and 2 g/L lignin
absorptivity at 280 nm of this spruce kraft lignin at pH 11 was determined to be 24.7
L·g-1·cm-1 (paper III) which is close to the value (24.6 L·g-1·cm-1) for a pine kraft
lignin reported by Fengel et al. (1981). The absorptivity was used to estimate the
lignin concentration in aqueous solutions. The content of carboxylic acidic groups in
lignin from industrial softwood kraft black liquors is reported to be in the range of 1520 per 100 C-9 units (Marton 1964; Mörck et al. 1986; Månsson, Öster 1988;
Gellerstedt, Heuts 1997), and the corresponding value for phenolic groups is 55-80
per 100 C-9 units (Marton 1964; Månsson 1983; Robert et al. 1984; Mörck et al.
1986).
The contents of both phenolic hydroxyls and carboxylic acids were higher in the
lignin remaining in solution after the addition of calcium ions and filtration, but the
increase in the content of carboxylic acids was almost twice the increase in phenolic
hydroxyls. The higher content of ionizable groups in the unprecipitated lignin
indicates that the low molecular mass fraction of the lignin, which is enriched in the
lignin remaining in solution after precipitation as shown in Fig. 6, contains a larger
proportion of charged groups than the high molecular mass part of the lignin. Mörck
et al. (1986) determined the contents of phenolic hydroxyls and carboxylic acids in
different molecular mass fractions from a softwood kraft lignin, and found higher
contents of both phenolic hydroxyls and carboxylic acids in the low molecular mass
fraction than in the rest of the lignin. This increase was of the same order as that
found in our data.
18
b)
Influence of different lignins
In Table 2, results from some of the precipitation experiments are summarized. In
most cases, about two thirds of the lignin was precipitated. The ccc for the lignin in
Indulin AT seems to be similar to the ccc for the lignin in the spruce black liquor. It is
uncertain whether the differences in ccc and in the fraction of remaining lignin in
solution are due to the lower lignin concentration, to differences in the lignin or to
other differences in composition between the samples. The pH-dependence of the ccc
of calcium ions in dilute birch black liquor was the same as that in dilute spruce black
liquor, but the ccc of Ca2+ was generally somewhat lower for birch kraft lignin than
for spruce kraft lignin. However, the part of the lignin that did not precipitate was
somewhat larger in the case of the birch kraft lignin. This may be because the birch
kraft lignin has a lower molecular mass than the spruce kraft lignin (Mörck et al.
(1988)), since it for softwood kraft lignin was the fraction of the lignin with a
molecular mass above 1000 g/mol that was precipitated (Fig. 7).
Table 2. Data for precipitation experiments with calcium ions and different kraft
lignin solutions.
Solution
pH
ccca)
Remaining
mmol/L
ligninb), %
Diluted spruce black liquor
9
3.5
29
0.45 g/L
11
2.5
35
13
7
38
Indulin AT
9
1.4
34
0.04 g/L
11
2.5
44
13
7.5
59
Diluted birch black liquor
9
2.6
48
0.5 g/L
11
1.7
38
13
5.2
44
a)
The critical coagulation concentration (ccc) was defined as the concentration of calcium ions at the
lignin concentration (C0+C1)/2, where C0 and C1 is the lignin concentration before and after
coagulation.
b)
Remaining lignin is the fraction of lignin remaining in solution at a calcium ion concentration equal
to twice the ccc.
19
Lignin concentration, g/L
0.6
¦¦¡
¡
0.5 „
¦
0.4
Dialysed, Mw<1000 g/mol
¡¡
„
¡
¦
¦
0.3
Undialysed
¡
¦
¡
¦
0.2
¡
¦
Dialysed, Mw>1000 g/mol
0.1
0
0
5
10
15
20
25
30
35
[Catot], mmol/L
Figure 8. The effect of calcium ion concentration on oxygen lignin in solution at pH
11 and 25°C. The solution was prepared from the liquor pressed out of a pulp after
an oxygen stage. The dialysed samples were prepared in dialysis tubes with a
molecular mass cut-off of 1000 g/mol immersed in water at pH 11 at room
temperature.
Fig. 8 shows the influence of calcium ion concentration on lignin from a liquor
pressed out of the pulp after an oxygen stage. The absorptivity of the dissolved lignin
after oxygen delignification (oxygen lignin) was determined to be 20.5 L·g-1·cm-1 at
pH 11; this value was used to estimate the lignin concentration in aqueous solutions.
The oxygen lignin with a molecular mass above 1000 g/mol was precipitated by
calcium ions. The sample denoted ”Dialysed, Mw<1000 g/mol” is from the liquor
passing the dialysis membrane. At 32 mmol/L Ca2+ only 15% of this lignin was
precipitated, while 50% of the lignin not passing the dialysis membrane was
precipitated. In the case of kraft lignin, no precipitation of the lignin passing the
membrane occurred (Fig. 6). This was interpreted as the precipitation was a
coagulation of lignin molecules or aggregates of colloidal size (1-1000 nm). This is
also most probable true for the oxygen lignin. The reason why some of the low
molecular mass lignin was precipitated may be either that the limit in size of lignin
molecules that behave like a colloid is lower for the oxygen lignin or that larger
molecules passed the membrane in the dialysis of oxygen lignin. Dialysis gives no
absolute separation at a given size but rather a distribution of molecular size around a
certain value.
20
The fraction of the oxygen lignin remaining in solution was about 1.7 times the
fraction of kraft lignin in spruce black liquor. The ccc for the dialysed oxygen lignin
was 4 mmol/L Ca2+ compared to 2.5 mmol/L for kraft lignin. The ccc(Ca2+ )for the
undialysed oxygen lignin was, however, twice the ccc for the dialysed oxygen lignin.
This may indicate the formation of a complex between calcium ions and the low
molecular mass oxygen lignin, since this would require a higher total calcium ion
concentration to yield a corresponding free calcium ion concentration. This effect was
not observed in the case of kraft lignin, but this difference can probably be explained
by a higher amount of carboxylic groups and a larger fraction of low molecular mass
lignin in the oxygen lignin in oxygen than in the kraft lignin. The number of
carboxylic groups in dissolved oxygen lignin found by Gellerstedt and Lindfors
(1987) was 50-60/100 C-9 units, while the content of phenolic groups was 25-40/100
C-9 units.
Magnesium ions precipitate kraft lignin in a similar way as calcium ions (paper I). If
this is also the case for oxygen lignin, then the addition of magnesium ions to an
oxygen stage could result in a precipitation of lignin in the pulp. Results from
measurements of magnesium concentrations in liquors from an industrial oxygen
stage (paper VI) indicate that the magnesium concentration is almost of the same size
or below the ccc(Ca2+) for undialysed oxygen lignin but above the ccc(Mg 2+ ) for kraft
lignin.
21
Precipitation of kraft lignin by metal cations during pulp
washing
(Paper II)
Influence of pH, [Ca2+] and temperature
Fig. 9 shows the relationship between kappa number of washed pulps and the
concentration of calcium ions during washing at different pH-levels. In washing
without calcium ions, that is, washing with pH-adjusted deionized water, the kappa
number was reduced from 25 at pH 9 to 21 at pH 13. This positive effect of increasing
pH on kappa number reduction is well known (Hagström-Näsi et al. 1987; Wilcox,
Goring 1990; Li, MacLeod 1993).
When washing was performed in the presence of Ca 2+ , there was a strong correlation
between the kappa number after washing and the Ca2+ concentration. At pH-values
between 9 and 12, the kappa number increased almost linearly with increasing
calcium ion concentration, but at pH 13 the effect was less strong, probably due to the
formation of CaOH+ and Ca(OH)2. This supports the conclusion that it is the divalent
Kappa number
40
35
¦
„
30
¦
„
25 ¦
„
²
¡
20
0
¦
„
²
pH 9
pH 11
pH 12
pH 13
¦
„
²
¡
¦
„
²
¡
¡
²
²
¡
1
¡
2
3
4
5
[Catot], mmol/L
Figure 9. The influence of calcium ion concentration on kappa number at different pH
values during the washing of a spruce kraft pulp. The pulp was produced by
laboratory cooking of spruce wood chips at 22% E.A., 30% sulfidity and 170 °C.
Washing at 25 °C in columns for 16 hours.
22
Table 3. Spruce kraft pulps used for used for oxygen delignification
Sample
H-factor
2+
[Ca ]tot
mmol/L
Kappa
a)
Klason lignin
number
%
c)
Klason lignin/
Light abs. coeff.
Kappa number
(457 nm) m2/kg
23.3
1100-W
1100
~0b)
31.9
3.8
0.12
1500-W
1500
~0
b)
23.4
2.6
0.11
21.1
1500-Ca
1500
5
32
3.8
0.12
29.1
a)
Concentration in washing liquor.
b)
Deionized water
c)
Acid-soluble lignin not included
calcium ions that lead to an increase in the amount of substances in the pulp that
contribute to the kappa number. Most of these substances are probably lignin, but
lignin-carbohydrate complexes as well as other permanganate-consuming substances
could also be a part of these.
In Table 3, three spruce kraft pulps are compared. One of the pulps (1500-Ca)
contained calcium-precipitated lignin. The ratios of Klason lignin to kappa number
for the three pulps in Table 3 were almost the same. This shows that the lignin
precipitated by calcium ions during washing consumed the same amount of
permanganate in the kappa number determination as the other residual lignin in the
three pulps. Differences in kappa number of spruce kraft pulps washed with different
calcium ion concentrations are thus caused by differences in the pulps’ lignin content,
and not due to other permanganate-consuming species.
The value of the ratio of Klason lignin to kappa number is somewhat lower than the
values commonly used. Tasman and Berzins (1957) reported a value of 0.13 and
Tasman (1959) reported two years later a value of 0.15. They did not, however,
remove the extractives before the determination of the Klason lignin content, which
means that the content of extractives is added to the lignin content. When the content
of the extractives of the pulps in Table 3 is added to the lignin content, a ratio of 0.14
is obtained.
In the study of the effect of metal cations on lignin in solution, the coagulated lignin
was separated by a filter with a pore size of approximately 200 nm. Since most of the
23
pores in pulp fibres are smaller than this (Li et al. 1993; Berthold, Salmén 1997), the
retention of coagulated lignin is probably even higher in the pulp than in the filter.
The coagulated lignin may also adhere to the outer fibre surface. When lignin was
precipitated at the end of a kraft cook by decreasing the pH by adding CO 2,
precipitated lignin was found both in the fibre wall and on the fibre surface (Hartler
1978). In paper I, we found that lignin was precipitated by calcium ions at calcium ion
concentrations at or above the critical coagulation concentration (ccc) which, at pH
11, was found to be 2.5 mmol Ca2+ /L. In Fig. 9, it is evident that lignin is retained in
the pulp by calcium ions at calcium ion concentrations lower than the ccc. There is
also no sharp transition concentration of Ca2+ above which lignin is retained in the
pulp. Instead, the relationship between calcium ion concentration and lignin content in
pulp is almost linear. These facts can possibly be explained by the formation of lignin
aggregates that are retained in the pulp, either by pores inside the fibre wall or by
adsorption to surfaces in/on the fibres, but not by the filter used in the previous study.
Another explanation may be that the negatively charged groups in the fibres lead to an
enrichment of calcium ions in the fibre compared to the surrounding solution
according to the Donnan theory. This could give rise to conditions where the
concentration of calcium ions in the solution contained in the fibre wall exceeds the
ccc(Ca2+), while the calcium ion concentration in the solution surrounding the fibres
is below the ccc(Ca2+).
Fibre swelling
One common explanation of the observed effect of electrolyte concentration on the
washing result is that it depends on the degree of fibre swelling (Favis et al. 1983a;
Hagström-Näsi et al. 1987). This is based on the application of the Donnan theory,
which predicts that the addition of cations reduces the difference in cation
concentration inside and outside the fibre. This reduction leads to a lower osmotic
potential inside the fibres, which results in less swelling of the fibres (Grignon,
Scallan 1980). The decrease in swelling means that the pores in the fibre wall are
smaller, resulting in an increase in resistance to material transport of lignin fragments
out of the wall.
24
1.8
1.6 ¦
„
²
¡
1.4
²
¡
¦
²
¡
¦
WRV
²
¦
„
¡
²
„
¦
¡
mL/g
1.2
1.0
ISEC
¦
0.8
¦
¦
¦
0.6
¦
„
²
¡
0.4
0.2
pH 9
pH 11
pH 12
pH 13
0.0
0
1
2
3
4
5
[Catot], mmol/L
Figure 10. Water contained (WRV) and total pore volume (ISEC) in spruce kraft
fibres washed at different pH values and calcium ion concentrations at 25 °C. WRV
was measured on pulps cooked to the same H-factor, but with different kappa
numbers after washing, due to the presence of different calcium ion concentrations.
The linked ISEC data points are for pulps cooked to different H-factor levels, but with
the same kappa number after washing. All pulps were unbeaten. Cooking conditions
as in Fig. 9.
Fig. 10 shows results of the fibre swelling measured as water retention value (WRV)
and by inverse size exclusion chromatography (ISEC). WRV gives a measure of all
the water contained, which includes bound water, pore water and to some extent
surface water, whereas ISEC gives a measure of the total pore volume in the fibre cell
wall (Berthold 1996). The values obtained by ISEC should therefore be lower than
those given by WRV.
The WRV value decreased from 1.6 at pH 9 to 1.5 mL/g at pH 11 and above when no
Ca2+ had been added to the washing solution ([Ca tot]=0). This decrease in WRV is
probably due to the larger concentration of Na+ present in the pulp after washing at
pH 11 than at pH 9 (Scallan 1983). For all other combinations of pH and Ca2+
concentration, except at pH 13 and 5 mmol/L Ca2+, the WRV values were almost the
same. This shows the negligible effect of Ca2+ concentration between 0 and 5 mmol/L
on the swelling of the fibres in these pulps. The ISEC measurements showed the same
tendency, indicating that differences in swelling of fibres cannot explain the effect of
Ca2+ on washing shown in Fig. 9.
25
Favis et al. (1983a) found that the leaching of lignin from fibres at neutral pH was
markedly reduced by the presence of some metal ions. This was explained by a
reduced swelling of the fibre wall, but no measurements of swelling were made. The
results here indicate that the effect of divalent metal cations on the material transport
of lignin out of the fibre is not due to any change in swelling of the fibre wall. It is
instead believed that the effect is due to a coagulation of the lignin according to the
findings in paper I. This is in accordance with the findings of Brogdon et al. (1996;
1997) in a study of the removal of lignin during an extraction stage after a chlorine
dioxide (D0) stage. They stated that the differences in the amounts of lignin removed
during the extraction stage could not be explained by differences in swelling, but
instead by differences in lignin solubility under different conditions.
It is well known that a higher temperature is favourable for washing (McKibbins
1960; Fogelberg, Fugleberg 1963; Favis et al. 1983b; Li, MacLeod 1993; Vilpponen
et al. 1993). This is also the case when the washing is carried out with Ca2+ ions
present. The relationship between kappa number and Ca2+ concentration at 90 °C was
almost the same as that at 25 °C, but the relationship was moved to a lower kappa
40
„
Kappa number
…
35
25 °C
30
…
„
„
25
…
„
90 °C
„
…
20
0
1
2
3
4
5
[Catot], mmol/L
Figure 11. The effect of the temperature on the kappa number during the washing of
spruce kraft pulps at different calcium ion concentrations at pH 11. Washing at 25 °C
in columns and at 90 °C in a flow-through reactor. Cooking conditions as in Fig. 9.
26
number level at 90 °C, as shown in Fig. 11. The same result was obtained when an
industrial cooked pulp was washed in the laboratory. A difference of about three
kappa number units was obtained between pulps washed at 25 and at 80 °C. This
difference in kappa number when washing at 25 and 80 °C, was the same for washing
in deionized water at pH 11 and for washing in a liquor containing 5 mmol/L Ca2+ at
pH 11.
In paper I, both the critical coagulation concentration (ccc) and the lignin
concentration remaining in solution after coagulation were found to decrease
somewhat with increasing temperature. This should give a higher kappa number after
washing at a higher temperature, but the positive effect of the higher temperature on
washing seems to dominate, and this leads to a lower kappa number after washing at a
higher temperature even when calcium ions are present.
Influence of wood species
Table 4 presents data showing the influence of calcium ion concentration during
washing on the kappa number for spruce and birch kraft pulps. Birch kraft lignin was
precipitated by calcium ions during washing and retained in the birch pulp to a similar
extent as in the spruce pulp. This is in accordance with paper I, where we found that
the concentration of calcium ions at which the lignin starts to coagulate was similar
for birch and spruce kraft lignins.
Table 4. Kappa number after washing of kraft pulps at pH 11 and 25 °C in columns
for 16 hours. Cooking at 22% E.A. and 30% sulfidity. The cooking temperature was
170 °C in the case of spruce wood chips, and 160 ° for birch wood chips.
Wood species
[Ca 2+]tot
Spruce
0
25
Spruce
5
38
Birch
0
23
Birch
5
38
Kappa Number
mmol/L
27
20
¦
ccc(Ca2+), mmol/L
18
16
14
12
10
¦
„
8
¦
6 ¦
4
„
2
0
0
2
4
6
8
10
12
14
16
Lignin concentration, g/L
Figure 12. The ccc of the calcium ion at different spruce kraft lignin concentrations at
pH 11 and 25 °C. (O) 0.3 mol/L Na+, (Q ) 0.01 mol/L Na+ at 0.5 g/L lignin and 0.1
mol/L Na+ at 5 g/L lignin, (----) calculated data according to Eq. 4.
Influence of lignin concentration
Fig. 12 shows that the ccc(Ca2+ ) increased with increasing lignin concentration. The
relation between ccc and lignin concentration was curved at lower lignin
concentrations, but seems to be linear at higher lignin concentrations. The increase in
ccc of Ca2+ when the sodium ion concentration increased from 0.01 mol/L to 0.3
mol/L at 0.5 g/L lignin may be due to competition between Na+ and Ca2+ for the
negatively charged groups in lignin. Even if Ca2+ has stronger affinity for the charged
groups, increasing amounts of sodium ions seem to exchange some of the calcium
ions, thus displacing the ccc to higher concentrations since the sodium ions have less
coagulation ability.
Fitting ccc0 and X in Eq. 4 by a least squares approach from experimental data, gives
the dotted line in Fig. 12. The parameter a in Eq. 4 represents the precipitable part of
the lignin, which was found to be 60-70%, see Table 2. The parameter X is the
calcium content (w-%) in the precipitated lignin. The factor 1000 converts from
mol/L to mmol/L and the factor 40 is the molar mass of calcium in g/mol. The model
seems to give reasonable results, except at low lignin concentrations and high ionic
strength. The fitted value of X was 6.7 w-% which corresponds to the calcium content
28
ccc = ccc 0 + a ⋅ [ Lignin ] ⋅
1000
X
⋅
100% 40
(mmol/L)
(4)
in the precipitated lignin. This is rather close to 6.3 w-%, which was found in the
isolated calcium-precipitated lignin analysed by ICP.
The simple model indicates that the calcium ion concentration needed for coagulation
is the amount of calcium ions needed for saturation of lignin with calcium plus a
certain concentration, which is independent of the lignin concentration. This value,
ccc0 in Eq. 4, was 2.5 mmol/L.
The results of washing a cooked pulp in three different stages are given in Fig. 13.
When Ca2+ was present only in the last stage, the influence of the calcium ions on the
kappa number was small. On the other hand, when Ca 2+ was present in the first
stage(s), the effect was large. This shows that the effect of calcium ions being present
during washing on the kappa number was larger at higher dissolved lignin contents,
which is reasonable since this means that more lignin is available to be precipitated.
This assumes, however, that the amount of calcium ions is sufficiently high to cause
coagulation at the given lignin-to-calcium ratio. As shown above, the ccc(Ca 2+ )
Step1
W
Step2
W
W
W
-
W
-
Ca
Ca
2+
Ca2+
Step3
W
Avj, Avj, Avj
Ca2+
Avj, Avj, Ca2+
Ca2+
-, Avj, Ca2+
2+
Ca
2+
Ca2+
W
-, Ca2+, Avj
W
Ca2+, Ca2+, Avj
Ca2+
Ca2+, Ca2+, Ca2+
20
25
30
35
40
Kappa number
Fig. 13. Washing of a laboratory-cooked spruce kraft pulp in three stages with
deionized water at pH 11 (W) or 5 mmol/L of Ca 2+ at pH 11 (Ca2+). The pulp
suspension was drained and diluted between each stage. Cooking conditions as in
Fig. 9.
29
increases with increasing lignin concentration. At a high lignin concentration, a larger
concentration of calcium ions may be present without causing precipitation, but if the
calcium ion concentration exceeds the ccc at the given lignin concentration, the effect
of precipitation will be considerable. At a low lignin concentration, on the other hand,
lignin will be precipitated at a lower calcium ion concentration, but the effect of the
lignin precipitation on the kappa number will in this case be smaller. The content of
dissolved lignin in the fibre wall, for the sample in row number 4 in Fig. 13, was
calculated to correspond to 1.8 kappa number units, using the WRV-value from Fig.
10 and assuming 20% pulp consistency of the pulp entering the W-stage and that all
dissolved lignin is evenly distributed in the solution. This is in agreement with the
increase in kappa number for the sample in row 4 compared to the pulp washed with
deionized water in all stages, and indicates that most of the dissolved lignin present in
the fibre wall at the beginning of the Ca2+-stage was precipitated. When a similar
calculation was made for the amount of dissolved lignin in the fibre wall initially
present in the first washing stage, a value of almost 90 kappa number units was
obtained. This means that if all dissolved lignin present in the fibre wall would
precipitate, the amount of precipitated lignin would correspond to 90 kappa number
units. This is obviously not the case, and indicates that a calcium ion concentration of
5 mmol/L is insufficient to precipitate the lignin at the start of the washing stage.
30
Influence of precipitation of lignin during pulp washing on
oxygen delignification and pulp bleaching
(Papers II and III)
Oxygen delignification
The results of the treatment of the three kraft pulps in Table 3 in an oxygen
delignification stage are summarized in Fig. 14. The selectivity, that is, the decrease
in viscosity at a given reduction in kappa number, was similar for the 1500-W and
1500-Ca samples, while the selectivity for the 1100-W sample was lower. A pulp
cooked and washed with calcium ions present, will yield a pulp after the oxygen stage
with either a lower viscosity or a higher lignin content than a pulp cooked in the same
way but washed in the absence of calcium ions. A higher lignin content in the
oxygen-delignified pulp will result in more lignin to be removed in the non-closed
Pulp viscosity, mL/g
part of the mill with, for example, an increase in COD-emissions as a result.
1200
¦
1100-W
1100
¦
1500-W
1000
„
900
„
¦
¡
¦
¦¡
„
1500-Ca
¡
„¦
¡
¡
¡
800
5
10
15
20
25
30
35
Kappa number
Figure 14. The change in viscosity during oxygen delignification of spruce kraft pulps
at 100 °C. The kraft pulps were washed after the cook with or without calcium ions
present. The 1500-Ca sample was washed with a calcium ion concentration of 5
mmol/L (Ca), pH 11 at room temperature. The 1500-W and 1100-W samples were
washed in deionized water (W) at room temperature. The 1500-W and 1500-Ca
samples were cooked to the same H-factor (1500); the difference in kappa number is
due to precipitation of lignin by calcium ions during washing. The point with the
highest kappa number in each series is for the cooked pulp, while the other points
represent the oxygen-delignified pulps. The time and the alkali charge in the oxygen
stage were varied in order to obtain pulps with different kappa numbers.
31
One significant difference between the 1100-W and 1500-Ca pulps is that the
viscosity of the pulp with precipitated lignin did not fall off rapidly at high degrees of
delignification, which is the normal case in an oxygen delignification stage (Dence,
Reeve 1996 p. 225). This may perhaps be due to the difference in chemical
composition of the lignin between the two samples, since the precipitated lignin
contains more phenolic groups than the other part of the residual lignin after the cook
(see below) and would thus offer more moieties in the lignin that are rather reactive in
an oxygen stage. This would extend the rapid initial delignification phase and
postpone the slower delignification phase in the oxygen stage. Another explanation
may be that there is a difference in location of the precipitated lignin in the fibre wall,
which could render the precipitated lignin more accessible to oxygen and facilitate
material transport of the dissolved lignin.
The pulp viscosity value of the 1500-Ca sample after the cook is somewhat lower
than that of 1500-W, see Fig. 14. This is unexpected, since the only difference in
treatment of the samples is differences in the washing liquor composition. However,
if the viscosity values are corrected for the lignin content in the samples, the
difference between the two samples is almost insignificant. The reductions in
viscosity of the 1500-W and 1500-Ca samples during the oxygen stage were similar,
while the reduction in the case of the 1100-W sample was somewhat larger, see Fig.
14. Thus, the precipitated lignin seems not to influence the degradation of cellulose
during the oxygen delignification stage. The greater reduction in viscosity of the
1100-W sample can, to some extent, be explained by the higher initial viscosity value
since this results in a larger reduction in viscosity for each chain scission of the
cellulose.
In Fig. 15, the kappa number after oxygen delignification is shown as a function of
the charge of sodium hydroxide in the oxygen stage. The amount of lignin removed
per unit amount of sodium hydroxide charged was initially greater for the two pulps
with a higher lignin content before the oxygen delignification, but the amounts were
similar for the three pulps after a certain degree of delignification.
32
35
Kappa number
30
25
¦
¡
1100-W
„
¦
20
1500-W
15
„
¡
¦
¡
„
10
¦
¡
„
¡
¦
¡
1500-Ca
¦
5
0
1
2
3
4
NaOH-charge, %
Figure 15. The kappa number after oxygen delignification as a function of charged
amount of sodium hydroxide in the oxygen delignification stage. See Fig. 14 for
explanation of sample designations.
During oxygen delignification, the contents of phenolic groups in both the residual
lignin and the dissolved lignin decrease, while the content of carboxylic groups in the
dissolved lignin increases (Gellerstedt, Lindfors 1987). However, the content of
phenolic groups is higher in the dissolved oxygen lignin than in the residual lignin
after the cook. Johansson and Ljunggren (1994) found that the reactivity of phenolic
compounds is considerably higher than for corresponding non-phenolic structures.
This indicates that a high content of phenolic groups in the residual lignin should give
a high reactivity in the oxygen stage. The contents of phenolic hydroxyl groups in
lignin in black liquor and in the remaining lignin in black liquor after precipitation of
lignin by calcium ions were found to be respectively 56 and 78 hydroxyl groups per
100 C-9 units (paper I). By difference calculation, taking into consideration the lignin
concentrations in the different solutions, an estimate of the content of phenolic
hydroxyls in the precipitated lignin is 50 hydroxyl groups/100 C-9 units. This is about
twice the content reported in residual lignin after a kraft cook by Gellerstedt and
Lindfors (Gellerstedt, Lindfors 1984a; Gellerstedt, Lindfors 1984b), who found 20-30
phenolic hydroxyl groups /100 C-9 units in the residual lignin in pulps with kappa
numbers in the 30-35 range. Thus, a greater reactivity with a faster degradation and a
more rapid dissolution of the precipitated lignin would be expected. The results in
33
35
Ca2+
k, m2/kg
30
|
|
33
|
¦
38
25
¦
20
Deionized water
¦
¦
43
15
Calc. brightness, %
28
40
48
10
10
20
30
40
50
60
Kappa number
Fig. 16. The relationship between the light absorption coefficient at 460 nm, k, and
kappa number for spruce kraft pulps washed with and without calcium ions present at
pH 11 and 25 °C. The brightness is calculated from k assuming a light scattering
coefficient of 50 m2/kg. Points connected with a dashed line are for pulps cooked to
the same H-factor. Cooking conditions as in Fig. 9.
Fig. 15 show a somewhat increased reactivity. One factor which may counteract the
expected increase in reactivity of the precipitated lignin is the calcium ions in the
precipitated lignin, provided that the calcium ions bound to the phenolic groups
decrease the electron density in the aromatic ring. Sultanov and Wallis (1991)
reported that an increase in the electron density in the aromatic ring increased the rate
of the cleavage of the aromatic ring by reactions with oxygen. Another factor may be
steric effects caused by the calcium ions bound to the lignin.
Chromofore content in kraft and oxygen-delignified pulps
It is evident in Fig. 16 that the lignin precipitated in the pulp during washing with
calcium ions present was darker than the rest of the residual lignin. The light
absorption coefficient of the precipitated lignin was approximately 1.8 times higher
per unit lignin content. This indicates a larger amount of chromophoric groups in the
lignin precipitated by Ca2+ , than in the other part of the residual lignin in the pulp.
Lignin dissolved during the cook was found to be darker than lignin remaining
undissolved in the fibres (Jansson, Palenius 1972). Pulps produced in flow-through
reactors are brighter than pulps produced in batch reactors (Jansson et al. 1975). This
34
30
¡
k(457 nm), m2/kg
1500-Ca
25
„
20
¦
¡
15
¡
¦
1500-W
1100-W
¡ „
¦
¦
¡„
¦
¡„
¦
10
5
5
10
15
20
25
30
35
Kappa number
Figure 17. Change in light absorption coefficient measured at an effective wavelength
of 457 nm during oxygen delignification. See Fig. 14 for an explanation of the sample
designations. The point with the highest kappa number in each series is for the cooked
pulp, while the other points represent the oxygen-delignified pulps.
difference may be explained by the lower concentration of dissolved lignin in the
flow-through system, which leads to less resorption or precipitation of dissolved
lignin on the fibres at the end of the cook. This may indicate that the dark pulps
obtained by washing in the presence of calcium ions are due to the dark colour of the
dissolved lignin that is coagulated and retained in the pulp.
A low residual alkali concentration in the cook is known to give pulps with low
brightness (Aurell, Hartler 1965). Recently, Gustavsson et al. (1999) found that a
lower residual alkali concentration in the cook gave a higher calcium content in the
pulp. They also obtained a higher light absorption coefficient for pulps with a high
calcium content. This may indicate that dissolved lignin is coagulated by calcium ions
on fibres also during the cook.
The increase in brightness during oxygen delignification is known to mainly be due to
delignification reactions and not to brightening reactions (Gustavsson, Swan 1974).
Fig. 17 shows the decrease in light absorption coefficient, k, measured at 457 nm at
different degrees of delignification during an oxygen stage. The relative difference in
k between the two pulps with the same cook kappa number (1500-Ca and 1100-W)
was the same for all degrees of delignification. The light absorption coefficient of the
35
k(457 nm)/Kappa number
1.1
¡
¡„
¡„
1500-Ca
¡
¡
1
„
0.9
¡
„ 1500-W
¦
0.8
¦
¦
¦
1100-W
¦
¦
0.7
0.6
5
10
15
20
25
30
35
Kappa number
Figure 18. The ratio of light absorption coefficient determined at an effective
wavelength of 457 nm to kappa number. See Fig. 14 for explanation of sample
designations. The point with the highest kappa number in each series is for the cooked
pulp, while the other points represent the oxygen-delignified pulps.
pulp with the lower cook kappa number (1500-W) did not decrease to the same
extent, and it approached the value for the 1500-Ca pulps at higher degrees of
delignification.
The ratio of light absorption coefficient to kappa number provides a measure of the
chromofore content in the residual lignin in the pulp. This assumes that the
contributions from carbohydrates and extractives to the k-value are constant and
small. This is the case for cooked softwood pulps (Hartler, Norrström 1969). In Fig.
18, this ratio (where the light absorption coefficient was determined at 457 nm) is
plotted at different degrees of delignification during an oxygen stage.
After the cook, the calcium-precipitated lignin was darker than the lignin in the pulp
cooked to the same kappa number (1100-W). However, the k/kappa number ratios
were similar for the pulp with precipitated lignin (1500-Ca) and for the pulp cooked
to the same H-factor (1500-W).
The ratio increased for the oxygen-treated pulps compared with the respective cooked
pulps at all degrees of delignification during the oxygen stage. The increase in the
ratio implies that the residual lignin is darker after the oxygen stage than after the
cook. This was also reported by Gustavsson et al. (1999) and by Sjöström (1999).
36
k/Kappa number
10
Kraft pulp; K no. 23
1.5% NaOH; K no. 16
2.3% NaOH; K no. 11
2.7% NaOH; K no. 10
1
0.1
0.01
400
450
500
550
600
650
700
Wavelength, nm
Figure 19. The ratio of light absorption coefficient to kappa number at different
wavelengths for the pulp 1500-W after the cook and after different degrees of
delignification in the oxygen stage. The NaOH charge in the oxygen stage and the
kappa number after the oxygen stage are given in the legend.
When optical properties are studied, an attempt is made to obtain a measure that
relates to what a human eye would perceive. The CIE Y-function is constructed to
approximate to the response of the human eye to light. This function covers the whole
visible wavelength interval, with an effective wavelength of 557 nm. The value of the
weighting function at 560 nm is about 6.6 times the value at 460 nm. The value of the
information provided by the light absorption coefficient at 457 nm (the same
wavelength at which pulp brightness is determined) is therefore rather limited with
respect to what the human eye perceives. Fig. 19 shows the ratio of the light
absorption coefficient over the visible wavelength range to the kappa number at
different degrees of delignification. At wavelengths below approximately 470 nm, the
k/kappa number ratio increased with increasing degree of delignification during the
oxygen stage, in accordance with Fig. 18. However, at wavelengths above 470 nm the
result was the opposite. At these wavelengths, the ratio decreased, which means that
the chromofore content in the residual lignin decreased during oxygen delignification.
Fig. 20 shows the ratio of k to kappa number during oxygen delignification. Here, the
light absorption coefficient is determined at 560 nm, close to the effective wavelength
of the Y-function, instead of at 457 nm as in Fig. 18. The residual lignin in the three
37
k(560 nm)/Kappa number
0.4
1500-W
0.3
¡
¡
„
¡
„
¦
¡
¡ „
„
0.2
¦
¦
1500-Ca
¡
1100-W
¦
¦
¦
0.1
5
10
15
20
25
30
35
Kappa number
Figure 20. The ratio of light absorption coefficient determined at 560 nm to kappa
number. See Fig. 14 for explanation of sample designations. The point with the
highest kappa number in each series is for the cooked pulp, while the other points
represent the oxygen-delignified pulps.
pulps became somewhat brighter during the oxygen stage. The pattern of the k/kappa
number ratios for the three pulps after the cook, that is, the three points farthest to the
right in the diagrams, was similar at 560 nm and at 457 nm.
Since the variation in the values of the k/kappa number ratio in Figs. 18 and 20 is
small, an error in the value of the kappa number may have a significant influence on
the ratio. The kappa numbers in Figs. 18 and 20 were not corrected for hexenuronic
acid content in the pulp. An estimate of the content of hexenuronic acid was obtained
using the model of Gustavsson and Wafa Al-Dajani (2000). This model indicated that
the content of hexenuronic acid corresponded to 0.9 kappa number units in the pulps
in Figs. 18 and 20. This leads to a somewhat larger increase in the k/kappa number
ratio in Fig. 18 and a somewhat smaller decrease in Fig. 20 with decreasing kappa
number than that shown.
38
Table 5. Pulps used for bleaching in the OQ(OP)Q*P sequence. The hydrogen
peroxide charge in the (OP) stage was 0.25 times the kappa number after the oxygen
stage. Regarding sample designation: 1100 and 1500 is the H-factor at which the
pulps were cooked. W stands for washing with deionized water after the cook and Ca
for washing with 5 mmol/L Ca2+ at pH 11. L and H stand for a Low (about 50%) and
a High (about 60%) degree of delignification during the oxygen stage.
After O
Sample
Kappa number
Viscosity
After OQ(OP)Q*
k(560 nm)
Kappa number
2
mL/g
m /kg
Viscosity
k(560 nm)
mL/g
m /kg
2
1500-WL
12.7
960
3.01
5.0
910
0.10
1100-WL
16.6
1000
3.41
5.9
860
0.07
1100-WH
11.9
950
1.90
5.3
860
0.08
1500-CaL
16.3
940
4.64
4.6
760
0.08
1500-CaH
12.2
890
2.92
4.4
815
0.07
Final bleaching
From the three cooked pulps in Table 3, five oxygen-delignified pulps were produced
for further bleaching in a Q(OP)Q*P sequence. The cooked 1500-Ca and 1100-W
pulps were delignified in the oxygen stage to two degrees of delignification, about
50% and about 60%. The 1500-W pulp was delignified to only about 50% degree of
delignification in the oxygen stage. This resulted in three pulps with kappa numbers
of about 12 (1100-WH, 1500-CaH, 1500-WL) and two with kappa numbers of about
16 (1100-WL, 1500-CaL). The data for these pulps are summarized in Table 5.
Fig. 21 shows the bleachability of the five different pulps, as brightness versus
consumed amount of hydrogen peroxide per kappa number. The OQ(OP)Q*P pulps
fell into two groups, with the pulps with precipitated lignin and having a higher
brightness in one group, and the other three pulps in another group. The calciumprecipitated lignin was easier to bleach than the other residual lignin in bleaching
stages containing hydrogen peroxide. Sjöström and co-workers found that lignin
precipitated in the cook at a low residual alkali level (Gustavsson et al. 1999) or at a
high sodium ion concentration (Sjöström 1999) resulted in a dark lignin that was
39
Brightness, %
90
‘
88
‘
86
¦|
| „
„
84
‘
¡
¡
¡
|
¦
¡
‘
¦
|
„
82
|
78
0.10
¦
¡
‘
¦
80
„
„
0.15
0.20
0.25
1500-CaL
1500-CaH
1100-WL
1100-WH
1500-WL
0.30
0.35
Consumed H2O2/Kappa number
Figure 21. The ISO-Brightness as a function of consumed hydrogen peroxide (%) per
kappa number for OQ(OP)Q* and OQ(OP)Q*P bleached pulps. The ratio of
consumed H2O2/kappa number is calculated as the sum of the consumption of
hydrogen peroxide in the OP and the P-stage divided by the kappa number after the
O-stage. The point with the lowest brightness value in each series is for the
OQ(OP)Q* pulp and the other points represent the OQ(OP)Q*P pulps. The letter L,
after the name of the cooked pulp sample, corresponds to the low degree of
delignification (about 50%) during the oxygen stage and the letter H corresponds to
the high degree of delignification (about 60%) during the oxygen stage. See also Fig.
14 for an explanation of sample designations.
difficult to bleach. The precipitation of lignin by calcium ions also yielded a dark
lignin, but showed a higher, instead of a lower, bleachability.
Brightness versus pulp viscosity is plotted in Fig. 22. The pulps were grouped in a
similar way to that in Fig. 21. Compared at a given brightness, the two pulps with
precipitated lignin obtained a 50-100 mL/g lower viscosity than the other three pulps.
For the three pulps delignified to a kappa number of about 12 after the oxygen stage,
the differences in viscosity between the calcium-washed pulp and the other two pulps
(1500-WL, 1100-WH) were about the same after final bleaching as after the oxygen
stage. However, in the case of the two pulps delignified to a kappa number of about
16, the difference was almost twice as large after final bleaching. The pulps washed in
the presence of Ca2+ after the cook and containing precipitated lignin seem to be more
reactive in bleaching stages with hydrogen peroxide, since not only the decrease
40
90
¡
¡
¡
Brightness, %
88
‘
‘
‘
¦
„
|
¦
|
¦
„
|
86
84
¡
¡
‘
¦
|
„
„
‘
1500-CaL
1500-CaH
1100-WL
1100-WH
1500-WL
¦
82
|
80
78
650
„
700
750
800
850
900
950
Pulp viscosity, mL/g
Figure 22. The ISO-Brightness as a function of pulp viscosity for OQ(OP)Q*- and
OQ(OP)Q*P-bleached pulps. The point with the lowest brightness value in each
series is for the OQ(OP)Q* pulp and the other points represent the OQ(OP)Q*P
pulps. See Figs. 14 and 21 for explanation of sample designations.
in kappa number and viscosity but also the increase in brightness were greater for
these pulps. This may possibly be due to the higher content of phenolic groups in the
precipitated lignin than in the other residual lignin, causing an increase of radical
reactions since the phenolate ion is a precursor of organic radicals in the pulp (Gierer
1982). The extended delignification in the oxygen stage for 1500-CaH seems to have
reduced the negative effect of hydrogen peroxide on the viscosity of pulps with
precipitated lignin. This is probably due to a lower content of precipitated lignin in the
pulp when it enters bleaching stages with hydrogen peroxide.
41
Studies on the interaction between kraft lignin and calcium
ions using an ion-selective electrode
(Papers IV and V)
Determination of end points
When a substance is titrated with a component that binds strongly to it, as in the
titration of calcium ions with EDTA at high pH, a characteristic jump can be seen at
the end point in the titration curve E=f(v) where v is the added volume of reagent. The
jump is however less pronounced the smaller the equilibrium constant is, and it is
completely invisible for reactions with an equilibrium constant, K, below 104
(Ringbom 1963). This is the case for the reaction of calcium ions with the lignin
compound catechol (Westervelt et al. 1982). For such weak reactions, traditional
techniques like differentiation or Gran-plots (Gran 1988) cannot be used. For that
reason, we developed a straightforward method to detect the end points of reactions
with equilibrium constants smaller than 104. The strategy is based on the assumption
that, if no reaction between Ca2+ and lignin can be detected, a plot of electrode
voltage against log[Ca2+ ] should give a straight line, that is, a calibration curve of
Electrode voltage, mV
100
y = 33x + 150
80
A
60
40
20
0
-20
-5
-4
-3
-2
-1
2+
log([Ca ]added)
Figure 23. Electrode voltage of a calcium ion-selective electrode as a function of
concentration of added calcium ions to a 0.5 g/L kraft lignin (Indulin AT) solution at
pH 11 (25 °C).
42
Electrode volt. residuals, mV
6
4
2
0
-2
-4
-6
-8
-5
-4
-3
-2
-1
2+
log([Ca ]added)
Figure 24. Residuals after linear regression of the data in Fig. 23.
Ca2+ in a lignin solution. As can be seen in Fig. 23, this is not the case. In fact, the
shape of the curve shows similarities to a titration curve for strong complexing agents,
although the end point can still not be detected by differentiation. A representative
plot of the residuals from the linear regression analysis of electrode voltage as a
function of log[Ca2+ ] is shown in Fig. 24. It is assumed that the left-hand part of the sshaped (sigmoidal) curve may be a sort of titration curve with the position of a
reaction-characteristic point exactly between the two curvatures.
3
|
[Ca2+]added, mmol/L
y = 0.51x + 0.04
R2 = 0.99
2
High Mw
|
1
|
|
Low Mw
||
|
|
|
|| |
0
0
1
2
3
4
5
Lignin concentration, g/L
Figure 25. Added calcium ion concentration at the end point for a calcium ionselective electrode with a membrane from Metrohm. Kraft lignin (Indulin AT)
solutions at pH 11 (25 °C).
43
Fig. 25 shows results from the detection of end points by a calcium ion-selective
electrode. The total Ca2+ concentration at which the end point is detected increases in
proportion to the lignin concentration in the studied region. This supports the
conclusion that the end point is a result of Ca2+ being bound to lignin molecules.
The data for both the high and low molecular mass fractions of the lignin that were
obtained by dialysis are consistent with the data for the rest of the samples in Fig. 25.
This appears to indicate that the molecular mass of the lignin does not influence the
end point, although this is somewhat surprising, since it is known (Mörck et al. 1986)
that low molecular mass lignin has a higher content of charged groups than the high
molecular mass fraction.
Fig. 26 shows the dependence of the critical coagulation concentration (ccc) on lignin
concentration. The total added concentration of calcium ions at which the
precipitation began was determined by measuring the light-absorbance of the liquor at
750 nm during the titration. The detected onset of precipitation was dependent on the
time between the additions of calcium ions. This indicates that the precipitation
reaction was rather slow at 25 °C, since a change from 2 to 5 minutes between Ca2+
additions influenced the result.
9
¡
7
ccc, mmol/L
‘
¡
Delay time 5 min
8
¡
6
¡
5
¡
Delay time 2 min
¡
‘
‘
‘
4
‘
3
2
1
0
0
1
2
3
4
5
Lignin concentration, g/L
Figure 26. The critical coagulation concentration (ccc) of calcium ions for different
kraft lignin (Indulin AT) concentrations at pH 11 (25 °C) and different times between
calcium ion additions during titration. The ccc was determined by turbidity
measurements at 750 nm.
44
A comparison of the results in Figs. 25 and 26 at a given lignin concentration shows
that the calcium ion concentration at the end point was always considerably lower
than the calcium ion concentration at the ccc. This shows that the complexation
reaction between lignin and Ca2+ and the coagulation reaction are two separate
processes. However, the coagulation is probably not independent of the complexation
reaction. At the point of coagulation, as detected by the turbidity, the signals from the
electrodes did not indicate any change in calcium ion concentration.
In a few experiments, the signals were measured a long time after the final addition of
calcium ions. The change in response of the electrode was small enough to be
considered as drift, while the turbidity increased significantly after the last addition.
This implies that the precipitation or ageing of the precipitate is much slower than the
uptake of calcium ions by lignin.
Determination of free calcium ion concentration during titrations
When the free calcium ion concentration was detected with calcium ion-selective
electrodes in lignin solutions during titration, several problems were encountered.
Large differences in the slope of the electrode voltage versus log[Ca2+ ] as well as in
the start potential between electrodes with different membrane types were obtained.
0
-1
-2
LogC
-3
Ca-L2-z
-4
-5
-6
simulated
calculated
Ca2+
-7
-8
-6
-5
-4
-3
-2
-1
0
2+
Log[Ca ]tot
Figure 27. Result on simulated dataset; a titration of a ligand Lz- with Ca2+ at pH 11
(25 ° C) with Lz- = 2 mmol/L when the internal calibration method is applied.
45
This indicates that the interaction between the lignin molecules and the membrane
must be taken into consideration if the value of the free calcium ion concentration is
to be determined. Thus, it seems difficult to determine the absolute free calcium ion
concentration with conventional methods.
To handle the problems associated with the determination of the free calcium ion
concentration, a new algorithm for calibration and for the estimation of the free
concentration of the same sample was developed. This method is referred to in the
following as the “internal calibration method”. In this method, we used the data from
the final part of the titration curve, that is, at a significant distance above the end
point. We were thus able to use data from the titration of a sample to construct a
calibration equation, so that the electrode had the same environment in the calibration
and measurement situations. In this way, it was also possible to correct for matrix
influences.
We made the assumption that each new addition of calcium ions results in a
corresponding increase in the free calcium ion concentration. That is, it was assumed
that there are no longer any free lignin molecules left to bind calcium ions. In order to
satisfy this assumption, the total added concentration of calcium ions has to be above
the end point. If the assumption is true, a plot of the free concentration (Cfree) against
the total concentration (Ctot) yields a straight line with a slope of unity. A simplified
version of the relationship between electrode potential (E) and free ion concentration
is given in Eq. 5.
E = m + k ⋅ log(Cfree )
(5)
This equation may be rewritten as:
C free = 10
E− m
k
(6)
The parameters m and k were adjusted by a least squares approach to yield a straight
line with a slope of 1 when Cfree was plotted against Ctot. Initial values of m and k in
46
the iteration were obtained by linear regression of the final part of the titration curve,
where E=f(logCtot) was plotted. The expression:

∆C free (i ) 
∑ 1 − ∆C 
tot(i )
2
(7)
was minimized numerically using the quasi-Newton methodology used by the
”Problem Solver” in Microsoft Excel 5.0. The calculations were repeated for different
numbers of data points in the final part of the titration curve.
When the free concentration has been estimated, the bound concentration can be
calculated as the difference between the total added concentration and the free
concentration.
In order to evaluate the method, simulation calculations were carried out in the
equilibrium calculation program Mineql. The complex equilibrium reaction between
Ca2+ and a ligand Lz- was used:
Ca
2+
+ Lz− ↔ CaL2− z
(8)
4
The equilibrium constant was set to 10 to obtain a system similar to the lignincalcium system. The total concentration of the ligand L was set to 0.002 mol/L and
the pH to 11. This solution was titrated with Ca2+ , and the titration curve obtained can
be seen in Fig. 27. A simulated electrode potential was calculated according to:
2+
Esim = 30 ⋅ log[Ca ]
(9)
Using this potential and applying the method described above, the concentrations of
Ca2+ and CaL2-z were calculated. The agreement between the original data and the
data calculated by the internal calibration method is evident in Fig. 27.
Fig. 28 shows estimations of the free calcium ion concentration by the internal
calibration method in a lignin solution, compared with estimations using the slopes
47
[Ca2+]calc.free, mol/L
0.02
0.015
0.01
EDT1 NaOH
Metrohm NaOH
EDT1 Internal
Metrohm Internal
0.005
0
0
0.005
0.01
0.015
0.02
[Ca2+]added, mol/L
Figure 28. Titration of a kraft lignin (Indulin AT) solution at pH 11 (25 °C).
Estimated free calcium ion concentrations for two different electrodes using the
internal calibration method or calibration curves from lignin free NaOH at pH 11.
and intercept from pure calibration solutions, that is, solutions of NaOH at pH 11. The
difference between the results of the two methods is smaller with the Metrohm
membrane than with the EDT membrane. In the figure, a line with a slope of one and
zero intercept is drawn, which corresponds to the assumption that all added calcium
ions form free calcium ions in the solution. The NaOH-calibration curve gave wrong
values for both the Metrohm and the EDT electrodes, and resulted in a higher free
calcium ion concentration than the total added concentration, which shows that it
impossible to use conventional calibration curves to determine the free calcium ion
concentration in these solutions. However, the internal calibration method gave
reasonable results; the free calcium ion concentration was slightly lower then the total
added concentration. This gives us a reason to believe that the iteration procedure
reduces the influence of different species on the electrode performance.
In Fig. 29, the logarithms of the free and bound concentrations for a sample are
plotted against the logarithm of the total Ca2+ concentration. When the slope of the
curve for the bound concentration levels out, the interpretation is that all calcium sites
in the lignin are occupied and that all calcium ions added then remain as free calcium
48
-1
ccc
-2
logC
bound
-3
-4
free
-5
-6
-6
-5
-4
-3
-2
-1
log[Ca2+]added
Figure 29. Estimated free and bound calcium ion concentration during titration of the
high molecular mass fraction of kraft lignin (Indulin AT) at pH 11 (25 °C) as a
function of added calcium ion concentration. The critical coagulation concentration
(ccc) determined from turbidity measurements is marked.
ions. At calcium ion concentrations below the end point almost all the added Ca2+ was
bound to the lignin.
The ccc, that is, the start of precipitation/coagulation of lignin, is also given in Fig.
29. The ccc was found to be higher than the concentration of calcium ions at which
the concentration of bound calcium ions had levelled off and reached, or almost
reached, its maximum value. This implies that the lignin is saturated with calcium
ions before or at the point where the lignin starts to precipitate. A probable
interpretation is that a number of the charged groups in the lignin need to be blocked
before precipitation can occur. This is in accordance with the interpretation of the
model presented earlier (Eq. 4) regarding the amount calcium ions needed for
coagulation at different lignin concentrations. Fig. 29 also shows that the lignin does
not bind any more calcium ions after the ccc has been reached.
The amount of calcium ions bound to lignin can be calculated from Fig. 30. In a 1 g/L
solution of lignin, the amount of bound calcium is 6.5-7.3% depending on the
membrane being used for the estimation. The lignin concentration used was the same
as that used in the preparation of isolated calcium-precipitated lignin. The calcium
49
[Ca2+]calc. bound, mmol/L
3
y = 1.39x + 0.21
R2 = 0.97
2
1
0
0
0.5
1
1.5
2
2.5
Lignin concentration, g/L
Figure 30. Estimated bound calcium ion concentration as a function of initial kraft
lignin (Indulin AT) concentration for three calcium ion-selective electrodes at pH 11
and 25 °C.
content in this lignin was determined by ICP. The value determined by the electrodes
can be compared with the value of 6.3 %, which was the amount found in the
precipitated lignin analysed by ICP. This value also includes some uncertainty due to
the sample preparation. Moreover, the calcium content estimated by the model given
in Eq. 4, which was 6.7%, is also in agreement with previous methods. Since the three
methods are distinctly different, the agreement between the results of the methods is
surprisingly good.
Recalculated, the amount of bound calcium found by the internal calibration method
is 30-32 Ca/100 C-9 units assuming a molecular mass of 182 g/mol for a C-9 unit in
the lignin. The corresponding value for the ”ICP-method” is 29 Ca/100 C-9 units. The
content of carboxylic acid groups in lignin from industrial softwood kraft black
liquors is reported to be in the range of 15-20 per 100 C-9 units (Marton 1964; Mörck
et al. 1986; Månsson, Öster 1988; Gellerstedt, Heuts 1997), and the corresponding
values for phenolic groups are 55-80 per 100 C-9 units (Marton 1964; Månsson 1983;
Robert et al. 1984; Mörck et al. 1986). Consequently, under the conditions
investigated here, all charged groups in the lignin do not bind calcium ions and this
may indicate that carboxylic groups and some of the phenolic groups that are
deprotonated at pH 11 (25 °C) bind the calcium ions.
50
Concentration of metal ions in industrial brownstock washing
and its influence on lignin precipitation
(Paper VI)
Pulp samples were collected from the blow-lines in kraft pulp mills with an oxygen
stage producing bleachable grades of softwood pulp (Mills A-C) and in the blow-line
in a kraft pulp mill producing unbleached softwood pulp of sackpaper grade (Mill D).
In Table 6, the effect of washing these industrially cooked softwood pulps with and
without calcium ions is shown. The differences in kappa number between the pulps
washed with and without calcium ions added to the wash liquor were four to six
kappa number units for the low kappa pulp (Mills A-C) and ten kappa number units
for the high kappa pulp (Mill D). This shows that lignin may precipitate in industrially
cooked pulps during washing.
The pulp in Mill C was washed with the mill’s own fresh water which resulted in a
similar difference in kappa number as that between the pulps from Mill A and Mill B
washed with and without calcium ions present. The concentration of calcium and
magnesium in the fresh water was 0.7 mmol/L.
The difference in kappa number between the industrially cooked pulps washed with
and without calcium ions present was considerably less than the difference obtained
Table 6. Kappa number for mill-cooked softwood pulps washed in the laboratory.
Sample
Kappa number
Kappa number
[Ca 2+]tot ≈0 mmol/L c
[Ca 2+]tot =5 mmol/Ld
Mill Aa
22.4
28.2
Mill Ba
18.0
22.2
Mill C
a
19.7
24.2e
Mill D
b
54.0
63.7
a. A mill producing bleached softwood kraft pulp.
b. A mill producing softwood pulp for unbleached kraft paper products.
c. Washing carried out at 80 °C, pH 11 and deionized water.
d. Washing carried out at 80 °C, pH 11 and 5 mmol/L Ca2+ added to the wash liquor.
e. Washing carried out at 80 °C, pH 11 with the mill’s own fresh water.
51
with laboratory-cooked pulps (Fig. 9). One explanation of this difference may be that
there is a higher sodium ion concentration in the mill system, giving rise to a
competition between sodium and calcium ions.
Another reason for the difference may be found in the experimental procedure. In the
case of mill-cooked pulp, a larger amount of black liquor followed the pulp to
washing. According to paper I, the free calcium ion concentration is lowered by the
consumption of calcium ions by lignin, before the lignin is coagulated. This decrease
in calcium ion concentration should lead to a smaller effect on kappa number, since it
would have the same effect as moving to the left in Fig. 9. Both different contents of
charged groups in the lignin and differences in molecular mass distribution of the
lignin may also cause different results.
Figs. 31 and 32 shows the metal concentration in liquor from five different positions
in a fibre line. The procedure used to determine the metal concentration yields a value
of the total concentration of the element in the liquor, including solid particles. The
filtration had a significantly decreasing effect on the calcium and magnesium
concentrations, probably due to the removal of solid particles of metal carbonates and
hydroxides. The difference between the filtered and unfiltered samples declined
2.0
Un-filtered
1.0
Filtered
[Ca]tot, mmol/L
1.5
0.5
0.0
End
digester
Blow
Blowtank
Wash
before O
Wash
after O
Figure 31. The calcium concentration in liquor from five different positions along a
fibre line (Mill A). A 0.45 µm filter was used for filtration. The filtration was
performed on warm samples shortly after the sampling.
52
12
Un-filtered
8
6
Filtered
[Mg]tot, mmol/L
10
4
2
0
End
digester
Blow
Blowtank
Wash
before O
Wash
after O
Figure 32. The magnesium concentration in liquor from five different positions along
a fibre line (Mill A). A 0.45 µm filter was used for filtration. The filtration was
performed on warm samples shortly after the sampling.
downstream in the fibre line. This is probably due to the decreasing pH, since the
formation of solid carbonates and hydroxides decreases with decreasing pH. The pH
was above 13 in the end of the digester and close to 11 before the oxygen stage. The
irregularities in total concentration in Figs. 31 and 32 may be due to differences in the
distribution of particles between the fibre and the liquor phase.
The large effect of filtration on the concentration of certain elements shows the
importance of sample preparation in combination with the analysis technique.
Moreover, the filtration of liquor samples before the determination of element
concentration by ICP does not ensure that the determination yields the concentration
of the metal ion, that is, the free concentration, since particles smaller than 0.45 µm
are probably present. In order to achieve this, other analysis techniques have to be
employed.
The concentration of metal ions in pulp mill liquors varies between different mills due
to a number of factors, including wood species, location of tree growth and pulp mill
operations. The concentration of metals from three different pulp mills are shown in
Figs. 33 and 34. The calcium concentration of the filtered samples was in the range of
53
1.2
|
[Ca]tot, mmol/L
1.0
0.8
|
²
0.6
Mill C
²
Mill B
0.4
0.2 ‘
¡
|
¡
‘
‘
|
¡
Blow
Blowtank
‘
¡
|
²
Mill A
¡
‘
0.0
End
digester
Wash
before O
Wash
after O
Figure 33. The calcium concentration in liquor from five different positions along a
fibre line in three different kraft pulp mills producing bleachable grades of softwood
pulp. For Mill A, two series of samples with a difference in sampling time of about 3
hours are shown. A 0.45 µm filter was used for filtration before the calcium
concentration was determined by ICP.
9
‘
¡
8
[Mg]tot, mmol/L
7
6
5
4
3
‘
2|
1
¡
0
End
digester
‘
Mill A
‘
¡
¡
|
²
|
|
|
²
²
Blow
Blowtank
Mill B
Mill C
Wash
before O
¡
‘
Wash
after O
Figure 34. The magnesium concentration in liquor from five different positions along
a fibre line in three different kraft pulp mills producing bleachable grades of softwood
pulp. For Mill A, two series of samples with a difference in sampling time of about 3
hours are shown. A 0.45 µm filter was used for filtration before the magnesium
concentration was determined by ICP.
0.2-0.8 mmol/L, except for the samples after the oxygen stage. In mill A, the calcium
concentration was fairly constant along the fibre line. In mill B and C, a trend towards
higher values close to the digester may be seen. In the case of magnesium the
54
opposite trend was observed, Fig. 34. The concentration of magnesium increased
from the end of the digester to the oxygen stage. The concentration of magnesium was
considerably higher than that of calcium, with up to 8 mmol/L magnesium in the
liquor in the washing before the oxygen stage. The cause of the high magnesium
concentration around the oxygen stage is the addition of magnesium ions in the
oxygen stage. A probable explanation of the difference in trend of the magnesium
concentration before and after the oxygen stage in mills A and B is the different
sampling positions. In mill A, the filtrate was collected from the last washing stage
after the oxygen stage, whereas the filtrate from the first washing stage was collected
in mill B. The decrease in magnesium concentration in filtered samples from the
oxygen stage to the digester is probably due to the formation of magnesium hydroxide
and carbonate (Figs. 32 and 34).
In order to estimate a possible effect of the calcium concentration levels in the mills
on the precipitation of lignin, a number of factors have to be taken into consideration,
for example:
• Calcium ion concentration and pH. Between pH 9 and 11 the relationship between
kappa number and calcium ion concentration is linear (paper II). At pH 12, there is
a slower increase in kappa number at low calcium ion concentrations.
• Temperature. In spite of the decrease in critical coagulation concentration (ccc)
with increasing temperatures (paper I), the difference in kappa number between
washing a pulp with and without calcium ions present seems to be independent of
the temperature at temperatures between 25 and 90 °C at pH 11 (paper II).
However, this may be dependent on pH since equilibrium reactions influencing the
concentration of calcium ions can change differently with temperature depending
on the pH.
• Molecular mass. In paper I, it was shown that it is the molecular mass fraction of
lignin above 1000 g/mol that is precipitated. Different pulping processes probably
degrade the lignin to different degrees, altering the fraction of the lignin being
precipitable.
55
• Origin of lignin. Calcium ions have a similar influence on softwood (spruce) and
hardwood (birch) kraft lignin regarding ccc, share of precipitable lignin (paper I)
and effect on kappa number (paper II). However, the calcium ion concentration
required to precipitate oxygen lignin is twice the concentration required to
precipitate kraft lignin. In addition, the proportion of precipitable oxygen lignin is
almost half that of kraft lignin (paper III). This may contribute to the lower
response of calcium ions when washing an industrial cooked pulp than when
washing a laboratory-cooked pulp, since oxygen lignin is brought countercurrently from the washing after the oxygen stage to the digester.
• Complexing agents. Added EDTA and DTPA or agents formed during the process,
for example degraded carbohydrate structures, can alter the concentration of
calcium ions ”available” for the precipitation of lignin. The presence of carbonate
has a similar effect.
At the calcium concentrations given in Fig. 33, a pH of 10-12 and a temperature
around 80 °C, which were the case in Mill A, lignin precipitation corresponding to 12 kappa number units can probably be expected, see Fig. 9. In the case of magnesium
ions, the estimation is even more uncertain than for calcium ions, since no pulp
washing trials corresponding to that referred to in the previous sentence have, to our
knowledge, been reported. The ccc(Mg) is almost the same as the ccc(Ca) at pH 11
(paper I), which indicates that there is a greater risk that lignin will be precipitated by
magnesium ions than by calcium ions, since the concentration of magnesium was
found to be comparatively high, Fig. 34.
56
Conclusions
The aim of this work was to study the influence of metal cations on the removal of
lignin during kraft pulp washing.
In the study of the influence of metal cations (Na + , Mg2+ , Ca2+, Al3+ ) on dissolved
lignin in solution, it was found that the colloidal part of the lignin (>1000 g/mol) was
precipitated by metal cations. The critical coagulation concentration (ccc) of the metal
cation increased in most cases with increasing pH and increasing lignin concentration,
but it decreased with increasing temperature and increasing valency of the metal
cation. The ccc(Ca2+) was similar for birch and spruce kraft lignins, but somewhat
higher for oxygen lignin, which may be explained by the different molecular mass
distribution of the lignin, the oxygen lignin having a greater fraction of low molecular
mass lignin with a high content of carboxylic groups.
The interaction of softwood kraft lignin and calcium ions was studied by the
application of ion-selective electrodes, and a method to estimate the free calcium ion
concentration in lignin solutions during titration with calcium ions was developed. An
end point was detected for the reaction between calcium ions and lignin. At calcium
ion concentrations below the end point, almost all added calcium ions were bound to
the lignin. The maximum content of calcium ions bound to the lignin was found to be
about 30 Ca2+ /100 C-9 units in the kraft lignin. This was verified by two other
methods. However, the use of calcium ion-selective electrodes in kraft lignin
solutions was complicated by a number of factors, including slow response, drift and
degradation of the ion-selective membranes in the electrodes.
In the study of the influence of metal cations on pulp washing, it was found that
calcium ions had a considerable effect on the lignin content of the pulp. In the
laboratory washing of a laboratory-cooked spruce kraft pulp, the presence of 5
mmol/L calcium ions resulted in a pulp with a kappa number 15 units higher than a
pulp washed in the absence of calcium ions. The corresponding difference in
laboratory washing of an industrial unbleached softwood kraft pulp was 5 kappa
57
number units for a pulp produced for bleached products and 10 kappa number units
for a pulp produced for unbleached products (sackpaper grade).
Lignin was retained in the pulp during washing at calcium ion concentrations lower
than the ccc(Ca 2+ ). One explanation of this may be an association of lignin molecules,
which are hindered from being transferred to the solution surrounding the fibre by the
porous structure of the fibre wall. Another explanation may be that the negatively
charged groups in the fibres cause an enrichment of calcium ions in the fibre
compared to the surrounding solution.
In the study of the properties of pulps containing calcium-precipitated lignin, it was
found that the precipitated lignin was darker than the other residual lignin in the pulp.
This is in agreement with the chromofore content of dissolved lignin being
precipitated at the end of the kraft cook.
Pulps containing calcium-precipitated lignin gave results on oxygen delignification
similar to those for pulps without precipitated lignin. In an alkaline hydrogen peroxide
stage, the precipitated lignin seemed to be more reactive than the other part of the
residual lignin. This is probably due to a higher content of phenolic groups in the
precipitated lignin. The higher bleachability of the calcium-precipitated lignin is
contrary to the lower bleachability of lignin precipitated in the cook by either a low
residual alkali concentration or a high sodium ion concentration. Nevertheless, when
the precipitation of lignin increases the lignin content in the pulp that enters the
bleach plant, an increased consumption of bleaching chemicals will arise.
An attempt to estimate the effect of precipitation of lignin by calcium ions during
industrial brownstock washing in a mill, where the calcium and magnesium
concentrations in a number of filtrated liquors were determined, resulted in 1-2 kappa
number units for a pulp produced for bleached products in a fibre line with an oxygen
stage. The estimate of the effect of magnesium is uncertain, but it may be
considerably higher due to the higher concentrations of magnesium in process liquors.
Increased closure of kraft pulp mills with, for example, recirculation of liquor from
58
the bleach plant to the brownstock area, probably leads to a higher concentration of
metal cations. This higher concentration would increase the degree of lignin
precipitation.
59
Acknowledgment
I would like to express my gratitude to:
My supervisor, Professor Nils Hartler, for his guidance and for being a fruitful
discussion partner throughout the work.
My co-author Malin Svedberg, who, in spite of all the difficulties we encountered,
was always positive and a pleasure to co-operate with.
Dr. Peter Axegård and Professor Ants Teder for valuable comments on the manuscript
of this thesis and the manuscripts of the papers.
Dr. Monica Ek for valuable comments on the manuscripts of the papers.
Dr. Johan Blixt for answering numerous questions, among other things, related to
inorganic chemistry.
Dr. Tord Eriksson for answering numerous questions related to wood chemistry.
Anders Törngren for the co-operation with the flow-through digester system, for the
assistance with the GPC-analysis and for all the jokes.
Hans Önnerud and Waleed Wafa Al-Dajani for assistance with experiments
performed at the division of Wood Chemistry.
Mona Johansson for assistance with kappa number determinations and for skilfully
performing the pulp viscosity determinations, and for, as unofficial head of the
laboratories, keeping a good working environment.
Brita Gidlund for being such a great administrator and taking care of all sorts of
practical aspects as well as the psycho-social environment at the department.
Dr. Hannes Vomhoff for the abundant encouraging conversations.
Birger Sjögren for his generosity in lending me laboratory equipment.
Fredrik Mörk for the arrangement of journeys and for being a great ”Reiseleiter”.
Toomas Kask for the trip to Tallinn.
Mattias and co-workers for indoor bandy, it was fun! and was good for my physical
well-being.
All friends at the department, among others: Mao - my room-mate, Lasse - the wine
supplier and joker, Micke & Christofer - the team, Ulrika - my neighbour, Martin the safety officer.
60
Dr. Anthony Bristow and Mary Roach for the linguistic revision of the thesis. Any
residual errors are the author’s entire responsibility.
The Swedish Pulp and Paper Research Foundation for financial support.
KAM, the ”Ecocyclic Pulp Mill” research program, financed by MISTRA, the
Swedish ”Foundation for Strategic Environmental Research”.
My parents Friedel and Sten for their constant and great support during these years.
My family, my wife Maria and my children Richard, Jonathan and Linnea, without
whose support I would never have been able to finish this work.
61
Literature
"paper" Encyclopædia Britannica Online :
<http://www.eb.co.uk:180/bol/topic?eu=59787&sctn=1&pm=1>. Accessed 29
August 2000.
Alén, R., Patja, P. and Sjöström, E. (1979): Carbon dioxide precipitation of lignin
from pine kraft black liquor, Tappi 62: 11, 108-110.
Allen, L.H. (1997): Pitch control in pulp mills, Minimum effluent mills symposium,
San Francisco, 211-224.
Allison, L., Ragauskas, A.J. and Hsieh, J.S. (2000): Metal profiling of southeastern
US SW and HW wood furnish, Tappi J. 83: 8, 97.
Andtbacka, S. (1998): The importance of washing in oxygen delignification and TCF
bleaching. Pulp washing must be optimized in fibrelines, Pulp Pap. Can. 99: 3, T93T96.
Annergren, G. and Sandström, P. (1996): A Bleach Plant Designed for Closure, Das
Papier 50: 10A, V12-V16.
Aurell, R. and Hartler, N. (1965): Kraft pulping of pine. Part 2. The influence of the
charge of alkali on the yield, carbohydrate composition and properties of the pulp,
Sven. Papperstidn. 68: 4, 97-102.
Back, E.L. and Salmén, N.L. (1982): Glass transitions of wood components hold
implications for molding and pulping processes, Tappi 65: 7, 107-110.
Bailey, J.H.E. and Reeve, D.W. (1994): Spatial Distribution of Trace Elements in
Black Spruce by Imaging Microprobe Secondary Ion Mass Spectrometry, J. Pulp Pap.
Sci. 20: 3, J83-J86.
Bailey, J.H.E. and Reeve, D.W. (1996): Determination of the Spatial Distribution of
Trace Elements in Jack Pine, Pinus banksiana Lamb., by Imaging Microprobe
Secondary Ion Mass Spectrometry, J. Pulp Pap. Sci. 22: 8, J274-J277.
Benko, J. (1964): The measurement of molecular weight of lignosulfonic acids and
related materials by diffusion. IV, Tappi 47: 8, 508-514.
Berthold, J. (1996): “Water Adsorption and Uptake in the Fibre Cell Wall as Affected
by Polar Groups and Structure”, Doctoral Thesis, Royal Institute of Technology,
Stockholm, Sweden.
Berthold, J. and Salmén, L. (1997): Effects of Mechanical and Chemical Treatments
on the Pore-Size Distribution in Wood Pulps Examined by Inverse Size-Exclusion
Chromatography, J. Pulp Pap. Sci. 23: 6, J245-J253.
Bird, R.B., Stewart, W.E. and Lightfoot, E.N. (1960): “Transport Phenomena”, John
Wiley & Sons,
Björklund-Jansson, M., Back, E.L. and Tufvesson, M.I. (1985): A review of kraft
pulp deresination and pitch problems, Transactions of the 8th Fundamental Research
Symposium, Oxford, 729-760.
62
Brogdon, B.N., Dimmel, D.R. and McDonough, T.J. (1996): The Influence of the
Bleaching Medium on Extraction Delignification Efficiency (II): Oxidized Lignin
Solubility, TAPPI Pulping Conference, Nashville, TN, 35-38.
Brogdon, B.N., Dimmel, D.R. and McDonough, T.J. (1997): The Influence of the
Bleaching Medium on Caustic Extraction Efficiency. Part 1: The Role of Fibre
Swelling, J. Pulp Pap. Sci. 23: 10, J499-503.
Bryant, P.S. and Edwards, L.L. (1996): Cation Exchange of Metals on Kraft Pulp, J.
Pulp Pap. Sci. 22: 1, J37-J42.
Crank, J. (1975): “THE MATHEMATICS OF DIFFUSION”, Second ed., Vol. 1.
CLARENDON PRESS, OXFORD.
Crotogino, R.H., Poirier, N.A. and Trinh, D.T. (1987): The principles of pulp
washing, Tappi J. 70: 6, 95-103.
Dence, C.W. and Reeve, D.W., ed. “Pulp Bleaching. Principles and Practice”,
Atlanta: Tappi Press, 1996.
Dong, D. and Fricke, A.L. (1995): Intrinsic viscosity and the molecular weight of
kraft lignin, Polymer 36: 10, 2075-2078.
Dong, D., Fricke, A.L., Moudgil, B.M. and Johnson, H. (1996): Electrokinetic study
of kraft lignin, Tappi J. 79: 7, 191-197.
Douek, M. and Allen, L.H. (1978): The distribution of calcium in kraft mill
brownstock systems, Pulping Conference, New Orleans, 17-26.
Edwards, L. and Rydin, S (1975): Washing of pulps. Part 3. Transfer rates of
dissolved sodium and lignin from pulps in a dilute, agitated suspension, Sven.
Papperstidn. 78: 16, 577-581.
Eriksson, G. and Grén, U. (1997): Pulp washing: Influence of temperature on lignin
leaching from kraft pulps, Nord. Pulp Pap. Res. J. 12: 4, 244-251.
Evans, D.F. and Wennerström, H. (1994): “The Colloidal Domain”, VCH
Publishers, Inc., New York.
Everett, D.H. (1988): “Basic principles of colloid science”, The Royal Society of
Chemistry, London.
Favis, B.D., Choi, P.M.K., Adler, P.M. and Goring, D.A.I. (1981): The leaching of
lignin from unbleached kraft fibres suspended in water, 67th Annual Meeting
Technical section CPPA, Montreal, Canada, A39-A45.
Favis, B.D. and Goring, D.A.I. (1984): A model for the leaching of lignin
macromolecules from pulp fibres, J. Pulp Pap. Sci. 10: 9, J139-143.
Favis, B.D., Willis, J.M. and Goring, D.A.I. (1983a): The effect of electrolytes on
the leaching of lignin from unbleached kraft pulp fibers, J. Wood Chem. Technol. 3:
1, 9-16.
Favis, B.D., Willis, J.M. and Goring, D.A.I. (1983b): High temperature leaching of
lignin from unbleached kraft pulp fibers, J. Wood Chem. Technol. 3: 1, 1-7.
63
Fengel, D., Wegener, G. and Feckl, J. (1981): Beitrag zur Charakterisierung
analytischer und technischer Lignine, Holzforschung 35: 3, 111-118.
Fogelberg, B.C. and Fugleberg, S. (1963): A study of factors influencing the amount
of residual alkali in sulphate pulp, Pap. Puu 45: 12, 675-680.
Fossum, T., Hartler, N. and Libert, J. (1972): The inorganic content of wood, Sven.
Papperstidn. 75: 8, 305-309.
Garver, T.M. and Callaghan, P.T. (1991): Hydrodynamics of Kraft Lignins,
Macromolecules 24: 2, 420-430.
Gellerstedt, G. and Heuts, L. (1997): Changes in the lignin structure during a totally
chlorine free bleaching sequence, J. Pulp Pap. Sci. 23: 7, J335-340.
Gellerstedt, G. and Lindfors, E.-L. (1984a): Structural changes in lignin during kraft
cooking. Part 4. Phenolic hydroxyl groups in wood and kraft pulps, Sven.
Papperstidn. 87: 15, R115-R118.
Gellerstedt, G. and Lindfors, E.-L. (1984b): Structural changes in lignin during kraft
pulping, Holzforschung 38: 3, 151-158.
Gellerstedt, G. and Lindfors, E.-L. (1987): Hydrophilic groups in lignin after oxygen
bleaching, Tappi J. 70: 6, 119-122.
Gierer, J. (1982): The chemistry of delignification, Holzforschung 36: 2, 55-64.
Goring, D.A.I. (1963): Thermal Softening of Lignin, Hemicellulose and Cellulose,
Pulp Pap. Mag. Can. 64: December, T517-T527.
Goring, D.A.I. (1971): "Polymer properties of lignin and lignin derivatives" in
“Lignins. Occurrence, Formation, Structure and Reactions”, Edited by K.V. Sarkanen
and C.H. Ludwig, John Wiley & Sons, 695-768.
Goring, D.A.I., Vuong, R., Gancet, C. and Chanzy, H. (1979): The Flatness of
Lignosulfonate Macromolecules as Demonstrated by Electron Microscopy, J. Appl.
Polym. Sci. 24, 931-936.
Gran, G. (1988): Equivalence volumes in potentiometric titrations, Anal. Chim. Acta
206, 111-123.
Grén, U. and Ström, K.H.U. (1985): Displacement washing of packed beds of
cellulose fibres, Pulp Pap. Can. 86: 9, T261-T264.
Grignon, J. and Scallan, A.M. (1980): Effect of pH and Neutral Salts upon the
Swelling of Cellulose Gels, J. Appl. Polym. Sci. 25: 12, 2829-2843.
Grähs, L.-E. (1976): Displacement washing of packed beds of cellulose fibres. Part 3.
A comparison of the washing behaviour of sodium and lignin, Sven. Papperstidn. 79:
4, 123-128.
Gullichsen, J and Fogelholm, C.-J., ed. “Chemical pulping”, Vol. 6A. Papermaking
Science and Technology. Fapet OY, 2000.
64
Gustavsson, C., Sjöström, K. and Wafa Al-Dajani, W. (1999): The influence of
cooking conditions on the bleachability and chemical structure of kraft pulps, Nord.
Pulp Pap. Res. J. 14: 1, 71-81.
Gustavsson, C.A.-S. and Wafa Al-Dajani, W. (2000): The influence of cooking
conditions on the degradation of hexenuronic acid, xylan, glucomannan and cellulose
during kraft pulping of softwood, Nord. Pulp Pap. Res. J. 15: 2, 160-167.
Gustavsson, R. and Swan, B. (1974): Evaluation of the degradation of cellulose and
delignification during oxygen bleaching, TAPPI Non-Sulfur Pulping Symposium,
Madison, Wisconsin, 43-51.
Hagström-Näsi, C., Järvinen, R. and Sundquist, J. (1987): Removal of lignin from
unbleached kraft pulp during washing, Pulp Washing Symposium, 213-225.
Hartler, N. (1978): Sorption cooking. Yield increase for unbleached alkaline pulps
through sorption of organic substance from the black liquor, Sven. Papperstidn. 81:
14, 457-463.
Hartler, N. and Norrström, H. (1969): Light-absorbing properties of pulp and pulp
components. III. Kraft pulp, Tappi 52: 9, 1712-1715.
Jacobson, S.A. and Nguyen, K.L. (1998a): Errata: Modelling the effect of fibre
properties and operating conditions on diffusion washing of different eucalypt kraft
fibres, Appita J. 51: 4, 312.
Jacobson, S.A. and Nguyen, K.L. (1998b): Modelling the effect of fibre properties
and operating conditions on diffusion washing of different eucalypt kraft fibres,
Appita J. 51: 3, 213-218.
Jansson, J. and Palenius, I. (1972): Aspects of the colour of kraft pulp and kraft
lignin, Pap. Puu 54: 6, 343-352.
Jansson, J., Palenius, I., Stenlund, B. and Sågfors, P.-E. (1975): Differences in
colour and strength of kraft pulps from batch and flow cooking, Pap. Puu 57: 5, 387396.
Johansson, E. and Ljunggren, S. (1994): The kinetics of lignin reactions during
oxygen bleaching, Part 4. The reactivities of different lignin model compounds and
the influence of metal ions on the rate of degradation, J. Wood Chem. Technol. 14: 4,
507-525.
Jurasek, L. (1995): Toward a three-dimensional model of lignin structure, J. Pulp
Pap. Sci. 21: 8, J274-J279.
Järvinen, R. and Välttilä, O. (1998): A practical method for studying NPEs in a kraft
mill, International Chemical Recovery Conference, 107-116.
Kerr, A.J. and Goring, D.A.I. (1975): The ultrastructural arrangement of the wood
cell wall, Cellul. Chem. Technol. 9: 6, 561-682.
Korhonen, J. and Lumme, P.O. (1978): Analyses of paper machine waters with ion
specific electrodes. Part 5. Sulphide determination using Ag+ ion specific electrode
and various measurement methods, Pap. Puu 60: 5, 373-379.
65
Korhonen, J., Lumme, P.O. and Kinnunen, L.J. (1973a): Analysis of paper machine
water with ion specific electrodes. Part 1. The effect of pH and ionic strength of
solution on Ca2+ , Cu2+ , Cl-, Na+ and NO3 - ion specific electrodes, Pap. Puu 55: 8,
559-570.
Korhonen, J., Lumme, P.O. and Kinnunen, L.J. (1973b): Analysis of paper machine
water with ion specific electrodes. Part 2. Ca2+ , Cu2+ , Cl-, Na+ and NO3 - ion specific
electrode potentials at various temperatures and in composite solutions, Pap. Puu 55:
9, 646-656.
Lenz, B. and Mold, J.R. (1971): Ion-selective electrode method compared to standard
methods for sodium determination in mill liquors, Tappi 54: 12, 2051-2055.
Li, J. and MacLeod, J.M. (1993): Alkaline leaching of kraft pulps for lignin removal,
J. Pulp Pap. Sci. 19: 2, J85.
Li, T.-Q., Henriksson, U. and Ödberg, L. (1993): Determination of pore sizes in
wood cellulose fibers by 2H and 1H NMR, Nord. Pulp Pap. Res. J. 8: 3, 326-330.
Lindström, M., Ödberg, L. and Stenius, P. (1988): Resin and fatty acids in kraft pulp
washing. Physical state, colloid stability and washability, Nord. Pulp Pap. Res. J. 3: 2,
100-106.
Lindström, T. (1979): The colloidal behaviour of kraft lignin. Part 1. Association and
gelation of kraft lignin in aqueous solutions, Colloid Polym. Sci. 257: 3, 277-285.
Lindström, T. (1980): The colloidal behaviour of kraft lignin. Part 2. Coagulation of
kraft lignin sols in the presence of simple and complex metal ions, Colloid Polym.
Sci. 258: 2, 168-173.
Lindström, T. and Westman, L. (1982): The colloidal behaviour of kraft lignin. Part
IV: Syneresis and hysteresis in swelling of kraft lignin gels, Colloid Polym. Sci. 260:
6, 594-598.
Mao, B. and Hartler, N. (1994): The effect of water quality on washing of
unbleached kraft pulps, Nord. Pulp Pap. Res. J. 9: 2, 134-135.
Marton, J. (1964): On the Structure of Kraft Lignin, Tappi 47: 11, 713-719.
Mathuthu, A.S. and Ephraim, J.H. (1993): Calcium binding by fulvic acids studied
by an ion selective electrode and an ultrafiltration method, Talanta 40: 4, 521-526.
McKibbins, S. (1960): Application of diffusion theory to the washing of kraft cooked
wood chips, Tappi 43: 10, 801-805.
McLaughlin, W.J., White, J.L. and Hem, S.L. (1993): The surface chemistry of
agglomerates formed in suspensions containing aluminum hydroxycarbonate and
magnesium hydroxide, J. Colloid Interface Sci. 158: 2, 333-341.
McLaughlin, W.J., White, J.L. and Hem, S.L. (1994): Effect of heterocoagulation on
the rheology of suspensions containing aluminum hydroxycarbonate and magnesium
hydroxide, J. Colloid Interface Sci. 167: 1, 74-79.
Milanovo, E. and Dorris, G.M. (1994): On the determination of residual alkali in
black liquor, Nord. Pulp Pap. Res. J. 9: 1, 4-15.
66
Miller, W.J., Shackford, L.D., Jiang, H. and Genco, J. (1991): Effect of entrained
black liquor carry-over on medium-consistency oxygen delignification, Tappi J. 74: 2,
117-122.
Mineql+ : 2000, <http://www.mineql.com>. Accessed Aug 10.
Morf, W.E. (1981): “The principles of ion-selective electrodes and of membrane
transport”, Elsevier Scientific Publishing Company, New York.
Månsson, P. (1983): Quantitative determination of phenolic and total hydroxyl
groups in lignins, Holzforschung 37: 3, 143-146.
Månsson, P. and Öster, R. (1988): Ozonization of kraft lignin, Nord. Pulp Pap. Res.
J. 3: 2, 75-81.
Mörck, R., Reimann, A. and Kringstad, K.P. (1988): Fractionation of kraft lignin by
successive extraction with organic solvents. III. Fractionation of kraft lignin from
birch, Holzforschung 42: 2, 111-116.
Mörck, R., Yoshida, H. and Kringstad, K.P. (1986): Fractionation of kraft lignin by
successive extraction with organic solvents. I. Functional groups, 13C-NMR-spectra
and molecular weight distributions, Holzforschung 40: Supplement, 51-60.
Niemelä, K. (1989): GLC-MS studies on pine kraft black liquors, Holzforschung 43:
2, 99-103.
Nyman, V., Rose, G. and Ralston, J. (1986): The colloidal behaviour of kraft lignin
and lignosulfonates, Colloids Surf. 21, 125-147.
Ohlsson, A. and Rydin, S. (1975): Washing of pulps. Part 2. The sorption of Na, Mg
and Ca on kraft pulp, Sven. Papperstidn. 78: 15, 549-552.
Papp, J. (1971a): Determination of sulphur compounds in alkaline pulping liquors
with a sulphide ion-selective electrode, Sven. Papperstidn. 74: 10, 310-315.
Papp, J. (1971b): Potentiometric determination of sulphur compounds in white, green
and black liquors with sulphide ion-selective electrode. 1. Determination of sulphide,
thiosulphate, sulphite, polysulphide sulphur and mercaptan in the presence of one
another by means of potentiometric titration, Cellul. Chem. Technol. 5: 2, 147-159.
Parks, G.A. (1965): The isoelectric points of solid oxides, solid hydroxides, and
aqueous hydroxo complex systems, Chem. Rev. 65, 177-198.
Pekkanen, M. and Nordén, H.V. (1985): Review of pulp washing models, Pap. Puu
67: 11, 689-696.
Ringbom, A. (1963): “Complexation in analytical chemistry: A guide for the critical
selection of analytical methods based on complexation reactions”, Interscience, New
York.
Robert, D.R., Bardet, M., Gellerstedt, G. and Lindfors, E.-L. (1984): Structural
changes in lignin during kraft cooking. Part 3. On the structure of dissolved lignins, J.
Wood Chem. Technol. 4: 3, 239-263.
Rosen, A. (1975): Adsorption of sodium ions on kraft pulp fibers during washing,
Tappi 58: 9, 156-161.
67
Rudatin, S., Sen, Y.L. and Woerner, D.L. (1989): "Association of Kraft Lignin in
Aqueous Solution" in “Lignin: properties and materials”, Edited by W.G. Glasser and
S. Sarkanen, Vol. 397, Washington, DC, 144-154.
Rydholm, S.A. (1965): “Pulping Processes”, John Wiley & Sons.
Salmén, L. (1990): On the interaction between moisture and wood fibre materials,
Mat. Res. Soc. Symp., 193-201.
Sandström, P, Ekebro, I., Pettersson, L. and Westerlind, B. (1998): Production and
use of TCF pulps, Pulping Conference, Montreal, Quebec, Canada, 1445-1456.
Sarkanen, S., Teller, D.C., Abramowski, E. and McCarthy, J.L. (1982): Lignin. 19.
Kraft lignin component conformation and associated complex configuration in
aqueous alkaline solution., Macromolecules 15, 1098-1104.
Sarkanen, S., Teller, D.C., Stevens, C.R. and McCarthy, J.L. (1984): Lignin. 20.
Associative interactions between kraft lignin components, Macromolecules 17, 25882597.
Scallan, A.M. (1983): The effect of acidic groups on the swelling of pulps: a review,
Tappi J. 66: 11, 73-75.
Sillén, L.G. and Martell, A.E. (1964): “Stability constants of metal-ion complexes.
Special publication no. 17”, 2 ed., The Chemical Society, London.
Sillén, L.G. and Martell, A.E. (1971): “Stability constants of metal-ion complexes.
Special publication no. 25”, The Chemical Society, London.
Sirén, K. (1998): Analytical methods for determination of metals, silicon and
phosphorous in kraft mill samples, International Chemical Recovery Conference,
1113-1119.
Sjöström, E. (1981): “Wood Chemistry. Fundamentals and Applications”, Academic
Press, Inc., Orlando.
Sjöström, K. (1999): Influence of ionic strength on kraft cooking and subsequent
TCF-bleachability, Nord. Pulp Pap. Res. J. 14: 3, 226-235.
Skogsindustrierna (2000): “The Swedish forest industries 1999. Facts and figures”,
Swedish Forest Industries Federation.
SSVL (1974): “The SSVL environmental care project. Technical summary”.
Sultanov, V.S. and Wallis, A.F.A. (1991): Reactivities of guaiacyl and syringyl lignin
model phenols towards oxidation with oxygen-alkali, J. Wood Chem. Technol. 11: 3,
291-305.
Swartz, J.L. and Light, T.S. (1970): Analysis of alkaline pulping liquor with sulphide
ion-selective electrode, Tappi 53: 1, 90-95.
Tasman, J.E. (1959): The permanganate consumption of pulp materials. IV. The
kappa correction coefficient, Pulp Pap. Mag. Can. 60: Convention Issue, 231-235.
68
Tasman, J.E. and Berzins, V. (1957): The permanganate consumption of pulp
materials. III. The relationship of the KAPPA number to the lignin content of pulp
materials, Tappi 40: 9, 699-704.
The Ecocyclic Pulp Mill (2000): “Final report KAM 1, 1996-1999”, KAM-report no.
A32.
Towers, M. and Scallan, A.M. (1996): Predicting the Ion-Exchange of Kraft Pulps
Using Donnan Theory, J. Pulp Pap. Sci. 22: 9, J332-J337.
Trinh, D.T. and Crotogino, R.H. (1987a): The rate of solute removal from kraft pulp
fibres during washing, J. Pulp Pap. Sci. 13: 4, J126-132.
Trinh, D.T. and Crotogino, R.H. (1987b): Sodium Equilibrium in Kraft Pulp
Washing, J. Pulp Pap. Sci. 13: 3, J93-J98.
Ulmgren, P. (1997): Non-process elements in a bleached kraft pulp mill with a high
degree of system closure - state of the art, Nord. Pulp Pap. Res. J. 12: 1, 32-41.
Ulmgren, P., Lindström, R. and Printz, M. (1994): The alkalinity of black liquor,
Nord. Pulp Pap. Res. J. 9: 2, 76-83.
Van Lee, R.C. (1987): Recommended method for designation bleaching stages, Tappi
J. 70: 11, 185-188.
Vilpponen, A., Gullichsen, J. and Lindholm, C.-A. (1993): Delignification by hot
alkali extraction, Tappi J. 76: 2, 134-138.
Westervelt, H.H., Frederick Jr., W.J., Malcolm, E.W. and Easty, D.B. (1982): The
determination and temperature-dependence of the stability constant of the calciumcatechol-4-sulfonate complex in alkaline aqueous media, Anal. Chim. Acta 138, 237243.
Wilcox, P.R. and Goring, D.A.I. (1990): The effect of pH on the leaching of lignin
from kraft and chemimechanical pulps, Cellul. Chem. Technol. 24: 6, 735-740.
Woerner, D.L. and McCarthy, J.L. (1987): Ultrafiltration and Light Scattering
Evidence for Association of Kraft Lignins in Aqueous Solutions, International
Symposium on Wood and Pulping Chemistry, Paris, France, 71-76.
69
Appendix 1
Appendices
Definition of technical terms and abbreviations
Black liquor is the liquor that exits the digester with the cooked chips at the end of
the kraft cook. In addition to the inorganic material that entered with the cooking
liquor, black liquor contains both organic and inorganic material removed from the
wood during the cooking process.
Brightness is the reflectivity, that is, the reflectance for an infinitely thick pile of
sheets, determined at an effective wavelength of 457 nm. It was determined according
to SCAN-P 3:93 and SCAN-CM 11:95.
Coagulation is the process of association of colloids to form a rather dense form of
aggregates.
Colloids may be seen as a dispersed or discontinuous phase distributed uniformly in a
finely divided state in a dispersion medium (or continuous phase). The term ”finely
divided state” refers to systems where at least one dimension is in the range 1-1000
nm.
Critical coagulation concentration (ccc) is the concentration of an electrolyte when
a colloid starts to coagulate.
Effective alkali (E.A.) is the percentage ratio of the sum of charged sodium
hydroxide and half the charged sodium sulfide expressed as NaOH, to charged dry
wood. It closely approximates the amount of charged hydroxide ions.
H-factor is a variable that is used to express the combined effect of time (t) and
temperature (T) on delignification.
t
H − factor = ∫ e a − b / T dt
t0
For the kraft process and most softwoods, a=43.2 and b=16113
Inductively coupled plasma (ICP) is an analysis technique to determine
concentration of elements. The plasma transfers the elements in the sample to excited
atoms or ions when atomic emission spectroscopy (AES) is used as detection method
and to ions when mass spectrometry (MS) is used as the detection method.
70
Appendix 1
Indulin AT is a commercially available precipitated pine kraft lignin produced by
acidification of black liquor.
Kappa number is defined as the number of millilitres of 0.02 mol/L potassium
permanganate consumed by one gram of dry pulp. The standard procedure followed in
this work is given in SCAN C1:77. The kappa number is related to the lignin content
of the pulp.
Klason lignin is a measure of a pulp’s lignin content. The lignin content is
determined gravimetrically after removal of extractives by acetone-extraction and
carbohydrates by acid hydrolysis.
Light absorption coefficient (k) is a measure of the concentration of light-absorbing
substances in a paper sheet. It was calculated from measurement of the reflectance
factor, which was determined for single sheets over a bright and dark background
using an Elrepho 2000 spectrophotometer. The sheets were prepared according to
SCAN-C 26:76.
Pulp viscosity is a measure of the molecular mass of cellulose in pulp. It was
determined according to SCAN-CM 15:88 where the intrinsic viscosity is calculated
using the equation:
[η] = lim
c→ 0
η − η0
η0 ⋅ c
(mL/g)
where η0 is the solvent viscosity, η is the sample solution viscosity and c is the
sample concentration, for which the pulp concentration is often used.
Sulfidity is a measure of the hydrogen sulfide concentration related to the
concentration of active alkali. The sulfidity is defined as:
Sulfidity =
mNa2 S
mNaOH + mNa2 S
⋅ 100 =
2 ⋅ [ HS − ]
⋅ 100
[OH − ] + [ HS − ]
(%)
where m is the mass of the chemical expressed as NaOH.
71
Appendix 2
Nomenclature in bleaching stages
The designation of bleaching stages in this thesis follows the method recommended
by TAPPI (Van Lee 1987).
O = oxygen
P = peroxide
W = wash, neutral or alkaline
Other letter symbols used in this thesis are the following:
Q = chelating agent
Q* = chelating agent with the addition of magnesium sulfate
72