A STUDY OF THE MECHANISM OF
MICHAEL REID SHAFFER
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
The author wishes to express his indebtedness to Dr. J.
Nolan, Professor of Chemical Engineering, for his keen interest and
valuable guidance in this research work. The following is a partial
list of persons who assisted in equipment fabrication, operating pro-
cedures, chemical testing, and literature translation: L. Hughlett,
K. Kalloway, K. Hutchins, G. Rose, J. Strickland, W. Ferguson, R. Baxley,
E. Warshyk, and S. I. Chang. The author also wishes to thank the other
members of his committee, Prof. R. D. Walker, Dr. R. Fahien, Dr. W. H.
Beisler, Dr. P. M. Downey, Dr. D. B. Wilcox, the late Dr. H. Meyer,
and Dr. R. G. Blake for their interest and participation.
TABLE OF CONTENTS
Acknowledgment . . . . . . .
List of Tables . . . . . . . .
List of Illustrations o . . . . .
Abstract of Dissertation . . . ...
Introduction . . . . . . .
* 0 0
I. A Brief History of Chemical Pulping . .
II. The Chemical and Physical Composition
of Wood . . . . . . . .
III. Sulfite Chemistry and the Sulfite
Pulping Process . . . . . .
IV. The Status of the literature . . . .
V. Definition of the Project . . . .
VI. Experimental Equipment . . . . . .
VII. Experimental Procedures and
Presentation of the Data . . . .
Wood Preparation . . . . . .
Liquor Preparation. . . . . .
Concentration Cooks . . . . .
Cooks with a Varied Ratio of "Free"
to "Combined" . . . . . .
Batch Type Cooks . . . . . .
Concentration . . ...
Concentration . . . . .
Pulp Handling and Treatment . . ..
VIII. Discussion of Results
General Equation for Reaction Rate . .
Prediction of Cellulose Degradation . .
The Effect of Varying the "Free" to
"Combined" Sulfur Dioxide . .
Batch Cooks . . . .. . ...
Observations Related to Pulping Theories
Conclusions and Recommendations . . . . .
Summary .. . . . . . . . .
References Cited . . . .. . .
Appendix At Numerical Solution of the Rate Equations
Biographical Items . . . . . . . .
LIST OF TABLES
1. Key to Apparatus Sketch Numbers . . . . . 37
2. Constant Temperature Constant
Concentration Cooks ............. 50
3. Cooks at 6.0% Total SO2 with 3.5%
Free SO2 and 2.5% Free SO2 Respectively . . 64
4. Batch Cooks . . . . .. . . 67
5. Constant Concentration Cooks with
Batch Cooking Temperature Cycle . . . . 70
6. Calculation of K . ...... ..... 79
7. Computation of the Rate Coefficients Using
Equation 14 . . . . .a .* . 91
8. Solution of the Rate Equation . . . . . 93
9. Test of Degradation Prediction Equation . . . 102
LIST OF ILIUSTRATIOUS
1. U. S. Chemical Pulp Production (1900 to 1960) . .. . 8
2. Flow Diagram, Sulfite Digester . . . . . . . 38
2a. Schematic Layout of Digester as Used in Constant
Concentration Cooking . . . . . . . ..
3. Sulfite Digester . . . . . . .... ... 39
4. Storage and Measuring Tanks, Sulfite Digester . ... 4.
5. Constant Temperature--Constant Concentration Cooks,
Nominal Concentration 2.5% Total S02 .. . . 54
6. Constant Temperature--Constant Concentration Cooks,
Nominal Concentration S.O. % Total SO2 ..... 55
7. Constant Temperature--Constant Concentration Cooks,
Nominal Concentration 7.5% Total S02 .. . 56
7a. Cross Plot of Data for Smoothing Purposes . .. . .57
8a. Cellulose Degradation as a Function of Cooking Time . . 9
8b. Cellulose Degradation as a Function of Cooking Time . . 60
8c. Cellulose Degradation as a Function of Cooking Time . . 61
9. The Effect of the Ratio of Free SO2 to Combined SO2 . . 6
10. Batch Cooks of Spruce ..... . . . . ... 68
11. Constant Temperature Cooks with Batch Temperature
Cycle . . . . . . . . . . . 71
12. Lignin Yield Related to Cooking Time . . ..... 77
13. Lignin Analysis Corrections . . . . . . . . 78
14. Constant Temperature--Constant Concentration Lignin
Related to Total Yield . . . . . . 85
15. Reaction Rate Constant Determination . . . . 87
16a. Simplified Representation of Degradation of
Cellulose with Cooking Time . . . . * 98
16b. Simplified Representation of Degradation of
Cellulose with Cooking Time . . . . . 99
16c. Simplified Representation of Degradation of
Cellulose with Cooking Time .... * . . 100
17. The Relationship Between the Reaction Rate
Coefficient and the Ratio of Free S02 to
Combined S2 . . . . 105
18. Lignin Yield Total Yield with a Varied Ratio
of Free SO2 to Combined SO2 . . . .. 106
19. Apparent Chemical Consumption and Waste Liquor
pH for Batch Pulping of Shredded Spruce . . 110
20. Lignin Yield of Batch Cooks . .. . 111
Abstract of Dissertation
Presented to the Graduate Council
in Partial Fulfillment of the Requirements of the
Degree of Doctor of Philosophy
A STUDY OF THE MECHANISM OF
Michael Reid Shaffer
Chairman Dr. William J. Nolan
Major Department: Chemical Engineering
An exploratory study of the mechanism of the bisulfite type
of sulfite pulping was made. By simulating constant concentration -
constant temperature conditions, kinetic data for pulping spruce wood
with sodium bisulfite over a sulfur dioxide concentration range of
from 2.5 to 7.5 per cent were obtained. Three temperature levels were
studied: 1600C, 1700C, and 1800C. From these data it is demonstrated
that a rate equation useful in reactor design or process control can
be developed. A method for predicting the accompanying degradation
of the cellulose has been shown.
Using batch type pulping methods, a measure of the chemical
consumption in bisulfite pulping was obtained. An interesting area
of further work was indicated by the rather unusual pH and chemical
consumption relationships observed.
A limited series of experiments was undertaken to observe the
effect of deviations from the bisulfite condition in the pulping chem-
ical. The amount of sulfur dioxide present for a given quantity of
cation was shown to have a very large effect on the reaction rate.
No difference in selectivity of lignin removal with varying ratios
of anion to cation was observed. This means that highly superior
special sulfite pulping processes are unlikely.
The over-all implications are that further development of the
moving interface theory of pulping should be undertaken. All observa-
tions made in this work indicate that more effort would be justified
in this area. Several methods of further testing and/or developing
the theory are suggested.
The timeliness of an engineering research project is always
of considerable concern since engineering, by its very nature, must be
built on a skeleton of economic considerations. After almost a century
of commercial application, the chemical pulping industry is again in a
state of technological upheaval. New processes are being proposed, old
processes are being restudied for new applications, and a new technology
The area in which the greatest turmoil and uncertainty seems
to exist is in the acid or "sulfite" pulping field. A broad, engineering-
based study in this area would do much to define both the variables in-
volved and the benefits to be obtained as well as the true limitations
of the process. In commercial sulfite pulping four different "bases"
or cations are used to buffer the pulping solutions. There are single-
and multi-stage processes. The pH range is from about one and a half
to about four and a half, which, in terms of hydrogen ion concentration,
is three orders of magnitude. Each user and each investigator claim
higher yield, superior pulp properties, or application to difficult
There are principally two reasons for all the concern. One is
that pulp properties are becoming more important. The revolution in the
container field resulting from such developments as steel foil, the com-
posite wall can, the multiwall bag with a coated polymer vapor barrier,
the printable display type shipping carton, the enormous cost reduction
in polymer films, to name a few, has forced a more vigorous evaluation
of pulp properties. The pulp and paper industry has had an undisputed
monopoly in the low cost packaging material line. To survive the on-
rushing competition it is having to evaluate and improve both properties
The other side of the sulfite pulping industry's problem is
steam pollution. Some years ago it was a generally held opinion in the
industry that the sulfite mill effluent in North America was sufficient
to render the combined flows of the Columbia, MLississippi, and Hudson
rivers unfit for fish. Inasmuch as the "clean water" movement has steadily
gathered strength, even mills which have demonstrated that the sulfite
effluent they discharge is not sufficient to be noxious or to be harmful
to marine life find themselves being attacked on esthetic grounds. Recov-
ery of the old traditional calcium base sulfite process waste liquors can-
not be done economically today. Once a change to a more readily adapted
process is considered, the entire question of process choice is thrown
open. There are no clear-cut guides based on a sound technology to rely
on. Even complete abandonment of the acid pulping process, and in some
cases the mill site, is advanced as a solution to the problem.
Over the past ten years the most interesting and apparently com-
mercially successful change in the sulfite process that has occurred has
been in the area of increased pH. The accompanying higher concentrations
of pulping agent are claimed to have some advantages in both pulp proper-
ties and the pulping of species not adaptable to the traditional sulfite
process (1) (2). Such processes, "Magnefite" with a magnesium bisulfite
pulping agent and "Arbiso" with a sodium bisulfite liquor, have received
extensive pilot plant testing and some commercial application. At the
time this work was begun there were no available data on kinetics and
reaction rates in the bisulfite pulping area.
It is the purpose of this project to demonstrate that rate equa-
tions suitable for the ordinary chemical engineering type of reactor
design calculations may be obtained. Further, it is hoped by this
demonstration to establish, at least over a substantial portion of the
pulping range, the effects of some of the important process variables
on both pulping rate and cellulose quality. It is intended that some
insight into the mechanism of pulping be obtained and that methods will
be outlined for further work. To accomplish these goals a specialized
piece of equipment was required. This was specified, designed, and
built as a part of the project. Some discussion of the variables and
an outline of the dissertation will complete the introduction.
The sulfite pulping process uses the sulfite and bisulfite ions
to dissolve the lignin, the material which binds together the fibers in
wood or other cellulosic materials. This dissolution of lignin must
be done in such a manner that the relatively undamaged fibers can be
later felted into a sheet of paper or paperboard or dissolved to form
cellulose solutions. From the standpoint of chemical equilibria it
can be seen that the sulfite pulping solutions must contain bisulfite
ions, sulfite ions, and dissolved sulfur dioxide since the pulping
agent is primarily a solution of sulfur dioxide to which a metallic
cation has been added.
The dissolved sulfur dioxide will approach an equilibrium with
any vapor phase present above the solution. Sulfurous acid has been used
without the metallic cation in laboratory scale pulp preparations, but in
commercial practice and most laboratory work a metallic oxide, hydroxide,
or carbonate is added. Among other things, this lowers the vapor pressure
of sulfur dioxide over the solution.
In commercial practice it is obvious that a minimum chemical re-
quirement would be sought for economic reasons. This consideration then
links together both concentration and the amount of pulping liquor pro-
vided. In the traditional sulfite process this consideration is compli-
cated by the fact that sulfur dioxide gas is released throughout the
process to avoid excessive pressures.
The temperature used in most commercial sulfite pulping processes
is similarly ill-defined. The steaming rate is usually gauged by the
pressure and this is interrelated to the amount of dissolved sulfur diox-
ide present and the chemical composition of the liquid phase. The cur-
rently widely accepted theory on the mechanism of sulfite pulping calls
for a period of slowly rising temperature, or "warm-up," to allow the
liquor to penetratedt the wood structure. This is followed by a period
of approximately constant temperature in which the reaction is assumed
to take place.
In addition to the variables of pulping liquor composition, ini-
tial concentration and amount supplied, and the temperature, a number of
variables such as wood species, chip size, and chip moisture are also
important. While there is much uncertainty about the relative importance
of the sulfite pulping variables, a generalized list could be s:
1) The physical and chemical properties of the wood or
other fiber source
2) The physical form or subdivision of the wood
3) The chemical composition of the digesting agent or
h) The ratio of "liquor" to wood, or the amount of
5) The temperature and its variation
6) The cooking time
In this project the first two general variables listed were
held approximately constant (within the best technique obtainable) and
the effects of the last four general variables studied.
The experimenter has selected bisulfite pulping as an area
of interest. He felt that sodium base, although currently limited in its
area of application by waste recovery problems, was the most promising
since relatively low cost sodium compounds usable in the process are
available. Further, there are no solubility limitations with sodium base
for any expected pulping conditions. Magnesium base is more readily re-
coverable, but the magnesium used must be low in calcium to avoid scaling
of the recovery equipment, and magnesium sulfite and bisulfite have very
definite solubility restrictions that fall well within the possible
Spruce wood was selected as a species. This selection was made
to allow comparison with the work of others. Methods for control of wood
properties and subdivision are treated in detail in the portion on pro-
The remainder of this work is divided in the following manner.
Chapters 2, 3, h, and 5 are background material intended to provide the
reader with sufficient grounding in the pulp and paper industry, its
nomenclature, literature, and principal processes such that an under-
standing of the project and its meaning and motivation may be achieved.
In Chapter 7 the apparatus designed and built for the project
and the procedures used are discussed in detail. Also, recommendations
for improvements that have resulted from experience are discussed. Pre-
sentation of experimental data is made in Chapter 8.
Chapter 9 is the discussion of the results and the conclusions
drawn from them. Recommendations for planning further work are included.
A BRIEF HISTORY OF CHEMICAL PULPING
The manufacture of paper is an ancient art. Developed by the
Chinese as a hand process, paper-making flourished for centuries almost
unchanged. The Industrial Revolution promoted change since it not only
supplied the technology necessary to mechanize the production of paper,
but also created an enormous new demand. Although the hand manufacture
of small quantities of paper for the communications of the era preceding
the Industrial Revolution met the needs of the day, the economic and
political changes that occurred in the Western world at the beginning
of the 19th Century brought about an enormous increase in both record
keeping and mass communication.
The development of the paper machine is a fascinating story,
but we are more concerned with one of the problems it created rather
than with those it solved.
Once paper-making was mechanized, the problem of pulp supply
became critical. The rag, straw, and bark sources were no longer ade-
quate. In the first 75 years of the 19th Century, three major processes
for preparing paper-making pulps from wood were developed. A mechanical
or "groundwood" process and two chemical processes, one alkaline and one
acid, appeared. A variation of the alkaline process gave pulps of high
strength but poor color. The acid process gave pulps of good color that
could be readily bleached where white papers were desired.
0 1 I ---- I I
1900 1910 1920 1930 1940 1950
Figure 1. U. S. Chemical Pulp Production (1900 to 1960, NSSC Included) (3)
Whereas the acid process used calcium bisulfite and sulfurous acid, the
alkaline process used sodium hydroxide and, in the improved process,
some sodium sulfide. withoutt recovery of the chemicals the acid process
was by far the more economical, and only the special pulp characteristics
and applicability to certain species of wood kept the alkaline process
In the 20th Century, two factors came into play that changed the
character of the chemical pulping industry. A chemical recovery system
for the modified alkaline or "Kraft" process was gradually developed, and
by about 1935 was good enough to change the economic picture completely.
This system not only recovered chemicals but also allowed the use of a
single lower cost material to supply the make-up chemicals, salt cake,or
crude sodium sulfate. A fortuitous distribution of losses allows a sin-
gle material, sodium sulfate, to supply both sodium and sulfur in proper
proportion, which is about 3:1 on a molar basis.
In addition to chemical recovery, the second factor that helped
change the chemical pulping industry is the fact that the Kraft process
has the advantage of being applicable to practically all wood species.
Pine was not readily pulped by the original acid process for reasons
which are presumed to be related to the presence of rosin and other
materials. But with the advent of the Kraft process, a tremendous impe-
tus was given to the production of chemical pulp in the United States
between 1930 and 1950. A whole new industry based on the Kraft process
grew up in the American South and West. Its effect on the national pulp
production figures is shown in Figure 1. A relatively modest growth is
shown for the acid or sulfite process while the expansion of Kraft is
Two other factors now enter the picture: the development of a
new pulping process and the broadening of the pulping base in the sulfite
process to include other cations in addition to calcium.
The new process is known as the Neutral Sulfite Semi-Chemical
process (NSSC) and is used for making pulps of moderate strengths in high
yields from hardwoods. This process is a combination of chemical and
mechanical pulping that has proven to be economical in spite of the use
of relatively expensive chemicals, sodium sulfite and sodium carbonate.
Its high yield keeps the wood cost down and mades the process practical.
Recently, however, a demand has risen for the development of a chemical
recovery system for this process so that it will be less expensive and
will not contribute to the stream pollution problem, about which there
'is growing concern. There has been some success in this development (4).
The change in the sulfite process to other bases, such as ammonia,
magnesium, and sodium for acid bisulfite pulping is also contributing to
the growth of the pulping industry. For the magnesium base there is a
relatively simple recovery process and for sodium base the systems being
developed for NSSC would be applicable. Recovery of ammonia base liquors,
however, is complicated by the thermal decomposition of ammonia in the
combustion process, and requires some special treatment such as ion
The chemistry of pulping, although much studied, is little
understood. Because of large areas of ignorance, many unsubstantiated
theories have been advanced that have interfered with a rational process
selection for industry growth. In addition, chemical pulping is pres-
ently undergoing a transition from a fundamentally batch process to a
continuous process. The necessary re-evaluation of the chemical pulping
processes that is accompanying this transition has resulted in the re-
assessment of much of the theory that has stood previously unchallenged.
There is sufficient concern for the relative merits of the vari-
ous processes to warrant such re-evaluation. In one case, a group of 30
companies retained one of the larger independent research organizations
to evaluate the future of the acid pulping process (5). Many other
studies are available in the literature. The consensus seems to be that
the old calcium base acid process is non-competitive. Almost all the
investigators recommend changing to a more soluble base and many recom-
mend eliminating excess sulfur dioxide beyond that necessary to form the
There is a wide divergence of opinion on the relative superior-
ity of the more soluble bases. Much of the work has consisted of a few
mill cooks or relatively few experimental cooks followed by a mill trial.
In no case have the process variables been studied completely enough to
allow an evaluation of the process through a study of its kinetics.
Practically all the chemical pulping agents in use attack cellu-
lose and cause a loss in fiber strength when their application is not
regulated within certain bounds. Another aspect of the problem of process
evaluation thus involves the selection of agents that will do the pulping
job without excessive attack on the cellulose. The literature and some
work done in this laboratory seem to indicate that, for an acid process,
the most likely areas of success may lie in the use of sodium bisulfite.
THE CHEMICAL AND PHYSICAL COMPOSITION OF WOOD
A very brief description of the chemical and physical structure
of wood, the principal paper-making material, will be given in an attempt
to clarify the background of the problem. For economic reasons, wood is
the principal source of paper-making fibers and some discussion of its
properties should be presented.
The chemistry of wood is complex and, as yet, little understood.
Much of the theory that does exist consists of definitions rather than
a clear chemical picture. The three chief constituents that make up about
90 per cent of the weight of wood are: cellulose; the hemicelluloses or
complex sugars, such as arabans, manans, and pentosans; and a material
called lignin, whose chemical structure is still largely unknown.
The chemistry of the cellulose is perhaps the best understood.
The structure of this polymer and the individual building blocks of cel-
lulosic units are fairly well established. The structure of the individual
units is given by various authors (6) (7):
CH2 OH 0 0
H H H
Ho 0--- o/ C
H CH / \ / H O/r
0' H CHOH H
These units form a linear chain. Basically cellulose is a typical
polymer with such attendant properties as the solution viscosity being
proportional to the chain length. Chains lying near one another inter-
act by a mechanism of "secondary valences" to form ordered regions or
micelless." These exhibit X-ray diffraction patterns similar to those
of crystals, and the groupings are often called crystallitess." This
sub-microstructure is further arranged in the form of fibrils visible
with an optical microscope. These fibrils are wound spirally to form
a hollow porous tube, or "fiber" skeleton. These hollow tubes are
closed at each end. The fibers form the grain of the wood and are
usually one to three millimeters in length, with a length to diameter
ratio on the order of 100 to one.
The function and location of the hemicelluloses are not so
clearly understood. Some of the hemicelluloses, such as pentosan, are
known to improve the strength characteristics of a paper web; however,
their function in the wood is open to discussion. The existence of
"tertiary" bonds has been proposed (8). For paper-making pulps, it is
generally advantageous to retain as high a hemicellulose content as
possible. A more detailed discussion and a review of the work done up
to 1953 may be found in a discussion by Nolan (9) on alkaline pulping.
Since that time much of the work has been similar to that of Galley (8),
an attempt to explain the observed strength improvement by various
While the arrangement and distribution of cellulose is well
understood, the distribution of the hemicellulose within the fiber wall
*is not so well known. Where some studies have fairly well determined the
distribution of lignin by dissolving the cellulose, locating the hemi-
cellulose by such direct methods is not possible. The fact that small
amounts of hemicellulose have such a marked effect on strength and the
fact that Findley (10) found that when the hollow fibers are impregnated
with pulping chemical in NSSC pulping, the yield of hemicellulose is
increased, indicate that the distribution of the hemicellulose is such
that it might be more concentrated immediately under the fiber's lignin
sheath or middle lamella.
The lignin, whose chemical composition is not yet well defined
(11), seems to.serve as a "cement" in wood, bonding the fibers together.
The areas between adjacent fibers are filled with lignin and there is
lignin in lesser quantities throughout the fiber wall. This material
makes up about 25 to 30 per cent of the dry weight of wood. It bonds
fibers to one another and cements the fibrils in the fiber wall, impart-
ing stiffness and strength to the fiber and, under conditions where the
lignin has not been modified by chemicals or temperature, it renders the
cell wall impervious to water.
The hollow fiber structure mentioned earlier is interconnected
by a series of openings in the fiber wall. Each fiber has openings re-
ferred to in the literature of wood structure as "bordered pits." The
name comes from the appearance of the openings under an optical micro-
scope. These openings are aligned with those in adjacent fibers. In addi-
tion there is a system of larger open-ended tubes called "vessels" in the
deciduous woods and "resin ducts" in the conifers. By applying a vacuum
or during successive heating and cooling, wood may be made to hold about
twice its dry weight of water. The specific gravity of the actual woody
material is almost twice that of water. A simple calculation shows that
approximately three quarters of the volume of wood is open space. A
further demonstration of the open structure of wood and the interconnec-
tions between openings is sometimes accomplished by connecting one end
of a log to a fire hydrant and showing the flow of water lengthwise
through the log. (Note: The arrangement is different in geometry from
most of the porous bed problems encountered by chemical engineers in
such applications as fixed beds of catalyst because the bulk of the
open space in wood is within the fibers rather than between them. A
realization of the actual physical arrangement is important to an under-
standing of the physical mechanism by which the pulping chemicals reach
the lignin with which they are to react.)
1ith such a structure, it is not surprising that a theory for
the pulping of wood should arise which would demand that the voids be
filled with the pulping agent as aprerequisite to the pulping reaction.
This "penetration" theory has been rather thoroughly refuted for Kraft
pulping (12). In the case of the acid pulping process, an extension of
the pepetration theory was used to account for "black" or "burned" cooks.
It was claimed that the insoluble dark material was "condensed" lignin
caused by excess acidity. This acidity was in turn supposed to have re-
sulted from a more rapid diffusion of sulfur dioxide gas than was possible
for the liquid phase. This sulfur dioxide redissolved in the water al-
ready present in the wood voids to form an unbuffered sulfurous acid.
Previous work done here at the University of Florida indicates that the
penetration theory is unsound even in the case of acid pulping (13) (14).
SULFITE CHEMISTRY AND THE SULFITE PULPING PROCESS
One of the chief problems in using sulfite pulping agents is the
instability of the sulfite ion. Even at room temperatures the metallic
sulfites and sulfurous acid slowly decompose by a sort of auto-oxidation-
reduction towards sulfur as the reduced form and sulfate as an oxidized
form (15). The complex reactions involved in such decomposition have
been studied by Marusawa (16) and Foerster (15) among others. A refer-
ence to a standard inorganic chemistry text (40) shows a large number of
possible sulfur acids that depend on the oxidation level of the sulfur.
The most reduced form of sulfur is the sulfide ion and the most oxidized
form of sulfur is the sulfate ion. The sulfite ion lies between these
states. Looking at the over-all oxidation-reduction picture, a portion
.of the sulfite is oxidized to the sulfate state and a part is reduced to
elemental sulfur. Reduction to the sulfide state does not seem to occur.
A number of intermediate levels of oxidation and the resulting formation
of ions seems to result. One of the principal intermediate products seems
to be the thiosulfate ion, but others seem to be present in smaller quan-
tities. In general, high temperature and high acidity promote the decom-
Another factor that must be considered is the gaseous phase in
equilibrium with sulfur dioxide solutions. The concentration of the
sulfurous acid in the solution and all equilibria such as those involving
the sulfite and bisulfite ions depend on the partial pressure of the
sulfur dioxide in the vapor phase. Considering the over-all solution
SO2 + H20 H2SO 3 + HSO;
there is a common ion effect with
NaHSO3 = Na+ + HS3-,
and the ionization of the sulfite,
Na2SO 2Na + SO,".
Also under consideration is the further dissociation of the bisulfit
HSOj H + SO .
The fact that there is considerable vapor pressure of sulfur dioxide
associated with bisulfite solutions, which by industry nomenclature have
no "free" sulfur dioxide, can be seen from Johnstone's correlation (17).
In general, there seems to be some confusion and ill-defined
nomenclature in this area in the industry literature. Perhaps it would
be well to discuss the situation since it will arise in referencing the
literature and again in the discussion of results. By definition, under
industry nomenclature, there is only one variable, concentration, that de-
fines the pulping agent used in bisulfite pulping once the cation or base
has been specified. This sort of thinking arises out of considering that
equimolal solutions (in the case of monovalent cations) of base and sulfur
dioxide result in a fixed and distinct form such as one would find in a
crystal. Considering the chemical equilibria involved one can see that the
bisulfite state is merely one point on a continuous surface of con-
centration of base and sulfur dioxide and attributing special proper-
ties to this point to distinguish it from its neighbors is a purely
artificial distinction. This point would be largely pedantic if it
were not for a second consideration. The decomposition of the solu-
tion changes the relation between the sulfite ion and the base because,
not only is the amount of sulfite ion reduced, but different equilibria
exist in the sulfate-cation system that results.
Having established an awareness of the considerations, it is
now intended to restrict the experimental work undertaken in this pro-
ject to an approximation of "bisulfite" pulping as defined in industry
nomenclature. The reasons are twofold. First and most importantly,
bisulfite pulping, by its industry definition, is the most active field
of interest in acid pulping. Secondly, such a restriction keeps the
experimentation within reasonable scope. Without the use of special-
ized statistical techniques recently published by Box (18) the intro-
duction of another variable would have tripled the number of experi-
The industrial background of the experimenter tended to em-
phasize the first consideration. The entire sulfite pulping industry
considers bisulfite pulping as a clearly limited project. Without the
knowledge of high temperature equilibria in a system which includes
the soluble lignin fragments and a method of tracking the shift in the
equilibria, such as with high temperature pH electrodes, the experi-
menter is fairly well limited to specifying a starting condition.
Specifying that this starting condition should be approximately the
bisulfite then becomes merely a convention. Useful work may be done
within this convention, but the over-all complexity of the problem must
not be forgotten or suppressed by such a device. Many of the paradoxes
and unexplained phenomena in the pulping technology result from the mis-
use of such conventions and definitions. As mentioned in the historical
sketch, there are, in industry nomenclature, three broad classes of
chemical pulping! acid, neutral, and alkaline. The distinction is made
strictly on the pH of the pulping agent. The best justification for
consideration of such a classification here is that the literature of
the pulp and paper industry is organized on this basis with respect to
indexing and bibliographies.
On such a pH scale, as mentioned previously, the process selected
for study is a modified, or relatively high pH, acid process. This is a
rather recent variation in the field of acid pulping. For reference, a
brief description of the conventional "acid bisulfite" pulping process
will be given. In this frame of reference the real and the supposed
significance of the increase in pH in the modified process will be more
In all chemical pulping processes, the debarked logs are mechan-
ically chipped and the chips are loaded into a "digester." The digester
is a large autoclave type vessel that contains normally from 3000 to
5000 cubic feet of chips. Loading is often accomplished with mechanical
loaders that use a steam jet or similar device to obtain a dense compact
filling to efficiently utilize the space available. After filling the
digester with chips, the operator or "cook" closes the digester with a
bolted or clamped cover. The cooking chemicals are pumped in and the
air which surrounded the chips is vented. The pulping process is start-
ed by heating. This heating may be by means of the introduction of
steam into the digester, "direct heating," or by circulating the pulp-
ing chemical, or liquor, through a surface-type heat exchanger and back
into the digester, which method is referred to as "indirect heating."
Since the pulping chemical in acid bisulfite pulping is a mix-
ture of the solution of metallic bisulfites and sulfurous acid, there is
a substantial partial pressure of sulfur dioxide in any gaseous phase
in equilibrium with the liquid phase. As the temperature is increased
to accelerate the reaction, this partial pressure rises along with that
of the water vapor. Air from within the chips is added to the mixture.
The vessel is suitably vented to prevent the total pressure from exceed-
ing the normal operating limits of the vessel.
This venting of the digester is called "relief." Considerable
care is given to relief schedules in an attempt to most effectively dis-
card the air and recover the sulfur dioxide for reuse in preparing more
pulping chemicals, or "cooking acid." The details of the schedules and
collection points are beyond the interest of this project. The concern
that is important is the fact that the release of sulfur dioxide in such
considerable quantities changes the entire character of the liquid phase
remaining, and conditions approach those of "bisulfite" pulping.
During the filling process, a compact, dense bed had formed, the
interstices of which were almost completely filled with the pulping agent.
The weight of this pulping agent is about five times that of the chips.
If an individual chip were examined, as mentioned earlier, one would fine
that from two-thirds to three-quarters of the volume was void. In the
usual case about half of this space is filled with water and the remain-
der with air. As the temperature within the digester rises, the total
pressure increases. If the liquid pulping agent is to penetrate into
the chips, the pressure in the space surrounding the chip must be higher
than the pressure within the chips. There is also diffusion of the ions
in solution into the water within the chip as a result of the concentra-
tion gradient. This flow and diffusion must follow the available chan-
nels, and for the bulk of the fibers within the chip, that is, those
not adjacent to either a surface formed in chipping or a naturally occur-
ring opening such as the resin ducts in the conifers and the vessels in
deciduous woods, this means passage through preceding fibers and inter-
connecting border pits. The fact that these fibers are partially filled
with water and partially with air tends to complicate the process. The
partial pressures of air and water are temperature dependent. In the
fully expanded penetration theory of the pulping mechanism, where it is
used to explain blackened or "burned" sulfite cooks, it is hypothesized
that the gas phase diffusion is faster than the combined flow and diffu-
sion transfer in the liquid phase. Hence, sulfurous acid without the
benefit of the metallic ions would be formed. This is presumed to bring
about an irreversible "lignin condensation" which renders the lignin dark
and forever insoluble in the pulping chemicals. This explanation of
"burning" neglects the fact that pulping can, under carefully controlled
conditions, be accomplished using sulfuruous acid alone (19). The very
long, slow, temperature rise periods of early commercial sulfite pulping
were supposed to allow the relatively slow ionic diffusion process in the
liquid phase to take place.
A theory developed in this laboratory, while having certain
shortcomings, gives a more reasonable explanation of the physical mech-
anism, This theory states that the chip temperature will be only slightly
below that of the surrounding pulping agent. Since the chip has both air
and water in its interior, a balanced pressure system will develop. In
the interior of the chip, the pressure will be the sum of the vapor pres-
sure of the water and the pressure contributed,by the air contained in
the chip. The-external pressure will be the vapor pressure of the water
and sulfur dioxide and any hydrostatic pressure. Flow into the chip will
occur only until the air within the chip is compressed so that the inter-
nal pressure balances the external pressure. In systems where no material
such as sulfur dioxide is used, such as in alkaline pulping, it is prob-
able that the net flow is always outward as air is released from the chip.
Ionic diffusion rates are too slow to account for the observed reaction
rates and a process of sequential diffusion and reaction does not appear
Even where there is a considerable partial pressure of sulfur
dioxide, a pressure equilibrium must be obtained at some point, and fur-
ther penetration can occur only by the diffusion mechanism. This fact is
supported by the very long penetration times required where manipulation
of the temperature and/or pressure is not used to effect rapid penetration.
For alkaline pulping, Nolan showed that the deliberate penetra-
tion of the hollow structure of the wood with pulping chemical by such a
temperature manipulation produced pulps markedly inferior in cellulose
quality to those obtained under similar conditions by normal procedures
(12). This might be the natural result of placing the pulping chemicals
where they must diffuse through the cellulose to attack the lignin. An
earlier detailed study of the alkaline pulping mechanism by Kulkarni (20)
led to the hypothesis that an interface of reaction with untouched lignin
on one side and an increasing gradient of delignification on the other
moves from the surface into the chip. This interface might be considered
to be marked by the occasional cleavage of a lignin molecule. Toward
the surface of the chip these cleavages increase. As pulping proceeds,
the interface moves toward the center of the chip with fresh chemical
diffusing from.the exterior to replace that consumed in the cleavage re-
action. The reaction products tend to diffuse outwardly countercurrent
to the diffusion of fresh chemical. When sufficient lignin has been
attacked, the fiber is freed from its neighbors although there may still
be a considerable amount of lignin remaining on the surface of the fiber.
By the time that the fibers at the center of the chip are free from each
other, the fibers at the outside of the chip, which are in contact with
the highest concentration of pulping chemical, must be rather completely
delignified. Notice that, in this theory, the principal path of the
active chemical and the reaction products is around the fibers rather
than through them.
In alkaline pulping, this moving interface theory has been ad-
vanced by Nolan and his coworkers (12)(22)(23) and strongly attacked by
a Swedish group (22). In acid bisulfite pulping, Nolan's view that there
is a very strong indication that the moving interface theory is more
satisfactory than the penetration theory has been presented in several
papers (13)(14) even though the experimental evidence has not yet been
as completely assembled as in the case of alkaline pulping. One should
note that this discussion centers around the physical mechanism rather
than the chemical.
Several of the studies made in this laboratory, while support-
ing the physical concept of the moving interface theory, resulted in
better empirical rate equations by considering that the reaction of lignin
dissolution is-carried out by two successive chemical reactions. This
concept was used by Findley in Neutral Sulfite Semi-Chemical pulping (10)
and by Chapnerkar in Kraft pulping (23). This idea might be contrasted
with the two-stage concept of the penetration theory in which the first
stage is penetration followed by the second, or reaction stage.
The chemical reactions themselves are complex in that they in-
volve the cleavage of a polymer chain of unknown specific structure. It
is usually hypothesized that there are points in the chain where cleav-
age may take place by sulfonation and that the dissolution of the lignin
is the result of such cleavages followed by the solution of the resulting
salt formed by the lignosulfonic acid and the cation supplied.
THE STATUS OF THE LITERATURE
There is an enormous literature of pulping, with sulfite pulp-
ing a major subject. Under these circumstances, since a complete re-
view would be sufficient material for an entire library based disser-
tation, only the material deemed most significant by the experimenter
will be included. This section then becomes a reflection of opinion
rather than a completely scholarly appraisal of a vast literature. Un-
fortunately there is no alternative.
The older but highly respected text of Hagglund (6) has 194 ref-
erences for the chapter on sulfite pulping. The textbook more recently
published by Casey (24) has 253 references for the corresponding chapter.
Each text has separate chapters on such subjects as the physical structure
of wood, the chemistry of cellulose, and other background material, with
each of these chapters having extensive references.
The late C. J. West was perhaps the industry's most extensive
bibliographer. On the use of ammonia, sodium, and magnesium in sulfite
pulping, he collected 304 references in the period before 1952 when this
was a relatively inactive field.
In addition to the extensive group of bibliographies compiled by
the late C. J. West, the Institute of Paper Chemistry issues monthly and
annual bibliographies under the direction of J. Weimer. For 1962 there
were 49 entries in the area of sulfite pulping.
The industry nomenclature, an appreciation of which is necessary
to an understanding of the literature, is based on stoichiometric con-
cepts. This results from building the system of nomenclature largely
from wet chemical analyses. As an example, "total sulfur dioxide" is
the result of an iodine titrationand "free sulfur dioxide" is a titra-
tion with standard base to the phenolphthalien end point. The difference
is called the "combined sulfur dioxide." Such a nomenclature suppresses
all dynamic considerations and equilibria. When even rather elementary
concepts of equilibria and dynamics are introduced, the nomenclature
becomes highly inconsistent. A serious effort to develop a more funda-
mentally correct system has been begun recently. As mentioned earlier,
there are extensive studies of the equilibria in the sulfur dioxide-
water system. There are two basic reasons why such studies have not
thrown more light on the fundamentals of the pulping reaction. With
such understanding, one could expect a more accurate terminology and a
better system of nomenclature.
The first reason is that there is not a clear understanding of
the reactions between the pulping agent and the wood. How these reactions
affect the sulfur dioxide-water system then cannot be quantitatively ex-
pressed. The lignosulfonic acids produced are strong acids (6) and cer-
tainly cannot be ignored.
The second reason is that only limited data on the decomposition
of the pulping agent are available. The decomposition reactions are
catalysed by the presence of selenium and elemental sulfur. In addition
the presence of an intermediate, thiosulfate, is also supposed to accel-
erate the reaction. The effects of the wood are unknown.
With reaction rate data for the pulping reaction, rate of de-
composition data on the sulfur dioxide system, and some reliable equilib-
rium relationships it should be possible to carry out a very interesting
theoretical study based on the changing distribution of sulfur during
the pulping process. A very brief attempt made with the limited informa-
tion available showed that the problem would not be practical for hand
calculation and would actually require the use of a good sized digital
computer, even with many simplifying assumptions which involve among
other things the important lignosulfonic acids and perhaps make the
study purely academic.
Several studies of the equilibria without reference to either
the reactions with wood or the decomposition problem are available. A.
recent example is the paper by Ingruber (25). This author has done ex-
tensive studies on pH in acid bisulfite pulping (26) (27) (28). An over-
all approach has not yet been accomplished, however.
DEFINITION OF THE PROJECT
It was the original intention of the author to satisfy himself
as to whether there was any good reason to continue work in sulfite pulp-
ing research and the related recovery of waste products, or if there were
such technical limitations to the sulfite process that it was obsolete.
As mentioned earlier, such evaluations have been made by others. Their
conclusions plus an observable trend in the industry support the view
that the classical calcium-base acid bisulfite pulping process is truly
obsolete. If sulfite pulping has any future, the initial conditions will
include a higher plH and a more soluble cation to achieve a concentration
range not possible with calcium.
The author's experience and a very superficial economic study
support the belief that from the point of view of both recovery of waste
liquor and solubility, sodium supplies the most practical cation or base.
Before proceeding more deeply into this discussion, it might be
well to review material already mentioned and introduce parenthetically
some further comments. To understand the problem completely, as seen
from the point of view of the experimenter, one must understand bisulfite
pulping defined both from a realistic point of view and in terms of in-
dustry nomenclature. The latter is necessary if one is to critically
evaluate this work in terms of the current literature of which it must
become a part.
In reality "bisulfite" pulping is merely relatively high pH sul-
fite pulping. It differs very little in a practical sense from the final
phases of the sulfite cook described earlier where much of the gaseous
sulfur dioxide has been vented to prevent excessive pressures. The
starting conditions are widely different. The pH in the so-called acid
bisulfite cook is usually below 2 at the start of the pulping process.
The concentration of sulfur dioxide in the cooking liquor is normally
between h and 6 per cent. There is enough cation or "base" so that on
a gross stoichiometric basis about 1 per cent sulfur dioxide might be
considered to be related to it, or "combined," as the metallic sulfite.
In the "bisulfite" pulping process considered in this study,
the initial pH was between 2.5 and h. A range of values of sulfur dioxide
between 1 and $ per cent was studied. Enough "base" (sodium ion) was
used to make the solution approximately the equivalent of sodium bisul-
fite with a slight tendency towards excess sulfur dioxide. In terms of
the equilibria this is merely a point on a continuum. In the literature
one finds that the bisulfite is a major bench mark in the terminology
of sulfite pulping to which special properties are assigned.
The sulfite pulping terminology is the result of taking a narrow
view of both wet analytical techniques and of stoichiometry. In general,
the amount of sulfur dioxide present in a solution is determined by
titration with iodine. The industry standard for analysis (32) tells
the analyst to report this as "total" sulfur dioxide. Then the entire
acid is titrated with standard base to the phenophthalein end point, or
a pH of about 8.3. The analyst is told to report the acid as "free"
sulfur dioxide. Further, he is told to subtract the "free" from the
"total" and report this as "combined."
In attempting to explain the results of analysis, as described
above, equilibria and many concepts so simple as to be included in a
basic chemistry course are ignored. It is said in the industry that
the "combined" represents the sulfite form of the base and that twice
the "combined" represents the bisulfite. Thus the condition with which
this project is concerned is the condition where the "combined" is
approximately equal to the "free." In a chemical sense, this point is
one of definition. In terms of the literature nomenclature, it is very
clearly fixed. From the experimental point of view, one degree of free-
dom, the relative amount of cation or base, is removed.
In reading the literature one becomes aware of the chaotic
condition of the nomenclature (29). The Technical Association of the
Pulp and Paper Industry, through the Sulfite Pulping Committee, as
mentioned earlier, is trying to rectify the problem. There is hope for
relief, but the current problem remains.
As conditions are changed from the bisulfite and the resulting
additional degree of freedom is used, (Two variables are required to
define concentration rather than one.) the terminology becomes confusing.
In addition to the terms "total," "free," and "combined" where three
terms are used to describe two variables and where the terms result from
the analytical method, some workers halve the value of the "free" and
call it "true free." The difference between the "total" and the "true
free" is the "true combined." This somewhat naive approach considers
that the "combined" represents the equivalent of the sulfite of the
"base" ion and that "true combined" represents the equivalent of the
bisulfite. To distinguish the original terms which resulted directly
from the analytical techniques, these may be referred to as "mill free"
and "mill combined," relating them to production terminology as opposed
to the more sophisticated terminology of research.
This situation has unfortunately been furthered by the technical
associations.- For example, The Technical Section, Canadian Pulp and Paper
Association in its Data Sheet C-OOc, June 1955, defines "total sulfur
dioxide" as the total amount of sulfur dioxide present in a sulfite
liquor and, although measured in grams per 100 milliliters, is expressed
as percentage. "Free sulfur dioxide" is the amount of free sulfur dioxide
present in a sulfite liquor and is equal to the "true free" plus one-half
of the sulfur dioxide in the form of bisulfite. "Combined sulfur dioxide"
is the difference between the "total sulfur dioxide" and the "free sulfur
dioxide" and represents one-half of the sulfur dioxide actually combined
with the base. This paraphrase of a longer definition illustrates the
manner in which the workers in the field can become entrapped in their
Obviously such definitions leave much to be desired from a the-
oretical point of view. Under such a system of nomenclature the pulping
solutions used in "bisulfite pulping" require only one variable in their
description and that is the amount of sulfur dioxide present. Although
commercially available, bisulfites are unstable, so that it is recom-
mended that one prepare one's own solutions (2). On preparing the solu-
tions one soon finds that not only are at least two concentration vari-
ables required to define the system, but that it is a dynamic system
with a very definite vapor pressure of sulfur dioxide. Closed storage
should be used for even such short periods of time as overnight. Under
the industry definition given above, such a bisulfite system has no "true
free" since the "free" or "mill free" is equal to the "combined" and is
also equal to one-half of the "total."
Even experienced workers in the field occasionally fall into
the nomenclature morass outlined when they try to compare their work with
that of others. A rational nomenclature with an accurate basis that is.:
simple enough in concept to achieve widespread acceptance would be an
enormous contribution to research in this field.
Returning to the definition of the project, of immediate need
in applying a pulping process is some method of predicting the effects of
changes in the independent variables on both reaction rate and pulp char-
acteristics. In general, this has been done in the industry by selecting
rather narrow working limits and carrying out sufficient experiments to
build a table. The chief shortcoming in this method is that often all
the variables have not been covered over a wide enough range to permit
interpolation and actual separation of effects has not been accomplished.
It might be better, especially in view of the advent of computer
control, if the tabulated results could be reduced to a set of equations
that would allow prediction. With such equations computer control can
take the measured history of a particular portion of the material in
process and adjust conditions to obtain a desired result. Such a'scheme
would be of value in the case of the present batch type operations or
in continuous digesters where large upsets or temporary stoppages might
occur. A method of approximating such a result is the subject of a recent
Within the limitations of university graduate research, such a
project may seem overly ambitious. With this in mind, some qualifications
should be made. The results are not expected by themselves to provide an
adequate basis for computer control in the modified sulfite pulping proc-
ess which is apparently coming into prominence in the pulping industry.
The results, however, should show the possibilities and provide a basic
structure on which more detailed studies can build the necessary refine-
ment. Thus it is hoped that an effective demonstration can be accomplished.
It is intended to obtain results in two forms. The first is to
demonstrate the practicality of a rate equation or equations which, when
integrated, will supply a predictor of the necessary time to accomplish
a given degree of delignification while pulping under a given set of
conditions. Since most pulping agents degrade cellulose while removing
lignin, a second or companion equation should be developed to predict the
effect on cellulose properties of any given set of conditions and reaction
The first type of equation will be a quasi-theoretical one and
will involve the common technique of testing various orders of reaction
until one is found to fit the data. The Arrhenius type equation will be
used to introduce the effect of temperature. It is hoped to use an empir-
ical relationship reported by Nolan in Kraft pulping (21) to obtain the
The digester system built for this study is very similar to the
experimental digester already in use in this laboratory and described
previously (31). The over-all arrangement is shown in Figure 2 and some
of the details are illustrated in Figures 3 and h.
The digester shell itself $ (2, 3)* is fabricated from a piece
of 14-inch steel pipe and two heads made from 11-inch steel plate. The
shell contains three tubes 6 (3) made of h-inch Nionel pipe. The mul-
tiple tubes allow several experiments to be run concurrently. An expan-
sion joint 7 (3) is provided to compensate for the difference in coefficient
of thermal expansion of Nionel and steel.
Each end of the digester tube is closed by a flange 8, 10 (3).
On one flange 10 (3) is mounted a reducing bushing 11 (3) and a 2-inch
standard pipe flange 12 (3). This provides a mounting for a Durco
2-inch plug valve 13 (2) with a Teflon sleeve and a body of Type 25 alloy.
This valve serves as the "blow" valve for emptying the digester. On the
opposite flange 8 (3) a i-inch pipe mounting 9 (2) is used for a i-inch
Durco plug valve. This half-inch pipe outlet is covered with a screen
cage to allow passage of liquid while retaining the chips and pulp.
The underlined numbers designate components and the numbers in
parentheses indicate the figure where the component is shown.
The liquor supply system consists of a 210-liter storage tank
1 (2, 4) and related piping. For simulated high speed continuous digester
operation, where hot cooking liquor is required, a return tube-type heater
2 (2, 4) is supplied with steam to preheat the liquor. Where the liquor
is forced in cold, a nitrogen pressure system is available to supply the
pressure. The storage tank gauge glass is shown as 21 (2).
For cases where an amount of liquor must be measured under pres-
sure,'a jacketed measuring tank 3 (2, 4) is supplied. This tank has a
calibrated gauge glass 20 (2) to allow measured quantities of liquor to
be withdrawn. An appropriate arrangement of valves and venting connec-
tions e.g., 23 (2) is provided to permit both tanks to be heated, with
circulation between the two. Appropriate valves may be closed and a
measured liquor charge withdrawn and introduced into the digester through
a flexible connection 22 (2). The digester may be positioned to allow the
connection of the hose 22 to the fitting 16.
Where liquor is to be stored cold and introduced under pressure,
rotameters 21 (2) are supplied to measure flow. Preheating is accom-
plished in jacketed coil-type heaters 15 (2). The flexible connections
to the digester, h (2), enter the body of the blow valves 13 (2) through
1/8-inch pipe fittings immediately upstream of the valve plug. This
insures liquor flow through the entire digester with minimum "dead"
space. When these connections are not used, the 1/8-inch openings in
the valve may be plugged.
The fitting 16 (2), which is a 1/2-inch pipe tee with the side
outlet reduced to 1/4-inch tubing, on the opposite end of the digester
from the blow valve, serves a number of purposes. First, when liquor
under pressure is supplied via the measuring tank ) (2), a flexible
connection provides the means of introducing the liquor through the
straight run of this fitting. As a second use, hot wash water to
clean the system after each blow is supplied through this connection.
Provisions are also made for mounting a pressure gauge at this point,
and a side tap leads waste liquor out through a coil cooler 17 (2) and
a sewer line 18 (2) when liquor is flowed through the system.
Emptying the digester tubes is accomplished through the two-
inch Durco "blow valves" and a flexible connection to a cyclone separator
14 (2) and hence to a canvas-lined trough 19 (2). Repeated "blows" with
hot wash water can be used through the provisions mentioned earlier to
insure that the system is clean.
Key to Apparatus Sketch Numbers, Figures 2, 3, 4
1 Liquor Holding Tank
2 Head of Bayonet Type Heater
3 Measuring Tank
4 Flexible Connections
$ Digester Shell
6 Digester Tubes
7 Expansion Joint
8 Closure Flange, Waste, Charging and Wash End
9 Liquor and Wash Connection
10 Closure Flange Blow Valve End
11 Reducing Fitting
12 Standard 2-inch Flange
13 Blow Valves
1i Cyclone Separator
15 Liquor Preheaters
16 Waste Liquor Outlet, Hot Liquor Charging and Wash Fittings
18 Cooling Water Outlet to Sewer
19 Receiving Trough
20 Measuring Tank Gauge Glass
21 Holding Tank Gauge Glass
22 Large Flexible Connection
23 Venting and Circulating Valves and Connections
Steam --X - - -
--X- -- -
Flow Diagram Sulfite Digester
Figure 32 Sulfite Digester
---s T 2'
Storage and Measuring Tanks Sulfite Digester
EXPERIMENTAL PROCEDURES AND PRESENTATION OF THE DATA
This chapter consists of descriptions of wood preparation,
liquor preparation, detailed operating procedures for each type of
experiment and the results obtained, and a description of the methods
used in pulp handling.
Wood Preparation The wood used throughout these experiments was spruce
obtained from the Augusta, Maine, mill of the Hudson Pulp and Paper Com-
pany. This wood was shipped to Hudson's Palatka, Florida, mill and chip-
ped July 7, 1962. The logs were generally straight and free from knots
and ranged from about 8 inches to about 20 inches in diameter. No mois-
ture samples were taken. This only affected chipping since the chips
were dried for storage, to prevent mold, and rewet before pulping. The
chipper supervisor stated that the logs appeared to be of normal moisture
content. The chipper used was a 12-knife Carthage chipper with "Norman"
blades. The chips were passed over Hudson's screens, which removed any
unchipped slabs or long slivers, and then removed from the conveyors
leading from the screens to the chip silos. After an air-drying opera-
tion, which probably resulted in some loss of fines in handling, the
screen analysis of the chips, obtained on a Tyler "Rotap," was as follows:
Screen Opening Per Cent Retained on Screen
+ 1.05 inches 2 .
1.05 inches + 0.742 inches 3.8
0.742 inches + 0,.25 53.6
Screen Opening Per Cent Retained on Screen
0.525 inches + 0.371 inches 27.4
0.371 inches + 3 mesh 10.3
3 mesh + 4 mesh 2.2
4 mesh 0.4
A little over half of the green chips were separated before dry-
ing and shredded in a vertical attrition mill. This operation has been
extensively described by Nolan (12) who has patents on this method of
chip preparation. The E. D. Jones "Vertiflex" used was operated with
3/8-inch shims lifting the upper housing for additional plate spacing.
The lower adjustable plate was backed down 120 turns or 0.480 inches
for a total gap of 0.855 inches. The feed gate was 3/8 open.
The product from the attrition mill was screened on a Link-Belt
vibrating screen with two decks. The top deck had slots 3/8-inches wide
by 1/2-inches long, and the bottom screen was 8-mesh screen. The use
of the bouncing action and the slotted screen gave a product whose cross-
sectional dimension in at least one direction was less than 3/8-inches.
The length of the product in the grain direction was substantially the
same as that of the original chips, about 5/8-inches. About 7 per cent
of the material passed through the 8-mesh screen and was discarded.
About 32 per cent of the material was oversize and was run through the
attrition mill again with the plate spacing reduced to 0.735 inches.
On this second pass, about 81 per cent was accepted, while 15 per cent
was oversize and about 4 per cent undersize.
The Tyler "Rotap" screen analysis of the combined product
was as follows
Screen Opening Per Cent Retained on Screen
+ 2 mesh 14 .7
+ 3 mesh 2 mesh 27.3
+ 4 mesh 8 mesh 25.0
+ 8 mesh 10 mesh 23.2
+ 10 mesh 16 mesh 6.8
+ 16 mesh 3.3
Through 16 mesh .7
Liquor Preparation It was felt that making up liquor by bubbling SO2
gas through a solution of the base would be the most practical method
of preparation. Although bisulfites are available, these are reported
to be unstable (2). There were available several open tanks in the
150-liter to 180-liter capacity range and it was felt that the use of
one of these tanks and a bubbler system for introducing the gas would
When making liquor for earlier acid bisulfite pulping studies
in this laboratory, a large portion of the water used (20-30 per cent)
had been supplied in the form of ice. The melting ice took up the
heats of solution and formation and kept the solution cool and the
vapor pressure of sulfur dioxide low. After several attempts to do
without the ice, because of an inadequate appreciation of the problem
of the vapor pressure of sulfur dioxide over a bisulfite solutions, the
use of ice was resumed. The same procedure was used in this research.
When the liquor was being prepared, the tank was mounted on a
scale and the calculated amount of ice and water added. Then the base
was dissolved. It did not prove to be an easy matter to dissolve large
quantities of sodium hydroxide. Often a solid cake of sodium hydroxide
would form in the bottom of the tankwhich would then require several
hours to dissolve. Besides being an inconvenience, failure to notice
this cake led to erroneous corrections in the amount of.base, which led
to errors in preliminary experiments.
In spite of the cooling, there is sufficient vapor pressure of
sulfur dioxide over bisulfite solutions to require that they not be left
open to the atmosphere for any extended period of time. In addition,
with such a method of preparation, mixing can be a serious problem. When
a batch failed to show the required chemical composition on analysis,
often merely stirring it again and resampling gave satisfactory results.
This sampling problem is similar to that encountered in commercial proc-
esses where calculated analyses often exceed in accuracy results obtained
by sampling and analysis. Then this type of work is undertaken, it would
seem that a closed tank with mechanical agitation would be a valuable
asset and would avoid much confusion and waste of effort.
The Palmrose method of liquor analysis (32) was used throughout.
This method has clearly defined endpoints in its titrations and is appli-
cable to waste liquors. The choice was based on earlier work in this
PROCEDURE AND DATA--CONSTANT TEMPERATURE AND CONCENTRATION OF BISULFITE
Constant Temperature--Constant Concentration Cooks These experiments
constituted the greater portion of the experimental work, and will be
covered by introductory comments, the description of some necessary
background work relating to the starting procedures, and a step-by-step
description of the procedure used.
Watr Liquor to Sewer Figure 2a. Schematic Layout
of Digester as used in Constant
!, Cooling water to Sewer Concentration Cooking
Cool Liquor y
2" Durco Valve
\ 2" Union
Liquor Hot Liquor /-
Tank -4 Condensate
To approximate constant temperature--constant concentration
conditions the digester tubes were loaded with the carefully prepared
and weighed charges of wood. The digester shell was heated rapidly.
Liquor from the storage tank 1 (all underlined equipment numbers refer
to Figure 2 in the preceding chapter, as well as to Figure 2a) was
brought through the rotameters 24 and preheaters where the liquor is
heated to the required temperature and introduced through flexible
connections to the body of the blow valve immediately inside of the
plug closure of the valve. This arrangement avoided "dead" spots.
After flowing up through.the digester tube, the used liquor, with a
relatively minor drop in concentration, passed out through a screens
which retained the chips and/or pulp, through the 1/2-inch connection
a coil type cooler, a needle valve, and then to the sewer. Col-
lection and periodic measurement at this point of each liquor stream
insured accurately controlled flows since the rotameters seemed to
have a tendency to stick.
By using a flowing stream of heated fresh liquor to give an
approximation of constant concentration and by approximating a con-
stant temperature environment, two of the principal pulping variables,
concentration and temperature, become known and fixed. By using one
species of wood and a common treatment, the important variables of
wood species and chip size, and the less important variables such as
chip moisture, etc., were fixed. The configuration or geometry of the
reaction zone, which includes actual lignin area exposed, diffusion
path lengths, etc., is a function of the starting conditions, the
reaction rates, the reaction times, etc. By selecting a range of known
times, it is possible to lump all these variables under the heading of
time in determining rate equations for any given starting conditions.
Perhaps the chief weakness in the procedure was in the method of approxi.
mating constant temperature. In the discussion of results in the next
chapter, several discrepancies will be pointed out which could have re-
sulted from a poor method of approximating constant temperature at the
start of the cooks.
Actually considerable care was given to developing a method of
approximating constant temperature, but the results indicate that further
work in this area might be needed, especially if the early stages of
pulping are to be studied. The method used in determining the starting
conditions involved a dummy run made with water so that temperature coulc
be directly inferred from pressure. A pressure gauge was installed on
the digester outlet. After filling the tube, the supply valve was closed
so that the pressure could be used as an indicator of temperature. The
time required to reach 100 psig starting from cold with a 150 psig heating
steam pressure was noted. A large number of such check runs should have
been made over the entire temperature range to be used and the tubes
should have been charged with wood. The indications from the results
are that either a higher temperature or a longer time should have been
used in preheating the digester.
The steps taken in making an actual experiment were as follows:
1) The storage and measuring tank were filled by the use of a
pump and hose arrangement with the prepared pulping chemicals.
2) Three batches of air-dried wood chips were weighed and a
moisture sample was taken to allow precise computation of the
weight of the wood used.
3) The batches were water soaked overnight by submergence
under a weighted screen.
4) The excess water was drained from the chips and the wet
drained chips were weighed to allow an estimate of the water
$) All three tubes of the digester were charged. Considerable
care was necessary since spillage would introduce errors in
the yield calculation.
6) A convenient starting time was selected.
7) Four minutes before the selected starting time, steam
was supplied to the jacket of the digester at 50 psig higher
pressure than desired during the pulping operation.
8) Che minute before the start, steam was supplied to the
liquor heaters to preheat them. Again, the higher pressure
steam was used to supply a steeper temperature gradient and
obtain more rapid heating.
9) At the preplanned starting time, the liquor (forced by
nitrogen pressure applied in the storage tank) was admitted
to the digester tubes. Approximately three minutes were
required to fill the tubes.
10) As soon as the air had been purged, the predetermined flow
rates were established7 using the rotameter provided, and check-
ing the flow by capturing timed samples at the sewer discharge.
The latter precaution was necessary since a few small pieces of
wood carried in the air from other work being done in the labo-
ratory seemed to find their way into the liquor, and these slivers
would jam the rotameter floats. (In any future work, a screen
in the liquor line would be useful.)
11) Five minutes after the start of the run, the steam pressure
was dropped back to the planned operating level.
12) The operation was monitored over the desired period of time
and any necessary adjustment made in pressure and flow rate.
13) At the end of the cook for each tube, the "blow line" was
attached and the contents of the digester "blown" into the
14) Two charges of hot water from the wash tank were "blown"
through the system to give a high degree of certainty that all
the pulp had been removed. Visual checking was used to insure
that all the pulp had been collected frcm the tubes.
15) The cyclone separator was washed into the collection trough
and the contents of the trough transferred to a canvas bag for
washing and pressing.
16) The product was screened using a Valley laboratory screen
with 0.008-inch slots and the screenings were dried and weighed.
17) To improve the accuracy of the yield determination, rather
than weighing the pressed accepted pulp and taking a moisture
sample, the entire quantity of pulp was made into heavy sheets
and dried on a sheet drier. This gave a very rapid drying proce-
dure and a minimum exposure to air and hence possible oxidation.
CONSTANT TEMPERATURE--CONSTANT CONCENTRATION COOKS
TEMP. TIME TOTAL FREE
COOK C MIN. S02,% S02,%
% ORIG. WOOD
TABLE 2 (Continued)
COMBINED % ORIG. WOOD
SO2,$ TOTAL SCREENED
77 170 210 4.98 2.51 2.47 50.8 49.1
TABLE 2 (Continued)
FREE COMBINED % ORIG. WOOD
SO,2 s02,% TOTAL
2.51 2.47 48.8 48.2
2.46 2.43 82.4 1.8
2.46 2.43 59.6 39.3
2.46 2.46 46.0 45.0
3.61 3.35 63.3 35.9
3.61 3.35 55.0 41.2
3.61 3.35 49.9 46.3
3.63 3.55 66.5 40.8
3.88 3.47 58.2 40.8
3.63 3.55 51.4 45.3
3.63 3.55 50.6 48.3
3.88 3.47 48.6 48.2
3.70 3.52 61.3 34.2
3.70 3.52 53.3 45.2
66 180 120 7.22 3.70 3.52 48.3 46.7
PER CENT YIELD
LD LIGNIN PER CENT
PULP SCREEN. ORIG. LIG.
0.1 22.3 1.9
17.1 23.5 77.6
3.1 18.6 21.1
0.8 17.4 2.4
4.7 22.3 30.9
0.8 20.1 12.3
0.9 15.5 3.9
7.2 19.7 36.8
The data obtained in the constant temperature--constant concen-
tration cooks are given in Table 2. In order to obtain the maximum in-
formation, cooking curves were plotted. In Figures $ through Figure 7,
the data are presented in this form, grouping according to the "target"
value of total sulfur dioxide used in the pulping agent and the tempera-
ture. Both the total and the screened yield are plotted against time.
The three temperatures, 1600C, 1700C, and 1800C, and the concentration
levels of 2.5 per cent, 5.0 per cent, and 7.5 per cent total sulfur
dioxide were selected with the idea of covering the possible operating
range. In the direction of the third independent variable, time, the
values were adjusted to achieve a wide range of yields, varying from
raw cooks with very small screened yields to completely cooked pulps.
This form of experimental design was selected since it was
considered desirable to have data that could be compared with the work
of others,and since previous experience in both acid and alkaline pulping
had proved that the pulping procedures used gave generally reproducible
results with minimum experimental error. Some cooks were carried out
in duplicate to confirm this experience and to check the specific pro-
cedures and the operator.
In addition, the fact that the cooking curves are by nature
smooth, with a fixed starting point and a known general form, made it
seem reasonable to carry out the experiments as indicated. A perhaps
more statistically elegant technique is that recently published by Box
(18). Box's approach is designed specifically for this type of work
to give maximum reliability with minimum experimentation.
Cooking Time Minutes
Constant Temperature--Constant Concentration Cooks
Nominal Concentration 2.5% Total SO2
5 Per Cent Total SO
80 ~- X 1700C
70 ~ ~_ ---- _ __------------,----,----
20 -0 -
0 100 200 300 400 500 600
Cooking Time Minues
Figure 6. Constant Temperature--Constant Concentration Cooks
Nominal Concentration 5.0% Total SO2
Figure 6. Contan-Tepe-tu----------Concntra-on -ook
200 300 b
Cooking Time Minutes
6.9% 7.h% Total SO2
Figure 7. Constant Temperature--Constant Concentration Cooks
Nominal Concentration 6.9-7.4% SO2
Concentration Grams per
Data from Figure 7
I I --
2.0 3.0 5.0 7.0
SO2 Concentration Grams per 100 Grams
Figure 7a. Cross Plot of Data for Smoothing Purposes
-- ~ ~ ~-------
The simulation or constant temperature--constant concentration
conditions was the most difficult problem undertaken. The apparatus used,
while capable of approximating these conditions, was a general purpose
experimental digester, rather than one designed specifically for this
purpose. For this purpose alone, a rather different design would be
superior. The problems associated with establishing the starting con-
ditions have been discussed and will be referenced again in the discussion
Using the raw data, an equation which was a reasonably good
predictor of over-all yield was developed. Before developing this
equation, the rather wide scatter of the data indicated that smoothing
of the raw data to arrive at reliable estimates of the cooking curves
might be desirable.
It was discovered that the plot of the logarithm of total
yield against concentration resulted in a family of parallel straight
lines. These plots (Figure 7a) not only gave an indication of the effect
of concentration, but provided a method of smoothing the data. The lines
drawn in Figures 5, 6, and 7 represent the result of this smoothing.
Included with the basic pulping data are measurements of the
degree of polymerization of the cellulose. These were made to serve
as guidesto the amount of cellulose degradation that accompanied each
experiment. The values given are computed from the Hercules Chart (33)
and the measurement of cupriethylene-diamine solution viscosities by
TAPPI Standard Method T230 SM 50. The relationships between D.P. num-
ber may be used as a relative guide.
Degradation of Cellulose as a Function of Cooking Time
30OC --------- --- ------------
Symbol Temp. Cone.
+ 1700 5.0
O 1800 5.0
100 200 300 400 500 600
Cooking Time Minutes
Figure 8b. Degradation of Cellulose as a Function of Cooking Time
Cooking Time Minutes
Degradation of Cellulose as a Function
of Cooking Time.
In Figures 8a, 8b, and 8c these data are plotted in a repre-
sentation of Log D.P. vs. Cooking Time. This type of presentation has
been used by Nolan (21). These data will be used later to develop an
equation for predicting cellulose degradation.
PROCEDURE AND DATA--CONSTANT TEMPERATURE, CONSTANT CONCENTRATION WITH
VARYING RATIO OF FREE TO COMBINED SULFUR DIOXIDE
Cooks with a Varied Ratio of "Free" to "Combined Since it was intended
to study primarily bisulfite pulping where the ratio of "free" to "com-
bined" is fixed, this study was considered supplemental. Some deviations
in liquor analysis from a bisulfite with exactly equal values of "free"
to "combined" occurred, and some idea of the effect of these errors was
desirable. Further, it was hoped that some general concepts useful in
the over-all evaluation of the process of sulfite pulping might be obtained.
As an example of the possible errors, consider a case where 5.0
per cent total sulfur dioxide is desired. The free sulfur dioxide would
be 2.5 and the combined SO2 2.5 per cent. After preparation and adjustment,
the total sulfur dioxide is 5.03 per cent, the combined 2.48 per cent, and
the free 2.55 per cent. With an open liquor system and hand agitation, no
further work is justified since losses in handling and unavoidable delay
in use, and errors in sampling which are associated with hand mixing pre-
vent a closer approach to the target values. Had minor deviations made
major changes in reaction rate, the closed liquor preparation system with
mechanical agitation, recommended elsewhere, would have been a necessity.
The problem was to plan experiments which would give a simple
clear answer to the question "What is the effect of a change in concentra-
tion of sulfur dioxide?" Since changing one of the values of "free,"
"combined" or "total" requires changing at least one of the other, some
planning is required. The most extensive data had been obtained in the
COOKS AT 6.0% TOTAL SO2 WITH 3.5% FREE SO2 AND 2.5% FREE S02 RESPECTIVELY
CONC. OF S02%
TOTAL FREE COMB. T
% ORIG. WOOD
) 200 300 400 o00 600
Cooking Time (min)
The Effect of the Ratio of Free SO2 with a Fixed
*Est. from 5.0 & 7.5% Total
vicinity of 5.0 per cent total sulfur dioxide, and additional data were
available at 7.5 per cent total sulfur dioxide. For the 5.0 per cent
total sulfur dioxide level, the "free" sulfur dioxide and the "combined"
sulfur dioxide were each 2.5 per cent. For one series of cooks,.the
"free" sulfur dioxide was increased to 3.5 per cent and in the other, the
"combined" sulfur dioxide was increased to 3.5 per cent.
This allows a direct comparison of the effect of increased
"free" and increased "combined" with the work done at 5.0 per cent total
sulfur dioxide. The increased total (6.0 per cent) can be accounted for
by comparison with a result interpolated from the actual experiments
made at 5.0 per cent total sulfur dioxide and at 7.5 per cent total
The results of the runs are given in Table 3 and plotted in
BATCH TYPE COOKS--INCREASING TEMPERATURE, FALLING CONCENTRATION
Batch- Type Cooks. Two types of cooks were run on a batch temperature
cycle, falling concentration and constant concentration. The first case
is the commonly encountered batch cook similar to commercial practice.
In such truly batch-type cooks, there was no liquor flow through the
digester tubes and the entire shell was rotated to provide agitation and
uniformity. The flexible lines from the three liquor preheaters to the
digester tubes were removed and the connections in the base of the blow
valves plugged. The liquor coolers were removed from the outlet end of
the digester and their attachment openings similarly plugged. Rotation
was at approximately two revolutions per minute.
BATCH COOKS 90 MIN. TO 1700C, LIQUOR 4.98% TOTAL, 2.59% FREE, 2.39% CCMB. SO2
0* MFill Size Chips
X Shredded Chips
*9 ____ ____
------- ~. ------ --------
2( 1 x
o 6C ------\ .---------
x x X
I I /
0 50 100 150 200 250 30)
Cooking Time Minutes
Figure 10. Batch Cooks of Spruce with 5.0% Total SO2
2.5 Combined SO2 and 2.5% Free SO2; A
Liquor Ratio of :1l, 90 Minutes to 1700C.
In this type of cooking, the liquor was stored in the pressure
tank to avoid changes in concentration due to the loss of sulfur dioxide.
Preliminary wood moisture value was obtained on the chips to permit more
accurate estimate of the liquor volume. The weighed chips were water
soaked overnight and loaded into the digester tubes. The measured amount
of liquor was added and the digester shell brought up to temperature in
small increments according to the desired schedule. A 4i-minute lag was
assumed to exist between the time a temperature change was made in the
shell and the time the change became effective in the digester. In other
words, steam pressures were set 41 minutes earlier than would be given
assuming an immediate temperature reaction. This was a convenient approxi-
mation and seems sufficiently good for this minor correction.
The batch type cooks were run with a bisulfite liquor of approxi-
mately 5 per cent total sulfur dioxide and, in the case of the falling
concentration, a liquor ratio of 4:1. This gave approximately 20 per
cent sulfur dioxide supplied on the basis of dry wood. On the basis of
the consumption figures obtained from these experiments, this was a little
higher than necessary.
The selected time to temperature of 90 minutes was chosen for
possible comparison with the work of others and with commercial appli-
cations. Nolan (14) had shown that, while there might be an optimum,
the effect of the length of the temperature rise period in acid bisulfite
pulping was small and even eliminating it completely was possible.
The data from the normal falling concentration batch cooks are
given in Table 4. The results are plotted in Figure 10 in the form of
CONSTANT CONCENTRATION COOKS 1VITH BATCH COOK TEMP. CYCLE (90 MIN. TO 1700C)
PER CENT LIGNIN
70- --- - ------------
0 60 .
20 ____ __ ---- I------ ------ --______
0 50 100 140 200 250 300
Cooking Time Minutes
Figure 11. Constant Concentration Cooks with Batch Temperature
Cycle (90 Minutes to 1700C)
BATCH TYPE COOKS--INCREASING TEPERATURE, CONSTANT CONCENTRATION
The batch temperature cycle cooks at constant concentration
combined the slow heating cycle with a constant supply of fresh liquor.
The tanks, heaters, and coolers were all used to supply fresh liquor at
approximately digester temperature to each tube and remove waste liquor
at a high enough rate to avoid any significant drop in concentration.
The procedure used in these cooks was identical to that used for the
constant temperature--constant concentration cooks except the same
temperature cycle used in the batch cooks was applied. The liquor pre-
heaters were operated at the same steam pressure as the digester shell
and, at the low flow rates needed, supplied liquor at approximately
These cooks were an attempt to provide a means of checking
the temperature relationship in any over-all equation. With the con-
centration variable fixed and a known temperature variation, it was
hoped that a partial test could be made of the equation which would
express, in differential form, yield as a function of temperature,
concentration, and time.
The results of the constant concentration cooks using the batch
temperature cycle are shown in Table 5 and the yield curves are plotted
as Figure 11.
Pulp Handling and Treatment After the digester had been blown, the pulp
was dipped from the canvas trough collector under the cyclone into a
canvas bag. The bag was hung on a rack and hot water at approximately
1700 F was sprayed into the bag and over the pulp for at least four
hours. The washed pulp was screened on a Valley laboratory screen using
an "eight cut" plate (slotted holes of 0.008 inch width).
The pulp was collected in another canvas bag and the screenings
were dried and weighed. The pulp was made up in the form of thick sheets,
using a laboratory sheet mold, and then dried on an electrically heated
sheet drier. This technique was used to minimize degradative oxidation
and still obtain the convenience and accuracy of oven dry pulps.
Although the capacity of the digester had been selected with
the possibility of beater tests in mind, these were not undertaken. There
is enough uncertainty in.such tests to require duplication and very close
controls. Since the desired objective was of a broad nature, not in-
volving application, it was felt that obtaining information on the degra-
dation of the cellulose would not only be more practical but more useful.
The viscosity test to determine the degree of polymerization of the cel-
lulose can be a more sensitive index of the results since considerable
degradation can occur before sheet strength tests show any effect (34).
The capriethylehediamine viscosities used to obtain the D.P.
numbers shown in Table 1 were obtained by use of the procedure described
in Technical Association of the Pulp and Paper Industry (TAPPI) Standard
Method T230 am50, using 0.5 molar cupriethylenediamine and Ostwald
Fenske viscometers. The pulps were completely delignified before dis-
solution by using a treatment with sodium chlorite and acetic acid.
The calculation of the D.P. number was accomplished by the use of the
Hercules Conversion Chart of Cellulose Viscosities, Form 803, AM 11-48
published by the Hercules Powder Company, Wilmington, Delaware.
The lignin content reported in the various tables was obtained
by the TAPPI Standard Method T222 m54 using 72 per cent sulfuric acid.
The values obtained may be slightly high since no alcohol benzene ex-
traction was used to remove any possible acid resistant rosins. These
were probably not present in any appreciable quantity since a few of
the pulps were shown to be almost completely lignin free. Had there
been any large amounts of acid resistant rosins, an analysis showing
a lignin free condition would not have been possible.
DISCUSSION OF RESULT'S
In general, the discussion of results will be treated in two
parts. In the first, those aspects that can be treated in a quantita-
tive manner will be developed. These are 1) the general equation for
reaction rate, 2) the computation of the Arrhenius constants and their
comparison with the work of others, and 3) the development of an equa-
tion for the degree of degradation of the cellulose. In the second
part, observations which do not involve specific measurements (something
did or did not happen) will be introduced and discussed with reference
to the moving interface and penetration theories of the pulping mechan-
General Equation for Reaction Rate The data used in developing this
equation are those presented in Table 1. In general, the cooking curves
developed from these data, shown as Figures 5 to 7, are about as ex-
pected. The "rule of thumb" that a ten degree Centigrade rise in tem-
perature doubles the reaction rate and hence cuts in half the time to
maximum screened yield seems to apply. As would be expected, increasing
the concentration of the pulping agent increases the reaction rate.
Previous work done in other pulping areas in this laboratory
(10) (20) (23) has resulted in the type of equation sought. While such
equations have a general theoretical basis, they are empirical in nature.
The basic concepts call for fitting the obviously non-applicable case
for the irreversible reaction in a homogeneous mixture to the data for
constant concentration--constant temperature pulping experiments. This
approach has a practical justification in that it gives results in a
The method used by Kulkarni (20) seems to be the most appli-
cable here. Practically any text dealing with chemical kinetics will
supply the form of the equation, for example, Smith (35). For a first
approximation assume a first order reaction with the equation of the
form for the constant temperature-constant concentration case
dL KL (1)
Where L = Lignin per cent remaining
9 0 Time
K = a constant,
C = concentration of sulfur dioxide
In the constant concentration case
K = kC (2)
and the effect of concentration may be introduced.
It should be noted that K and, subsequently, k are proportion-
ality constants that keep the equation dimensionally consistent. The data
may be grouped so that there are sub-sets for which both temperature and
concentration are at a given value. Using this device and integrating
-ln L = KG + B
90 2.5% Total, 160C
80 5_ .0% Total, 16000
70 Q17.5% Total, 1600C
60_ __x 2.5% Total, 17000C
0- + 5.0% Total, 1700C
S40 V 7.5% Total, 1700C
i2 2.5% Total, 1800C
30 0 5.0 Total, 18000
X\ 7.5% Total, 1800C
4. Cooking Time Minutei
\ \ \ *
10 \\__\ \ __
Cooking Time Minutes
Figure 12. Lignin Yield Related to Cooking Time
Corrected Screened Yield Per Cent
Corrected Screenings Per Cent
Figure 13. Lignin Analysis Corrections
CALCULATION OF K
CONSTANT CONCENTRATION CONSTANT TEMPERATURE COOKS
1 n 1/L
1.0 100.00 4.394
TABLE 6 (Continued)
TEMP. C 9 L YIELD
100/L 1 n 1/L
TABLE 6 (Continued)
where B = the constant of integration,
Considering L to be in percentage units and using the boundary
where O 0
L = 100
B = In 100 .()
Solving for K in (4) the desired rate equation is obtained
K lrn( ) I 2n (100) ($)
100 Q JJ
In Figure 12 are shown a number of the data sets with logarithm
of lignin plotted against time. Where there are enough data to establish
a relationship, i.e., 2.5 per cent total sulfur dioxide at 1600C, it can
be seen that an equation of the form of Equation 6 could very well rep-
resent the data. For the more rigorous cooking conditions, the lines
through the data have to be forced to pass through 100 per cent lignin
at zero time, indicating the problems in establishing the constant tem-
To calculate the constants in Equation 3 for all cases where
lignin data were available, Equation 3 was solved. As a first try, the
raw data were used and the results averaged. At the suggestion of Pro-
fessor Nolan, the smoothed yields (Figures 5, 6, ?) and smoothed lignin
data (Figure 13) were used. This technique made the rate coefficients
much more consistent but had little effect on the over-all averaged
results. The results are shown in Table 6. Compared to some of the
other areas investigated, the results of the constant-concentration--
constant temperature work are somewhat erratic.
An interesting paper on errors in linear approximations has been given
by Osburn (41) which shows why even with the difficulties shown, a rea-
sonable result can be obtained.
An interesting possibility is suggested by the difficulties en-
countered. Other workers have partially overcome a similar problem by
suggesting that two different rate controlling mechanisms exist, such
that in the course of the pulping reaction, control is transferred from
one mechanism to the other. A physical explanation of such a situation
could be made as follows. If the first mechanism had a chemical reaction
rate limitation,and the second was actually a question of diffusion,
which could be expressed as a chemical reaction, the two-mechanism proc-
ess would agree with the availability of lignin in the wood structure.
Findley (10) in NSSC pulping and Chapnerkar (23) in Kraft
pulping used two sequential reactions to correlate their data. If more
"early" or very high yield data had been taken here, such an approach
might have been profitable. Besides such a short range solution to the
problem, several attacks which might help separate diffusion and actual
chemical reaction would involve such things as deliberate prepenetration
of the pulping agent and the use of a range of bases such as lithium,
sodium, and potassium. The first technique was used by Nolan (12) to show
pulp damage in Kraft pulping resulting from the currently fashionable
"vapor phase" pulping. He also did some Kraft pulping with potassium
base and had an apparent slower rate of reaction. Such devices were not
considered in this exploratory approach and various techniques for hand-
ling the existing data will be explored.
In this type of approach, fractional orders of reaction are
allowed (35). In fact, in a complex mechanism of the type considered
here, fractional orders might also be expected. To ascertain if there
were any valid reasons for such an approach, a return to the general
equation was made.
dL kLnC, n 1. (6)
.(-n + 1)
L(n+ ) (-n + 1) Gk (7)
From this equation, it can be seen that if the device of linear-
izing such expressions by using the logarithmic form is applied,
(-n + 1) log L = Log 9 + log (-n +1) Ck (8)
then in such a plot, (-n + 1) will be the slope of the resulting straight
line relationship. Unfortunately, such a device failed when applied to
these data. When all the data were considered there was a distinct cur-
vature in the plot. At the higher lignin content conditions a value of
minus two for (-n + 1) would have been a fairly good approximation. How-
ever, the lower lignin content data represented a virtually flat area
with a slope of zero. About the only conclusion that could be drawn is
that the apparent order of reaction was between one and three.
In another attempt to evaluate the method and the data, the follow-
ing approach is used. The K values for the last four experiments at 1600
C and 2.5 per cent were averaged (this was selected because more data
were available for those conditions and the slower rates should better
Total Yield (%)
Constant Temperature-Constant Concentration Lignin Yield
Related to Total Yield
illustrate the issue.) and the yield calculated for the first two experi-
ments. The average value of the reaction rate coefficient (calculated)
from the data for 240, 360, 480, and 600 minutes) is 0.00593 and Equation
0.00593 In 100 (9)
L 100 e- (0.o00593)
Using this coefficient, the lignin yield for the 90-minute cook
would calculate to 57.1 per cent and 39.4 for the 160-minute cook. The
experimental results were 55.7 and 3$5.. If the lignin total yield re-
lationship shown as Figure H1 (and which will be discussed later) were
used, the predicted total yield in the first case would be 73 per cent
and in the second 66.5 per cent. The experimental values were 72.3 and
64.3 per cent.
No conclusive proof of a two-stage reaction results. It is
interesting to note that K is very sensitive to experimental error and
conversely, the use of average and hence approximate values of K gives
reasonably good predictions of the results.
To account for the temperature effect, the Arrhenius type equa-
tion will be used. Although this equation is empirical in nature, some
success has been achieved in obtaining the constants involved by purely
theoretical methods. A general form is
d In K =
dt RT_ (10)
Variation of K with Temperature
Total SO2 I
~x 7.5% Total S02
--- I __ -- _--
-r 4 ,S
0.00228 0.00230 0.00232
Figure 15. Reaction Rate Constant Determination
here A energy of activation
R = gas constant
T = absolute temperature
K = reaction rate coefficient.
If, as in many cases, A is approximately constant,
In K =-A D, (11)
where D = constant of integration.
Both A and D are related to the particular reaction being studied.
A method used in checking the validity of the application obtained is the
use of a plot of log K versus I/T for the experimental results. For this
case, the.plot is shown in Figure 15. Averaged values of K were used.
In computing these average values, specific values of K obtained from
using short cooks (such as 2.5 per cent total SO2 and 1800C for 90 minutes)
were not used due to the unreliability of the data in the very early stages
of the cook. It is interesting to note at this point that the results
could be approximated by three parallel lines with a different line for
each liquor concentration used. While not supported by any theoretical
considerations this condition can be used to introduce the liquor concen-
tration effect into the approximate prediction equation. Since a differ-
ent intercept value D seems to exist for each concentration, it is only
necessary to find D as a function of concentration and introduce this re-
lationship into Equation 11. Since plotting D against C results in a
straight line, D can be empirically expressed in terms of concentration
D = 0.064 C + 24.L8
Substituting into Equation 10,
In K -12-900 + 0.06 O + 24.48 (13)
This equation provides a predictor that considers variations in
temperature and concentration. It is written in exponential form as
K n expQ ( -12,900 + 0.064 + 24.48) (14)
This equation is checked and compared to experimental values
in Table 7. (* "cxp" is used to abbreviate the expression of e to the
The energy of activation found here is 25,700 calories per
gram mol. Findley (10), who treated NSSC pulping as a two-stage reaction,
found a value of 35,400 calories per gram mol for his first reaction and
25,200 for his second. In calcium-base acid bisulfite pulping, Calhoun
et al. (36) report a value of 20,200 calories per gram mol. In "soda"
type alkaline pulping Laroque and Maas (37) report 32,000 calories per
gram mol and Kulkarni (20) reports 24,160 calories per gram mol for
Kraft pulping. The recent paper of Kubelka et al. (38) reports 25,000
calories per gram mol for a "semi-acid" cooking condition that appears
to have a slightly higher starting pH than the work done for this
The second constant in the Arrhenius equation, D, or the constant
of integration in the nomenclature of this paper, varies with concentra-
tion in this result and ranges from 2.5 to 10 minutes for 7.5 per
cent total sulfur dioxide to 3.5 x 1010 minutes -1 for 2.5 per cent total
sulfur dioxide. Findley reports 4.4 x 1041 for his first reaction and
5.9 x 10 for his second. The other investigators reported gave results
the order of magnitude of which was 10I for "soda" pulping, 1010 for
Kraft pulping, and 109 for acid bisulfite. The "semi-acid" work resulted
in a value of 109.
Since both the wood species and the pulping conditions and agents
used by the investigators mentioned above differed from the conditions used
in this study, direct comparison is not possible. However, it certainly
appears as though the values obtained are reasonable.
COMPUTATION OF THE RATE COEFFICIENTS USING EQUATION 14
Temp. Cone. Calc. Avg. Expl.
TK C, _% K K
433 2.5 0.0058 0.0062
443 2.5 0.0112 0.0117
453 2.5 0.0230 0.0213
433 5.0 0.0067 0.0076
h43 5.0 0.0131 0.0141
453 5.0 0.0252 0.0242
733 7.5 0.0080 0.0092
h43 7.5 0.0155 0.0152
453 7.5 0.0290 0.0287
Sample Solution of the Rate Equation
Example: Cook #61
- (Y -. 8.7) e
= (Y 48.7) (
+ 0.064C + 2hL.8)
- In (Y 48.7) (0.00580) 9 + B1
In (Y 48.7) .538 + 3.938
Y = 48.7 + e3400
Y = 48.7 + 29.6 = 79.1