EP A-R 2-73-196
APRIL 1973 Environmental Protection Technology Series
Steam Stripping
Odorous Substances
from Kraft Effluent Streams
Office of Research and Monitoring
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and
Monitoring, Environmental Protection Agency, have
been grouped into five series. These five broad
categories were established to facilitate further
development and application of environmental
technology. Elimination of traditional grouping
was consciously planned to foster technology
transfer and a maximum interface in related
fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
1. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL
PROTECTION TECHNOLOGY series. This series
describes research performed to develop and
demonstrate instrumentation, equipment and
methodology to repair or prevent environmental
degradation from point and non-point sources of
pollution. This work provides the new or improved
technology required for the control and treatment
of pollution sources to meet environmental quality
standards.
-------
EPA-R2-73-196
April 1973
STEAM STRIPPING ODOROUS SUBSTANCES FROM
KRAFT EFFLUENT STREAMS
By
Bjorn F. Hrutfiord
Lennart N. Johanson
Joseph L. McCarthy
University of Washington
Seattle, Washington 98105
Project 12040 EXQ
Project Officer
Dr. H. Kirk Willard
Pacific Northwest Environmental Research Laboratory
National Environmental Research Center
Corvallis, Oregon 97330
Prepared for
OFFICE OF RESEARCH AND MONITORING
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
For sale by the Superintendent of Documents, U.S. £o»ernment J*rtatlng.0ffloe. Washington, D.C. 20402
Price $1.25 dwnwtlc postpaid or »1 QPO Bookiiore
-------
EPA Review Notice
This report has been reviewed by the Water Quality Office, EPA, and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
-------
ABSTRACT
Nature of Steam Volatile Components
The nature of the steam volatlle organic compounds which occur in kraft
pulp mill aqueous streams has been determined. In order of decreasing
concentration, these are alcohols, terpenes, ketones, sulfur bearing
compounds and phenolic compounds. Methanol is the main alcohol and
was found in concentrations from 280 to 8400 ppm, while ethanol
occurred at about 1/10 of these levels. Terpenes were found in ranges
from a few ppm to about 4500 ppm. Acetone is the main ketone and
occurs from 2 to 210 ppm. Sulfur compounds range from 2 to 800 ppm,
based upon prior studies. Combined-stream quantities of these components
in Ib/ADTare 11.5 to 15.9 for methanol, 0.9 to 2.6 for ethanol, 3.8 to
9.2 for terpenes, 0.07 to 0.4 for acetone.
Process Design Studies
The feasibility of combining steam stripping of black liquor issuing
from a continuous Kamyr digester, with steam stripping of condensates
was explored. Volatile compound release predictions were made for such
a process considering terpenes, terpineols, methanol, and the sulfur
compounds, hydrogen sulfide, methyl mercaptan, methyl sulfide and
dimethyl disulfide. Black liquor stripping would Increase overall
stripping costs about two-fold, but would have the advantage of
simplifying turpentine recovery and further decreasing odor emanations
within the pulp mill and from discharged condensates, as compared with
a one to two-column condensate stripping process. An exploratory study
has been made of predicting ternary and higher component systems of
volatile constituents with water, utilizing binary vapor-liquid and
solubiIity data.
Separation of 01 Is
The possibility of separation of the turpentine fraction of SEKOR oils
from the impurities, chiefly sulfur compounds such as methyl sulfide and
iii
-------
dimethyl disulfide and methanol was investigated. For low concentration
of impurities, present methods of oxidative destruction would appear
preferable. For higher concentrations fractional distillation may be
promising if only a-pinene is to be recovered. Solvent extraction does
not appear to be promising: selective adsorption, using silica gel
in preference to molecular sieves, may be feasible for higher
concentrations of impurities.
This report was submitted in fulfillment of project I2040EXQ under the
sponsorship of the Environmental Protection Agency in cooperation with
the University of Washington, Seattle.
iv
-------
CONTENTS
Sect 1 on
I Conclusions '
I I Recommendations *
III Introduction 5
IV Objectives and Plans 7
V Nature and Concentration of Steam Volatile Compounds 9
VI Process Design Studies 3I
VII Separation of SEKOR OiIs 65
VIII Acknowledgements '*
-IK
IX References IJ
X Publications and Patents 79
XI Glossary 8I
-------
FIGURES
I FORMATION OF METHANOL, ETHANOL AND ACETONE IN LABORATORY 18
KRAFT COOKS OF DOUGLAS FIR
2 SCHEMATIC OF SAMPLE SITES FOR BATCH AND CONTINUOUS PROCESSES 20
3 FLOW SYSTEM FOR COMBINED STEAM STRIPPING STUDY UTILIZING 34
HEAT FROM BLACK LIQUOR STRIPPING
4 FLASH CONCENTRATION AND OXIDATION STEPS IN BLACK LIQUOR 40
RECOVERY
5 MULTiEFFECT EVAPORATION STAGES AND DIRECT CONTACT EVAPORATION 42
6 TERNARY LIQUID EQUILIBRIUM RESULTS AT 25°C. WATER PINENE 60
METHANOL
7 TERNARY LIQUID EQUILIBRIUM RESULTS AT IOO°C. WATER PINENE 60
METHANOL
8 TERNARY LIQUID EQUILIBRIUM RESULTS AT IOO°C. WATER 61
DIMETHYL SULFIDE METHANOL
9 TERNARY LIQUID EQUILIBRIUM RESULTS AT IOO°C. WATER- 61
TERPINEOL-METHANOL
VI
-------
TABLES
No. Page
I Organic Compounds In Kraft Mill Condensate Streams 10
II Estimated Composition of Kraft Mill Streams 13
III Methanol Content of Some Condensate Streams from a 14
Pacific Northwest Kraft Mi I I
IV Total Yields of Methanol, Ethanol and Acetone from 16
Laboratory Kraft Cooks
V Concentration and Total Turpentine in Kraft Mill Conden- 22
sate Streams
VI Flow of Condensate Streams in the Kraft Pulp Mill Process 23
VII Material and Energy Balance for Kamyr Digester, Washers, 35
and Flash System
VIM Tower Height and Process Steam Requirements for Methyl 38
Mercaptan Removal by Steam Stripping of Black Liquor
IX Removal of Volatile Constituents by Steam Stripping of 38
Kraft Black Liquor
X Operating Parameters of Black Liquor Stripping Column 41
for Cases Extended to Steam Stripping of Condensates
XI Summary of Volatile Component Distribution - Case I 44
XII Summary of Volatile Component Distribution - Case Ma 45
vii
-------
TABLES (CONT'D)
NO» Page
^"•^"™ i—™»S*—
XIII Summary of Volatile Component Distribution - Case Mb 46
XIV Summary of Volatile Component Distribution - Case III 47
XV Case I la - Composition of Feed and Product Streams of 48
SEKOR A Columns
XVI Case Ma - SEKOR A Column Design For Condensate Stripping 49
XVII Case Ma - Composition of Feed and Product Streams, SEKOR 49
Column C Stripping Evaporator Condensate
XVIII Case I la - SEKOR Column C Design For Evaporator Condensate 50
Stripping
XIX Comparison of Contaminant Content of Condensate Without 50
and With Steam Stripping
XX Water-Organic Binary System Results 57
XXI Methanol-Organic Binary System Results 58
XXII Summary of Margules Equation Constants For Binary Systems 59
XXIM Summary of Renon Equation Constants For Binary Systems 59
XXIV Summary of Break-through Data for Selective Adsorption of 71
Dimethyl Disulfide from Terpenes
viit
-------
SECTION I
CONCLUSIONS
Characteristics of Steam VoI at Ile Compounds
I. Methanol, ethanol, acetone and traces of other water soluble
organic compounds are quantitatively the most important steam strippable
organic compounds in condensate streams.
2. Turpentine, characterized by a-pinene and especially a-terpineol is
the second most important group of steam volatile organic compounds
found in condensate streams.
3. Sulfur bearing compounds do not occur in large concentrations, but
are important owing to their odor characteristics.
4. Effective stripping of turpentine decanter underflow and of blow
condensate would control about three-fourths of the total steam
volatile compounds in kraft mill streams.
Design Studies
I. Steam stripping of black liquor under moderate pressure (under 100
psia) appears feasible on the basis of preliminary design study.
2. Terpenes, methyl mercaptan, and methyl sulfide can be largely
stripped out of black liquor in such a column.
3. Overhead vapors generated are sufficient to serve as a heat medium
for steam stripping of condensates through the first two multi-effect
evaporator stages.
4. Methanol is stripped out of black liquor and condensates with
difficulty, appearing in the evaporation train through the first two
multi-effect stages, and in the air effluent from weak liquor oxidation.
5. Incorporation of black liquor stripping in a kraft system utilizing
condensate steam stripping would increase overall costs approximately
two to three-fold, relative to costs of steam stripping only condensates,
-------
OiI Separations
I. The separation of the turpentine fraction of SEKOR oils from the
sulfur compounds by solvent extraction does not appear promising.
2. The separation of the turpentine fraction from sulfur compounds
utilizing selective adsorption was found to be technically feasible
on a laboratory scale.
3. Silica gel will selectively retain sulfur compounds allowing their
removal from terpenes in a packed column system.
4. Synthetic zeolites of 10 to 13 angstrom pore size also selectively
adsorb sulfur compounds relative to terpenes, but less effectively than
si 1ica gel.
-------
SECTION II
RECOMMENDATIONS
All decanter underflow streams as well as blow gas condensate from
batch kraft pulping processes and analogous streams from continuous
processes should be steam-stripped for removal and control of volatile
organic compounds.
Decisions on steam stripping of other condensate streams should be
made on an individual mill basis. For example, if methanol is to be
reduced to levels below the equivalent of 2.0 Ib methanol/ADT of pulp,
then two columns should be used, stripping the first two evaporator
stage condensates in addition to the higher concentration condensates.
Steam stripping will be superior to air stripping of condensates
for the majority of installations and should be used in all except the
unusual cases, in which the effluent air can be utilized in furance
combustion.
Steam stripping of black liquor would appear to be advantageous only
for extreme situations of removal and recovery of volatile constituents.
-------
SECTION III
INTRODUCTION
Study of the steam stripping of Kraft Pulp Mill effluent streams began
over a decade ago in the laboratories of the Chemical Engineering
Department of the University of Washington. These investigations were
conducted largely through the "Pulp Mills Research Program" of this
*
University, financed in part by the Northwest Pulp and Paper Association.
Results were reported to this sponsoring organization and in part have
since been published as three papers under the acronym SEKOR (Stripping
Effluents for Kraft Odor Reduction) as SEKOR I, II, and III (TAPPI
Magazine, 50_ No. 2, pp 82-85, February 1967; 5£ No. 2, pp 86-91,
February 1967; and 5£ No. 6 pp 270-275, June 1967). Part of the study
(reported in SEKOR II) was a cooperative venture with the St. Regis
Company at their Tacoma, Washington Kraft Mill, in which the steam
stripping of condensates was successfully demonstrated in a bench-scale
continuous pilot plant.
Fellicetta, Peniston and McCarthy first identified the major odorous
constituents in such condensate streams (TAPPI 36_ 425, 1953) as hydrogen
sulfide, methyl mercaptan, dimethyl sulfide, and dimethyl disulfide.
This finding has since been verified in mills and laboratories through-
out the world. The SEKOR process provides a means of concentrating
these and other non-aqueous constituents, to allow disposal or further
processing and thus remove them from effluent waters.
While initial emphasis in the steam stripping study was aimed at removal
and control of odorous compounds, it has been apparent for some time
that significant amounts of other steam volatile compounds are present
in kraft pulp mill aqueous streams. The SEKOR I paper pointed out the
presence of significant amounts of turpentine in the originally
-------
isolated oils, and analysis of aqueous condensate revealed the presence
of various alcohols and ketones of low molecular weight. These classes
of compounds are not of importance from the viewpoint of odor, but
are very important in mill effluent due to the major BOD load caused
by the methanol etc., and the turpentine is not particularly desirable
in that certain compounds in this class may be toxic to marine
organisms.
With this general background, enough was known to define what further
information was needed for design of removal equipment. Concerning
the compounds to be encountered, additional information on their nature
and properties was required as well as information on concentration
ranges in the several streams to allow decisions as to which streams
should be processed and what compounds to expect in the stripped
material. Sizing of equipment could proceed from this information,
as well as the somewhat more difficult problem of defining operating
procedures.
Earlier reported work demonsrrated the feasibility of stripping
condensates, and the present study extended these studies and included
as well an analysis of application of steam stripping to black liquor,
since this is the stream in which all volatile compounds are present.
Fractionation of the resulting stripped-out oils presented a special
problem, and basic investigations into separation methods were also
a part of this study.
-------
SECTION IV
OBJECTIVES AND PLANS
For the past three years, this work has been continuing at the
University of Washington under the sponsorship of the Research and
Monitoring Office of the EPA. The goal of this program was to develop
sufficient fundamental data and design information to allow the design
of SEKOR process modifications for a variety of industrial situations.
Inherent in the study was the improvement of process economics and
means of recovery of values such as hot reusable water and chemicals,
in order that the process may be made as attractive as possible for
adoption within the Kraft pulping industry.
The three specific aims of the research program were:
(a) To secure further information concerning the nature and concent-
ration of steam-voI atile substances present in Kraft pulp mill black
liquors and condensates arising from a number of species of wood and
under several process conditions;
(b) To conduct further laboratory experiments and SEKOR process
design studies in order to evaluate several alternative ways of
conducting the SEKOR process and to permit the optimum procedure or
procedures to be identified; and
(c) To conduct laboratory and process design studies directed toward
the development and evaluation of procedures by which SEKOR oils,
arising under various conditions, can be separated on an industrial
scale into components or fractions which may be sold to return a
significant income to offset the costs of conducting the SEKOR process,
A further part of the objectives and plans for the study comprised
collaboration with representatives of the Weyerhaeuser Company. A
joint plan of work-was developed and made known to representatives of
-------
Environmental Protection Agency. Many discussions and conversations
occurred and yet at the time of conclusion of the present study, it
had not yet been possible for representatives of the Weyerhaeuser
Company to proceed with development of the contemplated pilot plant
and thus, this phase of our original plan could not be carried out.
Meanwhile, complementary attention was devoted to other phases of
the activity, and this is set forth below.
8
-------
SECTION V
NATURE AND CONCENTRATION OF STEAM VOLATILE COMPOUNDS
IN KRAFT PULP MILL PROCESS STREAMS
Introduction
The general plan for this part of the study was to survey kraft pulp
mill aqueous streams for steam volatile organic compounds and to
identify as many of them as possible. Following identification, the
concentration range in which the more important of these organics
occurred was determined, especially in the several condensate streams
i n the mill.
Samples were obtained from a wide variety of sources, from a number of
Northwest pulp mills. Basically these covered the main variables of
individual streams in batch and continuous processes and also the
influence of wood species being pulped. Effort was concentrated on
condensate streams such as digester relief condensate, blow gas
condensate and evaporator condensate in batch processes, and their
equivalent flash tank vapors and evaporator condensates in continuous
processes. Some other more concentrated samples were also studied,
i.e. crude sulfate turpentine and black liquor from various mills.
Identification of Steam Volatile Compounds
A summary of the organic compounds identified in our studies is pre-
sented in Table I. Many of the compounds listed have also been
identified by others.
-------
TABLE I
Organic Compounds in Kraft Mill Condensate Streams
AIcohoIs
Methanol
Ethanol
l-propanol
2-propanol
Butanol
2-methyI-1-p ropanoI
Ketones
Acetone
2-butanone
3-pentanone
3-methyI-2-b utanone
4-methyl-2-pentanone
2-heptanone
Phenols
Guaiacol
Phenol
Syringol
o-Cresol
m-Cresol
p-Cresol
AcetovaniI lone
Sulfur Bear!ng
Methyl Mercaptan
Dimethyl Sulfide
Dimethyl disulfide
Thiophene
Terpenes and Related Compounds
a-pi nene
Camphene
6-pinene
Mycrene
A -carene
a-phellandrene
a-terpinene
Limonene
$-phellandrene
3-terpinene
Terpinolene
Fenchone
Linalool
FenchyI a IcohoI
Terpi nene-4-ol
a-terpineol
2-methyIfuran
Toluene
4(p-toIyI)-1-pentano I
10
-------
The sulfur containing compounds were well known and were verified as
present in oils isolated from condensates by steam stripping in our
earlier studies (I). The presence of the major alcohols, methanol
and ethanol, were also verified in the earlier work and additional
identification of trace components has been made by Bethge and
Ehrenborg (2). The major ketone acetone has long been known to be
present in condensates, and the next most important ketone quantitat-
ively, 4-methyl-2-pentanone, had been identified earlier by infra-red
methods (I). Ketones found in trace quantities and some additional
trace compounds are reported in the present study. Of the phenolic
compounds, guiacol is the main compound found and was reported earlier
(4). The remaining phenolics have been verified in our studies here
and by others (4, 5). The terpenes in condensate streams were
identified in our prior studies (I) and have been verified in detail
in more recent work (6). Numerous other organic compounds are present
in trace amounts.
These compounds have been isolated by steam distillation followed by
separation from water by solvent extraction or fractionation and
sometimes by selective chemical reactions. In earlier work infra-red
spectroscopy was used for identification and this has been replaced
by use of gas chromatography-mass spectrometry in more recent studies.
Details are presented in the references listed (2, 3, 6).
The formation of these compounds will be discussed later. At this
point it is useful to recognize that most of these classes of compounds,
i.e., sulfur bearing compounds, alcohols, ketones, and phenols are the
result of wood treatment and especially of reactions occurring during
the pulping process and will be found in all kraft mills. Only the
terpenes are characteristic of the wood species being pulped. Thus
the main qualitative differences will occur in this class of compounds.
Quantitative differences are more a function of process operation and
equipment design and will also occur.
II
-------
Concentration Ranges of Steam Volatile Compounds
Concentration ranges for sulfur bearing compounds in condensate
streams have been well documented in the literature and were not
studied (7). These ranges are methyl mercaptan (2-250 ppm), dimethyl
sulfide (10-800) and dimethyl disulfide (2-140).
The concentration of methanol, ethanol and acetone in condensate
streams from the pulping of pine and birch has been reported from
various mil Is and these values taken from the literature are
summarized in Table II. In order to estimate the total methanol,
etc., produced, the flow rates of the individual condensate streams
must be known. These were unavailable and therefore a calculation
was made based on the estimated quantities of 232, 2030 and 12,775
Ibs of water per air dried ton (ADT) of pulp for the digester relief,
blow, and evaporator condensates respectively. These represent
typical values from U.S. kraft mills (II). The total methanol yields
resulting from the concentrations reported in Table II are then about
M to 16 Ibs methanol/ADT from pine and about II Ibs methanol/ADT
from birch. Estimates of total ethanol and acetone are also presented
in Tab Ie II.
Methanol Content of Kraft Mill Condensate Streams
Methanol concentration was determined for condensate samples from a
mill pulping several wood mixes in order to provide further information
on the effect of wood species and on the total amount of methanol
formed (12). This study was limited to methanol since its quantity is
an order of magnitude greater than the next most abundant compound
(ethanol) and since it makes the greatest single contribution to
BOD (9). The concentration of methanol, based on the same material
balance used for calculating the data in Table II, are summarized in
Table III. The data show that the methanol content of digester relief
12
-------
TABLE I I
Estimated Composition of Kraft Mill Streams
Wood Species Pulped
Condense Stream ?£ B
Digester Rel ief
Methane 1 ppm
(Ib/ADT)*
Ethanol ppm
(Ib/ADT)*
Acetone ppm
(Ib/ADT)*
Blow Gas
Methanol ppm
(Ib/ADT)*
Ethano! ppm
(Ib/ADT)*
Acetone ppm
( Ib/ADT)*
Evaporators
Methanol ppm
(Ib/ADT)*
Ethanol ppm
(Ib/ADT)*
Acetone ppm
(Ib/ADT)*
Total Condensates
Methanol Ib/ADT
Ethanol Ib/ADT
Acetone Ib/ADT
5900-7500
( 1.4-1.7)
700- 1 500
(0.2-0.4)
60-210
(0.02-0.05)
390-960
(0.8-1.9)
60-670
(0.1-1.4)
10-60
(0.02-0.1)
725-960
(9.3-12.3)
50-60
(0.6-0.8)
2-12
(0.03-0.2)
1 1.5-15.9
0.9-2.6
0.07-0.4
7100
(1 .6)
1600
(0.4)
120
(0.03)
525
(I.I)
25
(0.05)
5
(0.01)
625
(8.0)
15
(0.2)
5
(0.06)
10.7
0.7
0.10
*Conversfon from parts per million to pounds per air dried ton of pulp
is based upon a nominal material balance giving 232 Ib/ADT relief
condensate, 2030 Ib/ADT blow condensate, and 12,775 Ib/ADT evaporator
condensate (II).
13
-------
TABLE I I I
Methane I Content of Some Condensate Streams from a
Pacific Northwest Kraft Mill
Condensate
Stream
Alder-Douglas fir Cedar-Douglas fir Douglas fir
(=4:1) (=4:1) (\00%)
Digester Re! ief
Condensate
methanol ppm
Ib MeOH/ADT
Blow Gas
Condensate
methanol ppm
Ib MeOH/ADT
Evaporator
Condensate
methanol ppm
Ib MeOH/ADT
Total Methanol
Ib/ADT
8400 ± 400 [2]* 2900 ± 100 [2] 2800 ± 400 [2]
2.0** 0.7 0.7
2000 ± 600 [3]
4.1
410 ± 20
5.2
11.3
1800 ± 100 [3] 2100 ± 300
3.7 4.3
410 ± 20 [2]
5.2
9.6
280 ± 20
3.6
8.6
*Methanol concentrations are reported in ppm (or mg/liter) together
with the approximate spread of data and the number of samples upon
which the result is based.
**Flow quantities were not available from the pulp mill. The ppm
values were converted to a total weight basis (Ibs/ton of air dried
pulp) by using a typical balance of 232, 2,030 and 12,775 Ibs of
water/ADT of pulp for digester relief condensate, blow gas condensate,
and evaporator condensate respectively (II).
14
-------
from alder cooks is similar to values from birch cited in Table II,
while prior values for pine are greater by a factor of two than values
reported here for Western red cedar and Douglas fir.
Blow gas condensate concentrations and amounts in Table Ml are
uniformly higher than previously reported values of Table II, while
evaporator condensate values are lower. This probably reflects a
difference in the process conditions relative to blow condensate.
Differences in extent of water recycle and direct contact evaporators,
for instance, are not taken into account in Tables II and III, and
these would obviously effect concentration. The total methanol
content of all condensates are of comparable magnitude in Table II
and III, of about 8 to 12 Ib/ADT.
The alcohols are quantitatively very important among the organic com-
pounds in mill condensates. Methanol is the main alcohol found and is
also the main organic compound, often reported in concentrations as
high as 0.5% in digester relief condensates. It is extremely important
in overall mill effluent BOD.
Effect of Wood Species on Methanol Formation
The effect of pulpwood species on the amount of methanol formed in
kraft pulping was studied in more detail by carrying out laboratory
cooks. Mill samples are unsuitable for this purpose for several
reasons, among which are the unavailability of condensates from
pulping of a single wood species, different cooking cycles with
different pulpwood mixes and difficulties in determining adequate
material balances on condensates. The results from the laboratory
cooks on the four woods are summarized in Table IV which also includes
results for ethanol and acetone. There are significant differences
in the total amount of methanol formed, ranging from about 28 Ibs/ADT
from the hardwood Red alder to about 14 Ibs/ADT for Douglas fir. The
trend in the quantity of methanol formed is similar to the total
15
-------
methoxyl content of the wood, which is reasonable since most of the
methanol is formed with 4-0-methyl glucuronic acid residues, typically
greater in hardwoods.
TABLE IV
Total Yields of Methanol, Ethanol and
Acetone from Laboratory Kraft Cooks
Wood Species
Red Alder
Western red cedar
Western hemlock
Douglas fir
Methoxy 1
Content, %
( + )
5.58
5.0
4.42
Methanol
27.8*
26.2
15.6
14.4
Ethanol
1.7*
1.4
1.7
2.3
Acetone
0.3*
0.4
0.3
0.4
* All values are in Ibs/ADT of pulp
+ Not available but considered higher than softwood methoxyl content
Ethanol does not show a trend which can be related to wood species,
which is expected since it is formed primarily by fermentation in the
wood after felling, and is thus dependent on chip pretreatment (13).
Generally, the ethanol forms rapidly under anaerobic conditions,
reaching a maximum In a week or so. The reaction is slower under
aerobic conditions. Ethanol is lost fairly rapidly on storage.
Acetone, the origin of which is not well defined, also shows no trend
with wood species. The ethanol and acetone by-products are present
in much smaller quantities than methanol.
Comparison of Methanol, Ethanol and Acetone Yields with Commercial
Results
The trends in the total amount of methanol formed from different wood
species in mill pulping agrees with that found in laboratory cooks.
The order of magnitude of the differences is, however, not as great as
16
-------
expected based upon the laboratory cooks. This may be explained in
part by the fact that the mill samples result from pulping of mixtures.
The estimated wood ratio used for the particular samples, shown in
Table Ml, will tend to decrease the magnitude of the species
differences. A more important factor is likely to be the difference
in pulping conditions used for softwoods versus hardwoods.
The reaction forming methanol is
R-OCH3 + NaOH •*• R-ONa + CH3OH Eq.l
and probably is a second order 3.2 type of reaction. The rate
expression for forming methanol would then be
Rate = k [R-OCH ] [OH~3 Eq.2
The rate constant has not been determined. However a brief study of
the rate of formation of methanol, ethanol and acetone was made by
pulping Douglas fir and determining the cumulative content of these
compounds at several time intervals during the cooking cycle. The
results are shown in Figure I. The curve for methanol formation from
wood is similar to the results from hemicellulose reported by
Clayton (14). Methanol formation is still occurring at the end of a
normal cooking time and shorter cooking cycles will result in less
methanol being formed. Lower temperatures and lower pulping pH also
have the same effect. In commercial pulping of hardwoods all of
these cooking parameters are reduced, resulting in less methanol
being formed despite the higher potential with hardwoods. The
ethanol concentration remains constant after about one hour of
reaction time in agreement with a fermentation route of formation prior
to the cook as the main source of this compound. Acetone, present in
low concentration, is actually declining in concentration after the
first hour of reaction time, indicating loss, in all probability, by a
condensation reaction.
17
-------
20
15
3
O-
o 10
<
v^
(fl
0
metHanoi
ethanol
acetone
30 60 90
Time in Minutes
120
FIGURE 1. Formation of Methanol, Ethanol and Acetone in Laboratory
Kraft Cooks.
17°C, 20* active alkali, 25% sulfidity, 6/1 1iquor/wood
18
-------
The difference in absolute magnitude between the methanol quantities
found in laboratory cooks and in mill streams may be due to several
factors, the most important of which is the significant difference in
pulping conditions. More complete recovery of volatiles possible
from the laboratory black liquor and departures of industrial process-
ing from the assumed material balance may also be important. The
large differences involved do suggest estimates of the methanol content
of kraft process streams are perhaps generally conservative.
Turpentine Concentration
Condensate water is often contaminated with turpentine. Usually
crude sulfate turpentine is collected from digester relief gas by
condensation and turpentine separation in a decanter. The decanter
underflow contains the greatest concentration of turpentine of any
stream in the process. There are many other turpentine containing
streams in a mi I I and large variations in concentration can occur in
these streams, dependent on design and especially on operation
variables.
A survey of turpentine concentrations was undertaken in which a variety
of water samples were collected from four mills in the Pacific North-
west and British Columbia. Usually a number of samples were collected
from the same source over a month's operating time so that what might
be called normal operating ranges could be established. Actual
sampling sites are shown in Figure 2, which is a composite schematic
of the essential features of the mills sampled. Both batch and
continuous systems were included in the study with the continuous
operation systems -including latest technology high turpentine yielding
systems (16, 17).
Sample sites I, 2 and 3 in Figure 2 are for batch processes and
represent digester relief, blow and evaporator condensate; sample were
actually taken from the turpentine decanter underflow, the heat
recovery accumulator, and evaporator condensate respectively.
19
-------
NJ
O
Condenser
Decanter
Condenser
15) Sewer
Decanter
(4) Underflow
>\
Batch
digester
*o
Blow tank
Eva|
Hot water accumulator
Underflow
Y
Washers
Kamyr
—*to furnace
'Condensate
] Steaming vessel
Flash
tank
1
Rash
tank
2
Waste heat
recovery
Flash tank
i
Condensate
Blow tank
-» —
Evaporator Concentration
condensate condensate
C?) (§)
Condenser
Hot well
decanter
Hot well decanter
Underflow
Figure 2. Sample sites for batch and continuous processes
-------
Sample sites 4 and 5 include condensate originating from the #1 flash
tank of a Kamyr system via the steaming vessel. Sample #4 is from a
turpentine decanter underflow and #5 is condensate from a system
where heat recovery is practiced but no turpentine is recovered.
Samples 6, 7, 8 and 9 originated from the further processing of black
liquor. Samples 6 and 7 were from early stages in the evaporators;
8 was from the final stages of evaporation and 9 was a decanter
underflow sample. Many interconnections exist in the system and
changes in operating variables may be expected to greatly influence
the concentration of turpentine in the several streams.
The results of this survey are summarized in Table V which lists con-
centrations of turpentine found in the various streams and also the
total turpentine represented by this stream in Ibs/ADT of pulp
produced. The latter figure is calculated from the concentration
and from the flow rate of the stream. We were able to obtain estimates
from mill personnel of flow rates for each of the streams sampled
and these along with the NCASI survey averages (II) are listed in
Table VI. Further details of survey techniques used may be found in
reference (15).
Batch Systems
The turpentine concentration has been determined in a number of indi-
vidual samples of decanter underflow water obtained from four different
mills. Each mill was fairly consistent, but very large differences
were found between mills. The highest concentration observed was
about 5400 ppm and all samples from this mill averaged 4500 ppm.
A second mill averaged about 1500 ppm and two other mills averaged
about 435 ppm. These mills all pulp similar wood mixes including
Western Hemlock, Douglas Fir, and some Ponderosa pine plus some other
species. Most pine is pulped at Mill A, (-25%) which has the highest
underflow concentration. The ten-fold difference found between these
mills cannot be explained based only on species differences and must
-------
TABLE V
Concentration and Total Turpentine in Kraft Mill Condensate Streams
Sample Mill Designation
No.* Process Stream A BCD
I Batch Decanter Underflow 4514 ± 950 (4)** 1490 ± 300 (4) 430 ± 90 (4) 444 ± 5 (2)
Turpentine, ppm (Ib/ADT) (2.02) (.37)
2 Hot Water Accumulator 665 (I) 208 (I) -
Turpentine, ppm (Ib/ADT) (2.98) (.63)
3 Evaporator Condensate 106 (I) 23
Turpentine, ppm (Ib/ADT) (.50) (.31)
4 Steaming Vessel Decanter 580 ± 155 (3)
Underflow (.23)
Turpentine, ppm (Ib/ADT)
5 Sewered Steaming Vessel 1493 (I)
Condensate (1.39)
Turpentine, ppm (Ib/ADT)
6 #2 Flash Tank Condensate 105 (I)
Turpentine, ppm (Ib/ADT) (.33)
7 Evaporation Condensate 146 (I)
(Kamyr) (2.49)
Turpentine, ppm (Ib/ADT)
8 Concentration Condensate 307 (I)
Turpentine, ppm (Ib/ADT) (.61)
9 HotwelI Decanter Under- 616 ± 350 (3)
flow (.04)
Turpentine, ppm (Ib/ADT)
*Sample sites shown in Figure 2
**Spread of data and number of samples analyzed
-------
TABLE VI
Flow of Condensate Streams in the Kraft
Pulp Mill Process, (Results in Ib/ADT)*
Sample MI I I Designation
No. A B C D NCASI
I Batch Decanter 448 253 230 230 232
Underflow
2 Hot Water 4,482 3,023 2,030
Accumulator
3 Evaporator 14,061 13,300 — 12,775
Condensate
4 Steaming Vessel 398 —
Decanter Underflow
5 Sewered Steaming Vessel 930
Condensate
6 #2 Flash Tank 3,186
Condensate
7 Evaporator Con- 17,074 —
densate (Kamyr)
8 Concentrator 1,991 —
Condensate
9 Hotwell Decanter 64
Underflow
'Flow data are based on average values from the mills involved
23
-------
be due to differences In equipment and operating practice. We were
able to obtain straight line correlations between sodium concentration
and turpentine concentration on batch decanter underflow, supporting
the idea of turpentine being emulsified by black liquor carryover.
Losses of turpentine via the underflow can be significant; e.g. mill
A may lose up to 2 Ibs/ADT in this stream.
As expected, the concentrations of turpentine found in blow condensate
are lower than those found in decanter underflow. Values of 665 and
208 ppm were found in two mills in the accumulater water. The amount
of turpentine found here will vary due to differences in digester
steaming practices and equipment design. For example, one mill
studied uses a vaporshere for control of noncondensables from the blow,
and condensate from this system is returned to the digester relief
turpentine decanter instead of to the hot water accumulater. On the
other hand, another mill was operating at reduced production rates
and was doing less relief gas venting, thus increasing the amount of
turpentine in the accumulator water. Samples of evaporator condensate
generally contain very little turpentine; 106 and 23 ppm were found in
two samples.
The total turpentine found in mills A and B in these three condensate
streams has been calculated, Table V, and is 5.5 Ibs/ADT in mill A
and 1.31 Ibs/ADT in mill B. It is important to recognize that these
values depend on accurate flow measurement which is difficult in an
operating mill. Referring to Table VI, the flow values reported for
the various streams In mill A are generally much higher than the
average values found in surveys done by the NCASI (II)-. Mi 11 A and
also mi 11 B tend to have high rates of condensate flow, and this may be
characteristic of Western mills as opposed to Southeastern mills.
It is apparent that the blow condensate is the more important stream
in total turpentine quantity despite the lower concentration. If both
decanter underflow and blow condensate were effectively steam stripped,
24
-------
75 to 90% of the turpentine lost could be recovered. Recovery of this
turpentine by air stripping is not feasible.
Continuous Processes
Decanter underflow, originating from the #1 flash tank and steaming
vessel condensate, was found in one mill to contain 580 ppm turpentine.
In a second mill, heat Is recovered from this stream but a turpentine
decanter is not used and the condensate was found to contain 1493 ppm
turpentine. In terms of total turpentine, the decanter underflow
represents a negligible 0.23 Ibs/ADT, while the unseparated condensate
represents 1.39 Ibs/ADT of turpentine.
In the continuous process, only a part of the available turpentine is
removed from the black liquor via the steaming vessel and recovery of
the remainder requires processing of a number of streams including
washer vents (17). Whatever turpentine is available from these vents
and from the blow tank is returned to the steaming vessel decanter
in the mill studied. Remaining turpentine in the black liquor is
routed to a separate decanter, condensate here originates from the
#2 flash tank, several points in the evaporators and other sources.
This decanter underflow contained 616 ppm turpentine, and this value
combined with the low flow rate shows very low toss of turpentine via
this route.
Samples of condensates from the first evaporator stages were analyzed
and these show relatively low turpentine concentrations in the 100 to
150 ppm range. The flow rates are high here, especially on Sample 7
and an appreciable amount of turpentine is lost in this condensate.
Concentrator condensate, Sample 8, analyzed 307 ppm, and another
high flow rate is reported for this stream. Altogether, the evaporator
condensates contain about 2.4 Ib/ADT of turpentine, and thus are about
as important in turpentine losses as batch blow condensate and decanter
underflow. Comparisons of this kind are not as meaningful as would be
25
-------
desired because of the extensive recycle and reuse of condensate water,
which are factors not considered in detail in this study.
Origin L_pf Steam Volatile Organic Compounds in Kraft MF I i Streams
Understanding of the origin of the organic compounds of interest may
lead to ways of controlling their formation and release. Of the main
classes identified (Table I) methyl mercaptan and its derivatives are
the most objectionable. The chemistry of their formation and physical
phenomena related to their release has been extensively studied and
is reviewed elsewhere (7). Briefly, methyl mercaptan is formed by a
reaction between lignin methoxyl groups and hydrosulfide Ion (Eq. 3),
dimethyl sulfide is formed by a similar reaction (Eq. 4) and the
disulfide is an oxidation product (Eq. 5).
Lig-O-Me + SH~ - > Lig 0* + MeSH Eq. 3
Llg-O-Me + MeS~ - > Lig 0* + MeSME Eq. 4
2MeS" + [02] - > MeSSMe Eq. 5
Methanol is believed to arise via alkaline hydrolysis of 4-0-methyl
glucuronic acid residues in hemicel lulose (Eq. 6), as discussed earlier.
OH' - > + Me OH
It has now been established that ethanol is present in wood chips prior
to pulping (13). Apparently anaerobic conditions develop in the log
as water transport Is disrupted and some ethanol is formed via
glycolysis as this condition develops. Usually the ethanol can be
detected a week or two after fel ling of the tree (18). Some ethanol
will be lost by volati I izat ion from the chip and the remainder wi I I
then be steam distilled off during pulping. Several higher alcohols
have been found in trace amounts (2), these alcohols are for the most
part normal components of fusel oil. This kind of alcohol is derived
from ami no acids, or their precursors, by a process of deam! nation,
decarboxy I at ion and reduction, Eq. 7 a, b, c.
26
-------
0 0
R-CHCO.H + RC-CCLH ,, transaminase ^ R-c-m H + R-CH-CCLH
, £ z V ^ t ^
Eq. 7a
0
R-C-COM decarb0xYlaSe> RCHQ +
Z Z tq. /b
RCHO + NADH + H
alcohol
,+ dehydrogenase ^ DPU .u . ,.,n+ Eq. 7c
"Fusel oil"
Component
All of the ketones reported from kraft mill streams are methyl ketones,
CH^CO-R, which are formed by air oxidation of extractives, followed by
decomposition of the extractive hydroperoxide to an a, 6-unsaturated
ketone, which in turn may undergo .a reversed aldol reaction under the
kraft pulping conditions, Eq. 8.
p a'r ^ Extractive decomposition ^ „
oxidatiolT Hydroperoxide w R-C-CH=CH-R
Eq. 8
0 0
11 * re\/pr«;fiH "
R-C-ONCHR .7; > R-C-CH, + other products
aldol 3
For the phenolic compounds, guaiacol is the main compound reported in
condensate streams. Others indicated as present are phenol, syringol,
cresols, vanillin and acetonvani I lone. Guaiacol and syringol are
more readily volatile with steam and usually are found in evaporator
condensates. They are probably formed from structural elements in
lignin having free phenolic and a-hydroxyl groups via a reversed
Lederer-Manasse reaction, Eq. 9.
H-4-
x^
OH H-C-OH H"9,>H
Tli OMe if OMe ^5i OMe lOMe
10 0 iO OH Eq. 9
27
-------
Guaiacol and related phenolIcs such as the cresols have not received
the attention they deserve in control of odor as well as toxicity.
Most people recognize that the overall odor of a kraft mill is
"different" from that of the well known sulfur bearing compounds, and
one can usually distinguish a "burned or typical" odor near the mill.
This is probably due to phenols in the evaporator condensate (5).
Guaiacol is the main phenolic component present in this stream,
and using the data of Marvel I and Wiman (4) a concentration of about
10 ppm may be estimated for evaporator condensate from the pulping of
Douglas fir. Since the odor threshold for guaiacol dissolved in
water is reported to be 0.021 ppm (19), guaiacol should be readily
detectable. Paracresol, probably present in much lower concentration,
since it is barely detectable, has an odor threshold value of 0.001 ppm
in air and probably is also important (20). These compounds are
described as having burned, smoky, phenolic, tar-like or pungent odors.
Chlorination of condensate streams is often used as a means of odor
control. This is usually done by mixing bleach plant effluent with
the condensate stream, This treatment may in fact give rise to
chlorinated compounds by reaction with guaiacoi (21). It has recently
been reported that tetrachloro-o-benzoquinone and related chloro-
catechols, which may be derived from guaiacol, are the principal
toxic components in bleached kraft chIorination effluent (22).
Control of the release of these phenolic compounds may be achieved by
using high alkalinity in the weak black liquor. All of these phenolics
have a pK of 10.0-10.5 so that their vapor pressure will be a function
a
of the pH of the system, very similar to the situation reported for
methyl mercaptan (23). Higher alkalinity will reduce the amount of
phenols in the condensate streams. Conversely, steam stripping of
condensates at lower pH values will aid in transferring the phenols
to the distiI late.
28
-------
Several terpenes are listed in Table I. These compounds as a class
are usually not considered in effluent problems since they are for
the most part collected and marketed. However, they do appear in
condensate streams, especially the digester relief condensate. This
is due mainly to incomplete phase separation in the turpentine
decantors which are never 100/K effective. In general, the terpenes
found in the condensate are similar to those present in the wood
before pulping. However, comparisons of turpentine from Douglas fir
wood with sulfate turpentine indicates the amount of a- and 6-pinene
decreases, while limonene as well as several terpene alcohols increase
during the pulping process. The changes are summarized in equation 10.
A,
1$)
ct-pinene
6-pinene limonene a-terpineol Fenchyl
alcohol
borneol
Eq. 10
Other degraded terpenes i.e. methyl furan and 4-(p-toly I )pentanol-l
are formed by more involved reaction pathways.
29
-------
SECTION VI
PROCESS DESIGN STUDIES
General Considerations
Prior studies have been reported from these laboratories (24, 25)
t
utilizing steam stripping to reduce terpene and sulfur compound loads
in kraft mil I condensate streams. These studies showed that by steam
stripping, reusable hot water can be obtained, or the odor levels and
oxygen demand can be greatly reduced in water to be discharged from
kraft recovery systems.
There remain associated with these operations, however, several
disadvantages which it would be desirable to minimize. These include
(I) the recycling of terpene and sulfur compounds back through the
digester to increase concentrations if "sour" condensates are utilized
for the steaming of chips as in the Kamyr system; (2) the gradual loss
or escape at numerous points in the recovery system, of terpenes and
odorous sulfur constituents, as black liquor is concentrated; (3) if
air oxidation of weak black liquor is utilized to minimize the above
losses, the air effluent from the oxidation system is itself a source
of odorous gases.
It is becoming increasingly clear that reduction of water contamination
by transferring the burden to air effluent streams is not a satisfactory
solution of such problems. Therefore, a modification of the steam
stripping process was sought to minimize the problems enumerated above.
The concept discussed in the following pages involves the steam
stripping of black liquor effluent from a continuous digester while it
is still under pressure, in a multi-stage stripping column. The
resulting overhead steam from the column, at about 77 psia, has
sufficient heating value to be reused as a stripping agent or indirect
heat source for steam stripping of various mill condensates. It would
-------
contain appreciable quantities of odorous constituents and terpenes,
however; and its utilization in this manner requires careful
consideration.
Since the completion of these prior steam stripping studies, our
knowledge of the volatile constituents of kraft black liquor has
improved considerably, as already reported, and the vapor-liquid
equilibrium distribution of some of these has also been clarified (23,
26), allowing for a more complete and more sound evaluation of the
overall process. In addition, a study has been completed of the
expected release of volatile terpenes, terpineols, and sulfur
compounds from kraft recovery systems (27). These results, together
with literature sources which have become available since initiating
this work, were utilized in the following report of design studies
of combined steam stripping of kraft mill recovery system streams.
The long term objective of the study is to find the optimum combination
of steam stripping column or columns so as to minimize combined capital
and operating costs, taking advantage of marketable by-products to
decrease the overall costs of controlling effluent water contamination
and air pollution.
Steam Stripping of Black Liquor
The steam stripping of black liquor as it leaves the digester would be
expected to reduce immediately the concentration of the most volatile
constituents, the terpenes, terpineols, methyl sulfide, and, if the pH
of the black liquor is 12.0 or below, of methyl mercaptan. These
volatiles would, therefore, not be appreciably present through the one
or more flash stages, dilute liquor oxidation and multi-effect
evaporation. Odors from leakage, etc., during these operations would
be decreased, and the resulting condensates would contain fewer
volatile constituents, and lower biochemical oxygen demand. In addition,
the terpenes should be more amenable to recovery for sale, if this
is desired, by isolating the condensate of black liquor stripping from
other condensates and controlling the amount of steam used for this
32
-------
purpose. The study of actual terpene content of various streams of
existing kraft mills. Section V of this report, illustrates that in
the absence of black liquor stripping, turpentine appears in the second
flash stage in continuous digester systems and to some extent, in
evaporator condensates as well.
Since methanol was not considered in the prior design studies, an
important purpose of the present calculations is to estimate its
distribution during black liquor evaporation.
Continuous Kamyr System Stripping Processes
The nature and quantity of terpenes and terpineols present in kraft
black liquor is dependent mainly upon the wood species and condition.
Methanol, methyl mercaptan and methyl sulfide yield are governed by the
methoxyl content of wood lignin and process kinetics. The assumptions
and procedures of Tsuchiya and Johanson (27) were adopted for this
study. These were modified as necessary to incorporate the possibility
for steam stripping of black liquor ("SEKOR stripping column B"), of
high-odor condensates ("SEKOR stripping column A"), and of low-odor
condensates ("SEKOR stripping column C"). A flow sheet illustrating
these possibilities is shown as Figure 3, which incorporates heat
recovery and disposal features.
Published material and energy balance information (28) for a continuous
Kamyr digester, washers, and flash system producing 550 ADT/D of pulp
was combined with additional information on concentrating black liquor
by evaporation (29,30) to give a general basis for calculation of
the removal of steam volatile substances from black liquor by steam
stripping. Pertinent data are shown in Table VII.
33
-------
Steaming Vessel
Column "B" Overhead Condensate
Black
To Flash Tank
FIGURE 3. Flow System For Combined Steam Stripping Study Utilizing Heat From Black Liquor Stripping.
-------
TABLE VII
Material and Energy Balance for Kamyr Digester,
Washers and Flash System
Digester Capacity 550 ADT/D
Black Liquor Flow Rate 481,000 Ib/hr
Dissolved Solids 73,200 Ib/hr
Per Cent Solids 15.256
B. L. Temperature 3I5°F
The major influence of wood species is in the nature and amount of
terpenes and in the methanol content. Terpene recovery is estimated
arbitrarily for a Western U.S. conifer yielding an assumed average of
1.0 gal terpene per air dried ton of pulp (31).
The methanol yield Is based on an average yield of 2.38 gal/ADT or
15.6 Ib/ADT as found for Douglas Fir (12), a low methanol-yield!ng
species, and has been superimposed on the original digester effluent
stream material balance for this calculation.
Prediction of Stripping Column Performance in Continuous Kraft Mill
Recovery Systems
Removal of 905? and 99% of methyl mercaptan from the black liquor
digester effluent by steam stripping were chosen for design bases.
Methyl mercaptan has an offensive odor which makes its removal from
black liquor highly desirable. Upon oxidation, any methyl mercaptan
remaining will be converted to dimethyl dtsulfide which also has an
objectionable odor. At a given alkaline pH, the vapor pressure of
methyl mercaptan is much higher than that of hydrogen sulfide.
However, the hydrosulfide ion concentration in black liquor is so high
that larger absolute amounts of hydrogen sulfide will ordinarily be
released.
35
-------
After selection of methyl mercaptan removed as the design basis, the
next variable to be considered was the feed to steam ratio, or liquid
to gas ratio CL/G). The minimum steam rate corresponds to equilibrium
conditions at the rich end of the stripping column (32). Equilibrium
data from work done previously at the University of Washington (23, 26)
were used for hydrogen sulfide and methyl mercaptan. Two pH levels,
II and 12, were considered as being representative for black liquor
from a continuous digester.
The stripping towers designated as Columns A, B and C in Figure 3, are
two-phase continuous contacting devices in which volatile components
within the liquid feed are transferred to a counter-current gas phase
which, in this case, is steam. For dilute solutions with no chemical
reaction occurring, the number of transfer units required for
separation is proportional to the logarithmic mean driving force. The
black liquor stripping tower "B" has been placed after the digester
and before the first flash tank in the process flow sheet. This
allows steam stripping under pressure to be utilized, which makes
the column overhead stream valuable as a heat medium for columns "A"
and "C".
Column height requirements were determined for a range of feed-to-steam
ratios of practical interest 20 < (L/G) < 55. At the lower end of
this range, steam costs are high and column height requirements are
minimized. At the upper end of the range, steam costs are reduced,
but column height requirements increase. Column height is evaluated
in terms of the theoretical "number of transfer units" (Ntu) (33).
For many common industrial column designs, a transfer unit corresponds
to about two actual plates in a plate-type tower, or to about 3 to 6
feet of packing in a packed-type tower.
Results summarized in Table VIII show that the removal of methyl
mercaptan is directly dependent upon the pH of the black liquor, and
36
-------
the column height or the number of transfer units within the stripping
column, and the feed to steam or L/G ratio. Dissociation of both
methyl mercaptan and hydrogen sulfide increases with increasing pH,
rendering them less volatile. For example, for feed steam ratio of
30:1, 90% removal of methyl mercaptan at pH 12 or 99% removal of
methyl mercaptan at pH II each require 5.3 transfer unfts. Or for the
same feed steam ratio and the same level of methyl mercaptan removal,
stripping black liquor at pH 12 requires at least double the number
of transfer units required for pH II. At a constant feed to steam
ratio of 30:1 and pH 12, increasing the desired removal of methyl
mercaptan from 90$ to 99% increases the number of transfer units from
5.3 to 19.1. The capital costs of the column would be almost
directly proportional to these numbers. The study, therefore, suggests,
based upon the removal of methyl mercaptan, that the pH and feed
steam ratio both must be low to minimize capital costs. However, a low
feed steam ratio is reflected in an increased steam cost, and a more
detailed optimization study would be required to select the actual
best conditions.
Utilizing the combinations of per cent removal of methyl mercaptan
(90% and 99%), feed to steam ratio (L/G), and column height (Ntu)
illustrated in Table VIII, the calculations were extended to include
the other major volatile components of interest, hydrogen sulfide,
methyl sulfide, methanol, and terpenes. Concentrations of these
constituents to be expected in the digester effluent black liquor were
known from the prior prediction study (27). Table IX summarizes the
results obtained.
37
-------
TABLE VIII
Tower Height and Process Steam Requirements
for Methyl Mercaptan Removal by
Steam Stripping of Black Liquor
Transfer Units Required for
L/G*
20
25
30
35
45
55
Steam Cost**
-------
Hydrogen sulfide is the most concentrated odorous compound in black
liquor and like methyl mercaptan, is also an ionizable substance which
shows increasing dissociation with increasing pH. As shown in Table IX,
the removal of hydrogen sulfide from black liquor by steam stripping
is very dependent upon both pH and feed to steam ratio. For steam
rates under consideration in this study, the removal of hydrogen
sulfide does not exceed \2% for pH I I or 2.5% for pH 12. Thus, steam
stripping of black liquor will not appreciably decrease H?S levels in
condensate streams from flash tanks and evaporators.
Published vapor equilibrium data for methanol-water (34) indicate that
the vaporization equilibrium constant at very low concentrations is
about 8, independent of pH and may be assumed Independent of ionic
strength, The methanol removal, then, depends mainly on the feed to
steam ratio. The removal of methanol for flow rates under consideration
varies from 22 to 36%. Thus, additional removal of methanol further
along the black liquor concentration path may be desirable.
Methyl sulfide and terpenes, which have much higher vaporization
equilibrium constants (25) are more than 99/6 removed in the black
liquor stripping column for all cases under consideration.
Four cases from those presented in Tables VIII and IX were chosen
for further calculations to extend the predicted distribution of these
volatile constituents through the remainder of the black liquor
concentration process. These are summarized as Table X.
The distribution of volatile compounds between the vapor and liquid
phases of the flash tanks, Figure 4, was calculated as a continuous
equilibrium vaporization (35). In the process scheme used here, the
overhead product from the first flash tank is fed to the steaming
vessel to preheat chips before they enter the digester. The vapor-
liquid distribution between condensed steam and steaming vessel relief
was also treated as an equilibrium vaporization (36).
39
-------
o
O
3
I
M
/! \
Relief
c
I
Chips
I
Steaming
vessel
1
I '
2
% o
» JT
0 (ft
•o
1
o
IT
M
O
3
Rash
steam
Air out
I
Weak
liquor
oxidation
Liquor to
evaporators
Back wash
Figure 4. Flash concentration and oxidation steps in black liquor recovery (27),
-------
TABLE X
Operating Parameters of Black Liquor Stripping Column
for Cases Extended to Steam Stripping of Condensates
Case
1
Ma
Mb
III
Meth Merc,
Removal
90%
90%
99%
99%
Feed: Steam
ratio
45
30
30
25
Ntu
5.4
5.3
5.3
9.2
pH
1 1
12
1 1
12
Steam
Cost tf/ADT
22
33
33
39
Weak liquor oxidation at 0%, 10% and 95% efficiency was considered for
each of the four cases and was subject to the assumptions of reference
(27), namely, cocurrent or complete mix flow, and phase equilibrium.
For no oxidation, the larger amount of feed to the first two evaporators
(#5 and #6, in parallel) was compensated for by arbitrarily increasing
the amount of evaporation from these two units.
Calculations of vapor-liquid distribution show that in this flow scheme
the only compounds lost in significant quantity with the exit air
from the oxidation tower are hydrogen sulfide and methanol. To
minimize hLS emission from this source, high black liquor pH and
oxidation efficiency of 95% or above is recommended. The methanol loss
is independent of oxidation efficiency and pH and depends only on the
flow rate of air through the oxidation system and the entering methanol
concentration. If recovery of methanol is to be an economic considera-
tion, the fact that 25% to 30$ of the methanol from the digester is
lost during oxidation may be very significant.
The oxidized black liquor is subjected to multi-effect evaporation in
six stages, followed by direct contact evaporation in the flow system
model utilized here, as shown in Figure 5. Calculations were extended
through each of the six evaporator stages for each of the four cases,
41
-------
M
ro
o>
0)
•o
00
o
Stm.
ro
o>
*.
o
CO
Flue gas
OUT
r stm.
i
1A
V
direct
contact
evaporator
IB
N
IN)
T>
(A
o
\ /
Condensed steam
Flue gas in
Black liquor in
Concentrated black liquor
co
to
o
03
\ /
03
•3.
lO
\
00
o
ro -n
I
\ /
•»
_to 9
condenser
ca
o
r\>
->i_
a
Evaporator condensates
03
o
n
,1
y
f2
y
3,
r
X
\
Skirr
tanh
/
^
«
i
i
/
^
fi
¥
-/
\_
,5
Figure 5. Multieffect evaporation stages and direct contact evaporators (27)
-------
and for each of the three levels of oxidation of 0$, 7055 and 95%.
These results are summarized and compiled as Tables XI, XII, XIII, and
XIV.
Note that the tables show quantitatively much of what has been
indicated earlier in a qualitative way. For example, terpenes and
methyl sulfide are removed in the black liquor stripping column.
Methyl mercaptan is removed to the extent fixed by the design basis
for the case under consideration (9Q% to 99%}. Hydrogren sulfide
continues to be removed steadily throughout the evaporation process,
depending upon the oxidation level and pH. The release of H S with
the oxidation tower air effluent is decreased about fourfold with
pH change from II to 12 (131.7 vs 35,4, or 28.2 vs 7.1). A change
in oxidation level from 10% to 95% decreases this release approximately
fivefold (131.7 vs 28.2, or 35.3 vs 7.1). Methanol volatilizes
steadily but at a diminishing extent throughout the multi-effect
evaporator train, though after the first two parallel effects (#6 and
#5), only about \0% of the original methanol remains and this is
almost entirely evaporated in the remaining five stages, to appear in
the condensates. A small quantity of dimethyl disulfide is found in
the oxidation tower from the remaining methyl mercaptan. This is.
released with the air from the oxidation system and negligible quantities
should appear in subsequent condensates.
Column Requirements for Steam Stripping of Selected Condensates together
with Black Liquor Stripping
Energy balances show that the vapors from black liquor stripping are
sufficient to strip all condensates depicted in Figure 3 for the 550
ton-per-day pulp mill used as a model. This is shown as being done
through indirect heat exchange in a reboiler because of the contaminants
in this steam source. The resulting SEKOR B overhead condensate,
steaming vessel relief condensate, and No. 2 flash condensate have
similar levels of contaminants, and are fed to the stripping column
43
-------
Table XI
Summary of Volatile Component Distribution
Case I
S S
Source
Digester
it
Sekor B overhead
Sekor B bottoms
FT-I overhead
(to steam vessel
it
St.Vess.Relf.
FT-I bottoms
*
FT-2 overhead
FT-2 bottom*
Oxidation Level
X
Ox tower V
Ox tower L
Evaps
*
6-overhead
6-bottomj
*
5-overhead
5-bottoms
4-overhead
it-bottoms
ft
J -over head
3-bot corns
2 -overhead
2-bottoms
IB-overhead
IB-bottoms
lA-overhead
lA-bottoms
Si
OL J>
t •—
O
277
277
0
0
0
0
0
0
C L.
a "o.^
(I —
1—
35.8
35.8
0
0
0
0
0
0
H2S
[Ib.S,
1 hr J
3589
269
3320
405
2915
255
2660
0
709
621
497
833
445
1009
379
630
310
320
159
160
43
117
196
70
131.7
666.3
179
154
126
207
106
254
99.1
155.3
80.8
74.5
33-0
35.5
9.4
26.1
95
28.2
04.8
28
24
19.9
32.5
16.7
39.9
15.6
24.3
12.7
11.6
6.1
5.5
1.5
4.0
CHjSH [^r-l
21.30
19.17
2.13
1.75
.38
.30
.08
1.65
0
0392
0008
.039
.001
70
.022
.002
95
.0037
,0003
HeOH [Ib/hr]
357.7
78.7
279
91
188
46
142
47.9
0
56.7
14.3
53.7
17.3
16.7
14.9
8.9
6.0
4.1
1.9
1.3
0.6
.25
.35
70 S 95
107.8
34.2
12.8
4.3
11.7
5.4
5.1
4.6
2.8
1.8
1.2
0.6
0.4
0.2
.08
.12
[il.i
hr
2.7
2.7
(CH3)2S2l!^
0
70
.055
.001
95
.075
.001
All column quantities are In Ib/hr (terpenes and methano!) or Ib sulfur/hr (sulfur cpds).
Case t - pH»ll (L/G)-^5, 90* methyl mercan. removal, Ntu>5.4.
Streams amenable to steam stripping.
44
-------
Table XII
Summary of Volatile Component Distribution
Case I la
8
Source
digester
ft
Sekor B overhead
Sekor 6 bottoms
FT-1 overhead
(to steam vesse
*
St.Vess.ftelf.
FT-1 bottoms
ft
FT- 2 overhead
FT-2 bottoms
Oxidation Level
t
Ox tower V
Ox tower L
Evaps
6-overhead
6-bottoms
5-overhead
5-bottoms
ft -overhead
4-bottoms
3-overhead
3-bottoms
2-overhead
2-bottoms
IB-overhead
IB-bottoms
lA-overhead
IB-bottoms
Ji
277
277
0
0
)
0
0
0
0
c u
?~
35.8
35.8
0
0
0
0
0
0
u c Ib.S-i
H2S -FT"1
3582
57
3525
96
3429
54
3375
94.1
0
236
1452
166
1522
205
2769
269
2500
378
2122
357
1765
12.4
164.1
70
35 A
977.1
69.8
418. £
49.3
439.3
56.7
801.1
81.0
720.4
119.6
600.8
109.1
491.7
33.4
458.3
95
7.1
161.;
11. <
69.:
8.2
72.3
9.3
132.7
13.:
119.2
19.9
99.3
18.
81.2
5.6
75.6
HjSH t^—1
20.4
18.36
2.04
1.53
.51
.36
.15
1.44
0
.065
.010
.064
.011
70
.039
.006
.004
0
95
.0067
.0008
HeOH [Ib/hr]
357.7
89.7
268
87
181
45
136
45.8
0
54.4
13.6
51.5
16.5
15-9
14.2
8.5
5.7
3.9
1.8
1.3
0.5
.21
.29
70 & 95
103.4
32.6
12.3
4.0
11.3
5-0
4.8
4.2
2.5
1.7
1.2
0.5
.3
,2
.08
.12
3 i.
3.6
3.6
~0
-o
-o
-0
-o
-o
««3>2V^
0
0
0
0
0
0
0
0
0
70
.103
.002
95
.141
.002
All column quantities are In Ib/hr (terpenes S methanol) or Ib sulfur/hr (sulfur
compounds).
Case Ha: pH-12, L/G-30, 90% methyl mercaptan removal, Ntu-5.3
Streams amenable to stream strlPP;n9-
45
-------
Table XIII
iry of Volatile Component Distribution
Case lib
Source
Digester
Sekor B overhead
Sekor B bottoms
FT-1 overhead
(to steam vessel
*St.Vess.Re1f.
FT-1 bottoms
FT-2 overhead
FT-2 bottoms
Oxidation Level
t
Ox tower V
Ox tower L
Evaps
6-overhead
6-bottoms
*
5-overhead
5-bottoms
*.
M-overhead
4-bottoms
3-overhead
3-bottomi
2-overhead
2 -bottoms
1 6 -overhead
IB-bottoms
*
lA-overhead
lA-bottoms
|i
O- -O
n
277
277
0
0
0
0
0
0
c •-
8 t £
35.8
35.8
0
0
0
0
0
0
H S l-^-^\
2 hr
3589
406
3183
388
2795
245
2550
94.1
0
679
596
476
799
427
968
364
604
298
306
153
153
42
111
70
126.2
638.8
171.8
147.6
121.2
198.2
101.9
243.9
95.0
148.9
77.5
71.4
27.4
44.0
11.7
32.3
95
27.0
100.5
27.1
23.2
19.05
31.2
16.1
38.3
15.0
23.3
12.2
11.1
5.8
5.3
1.4
3.9
CH.SH [^N
21.3
21.1
.213
.175
.038
.030
.008
0.144
0
0039
0001
0039
0001
70
0022
0002
95
0003!
oooo:
HeOH [Ib/hr]
357.7
78.7
279
87
181
*5
136
40
9
54.4
13.6
51.5
16.5
15-9
14.2
8.5
5.7
3.9
1.8
1.3
.5
.21
.29
70 & 95
103.4
32.6
12.3
4.0
11.3
5.0
4.8
4.2
2.5
1.7
1.2
0.5
0.3
0.2
0.08
0.12
3'2a
HM
hr
2.7
2.7
-0
-0
-0
-0
-0
-0
All column quantities In Ib/hr (terpenes, met Hanoi) or Ib sulfur/hr (sulfur
322 hr
0
0
0
0
0
0
0
0
0
70
0055
.0001
95
.0075
.0001
compounds) .
Case lib - pH-11, L/G-30, 99* methyl mercaptan removed, Ntu-5.3.
*
Streams amenable to steam stripping.
46
-------
Table XIV
Summary of Volatile Component Distribution
Case III
Source
Digester
Sekor B overhead
Sekor B bottoms
FT-I overhead
(to (team vessel
*St.Vess.Relf.
FT-I bottoms
FT-2 overhead
FT-2 bottoms
Oxidation Level
t
Ox tower V
Ox tower L
Evaps
6-overhead
6-bottoms
5-overhead
5-bottoms
4-overhead
4-bottoms
3-overhead
3-bottoms
2-overhead
2-bottoms
IB-overhead
IB-bottoms
lA-overhead
lA-bottoms
§i
«— *v
a. A
0 ~"
277
277
-0
-0
-0
-0
-0
-0
c >~
£f
35.8
35.8
-0
-o
-0
-0
-0
-0
H,S l!4-£-]
i nr
3582
68
3514
96
3418
54
3364
94.1
0
237
1445
165
1517
204
2758
268
2490
376
2114
356
1758
11.5
164.3
70
35.3
973.9
69. 5
417.5
49.1
437.9
56.5
798.9
80.7
718.:
119.3
599. C
108.;
490.;
32 .(
458.:
95
7.1
161.1
11.5
69.1
8.2
72.4
9.2
132.3
13.4
118.9
19.8
99.1
17.9
81.2
5.6
75.6
CHjSH [' *\£ ]
20.4
20.2
.204
.153
.051
.036
.015
0.144
0
0065
0010
0064
0011
70
.003*
.oooe
95
0006
0000
HeOH [Ib/hr]
357.7
76
157
39
118
40
0
47.2
11.8
44.7
14.3
13.8
12.3
7.4
4.9
3.3
1.6
1.1
0.5
.04
.06
70 & 95
89.9
28.1
10.7
3.35
9.8
4.25
4.0
3.6
2.2
1.4
1.0
0.4
0.3
0.1
Hj)2i
Ib.S]
hr
3.6
3.6
-0
-0
-0
-0
-0
-0
"S'z8!1^1
0
0
0
0
0
0
0
0
0
70
0103
0002
0001
0
95
.0141
.0002
.0001
0
All column quantities expressed as Ib/hr (methanot, terpenes) or Ib sulfur/hr (sulfur
compounds).
Case Ml: pH-12. (L/O-25, 99* methyl mercaptan removed, Ntu-9.2
Streams amenable to steam stripping.
47
-------
SEKOR A shown. Optional removal of turpentine from Column B
condensate and steaming vessel relief condensate is also shown.
The amount and composition of the resultant overhead "concentrate" and
of the decontaminated condensate bottoms are shown in Table XV.
Quantities of volatile components shown in this table are reported as
pounds per air-dried ton of pulp. The table footnote relates this
unit of measurement to that of Tables XI to XIV. Column dimensions
required for 15% and 90% removal of remaining methanol content are
shown in Table XVI. Column conditions are as depicted in Figure 3.
TABLE XV
Case Ma - Composition of Feed and Product
Streams of SEKOR A Column
"Volatile Component in Ib/ADT** or Ib. sulfur/ADT
U/a+^r- Tar-nan^e l-l C PUI CLI f PU ^ C PU fM-l
Wo I ol I CI pt? It Co n«O L/n -*Dn V wil« / /*3 Lfll -f\J\\
Feed Stream
SEKOR B Overhead
Steaming Vessel
Rel ief
2nd Flash Overhead
Overhead Product
Stripped Condensate
700
200
780
5 to 25
1655
13.7
nil
nil
13.7
ni 1
2.5
4.2
2.35
9.0
ni 1
0.8
0.072
0.016
0.89
ni 1
0.16
nil
ni 1
0.16
nil
3.9
2.0
2.0
5.9
2.0
* For Fd/stm-5.0, 75^ MeOH removal of Table X
** For a 550 ADT per day pulp production Ib/hr (.0436) = Ib/ADT for
any constituent.
Evaporator condensates resulting from concentration of the stripped
black liquor also contain some contaminants, chiefly hydrogen sulfide
and methanol, the latter chiefly appearing during evaporation from 1756
to 22% solids, represented by the No. 5 and No. 6 evaporator condensates.
These condensates are shown in Figure 3 as feed to a separate stripping
column employing SEKOR B overhead as heat medium. Tables XVII and
XVIII summarize resulting overhead and bottoms concentrations, and
48
-------
required column dimensions for the particular case of Q% oxidation
level of case 11-a. This column and its required conditions are also
shown in Figure 3.
TABLE XVI
Case I la - SEKOR A Column Design for
Condensate Stripping
Fd/Stm
(2.4)min
2.8
5.0
10.0
2.5
4.2
8.3
Steam
Ib/hr
16,000
13,700
7,700
3,900
15,600
9,200
4,600
Reflux
Ratio
70.4
40.6
19.8
(I5.2)min
70.4
40.6
No. Ideal
Stages
4.0
4.7
5.0
6.3
Column
Diameter
(ft)
5.5
4.5
6.0
4.5
3.0
% MeOH
Remova 1
75
ii
M
11
90
IT
II
TABLE XVII
Case Ma - Composition of Feed and Product
Streams SEKOR Column C Stripping Evap. Condensates
Feed Stream
#5 and #6
Evap. Cond.
Overhead Product
Stripped Condensate
"Volatile
Water
4500
3 to 15
5055
Component
Terpenes
ni 1
ni 1
nil
in Ib/ADT or Ib. sulfur/ADT
HC PU CU 1 PU \ C
«;> Un-.5n von,/«o
17.5 0.006 nil
17.5 0.006 nil
nil nil nil
CH3OH
4.62
3.47
1.15
* For Fd/Stm = 8.0, 75* MeOH removal.
49
-------
TABLE XVIII
Case Ma - SEKOR Column C Design for
Evaporator Condensate Stripping
Fd/Stm
7.2
8.0
Steam
Ib/hr
14,300
12,800
Ret 1 ux
Ratio
12.8
11.5
No. Ideal
Stages
8.2
1 1.0
Col umn
Diameter
(ft)
5.5
5.0
% MeOH
Remova 1
75
75
Table XIX compares the total final condensate load, as pounds of vola-
tile contaminants per ADT of pulp, with the corresponding condensate
load if no black liquor stripping and no condensate stripping is
practiced. The latter information is from the previous study of
Tsuchiya and Johanson (27).
TABLE XIX
Comparison of Contaminant Content of
Condensates Without and With Steam Stripping
Total Volatiles Content* Ib/ADT
Without Stripping With Stripping
Combined Condensates
to 17$ B. L. Solids
Combined Condensates
\1% to 22% B. L. Solids
29.0
23.9
2.0
I.I
* Includes methanol, terpenes, H?S, CH,SH, (CH )?S, with sulfur com-
pounds as pounds sulfur equivalent.
Improved Phase Equilibrium and Solubility Data
The prior detailed calculations of volatiles behavior during flash
evaporations and steam stripping, and Indeed all prior design studies of
this system are based upon a simplification of the vapor-liquid
equilibrium situation. Simplification is necessary because the actual
50
-------
number of volatile components is at least fifteen, consisting of several
terpenes in addition to a-pinene; several terpineols in addition to
a-terpineol; and acetone, ethanol, methyl isobutyl ketone in addition
to methanol; together with the four sulfur compounds—hydrogen sulfide,
methyl mercaptan, methyl sulfide and dimethyl sulfide. The simplifica-
tions were basically to treat the multi-component system as a collection
of binary systems with water, utilizing the vapor-liquid binary equili-
brium data with water. With all substances except water present only
in trace amounts, this simplification may be justifiable. Under
conditions of concentration of methanol and terpenes to higher levels,
as in the upper sections of stripping column A with reflux, the
assumptions involved may not be valid.
An assessment of the validity of these assumptions and an improved
method of developing multi-component correlations of equilibrium data
to supplant these techniques, has been made by Mr. Robert T. Ruggeri
in work supported by this contract and reported in detail as a Master's
thesis study (37). Results obtained in this work are discussed in the
foI lowing pages.
The technique consisted first of reducing the muIti-component system
to a five-component system, water, methanol, a-pinene, a-terpineol and
methyl sulfide, where each compound represents a "family" of compounds
similar in polarity and chemical nature. This in turn was shown to be
reducable to three ternary systems, because only water and methanol
appear in moderate to large concentrations in the industrial problem
of interest here. Mr. Ruggeri then developed a technique for
predicting the ternary equilibrium behavior for the systems from binary
equilibrium data available in the literature or obtained experimentally
in the thesis investigation. This technique was based essentially upon
the equations of Renon and Prausnitz (38). An added advantage of the
approach is that extrapolations of equilibrium relationships from the
region of available experimental data (such as 25°C) to the level of
-------
design interest (such as IOO°C) is much more sound than the direct
extrapolation of curves. Renon's equations are based on the molecular
pairwise interactions of closest neighbors. By utilizing such a model,
an expression for the activity coefficients of the components was
obtained which yielded good results for several molecularly dissimilar
systems. For binary systems, Renon's equations for the activity
coefficient of the first component are as follows:
In Y! =C62I(G2I/(XI+X2G2I))2 + 6|2G|2/(X2 + X|G|2)2^ *^ • (ll)
where G|2 = exp '-a^6^' and ( I2)
G2I = exp (-a|252|K (I3)
The quantities ct|2, 6._, 62 may be treated as empirical constants,
although a.^ is related to the coordination number and as such has some
physical significance. It is found experimentally, however, that smalI
variations in a 2 do not significantly effect the ability of Equations
(II), (12), and (13) to represent experimental data. Furthermore, the
value of a.2 is characteristic of the types of compounds involved in
the 1-2 interaction. This means that the system of Equations (II),
(12), and (13) can be considered a two constant equation for the
activity coefficient. The Renon equations also enjoy two additional
advantages over the Margules type equations: the activity coefficient
for the i component of a muIticomponent mixture can be written in
terms of binary data only, and the temperature dependence of the
empirical constants, 6 2 and 62|, is found to be linear with reciprocal
temperature. Both of these advantages result from the fact that only
pairwise nearest neighbor interactions are considered in the Renon
liquid model. The disadvantage of the Renon equations is the added
computational complexity, but the advantages are enough to justify their
use, especially for multicomponent work.
An important advantage the Renon equations hold over the more convention-
al methods of describing excess molal Gibbs energy (g ) is that the
52
-------
Renon constants can be found from linear functions of temperature.
From Renon!s work the delta constants were defined as follows:
/RT, (14)
62I = (g2l " 9||)/RT' (I5)
and g|2 = g2|, (16)
where g.. is the energy of interaction between a pair of molecules, i
and j. The functions of g, i.e., (g. . - g..), were then determined
•J J J
experimentally to be linear in temperature. Renon demonstrated this
temperature dependence for at least two systems over a temperature
range of 50°C.
One of the primary advantages of using Renon type equations to describe
the nontdeality of solutions is that these equations can be generalized
to multicomponent systems without making any additional assumptions.
The more conventional methods of describing the compositional dependence
of activity coefficients is to use power series in mole fraction. Wohl's
expansion for g , used to describe multicomponent mixtures, involves
terms related to molecular interactions. For instance, the constant
a|2 is related to the 1-2 molecular interaction and a..„ is related to
the three-body interaction involving two molecules of component one
and one molecule of component two. Wohl's equation for g can be
considered a power series In volume fraction, and when this equation
is fitted to experimental data, as many terms are used as are needed
to give the desired accuracy. All the two-body constants of the
Won I equation can be calculated by studying binary solutions, but
ternary system behavior may not be represented accurately if only the
two-body terms are included in the ternary Wohl expansion. The only way
to check the reliability of this equation is to study the ternary
system and evaluate the three-body constants. If the two-body constants
are large with respect to the three-body constants, the Wohl expansion
can be determined from binary data. Both the Van Laar equations and
the Margules equations can be derived from Wohl's expansion and,
53
-------
therefore, their use is governed by similar considerations.
Unlike the Won I type equations, the Renon expansion of g was derived
by considering only nearest neighbor two-body interactions. The Renon
equation, therefore, can be determined from binary data only, without
making any additional assumptions. The multicomponent Renon expression
for the excess molal Gibbs energy, for a system of N components is
as fol lo»s: N N N
9/RT^,x> % 4jiGjixj"(jiiGkixk)' (I7>
where 6.. = (g.. - g::)/RT, (18)
= (g - g..)/RT,
ajjS.,), (20)
and
By differentiating Equation (17) appropriately, the following
expression is obtained for the activity coefficient:
N N
In y. = D. + (x.G. ./ G. .x. ) (6. . - D.,), (22)
i I p] J U jjEy kJ k ij 2
N N
where D, = (S" «..G..x.)/( G. .x, ), <23)
I Jrj Ji Ji J £r| kl k
N N
and D, = ( x 6 .G .)/( G. .x. ). (24)
2 (^7 m mj mj ^j kj k
The primary goal of this project was to describe in some way the
chemical and phase equilibrium of the mixture of components which might
be present in a commercial steam stripper designed for odor reduction
of condensates in a kraft pulp mill. Since the number of components in
such a piece of equipment is large, a total analysis of the system
would require a great deal of time and labor. In order to simplify the
54
-------
problem, three compounds representing different classes of organic
chemicals were chosen: a-terpineol, dimethyl sulfide, and pinene, and
each of these three compounds was studied in a ternary system with
water and methanol. This fact should not be interpreted to imply that
water and methanol do not represent classes of compounds, but rather,
that these two chemicals were assumed to be present throughout the
column; therefore, they were considered as components of every ternary
system. Each ternary system was studied by analyzing only binary
systems. Thus, the system of water-methanol-pinene was studied by
analyzing the three binary systems of water-methanol, water-pinene, and
methanol-pinene. The systems were investigated at two temperatures
wherever possible. The glass apparatus used prohibited analysis at
pressures above atmospheric; therefore, some systems were analyzed at
only one temperature.
Once the binary systems had been analyzed, their behavior was character-
ized by the use of activity coefficients. The Renon equations were
used for this purpose. The Renon method was chosen because it can be
generalized from binary data to mult{component systems without making
any additional simplifying assumptions, and because the temperature
dependence of the constants can be represented as linear functions of
temperature. The basic plan of attack is, then, to study three ternary
mixtures which represent mixtures of three classes of organic compounds,
and to characterize the behavior of these mixtures with activity coef-
ficients which are functions of composition at constant temperature.
The data collected on all the systems were analyzed basically by
computer. Some systems involving only two data points were calculated
by hand, but all systems utilizing more than two data points were ana-
lyzed with the aid of a computer. Simple programs were written
utilizing, wherever possible, Boeing library subroutines available at
the University of Washington computer center. All the programs were
designed to give some kind of average fit to the data. Usually a
55
-------
least squares averaging process was used. The solubility data were fit
by least squares polynominals in temperature. The Renon equations were
fit to the data by the same type of procedure, but the calculations
utilized a library program designed to solve a system of non-linear
algebraic equations.
The results obtained from the work described previously can be
classified into two groups: the binary results calculated from experi-
mental data, and the ternary results predicted from the binary results.
The ternary results are of primary concern here, since they are to be
essentially derived from the binary results and would be of value in
making detailed design calculations for separation columns. Binary
system experimental results and computed equation constants based
upon the results are shown in Tables XX to XXII.
Figures 6, 7, 8, and 9 show the more important of these ternary
equilibrium relationships. Figure 6 shows the only direct ternary
equilibrium data available to date superimposed on the ternary
diagram results for 25°C. The tie-lines linking liquid phases in
equilibrium show reasonable agreement between the experimental and
theoretical results. Comparison of Figures 6 and 7 shows the influence
of temperature in this methanol-water-a-pinene ternary.
Appraisal of Approaches
Earlier methodology applied to studies of the removal of volatile and
odorous components from aqueous streams have been extended to include
consideration of more of the volatile components, stripping of black
liquor, stripping of condensates in dual columns, and the vapor-liquid
equilibrium relationships prevailing in the multi-component system.
Some guidelines may be useful as to the extent that one or more of
these approaches should be adopted in a given situation.
56
-------
TABLE XX
Water-Organic Binary System Results
H20 IN a-TERPINEOL @ 25°C
Average Solubility = 47.65 gml-LO/l. sin.
H0 IN a-TERPINEOL @ 58°C
Average Solubility = 27.56 gm HJVI. sin.
a = 1.76 = 6.4%
a-TERPINEOL IN HJD
Polynominal Order
I
2
3
Constants
C.
2.14
3.26
4.20
9 = P
b t,
C2 C3 C4
6.2 X I0~* 0.0 . 0.0
-4.2 X 10 , 3.6 X 10 -, 0.0 ,
-9.5 X 10 ^ 1.3 X 10 -4.8 X 10
Equations:
+ C2T + C3T2 + C4T3
T = Temperature °C
S = Solubility (gm a-Terpineol/l . soln. )
PINENE IN H20
Polynominal Order Constants
I I.14 X 10"^ I.7 X 10 ^ 0.0
2 9.95 X 10 2.4 X 10 -7.0 X 10
Equations: Same as for a-Terpineol solubility.
H20-DIMETHYL SULFIDE SYSTEM
Liquid Phase Mole Fraction Water Temperature
H20 rich 0.9998 25°C
DMS rich 0.0101 25°C
57
-------
TABLE XXI
Methanol-Organic Binary System Results
METHANOL-DI METHYL SULFIDE % 24°C METHANOL-DI METHYL SULFIDE @ 36°C
Mole Fraction
Liquid Phase
0.046
0.049
0. 1 15
0. 1 15
0.215
0.215
0.297
0.297
0.355
0.355
0.465
0.465
0.465
0,560
0.560
0.673
0.673
0.844
0.844
0.958
Liquid Mole
Fraction MeOH
0.773
0.788
0.803
0.872
0.883
0.891
0.949
0.953
0.958
0.974
0.976
0.978
Dimethyl Sulfide Mole Fraction
Vapor Phase Liquid Phase
0.405 0.041
0.422 0.041
0.581 0.041
0.588 0.203
0.692 0.203
0.685 0.203
0.729 0.874
0.736
0.727
0.758
0.822
0.829
0.823
0.835
0.829
0.851
0.853
0.885
0.890
0.897
Methanol-Terpineol System
Activity of Methanol
Activity of MeOH
at 26°C
0.671
0.705
0.727
0.813
0.827
0.840
0.915
0.928
0.928
0,956
0.964
0.964
Dimethyl Sulfide
Vapor Phase
0.435
0.493
0.413
0.664
0.664
0.685
0.874
Activity of MeOH
at 44°C
0.747
-
0.771
0.843
-
0.860
-
0.933
0.940
-
0.962
0.966
58
-------
TABLE XXI (Cont'd)
Methane I-Pinene @ 24°C
Liquid Phase Mole Fraction Methanol
MeOH rich
Pinene rich
0.913
0.193
TABLE XXI I
Summary of Margules Equation Constants For
Binary Systems
Water ( 1 )
Pinene ( 1 )
Methanol (
Methanol (
Methanol (
System
- DMS (2)
- Water (2)
1 ) - Pinene (2)
1) - DMS (2)
1 ) - ct-Terpineol (2)
Temperature
25°C
25°C
25°C
24°C
24°C
Margul es
B
6.561
10.47
2.870
1.442
-1.363
Constants
C
1 .951
-2.648
0.037
-0.235
-0.152
Equations:
lnv( = X^ B - C(4X2 -3)
Iny2 = X^ B + C(4Xj -3)
TABLE XXI I I
Summary of Renon Equation Constants
Binary Systems
For
Renon Constants
System
Water (1) - Methanol (2)
Water (1) - Methanol (2)
Water (1) - Pinene (2)
Water (1) - DMS (2)
Water (1) - Terpineol (2)
Water (1) - Terpineol (2)
Methanol (1) - Pinene (2)
Methanol (1) - Pinene (2)
Methanol (1) - DMS (2)
Methanol (1) - Terpineol (2)
Temperature
25°C
75°C
25°C
25°C
25°C
58°C
25°C
64°C
24°C
26°C
a!2
0.47
0.47
0.007
0.005
0.20
0.20
0.034
0.034
0.74
0.27
6!2
-0.162
-0.559
36.10
45.30
8.903
8.564
1.999
1.037
1.050
-1.241
52I
0.377
0.315
-19.85
-31.36
-0.283
0.275
1.000
0.802
2.00
- 1 . 397
59
-------
MltHAHOL
Mgurt
rvtutt*. Cooc«ntr»tl«\» «•• |t»*n
u
VATEK
ri«ur* 1. ternary UquW
»t 100°C.
in ml
METHAKOJ.
WATER
JHKEHE
60
-------
METHANOL
Figure I. Ternary liquid equilibrium
reiult* «t 100°C. concentration* «r«
given In aele fraction*.
NATU
Fliurt 9, Ternary liquid equilibrium
result* «t 100°C. Coneentratlotui are
given In ••!• fractions.
DIMETHYL S III/1 DC
MCTHANOL
HATH
a-TCRPINEOL
-------
First, it is the author's belief that In the great majority of applica-
tions, steam stripping will be superior to air stripping of condensates.
The major advantages are (I) the elimination of the air contamination
problem, (2) the possibility of readily and cheaply concentrating the
stripped volatile components for burning or other disposal, or for
subsequent recovery, and (3) the use of temperature of the condensate
in the column as a readily controllable variable, by means of steam
pressure. Neither air nor steam stripping are effective for removal of
non-volatile compounds.
The use of a single column to strip only those condensates of highest
concentration of volatile odorous components, or those to be directly
discharged rather than utilized within the mill, remains the most
advantageous application of the SEKOR process at minimum cost. The
economics of such a system, as described in prior work, Maas, et_. a I.
(25), represents a minimal installation of a column about one-third to
one-half the cross-sectional area considered in the present case. If
these capital costs are scaled up by a factor of about 3.0 to provide
for these differences, the overall costs should remain substantially
valid today if costs and turpentine pricing are updated. If methanol
is also to be removed to levels less than 2.0 pounds per ton of pulp,
then such a column will be inadequate, and two columns are recommended,
stripping flash and steaming vessel relief condensates in the first,
and about the first two evaporator stage condensates in the second.
This would correspond to Figure 3 without the SEKOR B column for
stripping black liquor. Costs for such a dual column system would be
about double a single column installation, with steam and other utility
costs also about double. A bonus in such a column operation is that the
evaporation condensates are much more free of hydrogen sulfide and other
odorous volatile components than if untreated, and more suitable for
reuse or discharge.
62
-------
The concept of steam stripping black liquor at digester pressure before
it is flashed appears to have considerable promise for some situations.
Such a system would require pilot scale investigations to corroborate
the design studies reported here. It would appear to offer the most
promise where recovery of turpentine in high yield is an important
consideration, and where extreme requirements for extent of removal
of methanol and odorous constituents from condensates to be discharged
warrants the added cost of black liquor stripping. Costs for such a
black liquor stripping column cannot be estimated with confidence with-
out pilot scale study of column operability, but it appears from this
design study that such a column would have a cross-sectional area about
equal to the combined areas of condensate stripping columns "A" and "C"
of Figure 3. Column height would be comparable to these columns.
Capital costs would be approximately double that of a two-column
condensate stripping system, but steam costs would be only slightly
higher, because of the steam re-use feature of the overall system.
63
-------
SECTION VI I
SEPARATION OF SEKOR OILS
General Considerations
In the basic research and development of McCarthy, Hrutfiord, Johanson
et al which established the SEKOR process, the observation was made that
an oiI phase would appear and could be collected separately from the
aqueous phase in the overhead stream if the fractionating column over-
head water phase was refluxed back to the column. The oil phase was
mainly terpene hydrocarbons and organic sulfides, while the aqueous
phase was mainly water together with smalI quantities of water-soluble
organics such as methanol, ethanol, and acetone derived from the
pulping process. Subsequent research and plans have been based upon
this observation. H. Maahs and D. Marsh, working with Professor L.N.
Johanson, established in preliminary studies that the terpene hydro-
carbons and organic sulfur compounds (methyl sulfide and dimethyi-
disulfide) could also be removed from condensates by counterflow solvent
extraction using petroleum hydrocarbon, as an alternate to steam
stripping. The economics in general appeared less attractive than
steam stripping, however.
QiI-Water Separations
Early bench scale experiments revealed that SEKOR oils consisted in
part of the terpenes ordinarily found in commercial turpentines, with
perhaps larger than expected amounts of terpineols. A large fraction,
however, (approximately 40/£) of oil derived from blow gas condensate
consisted of sulfur compounds, chiefly the methyl sulfides. Thus the
usual methods of sale or disposal of Kraft turpentines, consisting of
less than one per cent sulfur compounds—concentrated acid treatment,
chlorination, and caustic washing—seem not to be ideal for the high
sulfur concentrations found in SEKOR oils. Furthermore, the high cost
and sulfur losses associated with this approach would tend to discourage
65
-------
rather than encourage adoption of the SEKOR process by industry. The
terpene hydrocarbons are largely within commercial turpentine boiling
range and would have value in commerce as a credit toward the costs of
operation of the SEKOR process. The sulfur compounds (chiefly methyl-
mercaptan, dimethyl sulfide and dimethyl disulfide) are odorous,
unpleasant and noxious. If they can be released from SEKOR terpene oil
in concentrated form, they would have value as makeup sulfur in the
pulping process, particularly if they could be reduced to hydrosulfide
form.
Fractional Disti1lation would appear to be a logical method of separa-
tion of the oils into terpenes and sulfur compounds. However, the
several terpenes, terpineols, and sulfur compounds have overlapping
volatilities, which makes such separation complex. A partial separa-
tion could perhaps be conducted if only a-pinene were to be recovered
from all other constituents.
So I vent Ext ract i on of the sulfur compounds from the terpenes, or vice
versa, would require a solvent with limited, but different solubility
for each class of compound. Over twenty solvent systems were explored
as possibilities, with terpene-dimethyl sulfide mixtures ranging from
10-90 to 90-10 molecular ratio. The solvents were found to fall into
two groups, the larger group being those having complete miscibility
with both terpenes and methyl sulfide, and a smaller group with little
or no solubility for either terpenes or methyl sulfides.
Selective Adsorption Separations
This would appear to offer a possible alternative separation method for
the two classes of compounds terpenes and organic sulfides. The former
compounds are all ring structures, of molecular diameter exceeding 8
angstroms. The sulfur compounds H_S, CH^SH, (CH^)^ are all linear
molecules, with diameters approximately 4 to 8 angstroms. Methanol
would appear to associate with the sulfur components because of its
molecular shape, but its greater polarity makes prediction of its
66
-------
behavior in this respect difficult. Possible selective agents for this
purpose are zeolites or synthetic zeolites, silica gel, and activated
carbon. All have at least a substantial portion of their pore diameters
within the probably useful 3 to 30 angstrom range.
The zeolites which are crystalline a IuminosiIicates have long been known
in nature, and have long been in use as water softening agents. The
synthetic zeolites, however, are a relatively recent development. They
differ from the natural crystals in having a controlled structure and a
controlled composition. When used as "molecular sieves" (a term coined
by J.W. McBain in 1926 for natural zeolites with pores less than five
angstroms in diameter), the structure rather than the chemical composi-
tion is of most importance. Such zeolites are now available commercially
with pore diameters less than 3, 4, 5, 10, or 13 angstroms respectively.
Silica gel and activated carbon have much broader pore size distributions
than the synthetic zeolites. Activated carbon or charcoal was not
tested in this study.
Solvent Extraction Separation jtudies
The exploratory search for suitable solvents, or extractive distillation
agents has not been promising. The major difficulty appears to be that
the methyl groups of methyl sulfides and mercaptan make their solvent
characteristics similar to the terpene hydrocarbons. Consequently a
solvent which is miscible with terpenes is also found to be completely
miscible with the sulfur compounds. Such solvents tested have included
acetone, butyl alcohol, n-butyl ether, chloroform, ethyl alcohol
'methanol, n-hexane, dimethyl aniline, and pyridine. All of these
solvents were found completely miscible with a-terpene - dimethyl
sulfide mixtures ranging from \0% - 90% to 90% - \0% in composition.
Additional solvents tested with both dimethyl disulfide and dimethyl
sulfide, as well as terpene mixture were: carbon dlsulfide, dimethyl
sulfide, benzene, toluene, cyclohexane, carbon tetrachloride, phenol,
nitrobenzene and cresol. The same results of complete miscibility were
67
-------
found. At the other extreme, liquids which are essentially Immiscible
with terpenes, which include ethylene glycol and water, are also found
to dissolve dimethyl sulfide to an extent less than detectable by the
techniques used, approximately 0.5/K.
A search was conducted for solvents which would selectively remove
either the terpenes or the sulfur compounds for use in extraction
processes. Both literature search and laboratory experiments with a
variety of solvent classes proved negative, in that solubility of the
terpenes and of sulfur compounds were too similar in a given solvent
to allow for an economically attractive separation process.
Thus extraction systems of promise have not been found. Some possibil-
ities still exist, for example liquid HF and liquid S0_, as used in the
petroleum industry for desulfurizing. These are unattractive as
solvents for reasons of refrigeration requirements and toxicity of HF,
and are intended for thiophene class sulfur compound removal as well.
Selective Adsorption Separation of Terpenes and Sulfur Compounds
The SEKOR oils consist largely of the two classes of compounds, terpenes
and methyl sulfur compounds. The terpenes found in the volatile
fractions of Kraft turpentine are substantially all monocyclic (Iimonene,
a-terpineol), or dicyclic (pinenes, carene, camphene). Their molecular
diameters thus exceed five angstroms, whereas the molecular diameter of
methyl mercaptan is 5.0 angstroms. Methyl sulfides, being linear
molecules, should similarly have a molecular diameter normal to the bond
directions of about five angstroms. Thus it should be possible to
selectively remove the sulfur compounds from the terpenes in a batch or
semi-continuous process. One such process, utilized to separate straight
chain paraffins from iso-paraffins or naphthenes, is described in
reference 39. Experiments were conducted utilizing "Linde" brand molec-
ular sieves (synthetic crystal Itne alumino-siIicates), and silica gel.
The molecular sieves have narrow pore size distribution while the silica
gel has a broad spectrum of pore sizes. Molecular sieve 3A, 5A, IOA
68
-------
and I3X were tested. The 3A and 5A materials, with approximately 3
angstrom and 5 angstrom pores, were found non-selective with respect
to sulfur compound removal from terpenes. The 10-A and I3X materials,
of approximately 8 angstrom and 10 angstrom pores, respectively were
found to preferentially retain the sulfur compounds from the terpenes.
The extent to which this occurs was determined utilizing packed columns
of crushed and pelleted sieve material, and determination of break-
through curves using mixtures of a-pinene or commercial turpentine and
dimethyl sulfide. Analysis was by means of gas liquid partition chrom-
atography, for which calibration data were established for the above
constituents and methanol. Methanol was utilized as a column purge, with
regeneration by evaporation of the methanol remaining by heating the
co Iumn.
Silica gel was also found to retain selectively the sulfur compounds,
with techniques much as utilized with the molecular sieves.
Laboratory experiments were planned such as to furnish suitable design
data for larger scale application. Continuous adsorption techniques
have been developed and reported in the literature utilizing moving
bed or sequencing feed and withdrawal lines. It is likely, however, that
for the scale of operation required for processing SEKOR oils a batch
sequence technique would prove preferable. In either case, required
design data would include break-through curves for both adsorption and
regeneration steps, at a given temperature, flow rate, and column length.
The S shaped break-through curve for adsorption represents the transition
from purified terpene effluent and effluent having the composition of the
feed material, as a result of exhaustion of the column capacity.
Similarly, the break-through curve for regeneration represents the
transition between methanol-sulfur compound mixtures and pure methanol,
signifying regeneration is complete.
Break-through curves have been obtained for silica gel and for 10
angstrom molecular sieve, with a-pinene methyl sulfide mixtures. For
69
-------
the latter adsorbent, successful regeneration by elution with methanol,
followed by drying has also been demonstrated. This regenerated
molecular sieve is found to be essentially the same as the original
material, giving almost superposition of adsorption break-through
curves.
The most effective adsorbent found for selectively removing, for example,
dimethyI disuIf ide from a-pinene was si Iica geI, with 13 to 20 ml. of
solute-free terpene (ct-pinene) recovered from a 29 ml. (0.64 cm by 90
cm) gel-packed column before the appearance of the break-through curve
for dimethyl disulfide. Much less effective separation was obtained
o o
using eight Angstrom and ten Angstrom molecular sieves. The five
Angstrom sieve material was found to have no selectivity for sulfur
compound removal.
The finding that silica gel was superior as a selective adsorbent for
methyl sulfides compared to molecular sieves was unexpected. It was
earlier postulated that the narrow size ranges of pores in the sieves
should be ideal for removing the straight chain methyl sulfides from
the ring-structure terpenes. Silica gel has a broad range of pore
sizes, and its superior effectiveness suggests factors other than
molecular shape are more important in this separation.
Twenty-five sets of break-through data were obtained all at room tem-
perature, comprising four adsorbents, three flow rates and four sulfur
compound-terpene concentrations. The more significant data obtained
are summarized in Table XXIV. Additional data, together with a
discussion of equivalent mass transfer coefficients, and application to
possible scale-up of results, are available in the Master's Thesis of
Kap Kyun Kim (40). It is apparent from Table XXIV that silica gel is
superior as a selective adsorbent to any of the molecular sieves tested.
This is shown by the larger volume of terpenes which can be passed
through the column before the appearance of the sulfur compound (at 5%
70
-------
of its influent concentration). An additional indication of superiority
is the shorter break-through zone height for silica gel.
TABLE XXIV
Summary of Break-Through Data for
Selective Adsorption of Dimethyl Disulfide from Terpenes
^Adsorbent
Molecular
Sieve 5A
Molecular
Sieve IDA
if
it
Mo lecul ar
Sieve 1 3X
Si lica Gel
ii
it
it
ti
Pore
o
Radius A
5
8
8
8
10
10
15-100
1 5- 1 00
1 5- 1 00
1 5- 1 00
1 5- 1 00
Mole. %
(CH3)2S2
in Terpene
20
5
20
20
20
20
20
20
20
30
30
Flow Rate ml. to Breakthrough Zone
2
g./cm hr 5% 95% Ht., cm
85
35
35
85
35
85
35
85
170
85
170
0
2.5
1.0
0.7
0.3
0.3
20.0
17.0
15.0
13.2
13.1
1
5
3
5
3
4
25
24
24
16
16.5
___
50
80
1 10
105
130
18
30
45
15
17
* Adsorption column: glass, 0.64 cm x 90 cm, filled with crushed
(24-200 mesh) molecular sieve (20 grams) or with 28-200 mesh silica
gel (22 grams). All experiments conducted at room temperature (72°F).
Appraisal of Approaches
For only moderate contamination of the turpentine fraction by sulfur
compounds and methanol, the "brute force" method of chemical destruction
by oxidation of the impurities followed by water washing would appear to
be still preferable to the methods explored here. For higher concentra-
tions of contaminants, selective adsorption would appear to offer more
promise than solvent extraction and also more than fractional distilla-
tion unless a-pinene is the only major constituent to be recovered by
distiIlation.
71
-------
For selective adsorption the higher cost of molecular sieves, together
with the frequency of regeneration required (because of early break-
through of solute) suggests that economic feasibility for this adsorbent
is unlikely in this application. There may be some possibility of a
feasible process based upon silica gel, however. For example, utilizing
the data for silica gel adsorbent, and a 20% methyl sulfide 8Q% terpene
2
feed mixture flowing at 170 grams per hour per cm o-f column cross
section, 15 ml. of effluent is collected before break-through occurs.
This is equivalent to the purification of about 1000 gallons of SEKOR
oil charged over a period of six hours to a gel-packed column of 2-ft
diam. x 10 ft height, before regeneration is required. This would
represent the oil production of a large Kraft mill.
Regeneration of the adsorbent was conducted in the laboratory study
utilizing methanol to elute the dimethyl disulfide (as well as the
retained terpenes). This polar material was found to be highly
effective in displacing the dimethyl disulfide, with the concentration
of the latter in the silica gel column effluent falling from 65 mole
per cent to less than I mole per cent. Although methanol is not the
only possible eluent (heating, steam, air or water may be used) it is
of interest because it does not introduce a new air contamination
problem, is a possible by-product of the SEKOR process, and can be
separated from the eluted methyl sulfides by distillation. Alternatively,
a continuous industrial process (moving bed or multi-feed column) may
be visualized, though it is probable that for the application to steam
stripping the batch cyclic process would be simpler and more promising.
72
-------
SECTION VIII
ACKNOWLEDGEMENTS
The financial support of this Investigation by the Environmental
Protection Agency Office of Research Monitoring is gratefully
acknowledged.
During the course of this study useful discussions were held with a
number of people, these included Professor K.V. Sarkanen, Dr. Josef
Gratzl, Dr. Wolfgang Glasser and Dr. Kaj Forss of the University of
Washington as well as John Van Vessen and James Leonard of the
Weyerhaeuser Company. Collaborators on the project who have given
invaluable assistance to the Investigation were Ms. Juanita Collins,
H.R. Monahan, J.T. Ruggeri, Kap Kim and Donald Wilson.
Assistance of the member mi I Is of the Northwest Pulp and Paper
Association is also acknowledged.
73
-------
SECTION IX
REFERENCES
I. Hrutfiord, B.F. and McCarthy, J.L., "SEKOR I: Volatile Organic
Compounds in Kraft Pulp Mill Effluent Streams." Tappi 5Q, No. 2,
82-85 (1967).
2. Bethge, 0., and Ehrenborg, L., "Identification of Volatile Com-
pounds in Kraft Mill Emissions." Svensk Papperstid., 22, 347-350
(1967).
3. Wilson, D.F. and Hrutfiord, B.F., "SEKOR IV. Formation of Volatile
Organic Compounds in the Kraft Pulping Process." Tappi 54, No. 7,
1094-1098, (1971).
4. Marvel I, E.E. and Wiman, R.E., "4-(p-TolyI)-l-pentanol in Douglas
Fir Pulping Products." J. Org. Chem., 28_, 1542-1545 (1963).
5. Leonard, J., Weyerhaeuser Company, personal communication.
6. Cabauatan, E.Q., "Monoterpenes in Douglas Fir Needle Oil, Wood
Turpentine and Crude Sulfate Turpentine." M.S. Thesis, University
of Washington (1969).
7. Sarkanen, K.V., Hrutfiord, B.F., Johanson, L.N., and Gardner, H.S.,
Kraft Odor," Tappi, 53_, No. 5, 766-783 (1970).
8. Ruus, L., "Study on Pulp Mill Effluent." Svensk Papperstid., \9_
751-755 (1964).
9. Turner, B.G. and van Horn, J.T., Southeastern Section TAPPI
Meeting, March, 1969.
10. Seppovaara, 0. and Hynninan, D., "On the Toxicity of Sulfate-Mi II
Condensates." Paperi Ja Puu, J_, 11-23 (1970).
II. McDermott, G.N., Tech. Bull. No_. 72_, NCASi, November 1954, revised,
1970.
12. Wilson, D.F., Johanson, L.N., and Hrutfiord, B.F., "Methanol,
Ethanol and Acetone in Kraft Pulp Mil I Condensate Streams," Tappi,
j>5, No. 8, 1244-1246 (1972).
13. Cade, S.C., Hrutfiord, B.F., and Gara, R.I., "Identification of the
Primary Attractant for Gnathotrichus Sulcatus Isolated from Western
Hemlock Logs." J_. Econ. Entom., 63, 1014-15 (1970).
14. Clayton, D., "The Alkaline Degradation of Some Hardwood 4-o-Methyl-
D-Glucuronoxylans." Svensk Papperstid., 4, 115-124 (1963).
75
-------
15. Hrutfiord, B.F. and Wilson, D.F., "Turpentine Concentrations in
Kraft Mill Condensate Streams." MMS in EPA review, August (1972).
16. Ores, J., Russell, J., and Bajak, H.W., Sulfate Turpentine Re-
covery. Pulp Chemicals Association, New York (1971).
17. Liu, L.Y., "Turpentine Collected from Continuous Digester."
Pulp and Paper Int., 12, No. 3, 55-57 (1970)
18. Graham, K., "Anaerobic Induction of Primary Chemical Attractancy
for Ambrosia Beetles." Can. J_. Zoo I., 46, 905-908 (1968).
19. Wasserman, A.E., "Organoleptic Evaluation of Three Phenols Present
in Wood Smoke." J_. Food Sci., 31, 1005-1010 (1966),
20. Leonardos, G., Kendall, D., and Barnard, N., "Odor Threshold
Determinations of 53 Odorant Chemicals." J. Air Poll. Control
Assoc., ]9, 91-95 (1969).
21. Sarkanen, K.V., "The Chemistry of Delignification in Pulp Bleaching."
Pure and Applied Chemistry, 5_, 221-233 (1962).
22. Das, B.S., Reid, S.G., Betts, J.L., and Patrick, K., J. Fisheries
Res. Board Can., 26, 3055-2067 (1969). ~
23. Shin, T.T.C., Hrutfiord, B.F., Sarkanen, K.V. and Johanson, L.N.,
"Methyl Mercaptan Vapor-Liquid Equilibrium in Aqueous Systems as a
Function of Temperature and pH." Tapp!, 50, No. 12, 634-638 (1967).
24. Matteson, M.J., Johanson, L.N., and McCarthy, J.L., "Steam
Stripping of Volatile Organic Substances from Kraft Pulp Mill
Effluent Streams." Tappi, 5£, No. 2, 86-91 (1967).
25. Maahs, H.G., Johanson, L.N., and McCarthy, J.L., "SEKOR III. Pre-
liminary Engineering Design and Cost Estimates for Steam Stripping
Kraft Pulp Mill Effluents." Tappi, 50_, No. 6, 270-275 (1967).
26. Shih, T.T.C., Hrutfiord, B.F., Sarkanen, K.V., and Johanson, L.N.,
"Hydrogen Sulfide Vapor-Liquid Equilibrium in Aqueous Systems as a
Function of Temperature and pH." Tappi, 50_, No. 12, 630-634 (1967).
27. Tsuchiya, G.S. and Johanson, L.N., "Prediction of Generation and
Release of Odorous Gases from Kraft Pulp Mills." Tappi, 55, No. 5,
777-783 (1972). ~~
28. Hubbe> P.O., "Digester Problem Statement," Tappi, 49, No. 5, 6IA-
69A (1966). ""
76
-------
29. Wetherhorn, D., "The Calculation of Evaporator Heat Balances from
Operating Data." Tappl, 47_, No. 2, I68A-I7IA (1964).
30. Libby, C. Earl (Ed.), Pulp and Paper Science and Technology, Vol. j_.
McGraw HIM, New York, (1962).
31. Drew, J. and Plyant, G.D., Jr., "Turpentine from the Pulpwoods of
the United States and Canada." Tappl, 49_, No. 10, 430-438 (1966).
32. Bennett, C.O. and Meyers, J.E., "Momentum, Heat and Mass Transfer."
McGraw-Hill, New York (1962) p. 531.
33. Treybal, R.E., "Mass Transfer Operations, 2nd Ed.," McGraw-Hill
New York (1968) p. 252.
34. Chu, J.C., "Distillation Equilibrium Data." Reinhold Publishing
Corp., New York, New York (1950).
35. Perry, J. "Chemical Engineers Handbook, 4th Ed." McGraw-Hill,
New York (1963) p. 13-20.
36. Tsuchiya, G., "The Generation and Loss of Volatile Constituents in
the Kraft Pulping Process." M.S. Thesis, University of Washington
(1970).
37. Ruggeri, R.T., "Phase Equilibrium of Kraft Mill Effluent Streams."
M.S. Thesis, University of Washington (1971).
38. Renon, H. and Prausnitz, J.M., "Local Compositions in Thermodynamic
Excess Functions for Liquid Mixtures." AlChE Journal, j_4, 135-
144 (1968).
39. Broughton, D.B., "Molex: Case History of a Process." Chem. Eng.
Prog., 64_, No. 8, 60-65 (1968).
40. Kim, K.K., "Separation of Methyl Sulfides from Terpenes of Kraft
Pulp Mill Condensate Streams by Adsorption." M.S. Thesis, Univer-
sity of Washington (1969).
77
-------
SECTION X
PUBLICATIONS AND PATENTS
The following publications have been produced or are anticipated to be
produced as a result of this project:
SEKOR IV. Formation of Volatile Organic Compounds in Kraft
Pulping Process. Wilson, D.F. and Hrutfiord, B.F.,
Tappi, _54_, No. 7, 1094-1098, (1971).
SEKOR V. Methanol, Ethanol and Acetone in Kraft Pulp Mill
Condensate Streams. Wilson, D.F., Hrutfiord, B.F.,
Johanson, L.N., Tappi, 55, No. 8, 1244-1246 (1972)
SEKOR VI. Turpentine Concentrations in Kraft Mill Condensate
Streams. Hrutfiord, B.F., and Wilson, D.F. Accepted
for publication In Pulp and Paper Magazine of Canada
(1973).
Thesis Wilson, D.F., "VoI atile Organic Compounds in Kraft
Pulping." M.S. Thesis, University of Washington,
(1970).
Thesis Rugger!, R.T., "Phase Equilibrium of Kraft Mill
Effluent Streams." M.S. Thesis, University of
Washington (1971).
Thesis Kim, K.K., "Separation of Methyl Sulfides from
Terpenes of Kraft Pulp Mill Condensate Streams by
Adsorption." M.S. Thesis, University of Washington,
(1969).
It appears that no patentable developments have arisen from the present
study.
79
-------
SECTION XI
GLOSSARY
I. Black Liquor — Spent pulping chemicals and dissolved wood
components resulting from kraft pulping.
2. Blow Gas Condensate — Condensed steam resulting from discharging
digester contents at high temperature.
3. Condensate — General term for liquid resulting from condensing
steam etc. to water and organic liquids.
4. Crude Sulfate Turpentine — Mixture of terpenes and sulfur
compounds isolated from kraft digester.
5. Digester Relief Condensate — Condensate from condensing vented
steam and relief gas from kraft digesters.
6. Evaporators — Units which concentrate black liquor, giving
evaporator condensate as one product.
7. Kraft Pulping — Delignification of wood with sodium hydroxide
and sodium sulfide.
8. Liquid to Gas Ratio -- (L/G) The ratio on a weight basis of the
liquid being stripped in a column to the gas (steam) used for
stri pping.
9. SEKOR — Stripping Effluent for j
-------
SELECTED WA TER i. Report NO.
RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
2. 3. Accession No.
w
4. Title steam Stripping Odorous Substances From J- Report Date
Kraft Effluent Streams (SEKOR) 6.
7. Authors) Bjopn F> Hrutfiord, Lennart N. Johanson
Joseph L. McCarthy
9. Organization
University of Washington
Seattle, Washington
8. Performing Organization
Report No.
10. Project No.
12040 EXQ
11. Contract I Grant No.
13. Type of Report and
Period Covered
12. Sponsoring Organization
15. Supplementary Notes
Environmental Protection Agency report
number, EPA-R2-73-196, April 1973.
16. Abstract
Laboratory and design studies have been completed relating to volatile
constituents which appear In Kraft black liquor and condensate streams, and how these
can best be removed and recovered. In order of decreasing concentration, the
volatile constituents are alcohols, terpenes, ketones, sulfur bearing compounds, and
phenolic compounds. Methanol, the major alcohol contaminant, Is found in from 280 to
8400 ppm in condensate streams, amounting to I I to 16 pounds per ton of pulp produced.
Terpenes are found to range from a few ppm to about 4500 ppm in condensates, 4 to 9
pounds per ton of pulp. Acetone is present at concentrations of 2 to 200 ppm,
corresponding to 0.07 to 0.4 pounds per ton of pulp. In all, some 40 compounds were
found to be present in condensate streams. The feasibility of combining steam
stripping of Kraft liquor with steam stripping of condensates was explored, and the
conditions under which this may be warranted are reported. Under most present mill
situations, steam stripping of black liquor and the last stages of evaporator
condensates does not appear to be warranted except in unusual cases. Exploratory
type studies were made and are reported concerning Improved methods of predicting
vapor-liquid equilibria In such systems, and separation of the resulting volatile oils.
I7a. Descriptors
Steam Stripping, Pulp Condensates, Pollution Abatement, Water Reuse, Volatiles
Recovery, Black Liquor, Odor Control.
17b. Identifiers
Terpenes, SuI fur Compounds, Methyl Mercaptan, Methanol, Turpentine.
lie. COWRR Field A Group
It. Availability 19. Security Class.
(Report)
20. Security Class.
L.N. Johanson (P*ge)
21. No. of Send To:
Pages
22. Price WATER RESOURCES SCIENTIFIC INFORMATION CENTER
US, DEPARTMENT OF THE INTERIOR
WASHINGTON. D. C 20240
Abstractor Institution
WRSICI02(REV JUNE 1971) Sp0 ,|3-j,g(
------- |