&EPA
United States
Environmental Protection
Agency
Robert S. Kerr Environmental Research
Laboratory
Ada OK 74820
EPA-600 2-80-062
April 1980
Research and Development
Caprolactam
Recovery from
Aqueous
Manufacturing
Streams
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental 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.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-80-062
April 1980
CAPROLACTAM RECOVERY FROM
AQUEOUS MANUFACTURING STREAMS
by
John H. Dibble
Union Carbide Corporation
Tarrytown, New York 10591
Grant No. R-803737
Project Officer
Billy L. DePrater
Source Management Branch
Robert S. Kerr Environmental Research Laboratory
Ada, Oklahoma 74820
ROBERT S. KERR ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ADA, OKLAHOMA 74820
-------
DISCLAIMER
This report has been reviewed by the Robert S. Kerr Environmental
Research Laboratory, U. S. Environmental Protection Agency and approved
for publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U. S. Environmental Protection Agency,
nor does mention of trade names or commercial products constitute endorse,
ment or recommendation for use.
ii
-------
FOREWORD
The Environmental Protection Agency was established to coordinate admin-
istration of the major Federal programs designed to protect the quality of our
environment.
An important part of the Agency's effort involves the search for infor-
mation about environmental problems, management techniques and new technologies
through which optimum use of the nation's land and water resources can be
assured and the threat pollution poses to the welfare of the American people
can be minimized.
EPA's Office of Research and Development conducts this search through a
nationwide network of research facilities.
As one of these facilities, the Robert S. Kerr Environmental Research
Laboratory is responsible for the management of programs to: (a) investigate
the nature, transport, fate and management of pollutants in ground water;
(b) develop and demonstrate methods for treating wastewaters with soil and
other natural systems; (c) develop and demonstrate pollution control tech-
nologies for irrigation return flows; (d) develop and demonstrate pollution
control technologies for animal production wastes; (e) develop and demonstrate
technologies to prevent, control, or abate pollution from the petroleum re-
fining and petrochemical industries; and (f) develop and demonstrate technolo-
gies to manage pollution resulting from combinations of industrial wastewaters
or industrial/municipal wastewaters.
Recovery of process unit wastewater contaminants and reuse of these con-
taminants in the manufacturing processes is a feasible engineering method to
reduce environmental contamination as well as improve the efficiency of
manufacturing processes. One method of contaminant recovery from process
water streams is extraction of the aqueous waste with specific solvents. This
report addresses the potential of solvent extraction methodology for caprolac-
tam aqueous waste streams utilizing laboratory and pilot- scale technology on
actual plant samples. Bench and pilot scale data indicate a potential
reduction of caprolactam content in wastewater streams from over 1,000 mg/1 to
as little .as 30 mg/1 with an energy -savings when compared to the existing
extraction technology.
0
W. C. Galegar
Director
Robert S. Kerr Environmental Research Laboratory
111
-------
ABSTRACT
The project objective is to further develop the Union Carbide Corporation
extraction process for caprolactam recovery by testing actual plant samples,
conducting pilot plant runs, and obtaining physical property data useful for the
process design.
Pilot plant runs have demonstrated the feasibility of a novel extraction process
for caprolactam recovery from dilute aqueous solutions. Following extraction,
aqueous effluent caprolactam concentrations as low as 30 ppm were obtained.
Further effluent treatment by activated carbon adsorption reduced the level to less
than 2 ppm. In contrast, the commercial multi-effect evaporation process is less
economical because much more water is vaporized and the condensate typically
contains up to 0.1-0.2 weight percent caprolactam.
Actual commercial plant samples were used for part of the work and gave
similar results to pure component mixtures, indicating that the plant samples
tested have no obvious deleterious effects on the process.
Various physical properties, which are needed for this process design, were
determined. In particular, vapor pressures for the key components were deter-
mined using a special high-temperature, low-pressure (vacuum) apparatus
designed specifically for this application.
This report is submitted in fulfillment of Contract No. R-803737 between
Union Carbide Corporation and the U. S. Environmental Protection Agency. This
report covers the period May 1, 1975 to February 28, 1978; work was completed
as of May 3, 1978.
iv
-------
CONTENTS
Foreword. iii
Abstract iv
Figures vi
Tables vii
Abbreviations and Symbols viii
Acknowledgements . ix
1. Introduction 1
2. Conclus ions 2
3. Recommendations 3
4. Experimental Procedures 4
Pilot Plant Preparation 4
Pilot Plant Description 4
Analysis for Caprolactam and DDP 7
5. Results and Discussion 8
Initial Evaluation 8
Pilot Plant Runs 9
Modified Process Flow Diagram 14
Further Treatment by Activated Carbon 14
Solvent Decomposition 15
Physical Property Data 16
References 44
Appendix 45
Glossary 46
v
-------
FIGURES
Number Page
1 Basic Union Carbide Corporation process for caprolactam
recovery from aqueous solution 20
2 Pilot plant flow diagram for the Union Carbide Corporation
caprolactam extraction process 21
3 Apparatus for obtaining liquid-liquid equilibrium data for
the water- caprolactam-DDP system 22
4 Distribution coefficient (Kp) for caprolactam between DDP
and water at 80°C 23
5 DDP viscosity versus temperature 24
6 Solubility of water in DDP versus temperature 25
7 Viscosities of DDP-beptane mixtures 26
8 Modified Union Carbide Corporation process for caprolactam
recovery from aqueous solution « 27
9 DDP density versus temperature • 28
10 Solubility of DDP in water versus temperature 29
11 Interfacial tension of the water-caprolactam-DDP system
at 80°C 30
12 V.L.E. apparatus 31
13 Caprolactam vapor pressure 32
14 DDP vapor pressure 33
15 Vapor pressures of caprolactam-DDP mixtures 34
vi
-------
TABLES
Number Page
1 Gas Chromatograph Parameters for Samples Analysis 35
2 Laboratory Distillation Operating Parameters 36
3 Pilot Plant Operating Parameters 37
4 Summary of Pilot Plant Efficiency Problems 38
5 Pilot Plant Data for Caprolactam Recovery 39-41
6 Batch Carbon Adsorption of DDP and Caprolactam at 80°C 42
7 Continuous Column Carbon Adsorption of Pilot Plant Secondary
Extractor Effluent from Run 5C at 80°C ; 43
8 Thermal Stability Tests of Stripped DDP 44
9 V. L. E. Apparatus Equipment 45
vii
-------
ABBREVIATIONS AND SYMBOLS
DDP --dodecylphenol
H. E.T. S. --height equivalent to a theoretical stage
I. D. - - inner diameter
KD --distribution coefficient
L. L. E. - - liquid- liquid equilibrium
Ml --microliter
O.D. --outer diameter
Temp. --temperature
V. L. E. - -vapor-liquid equilibrium
xp --caprolactam concentration in an aqueous feed
Xj_£ --caprolactam concentration in a lean solvent
Xp --feed mass flow rate
- -lean solvent mass flow rate
viii
-------
ACKNOWLEDGEMENTS
The cooperation of Allied Chemical Company's Hopewell plant personnel is
gratefully acknowledged. We are particularly indebted to Mr. Joseph A. Smith,
Manager of Plant Technology, for providing aqueous caprolactam samples and
to Mr. Harold Billingsly, Manager of Quality Control, for sharing analytical
procedures for caprolactam.
We are also deeply indebted to Mr. Roger Kidwell of Monsanto Company, St.
Louis, for sharing some of his plant experience with dodecylphenol manufacture
which has assisted us in pilot plant operations.
A special gas-chromatographic procedure, devised and operated by Mr.
Michael Biller, has been an invaluable tool for samples analysis. His assistance
in the bulk of the pilot plant runs and related experiments is also appreciated.
We are also indebted to Mr. Peter C. Dodson for designing and operating
the V. L. E. apparatus and assisting in pilot plant runs and related experiments.
ix
-------
SECTION 1
INTRODUCTION
Final product aqueous wash streams having caprolactam concentrations in the
range of approximately 1 to 25 weight percent result from the manufacture of
caprolactam and from the polymerization of caprolactam to nylon- 6. The more
concentrated streams are economically recovered by multi-effect evaporation.
However, there is no economic incentive to recover the more dilute streams by
this energy-intensive method. A properly designed, high-boiling solvent
extraction system can be less energy-intensive and extend the lower concentration
limit of economically recoverable solutions. Conventional low-boiling solvents,
such as benzene, are not economical because large amounts of solvent have to
be distilled.
Prior to this project, Union Carbide had developed a laboratory-scale
process for caprolactam recovery from aqueous solution. Basically, caprolactam
is extracted by DDP (a high-boiling solvent) and recovered from the extract by
distillation. The process was further developed during the current project
through laboratory testing with plant samples, pilot plant runs simulating the
process, and the acquisition of critical physical property data.
-------
SECTION 2
CONCLUSIONS
Pilot plant runs and related experiments have demonstrated the feasibility of
the DDP extraction process for recovering caprolactam from the aqueous streams
of production plants. This process is more economical than the currently used
multi-effect evaporation because less water is distilled (vaporized). Commerciali-
zation of this extraction process should extend the lower concentration of
economically recoverable aqueous caprolactam solutions from the current 3-5
weight percent limit of the evaporation process down to 0. 5 weight percent and
very possibly as low as 0.1 weight percent.
The caprolactam concentration in the effluent from Union Carbide Corporation's
extraction process is much lower than that of the evaporation process. Evaporator
condensates typically can contain as much as 0.1-0.2 weight percent caprolactam,
while the extraction process yields an effluent having as little as 30 ppm capro-
lactam or, if adsorption with activated carbon follows, less than 2 ppm. Carbon
treatment of the evaporator condensate may not be practical because the concen-
tration is considerably greater.
-------
SECTION 3
RECOMMENDATIONS
Additional refined V. L. E. data may be required for determining economical
stripping column operation. The data accuracy required generally increases with
the column (process) size.
More information on size and composition of recoverable aqueous capro-
lactam streams could perhaps better define further development needs for this
process. For example Allied Chemical claims to recycle all its aqueous
caprolactam streams and, therefore, does not need a recovery process at its
Hopewell caprolactam production plant. The stream sample we received from this
plant had only 30 ppm caprolactam. Even if recovery was needed, this is too dilute
for economical recovery by extraction or any other means.
-------
SECTION 4
EXPERIMENTAL PROCEDURES
PILOT PLANT PREPARATION
Pilot plant preparation included repairing parts and repiping and adjusting
equipment to accommodate this process (Figure 1). Because of a recent flood,
all pumps, gauges and electrical equipment near the floor had to be replaced or
repaired and cleaned. In addition, seals and joints, loosened by several years
of dormancy, were tightened or, in some cases, replaced. The glass distillation
column joints needed particular attention because of the high vacuum distillation
required. A cracked flange near the bottom leaked excessively during attempts to
evacuate the column. The flange was repaired instead of replacing the column
because much time would have been required to disassemble the packed column.
This repair, which consisted of a wooden plug secured by silicone rubber cement
and sealed with vacuum sealing putty, performed satisfactorily for vacuum
operation.
Some repiping and valving was done to accommodate the particular flow
scheme for this process, including proper connections between the primary and
secondary extraction columns and the distillation column. The plate spacing and
agitation amplitude in the extraction columns were adjusted for this process to
the values in Table 1.
PILOT PLANT DESCRIPTION
After preparation, the pilot plant functioned according to the flow diagram
in Figure 2. The initial flow scheme is discussed here, and later modifications
will be discussed when appropriate. A typical set of operating temperatures and
pressures is included for clarity. Aqueous caprolactam and heptane are con-
tinuously pumped from feed vessels into the process lines while DDP is charged
batchwise (Stream 10) into the bottom of the distillation column which functions
as a holdup vessel.
A centrifugal pump (P-401) circulates the aqueous feed (Stream 1) in the
feed drum (D-401) to enhance heating by external band drum heaters. The
centrifugal pump also pumps the aqueous feed from the drum to the piston pump
(P-402). This pump regulates the feed flow rate into the top of the primary
-------
extractor (C-401). Line heaters (not shown) between the pump and the column
automatically control the feed temperature at the column temperature. A
separate system controls the column temperature.
A second piston pump (P-405) at the base of the distillation column regulates
the solvent flow rate as it cycles between the primary extractor (C-401) and the
distillation column (C-403). DDF in the cycle section (Stream 3) from the bottom
of the distillation column to the bottom of the extractor is called lean solvent since
its caprolactam content is low. Inside the extraction column (C-401), the solvent
phase, which is lighter than the aqueous phase, travels upward in droplet form,
extracting caprolactam from the downward-flowing continuous aqueous phase.
The reciprocating action of perforated plates promotes the mass transfer by
forming and maintaining the droplets and minimizing concentration gradients in
both phases. Upon reaching the water-extract interface and coalescing at the
column top, the DDP is rich in caprolactam and is called the extract. The
extract (Stream 4) exits at the column top and proceeds to the distillation column
(C-403). A valve (V-403) controls the extract flow rate and maintains a positive
extract line pressure. Without it, the distillation column would evacuate the
extract line, creating more places for air to leak into the column. This valve is
adjusted by visual observation of a pressure gauge. Manually controlled extract
line heaters raise the extract temperature to the distillation column feed point
temperature. A manually controlled heating jacket surrounds the column base and
supplies essentially all the heat for the distillation. Generally kept constant, "the
jacket control is varied only when the column operating parameters are changed.
Insulation and manually controlled heating tape minimize column heat losses
which otherwise would be very large for this high temperature distillation.
The packed distillation column fractionates caprolactam and water from
the higher boiling DDP. A solenoid-operated reflux splitter controls the
fraction of condensate refluxed to the column. The splitter is adjusted manually
according to visual observation of the sensitive feedpoint temperature. As the
temperature increases above its control point, more of the lower-boiling
condensate is refluxed to lower the temperature and vice versa. The column
vacuum, provided by a two-stage pump, is maintained to within 1 mm Hg by
mercury manometer-controlled solenoids.
The feedpoint is located at the upper section of the column to provide
sufficient staging for the difficult caprolactam-DDP separation which occurs
below it, Some of the water-DDP separation is achieved-here also. Purified DDP
exits the column bottom as lean solvent. Performance of this section and,
therefore, the lean solvent purity, is sensitive to water concentration at the feed-
point and in the column section above it. A concentration increase, due to an
increased extract flow rate or to extract water content fluctuations, results in a
higher caprolactam content of the lean solvent. Lean solvent samples are
obtained by applying vacuum suction to the lean solvent (Stream 5) coming from
the column bottom.
-------
A piston pump regulates the lean solvent flow rate (Stream 3) and pressurizes
it from the vacuum pressure (100 mm Hg) of the distillation column to the
extractor pressure (our centrifugal pumps do not function at a low suction
pressure of 100 mm Hg), returning it to the primary extractor and completing
its cycle through the pilot plant. No lean solvent cooling is necessary as line
heat losses easily cool it from the high distillation temperature (>250°C) to the
extractor temperature (80°C).
The aqueous phase, upon reaching the bottom of the extraction column, is
depleted of caprolactam and is called the primary raffinate (Stream 6). It
contains a small amount of dissolved and entrained DDP. Quite often, the
entrained fraction is the larger one. It consists of extremely small DDP droplets
whose upward settling (buoyant) velocities are less than that of the oppositely-
moving aqueous phase. Therefore, these droplets are carried downward (en-
trained) with the aqueous phase. A given degree of column agitation produces a
large range of droplet sizes, with the majority large enough to rise against the
aqueous flow as required for proper extractor operation. The droplet size
distribution broadens with increasing column agitation. In this way, more of the
smaller, entrainable droplets are produced, undesirably increasing the total DDP
content of the aqueous raffinate. The operating strategy is to agitate as much as
possible to enhance mass transfer without excessive entrainment.
The primary raffinate proceeds to the top of the secondary extraction column
(C-402) where heptane extracts DDP from it. A small amount of any remaining
caprolactam is also extracted. Upon reaching the column bottom, the aqueous
phase is depleted of DDP and is called the secondary raffinate (Stream 7). A
manually-adjusted valve (V-406) maintains the column pressure and regulates the
secondary raffinate flow rate from the column. This stream is the aqueous
effluent from this process and may contain as little as 10 ppm caprolactam, 10
ppm DDP, and 100 ppm heptane. Further concentration reduction of these
compounds by activated carbon will be discussed in Section 5, Further Treatment
of Activated Carbon.
Heptane (Stream 8) is gravity-fed to a piston pump that regulates its flow
rate to the bottom of the secondary extraction column. With the heptane phase
dispersed, this column operates similarly to the primary extraction column. At
the column top, the heptane droplets, high in DDP content, coalesce to form the
secondary extract (Stream 9). A manually-ope rated valve (V-404) maintains the
column pressure and regulates the flow rate from the column to a storage drum
(D-403), the terminal point for this stream. In an actual process (Figure 1),
the secondary extract would be distilled to recover heptane overhead for recycle
to the secondary extractor. The DDP bottoms from this distillation would be
recycled to the primary extractor. This distillation was not demonstrated in our
pilot plant because: a) it is a relatively simple one due to the large boiling point
difference between heptane (98°C) and DDP (325'C), and b) substantial pilot
plant modification is needed to include this distillation.
-------
ANALYSIS OF CAPROLACTAM AND DDP
A Perkin-Elmer gas chromatograph and an Autolab minigrator were purchased
for analyses of the hundreds of samples generated by this project. Allied
Chemical provided operating parameters for their caprolactam analysis which
aided in the establishment of procedures to analyze dilute aqueous solutions of
caprolactam and DDP and dilute solutions of caprolactam in DDP. Operating
parameters are shown in Table 2. Flame ionization detection and an internal
standard, acetophenone, provide the sensitivity and accuracy needed in the ppm
concentration range.
-------
SECTION 5
RESULTS AND DISCUSSION
INITIAL EVALUATION
Before pilot plant work was begun, an aqueous plant sample, dilute in capro-
lactam, gave extraction and distillation results in good agreement- with pure com-
ponent data. This evaluation was performed to determine whether other impurities
(if any) in the plant stream significantly affected the extraction or distillation parts
of this process. The sample was obtained from a process column overhead of an
Allied Chemical Company plant, a major U. S. caprolactam producer. Our
analysis showed this sample to contain 30 ppm caprolactam. To avoid the
analytical error at this low concentration, additional caprolactam was added to
increase the concentration to 600 ppm.
The extraction was performed at 80° C in the apparatus shown in Figure 3
wherein equal parts by weight of the aqueous solution and DDP were contacted and
allowed to reach equilibrium. Analysis of raffinate and extract phases by gas
chromatography showed that 95 percent of the caprolactam was extracted from
the aqueous phase. For this project, the primary extract and raffinate phases
are the DDP (solvent) and aqueous phases, respectively, at equilibrium with
respect to the caprolactam content. The distribution coefficient
_ weight percent caprolactam in the extract
D ~ weight percent caprolactam in the raffinate
for this extraction was calculated to be about 20. This value is in good agreement
with our laboratory data previously obtained using pure water, caprolactam, and
DDP (Figure 4). Therefore, it was concluded that impurities present in the plant
samples tested have no significant effect on the extraction part of this process.
The extraction distillation was performed in a 2"-I. D. Oldershaw column
at 45 mm Hg pressure. Operating parameters are given in Table 2. The
caprolactam content of the extract was reduced to 592 ppm. This is in good
agreement with laboratory distillation data using pure water, caprolactam, and
DDP, as shown in Table 3. With the plant sample evaluation completed and
satisfactory, the pilot plant phase was begun.
8
-------
PILOT PLANT RUNS
Solvent Stripping
Several batches of commercial DDP were stripped of light ends in the dis-
tillation column to remove lower-boiling alkylphenols which form maximum
boiling azeotropes with caprolactam. If not removed, they would reduce the
efficiency of the caprolactam-DDP distillation during the pilot plant runs. Each
batch was stripped at 40 mm Hg and 250°C for about 8-10 hours.
During these distillations, useful column operating experience was obtained.
Excepting the vacuum system, the column is controlled manually, which
requires some familiarity. At the low pressure of 40 mm Hg, we found that the
column can easily be flooded by high vapor rates and we employed a maximum
jacket heat input to prevent flooding. A preliminary heat balance revealed a
large heat loss through the insulated column wall. To help to compensate for this,
the column heaters were adjusted to near-maximum output.
The bottoms product from each distillation was saved for use as a solvent.
Gas chromatograms showed reduced light ends peaks of the bottoms products
similar to those of laboratory-stripped DDP, indicating that the pilot plant-stripped
DDP was suitable for use.
First Process Run Series
During this 168-hour run series, caprolactam recovery was good but less
than expected. A summary of problems and suspected causes is shown in Table
4. The distillation column was started up first, as it required several hours to
reach equilibrium. Then the primary and secondary extraction columns were
started, putting the process into full operation. The operating parameters and
stream conditions for this run series are summarized in Table 5, Columns
(Runs) 1A-1D. For each parameter set, the pibt plant was allowed to equilibrate
for at least 1 hour before samples were taken. This allowed each component
(e.g., distillation column) to equilibrate individually, but not with the whole
pilot plant. This would have required several hours and is neither practical
nor necessary for meaningful results.
For Runs 1A-1D, the caprolactam-supplemented 600-ppm plant sample
from Allied was used. The best caprolactam recovery is about 89 percent
(Run 1C). Using literature methods (1), the height equivalent to a theoretical stage
(H.E. T. S.: a measure of extractor efficiency) is calculated to be about 35 inches.
Literature data (2) for the Karr column show the H. E.T.S. is usually below 15
inches. The initially suspected cause of this low extractor efficiency was the
lean solvent which contained about 1500 ppm caprolactam -- about three times
that of the laboratory distillation bottoms. A high lean solvent caprolactam
concentration may reduce caprolactam recovery by limiting the theoretical
-------
equilibrium recovery attainable.
^H ~v^T S' T))
Theoretical Equilibrium Recovery = —— x 100%
XF
XF
When ~—
XLS
Where x^g = concentration of caprolactam in the lean solvent
Xp = concentration of caprolactam in the feed
XLS = lean solvent mass flow rate
X = feed mass flow rate
Initially suspected causes of the poor lean solvent purity were insufficient
solvent stripping of light ends and insufficient column staging. Attempting to
correct for the latter, column pressure and temperature were increased in Runs
1C and ID. Laboratory distillation experience prior to this project suggests the
relative volatility (ease of separation by distillation) between caprolactam and
DDP increases with increasing temperature. Doing so produced a small but
insufficient improvement in lean solvent purity.
Additional Solvent Stripping
To determine whether insufficient light ends removal had caused the poor
lean solvent purities (the mechanisms for which were discussed in Solvent
Stripping) in Run 1, the solvent was restripped semicontinuously. This differed
from the simpler batch stripping in that the bottoms product was drawn off during
the distillation rather than after the distillation had been shut down, which
allowed the column liquid, possibly containing some light ends, to drain into the
bottoms product.
After all the bottoms were removed, the column was shut down and cleaned
with isopropanol to wash away any remaining light ends. The restripped solvent
was charged to the column bottom for a second process run.
Second Process Run Series
In spite of additional solvent stripping and other parameter changes, results
from this series of runs are very similar to the first. For Runs 2A and 2B,
the feed caprolactam concentration is the same as in Run Series 1. The lean
10
-------
solvent purity and caprolactam recovery are the same as or less than in Run Series
1, which indicates that the additional solvent stripping has little effect. In contrast
to Run Series 1 results, a higher distillation column pressure (83. 5 mm Hg
versus 38. 5 mm Hg) did not improve the lean solvent purity in 2B over that of 2A,
which indicates that another parameter has a greater effect on the recovery.
To further investigate the caprolactam recovery problem, we switched to a
feed containing about 1 percent caprolactam prepared from pure water and capro-
lactam. If lean solvent purity had been limiting the recovery, then this higher
concentration should improve the recovery. For Run 2F, the recovery is 93
percent -- somewhat better, but still not as good as expected. For this run, the
column agitation is lower than for the previous runs because, at the higher
caprolactam concentration, the water-caprolactam-DDP system is more prone to
entrainment. This is evidenced by the high DDP content of the primary raffinate
in 2D and 2E where greater agitation was used. This condition is not practical
and is not recommended for proper process operation.
Changing the primary extractor dispersed phase had little effect on caprolactam
recovery. The high viscosity of the solvent (27 c. p. - Figure 5) was suspected to
cause low mass transfer and low recoveries. To provide more agitation to over-
come this effect, the solvent was made the continuous phase. However, Run 2G
shows essentially no improvement over 2F; as will be shown later, this method is
simply not sufficient to overcome the viscosity effect.
Third Process Run Series
Raising the column temperature is found to be an unsatisfactory method of
reducing the solvent viscosity. It had been planned to operate the column at
127°C where the solvent viscosity is about 5 c.p. (Figure 5), a viscosity at which
satisfactory literature data had been obtained with other systems (2). Upon
beginning this run, the column heaters were soon discovered to be inadequate to
achieve this temperature; therefore, we obtained Run 3A at 92°C and then
stopped this run series. Essentially no recovery improvement was noted with
this incremental temperature change. .The column heating capability was not
improved because subsequent laboratory solubility experiments showed that
the water solubility in DDP was too high at 127°C (Figure 6). Higher water
concentrations of the extract increase the load on the distillation column.
Fourth Process Run Series
Extraction Runs- -
The primary extractor efficiency increases significantly when the solvent
viscosity is reduced by adding heptane. Laboratory viscosity measurements of
DDP-heptane mixtures show that the addition of 10-30 weight percent heptane
markedly reduces the viscosity of DDP (Figure 7). A 30 percent heptane/70
11
-------
percent DDP mixture was employed as the solvent for Runs 4A-4F. The dis-
tillation column was not operated for these runs because there was no available
method for providing good and constant ratio mixing of the lean solvent with heptane.
Instead, a batch of this mixture was prepared before the run and fed from a drum
to the solvent flow controlling piston pump. The primary extract was simply fed
to the dormant distillation column where it was stored for a subsequent batch
distillation. The heptane-DDP flow rate was adjusted to be equivalent to 32 ml
DDP/min, the flow rate used in 2D-2G. Except for this adjustment, the operating
parameters of 4A, 4B and 4C are identical to those of 2D and 2E. The caprolactam
recovery is much higher, presumably due to heptane's viscosity-reducing effect.
At the lower temperature of 60°C, Runs 4D-4F gave similar caprolactam re-
coveries but resulted in much less DDP in the primary raffinate. This is because
the higher interfacial tension between the solvent and aqueous phases at the lower
temperature decreases the solvent phase breakup into the very small, entrainable
drops.
Extract Batch Distillation- -
This distillation yielded a much better lean solvent purity than did the previous
continuous ones. The improvement is attributed to the additional column staging
made available by the batch method and to the separate removal of water (and
any other low-boiling compounds present) which hinders the more difficult
caprolactam-DDP separation. This distillation, being batch, necessarily had to
be run following the extraction run. The objective was to demonstrate whether
distilling water (and heptane) first, and then caprolactam, would result in a better
lean solvent purity. This separation could also be performed continuously using
two columns, however the extra column was not available in our pilot plant.
The extract from Runs 4A-4F was, therefore, batch-distilled under the usual
operating conditions as shown in 4G-4M. Heptane and water were the first
components to come overhead. Initially, the pressure was maintained at 100 mm Hg.
to assure adequate condensation of the 30°C aqueous overhead. As the dis-
tillation progressed and the head temperature increased because of the buildup
of higher-boiling light ends, the column pressure was reduced to 75 mm Hg.
After most of the water and caprolactam was^ removed, an intermediate fraction
was removed which consisted of caprolactam, water, heptane, and light ends.
The light ends source would be the plant feed and/or oxidative decomposition
products from the distillation column (solvent decomposition will be discussed
later).
•4
In Run 4L, a pure caprolactam fraction was obtained as evidenced by the
constant head temperature of 187°C. Finally, the head temperature began rising,
indicating that most of the caprolactam was removed and the higher boiling DDP
solvent was beginning to come overhead. The distillation was continued until the
temperature rose to 235°C (Run 4M), significantly above the caprolactam
overhead temperature, assuring a good caprolactam removal from the column
bottoms. The bottoms analysis (4M) showed 466 ppm caprolactam, a large
12
-------
reduction below the 1000-3000 ppm range typical of the previous runs and agreeing
well with the laboratory value of 592 ppm.
Fifth Process Run Series
For this final process run, caprolactam recoveries were the highest of all
runs. However, a dispersal phase change for the secondary extractor failed to
improve its efficiency as expected. For these runs, the primary extractor was
operated as in Runs 4D-4F, the only difference being the lower lean solvent
caprolactam concentration. The greater caprolactam recoveries are attributed
to this. Although the total extracted amount is not much greater, the primary
raffinate concentration is about half that of Runs 4D-4F.
The secondary extractor dispersed phase was changed to aqueous in an
attempt to improve the efficiency. Earlier Runs 1C-ID gave good DDP recovery
when the heptane flow rate was at the "higher", but still very economical, rate of
9 ml/min, leaving as little as 9 ppm DDP in the secondary raffinate. However,
cutting the heptane rate to 7 ml/min greatly reduced the DDP recovery as seen in
the remaining runs. Extraction data obtained several years ago, and our
extraction experience, suggest that distribution coefficients for DDP between
heptane and water should be greater than 1000/1 in favor of the heptane. Therefore,
it was reasoned that the 1/18 heptane-to-water ratio (at 9 ml heptane/min) is
more than sufficient to extract DDP.
The secondary extraction column was suspected of being too large, thus
resulting in an insufficient interfacial area. Karr column literature data (2)
show a maximum combined phase flow rate of about 1000 gph/ft2. In contrast,
the secondary extractor combined rate is 123 gph/ft2. Of this, the individual
heptane rate is only 6.5 gph/ft2. Visually, the dispersed heptane droplets are
very sparse. In addition, the large density difference between heptane and
water causes the droplets to travel quite quickly up the column. Two methods
of slowing the droplet ascent are by producing smaller, slower-ascending
drops and by increasing the continuous phase velocity. Smaller droplets could
not be formed because extreme agitation is required to overcome the heptane/
water interfacial tension (36. 2 dynes/cm at 20°C). A smaller diameter column,
which is not available in our pilot plant, will increase the continuous phase
velocity and thereby reduce the droplet velocity to the column. In this way,
droplet residence time and column efficiency can be increased.
Instead, an efficiency improvement was attempted by changing the dispersed
phase. Having a flow rate 18 times larger than the heptane phase, the aqueous
phase, when dispersed, should provide much more interfacial tension area
and increase the column efficiency. However, as seen in Runs 5A-5D,
essentially no DDP was extracted. This was partially due to the preferential
wetting of the glass column walls by the aqueous drops. During the runs, the
bulk of the aqueous phase was seen cascading down the column wall in spite of
13
-------
maximum agitation. The droplets which were produced were large and, therefore,
resulted in small interfacial area (mass transfer area).
A secondary cause of the poor recovery was that water viscosity is about 5
times that of heptane. This reduces the mass transfer in the dispersed phase.
Modified Process Flow Diagram
The above results suggest an improved process flow scheme shown in Figure
8. Schematically, the primary and secondary extractors in Figure 1 are combined,
and an additional distillation column has been added. This process incorporates
heptane addition to DDP for increased extractor efficiency and distilling water
separately from caprolactam which improves the distillation column efficiency.
The flow scheme is similar to the original process. The aqueous caprolactam
feed enters at the extractor top and flows downward. DDP enters at the extractor
midsection as lean solvent and flows upward, extracting caprolactam from the
aqueous phase. Heptane enters at the extractor bottom and flows upward,
extracting DDP from the aqueous phase. The heptane-DDP extract combines with
the lean solvent at the extractor midsection and, in so doing, reduces the solvent
viscosity. The extract exits at the extractor top and contains DDP, heptane,
caprolactam, and some water. Heptane and water are distilled from this
extract in the drying column. Being very immiscible with one another, heptane
and water in the overhead are separated by decanting. Part of the heptane is
refluxed to the drying column; the remainder is recycled to the extractor column.
The water from the decanter is combined with the extractor raffinate, forming
the process effluent.
DDP and caprolactam constitute the drying column bottoms product which is
fractionated into caprolactam (overhead) and DDP (bottom) in the stripping column.
Further Treatmentby. Activated Carbon
Batch and continuous column tests show that activated carbon adsorption is a
good method for removing the trace amounts of caprolactam, DDP and heptane
in the secondary extractor raffinate. This polishing step can be used when very
low effluent concentrations are desired.
As discussed previously, pilot plant runs have produced an aqufous stream
containing as little as 30 ppm caprolactam, 10 ppm DDP, and 20 ppm heptane.
These levels are very low for a recovery process. Although it should be possible
to reduce the DDP concentration by increasing the secondary extractor length,
it offers little overall improvement since the caprolactam and heptane concentrations
are not reduced by this parameter change. The lean solvent purity (caprolactam
content) lower limit appears to be about 500 ppm. This, in turn, places a lower
14
-------
limit on the caprolactam concentration in the primary extractor raffinate. The lower
limit of heptane in the secondary raffinate is simply the solubility of heptane in
water.
To learn how effective carbon adsorption is in further reducing the above
concentrations, batch adsorption experiments were conducted on dilute aqueous
solutions of caprolactam and DDP. Heptane was not included here, as hydrocarbon
adsorption by activated carbon is well known. The results are shown in Table 6.
The distribution coefficient measures how well caprolactam and DDP are adsorbed.
It is the ratio of the weight percent material adsorbed on the carbon divided by the
weight percent material in the aqueous solution at equilibrium. The former is
calculated by dividing the total material recovered from the aqueous solution by
the sum of the total weight of carbon employed in the experiment and the removed
material weight. For most of the experiments, the solution concentrations are
below (probably significantly below) the detection limits as indicated. Although
this does not allow the calculation of exact distribution coefficients, it does show
that the amount of carbon needed to treat the secondary raffinate is very small.
A continuous column test (results shown in Table 7) shows similar results. Here
the starting solution is the secondary raffinate from pilot plant Run 5C.
Solvent Decomposition
Oxidative decomposition darkened the solvent but did not reduce its extraction
capacity. The glass distillation column had numerous joints and, although all
were tightened, air leakage into the column, combined with the high distillation
temperature, was a suspected cause of solvent discoloration.
Decomposition experiments verified the decomposition to be oxidative and not
simply thermal. To determine this, the air leakage into the distillation column was
measured and the total amount calculated for 168 hours' (1 week's) operation.
This amount was divided by the total amount of solvent in the pilot plant system.
A mixture of the above ratio was prepared using pure DDP and heated to 250°C
for 168 hours. A second sample of DDP and nitrogen was heated identically.
The results are impressive (Table 8). The air-containing sample turned dark
brown and was very similar in appearance and color to the darkened pilot plant
solvent. The second sample, containing no oxygen, was essentially unchanged
in color.
These results clearly indicate that DDP decomposition is caused by the
presence of oxygen at elevated temperatures. Interestingly, the air leakage rate
for our distillation column falls within the limits for typical "commercially tight"
units. However, it is possible to design a commercial unit to be practically
airtight. DDP is commercially purified by a distillation in such a column
(according to personal communication with Mr. Roger Kidwell, Monsanto).
15
-------
A laboratory extraction using pilot plant lean solvent having about 300 hours'
operating time showed results similar to those obtained with pure DDP (Figure 4).
Although these results do not guarantee that solvent oxidation, if allowed to occur
at rates experienced in our pilot plant over a longer period of time, will not reduce
the solvent extraction capacity, the results do indicate that any effects are very
gradual.
Physical Property Data
Other physical properties needed for process design that were measured are
DDP density, DDP solubility in water, and interfacial tensions for the water-
caprolactam-DDP system. The results are shown graphically in. Figures 9,
10 and 11, respectively. The experimental methods employed are given in the
Appendix.
In addition, consistent, but very abnormal, pressure data were obtained
for the caprolactam-DDP mixtures. Pure component and binary vapor pressure data
are needed for the design of distillation columns. For this process containing
water, caprolactam, DDP and heptane, there are four pure components and six
binary pairs. However, it is not necessary to obtain data for which sufficiently
accurate data are available in the literature or for pairs that have large boiling
point differences for which calculated ideal vapor pressure data are sufficient.
Heptane, water, caprolactam, and the caprolactam-water pair and heptane-water
pair fall into the first category. The heptane-caprolactam, heptane-DDP, and
water-DDP pairs fall into the second category. No literature data are available for
DDP or the caprolactam-DDP pair. These and caprolactam vapor pressure data
are the key components needed for the design of the system. Therefore, these
vapor pressure data were obtained, including those for caprolactam, since
precision can be lost when key vapor pressure data are obtained from different
sources.
A special apparatus was built for this purpose (Figure 12). The key
apparatus components are listed in Table 9. It was designed to operate up to
340°C and at pressures from 10"^ psia up to 100 psia. A differential pressure
transducer measures the pressure difference between the sample and a reference
pressure. For our experiments, the reference is a vacuum source that was low
enough to be considered 0 psia. Therefore, the transducer gave direct sample
pressure headings. The data for this project were taken at high temperatures
and low pressures, requiring that the apparatus used have several features. To
minimize air leakage into the system, most connections were welded. Special
vacuum fittings were used for the few connections which had to be separable.
To provide a uniform temperature, the apparatus was enclosed by an oven. The
differential pressure transducer sensor also had to be in the oven to maintain it
at the system pressure so that the sample vapors do not condense in the transducer
sensor. This condensation is undesirable because a) it allows fractionation of
16
-------
the sample which changes its composition and b) liquid(s) in the sensor, especially
polar ones, cause inaccurate, nonreproducible readings because they interfere
with the capacitor-type sensor. After operating the apparatus, we discovered
that, for the more polar caprolactam, maintaining the sensor at the system
temperature was not sufficient to prevent this sensor malfunction. Heating tape,
wrapped around the sensor, maintains the sensor temperature 40°C above the
system temperature, eliminating the problem.
During operation, once the system temperature is reached, the sensor
requires two hours to thermally equilibrate because the sensor generates some
heat and is temperature-sensitive. The system has to be evacuated for at least
30-60 minutes to remove all the air and, during operation, shorter evacuations
have to be performed about every 2 hours to remove air that has leaked into the
system. The evacuation time is kept approximately inversely proportional to the
system temperature to minimize sample loss. Generally, the evacuations are
done as long as possible without noticeable sample buildup (sample loss) in the
dry ice trap (not shown) located along the vacuum line between the oven and the
vacuum pump (not shown).
The pure component data for caprolactam and DDP are shown in Figures 13
and 14, respectively. The caprolactam data agree reasonably well with literature
data (8). The vapor pressure data for caprolactam-DDP mixtures shown in
Figure 15 are very consistent with one another. However, when compared to
the pure component data, the mixture data are unexpectedly high in the low
temperature range (100-175°C) and somewhat low in the high temperature range
(230-250°C). This does not necessarily mean that the data are invalid or that
they are accurate enough for distillation column design, however, as process
scale increases, so does the data accuracy required for economic process
design.
17
-------
Dodecylphenol,
Caprolactam.
FEED
Water,
Caprolactam
PRIMARY
RAFFINATE
Water.
. 01-. 1% Dodecylphenol
He
Do
EX
SECONDARY
RAFFINATE
Water,
< 100 ppm Heptane
T
Water
^v
*RTMA
XTR>
RY
CTOR
Caprolactam,
Water
PRIM; .RY
DISTD .LATION
COLU
Icodecylphenol- < solvent)
^ph
)ND,
?AC
enol
LRY
TOR
SEO
DIS1]
COI
f "N
)NDAI
ILLA1
UMN
Y
ION
Steam
Dodecylphenol
Heptane
Figure 1. Basic Union Carbide Corporation process for Caprolactam
recovery from aqueous solution.
18
-------
H
VO
EXTRACT FLOW
CONTROL VALVE
V-403
CONTROL VALVE
V-404
EXTRACT
PRESSURE
CONTROL
VALVE
V-«5
FEED PRESSURE
CONTROL VALVE
V-401
TO VACUUM
SYSTEM
'CUNDARY SECONDARY
DODECY
DRUM
D-404
.RAFFINATB
'FLOW CONTROL
VALVE
V-406
DISTILLATION
COLUMN
C-403
TO VACUUM SYSTEM
FEED FEED • j.
DRUM cmcuu riNO,—T]
D-401 PUMP
P-401
LEAN ' LEAR
SOLVENT SOLVENT
SAMPLER DRUM
D-407 D-40S
HEPTANE
FLOW CONTROL
PUMP
P-403
LEAN SOLVENT
FLOW CONTROL
Figure 2, Pilot plant diagram for the Union Carbide Corporation caprolactam
extraction process.
-------
NEEDLE VALVE
NEEDLE TAPER
-1/4" SS TUBE (SAMP. UPPER PHASE)
1/8" SS TUBE (SAMP. LOWER PHASE)-
GLASS TUBE (HIGH PRESS.)
PRESSURE UP TO 500 psi
TEMPERATURE UP TO 200°C
Figure 3. Apparatus for obtaining liquid-liquid equilibrium data
for the water-caprolactam-DDP system.
20
-------
100.
.
g 40 .
t 30 „
20 *
10 -
s -
4 •
3 .
2 -
A
Key:
pure component data
0 plant sample
A "decomposed" solvent
.001
10 ppm
I I
.005 .01
100 ppm
.05 .1 .5 1
1000 ppm
WT. % CAPROLACTAM IN RAFPINATE
10
Figure 4. Distribution coefficient CO for caprolactam between DDP and water at 80° C.
-------
ft,
cj
50 .
40 .
30 .
20
10
g
>
5
4
3 .
60
70
80
90
100
110
l3o~
130
140
TEMPERATURE, °C.
Figure 5. DDP viscosity versus temperature.
-------
§
2
fc
3
s
OS
w
2 -
1 -
20
40 60
Temperature °C
80
O
100
120
Figure 6. Solubility of water in DDF versus temperature.
23
-------
w
CO
O
i
s
40 60 80
WT. % HEPTANE
100
Figure 7. Viscosities of DDP-heptaue mixtures.
24
-------
Heptane
Dodecylphenol,
Caprolactam ,
Heptane ,
FKKD
Water.
Caprolacta
[Water
n
EXTRA<
TOR
i
— >l
^
J
/7i
/Jo'
>
DRYIN
COLU?
H
V.
3
IN
•*~»
KAFriXATE
Water,
< 100 ppm Hydrocarbon
^
Caprolactam
Dodecylphenol
STRIP PING
COLU
Caprolnctam
\IN
figure 8. Modified Union Carbide Corporation process for caprolactam
recovery from aqueous solution.
25
-------
H
53
.96.
.94
.92.
.90
.88.
.86
.84
.82
.80
.78
0
20 60 100 140 180 220 260 300
TEMPERATURE °C
Figure 9. DDP density versus temperature.
26
-------
K
a
8
Q
J
§
M
0,
160
140
120
100 .
80-
60 „
40
20-
20
©
40 60 80
TEMPERATURE°C
100
120
Figure 10. Solubility of DDP in water versus temperature.
27
-------
i
00
w
H
c;
CD
8 .
6 .
4 •
2 •
12345
EQUILIBRIUM WT. % CAPROLACTAM IN THE AQUEOUS PHASE
Figure 11. Interfacial tension of the water-
caprolactam-DDP system at 30°C.
28
-------
Electronic Manometer
Valve
:Air Dr;Lyen
Magnetic
Stirrer
Figure 12. V.L.E. apparatus.
29
-------
2
•
u
cs
CS
CS
2000.
1000-
500.
400.
300.
200.
100-
50.
40.
30.
20.
10.
5.
4.
3.
2.
100 120 140 160 180 200 240 280 320
\TEMPERATURE0C
Figure 13. Caprolac.ta*a vapor pressure.
30
-------
w
OS
o
200
100
50
40
30
20
10
5
4
3
2
1 J
.5
.4
.3
,2 4
O
100
120
180
240
140 160 180 200
TEMPERATURE°C
Figure 14. DDP vapor pressure.
31
280
320
-------
in
V)
at
OS
a,
o£
2
300
200
100 .
50
40
30
20
10 -
5
4
3
2 .
.5
.4,
.3
.2.
Key:
£
0
0.5 vrt. % caprolactam in dodecylpbenol
1.0 wt. % caprolactam in dodecytphenol
7.5 wt. % caprolactam in dodecylphenol
. pure caprolactam (Figure 13)
- pure dodecylphenol (Figure 14)
100 120 140 160 180 200 240 280 320
TEMPERATURE °C
Figure 15. Vapor pressures of caprolactam-DDP mixtures.
32
-------
TABLE 1. GAS CHROMATOGRAPH PARAMETERS
SAMPT.FS ANALYSTS
Sample Types
Column Material
Length
Diameter
Liquid Phase
Wt. %
Support
Mesh
Carrier Gas
Inlet Press psig
Rate, ml. /min.
Sample size, jil
Detector
Sensitivity
Hydrogen Press psig
Air Press psig
Detector Temp. °C
Inlet Temp. °C
Column Initial Temp. °C
Column Final Temp. °C
Rate
Solvent
Internal Standard
Raffinates, Feeds
Stainless Steel
18 in.
1/8 in. O.D.
CARBOWAX 20M
5
Gas Chrom Q
80/100
Helium
91
20
0.8
Flame lonization
X10
18
50
300
250
170
170
-
Isopropanol
Acetophenone
Extract"6!. Lean Solvents
Glass
6ft.
2 mm I. D.
-
-
Tenax GC
60/80
Helium
91
20
as
Flame lonization
X100
18
50
350
300
150
320
8% min.
Isopropanol
Acetophenone
33
-------
TABLE 2. LABORATORY DISTILLATION OPERATING PARAMETERS
Column Oldershaw, glass
Sieve plate spacing 2"
Number of plates 30
Column diameter 2"
Column head pressure 45 MM Hg
Final bottoms temperature 243 °C
Caprolactam content of the bottoms product 592 ppm
Caprolactam content from similar distillation
using pure Caprolactam, dodecylphenol & water 520 ppm
34
-------
TABLE 3. PILOT PLANT OPERATING PARAMETERS
Primary Extractor
Column Type
Shell Material
Plate Material
Column Diameter
Plate Spacing
Number of Hates
Agitation Amplitude
Secondary Extractor
Column Type
Shell Material
Plate Material
Column Diameter
Plate Spacing
Number of Hates
Agitation Amplitude
Distillation Column
Column Type
Shell Material
Packing
Packing Height
Column Diameter
Karr Reciprocating Plate
Glass
Teflon
2 Inches
2 Inches
45
. 75 Inch
Karr Reciprocating Plate
Glass
Teflon
2 Inches
llnch
63
.5 Inch
Q.V.F.
Glass
1/4 Inch extruded protruded stainless steel
15 Feet
4 Inches
35
-------
TABLE 4. SUMMARY OF PILOT PLANT EFFICIENCY PROBLEMS
Problems
Suspected Cause
Action Taken
Results
Extractor efficiency
H.E.T.S. - 35 in.
Lean solvent
purity
Secondary
extractor
efficiency
Solvent
"decomposition"
Poor lean solvent
purity 1000-3000 ppm
Solvent viscosity
Light ends
Insufficient staging
Oversized column-
insufficient inter-
facial contact
Air leakage
into column
Additional stripping, No change
semi-continuous
Increase feed cone.
600-5280 ppm
Change dispersed
phase
Increase tempera-
ture
Add 30% heptane
Batch distillation
Change dispersed
phase
Decomposition
tests
Laboratory
extraction with
"decomposed"
solvent
No change
No change
No change
98-99%
recovery
H.E.T.S.-16 in.
LS-466 ppm
No recovery-
column wetting,
higher viscosity
of the dispersed
phase
Oxidative
decomposition
Extraction
performance
unaffected
36
-------
TABLE 5. PILOT PLANT DATA FOR CAPROLACTAM RECOVERY
lo
Run Number
Primary Bgractor
Leu Solvent Flownte ml. /mln.
PPM Caprolactam In Leu Solveac
Feed Flownte mU /mln.
PPM CaproUctam In Peed
Diapered Phue
Column Temperature *C
Agitation Speed, etrokea/mln.
Agitation Amplitude, In.
PPM CaproUcum In Rafflnate
PPM OOP In Rafflnate
Wu t Water In Ettna
Wti % Caprolactam In Batraet
% Caprolactam Recovery
Secondary Extractor
Heptane Flownte, mU /mln.
Dtipereed Phase-
Column Temperature *C
Agitation Speed, atrokei/mln.
Agitation Amplitude, In.
PPM Ciprotactam In Rafflnate
PPM OOP In Roinnue
PPM Heptane In Rafflnate
PPM DDP In Extract
PPM Caprolactam In Extract
DlatUlatlon Column
Head Temperature *C
Feed Point Temperature *C
Bottom Temperature C
Head Prtnure, MM Hg
Reflux Ratio (reflux/total overhead)
Overhead
1A
15.2
1330
161
600
DDP
80
340
.4
87. S
1067
0.523
85
5.0
Heptane
SO
530
.5
57.8
21.2
-
•
-
35
50
249
40
.95
IB
20.0
248»
161
600
DDP
80
340
.4
109.5
1510
-
.590
82
7.0
Heptane
50
530
.5
118.2
1550
-
3630
67
34
205
238
38
.60
1C
20.0
1644
161
600
DDP
80
340
.4
68
124
-
.530
89
9.0
Heptanr
50
530
.5
242.5
30
-
2500
518
41.5
220
24)
49
.57
ID
20.0
1406
161
600
DDP
80
340
.4
655
94
-
.643
-
9.0
Heptane
SO
530
.5
176.8
<9.2
-
2650
549
42
218
246.7
61.5
.58
Water
2A
20.0
968
161
600
DDP
80
340
.4
115
1457
-
.525
81
7.0
Heptane
50
530
.5
48
305
-
High
270
45
206
241
38.5
.75
and Caprotaetam
28
20.0
1970
161
600
DDP
80
340
.4
138
5145
-
-
77
7.0
Heptane
50
530
.5
98.8
275
-
High
63.7
44.7
106
250
83.5
.70
2C
20.0
1689
161
100
DDP
80
340
.4
64
2535
-
1600
36
7.0
Heptane
SO
530
.5
102
39
-
•
•
51.5
52.5
> 250
100
.65
2D
32.0
2178
161
1.04*
DDP
80
340
.4
693
1.04ft
-
3.34%
93
7.0
Heptane
50
530
.5
713
538
-
•
-
51.1
178
> 250
100
.40
2E
32.0
-
161
1.04%
DDP
80
450
.4
808
1.587!
-
-
92
7.0
Heptane
50
530
.5
-
-
-
•
-
51.5
188.5
>2SO
100
.40
2F
32.0
1917
161
1.04%
DDP
80
250
.4
972
2137
-
-
91
7.0
Heptane
50
530
.5
923
325
-
-
-
51.4
171.1
>250
100
.40
(Continued)
-------
TABLE 5 (continued)
00
Run Number
Primary Extractor
Lean Solvent Flowrate roU /mln.
PPM Caprolactam In Lean Solvent
Feed Ftownte mU /mln.
PPM Caprolactam In Peed
Dispersed Phase
Column Temperature *C
Agitation Speed, stroke* Aiin,
Agitation Amplitude, In.
PPM CaproUctam In Raldnate
PPM DDP in Rafflnate
Wt % Water In Extract
Wt. % CaproUctam in Extract
% Caprolactam Recovery
Secondary Extractor
Heptane Flowrate, mU /mln,
Dispersed Phase
Column Temperature *C
Agitation Speed, strokes/ndn.
Agitation Amplitude, In.
PPM CaproUctara In Ralflsate
PPM DDP In Raffloate
PPM Heptane In Rafflnate
PPM DDP In Extract
PPM CaproUctam In Extract
Distillation Column
Head Temperature *C
Feed Point Temperatare *C
Bottom Tempontum 'C
Head Pressure. MM Hg
Reflux Ratio (reflux/total overhead)
Overhead
4H 41 4] 4K 4L 4M SA SB
Primary and secondary extractors not operated during 57 57
Runs 4-G--4-M
s 466 326 326
164.6 164.6
.493% .493%
DDP DDP
60 60
345 345
.4 .4
44
98
135
1.76
99
5.3 5.3
Aqueous Aqueous
SO 50
500 570
.5 .5
32
90
13
5C
57
326
164.6
.493%
OOP
60
385
.4
29
70
5.17
1.65
99.4
5.3
5D _^
57
326
164.6
.493%
DDP
60
420
.4
28
77
11.07
1.46
99.4
5.3
Aqueous Aqueous
55
570
.5
22
95
19
55
570
.5
23
122
13
30 28 45. S 183 187 235 Distillation column not operated during
31 10 188 190 190 2.17.5 Runs S A- -5 D
238 7520 249 249. S 249. S >250
100 100 75 75 75 75
.30 .30 .20 .50 .10 .35
Heptane Heptane Heptane Light CaproUc- DDP
* Water 4 Water A Water Ends tarn
(Continued)
-------
TABLE 5 (continued)
to
so
Run Number
Primary But rector
Leu Solvent Flownte mU /mln.
PFM Caprolactam In Lean Solvent
Feed Flownte ml. /mliu
PPM Caprolactam In Peed
Dispersed fliase
Column Temperature *C
Agitation Speed, nrokn/mln.
Agitation Amplitude, In.
PPM CaproUctam In Rafflnate
PPM DDP In RaffinaM
Wt. % Water In Sana
Wt. % Caprolactam In Extract
% CaproUctam Recover*
Secondary Extractor
Heptane Flowrate, ml. /mln.
Dlapersed Phase
Column Temperature *C
Agitation Speed, atrokes/mln.
Agitation Amplitude, In.
PPM Caprolactam In RaNInate
PPM DDP In RaKlnate
PPM Heptane In Ratflnate
PPM DDP In Extract
PPM Caprolactam In Extract
DtstUlatlon Column
Head Temperature C
Feed Point Temperature *C
Bottom Temperature *C
Mead Premwre. MM Hg
Redux Ratio ( reflux A«al overhead)
Overhead
2G
32.0
-
161
1.04%
Aqueous
80
340
.4
951
5662
-
3.2
91
7.0
Heptane
50
S30
.5
B61
288
-
-
•
.
152
>250
100
.40
Water
3A
35.0
3952
161
.528%
DDP
92
-
.4
549
3231
-
-
90
7.5
Heptane
SO
530
.5
492
1213
-
-
-
54
IS.1)
>2SO
115
.'15
and Caprolacum
4A
55
716
161
.437%
DDP
80
345
.4
102
200
7.77
1.53
98
7.0
Heptane
50
500
.5
-
-
-
-
-
4B
55
716
161
.437%
DDP
80
385
.4
304
.211%
4.53
L33
93
7.0
Heptane
SO
500
.5
142
84
-
-
-
4C
55
716
161
.437%
DDP
80
420
.4
63
.292%
6.70
1.30
99
7.0
Heptane
SO
SOO
.5
28S
73
-
-
-
4D
5S
716
161
.437%
DDP
60
385
.4
120
68
9.27
-
97
7.0
Heptane
50
555
.5
201
60
-
-
-
4E
55
716
161
.437%
DDP
60
420
.4
66
360
5.72
-
98
7.0
Heptane
50
555
.5
282
192
-
-
-
4F
55
716
161
.437%
DDP
60
345
.4
179
165
4.88
-
96
7.0
lleixane
SO
555
.5
-
-
-
-
-
Diminution column dot operated during Runs 4-A--4- F
4G
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
•
•
-
-
-
•
33
35
70
1.10
.50
Heptane and water
-------
TABLE 6. BATCH CARBON ADSORPTION OF DDP AND CAPROLACTAM AT 80 C
*-
o
'•
Experiment No. Carbon Type*
Starting solution
1 Columbia LCL
2
3
4 CalgonCAL
5
6
Solution/Carbon
(wt. /wt.)
-
10/1
50/1
100/1
10/1
50/1
100/1
PPM DDP in PPM caprolactam
aqueous phase in aqueous phase (wt. 3
at equilibrium at equilibrium (wt. $
38 103
N. D. ** <9 ppm N. D. < 2 ppm
(t H
2 ppm
N. D. < 2 ppm
M it
3 ppm
Kd DDP
I DDP in carbon)
a DDP in aqueous
DMSe)
-
>32
> 161
>322
> 32
> 161
>322
K505
>2525
5050
>505
>2525
3333
* mesh - 375
** not detected
-------
TABLE 7. CONTINUOUS COLUMN CARBON ADSORPTION OF PILOT PLANT
SECONDARY EXTRACTOR EFFLUENT FROM RUN 5-C at 80 C
Carbon Used: Columbia LCL, 12/28 mesh, 25 gms.
Column Glass: 1 1/2" I. D.
Carbon Bed Height: 10"
Flowrate: . 5 bed volumes/hr. (35 ml. /hr. )
Residence Time: 1 hour
Sample Analysis
Total ml. of solution
Sample No. PPM DPP Caprolactam Heptane passed through bed
Starting Solution 95 25 19
1 N.D. <9 N.D. <2 N.D. <10 130
2 " " " 210
41
-------
TABLE 8. THERMAL STABILITY TESTS OF STRIPPED DDP
Oxidation Test
Vessel
Sample size
Air purged, air atmosphere
Duration of test
Color of starting material
Color after test
Starting/residual pressure
Control Test
Vessel size
Sample size
Nitrogen purged, nitrogen
atmosphere
Duration of test
Color of starting material
Color after test
Starting/residual pressure
300 mis.
20 mis.
10 Lbs. pressure
168 Hours
Cloudy, yellow hue
Dark Brown
10 Lbs./lO Lbs. at 20°C
300 mis.
50 mis.
10 Lbs. pressure
168 Hours
Cloudy, yellow hue
Clear, slight yellow hue
11 Lbs./lO Lbs. at 20°C
42
-------
TABLE 9. V.L.E. APPARATUS EQUIPMENT
Oven:
Pressure Transducer:
Pressure Readout:
Vacuum Pump:
Valves:
Tube Couplings:
Flanges:
"Blue M" Horizontal air-flow
Mechanical convection oven
Model No. POM 206 B-l with a double
glass observation panel built into the
center of the oven door. Temperature
range ambient to +343° C.
Datametrics Barocel Pressure Sensor
Model No. 531D-100P-1B1-H7
Range - 0 to 100 psi differential,
Accuracy to + 0.03% of reading,
Temperature capability to 450° C.
Datametrics Electronic Manometer,
Model No. 1174-A1A-1A1-A7
Display - 3 1/2 digit (digital readout)
Accuracy - + 0.05% of reading
Readings are possible to as low as
1 x 10"^ psia.
Welch Duo-Seal two-stage pump,
Free air capacity, 25 liter/min. at 580 rpm,
Guaranteed pressure 1 x 10~^ torr. Hg.
Pump series 1405, 1/3 hp motor.
Nupro Bellow Valves
SS-4H-TSW
Cajon®VGR Vacuum Couplings
Varian Associates
Mini-conflat flanges using OFHC
copper gaskets.
43
-------
REFERENCES
1. Treybal, R. E. Liquid Extraction. McGraw-Hill, Inc., New York,
1963. pp. 246-253.
2. Karr, A. E. Performance of a Reciprocating-Plate Extraction Column.
AIChE J., 5 (4): 446-452, 1959.
3. Heat Exchange Institute. Standards for Jet Steam Ejectors. New York,
New York, 1956, p. 63.
4. Renon, H., and J. M. Prausnitz. Local Compositions in Thermodynamic
Excess Functions for Liquid Mixtures. AIChE J., 14 (1): 135-144, 1968.
5. American Petroleum Institute. Selected Values of Physical and Thermo-
dynamic Properties of Hydrocarbons and Related Compounds. Carnegie
Press, Pittsburgh, Pennsylvania. 1953. p. 336.
6. Manczinger, J., and K. Tettamanti. Phase Equilibria of the System
Caprolactam/Water. Polytechnical University of Budapest, Hungary.
1966. pp. 183-195.
7. Wichterle, Linck, and Hala. Vapor-Liquid Equilibrium Data Bibliography.
Elsevier Scientific Publishing Company, New York. 1973. p. 752.
8. Russel, G. E. Nylon-6: Report No. 41. Stanford Research Institute,
Menlo Park, California. 1968. p. 159.
44
-------
APPENDIX
EXPERIMENTAL METHODS USED IN OBTAINING PHYSICAL PROPERTY DATA
PHYSICAL PROPERTY MEASURED
Density
METHOD USED
Pycnometer in a
constant temperature
bath
Solubility of dodecylphenol in water
Solubility of water in dodecylphenol
Extraction apparatus
(Figure 1)
Interfacial tension of the water-
caprolactam-dodecylphenol system-
Tensomat ring
tensometer
Dodecylphenol viscosity
Heptane-dodecylphenol viscosity
Cannon-Fenske
capillary viscometer
COMMENTS
Can be calibrated
with water at
temperatures below
100°C and with
ethylene glycol at
temperatures above
100°C
Entrainment can
cause large error
especially below 100
ppm. Minimizing
system stirring helps
to prevent entrainment.
This problem could be
eliminated by filtering
the phases (at die
system temperature, of
course), however the
apparatus would be
more complex.
With skill and
patience, will give
good results at
temperatures below
the mixture boiling
point.
Standard, very
satisfactory
equipment
45
-------
GLOSSARY
extract: Solvent (dodecylphenol)* that has been used in an extraction and contains
the desired extractable component (caprolactam). Generally, a small
fraction of the second phase (water) is dissolved in the extract.
lean solvent: Solvent which has been regenerated from its extract mixture by
distillation or other means.
raffinate: The liquid phase (water) which has been depleted of one or more
components (caprolactam) by an extraction.
solvent: A liquid phase (dodecylphenol) which is used to extract a desired
component (caprolactam)from a second immiscible phase (water).
* compounds in parentheses apply to the process in this report
46
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-80-062
3. RECIPIENT'S ACCESSIOI*NO.
4. TITLE AND SUBTITLE
Caprolactam Recovery from Aqueous Manufacturing
Streams
5. REPORT DATE
April 1980 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
John H. Dibble
8. PERFORMING ORGANIZATION REPORT NO.
. PERFORMING ORGANIZATION NAME AND ADDRESS
Union Carbide Corporation
Tarrytown, New York 10591
10. PROGRAM ELEMENT NO.
1BB610/C33B1B
11. CONTRACT/GRANT NO.
Grant No. R-803737
12. SPONSORING AGENCY NAME AND ADDRESS
Robert S. Kerr Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Ada, OK 74820
13. TYPE OF REPORT AND PERIOD COVERED
Final, 5-1-75 to 2-28-78
14. SPONSORING AGENCY CODE
EPA/600/15
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Pilot-plant runs using plant samples have demonstrated the feasibility of a
novel extraction process'for caprolactam recovery from dilute aqueous solutions.
Following extraction, aqueous effluent caprolactam concentrations as low as 30 ppm
were obtained. Further effluent treatment by activated carbon adsorption reduced
the level to less than 2 ppm. In contrast, the commercial multi-effect evaporation
process is less economical because much more water is vaporized and the condensate
typically contains up to 0.1-0.2 weight percent caprolactam.
Various physical properties were determined. In particular, vapor pressures
for the key components were determined using a special high-temperature, low-pressure
(vacuum) apparatus designed specifically for this application.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Extraction
Separation
Dewatering
Materials Recovery
Solvent Extraction
Solvents
Sorption
Separators
Mixed Solvents
Caprolactam
Activated Carbon
High-temperature, low-
pressure apparatus
Dodecyl alcohol
Heptane
Solvents
Chemical
Engineering
Solvent Extrac-
tion
Phys. Chemistry
Organic Solvents
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)'
UNCLASSIFIED
21. NO. OF PAGES
57
20. SECURITV CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
47
ft US GOVERNMENT PRINTING OFFICE: 1990-657-146/5657
------- |