Grantee Research Project Results
Final Report: Phase Equilibria of CO2-Based Reaction Systems
EPA Grant Number: R824731Title: Phase Equilibria of CO2-Based Reaction Systems
Investigators: Brennecke, Joan F. , Stadtherr, Mark A.
Institution: University of Notre Dame
EPA Project Officer: Aja, Hayley
Project Period: October 1, 1995 through September 1, 1997
Project Amount: $200,000
RFA: Technology for a Sustainable Environment (1995) RFA Text | Recipients Lists
Research Category: Sustainable and Healthy Communities , Pollution Prevention/Sustainable Development
Objective:
The objective of this research project was to develop measurement and modeling techniques, as well as reliable computational tools, for high pressure phase behavior of CO2-based reaction systems. Carbon dioxide, which is nontoxic and nonflammable, has been shown to be a viable solvent for a wide variety of chemical reaction systems. However, the economics of replacing traditional solvents, many of which are hazardous, with CO2 is likely to hinge on the limited solubility of many compounds in CO2. Therefore, reliable measurement, modeling and computation of the high pressure CO2 phase behavior is vital to the evaluation of CO2 as a candidate replacement solvent. The overall research program involved: (1) the development of three experimental apparatuses to measure high pressure solid/fluid and vapor/liquid equilibria; (2) the measurement of binary and multicomponent phase equilibria for candidate reaction systems in CO2; (3) modeling of the phase behavior with appropriate equation of state models and fitting of the binary interaction parameters to the experimental data; and (4) the development of a reliable computational method to solve the phase stability and phase equilibria models. This project was jointly sponsored by EPA and NSF (tasks 1 and 4 sponsored by EPA and tasks 2 and 3 sponsored by NSF) and was conducted in collaboration with DuPont and Los Alamos National Laboratory. A second 1-year no-cost extension has been obtained from NSF to complete ongoing measurement and modeling work on the project. The EPA-sponsored experimental setup and development of computation techniques has been completed. This report covers the entire project.Summary/Accomplishments (Outputs/Outcomes):
The overall progress on this project includes the development of a completely reliable technique for the computation of high pressure phase behavior, the measurement and modeling of high pressure phase behavior associated with three representative, industrially relevant reaction systems, and the application of the reliable computational technique to prediction of the phase behavior of those reaction systems. The primary products of this research project are computer programs, based on interval arithmetic, that can compute, for the first time, high pressure phase behavior of multicomponent systems with complete reliability. These programs are available from the Principal Investigators. This advance is important because conventional computational methods available in all chemical process simulators are apt to fail to converge or to converge to the wrong answer at some conditions. Thus, prior to this work, there was a great deal of uncertainty, not only about the equation of state models being used, but also about the ability to obtain a reliable answer for a particular model. To demonstrate the new computational techniques, we have studied the allylic epoxidation of trans-2-hexen-1-ol, the acylation of naphthalene, and the hydrogenation of N-methyl succinimide (NMS) to N-methyl pyrrollidone (NMP). By taking binary data to fit binary interaction parameters that are needed in the equations of state, and measuring a few multicomponent state points, we have been able to successfully describe the high pressure systems for the three example reactions. We have been able to identify ranges of temperature, pressure, and compositions in which single-phase operation would be possible, and determine when the presence of multiple phases might be inhibiting the reaction rate. Thus, it is now possible to reliably evaluate CO2 as a possible replacement solvent and design processes using CO2 as the solvent for a wide variety of reactions.Progress also has been made on two related projects that have been outgrowths of the primary NSF/EPA project. These include (1) the application of the completely reliable computational technique to the calculation of homogeneous azeotropes using excess Gibbs free energy models; and (2) the exploration of the use of CO2/ionic liquid biphasic systems.
The primary accomplishment of this research project is the development of computer programs, based on interval arithmetic, that can compute, for the first time, high pressure phase behavior of multicomponent systems with complete reliability. To demonstrate the new computational techniques, we have modeled three representative reaction systems. To perform modeling of these systems required some experimental data?primarily binary systems of the various components with CO2 and some multicomponent systems. Three high pressure equilibrium apparatuses were constructed and/or modified to perform these measurements. The research results are organized as follows: (1) a description of the completely reliable computational technique; (2) a brief description of the experimental apparatuses; and (3) a description of the results for each of the three model reaction systems: (3a) allylic epoxidation of trans-2-hexen-1-ol; (3b) the acylation of naphthalene; and (3c) the hydrogenation of N-methyl succinimide (NMS) to N-methyl pyrrollidone (NMP). This is followed by: (4) short descriptions of the two outgrowth projects that have developed from this work: (4a) the application of the completely reliable computational technique to the calculation of homogeneous azeotropes using excess Gibbs free energy models; and (4b) the exploration of the use of CO2/ionic liquid biphasic systems. This is followed by: (5) a summary.
1. Completely Reliable Computational Technique
We have developed and implemented a completely reliable computational method for determining phase stability (i.e., will a given feed split into two or more phases?) and phase equilibrium of CO2-based reaction systems. The method that we have developed is based on an interval Newton/generalized bisection procedure, and we have implemented this technique in connection with a generalized cubic equation of state model. We chose the generalized cubic because it includes the van der Waals equation, the Soave-Redlich-Kwong equation, and the Peng-Robinson equation; all of which can be used to model high pressure systems containing CO2. We demonstrated that the new method is completely reliable and provides results with reasonable efficiency for a variety of systems. We have implemented the technique for conventional van der Waals mixing rules, as well as the more recently developed Wong-Sandler mixing rules. This technique, better than any other developed to date, is a general purpose method (that can be used for any thermodynamic model) to mathematically guarantee that the correct solution is found.
We have improved the method by taking advantage of monotonicity and the naturally constrained variable space afforded by the mole fraction weighted averages that frequently appear in equation of state models to improve the computational efficiency of the technique. Finally, we have combined the totally reliable phase stability interval method with conventional flash algorithms to produce a full phase equilibrium routine (Intflash) that provides complete reliability while maintaining attractive computational speed. These programs are available for general use from the PIs.
As mentioned above, the significance of this work is that high pressure phase behavior can now be computed with complete reliability, given any cubic equation of state and a variety of mixing rules. As a result, we can now reliably evaluate CO2 as a possible replacement solvent and design high pressure CO2-based reaction systems.
2. Experimental Apparatus
To demonstrate the new computational technique to calculate high pressure phase behavior, we have chosen the three example reactions discussed below. To obtain reasonable agreement between cubic equation of state models and the actual systems requires binary interaction parameters that must be fit to experimental data. Thus, it was necessary to measure binary, and some multicomponent, high pressure phase behavior. To complete this task, we constructed/ modified/purchased three experimental apparatuses, which are described below.
2a. Static binary equilibrium apparatus
This was a modification of
an existing piece of equipment. The apparatus consists of a calibrated glass
cell in which the heavy component (a liquid) is loaded. CO2
is then carefully metered into the cell from a calibrated reservoir at a known
temperature and measured pressure. Assuming that the gas phase is essentially
pure CO2, the composition of the liquid phase(s) can be
accurately determined by difference. Thus, this apparatus was used to determine
the composition of the liquid phase(s) of high pressure vapor (or fluid)/liquid
binary equilibrium systems.
2b. Dynamic flow apparatus
To determine the composition of the gas
(or fluid) phase in equilibrium with a liquid or solid, we used an ISCO SFX220.
The solid or liquid is loaded into a small extractor vessel. The high pressure
gas or fluid is delivered from an ISCO syringe pump (model 260D) so that it
flows slowly over the sample. The saturated gas or fluid phase is expanded
through a heated restrictor and then bubbled through a collection liquid, which
is analyzed by UV-vis spectroscopy. Thus, this apparatus was used to determine
the composition of the gas or fluid phase in equilibrium with a liquid or a
solid.
2c. Multicomponent sampling apparatus
This apparatus was
constructed to allow sampling of the liquid phase of a multicomponent vapor (of
fluid)/liquid equilibrium system. It consists of a Jerguson sight gauge, in
which the sample is loaded and pressurized with CO2. A
small sample of the liquid can be taken with a system of valves and injected
into a gas chromatograph for analysis. This apparatus was used to determine the
composition of the high pressure liquid phase of a multicomponent vapor (or
fluid)/liquid equilibrium system.
3. Reaction Systems Studied
To demonstrate the newly developed computational technique, we have studied the phase behavior of three reaction systems, described below. These reactions were chosen on the advice of our collaborators at Los Alamos National Laboratory and DuPont as representative industrially relevant reaction systems.
3a. Allylic epoxidation of trans-2-hexen-1-ol to
(2R,3R)-(+)-3-propyloxiramethanol
This is a homogeneously catalyzed
reaction that yields a high value agrochemical intermediate. Our collaborators
at Los Alamos have investigated it as a candidate for solvent substitution
because it occurs with very high enantiomeric selectivity in liquid CO2 and would normally be carried out in an organic solvent like
benzene. Using the apparatuses described above, we have measured the phase
behavior of binary mixtures of each of the reactants (trans-2-hexen-1-ol and
tert-butyl hydroperoxide (in decane)), products
((2R,3R)-(+)-3-propyloxiramethanol and tert-butyl alcohol) and catalysts
(titanium (IV) isopropoxide or vanadium (V) tri-i-propoxy oxide, and diisopropyl
L-tartrate) with CO2 at temperatures between 305.15K and
323.15K and pressures to 90 bar. We used these binary data to fit binary
interaction parameters in the Peng-Robinson equation of state. We also measured
the composition of the liquid phase of various multicomponent reactant and
product mixtures. Using Intflash (the completely reliable computational
technique developed as part of this project, and described above), we calculated
the binary and multicomponent phase behavior. We found that the binary data were
well represented, but that one of the binary parameters needed to be adjusted to
provide adequate representation of the multicomponent system. Using Intflash
with this model, we were able to identify ranges of temperatures and pressures
where the reaction system would remain single phase from initiation to full
conversion. Pressures as low as 100 bar could be used, which would greatly
reduce capital costs compared to the 350 bar used by Los Alamos in their
kinetics studies.
3b. Friedel-Crafts acylation of naphthalene
Friedel-Crafts
alkylations and acylations are an important class of industrial reactions and
they have been run successfully in CO2. Because CO2/naphthalene phase behavior has been studied extensively, we
have focused on the acylation of naphthalene with acetyl chloride. We have
measured the solubility of the 1'-acetonaphthone and 2'-acetonaphthone isomer
products in CO2 in binary and ternary mixtures and modeled
them with the Peng-Robinson cubic equation of state. Currently, we are studying
the reactant acetyl chloride and reactant/catalyst complex phase behavior with
CO2. Binary interaction parameters from the binary data
will be used to predict conditions that would be required to maintain a
single-phase system. In addition, we hope to identify temperatures and pressures
at which a separate product-rich might be formed. These continuing
investigations are part of a second 1-year no-cost extension to the NSF portion
of this project.
3c. Hydrogenation of N-methyl succinimide (NMS) to N-methyl pyrrollidone
(NMP)
This is a heterogeneously catalyzed reaction that we carried out
with our collaborators (Keith W. Hutchenson and Frank E. Herkes) at the DuPont
Experimental Station. Hydrogenation reactions of this type are typically done in
a mixed liquid and gas system with a solid supported catalyst, with mass
transfer of the hydrogen to the catalyst surface being the rate limiting step.
The goal was to conduct the reaction in a single-phase CO2-based system. Unfortunately, in CO2 we
experienced rapid catalyst deactivation, possibly due to contamination from CO,
that could be produced in the reverse water gas shift reaction. However, even in
supercritical pentane we observed relatively low conversions and large
variability in the data. Using the modeling and computational tools developed in
this project, we determined that the mixtures being investigated were very close
to the phase boundary and that the variability was likely due to the formation
of a small amount of product-rich liquid phase in the system. With this problem
solved, we still observed relatively low conversions to the NMP product, even at
long times. Therefore, we believe that this reaction may be a thermodynamically
equilibrium-limited reaction at the conditions investigated. Our evaluation of
the high pressure phase behavior of CO2-based reaction
systems addressed in this project assumed that the reactions went to completion
(i.e., we evaluate the phase behavior from all reactants [0 percent conversion]
to all products [100 percent conversion]. Reliable calculation of combined phase
and reaction equilibrium for reactions that do not go to completion cannot be
done with the tools developed as part of this project. This prompted us to
prepare a new proposal, in collaboration with Dr. Keith W. Hutchenson and Frank
E. Herkes at DuPont, on the topic of combined phase and reaction equilibria of
CO2-based high pressure systems. That proposal has been
selected for funding by EPA under the Technology for a Sustainable Environment
Program and will, in effect, be the continuation of the accomplishments secured
under this grant.
4. Outgrowth Projects
Two new projects have resulted from the research conducted as part of this grant and are described below.
4a. Calculation of homogeneous azeotropes
The identification of
homogenous azeotropes is of fundamental importance because azeotropes can cause
serious limitations in vapor/liquid separation processes. Like high pressure
phase equilibria, the reliable calculation of homogeneous azeotropes is
computationally challenging. Because the techniques based on interval
mathematics developed in this project provide a mathematical and computational
guarantee of reliability, we have applied these techniques to the calculation of
homogeneous azeotropes. We have successfully developed an interval method to
calculate azeotropes of liquid mixtures that are modeled with excess Gibbs free
energy equations, like the van Laar equation and non-random two-liquid theory
(NRTL).
4b. CO2/ionic liquid biphasic systems
Recently, room temperature ionic liquids (RTILs) have received
considerable attention as environmentally friendly solvents as they do not have
measurable vapor pressures. However, if RTILs are to be used as solvents for
reactions, the question of how to remove low volatility or thermally labile
products from the RTIL remains. In collaboration with Prof. Eric Beckman at the
University of Pittsburgh, we have performed a preliminary investigation of the
possibility of using CO2 to extract products from an RTIL.
We studied the phase behavior of I-butyl-3-methylimidazolium hexafluorophosphate
[BMIM][PF6] with CO2 and found that even though substantial
amounts of CO2 dissolve in the RTIL, essentially no RTIL
dissolves in the CO2. Mixtures of a few percent RTIL in
CO2 remained immiscible to pressures above 400 bar. This
type of phase behavior, vapor/liquid equilibria where the vapor phase is
essentially pure CO2, is quite unusual and extremely
desirable. Moreover, we were able to quantitatively extract naphthalene from the
RTIL with no contamination of the CO2 with the RTIL. As a
result, we believe there is tremendous potential for CO2/ionic liquid biphasic systems.
5. Summary
We have developed a new, completely reliable method to calculate multicomponent phase behavior from equation of state models. We have demonstrated this technique to model the phase behavior of three actual multicomponent reaction systems that use CO2 as a replacement solvent. We have measured and modeled the binary and multicomponent phase behavior and identified temperature and pressure regions in which the reactions could be safely operated as single-phase systems. As a result of the accomplishments of this grant, we can now reliably evaluate CO2 as a possible replacement solvent and design high pressure CO2-based reaction systems.
Journal Articles on this Report : 11 Displayed | Download in RIS Format
| Other project views: | All 32 publications | 12 publications in selected types | All 11 journal articles |
|---|
| Type | Citation | ||
|---|---|---|---|
|
|
Blanchard LA, Hancu D, Beckman EJ, Brennecke JF. Green processing using ionic liquids and CO2. Nature 1999;399(6731):28-29. |
R824731 (Final) |
Exit |
|
|
Gau C-Y, Brennecke JF, Stadtherr MA. Reliable nonlinear parameter estimation in VLE modeling. Fluid Phase Equilibria 2000;168(1):1-18. |
R824731 (Final) |
Exit Exit |
|
|
Hua JZ, Brennecke JF, Stadtherr MA. Reliable prediction of phase stability using an interval Newton method. Fluid Phase Equilibria 1996;116(1-2):52-59. |
R824731 (Final) |
Exit Exit |
|
|
Hua JZ, Brennecke JF, Stadtherr MA. Reliable phase stability analysis for cubic equation of state models. Computers & Chemical Engineering 1996;20(Suppl 1):S395-S400. |
R824731 (Final) |
Exit |
|
|
Hua JZ, Brennecke JF, Stadtherr MA. Enhanced interval analysis for phase stability: cubic equation of state models. Industrial & Engineering Chemistry Research 1998;37(4):1519-1527. |
R824731 (Final) |
Exit |
|
|
Hua JZ, Brennecke JF, Stadtherr MA. Reliable computation of phase stability using interval analysis: cubic equation of state models. Computers & Chemical Engineering 1998;22(9):1207-1214. |
R824731 (Final) |
Exit Exit Exit |
|
|
Hua JZ, Maier RW, Tessier SR, Brennecke JF, Stadtherr MA. Interval analysis for thermodynamic calculations in process design: a novel and completely reliable approach. Fluid Phase Equilibria 1999;158-160:607-615. |
R824731 (Final) |
Exit Exit |
|
|
Maier RW, Brennecke JF, Stadtherr MA. Reliable computation of homogeneous azeotropes. AIChE Journal 1998;44(8):1745-1755. |
R824731 (Final) |
Exit |
|
|
Stradi BA, Stadtherr MA, Brennecke JF. Multicomponent phase equilibrium measurements and modeling for the allylic epoxidation of trans-2-hexen-1-ol to (2R,3R)-(+)-3-propyloxiranemethanol in high-pressure carbon dioxide. Journal of Supercritical Fluids 2001;20(1):1-13. |
R824731 (Final) |
Exit Exit Exit |
|
|
Stradi BA, Kohn JP, Stadtherr MA, Brennecke JF. Phase behavior of the reactants, products and catalysts involved in the allylic epoxidation of trans-2-Hexen-1-ol to (2R,3R)-(+)-3-Propyloxiranemethanol in high pressure carbon dioxide. Journal of Supercritical Fluids 1998;12(2):109-122. |
R824731 (Final) |
Exit Exit |
|
|
Tessier SR, Brennecke JF, Stadtherr MA. Reliable phase stability analysis for excess Gibbs energy models. Chemical Engineering Science 2000;55(10):1785-1796. |
R824731 (Final) |
Exit Exit Exit |
Supplemental Keywords:
solid/fluid equilibria measurement, evaluation, Jacobian ranges, Interval Newton/Generalized Bisection method, phase stability, modeling, supercritical fluids, pollution prevention., RFA, Scientific Discipline, Sustainable Industry/Business, Sustainable Environment, Physics, Environmental Chemistry, cleaner production/pollution prevention, Technology for Sustainable Environment, Chemistry and Materials Science, carbon dioxide reaction systems, multiphase reactive equilibria, phase equilibria, Jacobian ranges, cleaner production, environmentally benign solvents, hazardous organic solvents, chemical reaction systems, innovative technology, CO2 - based systems, pollution prevention, green chemistryRelevant Websites:
http://www.nd.edu/~jfb/ Exit
http://www.nd.edu/~markst/ Exit
Progress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.