Solvent selection evaluation tools for an early stage at

17th European Symposium on Computer Aided Process Engineering – ESCAPE17
V. Plesu and P.S. Agachi (Editors)
© 2007 Elsevier B.V. All rights reserved.
Solvent selection evaluation tools for an early stage
at pharmaceutical process
Samuel PEREZ and Paul SHARRATT
School of Chemical Engineering and Analytical Science, The University of Manchester,
PO Box 88, Sackville St, Manchester, M60 1QD, UK
A first level of a methodology for the selection and evaluation of solvent at an
early stage is presented inside a programming tool for its systematic and
friendly application. This first level is based on properties and characteristics of
solvents for their selection. Different tools are integrated into the program to aid
the evaluation of the group of solvents selected. A case study is presented to
highlight the most important aspects of the methodology. As a result, this kind
of analysis can be useful to determine where the efforts related with solvent
selection at an early stage should be aim.
Keywords: solvents, system, systematic, early stage, selection, and evaluation.
1. Introduction
Throughout the years solvent selection has been a difficult activity to do in the
pharmaceutical industry. This is because there is not a complete understanding
of the effects that this decision has on the complete process. One of the
principal concerns of this industry is the lack of time to introduce a drug into the
market; as a result, there is a need to create methodologies to appreciate future
implications related with solvents at early stage.
S.Perez et al.
2. Problem Statement, background
Because of the regulations concerning the approval of pharmaceutical process,
early stage of development it is the best time to highlight potential issues
concerning solvents. There had been recently reviews [1] about approaches
available to aid the selection of solvents. Nevertheless, it is a fact that these
approaches are in a early stage of development [2]. Among these approaches we
find software for the selection of solvents. However, it is necessary that solvent
search software behave more specifically according with the requirements, of a
given production process [3]. Some property-based methodologies have been
introduced [4-6] for the selection and design of solvents. A common
characteristic of these approaches is the simulation of properties through
thermodynamic models when properties are not available. Because of the
molecular complexity of the pharmaceuticals the models have a limited
application [7]. Recently new models have been developed more according to
the needs of the pharmaceutical industry [8, 9]; nevertheless, there is a need to
produce knowledge about when these models can be specifically applied and
what is the reliability for its application for new molecules.
There is a need of software management tools where systematic evaluations for
the selection of solvents it is evaluated. At the same time, the tool should
exploit the reality of the conditions and information available at an early stage.
Also, solvent should be evaluated not only from operation perspective (reaction,
extraction, crystallisation, etc), but also from a system perspective (as a whole).
Evaluation tools should promote the participation of the stakeholders involved
(chemist, chemical engineer, contractors, suppliers, etc). Another characteristic
is that this kind of tools should tell the user what kind of tools and evaluations
are more convenient to apply in order to appreciate future cost implications
related with solvents in a complete system.
3. Paper approach
The methodology applied in this approach consists of three levels for the
evaluation. Each level start from information that can be provided from the
chemist at this stage: level 1 “What solvents are you planning to apply?”, level
2 “What lab procedures do you intend to follow? ”, and level 3 “What is the
synthesis recipe?”. Level 1 is presented on this paper with the aim of helping
the user to select and evaluate solvents from property-based knowledge. After
the solvent has been selected a series of evaluation are carried. Results at this
stage should be considered just as potential implications with solvents. Further
levels (2 and 3) will be in charge of producing more information for supporting
the advantages or disadvantages of solvents performance on a system.
Solvent characteristics, properties, and tools belonging to the first level of
evaluation are structured into a Visual Basic.Net framework. Fig. 1 shows some
Solvent selection evaluation tools for an early stage at pharmaceutical process
of the most important steps to follow at level 1 as well as some of the tools that
interact along the approach.
Look for
Visual Basic. Net Interface
Create a query
Query database
Query database
Display potential
Perform experiments
Consult suppliers
Query database
Perform experiments
Consult suppliers
Compatibility software
Figure 1. First level of evaluation “properties and characteristics based”
3.1. Methodology
The main goal at this stage is to develop a framework where the user can
establish different queries (Fig. 2) to search for a potential solvent for a specific
P1, P2, P3,
Determine value’s range
“Max” and “Min”
O1, O2, O3……
Properties related with
the characteristic: P1
Properties related with
the characteristic: P1
Make another
Predetermine value
are displayed
Determine value’s range
“Max” and “Min”
Add query
to table
Figure 2. Querying procedures
Fig. 2 represents the behaviour of the routines at the time of creating a query
based on properties selected. Query by property is available for the user to
determine desire properties. Query by characteristics it is aimed to aid the user
about properties related with certain characteristics. An example would be that
the user it is looking for a solvent “easy to handle” as a result the program will
display properties and parameters related with the easier handling of solvents
such as: boiling point, density, and flammability limits among others. Query by
operations is a procedure where the user might be interested in certain operation
such as: reaction, crystallisation, liquid extraction, and drying among others. To
develop this query routine it is important to analyse which properties are
relevant for certain unit operations; work has been presented about this kind of
properties[10]. After potential solvents are retrieved, evaluation procedures are
S.Perez et al.
3.2. Case study
The synthesis of propranolol was taken as a case study to highlight the principal
aspects of our methodology. The information was treated as an early
development stage.
Figure 3. Propranolol potential synthesis reaction
Fig. 3 displays the potential main reaction for the synthesis of propranolol. The
chemists assume that they require four solvents to achieve the drug synthesis.
Fig. 4 shows the form to develop the query procedures proposed in Fig. 2.
Figure 4. Query form
The query is added into the grid and the search into the database is performed.
As a result, potential solvents for the desired operations are retrieved from the
database (Fig. 5). In this case, the solvents selected were: isopropylamine,
diethyl ether, water, and cyclohexane. The next step is to take the solvents
selected into the evaluation assessment. Tools such as: databases, evaluation
programs, contractor contact, correlations, and web service are employed. The
criteria available for the evaluation of solvents are displayed in the bottom of
the form (Fig. 5). Because the chemist wants to know about potential
implications regarding the handling of these groups of solvents the “Handling”
check box is selected as is shown in Fig. 5.
Solvent selection evaluation tools for an early stage at pharmaceutical process
Figure 5. Evaluation form
After the user selects the criteria to evaluate, the tool selector is display (Fig. 6)
with the tools available for the evaluation of the criteria selected. As a result, the
user selects the tools to make the evaluation. In the case of our case study the
user selects query of database, use of software, and experiment.
Figure 6. Tools selector form
In the case of “query of database” the system retrieves relevant notes and
properties related with the handling of the solvents selected (if they are
available). One important aspect about these databases is that they must be
design in a friendly way in order to be useful for different applications.
Software such as the “chemical reactivity worksheet” created by the NOAA
[11] can be linked to consult an check reactivity and compatibility among the
solvents. Also, the program can suggest experiments such as calorimetric
evaluations in order to evaluate the compatibility and other potential issues
related with the SHE performance of the group of solvents. Since many of the
solvents used at the pharmaceutical industry are disposed, another important
tool to select might be to contact a waste disposal contractor. This with the aim
of assessing potential implications with the disposal of the solvents selected.
Regarding the evaluation of the solvents selected at this level and handling
characteristics; we found that solvents such as isopropylamine and diethyl ether
do not have very favourable handling properties and characteristics. Because of
this, more evaluations included on the levels 2 and 3 is required in order to
determine the principal cost implication related with such issues. An important
aspect it is that the friendly structure of the approach presented allows to guide
the chemist through the evaluations. Also, the approach permits to visualize
S.Perez et al.
very early potential issues related with solvents and where the research and
development efforts should be aimed.
4. Conclusions
A first level for the systematic selection and evaluation of solvents was
presented. The programming tool provides the user with a series of different
procedures to create queries and select desired properties for the solvents
required for an operation. Moreover, the program provides guidance for the
evaluation of the group of solvents selected and interaction with different tools
and stakeholders. The aim of this level would be to detect advantages and
disadvantages related with the use of a group of solvents in early stage. Future
work will integrate the second and third level in to the program with the aim to
evaluate solvent selection cost implications on a system.
Gani, R., Jimenez-Gonzalez, C., ten Kate, A., Crafts, P.A., Powell, M.J., Powell, L.,
Atherton, J.H., and Cordiner, J.L., A modern approach to solvent selection. Chemical
Engineering (Rockville, MD, United States). 113(3):(2006) p. 30-43.
2. Constable, D.J.C., Jimenez-Gonzales, C., and Henderson, R.K., Perspective on solvent
use in the pharmaceutical industry. Organic Process Research and Development.
11(1):(2007) p. 133-137.
3. Wypych, G., Important determinants of solvent selection. Chemical Engineering
(Rockville, MD, United States). 113(6):(2006) p. 54-60.
4. Modi, A., Aumond, J.P., Mavrovouniotis, M., and Stephanopoulos, G., Rapid plantwide sscreening of solvents for batch processes. Computers Chem. Eng., 20:(1996) p.
5. Li, M., Harten, P., F., and Cabezas, H., Experiences in designing solvents for the
environment. Ind. Eng. Chem. Res., 41:(2002) p. 5867-5877.
6. Gani, R., Jimenez-Gonzales, C., and Constable, D., J.C., Method for selection of
solvents for promotion of organic reactions. Computers and Chemical Engineering,
(29):(2005) p. 1661-1676.
7. BRITEST, Solvent selection current practice, in Solvent Selection, B.Q.W. Project,
Editor. (2004).
8. Chen, C.-C. and Crafts, P.A., Correlation and prediction of drug molecule solubility
in mixed solvent systems with the nonrandom two-liquid segment activity coefficient
(NRTL-SAC). Ind. Eng. Chem. Res. 45:(2006) p. 4816-4824.
9. Chau-Chyun, C. and Song, Y., Extension of nonrandom two-liquid segment activity
coefficient model for electrolytes. American Chemical Society:(2005)
10. Jaksland, C.A., Gani, R., and Lien, K., M, Separation process design and synthesis
based on thermodynamic insights. Chemical Engineering Science. 50(3):(1995)
11. ChemicalReactivityWorksheet, a database of reactivity information.
Download PDF