Computer, solvent selection

The problem of solvent selection is most difficult for high molecular-weight polymers such as thermoplastic acryHcs and nitrocellulose in lacquers. As molecular weight decreases, the range of solvents in which resins are soluble broadens. Even though solubihty parameters are inadequate for predicting ah. solubhities, they can be useful in performing computer calculations to determine possible solvent mixtures as replacements for a solvent mixture that is known to be satisfactory for a formulation. 

Odele, O., and Macchietto, S. (1993). Computer aided molecular design A novel method for optimal solvent selection. Fluid Phase Equilibria, 82,47-54. 

This type of solvent selection problem can be formulated and solved as a Computer Aided Molecular Design (CAMD) problem [22], Application of this method for solvent selection and design is highlighted in chapter 14 and is not discussed in detail in this chapter. The ProCAMD software [23], which is based on a hybrid CAMD method can be used to solve solvent selection problems of this type. 

Normal-phase liquid chromatography is thus a steric-selective separation method. The molecular properties of steric isomers are not easily obtained and the molecular properties of optical isomers estimated by computational chemical calculation are the same. Therefore, the development of prediction methods for retention times in normal-phase liquid chromatography is difficult compared with reversed-phase liquid chromatography, where the hydrophobicity of the molecule is the predominant determinant of retention differences. When the molecular structure is known, the separation conditions in normal-phase LC can be estimated from Table 1.1, and from the solvent selectivity. A small-scale thin-layer liquid chromatographic separation is often a good tool to find a suitable eluent. When a silica gel column is used, the formation of a monolayer of water on the surface of the silica gel is an important technique. A water-saturated very non-polar solvent should be used as the base solvent, such as water-saturated w-hexane or isooctane. 

Tie lines of the system can be generated from the equilibrium compositions for each run and selectivities computed. The results of measurements obtained for the 5% by volume of ammonia/ethylene are represented in the binodal diagram in Fig. 3. Butene is represented as the distributed component between the solvent phase and the butadiene-rich phase. The ammonia-solvent gas mixture was considered to behave as a pseudo-solvent of fixed composition. The ratio of the integrated peaks for butene(i) and butadiene(j) was used to compute the selectivity, B (beta), defined on a solvent-free basis, as ... 

For multivariable optimization as in the present case, so-called overlapping-resolution mapping (ORM) can be a useful approach to computer-assisted optimization of solvent selectivity 20,46). This will be... 

The present model predicts how solvent selectivity will vary with mobile-phase composition, and this allows the selection of extreme solvents for maximum differences in selectivity. This information plus the ability to calculate solvent strength versus composition of the mobile phase then allows development of a general strategy for optimizing retention of any sample, so as to maximize resolution. This four-solvent approach can be further refined by use of computer-assisted procedures, such as the overlapping-resolution-mapping technique. 

Table 1.2 Rules used in computer-aided solvent selection for organic reactions.

Solvent selection, as such, requires not only knowledge of solution behavior, but also of environmental regulations, solvent viscosity, flash point, odor, etc. Most major solvent suppliers and many larger companies have computer programs to help find optimum solvents. Cohesion energy parameter calculations based on the equations above are central in programs dealing with solvent selection."... 

A linear programming technique is described which selects mixed solvents based on specifications of the ffo, 8p, end SH solubility parameters, evaporation rates, and other significant parameters of a solvent blend. Suggestions are made for setting the various restrictions and for setting procedures of data processing. For simpler cases of solvent selection, recourse to a computer is not necessary. Use of a solvent improvement cost factor, 8 J cost, then leads to optimum formulations since one can determine which solvents increase solubility at least cost. 8 is given by y/ 8P2 -(- 8H2. 

The properties of the optimized solvent blend generally meet the restrictions of minimum solvency and either a lower or upper limit on average evaporation rate. The use of two equations to place upper and lower bounds on the evaporation rate is recommended. If the evaporation rates are too restrictive, the computer tends to select solvents which are more expensive and of better quality. It is possible to set such stringent or unreasonable restrictions that no solution is acceptable to the computer. It is instructive at times to eliminate the evaporation rate restrictions altogether evaporation rate can be controlled by the solvents the computer is allowed to consider. Our solvent selection programs have operated satisfactorily with the relative evaporation rate system based on n-butyl acetate as 100 although this is an arbitrary choice which has not been evaluated by comparison. 

Hansen (86) has modified the method of Nelson, et al. by incorporating solubility parameters for the three component forces in the solubility parameter, viz, dispersion, polar, and hydrogen bonding forces. The solvent selection procedure was designed for use by plant laboratories on a time sharing terminal. The enlistment of the computer in the selection of solvent blends has been a boon to the formulator, but the use of older methods is still very useful. 

Additional support for the proposed model was provided from the correlation obtained between the 2 computations of TK for the ethylene glycol, ethanol, NaA system when solvent selectivity data for the 2 systems, ethanol, water, NaA and ethylene glycol, water, NaA were employed in the model to evaluate ac and aeg. The a° values for ethylene glycol and ethanol, respectively, that were obtained (Table IV) were... 

The internal consistency between the 2 computed values of TKeeg is believed sufficient to provide additional meaningful support for the validity of the osmotic treatment in view of the error possible in evaluation of 8Keeg. The implication that the general shape of experimentally difficult adsorption isotherms can be approximated by the analysis of solvent selectivity data is of great importance. 

---