Column Optimisation

Once the column has schemes which provide effective energy and material balance and composition control there may still remain a number of variables which can be manipulated to improve profitability. It may be possible to adjust feed rate, feed composition or feed enthalpy. There is usually scope to adjust column pressure. And, if there is a large difference in product prices, compositions can be adjusted to be better than specification. 

In adjusting the operation several of a wide range of equipment constraints may be encountered. These include condenser duty which may limit because of high coolant inlet temperature (on air-fins in hot weather) or because there is a maximum permitted coolant exit temperature (e.g. corrosion by salt water). The condenser limit might be approached because the column pressure is too low such that the dew point at the top of the column approaches the coolant temperature. High feed enthalpy can similarly overload the condenser, as can fouling on either the tube or shell side. 

Parts of the unit may have hydraulic limits. These might apply to any vapour product and to pumps on feed, products and reflux. They can also apply to the flow of coolant through the condenser and heating fluid through the reboiler. As we saw at the beginning of this chapter the column may blow, weep or flood. 

In selecting the optimisation technology we need to determine what form the problem takes. If the number of available MVs exceeds the number of active constraints then there 

---

Note that the plate numbering in the program is slightly different to that shown in Fig. 1, owing to the use of the vector notation to include both the reflux drum and the column base. In the program, index number 1 is used to denote the reflux drum and product distillate, and index Nplate+1 is used to denote the reboiler and bottoms product. This is convenient in the subsequent plotting of the steady state composition profiles in the column. Both Nplate and the feed plate location Fplate are important parameters in the simulation of the resulting steady state concentration profiles and the resultant column optimisation. 

Figure 12.126 Sidestream added to preferred material balance scheme 12.20 Column Optimisation...

Post-column on-line derivatisation is carried out in a special reactor situated between the column and detector. A feature of this technique is that the derivatisation reaction need not go to completion provided it can be made reproducible. The reaction, however, needs to be fairly rapid at moderate temperatures and there should be no detector response to any excess reagent present. Clearly an advantage of post-column derivatisation is that ideally the separation and detection processes can be optimised separately. A problem which may arise, however, is that the most suitable eluant for the chromatographic separation rarely provides an ideal reaction medium for derivatisation this is particularly true for electrochemical detectors which operate correctly only within a limited range of pH, ionic strength and aqueous solvent composition. 

Table 3.42 lists the main factors influencing optimisation of SPE. When considering a specific extraction problem, many different aspects influence column selection, including nature of the analytes and of the sample matrix degree of purity required nature of major contaminants in the sample and final analytical procedure. Reversed-phase sorbents have nonpolar functional groups and preferentially retain nonpolar compounds. Thus, for a nonpolar analyte, to remove polar interferences using a polar sorbent phase, the sample... 

Cool on-column >250 pm column (i.d.) 1 ppm (FID) Reduced thermal degradation and discrimination Wide range of analyte concentrations High sample capacity (LVI) Autosamplers Direct quantification Excellent precision Control of operational conditions (initial oven temperature) Optimisation required Not applicable for polar solvents Column contamination by dirty matrices Poor long term stability... 

Cool sample transfer to column No discrimination Little optimisation needed Simple to use, robust Handles difficult samples Good repeatability High sensitivity Rugged... 

Packed column SFC and CE are both able to make inroads into the application area served by HPLC, but from opposite extremes of polarity and with little overlap. CE is likely to be more efficient and faster, but mostly applicable to very polar molecules and ions. SFC qualifies as a more reproducible, trace technique, with greater selectivity and multiple detection options. HPLC and CE have been compared [365], Owing to their orthogonality, CZE and SFC are worth developing, not in competition or as an alternative to HPLC, but as an additional method in order to augment the information obtained from the analysis. With the broad scope of possible eluents and stationary phases, HPLC has fewer constraints than SFC and CZE. The parameters influencing selectivity may be used as a guide to optimisation (Table 4.44). 

In the mid-to-late 1990s, SFC became an established technique, although only holding a niche position in the analytical laboratory. The lack of robust instruments and the inflexibility of such systems has led to the gradual decline of SFE-SFC. Only a small group of industrial SFE-SFC practitioners is still active. Also the application area for SFC is not as clearly defined as for GC or HPLC. Nevertheless, polymer additives represent a group of compounds which has met most success in SFE-SFC. The major drawbacks of SFE-SFC are the need for an optimisation procedure for analyte recovery by SFE (Section 3.4.2), and the fair chance of incompatibility with the requirements of the chromatographic column. The mutual interference of SFE and SFC denotes non-ideal hyphenation. 

LC-PB-MS is especially suited to NPLC systems. RPLC-PB-MS is limited to low-MW (<500 Da) additives. For higher masses, LC-API-MS (combined with tandem MS and the development of a specific mass library) is necessary. Coupling of LC via the particle-beam interface to QMS, QITMS and magnetic-sector instruments has been reported. In spite of the compatibility of PB-MS with conventional-size LC, microbore column (i.d. 1-2 mm) LC-PB-MS has also been developed. A well-optimised PB interface can provide a detection limit in the ng range for a full scan mode, and may be improved to pg for SIM analyses. 

LC-APCI-MS is a derivative of discharge-assisted thermospray, where the eluent is ionised at atmospheric pressure. In an atmospheric pressure chemical ionisation (APCI) interface, the column effluent is nebulised, e.g. by pneumatic or thermospray nebulisation, into a heated tube, which vaporises nearly all of the solvent. The solvent vapour acts as a reagent gas and enters the APCI source, where ions are generated with the help of electrons from a corona discharge source. The analytes are ionised by common gas-phase ion-molecule reactions, such as proton transfer. This is the second-most common LC-MS interface in use today (despite its recent introduction) and most manufacturers offer a combined ESI/APCI source. LC-APCI-MS interfaces are easy to operate, robust and do not require extensive optimisation of experimental parameters. They can be used with a wide variety of solvent compositions, including pure aqueous solvents, and with liquid flow-rates up to 2mLmin-1. 

Temperature control is important for the accurate measurement of retention data, and has to be used with refractometer detectors (Section 2.4.5). Increasing the temperature can increase the speed of the separation, especially in exclusion chromatography, and usually increases the efficiency of the column (though the gain in efficiency can be lost if the mobile phase is not properly equilibrated). Complicated separations can often be optimised by increasing the temperature, but this is done very much on a trial and error basis, and most work in hplc is still done without temperature control. 

---