G.B. Cox

Study of the retention mechanism for basic compounds on silica under “pseudo-reversed-phase” conditions

Abstract The performance characteristics of a set of nitogenous bases were studied on a number of silicas using aqueous organic mobile phases. The retention characteristics were complex functions of the organic solvent and buffer concentrations as well as pH. The retention mechanisms were shown to be a combination of ion-exchange and interaction with siloxane and silanol groups over the entire range of concentration of organic solvent. The differences in retention on silica were due largely to the differences in ion-exchange strength of the silanol groups and the surface concentration of the siloxane bridges. The same solutes were studied on a range of C8-bonded phases of varying surface coverage, where the same interactions were observed as seen on silica with the addition of reversed-phase interactions. The unexpected relation between capacity factor and bonded-phase coverage was explained by interactions of the several retention mechanisms involved. The retention of small proteins on pure silica was also related to its ion-exchange strength. The performance of the silicas was related to that of the bonded-phase packings prepared from them, the retention of the proteins again being related to the ion-exchange characteristics.

A simplified description of HPLC separation under overload conditions. A synthesis and extension of two recent approaches

SummaryKnox and Pyper [1] have recently proposed a simplified model of HPLC separation in an overload mode. We have now compared certain predictions of their model with experimental data reported by us earlier [2]. The Knox/Pyper model shows quantitative agreement with this data base, over a broad range in sample size and other experimental conditions. The neglect by Knox and Pyper of mixed-isotherm effects seems generally unimportant, except for the case where two adjacent bands begin to overlap appreciably at elution. The Knox/Pyper model and other treatments based upon a two-term expansion as an approximation to the sorption isotherm appear unreliable when the mass of any solute exceeds 5–10% of the column saturation capacity.We have combined certain features of the Knox/Pyper model with our own model of preparative HPLC, so as to allow convenient and reliable computer simulations to be carried out under conditions of mass-overload. The use of this program (PREPSIMX) in conjunction with the Knox/Pyper model has made it possible to draw a number of general conclusions relating to optimum conditions for preparative HPLC.

Preparative high-performance liquid chromatography under isocratic conditions. III: The consequences of two adjacent bands having unequal column capacities

Abstract Previous treatments of preparative high-performance liquid chromatography (HPLC) have generally assumed Langmuir isotherms, a fixed number of surface sites for the sample compounds being separated, and equal molecular weights for these compounds. This is consistent with assuming that values of the column saturation capacity (ws) are equal for the various sample components. However, data summarized in this paper indicate that values of ws can vary widely for small molecules. The possibility of a similar variation of ws values for larger molecules such as peptides and proteins may be even more likely. The practical consequences of such variations in solute column capacity can be striking in heavily overloaded HPLC separations. The separation can improve dramatically when the second-eluted component has a larger ws value than the first-eluted compound. Conversely, the separation can be much poorer when the ws values are reversed. When developing a procedure for preparative HPLC under conditions of heavy overloading (overlapping bands), it is therefore important to know the ws values of the compounds being separated.

Simplified description of high-performance liquid chromatographic separation under overload conditions, based on the craig distribution model. IV: Comparison of the model with experimental data for samples containing two or more solutes

Abstract The preceding paper presented a quantitative description of high-performance liquid chromatography (HPLC) separation under mass-overload conditions for samples that contain two or more solutes. Specifically it was suggested that such separations differ from corresponding single-solute separations in terms of a so-called “blockage” effect. The present paper describes the experimental verification of this model for samples containing two or more solutes. Examples of blockage are shown, and these are predicted quantitatively by means of our previous model. Mixtures of β-hydroxyethyl- and 7-β-hydroxypropyltheophylline were separated by reversed-phase HPLC on a 15 x 0.46 cm C 8 column. Total sample mass w was varied over the range 0.002 ⩽ w ⩽ 27 mg, leading to pronounced changes in separation. Similar studies were carried out with mixtures of methyl, ethyl, propyl and butyl parabens, varying total sample mass over the range 0.004 ⩽ w ⩽ 20 mg. Resulting separations were generally in good agreement with predictions (“model simulations”).

Preparative separation of peptide and protein samples by high-performance liquid chromatography with gradient elution. II. Experimental examples compared with theory.

Craig simulations of mass-overloaded gradient elution reported in the preceding paper have been extended to the case of non-Langmuir isotherms. Isotherms were selected that appear to be characteristic of peptide and protein samples in reversed-phase high-performance liquid chromatography. The dependence of bandwidth on sample size and gradient conditions was examined by Craig simulation and compared with experimental data for 13 different experimental systems involving four different proteins. There is a good correspondence between simulations and experimental data, and it seems possible to quantitatively predict bandwidth and resolution as a function of small-sample retention data, experimental conditions, and sample size. A systematic approach for designing the preparative or process-scale separation of protein mixtures by reversed-phase gradient elution is proposed.

Preparative high-performance liquid chromatography under gradient conditions : III. Craig simulations for heavily overloaded separations

Abstract Craig simulations of preparative high-performance liquid chromatography were carried out for heavily overloaded separations as a function of the separation conditions (small-sample retention and column efficiency, sample size). These data were used to derive conditions for a maximum production rate (grams per hour) of purified product, and the results were compared with the treatment of Knox and. Pyper. There is an optimum sample size and column plate number for every separation; these optimum conditions are related to the desired recovery of purified product and to the retention (capacity factor, k ′; separation factor, α) of a small sample under the same chromatographic conditions. Relative to the case of a 99.8% recovery of purified product from the feed, a 3-fold higher production rate is possible if sample size and column plate number are adjusted for 95% recovery of pure product; a 10-fold higher production rate is possible for conditions that give a 50% recovery of pure product. Teh required (optimum) plate number is halved on going from touching-band (99.8% recovery) separation to 95% recovery, and further halved for 50% recovery vs . 95% recovery. The maximum production rate also varies with sample retention ( k ′, α for a small sample); a maximum value of α is preferred and k ′ should be between 0.5 and 3.

Preparative high-performance liquid chromatography under gradient conditions : I. Band broadening in gradient elution as a function of sample size

Abstract Previous studies of gradient elution under mass-overload conditions are continued. Use of the Craig distribution model shows that there is a simple relationship between bandwidth, sample size and gradient conditions for single-solute samples. this allows the use of two or three experimental measurements to determine the saturation capacity of a column by a given sample. Several (reversed-phase) experimental studies with small-molecule samples provide quantitative confirmation of this model. For the case of protein samples the model works equally well, but it appears that only 20–44% of the column capacity (as measured by saturation uptake) is available under separation conditions. the column capacity for lysozyme was measured as a function of pore diameter and the accessibility of the surface under separation conditions was found to increase with increasing pore size.

Preparative high-performance liquid chromatography under isocratic conditions : II. The role of column variables

Abstract In order to understand better the complex interaction of different separation variables in preparative high-performance liquid chromatography, general equations were derived that relate production rate and run time plus optimum column length and flow-rate to maximum operating pressure, particle size and sample molecular weight. Computer simulation was also used to illustrate optimum conditions for representative cases.

Preparative high-performance liquid chromatography under gradient conditions : II. A computer program for the design of reversed-phase gradient-elution separations of peptide and protein samples

Abstract A commercially available computer program (BIOPREP) is described as an aid for developing preparative separations of peptide or protein samples using reversed-phase gradient elution. On the basis of four small-scale runs in the laboratory (with advice offered by the computer), experimental conditions for “touching-band” separations can be predicted. This in turn allows comparisons of the production rate of a purified product as a function of the gradient conditions and column dimensions. In this way, conditions can be selected that either maximize the production rate or provide an otherwise satisfactory separation.

Analytical performance of process-scale liquid chromatography columns

Abstract The design of process columns with efficiencies equal to those of analytical columns is discussed, including advantages of the use of the same packing material in both. A new design in column hardware has been developed which allows process columsn of 50 mm internal diameter to be packed with 10-μm, silica-based materials and to be routinely produced with reduced plate heights of 2.5 or less, yielding column efficiencies of > 40 000 plates/m in 25-cm-long columns. A description of the column design, focused on flow distribution and high-pressure safety considerations is presented. The scale-up from analytical to process chromatography is facilitated by the use of high-efficiency chromatographic materials.