Resolution of Carbohydrate Mixtures by Continuous Chromatographic Refining Techniques

Abstract

A review is given of the general chromatographic theory, the factors affecting the performance of chromatographic columns and the development of the chromatographic technique. Also included is a review of the industrial sources of fructose and a description of several industrial chromatographic processes. A reconstruction of the semi-continuous chromatographic refiner (SCCR4) was performed. The SCCR4 unit now has twelve 2.54 cm (I.D.) x 70.5 cm stainless steel columns, a temperature enclosure and a new timer control. It has more flexibility than previously and can be operated at liquid pressure of up to 1.68x103 kNm-2 (250 psi) at 80°C.  Counter current operation was simulated by sequencing a system of inlet and outlet ports around the twelve columns. Experiments with a batch glass column (1.14 cm I.D.) packed with a calcium charged zerolit SRC 14 resin (150-300 µm size range) were performed to determine the packed column characteristic and the effects of temperature and flowrate on the equilibrium distribution coefficient and the plate efficiency. Batch experiments were also conducted with the 2.54 cm I.D. stainless steel columns. Results from these experiments provided chromatographic data for selecting run parameters for the continuous operation of the SCCR4 unit, and in the simulation model. Continuous operation of the SCCR4 unit was performed using seven different flow arrangements to find the best operating conditions for the highest sugar throughput, product purity and solids concentration in the outgoing product. Separation of a 50% (w/v) glucose-fructose mixture was achieved, giving 90% (w/w) fructose rich (FR) and glucose rich (GR) products and a solid concentration of 2.5% (w/w) and 7.8% (w/w) respectively at a carbohydrate throughput of 75 gms/hour. Separation of a 70% (w/v) fructose rich, dextran mixture yielded a 95% (w/w) dextran free, FR product having a solid concentration of 8.9% (w/w). The carbohydrate throughput was 126 gms/hr. A 600 hours life test with a Fisons feed containing dextran, glucose and fructose was completed without noticeable deterioration in the performance of the chromatographic packing. Ions in the Fisons feed were monitored in the product by using atomic absorption spectrometer techniques. Improvements to the theoretical model to simulate the operation of the SCCR4 unit were made. Liquid hold-up in the SCCR4 unit and its effect on the performance was identified and allowed for in the model. The model was extended to cover multi-component systems and any number of columns. Due to data limitations the model could not be fully tested, but within the limits of the data, some general points of agreement between the experimental and simulated data were obtained.