Exhaustion of South African Final Molasses

A relationship between true purity, reducing sugar and ash was developed for South African molasses. This relationship was established by equilibrating molasses with sucrose at 40°C and a viscosity of approximately 1 000 poises. The formula proposed is p = 51,02 10,89 ' / a in which p = true purity r = reducing sugars a = sulphated ash Introduction Many investigations have been carried out attempting to correlate the composition of final molasses with ~ t s purity after obtaining equilibrium, under certain pre-defined conditions, in the system sucrose cane molasses. The aim of these relationships has been to provide a value, derived from a few simple analyses, by which the final obtainable purity of the molasses which could have been achieved in the factory could be judged. These correlations generally had no strong physical or theoretical basis and were merely statistical averages from which, in some cases, considerable deviations were possible'). Some formulae have been in use for a considerable time in a number of cane-sugar producing countries3. As it is doubtful whether one relationship will be valid for the wide variety of molasses compositions, most sugar producing areas have developed their own individual relationships between the target purity and some impurities which can be easily determined. Examples of individual formulae are those in use in Hawaii 9, Australia 5 , Taiwan 2, Mauritius 12 and India6. To derive their relationships, some investigators 2 3 ' 2 have used selected samples of factory molasses which were believed to have been exhausted and determined the various analytical parameters, after which a regression analysis was . carried out. Other relationships have been derived from laboratory exhaustion experiments5 9. In Australia the relationship between true purity, reducing sugar and ash was calculated from results using a laboratory scale crystalliser. The samples were equilibrated at a viscosity of 1000 poises and a temperature between 40' and 50°C. In a later stage of the investigation, equilibrium was obtained in copper vessels vibrating at 50 cycles 4. The viscosity was found to be the most important parameter and Foster5 stated that it should be sufficient to measure the viscosity of a molasses sample under controlled conditions to be able to judge the exhaustion of the sample. He however considered this to be too difficult for most sugar factory laboratories and prefered for this reason the determination of reducing sugars and ash content. Payne 9 also considered it necessary to obtain saturated molasses at a constant viscosity before using the analytical data, for a relationship between purity, reducing sugar and ash. The viscosity to which the data were corrected was 600 poises at 50°C. A number of investigators have tried to base these relationships on theoretical considerations. Ramaiah and Vishnu 10 postulated the existence of salt-sucrose complexes and were able to establish from a number of optical rotation measurements of sucrose in salt solutions of various concentration, the specific rotation and the nature of Na, K and Ca complexes with sucroseH. Assuming the existence of these salt-sucrose complexes, Gupta7 found, from a kinetic analysis, that an exhaustion relationship between molasses purity, reducing sugars and ash should have the form p = ~ B '/a in which P = true purity of the molasses r = reducing sugar concentration a = ash concentration The fact that a number of the published relationships show this form 9 l 2 tends to add support to his arguments. In some cases, the correlation coefficient of these relationships does not show a high value, as for example for the Mauritian formula12 where the correlation coefficient is 0.453. Others show much higher values: ~ u s t r a l i a 0,897 5, Hawaii 0.83 9. The correlation coefficient should however be considered with reservation, as in a number of these exhaustion relationships, the independent variable (true purity) is not completely independent of the dependent variables3 6. This is illustrated by the fact that for randomly generated values for the variables, a rather high value for the correlation coefficient can be obtained. This may be of the order of 0,5 Is. Development of an Experimental Exhaustion Techni,que: In the initial experiments, C massecuites from factories were equilibrated in laboratory crystallisers. The design of these crystallisers was similar to those used in Australia 5, differin only in length, being 275 mm long. The factory massecuites however were often found to be unsuitable, being either not "tight" enough, having insufficient crystal surface, or containing false grain. In a later stage, final molasses were used, which were boiled in a stirred laboratory vacuum 104 Proceedings of The South African Sugar Technologists' Association-June 1972 pan having a strike volume of 13 1.1 It was soon. found that unless the concentration of the molasses could be carried out to a defined viscosity suitable data were not obtained. For this reason, the laboratory pan was equipped with a viscosity measuring device. A current measurement of the stirrer motor on . the pan proved to be too insensitive. An electronic viscometer (Bendix Ultrason) was tried; this has a probe consisting of a reed vibrating longitudinally at 30 kc, with a small amplitude. It was found however that the ultrasonic vibrations induced false grain formation and the measurements showed a poor reproducibility. Finally, a torque measuring device on the main stirrer shaft was constructed. This device was similar in design to the one described by Nicklin and Beales. The motor was constructed in a fixed position in line with the stirrer shaft. The stirrer shaft was driven by a counter shaft fitted in a cage which could rotate freely around the stirrer shaft. The force which counter-balanced the torque on the cage was a weak spring at the end of a non-elastic cord. As the viscosity of the massecuite increased, the spring stretched and a pointer attached to the spring moved along an arbitrary scale. The latter was calibrated at various points on the scale by measuring the viscosity of the molasses by means of a Hoepller viscometer. The reproducibility of this device was found to be of the order of 100 poises. In an attempt to reduce the time necessary for obtaining equilibrium in the system sucrosemolasses, crystallisation in an ultrasonic bath was tried. However, the massecuite remained supersaturated and no crystallisation took place. Similar results were obtained using Foster's4 vibrating method Experimental Method Approximately 14 kg of the molasses under investigation was diluted to 650 Bx and pumped into the electrically heated vacuum pan described above. The diluted molasses was heated slowly at a vacuum of 80 mbar absolute (T = 45-50°C). By this means, trapped air was released. As soon as the boiling point was reached, the temperature of the calandria was adjusted to 90°C by regulating the energy to one of the calandria heating elements. Concentration of the massecuite was then continued until the torque viscometer showed a reading of approximately 100-200 poises. In order to allow further concentration with the object of reaching the super saturation zone, the boilingpoint temperature was increased to 65-700C. This was achieved by lowering the vacuum to 133 mbar absolute. Concentration was continued until a viscosity at 600C of 500-600 poises was obtained, when a sample was withdrawn from the pan and the viscosity checked by a Brookfield viscometer. At this point about 4,5 kg of baking sugar were introduced into the pan and the molasses was boiled for a further two minutes to thoroughly mix the introduced crystals and the molasses. Boiling was then stopped and the massecuite obtained was drawn into the thermostatically controlled crystalliser described previously. The massecuite was cooled down to 40°C and allowed to equilibrate for 24 hours at this temperature. This brought the viscosity up to approximately 1 000 poises. In separate tests, the cooling curves of Cmassecuites were investigated. A typical example is given in Fig. 1. Although the figure shows that complete saturation was not obtained in 24 hours, this time was considered to be sufficient for practical purposes and was therefore chosen for these experiments.