Measurement of glass transition temperatures in freeze concentrated solutions of non-electrolytes by electrical thermal analysis.

The electrical resistance (R) of frozen aqueous solutions was measured as a function of temperature in order to determine whether this technique can be applied for determination of glass transition temperatures of maximally freeze concentrated solutions (Tg') of non-electrolytes which do not crystallize during freezing. Electrical thermal analysis (ETA) thermograms of frozen solutions containing the solute alone show a gradual change in slope over the temperature range of interest, with no inflection point which corresponds to Tg'. However, addition of low levels (about 0.1%) of electrolyte changes the shape of the thermogram into a biexponential function where the intersection of the two linear portions of the log (R) vs. T plot corresponds to the glass transition region. The total change in log (R) over the temperature range studied increases as the ionic radius of the reporter ion increases. The sharpest inflection points in the log (R) vs T curves, and the best correlation with DSC results, were obtained with ammonium salts. Tg' values measured by ETA were compared with values measured by DSC. DSC thermograms of solutes with and without electrolyte (0.1%) show that the electrolyte decreases Tg' by about 0.5 to 1.0 degrees C. However, Tg' values measured by ETA are somewhat higher than those measured by DSC, and difference between the two methods seems to increase as Tg' decreases. Tg' as measured by ETA is less heating rate dependent than DSC analysis, and ETA is a more sensitive method than DSC at low solute concentrations and at low heating rates.(ABSTRACT TRUNCATED AT 250 WORDS)

Using fault analysis methods to improve bioreactor safety.

A simplified fault analysis algorithm has been described and applied to the analysis of containment loss in a model bioreactor system. Using approximate data for component failure rates, the relative merits of three proposed operating regimes were evaluated by means of a computer program that implements the algorithm described. With the data given, operator error is shown to be the dominant cause of possible failure in the proposed system. In view of these results, the simplified algorithm appears useful for the comparative evaluation of various design proposals in simple systems.