Study On Long-term Stability Of Glassy Small Molecule Lyophilized Protectant

Adding lyoprotectant is an effective way to reduce the lyophilization damage ofbiological materials.The stability of the lyophilized protective agent in the glass state is closely related to the storage temperature.After lyophilization,these protective agents form a lyophilized protective matrix in the glass state,and its molecular activity largely determines the storage period of the stored object.Studies have shown that the structural relaxation time at the T_k?Kauzmann temperature?point is usually3-5 years,which is a suitable temperature for lyophilization.At this temperature,the molecular activity of the lyophilized substance is greatly reduced and the stability reaches the maximum.However,the value calculated by the existing Adam-Gibbs-Vogel?AGV?model that can estimate the T_k value is not within its normal range,which will bring wrong guidance to lyophilization and reduce the activity of lyophilized substances.To this end,this paper explores the revised T_kestimation method,aimed at finding a method to revise the AGV to make the parameter values fitted by the model more accurate.First,using maltitol as a model,differential scanning calorimetry?DSC?was used to measure the specific heat capacity of maltitol during continuous temperature rise and annealing cycles,and curve fitting was performed.The results show that the physical rationality of the parameters fitted by the MAGV?modified AGV model?is significantly increased,and the T_kvalue is closer to other values calculated by the non-fitting method.A similar analysis of the other four substances in the article shows that this rule is generally established,not related to the selected substance.The main findings of the paper are summarized as follows:1. The three models of TNM,AGV,and MAGV can reproduce the specific heat capacity of the continuous cooling cycle,but for the annealing experiment,the enthalpy recovery peak distortion is caused by the increase of annealing time,and the fitting effect is poor.Therefore,when fitting the enthalpy relaxation phenomenon with obvious overshoot phenomenon,the existing phenomenological model is insufficient.The relaxation time calculation for maltitol shows that the results of TNM and AGV are not much different.Considering the pure empirical properties of the TNM model parameters,it is recommended to use the AGV model with more physically meaningful model parameters.2. It can be seen from the fitting results of maltitol that the model parameters fitted by the AGV model and MAGV are significantly different.The brittleness coefficient D given by the AGV model is very large and does not have physical rationality,which in turn leads to the estimated T₀ being very small and the rationality is questionable.And MAGV makes this phenomenon much improved.3. The AGV and MAGV models were used to fit the apparent specific heat capacity data of xylitol,xylose,sorbitol and glucose,respectively.The analysis results show that the calculation results of the AGV model all have unreasonable parameter values,and the parameter range given by the MAGV model has clear physical rationality.For the fitting of"virtual heat capacity",it is found that the improvement of the fitting accuracy of the MAGV model stems from the reasonable selection of the configuration ground state.When a crystal with a lower heat capacity is used instead of the glass state as the configuration ground state,it will significantly increase The rationality of the model parameters.4. In terms of fitting details,when the fitting step size adopts the literature recommended ?t_k as 0.1,the fitting effect is not ideal,and after reducing the step size to 0.05,as the fitting mean square error?decreases,the fitting The effect becomes better,indicating that for the fitting of such phenomenological models,under the condition of computing resources permission,appropriately reducing the step size is an effective measure to improve accuracy.