Using Differential Scanning Calorimetry to Investigate the Optimization of Frozen Solution Conditions for Efficient Freeze Drying

Introduction

Freeze-drying, also known as lyophilization, has become a standard process in the pharmaceutical industry for manufacturing biologically active substances [1-3]. Despite its widespread use, the process is not without limitations, including high capital and energy costs, long processing times, and challenges in selecting optimal parameters such as time, temperature, pressure (vacuum), and component concentration. These parameters must be carefully optimized to ensure full recovery of activity, complete reconstitution of the often-labile drug, acceptable appearance of the freeze-dried cake, and good storage stability [4].

The freeze-drying process relies on the vapor pressure of ice. Even at temperatures as low as −50 °C, ice sublimates, leaving behind a very porous, low-density cake containing the stabilized drug. The sublimation rate (drying rate) is highly temperature-dependent, approximately doubling with a 5 °C increase [5]. Therefore, using the highest possible temperature during primary drying maximizes drying efficiency and minimizes process costs. However, the process’s limitations arise in selecting the optimum drying temperature, which can vary throughout the process and depend on the other parameters mentioned above.

When defining the freeze-drying conditions, it is important to characterise the glass transition temperature (T_g) of the frozen solution. The ability to accurately measure T_g in the frozen solution or in the partially- and fully-dried lyophilized cakes greatly improves the cost effectiveness and the quality of the final product. Cake collapse should not occur at temperatures below T_g.

This technical brief builds on some analysis aspects discussed in an earlier brief [6].

Experimental, Results, and Discussion

In this experiment, a 40% w/w sucrose solution in distilled water was used as a model solution for lyophilization. For each measurement 20 μl of the solution was loaded into a TZero™ Hermetic Pan.

Several measurements were carried out using a TA Instruments™ Differential Scanning Calorimeter. In the first measurement the sample was crash cooled to -70 °C, held isothermal for 5 minutes and then ramped at 10 °C/min through to the melt. The same crash cooling was used in the second measurement; however, in this case this was followed by a modulated differential scanning heating profile (±1 °C, 60 second period, 2 °C/min ramp to 35 °C).

Figure 1 shows the 10 °C/min ramp after crash cooling. The subambient transitions can be seen but there is overlap, making analysis difficult. These show evidence of a plasticized glass transition of the sucrose followed by a crystallization (free water before the freezing point).

Figure 2 shows the Modulated DSC Technology response after the same crash cooling. The glass transition temperatures can be more easily characterized in the reversing heat flow signal, allowing a clearer analysis of thermal conditioning. The plasticized glass transition and the water recrystallization can be seen along with the main glass transition. This indicates the possibility to anneal the material at a suitable temperature to increase the ice content and therefore improve the lyophilization performance.

The effect of annealing was studied in the third measurement. Here, the same crash cooling was used; however, following this, the sample was heated to -40 °C (above the low glass transition step) and held isothermal for 60 minutes. The sample was then cooled again and reheated using the same Modulated DSC Technology profile as in measurement 2. Figure 3 shows the comparison between the sample without annealing and the sample with annealing. There is a clear increase in the glass transition temperature, which will allow an improvement in the efficiency of the lyophilization process.

As a further investigation a range of controlled cooling rates were used to study the effect on the glass transition temperature. Slower cooling rates allow more time for crystal formation so there should be a higher glass transition temperature with slower cooling rates.

In these measurements, the samples were cooled from room temperature at rates of 10, 5, 2, 1, and 0.5 °C/min to -70 °C. They were then held isothermal for 5 minutes followed by heating using the same modulated DSC conditions as previous measurements (±1 °C every 60 seconds, ramp at 2 °C/min).

Figure 4 shows the reversing heat flow response after the crash cooling rate and the controlled cooling rates. A clear trend can be seen with the increase in the plasticized glass transition temperature as a function of the cooling rate used during the freezing step.