D-mannitol is a common excipient used in the pharmaceutical formulation of tablets. It is often desirable to process the formula into an amorphous glassy state to improve some physical or biological properties of the drug. The glass property of D-mannitol, e.g., the glass transition temperature, Tg, therefore plays a key role in formulation development.
Differential scanning calorimetry (DSC) is the classical tool used to study the thermodynamic properties of materials, including the glass transition (Tg). The melting behavior of D-mannitol has been studied by DSC.1,2 When D-mannitol is cooled from its molten state, it can be easily crystallized. Therefore, it is difficult to get a glassy state of D-mannitol by cooling in conventional DSC because of the slow cooling rate. Previously, the D-mannitol glass was achieved by quenching the molten outside of the DSC, for example, by contact with a –80 to –85 °C metal block for approximately 30 sec. Subsequent heat scanning from –30 °C at 7 °C/min was used to detect the glass transition. However, it is still difficult to obtain a well-defined glass transition because D-mannitol shows an immediate cold crystallization right after the glass transition, which makes the accurate determination of glass transition impossible.3 Since the Tg of D-mannitol cannot be measured directly in conventional DSC, it has to be inferred from other data, for example, by estimation from the melting point using a variety of scaling rules,4 or by introducing a small amount of sorbitol as an impurity to delay the D-mannitol crystallization and extrapolating the Tg of the mixture to zero sorbitol concentration to obtain the Tg of the pure D-mannitol.3
One recent development of advanced DSC technologies is the fast-scan DSC technology pioneered by the HyperDSC® technique (PerkinElmer LAS, Shelton, CT). With HyperDSC, it is possible to scan at a much faster rate for both heating and cooling. The fast scanning rate not only increases the sensitivity of the measurement, but also suppresses many kinetic events such as cold crystallization, which may interfere with the measurement. Since introduction of the technique, many new applications in the polymer and pharmaceutical industries have been found that were previously impossible. 5–16 This article describes the use of HyperDSC to characterize the glass transition of D-mannitol.
D-mannitol also exhibits complex polymorphic behaviors. Polymorphism refers to the different crystalline forms of the same molecule. Many drugs or excipients can exist in multiple crystalline forms. The polymorphic form can have a profound impact on the physicochemical properties of the substance. D-mannitol has three common polymorphic forms with different thermodynamic stability.17 Polymorphs can be identified by several different techniques. Here, FTIR and Raman spectroscopy were used to identify different polymorphic forms of D-mannitol.
D-mannitol (99+%) was obtained from Sigma Aldrich (St. Louis, MO) and used without further purification.
The DSC experiment was performed on the PerkinElmer Diamond DSC instrument using the HyperDSC technique. The instrument was equipped with a liquid nitrogen cooling accessory with helium as the purge gas. The instrument was calibrated using indium at 10 °C/min. A D-mannitol sample was prepared in a standard aluminum pan. The sample was melted first at 200 °C and then cooled at several different rates to –30 °C and subsequently heated at several different rates up to 500 °C/min.
The FTIR experiment was performed using a PerkinElmer Spectrum™ 100 FTIR spectrometer. Samples were measured using attenuated total reflection (ATR) on a single bounce diamond/ZnSe ATR crystal.
All samples were measured between a frequency range of 4000 to 650 cm–1. Each was collected at 4 cm–1 resolution with Strong Beer-Norton apodization. A total of 16 background and sample scans were measured for each sample. Data were collected with a temperature-stabilized deuterated triglycine sulfate (DTGS) detector. The three different crystalline forms of D-mannitol were measured separately by placing the sample in contact with the ATR crystal and by applying force from the pressure applicator supplied with the ATR accessory. The application of pressure enabled the sample to be in intimate contact with the ATR crystal, ensuring a high-quality spectrum was achieved.
The Raman experiment was performed with a PerkinElmer RamanStation™ 400 benchtop Raman spectrometer, an Echelle-based system that has no moving part within the spectrograph. The system is able to achieve 4 cm–1 resolution across a measurement range of 95–3500 cm–1 Raman shift. The system uses an open electrode silicon base charged-coupled device (CCD, 1024 × 255 pixels) detector that is thermostatically controlled to –50 °C. The system is fitted with a 300-mW 785-nm near-infrared laser, which delivers 100 mW of power at the sample with a 100-μm spot size. The three different D-mannitol crystalline forms were measured directly using 100 mW of power and 100-μm spot size. Samples were measured between 200 and 3200 cm–1 Raman shift at 4 cm–1 resolution. A total of 16 sec of collection time was used.
Results and discussion
Figure 1 - Sample temperature signal vs time at several different cooling rates on Diamond DSC.
Figure 2 - DSC curve of D-mannitol heating at 10 °C/min.
HyperDSC is a premier fast-scan DSC technology performed only on power compensation DSC. With HyperDSC, valid DSC measurements can be made with a scanning rate up to 500 °C/min covering a broad temperature range, as seen in Figure 1. The figure shows measured sensor temperature (blue curve) and program temperature (red) as functions of time. Conventional DSC is incapable of heating at rates approaching 500 °C/min or realizing a constant cooling rate higher than 20 °C/min over a broad temperature range. Of course, both of these capabilities are important for accurate fast-scan measurements. In Figure 1, all heating ramps were performed at a rate of 500 °C/min. In every case, the sample temperature coincided with the programmed temperature very well. For the cooling experiment, several different cooling rates were used from 100 °C/min to 500 °C/min. The results obtained in the cooling scans demonstrate the exceptional cooling capability of HyperDSC. Only toward the end of the cooling scan did the sample temperature start to deviate from the program temperature.
The D-mannitol sample was first heated to 200 °C at 10 °C/min. It had a big melting peak with onset at 166 °C. The sample was then cooled down to –30 °C at 10 °C/min and heated again to check for glass transition around 10 °C. Clearly, there was no detectable Tg (Figure 2). The sample was crystallized fully during cooling at 10 °C/min, which is expected due to its strong tendency to crystallize.