Salbutamol sulfate crystallization speed as a function of temperature and humidity
Salbutamol sulfate is widely used for the treatment of bronchospasm. To ensure good penetration and deposition of the drug in the lung tissue, a certain particle size, in the order of a few microns, is required. One of the potential methods of production of salbutamol sulfate, with controlled particle size, is spray drying. Using this method an amorphous salbutamol sulfate with the desired particle size can be produced. However, at elevated temperature (T) and/or relative humidity (RH) the compound can crystallize, thus losing its penetration properties. The first step in determination of safe storage conditions of spray-dried salbutamol is investigation of its crystallization process as a function of temperature and relative humidity.
In situ X-ray diffraction (XRD) is a comprehensive and sensitive tool for this type of analysis. The crystallization process is studied both qualitatively and quantitatively providing, amongst other things, the ratio of amorphous to crystalline components.
Up until now in situ XRD experiments with accurate control of T and RH required a certain level of knowledge and experience from the operator. Furthermore, special care had to be taken in order to avoid water condensation on the sample or on the
X-ray windows of the humidity chamber while operating the setup at high T-RH values. If condensation were to happen, the experiment would be stopped and the chamber would require venting and drying before re-starting the experiment, often from the beginning.
To simplify user tasks and to prevent the costly downtime, Malvern Panalytical and Anton Paar have developed an automatic combined control of temperature and humidity ensuring smooth, condensa- tion-free in situ XRD experiments at temperatures 10-80 ºC and relative humidities 5-95%.
Automatic combined control of temperature and humidity is a result of an ongoing collaboration between Malvern Panalytical and Anton Paar. This tool enables in situ XRD experiments at controlled temperature-humidity in a completely automatic fashion and is applicable to both MHC-trans and CHC plus+ Anton Paar chambers. The algorithm accounts for the strong correlation between all involved parameters and automatically controls humidity and temperature of the sample. Furthermore, for the CHC plus+ setup, the temperature of the chamber (= the water bath temperature) is also automatically controlled. This unique algorithm not only ensures a smooth path through the T-RH parameter space without strong deviations from the programmed profile, but also guarantees condensation-free experiments.
The software integration of Anton Paar humidity chambers (MHC- trans and CHC plus+) is completed by the unique add-on, comprising automatic combined control of relative humidity and temperature. This add-on offers:
• Full automatic control of the entire experimental setup (including water bath of CHC plus+ chamber);
• Minimum deviations from the programmed profile through the entire temperature-humidity parameter space;
• Simultaneous change of both temperature and humidity;
• Condensation-free experiments.
Using the new control mechanism in combination with the Anton Paar MHC-trans chamber, a study of the crystallization process of salbutamol sulfate was performed.
Instrumentation and analytical approach
Partial least squares (PLS) regression
This is a statistical method used to predict any latent information (crystallinity, chemistry, etc.) from the raw data. The method treats XRD patterns as 2D images, not considering the physical meaning behind them. The first step is to build and cross-check a PLS calibration model using reference XRD patterns for which the value of a calibrated parameter (in this case crystallinity) is known. The second step is to verify the model on the reference patterns treating them as unknown. Once a good model is obtained it can be used to treat real unknown data sets. The implementation of the PLS method in HighScore is highly automated and the calibration and optimization steps require little user input. Furthermore, a PLS model can be used as a step of an automatic HighScore batch with an output into an Excel spreadsheet or an *. ascii file. With this configuration steps, data analysis can be a push-button process.
Figure 1. Programmed and measured non-ambient profile for the experimental series at 25 ºC and 70% RH. Thanks to the optimized control mechanism even large humidity steps are fast and smooth, without significant fluctuations, ensuring an accurate predictable path through T-RH space.
Spray-dried salbutamol sulfate was provided by Research Center Pharmaceutical Engineering, Graz, Austria. X-ray diffraction measurements were performed on a Malvern Panalytical Empyrean diffractometer equipped with PIXcel3D detector and an Anton Paar MHC-trans.
The new MHC-trans chamber for transmission geometry has an environmental heater which provides a homogeneous temperature field around the sample and the humidity sensor, thus reducing the temperature gradient across the sample surface and improving the accuracy of the humidity measurement. An internal sample changer allows for simultaneous temperature-humidity controlled XRD experiments on up to eight samples. A series of time-resolved experiments was performed at 25 ºC and 60%, 70%, 80% and 90% RH. Each run started at 25 ºC and RH between 10%-20%, followed by a fast (~10 minutes) change of RH to a preset value. As an example, one of the humidity profiles is shown in Figure 1. Once the required T-RH conditions were achieved a set of repeated 2θ scans was measured over a few hours monitoring the crystallization process.
It is known that salbutamol crystallizes faster at higher humidity values, therefore, the measurement times were adjusted accordingly. Full-range 10 minute scans were sufficient to monitor the slow crystallization process at 60% and 70% RH, while at 80% and 90% RH much faster measurements over a shorter 2θ range were required. In order to quantify the amorphous content, reference samples with known amorphous/crystalline ratio were prepared. For each sample loading the amount of sample placed in a sample cup was weighed to ensure the constant irradiated volume required for amorphous content quantification. The same sample preparation technique was applied to all T-RH runs.
Result and discussion
Calibration of PLS model using reference data set In Figure 2 the XRD patterns from eight reference samples including two sample loadings for each mixture with a known amorphous/crystalline ratio are shown. The two patterns from each mixture are almost identical, confirming the homogeneity of the reference mixtures. The three-parameter PLS calibration model built using long-range reference XRD patterns, and its verification are shown in Figure 2 (on the right).
This model was used to quantify the amorphous/crystalline ratio for the run at 70% RH and 25 ºC. Two more calibration models using short range reference scans were used to analyze data sets recorded at 80% and 90% RH.
Figure 2. XRD patterns from reference sample set used to build PLS model and the cross-validation of the PLS model
Crystallization of spray-dried salbutamol sulfate at variable humidity conditions
The crystallization speed of spray-dried salbutamol is a non-linear function of relative humidity. A 12-hour time- resolved series at 60% RH showed that salbutamol had changed, however, it still remained amorphous (Figure 3 top). At 70% RH the crystallization process is much faster. XRD patterns subsequently measured from salbutamol sulfate kept at 25 ºC and 70% RH showed that complete crystallization at these conditions occurred within 4 hours (Figure 3 bottom). The process was even faster at 80% and 90% RH.
Figure 3. Time-resolved monitoring of salbutamol crystallization at 60% and 70% RH at 25 ºC
Using the PLS method the amorphous- to-crystalline ratio for each measured XRD data set was extracted. The results are shown in Figures 4 and 5. It is clear that the crystallization speed of salbutamol increases with increasing humidity; and this dependence is non-linear (Figure 5).
Figure 4. The crystallization speed of salbutamol sulfate as a function of relative humidity at 25 ºC. Figure 5. Time required for the complete crystallization of salbutamol sulfate at a certain value of relative humidity at 25 ºC
Figure 5 shows the time required for crystallization in each of the three RH conditions plotted in Figure 4. The 60% RH sample showed no crystallization within the total measurement period. In order to identify the function RH =f(time of crystallization) a curve was drawn to be consistent with all four of these results suggesting an asymptote close to 60% RH.
The resulting non-ambient data sets were analyzed using Using the newly developed Anton Paar humidity chamber Malvern Panalytical's HighScore software. The partial least squares (MHC-trans) in combination with the fully automatic (PLS) regression method (implemented in HighScore 4.0) combined software control of both temperature and was used to quantify the amount of the amorphous relative humidity, the crystallization process of spray-dried salbutamol sulfate phase as a function of time. It was salbutamol sulfate was studied in situ. A series of time shown that crystallization of salbutamol sulfate strongly resolved experiments under the accurately controlled depends on humidity conditions and this dependence was temperature-humidity conditions was performed. shown to be non-linear.