Monitoring of phase transformations during processing of solid dosage forms

April 1, 2007
Jakko Aaltonen, Niklas Sandler
Pharmaceutical Technology Europe
Volume 19, Issue 4

When a dosage form or an API is introduced to the gastrointestinal tract, or dissolution media mimicking it, a transformation from a metastable to a more stable form with lower solubility and bioavailability is possible.

When developing active pharmaceutical ingredients (APIs) it is essential, first, to screen the most suitable polymorph for the formulation and, second, to be able to preserve its properties when the entire multistep manufacturing process is completed. However, it is well known that during processing, APIs and excipients can be subject to mechanical and/or thermal stress, and exposed to aqueous environments, which can cause transformations in the solid state.1 These process-induced transformations (PITs) cannot be neglected as they can lead to changes in processability, bioavailability and stability of the drug.2 The properties mentioned can change as a result of polymorphic interconversions, desolvation, alteration of the crystallinity or formation of hydrates.

Solid state characterization with various complementary techniques is required to generate meaningful supporting data throughout the manufacturing process to assess concerns regarding PITs. Among a wide range of analytical techniques, Raman spectroscopy (RS) and near infrared (NIR) spectroscopy have attracted growing attention as they have great potential in process analysis and control. The recent developments in optics and probe design have made timely measurements easier in dynamic process conditions, such as moving powder beds and agitated suspensions, and the number of papers dealing with these issues is constantly growing. This paper lists some examples of PITs and how they can be monitored, with a focus on recent applications of Raman and NIR spectroscopy in different processing steps of pharmaceutical solids (Figure 1).

Figure 1


The first unit operation during drug manufacture is often crystallization. To ensure the right solid form and avoid setbacks at later stages caused by undesirable solid forms, crystallization has to be performed in a highly controlled manner. As crystallization is a very complex process, several parameters have to be monitored to achieve a well-controlled process.

Different kinds of PITs may occur during crystallization; for example, if the API being crystallized is enantiotropic (a system in which the stable polymorphic form is not the same above and below a specific transition temperature), temperature changes during the crystallization process can induce solid phase transformations. Transformations during crystallization may also occur because the first solid form to crystallize is often a metastable form that converts to a more stable form as the crystallization process proceeds.

Hu et al. used RS to simultaneously measure the desupersaturation profile (drug concentration in the solution) and polymorphic form of an enantiotropic API, flufenamic acid, during the crystallization process.3 Raman spectra were recorded in the study with a fibre-optic immersion probe in the crystallization vessel. Peaks characteristic of different phases of flufenamic acid (polymorphic forms I and III, and the solute) were identified in the spectra and quantitative information of the liquid and the solid phase was derived from the same spectra. Quantification of the solid phases was based on ratios of characteristic peak intensities. The Raman spectral analysis of drug concentration in the solution was in agreement with ultraviolet (UV) spectroscopic verification, and a temperature-induced interconversion between the polymorphic forms during the crystallization process was quantitatively monitored.

Févotte et al. studied solid state transformations during crystallization and filtration processes with NIR spectroscopy.4 The study was performed with an API (SaC) appearing in two crystalline forms (SaC I and SaC II) and one amorphous form during processing. Crystallizations were performed in acetone solutions, and NIR spectra were recorded with a transreflectance probe immersed in the solution during crystallization. The downstream filtration process was monitored with a diffuse reflectance probe above the filter cake. Quantitative analysis was based on a multivariate method, principal component regression, using only a small number of calibration samples. Transformations from SaC I to SaC II (the stable form) were observed at different rates depending on various parameters (seeding, temperature and solvent composition). The transformations were successfully measured with NIR spectroscopy, and the coarse calibration with only few data points was found satisfactory for the measurement of polymorphic composition during processing.


Milling or grinding is an important manufacturing process for solid dosage forms. A possible unwanted milling-induced phase transformation is the production of small levels of disorder or amorphous material in the milled crystalline material. This is usually found predominantly at the surfaces of crystals or powders, and these can lead to significant chemical and physical instability and change the processability of the formulation. However, a change from crystalline to amorphous can also be deliberately made by grinding when the aim is to produce an amorphous drug.

Chikhalia et al. investigated the effect of morphology on milling induced changes.5 They showed that succinic acid with a plate-like morphology has more tendency to disorder than a needle-like morphology of the drug. It was also found that the type of mill influences the amount of change in crystallinity. Of the methods used, solution calorimetry, proved to be useful in quantifying small levels of amorphous material, which X-ray powder diffraction (XRPD) and disc-spinning calorimetry (DSC) could not detect. Mackin et al. used isothermal heat conduction microcalorimetry and dynamic vapour sorption to quantify low levels of amorphous material within a crystalline API after micronization.6 They found that amorphous material can be generated with extremely small changes in the operating conditions of the micronizer and were able to detect 0.5% levels of amorphous material with the above mentioned techniques.

The aim is to monitor changes from crystalline to amorphous during processing, but so far very few examples of real-time measurements of phase transformations during milling exist in the cited literature. Only very recently Savolainen et al. used on-line NIR and RS together with multivariate modelling tools to monitor process-induced changes from the crystalline to amorphous form for a crystalline starting material, lactose monohydrate.7 They concluded that spectroscopic techniques do not only provide a tool to monitor the changes in the amorphous content during milling, but they are also applicable as fast and uncomplicated methods to monitor amorphous content at different locations of the milling bowl. The variation in amorphous content in the bowl could also be used to describe the state of the process and to determine the end-point of milling.

Wet granulation and pelletization

APIs are exposed to water in many pharmaceutical processes and the subsequent API–water interaction may induce phase transformations. As a typical example of this, wet granulation processes provide an ideal environment for the crystallization of new phases of the API. In fact, a number of PITs can occur during different process steps of wet granulation, which include wetting, mechanical stress and drying.

Granulation or the manufacture of pellets using the extrusion–spheronization technique includes several process stages (blending of the dry mass, wet granulation of the mass, extrusion of the moist mass, rotation of the extrudate by spheronization and drying). The amount of wetting liquid in the powder mass in pelletization by extrusion–spheronization is relatively high compared with other granulation methods. Consequently, depending on the drug substance and excipients processed, solution-mediated polymorphic transformations are probable. Previously, Sandler et al. have investigated phase transitions occurring in nitrofurantoin and theophylline formulations during pelletization by extrusion– spheronization.8 An at-line monitoring approach was used to increase the understanding of the solid-state behaviour of the APIs during the process. RS, NIR spectroscopy and XRPD were used in the characterization of polymorphic changes during the process. Samples were collected at the end of each processing stage (blending, granulation, extrusion, spheronization and drying). Water induced a hydrate formation in both model formulations during processing, and NIR spectroscopy gave valuable real-time data about the state of water in the system, but it was not able to detect the hydrate formation in the theophylline and nitrofurantoin formulations during the granulation, extrusion and spheronization stages because of saturation of the water signal. Raman and XRPD measurements confirmed the expected pseudopolymorphic changes of the APIs in the wet process stages. However, a relatively low Raman signal of theophylline (10% [w/w] in the formulation) complicated the interpretation of spectra.

RS has been shown to be very helpful in controlling wet granulation processes. Jørgensen et al. successfully used (off-line) charged coupled device (CCD) RS and NIR spectroscopy to investigate solid phase transitions of caffeine and theophylline during wet granulation.9 They concluded that the symmetric vibrations in the drug molecules could be effectively detected and analysed with RS. As Raman bands were much narrower than those in the NIR spectra (which has dominating OH bands) it was easier to observe the changes in the Raman spectra. Wikström et al. successfully interfaced in-line RS to a high-shear granulation process and also monitored solid phase transformations of anhydrous theophylline.10 They also studied effects of several processing parameters, namely the mixing speed and monohydrate seeding. It was found that the mixing speed had the greatest effect on the phase transformation, showing that an increase in mixing speed shortened the onset time of conversion and increased the rate of transformation to the monohydrate. However, the seeding with monohydrate or a change in the incorporation of the binder to the system did have an effect on the transformation profile.


After wet granulation or pelletization, the granules/pellets are usually dried to a certain moisture content. The drying can be performed, for example, in an oven on a tray or in a fluidized bed dryer. Either way, heat is brought to the system to evaporate the excess moisture from the particles, but in fluidized bed drying moisture removal is made more effective by blowing an air stream through the fluidized particles. The drying processes can cause phase transformations, in particular dehydration of hydrate forms.

Recently, Aaltonen et al. used in-line NIR and RS to quantitatively monitor the solid phase of an API during fluidized bed drying.11 Theophylline monohydrate granules were fluidized in a microscale fluid bed dryer and NIR and Raman spectra were recorded through a quartz sight window in the dryer wall (Figure 2). The quantification of the solid phase was based on a multivariate method, partial least squares regression. During fluidization, theophylline monohydrate lost its water from crystallization. The loss of water was clearly visible in the NIR spectra and quantification of monohydrate content was accurate. However, the real-time Raman spectra of the fluidized granules exhibited apparent features not present in Raman spectra of pure monohydrate or anhydrate form. These features were linked to the intermediate metastable anhydrate form upon dehydration of theophylline monohydrate to theophylline anhydrate. It was concluded that as a method sensitive to water, NIR spectroscopy detects the dehydration, and as a complementary method sensitive to alterations in the crystal structure, RS detects the anhydrate formation.

Figure 2

Dissolution testing

Even though dissolution is not a manufacturing process, similar transformations can occur during dissolution of solid material. When a dosage form or an API is introduced to the gastrointestinal tract, or dissolution media mimicking it, a transformation from a metastable to a more stable form with lower solubility and bioavailability is possible. Such transformations have traditionally been detected as dissolution rate changes during intrinsic dissolution rate tests. However, the solid phase transformations during dissolution involve several overlapping factors affecting the dissolution rate that can hinder the detection of the transformation. For example, changes in the crystal habit and specific surface area as well as the concentration of the dissolution medium have an effect on the transformation kinetics and the overall dissolution rate.

Key points

Aaltonen et al. combined intrinsic dissolution rate testing with simultaneous in situ solid phase analysis of the dissolving sample to directly measure solid phase transformations during dissolution testing.12 The solid phase was quantitatively analysed by RS through a quartz window and the quantification was performed using ratios of characteristic peak intensities of each solid form in the Raman spectra. Dissolution behaviour of two hydrate-forming APIs (theophylline anhydrate and nitrofurantoin anhydrate) was investigated. Both APIs transformed to monohydrate forms during dissolution, but their dissolution behaviours were very different. Figure 3 shows results of an intrinsic dissolution rate test of theophylline anhydrate and, for comparison, theophylline monohydrate. The dissolution rate of theophylline anhydrate decreased clearly as the solid phase transformation to monohydrate form proceeded. The transformation was also very fast (~7 min). Dissolution rate changes of nitrofurantoin were more complicated and the transformation took much longer (~130 min). The incorporation of real time solid phase analysis to dissolution testing enables extensive knowledge of dissolution behaviour of drugs and is an improvement to prevailing practice, as it provides accurate information of the whole dissolution process, and not only the drug concentration in the dissolution medium.

Figure 3


This overview shows that process-induced transformations are likely to occur during manufacture and end-product analysis of solid dosage forms. RS and NIR spectroscopy often provide suitable solutions for process monitoring, but the choice of the process analytical technique has to be made carefully to obtain meaningful data from the process. Developments in analytical methods and the immense activity in the field will hopefully result in a large amount of new examples of PIT monitoring during processing of solid dosage forms in the near future.

Jaakko Aaltonen is a postdoctoral researcher at the Division of Pharmaceutical Technology, University of Helsinki (Finland). His research area is solid phase analysis, with a special interest on the development of spectroscopic tools for real-time process analysis.

Niklas Sandler is a senior scientist at AstraZeneca R&D (UK) in the area of material science supporting product development projects. He has experience in applications of image analysis in solid dosage forms process monitoring and solid state characterization with spectroscopic tools.


1. H.G. Brittain, J. Pharm. Sci., 91(7), 1563–1570 (2002).

2. G.G.Z. Zhang et al., Adv. Drug Del. Rev., 56(3), 371–390 (2004).

3. Y. Hu et al., Ind. Eng. Chem. Res., 44(5), 1233–1240 (2005).

4. G. Févotte et al., Int. J. Pharm., 273(1–2), 159–169 (2004).

5. V. Chikhalia et al., Eur. J. Pharm. Sci., 27(1), 19–26 (2006).

6. L. Mackin et al., Int. J. Pharm., 231(2), 227–236 (2002).

7. M. Savolainen et al., J. Pharm. Pharmacol., 59(2), 161–170 (2007).

8. N. Sandler et al., AAPS PharmSciTech., 6(2), E174–183 (2005).

9. A. Jørgensen et al., Pharm. Res., 9(9), 1285–1291 (2002).

10. H. Wikström, P.J. Marsac and L.S. Taylor, J. Pharm. Sci., 54(1), 209–219 (2005).

11. J. Aaltonen et al., Chem. Eng. Sci., CES 62(1–2), 408–415.

12. J. Aaltonen, et al., J. Pharm. Sci., JPS 95(12), 2730–2737.