An approach to characterising the cohesive behaviour of powders using a flow titration aerosolisation based methodology

Abstract

The purpose of this study was to characterise the cohesive behaviour of powders by measuring aerosolisation at a sequence of different flow rates, i.e. flow titration. Salbutamol sulphate and Lactohale 300 were model cohesive materials. Laser diffraction particle sizing of the aerosolised plume provided in-situ, real time particle size distributions of the aerosolised powder dispersed at increasing flow rates, from 30 to 180 L min⁻¹. Relative de-agglomeration was determined from the cumulative particle size of the aerosolised powder less than 5.4 μm compared with the full availability of particles in that size range. Relative de-agglomeration-flow rate profiles were modelled using a non-linear least squares regression to fit an empirical 3-parameter sigmoidal equation; the parameters of the sigmoid were estimated from the data. Relative de-agglomeration–flow rate titration profiles and their estimated parameters were obtained to define the different de-agglomeration mechanisms of the two cohesive powders. Salbutamol sulphate showed a maximum percent de-agglomeration of 54.9%, compared with Lactohale 300 which was only 29.5%. Lactohale 300 gave 50% of its maximum de-agglomeration at a flow rate of 48.0 L min⁻¹ while the equivalent for Salbutamol sulphate was 113.8 L min⁻¹. Salbutamol sulphate and Lactohale 300 demonstrated different patterns of dispersion in relation to de-agglomeration mechanism. This approach to characterising the aerosolisation processes has significant application in designing formulation and processing strategies for pharmaceutical inhalation drug delivery.

Introduction

Drug particles for respiratory delivery should form aerosols of less than about 5 μm aerodynamic diameter to allow deposition in the peripheral areas of the lungs (Qiu et al., 1997). Particles of this size are cohesive since the magnitude of their interactive forces generally exceeds the gravitational detachment force (Visser, 1989). Particle interaction mechanisms including contact potential, Coulombic, intermolecular, and capillary interactions have been extensively studied (Zimon, 1982). Theories such as the Johnson–Kendall–Roberts (JKR) model (Johnson et al., 1971) or the Derjaguin–Muller–Toporov (DMT) model (Derjaguin et al., 1975) have been advanced to explain such particle interactions.

Despite the range of fundamental studies of model cohesive powders, the delivery of real cohesive drug particles to the lungs is generally inefficient with most commercial devices delivering less than 20% of the emitted dose (Smyth and Truman, 2007) at the required aerosol size. This poor performance can in part be related to the complex surface nature of the drug particles which have highly non-homogeneous surface morphology and surface composition with variable crystallinity and adsorbed impurities (de Boer et al., 2003). Consequently, the network of interactions that occur in such powders is variable resulting in microstructures ranging from loosely adhered, open packed particles which readily aerosolise to strongly bound, closely packed agglomerates that are unlikely to aerosolise under air flow conditions found in commercial passive inhalation devices.

In recent years, research to predict drug aerosolisation of cohesive powders has been influenced by the studies of adhesion between individual particles using atomic force microscopy (AFM) and the colloid probe approach (Louey and Stewart, 2002, Young et al., 2006). This work has been useful in understanding particle interactions, but the interaction between particles has been determined under conditions removed from the complexities that occur in a powder bed. The AFM approach can only measure one contact interaction at a time, presenting problems in representatively sampling the total spectrum of interactions present in the powder bed. The AFM approach also may not produce a true representation of the interactive force because of the unknown orientation of the interacting particle surfaces, and because the interactive force determination has been conducted in isolation. Thus, the interactions are devoid of the influences of other particles within the powder packing and the interactions do not take into account the variability in packing within a powder bed. It is not unexpected that there are difficulties in determining representative relationships between particle interactions determined using such AFM and particle aerosolisation.

The purpose of the current study was to approach powder aerosolisation by considering the whole powder system. This approach would allow an understanding of the structure of the powder, i.e. the interactive relationship between particles in the whole powder bed produced under a defined set of conditions. The approach takes into account the distribution of inter-particular forces throughout the powder bed (and not just between individual particles) and the spatial orientations between particles within the bed. The functionality of a powder bed will be dependent on the micro-homogeneity of the inter-particular forces and the particle packing. Cohesive micronised particles will not exist as individually dispersed particles but will be part of a continuous networked structure. The distribution of both interactive forces and packing fractions of micro-regions is likely to be broad (and may even be multi-modal). Thus, when the powder structure is exposed to aerosol forming conditions (i.e. conditions of air flow causing shear, impact, etc.), the powder bed may be redistributed into particular particle size fractions dependant on its structure. Indeed, this may result in fully de-agglomerated primary particles or it may produce a highly developed matrix of non-dispersible particles.

In the current project, the aerosolisation behaviour of single model compounds (as the simplest systems) has been determined using laser diffraction. This is used to measure the particle size distributions of the aerosolised plume dispersed from a commercial inhaler device at a range of different flow rates, or as we term, “flow titration”. This approach is then developed to allow the calculation and modelling of the relative de-agglomeration flow rate profiles, enabling the de-agglomeration behaviour of the test powders to be characterised. The use and effect of different air pressures when atomising liquids is well known as established in spray drying (Masters, 1991) and for example in particle engineering (He et al., 2006). However, the liquid based systems are different from solid dry powders, where powders have a more complex and diverse magnitude of interactions. Different air flows are also used in assessing respirable fractions from inhalers, including pressurised metered dose inhalers (Guo et al., 2008) and dry powder inhalers (Zanen et al., 1992). While the use of flow rates at 60 L min⁻¹ is common to study solid powder dispersion, a few studies (Chew and Chan, 2001, Coates et al., 2005) have examined powder aerosolisation at varying air flow rates (30–120 L min⁻¹). In this study, flow rates titrated between 30 and 180 L min⁻¹ were employed to approximately coincide with the estimated peak inspiratory flow rates for human subjects (20–160 L min⁻¹), depending on the resistance of various commercial passive inhaler devices (Al-Showair et al., 2007, Clark and Hollingworth, 1993, Finlay, 2001, Newman and Busse, 2002). The behaviour of the powders under these dynamic flow conditions will be related to particle fractions in the particle size distributions.