(W1030-11-63) Crystalline Craftsmanship: Controlling Surface Roughness of Spray-Freeze-Dried Particles

Purpose: Optimizing surface roughness can significantly impact powder flow properties, dissolution rates, and interparticle interactions, influencing drug bioavailability and efficacy. This study delves into the intricacies of measuring and designing surface roughness in micrometre scale spray-freeze dried (SFD) powders, focusing particularly on the influence of surfactants. By investigating the effects of surfactant addition on surface morphology, our aim is to develop tailored approaches for controlling and optimizing surface roughness in SFD formulations.

Methods: We developed a novel method for quantifying surface roughness of small, spherical particles for measuring curved surfaces and micrometre-scale areas. Electron microscopy (SEM), with its widespread use, was chosen as the primary analytical tool (SU3500, Hitachi, Japan). Image processing techniques (Fiji/ImageJ2) were automated to streamline particle isolation and Region of Interest (ROI) processing and optimized in a design of experiment (DoE). For the calculation of surface roughness (RStudio) based on the grey value of SEM images, methods formerly developed by [1], [2] and [3] were adapted to enable the analysis of micrometre scale spherical powder grains. The root mean square average of the grey value deviations along an image axis, R_q, was chosen as roughness parameter and adapted for the use with spherical bodies. The SFD powders were produced by atomization of a solution into small droplets using a one-fluid nozzle (MTG-01-G1, FMP Technologies) and rapidly frozen in free fall in a cryogenic environment (SFD into cold gas). Subsequently, freeze drying (Alpha 1-4 LSCplus, Martin Christ, Germany) was performed to remove the solvent, leaving behind solid particles with controlled morphologies. Energy dispersive X-Ray spectroscopy was conducted with an EDAX Element detector system (Ametek, USA) to examine regions of surfactant enrichment. The excipients used for production were mannitol (Pearlitol SD200, Roquette), with or without the addition of acacia gum (Caelo), gelatin B 160 (Caelo) or sodium lauryl sulphate (SLS, Sigma Aldrich). All solutions were prepared in a 10 % (w/w) solid content.

Results: The developed method proved to be successful in automated particle isolation from SEM images, in determining surface roughness in micrometre scale spherical powder particles, and in quantifying the decrease of surface roughness with the addition of surface-active substances. Particle isolation resulted in highest success rates of correctly identified particles with stack contrast adjustment, a median blur of 6 px for the ROI mask, the thresholding algorithm formerly described by Li and Tam [4], a lower cut-off for circularity at 0.5 and an upper cut off for particle size at 250,000 µm. The isolated particles were subsequently cropped, the grey values analysed, a shadow correction via linear model applied and the R_q value calculated (fig. 1). The calculated roughness values were evaluated based on the mean variation of roughness values in one sample, the arithmetic average, and median values (fig. 2A). The mean variation within a sample provided information regarding heterogeneity of roughness values within the sample and are shown as a cumulative and density distribution function of roughness (fig. 2B). To specify, formulations produced with surfactant exhibited a decrease in R_q values from 3.8 to 3.1 with SLS, to 2.9 with acacia gum and to 2.4 with gelatin (mannitol, compared to mannitol-surfactant mixture). The addition of surfactant furthermore led to a decrease in in roughness heterogeneity within the sample. For mechanistic elucidation, the mannitol SLS samples were additionally cut before SEM imaging and analysed via EDS, which showed the introduction of SLS affected surface roughness without impacting inner particle structure (fig. 2 C and D), as confirmed by EDS analysis (fig. 3), revealing enrichment of SDS on the particle’s surface. These findings underscore the impact of surfactants and additives on surface roughness design in SFD formulations.

Conclusion: By successfully implementing a method for quantifying surface roughness in micrometre-scale spherical powder particles, our findings demonstrate the effectiveness of surfactants in reducing surface roughness. These results highlight the ease with which particle surfaces can be designed, emphasizing the potential for innovative approaches in pharmaceutical powder optimization.

References: [1] Chappard, D. et al. (2003) ‘Image analysis measurements of roughness by texture and fractal analysis correlate with contact profilometry’, Biomaterials, 24(8), pp. 1399–1407. doi: 10.1016/S0142-9612(02)00524-0.
[2] Hameed, N. A., Ali, I. M. and Hassun, H. K. (2019) ‘Calculating surface roughness for a large scale sem images by mean of image processing’, Energy Procedia. Elsevier B.V., 157, pp. 84–89. doi: 10.1016/j.egypro.2018.11.167.
[3] Sato, H., O-Hori, M. and Nakayama, K. (1982) ‘Surface Roughness Measurement by Scanning Electron Microscope’, CIRP Annals - Manufacturing Technology, 31(1), pp. 457–462. doi: 10.1016/S0007-8506(07)63347-2.
[4] Li, C. H. and Tam, P. K. S. (1998) ‘An iterative algorithm for minimum cross entropy thresholding’, Pattern Recognition Letters, 19(8), pp. 771–776. doi: 10.1016/S0167-8655(98)00057-9.

Acknowledgements: The authors thank the German Federal Ministry for Economic Affairs and Energy for financial support (ASG project, grant no. 03ET1643A).


Developed workflow of the surface roughness measurement including image processing, ROI processing and subsequent calculation of the Rq values with respective software used and exemplary images and plotted results


Roughness values for mannitol without (man) and with sodium lauryl sulphate (mSLS), acacia gum (mgum) and gelatin (mgel). Black line in boxplot for mean, grey diamond for median roughness value, + stand for the individual particles. B: cumulative (line) density distribution function (area) and histogram (bars) for surface roughness distribution in each sample. C and D: mannitol with SLS (C) and without (D) for the intact and cut particles


SE image (upper left) and EDS analysis (right panel) of a mannitol-SLS sample showing carbon (red) oxygen (blue), and sodium (yellow). The bar plot additionally depicts the relative weights including sulphur (green) for area 1 (the particles surface) and area 2 (the particles inside)