See discussions, stats, and author profiles for this publication at: [Link]

net/publication/335043417

Determination of Particle Size and Microstructure in Topical Pharmaceuticals

Chapter · August 2019

DOI: 10.1007/978-3-030-17355-5_3

CITATIONS READS

5 2,862

3 authors, including:

Kevin E Dahl Harshil M Parikh

Particlese LLC Teva Pharmaceutical Industries Ltd.
18 PUBLICATIONS 554 CITATIONS 4 PUBLICATIONS 8 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Harshil M Parikh on 09 September 2022.

The user has requested enhancement of the downloaded file.

Determination of Particle Size
and Microstructure in Topical
Pharmaceuticals

David W. Osborne, Kevin Dahl, and Harshil Parikh

Abstract This chapter focuses on techniques for determining API particle size in
topical semisolid products using manual microscopy, automated microscopy, mor-
phologically directed Raman spectroscopy, dynamic light scattering, and laser dif-
fraction. The use of optical microscopy, both bright field and polarized light, for the
determination of formulation microstructure is also discussed. For topical emul-
sions, this includes characterizing globule size and being able to identify polarized
light microscopy birefringence patterns typical of liquid crystal stabilized emul-
sions. Emphasis is placed on sample preparation techniques that have been found
successful in characterizing particle or globule sizes accurately without biasing the
results.

Keywords Particle size distribution · Globule size distribution · Microscopic view

· Morphologically directed Raman Spectroscopy · Polarized light microscopy ·
Bright field microscopy · Liquid crystal stabilized emulsions

The simplest pharmaceutical product applied to the skin is a topical solution with
dissolved active and insufficient microstructure for changes in processing to influ-
ence bioavailability. Skin permeation depends on the molecular structure of the
active and how the excipients influence skin barrier properties at the application
site. For this type of formulation, two products from different manufacturing batches
that are qualitatively and quantitatively the same will have the same bioavailability
regardless of manufacturing site or processing history. It must be remembered that
even the simplest topical solution can have a complex skin delivery profile when
complex interactions occur during drug product metamorphosis. Evaporative

D. W. Osborne (*)
Arcutis, Westlake Village, CA, USA
K. Dahl
KBI Biopharma, Boulder, CO, USA
H. Parikh
Teva Pharmaceutical, Salt Lake City, UT, USA

© AAPS (American Association of Pharmaceutical Scientists) 2019 89

N. Langley et al. (eds.), The Role of Microstructure in Topical Drug Product
Development, AAPS Advances in the Pharmaceutical Sciences Series 36,
[Link]
90 D. W. Osborne et al.

changes in product composition can dramatically alter thermodynamic driving force

of the active into the skin. Preferential evaporation of the best solvent for the active
pharmaceutical ingredient (API) can result in crystallization of active on the surface
of the skin, which can effectively “turn off” partitioning of API into intact skin.
If the topical solution described above is the simplest pharmaceutical product
applied to the skin, then a topical emulsion semisolid formulation containing sus-
pended API is arguably the most complex pharmaceutical product to develop for
application to the skin. The ratio of dissolved active to suspended active can influ-
ence skin permeation, especially for products applied to diseased skin having vari-
able barrier properties depending on lesion stage or severity (Osborne 1999). For
suspended actives, particle size distribution (PSD) can not only alter bioavailability
but also potentially alter the route that some actives use to cross the skin (Patzelt
et al. 2008). For actives with measurable solubility in both the dispersed and con-
tinuous phases of the emulsion, the ratio of active dissolved in the oil phase com-
pared to active dissolved in the aqueous phase can influence bioavailability. Thus,
changes in emulsion globule size not only indicates potential physical stability
issues but can also change systemic or localized bioavailability. Suspended actives
and emulsion globules are both prone to change over the shelf life of the product.
Sometimes this change in particle or globule size distribution can be explained
by Ostwald ripening. This phenomenon occurs when small particles in the topical
product dissolve and redeposit onto larger particles suspended in the same con-
tainer. Over time, Ostwald ripening shifts the particle size distribution toward larger
particles at the expense of the smaller particles. Other times the precipitation of a
less soluble polymorphic form of the API occurs and changes in both particle size
distribution and crystal habit occur. The status of the API in a topical product and
how excipients impact emulsion globule size and stability are the focus of other
chapters in this text. However, understanding that systemic and localized delivery
can be significantly altered by (1) lot to lot differences in particle/globule size dis-
tribution at the time of release and/or (2) changes in particle/globule size on stability
emphasizes the need for robust and reproducible techniques for measuring particle/
globule size.
Relevant techniques available to determine particle size are listed in Figure 1.
Each technique is associated with a working particle size range. If a single tech-
nique cannot accurately determine the entire particle size distribution, then a second
technique may be required, especially during development, to completely

10nm 100nm 1µ m 10µm 100µm 1mm

----------------Dynamic Light Scattering--------------

----------------------------------------Laser Diffraction----------------------------
------------MDRS/Raman microscopy---------
--------------------Optical Microscopy----------
10nm 100nm 1µ m 10µm 100µm 1mm

Fig. 1 Size range by technique

Determination of Particle Size and Microstructure in Topical Pharmaceuticals 91

c­ haracterize the topical drug product. The biggest challenge for any particle size
methodology is to develop a sample preparation technique that (1) has sufficient
contrast from other product components to allow for accurate particle or globule
counting and (2) does not alter the particle/globule size distribution due to dilution,
evaporation, or other manipulation of the sample. This chapter focuses on tech-
niques for determining API particle size and formulation microstructure, i.e.
emulsion globule size and microscopic view testing, of the formulated product. The
specific techniques are manual microscopy, automated microscopy, morphologi-
cally directed spectroscopy, dynamic light scattering, and laser diffraction.

1 Manual Microscopy and Imaging of Topicals

Manual microscopy is accepted as the most direct measure of particle size (above
1 μm) and morphology. Of all techniques used to determine particle size, optical
microscopy is particularly useful for characterizing non-spherical particles (USP
<776>) and for verifiable calibration accuracy (Morse and Loxley 2009). For emul-
sions of two liquids having different refractive indexes, optical microscopy provides
direct measure of globule size and globule size distribution, which may be an indi-
cator of the physical stability of the emulsion, i.e. a critical material attribute. In
April of 2016, the Food and Drug Administration (FDA) issued the guidance
“Technical Performance Assessment of Digital Pathology Whole Slide Imaging
Devices” (FDA 2016). Known as the WSI guidance, it describes the technical per-
formance assessment data that FDA believes will facilitate the regulatory evaluation
of this microscope-based technique. Based on this guidance, details concerning the
light source, condenser, and imaging optics should be stated in the experimental
section of documents reporting the use of optical microscopy to evaluate topical
semisolids.
Specific details for the light source include (1) bulb type (e.g. halogen, xenon arc,
LED), (2) manufacturer and model, (3) wattage, (4) output adjustment control (elec-
trical/electronic/mechanical), (5) optical filters (specifically polarization filters)
including manufacturer and model if relevant, and (6) capability of tracking inten-
sity and spectral degradation with lamp lifetime, if applicable. Information concern-
ing the condenser should include (1) illumination format (e.g. Kohler, critical), (2)
manufacturer and model, (3) numerical aperture, (4) focal length, and (5) working
distance. The description of imaging optics requires the following information for
the microscope objective: (1) manufacturer, (2) type, (3) magnification, (4) numeri-
cal aperture, (5) focal length, and (6) working distance. Any auxiliary lenses must
also be identified. For automated digital microscopic view testing, other compo-
nents detailed in the WSI guidance may be of relevance. Furthermore, the WSI
guidance instructs sponsors to complete ISO 13653:1996, ISO 9039:2008, and ISO
15795:2002 with the International Standards, if applicable.
As stated in USP <776> “…the particles must be representative of the distribu-
tion of sizes in the material and must not be altered during preparation of the mount.”
92 D. W. Osborne et al.

Sample preparation usually falls within one of the three techniques: (1) sandwich-
ing about 100 μl of sample between a slide and coverslip with gentle pressure to
achieve a sample thickness of about 25 μm, (2) applying product within the confines
of a spacer (25–125 microns or 1–5 mils) and applying gentle pressure to the cover-
slip making certain that the thickness of the sample is determined by the spacer and
not the sample, and (3) slowly coating the product to a known gap (0.1 mm or less)
onto the microscope slide using a doctor knife or functional equivalent. The need for
“gently pressing” and “slowly coating” is to minimize shearing of the sample that
could alter the microscopic view of the sample during preparation. Note that for
topical products having complex rheology, the sample may “spring back” after
preparation of the mount regardless of the technique used. Two additional sample
preparation considerations are temperature and evaporation. Both open and closed
thermal stages are available for samples that are particularly temperature sensitive
or to complete temperature ramping or temperature cycling studies to simulate
freeze-thaw testing or temperature excursion studies. For emulsions containing vol-
atile excipients, the microscopic view obtained with and without a coverslip should
be determined. If the coverslip causes globule deformation, then capturing the
image without a coverslip may be preferred. However, working without a coverslip
requires that the time between sample prep and image capture be validated with
regard to changes in the microscopic view due to evaporation of the sample. For
development work, the manual microscopy observations of a single, experienced
development scientist are often sufficient to obtain unbiased results, but for auto-
mated microscopy, sample preparation techniques must be validated (Morse and
Loxley 2009) and the scientist must have documented training.
Microscopy of topical pharmaceuticals is usually limited to bright-field micros-
copy and polarized light microscopy (PLM). In bright-field microscopy, white light
illuminates the sample from below and the sample is observed from above. Contrast
in the sample is caused by attenuation of the transmitted light in dense areas of the
sample. Bright-field microscopy is the best technique for visualizing globule size in
a pharmaceutical emulsion, i.e. cream or lotion.
For PLM the white light coming from the lamp passes through a fixed filter that
plane polarizes the light before it illuminates the sample. A second polarizing filter
that can be rotated is called the analyzer and is placed between the objective and the
ocular lenses or camera. Crossed-polarized light is produced when the polarizer and
analyzer are 90 degrees apart to produce extinction, i.e. when the field of view
becomes as dark as possible. It is important to not confuse cross-polarized light
microscopy with dark-field microscopy. Both techniques produce an image having
a dark background, but the term dark-field should be reserved for microscopy tech-
niques that block the central beam of light or electrons. For PLM contrast in the
sample is caused by rotation of polarized light through the sample. This allows the
observer to evaluate the three-dimensional structure of anisotropic specimens. PLM
is the best technique for visualizing dispersed crystalline drug particles or for com-
pleting microscopic view testing on certain topical semisolids such as liquid crystal
stabilized emulsions.
Determination of Particle Size and Microstructure in Topical Pharmaceuticals 93

1.1 Bright-Field Manual Microscopy

In the 1957 ACS Monograph Emulsions: Theory and Practice, Paul Becher wrote
only twelve sentences concerning direct microscopic measurement to determine
droplet (globule) size distribution of an emulsion. “This method has been widely
used in the past. While it is extremely laborious, it is the method in which there is the
most certainty in the results … It should be pointed out that, in order to obtain
meaningful results, a large number of droplets should be measured. Harkins and
Fischer (Fischer and Harkins 1931), for example, in work using this technique mea-
sured 50,000 droplets for each determination … With concentrated emulsions, it is
necessary to dilute the emulsion in order to be able to carry out the count. This may
introduce a serious error, since the very act of dilution may cause changes in the
distribution pattern … A more serious limitation on the determination by this
method is the impossibility of measurement of droplets in the fractional micron
range. This can introduce a serious error in the overall distribution” (Becher 1957).
Although the automated microscopy and imaging techniques described below have
reduced the “extremely laborious” nature of this technique, the challenges of direct
microscopic measurement of emulsion globules have not changed. Bright-field
manual microscopy of emulsion globules is primarily a technique used during
development to identify a lead cream or lotion formulation having superior pre-
dicted physical stability compared to other prototype formulations varying qualita-
tively (Q1) and/or quantitatively (Q2).
The breakdown processes in emulsions include creaming, sedimentation, floc-
culation, Ostwald ripening, coalescence, and phase inversion. These six different
breakdown processes are complex and depend upon surface forces (Tadros 2013).
For our purposes, the theories of emulsion stability can be reduced to three consid-
erations: (1) if you can see the globules using optical microscopy, then one or more
of these breakdown processes are occurring in the pharmaceutical emulsion, (2)
small globule size indicates greater emulsion stability, and (3) narrow globule size
distribution indicates greater emulsion stability. It is the microstructure of the emul-
sion that dictates the kinetics of the breakdown process and prevents phase separa-
tion for the shelf life of the product. Bright-field manual microscopy is the technique
used during development of a pharmaceutical emulsion to predict which Q1/Q2
combination of excipients (both with and without API) provides sufficient micro-
structure (Q3) to produce the cream or lotion having the greatest physical stability.
Two creams containing the same active but having different emulsifiers and oil
phases are shown in Fig. 2. Both creams were manufactured at laboratory scale (less
than 1 kg) with the processing parameters recorded but not controlled, i.e. first try
manufacturing of a series of prototype creams. The photomicrograph of the first
cream sample was taken about 6 months (177 days) after date of manufacture and
shows small uniform globules indicating a stable emulsion. An air bubble is readily
identified based on the dark refractive index ring at the air/cream interface. The
photomicrograph of the second cream was taken about 7 months (220 days) after
94 D. W. Osborne et al.

date of manufacture and shows much greater globule polydispersity with globules
showing both stages of the flocculation and coalescence breakdown processes. In
the first stage, flocculation, the globules of the dispersed phase form aggregates in
which the drops have not entirely lost their identity (this aggregation tends to be
reversible). The second stage, coalescence, each aggregate combines to form a sin-
gle drop (this is essentially an irreversible process). The largest, presumably
coalesced globule in this microphotograph has a diameter of approximately 35
microns. Once a sufficient number of large oil droplets form they will separate (usu-
ally by creaming) and phase separation of the cream or lotion will be visible. This
is the stability time point that the pharmaceutical emulsion fails appearance testing.
The bright-field manual microscopic evaluation of these two prototype creams
clearly indicates that the first cream is a more stable emulsion than the second
cream. Just as the development scientist can rank the first Q1/Q2 combination as
being more stable than the second, bright-field manual microscopy can also be used
during scale-up to rank one set of process parameters as producing a more stable
emulsion compared to another set of parameters.
It is noteworthy that after two years of ambient storage in an amber glass jar, both
formulations shown in Fig. 2 were smooth white creams without any sign of phase
separation, i.e. both creams would have passed appearance testing. This observation
raises two important points: (1) bright-field microscopy is a very sensitive technique

Fig. 2 Bright-field
photomicrographs of two
different cream
formulations taken using
an Amscope PZ300 Series
polarized light microscope
fitted with a 6 V/30 W
halogen bulb (Philips
5761). The Infinity Plan
achromatic objective had
40× magnification, a
numerical aperture of 0.10,
and a working distance of
7.8 mm. The air bubble in
the first photomicrograph
has a 0.026 mm outer
diameter of the refractive
index ring
Determination of Particle Size and Microstructure in Topical Pharmaceuticals 95

Fig. 3 Using an Amscope

PZ300 Series polarized
light microscope fitted
with a 6 V/30 W halogen
bulb (Philips 5761). The
Infinity Plan achromatic
objective had 4×
magnification, a numerical
aperture of 0.10, and a
working distance of
7.8 mm. This image was
captured about 4 weeks
after the manufacture of
the cream

for observing microstructural changes that shift globule size and polydispersity and
(2) topical product physical stability is determined by a combination of long-term
stability appearance testing at the labeled storage conditions and stress testing
(accelerated temperature, freeze-thaw testing, excursion studies) of the final product
manufactured using the final process.
Another bright-field manual microscopy observation that indicates a potential
physical stability problem is shown in Fig. 3. Here the continuous phase of the
emulsion is seen to be separating from the emulsion at the edge of the cream sample
that has been placed between microscope slide and coverslip. This liquid crystal
stabilized emulsion also has a high degree of globule size polydispersity (note that
the microphotograph was taken with a 4× objective rather than the 40× objective
used for Fig. 2). This cream failed appearance testing due to phase separation prior
to one-year storage at ambient conditions.

1.2 Polarized Light Microscopy

“The use of polarized light microscopy has a long history in FDA” (FDA 2003) with
the primary purpose being the rapid and accurate identification of crystalline drug
substances while avoiding elaborate and costly chemical analyses. Most FDA dis-
trict laboratories have a reference catalog of known optical properties of thousands
of crystalline substances, and both inspectors and investigators are trained in the use
of PLM to perform optical crystallography to identify crystalline drug substances
and perform filth analysis which includes identifying glass, struvite, urea, dixanthy-
lurea, and the starches. Despite numerous references to optical microscopy being
found on the FDA website and USP Chapter <776> Optical Microscopy being one
of the older monographs, very little information is provided specific to the use of
PLM to determine suspended API particle size and distribution or identify a liquid
crystal stabilized emulsions.
96 D. W. Osborne et al.

PLM Images of Suspended API During development of a topical product, it is

useful to determine which prototype formulations completely dissolve API and
avoid drug precipitation upon storage. Use of manual PLM is arguably the best
technique to identify undissolved or precipitated API in a series of topical formula-
tions having a range of qualitative and/or quantitative compositions. Precipitation of
the active ingredient upon storage can occur for various reasons. Some actives tend
to form supersaturated solutions at the time of manufacture. After days, weeks, or
months, this metastable topical product will equilibrate, and active ingredient par-
ticles will form. Products formulated too close to the saturation limit of the active
ingredient will have drug precipitation when stored below 25 °C or during shipping
excursion studies. If a topical product contains a volatile solvent such as ethanol,
then evaporation of the solvent upon storage could result in precipitation of the
active. A less soluble polymorph (Pudipeddi and Serajuddin 2005) may nucleate in
the topical product and form active ingredient particles that will not re-dissolve.
Figure 4 shows that the microscopic view of roflumilast can vary significantly when
precipitated out of different solvents typically used to formulate topicals (Osborne
2018). This is despite roflumilast being non-hygroscopic, having no asymmetric
centers and having no known polymorphs (EMA 2010). In general, microscopic
view of topical drug products is more complex than PLM images of drug
substances.

Fig. 4 PLM cross-polarized photomicrographs of roflumilast precipitated from three different

solvents taken using an Amscope PZ300 Series polarized light microscope fitted with a 6 V/30 W
halogen bulb (Philips 5761). The Infinity Plan achromatic objective had 4× magnification, a
numerical aperture of 0.10, and a working distance of 7.8 mm
Determination of Particle Size and Microstructure in Topical Pharmaceuticals 97

Birefringence Patterns of Liquid Crystal Stabilized Emulsions The publication

by Friberg et al. (Friberg et al. 1969) explained that one factor in the sudden change
of emulsion stability with relatively minor changes in composition was the forma-
tion of a lyotropic liquid crystal within the emulsion. This discovery firmly linked
the knowledge of composition, properties, and structures of liquid crystalline phases
comprised of amphiphilic compounds (Ekwall 1975) with emulsion stability.
Regarding microscopic view, blends of water, an amphiphilic compound, and an
organic compound (such as oil) are unique because the molecules of the solvent are
not uniformly distributed throughout the phase. Instead the water is concentrated in
those parts containing the polar groups of the amphiphile and the oil is concentrated
entirely or partially in the hydrocarbon parts of the phase. When a liquid crystal
stabilized emulsion is placed between the microscope slide and coverslip and
viewed using PLM, the image will contain a pattern of birefringence on a dark back-
ground. The birefringence results from the three-dimensional structured liquid crys-
tals in the cream sample having a diversity of orientations. The plane polarized light
entering this anisotropic cream sample is rotated to produce a birefringence pattern
that is characteristic of a specific cream’s microstructure. It should be noted that
many pharmaceutical creams are isotropic semisolids, thus any cream having a bire-
fringence pattern when viewed using PLM is most likely a liquid crystal stabilized
emulsion. Figure 5 shows birefringence patterns from four different liquid crystal

Fig. 5 PLM cross-polarized photomicrographs of four different liquid crystal stabilized emulsion
pharmaceutical creams taken using an Amscope PZ300 Series polarized light microscope fitted
with a 6 V/30 W halogen bulb (Philips 5761). The Infinity Plan achromatic objective had 4× mag-
nification, a numerical aperture of 0.10, and a working distance of 7.8 mm
98 D. W. Osborne et al.

stabilized creams. These birefringence patterns are consistent with microscopic

view images reported in the literature (Sutton et al. 2018, Zhang and Liu 2013).

2 Automated Microscopy and Imaging of Topicals

Static optical microscopy and PLM have previously been identified as traditional
methods for characterization of particle and/or globule size in topical products.
With the advent of computers and high-quality micro-positioning controls, the abil-
ity to automate these experiments has flourished. Automated microscopy offers the
same sensitivity over the same size range as manual methods, with the advantage of
removing operator variability and fatigue and greatly improving statistical robust-
ness. Instruments are available from a variety of manufacturers and have found use
in supplementing or replacing compendial methods such as USP<778> Method 2
testing.
The critical facets of performing good manual microscopy are unchanged for
automated microscopy; that is, sample preparation is of paramount importance, fol-
lowed by contrast generation. Samples are prepared with attention paid to keeping
particles/globules unmoving (static), typically as a thin plane (~25 μm) using a
microscope slide and coverslip that is lined with a low vapor pressure fluid such as
mineral oil. The oil prevents convection due to evaporation and ensures that particle
mobility is limited. Moving particles will lead to multiple counting and may also
lead to image distortion, depending on the speed of movement. Contrast may be
generated by any of the more common methods, including white light (color, dark
field, etc.), PL, fluorescence, and differential interference contrast (DIC), though
not all may be appropriate for a given sample. Additionally, adjustment of the illu-
mination aperture has been found to improve contrast in emulsions that allows for
measurement of otherwise low-contrast particles. Such a lighting geometry is
thought to provide a quasi-numerical aperture (NA) adjustment that significantly
increases particle contrast at the cost of reducing spatial resolution. Example images
of this type of approach are shown in Fig. 6. As can be seen from the images above,

Fig. 6 Representative images of white-light microscopy, low illumination angle white-light

microscopy, and the thresholded version of the low illumination angle image
Determination of Particle Size and Microstructure in Topical Pharmaceuticals 99

the contrast improvement is marked and allows for the characterization of the glob-
ules using image analysis algorithms.
Once a contrast generation mechanism appropriate for the sample has been cho-
sen, the thresholding mechanism becomes the determining factor in automated
image analysis results quality. Simply stated, the threshold operation is one that
separates the sample (globule or particle) from the background matrix. The human
eye and brain combination provides an exquisite natural image analysis system that
is only now being approached using computer processing. The standard method for
threshold setting is that of a scalar threshold value applied globally to the image.
Using Fig. 6 as an example, the scalar threshold method defines a cutoff value above
which is a sample globule and below which is matrix. While computationally cheap,
such an approach is only comparable to the eye-brain system for the highest-­contrast
systems, i.e., completely opaque particles. Additional thresholding methods, such
as local thresholding and histogram-based thresholding, have become relevant as
computational power has increased. These methods have been successfully applied
to low-contrast samples.
Critical to the robust performance of thresholding methods is the quality of the
images themselves, specifically sample region lighting uniformity. Light sources
typically have a Gaussian profile where the center of an illuminated region receives
a higher illumination density than the corners of the region. Likewise, the optics
used to collect light after interaction with the sample have a variable transmission
efficiency, decreasing from the center to the edges. This effect is termed vignetting
and is well known from digital photography. Illumination inhomogeneities and
vignetting may be easily corrected using a flat-field correction—the process of tak-
ing a defocused, but illuminated, image through the optical train of the instrument.
Flat-field correction also corrects for differences in pixel gain and dark response
(current) in a large-format array. In order to maximize the potential of any thresh-
olding mechanism, it is critical to correct for optical and electronic variations in the
images by flat-fielding.
Another place where the human eye-brain system excels over current technology
is in image segmentation (separation). From the image in Fig. 7, it is possible for a
human operator to accurately size and count smaller globules “riding” larger glob-
ules. Segregation algorithms such as the Watershed can be useful for segmenting
extended complexes of globules or particles, but often provide results that are less
than optimal. More complicated algorithms exist but are often computationally lim-
iting. When possible, the sample can be diluted to remove touching bodies.
Topical samples, as mentioned previously, are usually prepared with a finite
thickness (~25 μm) to control sample appearance. For the high magnifications often
employed during automated microscopy, this may lead to a mismatch between the
sample thickness and the objective depth- of-field. For example, at 20× magnifica-
tion for an optic with NA of 0.40, the depth of field is on the order of 4 μm, resulting
in the majority of the 25-μm-thick sample being out of the focal plane. Such a situ-
ation can be remedied by the use of a Z-stacking algorithm, where a series of images
is taken at different Z-axis positions, with the images subsequently compressed to
form a depth-of-field-free single image. This technique offers improved robustness
100 D. W. Osborne et al.

Fig. 7 Microscopic view

image showing smaller
globules “riding” on larger
globules. Automated
microscopy of this sample
required a segregation
algorithm

for viscous samples, such as topicals, that have a substantial distribution of globules
or particles in the Z dimension.
Perhaps the most important aspect regarding the use of automated methods is the
improvement in statistical robustness offered relative to that of manual microscopy.
Assuming that an automated experiment is properly executed and the sample inves-
tigated is representative of the bulk product, automated methods can provide parti-
cle morphology for hundreds of thousands of particles in less time than that required
for even a cursory manual examination. When considering international standards
for static image analysis, as summarized in ISO 13322-1, Annex A, automated
image analysis easily allows for the measurement of statistically-relevant numbers
of particles for almost any given sample size. For example, from Annex A, for a
volume-weighted distribution of particles with a distribution width (D90/D10) of ~3
(5% error), 100,000 particles would need to be counted to give an accurate represen-
tation of the particle size distribution. This is an untenable problem for a manual
method, even though the actual particle distribution is not excessively broad.

3 Morphologically Directed Raman Spectroscopy of Topicals

Morphologically directed Raman spectroscopy (MDRS) is a natural evolution of

automated image analysis, much as image analysis was an evolution of manual
microscopy. Once confidence had been established using automated microscopy to
characterize particle size distributions of particles/globules in topical products, the
explicit identification of discrete bodies in these products became possible. MDRS,
a combination of automated microscopy and Raman spectroscopy, was conceived,
developed, and became commercially important.
Determination of Particle Size and Microstructure in Topical Pharmaceuticals 101

The foundation of MDRS relies on high-quality image analysis: not only is par-
ticle morphology collected during automated (static) microscopy experiments but
also particle location. The ability of an automated microscope to return to a particle
as small as 1 μm in size allows for the spectroscopic interrogation and subsequent
identification of particles on this size scale. More importantly, the ability to provide
explicit identification of any particles in a topical product gives MDRS the capabil-
ity of measuring particle size distribution of multiple species from within a finished
sample. Recently, these techniques have received intense interest from regulatory
bodies and generic manufacturers, and in some cases allow for the bypassing of
clinical studies for generic product release (Lal 2016). There currently exist no other
automated techniques that offer the potential of collecting this type of chemically-­
specific information that is critical to proving, for example, the bioequivalence of a
generic product compared to a Reference Listed Drug (RLD).
The combination of automated optical microscopy with Raman spectroscopy is
a natural fit. Modern Raman spectrometers typically use visible (diode) laser excita-
tion, which does not necessitate specialty optics. Excitation spot size, and thus
energy density (proportional to Raman signal), is determined by the optical magni-
fication, and can be on the order of 1 μm for off-the-shelf objectives. The availabil-
ity of miniaturized spectrometers and fiber-optic coupling provides for simple
integration with existing optical microscopes.
Owing to strengths such as minimal sample preparation requirements, insensitiv-
ity to water, and high chemical specificity, Raman spectroscopy is an excellent
experimental match for topical products. The chemical specificity, certainly the
most important factor from an identification point of view, arises from the different
vibrational frequencies characteristic of functional groups present on a given mol-
ecule. For example, steroids, a common therapeutic agent delivered topically, have
a strong carbonyl stretch between 1600 and 1750 cm−1. This feature alone is often
enough to differentiate the steroid from any other materials present in a topical
formulation.
The chemical specificity of Raman spectroscopy allows detailed libraries of a
product’s input materials to be created for use in correlation matching of individual
particle Raman spectra for identification. Typically, Raman spectra are collected for
the pure input materials, which are then arranged into a reference library. Regions
of interest containing specific functional information may be highlighted by removal
of uninteresting regions of the spectrum and a correlation algorithm employed to
match each unknown particle to the library. While there are no theoretical limits to
the number of materials contained in a reference library, the more concise the library
the better the spectral sensitivity that can be obtained from a method.
Raman spectral collection is typically on the order of 1–15 s for a single particle,
which provides adequate signal-to-noise ratio for qualitative identification of parti-
cles down to the single micron range. This results in methods that are capable of
identifying greater than 1000–5000 particles in approximately 8–16 h.
102 D. W. Osborne et al.

4  article Characterization Using Laser Diffraction

P
and Light Scattering

Major factors that determine selection of the right particle sizing technique and
instrumentation are size range, sample availability, nature of the sample, and mor-
phological characteristics of the particulates/droplets. Laser diffraction (LD) and
dynamic light scattering (DLS) are extensively used techniques for particle size
analysis in the industry. LD measures particle size distributions by measuring the
angular variation in intensity of light scattered as a laser beam passes through a
dispersed particulate sample. Large particles scatter light at small angles relative to
the laser beam and small particles scatter light at large angles. The angular scatter-
ing intensity data is then analyzed to calculate the size of the particles responsible
for creating the scattering pattern, using the Mie theory of light scattering. The
particle size is reported as a volume-equivalent sphere diameter. DLS measures
Brownian motion and relates this to the size of the particles. Brownian motion is the
random movement of particles due to the bombardment by the solvent molecules
that surround them. Normally DLS is concerned with measurement of particles sus-
pended within a liquid.
It is important to select the right method based on the final product dosage form and
morphology of the particulate. There is no perfect instrument available for particle
size measurement. Sensitivity varies with the final product dosage form. The major
advantages and limitations of the laser diffraction and dynamic light scattering tech-
niques when used in the semisolid dosage form will be presented with a case study.

4.1 Laser Diffraction

The basic set up of an LD instrument consists of a laser beam that is expanded prior to
passing through the flow cell containing the sample. The sample must be sufficiently
fluid to move from the sample reservoir through the flow cell. A Fourier lens then
focusses the angle diffraction patterns onto the low-angle detector, and multiple wide-
angle detectors are positioned away from the central beam (Hackley et al. 2004). There
is very little scientific literature available on utilization of LD to evaluate topical dosage
forms or the effect on semisolid microstructure that occurs during PSD analysis. The
techniques and case studies discussed below come from the experience of the authors.
The laser diffraction technique is fast, reproducible, robust (Ma et al. 2000),
requires little sample, and is easy to transfer between labs or facilities. Hence, LD is
widely used in pharmaceuticals. However, it has some major limitations; LD model
assumes particles are spherical (Heffels et al. 1996) because only objects of simple
geometry can be unambiguously described by a single numerical descriptor (Burgess
et al. 2004). Irregularly shaped particles provide deceiving scatter pattern and are
converted by the LD instruments software into equivalent spherical diameters.
Various algorithms based on surface area, volume, or linear dimension can be used
to calculate equivalent spherical diameters. This often leads to inaccurate results
and lack of agreement for identical samples evaluated on different LD instruments.
Determination of Particle Size and Microstructure in Topical Pharmaceuticals 103

Difficulty in achieving agreement among several methods increases as the particle

shape differs from spherical, for example, when the particles have a long needle or
disk shape.

4.2 Multi-disperse Systems

Topical emulsions (creams and lotions) contain dispersed globules (oil or waxes).
The temperature and concentration of the dispersed globules are two key factors in
the equilibrium microstructure of a topically applied emulsion. When particulate
API is suspended in a topical emulsion, the formulation becomes a multi-disperse
system. It is important to understand API particle and emulsion globule size distri-
bution in the emulsion matrix for drug delivery into the skin layer. A major disad-
vantage of LD is to distinguish the API particulates from emulsion globules in a
multi-disperse system.
To evaluate particle/droplet size of a cream, the preferred sample preparation
technique requires minimal modification of the topical product. In most cases, addi-
tion of a neat sample of topical product to the reservoir media will not work because
the thick matrix of the cream or lotion will not flow through the sensing zone of the
instrument. To measure particle/droplet size, the sample must be significantly
diluted to introduce the sample into the LD flow cell. Blends of water and surfactant
are used to dilute the product and separate particles or droplets from the emulsion
matrix for the analysis. In most cases, this leads to breaking the emulsion and
destroying the microstructure of the product. Also, this sample preparation method
has the potential to break agglomerates of API particles that may have been gener-
ated during processing or storage at various stability conditions. If significant sam-
ple preparation is required, then the particle size distribution (PSD) results might
not be a true representation of the product.
Case Study The following case shows an industrial example where sub-­micronized
API (D90(v)~6 μm) was used in the formulation. When finished product was ana-
lyzed for PSD using LD, the Fig. 8 graph was obtained where D90(V) was found to

Fig. 8 LD of multi-disperse system

104 D. W. Osborne et al.

be ~70 μm. Analysis of the same sample using MDRS with Raman microscopy
determined that the D90(v) of dispersed API was ~8 μm. The graph clearly shows
the combination of two dispersed particles in the matrix. This example shows the
limitation of LD in the analysis of multi-disperse systems.

5  article/Droplet Size Measurement for Topical Gel

P
Dosage Form

A gel is a crosslinked polymer network swollen in a liquid medium where API may
be dissolved or dispersed. Neat samples of a gel containing dispersed API will be
too viscous to pass through the flow cell of the instrument. Since API particles typi-
cally sediment when yield value of the gel matrix drops by adjusting pH or adding
electrolyte, API will not remain uniformly dispersed in the LD’s reservoir system
after the breaking the gel matrix. Hence, sample preparation methods for LD deter-
mination of PSD for the gel-based topical drug delivery system often produces
results that are not a true representation of the product.
For both multi-disperse and gel systems, the FDA discourages the use of strong
surfactants or other steps that destroy the microstructure of the product prior to
particle size measurement. If these types of sample preparation techniques are used,
specific justification must be provided. The use of an orthogonal method during
validation of the LD method is one strategy to justify use of LD for characterization
of PSD for a topical product. This is especially relevant for generic drug develop-
ment where the generic formulation is seeking approval based on Q1/Q2/Q3 equiva-
lence rather than relying on a clinical end-point bioequivalence study. To ensure
therapeutic equivalence of the generic to the Reference Listed Drug (RLD), it is
important to match the physical aspects of test and reference formulation during
manufacturing and at storage conditions. PSD is critical to the local and systemic
bioavailability of the API, and FDA encourages matching the generic product
microstructure to the RLD unless it is proven otherwise.

5.1 Dynamic Light Scattering

DLS measures Brownian motion of particulates present at low concertation in the

dispersion media and therefore requires significant dilution. Although this method
is often used for injectable and ophthalmic pharmaceutical products, it has found
very limited utility for topical dosage forms having viscous or structured formula-
tions. To determine PSD the microstructure of the topical product must be com-
pletely broken which does not give true representation of the formulation
Determination of Particle Size and Microstructure in Topical Pharmaceuticals 105

characteristics. Rapid sedimentation of particles above 1 μm makes DLS unsuitable

for characterizing most topical products containing dispersed API.

6 Conclusion

The most direct method of measuring particle or globule size in a semisolid product
is to use optical microscopy. Bright field microscopy is used to measure globule size
while polarized light microscopy is typically used to determine size of crystalline API
particles. The primary advantage of optical microscopy is that sample preparation
techniques using minimal manipulation of the product can almost always be shown
to provide accurate, unbiased size measurements. The limitation of this technique is
that it cannot measure particles or globules smaller than 1 micron. If particles or
globules are present having fractional micron dimensions, then an error in the overall
particle size distribution will occur. The disadvantage of manual microscopy requir-
ing measurement of thousands of particles to obtain statistically valid particle size
distributions can be overcome using automated optical microscopy when adequate
contrast can be achieved. Morphologically directed Raman spectroscopy (MDRS) is
a natural evolution of automated image analysis that is very useful for the analysis of
topical semisolid products. Raman spectral collection is typically on the order of
1–15 s for a single particle, which provides adequate signal-to-noise ratio for qualita-
tive identification of particles down to the single micron range. This results in meth-
ods capable of identifying greater than 1000–5000 particles in approximately 8–16 h.
Both dynamic light scattering and laser diffraction can accurately measure smaller
particles (down to <10 nm and about 20 nm, respectively). However, for topical prod-
ucts containing microstructure, the samples must be modified in a way that will allow
them to pass through a flow cell. This usually means significant dilution which almost
always causes a change in both particle size and particle size distribution, invalidating
the method for use in characterizing topical semisolid products.
This chapter has been written with an emphasis on sample preparation tech-
niques that have been found successful in characterizing particle/globule sizes accu-
rately without biasing the results. These techniques must be adapted to each specific
topical semisolid based on the microstructure of the sample, particularly rheology
and volatility. Although most topical products developed in the last 15 years will
routinely utilize optical microscopy or MDRS during development, an exceedingly
small number of topical semisolid product sample preparation descriptions have
been published in the technical literature. It is recommended that topical product
development scientists publish technical notes detailing specific sample preparation
challenges and describe the critical technique that allowed the unbiased determina-
tion of particle or globule size in topical semisolids.
 106 D. W. Osborne et al.

References

P. Becher, Emulsions: Theory and Practice, Volume 135 of the American Chemical Society
Monograph Series (Reinhold Publishing Company, New York, 1957), p. 329
D. Burgess, E. Duffy, F. Etzler, A.J. Hickey, Particle size analysis: AAPS workshop report, cospon-
sored by the Food and Drug Administration and the United States Pharmacopeia. AAPS J. 6(3),
20 (2004)
P. Ekwall, Composition, properties and structures of liquid crystalline phases in systems of amphi-
philic compounds, in Advances in Liquid Crystals Volume 1, ed. by G. H. Brown, (Academic
Press, New York, 1975), pp. 1–142
EMA, EMA/464905/2010 Daxas CHMP Assessment Report issued 22 April 2010 (2010), http://
[Link]/docs/en_GB/document_library/EPAR_-_Public_assessment_report/
human/001179/[Link]. Accessed 5 Jan 2019
FDA, ORA Laboratory Manual, Volume IV, Section 4 Microanalytical & Filth Analysis. Section
4.4.6 Specialized Microscopy Techniques, Optical Crystallography (a.k.a. Polarized Light
Microscopy), and ID Spot Tests. Effective Date 10-01-03 Revised 02-14-13, (2003)
FDA, Guidance for Industry “Technical Performance Assessment of Digital Pathology Whole
Slide Imaging Devices” issued 20 Apr 2016, (2016)
E.K. Fischer, W.D. Harkins, Monomolecular films. The liquid-liquid interface and the stability of
emulsions. J. Phys. Chem. 36, 98–110 (1931). [Link]
S. Friberg, L. Mandell, M. Larsson, Mesomorphous phases, a factor of importance for the proper-
ties of emulsions. J. Colloid. Interface Sci. 29, 155–156 (1969)
V. Hackley, L.-S. Lum, V. Gintautas, C.F. Ferraris, Particle Size Analysis by Laser Diffraction
Spectrometry: Application to Cementitious Powders NISTIR 7097 (2004), p. 9
C.M.G. Heffels, P.J.T. Verheijen, D. Heitzmann, B. Scarlett, Correction of the Effect of Particle
Shape on the size distribution measured with a laser diffraction instrument. Part. Part. Syst.
Charact. 13(5), 271–279 (1996)
R. Lal, Interview with Dr. Bing Li: FDA embraces emerging technology for bioequivalence evalu-
ation of locally acting nasal sprays. FDA/CDER SBIA Chronicles, May 19 (2016), https://
[Link]/downloads/Drugs/DevelopmentApprovalProcess/SmallBusinessAssistance/
[Link]. Accessed 24 Jan 2019
Z. Ma, H.G. Merkus, S. de JGAE, C. Heffels, B. Scarlett, New developments in particle character-
ization by laser diffraction: size and shape. Powder Technol. 111(1), 66–78 (2000)
P. Morse, A. Loxley, Light microscopic determination of particle size distribution in an aqueous
gel. Drug Deliv. Technol. 9(5), 44–49 (2009)
D.W. Osborne, Compositions and methods for topical application of therapeutic agents US Patent
5,863,560 (1999)
D.W. Osborne, Inhibition of crystal growth of roflumilast US patent 9,884,050 (2018)
A. Patzelt, F. Knorr, U. Blume-Peytavi, J. Lademann, Hair follicles, their disorders and their
opportunities. Drug Discov. Today Dis. Mechan. 5(2), e173–e181 (2008)
M. Pudipeddi, A.T.M. Serajuddin, Trends in solubility of polymorphs. J. Pharm. Sci. 94(5), 929–
939 (2005)
M.J. Sutton, D.W. Osborne, K. Dahl, V. Bax, G.A. Schick, Characterization of a liquid crystal sta-
bilized pharmaceutical oil-in-water emulsion optimized for skin delivery. J. Cosmet. Dermatol.
Sci. Appl. 8(4), 207–217 (2018). [Link]
T.F. Tadros, Emulsion formation, stability, and rheology, in Emulsion formation and stability, ed.
by T. F. Tadros, (Wiley-VCH Verlag GmbH & Co., Weinheim, 2013), pp. 1–75
USP. <776> Optical Microscopy
USP. <778> Particulate Matter in Injections
W. Zhang, L. Liu, Study on the formation and properties of liquid crystal emulsion in cos-
metic. J. Cosmet. Dermatol. Sci. Appl. 3(2), 139–144 (2013). [Link]
jcdsa.2013.32022

View publication stats