Sponsored by

Optimizing Pharmaceutical Formulations:

The Role of Physicochemical Analysis
Contents
3 Introduction

4 Answering the Challenge of Setting Effective Particle Size Specifications

During Pharmaceutical Product Development

12 Discriminating Between Polymorphs of Acetaminophen Using

Morphologically-Directed Raman Spectroscopy

16 Laboratory PXRD Systems Can Address Pharmaceutical R&D Challenges

20 An Interview with Andrew L. Theophilus, R&D Manager -

Physical Properties, Chemo R&D

24 Comparing Your Material’s Structure to Literature and Patent

PXRD Patterns

26 Automated Morphological Imaging: Deeper Insight into your

Pharmaceutical Formulation or Product

30 An Interview with Paul Kippax, Pharmaceutical Sector Director -

Malvern Panalytical

2
Introduction
Developing drug products is a progressively competitive, expensive and high-risk venture that
when successful, can bring considerable humanitarian and financial reward. At all stages of the
pharmaceutical development process, developers need to increase efficiency while maintaining
quality and safety – beginning with drug discovery, through chemical and formulation
development, to manufacturing and QC.

In recent decades, formulation development and optimization has become an increasingly

systematic, knowledge-driven process. Developing a successful formulation relies on a detailed
scientific understanding of the behavior of individual components, and of the interaction and
combined effect of these constituents in the candidate formulation.

Malvern Panalytical’s unique range of physicochemical analysis techniques can be used to

accelerate the development of pharmaceutical formulations, enabling critical material attributes
such as particle size, shape, structure and composition to be measured, aiding reference listed
drug (RLD) deformulation, helping establish bioequivalence in vitro, and accelerating formulation
development efficiency for both innovator and generic products.

This eBook explores several ways in which Malvern Panalytical’s technologies and expertise can
assist pharmaceutical scientists in their quest to develop high quality, safe and effective products,
and accelerate their speed to market.

3
Answering the Challenge
of Setting Effective
Particle Size Specifications
During Pharmaceutical
Product Development
Introduction
Setting meaningful and realistic specifications is an
essential element of Quality by Design (QbD). Well-
defined specifications control product performance
since they derive from correlations between clini-
cal behavior and the variables measured routinely
during processing and for QC. This paper examines
the process of setting specifications, taking as an ex-
ample particle size, a critically important parameter
for many pharmaceutical formulations.
Multiple pressures on the pharmaceutical industry
and the regulatory bodies that control it are driving
the shift towards a risk-based approach to produc-
tion. Increasingly, regulators are focusing scarce re-
source on those less well-understood products and
processes with significant potential for harm, re-
sponding more flexibly to manufacturers who can
demonstrate sound understanding and control.
QbD is a key part of this change in emphasis, pro-
moting development of the knowledge through
which manufacturers will prove competence.
4
Setting appropriate product and processing spec-
ifications is fundamental to QbD. Clinical perfor-
mance is a function of key variables controlled by
and during the manufacturing process. Identifying
these Critical Quality Attributes (CQAs) and con-
trolling them within clearly-defined ranges delivers
acceptable product quality and the required in vivo
behavior. Getting specifications right is therefore es-
sential, but demanding. An overly-tight specification
will make manufacturing more difficult than it needs
to be, compromising production economics, yet one
that is too loose may result in product inconsisten-
cies that impact efficacy.
Here we consider the process of specification setting
in some detail, using particle size as an example pa-
rameter. Particle size influences the key performance
characteristics of many of the suspensions, inhaled
powders and tablets widely used in drug delivery.
Establishing the Need for
a Specification
FDA guidance enshrined in ICH topic Q6A states
that, ‘specifications are critical quality standards that
are proposed and justified by the manufacturer and
approved by regulatory authorities as conditions of
approval.’ Because specifications define product ac-
ceptability, the requirement for them is based on the
existence of a link between a variable and some as-
pect of performance. Within the framework of QbD,
such correlations establish a parameter as a CQA, the
control of which assures product quality. The regula-
tory guidance suggests these might include:
• Physicochemical properties such as pH,
refractive index, melting point
• Particle size
• Polymorphic form/amorphous content
• Chiral identity
• Water content
• Inorganic impurity levels
• Microbial content
Clearly if a variable has no influence on product qual-
ity there is no need to specify values for it; equally,
where there is a correlation, any associated specifica-
tion must provide adequate control. The best way to
define the specification will depend on the mecha-
nism by which the variable influences behavior; opti-
mal definition requires detailed understanding. QbD
emphasizes the need for relevant information at the
early stages of a project, and the importance of de-
veloping an adequate knowledge base. The pharma-
copoeial standards, development data, test data from
pre-clinical and clinical studies, and from stability tri-
als are all good sources of information.
Focusing on Particle Size
ICH topic Q6A contains a decision tree that deter-
mines the need for a particle size specification, rec-
ommending them for both solid dosage forms and
liquids containing undissolved drug suspensions
when particle size is critical to any of the following:
5
• Dissolution, solubility or bioavailability
• Processability
• Stability
• Product content uniformity
The dependence of dissolution behavior on particle
size, or more accurately surface area, is widely recog-
nized. For solid dosage forms, dissolution behavior
(including overall solubility), influences in vivo re-
lease and is often tightly controlled to obtain the re-
quired delivery profile. Finer powders dissolve more
rapidly, and those with a narrow size distribution all
dissolve at a similar rate, which may or may not be
the design intent. For OINDPs (orally inhaled and
nasal drug products), the link with bioavailability is
even more direct because incorrectly sized particles
will not deposit within the desired part of the respi-
ratory system, thus failing to enter the body by the
prescribed route.
Processability is important because it tends to cor-
relate with product consistency and quality, with er-
ratic or poorly controlled processes producing a low
quality output. The aim of QbD is to develop a man-
ufacturing process on the basis of sound science that
will deliver long-term optimal manufacture, reducing
the reliance on final QC testing. Particle size relates to
parameters such as powder flowability, segregation
behavior and, in the case of tableting processes, com-
pressibility. As such, well-defined size specifications
for in-process material or intermediates may provide
a means of reducing process variability.
Stability issues arise because the forces acting on
particles are a function of their size. Smaller parti-
cles are more likely to cluster or agglomerate be-
cause inter-particle forces are relatively high, while
with suspensions the propensity to settle increas-
es with size because of the effects of gravitational
forces. If settling does occur, then the chances of
consistent dosing are reduced - unless the active
ingredient is successfully resuspended before use.
Stability must be considered during the develop-
ment of particle size measurement methods, to
ensure that the state of dispersion during analysis
reflects that of the material.
Finally, particle size can influence blend uniformity
and, as a result, the dose content uniformity asso-
ciated with a product. For example, Figure 1 shows
a near-infrared (NIR) chemical image of a recalled
tablet with the active ingredient-rich domains
picked out in red. These areas are much larger on
the right side of the tablet than the left, showing
that composition is not uniform. The reason giv-
en for the recall of these tablets was that, ‘due to
larger particle size some of the unit doses may not
meet potency specifications.’ Here then is a direct
example of poor particle size control resulting in a
sub-standard product. Development of a particle
size specification may provide a means of control.
In summary, particle size specifications are often
required for solid dosage forms or suspensions be-
cause particle size influences so many aspects of
Figure 1. NIR image of a tablet showing
heterogeneous distribution of the active ingredient
6
product performance. Correlations with, for exam-
ple, dissolution profile, content uniformity, suspen-
sion stability, and/or mouth feel, form the basis for
specification development.
Measuring Particle Size
Distributions: Laser Diffraction
Before considering how particle size specification
can be set, it is worth reviewing the methods by
which particle size analysis can be carried out. There
are now a range of particle size analysis tools which
are available to support pharmaceutical develop-
ment. Selection of the most appropriate technique
is, in itself, an important consideration, with param-
eters such as the precision and accuracy of measure-
ment, the time for analysis, ease of method develop-
ment and transfer, and system cost being possible
selection criteria.
One of the most widely used techniques for parti-
cle size analysis within the pharmaceutical industry
is laser diffraction. Laser diffraction is a particle size
measurement technique with broad applicability,
suitable for both wet and dry systems with particles
in the size range 0.1 to 3500 microns. A particularly
attractive feature of the method is that it is an ide-
al Process Analytical Technology (PAT), so specifica-
tions developed in the laboratory can be transferred
into production, and if the right equipment is used.
How are Size Distributions
Represented?
Using laser diffraction as an example, the first issue
which should be understood in setting a specifica-
tion relates to how the particle size distribution is
represented. As an ensemble sizing method, laser
diffraction produces volume based distributions.
This means that data are generated simultaneously
for the whole sample (ensemble) rather than for in-
dividual particles, and are presented in terms of the
volume of material in a given size fraction. An alter-
native, used by imaging and microscopy methods
for example, is to display data on a number basis (i.e.
show the number of particles in each size fraction).
Both approaches are equally valid, but the results
produced are very different (see Figure 2).
Figure 2. Comparing volume and number
based distributions
The number and volume distributions shown in
Figure 4, clearly an extreme case, reflect different
attributes of the sample. The volume distribution
highlights those size fractions in which most of the
mass or volume lies, while the number distribution
captures the presence of fines. Which is the better
measure depends on the application; comparisons of
number and volume-based distributions are feasible
but best avoided if possible, because of the impact of
measurement errors. A better approach is to consider
this issue during analytical technique selection.
Defining Particle Size
Although the measurement of particle size distribu-
tions is extremely useful, specifications are seldom
7
based on the entire distribution, but instead nor-
mally consider specific distribution parameters or
statistics. Choosing the most relevant parameter
and making a judgement as to whether a single fig-
ure is adequate or whether multiple descriptors of
distribution are necessary are important steps in the
specification setting process.
Figure 3 shows an example size distribution, in this
case a mixture of two components, which has been
measured using laser diffraction. Using this, we can
consider which statistics relate to different regions
of the distribution.
Figure 3. Example size distribution, showing possible
parameters which could be used in a product specification
The median is the particle size in the middle of the
distribution. In the case of laser diffraction, this is
referred to as the Dv50, indicating that 50% of the
volume of particles are larger and 50% smaller than
this value (the notation ‘v’ shows that this param-
eter relates to a volume distribution). For the sin-
gle Gaussian distribution, median and mode (the
most common particle size) are identical; however
for the bimodal distribution such as the one shown
here, the mode reveals nothing about the smaller
peak. The median value on the other hand does re-
flect the presence of fines and is located between
the two populations.
A similar notation can be used to describe other
points within the size distribution. So, the Dv10 re-
lates to the particle size below which 10% of the vol-
ume of material exists, and is therefore positioned
towards the fine end of the distribution. In contrast,
the Dv90 is associated with the coarse end of the dis-
tribution. The Dv100 is also shown, and might seem
a logical choice if the aim is to avoid particles above
a certain size. However, using Dv100 makes the anal-
ysis sensitive to the presence of just one large par-
ticle, massively increasing the importance of repre-
sentative sampling. With modern instrumentation,
sampling is often the largest source of error, so this
approach is not recommended. Using Dv90 or even
Dv95 can provide a more robust parameter for as-
sessing the coarse fraction.
The other term commonly used for the description
of statistical data is mean or average. However, in
selecting a mean size parameter, it is important to
understand how the mean is calculated. The mean
may be ‘weighted’ according to different particle
parameters. Typical parameters include the num-
ber or arithmetic mean diameter, the D[1,0], surface
area moment mean, the D[3,2], or volume moment
mean, the D[4,3]. Clearly the different weightings for
each of these means will yield different results, as
shown in Figure 3.
The moment means are more commonly used for
specifications: the D[3,2] shows a greater sensitivity
to fines content, as the fine particles have a high
specific surface area, whereas the D[4,3] is partic-
ularly useful for the detection of larger particles,
as these contain the greatest volume of material.
Where the mean is the basis for a specification, se-
lecting an appropriately sensitive measure is there-
fore essential.
8
The Importance of Selecting an
Appropriate Mean
Consider a particle population consisting of ten par-
ticles of unit diameter and one with a diameter of
100 units. The number mean, or arithmetic average,
diameter or D[1,0] for this group of particles is calcu-
lated as follows:
D[1,0] = (10*1 + 1*100) / (10 + 1) = 10 units
This confirms that the D[1,0] is weighted towards the
fines. If the large particle breaks into two, each with
a diameter of 79.37 units (thereby conserving mass)
then the D[1,0] for the population becomes:
D[1,0] = (10*1 + 2*79.37) / (10 + 2) = 14.06 units
This is a 40.6% increase in size, and as such clearly
fails to reflect particle attrition.
On the other hand, we could consider the D[4,3].
Here, the proportion of particles at each size is
weighted according to their volume when calcu-
lating the mean size. So for the first population, the
D[4,3] is:
D[4,3] = (10*1*13 + 1 * 100 *1003) / (10 * 13 + 1 *
1003) ≈ 100 units
This confirms that the D[4,3] is weighted towards the
coarse material. The break-up of the large particle
into two results in the following D[4,3]:
D[4,3] = (10*1*13 + 2 * 79.37 *79.373) / (10*13 +
2*79.373) ≈ 79.37 units
This moment of distribution mean is clearly able to de-
tect the change caused by particle break-up, although
it shows little sensitivity to the fines population.
Selecting an Appropriate Parameter
Measuring how sensitive different particle size pa-
rameters are to changes in a sample is a good way
of determining the best basis for a specification.
Figure 4 shows particle size distributions, mea-
sured for the blending of two excipient grades,
generated using laser diffraction (Mastersizer 2000,
Malvern Panalytical).
Figure 4. Investigating the ability of a laser diffraction
analyzer (Mastersizer 2000) to detect fine and coarse
particles in a pharmaceutical blend
Here the aim was to assess the ability of the instru-
ment to detect small changes in the fines content.
Detecting the relative proportion of fines and coarse
material in a blend is a common application, since
particle size distribution often has an influence on
important properties such as flowability. In the for-
mulation of dry powder inhalers, for example, the
inclusion of fine excipient particles blended with
coarser carrier particles is an established way of
enhancing drug delivery, because it improves the
aerosolization process. Figure 5 shows how various
size descriptors change with increasing fine material
content in this case.
9
Figure 5. The impact of the percentage of fine material on
different size descriptors
For this example it is obvious that Dv10 and D[3,2]
can only track changes in the particle size at low
fines content (<10%), whereas Dv90 only shows
sensitivity at high fines contents (>40%). In contrast,
the Dv50 and D[4,3] show a linear change as the
percentage of fine material increases, and therefore
provide a means of differentiating between blends
across the range tested here. These parameters may
therefore be more appropriate for a specification, al-
though this does depend on the range of possible
blends which could occur in production. If the fines
content is never greater than 10%, the Dv10 or D[3,2]
may provide a more sensitive measure of change, as
the rate of change of these parameters is greatest
below 10% fines.
Setting Tolerances
Techniques such as laser diffraction offer excellent
repeatability, reproducibility and robustness, pro-
ducing good quality data with minimal manual in-
put or expertise. High repeatability means that the
same sample, run on the same system, produces
very similar results, so it is solely dependent on the
instrument itself and on reproducible sample dis-
persion. Reproducibility is a more demanding pa-
rameter that quantifies the variability introduced by
a change of operator, sample, time and instrument;
sampling methodology consequently becomes im-
portant. Measurement variability informs the final
step of specification setting: the determination of
acceptable tolerances.
Setting appropriate tolerances for a specification
requires an appreciation of measurement errors as
well as an understanding of the correlation between
the variable and product performance. Figure 6
shows some particle size distribution data, the red
line being a typical reading, the yellow and orange
lines quantifying reproducibility, the range of vari-
ability associated with the reading.
If the specification for this product is a Dv50 of 10
microns, then the plot shows that the measurement
variability will be +/- 5%. However, it would be wrong
to assume that the same tolerance can also be used
for the percentage less than 10 microns. Due to the
slope of the undersize distribution curve, a specifica-
tion requesting that 50% of the volume of material
Figure 6. Highlighting how accuracy is affected by the
wording or basis of a specification
10
must be 10 microns size or less would be subject to
a measurement variability of +/-14%.
This analysis highlights a fairly obvious point: as
measurement variability increases, confidence in
the reported result should reduce. Although this
concept is readily appreciated, it is not always prop-
erly accounted for in specifications. Here is a spec-
ification for a tableting blend [Evolutions in Direct
Compression, Douglas McCormick, Pharmaceutical
Technology, April 2005. Pg 52-62]:
Dv10 > 30 μm
D[4,3] > 80 μm
Dv90 < 1000 μm
The tolerances above do not take into account any
error introduced by measurement, but rather re-
flect the ideal material for optimal production. Dv10
should be greater than 30 microns, D[4,3] greater
than 80 microns, and the Dv90 should be less than
1000 microns. USP <429> [General Chapter <429>
(2008), “Light Diffraction Measurements of Particle
Size,” United States Pharmacopeia. www.usp.org]
provides acceptable reproducibility figures for par-
ticle size measurement, suggesting that the Dv50
and other central measures should be within 10%,
and noncentral values such as Dv10 and Dv90
should be within 15%. If the analytical technique
used delivers this performance, what does that
mean for the specification?
If the USP <429> reproducibility limits are applied, a
Dv10 measurement of 30 microns could be recorded
for a sample with a Dv10 of up to 34.5 microns (15%
error). If the aim is to ensure that the material actual-
ly has a Dv10 greater than 30 microns, the specifica-
tion will need to be changed accordingly. Applying
the same approach for each parameter produces the
following measurement specification which needs
to be followed if the manufacturing specification
above is to be met:
Dv10 > 34.5 μm
D[4,3] > 88 μm
Dv90 < 850 μm
So, tolerances should be reduced to incorporate the
errors introduced by the measurement technique,
with less accurate methods requiring tighter speci-
fications. This provides impetus for the procurement
of instruments with high reproducibility and the de-
velopment of robust methodologies.
In Conclusion
Setting specifications on the basis of product under-
standing is a key element of QbD and essential for
effective quality control. For many pharmaceuticals,
particle size is a CQA – a variable that must be tightly
controlled because it directly impacts performance
– so, specification and measurement is routine
11
within the industry. Recognizing the parameters
that different particle sizing techniques deliver and
the relevance of alternative size descriptors is crucial
for the development of a relevant specification, as is
the refinement of tolerances in line with the accura-
cy of the method.
Laser diffraction is an ensemble particle sizing tech-
nique that delivers volume-based distributions. It is
suitable for many applications within the sector and
is particularly sensitive for the detection of over-
sized material. Systems such as the Mastersizer 2000
deliver highly repeatable results (+/-1%) reducing
the need to tighten tolerances to account for mea-
surement inaccuracy. Furthermore, because laser
diffraction is well-established as a process analytical
technique, a specification developed in the labora-
tory can be transferred during scale-up. The success-
ful transfer of specifications that are as broad as pos-
sible, within the constraints of meeting performance
targets, is a route to optimized manufacture and re-
duced cost.
Discriminating Between
Polymorphs of Acetaminophen
Using Morphologically-Directed
Raman Spectroscopy
Introduction
Interest in studying mixed systems of pharmaceuti-
cal relevance is not a new concept. From intrinsic
polymorphism to complex final products, the chal-
lenges encountered by modern analytical scientists
have multiplied both in number and in complexity.
Established techniques are often unable to provide
both separation and identification of species in a
mixed system, as they are used to probe chemistry
(bulk spectroscopy), or individual particle character-
istics (microscopy and SEM). Combining techniques
capable of analyzing micron-scale particulates with
absolute chemical specificity provides a new win-
dow into describing the complicated mixtures com-
mon in today’s competitive pharmaceutical market.
Malvern Panalytical’s Morphologi® ID provides just
such a technology combination, uniting automated
microscopy and spectroscopy in an integrated pack-
age. This partnering of techniques has been termed
Morphologically-Directed Raman Spectroscopy
(MDRS®). Raman spectroscopy is sensitive to the
unique polymorphic forms of a substance, and
can be used to investigate single particles to assist
in determining the long-term stability of a prod-
uct. Correlations may be established between
12
morphological and spectroscopic results, where
morphologically-distinct particle groups are found
to share common spectra, and vice versa. The ability
of MDRS to target particles of distinct morphology
for Raman investigation provides added sensitivity
to the measurement, as spectra for particles that are
not of interest are not measured. Likewise, know-
ing that a group of particles with similar spectra may
be described by a common morphology allows the
analyst to assess particulate quantity in the absence
of measuring a spectrum for each particle in the
morphologically-similar group. Such correlations
between particle morphology and chemical identity
allow MDRS to provide an improvement in efficiency
over manual methods.
Materials and Methods
A polymorphic mixture of common pain-relief
medication was investigated using a Morphologi
ID instrument to characterize samples in terms
of particle size, shape, and chemical identity. The
polymorphic mixture consisted of the Type I and
Type II polymorphs of acetaminophen (also known
as paracetamol). Acetaminophen in its Type I form
(Fisher (98%, AC102332500)) was recrystallized by
slowly cooling a saturated solution in water at 60°C.1
The resulting material was recovered and dried at
75°C for eight hours, at which point half of the sam-
ple (Type I) was removed and the other half heated
to 170°C to effect the Type I to Type II polymorphic
conversion.2 A volumetric fraction consisting of 9
parts Type I to 1 part Type II by weight was milled
for 1 minute in a hand mill. 3 mm3 of the resulting
mixed powder was placed into the dry dispersion
accessory for the Morphologi ID instrument and dis-
persed at a pressure of 4 bar onto a quartz plate.
A 20 mm diameter area was scanned at 5x magnifi-
cation to detect acetaminophen particles, and 1375
particles were selected for Raman measurement
based on the following size categories: greater than
50 μm, 25 μm to 50 μm, and 10 to 25 μm.
Reference spectra for Raman analysis were obtained
from pure materials. Spectra were collected with a
785 nm laser (20 mW) illuminating a 3 μm spot size
at 50x magnification, and covering a range from
150 cm-1 to 1850 cm-1 at 6 cm-1 resolution.
Figure 1. Representative particle images from the polymorph acetaminophen mixture (left panel) and the size distribution
(volume weighted, right panel).
13
Results
Over 70,000 individual particles were detected for
the 9 (Type I):1 (Type II) acetaminophen mixture.
The sample was polydisperse, with some particles
greater than 400 μm detected. Examples of the par-
ticles analyzed and their final size distribution are
shown in Figure 1.
The morphological results show that the hand mill
produces a broad distribution of particle sizes. It
is also apparent that a few large particles (greater
than 200 μm) are still present in the final disper-
sion. The size distribution result is used to classify
the particles based on size (10 μm to 25 μm, 25 μm
to 50 μm, and greater than 50 μm) for Raman in-
vestigation. Particles below 10 μm were excluded
from Raman analysis.
The results of correlation against the library of pure
polymorph spectra revealed 1,220 Type I particles
and 144 Type II particles, out of 1,375 particles tar-
geted, corresponding to 10.5% Type II polymorph.
The pure reference spectra and examples of particle
spectra are shown in Figure 2. The correlation was
constrained to contain only the 1200 cm-1 to 1260
cm-1 and 1540 cm-1 to 1590 cm-1 regions (indicated
by the green areas in Figure 2) where the two spe-
cies are readily distinguishable. A correlation score
of greater than 0.85 was required for a positive iden-
tification as either Type I or Type II. Mixture spectra
(correlation to both Type I and Type II references
with scores greater than 0.85) were detected for 11
particles, an example is also shown in Figure 2.
The agreement between the recorded and library
spectra is outstanding, and allows for simple dis-
crimination between the two polymorph types
within the confines of the correlation analysis. The
Figure 2. Raman spectra collected from example
particulates for acetaminophen Type I (blue),
acetaminophen Type II (red), and a mixture spectrum
(black). The Type I and Type II spectra are superimposed
over reference spectra used to create the library (black
dotted). The portions of the spectrum in green are those
used for correlation to the reference library.
14
mixture spectra identified for a subset of particles
are most likely due to the formation of aggregates
during dispersion.
Once the spectroscopic data is available, the mor-
phologies of the two species can be reviewed to
determine if it is possible to segregate similar look-
ing materials based solely on their morphologies. In
the present example, results confirm that the Type I
particles have a higher Intensity Mean than the Type
II particles, meaning that Type I particles are more
transparent, as shown in Figure 3. However, while a
clear difference in the transparency of the two parti-
cle types exists, it should be noted that there is some
overlap between the Intensity Mean distributions
which suggests that it is not a definitive morpholog-
ical separator. No other morphological descriptors
indicated a significant difference between the par-
ticle types, most likely due to loss of morphological
information during the milling process. This situa-
tion emphasizes the importance of having a ‘referee’
technique, such as Raman spectroscopy, available to
discriminate between different chemical species.
Figure 3. Intensity Mean distribution for the Type I and
Type II particles as determined by the Raman spectroscopic
results. Insets display example images of particles from
each group.

Discussion
Polymorphic conversions are considered an ongo-
ing problem in the pharmaceutical industry. Process
conditions, improper storage and simply time can
cause the transition from a stable, efficacious form
of a drug substance into an unwanted structural
analog. Thus, establishing the presence of multi-
ple polymorphic forms is of great interest to both
innovator and generic pharmaceutical companies
alike. Vibrational spectroscopy has found wide use
as a tool with the requisite sensitivity to establish
the presence of multiple polymorphic forms in a
mixed system. Unfortunately, commonly-available
bulk spectroscopic techniques lack the sensitivity to
detect the low levels of polymorphic contaminants
that may be present in final products. Additionally,
bulk techniques provide only an overview of the ma-
terial that is present in the sample area and have no
ability to partition the sample into individual parti-
cles to provide specific characterization.
Conclusion
MDRS provides clear advantages over traditional
methods of vibrational spectroscopy and is well-suit-
ed to address low-level polymorphic contamination
15
directly. Spectra are collected for single particles,
which allow at least an order of magnitude increase
in detection sensitivity. Data collection can be con-
ducted for any number of particulates desired, and
unique spectra are easily identified. As mentioned
previously, the morphological characteristics of
particulates can also be used to further refine the
selectivity of the method. If different polymorphic
forms have distinct morphology, the method can
be modified to investigate only those that fit a par-
ticular morphological profile. These strengths posi-
tion MDRS as a technique that can handle the most
difficult polymorphic mixtures with unparalleled
sensitivity and flexibility, providing a new tool for
root cause analysis, product quality and formulation
development.
References
1. Lee T., Kuo C.S., and Chen Y.H. Oct. 2, 2006. Solubility,
Polymorphism, Crystallinity, and Crystal Habit of Acetaminophen
and Ibuprofen by Initial Solvent Screening. Pharmaceutical
Technology.
2. Di Martino P., Conflant P., Drache M., Huvenne J.-P., and Guyot-
Hermann A.-M. 1997. Preparation and Physical Characterization of
Forms II and III of Paracetamol. Journal of Thermal Analysis (48):
447-458.
Laboratory PXRD Systems
Can Address Pharmaceutical
R&D Challenges
Detection of trace polymorphs, determination of
amorphous content, and the study of the stability of
active pharmaceutical ingredients (APIs) subjected to
heat and humidity are some of the many applications
where a modern powder X-ray diffraction (PXRD) sys-
tem can add value to drug development. Advances
in beam conditioning and detector technology have
led to greatly improved detection limits, resolution
and range of applications now possible. Today’s mod-
ern, laboratory-based PXRD instrumentation and ad-
vanced software can greatly assist pharmaceutical
scientists in their quest to develop high quality, sta-
ble products and accelerate their speed to market.
All applications described and many more can be
performed on a single PXRD instrument, such as the
Malvern Panalytical Empyrean.
Detecting the Presence of Trace
Polymorphs and Impurities
The detection of trace polymorph phases is one of
the most important applications for PXRD in the
pharmaceutical industry. While classical PXRD con-
figurations were purported to have a limit of detec-
tion (LOD) in the range of 0.5- 2%1,2,3 modern systems
with specially designed optics can have detection
limits below 0.1%, up to an order of magnitude be-
low the commonly believed detection limit. These
16
optics achieve excellent LOD by suppressing the
continuous radiation (Bremsstrahlung) and there-
fore improving the signal to noise ratio in the data.
One such optic for greatly improved limits of de-
tection is the Malvern Panalytical BraggBrentanoHD
(BBHD) optic.
The following example shows part of the diffrac-
tion pattern for an Indomethacin (IMC) mixture
of two crystalline polymorphs in the excipient.
The mixture was made with 0.25% alpha-Indo-
methacin, 4.75% gamma-Indomethacin and two
excipients (98.5% alpha-lactose monohydrate,
1.5% Magnesium stearate). Figure 1a. shows the
alpha-Indomethacin peak measured with a BBHD
optic with a significant improvement of the LOD
to 0.12%, as compared to standard slit optics in
Figure 1b. BBHD offers further benefits to phar-
maceutical research, with its improved low angle
performance as compared to slit optics, and the
suitability to be used for several applications be-
yond powder diffraction, including small angle
X-ray scattering (SAXS) and micro-spot analysis.
Determining Amorphous Content
Accurate determination in the 1-10% amor-
phous content range is now possible.4 PXRD

Figure 1a. 0.25% alpha-indomethacin peak measured with BBHD optic; 1b. 0.25%
alpha-indomethacin peak measured with slit optics
instrumentation advances now
make it possible to produce high-
er quality data combined with
modern approaches to the quan-
tification of amorphous content.
Figure 2 shows an example
whereby 10 α-lactose mono-
hydrate standards created with
Figure 2. PXRD patterns for ten α-lactose monohydrate standards with 1-10%
amorphous content
amorphous content range be-
tween 1 and 10 percent; the re-
sulting data set was then analyzed
using two different methods to
quantify the amorphous content:
1. peak to background
ratio method
17
2. partial least squares
(PLS) regression
The PLS method is used by other
techniques as well, such as FT-IR
for pharmaceutical amorphous
determination, and is ideal for
cases where discernible peak(s)
attributed to a phase are not
found in the pattern. “PLS uses the
whole spectrum or a selected re-
gion of the spectrum to develop
calibration models.”5 As shown
in Figure 3, the PLS method pro-
duced a calibration curve with
a significantly better standard
deviation (0.035% vs 0.262% for
the peak to background meth-
od), higher coefficient of deter-
mination (R2) which equates to
a better fit of measured to pre-
dicted values, and a lower error
of intercept value (0.024% versus
0.179%). This shows that PLS can
be a more reliable and accurate
method to determine low amor-
phous content.
Exploring
Co-Crystallization
Parameters and
Polymorphs
The investigation of polymorphs,
chemically identical substanc-
es present in different crystal-
lographic forms, is important in
drug development. APIs often
occur in a variety of polymor-
phic forms, potentially displaying
 Figure 3. Comparison of confidence limits for two methods of amorphous content determination

different dissolution properties,

and as a result, different bioavail-
ability. To optimize critical prop-
erties of a drug, polymorph anal-
ysis and salt screening may aid in
the selection of a suitable solid
form of a drug. The discovery and
protection of all polymorphs by
patent allows drug inventors to
secure intellectual property and
market position.

PXRD systems equipped with

an X-Y-Z-stage for combinatori-
al screening of samples in a well
plate can perform high resolu-
tion, high throughput screen-
ing when using a transmission
configuration.
To compare, group and identi-
fy the crystalline (or amorphous)
phases formed in the wells, the
processing of data is greatly
Figure 4. Cluster analysis of PXRD patterns measured on a 96 well plate
aided by automated cluster analy- and temperature combinations.
sis software. Cluster analysis (Figure 4) of the
PXRD patterns collected from
In this example, cimetidine from
MilliporeSigma was dissolved in each well showed three clus-
various solvents and dispensed ters of closely related scans; fur-
into 96-well plates to recrystal- ther analysis of representative
lize under various concentration scans from the clusters could

18
be performed for indexing and
full pattern fits. Cluster analysis
greatly aids in the sorting and
analysis of large amounts of data.
Stability of Materials
Subjected to Varying
Heat and Humidity
Storage and transport of pharma-
ceutical substances can expose
them to heat and/or high relative
humidity (RH), which in turn can
induce unexpected phase trans-
formations that can affect a prod-
uct’s efficacy.
Modern PXRD systems can be
equipped with the latest tem-
perature-controlled humidity
Figure 5. RH - T phase diagram derived from constant temperature and constant
relative humidity experiments
stages, allowing the measure-
ment of PXRD data at RHs up to
95% at room temperature and
75% at 60 degrees C.
In this experiment, crys-
talline trehalose dihydrate
(C12H22O112H2O), an important
cryoprotectant used in the pro-
duction of protein-based drugs
that undergo lyophilization, was
exposed to a range of tempera-
ture and relative humidity com-
binations within a nonambient
stage on a Malvern Panalytical
Empyrean X-ray diffraction sys-
tem. The PXRD patterns were
analyzed to determine which
phases were present at each
condition, and a resultant RH-T
19
‘phase’ diagram could be devel-
oped. Trehalose dihydrate (Th)
was seen to be very stable over
a wide range of temperatures.
Two polymorphs of trehalose an-
hydrate were observed: a stable
form Tβ and an unstable form
Tα. An amorphous form of treha-
lose (Ta) was also observed in the
high temperature – low humidity
regime. Important to note is that
once Tβ formed, the phase did
not return to the Th form upon
cooling – a permanent transfor-
mation occurred.
References
1. http://www.nf-itwg.org/pdfs/ITWG-
INFL-PXRD.pdf, Page 6
2. http://www.xrd.us
3. http://old.vscht.cz/clab/RTG/
dokumenty/panalytical/xrd/X-ray%20
powder%20diffraction%20-%20
a%20powerful%20tool%20in%20
pharmaceutical%20analysis.pdf
4. http://www.
americanpharmaceuticalreview.
com/Featured-Articles/36758-
Approaches-to-Quantification-of-
Amorphous-Content-in-Crystalline-
Drug-Substance-by-Powder-X-ray-
Diffraction/
5. S. Byrn, G. Zografi, X. Chen, Solid State
Properties of Pharmaceutical Materials,
John Wiley and Sons, 2017, p.337
An Interview With...

Andrew L. Theophilus
R&D Manager - Physical Properties
Chemo R&D

During his time in academia, Andrew originally studied biological sciences in Cardiff, specializing in respiratory
physics. In 1991, Andrew joined Glaxo Group Research, formulating and analyzing inhaled medicines. Andrew went
on to specialize in the physical analysis of raw materials and products, expanding into oral and other formulation
types. During his 24 years with Glaxo, Glaxo Wellcome and GlaxoSmithKline, he took on roles of increasing respon-
sibility, ultimately managing a materials science group and representing GSK in various academic collaborations.

Three years ago, Andrew moved to Spain and joined Chemo R&D in Madrid. As R&D Manager - Physical Properties,
he is responsible for defining the physical characteristics of API, excipients and formulated materials, across a broad
portfolio, to ensure consistent product performance.

different medicines. Method development and val-

As someone who has worked on the
development of both generic and
innovator drug products, have you
noticed any differences between the
physicochemical analysis approaches
used in the generics industry
compared to those used by
innovators, and do you see any
opportunities for analytical learnings
to be shared between the two?
One of the main differences I have seen in generic
companies is that there are a greater variety of proj-
ects and the pace and throughput is much faster
compared to innovators. I encounter a large port-
folio of APIs (Active Pharmaceutical Ingredients),
excipients and final dosage forms across a lot of
idation needs to be efficient. Methods themselves
need to be rapid, automated if possible, yet still pro-
ducing high quality data.
Generics, it would seem, have the advantage of
background information on a product, and this is
true to some extent; no doubt there will be patents
and literature relating to the API and product. But as-
sessing this literature throws up problems. Take poly-
morphism, for example. There are sure to be patents
covering numerous polymorphs – I have seen some
materials with in excess of 20 forms reported. Some
of these are likely to be mixtures of forms, or dupli-
cates patented by different companies, and there
are bound to be huge gaps in the available data.
Seeing the wood for the trees is tricky. You need to
be a detective and find the clues.

20
For innovators the issues are different. Characterization
is much more targeted. More effort can be spent on
specific molecules. But it is necessary since you are
starting from scratch with a new material, perhaps
with very little material to work with - just a few mil-
ligrams in many cases. It’s likely that each small scale
batch will be different to the last, with changes in
particle size, shape, form, water content and so on.
The challenge of understanding a material that has
never existed before is exceptionally rewarding. The
use of complementary techniques is key. There is no
one technique that will tell you all you need to know.
X-ray diffraction, thermal analysis, Raman and IR spec-
troscopies are the minimum for understanding solid
state, with ssNMR and other techniques to try to fill
some of the gaps. Then surface technologies like va-
por sorption and microcalorimetry if any disorder or
amorphous material is involved. As more material
becomes available during scale up, particle size and
surface area go without saying (but careful develop-
ment of a particle size method is critical, even at this
stage). But at every stage the most important tech-
nique of all is microscopy. You have to visualize the
material, see it with your own eyes. A picture paints
a thousand words, but microscopy underpins a thou-
sand techniques.
22
You are involved in the development
of multiple dosage forms. Could you
comment on the challenges of
characterizing the physical
attributes of such a wide variety
of products?
Final dosage forms are often an interesting chal-
lenge. The properties of the API are clearly import-
ant, but the properties and functionality of the
medicine heading for the patient is critical for effi-
cacy and safety. The technologies for examining the
properties of these various dosage forms are limit-
ed in comparison to those for studying API. Various
microscopy-based imaging systems with image
analysis are available, as well as chemical imaging
technologies based on various spectroscopic tech-
niques. These have improved the situation a great
deal, but they have their limitations, particularly the
resolution of small particles.
Some of the more tricky formulations relate to
creams and ointments. Isolating or characterizing
the API, in situ, in complex multicomponent colloidal
systems require a variety of different technologies,
mainly microscopic, and the choice is highly depen-
dent on API concentration and the presence of addi-
tional components. When it is necessary to measure
the API particle shape, size and polymorphic form,
it helps to have an arsenal of possible techniques
at your disposal. Injectable suspensions present a
similar challenge. Again, analysis in situ would be
ideal, but many techniques are confounded by the
dispersion media or the glass container. Some anal-
ysis techniques can be used – Raman is always one
option as well as optical microscopy, but sometimes
isolation from the dispersing medium by drying or
filtration is the only approach. But clearly the con-
sequences of either approach include the possibility
of damaging or changing the API or crystallizing out
soluble components.
Are there any frustrations or areas
for improvement you can identify
with regard to the types of
analytical data required for
regulatory submission?
Certain physical techniques appear, almost by
default, in regulatory specifications. In the same
way that identity, assay and impurities will always
feature in a release specification for API, particle
size will normally feature as well, for example. I be-
lieve this is driven by an assumption that particle
size will always influence processability or tablet
dissolution, and also by the apparent ease and
availability of particle size equipment. In many
cases this is right, but particle size is not the only
characteristic that is important for a product’s
performance, and sometimes it is of very little im-
portance at all. Sometimes particle shape is more
important in processability, and, as theory tells us,
the surface area of an API should be the main driv-
er in dissolution. Correlation with product per-
formance is the key. It is so important to identify
which attributes of a material actually influence
23
the performance of a medicine, be it dissolution
or in vivo testing. If, for example, surface area is
the better parameter to control tablet dissolution,
there is no reason why it shouldn’t be the primary,
or indeed only critical quality attribute. I know that
regulatory agencies will support this approach if
they are provided with a good scientific argument
and the data to support it.
Can you comment on the use of any
emerging technologies which provide
additional information on a product’s
physical properties and thus its
predicted quality and performance?
There are some amazing new technologies for the
characterization of APIs and formulated materials.
There have been great strides forward in imaging
and related technologies such as Raman and X-ray
diffraction imaging, surface analysis technologies
such as various gas and vapor sorption techniques,
as well as huge advances in computing power, driv-
ing modelling and prediction technology. But per-
haps we should take time to think about what we
are actually analyzing. It is great that we can ana-
lyze samples with higher and higher resolution, and
greater and greater precision, but if the sample we
are analyzing is not representative of the bulk mate-
rial we are trying to characterize then these advanc-
es are missing important information. Something
every physical scientist must remember is that the
materials they work with are heterogeneous, and as
such it is important to understand the entire com-
position of these materials. For example, the correct
sampling of a 10 kg batch of API, is as important as
any characterization technology you subsequently
apply to it.
Comparing Your Material’s
Structure to Literature and
Patent PXRD Patterns
Patents and literature can be sources for digital or
printed diffraction patterns, which may be needed
to compare to material you are producing. A bitmap
converter found in Malvern Panalytical’s Highscore
(Plus) software1 allows digital or printed images to
be digitized into an XML file suitable for diffraction
analysis and comparison. This allows all related pat-
ented polymorph patterns to be added, along with
patterns calculated from single crystal structure
solution data, into one Search/Match database for
comparison purpose to ensure that synthesized APIs
do not infringe patents. In this example (see Figures
1 and 2), images of powder X-ray diffrac­tion (PXRD) Figure 2. Cluster analysis of the patterns found in
patterns in patent US 8217061 B22 were converted US 8217061 B2
into XML files and compared by cluster analysis to
look for similar patterns. The seven patterns formed
two clusters of two similar patterns, and three were
found to be statistically different from the clusters.
Digitization of PXRD data from patents and litera-
ture can be a powerful way to ensure that patent
violation is not occurring and can confirm a new
entity is unique.

References
1. T. Degen, M. Sadki, E. Bron, U. König & G. Nénert, The HighScore
Suite, Powder Diffr. Vol. 29, (2014), Supplement S2, S13 - S18.
Figure 1. Digitized PXRD patterns from
patent US 8217061 B2 2. https://patents.google.com/patent/US8217061B2/en

24
Automated Morphological
Imaging: Deeper Insight into
your Pharmaceutical Formulation
or Product
Fully-automated morphological imaging of a phar-
maceutical formulation or product can deliver particle
size and shape information in a single measurement,
and with minimal user intervention. Simultaneous de-
termination of chemical composition by Raman spec-
troscopy delivers the power to unlock complex particle
characterization problems.
A knowledge and understanding of particle shape as
well as particle size is essential in the development,
control and optimization of pharmaceutical formu-
lations and final products. Particle characterization
using image analysis-based systems can provide real
insight into the nature of particulate materials, sup-
porting each stage of pharmaceutical development,
from the selection of raw materials, via formulation
development, right through to QC. From an intel-
lectual property perspective, this information can
be used to tighten the specification on a particular
product and make it difficult to replicate. This type of
detailed analysis is now easily achievable, thanks to a
new generation of automated analyzers, the best of
which provide size and morphological information
26
by imaging individual particles and generating a va-
riety of quantitative shape data.
A concurrent measurement by Raman spectroscopy
of each particle can provide the complete picture of
what is present in a multicomponent system, or per-
haps what unwanted particle type or contaminant is
present in a product. This powerful combination of
technologies can also be used to speed the defor-
mulation of the most complex pharmaceutical prod-
ucts, helping to establish in vitro bioequivalence
with the reference product.
Why Particle Shape is Important
Even very small differences in particle size or shape
can have a significant effect on the properties and
performance of materials. In pharmaceuticals,
for example, bioavailability, flowability, stability,
blending or tabletting efficiency may all be affect-
ed. Measuring particle size alone may not be sen-
sitive enough to identify the important but subtle
differences between samples. Particles with the
same overall size but very different shapes may
be characterized as identical
by some techniques. Applying
image analysis is a highly effec-
tive way of simultaneously mea-
suring particle size and shape.
Furthermore, particle charac-
terization by image analysis is
complementary to other parti-
cle sizing methods such as laser
diffraction. So, not only does it
provide additional information,
it also offers a means of validat-
ing other techniques.

Size and Shape

Definitions
Describing a 3-dimensional par-
ticle can be much more complex
than it first appears. While it can
be convenient to use a single
number, if the particle is not a
perfect sphere there are multi-
ple ways in which its size could
be described (Figure 1). Image
analysis captures a 2-dimension-
al image of a 3D particle, from
which it calculates various size
and shape parameters. Principal
among these is circle equivalent
(CE) diameter, which can be used
to calculate particle size. The cap-
tured image is converted to a cir-
cle of equivalent area and while
differently shaped particles will
influence the CE figure, it has the
virtue of being a single number
Figure 1. Key particle dimensions
that becomes larger or smaller as
the particle does. Importantly, it
is both objective and repeatable.
Particle shape is an even more
complex challenge and requires
the measurement of many pa-
rameters to build a complete
picture. The calculation of multi-
ple shape parameters, including
convexity, elongation and so-
lidity, for every particle, and the
generation of distributions for
each, allows the identification
and quantification of even very
subtle differences.
Sample Preparation
Challenges
A significant challenge in making
particle characterization by im-
age analysis a routine operation
is the need for fast and consis-
tent sample preparation. Sample
preparation must be reproduc-
ible, and samples must be homo-
geneously distributed, usually on
a glass slide or plate which is then
placed on an x-y stage for imag-
ing. Reproducible sample prepa-
ration will ensure that results are

27
repeatable, while homogeneity allows the option
of analyzing a subsection, knowing that it is repre-
sentative of the whole. This translates into shorter
measurement times and higher throughput, making
image analysis suitable for routine use. Dry powders
pose a particular challenge, requiring strict control
of dispersion conditions. Reducing sample prepara-
tion times and eliminating user bias demands a high
level of automation. One highly successful solution
is the integration of a sample dispersion unit within
the particle characterization system itself.
Adding Component-Specific
Microstructural Insight
Identification of the different components with-
in a formulation or product completes the particle
characterization picture. Morphologically-Directed
Raman Spectroscopy, or MDRS®, combines the tech-
niques of automated particle imaging and Raman
microspectroscopy in an integrated platform, to en-
able rapid, automated chemical and morphological
characterization of the individual components in
a multi-component sample. Multiple applications
benefit from this, including:
• Establishing the particle size of the active
pharmaceutical ingredient (API) within a
formulation
• Understanding the impact of process steps or
changes on the API or excipients
• Pinpointing the origin and identity of any
contaminants
Comparing the particle size and shape of the API in
reference listed drug (RLD) or generic (test) formula-
tions – useful during deformulation and when inves-
tigating in vitro bioequivalence
28
The first stage of an MDRS measurement is automat-
ed image analysis of the dispersed sample, to de-
termine the size and an array of shape parameters
for each individual particle, together with its precise
position on the measurement slide. This information
is then used to direct the Raman spectroscopic anal-
ysis, the second stage of the technique.
At its simplest, Raman analysis is performed on a
randomly-selected subset of the particles, allowing
the user to chemically identify a sample’s multiple
components, create particle classes based on these
assignments, and determine the relative propor-
tions of each component. A morphological descrip-
tion of each chemically-identified component can
then be generated in the form of size and shape
distributions, using the data from the automated
image analysis.
If particular components in the blend are of interest,
the measurement can be refined by selectively tar-
geting those particles with a distinct morphology
for chemical analysis; for example, particles within a
certain size range, or with shape parameters above a
specific value. By focusing the Raman measurement
only on the components of interest, more of these
particles can be characterized in a given timeframe.
Accelerating Bioequivalence
Assessments
The application of MDRS in support of in vitro bio-
equivalence studies has recently been shown in the
approval of an abbreviated new drug application
(ANDA) for a locally-acting mometasone furoate na-
sal spray.1
MDRS enables determination of the size and shape
of specific components within a multi-component
Figure 2. Use of MDRS in establishing particle identity, componentspecific particle size distribution, and comparability
between reference and test formulations.
blend. For the purposes of the ANDA, this capabil-
ity was used to compare the form and particle size
of the API in the reference and test nasal sprays and
confirm their close similarity as seen in Figure 2.
The resulting data from a Malvern Panalytical
Morphologi system were accepted in lieu of a clinical
bioequivalence (BE) endpoint study, setting a prece-
dent for the inclusion of in vitro studies of bioequiv-
alence in future ANDAs – this could help reduce the
cost and complexity of tests associated with a sub-
mission and accelerate new products to market.
Conclusion
Knowledge and understanding of particle shape as
well as size is now essential for many applications.
The advent of automated image analysis-based
particle characterization systems, with powerful
29
measurement and data analysis capabilities, is al-
lowing the more widespread use of this technique
research, development, and also quality assurance.
Automated sample preparation, integrated as part
of the measurement process, has proven to be one
of the major keys to even higher throughput with
excellent reproducibility, especially for dry powders.
The integration of complementary technologies,
such as Raman spectroscopy, provides the compo-
nent-specific microstructural insight which is crucial
for a deeper understanding of both pharmaceutical
products and processes.
References
1. FDA/CDER SBIA Chronicles; FDA embraces emerging technology
for bioequivalence evaluation of locally-acting nasal sprays:
https://www.fda.gov/media/97705/download
 An Interview with
Paul Kippax
Pharmaceutical Sector Director at Malvern Panalytical
Can you tell us a little about your
journey with Malvern Panalytical?
I came to Malvern Instruments, as Malvern
Panalytical was then known, almost 23 years ago,
following the completion of my Ph.D. in physical
chemistry and colloid science. My first role was
as an applications engineer, but I moved into our
Product Management department in 2002, and
stayed there for 16 years!
I’d always been interested in the application of par-
ticle characterization technologies, but it was our
Spraytec spray particle and droplet sizing system,
and an understanding of how it was being used
in nasal spray and inhaler product development,
which first introduced me to the pharmaceutical in-
dustry. I recall attending the UK’s Drug Delivery to
the Lungs conference back in 1999, where I heard
how our customers were struggling to find a way to
control drug delivery from aerosol-based devices.
Realizing that Spraytec could help, I started work-
ing on collaborative projects to see if we could gen-
erate the relevant data to understand drug prod-
uct performance. The experience I gained enabled
30
me to move on to managing the development of
the company’s Mastersizer laser diffraction parti-
cle sizing system and, finally, to leading the prod-
uct management group associated with Malvern
Panalytical’s complete portfolio of morphological
characterization solutions.
We’ve long known that understanding our custom-
ers’ challenges is key to helping us understand where
to focus and innovate. Over the past couple of years,
we’ve made great strides in this area and grown into
a truly customer-centric organization by creating
dedicated teams focused on developing deep un-
derstanding of the industries we serve. As a result of
this, I jumped at the chance to move into my current
role as Pharmaceutical Sector Director, which, from
the beginning of this year, has enabled me to con-
centrate on pharmaceutical industry applications.
How does Malvern Panalytical ensure
it stays close to its customers and
understands their needs?
We’re constantly improving the way in which we work
with our customers and, through this, developing
our understanding of how we can support them in
being more effective or generating additional value
from their own projects.
We’ve always had an active account management
program, where we partner with key customers.
Here, we aim to become a trusted member of a cus-
tomer’s team, maintaining open lines of communica-
tion and support. Often this has been based on the
supply of a single instrument or solution within their
organization. However, we’ve always understood
that our technologies are not stand-alone but form
part of our customers’ workflows alongside other
analytical methods. We have therefore always tried
to give good advice on method selection in line with
our customers’ product and process development
and control requirements, even when this involves
specifying solutions we do not provide! This ensures
that when our solutions are applied, this is done in a
way which enables the identification of relevant crit-
ical material properties, enabling a QbD approach to
product development.
We are heavily engaged with our academic net-
work and often present applications work at sci-
entific conferences. We do this not to just promote
– we aim to provoke a response from the audience
and gain more insight into how our solutions are
currently applied or could be developed in the fu-
ture. Here, we’re interested in the realistic, relevant
and valuable application of our technologies – and
where they can be used synergistically.
As an organization, we involve our customers and
key opinion leaders is in the development of our
solutions – testing prototypes of software and
measurement workflows and helping us under-
stand what’s most important to them. Recently,
we’ve worked directly with our customers in the
31
development of our Morphologi 4 range of an-
alytical imaging solutions. Understanding the
market’s desire to establish the bioequivalence of
formulations in vitro, and also hearing direct feed-
back from the US FDA, helped us to recognize the
relevance of the Morphologi 4-ID, which provides
component-specific chemical identification. We
then knew how to build on the capabilities of this
technology so it could be applied more effectively
within our customers’ development projects.
For another of our solutions launched last year, the
Zetasizer Ultra, we received significant feedback on
our proposed analytics workflow from customers
very early on in development. Some of our custom-
ers then helped us beta-test the product, which was
a huge help with the prioritization of its software
development, shaping the solution into the success
story it is today.
How does Malvern Panalytical decide
which technologies to develop next?
We’ve been engaged in providing analytics to the
pharmaceutical industry for many years now, so
we keep a close eye on industry trends. During the
1990s, the use of Malvern Panalytical technologies
within the pharmaceutical industry grew dramati-
cally, as the importance of microstructure in defin-
ing product and process performance increased.
That growth gave us the foundation to provide our
userbase with a commitment that we would ensure
consistent access to the capability to measure crit-
ical material attributes such as particle size – this
translates into a promise to continually support our
customers, which feeds into how we maintain our
solution portfolios.
When we start to consider what drives our direc-
tion, there are several elements at play.
Firstly, we’re seeing an increase in interest in data in-
tegrity. This has been prompted by regulatory guid-
ance such as the US FDA’s 21 CFR Part 11 rule, relat-
ing to electronic records and signatures. However,
data integrity goes deeper than simply considering
how data is stored; it also challenges whether the
measurement data itself is relevant to controlling
product quality and whether it has been generat-
ed in a product-relevant way. So, although applica-
tion of data integrity principles is often considered
a QC issue, it actually extends all the way back to
decisions made and analytics applied during the
development process. Our customers’ focus is to
consider what is required for the patient and the
treatment of a given disease. Malvern Panalytical
can help with the selection of relevant physico-
chemical measurement parameters and methods
to ensure that realistic and reliable data are gener-
ated to guide good development decisions, which
in turn ensure that the product is safe and effective.
We’re also very aware of global requirements to
drive down the cost and increase the availability of
both existing and new medicines. Important within
this is encouragement of the development of more
generic products, particularly for complex drug
products and formulations. Regulators are consid-
ering the application of in vitro approaches to as-
sessing bioequivalence as a surrogate for clinical
endpoint studies, in order to aid generic product
development. These approaches consider the im-
portance of the microstructure of the dosage form
in defining the effectiveness and reproducibility
of drug delivery. Malvern Panalytical’s solutions
provide relevant data to support and direct these
32
in vitro bioequivalence studies. We’re also work-
ing with our customers to understand how micro-
structural understanding can be applied in innova-
tor product development workflows to speed the
identification of candidate formulations and en-
able new products to be brought to market more
quickly.
Finally, we value working with our customers to
guide us on appropriate routes forward. In fact, this
works both ways! We’ve learned how our analytical
solutions fit alongside other techniques to guide
new product development. We know that the solu-
tions we provide are being used to help eliminate
drug candidates or candidate formulations that are
unlikely to be viable at a much earlier point in de-
velopment than was possible in the past, optimiz-
ing and speeding the drug development pipeline.
Alongside this, we recognize the pressure on our
customers to rapidly assimilate multiple analytical
datasets in order to inform their development de-
cisions. With the desire to compress product devel-
opment times, and the requirement for scientists
to support multiple concurrent projects, there is
often not enough time for our customers to devel-
op and retain deep knowledge regarding the appli-
cation of specific techniques. We therefore partner
directly with development teams in order to aid
realistic method selection. Here, the recent addi-
tion of the contract research organization Concept
Life Sciences to the Malvern Panalytical platform is
helping us on many levels: from understanding the
industry and its requirements to implementing a
given technology in our customers’ workflows.