The Challenge of Die Filling in Rotary Presses-A Systematic Study of Material Properties and Process Parameters

pharmaceutics
Article
The Challenge of Die Filling in Rotary Presses—A
Systematic Study of Material Properties and
Process Parameters
Ann Kathrin Schomberg 1,2, * , Arno Kwade 1,2 and Jan Henrik Finke 1,2
1 Institute for Particle Technology, Technische Universität Braunschweig, Volkmaroder Str.5,
38104 Braunschweig, Germany; a.kwade@tu-braunschweig.de (A.K.); j.finke@tu-braunschweig.de (J.H.F.)
2 Center of Pharmaceutical Engineering (PVZ), Technische Universität Braunschweig,
Franz-Liszt-Str. 35A, 38106 Braunschweig, Germany
* Correspondence: a.schomberg@tu-braunschweig.de

Received: 25 February 2020; Accepted: 5 March 2020; Published: 10 March 2020
Abstract: For the efficient and safe production of pharmaceutical tablets, a deep process understanding
is of high importance. An essential process step during tableting is the die filling, as it is responsible
for a consistent tablet weight and drug content. Furthermore, it affects the results of subsequent
process steps, compaction and ejection, and thus critical quality attributes. This study focuses on
understanding the influences of process parameters and material properties on die filling on a rotary
tablet press. By the systematic variation in process parameters as the turret and paddle speeds as
well as the fill and dosing depths, five formulations with differing properties are processed. Analysis
of the normalized tablet weight, called filling yield, revealed different limitation mechanisms of the
filling process, i.e., incomplete filled dies for certain parameter settings. Kinetic limitations occur
due to a short residence time under the feed frame (filling time) caused by high turret speeds, which
additionally induce high tablet weight variation coefficients. Characteristic maximum turret speeds at
certain paddle speeds can be found to still achieve complete filling. At low turret speeds, densification
of the powder inside the dies takes place, induced by two mechanisms: either high paddle speeds or
high overfill ratios, or a combination of both. The challenge to fill the dies completely as well as avoid
densification is dependent on material properties as the flowability. The mass discharge rate from an
orifice was found to be in a linear correlation to the filling results of different formulations below
complete filling.
Keywords: tableting; die filling; rotary tablet press; process understanding; paddle feeder; turret
speed; paddle speed; filling efficiency; kinetic filling limitations; densification
1. Introduction
In the pharmaceutical industry, deeper process understanding is gaining importance. This is
emphasized by the Quality by Design (QbD) approach promoted by health authorities [1]. In the
context of continuous manufacturing, which is an aspirational aim, process understanding, among
others enabling a reliable process control, has even a greater priority. The present study focuses on
the systematic understanding of the die filling process directly in rotary tablet presses. The primary
objective of this work is to qualitatively describe the influence of the process parameters and material
properties on the die filling of selected model materials. Derived knowledge can be used to enable the
development of process models to forecast quantitative impacts, and thus shorten the development
time of novel products and increase process stability.
The tableting process can be divided into four steps: powder feeding, die filling, compaction
and ejection. The product quality fundamentally depends on die filling. It critically affects the tablet
Pharmaceutics 2020, 12, 248; doi:10.3390/pharmaceutics12030248 www.mdpi.com/journal/pharmaceutics

Pharmaceutics 2020, 12, 248 2 of 23
quality, as weight variations contribute to drug content variations. Furthermore, subsequent process
steps such as compaction and ejection depend on the result of filling. If die filling is not accurate, the
effective pressure acting on the powder bed, applied by the opposing punch movement, undercuts the
processing limits. Thus, quality attributes like tensile strength, porosity and dissolution behavior can
be influenced. To meet the required quality, it is necessary to ensure a complete, reproducible and
consistent die filling.
Different mechanisms have been identified for die filling, which involve gravity filling, suction
filling and forced feeding [2]. Gravity filling occurs for powder flowing into a die cavity under gravity [3].
Suction filling takes place due to an arising pressure gradient from the downward movement of the lower
punch under the powder bed inside the feed frame, pulling the powder into the die cavity. Furthermore,
improvement in the filling process is achieved by feeding systems with moved components, referred
to as forced feeding.
Industrial rotary tablet presses commonly use a forced feeding system consisting of a feed frame
equipped with one to three stirrers, called paddle feeder. In the most common two-wheel setup, the first
wheel, called the feeding wheel, delivers powder into the die while the second wheel, the metering
wheel, transfers the exceeded powder back into the feed frame. In some constructions, a third wheel
is positioned in the center above the feeding and metering wheel to supply powder equally to both
wheels [4].
The dosing depth results from the fill depth, which is defined by the fill cam, minus the overfilling
height, which is used to support complete die filling [5]. Thus, after previous overfilling, the lower
punch moves upwards to eject some powder for weight control. The mechanisms, including forced
feeding, suction filling and overfilling as well as disruptive factors like centrifugal forces, vibrations and
vertical stress caused by the above powder, acting simultaneously in a rotary tablet press, result in a
complex process [6]. Variable parameters, including turret speed, paddle speed, fill depth and dosing
depth, help to control the process. In addition, powder properties clearly affect the filling quality
and efficiency.
Several studies were carried out to understand the filling process using a linear filling
device [2,3,7–12]. Wu et al. examined the flow of metallurgical powders in stationary simple
and stepped dies from a moving feeding shoe in air and vacuum [3]. It was found that powder
properties, shoe speed and airflow affect the filling process. Sinka et al. also used a linear filling device
with gravity fill in air and vacuum to investigate the flow properties of pharmaceutical powders to try
to resemble die filling on a rotary press [12]. The critical velocity, which still allows completely filled
dies, was used for classifying the flow properties. The strong improvement in the filling result due
to suction fill, which can take place in a rotary press, was detected by Jackson et al. [2]. Besides the
experimental investigations, CFD/DEM-Simulations focused on the die filling process and analyzed air
effects [13–24]. A rotational filling device with gravity fill equipped with moved dies and a stationary
feeder, mimicking a rotary press more realistically, was used to correlate die filling performance and
critical material attributes [25]. Furthermore, rotational systems with forced feeding were selected to
investigate the powder residence time and the impact of feed frame speed and die speed on the die
filling [26,27]. With increasing feed frame speed, the separated powder mass was increasing, while
an increase in die speed led to a decrease in powder mass. In addition, the flowability of lubricated
powders was improved due to overlubrication caused by increasing feed frame speed. In addition to
the approaches with filling devices, computer modeling was used to investigate the flow behavior of
powders in feed frames and into dies [19–22].
Although there were several research studies describing the filling efficiency, these simplified
systems are able to investigate effects of selected parameters, but often important parameters were not
taken into account, which is why the general transfer of findings is limited [5]. There are only few
studies carried out on rotary tablet presses. The impact of flow behavior of powders on die filling
properties was examined by Yaginuma and a PLS model for thirty blends, predicting tablet weight
variability, was developed by van Snick, but the application of their findings on other materials and
tablet presses is limited [28,29]. Other studies focus on specific issues like feed frame design and
filling level inside a feed frame by laser triangulation and in-line NIR spectroscopy [4,30–33]. General
relations between powder properties, operating conditions and tablet weight and its variability are rare.
In contrast, Peeters et al. investigated the influence of turret speed, paddle speed, fill depth and
material properties with a Design of Experiments (DoE) by using microcrystalline cellulose (MCC) and
dibasic calcium phosphate dihydrate (DCP) directly on the rotary tablet press MODULE™ P (GEA) [5].
The interaction between the process parameters and both the tablet weight and weight variability were
examined, identifying the turret speed and the paddle speed of the feeding wheel as strong impact
factors. In addition, the weight variability of MCC is affected by the paddle speed of the metering
wheel. Contrasting the powders, the impact of the paddle speeds on tablet weight is more pronounced
for MCC due to its fairly flowing behavior. On basis of the DoE, the optimum combinations of process
parameters are presented to meet requirements for minimum tablet weight variability. However, these
determined optimum parameters are singular solutions for the investigated powders in combination
with the used rotary press. Accordingly, further advancement would be provided by developing
general and transferable deductions from such datasets instead.
Due to the lack of deep understanding, the aim of the current work was to systematically
investigate the influence of particular process parameters and powder properties to achieve an
improved understanding of the mechanistic relations regarding tablet weight and weight variation.
This is performed directly on a rotary press to avoid differences in operating modes that may finally
hinder the transferability of models from testing rigs to industrially relevant machines and processes.
The findings should also be transferable to other powders and formulations.
2. Materials and Methods
2.1. Materials
Microcrystalline cellulose (MCC, Vivapur® 102, JRS Pharma, Rosenberg, Germany) and anhydrous
dicalcium phosphate (DCPA, Emcompress® Anhydrous, JRS Pharma, Rosenberg, Germany) were
used as model materials. Magnesium stearate (MgSt, Magnesia GmbH, Lüneburg, Germany) was
used as a lubricant. The characteristic particle sizes of MCC and DCPA are in a comparable range with
slightly lower particle sizes for MCC (Table 1). The solid, bulk and tapped density of DCPA is higher,
almost double, than MCC.
Table 1. Characteristic particle sizes, solid, bulk and tapped density for microcrystalline cellulose
(MCC) and anhydrous dicalcium phosphate (DCPA).
Material x10 [µm] x50 [µm] x90 [µm] ρs [g/cm3 ] ρb [g/cm3 ] ρt [g/cm3 ]
MCC 28 110 243 1.56 0.327 0.449
25 wt % DCPA 29 117 258 0.387 0.514
50 wt % DCPA 31 136 284 0.483 0.596
75 wt % DCPA 31 148 298 0.581 0.698
DCPA 38 178 322 2.83 0.703 0.822
2.2. Methods
2.2.1. Powder Characterization

Particle size distribution was determined by laser light diffraction (Mastersizer 3000, Malvern
Panalytical, Kassel, Germany). Three measurements were performed, and the mean values were
calculated. The solid densities of the powders were measured in triplicate with the helium pycnometer
Ultrapyc 1200e (Quantachrome Instruments, Boynton Beach, FL, USA). The bulk and tapped density
were determined according to the Ph. Eur. 9.3 2.9.34 using a 100 mL cylinder, tapped by a volumetric
analyzer (Erich Tschacher Laboratoriumsbedarf, Bielefeld, Germany). Measurements for each sample
were carried out in triplicate, calculating the average. Based on the bulk and tapped density, the
Hausner ratio H and Compressibility index C were calculated as follows
ρt
H= (1)
ρb
ρt − ρb
C= × 100 (2)
ρt
Flow properties were determined using a granulate flow tester (ERWEKA GmbH, Langen, Germany)
by measuring the time required for 100 g of the powder sample to flow through an orifice of 25 mm
according to the Ph. Eur. 9.3 2.9.16. The orifice has an opening angle of 40◦ and is made of stainless
steel. From these results, the powders’ mass flow rate was calculated. Furthermore, the ring shear
tester RST-XS (Dr. Ing. Dietmar Schulze Schüttgutmesstechnik, Wolfenbüttel, Germany) was used
to determine flow properties. Therefore, preshear stresses of 1, 2, 4 and 8 kPa were used with
corresponding shear to failure stresses of 20, 50 and 80% (33% for 1 kPa). On basis of the consolidation
stress σ1 and the unconfined yield strength σc , the ffc is calculated as follows [34]
σ1
f fc = (3)
σc
2.2.2. Powder Blending

The three blends of MCC and DCPA were prepared using a cube mixer (ERWEKA GmbH, Langen,
Germany). The volume of 3.5 L was filled up to 40%. Initially, DCPA was mixed with 1 wt % magnesium
stearate (MgSt) for five minutes at 30 rpm, then MCC was added and mixed for another ten minutes.
The composition of the three blends processed on the XL 100 are displayed in Table 2.
Table 2. Composition of the blends of MCC and DCPA processed on the XL 100.
Blend wt % MCC wt % DCPA + 1% MgSt

25 wt % DCPA 75 25
50 wt % DCPA 50 50
75 wt % DCPA 25 75
2.2.3. Die Filling/Tableting Experiments

Die filling/tableting experiments were carried out with the rotary tablet press XL 100 (KORSCH
AG, Berlin, Germany) in lab scale. Therefore, tablets were produced at different process parameters,
determining their weight afterwards as a measure for the filling weight (n = 10) (Quintix® 224 – 1CEU,
Sartorius, Göttingen, Germany). For the investigation, a paddle feeder with one wheel—the feeding
wheel—was used. The press was equipped with four flat-faced round Euro-D punches with a diameter
of 9 mm. No external vacuum cleaner for aspiration was used. In order to investigate the influence of
different fill depths, three fill cams with fill depth of 10, 14 and 16 mm were used; the corresponding
dosing depth and operating speeds can be found in Table 3. For each fill depth a minimum and a
maximum dosing depth, resulting from a machine-specific maximum overfill of 5.5 mm as well as
a minimum overfill of 0.5 mm, were chosen. To enable the comparison of the different fill depths at
the same dosing depth, intermediate dosing depths corresponding to the minimum dosing depths of
the neighboring larger fill cam were applied. Low dosing depths were selected to evaluate possible
densification effects, while high dosing depths were chosen to investigate whether complete filling
could be achieved. For all experiments, three turret speeds as well as four paddle speeds were used,
including the machine-specific minimum and maximum as well as intermediates. In order to gain
more detailed information at the highest dosing depth, representing the highest challenge of filling,
turret speeds of 30 and 50 rpm are additionally investigated.
Pharmaceutics 2020,
Pharmaceutics 12, x248
2020, 12, FOR PEER REVIEW 55 of
of 23
23
Table 3. Process parameters for tablet production with XL 100.

Table 3. Process parameters for tablet production with XL 100.
Fill Depth (mm) Dosing Depth (mm) Turret Speed (rpm) Paddle Speed (rpm)
Fill Depth (mm) Dosing Depth (mm) Turret Speed (rpm) Paddle Speed (rpm)
4.5
4.5
8.5 20, 40, 60 5, 20, 40, 60
10
10 8.5 20, 40, 60 5, 20, 40, 60
9.5
9.5
8.5
8.5
14 10.5
10.5 20,
20,40,
40,60
60 5, 5,
20,20,40,
40,60
60
14
13.5
13.5
10.5
10.5 20,
20,40,
40,60
60 5, 5,
20,20,40,
40,60
60
16
16 13.5
13.5
15.5
15.5 20,30,
20, 30,40,
40,50,
50, 60 5, 5,
20,20,40,
40,60
60
3.
3. Theoretical
TheoreticalBackground
Background
3.1. Definition
3.1. Definition of
of Terms
Terms
In the
In thefollowing,
following,important
importanttermsterms are
are named,
named, andand equations,
equations, used used for evaluation
for the the evaluation below,
below, are
are defined.
defined. Under Under
the feedtheframe,
feed frame, the punch
the lower lower is punch
moved is moved
down bydownthe fillbycam
the to
filla cam to position,
certain a certain
position,
called depth fill
fillcalled hF depth hF (Figure
(Figure 1). To 1). To change
change the fill fill depth,
thedepth, another
another fill fill
cam,cam,a astationary
stationarymechanical
mechanical
component, must
component, must be be installed.
installed. AsAsdescribed
describedabove,
above,the thediedie cavity
cavity is
is overfilled,
overfilled, thus,
thus, for
for dosing,
dosing, the
the
lower punch
lower punch isislifted.
lifted.The
Theheight,
height,bybywhich the the
which punch is moved
punch upwards
is moved until scraping
upwards off the powder
until scraping off the
bed at the
powder bedendat of
thethe
endfeed
of frame,
the feedis frame,
named isoverfill
named ∆h.overfill
Therefore, the so-called
Δh. Therefore, thedosing depth
so-called hD is depth
dosing reached
hD
and
is the filling
reached and process
the fillingis completed by scrapingbypowder
process is completed scraping above the die
powder above(Figure 1). (Figure
the die The remaining
1). The
powder inside
remaining powderthe die is compacted
inside afterwards.afterwards.
the die is compacted
Figure 1. Schematic overview of the punch positions used for filling with perspective from the side –
Figure 1. Schematic overview of the punch positions used for filling with perspective from the side –
fill depth hF , overfill ∆h and the resulting dosing depth hD .
fill depth hF, overfill Δh and the resulting dosing depth hD.
Different dosing depths can be achieved by stepless variable height adjustment. Correspondingly,
Different
overfill can varydosing depths
in the can
range ofbe
0.5achieved
to 5.5 mmbyfor
stepless
each fill variable height
depth set. Theadjustment. Correspondingly,
kinematic volume reduction
overfill
duringcan thisvary in theisrange
process of 0.5by
described to 5.5
the mm for ratio
overfill each rfill, depth set.as
defined The kinematic
the ratio of volumetoreduction
overfill fill depth
o
during this
(Equation (4)).process is described by the overfill ratio r o, defined as the ratio of overfill to fill depth
(Equation 4). ∆h
ro = (4)
h𝛥ℎF
𝑟 = (4)
Regarding different dosing depths as well as powders ℎ with different bulk and tapped densities,
normalization is necessary to allow the comparison of the resulted tablet weight as the dimensionless
Regarding different dosing depths as well as powders with different bulk and tapped densities,
result of filling. Therefore, the filling yield ϕ is defined as the ratio of the tablet weight mT to the
normalization is necessary to allow the comparison of the resulted tablet weight as the dimensionless
theoretical maximum weight resulting from the dosing volume inside the die, calculated with the
result of filling. Therefore, the filling yield φ is defined as the ratio of the tablet weight mT to the
die radius rdie and the dosing depth hD , as well as the tapped density ρt of the respective powder
theoretical maximum weight resulting from the dosing volume inside the die, calculated with the die
(Equation (5)).
radius rdie and the dosing depth hD, as well as the tapped mT density ρt of the respective powder (Equation
5). ϕ = (5)
π·r2die ·hD ·ρt
𝑚
𝜑= (5)
𝜋∙𝑟 ∙ℎ ∙𝜌
Pharmaceutics 2020, 12, x FOR PEER REVIEW 6 of 23

The tapped density is used due to the assumption that maximum powder consolidation in the feed
frameTheandtapped
duringdensity
die filling is comparable
is used due to the to assumption
the impact ofthat the maximum
taps, whichpowder
leads toconsolidation
the greatest change
in the
in particle
feed framepacking
and duringin normal direction
die filling due to thetopowder
is comparable the impactweight. For
of the thiswhich
taps, reason, it is to
leads hypothesized
the greatest
that the tapped
change density
in particle can beinused
packing as a limit
normal for consolidation
direction due to the inside
powder theweight.
die without compaction,
For this reason, and
it is
accordingly the maximum filling yield amounts to 1.
hypothesized that the tapped density can be used as a limit for consolidation inside the die without
Corresponding
compaction, to the consolidated
and accordingly the maximum state, a yield
filling reference value
amounts to for
1. complete die filling in the
unconsolidated state, related to the bulk density, can be defined.
Corresponding to the consolidated state, a reference value for complete Therefore, the fillingdie
yield represents
filling in the
complete die filling
unconsolidated forrelated
state, the ratio toof
thethe weight
bulk resulting
density, can be from the dosing
defined. volume
Therefore, thefilled
fillingwith
yieldbulk density
represents
ρcomplete
b to the weight resulting
die filling for the from
ratiotheofdosing volume
the weight filled with
resulting fromtapped density,
the dosing unconsolidated
calledfilled
volume with bulk
complete die filling UCF (Equation (6)). The equation can be simplified to
density ρb to the weight resulting from the dosing volume filled with tapped density, called the ratio of bulk density to
tapped densitycomplete
unconsolidated which equals theUCF
die filling inverse of the Hausner
(Equation ratio. can be simplified to the ratio of bulk
6). The equation
density to tapped density which equals the inverse of the Hausner ratio.
π·r2Mat ·hD ·ρb ρb
UCF = 𝜋 ∙ 2𝑟 ∙ ℎ ∙= 𝜌 𝜌< 1 (6)
𝑈𝐶𝐹 = π·rMat ·hD ·ρt ρ=t <1 (6)
𝜋∙𝑟 ∙ℎ ∙𝜌 𝜌
The
The variation
variation in
in the
the filling
filling yield
yield isis described
described by the tablet
by the tablet weight
weight variation
variation coefficient
coefficientWVC
WVC which
which
equals
equals the relative standard deviation (Equation 7). It is calculated as the percentage of the standard
the relative standard deviation (Equation (7)). It is calculated as the percentage of the standard
deviation
deviation of
of the
the mean
mean value
value of of the
the filling
fillingyield.
yield.
11 Pn
q
ϕ)𝜑2
∑ (ϕ𝜑−−
𝑛n−1
− 1 i=1 i (7)
𝑊𝑉𝐶==
WVC ×× 100
100 (7)
ϕ𝜑
3.2. Detailed Process Description

3.2. Detailed Process Description
For a better understanding and explanations of the following experimental data, the filling process
For a better understanding and explanations of the following experimental data, the filling
is described in detail. Die filling can be divided into four process steps. The area in which the respective
process is described in detail. Die filling can be divided into four process steps. The area in which the
process step takes place is marked with the abbreviations in a schematic overview with plan view
respective process step takes place is marked with the abbreviations in a schematic overview with
(Figure 2):
plan view (Figure 2):
PS1—powder feeding from the hopper into the feed frame;
PS1 – powder
PS2—powder feeding
transport from
by the the hopper
feeding wheelinto
(intothe
thefeed frame;
filling area);
PS2 – powder
PS3—die filling; transport by the feeding wheel (into the filling area);
PS3 – die filling;
PS4—dosing out: reduction of die fill volume (by ∆h).
PS4 – dosing out: reduction of die fill volume (by 𝛥ℎ).
Figure
Figure 2. Schematicoverview
2. Schematic overviewofofthe
thefour
four process
process steps
steps forfor
diedie filling
filling in ainone-paddle
a one-paddle
feedfeed frame.
frame. The
The paddle rotates clockwise, the die table rotates counter-clockwise.
paddle rotates clockwise, the die table rotates counter-clockwise.
During
During PS1,
PS1, the powder flows
the powder flows from
from the
the feeding
feedingsystem,
system,aahopper,
hopper,into
intothe
thefeed
feedframe.
frame.Here,
Here,it
itenters
entersthe
theintersections
intersectionsbetween
betweenthe thepaddles
paddlesofofthe thepaddle
paddle wheel.
wheel. The
The mass
mass flow
flow from
from the hopper
the hopper
determines
determines the the filling
fillinglevel
levelofofthe
theinterspaces
interspaces in in combination
combination with
with thethe removal
removal massmass
flowflow bydies.
by the the
dies. Methods to measure the filling level are developed but not used under
Methods to measure the filling level are developed but not used under continuous production continuous production
conditions
conditions yetyet [30–32].
[30–32]. Hypotheses
Hypotheses about about the
the filling
filling level
level and
and the
the linked
linked flow
flow mechanisms,
mechanisms, which
which are
are
proposed to explain the powder flow, should be evaluated with these measurements in the future.
For poor flowing powders, the intersections might not be filled with as much powder as for free-
flowing powders. This could be a limiting factor for successful filling at a high turret speed. However,
a consolidation would lead to worse flow behavior for subsequent die filling.
Following the powder transport, the die is filled in PS3. The residence time under the feed frame
is one influencing parameter, which is set by the turret speed and the angle of the filling area, marked
as α (Figure 2). With increasing turret speed, the period of time shortens, which affects the results of
die filling (Figure 4). In parallel, a higher turret speed creates a higher pressure drop for suction filling
due to the increased vertical punch velocity (Figure 4). For experimental investigations using the
rotary tablet press XL 100, a feed frame with an installed external lubrication of the die was used,
which
proposed implies that thethe
to explain lower punch
powder is moved
flow, shoulddownwards
be evaluated before
withentering the feed frame.
these measurements in Therefore,
the future.
the
Forsuction fill effect
poor flowing mightthe
powders, be intersections
small or even negligible.
might The combination
not be filled with as muchof the process
powder parameters
as for free-flowing
turret
powders. and paddle
This could speedbe has a strong
a limiting influence
factor on the diefilling
for successful filling.
at Regarding
a high turret thespeed.
rotational speedsthe
However, of
the paddleof
diameter andthethe
pipeturret, two states
entering the feedfor die
frame filling can besodistinguished—either
is large, a sufficient powder volume the paddles flowovertake
is likely
the diesthe
unless in the fillingofarea,
bridging highlyrepresented
cohesive by a positive
powders relative velocity (Figure 3a), or the dies pass the
occurs.
paddles Theiffollowing
the relative velocity
process step,is PS2,
negative (Figure
describes the3b). On basis
powder of the hypothesis
movement induced bythat thethe powder
paddle wheel. is
accumulated
As Dühlmeyer and probably
et al. show in densified
their work, in front of the pushing
the powder paddles, moredistributed
is not homogenously powder would over theenter the
radius
die if the paddles push the powder over the die (Figure 3a). Wu et al.
as well as over the width of an interspaces [30–32]. With increasing paddle speed, the powder is moved investigated the powder flow
from
radiallya linear
towardsmoved the fill shoe
outer wallintoofstatic dies and
the filling chamberfoundand a landslide-like
the filling area flow of the2,front
(Figure lightregion for
blue) due
low relative velocities,
to centrifugal forces asnamed
a function noseofflow, which supports
the paddle speed squared.complete Thisdie fillingthat
means [3]. the
In the feed frame,
powder, which
nose flow can
is initially on an develop if the interspaces
inner radius, might enterare thenot filledarea
filling completely with supply
and therefore powdernew and powder
if the relative
which
speed
can bebetween
filled into thethe
powder
dies. and the die is
Dühlmeyer et low enough that,
al. present and positive.
for a fixed Otherwise,
powder mass especially
insideifthethefeed
die
overtakes
frame, thethe paddles
powder with powder
accumulates in frontin front, bulk andpaddles.
of the pushing intermittent flow for
Therefore, the poor
powder flowing powdersa
bed develops
may
bulkoccur
geometry[3]. Ifwith
a suction pressure
a triangular develops,
vertical cross which
area alongis unlikely in the used (Figure
the circumference experimental setup, as
3). Mateo-Ortiz
described, it has a significant
et al. also observed a wave-like influence
powderonflow the inside
powder theflow.
feedBesides
frame due the todifferent flowmovement
the paddle mechanisms, [4].
the absolute value of the relative velocity between the paddle, moving
This may influence the powder flow into the die. A change in bulk density due to possible densificationthe powder, and the die might
be
of athe
crucial
powder factor.
has A notlow
yetrelative velocity correlates
been investigated. with a longer
For a compressible periodsuch
powder, of time for the particles
a consolidation would to
flow
lead into the die
to worse flowwithout
behavior the for
shift of a paddle
subsequent dieover the die.
filling.
Figure 3. Schematic representation of the two possible states during die filling (PS3): (a) paddles overtake
Figure
dies, (b)3.dies
Schematic representation
overtake paddles in theoffilling
the two
area.possible states
The green during
arrows die filling
represent (PS3): (a)
the relative paddles
velocity.
overtake dies, (b) dies overtake paddles in the filling area. The green arrows represent the relative
velocity.
Following the powder transport, the die is filled in PS3. The residence time under the feed frame
is one influencing parameter, which is set by the turret speed and the angle of the filling area, marked
as α (Figure 2). With increasing turret speed, the period of time shortens, which affects the results
of die filling (Figure 4). In parallel, a higher turret speed creates a higher pressure drop for suction
filling due to the increased vertical punch velocity (Figure 4). For experimental investigations using
the rotary tablet press XL 100, a feed frame with an installed external lubrication of the die was used,
which implies that the lower punch is moved downwards before entering the feed frame. Therefore,
the suction fill effect might be small or even negligible. The combination of the process parameters
turret and paddle speed has a strong influence on the die filling. Regarding the rotational speeds of
the paddle and the turret, two states for die filling can be distinguished—either the paddles overtake
the dies in the filling area, represented by a positive relative velocity (Figure 3a), or the dies pass the
paddles if the relative velocity is negative (Figure 3b). On basis of the hypothesis that the powder is
accumulated and probably densified in front of the pushing paddles, more powder would enter the
die if the paddles push the powder over the die (Figure 3a). Wu et al. investigated the powder flow
from a linear moved fill shoe into static dies and found a landslide-like flow of the front region for low
relative velocities, named nose flow, which supports complete die filling [3]. In the feed frame, nose
flow can develop if the interspaces are not filled completely with powder and if the relative speed
between the powder and the die is low enough and positive. Otherwise, especially if the die overtakes
the paddles with powder in front, bulk and intermittent flow for poor flowing powders may occur [3].
If a suction pressure develops, which is unlikely in the used experimental setup, as described, it has a
significant influence on the powder flow. Besides the different flow mechanisms, the absolute value of
the relative velocity between the paddle, moving the powder, and the die might be a crucial factor.
resistance due to overhead powder. For high fill levels inside the feed frame, the resistance could lead
to densification of the powder inside the die, resulting in densities above the bulk density ρb.
Therefore, filling yields between 1 and the UCF can be reached. Possibly, for high fill levels, the
powder is already densified inside the feed frame, and thus during die filling. However, it is currently
not possible
Pharmaceutics to 12,
2020, quantify
248 the mass inside the die and its apparent bulk density before ejection of 8 ofthe
23
tablet. Accordingly, insights into the initial state of the powder inside the die can only indirectly be
deduced from systematic studies.
A lowTherelative velocity
different correlates
impacts with aweight
on the tablet longerasperiod of time
the result forfilling,
of die the particles to flow
which are into the
described die
in this
without the shift of a paddle over the die.
section, are displayed in the summarized representation (Figure 4).
Figure 4. Summarized influences of process parameters and material properties on the tablet weight
Figure die
during 4. Summarized
filling. influences of process parameters and material properties on the tablet weight
during die filling.
After die filling, the filled die volume is reduced in PS4 by a partial upwards movement of the
4. Results
lower punchandinDiscussion
the dosing out phase. The acceleration of the powder mass inside the die results in
particle rearrangement, and thus, in the reduction of porosity and, possibly, the ejection of excessive
4.1. Powder
powder Characteristics—Flow
(Figure 4). In dependence Behavior
on the fill level of the feed frame, ejecting powder experiences a
resistance due to overhead powder.
Since the flow property of a powder For high
is afill levelsfactor
crucial insideinthethefeed frame,
context of the
die resistance could
filling, different
lead to densification
techniques of the powder
for its determination wereinside
used. the die,
In the resulting
granular in tester,
flow densitiesthe above the bulk
mass flow density ρb .
was determined
Therefore, filling yields between 1 and the UCF can be reached. Possibly, for high fill
for all powders used in this study (Figure 5a). The progressive rise in the mass flow shows a constant levels, the powder
is already densified inside the feed frame, and thus during die filling. However,
improvement in the flow property with increasing DCPA mass fraction. Thus, this method indicates it is currently not
possible
that DCPA to quantify
mass flow theismass
moreinside
than the
fourdie and greater
times its apparent
than bulk
that density
of MCC.before ejection
The flow of the tablet.
functions of the
Accordingly, insights into the initial state of the powder inside the die can only
powders used in the study were calculated using a ring shear tester (Figure 5b). Great differences indirectly be deduced
can
from systematic
be found for thestudies.
pure substances DCPA and MCC. DCPA shows very high values of ffc in the range
The
of >20, sodifferent impactsare
flow properties on excellent.
the tablet Contrasting
weight as thethis,result
theof
ffc die
for filling,
MCC iswhich are described
comparably in this
low, between
section, are displayed in the summarized representation (Figure 4).
4 and 6, which indicates poor flow properties, still classified as easy-flowing [34]. The ffc for blends of
the two powders are systematically located between the pure substances. The order corresponds to
4. Results and Discussion
the mass fractions of DCPA and MCC. Therefore, the blends represent the gradation in the flow
properties
4.1. Powderdetermined by both tests
Characteristics—Flow applied.
Behavior
The Hausner ratio and the Compressibility index of the powders allow for an estimation of the
Since compressibility
materials’ the flow property of a powder
(Figure 5c). MCCishasa crucial
a highfactor in the context
compressibility, of is
which die filling,double
almost different
the
techniques for its determination were used. In the granular flow tester, the mass flow
value of DCPA. The blends show a compressibility between the pure substances, although it does not was determined
for all powders
decline linearly.used in this
Blends studya (Figure
below 5a). Theofprogressive
mass content rise in the
75 wt % DCPA mass aflow
possess shows a constant
considerably higher
improvement in the flow property with increasing DCPA mass
compressibility and Hausner ratio compared to powders with higher content.fraction. Thus, this method indicates
that DCPA mass flow is more than four times greater than that of MCC. The flow functions of the
powders used in the study were calculated using a ring shear tester (Figure 5b). Great differences can
be found for the pure substances DCPA and MCC. DCPA shows very high values of ffc in the range of
>20, so flow properties are excellent. Contrasting this, the ffc for MCC is comparably low, between 4
and 6, which indicates poor flow properties, still classified as easy-flowing [34]. The ffc for blends of the
two powders are systematically located between the pure substances. The order corresponds to the
mass fractions of DCPA and MCC. Therefore, the blends represent the gradation in the flow properties
determined by both tests applied.
Pharmaceutics
Pharmaceutics 2020,
2020, 12,12,
248x FOR PEER REVIEW 9 of923
of 23
(a) (b)
(c)
Figure
Figure 5. 5.Flow
Flowproperties
propertiesofofthe
the powders
powders used
used in in this
thisstudy;
study;(a)
(a)Mass
Massflow
flowofof
powders,
powders, measured
measured by by
granularflow
granular flowtester;
tester;(b)
(b)Flow
Flowfunction
function (ff
(ffcc)) measured
measured with
withring
ringshear
sheartester;
tester;(c)(c)Hausner
Hausnerratio and
ratio and
Compressibility index for the five formulations
Compressibility index for the five formulations used. used.
4.2.The
Impact of Process
Hausner Parameters
ratio on Filling Yield index of the powders allow for an estimation of the
and the Compressibility
materials’ The compressibility
impact of different (Figure
process5c). MCC hassuch
parameters a high compressibility,
as the turret speed, the which
paddleis speed,
almostthedouble
fillingthe
value of DCPA. The blends show a compressibility between the pure substances,
and dosing depth as well as the overfill are investigated in the following. A limited set of experiments although it does
notis decline
highlighted linearly. Blends
to point below
out the a mass
impact content
of each of 75 wt
parameter % DCPA
variation. Duepossess a considerably
to the greatest challengehigher
of
compressibility
filling, the highest and fill
Hausner
depth ofratio compared
16 mm and thetohighest
powders withdepth
dosing higherof content.
15.5 mm were chosen in most
cases to display the strongest and most fundamental effects.
4.2. Impact of Process
The turret and Parameters
paddle speed, on Filling Yield
as the dynamic parameters of the system, simultaneously influence
theThedie impact
filling byof different process parameters such is
their interaction. The turret speed asdirectly
the turretlinked
speed,to the
the paddle
tablet production
speed, therate
filling
and dosing depth as well as the overfill are investigated in the following. A limited set of maximum
(Figure 4), which is desired to be as high as possible due to economic reasons. The experiments
applicable speed is technically determined by the rotary press on one hand and by quality
is d to point out the impact of each parameter variation. Due to the greatest challenge of filling, the
requirements for the tablets on the other hand. The maximum tablet production rate which can safely
highest fill depth of 16 mm and the highest dosing depth of 15.5 mm were chosen in most cases to
meet the requirements depends on the formulations properties and the combination of the chosen
display the strongest and most fundamental effects.
values of the static process parameters (hF, hD, ∆h) and the paddle speed. Thus, in the following,
The turret and paddle speed, as the dynamic parameters of the system, simultaneously influence
general understanding of the impact and interplay of each process parameter is investigated based
theondie filling by their interaction. The turret speed is directly linked to the tablet production rate
the most extreme formulations of MCC and DCPA. As the blends show results between the pure
(Figure 4),
substances, which
theyisare
desired to be as high
not displayed as possible
separately due
in this to economic
section. reasons.
The influence of The maximum
the material applicable
properties
speed
of the is blends
technically determined
is investigated andbydiscussed
the rotaryinpress on4.3.
Section oneThere,
hand theandfilling
by quality requirements
yield of for the
all formulations
tablets on the other hand. The maximum tablet production
used is displayed, using dependence on the paddle speed as an example. rate which can safely meet the requirements
depends on the formulations properties and the combination of the chosen values of the static process
parameters (hF , hD , ∆h) and the paddle speed. Thus, in the following, general understanding of the
impact and interplay of each process parameter is investigated based on the most extreme formulations
of MCC and DCPA. As the blends show results between the pure substances, they are not displayed
separately in this section. The influence of the material properties of the blends is investigated and
discussed in Section 4.3. There, the filling yield of all formulations used is displayed, using dependence
on the paddle speed as an example.
4.2.1. Impact of Turret Speed on the Filling Yield
4.2.1. Impact of Turret Speed on the Filling Yield
The impact of the turret speed on the filling yield is analyzed at five rotational velocities for four
Thepaddle
different impactspeeds.
of the turret speed on
In addition tothe
thefilling
data, yield is analyzed
the graphs showatthefive rotational velocities
unconsolidated fordie
complete four
filling
different paddle speeds. In addition to the data, the graphs show the unconsolidated
(UCF) in order to compare the actual filling yield with the complete filling at an unconsolidated state. complete die filling
(UCF)
For MCC, in one
order to observe
can comparethat the actual
only atfilling yield with
low turret speedstheof
complete
20 and 30 filling
rpm,atand
an unconsolidated
in parallel highstate.
paddle
For MCC, one can observe that only at low turret speeds of 20 and 30 rpm, and in parallel high paddle
speeds of 40 and 60 rpm, UCF is achieved (Figure 6a). With increasing turret velocity, the filling yield
speeds of 40 and 60 rpm, UCF is achieved (Figure 6a). With increasing turret velocity, the filling yield
of MCC shows a continuous decrease in all paddle speeds. DCPA achieves complete die filling at every
of MCC shows a continuous decrease in all paddle speeds. DCPA achieves complete die filling at
turret speed if paddle speeds of 40 and 60 rpm are applied, while the filling yield decreases as well for
every turret speed if paddle speeds of 40 and 60 rpm are applied, while the filling yield decreases as
20 and 5 rpm paddle speed (Figure 6b). The considerably lower filling yields for MCC in comparison
well for 20 and 5 rpm paddle speed (Figure 6b). The considerably lower filling yields for MCC in
tocomparison
DCPA at thetosame turret und paddle speed can be related to the different flow properties. The initial
DCPA at the same turret und paddle speed can be related to the different flow
powder
properties. The initialdie
flow into the for MCC
powder flowmight
into thebedieconsiderably
for MCC might lowerbecompared
considerably to DCPA and mighttonot
lower compared
develop
DCPA and properly
mightin the
not short period
develop properlyof in
time. The hypothesis
the short is supported
period of time. by theislower
The hypothesis massby
supported flow
ofthe
MCC, investigated with the granulate flow tester (Figure 5a). In addition,
lower mass flow of MCC, investigated with the granulate flow tester (Figure 5a). In addition, regarding the previous
process stepsthe
regarding PS1 and PS2,
previous less powder
process steps PS1might
and be PS2,transported
less powder bymight
the paddles and by by
be transported centrifugal
the paddlesforces
forand
MCC into the filling
by centrifugal forcesarea
foras compared
MCC into theto DCPA
filling at as
area a given paddle
compared speed.atNonetheless,
to DCPA a given paddlethespeed.
fill level
inNonetheless,
the filling area
thewas not directly
fill level measureable.
in the filling area was not directly measureable.
(a) (b)
Figure6.6.Impact
Figure Impactof ofthe
the turret
turret speed on
on the
thefilling
fillingyield
yieldfor
for(a)(a)MCC
MCC and
and (b)(b) DCPA
DCPA at different
at different paddle
paddle
speedsatata afill
speeds filldepth
depthofof16
16mm
mm and
and a dosing depth
depthof of15.5
15.5mm.
mm.
Comparing
Comparingthe theprogression
progressionof of the withincreasing
the data with increasingturret
turretspeed,
speed,oneonecancan observe
observe a maximum
a maximum
turretspeed
turret speedup uptotowhich
which point
point thethe dies can
can be
be filled
filledcompletely
completelyatata acertain
certain paddle
paddle speed
speed forfor
eacheach
material.Above
material. Abovethethemaximum
maximum turretturret speed
speedininaafirst
firstapproximation,
approximation, the
thefilling yield
filling shows
yield shows a linear
a linear
decline.The
decline. Themaximum
maximumturret
turret speed
speed for UCF UCF asaswell
wellas asthe
theslope
slopemaymaybebe interpreted
interpreted as as
characteristic
characteristic
values influenced by the paddle speed and the powder properties. As
values influenced by the paddle speed and the powder properties. As described for PS3, described for PS3, forfor
higher
higher
turretspeeds
turret speedsthe thedie
dieremains
remains forfor aa shorter
shorter period
periodof oftime
timeunder
underthe thefeed
feedframe.
frame. ForFor
this reason,
this thethe
reason,
challenge of complete die filling is magnified with increasing turret speed, especially
challenge of complete die filling is magnified with increasing turret speed, especially for powders with for powders
with
poor poor
flow flow properties.
properties. CoincidentCoincident observations
observations related torelated to the
the decline candecline
be found caninbe
thefound in the
literature [5,28].
literature [5,28]. Peeters et al. presented different linear gradients for the tablet weight with increasing
Peeters et al. presented different linear gradients for the tablet weight with increasing turret speed at
turret speed at two different paddle speeds [5]. However, no further general findings and correlations
two different paddle speeds [5]. However, no further general findings and correlations were derived.
were derived.
The strong impact of the paddle speed can be observed by comparing the different rotational
The strong impact of the paddle speed can be observed by comparing the different rotational
speeds of the feeding wheel. The higher the paddle speed, the higher the maximum turret speed at
speeds of the feeding wheel. The higher the paddle speed, the higher the maximum turret speed at
which complete filling can generally be achieved. Further investigations on the impact of paddle
speed on the filling yield are discussed in the following section.
The decline in the filling yield is based on the discussed reasons for incomplete filling referred
to as a kinetic limitation of the system, which is a general finding for the process. The start of the
limitation,
which completethe filling
maximum turret speed
can generally to achieveFurther
be achieved. UCF, depends on theon
investigations material properties
the impact on one
of paddle speed
onhand, as well
the filling as on
yield arethe process parameters
discussed on thesection.
in the following other hand.
The decline in the filling yield is based on the discussed reasons for incomplete filling referred
4.2.2. Impact of Paddle Speed on the Filling Yield
to as a kinetic limitation of the system, which is a general finding for the process. The start of the
In order
limitation, to better clarify
the maximum the influence
turret speed to achieveof the paddle
UCF, dependsspeed,
on the
the filling
materialyields achieved
properties on at a fill
one hand,
as depth
well asofon 16the
mmprocess
and a dosing depthon
parameters ofthe
15.5other
mm,hand.
as the highest filling challenge, are presented for
MCC (Figure 7a) and DCPA (Figure 7b) as a function of the paddle speed. To show the impact of
4.2.2. Impact oftoPaddle
combination Speed
the turret on the
speed, Filling
three Yieldincluding the slowest, middle and the fastest, are
speeds,
chosen. The paddle speed shows a consistent influence for all experiments and powders—the filling
In order to better clarify the influence of the paddle speed, the filling yields achieved at a fill
yield increases in a degressive way with increasing paddle speed until achieving UCF. Similar
depth of 16 mm and a dosing depth of 15.5 mm, as the highest filling challenge, are presented for
observations were made by Mendez et al. [26]. The slope and, as already discussed in Section 4.2.1,
MCC (Figure 7a) and DCPA (Figure 7b) as a function of the paddle speed. To show the impact of
the paddle speed needed to achieve complete die filling are dependent on the material and its flow
combination to the turret speed, three speeds, including the slowest, middle and the fastest, are chosen.
properties. While the gradient for MCC is high at the beginning, due to the low filling yields at 5 rpm
The paddle
paddle speed,speed shows with
it flattens a consistent
increasing influence for allIn
paddle speed. experiments
contrast, DCPA and shows
powders—the filling yield
a lower gradient at
increases in a degressive way with increasing paddle speed until achieving
low paddle speeds, which declines only slightly. With increasing paddle speed, more powder is UCF. Similar observations
were made by
delivered to Mendez
the fillinget area
al. [26].
in aThe slopetime
certain and,(PS2),
as already
whichdiscussed in Section
can fill the 4.2.1, the paddle
dies. Additionally, a higherspeed
needed to achieve complete die filling are dependent on the material and
paddle speed reduces the relative speed between the dies and the powder moved in the feed frame, its flow properties. While
theasgradient
describedfor forMCC
processis high at the
step PS3 beginning,
(Figure due to the
4). Furthermore, thelow fillinginduced
increased yields atkinetic
5 rpmenergy
paddle speed,
might
it flattens
support with increasing
the powder paddle
flow into the speed.
dies. In contrast, DCPA shows a lower gradient at low paddle
speeds,As which declines
the graphs for only
DCPA slightly. With yields
show filling increasing
abovepaddle speed, moreinside
UCF, consolidation powder theisfeed
delivered
frame and to the
the die
filling areacan inoccur and time
a certain increase as awhich
(PS2), function
canof fillpaddle speed
the dies. under the same
Additionally, staticpaddle
a higher conditions
speed (hreduces
F, hD,
the∆h). On the
relative contrary,
speed between at the
theparameter
dies and the setting
powder displayed,
movedMCC in thedoes
feednot show
frame, asdensification
described foraboveprocess
UCF
step PS3but asymptotically
(Figure reachesthe
4). Furthermore, theincreased
UCF, indicating
induced that no consolidation
kinetic energy mightby dynamic
support theparameters
powder flow
was
into theachieved
dies. for MCC for this case.
(a) (b)
Figure
Figure 7. 7.Impact
Impactof
ofthe
thepaddle
paddle speed
speed on
on the
the filling
fillingyield
yieldfor
for(a)
(a)MCC
MCCand
and(b)(b)
DCPA with
DCPA varying
with varying
turret
turret speed.
speed.
AsInthegeneral,
graphsonefor can
DCPA findshow
a minimum paddle
filling yields speed
above at aconsolidation
UCF, certain turret inside
speed the
to overcome
feed frame theand
thekinetic
die canlimitation
occur and andincrease
to fill the
asdies completely.
a function The higher
of paddle speedthe turret
under thespeed
sameand theconditions
static poorer the (h
flow
F , hD ,
properties of the material, the higher the minimum paddle speed needs to be.
∆h). On the contrary, at the parameter setting displayed, MCC does not show densification above An increasing paddle
UCFspeed
but has a considerably
asymptotically higherthe
reaches impact
UCF,on the fillingthat
indicating yield
noofconsolidation
fairly flowingby (MCC) compared
dynamic to free-
parameters was
flowing powders (DCPA).
achieved for MCC for this case.
In general, one can find a minimum paddle speed at a certain turret speed to overcome the kinetic
4.2.3. Impact of Turret Speed and Paddle Speed on the Tablet Weight Variation Coefficient
limitation and to fill the dies completely. The higher the turret speed and the poorer the flow properties
Besides the
of the material, thetablet
higherweight, which is paddle
the minimum determined
speedbyneeds
the filling
to be.yield, the tablet paddle
An increasing weight variation
speed has
coefficient (WVC) is a crucial factor for tablet production. It is important to gain
a considerably higher impact on the filling yield of fairly flowing (MCC) compared to free-flowing reproducible tablet
weights(DCPA).
powders to ensure constant drug content. The WVCs of MCC and DCPA show strong differences as
4.2.3. Impact of Turret Speed and Paddle Speed on the Tablet Weight Variation Coefficient
Besides the tablet weight, which is determined by the filling yield, the tablet weight variation
coefficient (WVC) is a crucial factor for tablet production. It is important to gain reproducible tablet
Pharmaceutics
weights 2020, 12,
to ensure x FOR PEER
constant REVIEW
drug content. The WVCs of MCC and DCPA show strong differences 12 of 23as
function of the turret and paddle speeds (Figure 8a,b). As Yaginuma and Bellini also presented, the
function of the turret and paddle speeds (Figure 8a,b). As Yaginuma and Bellini also presented, the
WVC for both, MCC and DCPA, increases with rising turret speed [26,33]. The free-flowing DCPA has
WVC for both, MCC and DCPA, increases with rising turret speed [26,33]. The free-flowing DCPA
very low WVCs for paddle speeds of 60 and 40 rpm, while for 20 and 5 rpm paddle speed the value
has very low WVCs for paddle speeds of 60 and 40 rpm, while for 20 and 5 rpm paddle speed the
grows with increasing turret speed. Comparing the data to the WVCs of MCC, it becomes clear that
value grows with increasing turret speed. Comparing the data to the WVCs of MCC, it becomes clear
the weight variation in MCC tablets is about four times higher than for tablets consisting of DCPA, due
that the weight variation in MCC tablets is about four times higher than for tablets consisting of
to the
DCPA, differing
due toflow properties.
the differing flowThe WVCs forThe
properties. 20,WVCs
40 andfor6020,
rpm
40 rise
and with increasing
60 rpm rise withturret speed,
increasing
while the WVC for 5 rpm is approximately constant on a high level. Unexpectedly, for
turret speed, while the WVC for 5 rpm is approximately constant on a high level. Unexpectedly, for both powders,
theboth
variation coefficient
powders, for 20 rpm
the variation exceedsfor
coefficient the20lower paddle speed
rpm exceeds and achieves
the lower values
paddle speed of achieves
and about 11%
forvalues
MCC (Figure 8a) and 3% for DCPA (Figure 8b) at high turret speeds.
of about 11% for MCC (Figure 8a) and 3% for DCPA (Figure 8b) at high turret speeds.
(a) (b)
Figure
Figure 8. Effect
8. Effect of of turret
turret speedon
speed onthe
thetablet
tabletweight
weight variation
variation for
for(a)
(a)MCC
MCCand
and(b)
(b)DCPA
DCPAatat
different
different
paddle
paddle speeds.
speeds.
In In order
order to to clarify
clarify thisphenomenon,
this phenomenon,ititisis necessary
necessary to better
better differentiate
differentiatethe thelimitation
limitation that
thatis is
causing the variation in general. With higher turret speeds, the kinetic limitation
causing the variation in general. With higher turret speeds, the kinetic limitation leads to decreasing leads to decreasing
filled
filled powder
powder masses.
masses. InIncase
casethe
thefilling
fillingyield
yieldis
is considerably
considerably below
belowthe theUCF,
UCF,which
whichmeans
means the kinetic
the kinetic
limit
limit of complete
of complete filling
filling is widely
is widely exceeded,
exceeded, the amount
the amount of filled
of filled powder powder is sothat
is so low lowonly
thatthe
only the
volume
volume of air inside the die is reduced by dosing out. Therefore, the
of air inside the die is reduced by dosing out. Therefore, the resulting powder mass only depends on resulting powder mass only
thedepends
process on steptheofprocess step of
die filling. If die
thefilling.
kineticIflimitation
the kineticislimitation
reached,isthe reached, the filledmass
filled powder powder mass
leads to a
filling yield below UCF. However, the powder height inside the die may be between the fillingthe
leads to a filling yield below UCF. However, the powder height inside the die may be between and
filling and dosing depth, while the apparent density is below the bulk density. In this case, powder
dosing depth, while the apparent density is below the bulk density. In this case, powder ejection and
ejection and densification can occur simultaneously during dosing out. Here, the resulting filling
densification can occur simultaneously during dosing out. Here, the resulting filling yield depends
yield depends on two process steps, die filling and dosing out. This may hypothetically lead to higher
on two process steps, die filling and dosing out. This may hypothetically lead to higher WVCs in
WVCs in comparison to the filling yield, which only depends on the filling.
comparison to the filling yield, which only depends on the filling.
These two phenomena can be identified by comparing the total tablet weight for different dosing
TheseFor
depths. twoMCC, phenomena can beatidentified
tablet weight 5 rpm paddleby comparing
speed falls the total
on the tablet
same weight
masses forfor
all different dosing
dosing depths
depths.
(Figure For
9a).MCC, tablet weight
For DCPA, at 5 rpm paddle
this phenomenon can only speed falls on
be found the the
under same masses
limiting for all dosing
combination of 5depths
rpm
(Figure 9a). For DCPA, this phenomenon can only be found under the
paddle speed, turret speeds of 40 rpm and above, and dosing depths of 15.5 and 13.5 mm (Figure limiting combination of 59b).
rpm
paddle speed, turret speeds of 40 rpm and above, and dosing depths of
In these cases, based on the hypothesis above, no powder is ejected during dosing out. Contrasting 15.5 and 13.5 mm (Figure 9b).
In these cases,
this, the tabletbased on theespecially
weights, hypothesis for above,
DCPA,no powderatis20ejected
decrease duringspeed
rpm paddle dosing out.
but Contrasting
differ for the
this, the tablet
various dosingweights,
depths, especially for DCPA,
which leads decrease at that
to the assumption 20 rpm paddle
powder speed
is still but out.
dosed differ for the
These various
findings
dosing depths,towhich
contribute leads to the
the explanation forassumption
the WVC for that powderspeed
a paddle is stillofdosed
20 rpm out.exceeding
These findings
those ofcontribute
5 rpm
provided
to the above.for the WVC for a paddle speed of 20 rpm exceeding those of 5 rpm provided above.
explanation
Pharmaceutics 2020,
Pharmaceutics 12,12,
2020, 248x FOR PEER REVIEW 1323
13 of of 23
(a) (b)
(a) (b)
Figure
Figure 9. 9. Effectofofturret
Effect turretspeed,
speed,paddle
paddle speed
speed and dosing
dosingdepth
depthon
onthe
therelative
relativetablet weight
tablet variation
weight variation
Figure
for (a) 9.
MCCEffect
and of(b)
turret
DCPAspeed,
at paddle speed
different and
paddle dosing depth on the relative tablet weight variation
speeds.
for (a) MCC and (b) DCPA at different paddle speeds.
for (a) MCC and (b) DCPA at different paddle speeds.
4.2.4. Impact
4.2.4. Impactofofthe theFill
FillDepth
Depth
4.2.4. Impact of the Fill Depth
TheThe filling
fillingyield
yieldisisinvestigated
investigated for for the
the three
three different
different fillfill depths:
depths:10, 10,1414and and1616mm. mm. ToTo yield
yield
comparableThe filling
comparable data,
data, yield
the is investigated
thelinked
linked minimum
minimum for the three
overfill
overfill of 0.5
of different
mm
mm and andfill
thedepths:
the highest
highest 10, 14 and
turret
turret speed 16 of
speed mm.
6060
of Torpm
rpm yield
were were
comparable
chosen, data, the linked
respectively. For minimum
both overfillMCC
formulations, of 0.5and
mmDCPA,and the highest
the higher turret
the speed
fill depth, of the
60 rpm
lower were
the
chosen, respectively. For both formulations, and DCPA, the higher the fill depth, the lower the
chosen,
filling respectively.
yield For turret
both formulations,
(FigureMCC and DCPA, the higher the fill depth, the lower the
filling yield atat a certain
a certain turret speedspeed(Figure 10a,b).
10a,b). With
With increasing
increasingfill filldepth,
depth,the thechallenge
challenge forfor
diedie
filling
filling yield
increasesat a certain
andthe turret
therelative speed
relative (Figure
filling yield10a,b). WithFor
decreases. increasingthe fill depth, thebetween challenge for die
filling increases and filling yield decreases. ForMCC,MCC, thedifference
difference between thethefilling
filling
filling
yield increases and thedepth
relative filling yield decreases. For MCC, theatdifference between theoffilling
yield at at
9.59.5 mm mm dosing
dosing depth inincomparison
comparison to deeper
to deeper onesonesisishigher
higher ataalow lowpaddle
paddle speed
speed 5 rpm
of 5 rpm
yield at
compared 9.5 mm
to an dosing
increased depth in
speed.comparison
The to
free-flowing deeper ones
DCPA is higher
presents at a low
considerably paddle low speed of
differences5 rpm forfor
compared
compared to an
to an increased
increased speed. The
speed. especiallyfree-flowing
The free-flowing DCPA
DCPA presents
presents considerably
considerably lowlow differences
differences for
13.513.5
and and 15.5
15.5 mmmm dosing
dosing depths,
depths,especially
especiallyinin in contrast
contrast to the
to thelowest
lowest filling, i.e.,
filling, a dosing
i.e., depth
a dosing of
depth 9.5 of
13.5 and 15.5 mm dosing depths, contrast to the lowest
mm. DCPA achieves better results for die filling in comparison to MCC, especially at a high fill depth,filling, i.e., a dosing depth of 9.5
9.5mm.mm. DCPA DCPA achieves
achieves betterbetter results
results for die forfilling
die filling in comparison
in comparison toespecially
MCC, especially at adepth,
high fill
owing to their differences in flowability. In addition, for MCCtothe MCC, paddle speed at a high
needs tofill
be higher
depth,
owing owing to
to theirthe their differences
differences in flowability.
in flowability. In addition, for MCC the paddle speed needs to be
to overcome kinetic limitation and toInreachaddition,
UCF for MCC the paddle
in combination withspeed needs to be
the increasing fillhigher
depth.
higher
to to overcome
overcome the kinetic
the results
kinetic limitationlimitation and toUCF
and to unusually,
reach reach UCF in combination
in combination with the increasing fill
Contrasting the for MCC, DCPA, achieves UCF forwith the the
dosingincreasing
depth of fill13.5
depth.
and
depth. Contrasting
Contrasting the the results
results for for MCC,
MCC, DCPA, DCPA,
unusually, unusually,
achieves achieves
UCF for UCFthe for the dosing
dosing depth depth
of 13.5 of 13.5
and
15.5 mm at the same paddle speed. Although the overfill is low, at 0.5 mm each, densification inside
and15.515.5
mmmm
the dies
at the
at the
induced same same
paddle
by the
paddle
speed.
increased
speed.
Although
paddle
Although
speed thetakes the
overfill overfill
placeis low, is 0.5
at
for DCPA.
low, mm at 0.5
each,mm each, densification
densification inside
inside the induced
the dies dies inducedby theby the increased
increased paddlepaddle
speed takesspeedplacetakesforplace
DCPA. for DCPA.
(a) (b)
(a) (b)
Figure 10. Filling yield for different fill depths for (a) MCC and (b) DCPA.
Figure10.
Figure 10. Filling yield
yield for
fordifferent
differentfill
filldepths
depthsfor
for(a)(a)MCC
MCCand (b)(b)
and DCPA.
DCPA.
In general, the increased die volume due to a higher fill depth leads to higher required paddle
InIngeneral,
speeds general, thethe
the
to achieve increased
increased die volume
die
same fillingvolume
yield asdue
for to
due toaahigher
lowerhigherfill
filldepth
volumes. depthleads
Besidesleadstotohigher
a higher higher required
ratepaddle
required
transport ofpaddle
the
speeds
speeds toto achieve
achieve the
the same
same filling
filling yield
yield as for
for lower
lower volumes.
volumes. Besides
Besidesa higher
a higher transport rate
transport
powder, increased centrifugal forces result in a higher powder mass in the filling area, which might of
rate the
of the
powder,increased
powder, increasedcentrifugal
centrifugal forces
forces result in
in aa higher
higherpowder
powdermass
massininthe
thefilling
fillingarea, which
area, which might
might
support the filling. Further investigations are needed in order to take influences as the fill level inside
support the filling. Further investigations are needed in order to take influences as the fill level inside
the feed frame into account. This may enable the identification of possible proportionalities.
the feed frame into account. This may enable the identification of possible proportionalities.
4.2.5. Impact of the Dosing Depth
4.2.5. Impact of the Dosing Depth
Following the section of filling, dosing takes place by an upward movement of the lower punch.
Following the section of filling, dosing takes place by an upward movement of the lower punch.
Therefore, the filling yield is influenced by this process step, and by the dosing depth used. The filling
Therefore, the filling yield is influenced by this process step, and by the dosing depth used. The filling
yield of three different dosing depths (10.5, 13.5 and 15.5 mm) at a fill depth of 16 mm are presented for
yield of three different dosing depths (10.5, 13.5 and 15.5 mm) at a fill depth of 16 mm are presented
MCCfor MCC(Figure 11a,b,c)
(Figure and and
11a,b,c) DCPA DCPA(Figure 11d,e,f)
(Figure for for
11d,e,f) three turret
three speeds.
turret speeds. ForForMCC
MCCand andDCPA,
DCPA,ititcan
becan
observed
be observed that thethatdosing depth
the dosing has has
depth a considerable
a considerable influence
influence ononthe thefilling
fillingyield.
yield. Independent
Independent of
the turret speed, the lowest dosing depth of 10.5 mm performs
of the turret speed, the lowest dosing depth of 10.5 mm performs best in die filling, based best in die filling, based on the lower
on the
final
lower final volume after dosing out, which has to be filled with powder. Due to the normalization ofthis
volume after dosing out, which has to be filled with powder. Due to the normalization of
volume, filling filling
this volume, yieldsyields
are higher. In addition,
are higher. lowerlower
In addition, dosing depths
dosing lead lead
depths to a densification inside
to a densification the die
inside
fortheboth
die powders, MCC and
for both powders, MCCDCPA.andItDCPA.
is observable that thethat
It is observable densification
the densificationabove theaboveUCF,the which
UCF, is
caused
whichby the overfill,
is caused by the isoverfill,
considerably lower forlower
is considerably DCPAfor than
DCPA for MCC.
than for The MCC.difference can be explained
The difference can be
byexplained
the varying by the varying compression
compression behavior ofbehavior
the powders.of theThe powders.
increasingThe increasing
amount ofamount powder,ofwhichpowder,needs
towhich
be dosed needs to leads
out, be dosedto aout, leads to a considerably
considerably stronger increase stronger increaseof
in density in the
density
bulkofinside
the bulk
the inside
die for a
the die for a powder
compressible compressible(MCC, powder (MCC,
i.e., high i.e., high
Hausner Hausner
ratio, ratio,5).
see Figure seeTheFigure 5). The mechanisms
mechanisms are described are in
described in Section 3.2 for process step PS4. Densification by dosing out
Section 3.2 for process step PS4. Densification by dosing out is characterized by a parallel shift in the is characterized by a parallel
shift yield
filling in theto filling
higher yield to higher
values, values, for
applicable applicable for filling
filling yields above yields
UCF.above UCF.
In addition to the densification by dosing out, filling yields above UCFUCF
In addition to the densification by dosing out, filling yields above rise with
rise with increasing
increasing paddle
paddle speed. For MCC at 20 rpm turret speed (Figure 11a), the change
speed. For MCC at 20 rpm turret speed (Figure 11a), the change in density of the bulk is strong in density of the bulk is strong
above
above the UCF and levels off for high paddle speeds at a filling yield of about 0.85. At lower turret
the UCF and levels off for high paddle speeds at a filling yield of about 0.85. At lower turret speeds,
speeds, the increase in the filling yield is almost constant. However, the filling yields are still below
the increase in the filling yield is almost constant. However, the filling yields are still below 0.85, which
0.85, which might be due to lower acting stresses on the powder at the certain process parameters.
might be due to lower acting stresses on the powder at the certain process parameters. Less time for
Less time for consolidation due to the shorter filling time might also be a contributing factor. In
consolidation due to the shorter filling time might also be a contributing factor. In contrast, the filling
contrast, the filling yield for DCPA rises linearly with increasing paddle speed over the whole range
yield for DCPA rises linearly with increasing paddle speed over the whole range studied. However,
studied. However, this effect, and the slope of the filling yield above UCF, decreases with higher
this effect,
turret and the slope of the filling yield above UCF, decreases with higher turret speeds.
speeds.
InIngeneral,
general,higher
higherdosing
dosing depths require higher
depths require higherpaddlepaddlespeedsspeedstoto fillfill
thethe dies
dies completely.
completely. WithWith
increasing turret speed, the paddle speed which ensures complete
increasing turret speed, the paddle speed which ensures complete filling and induces densificationfilling and induces densification
rises
risesasaswell.
well.One Onecan canobserve
observe different behaviorsfor
different behaviors forthe
thechange
changeinindensity
density ofofthethe bulk,
bulk, depending
depending on on
thethedosing
dosingdepth depthand andthethepaddle
paddle speed. Furthermore,ititisisnoticeable
speed. Furthermore, noticeablethat that both
both powders
powders dodo
notnot reach
reach
thethetapped
tappeddensity,
density,according
according to to aa value of
of 11 for
forthethefilling
fillingyield
yieldatatthethe displayed
displayed parameter
parameter settings.
settings.
TheThe stress,
stress,applied
appliedneither
neitherby by paddle
paddle force feeding,
feeding,dosing
dosingout, out,nornorthethecombination
combination of of both,
both, is high
is high
enoughatatthe
enough thegiven
givenprocess
processparameters.
parameters.
(a) (d)
Figure 11. Cont.

Pharmaceutics
Pharmaceutics 2020,
2020, 12,12,248
x FOR PEER REVIEW 15 of
1523of 23
(b) (e)
(c) (f)
Figure
Figure 11.11.Filling
Fillingyield
yieldforfordifferent
different dosing
dosing depth
depthat
ataafill
filldepth
depthofof1616mm
mmatat
20,20,
4040
and 60 60
and rpm turret
rpm turret
speed
speed forMCC
for MCC(a,b,c)
(a, b, c)
andand DCPA
DCPA (d, e, f).
(d,e,f).
4.2.6.
4.2.6.Impact
ImpactofofOverfill
Overfill
InInorder
ordertotoelucidate
elucidate thethe influence
influence of the the overfill,
overfill, the
thefilling
fillingyield
yieldfor forMCC
MCC (Figure
(Figure 12a)
12a) andand
DCPA
DCPA (Figure12b)
(Figure 12b)isisdisplayed
displayed for for aafill
filldepth
depthofof1414andand16 16
mm,mm,both resulting
both in a in
resulting dosing depthdepth
a dosing of
of13.5
13.5mm.
mm.Comparing
Comparingthe thefilling
fillingyield
yieldbelow
belowUCF,
UCF, nonoconsiderable
considerable difference
difference between
between thethe
twotwo
fill fill
depths
depths isisobserved
observedfor forboth
bothpowders.
powders. The The filled
filled powder
powdermass massisisnormalized
normalized toto
thethe
same
same volume
volume for for
both
both datasets—thedosing
datasets—the dosingvolume.
volume.Thus,Thus,thethedata
datashow
show that
that the
the filled
filled powder
powder mass mass isis independent
independent of
of the fill depth as long as the filling is below the UCF. In this case, no powder is
the fill depth as long as the filling is below the UCF. In this case, no powder is ejected during dosing ejected during dosingout.
out. For 14 mm fill depth, the dies are filled completely at a certain paddle speed,
For 14 mm fill depth, the dies are filled completely at a certain paddle speed, but no densification is but no densification
is observable
observable for MCC.
for MCC. Contrasting
Contrasting this,
this, for 16for
mm16fillmm fill depth,
depth, fillingabove
filling yields yieldsthe above
UCFtheare UCF are
observable.
observable. Therefore, the apparent density inside the die after dosing out is above the bulk density.
Therefore, the apparent density inside the die after dosing out is above the bulk density. The higher
The higher filling volume at 16 mm enables more powder to enter the die during filling. In the
filling volume at 16 mm enables more powder to enter the die during filling. In the consecutive process
consecutive process step, the fill volume is reduced by dosing. Thus, the powder is moved upwards
step, the fill volume is reduced by dosing. Thus, the powder is moved upwards by the lower punch,
by the lower punch, resulting in a rearrangement of the particles, which may lead to a change in the
resulting in a rearrangement of the particles, which may lead to a change in the density of the bulk
density of the bulk (PS4). Furthermore, at higher paddle speeds, additional powder may be above
(PS4). Furthermore, at higher paddle speeds, additional powder may be above the die during dosing
the die during dosing out, acting as a resistance against powder ejection and leading to densification
out,
as acting as a As
well (PS4). resistance
MCC has against
a higherpowder ejection and
compressibility leading
based on thetoCompressibility
densification asindex well (PS4).
compared As MCC
to
has a higher compressibility based on the Compressibility index compared
DCPA (Figure 5c), the increase in the filling yield due to the overfill for MCC is considerably higherto DCPA (Figure 5c), the
increase
compared in the
to filling
DCPA.yield due tofor
Although theDCPA
overfill for MCC is considerably
a densification with increasing higher compared
paddle speed is to also
DCPA.
Although
observable foratDCPA a densification
a fill depth of 14 mm, with increasing
the influence paddle
of the speed
overfill is also
on the observable
filling at a fill depth of
yield is negligible.
14 mm, the influence of the overfill on the filling yield is negligible.
Pharmaceutics
2020,
Pharmaceutics 2020, 12,248
12, xx FOR
FOR PEER
PEER REVIEW
REVIEW 16 of
16 of 23of 23
1623
(a)
(a) (b)
(b)
Figure12.
Figure
Figure 12.Impact
12. Impactof
Impact ofthe
of the paddle
the paddle speed
paddle speed on
speed on the
onthe filling
thefilling yield
fillingyield for
yieldfor (a)
for(a) MCC
MCC
(a) and
and
MCC (b)(b)
(b)
and DCPA
DCPA with
with
DCPA varying
varying
with varying
turret
turret speed.
speed.
turret speed.
In
InIn order
order
order tototo deeper
deeper
deeper evaluate
evaluate
evaluate densification
densification
densification due due
due to toto the
the the overfill
overfill
overfill ratio,
ratio,
ratio, filling
filling
filling yields
yields
yields for
for for
MCC MCC
MCC andandand
DCPA
DCPA
atDCPA at every
at everyratio
every overfill overfill
overfill ratio
usedratio used are interpreted
used are interpreted
are interpreted (Figure 13). (Figure
(Figure
At every 13). At
13). fill every
At depth,
every fill fill
threedepth,
depth, three
threedosing
different different
different depth
dosing
dosing
were used,depth
depth which were
were led used,
used,
to three which
which led to
led
overfill toratios
three overfill
three overfill
each. To ratios
ratios each. To
each.
simultaneously To simultaneously
simultaneously
show the impact show the
show thethe
of impact
impact
paddle
of the
of
speed, thethe
paddle
paddle filling speed,
speed,
yieldthethe filling yield
filling
is displayed yieldfor is displayed
is displayed
20 rpm (Figure for 20
for 2013a,c)
rpm (Figure
rpm (Figure 13a,c)
and 6013a,c)rpm and and 60
(Figure60 rpm
rpm
13b,d) (Figure
(Figurepaddle 13b,d)
13b,d)
speed.
paddle
paddle speed.
speed. It
It can
can be
be observed
observed that,
that, for
for both
both powders,
powders,
It can be observed that, for both powders, MCC and DCPA, a linear correlation between the overfill MCC
MCC and
and DCPA,
DCPA, a
a linear
linear correlation
correlation
between the
between the overfill
overfill ratio
ratio and and the the densification—the
densification—the filling filling yield
yield above
above UCF—is
UCF—is assumable,
assumable, indicated
ratio and the densification—the filling yield above UCF—is assumable, indicated by the indicated dotted lines.
by the dotted lines. The slope is determined by powder
by the dotted lines. The slope is determined by powder properties. More susceptible MCC properties. More susceptible MCC displays
displays
The slope is determined by powder properties. More susceptible MCC displays a higher slope and,
aa higher
higher slope slope and, and, through
through that, that, aa higherhigher influence
influence of of overfill.
overfill. For For thethe investigated
investigated process process
through that, a higher influence of overfill. For the investigated process parameters, the paddle speed
parameters, the
parameters, the paddle
paddle speed speed has has no no considerable
considerable impact impact on on the
the slope.
slope. BothBoth powders
powders show show an an
has no considerable impact on the slope. Both powders show an increased filling yield for 60 rpm
increased filling
increased filling yield
yield for for 6060 rpm rpm paddlepaddle speed speed at at aa fillfill depth
depth of of 10
10 mm mm eveneven at at low
low overfill
overfill ratios,
ratios,
paddle
whichspeedmay result at a fill
result depth
from the of 10 mm
higher even at
powder low overfill
volume inside theratios,
the feed
feed which
framemay dueresult thefrom
to the lowerthe higher
outlet
which may from the higher powder volume inside frame due to lower outlet
powder volume
powder flow. flow. Whileinside
While for the
for MCC, feed
MCC, the frame
the filling due
filling yieldto the
yield at lower
at low
low overfill outlet
overfill ratio powder
ratio is
is in flow.
in accordance While
accordance to for
to the MCC,
the UFC, the
UFC, DCPADCPA filling
powder
yield
shows
shows at low overfill ratio
densification
densification evenisat
even atinlowaccordance
low overfill ratios
overfill ratios atUFC,
to theat aa paddle
paddle DCPA speed
speedshowsof 60
of densification
60 rpm. In
rpm. even at low
In comparison,
comparison, at aaoverfill
at low
low
ratios at a
paddle speed
paddle paddle
speed of speed
of 20
20 rpm
rpm theof 60 rpm.
the filling
filling yield In comparison,
yield of of DCPA
DCPA is at
is in a low
in good paddle
good accordance speed
accordance with of
with the 20 rpm
the UCF,
UCF, so the filling
so the
the powder yield of
powder
DCPA
is notisdensified.
is not in good accordance
densified. Therefore, the
Therefore, with
the the UCF, soof
densification
densification the
of powder with
aa powder is notexcellent
with densified.
excellent Therefore,
flowability
flowability the
is more
is more densification
affected
affected
ofbya powder
by the dynamic
the dynamic with parameters
excellent flowability
parameters like the
like the paddleis more
paddle affected
speed
speed than by
than by the
by the dynamic
the overfill ratio.
overfill parameters
ratio. Besides like
Besides the
the reduced
the paddle
reduced
speed than
interparticular by the overfill
forces for ratio.
DCPA, Besides
the the reduced
susceptibility interparticular
to centrifugal
interparticular forces for DCPA, the susceptibility to centrifugal acceleration is higher due to the forces
acceleration for DCPA,
is higherthe susceptibility
due to the
higher
tohigher true density
centrifugal
true density compared
acceleration compared is higher to MCC.
to MCC.
due toThis This mighttrue
the higher
might leaddensity
lead to an
to an increased
increased
compared radial
to MCC.
radial mass
mass Thisflow
flow into lead
might
into the to
the
filling
anfilling area,radial
increased
area, and thus
and thus
mass to flow
to higher
higher into filling yields area,
the filling
filling yields at low
at low andoverfill ratios.
thus ratios.
overfill to higherIn comparison,
In comparison,
filling yieldsfairly fairly
at low flowing
overfilland
flowing and
ratios.
compressible
Incompressible
comparison, powders powders
fairly such and
such
flowing as MCC
as MCC present aapowders
present
compressible strongersuch
stronger dependence
dependence
as MCC on on thethe overfill
present overfill
a stronger ratio
ratio for the
for
dependencethe
increase
onincrease
the overfillin the
in the apparent
apparent
ratio for the density
density
increase inside
insidein thetheapparent
the die during
die during dosinginside
dosing
density out. the die during dosing out.
out.
(a)
(a) (c)
(c)
Figure 13. Cont.

Pharmaceutics2020,
12,248
x FOR PEER REVIEW 17 of1723of 23
(b) (d)
Figure13.
Figure 13.Filling
Fillingyield
yieldatatoverfill
overfillratios
ratiosfor
forthe
thethree
threefill
filldepths
depthsat
ataaturret
turretspeed
speedofof20
20rpm
rpmfor
forMCC
MCCat a
at a paddle speed of 20 (a) and 60 rpm (b) and, for DCPA, at a paddle speed of 20 (c)
paddle speed of 20 (a) and 60 rpm (b) and, for DCPA, at a paddle speed of 20 (c) and 60 rpm (d).and 60 rpm (d).
4.3.
4.3.Impact
ImpactofofMaterial
Material Properties
Properties on the Filling
FillingYield
Yield
TheTheformulation,
formulation, and
and therefore
thereforethe
thematerial
materialproperties,
properties,have a strong
have influence
a strong on dieon
influence filling. As
die filling.
Asthetheflow
flowbehavior
behavior affects the
affects powder
the powder movement
movement inside thethe
inside feed frame
feed andand
frame thethe
flow intointo
flow the the
dies, the the
dies,
capabilityofofparticle
capability particlerearrangement
rearrangement at at low
lowstresses
stresseshas
hasananimpact
impact ononthethe
densification
densificationof the powder.
of the powder.
The investigation of the influence of the material properties on filling yield complements
The investigation of the influence of the material properties on filling yield complements the the analysis
analysis
ofofthe
theprocess
processparameters.
parameters.
4.3.1.Impact
4.3.1. Impactof
ofthe
theMaterial
Material Properties
Properties on
on the
theDensification
Densification
Besidesthe
Besides thechallenge
challenge toto fill
fill the
the dies
dies completely,
completely,there
thereareareprocess
processparameters
parameters which
whichinduce
induce
densification of the powder. Two different mechanisms are identified for densification:
densification of the powder. Two different mechanisms are identified for densification: densification densification
with increasing paddle speed and densification with increasing overfill ratio (Equation 3).
with increasing paddle speed and densification with increasing overfill ratio (Equation (3)).
Densification due to the overfill ratio is presented in Section 4.2.6 with a focus on the impact of
the process parameter. A strong dependence on material properties was already observed for DCPA
and MCC. For further investigations, the filling yield, depending on the overfill ratio, is fitted with a
and MCC. For further investigations, the filling yield, depending on the overfill ratio, is fitted with
linear equation for each fill depth and the slope, and the axis intercept are evaluated for the five
a formulations.
linear equation for each
While fill depth
the slope and thethe
characterizes slope, and the
influence of axis intercept
the overfill are the
ratio, evaluated
axis canfor
helpthetofive
formulations. Whileofthe
identify the impact slope characterizes
different paddle speedsthe on influence
the filling of the The
yield. overfill
dataratio, the axis can
are displayed overhelp
the to
identify
mass content of DCPA (Figure 14,15). As described in Section 4.2.6, the lowest turret speed of 20 rpm the
the impact of different paddle speeds on the filling yield. The data are displayed over
mass content
is used of DCPA
to ensure (Figures
complete 14 andand
die filling 15).toAs described
exclude in Section
dynamic filling 4.2.6, the lowest
limitations turret
(Figure speed
11a,d). To of
20differentiate
rpm is usedtheto impact
ensure of
complete
the paddledie speed,
filling data
and toforexclude
20 and 60dynamic
rpm arefilling limitations (Figure 11a,d).
investigated.
To differentiate the impact of the paddle speed, data for 20 and 60 rpm are investigated.
For the slope, one can observe hyperbolic courses over the mass content of DCPA at both paddle
speeds for all fill depths. MCC provides considerably high slopes, while the values decrease drastically
with increasing DCPA content. A high positive slope indicates a strong impact of the overfill ratio
on the apparent density inside the die. Unexpectedly, the slopes of the blends are almost as low as
for DCPA, which does not correlate with the Compressibility index (Section 4.1). Therefore, another
parameter, characterizing the reduction in density, is necessary to correlate this with the densification
of the powder inside the die due to the overfill ratio.
For the two paddle speeds of 20 and 60 rpm, no considerable and systematic differences can be
observed. The courses for fill depth of 14 and 16 mm are in good accordance, while, at 10 mm fill depth,
the values for the slopes differ slightly. For MCC and the blend with 25 wt % DCPA, lower slopes
are found, while the slopes for DCPA and the blends with a higher DCPA content are slightly higher.
This indicates a higher impact of the overfill ratio at low fill depth. However, this can be related to the
increased fill level inside(a)
the feed frame for free-flowing powders, which (b) is assumed to be higher with
lowering dosing depths. Thus, the higher slope might be due to an increased apparent density inside
and MCC. For further investigations, the filling yield, depending on the overfill ratio, is fitted with a
linear equation for each fill depth and the slope, and the axis intercept are evaluated for the five
formulations.
Pharmaceutics 2020,While
12, 248 the slope characterizes the influence of the overfill ratio, the axis can help 18 toof 23
identify the impact of different paddle speeds on the filling yield. The data are displayed over the
mass content of DCPA (Figure 14,15). As described in Section 4.2.6, the lowest turret speed of 20 rpm
the die and
is used probably
to ensure in the die
complete feedfilling
frame during
and fillingdynamic
to exclude or because of the
filling increased
limitations resistance
(Figure 11a,d).against
To
differentiate
powder ejectiontheduring
impact dosing.
of the paddle speed, data for 20 and 60 rpm are investigated.
(a) (b)
(c)
Figure
Figure 14.14.Slope
Slopeofofthe
thelinear
linear fits
fits of
of filling
filling yield
yield over
over the
the overfill
overfillratio
ratioinindependence
dependence onon
thethe
mass
mass
content
content ofof DCPA
DCPA atat a turretspeed
a turret speedofof20 20rpmrpmatatfill
fill depth
depth of
of 10
10 (a),
(a), 14
14 (b)
(b)and
and16
16mm
mm(c).
(c).
The Foraxis
theintercept,
slope, one investigated
can observe hyperbolic
at paddle courses
speeds overof 20the
and mass content
60 rpm, of DCPA
provides at both paddle
information about
thespeeds for all
theoretical fill depths.
apparent fillingMCC
yieldprovides
inside theconsiderably high ratio
die at an overfill slopes, whilesothe
of zero, no values
volumedecrease
reduction
drastically
takes with the
place. Thus, increasing
influence DCPA content.
of other processA high positive as
parameters, slope indicatesspeed,
the paddle a strong impact
on the of the
densification
overfill ratioBesides
is observable. on the apparent
the datadensity
for the inside the die. Unexpectedly,
axis intercept the slopes
at the paddle speeds of and
of 20 the blends
60 rpm, arethe
almost
UCF is
as low as for DCPA, which does not correlate with the
displayed to compare the data to the status of unconsolidated filling. Compressibility index (Section 4.1). Therefore,
another
For the parameter, characterizing
axis intercept of the fivethe reduction inthe
formulations, density, is necessary
data for a paddle to correlate
speed of 20 this
rpmwith the
are below
densification of the powder inside the die due to the overfill ratio.
60 rpm, which shows a higher densification for a higher paddle speed as expected. Furthermore,
For the two paddle speeds of 20 and 60 rpm, no considerable and systematic differences can be
the difference between the intercepts of the two paddle speeds increases with increasing DCPA mass
observed. The courses for fill depth of 14 and 16 mm are in good accordance, while, at 10 mm fill
content. The difference reflects the stronger influence of the paddle speed on the apparent density
depth, the values for the slopes differ slightly. For MCC and the blend with 25 wt % DCPA, lower
inside the die, and probably inside the feed frame, for excellently flowing powders. It easily enters the
slopes are found, while the slopes for DCPA and the blends with a higher DCPA content are slightly
die cavity during filling, and might firstly be densified by the high amount of powder provided by the
higher. This indicates a higher impact of the overfill ratio at low fill depth. However, this can be
paddles,
related which overtakefill
to the increased thelevel
diesinside
(PS3).the
Thefeedgood
frameflowability of these
for free-flowing formulations
powders, which is allows
assumedfasttoand
easybe rearrangement
higher with loweringof thedosing
particles, which
depths. can the
Thus, contribute to a denser
higher slope might be bulk.
dueFurther densification
to an increased apparenttakes
place during dosing out (PS4). Even MCC shows a higher axis intercept for
density inside the die and probably in the feed frame during filling or because of the increased a paddle speed of 60 rpm
at aresistance
fill depthagainst
of 10 mm, which might be explained
powder ejection during dosing. by the higher fill level inside the feed frame at such
a low fill depth.
The axis intercept, investigated at paddle speeds of 20 and 60 rpm, provides information about
theIntheoretical
addition,apparent
the axis filling
intercept atinside
yield 20 rpm theindie
comparison
at an overfill the UCF
to ratio is ofsospecial
of zero, interest.
no volume For the
reduction
puretakes place. Thus, the influence of other process parameters, as the paddle speed, on the densification of
substances MCC and DCPA, the intercepts are in particularly good agreement with the UCFs
theispowders.
observable. Besides
Thus, at athe data speed
paddle for theofaxis
20 intercept
rpm the diesat theare
paddle speeds filled
completely of 20 and
at an60apparent
rpm, the density
UCF
is displayed to compare the data to the status of unconsolidated filling.
For the axis intercept of the five formulations, the data for a paddle speed of 20 rpm are below
60 rpm, which shows a higher densification for a higher paddle speed as expected. Furthermore, the
difference between the intercepts of the two paddle speeds increases with increasing DCPA mass
content. The difference reflects the stronger influence of the paddle speed on the apparent density
equivalenttotothe
equivalent thebulk
bulkdensity
density without
without dosing
dosing out.
out. Densification
Densificationdoes
doesnot occur.
not occur.In In
contrast to the
contrast to the
results for the pure substances, the blends do not show a good agreement between the
results for the pure substances, the blends do not show a good agreement between the axis intercept axis intercept
and the UCF. The higher intercepts represent an apparent density higher than the bulk density of the
and the UCF. The higher intercepts represent an apparent density higher than the bulk density of the
blends. Thus, due to mixing effects, the blends show a different densification behavior in relation to
blends. Thus, due to mixing effects, the blends show a different densification behavior in relation to the
the bulk density compared to the pure powders MCC and DCPA. A possible explanation can be
bulk density compared to the pure powders MCC and DCPA. A possible explanation can be found by
found by regarding the bulk and tapped densities of the blends (Figure 15d). The bulk and tapped
regarding the bulk and tapped densities of the blends (Figure 15d). The bulk and tapped densities of the
densities of the mixtures are below a theoretical linear mixing rule. This means that the bulk porosity
mixtures are below apure
of the combined theoretical linear
substances mixing
with rule. This
equivalent means
ratios that the
is lower bulk
than porosity
after mixingof the
themcombined
to a
pure substances with equivalent ratios is lower than after mixing them to
homogenous blend. The bulk porosity of the blends might increase due to interacting particlea homogenous blend.
The bulk porosity of the blends might increase due to interacting particle properties.
properties. Hypothetically, the higher axis intercept of the blends compared to the UCF can be Hypothetically,
the higher axis
explained intercept
by the of the
reduction blends
in the compared
increased to the
porosity UCF
under can be
stress, explained
which by at
is applied thea reduction
paddle speedin the
increased
of 20 rpm.porosity under stress, which is applied at a paddle speed of 20 rpm.
(a) (b)
(c) (d)
Figure15.
Figure 15.Axis
Axisintercept
intercept of
of the
the linear fitsfits of
of the
the overfill
overfillratio
ratioinindependence
dependence ononthethe mass
mass content
content of of
DCPA
DCPA atat
a aturret
turretspeed
speedofof20
20rpm
rpm atat fill depths
depths of of10
10(a),
(a),14
14(b)
(b)and
and1616
mm (c);(c);
mm Bulk and
Bulk tapped
and density
tapped density
ofof thefive
the fiveformulations
formulationsusedused(d).
(d).
4.3.2.
4.3.2.Impact
ImpactofofFlow
FlowProperties
Properties
The
Thefindings
findingsfor
forthe
thehighly
highly differing flow properties
differing flow propertiesofofpure pureMCCMCC andand DCPA
DCPA areare extended
extended to to
the influence
the influenceofofthe
the material propertieson
material properties onbasis
basisofofthethe blends
blends of these
of these two.two. Forturret
For the the turret
speedspeed
of 60 of
60rpm,
rpm,one
onecan
canobserve
observeranked
rankedgraphs
graphs forfor
thethe
filling
fillingyields
yieldsof of
thethe
five formulations
five formulations used (Figure
used 16),16),
(Figure
accordingtotothe
according theranking
rankingofof the
the flow
flow properties
properties (Figure
(Figure5a,b).
5a,b).AtAta agiven
given paddle
paddlespeed, DCPA
speed, DCPA hashas
significantly higher values compared to the other formulations. DCPA first achieves
significantly higher values compared to the other formulations. DCPA first achieves complete filling, complete filling,
closely
closely followed
followed byby the
the blend
blend with
with 7575
wtwt%% DCPA.
DCPA. AtAt a turret
a turret speed
speed of of
6060 rpm,
rpm, MCCMCC and
and thethe blend
blend with
with 25 wt % DCPA do not reach the UCF, owing to their considerably worse
25 wt % DCPA do not reach the UCF, owing to their considerably worse flow properties. The higher flow properties. The
the mass
higher thecontent of MCC,ofthe
mass content higher
MCC, thethe minimum
higher paddle speed
the minimum to achieve
paddle speed tocomplete
achieve filling.
completeTherefore,
filling.
the kinetic
Therefore,
higher limitation
the the kinetic
mass is
content linked toisthe
limitation
of MCC, powder
linked tothe
higher properties.
the powder properties.
minimum paddle speed to achieve complete filling.
Therefore, the kinetic limitation is linked to the powder properties.
16. Impact of
Figure 16.
Figure of the
the paddle
paddle speed
speed on
on the
the filling
filling yield
yield for
for the five
five formulations
formulations used
used at aa turret
turret
Figure 16. Impact
Impact of the paddle speed on the filling yield for thethe
five formulations used at a at
turret
speed
speed of 60 rpm, a fill depth of 16 mm and a dosing depth of 15.5 mm.
speedof
of 60
60 rpm,
rpm, aa fill
fill depth
depthof
of16
16mm
mmand
anda adosing
dosing depth
depth of of 15.5
15.5 mm.mm.
In order to gain deeper understanding of the influence of the blend composition, the filling yield is
In
Inorder
order toto gain deeper
deeperunderstanding
understandingofofthe theinfluence
influence of of
thethe blend
blend composition,
composition, the the filling
filling yieldyield
presented in dependence on the mass content of DCPA (Figure 17a). At paddle speeds of 5 and 20 rpm,
isispresented
presented in in dependence
dependence on onthe
themass
masscontent
contentofofDCPADCPA (Figure
(Figure 17a).
17a). At At paddle
paddle speedsspeeds of 5of and5 and
20 20
the filling yield for all formulations is far below the UCF, which is marked for each formulation by an
rpm,
rpm,thethefilling
filling yield for
for all
all formulations
formulationsisisfar farbelow
below the
the UCF,
UCF, which
which is marked
is marked for for
each each formulation
formulation
orange
byan
dot. The
anorange
orange dot.
course of theoffilling
dot. The course
yield withwith increasing DCPA mass content is comparable to the
by course ofthe thefilling
fillingyield
yield withincreasing
increasing DCPA
DCPA massmasscontent
content is comparable
is comparable
onetheof the
onemass discharge rate (Figure 5a). To determine the relationship betweenbetween
the two the parameters,
totothe one of the
of the mass discharge
discharge rate rate(Figure
(Figure5a).5a).ToTodetermine
determine thethe
relationship
relationship between the two two
they are plotted
parameters, they against
they are each
are plotted other (Figure 17b). The graphic shows a linear correlation between the
parameters, plotted against
againsteach
eachother
other(Figure
(Figure 17b).
17b).TheThegraphic
graphic shows
shows a linear
a linear correlation
correlation
filling yield
betweenthe of the
the filling five different
filling yield
yield of formulations at paddle speeds of 5 and 20 rpm and the respective mass
between of the
thefive
fivedifferent
differentformulations
formulations at at
paddle
paddlespeeds
speeds of 5ofand
5 and20 rpm20 rpm andand the the
discharge
respective rate.
mass Therefore,
discharge at a specific
rate. paddle
Therefore, at a speed,
specificthe fillingspeed,
paddle yield the
of different
filling powders
yield of correlates
different
respective mass discharge rate. Therefore, at a specific paddle speed, the filling yield of different
with the mass
powders discharge
correlates with rate. mass
An increasing paddle speed is described byspeed
a parallel shift, accounting
powders correlates with the
the mass discharge
dischargerate.rate.AnAnincreasing
increasing paddle
paddle speedis described
is described by aby a
for the provision
parallel shift, of higher
accounting forpowder
the massesofby
provision the paddle
higher powder feeder.
masses by the paddle feeder.
parallel shift, accounting for the provision of higher powder masses by the paddle feeder.
(a) (b)
Figure 17.
17. Impact (a) (b)
Figure Impact ofofthe
themass
masscontent
contentofofDCPA
DCPA onon
thethe
filling yield
filling (a),(a),
yield andand
correlation of the
correlation of mass
the mass
discharge
discharge rate from
from aa hopper
hopper (25
(25 mm
mm orifice, Figure
orifice, Figure5a)
5a)with
withthe filling
the yield
filling of
yield the
of corresponding
the
Figure 17. Impact of the mass content of DCPA on the filling yield (a), and correlation of the mass corresponding
formulations
formulations
discharge atfrom
rateat paddle
paddle speeds
speedsof
a hopper of55and
(25 and20
mm 20rpm
rpm(b).
orifice, (b). 5a) with the filling yield of the corresponding
Figure
formulations at paddle speeds of 5 and 20 rpm (b).
Thegood
The goodcorrelation
correlationwith
withthe themass
massdischarge
dischargerate
ratemight
mightbebe duedueto to
thethe
factfact
thatthat
thethe effect
effect of
of suction
suction
filling filling is
is negligible negligible at
at the used the used setting. Therefore, the filling of the dies is a similar process to
The good correlation withsetting.
the mass Therefore,
dischargetherate
filling of the
might bedies
dueistoa similar
the factprocess
that thetoeffect
flowing
of
flowing an
through through analthough
orifice, orifice, although air effects
might might differ as it needs to pass the flowing powder in of
suction filling is negligible atair
theeffects
used setting. differ as
Therefore, it needs to pass
the filling the dies
of the flowing
is a powder in case
similar process to
case
die of die filling.
filling. However, However, it has
it has although
to to be considered
be considered that that the diameter
the diameter of the orifice used at thegranular
granularflow
flowing through an orifice, air effects might differ as of the
it needs orifice
to pass used
theatflowing
the powder in
flow tester
tester is 25 mm,
is 25 mm, whilewhile the dies
the dies havehave a diameter
a diameter of 9ofmm.
9 mm. Nevertheless,the
Nevertheless, thecorrelation
correlation between
between the
case of die filling. However, it has to be considered that the diameter of the orifice used at the granular
flow tester is 25 mm, while the dies have a diameter of 9 mm. Nevertheless, the correlation between
powder property and the filling yield of different powders enables first predictions of the filling yield
that are below the UCF for different powders.
5. Conclusions
The presented work focuses on a systematic evaluation of the impact of process parameters and
material properties on the result of die filling, expressed as the dimensionless filling yield. For an
increase in turret speed, and therefore in tablet production rate, the dies can be completely filled only
up to a critical velocity, dependent on the applied paddle speed and in particular on the material
properties such as the flowability. Exceeding this velocity, the kinetic limitation is reached, and the
filling yield declines approximately linearly. For fairly flowing powders (MCC), this critical turret
speed is lower in comparison to good flowing formulations (DCPA). In addition, the tablet weight
variation coefficient is considerably higher and rises with increasing turret speed. An increase in
paddle speed shows a degressive rise in the filling yield for all powders presented and allows higher
critical turret speeds, still achieving complete filling. The better the flowability, the lower the minimum
paddle speed required to achieve die filling at a certain turret speed. Furthermore, higher paddle
speeds lead to smaller variation coefficients. For low paddle speeds (5, 20 rpm) a limited powder mass,
which entered the die as of a certain turret speed, is observed for DCPA and MCC, independent from
the dosing depth.
Higher fill depth and therefore higher dosing depth increase the challenge of obtaining complete
filling naturally, especially for fairly flowing powders (MCC). Besides the challenge of filling the
dies completely, densification of the powder is observed, exceeding the unconsolidated complete die
filling. Two mechanisms are identified which lead to denser bulks during filling: the paddle speed
and the dosing out, quantified by the overfill ratio. With increasing paddle speed, especially at low
turret speeds and a low dosing depth, the powder in front of the paddles and inside the dies can
be densified up to the tapped density for good flowing powder (DCPA). Overfill ratio is identified
as an independent factor that causes a linear increase in density. The analysis of the slope and axis
intercept of the filling yield over the overfill ratio at different paddle speeds displayed the influence
of the process parameters on densification, dependent on the material properties. While the overfill
ratio is responsible for an increase in the apparent density for compressible powders such as MCC, the
paddle speed has a considerable impact on the densification of excellent flowing powders as DCPA
and blends with a high DCPA mass content. However, a correlation with measured powder properties
as the Compressibility index was not identified. Besides the densification, the filling yield below UCF
differs at the same process parameters for different materials. A linear correlation between the filling
yield and the mass discharge rate from an orifice was found for low paddle speeds of 5 and 20 rpm.
Based on the general correlations identified in these findings, it is necessary to find appropriate
process parameters to achieve complete filling without densifying the powder. Therefore, the paddle
speed needs to be high for fairly flowing powders, while the overfill ratio should be small if densification
is undesired. In contrast, for good flowing powders, the paddle speed should not be considerably
higher than needed to fill the dies completely, if densification is to be avoided.
In the future, a prediction of the filling yield should be searched for. Therefore, the courses of the
filling yield depending on the dynamic process parameters need to be mathematically described and
the influence of the material properties needs to be quantified and correlated to measurable properties.
The findings presented here deliver a basis for the model development. Moreover, a scale transfer to
larger productions presses and different feeding systems is envisaged.
Author Contributions: Conceptualization, A.K.S., J.H.F., A.K.; methodology, A.K.S., J.H.F.; formal analysis, A.K.S.;
investigation, A.K.S.; data curation, A.K.S.; writing—original draft preparation, A.K.S.; writing—review and
editing, J.H.F., A.K.; visualization, A.K.S.; supervision, J.H.F.; project administration, J.H.F. All authors have read
and agree to the published version of the manuscript.
Funding: This research was funded by the German Research Foundation (project number: 413141366).
We acknowledge support by the German Research Foundation and the Open Access Publication Funds of
the Technische Universität Braunschweig.
Acknowledgments: The authors acknowledge and thank KORSCH AG for the support and the provision of the
tablet press XL 100 and JRS Pharma GmbH & Co. KG for the provision of testing materials (DCPA, MCC).
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Chatterjee, S.; Food and Drug Administration CDER. Guidance for Industry, Q8 Pharmaceutical Development;
Food and Drug Administration CDER: White Oak, MD, USA, 2006.
2. Jackson, S.; Sinka, I.C.; Cocks, A.C.F. The effect of suction during die fill on a rotary tablet press. Eur. J. Pharm.
Biopharm. 2007, 65, 253–256. [CrossRef]
3. Wu, C.-Y.; Dihoru, L.; Cocks, A.C.F. The flow of powder into simple and stepped dies. Powder Technol. 2003,
134, 24–39. [CrossRef]
4. Mateo-Ortiz, D.; Colon, Y.; Romañach, R.J.; Méndez, R. Analysis of powder phenomena inside a Fette 3090
feed frame using in-line NIR spectroscopy. J. Pharm. Biomed. Anal. 2014, 100, 40–49. [CrossRef] [PubMed]
5. Peeters, E.; de Beer, T.; Vervaet, C.; Remon, J.-P. Reduction of tablet weight variability by optimizing paddle
speed in the forced feeder of a high-speed rotary tablet press. Drug Dev. Ind. Pharm. 2015, 41, 530–539.
[CrossRef]
6. Xie, X.; Puri, V.M. Uniformity of Powder Die Filling Using a Feed Shoe: A Review. Particul. Sci. Technol.
2006, 24, 411–426. [CrossRef]
7. Baserinia, R.; Sinka, I.C. Mass flow rate of fine and cohesive powders under differential air pressure. Powder
Technol. 2018, 334, 173–182. [CrossRef]
8. Baserinia, R.; Sinka, I.C. Powder die filling under gravity and suction fill mechanisms. Int. J. Pharm. 2019,
563, 135–155. [CrossRef]
9. Schneider, L.C.R.; Sinka, I.C.; Cocks, A.C.F. Characterisation of the flow behaviour of pharmaceutical powders
using a model die–shoe filling system. Powder Technol. 2007, 173, 59–71. [CrossRef]
10. Sinka, I.C.; Cocks, A.C.F. Evaluating the flow behaviour of powders for die fill performance. Powder Metall.
2009, 52, 8–11. [CrossRef]
11. Goh, H.P.; Heng, P.W.S.; Liew, C.V. Understanding effects of process parameters and forced feeding on die
filling. Eur. J. Pharm. Sci. 2018, 122, 105–115. [CrossRef]
12. Sinka, I.C.; Schneider, L.C.R.; Cocks, A.C.F. Measurement of the flow properties of powders with special
reference to die fill. Int. J. Pharm. 2004, 280, 27–38. [CrossRef]
13. Bierwisch, C.; Kraft, T.; Riedel, H.; Moseler, M. Die filling optimization using three-dimensional discrete
element modeling. Powder Technol. 2009, 196, 169–179. [CrossRef]
14. Bierwisch, C.; Kraft, T.; Riedel, H.; Moseler, M. Three-dimensional discrete element models for the granular
statics and dynamics of powders in cavity filling. J. Mech. Phys. Solids 2009, 57, 10–31. [CrossRef]
15. Guo, Y.; Wu, C.-Y.; Kafui, K.D.; Thornton, C. 3D DEM/CFD analysis of size-induced segregation during die
filling. Powder Technol. 2011, 206, 177–188. [CrossRef]
16. Guo, Y.; Wu, C.-Y.; Thornton, C. The effects of air and particle density difference on segregation of powder
mixtures during die filling. Chem. Eng. Sci. 2011, 66, 661–673. [CrossRef]
17. Nwose, E.N.; Pei, C.; Wu, C.-Y. Modelling die filling with charged particles using DEM/CFD. Particuology
2012, 10, 229–235. [CrossRef]
18. Wu, C.-Y.; Cocks, A.C.F. Numerical and experimental investigations of the flow of powder into a confined
space. Mech. Mater. 2006, 38, 304–324. [CrossRef]
19. Ketterhagen, W.R. Simulation of powder flow in a lab-scale tablet press feed frame: Effects of design and
operating parameters on measures of tablet quality. Powder Technol. 2015, 275, 361–374. [CrossRef]
20. Furukawa, R.; Shiosaka, Y.; Kadota, K.; Takagaki, K.; Noguchi, T.; Shimosaka, A.; Shirakawa, Y. Size-induced
segregation during pharmaceutical particle die filling assessed by response surface methodology using
discrete element method. J. Drug. Deliv. Sci. Technol. 2016, 35, 284–293. [CrossRef]
21. Wu, C.-Y.; Guo, Y. Numerical modelling of suction filling using DEM/CFD. Chem. Eng. Sci. 2012, 73, 231–238.
[CrossRef]
22. Hildebrandt, C.; Gopireddy, S.R.; Scherließ, R.; Urbanetz, N.A. Investigation of powder flow within a
pharmaceutical tablet press force feeder—A DEM approach. Powder Technol. 2019, 345, 616–632. [CrossRef]
23. Tsunazawa, Y.; Shigeto, Y.; Tokoro, C.; Sakai, M. Numerical simulation of industrial die filling using the
discrete element method. Chem. Eng. Sci. 2015, 138, 791–809. [CrossRef]
24. Gopireddy, S.R.; Hildebrandt, C.; Urbanetz, N.A. Numerical simulation of powder flow in a pharmaceutical
tablet press lab-scale gravity feeder. Powder Technol. 2016, 302, 309–327. [CrossRef]
25. Zakhvatayeva, A.; Zhong, W.; Makroo, H.A.; Hare, C.; Wu, C.Y. An experimental study of die filling of
pharmaceutical powders using a rotary die filling system. Int. J. Pharm. 2018, 553, 84–96. [CrossRef]
26. Mendez, R.; Muzzio, F.; Velazquez, C. Study of the effects of feed frames on powder blend properties during
the filling of tablet press dies. Powder Technol. 2010, 200, 105–116. [CrossRef]
27. Goh, H.P.; Sia Heng, P.W.; Liew, C.V. The Effects of Feed Frame Parameters and Turret Speed on Mini-Tablet
Compression. J. Pharm. Sci. 2019, 108, 1161–1171. [CrossRef]
28. Yaginuma, Y.; Ozeki, Y.; Kakizawa, M.; Gomi, S.-I.; Watanabe, Y. Effects of powder flowability on die-fill
properties in rotary compression. J. Drug. Deliv. Sci. Technol. 2007, 17, 205–210. [CrossRef]
29. van Snick, B.; Grymonpré, W.; Dhondt, J.; Pandelaere, K.; Di Pretoro, G.; Remon, J.P.; de Beer, T.; Vervaet, C.;
Vanhoorne, V. Impact of blend properties on die filling during tableting. Int. J. Pharm. 2018, 549, 476–488.
[CrossRef]
30. Dühlmeyer, K.P.; Özcoban, H.; Leopold, C.S. A novel method for determination of the filling level in the feed
frame of a rotary tablet press. Drug Dev. Ind. Pharm. 2018, 44, 1744–1751. [CrossRef]
31. Dühlmeyer, K.P.; Özcoban, H.; Leopold, C.S. Inline monitoring of the powder filling level within a rotary
tablet press feed frame. Powder Technol. 2019, 351, 134–143. [CrossRef]
32. Dühlmeyer, K.P.; Özcoban, H.; Leopold, C.S. Comparison of two paddle wheel geometries within the filling
chamber of a rotary tablet press feed frame with regard to the distribution behavior of a model powder and
the influence on the resulting tablet mass. Drug Dev. Ind. Pharm. 2019, 45, 1233–1241. [CrossRef]
33. Grymonpré, W.; Vanhoorne, V.; van Snick, B.; Blahova Prudilova, B.; Detobel, F.; Remon, J.P.; de Beer, T.;
Vervaet, C. Optimizing feed frame design and tableting process parameters to increase die-filling uniformity
on a high-speed rotary tablet press. Int. J. Pharm. 2018, 548, 54–61. [CrossRef]
34. Schulze, D. Powders an Bulk Solids. Behavior, Characterization, Storage and Flow; Springer: Berlin/Heidelberg,
Germany, 2007; ISBN 978-3-540-73767-4.
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).

The Challenge of Die Filling in Rotary Presses-A Systematic Study of Material Properties and Process Parameters

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Challenge of Die Filling in Rotary Presses-A Systematic Study of Material Properties and Process Parameters

Uploaded by

Copyright:

Available Formats

pharmaceutics

Pharmaceutics 2020, 12, 248; doi:10.3390/pharmaceutics12030248 www.mdpi.com/journal/pharmaceutics

2. Materials and Methods

2.2.1. Powder Characterization

2.2.2. Powder Blending

Blend wt % MCC wt % DCPA + 1% MgSt

2.2.3. Die Filling/Tableting Experiments

Table 3. Process parameters for tablet production with XL 100.

Pharmaceutics 2020, 12, x FOR PEER REVIEW 6 of 23

3.2. Detailed Process Description

Pharmaceutics 2020, 12, x FOR PEER REVIEW 14 of 23

Figure 11. Cont.

Figure 13. Cont.

Pharmaceutics 2020, 12, x FOR PEER REVIEW 18 of 23

You might also like