See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/324488443

A review of current techniques for the evaluation of powder mixing

Article  in  Advanced Powder Technology · April 2018

DOI: 10.1016/j.apt.2018.03.031

CITATIONS READS

25 9,242

3 authors:

Maryam Asachi Ehsan Nourafkan

University of Leeds Queen Mary, University of London
13 PUBLICATIONS   111 CITATIONS    39 PUBLICATIONS   532 CITATIONS   

SEE PROFILE SEE PROFILE

Ali Hassanpour
University of Leeds
156 PUBLICATIONS   2,558 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Flowability Assessment of Weakly Consolidated Powders View project

Special Issue "Solid acid catalysts for biodiesel production" View project

All content following this page was uploaded by Ehsan Nourafkan on 18 September 2018.

The user has requested enhancement of the downloaded file.

 Advanced Powder Technology 29 (2018) 1525–1549

Contents lists available at ScienceDirect

Advanced Powder Technology

journal homepage: www.elsevier.com/locate/apt

Review paper

A review of current techniques for the evaluation of powder mixing

Maryam Asachi, Ehsan Nourafkan ⇑, Ali Hassanpour ⇑
School of Chemical and Process Engineering, University of Leeds, Leeds LS2 9JT, UK

a r t i c l e i n f o a b s t r a c t

Article history: Blending a mixture of powders to a homogeneous system is a crucial step in many manufacturing pro-
Received 9 August 2017 cesses. To achieve a high quality of the end product, powder mixtures should be made with high content
Received in revised form 22 March 2018 uniformity. For instance, producing uniform tablets depends on the homogeneous dispersion of active
Accepted 30 March 2018
pharmaceutical ingredient (API), often in low level quantities, into excipients. To control the uniformity
Available online 12 April 2018
of a powder mixture, the first required step is to estimate the powder content information during blend-
ing. There are several powder homogeneity evaluation techniques which differ in accuracy, fundamental
Keywords:
basis, cost and operating conditions. In this article, emerging techniques for the analysis of powder con-
Powder mixing evaluation
Powder blend homogeneity
tent and powder blend uniformity, are explained and compared. The advantages and drawbacks of all the
Wet techniques techniques are reviewed to help the readers to select the appropriate equipment for the powder mixing
Dry techniques evaluation. In addition, the paper highlights the recent innovative on-line measurement techniques used
for the non-invasive evaluation of the mixing performance.
Ó 2018 The Society of Powder Technology Japan. Published by Elsevier B.V. and The Society of Powder
Technology Japan. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1526
2. Wet techniques to assess powder blend uniformity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1529
2.1. Ultra-Violet (UV)-visible absorbance spectrophotometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1529
2.2. HPLC (High-Performance Liquid Chromatography) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1529
3. Dry techniques to assess powder blend uniformity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1531
3.1. Image analysis technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1531
3.2. Assessment of powder blend uniformity by bulk particle properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1532
3.2.1. Pressure drop method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1532
3.2.2. Electrical conductivity method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1533
3.2.3. Tribo-electrification method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1533
3.2.4. Thermal analytical method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1534
3.3. Tomographic techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1536
3.3.1. X-ray microtomographic method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1536
3.3.2. Positron emission particle tracking (PEPT) method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1537
3.3.3. Electrical capacitance tomography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1539
3.3.4. Magnetic resonance imaging tomography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1541
3.4. Spectroscopic techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1541
3.4.1. Near-infrared spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1541
3.4.2. Raman spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1543
3.4.3. Acoustic emission spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1544
3.4.4. Fluorescence spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1546
4. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1546
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1547

⇑ Corresponding authors.
E-mail addresses: [email protected] (E. Nourafkan), [email protected] (A. Hassanpour).

https://doi.org/10.1016/j.apt.2018.03.031
0921-8831/Ó 2018 The Society of Powder Technology Japan. Published by Elsevier B.V. and The Society of Powder Technology Japan. All rights reserved.
1526 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549

Nomenclature

AE acoustic emission PLSR partial least squares regression

API active pharmaceutical ingredient PT pressure transducer
ASA acetyl salicylic acid RAM resonant acoustic mixing
CCD charge coupled device RSC royal society of chemistry
CV coefficient of variation RSD relative standard deviation
DSC Differential Scanning Calorimetry SEM scanning electron microscope
ECT electrical capacitance tomography SiC silicon carbide
FT Fourier-transform SRS stimulated Raman scattering
HPLC High-Performance Liquid Chromatography U velocity
LIF light induced fluorescence UV Ultra-Violet
MCC microcrystalline cellulose mCT X-ray computed microtomography
MRI magnetic resonance imaging
NIR near-Infrared
PEPT positron emission particle tracking

1. Introduction To apply a proper sampling procedure on bulk powders, the

population and sample size, sample collection and sample size
Powder mixing unit operation is a very common step in partic- reduction method, and statistical analysis confirming the stated
ulate processes and has significant impact on the quality of the end level of acceptance of the sampling plan must be fully addressed
product; examples are in industries such as pharmaceutical, agro- [16,17]. A general framework of the sampling used in powder mix-
food, cement and plastics [1–8]. The final characteristics of the ing processes is illustrated in Fig. 1a.
powdered products nowadays are becoming more complex. In The scale of scrutiny according to the product specification is
some cases, a mixture of up to 20 powder ingredients is necessary mainly used to determine the ideal size of samples. The taken sam-
to meet acceptable quality standard of the final product. Powder ples having the required size of scale of scrutiny must be analysed
segregation, a phenomenon which is described as the opposite of for the composition evaluation [12]. Generally, sample size should
mixing, or reverse mixing, takes place as a result of powder in- be close to the size of the final product, e.g. tablet size in pharma-
homogeneity during blending process or during secondary pro- ceutical industry. Moreover, the sample size should be larger than
cessing steps such as packaging and transportation [9–11]. A batch the minimum amount needed for the analytical analysis. It should
of pharmaceutical tablets at a cost of thousands of dollars could be be noted that if the scale of scrutiny is defined too small, the homo-
rejected due to powder in-homogeneities arising from segregation geneity of the blend would wrongly appear to be unacceptable. In
of active component. As another example, layer by layer deposition contrast, if the sample size is considered too large, the homogene-
of the segregated powders before getting fused together by laser, ity of the blend would wrongly appear to be acceptable resulting in
could adversely affect the quality of products in 3D printing an overestimation of the homogeneity (Fig. 1b). If the scale of scru-
manufacturing. tiny is much smaller than the minimum amount of sample that the
Mixing processes of granular materials often aim at producing a sampler provides, reliable sample reduction techniques, which
product with a suitable degree of homogeneity. There are two main have been fully reviewed by Allen [18], must be applied. The gran-
types of equipment available for the mixing of granular materials. ules segregation during sample reduction process must be pre-
In the first group, the container is rotated around an axis to move vented to have representative sub-samples which meet the scale
around the materials inside to mix them with a dominant shear of scrutiny. Cone and quartering, scoop sampling, rotary riffler,
and diffusion mixing mechanisms. On the other hand, convection chute riffler and table riffler are among the most well-known and
and/or shear could be the main mechanism in the second group applied methods for sample splitting. Allen reported the represen-
in which the container is stationary and an internal rotor causes tivity of all the methods (based on standard deviation in size distri-
a mild or fast agitation. For the mixing of granular materials, the bution observed between different subsamples generated from the
choice of mixer type is vital as the quality of the mixture highly same primary sample) and introduced the rotary riffler as the most
depends on the mixer selection [12]. In addition, the quality of representative method for sample size reduction [18].
the mixture and the end-point of the process should be interpreted Powder sampling in a dynamic mixing process (Fig. 1a) must be
by evaluating the samples taken from the mixture. The goal of performed based on Golden Rules as follows [19]:
powder sampling is to collect a small amount of sample from a
bulk powder in such a way that the sample represents the physical – Rule 1: Powder mixtures should be sampled when in motion.
and chemical characteristics of the entire bulk. It should be noted – Rule 2: The whole stream of the powder mixture should be
that acquiring a representative sample from the bulk powder is taken for many short increments of time, rather than part of
crucial because further analyses and data interpretation regarding the stream being taken for the whole time.
the mixing process performance depends on the sampling repre-
sentativity [13]. Acquisition of samples at the outlet of a continuous mixer
It should be noted that developing a general guidance for should be performed at a regular frequency (starting at 3 times
obtaining a representative sampling is challenging and beyond the residence time of the mixer when the system has reached a
the scope of this article since the conditions of mixing processes steady state). Full stream samplers at the outlet of the mixer are
are different from case to case. However, a brief description of sam- used in continuous mixers [19]. In a batch mixer, the mixer needs
ple representativity procedure has been provided here and the to be stopped at different blending times and then the sampling is
readers are referred to the additional sources on this topic carried out at various locations (by dividing the space into equal
[7,14,15]. regions). Regarding the location of sampling, the whole powder
 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549 1527

(a) Blending Loading of raw Mixing and End point

vessel materials sampling mixing time

Representativity
of sampling
-Smaller than scale of scrutiny Size reduction
Sample size: (Not acceptable) methods
Sampling system -Cone and quartering
-Larger than scale of scrutiny -Sample splittor
Off-line
(size reduction is required) -Rotary riffler
At-line
-Chute riffler
In-line
-Table riffler
On-line
Sample number: Golden rules
Samples location:
Invasive
-Cross-cut sampler Non-invasive
-Powder thief -NIR probe
-FreeGlide -Raman probe
Sample collection: -Slot, tip, pocket or -Light induced
and sleeve samplers fluorescence

(b)
Too small scale of scrutiny (poorly mixing evaluation)

Proper scale of scrutiny

Too large scale of scrutiny (overestimated mixing)

(c)

Fig. 1. (a) A general framework of sampling in powder mixing processes, (b) effect of the scale of scrutiny on mixing evaluation, (c) different powder mixing evaluation
techniques reviewed in the current study.
1528 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549

(a)

(b)

30.0 ELSD Test #23 [modified by chmhplc] 0.05 Hexane DAD_Signal_C

mAU

2 - 21.807
25.0
3 - 23.266

20.0

15.0

1 - 3.280

10.0
5 - 30.562

5.0 4 - 29.971

0.0

min
-5.0
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 44.9

Chromatogram of HPLC

Fig. 2. (a) Schematic of UV–Visible spectrometer, (b) High-pressure liquid chromatography (with permission from University of Leeds).

stream should be sampled for numerous short increments of time troscopy (NIR), Raman spectroscopy and Electrical Capacitance
[19]. Tomography with no interference with the blending process [23].
The traditional samplers, e.g. thief and cross-cut, are most com- To control the uniformity of a mixture and reduce its inhomo-
monly used instruments for powder sampling [20]. Using samplers geneity during blending, it is necessary to monitor the component
is mostly an invasive approach which causes disturbance in the concentration inside the mixer. This helps to achieve the optimum
mixture, hence special care and designs are needed to mitigate it mixing conditions, such as end-point mixing time. Process moni-
[15,21,22]. However, in recent decades, non-invasive analytical toring can be achieved by performing routine testing of the process
technologies have been developed such as Near-Infrared spec- samples using off-line, at-line, inline and on-line instrumentations.
 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549 1529

Off-line analyser is a discontinuous, invasive, slow and time con- samples, spectrum background correction using a buffer blank is
suming traditional laboratory method that is performed in a con- needed. According to Fig. 2a, the unknown sample is placed
trolled location by a technically trained person. For this purpose, between a light source and a photodetector. The intensity of the
the samples removed from the process are transported to a labora- beam of UV–visible light, before and after passing the light through
tory for further analysis. Its advantage lies in the fact that it pro- the sample, is then used to measure the unknown concentration
vides the greatest flexibility in selecting the measurement [28].
method, sample preparation and the most accurate method [24]. Determination of the active ingredient content of powder mix-
Procedure of at-line analyser is similar to that of off-line, however, ture samples (especially in pharmaceutical products e.g. tablets
the main difference is the time duration of the analysis. Usually, at- and capsules) with UV–visible spectroscopy is a frequently useful
line analysis can be performed with automatic facilities quicker technique [29,30]. In a work done by Mendez et al. [20] pharma-
than off-line analysis by a mediocre operator. For at-line analysis, ceutical powder mixing was successfully evaluated in a V-
a defined device is placed close to the line to analyse the samples blender during batch production using UV–visible spectrophotom-
which may be occasionally or continuously extracted from the etry. The acetaminophen concentration of tablets, produced from
stream. Although the devices used in at-line analysis are mainly samples from different regions of the blender at different times,
robust, they rely on standardized procedures and fixed parameter was measured using UV–visible spectrophotometry (Agilent UV–
settings [24]. VIS 8453E double-beam spectrophotometer at a wavelength of
In-line analysers are simple sensors or measuring devices that 244 nm). 3 samples of approximately 3.0 g were collected each
are placed directly into a process stream. Therefore, there is no time using a thief probe at three different locations (right, left
need to extract a sample from the process for examination. In- and the bottom of blender). To perform quantitative determina-
line process measurement performs the analysis either invasively tion, different standard solutions of acetaminophen were prepared
by using a probe positioned in the process stream or non- to build a calibration curve.
invasively in which the probe does not have physical contact with Moreover, the UV–visible spectrophotometer is widely used as
the sample [25]. However, using in-line method, achieving a repre- reference method for the determination of active ingredient during
sentative sample might be difficult since the measurements could powder mixture as it is fast and simple method [31–33]. However,
be influenced by immediate process fluctuations. Temperature, pH, the disadvantages of this analytical method are as follow [34–36]:
pressure, and flowrate are the usual process parameters which are
measured in-line. On-line analysers are fully automated systems – The main disadvantage of UV–visible spectrophotometry is that
used to closely monitor the parameters that are critical in the pro- the absorbance spectra of soluble components are measured all
duction process. On-line methods have this ability to control the together. Therefore, the separate measurement of component
mixing processes by automatically changing the status of devices, fractions within a mixture is not applicable in most cases using
such as valves, as part of their analysis sequence. Using an on- this technique.
line measurement, continuously extracted samples from the pro- – The impurities could influence the absorption spectra of the tar-
cess stream (by the means of a bypass) are transported to a special- get component. Therefore, UV–visible spectrophotometry could
ized analyser and then returned to the process stream. This not properly discriminate between the sample of interest and
eliminates many preparation possibilities (i.e., all forms of destruc- the contaminants absorbing at the same wavelength. Moreover,
tive testing methods) and allows a large fraction of the product the detectors of the spectrophotometers are sensitive to the
stream to be analysed [24]. light. If any impurity in the sample reflects the light, an erro-
Current research provides useful information on analytical neous reading may be recorded by the detectors.
methods for uniformity analysis of powder mixtures (Fig. 1c). Fast – Absorption values could be influenced by parameters such as
and accurate uniformity evaluation of powder blends is one of the temperature and pH leading to inaccurate result analysis.
challenges faced by many industries. This review article provides – The sensitivity of a spectrophotometer is often inadequate at
comprehensive background information on different powder mix- low concentrations.
ing evaluation techniques and introduces the readers to the pros
and cons of all the techniques in terms of cost, functionality, preci- 2.2. HPLC (High-Performance Liquid Chromatography)
sion and operation conditions.
High-pressure liquid chromatography (HPLC) is a traditional
off-line wet technique to separate, identify, and quantify the com-
2. Wet techniques to assess powder blend uniformity
ponents in a mixture (Fig. 2b). HPLC differs from the standard col-
umn chromatography method by the fact that it applies pressure to
2.1. Ultra-Violet (UV)-visible absorbance spectrophotometry
force the solution through the column, and therefore provides
quicker, and more accurate results [37,38]. As shown in Fig. 2b,
By illuminating the sample with UV radiation and obtaining the
the solvent flows through a column with the help of a high pres-
absorbance, it is possible to estimate the concentration of a speci-
sure (up to 400 atm). The sample, injected to the mobile phase, is
men. According to the Beer-Lambert law (Eq. (1)), the absorbance
transported to the HPLC column for content determination as
information has a linear relationship with the concentration data
schematically shown in Fig. 2b.
[26,27].
The separation of components takes place in a stationary phase,
A ¼ aðkÞ  b  C ð1Þ fixed inside the HPLC column, because of difference in relative
affinities of the constituents. Different retention times of the con-
where A, b, C and aðkÞ are the measured absorbance, path length, stituents in the outlet of HPLC column enables the estimation of
specimen concentration and wavelength dependant absorptivity different component fractions by a detector [39,40]. A broad range
coefficient, respectively. Thus, UV–visible spectrophotometry is of detectors such as UV/Vis, florescence (FL), Evaporative Light
able to determine the concentration of the absorber in a solution Scattering (ELSD) and Photo Diode Array (PDA) are available for
based on Beer-Lambert law after selection of a proper wavelength. the detection of different types of constituents in the mobile phase.
Quantifying the trend of the absorbance values versus different The description of different detector types is beyond the scope of
standard concentrations is necessary for the measurement of the this article and the readers are referred to the additional sources
unknown concentrations. Before spectral analysis of the unknown on this topic [26,37]. For the content analysis of the unknown
1530 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549

samples, the standard solutions are first analysed by HPLC to phosphate mixture were used. Quantification of naftazone was
obtain the calibration plot, from which it is possible to measure performed by making a calibration curve obtained from different
the chromatogram areas as a linear function of the concentrations standards containing different concentrations of naftazone. Good
of each constituent. After analysing the chromatogram of the correlation over the concentration range of 0.1–10.0 lg/mL has
unknown sample (Fig. 2b), the concentration of each constituents been shown by this method with a lower detection and quantifica-
could be specified [41]. tion limits of 0.032 and 0.096 lg/mL, respectively. Tanaka et al.
For content uniformity testing of tablets, HPLC method was [43] compared the efficiency of resonant acoustic mixing (RAM)
used by Walash et al. [42] for the determination of naftazone com- technology for with ordinary mixing method for pharmaceutical
ponent. For this purpose, separation was performed using a Merck blending process. The mixing of theophylline powder and lactose
Hitachi L-7100 Chromatograph. A Nucleosil 100-5 phenyl column or magnesium oxide and lactose were carried out in a modified
and a mobile phase comprising methanol-sodium dihydrogen V-shaped mixing device and the APIs content uniformity in the

Fig. 3. (a) Schematic of two setups applied for capturing images during mixing process and (b) different required steps for homogeneity assessment of images (Reprinted
from Rosas and Blanco [50]).
 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549 1531

mixture was quantified by HPLC method (LC-10vp series, Shi- the material is lost by these methods because they the solid needs
madzu). The API and lactose powders were mixed for 0.5 h at 30 to be dissolved, hence, the studied material is not recoverable.
rpm by ordinary method, and the powders were mixed for 0.01 h Moreover, they are not classed as green method since they are
and 0.03 h at 100 G, 60 Hz by the resonant acoustic method. The wet-based techniques [47]. Therefore, other methods for analysing
coefficient of variation (CV) value of ordinary method was the lar- powder content are required which are briefly reviewed below.
gest and indicated non-uniformity. However, the CV values of res-
onant acoustic were smaller than ordinary mixing. The results
showed that the resonant acoustic method could obtain content 3. Dry techniques to assess powder blend uniformity
uniformity approximately 900 times more rapidly compared with
ordinary methods [43]. 3.1. Image analysis technique
Overall, as stated earlier UV–visible spectrophotometer method
has its own limitation regarding the concentration measurement of With image processing of a captured photo from a powder mix-
each constituent in the mixture separately [42]. However, HPLC is ture, it is possible to assess the uniformity of particles, providing
frequently used for the separation, identification and quantifica- that particles differ in colour. The 2D photo taken from the mixture
tion of different constituents in a mixture. Moreover, HPLC resolves of powders can be divided into several sections, where a compo-
the problem of the interference of active impurities and thus pro- nent of interest can be quantitatively analysed using a potential
vides more accurate analysis compared to the UV–visible spec- image processing method. Then the fraction distribution of the
trophotometer method [44,45]. Eraga et al. [46] analysed the desired component can be estimated in order to evaluate its homo-
content of ibuprofen ingredient, in ibuprofen tablets by UV–visible geneity throughout the mixture [48,49].
and HPLC method (Agilent Infinity 1260, Agilent Technologies Inc., Rosas and Blanco [50] applied image processing technique to
USA). The standard samples was prepared by dissolving 100 mg of evaluate the mixing homogeneity of coloured sand (SiO2) of differ-
ibuprofen powder reference standard in a 100 mL volumetric flask ent particle sizes in a blender. Two different set-ups were used to
with 0.1 mol/L NaOH solution. A 1 mL aliquot of the solution was take the images during the mixing processes. Set-up 1 was based
further diluted to 100 mL to give the desired concentration. The on at-line mode in which the samples were manually taken and
main samples were prepared by dissolving of 100 mg of crushed photographed (using a Video meter Lab camera) and set-up 2
ibuprofen tablets in water with similar procedure of standard sam- was based on a non-invasive image recording mode (Fig. 3a). Dif-
ple preparation. The results showed that the HPLC method is more ferent stages for the evaluation of the mixing quality of the cap-
sensitive and reliable assay for ibuprofen detection compare to tured photos are shown in Fig. 3b. The quality of the raw image
UV–visible spectroscopy and hence the HPLC technique suggested was first improved by some preliminary image pre-processing.
to be used for verification of the UV–visible method. After obtaining the segmented images using Matlab software, they
Beside several advantages of HPLC, there are several disadvan- were split into small subsections. Different mixing indices such as
tages for this analytical analysis technique. HPLC is time consum- modified Lacey index, ML (0 for totally segregated and 1 for totally
ing as it requires two steps of 1-finding the optimum conditions random), modified Poole index, Mp (tends to unity in totally ran-
(mobile phase composition, flowrate, injection volume and detec- dom images), homogeneity ratio defined by HL and HP% (100 for
tor type) and 2-performing sample analysis at optimum condi- near-random and homogeneous mixture) were measured to inves-
tions. Moreover, the complete package of HPLC facilities is tigate the quality of the powder mixtures. The homogeneity ratio
expensive. was defined as the ratio of the mixing index (Mimage) for an image
It should be noted that there are several issues with the above- and its randomized version (Mrandom_image). The mathematical cor-
mentioned traditional wet-based analytical methods. First, wet relation of different mixing indices has been presented in the study
analytical methods are deemed tedious and time-consuming. Also, by Rosas and Blanco [50] in detail. Moreover, a review of different

Fig. 4. The combined mixer/image-processing for on-line monitoring of continuous mixing (Reprinted from G. Ammarcha et al. [51]).
1532 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549

mixing indices used for the evaluation of powder mixing perfor-

mance has been provided by Poux et al. [10].
In a recent work, mixing of couscous particles of nearly spheri-
cal shape (with specific size ranges) in a continuous system has
been assessed by on-line image-processing during steady and
unsteady state regimes [51]. Using the on-line image-processing
technique described by Ammarcha et al. [51], homogeneity of mix-
tures was evaluated by capturing all images of particles at the dis-
charge of the continuous pilot-scale mixer (Fig. 4). The Charge
Coupled Device (CCD, one pixel width-5000 pixels length) camera
of the set-up was placed vertically with respect to the belt to film
the passing mixture at different times. Each pixel represented a 60
 60 lm2 surface on the belt which was much smaller than the
size of each particle. The pictures were captured by grouping 200
consecutive one-pixel-width lines (500  200 pixel2 image or 30 Fig. 5. The top view of the sliced samples after solidification of a homogenised
 1.2 cm2 surface). Quality of the mixtures was evaluated by the mixture (Reprinted from A. Realpe et al. [60]).

concept of coefficient of variation obtained for the key ingredient.

There have been numerous reported works on this topic and the other techniques. However, this technique cannot be easily
Table 1 summarises some recent works based on image analysis applied for the differentiation of components of similar colours.
technique for the evaluation of powder mixing performance. As the lighting conditions may not be always stable through the
In all the above-mentioned research studies, only the surfaces images, further background correction is also required for the anal-
of mixtures were scanned for homogeneity characterisation. Slic- ysis of the raw images. Moreover, unless slicing technique is used,
ing samples while preserving its properties, (e.g. using solidifica- image processing will provide only the surface properties of the
tion of a mixture by a binder such as gelatin as solidifier and mixture.
then refrigeration), could provide more content information from
a discretized mixture. Measurement of the composition of a mix-
3.2. Assessment of powder blend uniformity by bulk particle properties
ture, containing two coloured particles of granite stone, was per-
formed by Realpe et al. [60] using a combination of slicing
Bulk powder property can be a good indicator of powder homo-
technique and image processing. Homogeneity of the mixture
geneity status during the mixing process. With determination of
was assessed with image processing of the samples collected after
the bulk powder properties during blending, (e.g. monitoring pres-
slicing. After image processing of the captured pictures of the
sure drop, conductivity, interaction charges or thermal behaviour),
entire sliced samples (Fig. 5), mixing quality of the mixture could
it could be possible to evaluate the mixing performance, which is
be evaluated. This technique could be beneficial to a broad range
briefly reviewed below.
of industries such as metallurgy, pharmaceuticals, food processing
and ceramics. Based on slicing technique, the internal structure of
a mixture can be characterised off-line or at-line. In addition, the 3.2.1. Pressure drop method
best binder should be investigated in such a way that it does not Segregation of binary mixture of silica sand and silica gel was
destroy the chemical structure while the internal physical proper- investigated by Olivieri et al. [61] in a Fluidization system. The
ties are preserved. pressure was monitored at different locations along the bed using
Image analysis method is widely used for powder homogeneity pressure transducers to investigate the segregation pattern
evaluation due to its relative simplicity and low cost compared to (Fig. 6a). The plot of pressure gradient against the gas superficial

Table 1
Research works based on image analysis for powder blend uniformity evaluation.

Objectives Instrument for the image analysis Reference

Investigation of a double cone system for the mixing of glass beads of different colours: Camcorder at the rate of 40 frames per [52]
Mixing performance has been done by computing the concentration of each species in each discretized second
square
MATLAB software was used to determine the number of particles in discretized images
– Investigation of the component distribution of counterfeit Viagra and Cialis tablets using image processing: Video Spectral Comparator (A high [53]
Bhattacharyya distance was performed as a measure of mixture quality resolution VSC 5000)
– Evaluation of the segregation in pseudo-2D beds using a state-of-the-art digital colour camera: A digital image analysis technique (AT- [54]
Sample colour variance was performed as a measure of mixture quality 200 GE)
– Investigation of the component distribution of dropping glass beads with two different sizes using image Nikon D70 digital camera [55]
analysis method:
Image obtained by the camera was converted to an indexed image using MATLAB softwareTwo indices
(I\ and IH) were used to distinguish between the segregation and stratification mechanisms
– Segregation analysis of horizontally shaken monolayers of different binary mixtures on a vibrating tray: CCD Nikon camera [56]
The position of individual components was estimated using image processing, developed in MATLAB
software
– Evaluation of the mixing time of coloured particles in a rotary drum: Video camera at a frame rate of 25 fps [57]
RGB colour analysis was provided using image Processing Toolbox provided by MATLAB software
Degree of particle mixing was described using variance method
– Evaluation of the effect of paddle rotational speed on the mixing performance of the agitation process in an High speed camera (MotionPro HS-4) [58]
electrophotographic system:
The beads velocity field during agitation was used to describe the mixing extent
– Investigation of the mechanism of dead zone formation:The effect of different mixing parameters (size, Canon EOS 600D camera [59]
packing of particles, speed and shape of the mixer)
on the degree of mixing was evaluated
 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549 1533

samples was measured by a conductivity meter instrument

(SG78, Mettler Toledo). Using a conductivity calibration curve of
salt obtained from analysis of different known standard concentra-
tions, the unknown salt concentration of each sample could be
determined.
In another study, Shenoy et al. [64] compared the image pro-
cessing with salt conductivity method for the quality analysis of
food powder mixtures. The powder mixtures were first located
next to each other inside of the mixer, representing a fully segre-
gated medium. Using a bent spoon (a kind of thief probe sampler),
nine samples were extracted at five different mixing times. The
extracted samples were then positioned in small containers
(4 cm diameter and 1 cm height). Digital colour imaging (DCI)
and salt conductivity method were used for measuring the salt
concentration of each sample. It was demonstrated that the image
processing technique could be a better option for mixing evalua-
tion of multicomponent samples when the components had differ-
ent colours. Using conductivity method, mixing performance of
components could only be assessed in binary mixtures, containing
conductive component (e.g. salt) and other non-conductive food
components. For multicomponent food samples, only the mixing
quality of salt would be determined. On the other hand, it was
found that the image processing technique could not be a good
indicator of mixing performance of components in the case of
strongly segregated mixture of oregano and salt as it could not
effectively detect the salt particles of samples which were sieved
through the voids of oregano. In this case, conductivity method
was reported to be a better candidate as it is volume sensitive.
In general, the conductivity method is a simple and inexpensive
tool to study the powder mixture uniformity. This technique only
Fig. 6. (a) Schematic of fluidized bed equipped with pressure transducer (PT) and
needs a conductivity meter to measure the conductivity of samples
(b) pressure gradient versus gas superficial velocity at different heights of the bed
(Reprinted from Olivieri et al. [61]). in order to evaluate the mixing. The deficiency of this method is
that it can only be applied for the differentiation of a component
with high conductivity mixed with other low conductivity
velocity was used to discover the occurrence of segregation components.
(Fig. 6b). Velocity (U) between 2.2 and 9.1 cm/s demonstrated
the bubble formation and the onset of segregation of sand particles 3.2.3. Tribo-electrification method
at the bottom of the bed. The decreasing pattern of pressure gradi- Contact friction between particles causes a phenomenon called
ent in upper region of the bed at this range of velocity, was tribo-electrification or particle charging. The tribo-electrification
reported to be as a result of the progressive accumulation of low phenomenon is due to particle-particle and/or particle-surface
density silica-gel particles in this region. A uniform pattern could interactions, usually creating bi-polar charging, which allows the
be observed at U > 9.1 cm/s. creation of attractive or repulsive forces between individual parti-
In other work performed by Yudin et al. [62], different distribu- cles. Hao et al. [65] investigated the relationship between the elec-
tor configurations, such as perforated plate, circular edged slotted trostatic properties of different pharmaceutical powder
type (90°) and novel swirling type (45°), were used for the investi- formulations and the blending homogeneity in a V-blender. A Fara-
gation of the mixing performance of a fluidized bed system using day Cup was used for the measurement of the electrostatic charge
the pressure drop concept. The differential pressure readings of particles (Fig. 7a). The electrometer was zeroed after warming
across the column were logged using AZ InstrumentTM 82,012 dif- up when it was connected to the Faraday Cup. A glass beaker
ferential digital manometer (resolution of ±1 mmH2O and ±1.0% instead of metallic scoop/sampling thief was applied for transfer-
accuracy). Excellent particulate mixing was reported to be ring the samples to the container in order to mitigate the charge
achieved by the novel swirling distributor without the need to transfer from the operator and the container. Nearly 2 g of sample
apply mechanical rotation. was enough to entirely coat the floor of the Faraday Cup.
Using pressure drop information of a system, segregation pat- The plot of particle charge versus blending time for six different
terns can be predicted only qualitatively and the technique could samples taken from six different spots (three spots from each side)
be potentially used to identify trend changes. Robust and precise inside of the V-blender is shown in Fig. 7b. The measured charge
quantitative analysis is needed for content analysis and mixing quantities were close to each other at 4 and 12 min (highlighted
performance evaluation which will be described later. in red1) indicative of a better uniformity of extracted powder sam-
ples under these conditions. Based on the observed results, existence
of the cyclic nature of the charge change events which occurred with
3.2.2. Electrical conductivity method certain periodicity patterns upon continuous powder blending was
Shenoy et al. [63] developed an at-line powder uniformity reported.
assessment method based on measuring the conductivity of pow- Measurement of the bulk electric charge of samples may offer a
der samples. The effects of particle density, size and shape on the desirable and economical method for qualitatively assessing the
mixture uniformity in a lab-scale paddle mixer were investigated.
The binary mixtures containing salt and other food seasoning pow- 1
For interpretation of color in Fig. 7, the reader is referred to the web version of
ders were prepared and then the conductivity of different taken this article.
1534 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549

powder mixing uniformity. However, it could be an unreliable tool

for evaluating the mixing performance because the measurement
is also affected by the variations of humidity, temperature and
other environment factors [66].

3.2.4. Thermal analytical method

Effusivity has been used recently to evaluate the degree of
homogeneity of powder mixtures by analysing the extracted sam-
ples of different locations of a mixer. This technique is based on
evaluating the ability of powders to transfer heat, which depends
on the particle size, shape, porosity and the composition of a mix-
ture as well as the phase surrounding the particles. Effusivity has a
proportional relationship with thermal conductivity (k), heat
capacity (Cp) and density (q) of a substance as defined by Eq. (2).
The sensor for this technique must be able to track the temperature
changes within the powder samples in order to obtain the effusiv-
ity information [67].
qffiffiffiffiffiffiffiffiffiffiffi
Effusiv ity ¼ kqC P ð2Þ

Leonard et al. [68] investigated the uniformity of pharmaceuti-

cal powders at-line by the effusivity technique. Experiments were
carried out on three pharmaceutical binary mixtures (containing
active and excipient ingredients) mixed in a V-blender. The sam-
pling operation at different times was carried out using a thief
probe after stopping the blender and collecting three samples of
approximately 5 g from different corners of the blender. The con-
centration of active ingredient in the samples was estimated using
Fig. 7. (a) A simple schematic of Faraday Cup and (b) charge versus blend time standard calibration curves obtained from multiple blends of
plotted for Blend 1 (200 g batch, 98.75% Avicel PH200, 0.50% Cab-o-sil, 0.75% known active concentrations. Powder effusivity was measured
Magnesium Stearate, Reprinted from Hao et al. [65]). using a BT-01TM unit from Mathis Instruments (Fredericton, NB,
Canada, Fig. 8a), working based on the hot wire technology and

Fig. 8. (a) Schematic of BT-01TM unit and (b) evaluation of Acetaminophen Fig. 9. (a) Average enthalpy changes of top, middle and bottom samples in the mixer
homogeneity at three sampling positions of a V-blender using effusivity measure- and (b) % RSD of top, middle and bottom samples of microcrystalline cellulose-
ment (Reprinted from Leonard et al. [68]). Atenolol blend at different time points (Reprinted from Bharvada et al. [70]).
 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549 1535

Fig. 10. (a) The schematic of mixing process, (b) images taken at different times of rotation and (c) evaluation of the mixing performance by normalized frequency
distribution (Reprinted from Liu et al. [73]).

tracking temperature changes within a sample during a given time reached a plateau after a period between 6 and 8 min, indicating
interval. The uniformity patterns of active ingredient of different the end-point of mixing. The estimated error of this technique
blends can be observed in Fig. 8b. According to Fig. 8b, all curves was reported to be ±3.4% which was higher than that obtained
1536 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549

from UV–visible spectrophotometry (±0.7%). The work reports that using a numerical index, S0 , based on the sphericity of particles,
more accurate analysis can be performed by UV–visible spec- estimated by following equations:
trophotometry, however, it is destructive, time consuming and
PNl
not applicable as an in-line measurement mode. On the other hand,  sÞ2
i¼1 ðsi
effusivity technique could potentially be used as an effective and r2 ¼ ð3Þ
ðNl  1Þ
fast tool for determining the end-point time of powder blending
processes. PN P
Differential Scanning Calorimetry (DSC) is another method for s ¼ i¼1 si
ð4Þ
the evaluation of powder uniformity which works based on heat Np
transfer phenomena. This method uses enthalpy values to predict
the sample content and offers a simple and cost-effective means r
S0 ¼ ð5Þ
of monitoring powder blending [69]. Bharvada et al. [70] evaluated s
the mixing of binary pharmaceutical mixtures, containing Micro-
crystalline cellulose (MCC) and active component of Atenolol, in where s, Np, si and Nl are the mean sphericity measured over the
a high shear mixer using the DSC technique. For the analysis of totality of the particles in all three levels of the container, particles
the mixing performance, different samples (equal to 3 the quan- in the upper, middle and lower levels in total, the ith value of s
tity of drug) were extracted from different locations of the mixer which represents a sphericity of the particle and particles in one
(top, middle and bottom) using a sampling thief. The enthalpy of level, respectively.
samples was measured by an Indium calibrated Auto DSC (TA Components were displayed with different colours depending
Instruments, USA). The relative standard deviation (RSD) of on the sphericity of the particles. The increased degree of coinci-
unknown samples (Fig. 9a and b) extracted from several positions dence of the normalized frequency distributions with vibration
of the blender was estimated using standard calibration curve, time obtained for the upper, middle and lower levels indicated a
obtained from analysing several known active component fractions better uniformity at higher vibration times (Fig. 10c).
with known enthalpies. The RSD was found to be low after 15 min,
indicating the optimum time of mixing for the binary system of
MCC-Atenolol blend. From the comparison of the results of DSC
technique with HPLC method, it was concluded that the relative
standard deviation measured by DSC was higher than that
obtained using HPLC for lower levels of drug, i.e. 0.5%, 1% and 2%.
However, the results of DSC and HPLC techniques were nearly sim-
ilar for the concentrations above 5%.
The above-mentioned methods based on thermal characterisa-
tion are convenient and simple for the evaluation of powder uni-
formity. However, a sampling operation, using thief probe is
needed for the extraction and analysis of thermal behaviour of
samples which might disturb the powder bed, and therefore
results may be biased [10,15]. Although the uniformity assessment
results derived from thermal characterisation methods shown ear-
lier seem to be reliable and accurate, these methods should be
applied to more powder systems with complex thermal behaviour.

3.3. Tomographic techniques

There are several tomographic techniques [71] such as X-ray,

positron emission particle tracking, electrical capacitance and
magnetic resonance imaging, used for the evaluation of powder
blend uniformity which are reviewed in this section.

3.3.1. X-ray microtomographic method

X-ray computed microtomography (mCT) is a non-invasive tech-
nique performed for the evaluation of powder homogeneity which
could provide high resolution images (typically 50 mm or less) [72].
By projecting an X-ray beam through the material and the mea-
surement of the energy debilitation of the beam received on a
detector, a three-dimensional structure of an object could be con-
structed. Liu et al. [73] used X-ray microtomography to evaluate
the mixing and segregation of a binary system containing spherical
and non-spherical particles at different times of vibration inside a
cylindrical container. Since the size of the container (9 mm in
diameter and 28 mm in height) was too large for the image acqui-
sition process, different samples at the upper, middle and lower
levels of the cylindrical container were exposed to the X-ray light
beam (Fig. 10a) and the total projected images were reconstructed
at different vibration times (Fig. 10b). The pixel size, the exposure Fig. 11. (a) Comparison of dispersion coefficients, Dx and Dr, as a function of
time and the sample-to-detector distance were 13.0 lm, 1.0 s and rotation speed and (b) occupancy plots for 20 min measurement time (Reprinted
25.0 cm, respectively. The uniformity of mixture was analysed from Marigo et al. [81]).
 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549 1537

Surface imaging tools such as Scanning Electron Microscope movement of the tracer closely follows the components in the
(SEM) are only able to scan the outer surface of objects and there- mixer. Marigo et al. [81] studied the mixing behaviour of glass
fore they cannot be applied for full-scale evaluation of powder spheres inside a cylindrical Turbula mixer with the use of PEPT
mixtures. On the other hand, mCT has a great potential for deep technique. Glass particle tracer as well as two detectors for track-
understanding of the structural properties of particulate materials ing the gamma rays of the tracer were applied. By the measure-
[74,75]. Precise data analysis could be provided from rendered high ment of axial and radial diffusion of the tracer (Eqs. (6) and (7),
resolution images using this technique. However, material with respectively), the mixing efficiency was evaluated for different
similar structural properties cannot be easily differentiated using operating conditions in axial and radial directions (Fig. 11a).
this technique. Also, this method is an expensive tool for the eval-
0
1 NX1 2
uation of powder homogeneity. ðxkþ1  xk Þ
Dx ¼ 0 ð6Þ
N  1 k¼1 ðt kþ1  t k Þ
3.3.2. Positron emission particle tracking (PEPT) method
Positron emission particle tracking tomography is an imaging 0
1 NX 1 kþ1 2
technique that produces a three-dimensional image of a process ðr  rk Þ
Dr ¼ 0 ð7Þ
by detecting pairs of gamma rays emitted by a positron-emitting N  1 k¼1 ðtkþ1  t k Þ
radio nuclide (tracer) into the system [76–79]. The spatial accuracy
in a range of 0.5 mm (observed at 500 times per second) was where xk and xk+1, N 0 , rk and rk+1 are the axial positions of the parti-
reported for particle velocity of 1 m/s [80]. The accuracy reduces cle at time tk and at time tk+1, the total shaft rotation and the radial
as particle velocity is increased because the particle moves further positions of the particle at time tk and at time tk+1, respectively.
and therefore more detectors are needed to spot the tracer [80].
Also, the accuracy of this method depends on the measurement
conditions, specifically the mass of material between the tracer (a)
particle and the detectors. The size and density of the tracer are
considered to be approximately the same as the components inside
the mixer. Therefore, it would be possible to assume that the

(a)

(b)
Upper funnel Probe

(b)
Lower funnel

Fig. 12. (a) Schematic of the experimental rig comprising the initial mixture (1),
vibrating channel (2), upper funnel (3), sensors (4), static mixer (5), lower funnel Fig. 13. (a) Schematic of fluidization experiment and the structure of capacitance
(6), belt conveyor (7) and capacimeter (8), (b) discharge profiles through a funnel probe and (b) fraction distribution of sand in polypropylene plastic and quartz sand
(Reprinted from Ehrhardt et al. [85]). mixture (Reprinted from Huang et al. [86]).
1538 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549

Fig. 14. MRI slices through a cylindrical powder sample at different time steps (Reprinted from Hardy et al. [88]).

The tracer motion was further discussed by occupancy con- speed of cylindrical Turbula mixer up to 46 rpm, a clear core-
cept which is defined as the ratio of time that the tracer spends shell pattern was observed in the transverse and axial direction
at a given position to the total tracking time. By increasing the (Fig. 11b). This event represented the tendency of the tracer par-
 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549 1539

ticle to stay in two core regions of the bed in the axial direction PEPT is a non-invasive technique for evaluating the flow pro-
which hindered the tracer particle to cross the middle point. This cesses by tracking the motion of a radioactive tracer particle.
core-shell pattern represented the inadequate mixing operation, This method allows looking through opaque systems without
leading to a decrease in the mixing efficiency in the axial interfering with flow [83]. However, erroneous locations may
direction. also be calculated using this technique due to scattering of
Gundogdu [82] described a technique based on a clustering gamma rays. In addition, in order to obtain reliable data long
method to track multiple particles rather than single tracer parti- run experiments have to be carried out, allowing the tracer (s)
cle, enabling exploration of more comparative information about occurrence in different regions of a mixer, except in the stagnant
the mixing in a system. The study of Gundogdu offers a unique sections [84].
solution to track more than one particle. He found that 18F radioac-
tive particles provided better results than 22Na particles. All exper- 3.3.3. Electrical capacitance tomography
iments of Gundogdu were carried out in open air and calculations Electrical capacitance tomography (ECT) provides a non-
were performed in off-line mode. It should be noted here that more intrusive cross-sectional view of a stream which can be obtained
intensive investigation of this method is needed due to the com- by the measurement of variations in dielectric properties (relative
plexity of its procedure [82]. permittivity) of the materials inside the vessel. The main advan-

Table 2
Summary of different tomographic techniques for powder mixing assessment.

No. Objective Materials Technique Reference

Ò
1 Mixing investigation of binary mixtures of free flowing particles in a Turbula Sugar beads MRI [93]
mixer
2 Characterization of the kinematics of mixing/size segregation of dried binary Poppy seeds and sugar beads MRI [94]
mixtures
3 Investigation of the solid concentration changes during granular flow in a Different type of sands ECT [95]
cylindrical model silo
4 Evaluation of the size segregation in a cold fluidized bed at atmospheric Spherical glass particles ECT [96]
pressure
5 Evaluation of the mixing and segregation in a fluidized bed Ground walnut shell particles and glass mCT [97]
beads
6 Assessment of the mixing in an industrially static mixer geometry Glycerol Combined PEPT and [98]
MRI
7 Estimation of the spatial distribution of components throughout the tablets Potassium chloride, spray-dried lactose mCT [99]
8 Investigation of the axial mixing and segregation of fuel at different operational Tracer particles with solid density Magnetic particle [100]
conditions of bubbling fluidized bed representing biomass char tracking (MPT)

Table 3
Current research studies investigating the powder blend uniformity using NIR system.

No. Objective Instrument Blending materials Reference

1 End-point determination of a pharmaceutical blending Sentronic GmbH NIR API (Active Pharmaceutical Ingredient), [103]
spectrometer crospovidone, lactose, and microcrystalline
cellulose
2 Application of Near Infrared hyperspectral imaging for the evaluation of the OPOTEK Inc. NIR Tolmetin sodium dehydrate, anhydrous [104]
spatial distribution of drugs in tablets system lactose, magnesium stearate
3 Evaluation of the effects of granule cohesion and size on the in-homogeneity FT-NIR spectrometer Aspirin, lactose, microcrystalline cellulose, [105]
tendency of pharmaceutical powders using NIR spectroscopy, bench scale magnesium stearate
sifting and fluidization segregation testers
4 Application of an in-line NIR spectroscopic method for the measurement of Visio NIR Acetaminophen, lactose and magnesium [106]
drug content of tablets during a continuous tableting process stearate
5 Implementation of a multi-point fiber optic based on NIR set-up to control Multipoint NIR Acetaminophen, magnesium stearate and [107]
the drug concentration at the discharge of a continuous mixer, comparison system Avicel PH-200
between in-line and off-line measurements
6 Investigation of an in-line NIR spectroscopy for content uniformity analysis NIR spectrometer API, EX1 (excipient1) and EX2 [108]
in an industrial tablet press
7 Non-contact monitoring of the blending process in pharmaceutical powder NIR spectrometer APAP, Avicel, Lactose [109]
blends
8 Evaluation of powder blend uniformity of a cohesive model formulation Integrated Corona Naproxen (API), microcrystalline cellulose, [110]
using a real-time NIR spectroscopy and an at-line NIR chemical imaging NIR spectrometer croscarmellose and free-flowing mannitol
approach from Zeiss
9 Rapid determination of four structurally similar active pharmaceutical A Bruker multi- Four APIs, microcrystalline cellulose as [111]
ingredients using a developed at-line Near Infrared method purpose FT-NIR excipient
analyzer
10 Investigation of the mixing performance of a laboratory-scale Resonant Transform Near- Micronized acetaminophen, granulated [112]
AcousticÒ Mixer (LabRAM). The effect of process parameters (fill level, Infrared (FT-NIR) acetaminophen and caffeine
acceleration, and blending time) was considered spectrometer
11 Evaluation of the blend mixing in a continuous/dynamic drug product Fourier transform Acetaminophen, microcrystallinecellulose, [113]
manufacturing process spectrometer mannitol, croscarmel-lose sodium,
magnesium stearate
12 Investigation of NIR spectroscopy for overall quantitative analysis of all Bruker Multi Purpose Washing powder ingredients [114]
ingredients of laundry detergents Analizer
13 Evaluation of NIR spectroscopy technique for the quality control of semi-solid NIR spectrometer Nonionic hydrophilic cream, Salicylic acid [115]
pharmaceutical formulations and Erythromycin
1540 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549

Fig. 16. (a) Experimental of an in-line NIR setup with a four-bladed mixer
connected to a controllable mixing device and (b) blending plots for the mixer,
light grey represents acetyl salicylic acid (ASA) and dark grey shows a-lactose
Monohydrate (Reprinted from Koller et al. [119]).

to estimate the capacity of particles. It is found that in the lower

funnel, silicon carbide was mixed with sugar at the beginning of
the experiment, whereas segregation was observed after a while
(Fig. 12b). This was reported to be as a result of density segregation
of silicon carbide.
In new research done by Huang et al. [86], a new measurement
method for the mixing of binary mixtures was developed based on
capacitance probe in a bubbling fluidized bed. A capacitance mea-
surement instrument (MTI Accumeasure) was used to measure the
signals of capacitance probes from different measurement points
of the fluidized bed (Fig. 13a). A linear relationship between the
signal of the probe and fraction distribution of solids in binary mix-
tures was considered (Eqs. (8)–(10)).
C1 þ C2 þ Ca ¼ 1 ð8Þ

Fig. 15. (a) Concentration profiles of segregation for a mixture of three bird seeds C1  v 1 þ C2  v 2 þ Ca  v a ¼ v ð9Þ
and (b) spectra of pure components and mixture and (c) segregation intensity as a
function of viewport size (Reprinted from Johanson [116]).
v v 2 ð1  C a Þ  C a v a
C1 ¼ þ ð10Þ
tage of this technique is that it could provide a cross-sectional view v1  v2 v1  v2
of a stream, containing liquid, solid and/or gas, in a non-intrusive where v, v1, v2, va, C1, C2 and Ca are the measured probe voltage sig-
way. However, as compared to X-ray, c-ray, as well as Magnetic nal, voltage signal corresponding to pure substance of solid 1, solid
Resonance Imaging (MRI), this low-resolution technique (spatial 2, air, the volume fraction of solid 1, solid 2 and air in the mixtures,
resolution typically 3–10% of a pipe diameter) has limited applica- respectively.
tion to powder mixtures that require an analysis at very fine scale. The fraction of quartz sand particles at different bed heights of a
Ehrhardt et al. [85] used an electrical capacitance sensor consisting binary mixture of polypropylene plastic and quartz particles is
of two copper electrodes placed around a glass tube to assess the shown in Fig. 13b. It can be observed that the fraction of quartz
segregation problems of discharging silicon carbide (SiC) and sugar in the lower part of fluidized bed (30 cm<) is high, while a decrease
mixture (sugar at the bottom and SiC at the top) through two fun- at heights above 30 cm is observed. A successful segregation mea-
nels (Fig. 12a). The sensor was connected to a capacitance meter surement of binary mixture in the fluidized bed using capacitance
and further to a computer through an analogue/digital converter probe was confirmed using this method.
 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549 1541

nX
ECT shows great potential in many harsh conditions such as comp

high-pressure or high temperature. However, this technique has FSmix ðkÞ ¼ ðxi  FSi ðkÞÞ ð11Þ
i¼1
a relatively low spatial resolution. Moreover, it is only applicable
for the materials with noticeable variations in dielectric properties.
nX
wav e

Error ¼ ðFSmix ðkk Þ  F mix ðkk ÞÞ2 ð12Þ

3.3.4. Magnetic resonance imaging tomography k¼1
Magnetic resonance imaging (MRI) is a tomography system that
provides a non-invasive 3D images of a mixture using strong mag- where FSmix(k), FSi (k) and xi are average intensity of the mixture,
netic field generated by the surrounding powerful magnets [87]. average intensity of pure components and fraction of components,
The signal intensity of one voxel in MRI is proportional to the respectively.
atomic nuclei number of observable elements such as hydrogen,
phosphorus or fluorine. Hardy et al. [88] studied the mixing perfor-
mance of oil-filled melamine microcapsules (hence visible to MRI)
and solid melamine spheres in a cylindrical container using an MRI
method. A tomographic system (Bruker Avance 200 SWB, Bruker
Biospin MRI GmbH, Ettlingen, Germany) was used to track the vis-
ible particle signal intensity. A 1 cm3 mixture volume was imaged
with an isotropic spatial resolution of 235 lm. The effectiveness of
the MRI technique for quantitative characterization of fine powder
mixtures (size range of about 10 lm) has been shown in this arti-
cle. Three orthogonal slices through the 3D mixture data set at
each mixing step are shown in Fig. 14. Bright and dark regions rep-
resent the presence of MRI active particles and non-visible parti-
cles, respectively. It is evident from Fig. 14 that the uniformity
was achieved at higher time steps, referring to the end-point of
mixing identified using the MIR method.
The limitation of MRI method lies in the fact that the particles
mainly need to be coated by substances, such as oil, to be visible
by MRI. Coating of particles may change the powder flow charac-
teristic which could influence their mixing behaviour [89–91].
Also, the spatial resolution is less than the mCT method [92]. A sum-
mary of research work based on tomographic techniques for pow-
der homogeneity assessment is presented in Table 2.

3.4. Spectroscopic techniques

Spectroscopic techniques have already found a wide range of

applications in industrial sectors [101]. In particular, component
composition of samples could be determined using their spectral
information and used to evaluate the mixing performance. In this
regard, different spectroscopic methods along with their function-
alities are explained in the following sections.

3.4.1. Near-infrared spectroscopy

NIR spectroscopy is a molecular vibrational spectroscopic
method for the detection of vibrational transitions in the molecules
that can be performed either by diffuse reflectance or transmission
mode. The intensity ratio of the scattered light from the sample
compared to the light reflected from a reference surface can be
measured by the diffuse reflectance mode. On the other hand,
decrease by radiation intensity can be detected by the transmission
mode when radiation passes through the sample [102].
There are numerous research works investigating the powder
uniformity using Near-Infrared (NIR) spectroscopy (Table 3).
In order to monitor the mixing of powders, a segregation tester
device was developed by Johanson [116] which works based on a
relationship between the mixture NIR spectral intensity and the
spectral intensity of pure components. Based on a similar concept,
Asachi et al. [117] and Oka et al. [118] evaluated the mixing perfor-
mance of powders, where component fractions are determined by
minimization of the difference between the computed intensity Fig. 17. (a) Schematic of multi-probe spectroscopy setup, (b) dependence of the
curve of the mixture (Eq. (11)) and one measured using a spec- sample volume size to the number of accumulated spectra and the moving speed in
troscopy tool as shown in Eq. (12) [117]. The distribution of com- front of the sensor window (f = 1 is the tip speed of the blade, f = 0.5 is half of this
ponent fractions through the mixture could then be used for the speed) and (c) continuous blend monitoring of powder streams using transmission
NIR (Reprinted from Scheibelhofer et al. [121] and Reprinted from Alam et al.
evaluation of powder mixing uniformity.
[122]).
1542 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549

Fig. 18. IR and NIR absorption, the Raman effect and fluorescence (Reprinted from De Beer et al. [25]).

(a) content level enzyme granules (1.85% w/w) in the mixture of

Incre
asing
washing powders; the percentage error for the quantification of
mass segregation of low content level enzyme in a heap of powder mix-
tures was reported to be around 10%.
Concentration profiles of three types of bird seed in a pile of
powders, estimated using the segregation tester, are shown in
Fig. 15a. Spectra of the pure components and those obtained for
the mixture are shown in Fig. 15b. Good agreement between the
actual values and the predicted values is observed in Fig. 15b.
The smallest representative size (view port) of each sample, was
deduced by plotting the segregation intensity factors as a function
of view port size (Fig. 15c). According to Fig. 15c, a view port size of
about 4500 lm was suitable for sand with an average particle size
of 1500 lm. The expected error of this method was reported to be
within 7% and 0.5% for a badly segregating and moderately segre-
gating materials, respectively. Moreover, suitable range of particle
(b)
size for this device was reported to be between up to 3 mm.
Koller et al. [119] applied a FT-NIR spectroscopy with a fiber
optical reflection probe in a four- bladed mixer to assess the mix-
ing process in pharmaceutical powder mixtures. The NIR spectra
were recorded using a PerkinElmer FT-NIR400 Spectrometer. A
penetration depth of approximately 0.3–0.5 mm and a spot diam-
eter of 4 mm (sampling volume of approximately 6 mm3) was
reported for the studied probe. Fiber optical probe of the spectrom-
eter was introduced in direct contact with the powder bed during
mixing process (Fig. 16a). The pure and mixture spectra were anal-
ysed by Partial Least Squares Regression (PLSR) to estimate the
concentration change in the mixtures (Fig. 16b). It is shown that
locating a NIR probe in powder beds is a suitable method to deter-
mine the end-point time of blending. The Root-Mean-Square Error
of Prediction (RMSEP) was reported to be in the order of 2–3% for
the API content.
Fig. 19. (a) Schematic of Raman probe and sample set-up and (b) Raman probe Other configurations of NIR set-ups have also been applied to
mixing profiles at 1606 cm1 for the case of the addition of different amounts of monitor greater quantities of samples. By placing several NIR
aspirin to Avicel (Reprinted from Allan et al. [125]).
probes at several positions of a blend, it is possible to monitor
the entire blend [32,120]. Scheibelhofer et al. [121] applied multi-
ple NIR probes, connected to a Fourier-transform NIR spectrome-
Asachi et al. [117] investigated the effect of using different pre- ter, to quasi-simultaneously investigate the pharmaceutical blend
processing methods on the measured component fractions accord- uniformity at multiple positions (Fig. 17a). In contrast to thief sam-
ing to the above-mentioned approach; scatter correction and the pling, the samples are analysed here by applying probes. In this
combined smoothing-derivatives. The second derivative of the work, it is shown that the size of sample volume (defined as the
smoothing technique of Norris-Williams method was reported to number of particles contained in the sample) was dependent on
be the best pre-processing method for the quantification of low the particle speed in front of the sensor window and the number
 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549 1543

Fig. 20. (a) Dispersive macro-Raman Set-up, (1) UV fused silica broadband plate splitter, (2) power meter, (3) plano-convex lens, (4) aluminum block, (5) square pin stub
holder, (6, 7) motorized X-Y-Z stage, (8) illuminating lamp, (9) monitoring camera, (10) microscope, (11) narrow-band notch filters, (12) Nikon lenses, (13) manual slit, (14)
filters. (b) different sampling methods: single spot (I), scanning across the cavity (II), scanning along the groove (III) and (c) cumulative distributions of the measured leucine
mass fraction during mechanical mixing (Reprinted from Wang et al. [126]).

of accumulated spectra (Fig. 17b). The measured volume, at opti- using transmission NIR is illustrated in Fig. 17c. As shown in
mum case, was reported to be about 16 mm3 which was less than Fig. 17c, a tube connected to the powder feeder was used to deliver
the final dosage form (75 mm3). Standard deviation of the model the samples to the NIR spectrometer, where the spectra could be
predictions was estimated to be about ±5% using this set-up. In a recorded in a non-contact manner.
recent work, the potential of NIR spectroscopy for continuous NIR offers a fast chemical imaging of multicomponent mixtures
monitoring of powder flow was demonstrated by Alam et al. and continuous monitoring of samples at different modes of at-
[122]. The information from the bulk phase of the powder stream line, in-line and/or on-line [123]. However, it is difficult to obtain
could be obtained using transmission NIR spectroscopy (NIR spec- a very high-resolution visualization to the internal field of particu-
trometer with a diode array detector). According to the analysis, late systems using this method. Furthermore, overlapped spectra
the proposed NIR device could detect the NIR signal intensity of of different components could pose difficulty and challenge when
powder beds up to 5 mm in thickness. The analysed volume of each using the NIR technique to assess homogeneity of a mixture.
sample using this transmission NIR method was estimated to be
around 0.25 mm3. Quantitative determination of API concentration 3.4.2. Raman spectroscopy
in the developed continuous stream of powder was carried out at The irradiation of materials cause different phenomena includ-
different process parameters (such as powder flow rate and tube ing scattering, absorption and fluorescence (Fig. 18). The Raman
angle). A schematic of on-line monitoring of continuous stream effect is a scattering process that alters the frequency of an incom-
1544 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549

ing monochromatic light beam, mainly from a laser in the visible, of particles from the acoustic emission signals created from parti-
Near-Infrared, or near Ultraviolet range [124]. cle–particle collisions. Allan et al. [132] investigated the blend pro-
Allan et al. [125] used an on-line Kaiser Raman spectrometer cess monitoring of aspirin or Avicel added to Avicel particles in a
with a non-contact optic to obtain the spectra of powder mixture scaled-down convective mixer with the use of powder acoustic
in a high shear convective blender (Fig. 19a). A sampling depth emission (AE). Mixing profiles which were produced from the mea-
of over 3.5 mm and an estimated sampling volume of about surement of AE spectra in this system, are represented in Fig. 21a.
11.095 cm3 (equated to a mass of 4.99 g) was reported for the mea-
surement of aspirin in Avicel using this system. The mixing end-
point of different amounts of aspirin and Avicel was investigated
using this technique (Fig. 19b). As shown in Fig. 19b, by increasing
the amount of aspirin, more time was needed to achieve a homo-
geneous state. The detection limit of aspirin in Avicel using the
proposed Raman system (2nd derivative at 1606 cm1) was
reported to be around 1.1% (w/w).
In a recent work, Wang et al. [126] developed a custom-
designed macro-Raman based system for homogeneity analysis
of multi-component bulk of pharmaceutical powders (Fig. 20a).
The studied spectroscopy system was equipped with a motorized
translational sample stage (Fig. 20a). For the quantitative analysis
of component fraction, a correlation based on spectral information
was applied as follows:

Yi Ii
¼ cij  ð13Þ
Yj Ij

where I and Y are the spectral intensity of the component i or j at all

wavenumbers to the measured raw spectrum of the mixture and
the mass fraction of components, respectively.
In the first part of the analysis, a direct relationship between the
errors of the measured compositions and the analysed sample vol-
ume and sampling methods was demonstrated. Larger sample vol-
ume scanning, across cavity or groove (Fig. 20b), was shown to be
more suitable for bulk composition analysis of inhomogeneous
samples. Therefore, more representative bulk compositions of
inhomogeneous samples were analysed by the cavity or groove
scanning methods. A large error was reported for the single spot
sampling method (sample volume of 0.4 lL) which was mostly
due to its relatively smaller sample volume. Binary mixtures of leu-
cine and mannitol with equal masses were mixed for different
times using a wrist action shaking machine. From analysis of sev-
Fig. 21. (a) Acoustic emission mixing profiles and (b) mixing profiles for the
eral samples at different times, a cumulative distribution of mass addition of 30 g aspirin to 75 g Avicel (Reprinted from Allan et al. [125,132]).
fraction of leucine was obtained as shown in (Fig. 20c). According
to Fig. 20c, mixing deteriorates in the binary mixture of leucine
and mannitol during blending (as standard deviation of the mea-
surements was increased by time).
Raman imaging has less overlapped spectra and higher spatial
resolution as compared to the NIR and therefore it has a better sen-
sitivity to detect minor components [127,128]. Thus, Raman elim-
inates the limitation of NIR in which the overlapped spectra make
it difficult to perform low dose quantitative analysis. However, this
technique is more expensive than NIR. In addition, substantial
interferences in Raman spectra can be produced by fluorescence
when the molecule is excited to an elevated state [25,47,129].
For the elimination of the fluorescence contribution, some
researchers proposed stimulated Raman scattering (SRS)
[25,130,131]. The advantage of SRS is that it removes the non-
resonant background via heterodyne detection. Therefore, SRS is
free of non-resonant background and fluorescence contribution
and it allows selective imaging of the molecules.

3.4.3. Acoustic emission spectrometry

The impact of particles on the inner surface of a vessel, which
generates the acoustic emission, can be used to non-invasively
monitor the particulate process. The attachment of a transducer
to the outside of a vessel is the most common method to detect Fig. 22. (a) LIF instrument and (b) triamterene concentration obtained using LIF
acoustic emission of particles. It is also possible to estimate the size and UV analysis (Reprinted from Domike et al. [135]).
 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549 1545

Table 4
Powder uniformity assessment techniques with their advantages and disadvantages.

Method Advantage Disadvantage

UV–visible absorbance Simple to implement: Spectrophotometers are fairly – Material is not recoverable using this method as it is dissolved in a solution
spectrophotometry straightforward instruments with very few moving (wet technique)
parts – The sensitivity of a spectrophotometer is often inadequate at low
– Fast analysis concentrations
– Low cost maintenance – Absorption results can be influenced by other parameters, e.g. impurities, tem-
perature and pH, leading to inaccurate results
– Not a green method, as it produces liquid waste
– Separate measurement of multi component fractions is not applicable
– Error due to off-line sampling and disturbance of the powder bed (invasive)
High-Performance Accurate and sensitive for multi component – Costly and time consuming
Liquid quantification – Material is not recoverable using this method as it is dissolved in a solution
Chromatography (wet technique)
– Not a green method, as it produces liquid waste
– Error due to off-line sampling and disturbance of the powder bed (invasive)
Image analysis – Simple to use – No information on 3D structure of powder mixtures (this method is surface
– Green method (dry technique) sensitive). Nevertheless, if coupled with the slicing technique, the internal
– Low cost structural analysis of mixture could be provided
– Possible for blend uniformity assessment of multi- – Cannot be used for uniformity assessment of powder components with the
component mixtures (providing that particles differ similar colours
in colour) – Because of lighting conditions, the raw images need background correction
– Sample analysis can be performed non-intrusively
Electrical conductivity – Providing information on the whole volume of – Only applicable for blend uniformity assessment of a conductive material in
method powder mixture the mixture of powders
– Easy to use – Cannot be used for multicomponent mixing assessment
– Low cost
– Good for uniformity analysis of conductive
materials
Tribo-electrification – Providing information on the whole powder – Error due to off-line sampling and disturbance of the powder bed (invasive)
method mixtures – Unreliable tool because it may be exposed to the variations in humidity, tem-
– Easy to use perature and other environmental factors
– Economic – Cannot be used for multicomponent mixing assessment
Thermal analytical – Easy to use and Low cost – Error due to off-line sampling and disturbance of the powder bed (invasive)
method – Providing information on the whole powder – Measurement must be performed quickly after sampling to reduce the heat
mixture exchange effect
– Good for the determination of the end-point of – This method should be checked on several powder systems with complex ther-
mixing mal event to judge whether it is suitable for powder uniformity assessment in
general
X-ray – Providing the three-dimensional structure of the – Expensive
microtomographic objects non-invasively – Cannot be readily used for multicomponent mixing assessment
method – Higher special resolution as compared to other – Safety issues due to X-rays
tomographic techniques – Material with similar structural properties cannot be differentiated easily
Positron emission – Produces a three-dimensional image of a process – Expensive
particle tracking – Non-invasive uniformity evaluation of powders is – This method mainly provides the tracking of a single particle as a representa-
achievable tive of other particles
– Cannot be used for multicomponent mixing assessment
– Erroneous locations could be spotted using this technique due to scattering of
gamma rays
Electrical capacitance – Providing the 3D cross-sectional view of a stream – Expensive
tomography non-intrusively – Hard to detect the exact component distribution due to low spatial resolution
– Applicable in many harsh conditions such as high- – This method is applicable just for powders with noticeable variation in permit-
pressure or high temperature tivity or dielectric constant
– Cannot be used for multicomponent mixing assessment
Magnetic resonance – Providing information on 3D powder mixtures by – Expensive
imaging constructing cross-sectional images of raw data – Many particles get visible by magnetic field via getting coated by detectable
tomography non-invasively substances. This could change the surface properties of particles such as
flowability
– Cannot be readily for multicomponent mixing assessment
Near-infrared – Fast chemical imaging technique – Overlapped spectra cause the minor component detection difficult
spectroscopy – On-line/continuous monitoring of samples is possi- – Light penetration to the sample is limited and therefore composition charac-
ble using this technique terisation of the bulk of powders is hard to be achieved
– Applicable for multicomponent mixing assessment
– Reasonable in price
Raman spectroscopy – Less overlapped spectra resulting in precise minor – Sometimes, interference by florescence may cause erroneous results
component detection (high spatial resolution) – Expensive
– Applicable for multicomponent mixing assessment
Acoustic emission – Non-invasive/on-line particulate blend monitoring – Representative for changes in powder composition at wall, rather than in the
spectrometry – Economic and easy to use bulk of materials
– For on-line monitoring, no need of optically trans-
parent window in the process

(continued on next page)

1546 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549

Table 4 (continued)

Method Advantage Disadvantage

Fluorescence – Low cost and fast in data acquisition – Sensitive to fluctuations in pH and temperature.
spectroscopy – Sensitivity of the method is higher than the absorp- – Applicability of the method depends on the strength of fluorescence of the
tion spectroscopy studied component as compared to the fluorescence of other components
within the mixture

No change could be observed in the mixing profile when Avicel purpose, the data obtained from LIF sensor placed on the surface
was added to Avicel particles. However, the AE signal increases of collected samples containing fluorescent API was analysed. A
in the case of the addition of aspirin, and after a while it reaches stainless-steel grain type sampling thief was used for extracting
a plateau as the mixture becomes homogeneous (about 700 s). In the samples. The excitation wavelength of the fluorescent API
another part of this study, the effect of impeller speed on the AE was reported to be around 330–360 nm. Five LIF measurements
response was investigated. It was found that at low impeller were made on each sample. Using fluorescence reference stan-
speeds (<50 rpm), the movement and velocity of powders near dards, the calibration of LIF sensor to estimate the final API concen-
the glass wall was not sufficient, thus a small acoustic signal was tration was performed. The results showed that the blend met the
produced. %RSD  4% criterion at 2–20 min, indicating to an acceptable mix-
Following the previous research, Allan et al. [125] carried out a ing status during this blending interval. A linear relationship (R2
comparison between different techniques of NIR, acoustic emission = 0.97, p < 0.0001) was observed between LIF%RSD and HPLC%
and Raman spectroscopy for obtaining mixing profiles. The end- RSD, suggesting that LIF can be successfully used to qualitatively
point identified for the blending process as well as the trends of evaluate the blend homogeneity.
the mixing profiles were similar for the three techniques Sensitivity of fluorescence is greater than absorption spectro-
(Fig. 21b). However, the detection limit for aspirin in Avicel using scopic methods because it uses both excitation and emission wave-
the acoustic emission (AE) system was reported to be around lengths. It is available with relatively low cost and it is fast and
5.2% (w/w) which was poorer than that reported for the Raman sensitive to differentiate active content at low concentrations.
and NIR systems. However, fluorescence intensity is sensitive to fluctuations in pH
Acoustic emission has the advantage of being a non-invasive and temperature. Also, the applicability of this technique depends
monitoring method for powder mixing. Unlike many techniques, on the strength of fluorescence of the component to be investi-
such as NIR, which needs an optically transparent window in the gated relative to the fluorescence of other components within
process vessel, this method is able to collect the spectral informa- the mixture.
tion of blends behind any type of wall material. Also, the cost of an
AE measurement system is less than NIR spectroscopy. However, 4. Conclusions
some authors [125,132] reported that this method would only be
representative for powder composition at the interface rather than To obtain a high-end product quality, mixture of powders
in the bulk material as AE spectra only could be obtained by the should be made with high content uniformity to reduce powder
collisions of particles with the vessel wall [125,132]. The depth segregation and product failure. Numerous uniformity assessment
information for AE spectrometry is reported to be less than other methods for particulate systems have been developed and studied
spectroscopic techniques such as NIR (typically 2–3 mm). More- over the last few decades. In this study, different techniques for the
over, the sensitivity of AE is found to be poorer than NIR and assessment of blend uniformity have been examined. Spectro-
Raman spectroscopy. scopic approaches such as NIR and Raman have been broadly used
for the assessment of powder mixture uniformity. These tech-
3.4.4. Fluorescence spectroscopy niques attracted more attention due to the fact that the uniformity
Materials that absorb ultraviolet or visible light energy, may of several components can be monitored either at-line, in-line and/
dissipate a part of the energy as heat or electromagnetic radiation. or on-line using these techniques. However, using tomography sys-
The latter phenomenon is called fluorescence. The fluorescence tems or methods based on particle properties, such as conductivity,
detection of materials can be useful to study the behaviour of pro- one or two ingredients could be monitored usually in an at-line
cesses and therefore is a potential tool for the assessment of con- mode. Among different spectroscopic tools, Raman and fluores-
tent blend uniformity [133,134]. Domike et al. [135] used a light cence showed a high sensitivity to detect minor components,
induced fluorescence (LIF) instrument to evaluate the total content therefore they can be applied for the evaluation of powder unifor-
of fluorescent active pharmaceutical ingredient (API) in tablets. mity with higher accuracy. However, NIR spectroscopy is more
Ten tablets from three batches containing different weights of tri- commonly used for the evaluation of powder homogeneity due
amterene have been tested using this technique and the results to its relatively low cost as compared to other spectroscopic instru-
were compared with those obtained by UV–visible spectrophotom- ments. In general, the suitable technique must be chosen according
etry. The position of tablet relative to the LIF instrument is shown to the application, material and device specifications, budget as
in Fig. 22a. The concentration of the API of tablets was obtained by well as the required precision and sensitivity. The advantages
plotting the calibration curve of LIF signal against different stan- and disadvantages of all the reviewed techniques are briefly sum-
dard API concentrations. Fig. 22b shows a relatively good agree- marised in Table 4.
ment between the results of fluorescence spectroscopy and UV– In certain circumstances, different techniques could lead to
visible spectrophotometry analysis. The Root Mean Standard Error false conclusions. For example, wrong choice of dilution technique
of Prediction (RMSEP) for LIF was reported to be in the range of may lead to inaccurate results obtained by HPLC and/or UV–visible
4.40–7.93%. absorbance spectrophotometry [69,137]. Improper light condition,
In another study, Karumanchi et al. [136] used a LIF sensor dirty camera lenses, a low-resolution camera and unreliable
(with a sample volume of approximately 0.117 cm3) to evaluate image-processing softwares could influence the accuracy and reli-
and monitor the blend homogeneity of fluorescent API. For this ability of the image analysis. For instance, Abdelrahman et al. [138]
 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549 1547

evaluated the flow detection of particles using optical imaging. It [23] C. Benedetti, N. Abatzoglou, J.S. Simard, L. McDermott, G. Leonard, L. Cartilier,
Cohesive, multi component, dense powder flow characterization by NIR, Int. J.
was shown that changing the camera focus on the point of interest
Pharm. 336 (2007) 292–301.
could affect the accuracy of the particle detection. In spectroscopic [24] S. Laske, A. Paudel, O. Scheibelhofer and the Author Team, A review of PAT
techniques, scanning cannot be achieved properly when a sticky strategies in secondary solid oral dosage manufacturing of small molecules, J.
lump of powders adheres to an optical sensor during the process. Pharm. Sci. 106 (2017) pp. 667–712.
[25] T. De Beer, A. Burggraeve, M. Fonteyne, L. Saerens, J.P. Remon, C. Vervaet, Near
In addition, the efficiency of spectral acquisition and the related infrared and Raman spectroscopy for the in-process monitoring of
data analysis is highly dependent on the distance of the probe from pharmaceutical production processes, Int. J. Pharm. 417 (2011) 32–47.
the powder mixture. For example, 3 mm is reported to be the most [26] D.A. Skoog, F.J. Holler, S.R. Crouch, Principles of Instrumental Analysis, 6th ed.,
Brooks Cole, 2007.
efficient distance for the MicroNIR1700Ò probe in the related study [27] D. Harvey, Modern analytical chemistry, Chapter 10, McGraw-Hill Higher
of Asachi et al. [117]. Other factors which influence the light Education, 1999, pp. 368–461.
absorption such as porosity, particle morphology, particle size dis- [28] D.G. Watson, Pharmaceutical Analysis: A Textbook for Pharmacy Students
and Pharmaceutical Chemists, fourth ed., Elsevier, 2016.
tribution and powder layer thickness must be precisely considered [29] S. Gorog, Ultraviolet-Visible Spectrophotometry in Pharmaceutical Analysis,
during spectral analysis [139]. Environmental conditions, such as CRC Press, 2018.
humidity and temperature, must also be taken into consideration [30] A.C. Tella, O.M. Olabemiwo, M.O. Salawu, G.K. Obiyenwa, Developing a
Spectrophotometric method for the estimation of Albendazole in solid and
for an accurate result analysis in different techniques such as NIR suspension forms, Int. J. Phys. Sci. 5 (2010) 379–382.
spectroscopy, tribo-electrification and electrical conductivity [31] M. Jamrogiewicz, K. Cal, M. Gruszecka, A. Ciesielski, Determination of API
[140–143]. content in a pilot-scale blending by near-infrared spectroscopy as a first step
method to process line implementation, Acta Pol. Pharm. 70 (2013) 419–429.
[32] A.S. El-Hagrasy, H.R. Morris, F. D’Amico, R.A. Lodder, J.K. Drennen, Near-
infrared spectroscopy and imaging for the monitoring of powder blend
References homogeneity, J. Pharm. Sci. Sep. 90 (2001) 1298–1307.
[33] H. Ma, C.A. Anderson, Characterization of pharmaceutical powder blends by
NIR chemical imaging, J. Pharm. Sci. 97 (2008) 3305–3320.
[1] B. Remy, J.G. Khinast, B.J. Glasser, Discrete element simulation of free flowing [34] S.L. Upstone, Ultraviolet/visible Light Absorption Spectrophotometry in
grains in a four-bladed mixer, AIChE J. 55 (2009) 2035–2048. Clinical Chemistry, Encyclopedia of Analytical Chemistry, John Wiley and
[2] S.I. Badawy, T.J. Lee, M.M. Menning, Effect of drug substance particle size on Sons Ltd., 2013.
the characteristics of granulation manufactured in a high-shear mixer, AAPS [35] Y.K. Jung, M.J. Kim, Y.J. Kim, J.Y. Kim, Limitation of UV-Vis absorption analysis
Pharm. Sci. Tech. 1 (2000) 55–61. for determination of aqueous colloidal fullerene (nC60) at high ionic strength,
[3] R.K. Thakur, C. Vial, K.D.P. Nigam, E.B. Nauman, G. Djelveh, Static mixers in the KSCE J. Civ. Eng. 17 (2013) 51–59.
process industries-A review, Chem. Eng. Res. Des. 81 (2003) 787–826. [36] S. Krukowski, M. Karasiewicz, W. Kolodziejski, Convenient UV-
[4] H. Hoornahad, E.A.B. Koenders, Towards simulation of fresh granular-cement spectrophotometric determination of citrates in aqueous solutions with
paste material behavior, Adv. Mater. Res. 295 (2011) 2171–2177. applications in the pharmaceutical analysis of oral electrolyte formulations, J.
[5] X.Y. Zhou, G.J. He, Y.L. Fan, Y. Xiao, S.K. Kunnath, G. Monti, Advances in Civil Food Drug Anal. 25 (2017) 717–722.
Infrastructure Engineering, Trans Tech Publications Ltd, 2013, pp. 1287–1294. [37] S. Moldoveanu, V. David, Selection of the HPLC Method in Chemical Analysis,
[6] J.K. Prescott, T.P. Garcia, A solid dosage and blend uniformity trouble shooting first ed., Elsevier, 2016.
diagram, Pharm. Technol. 25 (2001) 68–87. [38] M. Thammana, A review on high performance liquid chromatography (HPLC),
[7] A.N. Huang, H.P. Kuo, Developments in the tools for the investigation of RRJPA. 5 (2016) 22–28.
mixing in particulate systems-A review, Adv. Powder Technol. 25 (2014) 163– [39] I.D. Kamalanathan, P.J. Martin, Competitive adsorption of surfactant-protein
173. mixtures in a continuous stripping mode foam fractionation column, Chem.
[8] P. Tang, V.M. Puri, Methods for minimizing segregation: a review, particulate Eng. Sci. 146 (2016) 291–301.
science and technology, Part. Sci. Technol. 22 (4) (2004) 321–337. [40] X. Liu, M. Tracy, C. Pohl, New Development in Surfactant Analysis by HPLC,
[9] M.E. Aulton, Aulton’s pharmaceutics: the design and manufacture of Technical report, Dionex Corporation, Sunnyvale, CA USA, 2014.
medicines, Chapter 12 written by Andrew Twitchell, 3rd ed., Edinburgh, [41] E. Nourafkan, Z. Hu, D. Wen, Controlled delivery and release of surfactant for
Churchill Livingstone, 2007, pp. 152–167. enhanced oil recovery by nanodroplets, Fuel 218 (2018) 396–405.
[10] M. Poux, P. Fayolle, J. Bertrand, D. Bridoux, J. Bousquet, Powder mixing: some [42] M.I. Walash, F. Belal, N. El-Enany, M. Eid, R.N. El-Shaheny, Stability-indicating
practical rules applied to agitated systems, Powder Technol. 68 (1991) 213– HPLC method for determination of naftazone in tablets. Application to
234. degradation kinetics and content uniformity testing, J. Chromatogr. Sci. 49
[11] A. Levy, C. Kalman, Handbook of Conveying and Handling of Particulate (2011) 495–501.
Solids, Chapter 7, Volume 10, 1st ed., Elsevier Science, 2001, pp. 589–603. [43] R. Tanaka, N. Takahashi, Y. Nakamura, Y. Hattori, K. Ashizawa, M. Otsuka,
[12] J. Bridgwater, Mixing of powders and granular materials by mechanical Verification of the mixing processes of active pharmaceutical ingredient,
means-A perspective, Particuology. 10 (2012) 397–427. excipient and lubricant in a pharmaceutical formulation using a resonant
[13] H.G. Brittain, Particle-size distribution II: the problem of sampling powdered acoustic mixing technology, RSC Adv. 6 (2016) 87049–87057.
solids, Pharm. Tech. 2 (2002) 67–73. [44] S.M. Dhole, P.B. Khedekar, N.D. Amnerkar, Comparison of UV
[14] E.Z. Dahmash, A.R. Mohammed, Review foundation: characterisation and spectrophotometry and high performance liquid chromatography methods
surface-profiling techniques for composite particles produced by dry powder for the determination of repaglinide in tablets, Pharm. Methods 3 (2012) 68–
coating in pharmaceutical drug delivery, Drug Discov. Today. 21 (2016) 550– 72.
561. [45] F. Eeni, Use of high-performance liquid chromatography in the
[15] F.J. Muzzio, P. Robinson, C. Wightman, D. Brone, Sampling practices in pharmaceutical industry, J. Chromatogr. A 507 (1990) 141–149.
powder blending, Int. J. Pharm. 155 (1997) 153–178. [46] S.O. Eraga, M.I. Arhewoh, R.N. Chibuogwu, M.A. Iwuagwu, A comparative
[16] R.W. Gerlach, J.M. Nocerino, Guidance for Obtaining Representative UVHPLC analysis of ten brands of ibuprofen tablets, Asian Pac. J. Trop.
Laboratory Analytical Subsamples from Particulate Laboratory Samples, U.S. Biomed. 5 (2015) 880–884.
Environmental Protection Agency, EPA/600/R-03/027, 2003. [47] V. Baeten, Spectroscopy: developments in instrumentation and analysis, J.
[17] Applying to Standards, Tests, Assays, and Other Specifications of the United Grasas. Y. aceites. 53 (2002) 45–63.
States Pharmacopeia, Bulk Powder Sampling Procedures, Copyright Ó The [48] V. Mizonov, I. Balagurov, H. Berthiaux, C. Gatumel, Intensification of vibration
United States Pharmacopeial Convention, ISBN: 978-1-936424-41-2, 2015, mixing of particulate solids by means of multi-layer loading of components,
pp. 523–536. Adv. Powder Technol. 28 (2017) 3049–3055.
[18] T. Allen, Particle size measurement, Powder sampling and particle size [49] M. Yamamoto, S. Ishihara, J. Kano, Evaluation of particle density effect for
Measurement, Volume 1, 5th ed., Springer, 1968. mixing behavior in a rotating drum mixer by DEM simulation, Adv. Powder
[19] E.L. Paul, V.A. Atiemo-Obeng, S.M. Kresta, Handbook of Industrial Mixing, Technol. 27 (2016) 864–870.
Science and Practice, Volume 15, Wiley-Blackwell Publication, 2003, pp. 887– [50] J.G. Rosas, M. Blanco, A criterion for assessing homogeneity distribution in
987. hyperspectral images. Part 1: Homogeneity index bases and blending
[20] A.S.L. Mendez, G. de Carli, C.V. Garcia, Evaluation of powder mixing operation processes, J. Pharm. Biomed. Anal. 70 (2012) 680–690.
during batch production: application to operational qualification procedure [51] C. Ammarcha, C. Gatumel, J.L. Dirion, M. Cabassud, H. Berthiaux, Continuous
in the pharmaceutical industry, Powder Technol. 198 (2010) 310–313. powder mixing of segregating mixtures under steady and unsteady state
[21] F.J. Muzzio, C.L. Goodridge, A. Alexander, P. Arratia, H. Yang, O. Sudah, G. regimes: Homogeneity assessment by real-time on-line image analysis,
Mergen, Sampling and characterization of pharmaceutical powders and Powder Technol. 315 (2017) 39–52.
granular blends, Int. J. Pharm. 250 (2003) 51–64. [52] J. Cho, Y. Zhu, K. Lewkowicz, S. Hee Lee, T. Bergman, B. Chaudhuri, Solving
[22] L. Susana, P. Canu, A.C. Santomaso, Development and characterization of a granular segregation problems using a biaxial rotary mixer, Chem. Eng.
new thief sampling device for cohesive powders, Int. J. Pharm. 416 (2011) Process. 58 (2012) 42–50.
260–267.
1548 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549

[53] C.R. Jung, R.S. Ortiz, R. Limberger, P. Mayorga, A new methodology for Discrete Element Method (DEM) model of granular flow and mixing in the
detection of counterfeit Viagra1 and Cialis1 tablets by image processing and Turbula mixer, Int. J. Pharm. 446 (2013) 46–58.
statistical analysis, Forensic Sci. Int. 216 (2012) 92–96. [82] O. Gundogdu, Positron Emission Tomography Particle tracking using cluster
[54] O.O. Olaofe, K.A. Buist, N.G. Deen, M.A. van der Hoef, J.A.M. Kuipers, Improved analysis, Nucl. Instrum. Methods. A. 534 (2004) 562–576.
digital image analysis technique for the evaluation of segregation in pseudo- [83] D.J. Parker, C.J. Broadbent, P. Fowles, M.R. Hawkesworth, P. McNeil, Positron
2D beds, Powder Technol. 244 (2013) 61–74. emission particle tracking-a technique for studying flow within engineering
[55] J.G. Benito, I. Ippolito, A.M. Vidales, Novel aspects on the segregation in quasi equipment, Nucl. Instrum. Methods 326 (1993) 592–607.
2D piles, Powder Technol. 234 (2013) 123–131. [84] M. Pasha, A. Hassanpour, H. Ahmadian, H.S. Tan, A. Bayly, M. Ghadiri, A
[56] D. Pihler-Puzovic, T. Mullin, The timescales of granular segregation in comparative analysis of particle tracking in a mixer by discrete element
horizontally shaken monolayers, Proc. R. Soc. 469 (2013) 1–20. method and positron emission particle tracking, Powder Technol. 270 (2015)
[57] X. Liu, C. Zhang, J. Zhan, Quantitative comparison of image analysis methods 569–574.
for particle mixing in rotary drums, Powder Technol. 282 (2015) 32–36. [85] N. Ehrhardt, M. Montagne, H. Berthiaux, B. Dalloz-Dubrujeaud, C. Gatumel,
[58] H. Mio, R. Higuchi, W. Ishimaru, A. Shimosaka, Y. Shirakawa, J. Hidaka, Effect Assessing the homogeneity of powder mixtures by on-line electrical
of paddle rotational speed on particle mixing behavior in electrophotographic capacitance, Chem. Eng. Process. 44 (2005) 303–313.
system by using parallel discrete element method, Adv. Powder Technol. 20 [86] J. Huang, Y. Lu, H. Wang, new quantitative measurement method for mixing
(2009) 406–415. and segregation of binary-mixture fluidized bed by capacitance probe, Chem.
[59] R.K. Soni, R. Mohanty, S. Mohanty, B.K. Mishra, Numerical analysis of mixing Eng. J. 326 (2017) 99–108.
of particles in drum mixers using DEM, Adv. Powder Technol. 27 (2016) 531– [87] A.J. Sederman, L.F. Gladden, M.D. Mantle, Application of magnetic resonance
540. imaging techniques to particulate systems, Adv. Powder Technol. 18 (2007)
[60] A. Realpe, K. Barrios, M. Rozo, Assessment of homogenization degree of 23–38.
powder mixing in a cylinder rotating under cascading regime, Int. J. Eng. [88] E.H. Hardy, J. Hoferer, G. Kasper, The mixing state of fine powders measured
Technol. 7 (2015) 394–404. by magnetic resonance imaging, Powder Technol. 177 (2007) 12–22.
[61] G. Olivieri, A. Marzocchella, P. Salatino, Segregation of fluidized binary [89] R. Stannarius, Magnetic resonance imaging of granular materials, Rev. Sci.
mixtures of granular solids, AIChE J. 50 (2004) 3095–3106. Instr. 88 (2017) 051806.
[62] A.S.M. Yudin, S. Anuar, A.N. Oumer, Improvement on particulate mixing [90] L. Pernenkil, C.L. Cooney, A review on the continuous blending of powders,
through inclined slotted swirling distributor in a fluidized bed: An Chem. Eng. Sci. 61 (2006) 720–742.
experimental study, Adv. Powder Technol. 27 (2016) 2102–2111. [91] M. Nakagawa, S.A. Altobelli, A. Caprihan, E. Fukushima, E.K. Jeong,
[63] P. Shenoy, M. Viau, K. Tammel, F. Innings, J. Fitzpatrick, L. Ahrne, Effect of Noninvasive measurements of granular flows by magnetic-resonance
powder densities, particle size and shape on mixture quality of binary food imaging, Exp. Fluids 16 (1993) 54–60.
powder mixtures, Powder Technol. 272 (2015) 165–172. [92] M.M. Khalil, J.L. Tremoleda, T.B. Bayomy, W. Gsell, Molecular SPECT imaging:
[64] P. Shenoy, F. Innings, K. Tammel, J. Fitzpatrick, L. Ahrne, Evaluation of a digital an overview, Int. J. Mol. Imag. 2011 (2011) 1–15.
colour imaging system for assessing the mixture quality of spice powder [93] N. Sommier, P. Porion, P. Evesque, B. Leclerc, P. Tchoreloff, G. Couarraze,
mixes by comparison with a salt conductivity method, Powder Technol. 286 Magnetic resonance imaging investigation of the mixing-segregation process
(2015) 48–54. in a pharmaceutical blender, Int. J. Pharm. 222 (2001) 243–258.
[65] T. Hao, J. Tukianen, A. Nivorozhkin, N. Landrau, Probing pharmaceutical [94] P. Porion, N. Sommier, A.M. Faugere, P. Evesque, Dynamics of size segregation
powder blending uniformity with electrostatic charge measurements, and mixing of granular materials in a 3D-blender by NMR imaging
Powder Technol. 245 (2013) 64–69. investigation, Powder Technol. 141 (2004) 55–68.
[66] S.V. Hammond, F.J. Muzzio, K.C. Pingali, T. Shinbrot, An observed correlation [95] M. Niedostatkiewicz, J. Tejchman, Z. Chaniecki, K. Grudzien, Determination of
between flow and electrical properties of pharmaceutical blends, Powder bulk solid concentration changes during granular flow in a model silo with
Technol. 192 (2009) 157–165. ECT sensors, Chem. Eng. Sci. 64 (2009) 20–30.
[67] L. Mathews, C. Chandler, S. Dipali, P. Adusumilli, S. Lech, S. Daskalakis, N. [96] C. Rautenbach, M.C. Melaaen, B.M. Halvorsen, Possible identification of size
Mathis, Monitoring blend uniformity with effusivity, Pharm. Technol. 80–84 difference segregation using electrical capacitance tomography and statistical
(2002). analysis, Eur. J. Sci. Res. 116 (2013) 351–364.
[68] G. Leonard, F. Bertrand, J. Chaouki, P.M. Gosselin, An experimental [97] W. Bai, N.K.G. Keller, T.J. Heindel, R.O. Fox, Numerical study of mixing and
investigation of effusivity as an indicator of powder blend uniformity, segregation in a biomass fluidized bed, Powder Technol. 237 (2013) 355–366.
Powder Technol. 181 (2008) 149–159. [98] O. Mihailova, V. Lim, M.J. McCarthy, K.L. McCarthy, S. Bakalis, Laminar mixing
[69] S. Mahmood, N.N.B. Hilmi, N.K.B. Husain, B. Chatterjee, U.K. Mandal, in a SMX static mixer evaluated by positron emission particle tracking (PEPT)
Differential scanning calorimetric characterization of pharmaceutical and magnetic resonance imaging (MRI), Chem. Eng. Sci. 137 (2015) 1014–
powder blend uniformity in a laboratory-scale V-blender, Powder Technol. 1023.
287 (2016) 152–159. [99] Z. Csoban, B. Kallai-Szabo, N. Kallai-Szabo, T. Takacs, T. Hurtony, P. Gordon, R.
[70] E. Bharvada, V. Shah, M. Misra, Exploring mixing uniformity of a Zelko, I. Antal, Assessment of distribution of pellets in tablets by non-
pharmaceutical blend in a high shear mixture granulator using enthalpy destructive microfocus X-ray imaging and image analysis technique, Powder
values obtained from DSC, Powder Technol. 276 (2015) 103–111. Technol. 301 (2016) 228–233.
[71] J. Jamaludin, M.Z. Zawahir, R.A. Rahim, F.R.M. Yunus, N.M.N. Ayob, M.S.R. Aw, [100] A. Kohler, A. Rasch, D. Pallares, F. Johnsson, Experimental characterization of
N.S. Fadzil, Z. Zakaria, M.H.F. Rahiman, A review of tomography system, axial fuel mixing in fluidized beds by magnetic particle tracking, Powder
Jurnal. Teknologi. 64 (2013) 47–51. Technol. 316 (2017) 492–499.
[72] M.J. Paulus, S.S. Gleason, S.J. Kennel, P.R. Hunsicker, D.K. Johnson, High [101] A.A. Gowen, C. Donnell, P. Cullen, S. Bell, Recent Applications of chemical
resolution X-ray computed tomography: an emerging tool for small animal imaging to pharmaceutical process monitoring and quality control, Eur. J.
cancer research, Neoplasia 2 (2000) 62–70. Pharm. Biopharm. 69 (2008) 10–22.
[73] R. Liu, X. Yin, H. Li, Q. Shao, P. York, Y. He, T. Xiao, J. Zhang, Visualization and [102] B. Stuard, Infrared Spectroscopy: Fundamentals and Applications, John Wiley
quantitative profiling of mixing and segregation of granules using & Sons, 2004.
synchrotron radiation X-ray microtomography and three dimensional [103] Y. Sulub, M. Konigsberger, J. Cheney, Blend uniformity end-point
reconstruction, Int. J. Pharm. 445 (2013) 125–133. determination using near-infrared spectroscopy and multivariate
[74] S. Poutiainen, J. Pajander, A. Savolainen, J. Ketolainen, K. Jarvinen, Evolution of calibration, J. Pharm. Biomed. Anal. 55 (2011) 429–434.
granule structure and drug content during fluidized bed granulation by X-ray [104] D.R. Ely, M. Teresa, Carvajal, Determination of the scale of segregation of low
microtomography and confocal Raman spectroscopy, J. Pharm. Sci. 100 dose tablets using hyperspectral imaging, Int. J. Pharm. 414 (2011) 157–160.
(2011) 5254–5269. [105] X. He, X. Han, N. Ladyzhynsky, R. Deanne, Assessing powder segregation
[75] I. Akseli, S. Iyer, H.P. Lee, A.M. Cuitino, A quantitative correlation of the effect potential by near infrared (NIR) spectroscopy and correlating segregation
of density distributions in roller-compacted ribbons on the mechanical tendency to tabletting performance, Powder Technol. 236 (2013) 85–99.
properties of tablets using ultrasonics and X-ray tomography, AAPS Pharm. [106] K. Jarvinen, W. Hoehe, M. Jarvinen, S. Poutiainen, M. Juuti, S. Borchert, In line
Sci. Tech. 12 (2011) 834–853. monitoring of the drug content of powder mixtures and tablets by near-
[76] D.L. Bailey, D.W. Townsend, P.E. Valk, M.N. Maisey, Positron Emission infrared spectroscopy during the continuous direct compression tableting
Tomography: Basic Sciences, Springer, 2005. process, Eur. J. Pharm. Sci. 48 (2013) 680–688.
[77] P.M. Portillo, A.U. Vanarase, A. Ingram, J.K. Seville, M.G. Ierapetritou, F.J. [107] A.U. Vanarase, M. Jarvinen, J. Paaso, F.J. Muzzio, Development of a
Muzzio, Investigation of the effect of impeller rotation rate, powder flow rate, methodology to estimate error in the on-line measurements of blend
and cohesion on powder flow behavior in a continuous blender using PEPT, uniformity in a continuous powder mixing process, Powder Technol. 241
Chem. Eng. Sci. 65 (2010) 5658–5668. (2013) 263–271.
[78] H.P. Kuo, P.C. Knight, D.J. Parker, M.J. Adams, J.P.K. Seville, Discrete element [108] P.R. Wahl, G. Fruhmann, S. Sacher, G. Straka, S. Sowinski, J.G. Khinast, PAT for
simulations of a high-shear mixer, Adv. Powder Technol. 15 (2004) 297–309. tableting: Inline monitoring of API and excipients via NIR Spectroscopy, Eur. J.
[79] D.J. Parker, A.E. Dijkstra, T.W. Martin, J.P.K. Seville, Positron emission particle Pharm. Biopharm. 87 (2014) 271–278.
tracking studies of spherical particle motion in rotating drums, Chem. Eng. [109] J.G. Osorio, G. Stuessy, G.J. Kemeny, F.J. Muzzio, Characterization of
Sci. 52 (1997) 2011–2022. pharmaceutical powder blends using in situ near-infrared chemical
[80] C.T. Crowe, Multiphase Flow Handbook, volume 14, CRC Press Taylor and imaging, Chem. Eng. Sci. 108 (2014) 244–257.
Francis, 2006. [110] B. Bakri, M. Weimer, G. Hauck, G. Reich, Assessment of powder blend
[81] M. Marigo, M. Davies, T. Leadbeater, D.L. Cairns, A. Ingram, E.H. Stitt, uniformity: Comparison of real-time NIR blend monitoring with stratified
Application of Positron Emission Particle Tracking (PEPT) to validate a
 M. Asachi et al. / Advanced Powder Technology 29 (2018) 1525–1549 1549

sampling in combination with HPLC and at-line NIR Chemical Imaging, Eur. J. analysis of multi-component pharmaceutical powders, J. Pharm. Biomed.
Pharm. Biopharm. 97 (2015) 78–89. Anal. 141 (2017) 180–191.
[111] Y. Lin, W. Li, J. Xu, P. Boulas, Development of a NIR-based blend uniformity [127] S.J. Hudak, K. Haber, G. Sando, L.H. Kidder, E.N. Lewis, Practical limits of
method for a drug product containing multiple structurally similar actives by spatial resolution in diffuse reflectance NIR chemical imaging, NIR News 18
using the quality by design principles, Int. J. Pharm. 488 (2015) 120–126. (2007) 6–8.
[112] J.G. Osorio, F.J. Muzzio, Evaluation of resonant acoustic mixing performance, [128] J.M. Amigo, C. Ravn, Direct quantification and distribution assessment of
Powder Technol. 278 (2015) 46–56. major and minor components in pharmaceutical tablets by NIR-chemical
[113] Z. Shi, K.C. McGhehey, I.M. Leavesley, L.F. Manley, On-line monitoring of imaging, Eur. J. Pharm. Sci. 37 (2009) 76–82.
blend uniformity in continuous drug productmanufacturing process-The [129] S. Svanberg, Atomic and molecular spectroscopy: basic aspects and practical
impact of powder flow rate and the choice of spectrometer: Dispersive vs. FT, applications, third ed., Chapter 10, Springer-Verlag, Berlin Heidelberg, 2001,
J. Pharm. Biomed. Anal. 118 (2016) 259–266. pp. 389–454.
[114] L.K. Bittner, S.A. Schönbichler, M. Schmutzler, O.M.D. Lutz, C.W. Huck, [130] C.J. Strachan, M. Windbergs, H.L. Offerhaus, Pharmaceutical applications of
Vibrational spectroscopic methods for the overall quality analysis of washing non-linear imaging, Int. J. Pharm. 417 (2011) 163–172.
powders, Talanta 148 (2016) 329–335. [131] M.N. Slipchenko, H. Chen, D.R. Ely, Y. Jung, M.T. Carvajal, J.X. Cheng,
[115] L.B. Schlegel, M. Schubert-Zsilavecz, M. Abdel-Tawab, Quantification of active Vibrational imaging of tablets by epi-detected stimulated Raman scattering
ingredients in semi-solid pharmaceutical formulations by near infrared microscopy, Analyst 135 (2010) 2613–2619.
spectroscopy, J. Pharm. Biomed. Anal. 142 (2017) 178–189. [132] P. Allan, L.J. Bellamy, A. Nordon, D. Littlejohn, Non-invasive monitoring of the
[116] K. Johanson, Review of new segregation tester method by Dr, Kerry Johanson, mixing of pharmaceutical powders by broadband acoustic emission, Analyst
Powder Technol. 257 (2014) 1–10. 135 (2010) 518–524.
[117] M. Asachi http://www.sciencedirect.com/science/article/pii/S003259101 [133] C.K. Lai, C.L. Cooney, Application of a fluorescence sensor for miniscale on-
7305405?via%3Dihub - !, A. Hassanpour, M. Ghadiri, A. Bayly, Assessment line monitoring of powder mixing kinetics, J. Pharm. Sci. 93 (2004) 60–70.
of Near-Infrared (NIR) spectroscopy for segregation measurement of low [134] C.K. Lai, D. Holt, J.C. Leung, G.K. Raju, P. Hansen, C.L. Cooney, Real time and
content level ingredients, Powder Technol. 320 (2017) pp. 143–154. noninvasive monitoring of dry powder blend homogeneity, AIChE J. 47
[118] S. Oka, A. Sahay, W. Meng, F. Muzzio, Diminished segregation in continuous (2001) 2618–2622.
powder mixing, Powder Technol. 309 (2017) 79–88. [135] R. Domike, S. Ngai, C.L. Cooney, Light induced fluorescence for predicting API
[119] D.M. Koller, A. Posch, G. Hörl, C. Voura, S. Radl, N. Urbanetz, S.D. Fraser, W. content in tablets: sampling and error, Int. J. Pharm. 391 (2010) 13–20.
Tritthart, F. Reiter, M. Schlingmann, J.G. Khinast, Continuous quantitative [136] V. Karumanchi, M.K. Taylor, K.J. Ely, W.C. Stagner, Monitoring powder blend
monitoring of powder mixing dynamics by near-infrared spectroscopy, homogeneity using light-induced fluorescence, AAPS Pharm. Sci. Tech. 12
Powder Technol. 205 (2011) 87–96. (2011) 1031–1037.
[120] Z. Shi, R.P. Cogdill, S.M. Short, C.A. Anderson, Process characterization of [137] A.S. El-Hagrasy, F.D. Amico, J.K. Drennen, A process analytical technology
powder blending by near-infrared spectroscopy: blend end-points and approach to near-infrared process control of pharmaceutical powder
beyond, J. Pharm. Biomed. Anal. 47 (2008) 738–745. blending. Part I: D-optimal design for characterization of powder mixing
[121] O. Scheibelhofer, N. Balak, D.M. Koller, J.G. Khinast, Spatially resolved and preliminary spectral data evaluation, J. Pharm. Sci. 95 (2006) 392–406.
monitoring of powder mixing processes via multiple NIR-probes, Powder [138] M. Abdelrahman, E.W. Reutzel, A.R. Nassar, T.L. Starr, Flaw detection in
Technol. 243 (2013) 161–170. powder bed fusion using optical imaging, Addit. Manuf. 15 (2017) 1–11.
[122] M.A. Alam, Z. Shi, J.K. Drennen, C.A. Anderson, In-line monitoring and [139] A. Rinnan, F.V.D. Berg, S.B. Engelsen, Review of the most common pre-
optimization of powder flow in a simulated continuous process processing techniques for near-infrared spectra, Trends Anal. Chem. 28
using transmission near infrared spectroscopy, Int. J. Pharm. 526 (2017) (2009) 1201–1222.
199–208. [140] J. Zumba, J. Rodgers, M. Indest, Impact of temperature and relative humidity
[123] L. Zhang, M.J. Henson, S.S. Sekulic, Multivariate data analysis for Raman on the near infrared spectroscopy measurements of cotton fiber micronaire, J.
imaging of a model pharmaceutical tablet, Anal. Chim. Acta 545 (2005) 262– Text. Res. DOI: 10.1177/0040517517720499, (2017) pp. 1–13.
278. [141] Y.E. Prawatya, M.B. Neagoe, T. Zeghloul, L. Dascalescu, Optimization of
[124] E. Smith, Modern Raman Spectroscopy-A Practical Approach, John Wiley & continuous triboelectrification process for polymeric materials in dry contact,
Sons, Manchester, 2005. 13th International Conference on Tribology, IOP Publishing, 174 (2017) 012067.
[125] P. Allan, L.J. Bellamy, A. Nordon, D. Littlejohn, J. Andrews, P. Dallin, In situ [142] W.D. Greason, Investigation of a test methodology for triboelectrification, J.
monitoring of powder blending by non-invasive Raman spectrometry with Electrostat. 49 (2000) 245–256.
wide area illumination, J. Pharm. Biomed. Anal. 76 (2013) 28–35. [143] A.K. Kamra, C.G. Deshpande, V. Gopalakrishnan, Effect of relative humidity on
[126] H. Wang, D. Barona, S. Oladepo, L. Williams, S. Hoe, D. Lechuga-Ballesteros, R. the electrical conductivity of marine air, Q. J. Roy. Meteor. Soc. 123 (1997)
Vehring, Macro-Raman spectroscopy for bulk composition and homogeneity 1295–1305.

View publication stats