Effect of Shear Applied During a Pharmaceutical Process on Near Infrared Spectra

Abstract

This study describes changes observed in the near-infrared (NIR) diffuse reflectance (DR) spectra of pharmaceutical tablets after these tablets were subjected to different levels of strain (exposure to shear) during the mixing process. Powder shearing is important in the mixing of powders that are cohesive. Shear stress is created in a system by moving one surface over another causing displacements in the direction of the moving surface and is part of the mixing dynamics of particulates in many industries including the pharmaceutical industry. In continuous mixing, shear strain is developed within the process when powder particles are in constant movement and can affect the quality attributes of the final product such as dissolution. These changes in the NIR spectra could affect results obtained from NIR calibration models. The aim of the study was to understand changes in the NIR diffuse reflectance spectra that can be associated with different levels of strain developed during blend shearing of laboratory samples. Shear was applied using a Couette cell and tablets were produced using a tablet press emulator. Tablets with different shear levels were measured using NIR spectroscopy in the diffuse reflectance mode. The NIR spectra were baseline corrected to maintain the scattering effect associated with the physical properties of the tablet surface. Principal component analysis was used to establish the principal sources of variation within the samples. The angular dependence of elastic light scattering shows that the shear treatment reduces the size of particles and produces their uniform and highly isotropic distribution. Tablet compaction further reduces the diffuse component of scattering due to realignment of particles.

Keywords

Near infrared (NIR) spectroscopy Light scattering Diffuse reflectance Pharmaceutics Tablet compaction Shear

Introduction

Analytical methods capable of monitoring critical quality parameters are valuable in the improvement of pharmaceutical processes. Near-infrared (NIR) spectroscopy has become one of the most used analytical techniques to monitor pharmaceutical processes since it does not require sample preparation, is capable of obtaining spectra in seconds, and is also a non-invasive technique.¹ Near-infrared spectroscopy applications have been developed to determine the drug and excipient concentration during mixing,^2–5 and moisture content during drying.⁶ Near-infrared spectroscopy spectra depend on the chemical composition of the materials analyzed and also on their physical properties such as particle size. Thus, the development of NIR methods to determine chemical composition usually requires mathematical transformations (called spectral pretreatments) to remove multiplicative and additive effects observed in NIR spectra that often occur due to particle size differences. The larger surface area to volume ratio from smaller particles leads to more light remitted to the detector and a diffuse reflectance (DR) spectrum with reduced absorbance.⁷ Near-infrared spectra obtained from pharmaceutical processes also show significant differences in their baseline. Without spectral pretreatments, the changes in the NIR spectra related to the physical properties of materials would be confused with chemical changes and could lead to errors in the determination of chemical composition.

This study shows that shear also affects NIR spectra. Powder shearing is important in mixing of powders that are cohesive. Shear stress is created in a system by moving one surface over another to cause displacements in the direction of the moving surface.⁸ In solid dosage manufacturing, powder is subjected to different unit operations, and thus to different levels of normal and abrasive stresses and strain. Lacey et al. reviewed the fundamental mechanisms (convective and diffusive) in powder mixing and indicated that agitation and shear help both in convective (macro-mixing) and dispersive (micro-mixing)^9,10 mixing. Shearing of powder is essential to break down agglomerates in cohesive powders, promoting micro-mixing, and thus uniform blends.¹¹ Most drug particles like acetaminophen and ibuprofen are intended to be of smaller size to allow a larger surface area for dissolution. However, decreasing the particle size also increases their cohesive nature. Mechanical shear also affects the hydrophobicity of pharmaceutical powder blends and its impact on drug release from tablets.¹² Typically, dissolution¹³ and hardness¹⁴ are adversely affected by excessive shear. This phenomenon is known as over-lubrication.^15,16 The work in this manuscript was done with acetaminophen, which is a cohesive material and a widely sold active pharmaceutical ingredient.

The effect of shear on NIR spectra is investigated in this study. Several studies describe the effect of compaction on NIR spectra.^17–22 However, the authors understand that this is the first report on the effect of shear on NIR spectra. This finding is important because the effect of shear could be a source of error in NIR methods for chemical composition. The effect of shear on NIR spectra could also be important for gaining better understanding of pharmaceutical processes and could also serve as a process signature to evaluate the authenticity of a commercial product. In this work, the effect of shear has been studied by comparison of NIR spectra obtained after subjecting blends to different strain conditions. In addition to this, NIR chemical imaging and angle-resolved elastic light scattering were used to study and explain these observed spectral differences. The work in this manuscript investigates the effect of strain on the NIR spectra for two different formulations. These formulations will be addressed herein as Case II and Case III.

Experimental Details

Materials and Sample Preparation

Case II Formulation: Powder blends were prepared by mixing microcrystalline cellulose (Avicel PH102, FMC biopolymer, Newark, DE, USA; mean particle size 100 µm) and lactose monohydrate NF (Foremost Farms, Rothschild, WI, USA; mean particle size 90 µm) as excipients, semi-fine acetaminophen (Mallinckrodt Inc., Raleigh, NC, USA; lot 008111H411) as an active ingredient (mean particle size 42 µm), and magnesium stearate NF (Mallinckrodt Inc., St. Louis, MO, USA; lot P09247, mean particle size 38 µm) as a lubricant. Each of these materials was passed through different sieves. Sieve no. 18 was used for acetaminophen (resulting mean particle size 45 µm), and sieve no. 30 for both lactose (resulting particle size 70 µm) and magnesium stearate (8 µm). Sieving was done to break agglomerates, if any. Case III Formulation: Powder blends were prepared with lactose monohydrate NF (Foremost farms, Rothschild, WI, USA; mean particle size 90 µm) as excipient material, semi-fine acetaminophen (Mallinckrodt Inc., Raleigh, NC, USA; lot 008111H411) as an active ingredient (mean particle size 42 µm), and magnesium stearate NF (Mallinckrodt Inc., St. Louis, MO, USA; lot P09247, mean particle size 38 µm) as a lubricant. Each of these materials were subjected to sieving to break agglomerates as in the Case II formulation. Table 1 summarizes the composition of the formulations known as Case II and Case III.

Table 1.

Tablet composition description for Case II and Case III.

Case study	Component	Concentration % (w/w)	Total % (w/w)
II	Acetaminophen	9	100
	Lactose	45
	Microcrystalline cellulose	45
	Magnesium stearate	1
III	Acetaminophen	9	100
	Lactose	90
	Magnesium stearate	1

Blending and Shearing Procedure

Case II: The blending was done in a 1.87 L V blender (Patterson-Kelley). One kilogram of the blend was made wherein the major ingredients (lactose, acetaminophen, and microcrystalline cellulose) were layered in a top–bottom fashion and rotated at 15 rpm for 15 min. A total of 10 g of the lubricant, magnesium stearate, was added to this premix and mixed for 25 revolutions. This step was undertaken so that sufficient macro-mixing and de-agglomeration was attained and the lubricant was well distributed. Further micro-mixing was achieved in the Couette Cell (Measurement Control Corporation, East Hanover, NJ, USA). The mixing time of lubricant in the blender was kept low and the intensifier bar of the blender was not operated to have minimum exposure to uncontrolled shear in the blender prior to exposing it to controlled shear in the shearing device (Couette Cell).²³ The blended powder was then removed from the V-blender and was split into three sets of 300 g each. For Case III a similar procedure was followed and an additional strain level of 2560 revolutions was introduced where the blend was sheared at 80 rpm for 32 min.

The Case II and III blends were subjected to controlled and uniform shear environment in a modified Couette Cell equipped with equidistant baffles. The Couette cylindrical cell consists of two concentric cylinders which rotate relative to each other and 300 g of the blend is filled in the annular region of the cylinder. The equidistant pins are uniformly spaced and the entire setup creates a uniform shear environment on the powder.²³ The shear rate was kept constant at 80 rpm for all the sets. The first set was subjected to a no strain level, the second sample was subjected to a strain level of 160 revolutions (sheared for 2 min), a third sample was subjected to a strain of 640 revolutions (8 min), and a fourth sample was subjected to a strain of 2560 revolutions (32 min). The 2560-revolution samples were only with the Case III formulation. Samples within the range of 160–640 revolutions are considered representative of a continuous mixing process based on the authors’ previous experience, while the 2560-revolution samples are considered an extreme case. However, the shear applied to these formulations is more uniform than the shear in industrial processes, because of the use of the Couette Cell which was necessary to perform reproducible experiments and promote micro-mixing of the ingredients by dispersion.

Tableting Procedure

Sheared powder samples were compressed into tablets using a Presster, a tablet press emulator (The Presster, Measurement Control Corporation, East Hanover, NJ, USA) designed to match compaction force and dwell time of any tablet press. The simulated press is a Kikusui Libra 2 with both pre-compaction and compaction rolls of 250 mm in diameter. The speed was kept constant at 20 rpm. Flat-faced punches and round dies were used to make round tablets of 10 mm diameter. The dosing position was adjusted and fixed to give tablets of 350 mg weight. Tablets were made at an average force of 8 kN, 12 Kn, and 16 kN by changing the compaction gap between the punches. For Case III all tablets had a constant compaction force of 24 kN. To avoid tablet weight variance, tablets at the start of the batch were discarded until the stable average pressure and weight was obtained. Finished tablets were kept inside a sealed bag to prevent exposure to moisture. A total of 54 tablets were made. Table 2 describes the tablets prepared.

Table 2.

Conditions and description of Case II tablets used in study.

Shear and strain treatment (rev)	Compaction force (kN)	Weight (mg)	Thickness (mm)
0	8.5	350	3.45
	12	360	3.40
	16	350	3.33
160	8.3	350	3.14
	11.7	350	3.34
	15.8	350	3.28
640	7.8	350	3.58
	11.2	350	2.34
	15.7	340	3.16

Near-Infrared Spectral Acquisition Parameters and Data Analysis

All spectra were obtained with a Multipurpose Analyzer (MPA) with a high intensity source from Bruker Optics (Billerica, MA, USA). The diffuse reflectance spectra were obtained with a resolution of 32 cm⁻¹. A total of 256 background scans were obtained and 128 sample scans. The resolution for the transmission measurements was also 32 cm⁻¹. A total of 256 background scans were obtained and 256 sample scans. Near-infrared data analysis was performed using The Unscrambler X v10.3. All spectra were pretreated using baseline correction at the first spectral point.

Near-Infrared Chemical Imaging Instrument Setup

Near-infrared chemical imaging was used to investigate the slope of the NIR spectra obtained across the tablet surface. A previously developed method to create an image based on the slope of the NIR spectra was used.¹⁹ A Malvern Spectral Dimensions SyNIRgi chemical imaging spectrometer with a focal plane array detector was used. All samples were analyzed with a 40 µm objective with a field of view (FoV) of 12.8 × 10.2 mm. Spectra were obtained in reflectance mode by placing the tablets over a white ceramic disk with a diameter of 28 mm. The reflectance spectra were converted to absorbance, log (1/R), followed by pixel correction to remove areas of non-uniform illumination and the effect of unresponsive pixels. The images consisted of 320 × 256 pixels, with a spectral range of 1200–2400 nm, forming the spectral hypercube. This disk was used as a reference for the acquisition system. Image analysis was performed with the Malvern ISys 5.0 and Matlab 2013A.

Scattering Measurements

Angle-resolved elastic light scattering measurements were used to analyze how the light scatters from the tablets. These measurements were obtained by the scatterometer shown in Figure 1. In this experiment a 15 mW He–Ne linearly polarized laser with wavelength λ = 632.8 nm was used to illuminate samples. The beam was focused on the sample surface into a spot of approximately 300 µm at normal incidence. The scattered light within the hemisphere was collected by an elliptical mirror of 20 cm diameter and projected onto a charge-coupled device (CCD) (SBIG ST-7MXEI-Camera, Santa Barbara, CA, USA). The CCD exposure time for each measurement was 300 ms. Integral light scattering was measured by a calibrated Si photodiode.

The scattering indicatrix images were analyzed using the bidirectional scatter distribution function (BSDF). The BSDF is a mathematical distribution which describes the way the light is scattered from a surface (Eq. 1). The BSDF is a superset of bidirectional transmittance distribution function (BTDF) and bidirectional reflectance distribution function (BRDF),²⁴ and is defined as:

BSDF (θ, ϕ) = (\frac{{dP}_{r} (θ, ϕ)}{d Ω}) / P_{i} \cos θ

(1)

θ and ϕ are the polar and azimuthal angles,

{dP}_{r}

is radiance light differential,

d Ω

is a solid angle differential, and P_i is the irradiance light. This equation can be applied for rapid evaluation of diffuse component of scattering indicatrix. Since fully diffuse scattering obeys Lambertian cosine law,^25,26 the BSDF in this case will be constant over all scattering angles.

Results and Discussion

Near-Infrared Spectroscopy

Figure 2 shows the pure component spectra used for Case II (Figure 2a) formulation (acetaminophen (APAP), lactose, microcrystalline cellulose (MCC), and magnesium stearate (MgSt)) and Case III (Figure 2b) formulation (APAP, lactose, and MgSt). Figure 3a shows the differences between the diffuse reflectance spectra of tablets obtained after compaction at 8 kN for blends of similar composition subjected to the three strain levels: no shear, 160 revolutions, and 640 revolutions, where strain refers to the duration of exposure to shear rate.²⁷ Broad bands are obtained at no shear conditions. As the strain is increased the bands became more defined as observed in Figs 3a to c. These spectra show significant differences in two spectral regions: 7450–7000 cm⁻¹ and 5600–5100 cm⁻¹. Figure 3b was obtained for tablets at 12 kN and Figure 3c for 16 kN does not show the differences between the second and third strain levels (160 and 640 revolutions, respectively), but the broad band in the no-shear-sample was still observed and two clearly defined bands are observed. Maxima are obtained at 7266 cm⁻¹ and 7189 cm⁻¹ after compaction forces of 12 kN and 16 kN. Near-infrared tablet transmission spectra did not show variation associated with shear. The spectra were baseline corrected to maintain scattering effects on the spectrum that could be related to the differences in porosity (physical differences) between the tablets.⁷ However, the spectral differences associated with variation in shear level are still observed after the SNV transformation or the second derivative is applied. The differences observed in terms of compaction force were expected,^17–22 but not the differences in the 7450–7000 cm⁻¹ and 5600–5100 cm⁻¹ regions. The 7450–7000 cm⁻¹ bands have been assigned to combinations of the first overtones of the C–H stretching modes and the C–H bending modes of cellulose.^28,29

Previous studies have shown that at lower compaction forces, the irregular surface of the particles combined with their random orientation results in the radiation becoming diffuse. At higher compaction forces, the surface area is reduced and the transmission of radiation increases with higher compaction force.^17,18 Thus, higher compaction results in higher absorbance results in the diffuse reflectance spectrum and lower absorbance in the transmission spectra. However, in this study the transmission spectra did not indicate significant differences between the tablets analyzed. The changes that occur in the NIR spectra as a result of differences in compaction were observed throughout the entire NIR spectrum, and cause changes in both spectral baseline and slope. Thus, the changes in compaction do not fully explain the spectral changes observed in this study.

Figure 4 shows a principal component analysis (PCA) scores plot of the spectra obtained with the entire set of 54 tablets. The PCA scores plot shows that spectra are grouped by compaction force and also by the level of strain applied to the blends. The compaction force is the principal source of variation as indicated by PC1. The differences in compaction force were expected since several studies have also shown that compaction pressure affects NIR spectra.^17–22 However, the effect of strain is a new observation discussed in this study. At 8 kN the spectra from the three strain levels are clearly separated confirming the results shown in Figure 3(a). At 12 and 16 kN the scores from the 160 and 640 revolutions are very similar, but they are separated from those obtained for no shear tablets. The PC1 loading represents the compaction force as first source of variation. This loading has a correlation coefficient with the NIR spectra of 0.9954 for 8 kN, 0.9962 for 12 kN, and 0.9967 for 16 kN tablets describing that the spectra are changing with respect to the compaction force. The PC2 loading shows that shear is a second source of variation. This loading presents two bands at 7450–7000 cm⁻¹ and 5600–5100 cm⁻¹ regions describing the change in the spectra related to the shear shown in Figure 5 and does not correlate with any of the pure component in the formulation. Table 3 shows the correlation coefficients of PC1 and PC2 with the pure components and tablets with different compaction forces. The spectra discussed were all obtained from tablets prepared under the conditions described in Table 1 for Case II. Near-infrared spectra of blends for Case II were also obtained. These blends did not reveal the spectral variation observed in the 7450–7000 cm⁻¹ and 5600–5100 cm⁻¹ regions.

Table 3.

Correlation coefficients of PC1 and PC2 loadings with Case II pure components and tablets with different compaction forces.

	Pure components				Tablets (kN)
	APAP	Lactose	MCC	MgSt	8	12	16
PC1	0.9387	0.9834	0.9805	0.8536	0.9954	0.9962	0.9967
PC2	−0.0828	−0.0975	−0.0642	−0.0393	−0.0620	−0.0578	−0.0684

APAP, acetaminophen; MCC, microcrystalline cellulose; MgSt, magnesium stearate.

The effect of a number of pretreatment methods on the NIR spectra was also evaluated for Case II. Spectral pretreatments are often used to remove or reduce the effects of light scattering on NIR spectra.²² The different spectral pretreatments removed the effect of tablet compaction forces, but did not remove the effect of strain level from the NIR spectrum. Table 4 shows the results obtained for several spectral pretreatments.

Table 4.

Different spectral pretreatment of Case II spectra evaluated with PCA.

Pretreatment (all spectral regions)	PC1		PC2
Pretreatment (all spectral regions)	Property	Explained variance	Property	Explained variance
Baseline corrected	Compaction	94%	Shear	5%
SNV	Shear	79%	No clear grouping	12%
SNV + 1st Der. (15 points)	Shear	91%	Compaction	6%
SNV + 2nd Der. (15 points)	Shear	96%	Compaction	2%

The effect of strain was also evaluated in the Case III formulation. In this case the differences observed in the 7500–7000 cm⁻¹ and 5600–5100 cm⁻¹ regions were not observed as shown in Figure 6. However, subtle differences were observed in the diffuse reflectance region of 7000–5000 cm⁻¹ and PCA showed four distinct groups of scores associated to the level of strain. Once again transmittance spectra were very similar and did not reveal spectral differences as the level of strain varied. Figure 7 shows the PCA scores plot for the spectra from the Case III tablets. The PC1 loading shows the variation for changes in shear. This loading has a correlation with the tablet spectra of 0.9943 describing the changing spectra with respect to shear. This second formulation was only compacted at 24 kN and an additional shear level was introduced (2560 revolutions). However, the effect of strain is clearly observed in the NIR spectra. The spectral differences between Cases II and III are obviously related to the chemical differences since Case III does not have MCC incorporated in the formulation. Hydrogen bonding affects NIR spectra²⁸ and a recent study indicates that MCC has a strain rate sensitivity as a viscoelastic excipient increasing the elastic effect at high-speed direct compression.³⁰ This critical material attribute of MCC was observed in a mechanical process and now observed using NIR spectroscopy. Microcrystalline cellulose appears to be a material that is affected by the shearing process and these effects are observed in the NIR spectra.

Near-Infrared Chemical Imaging

Near-infrared chemical imaging was used to investigate whether changes in the slope of the spectra are related to the shear effect in the tablets. The slope image was calculated using the first image plane of the spectral hypercube and the correspondent absorbance intensity with the last image plane and correspondent intensity. The slope calculation was used to determine the slope value with Eq. 2 and image statistics are summarized in Table 5.

m = \frac{A_{2} - A_{1}}{λ_{2} - λ_{1}}

(2)

Table 5.

Results of the slope calculation for three shear levels and three different compaction forces.

Shear (rev)	Compaction force (kN)	Mean slope (10⁻³)	Std. dev (10⁻⁴)	Skew	Kurtosis
0	8	2.47	2.63	0.43	6.19
0	12	2.79	2.95	0.44	6.11
0	16	3.07	2.95	0.01	6.88
160	8	2.45	3.29	0.62	5.17
160	12	2.74	3.36	0.61	5.51
160	16	3.03	3.35	0.58	5.9
640	8	2.4	3.27	0.67	5.3
640	12	2.77	3.33	0.49	7.04
640	16	3.01	3.14	0.17	5.69

When A₁ = intensity of first image plane, A₂ = intensity of last image plane, λ₁ = first wavelength, (1340 nm), and λ₂ = last wavelength (2220 nm). In this way, differences in baseline do not affect the calculations.

A two-factor analysis of variance (ANOVA) was performed to evaluate the variation in the slope obtained throughout the tablet surface. Table 6 summarized ANOVA results that show significant differences of the mean slope of the tablets according to the level of compaction.¹⁹ However, the evaluation showed that the spectral changes were not related to changes in slope.

Table 6.

Results of ANOVA two-factor analysis for three levels of shear and three different compaction forces.

Source of variation	SS	DF	MS	F	P value	F_crit
Compaction	3.77 × 10⁻⁰⁷	2	1.89 × 10⁻⁰⁷	22.03	0.01	6.94
Shear	1.36 × 10⁻⁰⁸	2	6.8 × 10⁻⁰⁹	0.79	0.51	6.94
Error	3.43 × 10⁻⁰⁸	4	8.57 × 10⁻⁰⁹
Total	4.25 × 10⁻⁰⁷	8

Angle-Resolved Elastic Light Scattering

Hemispherical measurements of scattered light were performed in order to gain knowledge about material morphology. Pharmaceutical tablets are strongly diffuse scatterers with particles of different size. For particles larger than the wavelength of incident light, the analysis of scattering pattern is more complex because the phase and amplitude of the scattered light changes as a function of particle geometry and orientation. Nevertheless, the angular distribution of scattered light and its integral intensity are directly related to the size and distribution of particles in the sample. In terms of single scattering approach, the scattering signal can be mapped versus spatial frequency f = sin θ/λ of surface relief Fourier decomposition.^24,31 Here, the surface relief is represented as a sum of diffraction gratings with different periods, amplitudes, phases, and orientations. Therefore, statistical information about grain size and local morphology can be obtained from angle-resolved light scattering data. For pharmaceutical tablets the scattering problem is complicated by multiple scattering processes inside the material, in which the correlation between directions of final scattering act and incident light is partly lost.

Figure 8 shows log (BSDF $(ϕ, θ)$ ) scattering indicatrices for the tablets with different compaction force and shear treatment. Here the scattering intensity gradually decreases as the rotation or compression is applied. Also, these treatments noticeably increase the degree of scattering isotropy.

The data in Figures 8 and 9 indicate that the compaction modifies the sample morphology. The compaction reduces the average distance between particles and aligns each particle in respect to the others. Since the orientation of particles becomes strongly dependent on surrounding and on direction of compaction force, a fluctuation of material density decreases. As a result, scattering intensity also decreases due to higher sample homogeneity. The most significant influence of the compaction force on scattering indicatrix can be observed for the samples without shear treatment. Thus, the cross-section of scattering indicatrix for 8 kN compaction shows nearly constant value of BSDF (θ) up to θ = 75° as show in Figure 9(a). This fact indicates that the untreated tablet scatters light highly diffusively, almost as Lambertian source.^25,26,31 As compaction increases, the scattering signal diminishes with noticeable change at larger polar angles (θ > 40°) due to mutual ordering of smaller particles with spatial frequencies above f ≈ 1 µm⁻¹. We note that for all types of samples (Figure 9a to c) the tilt of BSDF (θ) gradually increases with compaction force, indicating the suppression of diffuse component of scattering.

The shear treatment de-agglomerates the constituent powder, reducing the average size of microparticles.²⁷ As a result, this treatment reduces the scattering intensity. Thus, 160 revolutions (Figure 9b) increase the tilt of BSDF (θ), and after 640 revolutions the BSDF (θ) drops to its minimal level (Figures 8 and 9c). Moreover, the scattering angular isotropy increases with the number of revolutions (Figure 8). This indicates that the shear treatment not only reduces the particle size but also provides better mixing of constituent powder. The posterior compaction produces further ordering of the structure, resulting in gradual decrease of scattering intensity.

Conclusion

This study shows that shear stress applied to the powder mixtures affects diffuse reflectance spectra and the angle-resolved light scattering of tablets. The spectral changes related to shear stress were not expected, unlike spectral changes related to compaction force which have been observed by several researchers. The NIR diffuse reflectance spectra of blends (prior to compaction) do not show changes associated with shear. However, for tablets the reflectance spectra noticeably change when the shear stress is applied, while the transmission spectra remain the same. In the transmission spectra the NIR radiation interacts with a greater fraction of the tablet. Therefore, the spectra are less affected by the tablet surface. Thus, the NIR diffuse reflectance spectra indicate differences in the tablet surface which were then investigated by angle-resolved light scattering.

The angle-resolved light scattering measurements show significant changes in the material morphology after different levels of shear were applied. The compaction realigns microparticles and reduces the distance between them. As a result, scattering intensity decreases with noticeable change in the scattering indicatrix. The shear stress treatment reduces the particle size and provides better mixing of constituent powder which reduces scattering but increases its isotropy. Angle-resolved light scattering confirms the differences in tablet properties first observed by NIR spectroscopy.

Footnotes

Acknowledgments

The authors thank Drs. Zhenqi Shi and Don Dahm for helpful discussion on light scattering and NIR spectroscopy throughout the development of the manuscript.

Conflict of Interest

The authors report there are no conflicts of interest.

Funding

The authors acknowledge the support of the National Science Foundation through grant no. 1237873 and Grant ECC-0540855.

References

Alcalà

Blanco

Menezes

J.C.

Felizardo

P.M.

Garrido

Pérez

Zamora

Pasquini

Romañach

R.J.

“Near-Infrared Spectroscopy in Laboratory and Process Analysis”. In: Meyers

R.A.

(eds). Encyclopedia of Analytical Chemistry, Chichester, UK: John Wiley and Sons, Ltd, 2012.

Vanarase

Alcalà

Jerez Rozo

J.I.

Muzzio

F.J.

Romañach

R.J.

. “Real-Time Monitoring of Drug Concentration in a Continuous Powder Mixing Process Using NIR Spectroscopy”. Chem. Eng. Sci 2010; 65(21): 5728–12.

Ward

H.W.

Blackwood

D.O.

Polizzi

Clarke

. “Monitoring Blend Potency in a Tablet Press Feed Frame Using near Infrared Spectroscopy”. J. Pharm. Biomed. Anal 2013; 80: 18–23.

Shi

Cogdill

R.P.

Short

S.M.

Anderson

C.A.

. “Process Characterization of Powder Blending by Near-Infrared Spectroscopy: Blend End-Points and Beyond”. J. Pharm. Biomed. Anal 2008; 47(4–5): 738–745.

Igne

Zacour

B.M.

Shi

Talwar

Anderson

C.A.

Drennen

J.K.

. “Online Monitoring of Pharmaceutical Materials Using Multiple NIR Sensors—Part I: Blend Homogeneity”. J. Pharm. Innov 2011; 6(1): 47–59.

Green

R.L.

Thurau

Pixley

N.C.

Mateos

Reed

R.A.

Higgins

J.P.

. “In-Line Monitoring of Moisture Content in Fluid Bed Dryers Using Near-IR Spectroscopy with Consideration of Sampling Effects on Method Accuracy”. Anal. Chem 2005; 77(14): 4515–4522.

Barajas

M.J.

Cassiani

A.R.

Vargas

Conde

Ropero

Figueroa

Romañach

R.J.

. “Near-Infrared Spectroscopic Method for Real-Time Monitoring of Pharmaceutical Powders During Voiding”. Appl. Spectrosc 2007; 61(5): 490–496.

Kaye

B.H.

. Powder Mixing, Amsterdam, The Netherlands: Springer, 1997.

Lacey

P.M.C.

. “Developments in the Theory of Particle Mixing”. J. Appl. Chem 1954; 4(5): 257–268.

10.

Osorio

J.G.

Stuessy

Kemeny

G.J.

Muzzio

F.J.

. “Characterization of Pharmaceutical Powder Blends Using in Situ Near-Infrared Chemical Imaging”. Chem. Eng. Sci 2014; 108: 244–257.

11.

Hogg

. “Mixing and Segregation in Powders: Evaluation, Mechanisms and Processes”. KONA Powder Part. J 2009; 27: 3–17.

12.

Pingali

Mendez

Lewis

Michniak-Kohn

Cuitino

Muzzio

. “Evaluation of Strain-Induced Hydrophobicity of Pharmaceutical Blends and its Effect on Drug Release Rate Under Multiple Compression Conditions”. Drug. Dev. Ind. Pharm 2011; 37(4): 428–435.

13.

Fukui

Miyamura

Kobayashi

. “Effect of Magnesium Stearate or Calcium Stearate as Additives on Dissolution Profiles of Diltiazem Hydrochloride from Press-Coated Tablets with Hydroxypropylmethylcellulose Acetate Succinate in the Outer Shell”. Int. J. Pharm 2001; 216(1–2): 137–146.

14.

Aoshima

Miyagisnima

Nozawa

Sadzuka

Sonobe

. “Glycerin Fatty Acid Esters as a New Lubricant of Tablets”. Int. J. Pharm 2005; 293(1–2): 25–34.

15.

Willetts

J.P.

Robbins

P.T.

Roche

T.C.

Bowley

Bridson

R.H.

. “Exploring the Effects of High Shear Blending on Lactose and Drug Using Fluidized Bed Elutriation”. Int. J. Pharm 2012; 434(1–2): 272–279.

16.

Hussain

M.S.H.

York

Timmins

. “A Study of the Formation of Magnesium Stearate Film on Sodium Chloride Using Energy-Dispersive X-Ray Analysis”. Int. J. Pharm 1988; 42(1–3): 89–95.

17.

Short

S.M.

Cogdill

R.P.

Wildfong

P.L.

Drennen

J.K.

Anderson

C.A.

. “A Near-Infrared Spectroscopic Investigation of Relative Density and Crushing Strength in Four-Component Compacts”. J. Pharm. Sci 2009; 98(3): 1095–1109.

18.

Otsuka

Tanabe

Osaki

Otsuka

Ozaki

. “Chemoinformetrical Evaluation of Dissolution Property of Indomethacin Tablets by Near-Infrared Spectroscopy”. J. Pharm. Sci 2007; 96(4): 788–801.

19.

Ropero

Colon

Johnson-Restrepo

Romañach

R.J.

. “Near-Infrared Chemical Imaging Slope as a new Method to Study Tablet Compaction and Tablet Relaxation”. Appl. Spectrosc 2011; 65(4): 459–465.

20.

Ellison

C.D.

Ennis

B.J.

Hamad

M.L.

Lyon

R.C.

. “Measuring the Distribution of Density and Tableting Force in Pharmaceutical Tablets by Chemical Imaging”. J. Pharm. Biomed. Anal 2008; 48(1): 1–7.

21.

Blanco

Alcalá

. “Content Uniformity and Tablet Hardness Testing of Intact Pharmaceutical Tablets by Near Infrared Spectroscopy”. Anal. Chim. Acta 2006; 557(1–2): 353–359.

22.

Alcalà

Ropero

Vázquez

Romañach

R.J.

. “Deconvolution of Chemical and Physical Information from Intact Tablets NIR Spectra: Two- and Three-way Multivariate Calibration Strategies for Drug Quantitation”. J. Pharm. Sci 2009; 98(8): 2747–2758.

23.

Mehrotra

Llusa

Faqih

Levin

Muzzio

F.J.

. “Influence of Shear Intensity and Total Shear on Properties of Blends and Tablets of Lactose and Cellulose Lubricated with Magnesium Stearate”. Int. J. Pharm 2007; 336(2): 284–291.

24.

J.C. Stover. “Optical Scattering: Measurement and analysis”. In: Society of Photo-optical Instrumentation. Bellingham, WA: SPIE Optical Engineering Press, 1995.

25.

P.R. Griffiths, D.J. Dahm. “Continuum and Discontinuum Theories of Diffuse Reflection”. In: D.A. Burns, E.W. Ciurczak, editors. Handbook of Near-Infrared Analysis. Boca Raton, FL: CRC Press, 2007, Ch. 3. Pp. 21.

26.

Born

Wolf

Bhatia

A.B.

Gabor

Stokes

A.R.

Taylor

A.M.

Wayman

P.A.

Wilcock

W.L.

. Principles of Optics: Electromagnetic Theory of Propagation, Interference and Diffraction of Light, Cambridge, UK: Cambridge University Press, 2000.

27.

Llusa

Levin

Snee

R.D.

Muzzio

F.J.

. “Shear-Induced APAP De-Agglomeration”. Drug Dev. Ind. Pharm 2009; 35(12): 1487–1495.

28.

Watanabe

Morita

Ozaki

. “Temperature-Dependent Structural Changes in Hydrogen Bonds in Microcrystalline Cellulose Studied by Infrared and Near-Infrared Spectroscopy with Perturbation-Correlation Moving-Window Two-Dimensional Correlation Analysis”. Appl. Spectrosc 2006; 60(6): 611–618.

29.

Awa

Shinzawa

Ozaki

. “Monitoring of Recrystallization of Microcrystalline Cellulose Inside Pharmaceutical Tablets During Storage Using Near Infrared Diffuse Reflectance Spectroscopy”. J. Near Infrared Spectrosc 2014; 22(3): 205–210.

30.

Thoorens

Krier

Leclercq

Carlin

Evrard

. “Microcrystalline Cellulose, a Direct Compression Binder in a Quality by Design Environment - a Review”. Int. J. Pharm 2014; 473(1–2): 64–72.

31.

C.F. Bohren, D.R. Huffman. Absorption and Scattering of Light by Small Particles. New York, NY: Wiley, 1983.