Gaining insight into tablet capping tendency from compaction simulation

Accepted Manuscript Title: Gaining insight on tablet capping tendency from compaction simulation Authors: Shubhajit Paul, Changquan Calvin Sun PII: DOI: Reference:

S0378-5173(17)30255-7 http://dx.doi.org/doi:10.1016/j.ijpharm.2017.03.073 IJP 16545

To appear in:

International Journal of Pharmaceutics

Received date: Revised date: Accepted date:

9-11-2016 14-3-2017 26-3-2017

Please cite this article as: Paul, Shubhajit, Sun, Changquan Calvin, Gaining insight on tablet capping tendency from compaction simulation.International Journal of Pharmaceutics http://dx.doi.org/10.1016/j.ijpharm.2017.03.073 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Gaining insight on tablet capping tendency from compaction simulation

Shubhajit Paul and Changquan Calvin Sun*

Pharmaceutical Materials Science and Engineering Laboratory, Department of Pharmaceutics, College of Pharmacy, University of Minnesota, 9-127B Weaver-Densford Hall, 308 Harvard Street S.E., Minneapolis, MN 55455

*Corresponding author Changquan Calvin Sun, Ph.D. 9-127B Weaver-Densford Hall 308 Harvard Street S.E. Minneapolis, MN 55455 Email: [email protected] Tel: 612-624-3722 Fax: 612-626-2125

1

Graphical abstract

ABSTRACT

Capping or lamination is an unsolved common problem in tablet manufacturing. Knowledge gaps remain despite an enormous amount of effort made in the past to better understand the tablet capping/lamination phenomenon.

Using acetaminophen - containing

formulations, we examined the potential use of a compaction simulator as a material-sparing tool to predict capping occurrence under commercial tableting conditions. Systematical analyses of the in-die compaction data led to insight on the potential mechanism of tablet capping/lamination. In general, capping strongly correlates with high in-die elastic recovery, high Poisson’s ratio, low tensile strength, and radial die-wall pressure. Such insight can be used to guide the formulation design of high quality tablet products that are free from capping problems for challenging active pharmaceutical ingredients.

2

Keywords: Capping, lamination, tableting, mechanical property, compaction simulation, formulation development.

3

1. Introduction Over 70% of the marketed drugs have been formulated as tablets as it is a simple and cost effective solid dosage form (Gupta et al., 2009). Although the process of tablet die compression is straightforward, successful manufacturing of intact tablets with sufficient mechanical strength is not always possible.

Mechanical strength of a tablet depends on the interplay between

bonding area and bonding strength among particles (Osei-Yeboah et al., 2016).

The

development of inter-particulate bonding area is, in turn, influenced by compaction pressure, tableting speed, and the mechanical properties of constituent particles (Bag et al., 2012; Hiestand, 1997; Krishna et al., 2015; Tye et al., 2005). In the process of tablet compression, the tablet is compressed by the upper and lower punches in the vertical (axial) direction but radially confined by a rigid die wall. Factors that can influence the mechanical properties of finished tablets include mechanical properties of the particles, (e.g., plasticity, elasticity, and brittleness) (Sun, 2009), environmental factors, e.g., temperature and humidity (Osei-Yeboah et al., 2016; Sun, 2008), and process conditions, e.g., compression speed, pressure, and tooling design (Elowe et al., 1954; Tye et al., 2005). In general, the irreversible plastic deformation and brittle fragmentation favor tablet formation but the reversible elastic deformation does not. The stored elastic energy during the compression phase is recovered axially during decompression and radially when tablet exits the die during ejection.

Due to anisotropic properties of most

pharmaceutical crystals and the non-uniform stress distribution in the powder bed during compression, tablet structure is usually anisotropic (Sun, 2016). If not properly treated, such structure anisotropy may lead to undesired consequence in tablet integrity and performance. Common problems in tablet manufacturing include insufficient mechanical strength (Hiestand, 1997), sticking to punches (Booth and Newton, 1987), uncontrolled dissolution 4

(Basalious et al., 2011), and capping or lamination (Fassihi and Parker, 1986). The typical symptom of capping is the complete removal of top part of the tablet upon ejection or in subsequent handling and physical testing (Garr and Rubinstein, 1991). A closely related tablet manufacturing problem is lamination, ranging from a presence of micro cracks visible on the side of a tablet to multiple separated layers. With the current emphasis on quality by design of pharmaceutical products (Yu et al., 2014), the need for mechanistic understanding of tablet capping is now more urgent. The problem of capping or lamination has been recognized since the birth of powder tableting technology. Various factors have been suggested to play a role in tablet capping, including tableting speed (Ruegger and Celik, 2000), tooling design (Sugimori and Kawashima, 1997; Sugimori et al., 1989), air entrapment (Mazel et al., 2015), uneven stress distribution within tablet (Wu et al., 2008), and deformation characteristics of the powder (Akseli et al., 2013; Akseli et al., 2014; Shotton and Ganderton, 1961). Computational modeling of powder compaction has also been employed to elucidate capping mechanisms based on the changes in tablet microstructure and stress distribution (Sinka et al., 2004; Wu et al., 2008). Despite the extensive research efforts, gaps in our understanding of this phenomenon still exist. There remains a strong need of a reliable, material-sparing, and easy-to-perform method, which can be used to gain more insight into the phenomenon of tablet capping at an early stage of tablet product development. One technique that likely meets that need is compaction simulation, which can be used to study powder compaction behavior on different tablet presses and using different compression parameters while using only a small amount of powder (Michaut et al., 2010). Using an instrumented die, useful information from in-die data could be obtained to characterize a number of elastic properties of powders (LaMarche et al., 2014; Mazel et al., 2012; Wu et al., 2008; Wu 5

et al., 2005), which are expected to be relevant to study capping propensity of tablets. This study is aimed at evaluating the suitability of compaction simulation for predicting tablet capping propensity under realistic compaction conditions and identifying its possible correlation with indie parameters obtained from compaction simulation. Such knowledge can be used to guide formulation optimization to eliminate the tablet capping problem.

2. Materials and methods 2.1. Materials Acetaminophen (ACM, Form I) (Sigma Aldrich, St Louis, MO), a well-known poorly compressible drug, was used as a model drug to study the capping phenomenon. Microcrystalline cellulose (MCC, Avicel PH102; FMC Biopolymer, Philadelphia, PA) and magnesium stearate (Mallinckrodt, St Louis, MO) were used as tablet excipients.

2.2. Methods 2.2.1.

Preparation of direct compression formulation Binary mixtures of ACM and MCC were prepared with 25% increments in ACM loading.

All powders were lubricated with a fixed concentration of magnesium stearate (0.5%, w/w). All ingredients were passed through a #30 mesh sieve and blended in a mixer (Turbula, Glen Mills Inc., Cliffton, NJ) for 2 min at 100 rpm. 2.2.2.

Determination of bulk and tapped density Each powder blend (̴ 10 g) was poured in a graduated glass cylinder and its volume was

recorded. The cylinder was dropped on to a padded bench top from a height of approximately 2 cm. Powder volume was determined after 50 taps. The process was repeated until the difference

6

in tapped volume was < 2.0% between two consecutive volume readings. Bulk and tapped density were calculated based on the untapped and tapped volumes, respectively. All measurements were triplicated. The ambient relative humidity ranged 32-35% throughout this study. 2.2.3.

Tableting parameters The influence of both ACM loading and tableting speed, measured by dwell time, on

capping tendency was studied using a fully instrumented compaction simulator, equipped with an instrumented die (Presster, Metropolitan Computing Corporation, NJ) simulating a Korsch XL 400 press (29 stations) using 9.5 mm round flat-faced punches. For each batch, 10-12 tablets were obtained over a pressure range of 25-300 MPa. Maximum die-wall pressure (MDP), residual die-wall pressure (RDP), and ejection force (EF) for each tablet were recorded. Initially, formulations containing ACM in 25% increments were studied at 10 ms dwell time (103,000 tablets/h), using pure MCC as a control. Subsequently, the 75% ACM loaded formulation was further studied at three additional tableting speeds, i.e., dwell times of 15 ms, 25 ms, and 100 ms. 2.2.4.

Determination of friction coefficient The interaction between tablet and die-wall during ejection was quantified in term of

friction coefficient (µ) according to Eqn. (1) (Sun, 2015):

𝜇=

𝐸𝐹 𝜋 ∙ 𝑅𝐷𝑃 ∙ 𝐷 ∙ ℎ′

(1)

where D and h' are the tooling diameter and in-die tablet thickness at the end of decompression phase, respectively.

It is possible that tablet can still slightly expand axially after the

7

decompression phase. If this happens, the actual µ would be lower than that calculated using Eqn. (1). However, this effect is expected to be small. 2.2.5.

Determination of radial tensile strength Tablets obtained after compaction were immediately subjected to diametrical breaking

test using a texture analyzer (Texture Technologies Corp., Surrey, UK). The radial tablet tensile strength (TS) was calculated according to Eqn. (2) (Fell and Newton, 1970).

𝑇𝑆 =

2𝐹 𝜋∙𝑑∙𝐻

(2)

where F, d, and H are the breaking force, tablet diameter, and thickness, respectively. The TS of capped tablets were also determined by gently rubbing the intact portion of the tablets on a fine sand paper (superfine grade - P400, 3M Inc., Saint Paul, MN) to obtain approximately cylindrical tablets. Intact tablets were also polished by the sand paper to remove possible flashing for accurate thickness determination. The sand paper treatment is critical for calculating accurate tablet porosity, especially for plastic materials compressed at a high pressure (Paul et al., 2017). 2.2.6.

Determination of true density of pure components and mixtures The true density of ACM (1.294 g/mL) was calculated from its room temperature single

crystal structure (Haisa et al., 1976). The true density of MCC equilibrated at 30% RH (1.467 g/mL) was obtained from the literature (Sun, 2008). The true density of mixtures (ρn) was obtained according to Eqn. (3) (Sun, 2004): 𝑛

1 𝑥𝑖 =∑ 𝜌𝑛 𝜌𝑖

(3)

𝑖=1

8

where x and n refer to the weight fraction and number of constituents in the mixture, respectively. 2.2.7.

In-die Heckel analysis Heckel analysis was performed using the in-die data since out-of-die data could not be

obtained for capped tablets. The Heckel equation (Eqn. 4) describes the change of compact porosity (ε) with compaction pressure (P) (Heckel, 1961b): −𝑙𝑛ε = 𝑘 ∙ 𝑃 + 𝐴

(4)

where the reciprocal of the slope, known as mean yield pressure (Py), has been used to assess plasticity of a material (Heckel, 1961a). Linear portions of the Heckel plots in the high pressure region (R2 > 0.99) were used for calculating Py. 2.2.8.

Determination of elastic parameters from in-die data For determination of elastic parameters, an instrumented die was used. Die-wall stress

was measured using a strain gauge, which was calibrated using poster putty (3M, Minneapolis, MN), for which the axial to radial stress ratio was assumed to be the same.

Instrument

deformation as a function of pressure was measured by compressing a rigid metal tablet, which was then automatically applied to correct collected raw data. A fixed punch gap (3.0 mm) was maintained to minimize error in die wall stress measurements while pressure was varied by controlling the powder fill weight. The full form of the linear elasticity equations, Eqns. (5) and (6), were used for extracting elastic parameters from the in-die powder bed height, axial stress, and radial stress data, assuming the linear stress-strain relationship is followed during unloading. (Mazel et al., 2012).

9

𝜀𝑎𝑥 =

1 (𝜎 − 2𝜈 · 𝜎𝑟𝑎𝑑 ) 𝐸 𝑎𝑥

𝜀𝑟𝑎𝑑 =

(5)

1 [𝜎 − 𝜈 · (𝜎𝑎𝑥 + 𝜎𝑟𝑎𝑑 )] 𝐸 𝑟𝑎𝑑

(6)

Where 𝜈 is Poisson’s ratio, E is Young’s modulus, and G is shear modulus, εax and εrad are axial and radial strains, σax and σrad are axial and radial stresses, respectively. Rearranging Eqn. (6) gives Eqn. (7) 𝜎𝑟𝑎𝑑 = 𝜈 · (𝜎𝑟𝑎𝑑 + 𝜎𝑎𝑥 ) + 𝐸 · 𝜀𝑟𝑎𝑑

(7)

Thus, the ν value may be obtained from the slope of σrad vs. (σrad + σax) plot according to Eqn. (7). Rearranging Eqn. (5) gives Eqn. (8) (𝜎𝑎𝑥 − 2. 𝜐. 𝜎𝑟𝑎𝑑 ) = 𝐸. (

ℎ0 −ℎ ℎ0

ℎ

) = 𝐸 − 𝐸. ℎ

0

(8)

Where h and h0 are the in-die thickness under pressure and the minimum thickness at the maximum compaction pressure, respectively. Therefore, the E value may be calculated from the intercept of the plot of (𝜎𝑎𝑥 − 2𝜐𝜎𝑟𝑎𝑑 ) vs. h, according to Eqn. (8). To satisfy the condition of linear elastic deformation, only data giving R2 ≥ 0.999 were used for linear fitting to calculate E and ν. The value of G was calculated from E and ν from Eqn. (9) (Wu et al., 2005):

𝐺=

𝐸 2(1 + 𝜈)

(9)

10

3. Results and discussion The size of ACM particles was 10 - 150 μm (Fig. 1a). It was shown that Form I ACM did not exhibit acceptable tabletability, due to a lack of plasticity resulting from its crystal structure of stacking corrugated H-bonded rigid layers (Shi and Sun, 2011). In this study, capping was observed for the mixture containing 75% of ACM at the highest tableting speed, i.e., 10 ms (Fig. 1b), while both capping and lamination was observed for pure ACM. 3.1. Compression phase Parameters collected from compression phase were examined for possible correlation with tablet capping. Poor powder consolidation could result in mechanically weak tablets, which is expected to aggravate capping tendency. The compressibility profiles of formulations with varying ACM loadings at 10 ms dwell time are shown in Fig. 2a. The in-die porosity at zero compaction pressure tends to decrease as ACM loading increases (Fig. 2a). This is consistent with the decreasing porosity of tapped powders with increasing ACM loading (Fig. 2b), which indicates improved packing efficiencies of the powder containing more ACM. It is interesting that bulk densities of these powders were nearly identical but tapped densities increased at higher ACM loading (Fig. 2c). The calculated tapped porosity of the powder bed decreased with increase ACM loading since their true densities did not differ significantly (Fig. 2c). The in-die tablet porosity of 75% loaded ACM formulation consistently increased with decreasing dwell time, i.e., increasing tableting speed (Fig. 2d). Hence, compressibility was poorer at higher tableting speed, which is expected to favor the occurrence of capping.

11

When ACM content was varied, the initial linear portion of the in-die Heckel profile shifted vertically because of difference in packing due to different bulk densities (Fig. 3a). When pressure increased, crossover was observed which gave rise to different slopes when the linear portion of the plot was considered. Py increased with increasing ACM concentration, implying poor consolidation (Table 1). For the 75% ACM formulation, Heckel profiles were also sensitive to tableting speed (Fig. 3b), where Py decreased with longer dwell time (slower speed) (Table 1).

Thus, both ACM concentration and tableting speed affect the powder

consolidation behavior and play a critical role on capping tendency during the compression phase.

Die-wall stress is a useful parameter for characterizing powder compaction behavior (Doelker and Massuelle, 2004). Surprisingly, the MDP (peak die-wall pressure) profiles of formulations containing different concentrations of ACM or dwell times were identical (Fig. 4), implying that stress transmission within the tablet was independent of composition or speed during the compression phase. This observation suggests that tablet capping cannot be attributed to different degrees of stress transmission during compression phase. Thus, other parameters must be sought to explain different capping propensity of powders.

12

3.2. Decompression phase One parameter to characterize decompression phase is the in-die elastic recovery (IER), which is the percent increase of in-die tablet thickness when the axial pressure varied from the maximum value to zero. IER may be used to quantify the extent of elastic deformation of the compact in confined die at the end of decompression phase. IER increases with increasing compaction pressure in all cases (Fig. 5). Higher IER was observed at both higher ACM loading (Fig. 5a) and faster tableting speed (Fig. 5b), which corresponds well with higher propensity to capping (Table 1). Thus, higher in-die elastic relaxation, which leads to weaker inter-particulate strength in the tablet, appears to strongly correlate with tablet capping.

Another important parameter relevant to powder tableting performance is RDP. It has been observed that plastic or elastic materials tend to exhibit much lower RDP than brittle materials (Abdel-Hamid et al., 2012). A shorter dwell time (faster tableting) corresponded to a lower RDP (Fig. 6a) in this work despite the identical MDP at different speeds (Fig. 4b). Thus, it translates to greater elastic recovery during decompression at a higher speed since MDP is independent of speed (Fig. 4b). The change in RDP with ACM loading did not follow a monotonic trend. For powders containing up to 50% ACM, RDP was insensitive to increasing ACM content below 200 MPa. When pressure was greater than 200 MPa, RDP increased with increasing ACM loading.

However, RDP of the powders containing 75% and 100% ACM is

noticeably lower (Fig. 6b), which is likely a result of the occurrence of capping inside the die during decompression (Sugimori et al., 1990). Since MDP is also not affected by ACM loading (Fig. 4a), a lower RDP for formulation containing more ACM (75% and 100%) implies their

13

compacts underwent more elastic deformation during decompression.

This is consistent with

the higher IER under these conditions (Fig. 5a).

Interestingly, the powders containing 75% and 100% ACM exhibited both higher capping propensity and lower RDP among these powders (Fig. 6b). Thus, in addition to elastic recovery, capping during the decompression phase may also reduce RDP. The tablet strength encompasses the net effect of inter-particulate bond formation during compression and bond breakage due to elastic recovery during decompression. The dominance of the former or latter process will lead to strong and weak tablets, respectively. Tabletability worsened with increasing ACM concentration (Fig. 7a). In the case of 100% ACM, no intact tablet could be formed at all pressures. The tensile strength (TS) of 75% ACM formulation at 10 ms was obtained by first gently polishing the body of the capped tablets on a sand paper into an approximately cylindrical specimen and then tensile strength was determined. At lower tableting speeds, i.e., 15 ms, 25 ms, and 100 ms, the 75% ACM formulation did not show capping and improved tabletability profiles were obtained (Fig. 7b). It was shown before that a change in tableting speed affects tablet compressibility, i.e., tablet porosity as a function of compaction pressure (Tye et al., 2005). Thus, the higher TS at slower tableting speed (100 ms dwell time) in this study could have been attributed to lower tablet porosity. Compressibility plots at the three tableting speeds, at which intact tablets could be made, are shown in Fig. 8. Tablet porosities were comparable at pressures below 125 MPa. In this pressure range, TS is also independent of tableting speed. At pressures above 150 MPa, compressibility plots diverge as tablet porosity is lower at lower speeds when the pressure is the same. At 100 ms dwell time, the porosity is the lowest and the TS is the highest. Thus, the higher TS at this speed may be attributed to larger 14

bonding area. In this pressure range (> 150 MPa), tablet porosity at 15 ms dwell times is higher than that at and 25 ms dwell time (Fig. 8). However, TS does not appear significantly different (Fig. 7b). This can be explained if macro cracks are present in these tablets. If so, the maximum strength of these tablets was dictated by the cracks not bonding area. This speculation is supported by two observations: a) capping was visually observed at 10 ms dwell time, and b) TS decreased when pressure exceeded 300 MPa at 25 ms dwell time (Fig. 7b). In all cases, higher capping tendency invariably corresponds to weaker tablet.

Hence, mechanical strength is

another factor that correlates with the capping tendency.

3.3. Ejection phase The relevant parameters in the ejection phase, such as percent radial expansion (RE), ejection force (EF), and coefficient of static friction (µ) were also studied. RE increased with increasing ACM concentration (Fig. 9).

For each powder, RE gradually decreased with

increasing pressure (Fig. 8). This corresponds well with the higher capping propensity at higher ACM loadings (Table 1). Although RDP initially increased and then decreased for 0-75% ACM formulations, implying that the trend of radial elastic recovery was not monotonic, the corresponding proportionately greater RE suggest that compacts might had undergone greater shear failure after ejection. When tableting speed was varied for the powder containing 75% ACM, the data was too variable to draw a meaningful conclusion on its impact on RE (data not shown). Such erratic behavior likely reflected the high capping propensity of this powder (Table 1).

15

The integrity of the ejected tablets could be compromised if the EF is too high. Since EF is influenced by tablet size, such as diameter, thickness, and RDP, μ is a preferred parameter for characterizing tablet – die wall interactions over EF (Sun, 2015). Effects of increasing ACM loadings and tableting speed on μ are shown in Fig. 10. μ gradually increased from 0% to 50 % ACM loading, and then steadily decreased for 75% and 100% ACM powders (Fig. 10a). This trend is similar to that of RDP profiles (Fig. 6b). For tablets that did not cap, μ was lower at slower tableting speed when the dwell time was varied from 15 ms to 100 ms (Fig. 10b). The improved lubrication efficiency at lower speed could be attributed to the higher degree of migration of magnesium stearate to the tablet-wall interface (Sun, 2015).

3.4. Elastic parameters The relationship between important elastic properties of materials, i.e., ν, E and G, and capping was also examined. It should be noted that these parameters were calculated using Eqns. 7, 8, and 9, based on the assumption that the tablets behave isotropically within a plane. However, in reality, most of the pharmaceutical tablets demonstrate anisotropy, particularly at high pressures (Cunningham et al., 2004). Nevertheless, such information can still be useful in understanding deformation behavior during tableting (Cunningham et al., 2004; Wu et al., 2005). Even though errors in calculated parameters are expected, the impact of such systematic errors on the assessment of relative contributions by different variables, such as ACM loading and speed, is much less. Hence, it is useful to examine their possible connections with capping phenomenon. At each ACM concentration, ν gradually increased with increasing compaction pressure (Fig. 11a). At low pressures, higher ACM concentration corresponded to greater ν. However, ν 16

converged with increasing pressure because the rate of increase is lower at higher ACM concentrations. With increasing tableting speed (shorter dwell time), ν also increased over the whole pressure range (Fig. 11b).

Thus, RDP and ν follow an inverse relationship since

increasing tableting speed led to both lower RDP (Fig. 6a) and higher ν (Fig. 11b). Higher ACM loading or higher tableting speed led to both larger ν and greater propensity to capping. Therefore, capping propensity positively correlates with ν.

Both E and G depended on ACM loading and tableting speed in similar ways (Fig. 12). The trend of E data for different materials at 10 ms dwell time is complex (Fig. 12a). To explain the data, we recognize that E of MCC (6.7 GPa at zero porosity determined by nanoindentation (Govedarica et al., 2012)) is lower than E of ACM (8.4 GPa by microindentation or 8.87 GPa by nanoindentation of single crystals (Duncan-Hewitt and Weatherly, 1989; Liao and Wiedmann, 2005)), but MCC is also more plastic and compressible (lower porosity at the same pressure) than ACM. The measured E from in-die data depends on the interplay between these two factors. The observation that MCC tablets exhibited higher E than all other powders (Fig. 12a) indicates that its superior compressibility (Table 1) dominated the interplay by forming compacts with lower porosity to compensate its intrinsically lower E than ACM. The rank order of E of ACM containing powders reflected the interplay between tablet porosity and different E values of ACM and MCC, which depends on compaction pressure and percentage of ACM.

At

pressures higher than 150 MPa, E of powders containing 50% and 75% ACM were similar but lower than other powders (Fig. 12a). For the powder containing 75% ACM, higher tableting speed corresponded lower E. This may be explained by the fact that powder consolidation is less

17

effective at higher speeds. Hence, tablets are more porous. This is consistent with the higher Py at higher tableting speed (Fig. 3b). Shear modulus, G, quantifies a material’s ability to resist deformation under a shear stress in the transverse direction.

The G value followed similar trends as those of E as discussed above.

Therefore, similar explanations may be given. Although rarely discussed, G could be a useful parameter to characterize the ejection phases. During ejection, shear stress develops when the part of tablet outside the die expanded radially while the part inside the die is constrained by the die wall. The rise in RE of the ejected tablets with increasing ACM loadings (Fig. 9) is consistent with this explanation. Shear failure is more likely to occur when the radial expansion of an ejected tablet is larger. A non-destructive approach of evaluating capping propensity of tablets based on acoustic wave propagation was reported (Akseli et al., 2013; Akseli et al., 2014). The capping index, i.e., EG (= Eaxial/Eradial), was obtained from the E of the tablet in the axial and radial directions. It was concluded that anisotropy of E in axial and radial directions could be the cause of capping. In the acoustic wave propagation method, all the tablets obtained were intact after tableting but capping was observed during tablet diametric breaking test. Thus, this method does not consider effects of more challenging process conditions that lead to capping immediately after ejection. Consequently, it cannot be used to predict capping occurrence under those circumstances. When formulation and process parameters are varied, capping behavior can drastically change as observed in this work. Thus, the in-die data analysis method presented here is complimentary to the out-of-die acoustic method. The in-die approach is relatively simple to implement, provided a fully instrumented compaction simulator is available. However, before its adoption for routine

18

analysis, it would be prudent to examine the usefulness of this method in predicting capping behavior using a larger number of powders with varying intrinsic properties.

The results so far suggest that IER, RDP, TS, RE, ν, E, and G are important parameters that correlate with capping. However, due to the complexity of the capping phenomenon, some of these parameters could not clearly distinguish the presence or absence of capping, particularly at low compaction pressures. To identify the most significant factors among these, a linear regression analysis was conducted using all the data points (tablets) collected at various combinations of tableting speeds and ACM concentrations. For this purpose, a score of 10 was assigned for capped formulations while a score of 1 was assigned to capping-free formulations to quantify capping propensity (CP). This score values were similar to the range of cap values designated for formulations, ranging from capping to capping-free, in an earlier study (Akseli et al., 2013). A two-factor interaction model was found appropriate, which reveals that IER, RDP, TS and ν, are most statistically significant parameters at p < 0.05 level (Eqn. 10, adjusted R2 of 0.74). 𝑪𝑷 = −2.29 + 0.65 · 𝐼𝐸𝑅 − 8.5 · 𝑅𝐷𝑃 − 5.73 · 𝑇𝑆 + 6.24 ∙ 𝜈 − 5.44 · 𝑅𝐷𝑃 ∙ 𝑇𝑆 − 0.44 · 𝑇𝑆 ∙ 𝜈

(10)

The statistically significant terms identified in Eqn. (10) are consistent with the observed relationship between individual factors and capping as discussed above, where a greater IER and ν, and smaller RDP, or lower TS correlates to higher capping propensity.

Fig. 13 shows the

contour plots of the significant parameters, which clearly show the capping and non-capping regions in the IER - RDP (Fig. 13a) for tablets with TS of 2.0 MPa, which is a recommended tensile strength for tablets to avoid gross problem during conventional handling, shipping, and 19

storage (Gong et al., 2015). Fig. 13b shows the capping and non-capping zones in the TS-ν space when IER and RDP are kept unchanged (IER = 11% and RDP = 12 MPa).

It is generally known that the propensity of capping is high for oval, concave, or caplet shaped tablets because of non-uniform stress transmission during compression (Wu et al., 2008). Materials of high capping tendency could still cap even for cylindrical tablets. In such cases, which cannot be studied using the acoustic approach, investigation of the relevant parameters of capping could be effectively accomplished by the compaction simulation approach outlined here. In addition to providing insight on tablet capping, elastic parameters generated from this analysis are also of general importance for better understanding of powder compaction behaviors. 4. Conclusion Capping is a problem that affects tablet quality.

In this work, we have examined

potential connections between several parameters, accessible using a compaction simulator, and capping behavior of round flat-faced tablets. Using formulated ACM powders, we have shown that capping positively correlates with high IER and ν and low TS and RDP. Thus, compaction simulation can provide mechanistic insight on the capping problem. This supports the use of compaction simulation to guide the formulation optimization in an early stage of development to design high quality tablet products.

References Abdel-Hamid, S., Koziolek, M., Betz, G., 2012. Study of radial die-wall pressure during high speed tableting: effect of formulation variables. Drug Dev. Ind. Pharm. 38, 623-634. 20

Akseli, I., Ladyzhynsky, N., Katz, J., He, X.R., 2013. Development of predictive tools to assess capping tendency of tablet formulations. Powder Technol 236, 139-148. Akseli, I., Stecula, A., He, X.R., Ladyzhynsky, N., 2014. Quantitative Correlation of the Effect of Process Conditions on the Capping Tendencies of Tablet Formulations. J. Pharm. Sci. 103, 1652-1663. Bag, P.P., Chen, M., Sun, C.C., Reddy, C.M., 2012. Direct correlation among crystal structure, mechanical behaviour and tabletability in a trimorphic molecular compound. Cryst Eng Comm 14, 3865-3867. Basalious, E.B., El-Sebaie, W., El-Gazayerly, O., 2011. Application of Pharmaceutical QbD for Enhancement of the Solubility and Dissolution of a Class II BCS Drug using Polymeric Surfactants and Crystallization Inhibitors: Development of Controlled-Release Tablets. AAPS PharmSciTech 12, 799-810. Booth, S.W., Newton, J.M., 1987. Experimental Investigation of Adhesion between Powders and Surfaces. J. Pharm. Pharmacol. 39, 679-684. Cunningham, J.C., Sinka, I.C., Zavaliangos, A., 2004. Analysis of tablet compaction. I. Characterization of mechanical behavior of powder and powder/tooling friction. J. Pharm. Sci. 93, 2022-2039. Doelker, E., Massuelle, D., 2004. Benefits of die-wall instrumentation for research and development in tabletting. Eur. J. Pharm. Biopharm. 58, 427-444. Duncan-Hewitt, W.C., Weatherly, G.C., 1989. Evaluating the hardness, Young's modulus and fracture toughness of some pharmaceutical crystals using microindentation techniques. J. Mater. Sci. Lett. 8, 1350-1352. Elowe, L.N., Busse, L.W., Higuchi, T., 1954. The physics of tablet compression. V. Studies on aspirin, lactose, lactose–aspirin, and sulfadiazine tablets. J. Am. Pharm. Assoc. 11, 685-689. Fassihi, A.R., Parker, M.S., 1986. Formulation Effects on Capping Tendencies. Int. J. Pharm. 31, 271-273. Fell, J.T., Newton, J.M., 1970. Determination of tablet strength by the diametral-compression test. J. Pharm. Sci. 59, 688-691. Garr, J.S.M., Rubinstein, M.H., 1991. An investigation into the capping of paracetamol at increasing speeds of compression. Int. J. Pharm. 72, 117-122. Gong, X.C., Chang, S.Y., Osei-Yeboah, F., Paul, S., Perumalla, S.R., Shi, L.M., Sun, W.J., Zhou, Q., Sun, C.C., 2015. Dependence of tablet brittleness on tensile strength and porosity. Int. J. Pharm. 493, 208-213. Govedarica, B., Ilic, I., Sibanc, R., Dreu, R., Srcic, S., 2012. The use of single particle mechanical properties for predicting the compressibility of pharmaceutical materials. Powder Technol 225, 43-51. Gupta, H., Bhandari, D., Sharma, A., 2009. Recent trends in oral drug delivery: a review. Recent patents on drug delivery & formulation 3, 162-173. Haisa, M., Kashino, S., Kawai, R., maeda, H., 1976. The Monoclinic Form of pHydroxyacetanilide. Acta Crystallography B32, 1283-1285. Heckel, R.W., 1961a. An analysis of powder compaction phenomena. Trans Metall Soc AIME 221, 1001-1008. Heckel, R.W., 1961b. Density-Pressure Relationships in Powder Compaction Trans Metall Soc AIME 221, 671-675. Hiestand, E.N., 1997. Mechanical properties of compacts and particles that control tableting success. J. Pharm. Sci. 86, 985-990. 21

Krishna, G.R., Shi, L.M., Bag, P.P., Sun, C.C., Reddy, C.M., 2015. Correlation Among Crystal Structure, Mechanical Behavior, and Tabletability in the Co-Crystals of Vanillin Isomers. Crystal Growth & Design 15, 1827-1832. LaMarche, K., Buckley, D., Hartley, R., F., Q., Badawy, S., 2014. Assessing materials' tablet compaction properties using the Drucker–Prager Cap model. Powder Technol 267. Liao, X., Wiedmann, T.S., 2005. Measurement of process-dependent material properties of pharmaceutical solids by nanoindentation. J. Pharm. Sci. 94, 79-92. Mazel, V., Busignies, V., Diarra, H., Tchoreloff, P., 2012. Measurements of elastic moduli of pharmaceutical compacts: a new methodology using double compaction on a compaction simulator. J. Pharm. Sci. 101, 2220-2228. Mazel, V., Busignies, V., Diarra, H., Tchoreloff, P., 2015. Lamination of pharmaceutical tablets due to air entrapment: Direct visualization and influence of the compact thickness. Int. J. Pharm. 478, 702-704. Michaut, F., Busignies, V., Fouquereau, C., de Barochez, B.H., Leclerc, B., Tchoreloff, P., 2010. Evaluation of a rotary tablet press simulator as a tool for the characterization of compaction properties of pharmaceutical products. J. Pharm. Sci. 99, 2874-2885. Osei-Yeboah, F., Chang, S.Y., Sun, C.C., 2016. A critical Examination of the Phenomenon of Bonding Area - Bonding Strength Interplay in Powder Tableting. Pharm. Res. 33, 11261132. Paul, S., Chang, S.-Y., Sun, C.C., 2017. The phenomenon of tablet flashing — Its impact on tableting data analysis and a method to eliminate it. Powder Technol 305, 117-124. Ruegger, C.E., Celik, M., 2000. The effect of compression and decompression speed on the mechanical strength of compacts. Pharm. Dev. Technol. 5, 485-494. Shi, L., Sun, C.C., 2011. Overcoming poor tabletability of pharmaceutical crystals by surface modification. Pharm. Res. 28, 3248-3255. Shotton, E., Ganderton, D., 1961. The strength of compressed tablets.III. The relation of particle size, bonding and capping in tablets of sodium chloride, aspirin and hexamine. J. Pharm. Pharmacol. 13, 144-152. Sinka, I.C., Cunningham, J.C., Zavaliangos, A., 2004. Analysis of tablet compaction. II. Finite element analysis of density distributions in convex tablets. J. Pharm. Sci. 93, 2040-2053. Sugimori, K.-i., Kawashima, Y., 1997. A new practical index to predict capping occurring during the tableting process. Eur. J. Pharm. Biopharm. 44, 323-326. Sugimori, K.-I., Mori, S., Kawashima, Y., 1990. Application of a newly defined capping index in evaluation of the compressibility of pharmaceutical powders. Advanced Powder Technology 1, 25-37. Sugimori, K., Mori, S., Kawashima, Y., 1989. Characterization of die wall pressure to predict capping of flat- or convex-faced drug tablets of various sizes. Powder Technol 58, 259-264. Sun, C.C., 2004. A novel method for deriving true density of pharmaceutical solids including hydrates and water-containing powders. J. Pharm. Sci. 93, 646-653. Sun, C.C., 2008. Mechanism of moisture induced variations in true density and compaction properties of microcrystalline cellulose. Int. J. Pharm. 346, 93-101. Sun, C.C., 2009. Materials Science Tetrahedron-A Useful Tool for Pharmaceutical Research and Development. J. Pharm. Sci. 98, 1671-1687. Sun, C.C., 2015. Dependence of ejection force on tableting speed-A compaction simulation study. Powder Technol 279, 123-126.

22

Sun, C.C., 2016. Microstructure of Tablet-Pharmaceutical Significance, Assessment, and Engineering. Pharm. Res. Tye, C.K., Sun, C.C., Amidon, G.E., 2005. Evaluation of the effects of tableting speed on the relationships between compaction pressure, tablet tensile strength, and tablet solid fraction. J. Pharm. Sci. 94, 465-472. Wu, C.Y., Hancock, B.C., Mills, A., Bentham, A.C., Best, S.M., Elliott, J.A., 2008. Numerical and experimental investigation of capping mechanisms during pharmaceutical tablet compaction. Powder Technol 181, 121-129. Wu, C.Y., Ruddy, O.M., Bentham, A.C., Hancock, B.C., Best, S.M., Elliott, J.A., 2005. Modelling the mechanical behaviour of pharmaceutical powders during compaction. Powder Technol 52, 107-117. Yu, L.X., Amidon, G., Khan, M.A., Hoag, S.W., Polli, J., Raju, G.K., Woodcock, J., 2014. Understanding pharmaceutical quality by design. AAPS J 16, 771-783.

23

100 μm

Fig. 1. (a) Polarized light microscope image of ACM and (b) capped tablets (75% ACM formulation compressed at dwell time of 10 ms).

24

(b) 0% ACM 25% ACM 50% ACM 75% ACM 100% ACM

Porosity

0.6

0.4

0.2

Tapped porosity of powder bed

0.75

10 ms

0.0 50

0

100

150

0.60

0.45

0.30

0.15

0.00

200

0

20

Compaction pressure (MPa)

40

60

80

100

ACM content (%)

0.6

Bulk density Tapped density

75% ACM

0.6

0.5

0.5

0.4

Porosity

Density values (g/cc)

0.7

0.4 0.3

10 ms 15 ms 25 ms 100 ms

0.3

0.2 0.2

0.1

0.1 0.0 0%

25%

50%

ACM content

75%

100%

0.0 0

50

100

150

200

Compaction pressure (MPa)

Fig. 2. a) Compressibility profiles of binary mixtures at different ACM contents; (b) tapped powder porosity; (c) bulk and tapped density of powders at different ACM contents; and (d) effects of dwell time on in-die compressibility.

25

7 0% ACM 25% ACM 50% ACM 75% ACM 100% ACM

6

6

4 3 2

75% ACM

10 ms 15 ms 25 ms 100 ms

8

-ln(porosity)

-ln(porosity)

5

10 ms

4

2 1

0

0 0

50

100

150

0

200

50

Compaction pressure (MPa)

100

150

200

Compaction pressure (MPa)

Fig. 3. In-die Heckel profiles of (a) mixtures containing different ACM contents at 10 ms dwell time, and (b) a mixture containing 75% ACM at different dwell times.

0% ACM 25% ACM 50% ACM 75% ACM 100% ACM

150

10 ms

75% ACM

10 ms 15 ms 25 ms 100 ms

200

150

MDP (MPa)

MDP (MPa)

200

100

100

50

50

0

0 0

50

100

150

200

250

Compaction pressure (MPa)

300

0

50

100

150

200

250

300

Compaction pressure (MPa)

Fig. 4. Peak die-wall pressure for a) different powders at 10 ms dwell time, and (b) a powder containing 75% ACM at different dwell times.

26

20

20

10 ms

0% ACM 25% ACM 50% ACM 75% ACM 100% ACM

18 16

16 14

IER (%)

14

IER (%)

75% ACM

10 ms 15 ms 25 ms 100 ms

18

12 10

12 10 8

8

6 6

4 4

2 0

50

100

150

200

250

300

0

50

Compaction pressure (MPa)

100

150

200

250

300

Compaction pressure (MPa)

Fig. 5. Variations in in-die elastic recovery (IER) for binary mixtures at different (a) ACM contents and (b) dwell times.

18

22

75% ACM

10 ms 15 ms 25 ms 100 ms

20 18

0% ACM 25% ACM 50% ACM 75% ACM 100% ACM

16 14

RDP (MPa)

RDP (MPa)

16 14 12 10

10 ms

12 10 8

8

6 6

4

4 2

2 0

50

100

150

200

250

Compaction pressure (MPa)

300

0

50

100

150

200

250

300

Compaction pressure (MPa)

Fig. 6. Residual die-wall pressure of powders at different (a) dwell times and (b) ACM contents.

27

1.0

10

Tensile strength (MPa)

8

Tensile strength (MPa)

0% ACM 25% ACM 50% ACM 75% ACM

6

4

2

0 0

50

100

150

200

250

300

350

Compaction pressure (MPa)

10 ms 15 ms 25 ms 100 ms

0.8

75% ACM

0.6

0.4

0.2

0.0 50

100

150

200

250

300

350

Compaction pressure (MPa)

Fig. 7. The dependence of tabletability on a) ACM concentration (at 10 ms dwell time), and (b) dwell time (for a formulation containing 75% ACM). Tabletability of pure ACM is not shown since intact tablets could not be prepared under these compression conditions.

75% ACM

0.24

15 ms 25 ms 100 ms

Porosity

0.20

0.16

0.12

0.08

50

100

150

200

250

300

Compaction pressure (MPa)

Fig. 8. Effects of tableting speed on compressibility of the formulation containing 75% ACM.

28

Fig. 9. Tablet radial expansion for powders containing different levels of ACM.

Friction coefficeint (

0% ACM 25% ACM 50% ACM 75% ACM 100% ACM

0.25 0.20 0.15 0.10

75% ACM

0.24

Friction coefficeint (

10 ms

0.30

10 ms 15 ms 25 ms 100 ms

0.20

0.16

0.12

0.05

0.08 0.00 0

50

100

150

200

250

Compaction pressure (MPa)

300

0

50

100

150

200

250

300

Compaction pressure (MPa)

Fig. 10. Friction coefficient (µ) for binary mixtures at different (a) ACM contents and (b) dwell times.

29

0.38 10 ms

75% ACM

0.36

0.36 0.34

Poisson's ratio ()

Poisson's ratio ()

0.34 0.32 0.30 0.28 0.26

0% ACM 25% ACM 50% ACM 75% ACM 100% ACM

0.24 0.22

0.32 0.30 0.28

10 ms 15 ms 25 ms 100 ms

0.26 0.24 0.22

0.20 0

50

100

150

200

250

Compaction pressure (MPa)

300

350

0

50

100

150

200

250

300

350

Compaction pressure (MPa)

Fig. 11. Poisson’s ratio for binary mixtures at different (a) ACM content and (b) dwell time.

30

3.0

2.5

2.5

2.0

E (GPa)

E (GPa)

75% ACM

3.0

10 ms

1.5

2.0

1.5 0% ACM 25% ACM 50% ACM 75% ACM 100% ACM

1.0

10 ms 15 ms 25 ms 100 ms

1.0

0.5

0.5 0

50

100

150

200

250

0

300

50

100

150

200

250

300

Compaction pressure (MPa)

Compaction pressure (MPa)

1.2 75% ACM

10 ms 1.0

1.0

G (GPa)

G (GPa)

0.8

0.6 0% ACM 25% ACM 50% ACM 75% ACM 100% ACM

0.4

0.8

0.6

10 ms 15 ms 25 ms 100 ms

0.4

0.2 0

50

100

150

200

250

Compaction pressure (MPa)

300

0

50

100

150

200

250

300

Compaction pressure (MPa)

Fig. 12. Young’s modulus (a-b) and shear modulus (c-d) for binary mixtures at different ACM content and dwell time.

31

Design-Expert® Software

RDP (MPa) B: RDP

Capping

10.4

10

Non-capping

A: IER B: RDP

al Factors S = 2.00 = 0.31 = 1.97 = 1.13

Factor Coding: Actual Capping Capping

17.0

9.1

1

14.0

7.8

Actual Factors A: IER = 11.00 B: RDP = 9.39 E: E = 2.12 F: G = 0.71

0

11.1 5

8.1

Non-capping 0

X1 = D: v X2 = C: TS

C: TS TS (MPa)

gn-Expert® Software or Coding: Actual ping 0

6.5 5.2 5

3.9 2.6

10

10

1.3

Capping

Capping

5.1 4.7

8.0

11.2

14.5

IER (%) A: IER

17.7

21.0

0.0 0.22

0.25

0.29

0.33

0.36

Poisson’s D: vratio (ν)

Fig. 13. Contour plots of RDP vs. IER at (a) TS = 2.0 MPa and ν = 0.3, and (b) TS vs. ν at IER = 11% and RDP = 12 MPa.

32

Table 1. Py and observed capping or laminating behavior of different formulations at a fixed tableting speed (10 ms dwell time) and the same formulation under various tableting speeds. ACM % a 0%

Py c 57.1

Dwell time b

Capping/Lamination No

10 ms

(0.27) 25%

74.6

80.6

No

15 ms

90.9

No

25 ms

94.3

Capping (all pressures)

66.0

Capping (> 300 MPa)

62.1

Capping (> 300 MPa)

(0.47) Capping (all pressures)

(0.84) 100%

90.9

(0.94)

(0.61) 75%

Capping/Lamination

(0.84)

(0.35) 50%

Py c

100 ms

58.0

No

(0.8) Capping (all pressures)

d

(0.93) a

at 10 ms dwell time for a formulation containing 75% ACM. c standard errors of fitting is shown in parenthesis. d lamination at > 250 MPa b

33