Calorimetric Investigation on Heat Release During the 1 Disintegration Process of Pharmaceutical Tablets

9 The compendial USP < 701 > disintegration test method offers a crucial 10 pass/fail assessment for immdiate release tablet disintegration. However, its 11 single end-point approach provides limited insight into underlying mechanisms. 12 This study introduces a novel calorimetric approach, aimed at providing com-13 prehensive process profiles beyond binary outcomes. We developed a novel dis-14 integration reaction calorimeter to monitor the heat release throughout the dis-15 integration process and successfully obtained enthalpy change profiles of placebo 16 tablets with various porosities. The formulation comprised microcrystalline cel-17 lulose (MCC), anhydrous lactose, croscarmellose sodium (CCS), and magnesium 18 stearate (MgSt). An abrupt temperature rise was observed after introducing 19 the disintegration medium to tablets, and the relationship between the heat 20 rise time and the tablet’s porosity was investigated. The calorimeter’s sensi-21 tivity was sufficient to discern distinct heat changes among individual tablets, 22 and the analysis revealed a direct correlation between the two. Higher porosity 23 corresponded to shorter heat rise time, indicating faster disintegration rates. 24 Additionally, the analysis identified a concurrent endothermic process alongside 25 the anticipated exothermic phenomenon, potentially associated with the disso-26 lution of anhydrous lactose. Since lactose is the only soluble excipient within 27 the blend composition, the endothermic process can be attributed to the ab-28 sorption of heat as lactose molecules dissolve in water. The findings from this 29


Introduction
Disintegration in pharmaceutical science refers to the mechanical break up of an intact tablet into fragments to enlarge the contact surface area with the dissolution medium to facilitate the drug substance dissolution (Markl and Zeitler, 2017).Therefore, it plays a significant role in the quality assurance of the final product's bioavailability and efficacy, and the associated attributes can be employed as a surrogate measure in place of the dissolution test, particularly for fast-disintegrating solid forms with highly soluble drug substance (ICH, 2009).
The disintegration process occurs in two steps.First, upon contact with the tablet matrix, the dissolution medium penetrates the pore network of the matrix.Second, the inter-particular bonds between the individual particles are weakened by the hydrogen bonds of the dissolution medium, inducing the matrix fragmentation (Alder and Frenning, 2022).As the wetting and wicking of the pore network is the initial step of disintegration, it is crucial to understand how the liquid front moves in the tablet matrix to design and reliably achieve specific disintegration performance.This approach is sensible because the disintegration process is governed by the intricate interplay of a series of complex mechanical and chemical phenomena, including wicking, swelling, gelation, dissolution, and erosion, to name a few.
Using terahertz pulsed imaging (TPI) coupled to a flow cell or open immersion cell, it is possible to capture the advancing liquid front profiles within a tablet during disintegration (Yassin et al., 2015;Markl et al., 2018a,b;Al-Sharabi et al., 2020, 2021;Lee et al., 2023).Such studies made it possible to measure the kinetics of the liquid imbibition quantitatively.It is well-known in porous media, such as pharmaceutical tablets, that the capillary force is es-sential in initiating the liquid ingress.This force is generally assessed using surface tension measurement methods, such as sessile drop or pendant drop analysis (Wenzel, 1936;Kwok and Neumann, 1999;Berg, 2009).The principle of these methods is analysing the geometry of the interfaces between vapour and liquid, liquid and solid, and solid and vapour at its equilibrium state, through which the wettability of the solid can be determined.
Thermodynamically, imbibition, i.e. the ingress of the dissolution medium into the dry tablet matrix, is spontaneous, as the free energy of the systems is lowered to a more stable state.During this process, the dry state's vapoursolid and vapour-liquid interfaces are replaced by the wet state's solid-liquid interfaces.Each interface has its own specific interfacial potential, and the accompanying net free energy change determines the favourability of the process and thus can serves as a measure of the wetting characteristic (Wenzel, 1936).
Bangham et al. also explained that the wettability of a solid to a liquid is determined by the interfacial potential difference between the dry and wet surface total energy (Bangham and Razouk, 1937).These interfacial potential differences are released as heat energy, increasing temperature at the transitioning interface.Claxton reported this temperature rise in 1959 (Claxton, 1959).They filled a capillary tube with silica gel and introduced various organic liquids in the capillary while measuring the resultant temperature changes using thermocouples.They resolved temperature changes at the boundary between liquid components and explained that the evolution of heat resulted from a free energy decrease in the system.More recently, Aslannejad et al. investigated temperature spikes at a wetting front as water propagated along paper sheets of various thicknesses and compositions using an IR camera.The authors concluded that the net decrease in interfacial potential caused the spontaneous imbibition of water into the dry matrix, and the released energy was developed into heat, resulting in a temperature spike at the water front (Aslannejad et al., 2017).
Additionally, during a feasibility test for this study, a direct compression tablet with 99 wt % microcrystalline cellulose (MCC) and 1 wt % magnesium stearate (MgSt) formulation showed temperature increases at the water front during spontaneous imbibition of deionised water.Infrared images were taken at room temperature, and approximately 2 • C of temperature increase was observed at the advancing water front.
Calorimetry has been an essential technique for the characterisation of drug substances and excipients for many decades.The thermal response of pharmaceutical materials has been extensively studied using differential scanning calorimetry (DSC) (Chiu and Prenner, 2011;Müllertz et al., 2016).However, whilst solution calorimetry is another important characterisation technique, the thermodynamics of complex reaction processes, such as the dissolution or swelling of entire tablets, has received less attention.Such lack of published work may result from the absence of suitable experimental instruments and the inherently non-specific nature of the calorimetric investigation of complex systems such as tablets.There are, however, a few studies regarding the enthalpy change during the dissolution and swelling of pharmaceutical actives and excipients: Buckton et al. observed the temperature changes during in vitro dissolution to study the interactions of tetracycline hydrochloride with milk, artificial liquid meals, and other food components using a flow-through micro-calorimetry (Buckton et al., 1989).They reported both exothermic and endothermic changes without providing a detailed interpretation of the phenomena.A solution calorimeter was also exploited to study whether it could be used to quantify the swelling of polymeric excipients.Conti et al. placed 3 mg of hydroxypropyl-methylcellulose (HPMC) powder in a sample cartridge and broke it in 15 mL of deionised water, monitoring the enthalpy changes (Conti et al., 2006).The results showed that a solution calorimeter has sufficient sensitivity to discriminate different grades of polymers as distinct enthalpy profiles were observed across three HPMC grades.Regardless, due to the complexity of the polymer powder dissolution process, the associated mechanisms were not interpreted.The dissolution of nanosuspension was also studied using a solution calorimeter by Kayaert et al.The authors investigated thermal changes during the dissolution of 60 mg to 800 mg of nanosuspension samples (naproxen, cinnarizine) in 100 mL of a dissolution medium and quantified the heat absorbed or produced during the process (Kayaert et al., 2010).All three studies have successfully demonstrated the sufficient sensitivity of the calorimeters to differentiate polymer compositions and the associated dissolution and swelling processes.Nevertheless, these techniques measured the total heat changes, making it challenging to explain the process mechanistically.
Building upon the proof of concept and the results of previous studies, we hypothesised that the rate of heat release during the wicking process of compacted tablets could be used as a sensitive metric to quantify the total interfacial area that is transitioning from vapour-solid interface to liquid-solid interface and that the corresponding enthalpy change could hence be utilised to rationalise the disintegration process further.To address the limitations outlined earlier, we developed a novel reaction calorimeter designed to deliver a dissolution medium to the reaction vial, where a sample tablet is placed, emulating the disintegration process under adiabatic conditions.Subsequently, we conducted a series of controlled experiments using five batches of tablets that ranged in porosities but were composed of the same formulation, emphasising the correlation between the enthalpy change profiles and the tablet porosities.

Placebo Tablets
Formulation.The composition of the powder blend was 64 wt % of microcrystalline cellulose (MCC), 32 wt % of anhydrous lactose, 3 wt % of croscarmellose sodium (CSS), and 1 wt % of magnesium stearate (MgSt), as described in Table 1.The excipients were screened through a 1.0 mm sieve to remove any agglomerates before they were blended, followed by two steps of the blending process: 1) pre-lubrication and 2) lubrication.For pre-lubrication processing, all ingredients except for the lubricant, MgSt, were loaded in a 500 mL container at 60 % fill level (40 % headspace) and blended at 49 rpm for 20 min using a commercial Turbula T2F mixer (WAB, Muttenz, Switzerland).After the mixing, the bulk density of the blend was measured at 0.442 g mL −1 .This value was utilised to determine the accurate weight for 60 % fill level in the container, and hence, the lubrication time can be determined based on the Kushner-Moore model (Kushner and Moore, 2010).For lubrication processing, an amount corresponding to 1 % of MgSt was added without sieving to avoid shear and then blended at 49 rpm for the appropriate time, which was 257 s for this study.
Direct Compression and Physical Properties of Tablets.A commercial compaction simulator, HB50 (Huxley Bertram, Cambridge, UK), was used to make tablets from the prepared powder blend.The Fette P1200 profile using a turret speed of 35 rpm was employed for the press profile, and the target pressures were set from 50 MPa to 250 MPa with 50 MPa intervals by adjusting the tooling gaps for the main compression load.The powder blend was filled into the die with a nominal tablet mass of 400 mg. 10 mm round concave toolings were used.The toolings' cap depth was 1.05 mm.Twenty tablets were produced for each batch, followed by disintegration tests and tablet hardness tests using a Sotax DT50 (Sotax AG, Aesch, Switzerland) and a Dr. Schleuniger Pharmatron 8M (Sotax AG, Aesch, Switzerland), respectively.The disintegration test complied with the USP⟨701⟩, and the end-points were determined by visual inspection due to considerably large deviations observed in end-points determined by the instrument's contact sensors.Tablet hardness was measured in newtons (N), and subsequently, it was converted to tensile strength in megapascals (MPa) using the following equation outlined by Pitt et al. (1988): where σ t is the tensile strength of the biconvex tablet, P is the measured fracture load, D, T , and W are the diameter, the overall thickness, and the cylinder length of the tablet, respectively.
The nominal porosity of tablets, Φ nominal , was calculated based on the measured tablet volume, V tablet , weight, W tablet , and the harmonic mean of true densities of the associated excipients as described in Equations 2 and 3 (Bawuah et al., 2020;Rowe et al., 2009).The derived true density of the blend and the tablet porosity ranged from 1.469 g cm −3 to 1.499 g cm −3 and from 7 % to 32 %, respectively.
where ρ bulk and ρ true are the bulk density and true density of the tablet.
The cracking patterns observed in the tablets after the hardness test exhibited mid-line fractures except for 250 MPa batch tablets, which all capped as exemplified in Figure 1.The physical properties of the five batches were tabulated in Table 2.

Reaction Calorimeter
Disintegration Reaction Calorimeter.Investigating the disintegration process of pharmaceutical tablets using a solution calorimeter may pose challenges because most sample cartridges of solution calorimeters are not sufficiently large to accommodate an intact tablet.Additionally, this study required a controlled feeding of a dissolution medium into the reaction vials, ideally without transferring heat outside the experimental chamber.It was also essential to prevent any additional mechanical stress caused by the stirrer of the solution calorimeter.
To meet all these requirements, we developed a novel reaction calorimeter, as   shown in Figures 2 and 3.The principle of the calorimeter design is facilitating a water reservoir, micro-peristaltic pumps (Williamson, Brighton, UK), and the reaction vials in a thermally insulated chamber while keeping heat-generating components out, which is a stepper motor for this case.Three 5 mL glass reaction vials were installed in the chamber and connected to the water reservoir through three parallel micro-peristaltic pumps using 1 mm inner diameter silicone tubes.The stepper motor was located on the lid of the chamber and connected to the micro-peristaltic pumps through 3 mm D-shaft, controlling the speed and amount of water feeding into the vials.Each vial had two K-type thermocouples with their exposed junctions positioned on the opposite side of the inner space.These junctions were immersed in the water during the experiments.One thermocouple was also immersed in the water reservoir, measuring the reference water temperatures.All temperature data were collected using a commercial data logger, TC-08 (Pico technology Ltd, St Noets, UK) at 100 ms intervals.8. Weigh the vials and calculate the amount of added water.9. Before cleaning the vials, stir the tablet residue with a stirring rod and check if the tablet is fully hydrated without any remaining palpable cores.
Data Analysis.Fundamentally, the temperature changes in the experiment were predominantly influenced by the reaction vial, added water, and the test tablet.Accordingly, these factors were accounted for by the enthalpy change calculations, as described in Equation 4. The specific heat capacities for water (C water ) and glass vials (C vial ) at room temperature were taken as 4.184 J g −1 • C and 0.840 J g −1 • C, respectively.These values were adopted from the literature (Cutnell et al., 2018) and may exhibit variations with different references.
An estimate of 0.800 J g −1 • C was employed for the specific heat capacity of the tablet.This estimate was based on the specific heat of microcrystalline cellulose reported by Ketolainen et al. (1995).
where ∆H T is the total enthalpy change in joules, and m and C represent the mass and the specific heat capacity of the glass vial, added water, and the tablet in the hydration reactions, respectively as specified by their subscripts.
∆T denotes the temperature change in Kelvin.

Results and Discussion
Temperature and Enthalpy Changes.As soon as the hydration was triggered by introducing approximately 1.5 mL of water to the reaction vials, a rapid temperature increase was observed across all measurements, followed by a gradual decrease after reaching its peak temperature.Tablets of higher porosity exhibited faster temperature increases, as depicted in Figure 4. Interestingly, the temperature profiles of the 200 MPa and 250 MPa batches appeared qualitatively similar, and the influence of compression pressure was less pronounced between the two compaction conditions.This similarity was attributed to the similar porosities of the tablets from both batches, as detailed in Table 3. Notably, the tensile strength of the 250 MPa batch measured a lower value compared to that of the 200 MPa batch (see Table 2).This suggests that an excessive compression force was applied to the former, resulting in a subsequent reduction in tensile strength (Adolfsson and Nyström, 1996).
Utilising the acquired temperature profiles and the mass of water, vials, and test tablets, which are tabulated in Table 3, enthalpy changes were derived from Equation 4, and the resulting profiles are presented in Figure 5.

Disintegration Time Comparison. Based on the hypothesis presented in
Section 1, the initial surge in enthalpy was postulated to correspond to water ingress into the tablet matrix.We defined the enthalpy rise time as the duration required to reach 90 % of the peak value to quantify the surge, as outlined in Table 3.The selection of 90 % as a criterion value was based on two considerations: 1) adhering to the convention of defining rise time as the interval between 10% and 90% of the peak, and 2) its smallest relative standard deviation among 80 %, 85 % and 90 %.
Comparisons with USP⟨701⟩-compliant disintegration times are summarised in Table 4 to elucidate our methodology.Due to differences in experimental conditions and heat conduction delays affecting calorimetric measurements, the calorimetry times consistently exceeded the disintegration times across all measurements.The enthalpy change profiles displayed distinct exothermic patterns among different porosity groups.An intriguing finding was that all profiles concluded with a negative value (refer to Figures 4 and 5), implying an endothermic process occurred following the initial exothermic process during measurements.
According to the heat release observations of individual excipients in powder forms, lactose showed a temperature drop while others exhibited a rise.Since lactose is the only water-soluble material in the composition, the associated dissolution enthalpy and entropy changes might be the main cause of the heat absorption (Petrucci, 2007).
While the kinetics of the enthalpy change varied between different batches, the total surface area that transitions from a solid-vapour interface to a solid-  liquid interface was expected to be similar.The cumulative enthalpy changes were derived and compared to the peak values across the batches to validate this assumption.The integration revealed that all batches exhibited a peak around 10 kJ s g −1 , as illustrated in Figure 6.However, it is important to note that these peaks, although seemingly close, exhibit considerable deviations, and drawing a definitive conclusion at this stage could be premature.
Kinetics Analysis 1.The obtained enthalpy change profiles could be divided into two phases: an initiating increasing phase and the subsequent decreasing phase.These two phases were termed the wetting and swelling phases to align with the associated dominant processes; however, this does not imply that wetting or swelling occurs exclusively within each phase.The first phase involves a combination of liquid ingress and subsequent phenomena, such as swelling and dissolving, while the second phase is dominated by swelling and dissolving as the tablet is fully hydrated.For the analysis of their kinetics, profiles were plotted Note: 250 MPa batch details can be found in Table 6.
using a base-10 logarithm scale on the y-axis, as shown in Figure 7. Common logarithms were initially used for easier data handling and interpretation.However, the rate constant of the swelling phase, k s , was obtained by multiplying the slope of the profile by a scaling factor of 2.3 to transform it into natural logarithm scale.The results, tabulated in Table 5, revealed strong linearity for this phase, indicating first-order kinetics, except for the 250 MPa batch.While the mean 250 MPa profile did not exhibit a clear linear pattern, individual profiles within the batch followed the same pattern observed in other batches (see Figure 8).Large deviations in the individual profiles made the mean profile appear non-linear.More importantly, fluctuations around 700 s (depicted by red dashed circles in Figure 8) were identified as the main contributors to the irregularity.Such fluctuations could indicate an abrupt water infiltration into the matrix through fractures.Excessive compression pressure during tabletting led to the accumulation of elastic energy stored in the compaction, which ultimately causes the breakage of bonds between tablet particles, as reported, for example, by Adolfsson and Nyström (1996).
Individual profiles were also analysed to assess the relationship between the compression pressure and the rate constants.The result suggests a moderate correlation with a 0.4757 correlation coefficient.Interestingly, the rate constants of the 150 MPa batch exhibited relatively narrower deviations compared to other groups, as shown in Figure 9.The mechanistic and physical implications of the   Kinetics Analysis 2. The preceding section identified a clear first-order process during the swelling phase.The extrapolated line from this process can be utilised to extract the kinetics of the wetting phase by subtracting it from the overall profiles.It is a good idea to approach such calculation carefully, especially when a comprehensive understanding of the mechanism behind the swelling phase kinetics is still lacking.However, this approach is not unreasonable.Accordingly, we examined the commonly used method of feathering residuals in pharmacokinetics and compared the residuals with the associated linear regressions, as illustrated in Figure 10.The residuals in all plots formed a convex line (upward curvature), which could be interpreted as a concave shape (downward curvature) in the wetting phase kinetics on a semi-logarithmic scale.
This downward deviation from simple first-order kinetics indicates that the reaction rate slowed as the hydration progressed.Such reduction in rate could be due to the influence of reaction inhibiting factors, such as the tablet's geometry, gel formation in the tablet matrix that hinders further liquid penetration, and increase in liquid viscosity over time.
In terms of the tablet's geometry, the heat release caused by the tablet hydration is proportional to the wetted volume.Accordingly, the exothermic process of the wetting phase is dependent on the geometry and the associated liquid imbibition.In this study, the biconvex geometry was simplified as a cylindrical shape with radius r 0 and height h 0 , and the volume of the wetted portion, V wet , was expressed with a function of the liquid ingress, L(t), defined by: V The propagation of liquid L(t) was assumed to occur isotropically and modelled as a linear function of kt + d based on empirical data from terahertz measurements, as illustrated in Figure 11.Further details regarding the liquid ingress function can be found in the study by Lee et al. (Lee et al., 2023).
Substituting L(t) with kt + d from Equation 5 yields a third-degree polynomial equation.Using this equation, the portion of the wetted matrix was derived and plotted along with the associated enthalpy change on a normalised y-axis, as shown in Figure 12.Despite differences in the experimental conditions, including the amount and temperature of water, comparing the two datasets could offer valuable insights into conceptualising the underlying mechanisms behind the observations.Firstly, a noticeable time delay was observable with the enthalpy changes.This delay primarily provides time for heat to transfer through the tablet matrix and the surrounding water before being probed by the thermocouples.However, the mechanism seems more complex given that the heat rise was instantly observable upon introducing the liquid in the vials (refer to Figure 4,

Conclusions
This work investigated five batches of direct compression tablets using a novel disintegration reaction calorimeter.The primary objective was to ascertain whether the porosity of pharmaceutical solid dosage forms could be distinguished through calorimetric analysis, and the results substantiated the feasibility of this approach.Specifically, the heat rise time proved to be a valuable quantitative metric for assessing the disintegration process of complex formulation tablets.While the study focused on the heat release during the wetting process, a concurrent endothermic phenomenon was also observed, prompting further exploration into the role of each excipient.
The inherently non-specific nature of calorimetric investigation methods could pose disadvantages in analysing complex systems.However, it offers a ubiquitous approach to evaluating intricate microstructures of complex compounds within well-designed experiments.In this context, the approach proposed in the study demonstrates promising potential for quantifying the wettability of complex pharmaceutical tablets.

Figure 1 :
Figure 1: Cracked tablets after the hardness test.Capping observed in tablets from the 250 MPa batch indicates that excessive compression forces were applied.

Figure 2 :
Figure 2: Schematic of the disintegration reaction calorimeter.The calorimeter was designed to accommodate an intact tablet in the reaction vials and prevent the transfer of heat while introducing a dissolution medium.

Figure 3 :
Figure 3: Images of the disintegration reaction calorimeter: a) the outlook and b) three reaction vials in the chamber.

Figure 4 :
Figure 4: Mean temperature changes and shaded standard deviations over time.The zero on the x-axis corresponds to the onset of the disintegration process.The plot illustrates a significant correlation between the temperature increase at t=0 and the tablet porosity.

Figure 5 :
Figure 5: Mean enthalpy changes and shaded standard deviations over time.The dashed rectangular boundary outlines the first 400 seconds of hydration, and the inset zooms in on this section, emphasising the initial exothermic process.

Figure 6 :
Figure 6: Integrated enthalpy changes and shaded standard deviations over time.All batches exhibited similar peak values around 10 kJ s g −1 .

Figure 7 :
Figure 7: Mean enthalpy changes on a common logarithmic scale over time.The plot reveals first-order kinetics during the decreasing phases across all batches, except for the 250 MPa batch.

Figure 8 :
Figure 8: Individual enthalpy changes of 250 MPa batch tablets on a logarithmic scale over time.Abrupt increases in the red circles can be interpreted as tablet matrix fractures.

Figure 9 :
Figure 9: A plot illustrating the relationship between compression pressure (x-axis) and swelling phase rate constant (y-axis).The correlation coefficient (0.4757) may support a weak to moderate correlation between the two factors.Notably, the 150 MPa batch tablets exhibit relatively consistent rate constants compared to other groups.

Figure 10 :Figure 11
Figure 10: Enthalpy change profiles focused on the wetting phase with the extrapolation of swelling phase kinetics on a semi logarithmic scale: a) 50 MPa, b) 100 MPa, c) 150 MPa, and d) 200 MPa batch means.The differences between the profile and the extrapolated line form an upward curvature from a linear dashed eye guide.

Figure 12 :
Figure 12: Plots illustrating the wetted tablet volume portions and the associated enthalpies across four batches: a) 50 MPa, b) 100 MPa, c) 150 MPa, and d) 200 MPa on a normalised scale.The wetted volumes were derived from terahertz liquid ingress measurements.Time delays between the two profiles can be attributed to different experimental conditions and heat diffusion times.

Table 1 :
Powder blend composition

Table 2 :
Physical properties of the placebo tablets

Table 3 :
Reaction vial, added water, and tablet masses with porosity and heat rise time

Table 4 :
Disintegration time comparison

Table 5 :
Swelling phase kinetics rate constants