0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 Average Relative Density-( ) 0 50 100 150 200 250 300 Average Main Compression Pressure[MPa]-P • Consolidate and compact powders (tableting) • Achieve continuous control of pharmaceutical solid products’ CQAs (e.g., tensile strength, weight, disintegration time) • Model for Quality by Design (QbD) and for Process Analytical Technology (PAT) • Utilize multi-scale results involving varying length and time scales to characterize the materials used in tablet manufacturing • Use simulation tools to verify and validate the reduced order models • Predictive constitutive models of inter-particle interactions for a variety of physical mechanisms + Predictability at high levels of confinement remains an open problem • Concurrent and efficient multi-scale strategies which are fully-descriptive at the granular scale + Based on a particle mechanics description MULTI-SCALE RESULTS Multi-scale characterization of powder compaction spanning single particle, single tablet fabrication and industrial rotary tablet press PARTICLE CHARACTERIZATION Goals: • Assist the development of constitutive models of inter-particle interactions with direct experimental testing of individual particles at high levels of deformation and confinement. • Create a protocol for measuring mechano-chemical properties at the particle scale. • Create a protocol for predicting tableting (or another unit operation) based on single particle measurements. • Adapt the protocol for non-spherical particles (chase towards the actual shape of pharmaceutical powders). Approach: • Single particle under controlled loading patterns. • Image acquisition of deformed configurations and processing to reconstruct deformation fields. • Initial focus: Excipients and APIs used in continuous line From particle behavior to table properties: By developing mechanistic predictive models at the particle scale and numerical multi-scale techniques at the powder/tablet scale (R&D effort!) Shimadzu MCT-510 PMMA spherical particles Side camera records the shape of a particle deformed under diametrical compression Capabilities : + particle diameter: 1-500 μm + displacement range: 0-100 μm + displacement increment: 1nm + loading/unloading cycles Challenges : + To position the particle and focus the camera REDUCED ORDER MODELING IN CONTINUOUS LINE PARTICLE MECHANICS STRATEGIES POWDER BLEND CHARACTERIZATION Dominant mechanisms: - Elastic deformations - Plastic deformations - Bonding - Strain-rate mechanisms - Friction and fracture - Water intake and swelling cp3 National Science Foundation www.nsf.gov Center for Structured Organic Particulate Systems www.csops.org U.S. Food and Drug Administration www.fda.gov Center for Particulate Products and Processes engineering.purdue.edu/CP3 Yasasvi Bommireddy, Prof. Marcial Gonzalez Mechanical Engineering § Test Modes • Target thickness • Target peak compaction force • Fracture test § Measurements • Compaction profile • Detachment force • Ejection force • Breaking force PILOT PLANT EXPERIMENTS § Upstream Material Properties • Particle size distribution (PSD) • True density of blend • Mass flow rate § Variable Parameters in Tablet Press • Production rate • Filling height of powder blend in dies § Measurable Critical Quality Attributes (CQAs) of Tablets by AT4 • Weight • Thickness • Diameter • Breaking Force CONCLUSION - Filling depth and gap height - Main & Pre-compression Force - Ejection & Take off Force - Tablet & Compartment Temp. - Compartment Humidity - Table Weight (Mass), thickness, diameter - Tablet hardness Micro compression testing Continuous manufacturing line : tablet press and AT4 Input Data with Uncertainty Calibration/Training of Bayesian Models Process control & optimization Product design Cost effective Science-driven Natoli BLP-16 Press Reduced Order Models with uncertainty quantification and built-in dimensionalityreduction 0 Probability −40 −20 0 20 40 0 0.01 0.02 0.03 0.04 0.05 0 Probability 0.03 0.04 0.05 0.06 0.07 0 0 0 Probability −40 −20 0 20 40 0 0.01 0.02 0.03 −20 −10 0 10 20 0 0.05 0.1 0.15 0 Probability −40 −20 0 20 40 0 0.01 0.02 0.03 0.04 0.05 0 Tablet strength Dissolution profile, etc. In-linemeasurements Model-form uncertainty Material properties, etc. Goal: • Development of first-principles predictive models to understand the attributes of powder blends and subsequent compacts Leuenberger → , = − − − , / , Goals: • Control of CQAs of tablets using predictive models to achieve QbD of the continuous line • Development of reliability strategies using PAT tools to have a robust control of the entire plant Kawakita → 1− 4 ⁄ = + ; 4 ⁄ Weight Prediction in-line → ? = @ A BC D E (1 −cTS) MCT can be used to characterize only one particle while the compaction simulator can produce only one tablet at a time while the tablet press generates hundreds of tablets per min. Hence the models describing each stage would be different and of varying orders. 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 Relative Density 0 20 40 60 80 100 120 Compaction Pressure[MPa] 0 ms 300 ms 600 ms 900 ms 1200 ms • At a time only a single part can be compacted • The particle size is only in micro-meters • The time of compaction is longer • The size of particles and their uniformity play a role in the characterization of particle • At a time only a single tablet can be fabricated • The time required can be changed by changing the punch speeds from very slow to very high • Different compositions of blend can be changed quite easily and hence varying ranges of properties can be observed • Hundreds of tablets can be used within a minute, hence the time required to produce one tablet is very low • The composition of the blend cannot be changed frequently as segregation and mixing could take place in various stages of the press • The models would need to be scaled up and reduced in order to be predictive for real time control Size of compacted unit Compaction Time MOTIVATION Powder blend characterization Note: Pilot plant is located at Purdue University Gamlen D Series Compaction Simulator