Particle Size Meets Chemistry

In oral solid dosage forms, particle attributes govern behavior throughout blending, granulation, drying, transfer and compaction. Dissolution rate — and ultimately bioavailability — is closely tied to particle size and surface area, while content uniformity and flow depend on distribution breadth, the prevalence of fines and the tendency of particles to adhere or agglomerate.

For single-component raw materials, particle sizing alone may be sufficient to qualify incoming lots. In multi-component blends, however, the critical question is not only particle size, but which component occupies which region of the particle size distribution (PSD). Without chemical attribution, a shift in median size, the emergence of a shoulder or a persistent coarse tail can be difficult to interpret, diagnose or control within a risk-based framework.

Conventional particle-size techniques — laser diffraction (LD), dynamic image analysis and static optical microscopy — quantify size and distribution but do not identify composition. Spectroscopic tools such as Raman and IR provide molecular identity yet often lack particle-level context. Particle Correlated Raman Spectroscopy (PCRS) addresses this gap by combining automated optical imaging, particle segmentation, size measurement and Raman-based chemical identification to generate spatially resolved, chemically annotated particle maps and component-specific size information. The central concept is straightforward: Use bulk PSD data to define the physical landscape, and PCRS to make that landscape chemically interpretable.

Materials and Methods

Bulk PSD by Laser Diffraction

LD infers particle size from angle-dependent scattering and reports a volume-weighted PSD (e.g. D10, D50, D90). Method settings (dispersion conditions, optical model, refractive indices) should be selected and reported consistently with recognized practices for qualification and measurement. These practices are well established in ISO 13320:2020 and related education on interpretation of D-values and volume weighting.1,2 (Analytical note: Because LD is volume-weighted, a small number of large particles can disproportionately influence upper percentiles, e.g. D90, making component attribution of coarse modes particularly important for risk assessment.)

PCRS

PCRS integrates high-resolution optical imaging, automated particle detection, and Raman-based chemical identification into a single, traceable workflow. After a powder is dispersed onto a Raman-compatible substrate, the system first acquires optical images that locate and visualize individual particles. Automated segmentation routines then detect particle boundaries using threshold-based approaches, reducing operator variability and ensuring consistent identification across large particle sets. Once particles are located, morphology metrics — including size, circularity and aspect ratio — are calculated for each particle, allowing users to filter or prioritize particles prior to Raman analysis.

Raman spectra are subsequently collected from each selected particle using material-optimized acquisition parameters. Each spectrum is compared against validated reference libraries to assign chemical identity, ensuring that the optical image, Raman spectrum, and classification are directly correlated at the particle level. This process produces chemically annotated particle maps, along with per-component particle counts, number-based size distributions, and morphology statistics. The result is a structured, reproducible dataset suitable for formulation development, contamination investigations, and broader particulate characterization workflows.

Because PCRS produces number-based particle-size metrics, qualitative comparison with a volume-weighted bulk PSD may require estimating particle volumes under a declared shape model (e.g., equivalent sphere). These estimated volumes can then be binned by chemical class to generate component-specific, volume-weighted distributions. All assumptions regarding particle shape, morphological variability, and conversion methodology should be explicitly stated to maintain transparency and interpretability.

Study Design and Compositions

A two-component system was evaluated at five compositions: pure excipient (0% API), pure API (100% API), and binary blends containing 25%, 50%, and 75% API (wt/wt%). Each mixture was characterized by laser diffraction to establish the bulk PSD, and by PCRS to obtain chemistry-resolved particle maps and particle-level size measurements. Evaluating the full series of compositions allows assessment of whether each component’s size behavior remains consistent across blend ratios and whether PCRS-derived, identity-resolved particle counts track predictably with formulation targets.

Results

Bulk PSDs Across Pure Components and Blends

LD produced multimodal distributions for the pure components and for the three blends, capturing fine, mid-range, and coarse features across a broad size range. As the API fraction increased, the mixture’s fines fraction declined and the median particle size increased — patterns consistent with the pure-component characteristics (Figure 1).

Pure-Component Benchmarks Anchor Interpretation

The excipient displayed a fine-rich distribution, with approximately 32% <10 µm and 98% <100 µm, whereas the API showed only ~2% <10 µm and ~67% <100 µm. These benchmarks shape expectations: as API loading rises, fines decrease and median size increases (Table 1, Figure 2).

PCRS Provides Chemically Annotated Particle Maps and Component-Specific Behavior

Optical mosaic images, constructed by stitching together a series of adjacent, small fields of view, were generated to cover a larger analysis area. These mosaics were then processed using the PCRS workflow, in which automated particle detection defined individual particle regions, morphology metrics were recorded, and Raman spectra were acquired with spatial correlation preserved between each particle and its optical image. Spectra were matched against validated reference libraries to assign chemical identity, producing chemically annotated maps that reveal how the API and excipient populate the imaged area (Figure 3). Following identification, particle counts and size statistics were compiled by component, and API-identified particle counts increased systematically from 25% to 75% API, consistent with blend composition and confirming reliable detection and classification across the series (Figure 3).

Deconvolving the Bulk PSD

Converting PCRS particle sizes to estimated volumes under a declared shape model and binning by component enables component-specific volume PSDs and qualitative overlays with the bulk LD curve to interpret peaks, shoulders, and tails. Even without a formal overlay, PCRS showed the excipient dominating the fine fraction and API dominating the coarse fraction — the causative explanation behind the bulk PSD trends. When morphologies deviate from spheres (e.g., plates, needles), shape-aware checks or sensitivity analyses strengthen confidence.

Discussion

Why Size and Chemistry Are Both Necessary

In multicomponent powders, an apparent coarse tail can reflect API, excipient, or mixed agglomerates — each demanding different controls (milling/sieving vs. deagglomeration vs. supplier oversight). LD quantifies what the blend does in bulk; PCRS identifies who occupies each size regime, enabling targeted, defensible decisions for efficient troubleshooting, robust process development, and risk-based control strategies.

Practical Implications for Development and Manufacturing

• Raw material qualification: Pure-component PSDs provide anchors for interpreting blend behavior. A coarser excipient lot, or finer-than-usual API lot, will shift the blend’s PSD in predictable ways.
• Blend design and dissolution control: If dosage performance depends on API fines, PCRS confirms whether shifts in the <10 µm population are API-driven. In this system, higher API loading reduced fines, indicating that API milling may require tighter control to maintain target fines levels, or that excipient grade selection and blend energy may need adjustment to compensate.
• Segregation, flow, and coarse modes: Coarse fractions increase segregation risk during transfer/compression. PCRS distinguishes true single-component coarse particles from mixed clusters, enabling appropriate interventions (tighten sieve cut vs. deagglomeration/humidity control).
• Troubleshooting deviations:
• Median ↑ + API-classified fines ↓ → investigate API milling or supplier change.
• Median ↑ + stable API fines but reduced excipient fines → examine excipient lot/handling.
• Coarse tail ↑ + mixed clusters on PCRS → adjust blend energy, order of addition, electrostatic controls.

Methodology, Validation and Reporting Considerations

• Sampling and dispersion: Standardize splitting and dispersion to deagglomerate without fragmenting friable particles; ensure thin, uniform deposition for PCRS to avoid multilayer overlap.
• Raman acquisition and CLS classification: Optimize SNR and manage fluorescence (wavelength choice, baseline handling). Validate reference spectra from pure materials and document CLS thresholds/match criteria; discuss potential confusion between components with overlapping bands and how resolved.4,5
• Reconciling number- and volume-based results: State shape model, bin widths, and normalization. If a formal conversion is not pursued, present PCRS as an attribution tool while leaving LD as the bulk descriptor.
• Linearity and system suitability: A blend series (e.g., 25/50/75% API) doubles as a linearity check for PCRS enumeration/classification; monotonic increases in API-classified counts bolster confidence in responsiveness.
• Statistics/reporting: Report particles per field, number of fields, within-blend variability, and confidence intervals for fines/coarse fractions.

Common Pitfalls and How to Avoid Them

1. Agglomerates vs. true coarse particles: PCRS separates intrinsic coarse particles from mixed clusters for appropriate interventions.
2. Over-reliance on number-weighted data: When aligning with LD, convert appropriately or be explicit that PCRS is used for attribution rather than exact overlay.
3. Limited fields of view: Use multiple randomized fields and replicate as feasible to support representativeness.
4. Opaque classification: Prefer transparent classifiers (e.g., CLS), report thresholds, and discuss ambiguous cases.
5. Missing pure-component anchors: Always establish pure-component PSDs and Raman references before analyzing blends.

QbD and Control Strategy Implications

The correlative LD-PCRS approach aligns naturally with QbD/PAT thinking. LD supports physical CMAs and specifications; PCRS provides specificity to link CMAs to CQAs by component. Monitoring component-specific thresholds (e.g., <10 µm, <100 µm) strengthens control strategies, supplier oversight, and risk management.

While presented through a two-component example for clarity, the same principles extend to more complex mixtures. Practical considerations include fluorescence (wavelength selection, preprocessing), anisotropic shapes (shape-aware volume estimation), and cohesion/electrostatics (humidity or charge mitigation during dispersion and mapping). Despite these caveats, the guiding principle stands: use LD to define the physical landscape, and PCRS to render that landscape chemically interpretable.

Conclusions

Combining bulk LD with PCRS provides a practical and scientifically rigorous way to decode multicomponent powder behavior. Bulk PSDs quantify how the powder behaves as a whole; PCRS shows which components occupy specific size regimes and whether coarse features represent intrinsic particles or agglomerates. In this model system, the excipient was finer (~32% <10 µm; ~98% <100 µm), while the API was coarser (~2% <10 µm; ~67% <100 µm). As API fraction increased, fines decreased and medians increased, and PCRS counts scaled with formulation targets. As material complexity and regulatory expectations grow, LD-PCRS correlative characterization will support robust formulation design, process control, and risk-based decision-making.

References

1. ISO 13320:2020 — Particle Size Analysis: Laser Diffraction Methods. (2020). International Organization for Standardization (ISO).
2. Understanding and Interpreting Particle Size Distribution Calculations. HORIBA. [accessed February 12, 2026]
3. Analysis of Morphology and Chemical Identification of Pharmaceutical Micro-Particles Using Particle-Correlated Raman Spectroscopy. HORIBA. [accessed February 12, 2026]
4. Workman, J. and Mark, H. (2010). Classical Least Squares: Mathematical Theory. Spectroscopy. 25(5):26-33
5. Marin, E., et. al. (2024). Statistical approaches to Raman imaging: principal component score mapping. Anal Methods. 16:2707-2720.
6. Ntziouni, A., et. al. (2022). Review of existing standards, guides, and practices for Raman spectroscopy. Appl Spectrosc. 76(7):747-772.
7. Particle Characterization of Pharmaceutical Powder Mixtures. (2023). HORIBA Application Note. [accessed February 12, 2026]