SpectraRx PackAnalyst

1. Results and Discussion

This study evaluates the capability of point near-infrared (NIR) spectroscopy to perform rapid, non-destructive chemical evaluations directly through polymer blister package barriers. In modern pharmaceutical supply chains, fast authentication and environmental degradation monitoring are crucial for public safety. Standard off-line methods (HPLC, Karl Fischer titration) are inherently destructive and slow, requiring sample extraction. By utilizing diffuse reflectance NIR scanning, we successfully traversed the protective blister barrier to read solid-state active pharmaceutical ingredients (APIs) and hydration shifts in real-time.

1.1. Spectral Characterization of Brands and Types (Study 1)

Figure 1 presents the raw and pretreated near-infrared absorption profiles for 15 samples comprising three commercial categories: Panadol (Standard Adult, 500 mg), Panadol Kids (pediatric formulation), and Setamol. The spectra exhibit strong absorption features characteristic of paracetamol (acetaminophen) chemical structure. Notably, the amide first overtone stretch is visible around $1650$ nm, alongside prominent aromatic $C-H$ overtones and intermolecular $N-H$ stretching combinations in the $1100$--$1400$ nm range.

Although standard adult white tablets of Panadol and Setamol share similar API concentrations, subtle baseline shifts and secondary structural peaks are visible in the $1100$--$1400$ nm region. These differences represent excipient composition variances, such as starch binders, microcrystalline cellulose, and lubricant compressives. In the case of Panadol Kids, distinct spectral adjustments (particularly baseline shifts) are observed, attributed to pediatric binders, flavoring agents, and coloring pigments.

1.2. Principal Component Analysis for Brand Classification

To convert these complex spectral differences into clear classification criteria, we applied Principal Component Analysis (PCA) to the $900$--$1700$ nm wavelength matrix. As shown in the scores scatter plot in **Figure 2**, the first two principal components explain over $94\%$ of the total chemical variance. The three paracetamol types segment into completely separated clusters:

• Class 4 (Panadol, red crosses): Clusters tightly on the upper-right quadrant of the projection space (PC2 $\approx 2.7$, PC1 $\approx -2.0$).
• Class 3 (Panadol Kids, green diamonds): Clusters on the upper-left quadrant (PC2 $\approx 1.8$, PC1 $\approx 2.3$).
• Class 2 (Setamol, orange squares): Resolves securely in the lower-central quadrant (PC2 $\approx -4.5$, PC1 $\approx -0.3$).

This clean separation demonstrates that despite identical nominal API identities, handheld NIR sensors paired with PCA-driven projection models can easily differentiate standard consumer brands based on localized excipient matrices and compression densities.

1.3. Identification of Blister Compromise and Moisture Exposure (Study 2)

Blister packages can occasionally split or be compromised during shipping, exposing tablets to atmospheric moisture. This environmental exposure leads to the hydrolytic degradation of paracetamol into 4-aminophenol. We simulated this by comparing properly sealed white tablets of Panadol with identical blister packs where the foil was intentionally compromised.

Over the full range (Figure 3), the raw profiles remain highly similar due to matching API contents. However, close inspection of the $1340$ to $1520$ nm wavelength region (Figure 4) reveals distinct structural differences. Moisture-exposed paracetamol displays a clear shift and baseline inflation at the water absorption peak, which corresponds to the first overtone of the symmetric and asymmetric $O-H$ stretching vibrations of absorbed liquid water around $1450$ nm. This spectral signature directly measures the ingress of environmental hydration.

1.4. PCA Score Segregation of Hydration Status

The PCA score matrix of the water-absorbance band reveals a distinct, highly practical pattern of separation (Figure 5). Properly packaged, unexposed tablets (Class 1, orange circles) form an exceptionally compact cluster on the left (PC1 $\approx -0.4$, PC2 $\approx 0$). This tight grouping reflects low internal variation in moisture content among untouched factory-sealed products.

In contrast, compromised moisture-exposed tablets (Class 2, yellow squares) migrate to the right side of the plot along PC1 (extending from 0 up to $+0.7$). The broad dispersion of the exposed group reflects different levels of local moisture absorption across the exposed tablet surfaces. This score shift along the PC1 axis serves as a reliable, non-destructive proxy for packaging degradation, allowing immediate identification of damaged packages without opening the blisters.

2. Critical Evaluation & Peer Review

3. Conclusions

This study demonstrates that handheld point Near-Infrared (NIR) spectroscopy, paired with PCA models, is a highly effective, non-destructive tool for evaluating pharmaceutical quality. The system achieved 100% classification accuracy when distinguishing standard consumer brands, and successfully isolated blister packages compromised by moisture exposure. The key advantage of this approach is its ability to perform high-resolution chemical assays directly through intact packaging, providing a fast, cost-effective, and practical solution for quality control and anti-counterfeiting in global pharmaceutical supply chains.