Blend Uniformity Verification in Active Pharmaceutical Blending

The Challenge of Powder Blending

In solid dosage manufacturing (tablets and capsules), blending active pharmaceutical ingredients (a critical stage in chemical and pharmaceutical blending) (APIs) with excipients—such as binders, fillers, and lubricants—is a crucial process block. Hitting complete homogeneity is essential to guarantee uniform dosage strength across every single tablet in a batch. Over-blending can lead to particle segregation or changes in dissolution rates, while under-blending results in out-of-specification batches.

Traditionally, plants verify blending using "sample thief" devices. This involves pausing the rotating V-blender or double-cone blender, opening the vessel, and inserting a metal probe to extract physical samples for HPLC testing in the lab. This process has major drawbacks:

• Segregation Risk: Inserting the sampling thief can disturb the powder bed, introducing sampling errors.
• Cycle Delays: Waiting for HPLC results takes 4 to 24 hours, during which the blender sits idle.
• Safety Concerns: Opening the blender vessel exposes operators to active chemical compounds and increases contamination risks.

"The FDA's PAT (Process Analytical Technology) initiative encourages non-destructive, real-time measurements directly inside the process loop to guarantee quality at the source."

Wireless FT-NIR Implementation

To eliminate sampling delays, a plant installed a **ProLine2550 FT-NIR Analyzer** configured with a wireless, battery-powered diffuse reflectance probe. The probe was mounted directly into the lid of a rotating bin blender, viewing the powder through a sanitary sapphire window.

The spectrometer collected a scan every time the blender rotated to the bottom position, ensuring the powder fell directly onto the window. A wireless transmitter sent the spectral data in real-time to a receiver connected to the **ProChem process control HMI** platform in the control room.

Real-Time Blend Endpoint Determination

Rather than mapping absolute chemical concentrations, the blending endpoint is determined by monitoring the variation between consecutive scans. As blending progresses, the differences between spectra diminish. The **caliX Suite** processes the spectra using Standard Normal Variate (SNV) and first-derivative filters to remove scattering differences, then calculates the Relative Standard Deviation (RSD) of the active ingredient absorption band:

Blending Stage	Time Elapsed	Active Ingredient RSD (%)	Homogeneity Status
Startup	0 - 2 mins	12.5% — 18.0%	Highly heterogeneous
Intermediate	4 - 8 mins	3.8% — 6.5%	Mixing active and excipients
Target Endpoint	10 - 12 mins	<1.80% (Plateau)	Homogeneous (End Process)

Plant Operations Impact & FDA Compliance

By moving to real-time verification, the plant realized immediate advantages:

• Cycle Time Reduction: Average blending time was reduced by 30%, as operators shut down the blender immediately upon hitting the RSD target rather than mixing for a fixed, conservative duration.
• Elimination of Re-Work: Out-of-spec batches due to over-blending or segregation were reduced to zero.
• Automated Batch Records: The **ProChem process control HMI** compiles the RSD trend and final spectra directly into an electronic batch record, ready for regulatory review under 21 CFR Part 11.

Conclusion

Deploying wireless FT-NIR spectroscopy in rotating blenders shifts pharmaceutical manufacturing from offline testing to active process control. By tracking the RSD decay directly inside the blending vessel, manufacturers can safely compress cycle times, guarantee dosage uniformity, and satisfy regulatory requirements without stopping production.