Dataset on the Impact of Acquisition Temperature Variations on Quantitation Models of a Miniaturized NIR Spectrometer: A Platform for Calibration Transfer
Description
The quantification capabilities of miniaturized Near-Infrared (NIR) devices still lag behind those of mature benchtop spectrometers, often failing to meet the required accuracy standards for pharmaceutical analysis unless transferred to a robust calibration on a benchtop device. Several factors, including instrument size, lack of thermal management systems, and operational conditions, make miniaturized devices more susceptible to temperature increases during continuous measurement mode. These temperature variations can contribute to the limited quantitation performance of these devices. While some vendors suggest performing frequent reference scans to account for temperature variations, this solution may not be feasible for inline applications. To address this issue, this dataset offers a comprehensive examination of the impact of varied spectrometer temperature during the acquisition of reference scans and sample measurements on a pharmaceutical quantitation model. The experimental samples includes nine pharmaceutical formulations prepared following a constrained mixture design. NIR acquisitions were performed using a MicroNIR PAT 1700 Wireless spectrometer from Viavi Solutions, located in Milpitas, USA, which captures data within the 950–1650 nm range. Reference scans were initially acquired at 40°C, followed by continuous scans of each sample at increasing acquisition temperatures over time, ranging from 38°C to 55°C. Subsequently, reference scans were repeated at 52°C, followed by sample scans within the same temperature range. Based on the above conditions, the dataset was divided into four subsets, labeled as X1, X2, X3, and X4: Condition 1 (X1): Dark and reference scans at 40°C. Sample scans at probe temperatures below 46°C. Condition 2 (X2): Dark and reference scans at 40°C. Sample scans at probe temperatures above 46°C. Condition 3 (X3): Dark and reference scans at 52°C. Sample scans at probe temperatures below 46°C. Condition 4 (X4): Dark and reference scans at 52°C. Sample scans at probe temperatures above 46°C. In pharmaceutical applications, particularly inline processes, the calibration transfer of one subset to another, or to a combination of other subsets, is highly crucial. Therefore, these datasets provide a valuable platform for applying existing or new calibration transfer algorithms, enabling enhanced accuracy and reliability in pharmaceutical analysis.