Infrared Fingerprinting Reveals Hidden Batch Differences in Popular Ayurvedic Immunity Syrup

A drop of herbal syrup, a beam of infrared light, and a few minutes of computation have exposed something that conventional quality control has long struggled to see: measurable chemical differences between production batches of one of India’s most widely used Ayurvedic immunity tonics. In a study published in Frontiers in Natural Products, researchers from Parul University, the National Forensic Sciences University, and King’s College London applied attenuated total reflectance–Fourier transform infrared (ATR–FTIR) spectroscopy to twelve commercially procured samples of Septilin syrup drawn from five distinct manufacturing batches, and found that while samples within a single batch were remarkably uniform, the batches themselves could be cleanly told apart by their chemical fingerprints alone.

Septilin is a polyherbal formulation built from botanicals including Tinospora cordifolia (Guduchi), Glycyrrhiza glabra (Yashtimadhu), Rubia cordifolia (Manjistha), Emblica officinalis (Amlaki), and Commiphora wightii (Guggulu). It is marketed as an immune-supporting adjuvant for respiratory tract infections, ENT disorders, and dental infections, and its therapeutic reputation rests on reported antimicrobial, anti-inflammatory, and immunomodulatory activities. But that same multi-component complexity is what makes quality assurance so difficult. Unlike a synthetic drug with a single defined active molecule, a polyherbal syrup contains hundreds of phytoconstituents whose concentrations can shift with the geographical origin of the plants, seasonal variation, cultivation practices, harvesting time, and processing conditions. Two batches that pass every individual marker test may still differ meaningfully in their overall chemical makeup.

The analytical logic of the new study is elegant in its simplicity. ATR–FTIR spectroscopy works by pressing a small amount of sample directly against an infrared-transparent crystal and measuring which frequencies of mid-infrared light the molecules absorb. Each absorption band corresponds to the vibration of a specific chemical bond: O–H stretches from phenols and hydrogen-bonded alcohols, C–H stretches from aliphatic groups, C=O stretches from carbonyl-bearing esters and organic acids, and a dense thicket of overlapping absorptions in the 1000–1300 cm⁻¹ fingerprint region from carbohydrates, glycosides, and other bioactive compounds. Because the technique requires no extraction, no dilution, no solvents, and no sample destruction, a complete chemical fingerprint of the syrup can be recorded in its native marketed form in under a minute.

The researchers scanned each of the twelve samples in triplicate over the 400–4000 cm⁻¹ range, co-adding sixteen scans at 4 cm⁻¹ resolution and automatically subtracting background spectra to eliminate atmospheric interference from carbon dioxide and moisture. The replicate spectra were then averaged to yield one representative fingerprint per sample. What emerged was a set of complex but broadly similar profiles, confirming the presence of a common formulation matrix, yet peppered with subtle variations in peak intensities and minor band shifts that differed systematically from batch to batch. The broad O–H stretching band at 3200–3500 cm⁻¹, the carbonyl band near 1700 cm⁻¹, and above all the carbohydrate-rich fingerprint region between 1000 and 1300 cm⁻¹ carried the strongest batch-specific signatures.

Raw spectra, however, are far too information-dense for the human eye to interpret reliably, which is where chemometrics enters the picture. The team applied Principal Component Analysis (PCA), an unsupervised multivariate technique that compresses thousands of correlated spectral variables into a small number of independent principal components while retaining most of the original variance. The results were striking: the first principal component alone accounted for 94 percent of the variance in the dataset, and together with the second component (4.3 percent) captured 98.3 percent of the total variation among samples. On the resulting score plot, samples from the same batch formed tight, well-defined clusters, while samples from different batches separated distinctly along the principal component axes.

Crucially, the researchers took steps to rule out the possibility that the clustering was an artefact of instrumental variation rather than genuine chemical difference. They repeated the PCA on Standard Normal Variate (SNV)-normalized spectra, a preprocessing transformation that removes multiplicative scaling effects and additive baseline offsets from each spectrum, thereby stripping out variations attributable to ATR contact pressure or sample thickness. Batch separation along PC1 survived this normalization, supporting a compositional rather than purely instrumental basis for the observed discrimination. The team also smoothed the PC1 and PC2 loading profiles with a Savitzky–Golay filter purely for display purposes, confirming that the fingerprint region between 1300 and 950 cm⁻¹ contributed most strongly to the first component, while the second component captured a different combination of O–H and fingerprint-region features that helped separate batches PC1 alone could not fully resolve.

Hierarchical Cluster Analysis (HCA) provided an independent verification of the PCA findings. Using Euclidean distance as the similarity measure and complete linkage as the clustering algorithm, the technique builds a dendrogram that reveals the hierarchical relationships among samples. The resulting tree told a consistent story: samples 1, 3, and 12 from batch 111,241,072 clustered tightly together, as did samples 2 and 11 from batch 111,240,082, and samples 4 and 5 from batch 111,230,952. Samples 9 and 10 from batch 111,230,238 and samples 6, 7, and 8 from batch 111,230,233 likewise grouped by batch before merging with other clusters. In every case, samples from the same manufacturing batch joined together before joining samples from other batches, demonstrating excellent intra-batch uniformity alongside moderate but reproducible inter-batch variation.

What makes the study particularly timely is its alignment with the principles of green analytical chemistry. The researchers evaluated their method using a comprehensive battery of sustainability metrics: the AGREE (Analytical GREEnness) metric, which quantifies compliance with the twelve guiding principles of green analytical chemistry; the Complex Green Analytical Procedure Index (Complex GAPI), which assesses environmental burden across the entire workflow from sample collection to final measurement; the Blue Applicability Grade Index (BAGI), which evaluates practical feasibility in terms of sample throughput, runtime efficiency, and instrument accessibility; and the RGB12-based White Analytical Chemistry framework, which integrates environmental impact, analytical performance, and practical usability into a single balanced score. The ATR–FTIR method performed strongly across all three dimensions, confirming its suitability for routine quality control without the solvent consumption, waste generation, or labour intensity associated with conventional chromatographic techniques such as HPTLC.

The implications extend well beyond a single syrup. Polyherbal formulations are a hallmark of Ayurveda and other traditional medicine systems, and regulatory agencies worldwide are grappling with how to standardize products whose chemical complexity defies conventional single-marker approaches. Holistic fingerprinting coupled with chemometrics offers a way to assess the overall chemical integrity of such formulations rapidly, non-destructively, and sustainably, capturing compositional differences that individual marker compounds may miss entirely. The approach aligns with a growing body of work applying ATR–FTIR and multivariate analysis to other Ayurvedic preparations, suggesting a broader shift toward fingerprint-based quality assessment in the herbal pharmaceutical industry.

The authors are careful to note the limits of what their findings demonstrate. The observed spectral variations indicate measurable differences in chemical composition between batches, but they do not automatically translate into meaningful differences in therapeutic effectiveness or biological activity, and the study was not designed to establish predefined acceptance limits for batch conformity. Future work, they suggest, should focus on developing and validating supervised chemometric models such as SIMCA and PLS-DA using larger sample sets encompassing multiple manufacturing lots, establishing statistically defined acceptance limits for routine batch qualification, and incorporating storage conditions and product ageing into experimental designs to distinguish manufacturing-related variability from post-production changes. Pharmacological and clinical studies will ultimately be needed to determine whether the chemical differences detected here have any bearing on therapeutic performance. What the study does establish, however, is a practical, reproducible, and environmentally conscious strategy for monitoring batch-to-batch variation in complex herbal medicines, one that could bring the consistency standards of modern pharmaceutical analysis to formulations whose chemical complexity has historically resisted them.

Subject of Research: Sustainable and rapid ATR-FTIR fingerprinting coupled with a chemometric approach for quality-based batch differentiation of marketed septilin syrup formulations

Article Title: Sustainable and rapid ATR-FTIR fingerprinting coupled with a chemometric approach for quality-based batch differentiation of marketed septilin syrup formulations

Article References: Patel, R., Vegad, K., & Patel, N. (2026). Sustainable and rapid ATR-FTIR fingerprinting coupled with a chemometric approach for quality-based batch differentiation of marketed septilin syrup formulations. Frontiers in Natural Products, 5, Article 1925392. https://doi.org/10.3389/fntpr.2026.1925392

Image Credits: AI Generated

DOI: 10.3389/fntpr.2026.1925392

Keywords: Sustainable, rapid, ATR-FTIR, fingerprinting, coupled, chemometric, approach, quality-based, batch, differentiation, marketed, septilin