Introduction:
The electronic tongue (e-tongue) is a bioinspired multi-sensor system that mimics human gustatory perception through an array of non-specific cross-reactive chemical sensors coupled with multivariate pattern recognition algorithms. The Alpha Astree II e-tongue (Alpha MOS, France) has been widely applied to objective palatability assessment of pharmaceutical formulations. However, reliable taste assessment of viscous oral suspensions is critically compromised by sample carryover arising from inadequate sensor probe washing, a challenge exacerbated by the elevated viscosity of pharmaceutical suspensions, which leads to greater retention of sample material on sensor probe surfaces. No published study has applied response surface methodology (RSM) to optimize an e-tongue washing protocol for viscous pharmaceutical products, representing a significant methodological gap in the field.
Methods:
An I-optimal RSM experimental design was applied to develop and optimize a sensor probe washing protocol for the Alpha Astree II e-tongue to analyze three commercially available ibuprofen oral suspensions with dynamic viscosities ranging from 55 to 430 cP. Anhydrous caffeine was used as a standardized bitter reference and test solution to quantify carryover via the seven-sensor ChemFET array. Three phases were conducted: Phase 1 identified the ibuprofen brand producing the greatest carryover using the standard one-beaker protocol (1000 rpm, 60 s); Phase 2 applied an I-optimal design across 10 experimental runs, optimizing stirring rate (1,000–15,000 rpm) and duration (300–900 s); and Phase 3 evaluated one-beaker and two-beaker washing configurations under standard and optimized conditions. Principal component analysis (PCA) and RSM numerical optimization were used for data analysis.
Results:
The standard one-beaker washing protocol (1000 rpm, 60 s) was insufficient to prevent carryover for all three ibuprofen suspension brands. RSM numerical optimization identified an optimized washing protocol of 5000 rpm for 300 seconds. A two-beaker configuration using these optimized parameters (IBUWASH-BOPT) produced caffeine sensor responses closest to the reference solution, ranking superior to all other protocols evaluated. PCA confirmed clear discrimination between standard and optimized washing conditions, demonstrating the critical impact of inadequate washing on data integrity and on the accuracy of taste profile measurements.
Conclusions:
A two-beaker sensor probe washing protocol at 5000 rpm for 300 seconds per beaker effectively prevents cross-contamination and sample carryover during e-tongue analysis of viscous ibuprofen oral suspensions. This is the first RSM-guided, I-optimal design-based optimization of an e-tongue washing protocol for viscous pharmaceutical products, providing a reproducible and transferable methodology for palatability assessment of pediatric oral suspensions using bioinspired taste sensing technology.