Nanomolar electrochemiluminescence via smartphone: Image encoding matters

Smartphone electrochemiluminescence detection is a practical route to decentralized diagnostics, but three error sources that the literature has largely overlooked limit sensitivity: unconstrained acquisition parameters, nonlinear gamma encoding in consumer image pipelines, and lossy JPEG compression. We show that addressing all three simultaneously yields nanomolar detection limits with a mid-range Android device. Long-exposure RAW images captured under fixed conditions are analyzed by full-histogram integration of the linearized red channel, applying the inverse sRGB transfer function to restore proportionality between pixel intensity and emitted radiance. On screen-printed glassy carbon electrodes with the tris(2,2'-bipyridine)ruthenium(II)/tripropylamine system, this workflow gives a detection limit of 16 nM and a quantification limit of 53 nM in solution, below values reported with any previous smartphone-based system. We also show that lossy JPEG export suppresses blank pixel variance by 16-25%, generating detection limits that appear competitive but carry no analytical meaning; lossless image formats are therefore a requirement for reliable low-concentration quantification. The complete pipeline runs in ECLiCare, an Android application that operates entirely on-device without cloud connectivity. Performance was validated against professional image-processing software on solution-phase data and on a bead-based sandwich immunoassay for epidermal growth factor receptor, where the linearized workflow gives a detection limit of 2.31 μg mL^-1. This work provides a quantitative framework for smartphone electrochemiluminescence imaging that is ready for field and point-of-care deployment.