Electrochemical biosensors are increasingly interfaced with dedicated CMOS front-ends, whose achievable resolution is determined jointly by the electronics and by the sensing interface itself. In practice, however, front-end design normally proceeds without a quan- titative, circuit-level characterisation of the noise generated by the electrode–electrolyte interface. The consequence is practical: the designer must either assume that amplifier noise dominates, and over-design the input stage in area and power for a noise perfor- mance that yields no benefit, or under-design it and discover the true noise limit of the sensor only after fabrication. The detection chemistry of analytes such as dopamine is well established and low-noise amperometric readout is a mature discipline, but no compact, measured, circuit-level description bridges the two. This thesis addresses that gap by building such a description for a screen-printed carbon electrode detecting dopamine, validating it, and stating the procedure that produces it. Impedance spectroscopy, cyclic voltammetry and chronoamperometric recordings of the working-electrode current and of the reference-electrode potential yield an equivalent circuit, an operating point and a pair of measured noise spectra at different dopamine concentrations. Each spectrum is decomposed into a flicker-like (1/f) component and a white floor, and the current channel is further separated into an ionic background and a faradaic excess carrying the entire concentration dependence. The resulting parametric model is implemented as behavioural current and voltage gener- ators in Verilog-A and simulated in Cadence Virtuoso; the simulated records are processed through the same analysis chain as the measured ones and compared with them through a criterion derived from the statistics of the spectral estimator. All comparisons agree within the precision of the measurement, with no parameter adjusted on the comparison, and the measured noise floor exceeds the thermal and shot-noise limits of the same cell by tens of decibels. The principal contribution is the measurement-to-model procedure itself, rather than the parameter values for this electrode.
I biosensori elettrochimici vengono interfacciati sempre più spesso a front-end CMOS ded- icati, la cui risoluzione è determinata congiuntamente dall’elettronica e dall’interfaccia di misura. La progettazione di tali front-end viene tuttavia condotta senza una caratteriz- zazione quantitativa, a livello circuitale, del rumore generato dall’interfaccia elettrodo– elettrolita. La conseguenza è pratica: il progettista deve assumere che domini il rumore dell’amplificatore, sovradimensionando lo stadio d’ingresso in area e potenza senza alcun beneficio, oppure sottodimensionarlo e scoprire il limite reale del sensore solo dopo la fabbricazione. Questa tesi costruisce la descrizione mancante per un elettrodo serigrafato in carbonio per la rilevazione di dopamina, la valida ed esplicita la procedura che la produce. Spet- troscopia di impedenza, voltammetria ciclica e acquisizioni cronoamperometriche della corrente di elettrodo di lavoro e del potenziale dell’elettrodo di riferimento forniscono un circuito equivalente, un punto di lavoro e una coppia di spettri di rumore misurati a di- verse concentrazioni di dopamina. Ogni spettro è decomposto in una componente di tipo flicker (1/f) e in un pavimento bianco; il canale di corrente è ulteriormente separato in un fondo ionico e in un eccesso faradico che porta l’intera dipendenza dalla concentrazione. Il modello parametrico è implementato come generatori comportamentali di corrente e tensione in Verilog-A e simulato in Cadence Virtuoso; le tracce simulate sono elaborate con la stessa catena di analisi delle misure e confrontate con esse mediante un criterio derivato dalla statistica dello stimatore spettrale. Tutti i confronti risultano in accordo entro la precisione della misura, senza alcun parametro aggiustato sul confronto, e il pavi- mento di rumore misurato supera di decine di decibel i limiti termico e shot della stessa cella. Il contributo principale è la procedura stessa che porta dalla misura al modello, più che i valori dei parametri ottenuti.
Electrical modeling of the noise in electrochemical biosensors for CMOS co-design
GRECO, GIORGIO
2025/2026
Abstract
Electrochemical biosensors are increasingly interfaced with dedicated CMOS front-ends, whose achievable resolution is determined jointly by the electronics and by the sensing interface itself. In practice, however, front-end design normally proceeds without a quan- titative, circuit-level characterisation of the noise generated by the electrode–electrolyte interface. The consequence is practical: the designer must either assume that amplifier noise dominates, and over-design the input stage in area and power for a noise perfor- mance that yields no benefit, or under-design it and discover the true noise limit of the sensor only after fabrication. The detection chemistry of analytes such as dopamine is well established and low-noise amperometric readout is a mature discipline, but no compact, measured, circuit-level description bridges the two. This thesis addresses that gap by building such a description for a screen-printed carbon electrode detecting dopamine, validating it, and stating the procedure that produces it. Impedance spectroscopy, cyclic voltammetry and chronoamperometric recordings of the working-electrode current and of the reference-electrode potential yield an equivalent circuit, an operating point and a pair of measured noise spectra at different dopamine concentrations. Each spectrum is decomposed into a flicker-like (1/f) component and a white floor, and the current channel is further separated into an ionic background and a faradaic excess carrying the entire concentration dependence. The resulting parametric model is implemented as behavioural current and voltage gener- ators in Verilog-A and simulated in Cadence Virtuoso; the simulated records are processed through the same analysis chain as the measured ones and compared with them through a criterion derived from the statistics of the spectral estimator. All comparisons agree within the precision of the measurement, with no parameter adjusted on the comparison, and the measured noise floor exceeds the thermal and shot-noise limits of the same cell by tens of decibels. The principal contribution is the measurement-to-model procedure itself, rather than the parameter values for this electrode.| File | Dimensione | Formato | |
|---|---|---|---|
|
2026_10_Greco_Tesi.pdf
non accessibile
Descrizione: testo tesi
Dimensione
10.69 MB
Formato
Adobe PDF
|
10.69 MB | Adobe PDF | Visualizza/Apri |
|
2026_10_Greco_ExecutiveSummary.pdf
non accessibile
Descrizione: executive summary
Dimensione
2 MB
Formato
Adobe PDF
|
2 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/266487