Towards Real-time Quality Control in Spark-Assisted Chemical Engraving (SACE) Process

Glass is widely utilized in microfluidics owing to its inherent properties, transparency, and biocompatibility. Among various techniques for fabricating glass microchannels, Spark-Assisted Chemical Engraving (SACE) emerges as a non-traditional effective method. SACE operates through an electrochemical phenomenon, generating high-energy discharges by immersing a tool electrode within an electrolyte and applying a high voltage, facilitated by a hydrogen gas film that isolates the tooltip. Discharges cause local heating to the temperature of around 650oc, causing high-temperature etching. SACE has demonstrated the capability to produce microchannels with surface roughness as low as 0.3 µm at an acceptable machining speed that reaches 200 µm/s. Critical properties such as microchannel texture, dimensions, and geometric consistency significantly influence their functionality in fluid dynamics and heat transfer applications. Despite the notable advantages of SACE, the technique exhibits inherent uncertainties and nonlinear behaviors, and there remains potential for further improvement to obtain more predictable outputs that meet specific shape and texture requirements. Addressing the unpredictable nature and irregular relationship between process parameters and machining quality, this study extracted SACE process signatures (gas film formation time and mean discharge energy) by current signal processing with time series classification to establish a correlation between these parameters so that a more robust predictive model is developed. This research introduces a model that controls the process signatures by adjusting the machining voltage parameters, including amplitude, period, and duty cycle. By utilizing a precision machine that maintains a constant tool travelling rate and gap during microchannel machining and is equipped with real-time hardware, the SACE machine with controlled process parameters can produce microchannels with uniform surface roughness and consistent geometries. In the future, developing a closed-loop control model incorporating more robust techniques and additional control elements could compensate for the SACE’s uncertainties, making a faster and more efficient method for glass microchannel fabrication.