Process-property correlations in laser-induced graphene electrodes for electrochemical sensing

Abstract

Laser-induced graphene (LIG) has emerged as a promising electrode material for electrochemical point-of-care diagnostics. LIG offers a large specific surface area and excellent electron transfer at low-cost in a binder-free and rapid fabrication process that lends itself well to mass production outside of the cleanroom. Various LIG micromorphologies can be generated when altering the energy input parameters, and it was investigated here which impact this has on their electroanalytical characteristics and performance. Energy input is well controlled by the laser power, scribing speed, and laser pulse density. Once the threshold of required energy input is reached a broad spectrum of conditions leads to LIG with micromorphologies ranging from delicate irregular brush structures obtained at fast, high energy input, to smoother and more wall like albeit still porous materials. Only a fraction of these LIG structures provided high conductance which is required for appropriate electroanalytical performance. Here, it was found that low, frequent energy input provided the best electroanalytical material, i.e., low levels of power and speed in combination with high spatial pulse density. For example, the sensitivity for the reduction of K₃[Fe(CN)₆] was increased almost 2-fold by changing fabrication parameters from 60% power and 100% speed to 1% power and 10% speed. These general findings can be translated to any LIG fabrication process independent of devices used. The simple fabrication process of LIG electrodes, their good electroanalytical performance as demonstrated here with a variety of (bio)analytically relevant molecules including ascorbic acid, dopamine, uric acid, p-nitrophenol, and paracetamol, and possible application to biological samples make them ideal and inexpensive transducers for electrochemical (bio)sensors, with the potential to replace the screen-printed systems currently dominating in on-site sensors used.

Keywords: Chemical sensor; Fabrication parameters; Laser-induced graphene; Porous carbon; Voltammetry.

Graphical abstract

Fig. 1

Electrode fabrication scheme (a), sketches of used electrode designs (WE working electrode, CE counter electrode, RE reference electrode), top = “simple” design, bottom = 3-electrode design (b), photo of prepared LIG electrodes (c)

Fig. 2

Heatmap of electrode outcome vs. power and speed settings at a pulse density of 1000 × 1000 (x by y) (color code as indicated in the inset: green = ok, darker brown = partial scribing (PS), lighter brown = no effect (NE), orange = LIG peeled off from substrate (PO), red = laser burned through substrate (B))

Fig. 3

a CV of 5 mM K₃[Fe(CN)₆] in PBS with electrode leads bare or covered with silver paint. Higher speed setting correlated with larger ΔE_p (laser power = 30%, 1000 × 1000 PPI). b Peak-to-peak separation in CV. c sheet resistance of LIG vs scribing speed

Fig. 4

Micrographs showing the active region of LIG electrodes made with the same pulse density but different power and speed settings (a, d, g). Square-wave voltammograms (b, e, h) and dose-response plots (c, f, i) of ferricyanide standards on those electrodes. E_begin = 0.7 V, E_end = −0.2 V, step = 5 mV, pulse = 50 mV, f = 10 Hz. All standards (V = 50 μL) were measured on the same electrode in order of rising concentration. 3-electrode design LIG electrodes were used

Fig. 5

Peak-to-peak distance obtained with LIG electrodes in CV (a) and sheet resistance of LIG samples (b) (n = 3)

Fig. 6

Photograph and SEM micrographs of LIG (1/10/1000 × 1000) scale bars in pictures a–f are 500, 20, 10, 20, 5, and 1 μm

Fig. 7

a CV in 5 mM K₃[Fe(CN)₆] in 0.1 M KCl on 1/10/1000 × 1000 electrodes at various scan rates. b Plot of cathodic peak current vs v^1/2. The slope was divided by the factors in the Randles-Sevcik equation (0.4463 F^3/2 (D/R T)^1/2 C), with F = 96,485 A s mol⁻¹, D = 7.63E−6 cm² s⁻¹, R = 8.314 J K⁻¹ mol⁻¹, T = 296.15 K, C = 5E−6 mol cm⁻³, to yield 0.107 cm². Division by the geometric surface of 0.06 cm² yielded a ratio of 1.8. c Nicholson plot based on the same data: slope = k^0,eff = 0.003 cm s⁻¹. 3-electrode design LIG electrodes were used

Fig. 8

Dose-response experiments to [Fe(CN)₆]³⁻ on LIG electrodes (1/10/1000 × 1000) using chronoamperometry (a + b, step from 0.4 V to 0.0 V), cyclic voltammetry (c + d), and square-wave voltammetry (e + f). Each electrode was used for only a single measurement and three electrodes were used per concentration level. In SWV, each fresh electrode displayed a slightly different background current and so, for clarity, each concentration level in e is represented by one voltammetric curve only. 3-electrode design LIG electrodes were used