3. Results and Discussion
The interface engineering of current collectors significantly affects the electronic and ionic transport behaviors in the active material layer of electrochemical electrodes, enhancing their performance [
18,
19]. The energy storage mechanism of EDLCs relies on the physical surface adsorption/desorption of electrolyte ions, which is driven by the electrostatic force generated from electron transfers within the active material. Therefore, the interfacial electron transport between the active material layer and current collector can significantly affect the energy storage performance of EDLCs, including ultrafast charge/discharge and high specific capacity, compared with those of other energy storage devices (Li-ion batteries, pseudocapacitors, and fuel cells).
Figure 1a illustrates the simple method used in this study for forming uniform microgrid-patterns on Ni foams using an SUS mesh as a rigid stamp during roll-pressing. Notably, compared to other patterning methods (e.g., chemical etching, lithography, and laser patterning), our stamp micro-patterning has distinct advantages, such as continuity of the roll-to-roll-based process, pattern uniformity, economic feasibility, and the absence of extra equipment and product lines [
20]. Moreover, the MP morphology of the Ni foam surface was effectively designed by differentiating the mesh size; that is, by employing 24, 40, and 60 standard mesh sizes (ASTM E11), as shown in
Figure 1b–d. The mesh size is defined as the number of openings per square inch (1 in = 2.54 cm). As the mesh number increases, the number of openings increases, and the opening size decreases. In addition, the diameter of the mesh wires decreases with an increase in the mesh number. Therefore, the surface morphology of the Ni foam, including its microgrid size, number of microgrids, and width/depth of the dented vertical stripes, can be systematically varied by changing the mesh size.
Figure 1e−h display the four types of Ni foams: bare Ni, 24MP-Ni, 40MP-Ni, and 60MP-Ni. Bare Ni does not have a specific surface pattern. In contrast, 24MP-Ni, 40MP-Ni, and 60MP-Ni exhibit uniform microgrid-patterns on their surfaces. The enlarged images of the different MP-Ni foams (inset images with orange dotted borders) revealed that these have a rough surface with prominent microgrid-patterns compared with the bare Ni foam with a smooth surface.
To investigate the MP morphology of the MP-Ni foams, their 3D microtopography images were obtained (
Figure 2). As shown in the 3D image in
Figure 2a, bare Ni had no specific micro-roughness; the slight surface roughness (<10 μm) is attributed to the inherent morphology of the foam itself. In contrast, the MP-Ni samples (24MP-Ni, 40MP-Ni, and 60MP-Ni) exhibited specific surface roughnesses, and their 3D topographic images presented protruding 3D microgrids of various scales and densities (
Figure 2b−d). The density of the 3D microgrids increased, whereas the microgrid size decreased in the order of 24MP-Ni, 40MP-Ni, and 60MP-Ni.
Figure 2e−l present the precise scales of the 3D microgrids of the Ni foams, as determined by 3D microtopography.
Figure 2e−h show the top-view 2D roughness images of bare Ni, 24MP-Ni, 40MP-Ni, and 60MP-Ni, respectively, and
Figure 2i−l show the corresponding line roughness data. Bare Ni exhibited no specific height variation within a height of 1600 μm from the sample surface, thus confirming its smooth surface morphology and the absence of a 3D micro-structure. In contrast, the MP-Ni samples exhibited specific height variations, confirming the formation of different 3D microstructures. The depth and width of the 3D microgrid-pattern decreased with increasing size of the mesh used for preparing Ni foam samples (150.8 and 754.5 μm for 24MP-Ni, 86.3 and 638.9 μm for 40MP-Ni, and 60.9 and 355.2 μm for 60MP-Ni, respectively), confirming that the surface roughness of the MP-Ni sample can be tuned simply by changing the mesh size. In summary, the 3D microtopography results of the bare Ni and MP-Ni foam samples confirm that the depths and widths of their surface features can be tuned.
The surface geometry of each Ni sample was examined in terms of the width and depth of the 3D microgrid-pattern, root-mean-square roughness (R
rms), skewness roughness (R
sk), and the penetration depth of the active material (see
Figure 3a and
Table 1). Compared to bare Ni, all MP-Ni samples exhibited specific width and depth scales owing to their 3D microgrid-patterns, resulting in microscale surface roughness. Accordingly, all the MP-Ni samples had a higher R
rms value than the of bare Ni foam. A significant increase in the R
rms of the Ni foam results in its increased contact area with the active material layer, resulting in decreased interfacial resistance between the Ni foam current collector and active material. Consequently, the MP-Ni-based electrode accommodates (adsorbs/desorbs) a larger amount of charge during charge/discharge compared with that of the flat Ni foam-based electrode. Moreover, the R
sk value, defined as the asymmetry of the surface profile around its mean value, should be considered in describing the surface geometry obtained by 3D microgrid-patterning. As illustrated in
Figure 3b, a highly negative R
sk represents a surface with predominantly sharp valleys, whereas a slightly negative R
sk represents a round surface with blunt valleys. Given that a viscous slurry mixture was used to cast the active layer, the surface geometry of the Ni foam with a highly negative R
sk induces shallow infiltration of the slurry mixture, creating voids between the active layer and Ni foam after drying [
21]. These voids block electron transfer under ultrafast charge/discharge conditions and weaken the interfacial adhesion between the Ni foam and active material following repetitive charge/discharge cycles [
22]. In comparison, the surface geometry of MP-Ni with a slightly negative R
sk facilitates deeper infiltration of the slurry mixture, leading to more complete contact between the active layer and Ni foam. Therefore, MP-Ni samples with slightly negative R
sk values (−0.46 for 24MP-Ni, −0.19 for 40MP-Ni, and −0.13 for 60MP-Ni) possibly enhance specific capacitance, even under high-rate charge/discharge conditions relative to that achieved using the bare Ni foam with a highly negative R
sk (−2.02). The gradual decrease in the R
sk in order of 24MP-Ni > 40MP-Ni > 60MP-Ni is attributed to the decrease in the mesh wire diameter, in accordance with the mesh size. This is because a smaller mesh diameter can alleviate the pressure exerted on the Ni foam during roll-pressing compared with that of a larger mesh diameter. Similarly, the width and depth of the 3D microgrid-pattern gradually decreased in the order of 24MP-Ni > 40MP-Ni > 60MP-Ni (
Table 1). The surface geometry of the 3D MP-Ni foam is likely to affect the electron transfer characteristics of the active material layer, which is directly related to the charge adsorption/desorption kinetics and reversibility of the EDLC electrodes.
The Raman spectrum of the active material, YP50F, in
Figure 4a exhibits broad peaks near 1341.5 and 1598.7 cm
−1, corresponding to the D- and G-bands of carbon, respectively. Considering that the D-band of carbon originates from lattice defects, and the G-band is due to sp
2-bonded carbon atoms, the high intensity ratio of the D- to G-band (I
D/I
G = 1.25) confirms the amorphous nature of the carbon material (i.e., YP50F) used in this study.
Figure 4b shows the nitrogen adsorption/desorption curves of YP50F. In the low relative pressure range, the volume change curve shows a sharp change due to rapid nitrogen adsorption/desorption from the micropores. Further, the hysteresis loop in the relative pressure range of 0.2–1.0 indicates the presence of mesopores. The adsorption data reveal a specific surface area of 1760.9 m
2/g of YP50F, indicating that it is a suitable active material for EDLC electrodes.
Figure 4c presents the XRD patterns of bare Ni, 24MP-Ni, 40MP-Ni, and 60MP-Ni in the 2
range of 20−80°. All Ni foams exhibited common diffraction peaks at approximately 44.8°, 52.3°, and 76.7°, corresponding to the (111), (200), and (220) planes of Ni, respectively (JCPDS #04-0850) [
23]. The XRD patterns of the MP-Ni samples without any peaks attributable to specific impurities or phase conversion confirmed the mild processability of the roll-to-roll mesh-patterning method. The XRD patterns of the EDLC electrodes prepared by casting the slurry of YP50F on the bare and MP-Ni foams (i.e., YP50F/bare Ni, YP50F/24MP-Ni, YP50F/40MP-Ni, and YP50F/60MP-Ni) were also analyzed (
Figure 4d). Consistent with the XRD results of the Ni foam (
Figure 4c), all electrodes yielded diffraction peaks at 44.5°, 51.9°, and 76.3° for Ni. Considering that YP50F has a highly porous and amorphous carbon structure, no specific diffraction peaks of carbon were observed.
Figure 4e−h show the OM images of bare Ni, 24MP-Ni, 40MP-Ni, and 60MP-Ni, respectively. All Ni foams were found to have net-like morphology with a Ni wire diameter of 61.3−93.5 μm; this morphology facilitates a larger contact area with the active material layer and thus enhances electron transfer and adhesion compared with the case of the flat Ni foam. Notably, compared with the continuous foam morphology of the other Ni samples, 24MP-Ni revealed discontinuous Ni wires, which can be attributed to excessive pressure induced by the 24 mesh stamp with a larger mesh diameter during stamp micropatterning. The discontinuity of the Ni wires in 24MP-Ni is likely to inhibit facile electron transfer to the active material. On the contrary, 40MP-Ni and 60MP-Ni with a continuous net-like foam morphology allows the electrons to flow rapidly and uniformly toward the active material during ultrafast charge/discharge.
To evaluate the electrochemical performances of the EDLC electrodes fabricated using bare Ni, 24MP-Ni, 40MP-Ni, and 60MP-Ni as current collectors, their gravimetric charge–discharge curves and specific capacitances were evaluated at different current densities in the range of 0.2–20 A/g (
Figure 5). All samples exhibited symmetric linear curves during charge/discharge, which is attributable to the physical adsorption/desorption-based charge transport mechanism of these EDLC electrodes. Among the samples, YP50F/40MP-Ni showed the longest charge–discharge times at all current densities. Based on the gravimetric charge–discharge data in
Figure 5, the specific capacitances were calculated for all electrodes (see
Figure 6a). Notably, the YP50F/MP-Ni samples had a higher specific capacitance than the YP50F/bare Ni sample at all current densities. The increased specific capacitances of the YP50F/MP-Ni samples can be attributed to the MP morphology of the Ni foam current collector, which promoted considerable electron transport to the active material layer, thus increasing the amount of charge adsorbed/desorbed during charge/discharge processes. More importantly, among the EDLC electrodes, YP50F/40MP-Ni showed exceptional specific capacitance retention, even at an ultrafast charge/discharge rate of 20 A/g. As summarized in
Table 2, YP50F/40MP-Ni exhibited the greatest specific capacitance retention (87.1%) in the range of 0.2–20 A/g, followed by YP50F/60MP-Ni (77.0%), YP50F/24MP-Ni (75.0%), and YP50F/bare Ni (72.1%). Thus, YP50F/40MP-Ni exhibits ultrafast-rate capability because the Ni foam (40MP-Ni) facilitates electron transfer at its interface with the active carbon material, even under ultrafast charge/discharge conditions. The Coulombic efficiency (CE), defined as the ratio of the discharge capacity to charge capacity, indicates the charge adsorption/desorption reversibility during the charge–discharge processes. Considering that the charge storage mechanism of the EDLC electrode is based on the surface adsorption/desorption of charge pairs (electrons and electrolyte ions), the CE is significantly influenced by the electrical conductivity of the electrode. Thus, YP50F/40MP-Ni exhibited higher CE values than the other electrodes at all current densities, especially at ultrafast-rate current densities of 10, 15, and 20 A/g (
Table 3).
Figure 6b shows the CV profiles of bare Ni, 24MP-Ni, 40MP-Ni, and 60MP-Ni in the voltage range of 0–1 V. All Ni foam-based electrodes yielded a broad closed curve without specific redox peaks owing to the typical electrochemical behavior of the EDLC electrode. The absence of specific peaks in the CV curves indicates that the electrode does not undergo the Faradaic process—a common charge storage mechanism in PCs—in the applied voltage range. The CV curve areas of all the MP-Ni-based electrodes were larger than that of the bare Ni-based electrode, implying that the 3D MP interfacial geometry of the MP-Ni foams results in enhanced electrochemical activity with greater charge adsorption/desorption compared to that of the bare Ni foam. Remarkably, YP50F/40MP-Ni exhibited the largest CV curve area among the electrodes, thus indicating the highest electrochemical activity. This can be attributed to the optimized surface geometry of the MP-Ni foam, which leads to deep infiltration of the YP50F slurry, facilitating the full utilization of the active material/Ni foam interface. More importantly, the robust contact geometry and increased contact area at the active material/Ni foam interface enhanced the electrochemical reversibility of the system and promoted fast-rate charge transport during repetitive and ultrafast charge/discharge cycles.
Figure 6c compares the capacitance retention and high-rate capability of YP50F/40MP-Ni with those of YP50F/bare Ni for different cycle numbers and current densities, respectively. YP50F/40MP-Ni exhibits a capacitance retention of 98.98% after 1000 consecutive charge/discharge cycles at 1 A/g, outperforming the YP50F/bare Ni electrode (92.99%). Furthermore, YP50F/40MP-Ni features a high-rate capability—defined as the specific capacitance retention at a given current density relative to that at 0.2 A/g—of 87.1% at 20 A/g, while YP50F/bare Ni exhibits an inferior high-rate capability of 72.1% at 20 A/g. To evaluate the ultrafast longevity of the MP-Ni-based EDLC electrodes, we compared the specific capacitance variations of the electrodes based on bare Ni and 40MP-Ni during 5000 consecutive charge/discharge cycles at 5 A/g (
Figure 6d,e). YP50F/40MP-Ni showed a constant specific capacitance (106.0 F/g after 5000 cycles) with a superior retention of 101.9% compared with that of YP50F/bare Ni (specific capacitance of 74.0 F/g and retention of 92.5% after 5000 cycles). These results demonstrate the remarkable ultrafast longevity of 40MP-Ni as a current collector for EDLC electrodes.
Figure 6f presents the energy density plots at power densities in the range of 360−36,000 W/kg for all Ni foam samples. The energy and power densities were calculated as follows:
where dV and dt denote the potential window (V) and discharge time (s), respectively. YP50F/40MP-Ni presented a notably higher energy density of 12.0 Wh/kg even at a high power density of 36,000 W/kg compared with those of bare Ni (8.0 Wh/kg), 24MP-Ni (10.0 Wh/kg), and 60MP-Ni (9.0 Wh/kg) (
Table 2). Furthermore, YP50F/40MP-Ni exhibited better energy storage capability at high power densities than those reported previously for other EDLCs [
24,
25,
26]. Therefore, the evaluated energy storage performances demonstrated the improved charge transport behaviors of our electrodes, including an ultrafast kinetics, reversibility, and electrochemical activity, owing to the 3D MP surface geometry of the Ni foam current collector. In particular, the optimized surface geometry of 40MP-Ni with a slightly negative R
sk (−0.19) and high R
rms (35.4 μm) facilitates its complete adhesion to the active material and increases the contract area between the two, facilitating accelerated electron transfer during ultrafast and repetitive charge/discharge cycles.
Figure 7a,b show the Nyquist plots of the EDLC electrodes based on bare Ni, 24MP-Ni, 40MP-Ni, and 60MP-Ni. EIS, represented as measured impedance plots in the applied frequency range, is considered a suitable tool to clarify the interfacial resistance of electrochemical electrodes. Typically, EIS results are categorized into two domains: (ⅰ) the low-frequency region, represented by a straight line, referred to as the Warburg impedance, and (ⅱ) the high-frequency region, represented by a semicircle [
27]. The low-frequency region is dominated by ion diffusion resistance at the electrode/electrolyte interface, whereas the high-frequency region represents from electron transfer resistance across the electrode/electrolyte interface [
28,
29]. As shown in
Figure 7a, YP50F/40MP-Ni exhibited the steepest slope with a value of 30.96 (YP50F/bare Ni: 20.35, YP50F/24MP-Ni: 25.86, and YP50F/60MP-Ni: 28.37), confirming that it has the lowest ion-diffusion resistance for electric double layer formation at the electrode/electrolyte interface [
30,
31]. Additionally, it exhibits the smallest semicircle with a charge transfer resistance of 0.128 Ω (YP50F/bare Ni: 0.293 Ω, YP50F/24MP-Ni: 0.142 Ω, and YP50F/60MP-Ni: 0.236 Ω), demonstrating the most active electron transfer across the electrode/electrolyte interface, as seen in the enlarged EIS profiles verifying the impedance plots in the high frequency region (
Figure 7b). The extremely small semicircle (<1 Ω) implies a double layered charge storage mechanism based on the physical adsorption/desorption of charged species at the electrode/electrolyte interface for all electrodes [
32]. The EIS result with a gentle slope and the largest semicircle for YP50F/bare Ni suggests the most resistive interfacial charge transport, owing to the absence of a specific surface geometry for the bare Ni foam. The scheme in
Figure 7c illustrates the interfacial charge transport with cross-sectional views of all EDLC electrodes. Despite the smallest microscopic features of the patterned 60MP-Ni, a gentler slope and a larger semicircle were generated compared with those of 24MP-Ni and 40MP-Ni, indicating a higher interfacial charge transport resistance at the electrode/electrolyte interface. The degraded interfacial charge transport in 60MP-Ni can be attributed to the excessive contact area between the electrode and Ni foam, which possibly causes electron scattering and aggregation, leading to a fast-rate and inhomogeneous electron transfer [
33]. Conversely, the relatively large microscopic features in 24MP-Ni can provide a relatively low contact area between the active material and Ni foam and more interfacial voids, preventing ultrafast electron transfer during high-rate charge/discharge [
34]. Owing to its optimized surface geometry with a 3D microgrid-pattern, 40MP-Ni enhances charge transport toward ultrafast kinetics and reversibility even under an ultrafast-rate charge/discharge process through the following synergistic effects: (ⅰ) a significant increase in the R
rms increased its contact area with active material, thus accommodating considerable charge transfer during ultrafast charge/discharge process and thereby improving ultrafast-rate capability; (ⅱ) the slightly negative R
sk, which is almost zero, promotes deep infiltration of the carbon slurry, thus leading to rigid adhesion with minimized interfacial defects (voids and electron scatters) at the rough surface, encouraging reversible charge transport during repetitive fast-rate cycles. Owing to its rigid interface adhesion structure, the EDLC electrode fabricated using 40MP-Ni exhibited notable flexibility, especially under ultrafast charge/discharge conditions.
Figure 8a shows photographs of symmetric EDLC electrodes fabricated from 40MP-Ni in the straight and bent states, respectively. Based on the measurement conditions in
Figure 8a, gravimetric charge–discharge curves of YP50F/bare Ni and YP50F/40MP-Ni were measured in the straight and bent states, respectively. YP50F/40MP-Ni exhibited excellent charge/discharge capability, even at 5 and 10 A/g (
Figure 8c), whereas the YP50F/bare Ni electrode showed a significantly lower charge/discharge capability at 5 and 10 A/g (
Figure 8b). Thus, YP50F/40MP-Ni delivered specific capacitances of 106.1, 90.0, and 88.1 F/g at 0.2, 5, and 10 A/g, respectively, which are comparable to those observed in the straight state (
Figure 8d). However, the YP50F/bare Ni electrode showed specific capacitances of 86.6, 66.0, and 60.1 F/g at 0.2, 5, and 10 A/g, respectively, which are lower than those observed in the straight state. The 3D microgrid-patterns on the Ni foam current collector significantly enhanced the interfacial adhesion of the foam with the active material layer, enabling rapid electron transfer, even under ultrafast charge/discharge conditions.