**3. Results**

The prepared Ag electrodes via mechanical stretching exhibited a matt dark gray color, because the cracked metal structures cause diffuse reflections rather than specular reflections (see Figure 1d). To confirm the overall shapes of the cracks formed on the Ag electrodes, the samples after completing the cold-welding processes were investigated using SEM as shown in Figures 3 and 4. For the UA120 and UA140 samples (i.e., uniaxially stretched to 120% and 140%, respectively), the Ag films tended to crack in the form of a one-dimensional line. With an increase in the uniaxial stretching ratio, the overall crack size increased, and the edges of the cracks were more clearly observed. Particularly, in the UA140 sample, minor cracks were further observed between the major cracks, leading to a higher crack density compared to UA120 (see Figure 3d). For the BA120 sample, which was biaxially stretched to 120%, the orientations of the cracks were significantly diversified in comparison with the uniaxially stretched samples. The overall shape and density of the cracks were comparable to the minor cracks of UA140; however, large cracks similar to the major cracks of UA140 with the well-defined edges were not observed in BA120. As the biaxial stretching ratio increased up to 140% (i.e., BA140), the Ag

film was eventually divided into island forms by major cracks, and an average area of the islands was found to be less than 3 μm2. In comparison with the other samples, minor cracks with a higher density were clearly observed within the Ag islands as shown in Figure 3h.

**Figure 3.** SEM images of the cracked Ag electrodes after completing the cold-welding processes. (**<sup>a</sup>**,**b**) SEM images of the UA120 sample. No further deformation of Ag was observed between crack lines as in (**b**). (**<sup>c</sup>**,**d**) SEM images of the UA140 samples. Minor cracks were further observed between major crack lines as in (**d**). (**<sup>e</sup>**,**f**) SEM images of the BA120 sample. Orientations of cracks were significantly diversified in comparison with the uniaxially stretched samples as in (**f**). (**g**,**h**) SEM images of the BA140 sample. The Ag film was divided into island forms as in (**g**), and minor cracks with a high density were clearly observed within the Ag islands as in (**h**).

Figure 4 also shows cross-sectional SEM images of the BA140 sample. For these SEM measurements, platinum (Pt) was coated onto the sample with a thickness of ~5 nm to clearly observe the non-conductive glass substrate. It was confirmed that the cracked Ag film was neatly transferred and welded on the intact Ag film through the cold-welding process. It is worth noting here that metal oxide layers could not be easily transferred and welded onto other substrates in our experimental setup, and thus binding materials would be required to enhance adhesion.

The J-V characteristics for the devices were measured in response to a voltage sweep from 0.0 to +1.0 V, as shown in Figure 5. It should be noted that the charge carriers are normally accumulated in the organic layers in high bias voltage region, due to the relatively low charge carrier mobility. Such charge carrier accumulations in the organic layers give a rise to changes not only in the electric field distributions, but also in the J-V characteristics (i.e., from ohmic currents to space-charge-limited currents) [17,26,28,48,49]. Thus, changes in the charge injection properties according to the local electric field enhancement can be more clearly examined in low bias region (i.e., ohmic current region). As indicated in Figure 5, the current density was gradually increased as the density of the cracks on the Ag electrode increased. In particular, it was found that the slope ratio of the device with the BA140 electrode to the reference devices was ~170 in the J-V characteristics. At a fixed bias voltage of +1.0 V, the current densities were measured to be 6.68 × <sup>10</sup>−8, 6.60 × <sup>10</sup>−7, 1.08 × <sup>10</sup>−6, 2.20 × <sup>10</sup>−6, and 1.16 × 10−<sup>5</sup> A cm<sup>−</sup><sup>2</sup> for reference, UA120, BA120, UA140, and BA140, respectively. It is worth noting that EHCz is an intrinsically p-type material, of which conduction is entirely governed by holes [1], and the devices presented in this work can be described as hole-only devices.

All the devices showed linear J-V characteristics within the bias voltage range (i.e., 0.0 to +1.0 V), indicating the ohmic behaviors of the devices in low bias region. If there is no change in charge carrier mobility, the ohmic currents of organic electronic devices at constant temperature are normally enhanced as the initial concentrations of charge carriers within the semiconducting layers increase using ionic dopants [31,32]. However, in this work, the increases in the ohmic currents were solely induced by the injected charges from the electrodes without using any extra dopants. These results strongly sugges<sup>t</sup> that efficient charge injection via local electric field enhancement can exert similar effects to the introduction of ionic dopants on J-V characteristics in terms of charge carrier concentrations. It is notable that the maximum ohmic current density of 1.16 × 10−<sup>5</sup> A cm<sup>−</sup><sup>2</sup> in this work is relatively lower than those of other solid-state devices based on carbazole derivatives or similar organic semiconducting materials. In ITO/undoped organic semiconductor/metal structures, previous works reported ohmic current densities of ~10−<sup>4</sup> A cm<sup>−</sup><sup>2</sup> for 4,4,4-tris(N-3-methylphenyl-N- phenyl-amino)-triphenylamine [32], poly(2,6-diphenyl-4-((9-ethyl)-9*H*-carbazole)-pyridinyl-*alt*-2,7-(9,9-didodecyl)-9*H*-fluorenyl) [50], and N,N-bis(3-methylphenyl)-N,N-diphenylbenzidine [51], and even higher ohmic current densities

were also found with a few of carbazole derivatives [52]. In addition, a few photoluminescent devices were successfully demonstrated using EHCz as a host material in the previous studies [5,6,8]; however, at the current stage, we could not observe electroluminescence from dye-doped EHCz materials due to the imbalance of electrons and holes.

**Figure 5.** (**a**) Current density-voltage (J-V) characteristics of the devices using the reference, UA120, UA140, BA120, and BA140 electrodes, in a linear scale. The characteristics were measured in low bias region (i.e., ohmic current region). All the characteristics showed a linear relationship, and that the slope ratio of the device with the BA140 electrode to the reference devices was ~170. (**b**) Electric field dependence of the current density. All the slopes were measured to be ~1, indicating ohmic current in each device.

The field intensity calculation was also performed to investigate the local electric field enhancement induced by a highly cracked metal structure as shown in Figure 6. The cracked Ag structure with a depth of 30 nm was considered for the calculation, on the basis of the observed SEM image (see Figure 6b). In the simulation model, the gap distance between the lower Ag and upper ITO electrodes was fixed at 5 μm, and the gap was filled with a dielectric material corresponding to EHCz. For the dielectric material, dielectric constant of 3.02, electrical conductivity of 2.1 × 10−<sup>9</sup> S cm<sup>−</sup>1, and density of 1.004 g mL−<sup>1</sup> were used as material parameters to simulate EHCz. It should be noted that the dielectric constant of EHCz was experimentally measured for this work. For the measurement, EHCz was injected into the gap between two Ag electrodes with a gap distance of 5 μm, of which the capacitance was monitored at a frequency of 1 kHz. The measured capacitance was converted into the dielectric constant in consideration of active area and thickness.

Static analysis was performed to clarify the spatial distribution of local electric fields within the device at a fixed bias voltage of +1.0 V. According to the distance from the Ag bottom, the local electric field intensities were calculated for the selected cross sections, as shown in Figure 6a. The local electric field intensities were significantly enhanced within the cracks of the Ag electrode, of which apexes were found at the upper edges. In particular, the intensity of the local electric field increased up to ~4000 V cm −1 at the upper edges. To clearly show the local electric field enhancement, the field intensities for the selected cross section were further visualized as shown in Figure 6c. These calculation results were in good agreemen<sup>t</sup> with the experimental results, where charge injection properties were dramatically improved by the high-density cracks contributing to local electric field enhancement.

**Figure 6.** (**a**) Local electric field intensities according to the distance from the Ag bottom, based on the field intensity calculation at a fixed bias voltage of +1.0 V. Blue circles indicate the local electric field intensities in the device with a highly cracked Ag structure (30-nm depth), and a red dashed line indicates the electric field in the reference device. (**b**) Model structure configured for the field intensity calculation in (**a**), and SEM image of BA140 used to prepare the model structure. (**c**) Local electric field intensities visualized for a selected cross section. The boundary between Ag and EHCz is indicated by a white line. The intensity level, presented by colors, increases from blue to red (i.e., blue-green-yellow-red).
