Supporting Information Inverse I-V injection characteristics of ZnO nanoparticle based diodes Paul Mundt 1,2, Stefan Vogel 3, Klaus Bonrad 2,4, Heinz von Seggern 1 * Technische Universität Darmstadt 1 Institute of Materials and Geo Sciences, Electronic Materials Group, Alarich-Weiss-Str. 2, Darmstadt, Germany 2 MerckLab TU-Darmstadt, Germany 3 Institut of Materials Science, Advanced Thin Film Technology, Alarich-Weiss-Str. 2, Darmstadt, Germany 4 Merck KGaA, Darmstadt, Germany * corresponding author Email: seggern@e-mat.tu-darmstadt.de S-1
In this paper the I-V characteristics of ZnO nanoparticle (ZnO(NP)) based diodes are compared to diodes with a ZnO layer produced by reactive sputtering (ZnO(SP)). Both ZnO materials exhibit n-type semiconducting behavior. This was confirmed by thin film transistor measurements of ZnO(NP) based devices. The transfer characteristic at V S/D = 30V of such a thin film bottom gate/bottom contact device is shown in Fig. S1. A p-type transport mechanism cannot be seen since at negative applied gate voltages no increase in the device current is observed. Figure S1: Transfer characteristic of the ZnO-NP based thin film Transistor at V S/D = 30V. Increased current for positive V G reveals n-type transport. No increase in device current for negative V G indicates nonexistence of p-type transport. Based on the fact that both ZnO materials exhibit n-type transport and the energetic position of the contact metals, the ZnO(NP) diode exhibits an inverse I-V characteristic compared to the ZnO(SP) diode. This inverse behavior is attributed to the different interface morphologies of ZnO(NP) and ZnO(SP) at the electrodes. The ZnO(NP) is rough with a lot of spikes, where Au tips are formed during the evaporation process of the top electrode, the ZnO(SP) is, however, smooth with no spikes. Pure ZnO(NP) and ZnO(SP) films were measured with the Al electrode on the rear side of the ZnO layer using an Asylum Research MFP-3D and HQ:NSC19/AlBS tips. Line profiles and statistics were done using Gwyddion software. Fig. S1 shows the AFM scan of the ZnO(NP) layer and Fig. S2 shows the one of the ZnO(SP) layer. S-2
Figure S2: AFM measurement of the ZnO(NP) layer. The Ra and Rms values are extracted using the complete measurement area before image enhancing technique was used. Figure S3: AFM measurement of the ZnO(SP) layer. Scan area is 1x1 µm. The Ra and Rms values are extracted using the complete measurement area before image enhancing technique was used. For better comparison of the two surfaces both profile lines are plotted together in Fig. S3. Au evaporated onto the ZnO(NP) (black line) can adapt the shape of the nanoparticle surface and forms sharp tips. The ZnO(SP) surface (red line) on the other hand is much smoother without any large tips. S-3
Figure S4: Line profiles of ZnO(NP) (black) and the ZnO(SP) (red) extracted from AFM measurements. To investigate the contact formation between the bottom Al electrode and the sputtered ZnO layer, SEM cross section pictures of a ZnO(SP) layer is shown in Fig. S5. No voids are visible neither in the film nor at the interfaces. Combined with the AFM measurements the results indicate good wetting of the ZnO(SP) layer with the bottom and top electrode. Figure S5: SEM cross section image of the ZnO(SP) layer structure. To investigate the thickness dependency of the ZnO(NP) diodes the measured currents in Fig. 2a have to be corrected. As mentioned in the article the measured currents for voltages U > -1 V have to be interpreted as leakage currents. Since these leakage currents are Ohmic they should differ by their thickness ratio which is 500/300 = 1,666. This value is in the range of the statistical fluctuation of the diode characteristics. To elaborate this, the data of Fig. 2a was corrected, assuming that the current of both diodes after the shape transition (U < -1V) are similar. In this case the thinner diode (green curve) shows higher currents for U > -1 V as can be seen in Fig. S6. That the measured current (red curve) for U < -1 V is not the same for both S-4
diodes has most probably its origin in the layer and interface formation of the nanoparticles which is a statistical process and therefore statistical fluctuations can be expected. Figure S6: I-V characteristics of ZnO(NP) layers with different thicknesses. The current of the 300 nm thick ZnO(NP) diode is corrected (green curve) so that the currents for voltages U < -1 V are similar to the one of the diode with a 500 nm ZnO(NP) layer. For an easier understanding both diode types without bias, under forward bias and under reversed bias are schematically shown in Fig. S7. The voltage refers to the voltage applied at the top Au electrode, as indicated in Fig. S7 as well as mentioned in the paper. For the sputtered ZnO diode a lower injection barrier occurs for positive voltage applied at the Au electrode. In case of the ZnO(NP) diode an energetic barrier, which is caused by the particle character of the ZnO material, occurs at the surface of the bottom Al electrode. This barrier hinders electron injection from the bottom Al electrode. For negative voltage applied it leads to a pile up of charge which generates a sufficient potential difference between the ZnO(NP) and the Al leading to a very strong field in the barrier and therefore to a shape transition thereof. The facilitated electron injection at the top Au electrode caused by the electric field enhancement at the tips is indicated as a lowered injection barrier between Au and ZnO(NP). S-5
Figure S7: Schematic diagram of a) the ZnO(SP) diode and b) the ZnO(NP) diode each without, negative and positive external bias. S-6