Functional Printing of Conductive Silver-Nanowire Photopolymer Composites

We investigated the fabrication and functional behaviour of conductive silver-nanowire-polymer composites for prospective use in printing applications. Silver-nanowires with an aspect ratio of up to 1000 were synthesized using the polyol route and embedded in a UV-curable and printable polymer matrix. Sheet resistances in the composites down to 13 Ω/sq at an optical transmission of about 90% were accomplished. The silver-nanowire composite morphology and network structure was investigated by electron microscopy, atomic force microscopy, profilometry, ellipsometry as well as surface sensitive X-ray scattering. By implementing different printing applications, we demonstrate that our silver nanowires can be used in different polymer composites. On the one hand, we used a tough composite for a 2D-printed film as top contact on a solar cell. On the other hand, a flexible composite was applied for a 3D-printed flexible capacitor.

: Light microscopy images (Olympus light microscope BX 51) of (a) Ag-NW applied via dip coating, (b) Ag-NW layer produced by spin coating, (c) Ag-NW applied with a glass rod and (d) Ag-NW percolation network fabricated via drop casting. S3 at the edge was observed. In the middle of the samples, uniform Ag-NW percolation networks could be produced (Fig. S1 (d)). By using a frame of 3 cm 2 as a template during the drop casting process, an Ag-NW layer with well defined area could be produced.

S1.3. Argon Plasma Treatment
For the plasma treatment with Argon a Harrick Plasma Cleaner PDC-002 (230 V) was used at 22 °C.
Before starting the Argon stream, the sample chamber was evacuated to a vacuum in the range of 1.7×10 -4 -3.2×10 -4 mbar. During the plasma treatment, the pressure with flow of Argon in the chamber was between 1.0×10 -2 and 1.2×10 -2 mbar.
By using Argon plasma treatment, the polyvinylpyrrolidone (PVP) shell, which stabilizes the nanowires during the synthesis, was removed resulting in a decrease in sheet resistance (see inset of Fig. S2 (a)). The PVP shell can be clearly seen in the SEM image shown in Fig. S2 (a). At the interconnects of the Ag-NWs, partial welding occur due to the Argon plasma treatment. 2

S1.4. UV-curable Resin
For the optimization of the resin the concentrations of the different components (initiator, monomer and cross-linker) were varied. The aim was the minimization of the layer thickness and surface roughness. A cross-linker was added to increase the viscosity and reduce the sedimentation. Beside the fabrication of layered structures (sandwich structure: polymer-Ag-NW-polymer), Ag-NWs were mixed with the polymer and the filler material could be kept in S4 suspension for several hours. Fig. S3 shows a light microscopy image of the produced Ag-NWpolymer composite.

Tab. S1:
Components of the used resin and their molecular formula und molecular weight.

S1.5. Curing Setup
The resin was cured in a curing setup -a schematic representation is shown in Samples with silicon substrate were exposed from above.

S1.6. Sample Overview
Tab. S2 shows a summary of the produced Ag-NW networks and composite materials and their key properties like sheet resistance, transmission, film thickness and roughness. Film thickness and roughness were determined by profilometry (S2.2).

S2.1. Scanning Electron Microscopy (SEM) and Energy Dispersive X-ray Spectroscopy (EDX)
The study of the surface morphology and shape of the samples were carried out on a commercial

S7
The chemical composition of the silver nanowire samples was investigated via energy dispersive X-ray spectroscopy (EDX, Zeiss). The spectrum of a plasma untreated Ag-NW percolation network shows a silicon peak (substrate) and silver peaks (Ag-NW). Additionally, carbon and oxygen were detected because of the presence of polyvinylpyrrolidone (PVP, stabilizing ligand) ( Fig. S7 (a)).
After argon plasma treatment (5 h), the PVP shell was completely removed and no carbon and oxygen were detected by EDX ( Fig. S7 (b)).

S2.2. Profilometry and Atomic Force Microscopy (AFM)
Profilometry measurements were performed with a Dektak XT equipment from Bruker company to determine layer thickness and surface roughness. Furthermore, the surface topography of the samples was investigated with an NT-MDT Solver NEXT tabletop scanning probe microscope in S8 semi contact (tapping) mode and an Anfatec Level AFM in tapping mode. Fig. S8 (a) shows an AFM image of a silver nanowire percolation network (c = 58 µg/cm 2 ) on a silicon substrate. The Ag-NW layer thickness is ~600 -700 nm. Exemplary AFM images for a pure polymer layer, an Ag-NW polymer composite and a sandwich sample are shown in Fig. S8 (b)-(d). The mean roughness (Ra) and root-mean-squared roughness (Rms) could be determined. The resulting roughnesses are similar to the mean roughnesses determined via profilometry measurements of the same samples (compare Fig. 1 (c) of the main text). Via AFM, a higher resolution was achieved, but a smaller sample area could be measured.

S2.3. Conductivity Measurements (Van-der-Pauw)
For resistivity measurements in Van der Pauw geometry, DPP 105-M/V-Al-S positioners (CascadeMicrotech, USA), a DC voltage/current source GS200 (Yokogawa, Japan) and the 34401A 6 ½ Digit Multimeter (Keysight, USA) were used. Four points of silver lacquer were applied on the edge of the sample in order to achieve a good contact to the Ag-NW network. An input current was applied at the points A and B (Fig. S9) and was varied between -0.6 mA and 0.6 mA. The voltage drop at the points on the opposite side (C and D) was measured. The measurement was repeated with an input current at A and D and the voltage was measured between B and C. The sheet resistance was calculated according to the Van der Pauw approximation 6 . A photograph of the setup is shown in Fig. S9 (b). The green line is a guide to the eye. Each point represents one deposition/measurement.

S9
As seen in Fig. S9 (c) showing our results of washed Ag-NW samples, we find a critical nanowire concentration of (25 ± 10) µg/cm 2 above which we reproducibly find conducting nanowire networks. In this range, the sheet resistance drops dramatically reaching values of down to 10 Ω/sq. At the critical nanowire concentration, we also observe some fluctuations in sheet resistance, which relate to the fact that at low concentration the probability of good electrical connection decreases and sensitivity towards the exact results of the washing procedure increases. However, the Ag-NW network showing a sheet resistance of around 500 Ω/sq at a concentration of only 10 µg/cm 2 indicates that with further optimization better conductivities at even lower Ag-NW concentration could be feasible.

S2.4. Ellipsometry
In addition to the characterization methods mentioned in the main text, ellipsometry was used to investigate the homogeneity and anisotropy of silver nanowire layers. A spectroscopic ellipsometer (SE-850) from Sentech was used. The following three samples were investigated: W1 (15 µL Ag-NW suspension), W2 (2 x 15 µL Ag-NW suspension) and W3 (3 x 15 µL Ag-NW suspension). For the samples W2 and W3 a second (and a third) Ag-NW layer was applied after the previous layer was dried. The homogeneity test was carried out at an incidence angle of 70° and different beam diameters. The phase difference Δ was determined. In Fig. S10 (a) the difference spectra are shown. Sample W1 (one layer Ag-NW) is homogeneous, sample W2 (2 layers) shows a slight inhomogeneity and W3 (3 layers) a significant inhomogeneity. To S10 investigate the anisotropy, the samples were rotated 45° for the second measurement. The incidence angle was 70° and the beam diameter was kept constant. The difference spectra are shown in Fig. S10 (b). Sample W1 is nearly isotropic. With increasing layer number, the anisotropy is increasing.

S2.5. Grazing Incidence Small Angle X-ray Scattering (GISAXS)
GISAXS measurements were carried out at the micro-and nanofocus X-ray scattering beamline MiNaXS/P03 at PETRA III (DESY) 8 . We used a sample-detector-distance (SDD) of 4990 ± 1 mm (polymer sample: 3600 ± 1 mm) and a wavelength of 0.9724 Å. The schematic illustration of the GISAXS setup is shown in Fig. S11 (a). The direct beam was located at x = 545 pixel and y = 15 pixel (polymer sample: x = 532 pixel, y = 72 pixel). The incidence angles are listed in Tab. S3. After height, tilt and angle adjustment, the samples were scanned through the beam with a step size of 50 µm and an acquisition time of 1 s in order to avoid radiation damage (x-scan). An example for a stack plot of the GISAXS cuts at y = 365 pixel (corresponding to qz = 0.78 nm -1 , width = 10 S11 pixel) is shown in Fig. S11 (c). In order to increase in statistics, all GISAXS images of one scan were summed up. The summed-up images are shown in the main text of this manuscript. For analysis of the data, we used the software DPDAK 9 (v1.3.0). The horizontal cuts were carried out at y = 300 pixel (qz1 = 0.63 nm -1 ), y = 365 pixel (qz2 = 0.78 nm -1 ) and y = 445 pixel (qz3 = 0.96 nm -1 ) with a height of 10 pixel, respectively. By following the qy-position of the peaks caused by the flares (arrows in Fig. 2(g)-(i)), an angle of the flares of around 36° was found. The positions of the horizontal cuts are marked in Fig. S11 (b).

Tab. S3:
Incidence angles of the measured samples.

S12
The influence of Argon plasma treatment on the nanowire geometry was studied using GISAXS. In Fig. S12, the flares have been strongly reduced due to the Argon plasma treatment of the Ag-NWs in comparison to Fig. 2(a) in the manuscript. Argon plasma treatment was used to remove the PVP layer and to improve the conductivity of the Ag-NW network (inset SI S1.3).
However, during the plasma treatment, the surface of the Ag-NWs becomes rougher due to defects 4,5 and the shape more round due to the removal of edges. 10 Furthermore, at the interconnects of the Ag-NWs, partial welding and sputtering occurs 2,3 (see SI Fig. S2), which also gives rise to additional scattering from spherical structures as observed by the increased diffuse scattering in Fig. S12.
Furthermore, plasma treatments are rather incompatible with a 3D-printing approach. For this reason, we concentrate on the removal of PVP through the controlled washing of the nanowires.

S2.6. GISAXS Simulation
We used the software IsGISAXS 11 V1.6 for simulating the GISAXS pattern. The actual sample morphology in terms of thin film is a 3D percolated network of Ag-NW with a film thickness of ~600 nm. The mean diameter and mean length of an individual Ag-NW is 130 nm and 10 µm (see a 3D network is present, where the Ag-NWs touch each other at interconnects (see Fig. S6), while at the same time the porosity is around 11%. Reflection and refraction effects were included by using the distorted-wave Born approximation (DWBA) (comparable to previous work, see 12 ). This choice of parameters allowed for reproducing the key scattering features in a reasonable way.