Optical characterization of highly conductive single-wall carbon-nanotube transparent electrodes

We report on a complete characterization of the optical dispersion properties of conducting thin films of single-wall carbon nanotubes (SWCNTs). The films studied exhibit sheet resistances between 50 and $1000\Omega ∕\mathrm{sq}$ and optical transparencies between 65% and 95% on glass and quartz substrates. These films have the potential to replace transparent conducting oxides in applications such as photovoltaics and flat-panel displays; however, their optical properties are not sufficiently well understood. The SWCNT films are shown to be hole conductors, potentially enabling their use as hole-selective contacts and allowing alternative device designs. The fundamental optical, morphological, and electrical characteristics of the films are presented here, and a phenomenological optical model that accurately describes the optical behavior of the films is introduced. Particular attention is paid to ellipsometry measurements and thorough evaluation of the reflection and absorption spectra of the films.

Figure 1

AFM image of a typical $50\Omega ∕\mathrm{sq}$ SWCNT network on quartz. This film is $40\mathrm{nm}$ thick with an rms roughness of $31\mathrm{nm}$.

Figure 2

(a) Resistivity as a function of temperature indicates a minimum resistivity at $225\mathrm{K}$ with a sharp increase at elevated temperatures. (b) Seebeck voltage as a function of temperature variation across a $2\mathrm{cm}$ substrate for a $50\Omega ∕\mathrm{sq}$ film on glass. The positive Seebeck coefficient indicates a hole-conducting material.

Figure 3

(a) Optical transmission of bare SWCNT networks on glass as a function of sheet resistance varying from $50\text{to}1000\Omega ∕\mathrm{sq}$ and wavelength. (b) shows the approximated absorbance calculated from $1-T$ for a $50\Omega ∕\mathrm{sq}$ film on quartz. The line was used to perform a linear subtraction for peak fitting as described in Ref. 22.

Figure 4

Experimentally measured and modeled (a) $\Psi$ and (b) $\Delta$ as a function of energy at 50°, 60°, and 70° angles of incidence for a $50\Omega ∕\mathrm{sq}$ SWCNT films on quartz. These fits were used to generate the optical constants, and the predicted transmission, reflection, and absorption spectra in WVASE32 that are shown in Figs. 5, 6, 7.

Figure 5

Real $\left({\epsilon }_1\right)$ and imaginary $\left({\epsilon }_2\right)$ dielectric constants as a function of energy for 50 and $100\Omega ∕\mathrm{sq}$ films. The values of the optical constants completely overlap for the 50 and $100\Omega ∕\mathrm{sq}$ films. The optical constant values were determined from fits to spectroscopic ellipsometry and plane-polarized transmission measurements. Peak assignments are from Refs. 20, 21.

Figure 6

(a) Reflection, (b) transmission, and (c) absorption spectra for a $50\Omega ∕\mathrm{sq}$ SWCNT film on quartz. (a) Measured reflection spectra (solid line) with the WVASE-model-predicted reflectance (dotted line) as a function of energy. (b) Measured transmission spectra with the WVASE-model-predicted transmission spectrum. (c) Measured absorption spectrum $(A=1-T-R)$ with the WVASE-model-predicted spectrum. The WVASE model data were generated from the optical constants shown in Fig. 5.

Figure 7

Angle-dependent reflectance spectra for a $50\Omega ∕\mathrm{sq}$ film on quartz. Measured reflection spectra were taken on the M-2000 ellipsometer using (a) $p$- and (b) $s$-polarized light at angles varying from 15° to 70°. The modeled data points for (c) $p$- and (d) $s$-polarized reflection spectra were calculated using WVASE from the $\Psi$, $\Delta$, and $T$ data and without fitting to the reflection data.