Quantum Transport in Ultrahigh-Conductivity Carbon Nanotube Fibers

TL;DR

This work probes quantum transport in ultrahigh-conductivity, solution-spun CNT fibers, revealing WL as a central mechanism at low temperature. By testing 1D, 2D, 3D WL theories and introducing a hybrid 3D+1D framework, the authors show that transport occurs along quasi-1D CNT bundles within a 3D network, with a temperature-driven crossover captured by a crossover parameter $M(T)$. The analysis yields long phase coherence lengths, notably $L_{\varphi,3D} \approx 180$ nm at 1 K, and demonstrates that conventional WL models alone cannot describe the data without unphysical scaling, necessitating a hierarchical, mixed-dimensional approach with a sizable magnitude adjustment (~$200$) for the 3D term. High-temperature transport is reconciled with phonon scattering via a Pop-like metallic regime, completing a comprehensive view of both quantum corrections and phonon-limited conduction. Collectively, the results provide a physically grounded framework for optimizing macroscopic CNT conductors through controlled alignment, doping, and interbundle coupling for flexible, low-loss power transmission.

Abstract

We investigate quantum transport in aligned carbon nanotube (CNT) fibers fabricated via solution spinning, focusing on the roles of structural dimensionality and quantum interference effects. The fibers exhibit metallic behavior at high temperatures, with conductivity increasing monotonically as the temperature decreases from room temperature to approximately 36 K. Below this temperature, the conductivity gradually decreases with further cooling, signaling the onset of quantum conductance corrections associated with localization effects. Magnetoconductance measurements in both parallel and perpendicular magnetic fields exhibit pronounced positive corrections at low temperatures, consistent with weak localization (WL). To determine the effective dimensionality of electron transport, we analyzed the data using WL models in 1D, 2D, and 3D geometries. We found that while the 2D model can reproduce the field dependence, it lacks physical meaning in the context of our fiber architecture and requires an unphysical scaling factor to match the experimental magnitude. By contrast, we developed a hybrid 3D+1D WL framework that quantitatively captures both the field and temperature dependences using realistic coherence lengths and a temperature-dependent crossover parameter. Although this combined model also employs a scaling factor for magnitude correction, it yields a satisfactory fit, reflecting the hierarchical structure of CNT fibers in which transport occurs through quasi-1D bundles embedded in a 3D network. Our results establish a physically grounded model of phase-coherent transport in macroscopic CNT assemblies, providing insights into enhancing conductivity for flexible, lightweight power transmission applications.

Figures (9)

• Figure 1: Hierarchical structure of CNT fibers. (a–c) Schematic illustration of the multiscale organization formed during spinning: (a) individual CNTs, (b) aggregation into bundles (each slender cylinder represents one CNT), and (c) alignment of bundles into a macroscopic fiber (each elongated element denotes one bundle). (d–f) Corresponding microscopy images: (d) TEM image of raw CNTs before fiber spinning (the inset shows a zoomed-in region illustrating the double-walled tubular structure), (e) cross-sectional TEM image showing densely packed bundles, and (f) SEM image of the fiber surface exhibiting high alignment. Together, these images demonstrate how van der Waals interactions and flow-induced alignment give rise to long-range structural order across scales.
• Figure 2: Temperature dependence of the electrical conductance for the ICl-doped fiber sample. The conductance increases upon cooling, reaches a maximum near 36 K, and decreases at lower temperatures. This nonmonotonic behavior is consistent with previous reports on doped carbon nanotube materials Behabtu2013. The curve shows densely sampled experimental data points directly connected without smoothing.
• Figure 3: (Color online) Differential conductance as a function of magnetic field for (a) perpendicular and (b) parallel orientations, measured at various temperatures. The conductance is expressed in units of $e^2/h$. Data were taken at $T = 1~\mathrm{K}$ (blue), $10~\mathrm{K}$ (red), $25~\mathrm{K}$ (yellow), and $50~\mathrm{K}$ (purple).
• Figure 4: Fit of the 1D WL model to the measured $\Delta G(B)$ at 1 K. The model captures the overall scale of the conductance correction but fails to reproduce the detailed field dependence, overestimating the curvature at low fields. The bundle cross-sectional area used in the fit is $A_\text{b} = (40~\mathrm{nm})^2$.
• Figure 5: Two-dimensional weak localization fits to the differential conductance data under perpendicular magnetic fields at four representative temperatures: 1 K, 10 K, 25 K, and 50 K. The experimental data (circles) are fitted using the model from Ref. Piraux2015, as given in Eq. \ref{['eq:2D_WL']}, with $L_{\varphi, 2\text{D}}$ as the sole free parameter. While the fits qualitatively reproduce the field dependence, they require a prefactor $N_{\text{LAYER}}$ to match the magnitude of the measured $\Delta G$, and the extracted coherence lengths are shown in each panel.