[Paper Review] Is CrO$_2$ Fully Spin-Polarized? - Analysis of Andreev Spectra and Excess Current

Tomas Löfwander, Roland Grein+1 more|arXiv (Cornell University)|Jul 19, 2010

X-ray Diffraction in Crystallography3 citations

TL;DR

This paper reanalyzes point-contact Andreev reflection (PCAR) data on CrO2 using a spin-active interface model, demonstrating that fully spin-polarized CrO2 bands combined with spin-dependent scattering at the contact can explain all experimental spectra—offering a consistent alternative to extended Blonder-Tinkham-Klapwijk (BTK) models that yield inconsistent spin polarization values between 50% and 100%. The authors propose that temperature-dependent zero-voltage conductance and excess current measurements can experimentally distinguish between models and confirm full spin polarization in CrO2.

ABSTRACT

We report an extensive theoretical analysis of point-contact Andreev reflection data available in literature on ferromagnetic CrO$_2$. We find that the spectra can be well understood within a model of fully spin-polarized bands in CrO$_2$ together with spin active scattering at the contact. This is in contrast to analyses of the data within extended Blonder-Tinkham-Klapwijk models, which lead to a spin polarization varying between 50~\% and 100~\% depending on the transparency of the interface. We propose to utilize both the temperature dependence of the spectra and the excess current at voltages above the gap to resolve the spin-polarization in CrO$_2$ in a new generation of experiments.

Motivation & Objective

• To resolve the inconsistency in reported spin polarization values (50–100%) from extended BTK model fits to PCAR data on CrO2.
• To test whether fully spin-polarized CrO2 bands with spin-active interface scattering can explain the same data more consistently.
• To propose new experimental observables—temperature dependence of zero-voltage conductance and excess current above the gap—to distinguish between competing models.
• To provide a theoretically consistent framework for interpreting PCAR data in half-metallic ferromagnets like CrO2.
• To challenge the conventional interpretation of PCAR spectra that attributes spectral features to variable bulk polarization, instead attributing them to interface scattering effects.

Proposed method

• The authors apply a spin-active interface model to fit PCAR spectra, assuming fully spin-polarized bands in CrO2 (P = 100%) and introducing two interface parameters: barrier strength Z and spin mixing angle ϑ.
• They compare this model with the extended BTK model, which assumes variable bulk polarization P and includes a barrier term Z, a gap parameter Δ, and a series resistance rs as fit parameters.
• The fitting procedure uses experimental conductance-voltage curves from multiple studies, with rs normalized to the normal contact resistance Rn.
• The models are evaluated using reduced chi-squared (χ²) statistics to assess goodness of fit, with emphasis on low-voltage and high-voltage spectral regions.
• Theoretical predictions for temperature-dependent conductance and excess current are derived and compared between models to identify experimentally measurable discriminators.
• The authors use analytical expressions for excess current I_exc = lim_{V→∞} [I(V) - I_n(V)] to predict observable signatures in the high-voltage regime.

Experimental results

Research questions

• RQ1 Can the observed PCAR spectra of CrO2 be consistently explained by a model assuming fully spin-polarized bands and spin-active interface scattering, rather than variable bulk polarization?
• RQ2 Why do extended BTK model fits yield a wide range of spin polarization (50–100%) despite CrO2 being a candidate half-metal?
• RQ3 How can the temperature dependence of zero-voltage conductance serve as a diagnostic tool to distinguish between the spin-active interface and extended BTK models?
• RQ4 Can excess current measurements above the superconducting gap provide a direct experimental test for the spin-active interface model?
• RQ5 What role does the series resistance rs play in fitting PCAR spectra, and why is it more critical in the extended BTK model than in the spin-active interface model?

Key findings

• The spin-active interface model with fully spin-polarized CrO2 (P = 100%) provides a consistent and improved fit to all available PCAR data, with reduced χ² values comparable to or better than the extended BTK model.
• The extended BTK model yields inconsistent spin polarization values (50–100%) depending on the fit, suggesting it is not a reliable probe of bulk polarization in CrO2.
• The zero-voltage conductance G(0) in the spin-active interface model vanishes as T → 0, a unique feature not present in the extended BTK model, offering a key experimental test via temperature dependence.
• Excess current predictions differ significantly between models: the spin-active interface model predicts a non-monotonic dependence on barrier strength Z, while the extended BTK model shows a different trend, making it measurable in future experiments.
• The fit quality improves when including series resistance rs as a parameter, but this parameter is more critical in the extended BTK model, highlighting its role in accounting for experimental artifacts.
• The study identifies that the low-voltage region of spectra is the most challenging to fit due to sparse data points, suggesting that improved experimental setups with better low-V resolution are needed.