Abstract
We present Raman measurements on the iron-pnictide superconductors ${\text{CeFeAsO}}_{1\ensuremath{-}x}{\text{F}}_{x}$ and ${\text{NdFeAsO}}_{1\ensuremath{-}x}{\text{F}}_{x}$. Modeling the Fe-As plane in terms of harmonic and a cubic anharmonic Fe-As interactions, we calculate the temperature dependence of the energy and lifetime of the Raman active $\text{Fe}\text{ }{B}_{1g}$ mode and fit to the observed energy shift. The shifts and lifetimes are in good agreement with those measured also in other Raman studies which demonstrate that the phonon spectrum, at least at small wave numbers, is well represented by phonon-phonon interactions without any significant electronic contribution. Even at zero temperature there is a non-negligent effect of interactions on the phonon energy, which for the $\text{Fe}\text{ }{B}_{1g}$ mode corresponds to $6\text{ }{\text{cm}}^{\ensuremath{-}1}$ or 3% of the total energy of the mode. We also estimate the anharmonic expansion from Fe $(56\ensuremath{\rightarrow}54)$ isotope substitution to $\ensuremath{\Delta}a\ensuremath{\approx}5.1\ifmmode\times\else\texttimes\fi{}{10}^{\ensuremath{-}4}\text{ }\text{\AA{}}$ and $\ensuremath{\Delta}{d}_{\text{Fe-As}}\ensuremath{\approx}2.5\ifmmode\times\else\texttimes\fi{}{10}^{\ensuremath{-}4}\text{ }\text{\AA{}}$ and the shift of harmonic zero-point fluctuations of bond lengths $⟨\ensuremath{\Delta}{x}^{2}⟩\ensuremath{\lesssim}3\ifmmode\times\else\texttimes\fi{}{10}^{\ensuremath{-}5}\text{ }{\text{\AA{}}}^{2}$, giving a total relative average decrease in electronic hopping integrals of $|\ensuremath{\delta}t|/t\ensuremath{\lesssim}2.0\ifmmode\times\else\texttimes\fi{}{10}^{\ensuremath{-}4}$. For a nonphonon-mediated weak-coupling superconductor this gives an isotope exponent $\ensuremath{\alpha}\ensuremath{\sim}{10}^{\ensuremath{-}2}$. The results pose a serious challenge for any theory of superconductivity in the pnictides that does not include electron-phonon interactions to produce a sizable Fe isotope effect.
Introduction
In the study of high-temperature superconductivity in iron-pnictides, a central debate concerns the nature of the pairing mechanism: whether it is driven by magnetic fluctuations or lattice vibrations (phonons). Experimental evidence has shown a significant iron isotope effect ($\alpha \approx 0.4$), where substituting $^{56}\text{Fe}$ with the lighter $^{54}\text{Fe}$ shifts the superconducting transition temperature ($T_c$). However, many theoretical models suggest that electron-phonon coupling in these materials is too weak to explain such high $T_c$ values. This paper investigates whether the observed isotope effect could be an indirect result of anharmonic lattice expansion rather than direct electron-phonon coupling.
Experimental Observations
Raman spectroscopy was performed on polycrystalline samples of $\text{CeFeAsO}_{0.84}\text{F}_{0.16}$ and $\text{NdFeAsO}_{0.88}\text{F}_{0.12}$ across a temperature range of 20 K to 300 K. The researchers specifically tracked the $\text{Fe } B_{1g}$ mode, which involves the movement of iron atoms within the Fe-As plane.
As shown in the spectra, the phonon peaks exhibit softening (shifting to lower frequencies) and broadening as the temperature increases. These changes are characteristic of anharmonic lattice vibrations.
Theoretical Modeling and Results
The authors modeled the Fe-As plane using a harmonic spring model supplemented by a cubic anharmonic interaction term. Using Green’s function methods, they calculated the phonon self-energy to determine how temperature and atomic mass affect the phonon's energy shift and lifetime (linewidth).
The model provides an excellent quantitative fit for both the energy shift and the linewidth of the Raman modes, with an estimated cubic anharmonic coupling strength of $g \approx 100\text{--}116 \text{ eV/\AA}^3$. This demonstrates that the temperature-dependent behavior of these phonons is driven by phonon-phonon scattering (anharmonicity) rather than electronic effects. Notably, even at absolute zero, anharmonic interactions contribute approximately $6 \text{ cm}^{-1}$ to the energy of the $\text{Fe } B_{1g}$ mode.
The Isotope Effect Challenge
A critical finding of this study relates to the iron isotope effect. The researchers calculated that substituting $^{56}\text{Fe}$ for $^{54}\text{Fe}$ results in a tiny lattice expansion of $\Delta a \approx 5.1 \times 10^{-4} \text{ \AA}$ due to anharmonicity. While this expansion changes the electronic hopping integrals ($t$), the resulting shift in $T_c$ (the isotope exponent $\alpha$) is calculated to be only $\sim 10^{-2}$.
This calculated value is an order of magnitude smaller than the experimentally reported $\alpha \approx 0.4$. Even with generous assumptions, the upper bound for $\alpha$ from this indirect mechanism is approximately 0.06.
Conclusion
The results present a serious challenge to theories of superconductivity in iron-pnictides that dismiss the role of electron-phonon interactions. Because the indirect effect of lattice expansion is insufficient to explain the large observed isotope shift, the study suggests that electron-phonon coupling must play a more significant and direct role in the pairing mechanism than currently assumed by many weak-coupling models.
Summary updated