Reexamining the strange metal charge response with transmission inelastic electron scattering

This study utilizes high-resolution transmission electron energy-loss spectroscopy on Bi-2212 to demonstrate that the strange metal's charge response at large momenta is an incoherent, non-dispersing continuum, thereby resolving decades of conflicting results by refuting the existence of RPA-like plasmons in favor of strongly damped excitations.

The Gist

The Big Mystery: The "Strange Metal"

Imagine you have a crowd of people at a concert. In a normal metal (like copper wire), the electrons are like well-behaved dancers. They move in a predictable rhythm, bumping into each other occasionally but mostly following the rules of physics we've known for a century. If you turn up the heat (temperature), they get a bit more chaotic, but you can still predict exactly how messy they will get.

Now, imagine a Strange Metal. This is a special state of matter found in materials like high-temperature superconductors (the stuff that makes magnets float). In a strange metal, the electrons are like a mosh pit at a heavy metal concert. They are so entangled and chaotic that the usual rules of dancing (physics) break down. They don't just bump into each other; they seem to move as one giant, confused blob.

For 40 years, scientists have been trying to understand this mosh pit. The key to understanding it is to watch how the electrons react when you push them. Specifically, scientists want to see if they form a Plasmon.

The Analogy: Think of a plasmon like a "wave" in a stadium. If you push the crowd in one spot, a ripple should travel across the stadium. In a normal metal, this wave is crisp, clear, and travels far. In a strange metal, scientists have been arguing for decades: Does this wave exist? Is it a clear ripple, or does it just dissolve into noise immediately?

The Great Debate: The "Ghost" Wave

For decades, scientists have been taking pictures of this electron wave in a material called Bi-2212 (a type of superconductor). But the pictures didn't match.

• Group A (The Optimists): In the 1980s and 90s, some scientists used a tool called Transmission EELS (which shoots electrons through a thin slice of the material) and claimed, "We see a clear, traveling wave! It behaves just like a normal metal."
• Group B (The Skeptics): Later, other scientists used a different setup or better tools and said, "No, we don't see a wave. We just see a fuzzy, messy blur. The wave dies instantly."

This was a huge problem. If you can't agree on what the electron wave looks like, you can't build a theory to explain how strange metals work. It's like two photographers taking pictures of the same ghost; one sees a clear face, and the other sees nothing but fog.

The New Investigation: A High-Definition Re-Do

The authors of this paper decided to settle the argument. They said, "Let's go back and take the picture again, but this time, let's use the best camera money can buy and make sure we don't mess up the photo processing."

They used a state-of-the-art electron microscope (a super-powerful camera) to shoot electrons through ultra-thin flakes of Bi-2212. They did this ten times on five different samples to make sure the results weren't a fluke. They also took a picture of Aluminum (a normal metal) to prove their camera was working correctly.

The Results:

1. The Aluminum Test: When they shot electrons through Aluminum, they saw a perfect, crisp wave. This proved their camera was working perfectly.
2. The Bi-2212 Test: When they shot electrons through the Strange Metal, they saw something very different.
   • At low energy (gentle pushes), they saw a faint ripple, but it was very blurry and died out quickly.
   • As they pushed harder (higher momentum), the ripple didn't travel further. Instead, it completely vanished into a fuzzy, featureless cloud.

The Verdict: The "clear, traveling wave" that Group A saw in the 1980s does not exist in the way they thought.

Why Did the Old Photos Look Different?

The paper suggests that the old scientists weren't lying, but they were likely tricked by their own tools.

The "Subtraction" Trap:
Imagine you are trying to hear a whisper in a room with a loud air conditioner.

• The Old Method: They recorded the sound, then used a computer to "subtract" the air conditioner noise. But because the air conditioner was so loud, the computer accidentally subtracted parts of the whisper too, or created a fake "ghost" sound that looked like a whisper.
• The New Method: The authors of this paper kept the raw data. They didn't try to "fix" the noise. They looked at the raw sound and realized: "Oh, there is no whisper. Just a lot of static."

The paper argues that the "clear wave" seen in the 1980s was likely an artifact—a fake signal created by the way they processed the data (specifically, how they removed the "zero-loss" peak, which is the main beam of electrons that didn't hit anything).

The Conclusion: The "Incoherent" Metal

So, what is a Strange Metal?

Based on this new, high-definition evidence, the authors conclude that Bi-2212 is an Incoherent Metal.

• Normal Metal: Like a marching band. Everyone steps in time. If you push them, a wave travels perfectly.
• Strange Metal: Like a crowd of people running in panic. If you push them, they don't form a wave. They just scatter and dissolve into chaos immediately.

The electrons in this material are so strongly linked and chaotic that they cannot sustain a collective wave. They are "damped" (killed off) almost instantly.

Why Does This Matter?

This might sound like a small technical detail, but it's huge for science.

1. Solving the Mystery: It tells us that the "Marginal Fermi Liquid" theory (a popular idea about how these metals work) needs to be adjusted. The electrons aren't just slightly messy; they are fundamentally chaotic.
2. Superconductors: Understanding how these electrons behave is the key to unlocking room-temperature superconductors (materials that conduct electricity with zero resistance at normal temperatures). If we know the electrons are this chaotic, we can stop trying to force them to behave like normal dancers and start designing materials that embrace the chaos.

In a nutshell: Scientists took a 40-year-old photo of a mysterious electron wave, realized the old photo was blurry and processed incorrectly, and took a new, crystal-clear photo. The new photo shows that the wave doesn't travel; it just dissolves into a fog. This changes how we understand the most mysterious materials in physics.

Technical Summary

1. Problem Statement

The "strange metal" phase, particularly in high-temperature cuprate superconductors like Bi$_2$Sr$_2$CaCu$_2$O$_{8+x}$ (Bi-2212), represents a fundamental unsolved problem in condensed matter physics. Unlike ordinary metals described by Fermi liquid theory, strange metals exhibit linear-in-temperature resistivity and scattering rates that violate the Mott-Ioffe-Regel limit, suggesting a breakdown of the quasiparticle picture.

Theoretical progress relies on the Marginal Fermi Liquid (MFL) phenomenology, which posits specific forms for the dynamic charge susceptibility, $\chi(q, \omega)$, particularly at finite momentum ($q$). However, experimental determination of $\chi(q, \omega)$ via Electron Energy-Loss Spectroscopy (EELS) has yielded contradictory results over the past four decades:

• Early Transmission EELS (T-EELS) studies (1989–1991): Reported a well-defined, dispersing plasmon mode consistent with Random Phase Approximation (RPA) theory, persisting up to high momenta ($q \approx 0.35$ Å$^{-1}$).
• High-resolution T-EELS studies (1995–1999): Observed no distinct plasmon, instead finding a featureless, incoherent continuum.
• Reflection EELS (M-EELS) and Infrared Optics: Generally agree on a plasmon-like feature at low $q$, but reflection techniques probe surface responses, and infrared optics are limited to $q \approx 0$.

The core issue is the lack of reproducibility and the inability to reconcile these datasets, hindering the understanding of Planckian dissipation and the nature of the strange metal.

2. Methodology

The authors employed state-of-the-art Transmission EELS (T-EELS) to resolve these discrepancies, focusing on simultaneous high energy and momentum resolution with minimal data processing.

• Instrumentation: Measurements were performed using a Nion HERMES monochromated, aberration-corrected Scanning Transmission Electron Microscope (STEM) at Oak Ridge National Laboratory.
  • Energy Resolution: Achieved $\Delta E \approx 30$ meV (post-sample $\approx 38$ meV).
  • Momentum Resolution: Achieved $\Delta q \approx 0.01$ Å$^{-1}$.
  • Geometry: Transmission geometry with the beam incident along the (0,0,1) zone axis, ensuring momentum transfer lies parallel to the CuO$_2$ planes ($q_z \approx 0$).
• Samples:
  • Bi-2212: Optimally doped single crystals ($T_c = 92$ K) were exfoliated into thin flakes. To ensure reproducibility, 10 distinct measurements were performed on 5 different flakes.
  • Control Sample: High-purity polycrystalline Aluminum was measured under identical conditions to validate the instrument's ability to detect standard Lindhard-RPA plasmons.
• Data Processing:
  • Data was analyzed in its raw, unprocessed form.
  • Crucially, the authors did not subtract the zero-loss peak (elastic line) or apply multiple-scattering corrections, which were common in earlier studies and suspected to artificially generate plasmon peaks.
  • The loss function, $\text{Im}[-1/\epsilon(q, \omega)]$, was calculated by dividing the cross-section by the Coulomb matrix element ($|q|^{-2}$) and the Bose factor.

3. Key Contributions

1. Resolution of Historical Discrepancies: The study provides the first T-EELS dataset with both high energy ($\sim 30$ meV) and high momentum ($\sim 0.01$ Å$^{-1}$) resolution, bridging the gap between earlier low-resolution T-EELS and high-resolution but low-momentum-resolution studies.
2. Reproducibility Benchmark: By repeating measurements on multiple samples and comparing against a known Fermi liquid (Aluminum), the authors establish a robust baseline for the charge response of Bi-2212.
3. Identification of Artifacts: The paper demonstrates that the "dispersing plasmon" reported in early studies (e.g., Nücker et al., 1989) is likely an artifact of zero-loss peak subtraction and poor momentum resolution, rather than a physical excitation.

4. Key Results

• Comparison with Aluminum: The control aluminum sample showed a sharp, well-defined plasmon with textbook quadratic dispersion ($E \propto q^2$), confirming the instrument was functioning correctly and capable of resolving coherent excitations.
• Bi-2212 Charge Response:
  • Low Momentum ($q < 0.15$ Å$^{-1}$): A broad, highly damped excitation is observed near 1 eV. However, unlike a Fermi liquid, its linewidth is comparable to its energy ($\Gamma \approx E$), indicating it is a marginally defined excitation.
  • High Momentum ($q > 0.15$ Å$^{-1}$): The excitation does not disperse. Instead, it rapidly damps out and evolves into a featureless incoherent continuum. No RPA-like dispersion is observed.
  • Reproducibility: The results were consistent across all 10 measurements on 5 different flakes.
• Re-evaluation of Past Data:
  • When comparing raw data, the authors' results align more closely with the "Terauchi" studies (which saw no plasmon) and reflection M-EELS data than with the early "Nücker" or "Wang" studies.
  • The authors show that subtracting the elastic line from their raw data (mimicking older processing techniques) can artificially generate a peak, explaining the false positives in earlier literature.
• Consistency with RIXS: The observed highly damped continuum is consistent with Resonant Inelastic X-ray Scattering (RIXS) data on Bi-based cuprates, which also report broad, overdamped charge excitations.

5. Significance

• Nature of the Strange Metal: The results strongly support the view that Bi-2212 is an incoherent metal where charge excitations are strongly damped and lack well-defined quasiparticle character, even at low momenta. This contradicts the RPA-like behavior expected in conventional metals.
• Validation of MFL Phenomenology: The absence of a dispersing plasmon and the presence of a broad continuum align with the Marginal Fermi Liquid hypothesis, which predicts a specific form of dynamic susceptibility that does not support long-lived collective modes at finite $q$.
• Methodological Impact: The paper highlights the critical importance of minimal data processing in EELS. It demonstrates that subtracting the zero-loss peak can create spurious features, urging the community to re-evaluate historical data and prioritize raw spectral analysis.
• Future Directions: The findings suggest that the strange metal state is fundamentally distinct from the Fermi liquid, characterized by a breakdown of the quasiparticle picture and a "Planckian" dissipation rate that prevents the formation of coherent plasmons at finite momentum.