Asymmetry and coverage dependence in two-pulse correlation measurements of CO photodesorption from Pd(111): Insights from theory

This study utilizes advanced molecular dynamics simulations incorporating temperature-dependent electronic properties to successfully reproduce the pulse-order asymmetry and significantly improve the quantitative accuracy of CO photodesorption probabilities on Pd(111) observed in two-pulse correlation experiments.

The Gist

The Big Picture: A High-Speed Dance on a Metal Stage

Imagine a ballroom where the floor is made of Palladium metal (a shiny, silver-colored metal) and the dancers are Carbon Monoxide (CO) molecules. Normally, these dancers are glued to the floor, sitting still.

Scientists want to make them jump off the floor (a process called photodesorption). To do this, they use a super-fast, intense laser pulse—like a sudden, powerful flash of light that hits the floor.

The big question the scientists are asking is: What actually makes the dancers jump?

1. Does the laser hit the electrons in the metal, which then kick the dancers? (The "Electron" theory).
2. Or does the laser heat up the metal floor itself, causing it to vibrate and shake the dancers off? (The "Phonon" or "Vibration" theory).

To figure this out, the experimentalists used a trick called Two-Pulse Correlation (2PC). Instead of one flash, they used two flashes of light, one strong and one weak, arriving at slightly different times.

The Mystery: The "Order Matters" Puzzle

In the real experiments, the scientists noticed something weird and asymmetrical:

• Scenario A: If the Weak pulse hits first, followed quickly by the Strong pulse, the dancers jump off easily.
• Scenario B: If the Strong pulse hits first, followed by the Weak pulse, the dancers are much less likely to jump.

It's as if the dancers have a memory. They react differently depending on which "kick" they feel first. This difference was especially huge when there were fewer dancers on the floor (low coverage).

The experimentalists were confused. Why does the order matter? And why does it matter so much when the floor is less crowded?

The Simulation: Building a Digital Twin

The authors of this paper are the "theorists." They built a super-advanced computer simulation (a "digital twin") of this ballroom to see if they could recreate the mystery.

They used a method called Molecular Dynamics, which is like a high-speed movie where they track every single atom. To make the movie realistic, they had to model how the laser energy moves through the system.

The Old Map vs. The New Map (The 2TM)

To describe how the laser energy heats up the metal, they used a model called the Two-Temperature Model (2TM). Think of this as a map of how heat flows.

• The Old Map (2TM-1): Used standard, "textbook" numbers for how much heat the electrons can hold and how fast they talk to the metal floor. These numbers were based on experiments done at normal temperatures.
• The New Map (2TM-2): The authors realized that during a laser blast, the electrons get extremely hot (thousands of degrees), far hotter than normal. They used new, cutting-edge calculations to update the map with how heat behaves at these extreme temperatures.

The Result: When they used the New Map, their simulation finally matched the real experiment! They could reproduce the "Order Matters" asymmetry.

• The Analogy: Imagine trying to predict how a car accelerates. If you use a map designed for a gentle drive, you get it wrong when the driver floors the gas pedal. The "New Map" accounted for the extreme "flooring of the gas pedal" (the intense laser), showing that the electrons get hotter and transfer energy differently than we thought.

The Missing Piece: The "Hot Friction" Problem

Even with the New Map, there was still a problem. In the real experiment, right when the two pulses hit at the exact same time (zero delay), there was a huge spike in dancers jumping off. The simulation, however, showed a dip (fewer dancers jumping).

Why?
The scientists realized they were treating the "friction" (the resistance the dancers feel when moving through the electron sea) as if the electrons were cold.

• The Analogy: Imagine running through a crowd. If the crowd is calm (cold electrons), it's hard to push through. But if the crowd is panicking and moving wildly (hot electrons), it's actually easier to slip through, or the interaction changes.
• The simulation assumed the crowd stayed calm. But in reality, at the moment of the laser blast, the electrons are "panicking" (very hot).

The authors updated their simulation to include "Hot Friction" (friction that changes based on how hot the electrons are).

• The Result: This boosted the number of dancers jumping off at zero delay by ten times! It got much closer to the real experiment, though a small gap remained.

The Conclusion: Why This Matters

This paper teaches us three main things:

1. Context is King: You can't use "normal" physics rules to describe extreme events. When you hit metal with a super-fast laser, the properties of the metal (like how much heat it holds) change drastically. You need a "New Map" (ab-initio calculations) to understand it.
2. The Order of Events: The asymmetry (why the order of pulses matters) is caused by how the electrons and the metal floor exchange energy over time. The "New Map" captures this timing perfectly.
3. The "Hot" Connection: At the very instant of the laser hit, the connection between the metal and the molecules is much stronger because everything is so hot. Ignoring this "hot friction" leads to big errors.

In a nutshell: The scientists built a better digital simulation of a laser hitting metal. By updating their "heat map" and realizing that "friction" changes when things get super hot, they finally solved the mystery of why the order of laser pulses changes how molecules jump off a surface. It's a reminder that in the world of ultra-fast physics, things behave very differently than they do in our everyday, slow-motion world.

Technical Summary

1. Problem Statement

The study addresses a discrepancy between experimental observations and theoretical simulations regarding the photodesorption of Carbon Monoxide (CO) from a Palladium (111) surface induced by intense femtosecond laser pulses.

• Experimental Context: Previous two-pulse correlation (2PC) experiments by Hong et al. revealed two puzzling phenomena:
  1. Asymmetry: The desorption probability ($P_{des}$) depends on the order of arrival of two laser pulses (strong vs. weak). Specifically, the half-width-at-half-maximum (HWHM) of the desorption curve differs significantly between positive delays (weak pulse first) and negative delays (strong pulse first).
  2. Coverage Dependence: This asymmetry is much more pronounced at low CO coverage (0.24 ML) compared to saturation coverage (0.75 ML).
  3. Zero-Delay Discrepancy: Experiments show a sharp peak in desorption yield at zero time delay (sub-picosecond regime), whereas standard theoretical models typically predict a dip or underestimation at this point.
• Theoretical Challenge: Existing models, often relying on the Two-Temperature Model (2TM) with constant or low-temperature parameters, fail to reproduce the magnitude of the asymmetry and the zero-delay peak. The authors aim to disentangle whether the process is governed by excited electrons or phonons and identify the missing physical mechanisms in current simulations.

2. Methodology

The authors employed Langevin Molecular Dynamics (MD) simulations coupled with advanced electronic structure modeling.

• Potential Energy Surface (PES): They utilized a multicoverage Neural Network (NN) potential energy surface trained on ab initio data (DFT with van der Waals corrections). This PES is valid for CO coverages ranging from 0.33 ML to 0.75 ML on Pd(111).
• Dynamics Framework: The system was modeled using the $(T_e, T_l)$-MDEF approach:
  • Langevin Equations: The motion of atoms (adsorbate and surface) is governed by Langevin dynamics, incorporating adiabatic forces from the NN-PES, electronic friction, and stochastic forces.
  • Two-Temperature Model (2TM): The laser-excited electronic ($T_e$) and lattice/phononic ($T_l$) temperatures are calculated via the 2TM equations to provide the time-dependent thermal environment.
• Key Improvements over State-of-the-Art:
  1. Temperature-Dependent 2TM Parameters: Instead of using constant values or low-temperature fits, the authors incorporated $T_e$-dependent electronic heat capacity ($C_e$) and electron-phonon coupling constant ($G$) derived from ab initio calculations (DFT) valid up to 20,000 K.
  2. Temperature-Dependent Friction: They calculated the electronic friction coefficient ($\eta$) for CO atoms as a function of $T_e$. This was done using the Local Density Friction Approximation (LDFA) and mode-resolved friction tensors derived from electron-phonon coupling matrix elements (using Quantum ESPRESSO and EPW codes).
• Simulation Setup: Simulations were run for two coverages (0.33 ML and 0.75 ML) with time delays ($\delta t$) ranging from -20 to +20 ps, matching the experimental laser fluences and pulse profiles.

3. Key Contributions

1. Refined 2TM Parameterization: The paper demonstrates that using ab initio-derived, temperature-dependent $C_e(T_e)$ and $G(T_e)$ is crucial for accurately modeling the energy balance between electrons and phonons under extreme laser excitation.
2. Inclusion of $T_e$-Dependent Friction: The authors explicitly modeled the increase in adsorbate-electron coupling strength at high electronic temperatures ($T_e \ge 5000$ K), a factor previously neglected in standard simulations.
3. Mechanistic Insight into Asymmetry: The study provides a clear physical explanation for the experimental asymmetry in 2PC measurements, linking it to the non-linear response of the electronic system to pulse ordering and coverage.

4. Key Results

A. Reproducing Asymmetry and Coverage Dependence

• Standard Model (2TM-1): Using constant or low-temperature fitted parameters failed to reproduce the experimental asymmetry, particularly the sharp drop in desorption probability at negative delays for low coverage.
• Improved Model (2TM-2): Incorporating $T_e$-dependent $C_e$ and $G$ significantly improved the results.
  • Mechanism: At negative delays (strong pulse first), the initial strong pulse creates a high $T_e$, but the subsequent weak pulse arrives when the system has partially cooled. At positive delays, the weak pulse arrives first, followed by the strong pulse which acts on a system that has not yet fully equilibrated.
  • Coverage Effect: The asymmetry is more pronounced at low coverage (0.33 ML) because the desorption probability is more sensitive to changes in $T_e$ at lower coverages. The improved 2TM-2 model successfully reproduced the larger asymmetry observed at low coverage compared to saturation coverage.

B. Addressing the Zero-Delay Discrepancy

• The Problem: Standard simulations consistently underestimate $P_{des}$ at zero delay (sub-picosecond regime).
• Role of Friction: When the authors included the $T_e$-dependent friction coefficient (which increases by factors of 1.4 to 1.8 for O and C atoms at $T_e \approx 6000$ K), the predicted desorption probability at zero delay increased by an order of magnitude.
• Outcome: While this reduced the discrepancy with experimental values significantly, the simulations still predicted a "dip" at zero delay, whereas experiments show a "peak."
• Conclusion on Zero Delay: The underestimation of the adsorbate-electron coupling at high temperatures is a major factor, but not the sole cause. The remaining discrepancy suggests that the standard 2TM assumption (that electrons thermalize instantly into a Fermi-Dirac distribution) may break down at sub-picosecond timescales, where non-thermal electron distributions might play a role.

5. Significance and Conclusion

• Importance of Electronic Structure: The study highlights that describing extreme non-equilibrium conditions (high $T_e$) requires accounting for the temperature dependence of electronic properties (heat capacity, coupling constants, and friction). Ignoring these leads to qualitative and quantitative errors.
• Pd(111) Specificity: The sensitivity of the chemical potential in Palladium to temperature changes makes it a critical case for testing these models.
• Future Directions: The inability to fully capture the sub-picosecond peak suggests the need for methods beyond the standard 2TM, such as Time-Dependent DFT (TDDFT), to capture non-thermal electron distributions and their direct coupling to adsorbates.
• Overall Impact: This work bridges the gap between complex surface science experiments and theoretical modeling, providing a more robust framework for understanding ultrafast laser-induced surface reactions and the interplay between electronic and phononic excitations.