Surface diffusion: The intermediate scattering function seen as a characteristic function of probability theory

This paper demonstrates that the intermediate scattering function measured in surface diffusion experiments can be interpreted as a characteristic function in probability theory, enabling the analytical derivation of position distribution moments and the diffusion coefficient, which is illustrated through the incoherent tunneling of hydrogen and deuterium on a Pt(111) surface.

The Gist

The Big Picture: Watching Molecules Dance on a Stage

Imagine a microscopic stage made of platinum atoms (a surface). On this stage, tiny actors—Hydrogen (H) and Deuterium (D) atoms—are trying to move from one spot to another. This movement is called diffusion.

Usually, when scientists want to watch these actors, they use a "flashlight" made of neutrons or helium atoms. In this paper, the authors focus on a special technique called Helium Spin Echo (HeSE). Think of this as a high-speed camera that doesn't just take a picture, but records a "movie" of how the atoms jitter and jump over time.

The data this camera produces is called the Intermediate Scattering Function (ISF). For decades, physicists have treated this data as a complex physics equation.

The Big Discovery:
The authors of this paper realized something brilliant: This "physics movie" is actually just a standard probability math problem in disguise.

They showed that the ISF is exactly the same thing as a Characteristic Function in probability theory.

---

The Analogy: The "Magic Dice"

To understand why this matters, let's use an analogy.

Imagine you are rolling a strange, magical die. You don't know what numbers are on the faces, but you want to know the average result and how spread out the results are (variance).

1. The Old Way (Physics): You try to calculate the physics of every single atom, the friction of the surface, and the quantum tunneling effects. It's like trying to calculate the trajectory of every grain of sand in a sandstorm to predict where the storm will go. It's messy and hard.
2. The New Way (Probability): The authors say, "Stop looking at the sand. Just look at the Characteristic Function."

In math, a Characteristic Function is like a secret decoder ring. If you have this ring (the ISF data), you can instantly unlock the answers to your questions without doing all the heavy lifting:

• The Average Position: Where is the atom likely to be?
• The Spread (Variance): How far has it wandered?
• The "Jumps": How often is it hopping?

The paper shows that the ISF is this decoder ring. By treating it as a probability tool, they can instantly calculate the Diffusion Coefficient (a number that tells us how fast the atoms are spreading out) just by doing a simple math trick (taking a derivative) on the data.

The Specific Experiment: The "Quantum Hopper"

The authors tested this theory on Hydrogen and Deuterium atoms hopping on a Platinum (Pt) surface.

• The Setting: The atoms are cold (between 80K and 250K). At these temperatures, they don't just "walk" over the bumps; they sometimes tunnel through them (a quantum effect where they pass through walls like ghosts).
• The Movement: They mostly jump to their immediate neighbors (like stepping from one tile to the next on a floor).
• The Result: Using their new "probability lens," they calculated how fast these atoms were moving.

The Surprise:
When they compared their new calculation to previous experimental reports, they found the atoms were moving three times faster than previously thought.

Why the difference?
Previous methods were like counting the steps of a person walking in a circle and assuming they walked in a straight line. The new method realized that the atoms were only moving in specific directions (two out of three possible paths). By correcting for this "directional bias," the true speed of the diffusion was revealed to be much higher.

Why This Matters (The "So What?")

1. Simplicity: It turns a complex quantum physics problem into a straightforward statistics problem. Instead of needing a supercomputer to simulate every force, you can use simple math formulas to get the answer.
2. Accuracy: It fixes errors in how we interpret experimental data. If you think you are measuring a slow walk, but you are actually measuring a fast run because you missed a directional detail, you get the wrong answer. This paper fixes that.
3. Future Applications: This method can be used for any surface diffusion problem. Whether it's hydrogen fuel cells, catalytic converters in cars, or battery technology, understanding exactly how fast atoms move is crucial.

Summary in One Sentence

The authors discovered that the complex data used to track moving atoms on a surface is actually just a standard probability math tool, allowing them to calculate the speed of these atoms much more easily and accurately than before, revealing that they move three times faster than we previously believed.

Technical Summary

1. Problem Statement

Surface diffusion is a fundamental process in surface science, typically probed using techniques like Quasi-Elastic Helium Atom Scattering (QHAS) and Helium Spin Echo (HeSE). While the Intermediate Scattering Function (ISF), denoted as $I(K, t)$, is the primary observable measured in HeSE experiments, its interpretation often relies on complex numerical simulations (e.g., Molecular Dynamics, Caldeira-Leggett formalism, or Lindblad master equations) to extract physical parameters like the diffusion coefficient.

The authors identify a gap in the direct analytical connection between the experimental ISF and the underlying probability distribution of the adsorbate's position. Specifically, they aim to demonstrate that the ISF is mathematically equivalent to the characteristic function in probability theory. This equivalence would allow for the direct, analytical extraction of statistical moments and cumulants (including the diffusion coefficient) from experimental data without relying solely on heavy numerical modeling.

2. Methodology

The paper employs a theoretical framework that bridges quantum statistical mechanics and probability theory:

• Theoretical Foundation: The authors start from the definition of the ISF in terms of the reduced density matrix diagonal elements, $\rho(R, t)$:
  $$I(K, t) = \int dR \, e^{iK \cdot R} \rho(R, t) = \langle e^{iK \cdot R(t)} \rangle$$
  They argue that this expression is identical to the definition of a characteristic function (the Fourier transform of a probability distribution function).
• Statistical Properties: By treating $I(K, t)$ as a characteristic function, the authors utilize standard probability theory properties:
  • Moments ($\langle X^N(t) \rangle$) are obtained via Taylor expansion derivatives at $K=0$.
  • Cumulants ($\langle X^N(t) \rangle_c$) are obtained from the derivatives of $\ln I(K, t)$.
• Model Application: The theory is applied to the specific case of incoherent tunneling of Hydrogen (H) and Deuterium (D) on a Pt(111) surface.
  • The system is modeled as a 1D diffusion along a symmetry direction ($[11\bar{2}]$) on a periodic lattice.
  • The dynamics are described using a Pauli Master Equation for the probability $P_n(t)$ of finding the adsorbate in the $n$-th potential well.
  • Two scenarios are analyzed:
    1. Nearest-Neighbor Hopping: Restricted to adjacent sites (Chudley-Elliott model).
    2. Long-Range Hopping: Extension to jumps involving multiple lattice sites.

3. Key Contributions

• Mathematical Identification: The primary contribution is the rigorous identification of the ISF as a characteristic function. This establishes that the ISF satisfies all properties of a characteristic function (e.g., $I(K,0)=1$, uniform continuity, existence of Fourier inversion).
• Analytical Extraction of Transport Parameters: The authors demonstrate that the diffusion coefficient ($D$) and higher-order statistical properties can be derived analytically from the ISF using moment and cumulant generating functions.
  • For the nearest-neighbor case, the ISF is derived as $I(K, t) = e^{-\Gamma(1-\cos K_{||})t}$.
  • The second-order moment/cumulant yields the diffusion coefficient: $D = \Gamma a^2 \cos^2 \beta$, where $\Gamma$ is the total jump rate, $a$ is the lattice constant, and $\beta$ is the angle between the observation direction and the diffusion axis.
• Generalization to Long-Range Jumps: The framework is extended to include jumps beyond nearest neighbors. The ISF is expressed as a product of forward and backward diffusion terms, allowing for the calculation of cumulants when jump probabilities decay exponentially with distance.

4. Results

• Re-evaluation of Experimental Data: The authors applied their analytical formula to experimental data for H and D on Pt(111) (previously analyzed in Ref. [19]).
  • Correction of Diffusion Coefficients: The study revealed that previous experimental analyses underestimated the diffusion coefficient by a factor of 3. This discrepancy arose because the original analysis did not fully account for the geometry of the active diffusion directions relative to the momentum transfer vector $K$.
  • New Calculation: Using the corrected formula $D = \Gamma a^2 \cos^2 \beta$ and adjusting for the fact that only two equivalent symmetry directions are active (instead of three), the authors recalculated the diffusion coefficients.
• Temperature Dependence: Figure 1 in the paper plots the corrected diffusion coefficients against the inverse temperature ($1/T$). The results show that the diffusion coefficients are three times larger than those reported in the original experimental work, providing a more accurate description of the quantum tunneling and thermal activation regimes (crossover temperatures estimated at 66 K for H and 63 K for D).
• Cumulant Behavior: The analysis confirms that for this diffusive process, the first moment and first cumulant are zero (no net drift), and the second moment and second cumulant coincide ($\langle X^2 \rangle = \langle X^2 \rangle_c = Dt$), which is characteristic of normal diffusion.

5. Significance

• Simplification of Analysis: This work provides a powerful, simplified analytical tool for interpreting surface diffusion data. Researchers can now extract the full probability distribution characteristics (moments and cumulants) directly from the ISF without needing complex numerical inversions or full density matrix simulations.
• Accuracy in Physical Parameters: By correcting the geometric factors in the diffusion coefficient calculation, the paper resolves a significant quantitative discrepancy in the literature regarding H/D diffusion on Pt(111). This leads to a more precise understanding of quantum tunneling rates and friction effects at surfaces.
• Theoretical Unification: The paper successfully unifies the concepts of scattering theory in condensed matter physics with the mathematical formalism of probability theory, offering a new perspective on how to interpret dynamic structure factors and intermediate scattering functions.

In summary, the paper establishes that the Intermediate Scattering Function is not just a response function but a fundamental characteristic function of the adsorbate's position probability distribution, enabling a direct, analytical route to determining diffusion coefficients and higher-order statistical properties from experimental HeSE data.