Cumulative Fidelity of LMT Clock Atom Interferometers in the Presence of Laser Noise

This paper resolves concerns regarding laser frequency noise in large momentum transfer (LMT) clock atom interferometers by demonstrating that population error scales linearly rather than quadratically with the number of pulses and that parasitic path contributions are negligible, thereby confirming the viability of high-fidelity LMT sensors.

The Gist

The Big Picture: The Ultimate Quantum Ruler

Imagine you are trying to measure the distance between two points with a ruler so precise it can detect the movement of a single atom. This is what atom interferometers do. They act like ultra-sensitive scales or rulers for gravity, dark matter, and even gravitational waves.

To make these rulers more sensitive, scientists use a trick called Large Momentum Transfer (LMT). Think of this like giving the atom a series of "pushes" to make it fly faster and further apart from its twin. The more pushes (or "kicks") you give, the more sensitive the ruler becomes.

The goal of this paper is to answer a scary question: If we give the atom thousands of these pushes, will the noise in our laser equipment ruin the measurement?

The Problem: The "Bad Copy" Fear

In the past, scientists were worried that if they tried to push an atom 10,000 times (an LMT of $10^4$), the tiny imperfections in the laser would add up like a snowball rolling down a hill.

• The Old Fear: Imagine you are trying to copy a document 10,000 times. If the copier makes a tiny smudge on the first copy, and you use that smudged copy to make the second, the smudge gets worse. By the 10,000th copy, the document is unreadable.
• The Specific Fear: Scientists thought that laser noise (tiny jitters in the laser's frequency) would act like that smudge. They feared that if you pushed the atom $n$ times, the error would grow as $n^2$ (a snowball effect). This would mean we could never build these super-sensitive sensors because our lasers aren't perfect enough.

The Discovery: The "One-Way Street" Advantage

The authors of this paper (Jiang, Rudolph, and Hogan) realized that the "snowball" fear was based on a misunderstanding of how these specific atom clocks work.

They found that in a Clock Atom Interferometer, the atom isn't just being pushed back and forth between two states (like a light switch being flipped on and off). Instead, every push moves the atom to a new, unique state (like moving to a new floor in a skyscraper).

Here is the analogy:

• The Old Way (Two-Level System): Imagine a person walking back and forth between two rooms (Room A and Room B). If they trip once, they are still in the same two rooms, and every time they walk, they might trip again in the same spot. The errors pile up quickly.
• The New Way (LMT Clock Interferometer): Imagine the person is walking up a spiral staircase. Every time they take a step (a laser pulse), they move to a brand new step they have never been on before.
  • If they trip on step 5, they fall a little bit.
  • But when they take the next step to step 6, they are on a completely new level. The "trip" from step 5 doesn't make them trip again on step 6. The error stays behind on step 5.

The Result: Instead of errors piling up like a snowball ($n^2$), the errors just add up one by one, like counting coins ($n$). This is a massive difference. It means the system is much more forgiving of imperfections.

The "Ghost Paths" (Parasitic Paths)

The paper also looked at "ghost paths." When a laser pulse isn't perfect, a tiny fraction of the atom might not move at all, or move the wrong way. This creates a "ghost" version of the atom that wanders off.

• The Fear: Maybe these ghosts will wander back later and crash into the main atom, ruining the measurement.
• The Reality: The authors did the math and found that these ghosts are like people who took a wrong turn in a massive city. Because the main atom is moving so fast and changing directions so often, the ghosts get left behind in different neighborhoods. They never catch up to the main atom at the finish line. Even if they do, there are so few of them that they don't matter.

The Conclusion: We Are Good to Go

The paper concludes that laser noise is not a deal-breaker.

Even with current laser technology (which has some jitter), we can build these super-sensitive sensors with LMT factors of 10,000 or more. The errors grow slowly (linearly), not explosively (quadratically).

In simple terms:
If you want to build a quantum sensor that can detect the faintest ripples in spacetime or the presence of dark matter, you don't need a "perfect" laser. You just need a "good enough" one. The design of the experiment naturally protects itself from the noise, allowing us to push the boundaries of physics further than we thought possible.

Summary of Key Takeaways

1. The Goal: Build ultra-sensitive sensors by giving atoms thousands of "kicks" (LMT).
2. The Worry: Would laser noise ruin the measurement? (Old theory said yes, errors grow fast).
3. The Fix: Because the atom moves to a new state every time, errors don't pile up. They just add up slowly.
4. The Verdict: Laser noise is manageable. We can build these next-generation sensors with today's technology.

Technical Summary

1. Problem Statement

Large Momentum Transfer (LMT) atom interferometry is a critical technology for next-generation precision measurements, including gravitational wave detection, dark matter searches, and tests of fundamental physics. To achieve the required sensitivity, these sensors aim for LMT enhancement factors beyond $10^4$ (momentum separations of thousands of photon recoils).

A primary concern regarding the viability of such high-order LMT systems is the impact of laser frequency noise.

• The Conflict: While single-photon clock transitions offer superior common-mode rejection of laser noise, recent literature (specifically Chiarotti et al., 2022) claimed that in LMT clock interferometers, population loss due to laser frequency noise scales quadratically with the number of pulses ($n^2$). This would impose prohibitive stability requirements on lasers, potentially rendering $10^4$ LMT orders impossible with current technology.
• The Gap: The conflicting claim assumes the atom is probed $n$ times from the same direction within a two-level Hilbert space. However, LMT clock interferometers utilize alternating pulse directions and a significantly expanded Hilbert space involving external momentum states. The validity of the $n^2$ scaling law for this specific architecture was unverified.

2. Methodology

The authors employed a combination of analytical derivation, combinatorial analysis, and numerical simulation to resolve the discrepancy.

• System Model: They modeled a narrowband LMT clock atom interferometer where atoms are interrogated by a sequence of $\pi$ pulses from alternating directions. The system's Hilbert space includes both internal atomic levels (ground $|g\rangle$, excited $|e\rangle$) and external momentum states ($p\hbar k$).
• Fidelity Analysis:
  • They derived the cumulative fidelity of sequential state inversions.
  • They distinguished between main path population loss (atoms failing to transfer momentum) and parasitic path interference (atoms taking incorrect momentum paths that eventually recombine with the main signal).
• Parasitic Path Counting: Using a combinatorial approach (Appendix A), they analyzed the number of parasitic paths generated by pulse errors. They applied constraints based on:
  1. Momentum: Only paths ending in the final momentum states ($0$ or $1\hbar k$) contribute to interference.
  2. Position: Only paths overlapping spatially with the main wavepacket at the detection region contribute.
• Noise Transfer Function: They derived an analytic expression for the frequency noise transfer function ($H_n(f)$) relating laser frequency noise power spectral density ($S_\nu(f)$) to population loss. They integrated this over realistic laser noise spectra to calculate sequence fidelity.

3. Key Contributions

A. Correction of Scaling Law ($n$ vs. $n^2$)

The paper's most significant theoretical contribution is proving that for LMT clock interferometers with alternating pulse directions, the population error scales linearly ($n$), not quadratically ($n^2$).

• Mechanism: In an alternating-direction sequence, a pulse error leaves a fraction of the wavefunction in a state that is Doppler suppressed from interacting with the next pulse (which is frequency-tuned to a different momentum state). Consequently, the error does not coherently accumulate in the same way it does in a two-level system probed from a single direction.
• Result: The fidelity loss is proportional to $n \delta P$ (where $\delta P$ is the per-pulse loss), rather than $n^2 \delta P$.

B. Analysis of Parasitic Paths

The authors rigorously bounded the contribution of "parasitic paths" (untransferred atoms that drift and re-interfere).

• They demonstrated that while the number of possible error combinations grows with $n$, the amplitude of these paths is suppressed by the error probability ($\alpha^m$).
• Crucially, they showed that the number of parasitic paths satisfying both momentum and position constraints at the output is bounded by a constant independent of $n$ (specifically scaling as $O(1)$ for large $n$).
• Conclusion: The interference background from parasitic paths is negligible compared to the main path amplitude loss, even for very high $n$.

C. Analytic Transfer Function

They derived the single-pulse and $n$-pulse frequency noise transfer functions ($H_1(f)$ and $H_n(f)$).

• $H_n(f) = n H_1(f)$, confirming the linear accumulation of noise-induced loss.
• This allows for precise calculation of fidelity given a specific laser noise spectrum.

4. Results

• Fidelity Scaling: Simulations and analytical bounds confirm that for a laser with an RMS frequency noise of $\Delta\nu = 10$ Hz and a Rabi frequency of $\Omega = 2\pi \times 1$ kHz, the sequence fidelity remains high even for $n \approx 10^4$.
• Parasitic Error Bound: For densely packed pulses ($\delta t = \pi/\Omega$) and a detection radius of 1 mm, the wavefunction error induced by parasitic paths ($\varepsilon$) saturates at a constant value ($\approx 5 \times 10^{-3}$) as $n$ increases, causing a contrast degradation of at most 1%.
• Optimal LMT Order: The study identifies an optimal LMT order ($n$) that maximizes the sensitivity enhancement ($n \times \text{Contrast}$).
  • For $\Delta\nu = 10$ Hz, the peak enhancement occurs at $n \approx 8.8 \times 10^3$, achieving an LMT enhancement factor of $3.3 \times 10^3$.
  • This is consistent with the requirements for the MAGIS-100 detector and future long-baseline interferometers.
• Comparison to Previous Claims: The results directly refute the $n^2$ scaling claim, showing that the previous analysis incorrectly applied a two-level model to a multi-level momentum system.

5. Significance

• Viability of Next-Gen Sensors: The paper establishes that laser frequency noise is not a practical limitation for developing high-fidelity LMT clock atom interferometers with enhancement factors of $10^3 - 10^4$.
• Technology Roadmap: It validates the feasibility of using current laser stabilization technologies (achieving $\sim 10$ Hz RMS noise) to support ambitious experiments like MAGIS-100 and space-based gravitational wave detectors (e.g., AEDGE, SAGE).
• Design Guidance: The work provides a robust mathematical framework for designing pulse sequences and estimating error budgets, confirming that the primary limitation for LMT scaling is spontaneous emission (in narrowband transitions) rather than laser noise accumulation.
• Broader Impact: The findings support the continued development of quantum sensors for fundamental physics tests, including the search for ultralight dark matter and the detection of mid-band gravitational waves.

In summary, Jiang et al. resolve a critical theoretical bottleneck by demonstrating that the unique architecture of LMT clock interferometers (alternating directions and expanded momentum space) inherently suppresses the accumulation of laser noise errors, enabling the realization of ultra-precise quantum sensors.