Direct Measurement of the $5s5p\,{}^1P_1 \to 5s4d\,{}^1D_2$ Decay Rate in Strontium

This paper reports the first direct experimental determination of the branching ratio and decay rate for specific transitions in neutral strontium, revealing values that significantly deviate from established theoretical predictions and providing a crucial benchmark for modeling loss processes in laser cooling and single-atom fluorescence detection.

The Gist

Imagine you are running a busy, high-tech factory where tiny, invisible workers (atoms) are being assembled and sorted. This factory is a Magneto-Optical Trap (MOT), and the workers are Strontium atoms.

The goal of this factory is to keep these atoms cool, organized, and ready for super-precise jobs like building the world's most accurate clocks. To do this, the factory uses a specific "conveyor belt" system made of laser light.

The Problem: The Leaky Conveyor Belt

In a perfect world, the atoms would stay on the main conveyor belt forever, doing their job. But in reality, the system has a tiny, annoying leak.

1. The Main Loop: The atoms are usually zipping around on a fast track called the $5s5p \ ^1P_1$ state. This is the "cooling" track.
2. The Leak: Occasionally, an atom slips off this fast track and falls into a side room called the $5s4d \ ^1D_2$ state.
3. The Fork in the Road: Once in this side room, the atom has to make a choice. It can either:
   • Option A: Take a quick exit back to the main factory floor (the ground state) and get back to work.
   • Option B: Fall into a deep, dark dungeon called the $5s5p \ ^3P_2$ state. This is a "metastable" trap. Once an atom falls here, it stays there for a very long time (about 1,000 seconds) and is effectively lost from the cooling process. The factory loses a worker.

For over 40 years, scientists knew this leak existed, but they didn't know exactly how big the hole was or how often the atoms chose the dungeon. They had to rely on computer simulations (theories) to guess the numbers.

The Old Guesses vs. Reality

For decades, the scientific community relied on a famous computer calculation from 1985. It predicted:

• The Branching Ratio: That about 32% of the atoms falling into the side room would choose the dungeon (Option B).
• The Leak Rate: That atoms were slipping off the main track at a certain speed.

However, newer, more powerful computers (from 2018) started suggesting the leak might be much bigger—perhaps 2.5 times larger than the old guess. This created a conflict: Is the factory losing 1 worker every 50,000 tries, or 1 worker every 20,000 tries?

This matters a lot. If you are trying to catch a single atom in a tiny "optical tweezer" (a laser trap) to do quantum computing, and the leak is bigger than you think, your atom might disappear before you can use it.

The Experiment: Catching the Leak in Action

The researchers in this paper (from the University of Tokyo) decided to stop guessing and measure it directly. They didn't want to rely on computer theories; they wanted to see the atoms with their own eyes.

Here is how they did it, using a clever trick:

1. The Setup: They filled their factory with millions of Strontium atoms and kept them running on the main cooling track.
2. The "Pump": They turned on a special laser (448 nm) that acts like a vacuum cleaner. This laser grabs any atom that has fallen into the "side room" ($^1D_2$) and instantly zaps it back to the main floor.
   • Result: The side room is empty. The factory is running at 100% efficiency.
3. The Switch: Suddenly, they turned off the vacuum cleaner laser.
4. The Observation: Now, the side room starts to fill up again. Atoms fall off the main track, get stuck in the side room, and then have to decide: "Do I go back to work, or do I fall into the dungeon?"
   • If they fall into the dungeon, they are gone forever (from the cooling cycle).
   • The researchers watched the total number of atoms in the factory drop over time.

By analyzing exactly how fast the atom count dropped and how the population shifted between the rooms, they could calculate the exact probabilities.

The Big Discovery

The results were surprising and overturned the 40-year-old theory:

1. The Branching Ratio (The Fork in the Road):

   • Old Theory: 32% of atoms go to the dungeon.
   • New Reality: Only 17.7% go to the dungeon.
   • Analogy: Imagine a fork in the road where people usually think 1 out of 3 people take the wrong path. The researchers found that actually, only about 1 out of 6 people take the wrong path. The "leak" is smaller than we thought!

2. The Decay Rate (How fast they slip):

   • They measured exactly how fast atoms slip off the main track. Their result matched an old experiment from 1986 but was significantly lower than the new, fancy computer predictions from 2018.

Why Does This Matter?

Think of this like fixing a blueprint for a skyscraper.

• If you thought the building had a 30% chance of collapsing (the old theory), you would build it with massive, expensive reinforcements.
• If you thought it had a 90% chance of collapsing (the new 2018 theory), you might panic and redesign the whole thing.
• This paper says: "Actually, the building is much safer than the new scary theories suggested, but slightly different from the old blueprints."

The Impact:

• Better Clocks: Strontium clocks are the most accurate timekeepers on Earth. Knowing the exact "leak rate" helps engineers build even better clocks.
• Quantum Computing: When scientists try to catch single atoms in laser traps (optical tweezers) to build quantum computers, they need to know exactly how long an atom will survive. This new data gives them a more accurate "survival guide."
• Science Check: It proves that even our best modern computer models can get things wrong. Sometimes, you just have to go out and measure the real world.

In short, these scientists opened the factory door, watched the workers for a while, and realized the "leak" in the system was smaller and different than anyone had guessed for the last 40 years.

Technical Summary

1. Problem Statement

Strontium (Sr) atoms are pivotal for precision metrology, including optical lattice clocks and quantum simulation. In standard magneto-optical traps (MOTs) using the 461 nm cooling transition ($5s^2$ $^1S_0$ – $5s5p$ $^1P_1$), a small fraction of atoms decays into the metastable $5s4d$ $^1D_2$ state. From this state, atoms can decay via spin-forbidden transitions to the long-lived $5s5p$ $^3P_2$ state (lifetime $\sim$100 s), causing them to be lost from the cooling cycle.

The effective loss rate depends on two critical parameters:

1. The decay rate of the $5s5p$ $^1P_1$ → $5s4d$ $^1D_2$ transition ($A_{^1P_1 \to ^1D_2}$).
2. The branching ratio of the $5s4d$ $^1D_2$ → $5s5p$ $^3P_2$ transition ($B_{^1D_2 \to ^3P_2}$).

The Gap: For over four decades, these values lacked independent experimental determination.

• Previous experimental values (e.g., Hunter et al., 1986) relied on theoretical inputs for other transition rates, making them not "purely" experimental.
• Theoretical values (e.g., Bauschlicher et al., 1985; Cooper et al., 2018) have diverged significantly. Specifically, recent theoretical predictions suggest a branching ratio of $\sim 1/3$ and a decay rate of $\sim 9.25 \times 10^3$ s$^{-1}$, which is roughly 2.5 times higher than older experimental estimates. This discrepancy creates uncertainty in modeling atom loss in laser cooling and single-atom fluorescence detection in optical tweezers.

2. Methodology

The authors performed a direct experimental measurement using an $^{88}$Sr MOT, observing the transient population dynamics of the $^1D_2$ state.

Experimental Setup:

• Trapping: Sr atoms were trapped in a MOT using the 461 nm cooling transition.
• Repumping: 481 nm and 483 nm lasers were used to repump atoms from the $^3P_2$ and $^3P_0$ states back to the cooling cycle.
• Optical Pumping: A 448 nm laser, resonant with the $5s4d$ $^1D_2$ – $5s8p$ $^1P_1$ transition, was used to optically pump atoms out of the metastable $^1D_2$ state back to the ground state.
  • The upper state of the 448 nm transition has a very short lifetime ($\sim$30 ns) compared to the $^1D_2$ state ($\sim$400 $\mu$s), ensuring efficient emptying of the $^1D_2$ population while the laser is on.

Measurement Protocol:

1. Steady State: The system reaches equilibrium with all lasers (461, 481, 483, and 448 nm) on.
2. Transient Response: The 448 nm pumping light is switched off. Atoms begin to accumulate in the $^1D_2$ state via the $^1P_1$ decay.
3. Observation: The total number of trapped atoms ($N(t)$) is monitored via 461 nm fluorescence. As atoms accumulate in the "dark" $^1D_2$ state (and subsequently leak to the long-lived $^3P_2$ state), the fluorescence signal decays exponentially.

Data Analysis:

• The authors derived rate equations describing the population dynamics of the $^1D_2$ and $^3P_1$ states.
• By fitting the exponential decay of the trapped atom number $N(t)$ after switching off the 448 nm light, they extracted the total decay rate of the $^1D_2$ state ($\gamma_{^1D_2}$) and the parameter $\alpha$, which relates to the product of the excitation fraction ($f$) and the decay rate ($A_{^1P_1 \to ^1D_2}$).
• A secondary measurement involved switching off the 481 nm repumper to isolate the loss rate specifically due to the $^1D_2 \to ^3P_2$ pathway.
• Combining these measurements allowed the independent extraction of $A_{^1P_1 \to ^1D_2}$ and $B_{^1D_2 \to ^3P_2}$ without relying on theoretical branching ratios from previous literature.

3. Key Contributions

• First Direct Measurement: This is the first experiment to determine the $5s5p$ $^1P_1$ → $5s4d$ $^1D_2$ decay rate and the $5s4d$ $^1D_2$ → $5s5p$ $^3P_2$ branching ratio independently of theoretical calculations.
• Resolution of Discrepancy: The study directly addresses the conflict between older experimental estimates and recent high-precision theoretical predictions (e.g., Cooper et al., 2018).
• Methodological Rigor: The approach utilizes transient population dynamics in a MOT to decouple the decay rate and branching ratio, providing a robust benchmark for atomic physics models.

4. Key Results

The authors report the following values:

• Branching Ratio ($B_{^1D_2 \to ^3P_2}$):

  • Measured: $0.177(4)$
  • Comparison: Significantly lower than the widely cited theoretical value of $0.322$ (Bauschlicher et al., 1985).
  • Implication: The probability of an atom in the $^1D_2$ state decaying to the lossy $^3P_2$ state is roughly half of what was previously assumed in theory.

• Decay Rate ($A_{^1P_1 \to ^1D_2}$):

  • Measured: $5.3(5) \times 10^3$ s$^{-1}$
  • Comparison:
    • Consistent with the indirect experimental value reported by Hunter et al. ($3.9(1.5) \times 10^3$ s$^{-1}$).
    • Significantly lower than recent theoretical predictions (e.g., Cooper et al.: $9.25(40) \times 10^3$ s$^{-1}$; Werij et al.: $1.7 \times 10^4$ s$^{-1}$).

• Derived Branching Ratio for Cooling Loss:

  • The branching ratio for the $5s5p$ $^1P_1$ → $5s4d$ $^1D_2$ transition is determined to be $1 : 36,000 \pm 3,000$.
  • This lies between Hunter's estimate ($1:50,000$) and Cooper's estimate ($1:20,000$).

• Total Decay Rate of $^1D_2$:

  • Measured as $\gamma_{^1D_2} = 2.37(1) \times 10^3$ s$^{-1}$, consistent with previous experimental results.

5. Significance

• Benchmark for Theory: The results challenge the validity of recent theoretical frameworks used to calculate spin-forbidden transition rates in alkaline-earth atoms. The discrepancy suggests that current many-body calculations may overestimate the mixing between singlet and triplet states.
• Optimization of Laser Cooling: Accurate knowledge of the loss rate is essential for optimizing the efficiency of Sr MOTs and understanding the limits of atom number in optical lattice clocks.
• Single-Atom Detection: The findings are critical for evaluating the survival probability of single atoms trapped in optical tweezers. Since the loss rate determines how long an atom remains visible before falling into a dark state, these new values will refine the parameters for quantum information processing and quantum simulation experiments using Sr.
• Reevaluation of Models: The paper calls for a re-evaluation of theoretical models for the $5s5p$ $^1P_1$ → $5s4d$ $^1D_2$ and $5s4d$ $^1D_2$ → $5s5p$ $^3P_{1,2}$ transitions to reconcile theory with the new experimental reality.