The Electron-Recoil Response Tail and the LZ High-Energy Event Background

summary

The gist

We reconstruct the continuum electron-recoil contribution to the LZ background expectation from the collaboration’s released simulation configuration The published characterisation of the LZ…

In short

The study reconstructs electron-recoil background expectations from a simulation, finding that published response characterizations do not determine the crucial probability of interest. Different distribution families matching model moments can differ wildly in their extreme tails by many orders of magnitude. This suggests that only precise measurements at specific high quanta levels can constrain the shape of the response distribution.

Key concepts

Central Finding
The core result is that the published description of the electron-recoil response does not fix which probability matters for discrimination. The authors created several distribution families matching key model statistics, yet these families diverge significantly in their extreme tail probabilities, proving the published characterization is insufficient.
Constructive Counterexamples
The constructed distribution families are not competing physical models of the liquid-xenon response. Instead, they serve as counterexamples showing that distributions can perfectly match the executable model's mean and variance but yield vastly different probabilities in the far tail, highlighting a limitation in relying solely on central moments.
Tail Constraint
The paper argues that constraining the *shape* of the response distribution, rather than just its width, requires measuring the low-side tail of the electron-recoil distribution at a high quanta level (Nq ≈ 5 x 10^3) with high precision (10^-5). This measurement is necessary to narrow down which theoretical shapes are physically admissible.
Implementation Divergence
The simulation used for the analysis differs from the printed specification. The implementation value must be chosen to match the published expectation, leading to a specific ratio between printed and implemented parameters, confirming that discrepancies in modeling assumptions affect the final result.

Terminology used across episodes

This episode discusses

The paper

The Electron-Recoil Response Tail and the LZ High-Energy Event Background · Read on arXiv

Triparno Bandyopadhyay, Rajarajeswari Ramasamy, Nishitha TP

Nayanta University · SRM Institute of Science and Technology

Transcript

Introduction to the show: ident: High Energy Physics: Experiment. The day in high-energy experiment: detectors, searches and limits.

Kai: We're looking at the paper: "The Electron-Recoil Response Tail and the LZ High-Energy Event Background".

Mira: Here's the gist of the paper. We reconstruct the continuum electron-recoil contribution to the LZ background expectation from the collaboration’s released simulation configuration The published characterisation of the LZ electron-recoil response fixes the centre of that response and does not determine the probability on which the background expectation for an event’s region depends Three families agree on this within a factor of about two at matched quantile deviation, and span two distinct construction mechanisms The ladder is short: a tail agreeing with that central characterisation to 0.47% makes it comparable to the entire background quoted for the comparison region The Central Finding The central result is that the published characterisation of the response does not determine the probability that matters, and carries no stated uncertainty on the quantity that does The authors construct families of distributions that reproduce the executable model’s mean, variance and standardised skewness exactly, and its event-conditioned 10, 50, and 90% Ne quantiles to about one per cent These families differ in the relevant tail probability by more than five orders of magnitude The thresholds that follow are correspondingly close: agreement to 0.47% makes it comparable to the entire background quoted for the region.

Kai: Constructive Counterexamples The families constructed are not offered as competing physical models of the liquid-xenon response, and the range they span is not a confidence interval on the leakage They are constructive counterexamples: distributions that satisfy the central characterisation defined in section 3—our summary of the released model, not an accuracy the collaboration claims—while giving different probabilities in the extreme tail.

Mira: Constraint and Measurement Requirements The question this section asks is what the published characterisation of that response determines about the probability at such a displacement The families are separated by their shape at a displacement of 4.1 standard deviations, not by their first three moments, which they share To constrain the shape and not only the width, a measurement at Nq ≃ 5 × 103 reaching the 10−5 level would substantially constrain the shapes admissible in section 3.

Kai: Implementation Divergences The implementation value is the one compatible with the published expectation: propagated through the parent spectrum and acceptance reconstructed here, Eq.

Title and authors: Kai: So, we're looking at this paper from arXiv titled "The Electron-Recoil Response Tail and the LZ High-Energy Event Background." Mira, what's the main vibe of that title?

Mira: It sounds really technical. It’s basically talking about how the tail of an electron recoil response affects a background expectation in LZ experiments. It suggests there's a big question about how much uncertainty we actually have on that tail because of how the collaboration characterized it.

Lev: I think the core idea is separating two hypotheses based on the expected rate, not just looking at the event itself or its shape. That's what this paper is zeroing in on, right?

Kai: Exactly. It’s about moving past just looking at one event and focusing on that total expected background count. This paper seems to be a deep dive into why the published characterization of that response doesn't actually tell you the probability that matters for discrimination.

Mira: Right, it points out that different ways of building the distribution can give you wildly different probabilities in the extreme tail, even if they match some basic statistical properties like mean and variance.

Lev: That’s a lot of freedom for models to have when you’re talking about physics at these scales. It means there isn't one single physical model being pinned down by the current characterization.

The paper's summary: Kai: So, what the authors are actually doing is constructing families of distributions that match the executable model perfectly—mean, variance, and standardised skewness—and they get their event conditioned ten percent, fifty percent and ninety percent quantiles to about one per cent. Mira, what’s the big takeaway from that construction?

Mira: The main point is that three different families of distributions agree on those central statistics within a factor of about two when you look at matched quantile deviation. But then, they span different tail probabilities by more than five orders of magnitude.

Lev: So they are all technically valid descriptions of the executable model's core behavior, but they disagree wildly on what happens in the far tail where you actually need to discriminate between things. That’s a huge gap for real hardware runs.

Kai: It seems like the published characterization doesn't constrain which of these five orders of magnitude we are dealing with. The paper says the ladder is short, meaning a tail agreeing with that central characterization to zero point four seven percent makes it comparable to the entire background quoted for the comparison region.

Mira: And then there’s this idea that you need a measurement at Nq about five times ten cubed reaching the one minus five level to constrain those shapes and not just their widths. That sounds like a very specific experimental prescription.

Lev: If we want to pin down the shape of that tail, we aren't looking at just the first few moments like mean or variance; we need data deep in the extreme tail to see how that distribution behaves.

The paper's improvements: Kai: The authors are setting up these constructive counterexamples to show what they’re arguing. They aren't presenting them as competing physical models of the liquid xenon response, which is important because it’s not claiming a new physics model. Mira, what do you think that means for how we interpret their results?

Mira: It means they aren't trying to tell us which physical process is right; they are just showing that based on the current characterization, you can have many different physical scenarios that all look good at the center but behave very differently in the extremes.

Lev: For someone building real hardware, this is a warning. It means whatever response shape we assume based on the simulation might be wildly off if we're operating in that extreme tail probability region. We need to know what constraints are needed to rule out those other five orders of magnitude possibilities.

Kai: And they suggest that the implementation value used in their code, which is compatible with the published expectation, actually doesn't match the one propagated through the parent spectrum and acceptance calculation in Eq. four <ref:2610.27105#pg2>. That ratio is about one point four seven five zero zero.

Mira: That mismatch between the implemented value and what’s calculated through propagation suggests that just using what was written down in a simple way isn't enough; you need to be careful about how you handle the physics of recombination and excitation when you build your detector model.

Lev: From an engineering side, if the implementation value is different, it means our noise modeling or our acceptance reconstruction might be picking up something that isn't fully accounted for by the simple published expectation.

Conclusion: Kai: So to wrap up this paper on "The Electron-Recoil Response Tail and the LZ High-Energy Event Background," the main point is that what settles whether a hypothesis is correct isn't just observing one event, it's about measuring that response distribution.

Mira: It’s not about finding the right shape based on just looking at the mean or variance; we need a measurement of the low-side response tail at Nq about five times ten cubed reaching the one minus five level to constrain the shape itself.

Lev: That measurement would be crucial because it tells us what kind of physics is allowed in that extreme region, and it lets us move from just checking if our model fits to actually constraining the physics.

Kai: They suggest using event level search data from S1c and log ten S2c to make that test reproducible, which is useful because we already have those measurements tabulated.

Mira: It’s a way to extend the response characterization by looking at that relevant tail, without needing an entirely new measurement campaign for something completely different.

Lev: For me, what this means is that we need better ways to handle uncertainty propagation when we extrapolate these response distributions into those deep tails.

Kai: We’ll keep an eye on how this constrains the possible models for the LZ background expectation. That’s where it leads us next.

More episodes

← Home