Scientific Seminar: MicroBooNE finds no evidence for a single sterile neutrino

0:00 Okay, welcome to Ramsey,

0:01 and on behalf of the Colloquium Committee, I'm Tom Junk,

0:05 and one of the reasons why we're here today is because

0:08 of the scheduling of this result with the publication ended up on a Wednesday,

0:13 so now it's a colloquium instead of a wine and cheese.

0:16 And because the Quantum people took One West,

0:18 here we are in Ramsey Auditorium, which is a really nice place.

0:22 So many thanks, Young-Kee, for providing this, and it's great.

0:25 So without further ado, let's introduce our very own Matt Toups,

0:29 co-spokesperson of MicroBooNE, along with Justin.

0:33 Matt?

0:33 I don't see Matt.

0:37 Yes, Justin, will you please be Matt?

0:40 I will be Matt.

0:42 Matt's up there.

0:43 Are you coming down, Matt?

0:46 Can I hold fort for a minute while we bring you down here?

0:53 Certainly, on behalf of myself and Matt,

0:54 we're very pleased to see you all here today.

0:57 Thank you for coming to see the release of this result.

0:59 Matt, I'll hand over to you.

1:03 It's great to be here.

1:06 Should I do another lap?

1:08 Thank you so much for coming.

1:09 We are so excited to share with you

1:11 this result published this morning on Nature.

1:13 Before we get started, I'd like to invite the lab director,

1:16 Young-Kee Kim, to say a few words of welcome,

1:19 and maybe something about neutrinos and the importance

1:21 of this result for the lab.

1:25 Young-Kee?

1:25 Well, I don't want to talk about

1:27 neutrinos because the speakers will talk about neutrinos,

1:29 but this is super exciting.

1:32 I'd like to thank Justin and Matt to lead

1:36 this collaboration and coming to this outstanding result.

1:40 I have to look at it.

1:41 I have to listen to it, you know, and judge.

1:45 But this anomaly has been going on for many, many days— decades.

1:53 So I'm super-excited that we have some conclusive— the conclusion.

1:59 Conclusive conclusion sounds funny.

2:01 And this is, again, very impactful to our science, and in particular, physics.

2:09 So I very much look forward to listening,

2:14 and congratulations to MicroBooNE Collaboration.

2:16 And this is our 10th anniversary, so it's a perfect time, so well done.

2:22 Thank you very much, Young-Kee.

2:25 Applause] One quick note.

2:28 We will have a reception following the colloquium.

2:31 So we'll have the colloquium, then a Q&A,

2:34 and then a reception afterwards, so please stick around for that.

2:38 And so, on behalf of Justin and myself, I'd like to invite our first speaker,

2:43 Hanyu Wei, professor at LSU, up to start our presentation.

2:47 Thank you.

3:01 All right.

3:02 Good afternoon, everybody.

3:03 I'm Hanyu Wei.

3:05 I'm an Assistant Professor at the Louisiana State University.

3:08 This is my colleague, Sergey Martynenko, from Brookhaven National Lab.

3:13 We'll give today's talk together.

3:28 Why doesn't it work?

3:34 Okay.

3:35 Good.

3:36 As advertised this morning on social media and various venues,

3:42 MicroBooNE has reported a landmark sterile neutrino

3:45 search using the first-ever two-beam and one-detector setup.

3:49 This result was published in Nature earlier today,

3:54 and it's a pleasure and honor for Sergey and me

3:59 to present this result on behalf of the whole MicroBooNE collaboration.

4:04 Okay.

4:04 Neutrinos, why study neutrinos?

4:08 Neutrinos remain some of the most intriguing particles in the universe

4:11 and are constantly challenging what we

4:13 think we know and pushing physics forward.

4:17 They open a unique window for probing some big questions,

4:20 such as if the standard model of elementary particles complete,

4:24 possible extensions,

4:25 and why is there much more matter than antimatter in the universe?

4:30 And how does a supernova and its energy, that feed the next generation star?

4:35 And how does a black hole emerge from collapse of a massive core.

4:39 And all beyond that, could there be

4:41 an entire sector of the universe, dark particles,

4:45 dark forces, that we've never seen and shape

4:48 the cosmic evolution from behind the scenes?

4:51 And neutrinos sit right in the middle of these mysteries.

4:55 They interact rarely.

4:58 Therefore, they are able to carry

5:00 critical information from extreme environments in nature

5:03 and provide one of the cleanest pathways

5:06 to new physics beyond the standard model.

5:09 So that's why we study them and use them

5:11 as the most powerful tool for discovering what comes next.

5:14 In the standard model of elementary particles,

5:18 neutrinos are fermions and matter particles coming in three flavors,

5:23 electron neutrino, muon neutrino, and tau neutrino,

5:27 and they are the second-most abundant in the universe.

5:31 They have no charge, no mass, and interact.

5:34 They are only weak forces, and one of the features that makes

5:40 neutrinos especially fascinating is they oscillate,

5:44 transform from one flavor to another as they travel.

5:47 This implies they are not massless and inconsistent

5:52 with what the standard model tells us.

5:54 And this is a simple two-flavor neutrino oscillation probability,

5:59 sinusoidal, as a function of L over E.

6:03 And L is the neutrino travel distance

6:06 or detector baseline relative to the neutrino source.

6:09 And E is the neutrino energy.

6:10 In the two-flavor picture,

6:12 let's assume only electron neutrino and muon neutrino.

6:16 Flavor eigenstates are superpositions of the two mass eigenstates,

6:20 and mixing can be described

6:22 by this two-by-two rotation matrix with one parameter,

6:26 theta, a so-called mixing angle.

6:29 And each mass eigenstate will pick up a different fifth value,

6:33 M squared times L over E.

6:36 And the fifth difference will be proportional

6:38 to delta M squared times L times over E.

6:42 That's the oscillation probability shown here.

6:44 And this factor is the sine squared 2 theta.

6:47 That is the mixing angle.

6:49 Back to this plot, assume only muon neutrino is produced at the source,

6:54 and the blue curve is the mu disappearance probability,

6:59 and the red curve is the appearance probability of the other flavor;

7:04 if the detector stays at the right baseline and right energy,

7:08 it can observe either an excess— one flavor showing up unexpectedly,

7:13 or a deficit of the flavor produced— at the neutrino source.

7:19 Okay.

7:19 This oscillation pattern can let us extract the mixing angles,

7:24 and the amplitude gives us the mixing angles,

7:27 and the frequency lets us extract the underlying mass squared difference,

7:31 which is the delta M squared term here.

7:35 For example, the in SuperK experiments,

7:38 the spectrum from atmospheric neutrino oscillation,

7:41 so the 1 over 2 deficit let us

7:43 extract the corresponding sine squared 2 theta value.

7:45 And the NOvA experiments, the spectrum from accelerator neutrinos,

7:49 and there's a dip here.

7:51 And the position of the dip corresponds to oscillation maximum,

7:54 you know, helped to review the underlying mass squared difference.

7:58 The journal results, the first journal results released two weeks ago,

8:02 there is also a dip, a different position,

8:04 and we viewed a different mass squared difference.

8:08 Okay.

8:09 In the full three flavor picture, flavor states and mass eigenstates mixed,

8:17 and can be described by a 3-by-3 unitary matrix, and we call it a PMNS matrix.

8:24 And each pair of mass eigenstates contribute to a two-flavor mixing like parts,

8:30 something like that.

8:31 The SuperK, NOvA contribute to this part.

8:35 And other experiments using reactor neutrinos and solar neutrinos,

8:39 they map out the other two parts extremely well.

8:42 Put them together, three mixing angles, one, three; one, two; two, three;

8:46 and two distinct mass squared difference corresponds

8:50 to these two characteristic L over E scales

8:54 of 10 to 3 or 10 to 5 kilometer per GeV or meter per MeV.

9:01 All right.

9:01 Over the past decades, extensive experimental results align beautifully

9:06 with the standard three flavor framework,

9:09 observing oscillations at the two characteristic

9:12 L over E scales we just discussed.

9:15 However, a few experiments have reported

9:18 anomalies due to possible oscillations at much, much smaller L over E.

9:23 This reminds us of the solar neutrino anomaly 60 years ago,

9:28 and where detectors observed a 30 to 50% deficit in solar neutrino flux,

9:33 which led to discovery of neutrino oscillations;

9:36 but what would this anomaly tell us?

9:38 And let me briefly review them first.

9:40 LSND and the MiniBooNE, V of e appearance anomalies,

9:46 electron neutrinos showing up unexpectedly.

9:50 And in this talk, the NuE may refer

9:52 to both electron neutrino NuE or electron anti-neutrino NuE bar.

9:56 But in this slide, I'll distinguish them.

9:59 The LSND experiment was to measure NuE bar

10:02 from a mu plus decay at rest assortment,

10:05 and the muons come from the decays of the secondary

10:09 mesons and generated by a proton beam striking target.

10:13 But in this neutrino source, the flux,

10:16 neutrino flux, is NuE or NuMI bar dominated,

10:20 electron neutrino, red, and the mu anti-neutrino, the cyan part.

10:24 And this muon energetic corresponds to mu neutrino from pion decay.

10:28 So as we see here, and it's negligible, the electron anti-neutrino component.

10:32 So, however, in the LSND experiment, this measurement,

10:36 so it observed a significant excess of electron anti-neutrino,

10:41 and this excess, where do they come from?

10:46 So it can be interpreted as a NuMu bar to NuMI bar oscillation,

10:52 where the NuMu bar comes from.

10:55 And this excess covers the L over E range from 0.4 to 1.4 meter per MeV,

11:01 much, much smaller than the two characteristic L

11:05 over E scales we just discussed in our scenario.

11:09 Okay, MiniBooNE appearance anomaly.

11:12 So MiniBooNE experiment, motivated by the LSND anomaly,

11:16 was to measure electron neutrinos from a different neutrino source,

11:20 decaying flight source.

11:21 So the neutrinos come from meson decays in flight, rather than at rest.

11:28 So this neutrino has much higher energy, so GeV or sub-GeV scale.

11:33 So this is the flux of the MiniBooNE experiment,

11:36 which is also the flux used by MicroBooNE, which we'll discuss later.

11:40 So the y-axis is log scale.

11:43 So the red curve represents the NuE component, which is very, very small,

11:47 just a sub-percent level,

11:49 and the MiniBooNE observed a significant excess in this measurement.

11:53 So the data represented by the dots, error bars,

11:57 and stacked histograms represent the prediction

12:00 of the expected number of events.

12:02 And this excess covers a LV range and similar to the LSND anomaly,

12:09 and also, can be interpreted as the oscillation

12:13 from NuMu from the source to NuE.

12:17 If we put together the NuE/NuE bar

12:22 excess to prediction ratios from different energy bins,

12:28 from different baselines of these experiments,

12:31 and expressing them as an appearance probability,

12:33 we get this plot as a function of L over E.

12:37 And the black curve represents the underlying appearance,

12:41 NuMu to NuMu appearance oscillation probability, right, something like this.

12:45 And this angle makes the angle.

12:48 Sine squared 2 theta NuE is dedicated to NuMu to NuE appearance channel.

12:53 All right, so if we map these parameters onto

12:56 a 2D plane of sine squared theta and delta M squared,

13:00 that is what they look like.

13:02 And we call these extended areas allotted regions,

13:05 and they are extended because they reflect the uncertainties in the measurement.

13:11 And a lot of regions of LSND in a MiniBooNE

13:15 cluster together and at EV scale delta M squared.

13:19 In addition to NuE appearance anomalies, there are NuE disappearance anomalies

13:25 from gallium experiments or certain reactor experiments.

13:29 The gallium experiments— GALLEX, SAGE,

13:31 BEST— they use radiochemical detectors based on gallium

13:35 that can only measure total reaction rate.

13:38 So they measure electron neutrinos from some active

13:43 sources corresponding to a few discrete sub-MeV energies.

13:47 And they constantly observe a 20% deficit

13:51 of the electron neutrinos produced from radioactive sources.

13:56 And this deficit can be interpreted as a disappearance

13:59 effect of the NuE produced at the source.

14:02 And L over E values correspond to a few discrete points,

14:04 ranging roughly from 0.5 to 2.5 meters per MeV.

14:08 And the Neutrino-4 experiment,

14:10 mirroring electron anti-neutrinos from the reactor core,

14:13 observed an oscillation pattern in a similar L over E region.

14:18 So this is a lot of regions from Neutrino-4 gallium anomaly,

14:24 the X-axis, and sine squared 2 theta ee.

14:27 So we use theta ee for NuE to NuE disappearance.

14:31 If we put back the appearance channel,

14:34 we notice that they point to a similar delta M squared scale, an ee scale.

14:40 This is 0.1.

14:42 This is 10.

14:43 So the oscillation hypothesis provides a plausible, testable,

14:49 and very natural explanation for all of them.

14:53 Okay, back to this slide,

14:55 so several anomalies due to possible oscillations at a small L over E,

15:01 and it corresponds to a very high delta M squared,

15:05 inconsistent with these two known mass splittings,

15:09 and imply existence of extra, heavier, flavor,

15:12 and mass eigenstates, and commonly called a sterile neutrino,

15:16 which has no standard model interactions.

15:19 And this talk will focus on such— any scale,

15:24 light sterile neutrinos, producing small L over E oscillations.

15:29 And in the 3+1 framework, this is additional flavor,

15:33 nu-s sterile neutrino, and this is additional mass eigenstates, M4.

15:37 And the delta M squared we discussed in previous slides,

15:40 is actually delta M squared 41, and ignoring the difference amount, M123,

15:44 and they are very, very small, compared to N4.

15:48 And if such a sterile neutrino can be found,

15:51 it will be a beyond a standard model

15:54 particle and reshape our understanding of mass generation,

15:58 lepton number violation, that kind of thing,

16:01 and profoundly impact particle physics,

16:03 extra-particle physics, as well as cosmology.

16:08 Okay, MicroBooNE, the focus of today's talk,

16:11 has already contributed to the sterile

16:14 neutrino interpretation of these anomalies,

16:16 by searching for sterile neutrino-induced oscillations.

16:20 This is our 2023 result, an old result,

16:23 and we set a constraint in the 2D primary space.

16:27 And this constraint corresponds to a exclusion curve, and meaning,

16:32 the oscillation parameters to the right side

16:36 of this curve corresponds to stronger oscillation signals,

16:40 and would have stood out clearly in our data.

16:43 But because we saw no oscillation in our data,

16:47 this region was excluded, this exclusion curve.

16:49 And in addition to this MicroBooNE old result, earlier experiments,

16:53 mostly accelerator experiments and short baselines, or low L over E,

16:58 had set their respective constraints in this 2D plot,

17:02 because they saw no oscillation in their data.

17:06 And as we see, this is the accumulative excluding curve,

17:09 represented by a solid black line.

17:12 As we see, the 3+1 sterile neutrino model is still an open possibility.

17:16 And there is another channel, okay, new MicroBooNE result.

17:22 So it'll be very exciting to see a new MicroBooNE result

17:26 and see how much it can contribute to clarifying this picture.

17:31 There's another channel, NuE to NuE disappearance channel, just to remember,

17:36 and the constraints from the earlier reactor neutrino experiments, the PROSPECT,

17:41 STEREO, have ruled out most regions allowed by the neutrino 4

17:45 and the gallium— and with the most recent result from KATRIN.

17:49 And KATRIN released its latest sterile neutrino search

17:53 result earlier today in the same issue of Nature,

17:57 together with the new MicroBooNE result, but used a different method.

18:01 It's not an oscillation-based method.

18:03 It's based on the beta decay spectrum kink.

18:06 All right, but KATRIN is more sensitive to high stratum,

18:10 four regions above the scale in this plot.

18:13 But with this new KATRIN result, I think,

18:16 if I'm not mistaken, the entire regions are excluded.

18:19 So this channel is very interesting, a very interesting situation.

18:23 But it will be also interesting to see whether new MicroBooNE

18:26 results can see something on it using a very different neutrino source.

18:33 Okay, where is MicroBooNE and what is MicroBooNE?

18:36 So MicroBooNE sits on the Booster Neutrino Beam Line in Fermilab,

18:40 the Booster Neutrino Beam, this 8 GeV proton beam with a target somewhere here,

18:46 producing neutrinos and passing through our detector.

18:49 And the MicroBooNE detector is 70 meters upstream

18:53 of the MiniBooNE detector here along the same beam line.

18:57 So it can provide a direct test of the MiniBooNE anomaly.

19:01 But MicroBooNE uses a very different detector technology, liquid argon TPC,

19:05 a combination of liquid argon perimeter,

19:09 high-resolution, facing a 10-projection chamber.

19:12 So liquid argon TPC enables

19:14 a powerful neutrino flavor reconstruction and selection,

19:18 and MicroBooNE's L over E ranges from 0.2 to 2,

19:21 so it can probe the EV-scale sterile neutrino

19:26 oscillations for all the anomalies we just discussed.

19:33 Okay, the strategies of the last known neutrino search at MicroBooNE 3+1.

19:39 So one additional sterile flavor and full 3+1 analysis,

19:43 meaning all detectable oscillation effects are considered in our data,

19:48 including NuE/NuE appearance,

19:49 NuE/NuE disappearance, this is what we just discussed,

19:53 and also, NuMu/NuMu disappearance,

19:55 and a NuE to NuMu oscillation, NuE/NuMu to mu-tau or new-sterile.

20:00 But these two channels are negligible in our data,

20:02 and they correspond to a very small number

20:04 of electron neutrino events or neutral current events.

20:08 So the three main oscillation effects are considered here.

20:13 We perform a simultaneous feed on all available

20:16 NuE and NuE interaction channel in our data, and this way we can also reduce

20:21 the shared systematic uncertainties and improve the sensitivity.

20:25 So this is, actually, a multi-parameter problem,

20:28 but we can project our result onto different

20:33 2D planes and directly compare it to the available,

20:36 you know, a lot of regions reported by these other experiments.

20:43 So in addition to the NuMu to NuE channel that we just mentioned,

20:47 we also set constraints in the NuE to NuE channel,

20:50 though this sensitivity is not very competitive.

20:53 Okay, so the sensitivity of the old result, 2023, is much worse than expected,

21:02 and mainly due to a degeneracy in oscillation parameters.

21:07 And this is a challenge that we need to deal with and what it is.

21:11 The basic idea is the appearance and the disappearance

21:15 effect in NuE spectrum can cancel each other.

21:17 So the oscillated spectrum is indistinguishable from non-oscillated spectrum.

21:22 For example, this is non-oscillated spectrum,

21:26 NuE, and this is the energy spectrum.

21:31 If you only consider disappearance effect,

21:33 a small fraction of NuE also is missing,

21:35 so times Pee, that's NuE/NuE probability.

21:39 If you only consider appearance effect,

21:42 a small fraction of mu-neutrino oscillating to NuE

21:45 is added on top of the NuE events.

21:47 So if the appearance effect happens to offset the disappearance effect,

21:51 that is what we have.

21:53 The net oscillated spectrum is the same as non-oscillated.

21:57 That is the region of the degeneracy,

22:00 but we have NuMu spectrum, if you remember.

22:02 We fit to all NuE and new NuMu, but here,

22:06 the NuMu spectrum provides little help with this situation.

22:09 And the reason is, even if the appearance effect is

22:13 as large as the total NuE events from the beam,

22:16 this p-MuE/NuMu/NuE oscillation probability is still a very, very small number,

22:20 as there is only a tiny NuE component

22:23 in the accelerated neutrino beam used through a microbeam.

22:25 And also, in general, in an accelerated neutrino beam,

22:29 the NuE component is very small.

22:31 So that will result in, basically,

22:34 invisible disappearance effect in the NuMu spectrum.

22:36 All right, so degeneracy arises.

22:40 These two effects cancel each other,

22:42 meaning the ratio of these two mixing angles

22:45 is equal to the beam NuE to NuMu ratio,

22:49 which is almost a fixed value for different energies in a beam,

22:54 meaning there's also no shape information that we

22:57 can use to disentangle the disappearance of this beam.

23:02 All right.

23:02 So MicroBooNE has two beams,

23:05 so this is where the MicroBoonNE's unique capability comes in.

23:10 So in addition to the booster neutrino beam on access,

23:13 we see neutrinos from a different beam.

23:16 The neutrinos, that main injector, the degree of access,

23:20 MicroBooNE gives a very different proton beam in respect to the target.

23:26 These two beams have very different NuE component.

23:29 So this is the flux from the BNB.

23:32 Only about 0.5% NuE, BNB beam, and for NuMI, about 4%.

23:38 So if we calculate the beam, NuMu to NuMI ratio,

23:43 BNB, 200, NuMI, 25, so one order of magnitude smaller.

23:47 So meaning they have quite different degeneracy points if we use

23:51 a single BNB or single NuMI beam to do the oscillation search.

23:56 So we have two handles, lift appearance and disappearance degeneracy.

24:02 Now what happens if we combine these two beams?

24:05 So we made this plot to demonstrate this.

24:08 So this is according to frame.

24:11 X and Y axes are the ratios of oscillated

24:16 and to non-oscillated NuE events in BNB and NuMI.

24:21 So these two dashed lines mark the degeneracy point.

24:26 This is for BNB.

24:27 This is for NuMI.

24:28 They correspond to different ratios.

24:30 This is NuMI to NuMI ratio from the beam.

24:33 So where do those mixing angles sit in this plot?

24:36 Here we go.

24:37 So we add some curves,

24:39 and each curve represents a constant value of the oscillation parameter.

24:44 For example, the red one corresponds to sine squared

24:47 2 theta ee dedicated to NuE to NuE disappearance effect.

24:51 And this greenish curve represents a constant value for another angle,

24:56 sine squared 2 theta NuE, NuE to NuE appearance effect.

24:59 And the NuMI to NuMI disappearance effect mixing angle

25:04 can be determined if we know these two values, and they are correlated.

25:10 Okay.

25:11 So if you take a look at this intersection

25:13 where we have sine squared 2 theta e 0.1,

25:15 sine squared 2 theta NuE 0.002, that is what we have.

25:19 This set of parameters will produce a stronger appearance effect

25:23 in the NuMI spectrum resulting in a net excess in the BNB/NuMI spectrum.

25:30 The same set of parameters will produce a stronger disappearance effect,

25:34 resulting in a net deficit in the NuE NuMI spectrum.

25:38 But we don't need to do this calculation,

25:40 we can directly read this information out from the coordinate frames.

25:43 So here, the BNB sits at 1.1, meaning 10% excess,

25:48 and the BNB 0.96 is a 4% deficit.

25:53 So if we place the full 2D grid, including the values of the mixing angles,

25:59 and X and Y-axis tell us the NuMI

26:01 and the BNB excess or deficit with the corresponding solution effect,

26:06 then we can, relatively straightforward to, sense how the degeneracy is broken.

26:11 For example, we can scan along this line.

26:14 It corresponds to BNB degeneracy point because the ratio

26:18 is always one for these oscillation parameter values.

26:22 The BNB, no change.

26:23 The NuMI deficit, if you take a look at the Y-axis value,

26:27 it rapidly grows with an increase in sine squared 2 theta e,

26:30 meaning we can disentangle the disappearance effect from the appearance effect.

26:34 And also, based on this plot,

26:36 we can sense the BNB is more sensitive to appearance

26:41 effect because the gradient of this curve is roughly along X-axis,

26:46 and the gradient of the sine squared 2

26:48 theta ee disappearance effect is roughly along Y-axis.

26:51 So NuMI is more sensitive to disappearance.

26:54 The survey will discuss further about that.

26:57 Okay.

26:57 We can use the more accurate

26:59 and predictive NuMI spectrum to cross-check this understanding.

27:04 So NuMI spectrum, BNB/NuMI spectrum, NuMI,

27:06 this is where we start, so no oscillation effect.

27:09 So we scan different sine squared 2 theta ee,

27:13 sine squared 2 theta NuE change accordingly.

27:16 That is what we have.

27:17 So the BNB/NuMI spectrum, barely shift, but NuMI develops a pronounced deficit

27:25 with an increasing sine squared 2 theta ee,

27:27 meaning we could use NuMI to constrain this appearance effect.

27:31 Subtracting the disappearance effect in BNB,

27:33 then we can probe appearance effect, so that is what we do.

27:37 All right.

27:38 So the new result, at least in today's Nature, used the first three years of BNB

27:44 and NuMI data and combined these two beams together,

27:48 unlocked the full power to test the 3+1 model

27:52 and compared to a lot of regions by the reported anomalies.

27:57 So the dotted blue curve represents the new sensitivity of today's new result,

28:05 and we achieved 0.1% sensitivity in this NuMI

28:08 appearance channel capable of testing these anomalies,

28:11 and in the NuMI to NuMI disappearance channel,

28:14 we also achieved a competitive sensitivity compared to KATRIN and PROSPECT.

28:19 But we used a very different neutrino source, so much higher energy;

28:23 ee sub-GeV compared to MeV1 or sub-Mev are usually used in this.

28:28 All right.

28:29 I'll stop here and hand it over to Sergey.

28:32 He will walk you through the analysis and the results.

28:38 Fantastic.

28:39 All right.

28:40 Thanks, Hanyu.

28:44 Let's talk about MicroBooNE 3+1 result.

28:49 But before talking about results, we need to discuss several pieces that are,

28:54 actually, in combination, give us a better understanding of the results we have.

28:57 So we need to discuss our data, our event selection,

29:00 systematics, and statistical analysis, starting with our data.

29:05 As Hanyu already mentioned, we do have two beams, and across two beams,

29:10 we use half of available MicroBooNE data for this analysis.

29:15 With BNB beam, we reached about 6.4, 10 to 20 POTs,

29:19 and for NuMI, we reached about 1, 10 to 21 POTs.

29:25 The most important part about these two beams for our result is that BNB,

29:29 it's almost pure NuMu, and it's very sensitive to NuE appearance.

29:33 And NuMI, it has much larger NuE

29:37 content and more sensitive to NuE disappearance.

29:41 That's our data.

29:42 For event selection, our event selection starts with our detector.

29:46 And for a detector, we used a low-energy threshold,

29:50 fully active liquid argon calorimeter,

29:52 and a high energy and position resolution time projection chamber.

29:57 This detector is excellent at identifying different species

30:01 of particles and reconstructing 3D images with fine-grained precision.

30:06 Here on the right, you can see the event display from our detector.

30:09 You can see a vertex of— neutrino vertex very well.

30:13 You can see a couple protons coming out,

30:15 a shower from pion decay, maybe delta ray and pion,

30:19 but one of the important qualities of the detector,

30:24 it's our electron-photon separation power, which we do in two steps.

30:31 First is dQ/dx, which is ionization energy loss.

30:34 It is two times different between one electron and a photon.

30:39 And also, we are looking for a gap between the vertex,

30:43 neutrino vertex, and the start of the shower.

30:46 This enables our high-performance NuE selection out

30:50 of an overwhelming NuMu charge and neutral current events.

30:54 All right.

30:56 The reconstruction in our detector is performed

30:59 using the Wire-Cell package in four stages.

31:02 It starts with noise filtering and signal processing.

31:07 Then it comes to 3D imaging, clustering, and charge-light matching.

31:12 Then we do 3D trajectory and dQ/dx fitting and cosmic muon tagging.

31:18 And lastly, we do multitrack fitting, 3D vertexing, and particle identification.

31:25 For neutrino energy, we use calorimetric reconstruction,

31:30 which means we sum the reconstructed kinetic

31:32 energies of all the particles in the event.

31:35 We add the rest mass values for muons, electrons, and pions,

31:41 and we also add the average binding energies per

31:44 nucleon for each proton we see in the event.

31:49 With this, we achieve a good resolution of about 15% and bias

31:53 of about 10% for our neutrino energy reconstruction for our NuEs and NuMus,

31:59 and this performance is similar for both BNB and new medians.

32:04 Right.

32:04 With this selection, we can take a look at our target channels,

32:09 which are obviously NuEs, and you don't see the data points just yet.

32:15 We will put them on later, but what you can see here is the green colors,

32:20 which is our NuE and anti-NuE CC selection

32:23 for BNB on the left and NuMI on the right.

32:26 You can also see these blue patches here.

32:29 It's our NuMu and anti-NuMu CC selection.

32:32 In the other shades of green, it's our CC pi-naught and NC pi-naught selection.

32:37 So the main takeaway here is that we have very few selection of NuE's.

32:41 and the main backgrounds are NuMus and pi-naughts.

32:44 All right.

32:46 But it's not just these channels that we use.

32:49 We have a lot more, so let's walk through all of them.

32:53 First of all, we split our NuE

32:55 CC channel into fully contained and partially contained.

32:58 The reason is based on particularly constructed

33:01 particles of fully in fiducial volume and up,

33:05 and that gives us an increased statistics channel.

33:08 We also do have sidebands.

33:11 That's our NuMus, also split and fully and partially contained.

33:15 And our NuMu channels are meant to constrain our NuMI prediction,

33:19 due to universality of cross-section modeling

33:22 for CC interaction and common hadronic parentages.

33:27 And also, we have three pi-naught channels,

33:30 CC pi-naught fully contained, partially contained, and NC pi-naught,

33:34 and pi-naughts are a major background for our NuMI

33:37 search because it mimics the NuMI signature.

33:41 All of these channels, which you can see seven of them here,

33:44 but in total there are 14 because there are two beams,

33:48 are fit simultaneously in our analysis.

33:52 Let's go through these sideband channels one by one, starting from our NuMu's.

33:57 So here on the left, I always show BNB.

34:02 On the right is NuMI version of our NuMu selection,

34:06 with good chi squares for both of them.

34:09 But also, you can see that both of them have— you can see under prediction,

34:15 which is still consistent at one sigma level.

34:19 I'll get back to the NuMI part of this slide later in the talk.

34:25 But for now, let's move to our CC pi-zero selection,

34:29 again with BNB and NuMI having a good data Monte Carlo agreement.

34:33 And last but not least, our sideband, it's NC pi-naught.

34:38 And again, BNB on the left and NuMI

34:40 on the right have a good data Monte Carlo agreement.

34:44 All right.

34:47 That concludes our event selection.

34:49 So let's talk about our systematics.

34:51 We have three main sources of systematics.

34:54 It's detector systematics, cross-section, and flux.

34:57 Let's start with detector.

34:59 So we have a range of tools that we use to characterize our detector.

35:04 It's a cosmic muons, protons, laser.

35:06 Like an example, you can see here on the right,

35:09 we can plot the reconstructed entry and exit position of cosmic muons.

35:15 But then, we evaluate the range of systematic effects,

35:18 like covering light yield, space charge, combination, wire waveform simulation.

35:24 On the right, again, you can see the effect of space charge.

35:28 So in a perfect world, the positions should lie nicely on the angle here,

35:32 but they are not because of space charge.

35:34 So we compare our data to Monte Carlo

35:38 to assess the magnitude from each of the effects,

35:42 and we use modified Monte Carlo samples to assess

35:45 the impact of these uncertainties on the analysis.

35:48 All right.

35:50 The second one is the cross-section, but before the uncertainty,

35:54 a couple words about our cross-section model.

35:56 So for the cross-section model, we have a base cross-section model,

35:59 which is GENIE v3, and it's shown as a blue line here on the plot.

36:04 But when we do tune it to T2K,

36:07 NuMu CC 0 pi cross-section, which are data points here on the right,

36:12 that's how we obtain our base MicroBooNE 2 model, which is the red line here.

36:20 Then we can use this model

36:22 and various sets of cross-section parameters around it,

36:26 like lines here, and the various set of parameters,

36:29 which is a total of 57 of them.

36:31 We obtain this one sigma band, and we can propagate the same parameter variation

36:37 to the event rates to build the covariance matrices.

36:41 And here, the example of, actually,

36:44 a correlation matrix because it's easier to read,

36:49 how you can read this, BNB correlations are

36:52 bottom left and NuMI on the top right.

36:55 And they split in NuEs and NuMus.

36:57 And the one, which is yellow, is, basically, the highest correlation, right?

37:03 So you can see immediately that NuEs

37:06 and NuMus are highly correlated for both beams.

37:09 And that's, as I mentioned,

37:12 because of universality of cross-section modeling for CC interactions.

37:19 We can leverage these correlations to constrain

37:22 the cross-section model and its uncertainties.

37:25 So as an example, let's take a look

37:29 at our uncertainty band for BNB NuE selection,

37:33 and here, only cross-section systematics.

37:36 And we can constrain our BNB NuEs with our BNB NuMus, okay?

37:43 And you can immediately see that due to high correlation,

37:46 the uncertainties significantly shrunk.

37:48 All right.

37:49 The last but not least, systematics of a flux,

37:54 and before talking about flux systematics itself,

37:57 a couple words about our flux model that we updated prior to this analysis,

38:02 why we started looking at updating it.

38:05 So at 8 degrees of axis,

38:08 the NuMI flux is quite different from the traditional on-axis.

38:13 Here on the left, you can see BNB flux for NuMus,

38:18 and on the right is NuMu flux for NuMus.

38:22 For BNB, most of NuMus come from pi-models.

38:27 But for NuMI, there is a higher fraction of kaons later on.

38:32 The model must account for kaon production very carefully.

38:36 But then, we started looking at available models.

38:39 We understood that they disagree significantly.

38:43 For example, here, there are some available models,

38:47 which are Geant 4.10, 4.10, 4.11, and 4.9, shown in different colors,

38:53 and there is not much data that exists

38:57 to constrain the K+ production for a new medium.

39:01 So what we did, we used the base

39:04 model that agrees best with available data for us,

39:09 which is Geant 4.10.

39:11 And we applied the very conservative uncertainty up to 40%,

39:16 where there is no coverage based on the model spread.

39:20 All right.

39:22 This slide, basically, shows the summary of what went into our updated flux.

39:28 We did update some inline geometry.

39:31 We updated our baseline model.

39:32 We introduced constraints from NA49 and others similar to NOvA and MINERvA,

39:38 and we did very conservative treatment

39:40 of uncertainties outside the data coverage.

39:43 On the right, you can see the ratio of old to new

39:46 flux for NuMu's on the left and NuE's on the right.

39:53 With new flux, we can, again,

39:55 take a look at our correlation matrix for two beams,

39:58 for NuEs and NuMus, and the same thing here.

40:02 BNB is at bottom left, and NuMI is at top right.

40:06 You immediately can see that between two beams,

40:09 the correlation part here is basically

40:11 zero because we treat them as uncorrelated,

40:15 because they have different hydronic processes and beam energies,

40:19 and also, for NuMI beam,

40:22 correlations are much stronger because they have a higher fraction of K pluses,

40:27 as we discussed, and it's also a dominant decay for NuEs.

40:30 So that's why the correlations are very high.

40:32 And also, for off-axis beam, there is a high number of re-interactions,

40:36 which also contributes to this high— So as I promised,

40:42 we are kind of getting back to our NuMu sidebands, and there is a way how we can

40:48 validate our NuMu flux modeling using our BNB data.

40:52 So as we discussed, BNB and NuMI

40:56 correlated in detector and cross-section systematics,

41:00 but their flux matrix is uncorrelated.

41:04 So what we can do, we can

41:07 constrain our NuMI/NuMu selection with BNB NuMus, right?

41:11 What it's going to give us,

41:15 the detector and cross-section systematics would largely cancel out,

41:18 leaving us mostly with flux.

41:21 And also, the NuMu prediction will be updated based on BNB data.

41:26 So let's take a look at this constrained NuMI/NuMu spectrum,

41:32 which is here in blue.

41:35 So you immediately can see that there is a good data Monte Carlo agreement.

41:40 The updated prediction agrees well with data,

41:43 and uncertainty here is dominated by the flux.

41:47 So with that procedure, we consider our NuMI flux model validated.

41:53 All right.

41:55 So we kind of discussed these three pieces,

41:57 all of them come into our statistical analysis.

42:00 And to understand the result, we need to remind ourselves of this equation,

42:06 which is the probability of oscillation

42:09 at short baseline for different neutrino flavors.

42:12 But the most important part for us here

42:15 is this sine squared, 2 theta alpha beta,

42:18 which, obviously, can be different based on what you are trying to calculate.

42:22 For NuMI appearance, it's sine squared 2 theta NuE; and for NuMI disappearance,

42:26 it's sine squared 2 theta ee;

42:27 and for the NuMu disappearance, it's sine squared 2 theta NuMu.

42:31 In our analysis, we explore, actually, the right part of these equations.

42:37 So we explore 3D parameter space in delta M squared for 1,

42:41 sine squared theta 14, and sine squared theta 24.

42:45 But all the results are given in 2D,

42:48 in delta M squared for 1, sine squared 2 theta NuE, or ee.

42:53 Right.

42:54 So to do so, we do profiling over sine squared theta 24, which is, basically,

43:00 minimizing chi squared over the one axis and calculating,

43:03 then, the result in two dimensions.

43:07 Right.

43:07 This is the full chi square that we use.

43:10 We discussed our measurements, we discussed our prediction,

43:12 and the covariance matrix that's coming to it.

43:15 It's also worth reminding that this chi square that we use,

43:19 it includes all 14 channels simultaneously.

43:22 And how can we use it?

43:24 Here is the pipeline that we follow in our analysis.

43:28 First, we can get best fit value

43:31 for oscillation parameters in four new hypotheses.

43:35 Then, with best fit, we can do a data consistency test

43:40 against three new hypotheses using the Feldman-Cousins procedure.

43:44 And after, we can set limits via

43:48 Frequentist CLs method if the data consistency passed.

43:55 Okay.

43:55 So now, all these pieces come together into our results,

44:01 and we can start with putting data points on our target channel,

44:07 starting from the BNB.

44:09 So I'll show the NuE CC channels here on the left

44:13 and its NuMu counterpart for this beam on the right.

44:18 This one is not new.

44:19 We published it in 2023.

44:21 But still, it's worth reminding us that we do have

44:25 a small deficit here in the BNB NuE CC channel,

44:30 and we do have access in NuMu CC.

44:36 But the interesting thing that happens in our 14-channel simultaneous

44:40 fit is that our NuMus will be constraining our NuE channel,

44:46 updating the prediction, and potentially, pulling it up based on new data.

44:51 Okay.

44:52 This one is new.

44:53 So that's our new NuMI selection.

44:57 On the left, again, it's NuE CC channel.

45:00 On the right, it's our NuMus, and both of them has a small data Monte Carlo set.

45:08 But in the same sense as BNB,

45:12 in our simultaneous fit, NuMus will constrain our NuEs, updating the prediction,

45:18 and reducing the normalization of that, as you see here.

45:21 Let's, actually, illustrate how this constraining procedure works.

45:26 Let's isolate these two channels, BNB NuEs and NuMI NuEs here.

45:31 And without constraint, chi square is about 37.9 for 40 degrees of freedom.

45:38 But then, we can constrain it with all non-NuE channels that we got,

45:43 and we have this blue line here.

45:46 And with constraint, chi square increases a little bit to 41,

45:50 which makes sense because the uncertainty is strong.

45:54 But this pool of the prediction up is coming from NuMus, as we discussed,

46:00 because most constraining power comes from NuMus,

46:04 and that's how our constraining procedure works in our NuMI channel.

46:09 The other interesting thing to look at here is how

46:13 our fractional uncertainties change in our NuE CC selection, because after all,

46:18 we are a statistics limited analysis

46:20 with dominant systematics being in flux and cross-section,

46:23 and what we really need is to reduce our uncertainties for better results.

46:29 So here is fractional uncertainty without constraint for BNB and NuMI NuEs.

46:35 Then we shrink the systematic uncertainties, and we get this black line,

46:40 which is already significantly reduced fractional uncertainty.

46:44 But also, as you remember, the prediction for both BNB and NuMI is pulled up,

46:50 which is so that it reduces the systematic uncertainty,

46:52 and higher for NuMIs because the pull-up is higher for NuMIs.

46:57 Okay.

46:57 So with this data, we finally can

47:01 go and follow our statistical analysis pipeline throughout.

47:07 So starting with best fit values, here it is.

47:12 So our best fit value is here on the left,

47:16 and we compare this best fit line, which is for new best fits,

47:22 red line here, with non-oscillation prediction, which is, in fact,

47:26 a histogram for BNB on the left and NuMI on the right.

47:30 Basically, you can see almost no difference between them.

47:34 The reason is because we got this very low delta M squared for one,

47:39 which leads to negligible oscillation effect.

47:42 With this best fit, we can get a test against a new hypothesis.

47:51 To do so, we use the Feldman-Cousins procedure.

47:53 So we throw about 10,000 toys and we

47:56 get this nice distribution of delta chi squared, and we can put our data,

48:02 delta chi squared, which is around 0.2, very close to zero here,

48:06 and that gives us a p-value of about 0.96.

48:10 What does p-value mean?

48:11 It means that BNB and NuMI data is consistent with three new hypotheses.

48:16 All right.

48:19 Now we can go ahead and set the limits

48:23 and know that the data is consistent with what we knew,

48:28 and we can start discussing our limits with our— in this channel.

48:33 So first, there is our sensitivity.

48:38 So we compare our sensitivity, which is this blue line here,

48:41 with our previous BNB-only result, which is our red line here.

48:45 So sensitivity, obviously,

48:46 improves with respect to BNB-only case because we mitigated the degeneracy

48:52 of oscillation parameters and also NuMI and NuE contributions in this channel.

48:59 But now, where actually our data stands?

49:02 So let's take a look at our data exclusion.

49:05 So that's our data exclusion.

49:07 This is solid red line here,

49:10 and you can see that MicroBooNE at 95% confidence level

49:15 in this channel excludes most of gallium-allowed

49:18 regions and parts of neutrino-4-allowed regions.

49:23 We also can put our data result in respect

49:27 of our Brazil band around our median sensitivity.

49:31 So here, blue line is median sensitivity.

49:34 Green band is one sigma, and orange band is two sigma bands around it.

49:39 So you can see that data exclusion

49:42 is mostly weaker than the sensitivity and lays,

49:45 primarily, inside the one sigma band.

49:48 Why is that?

49:49 So we need to remember that disappearance channel is more influenced

49:54 by the NuMI because it has four times more NuEs than BNB.

49:59 And NuMI, after constraint, data matches prediction very well.

50:06 But still, BNB data has this deficit that weakens the NuE disappearance limit.

50:15 Okay.

50:15 Now, we can go through he same logic with our appearance channel,

50:20 first, comparing our sensitivity with our previous result.

50:25 And again, our sensitivity improves with respect

50:29 to BNB-only case because of degeneracy mitigation.

50:33 But now, we can finally take a look

50:38 at our exclusion limit at appearance channel, and that's how it looks.

50:43 This solid line is a huge jump in chi delta-M square,

50:47 and we exclude at 95% confidence level LSND 99% allowed region.

50:55 And also, we exclude vast majority of minimal 95% allowed region.

51:02 Right.

51:02 Same logic with disappearance.

51:05 We can put it in context of our Brazil band around median sensitivity.

51:11 You can see, again, that our exclusion

51:14 limit is stronger than median sensitivity,

51:17 and it lies on the border of two sigma regions.

51:22 Why this is happening?

51:24 Let's take a look.

51:27 Again, we get back to this constraint, NuE predictions for BNB and NuMI.

51:33 And what we need to know about this appearance channel is that it's dominated

51:38 by the BNB beam because it has four times less NuEs and two times more NuMIs.

51:43 And this BNB, it has a data deficit that boosts the appearance limit,

51:50 which basically favors smaller appearance angles.

51:55 But also, there is another way to look at this result.

51:59 For this, we need to step back and remember our BNB only result, so here it is.

52:06 These red lines show BNB only full 3+1 result that considers appearance,

52:12 NuE appearance, NuE disappearance, and NuMI disappearance all together,

52:16 the solid line being data result and the dashed line being sensitivity.

52:23 And obviously, the data result is slightly

52:26 stronger than sensitivity because of the BNB deficit.

52:30 But what we also published back in 2023 is appearance only result.

52:36 It is unphysical in our case because it

52:39 disregards any effects from NuE disappearance and NuMI disappearance,

52:43 this black solid line here,

52:46 and it's obviously stronger than our full result because

52:51 it doesn't have any degeneracy coming from this channel.

52:55 But what it also shows, it shows, basically,

52:58 the limit that can be reached with BNB only data if there

53:03 would be no degeneracy of oscillation

53:05 parameters coming from other oscillation channels.

53:08 So let's overlay this appearance only result with our new two-beam full result.

53:18 So you can see that they are almost

53:20 similar except this small part around the— [inaudible] squares.

53:24 And the reason for this is that basically, NuMI provides the maximum degeneracy

53:31 mitigation to the BNB-dominant appearance channel,

53:34 and that's why we reach this limit.

53:38 All right.

53:38 So just to summarize what I've discussed right now.

53:43 So we saw no evidence for 3+1 sterility in the model.

53:47 We exclude, at 95% confidence level,

53:50 the 3+1 explanation for LSND and MiniBooNE anomalies

53:55 and exclude most regions allowed by the gallium and neutrino-4.

54:01 And with that, we close the window for 3+1

54:04 model as an explanation for long-standing LSND and MiniBooNE anomalies.

54:09 And here, I'll give the torch back to Hanyu

54:13 to put these results into a broader perspective.

54:23 Okay.

54:23 Thank you, Sergey, for the presentation.

54:26 Good to see you again.

54:28 So now, I'm going to wrap up today's

54:32 talk with an outlook where this takes us next.

54:36 So as Sergey just summarized, the 3+1 framework no longer provides a variable

54:41 explanation for the LSND and MiniBooNE anomalies.

54:44 But those anomalies, the observed access data, haven't gone away;

54:49 so do the reactor and gallium anomalies, and they also remain unexplained.

54:56 And these anomalies continue to draft a global program

55:02 of current next-generation experiments

55:04 in that resolving the remaining possibilities.

55:07 I just list some of them, not exhaustive,

55:11 and many of these experiments can provide a direct

55:14 test of these anomalies as they employ similar neutrino sources,

55:18 similar detector configurations, or in a similar LRE range.

55:22 And many of them incorporate more

55:24 advanced detection techniques or detector technologies,

55:26 so more sensitive investigation of these anomalies.

55:30 And among these experimental efforts,

55:32 a central program is a strong baseline neutrino program at Fermilab,

55:36 which is an extension of MicroBooNE with two

55:40 additional liquid argon TPC detectors at different baselines,

55:45 so the near-detector SBN and the far-detector ICARUS.

55:49 So the near-far detector configuration can substantially reduce systematic

55:54 uncertainty in flux in the cross-section detector and improve sensitivity.

56:01 And also, the SBND detector,

56:02 in close proximity to the beam target and ICARUS is very massive,

56:07 and both fully operational,

56:09 and have already collected the largest statistics of data

56:14 in the [inaudible] band L over E region.

56:16 So this enabled the SBN program to provide a high statistics,

56:23 low-systematic exploration of the last far neutrino oscillation.

56:28 And not only NuE-appearance or NuE-disappearance, as we just discussed,

56:33 but also the NuMu-disappearance channel,

56:35 and to which the MicroBooNE alone has very limited assistance.

56:40 All right.

56:40 In addition to sterile neutrino oscillation search,

56:43 the SBN program can provide a more comprehensive and sensitive

56:47 preparation of a broader landscape of beyond-the-standard model physics as well,

56:52 some of which will be highlighted in the following

56:57 slides and with relevance to the LSND and the MiniBooNE.

57:01 All right.

57:02 For example, expanded sterile neutrino models,

57:07 this includes possibilities like 3+1 plus additional sterile flavors,

57:13 plus 2, for example,

57:16 or 3+1 and augmented by new dynamics involving the sterile state.

57:22 For example, non-standard interactions,

57:24 decaying sterile neutrinos, and some other examples.

57:29 So MicroBooNE has begun exploring the 3+2 framework

57:34 and is actively investigating the 3+1 plus decay scenario.

57:40 In addition to sterile neutrino oscillation searches,

57:42 non-oscillation searches targeting photon final states are making significant

57:48 progress towards alternative explanations for LSND and minimum anomalies.

57:54 So because the electromagnetic signature in the excess

57:58 events may arise from photon final states,

58:01 which convert into a pair of electrons,

58:04 a positron, rather a single electron from oscillated mu.

58:10 And the excess photon final states may

58:13 come from unexpected or underestimated standard model processes.

58:17 For example, the delta-related decay, single photons,

58:21 coherent single photon production from current interactions,

58:25 and the MicroBooNE recent analysis saw no sign

58:28 of excess in these two standard model processes,

58:32 but we do see an interesting two-sigma excess

58:36 in a model of agnostic inclusive single photon search,

58:41 particularly in this zero-proton channel, as shown here.

58:45 These two bins, 100 to 300 MeV, and this is a two-sigma.

58:50 So this motivates further investigations in MicroBooNE and SBN.

58:56 All right, another active frontier is a broader landscape of dark sector models.

59:03 And for example, the neutrino-induced upscattering,

59:06 dark matter product-induced upscattering,

59:08 or heavy neutral decay generated from the same target,

59:12 and many of these models predict photon final state

59:16 or direct the E plus or E minus final state.

59:19 So signatures that MicroBooNE or SBN can identify with high precision,

59:25 as we just discussed in the photon studies.

59:28 This is the event at play of an e plus/e

59:31 minus signal from a dark sector model simulation in MicroBooNE,

59:35 with the argon-TPC detector.

59:38 And recently, MicroBooNE released the world's

59:41 first direct search for dark sector

59:45 e plus/e minus signal as a possible explanation to the MiniBooNE anomaly.

59:49 This is the result.

59:49 This is data with error bars.

59:52 We saw no sign of excess of such signals compared to a green dashed,

59:57 the histogram, representing typical e plus/e

59:59 minus signal from a dark sector model.

1:00:03 All right, but alternative dark sector models,

1:00:06 particularly those with very different characteristic angular distributions

1:00:10 of e plus/e minus pair or energy distributions,

1:00:14 may remain compelling targets for further

1:00:18 investigation of MicroBooNE and the SBN program.

1:00:21 All right.

1:00:23 I'll finish up here and back to the main results discussed today.

1:00:28 So using the first two beams,

1:00:30 one detector search of it can— MicroBooNE sees no sign

1:00:33 of sterile neutrino-induced oscillations in the relevant L over E regime,

1:00:37 and closing the 3+1 window associated with the anomalies.

1:00:41 With this chapter settled, the path is now open for new ideas

1:00:47 and discoveries in future theoretical and experimental efforts.

1:00:52 More to come.

1:00:53 Stay tuned.

1:00:54 Thank you for listening.

1:00:59 Applause] Great.

1:01:06 Thank you, Hanyu and Sergey.

1:01:08 Even though I've known about those results for quite a long time,

1:01:12 it's still really nice to see them up there on the screen at last.

1:01:16 And it's really nice, now,

1:01:18 to actually get the opportunity to start hearing some feedback from all

1:01:22 of you outside of a collaboration on these results and to get your questions.

1:01:27 So I think we now have about 10 minutes to take questions, so please.

1:01:31 I see one up there, three up this side.

1:01:34 Go for it, Tom.

1:01:35 Yes.

1:01:42 Okay.

1:01:42 So first of all, a huge congratulations,

1:01:44 and thanks for beautiful results and also beautiful talks.

1:01:49 I really like the results in your talks.

1:01:53 I have a question.

1:01:54 So you said that you have very negligible detector systematics,

1:01:59 regardless of, you know, you have the NuMI beam or not, right?

1:02:03 Did you say that?

1:02:05 You have small detector systematics?

1:02:09 Negligible systematics for detector?

1:02:11 Not negligible, but not dominant.

1:02:15 They're not dominant, yeah, compared to the other two.

1:02:18 So that's true even though you don't use any NuMI beam.

1:02:24 So detector systematics is still small,

1:02:27 no matter what you use, NuMI beam or not?

1:02:31 Yes.

1:02:32 Okay.

1:02:32 So how did you achieve such good, you know, detector systematics?

1:02:39 It's a good detector.

1:02:40 [Laughter] I think that the detector design, the whole collaboration,

1:02:46 tremendous effort, with the KATRIN detector,

1:02:49 we better understand detector response.

1:02:51 We also produce some [inaudible] version samples,

1:02:55 something like that, to help us

1:02:57 to estimate detector systematic uncertainty in high precision.

1:03:01 Okay.

1:03:01 So that's one thing we need to learn from you,

1:03:04 and another one quick question is that, so the last slide of your— Hanyu's talk,

1:03:10 so you compared the sensitivity with 2023 and the current sensitivity.

1:03:19 So I see that the bigger science care data values,

1:03:25 so your sensitivity degraded with the NuMI beam, so why is that?

1:03:33 Multiple Speakers] You mean at the bottom?

1:03:36 Yeah, yeah, bottom.

1:03:37 Yeah, those, you know, higher [inaudible] UE.

1:03:41 So you see the current sensitivity is degraded compared to the 2023?

1:03:49 Because it's not exactly an apple-to-apple comparison, right,

1:03:53 because here we compare data results with the sensitivity results.

1:03:57 So 2023 is the data results?

1:04:00 The data, so it would be more fair to compare sensitivity— Oh, okay.

1:04:03 I see.

1:04:04 I thought it was—

1:04:04 then they would be converging at this point and that's going to be fine.

1:04:07 Yes.

1:04:07 It would be good to see some apple-to-apple as well.

1:04:10 But this is data, so it's kind of a little bit to the left.

1:04:13 Okay.

1:04:13 So, okay.

1:04:15 Thank you.

1:04:23 In the future on the SBN program where you have two detectors on the BNB beam,

1:04:29 how useful will the NuMI beam be, or will you need it at all?

1:04:34 Right.

1:04:34 To me, I think that NuMI beam will

1:04:39 further boost the sensitivity in that beam program.

1:04:42 ICARUS can see the neutrino from NuMI for sure.

1:04:46 And also, though we have near-far detector,

1:04:49 but there's still sort of degeneracy.

1:04:52 NuMIcan help, but this degeneracy will be weaker as we have,

1:04:58 you know, a near-far detector and different devices still, but NuMI can help.

1:05:10 Another one up there, Tom?

1:05:17 All right.

1:05:18 Thanks for the talk.

1:05:19 This is great.

1:05:20 It's great to use NuMI beam to do the searches.

1:05:23 I had two quick but highly correlated questions.

1:05:26 One is, what is the NuMu bar contamination in NuMI fluxes?

1:05:36 Do you have that?

1:05:39 This is negligible.

1:05:40 Let me find the flux somewhere.

1:05:51 I'm not sure we have a breakdown for NuMu and anti-NuMI,

1:05:55 but it's a small fraction.

1:05:57 Yeah.

1:05:57 Percent level?

1:05:59 Oh, percent level?

1:06:00 Yeah.

1:06:00 Yeah.

1:06:01 Oh, okay.

1:06:01 And in our selections, it's also very small.

1:06:03 It's bigger than NuE, but much, much smaller than NuMu.

1:06:08 Wait, but NuE contamination is 5%.

1:06:10 So then, NuMu can't be a percent level.

1:06:13 It needs to be like, what, 20%?

1:06:16 I think my question is, if anti-NuMus are relevant for this, and if they are,

1:06:22 how do you correlate the cross-sections,

1:06:25 because the anti-NuMu cross-section will

1:06:27 have a very different Q-squared dependence,

1:06:29 and it will lead to very different topologies in your detector, right?

1:06:34 So your NuMu sample may have a much larger contamination from the NuMu bar,

1:06:39 which will change the efficiencies, I would expect, due to topology.

1:06:43 Do you have the NuMI spectrum somewhere?

1:06:47 We don't have a breakdown for NuMu anti-NuMu here,

1:06:52 but our cross-section model, you know, treats NuMu and the NuMI bar differently,

1:06:59 and that information will go into prediction, for example,

1:07:03 in this predictive electron,

1:07:05 as well as in the systematic uncertainty estimation.

1:07:08 Yeah, the key thing is that that will be in the covariance matrix.

1:07:12 Right.

1:07:12 It will be accounted for.

1:07:14 We just did not distinguish NuMu/NuMI bar here in this plot.

1:07:17 Yes, but in general, we do.

1:07:19 Yeah, okay, thank you.

1:07:26 Tom, there's a couple of questions down here.

1:07:31 There's one there, and then Young-Kee has one over here.

1:07:37 Background Discussion] Can I ask something?

1:07:41 It's not directly related to MicroBooNE.

1:07:45 Go ahead.

1:07:46 Or KATRIN, how— what's going on?

1:07:50 Do we have any results with the mass, neutrino mass?

1:07:55 Neutrino mass?

1:07:56 Yeah, that was the main goal for this.

1:07:58 That was the main goal.

1:07:59 Yeah, but they did release something at Neutrino didn't they?

1:08:02 Now, what is the new KATRIN result?

1:08:05 You mentioned about their— Well, yes.

1:08:08 The paper that came out today was a stab on neutrinos.

1:08:11 I understand, but I was curious what's going on with their main goal.

1:08:15 Ah, yeah.

1:08:16 Main goal is to do mass.

1:08:25 0.40.

1:08:26 0.45.

1:08:28 0.45.

1:08:29 EV.

1:08:30 That is the current limit.

1:08:32 0.45 EV.

1:08:38 My voice may or may not support this question.

1:08:42 I was watching as it went by.

1:08:44 You showed us more than a dozen fit results

1:08:47 with chi-squared over a number of degrees of freedom.

1:08:51 They were all less than one.

1:08:53 Statistics doesn't do that.

1:08:56 So somehow, the errors are too big.

1:09:07 I know exactly where to go.

1:09:10 I mean, yes, the errors are big.

1:09:12 The errors are concerning, but what we are doing in the end—

1:09:17 Your constraint plot?

1:09:19 Yeah, I'll show you a constraint plot.

1:09:20 Yeah, a constraint plot.

1:09:23 Far, far back.

1:09:24 Yeah, here we go.

1:09:27 Here we go.

1:09:28 This one.

1:09:29 Yeah.

1:09:30 Yeah.

1:09:31 So, yeah, because after constraints, the uncertainty shrinks.

1:09:35 So, yeah, we do have a concerted uncertainty that in a simultaneous

1:09:39 14-channel fit would be changing due

1:09:41 to the correlations from our chi-statistic sidebands.

1:09:46 In many of these plots— all of these plots are correlated with each other.

1:09:53 They're all showing much the same distribution in different ways.

1:09:57 So I think in a feature of a, you know,

1:10:00 a somewhat systematically limited analysis,

1:10:02 once one of your plots has a chi-squared degree of freedom below 1,

1:10:09 given that a lot of these error bars are highly correlated across their length,

1:10:15 any distribution you make on that data is going to be,

1:10:19 sort of, similarly correlated in a way.

1:10:31 Even the top-left plot, is a little— 9 out of 25,

1:10:34 even by itself, all in one, is a little unlikely, but— Yeah.

1:10:39 we'll give you one.

1:10:40 Oh, yes.

1:10:40 No, it's a systematically dominated analysis

1:10:44 with a highly correlated bin-to-bin systematics.

1:10:51 Thank you very much for the nice talk.

1:10:53 Could you go back to the slide with a neutral energy estimate?

1:11:00 I have two questions.

1:11:03 So, first of all, so what is the source of the binding energy value,

1:11:07 the eight-point something, maybe, I think?

1:11:10 And the second thing, if you could produce better energy estimate for neutrinos,

1:11:15 including neutrons, could you also get further improvement in the sensitivity?

1:11:23 For your second question, the answer is yes,

1:11:26 and I think there is some work on deep learning energy estimation for MicroBooNE

1:11:32 that actually might go later

1:11:34 in the analysis and potentially boost the sensitivity.

1:11:38 And for the second question, I missed it.

1:11:41 The binding energy?

1:11:43 Second question.

1:11:43 For the first question, I missed the first question.

1:11:45 Binding energy.

1:11:47 What's the source of the binding energy?

1:11:49 What's the source of the binding energy?

1:11:54 Some paper.

1:11:55 Some measurement.

1:11:56 A test.

1:11:57 Yeah.

1:12:07 Okay.

1:12:08 I see.

1:12:09 I see.

1:12:10 Yeah.

1:12:10 Because the values keep changing, so I was curious about it.

1:12:13 Right.

1:12:13 Thank you very much.

1:12:14 But that one, we also did some study.

1:12:15 We can vary that a little bit, so the impact is not very significant.

1:12:18 Oh, sure.

1:12:19 Thank you.

1:12:24 Any more questions?

1:12:26 Oh, there's one.

1:12:31 Thanks, Pedro.

1:12:32 Okay.

1:12:33 Sorry for like dominating, but just one more question, I promise.

1:12:36 Do you have sensitivity to NuMu disappearance?

1:12:40 Because if you can do the appearance at 10 to the minus 3,

1:12:44 and the NuMI disappearance at the 10%,

1:12:46 it feels like you should be able to get NuMu disappearance at the percent level,

1:12:51 right, so did you check?

1:12:53 Yes.

1:12:53 We checked that.

1:12:55 It's not that sensitive, but we can provide some results in print,

1:13:02 but we didn't publish that result.

1:13:06 But is that, like, at the 10% level or the 1% level?

1:13:10 Ten-percent level.

1:13:11 Okay.

1:13:11 Yeah.

1:13:11 So that is, basically, the uncertainty in the NuMu spectrum.

1:13:16 But in the NuE, that is something different,

1:13:18 because that sensitivity is basically shrank,

1:13:20 because the NuMu to NuE ratio is very, very high.

1:13:24 We measure the NuE, but it comes from NuMu,

1:13:27 but NuMu disappearance is still in NuMu channel,

1:13:30 so that sensitivity will correspond to that uncertainty.

1:13:35 Okay.

1:13:36 NuMu disappearance is very hard with a single detector.

1:13:41 It's a dead record of your spectrum.

1:13:43 You don't have this NuMu to NuE constraint.

1:13:46 You can't use that in the same way to shrink your uncertainty,

1:13:50 so that is somewhere where you really need

1:13:53 a multi-detector SBN program to do a good job there.

1:13:58 All right.

1:13:59 I don't think I saw any hands, but if there is one— oh.

1:14:06 So if I'm following what you said earlier,

1:14:08 then you have— your systematics are correlated bin to bin,

1:14:13 and you've not taken out a separate global systematic.

1:14:19 If that's the case, then, in fact, if you had accounted for that as a global,

1:14:24 you would have had much worse chi squares.

1:14:28 That, in turn, would have decreased your sensitivity in each of your points.

1:14:40 And, basically, it sounds to me, you know, that, in fact,

1:14:44 what you've done is you have over-coverage,

1:14:46 and your limits, actually, should be weaker.

1:14:57 I'll leave that to you to think about.

1:15:00 Yeah, I mean, it's not clear why that would

1:15:03 make the limits weaker at all, certainly to me.

1:15:06 I mean, the sequential CLS method that we do at the end,

1:15:10 it takes into account all

1:15:12 of the possible uncertainties in the covariance matrix,

1:15:16 as opposed to fake data experiments,

1:15:18 and that's how we calculate the exclusion region.

1:15:24 And, also, there's a strong

1:15:26 normalization systematic uncertainty for something like

1:15:28 that, but if you remember the preceding band of the sensitivities,

1:15:34 and we add the one-sigma,

1:15:37 two-sigma band around the median sensitivity, we do see, you know,

1:15:41 where data exclusion result is consistent

1:15:45 with any reasonable result or uncertainty.

1:15:54 The uncertainties are correlated.

1:15:55 You did it the same way.

1:15:56 That's fine, but I'm questioning whether the methodology is right,

1:15:59 if you have a global correlated systematic,

1:16:01 which needs to be accounted for only once.

1:16:04 But the constraint, basically, reduced the systematic.

1:16:07 It doesn't matter.

1:16:08 There's still a global effect.

1:16:11 If there's a global effect that's going to shift the whole thing up and down,

1:16:14 then that should be counted once and not as part of a broader bin-to-bin.

1:16:18 So you can account for a global normalization shift in a covariance matrix.

1:16:25 It just looks like a flat covariance matrix,

1:16:27 but you can encode that type of uncertainty in a covariance matrix.

1:16:31 And so, the normalization uncertainties that are a part

1:16:33 of this analysis are encoded in that covariance matrix.

1:16:37 You get the same result as if you took it out, so it's included.

1:16:49 I think one more question down here, and then, I think we have to wrap up.

1:16:54 So we've got time for one more question, and then we,

1:16:57 actually, have to leave and get to the reception.

1:17:04 Should we go online?

1:17:05 So if you want one more question, Tony, go for it.

1:17:10 Okay.

1:17:11 So I just want to make sure that I understood you correctly.

1:17:15 So the reason why you have much better, you know,

1:17:18 sensitivity with the NuMI beam is not, actually, because of the NuMI beam.

1:17:23 That's because you constrain with even the BNB beam.

1:17:25 You constrain with the other in the channel,

1:17:28 the disappearance channel to appearance channel, right?

1:17:31 That gives a huge reduction.

1:17:33 So this plot is a little bit— I mean, not this plot.

1:17:35 I mean, so 2023 versus 2025 results comparison could be a little

1:17:41 bit misleading because the main reason could be due to NuMI beam,

1:17:46 but it's not NuMI beam.

1:17:48 Even the BNB beam, you know,you can

1:17:51 reduce a lot with constraining; is that correct?

1:17:55 I would say mostly correct.

1:17:56 I think the power of this channel comes from the BNB beam,

1:18:00 but the question is if we just do a single BNB beam study,

1:18:05 if we do 2D comparison, there will be degeneracy,

1:18:08 but NuMI helped to break this channel.

1:18:10 NuMI policy unlocked the power from BNB.

1:18:13 As mentioned, so here this sensitivity

1:18:15 comes from the very large NuMI-NuMI ratio,

1:18:17 but that is only true for BNB, not NuMI.

1:18:21 Okay.

1:18:22 Maybe I want to talk to you more later.

1:18:26 Great.

1:18:26 Thank you.

1:18:26 Right.

1:18:27 So the reception, I think, is outside One West.

1:18:30 So, yes.

1:18:31 Yes, please do now all head up to the atrium,

1:18:33 and thank you very much for coming.

1:18:35 Applause] Thank you.

1:18:39 Applause]