The plots have been reported in a DP note with number DP-2021-001 .
Contact: cms-phys-conveners-jme@cernNOSPAMPLEASE.ch
Pileup mitigation techniques are important for performing physics analysis in high pileup scenarios, such as in the data recorded by the CMS Collaboration during Run 2 of the LHC. The pileup per particle identification (PUPPI) algorithm was commissioned with 13 TeV data. While PUPPI generally showed improved performance compared to the charge hadron subtraction (CHS) algorithm in many aspects, the initial version of PUPPI yielded worse jet energy resolution for high-pT jets. In order to improve this behaviour of PUPPI, a new tune was developed. The results of this new tune, the achieved jet energy resolution, efficiency and purity, pileup jet rate, and the impact on jet substructure variables and missing transverse momentum resolution are shown in this presentation. All studies are done using the recently reconstructed 2017 "Legacy" data. Previous and latest PUPPI tune are compared using the same data. The label "PUPPI v11a" represents the configuration used in Ref [1], the label "PUPPI v15" represents the new configuration.
[1] CMS Collaboration, “Pileup mitigation at CMS in 13 TeV data”, JINST 15 (2020) P09018, doi: 10.1088/1748-0221/15/09/P09018, arXiv:2003.00503.
JER The jet energy resolution is defined as the spread of the response distribution, which is Gaussian to a good approximation. The width (σ) of a Gaussian fit to the response in the range [m - 1.5σ, m + 1.5σ], where m is the mean, determined with an iterative procedure. Jet energy corrections are applied to the reconstruction-level jets such that the ratio of reconstruction and particle-level jet p_{T} (the response) is on average 1 (response-corrected). Expected change: PUPPI v15 is designed to improve the JER especially at high p_{T} and to result in a similar performance then CHS.
The glossary can be found at the end of the document.
No | Link | Figure | Caption |
---|---|---|---|
1 a | Figure 1: Jet energy resolution as a function of the particle-level jet p_{T} for PF jets with CHS applied (red open triangles), PF jets with PUPPI v11a applied (blue open squares) and PF jets with the new tune of PUPPI v15 (black filled circles) in QCD multijet simulation in different η regions. | ||
1 b | Figure 1: Jet energy resolution as a function of the particle-level jet p_{T} for PF jets with CHS applied (red open triangles), PF jets with PUPPI v11a applied (blue open squares) and PF jets with the new tune of PUPPI v15 (black filled circles) in QCD multijet simulation in different ηregions. | ||
1 c | Figure 1: Jet energy resolution as a function of the particle-level jet p_{T} for PF jets with CHS applied (red open triangles), PF jets with PUPPI v11a applied (blue open squares) and PF jets with the new tune of PUPPI v15 (black filled circles) in QCD multijet simulation in different η regions. | ||
1 d | Figure 1: Jet energy resolution as a function of the particle-level jet p_{T} for PF jets with CHS applied (red open triangles), PF jets with PUPPI v11a applied (blue open squares) and PF jets with the new tune of PUPPI v15 (black filled circles) in QCD multijet simulation in different ηregions. | ||
1 e | Figure 1: Jet energy resolution as a function of the particle-level jet p_{T} for PF jets with CHS applied (red open triangles), PF jets with PUPPI v11a applied (blue open squares) and PF jets with the new tune of PUPPI v15 (black filled circles) in QCD multijet simulation in different η regions. | ||
1 f | Figure 1: Jet energy resolution as a function of the particle-level jet p_{T} for PF jets with CHS applied (red open triangles), PF jets with PUPPI v11a applied (blue open squares) and PF jets with the new tune of PUPPI v15 (black filled circles) in QCD multijet simulation in different ηregions. | ||
2 a | Figure 2: Jet energy resolution as a function of PU for p_{T} = 30 GeV (a) and p_{T} = 500 GeV (b) for PF jets with CHS applied (red open triangles), PF jets with PUPPI v11a applied (blue open squares) and PF jets with the new tune of PUPPI v15 (black filled circles) in QCD multijet simulation in |η|<1.3 . | ||
2 b | Figure 2: Jet energy resolution as a function of PU for p_{T} = 30 GeV (a) and p_{T} = 500 GeV (b) for PF jets with CHS applied (red open triangles), PF jets with PUPPI v11a applied (blue open squares) and PF jets with the new tune of PUPPI v15 (black filled circles) in QCD multijet simulation in |η|<1.3 . | ||
3 a | Figure 3: Jet energy resolution as a function of PU for p_{T} = 30 GeV (a) and p_{T} = 500 GeV (b) for PF jets with CHS applied (red open triangles), PF jets with PUPPI v11a applied (blue open squares) and PF jets with the new tune of PUPPI v15 (black filled circles) in QCD multijet simulation in 2 < |η|<2.5. | ||
3 b | Figure 3: Jet energy resolution as a function of PU for p_{T} = 30 GeV (a) and p_{T} = 500 GeV (b) for PF jets with CHS applied (red open triangles), PF jets with PUPPI v11a applied (blue open squares) and PF jets with the new tune of PUPPI v15 (black filled circles) in QCD multijet simulation in 2 < |η|<2.5. |
The efficiency is defined as the fraction of particle-level jets with p_{T} > 30 GeV that match within ∆R < 0.4 with a reconstruction-level jet with p_{T}
> 20 GeV. The purity is defined as the fraction of reconstruction-level jets with p_{T} > 30 GeV that match within ∆R < 0.4 with a particle-level jet
with p_{T} > 20 GeV from the main interaction. The p_{T} cuts at reconstruction and generator level are chosen to be different to remove
any significant JER effects on this measurement.
Expected change: While no big difference in |η| < 2.5 are expected, an
increase in efficiency for higher |η| regions is observed. This is due to
the fact that PUPPI v15 has a lower N_{vertices} -dependent p_{T}
requirement. This means that less low p_{T} particles are rejected and more PU jets can be formed.
The glossary can be found at the end of the document.
No | Link | Figure | Caption |
---|---|---|---|
4 a | Figure 4: The LV jet efficiency (upper) and purity (lower) in Z+jets simulation as a function of the number of interactions for PUPPI v11a (blue open squares), CHS (red open triangles), and the new tune of PUPPI v15 (black filled circles). Plots are shown for AK4 jets with p_{T} > 20 GeV, and with (a,d) | η| < 2.5, (b,e) 2.5 < |η| < 3 and (c,f) |η| > 3. | ||
4 b | Figure 4: The LV jet efficiency (upper) and purity (lower) in Z+jets simulation as a function of the number of interactions for PUPPI v11a (blue open squares), CHS (red open triangles), and the new tune of PUPPI v15 (black filled circles). Plots are shown for AK4 jets with p_{T} > 20 GeV, and with (a,d) | η| < 2.5, (b,e) 2.5 < |η| < 3 and (c,f) |η| > 3. | ||
4 c | Figure 4: The LV jet efficiency (upper) and purity (lower) in Z+jets simulation as a function of the number of interactions for PUPPI v11a (blue open squares), CHS (red open triangles), and the new tune of PUPPI v15 (black filled circles). Plots are shown for AK4 jets with p_{T} > 20 GeV, and with (a,d) | η| < 2.5, (b,e) 2.5 < |η| < 3 and (c,f) |η| > 3. | ||
4 d | Figure 4: The LV jet efficiency (upper) and purity (lower) in Z+jets simulation as a function of the number of interactions for PUPPI v11a (blue open squares), CHS (red open triangles), and the new tune of PUPPI v15 (black filled circles). Plots are shown for AK4 jets with p_{T} > 20 GeV, and with (a,d) | η| < 2.5, (b,e) 2.5 < |η| < 3 and (c,f) |η| > 3. | ||
4 e | Figure 4: The LV jet efficiency (upper) and purity (lower) in Z+jets simulation as a function of the number of interactions for PUPPI v11a (blue open squares), CHS (red open triangles), and the new tune of PUPPI v15 (black filled circles). Plots are shown for AK4 jets with p_{T} > 20 GeV, and with (a,d) | η| < 2.5, (b,e) 2.5 < |η| < 3 and (c,f) |η| > 3. | ||
4 f | Figure 4: The LV jet efficiency (upper) and purity (lower) in Z+jets simulation as a function of the number of interactions for PUPPI v11a (blue open squares), CHS (red open triangles), and the new tune of PUPPI v15 (black filled circles). Plots are shown for AK4 jets with p_{T} > 20 GeV, and with (a,d) | η| < 2.5, (b,e) 2.5 < |η| < 3 and (c,f) |η| > 3. | ||
5 a | Figure 5: Rate of jets in the PU-enriched region divided by the rate of jets in the LV-enriched region as a function of the number of vertices for CHS (red triangles), PUPPI v11a (blue squares) and the new tune of PUPPI v15 (black circles) jets in Z+jets simulation (open markers), and data (full markers). Jets that overlap with ΔR < 0.4 with one of the two muons from the Z boson decay are removed. For reference, the rate of jets that are matched to a particle-level jet in simulation is also shown for CHS jets (red solid line). The plots show the ratio for events with |η|< 2.5 (a) and |η| > 2.5 (b). The lower panels show the data-to-simulation ratio. | ||
5 b | Figure 5: Rate of jets in the PU-enriched region divided by the rate of jets in the LV-enriched region as a function of the number of vertices for CHS (red triangles), PUPPI v11a (blue squares) and the new tune of PUPPI v15 (black circles) jets in Z+jets simulation (open markers), and data (full markers). Jets that overlap with ΔR < 0.4 with one of the two muons from the Z boson decay are removed. For reference, the rate of jets that are matched to a particle-level jet in simulation is also shown for CHS jets (red solid line). The plots show the ratio for events with |η| < 2.5 (a) and |η| > 2.5 (b). The lower panels show the data-to-simulation ratio. |
Jet substructure information is important to identify Lorentz-boosted decays. The performance is tested on a simulated sample of events containing a graviton decaying into a pair of scalar bosons subsequently decaying into two quarks. Scalar masses between 70 and 90 GeV are considered to mimic the W boson kinematics. With a Gaussian fit the mean and width of the Soft Drop Mass distribution is determined iteratively. The median of τ_{21} is estimated on the same set of events. Expected change: PUPPI v15 allows more particles to end up in an AK8 jet. This will especially influence the jet mass. The jet mass and τ_{21} value should be increased a little bit in the absolute value, however, the PU stability should remain.
The glossary can be found at the end of the document.
No | Link | Figure | Caption |
---|---|---|---|
6 a | Figure 6: Mean Soft Drop jet mass (a) and Soft Drop jet mass resolution (b) for AK8 jets from boosted bosons with 400 < p_{T} < 600 GeV for CHS (red triangles), PUPPI v11a (blue squares) and the new tune of PUPPI v15 (black circles) jets in a bulk graviton decaying to two scalar bosons signal sample, as a function of the number of vertices. The error bars correspond to the statistical uncertainty in the simulation. | ||
6 b | Figure 6: Mean Soft Drop jet mass (a) and Soft Drop jet mass resolution (b) for AK8 jets from boosted bosons with 400 < p_{T} < 600 GeV for CHS (red triangles), PUPPI v11a (blue squares) and the new tune of PUPPI v15 (black circles) jets in a bulk graviton decaying to two scalar bosons signal sample, as a function of the number of vertices. The error bars correspond to the statistical uncertainty in the simulation. | ||
7 | Figure 7: Median τ_{21} for AK8 jets from boosted bosons with 400 < p_{T} < 600 GeV for CHS (red triangles), PUPPI v11a (blue squares) and the new tune of PUPPI v15 (black circles) jets in a bulk graviton decaying to two scalar bosons signal sample, as a function of the number of vertices. The error bars correspond to the statistical uncertainty in the simulation. |
The performance of the transverse missing energy is estimated in Z+jets events, where the Z boson is reconstructed by two muons. The only source for p_{T}^{miss} is the miscalibration of physics objects. The hadronic recoil is used to estimate the amount of miscalibration. The hadronic recoil is split into two components the parallel (u_{ || }) and perpendicular (u_{⟂}) component. The parallel component is a measure for the response, hence the jet resolution. The RMS of the perpendicular and the parallel component is a measure for the resolution of p_{T}^{miss}. Expected change: The response (μ(−u∥/pT(Z))) of the hadronic recoil is correlated with the jet energy resolution. Therefore, an improvement is expected compared to the improvement of JER. The resolution (σ(u∥ − pT(Z))) is expected to improve at high pT due to the improvement in JER at high p_{T} . The resolution of the perpendicular component (σ(u⊥)) is affected by the additional particles added by the v15 tune and is worse.
The glossary can be found at the end of the document.
No | Link | Figure | Caption |
---|---|---|---|
8 a | Figure 8: The hadronic recoil response (a) and resolution (b) for the component u∥ for PUPPI v11a, PUPPI v15 and PF pmiss as a function of the p in Z → μμ events in collision data. | ||
8 b | Figure 8: The hadronic recoil response (a) and resolution (b) for the component u∥ for PUPPI v11a, PUPPI v15 and PF pmiss as a function of the p in Z → μμ events in collision data. | ||
9 | Figure9: The hadronic recoil resolution for the component u_{ ⟂} for PUPPI v11a, PUPPIv15 and PF p_{T}^{miss} as a function of the number of vertices in Z → μμ events in collision data. |
The impact of the improved track-vertex association for the PUPPI algorithm (PUPPI v15) on object reconstruction performance is shown. The studies are performed on data collected in 2017 and reconstructed with the most recent version of the CMS software and the final detector alignment and calibration (legacy reprocessing). The only difference is the tune of PUPPI. In addition, a comparison to charged hadron subtraction is presented. The jet energy resolution is improved for PUPPI v15 to the level of the jet energy resolution of jets with CHS applied. Substructure variables like soft drop mass and mass resolution show the same stability against pileup as before. The PUPPI p_{T}^{miss} resolution is also found to be improved compared to PF p_{T}^{miss} for hadronic recoils up to more than 500 GeV.
[1] CMS Collaboration, “Pileup mitigation at CMS in 13 TeV data”, JINST 15 (2020) P09018, doi: 10.1088/1748-0221/15/09/P09018, arXiv:2003.00503.
[2] D. Bertolini, P. Harris, M. Low, and N. Tran, “Pileup per particle identification”, JHEP 10 (2014) 059, doi:10.1007/JHEP10(2014)059, arXiv:1407.6013.
[3] M. Cacciari, G. P. Salam, and G. Soyez, “The anti-kT jet clustering algorithm”, JHEP 04 (2008) 063, doi:10.1088/1126-6708/2008/04/063, arXiv:0802.1189.
[4] M. Cacciari, G. P. Salam, and G. Soyez, “FastJet user manual”, Eur. Phys. J. C 72 (2012) 1896, doi:10.1140/epjc/s10052-012-1896-2, arXiv:1111.6097.
[5] CMS Collaboration, “Description and performance of track and primary- vertex reconstruction with the CMS tracker”, JINST 9 (2014) P10009, doi:10.1088/1748-0221/9/10/P10009, arXiv:1405.6569.
[6] CMS Collaboration, “Particle-flow reconstruction and global event description with the CMS detector”, JINST 12 (2017) P10003, doi:10.1088/1748-0221/12/10/P10003, arXiv:1706.04965.
[7] M. Cacciari, G. P. Salam, and G. Soyez, “The catchment area of jets”, JHEP 04 (2008) 005, doi:10.1088/1126-6708/2008/04/005, arXiv:0802.1188.
[8] M. Cacciari and G. P. Salam, “Pileup subtraction using jet areas”, Phys. Lett. B 659 (2008) 119, doi:10.1016/j.physletb.2007.09.077, arXiv:0707.1378.
[9] CMS Collaboration, “Jet energy scale and resolution in the CMS experiment in pp collisions at 8 TeV”, JINST 12 (2017) P02014, doi:10.1088/1748-0221/12/02/P02014, arXiv:1607.03663.
[10] A. J. Larkoski, S. Marzani, G. Soyez et al., “Soft Drop”, JHEP 1405 (2014) 146.
[11] J. Thaler and K. Van Tilburg, “Identifying Boosted Objects with N- subjettiness”, JHEP 03 (2011) 015, doi:10.1007/JHEP03(2011)015, arXiv:1011.2268.
[12] CMS Collaboration, "Performance of missing transverse momentum reconstruction in proton-proton collisions at s = 13 TeV using the CMS detector", JINST 14 (2019) P07004, doi: 10.1088/1748-0221/14/07/P07004, arXiv: 1903.06078."
Vertices [5] are reconstructed from charged-particle tracks. The physics objects considered for selecting the primary pp interaction vertex are track jets, clustered using the anti-k_{T} algorithm [3, 4] with the tracks assigned to the vertex as inputs, and the associated negative vector sum of those jets. The vertices are sorted by the value of summed physics-object p_{T}. The reconstructed vertex with the largest value of summed physics-object p_{T} is selected as the leading vertex (LV). Other reconstructed collision vertices are referred to as pileup (PU) vertices.
Pileup (PU) At the LHC multiple pp collisions happen at the same bunch crossing. All collisions that happen in addition to the LV collision are known as pileup (PU).
PUPPI The pileup per particle identification algorithm [1,2] proposes a method for pileup mitigation by assigning a weight to every particle depending on its probability to originate from a leading or pileup vertex.
Mean number of interactions per crossing (abbreviated "number of interactions" and denoted μ ) is used to estimate the amount of PU in simulation.
Number of vertices (denoted N_{Vertices}) are reconstructed through track clustering using a deterministic annealing algorithm [5]. This number can be determined in both data and simulation. Please note that the difference between number of interactions and number of vertices differ by ~30% which is due to the reconstruction efficiency. A detailed explanation can be found in Ref. [1].
CHS Charged Hadron Subtraction removes charged PF [6] candidates that are used in the fit of one of the PU vertices. More information on the reconstruction of the vertices can be found in Ref. [5].
PF candidates Particle-flow algorithm reconstructs and identifies each individual particle in an event, with an optimised combination of all subdetector information. More information can be found in Ref. [6].
Particle-level jet A jet clustered from stable (lifetime c𝜏 > 1 cm) particles, excluding neutrinos, before any detector simulation. Reconstruction-level jet A jet clustered from reconstructed PF candidates
AK4 jets/AK8 jets Jets are clustered from PF candidates using the anti-k_{T} algorithm [3] with the FASTJET software package [4]. A distance parameter of 0.4 (0.8) is used. Before the clustering one of the PU mitigation techniques is applied to the PF candidates. To jets clustered from PF candidates with CHS applied (CHS jets) an event-by-event jet-area-based correction [7-9] is applied to subtract the neutral PU component. Jet energy corrections are applied such that the measured response of the jets is on average one.
Response is the ratio of reconstruction-level jet p_{T} to particle-level jet p_{T}, while the jets are matched with ΔR < 0.2.
Soft drop algorithm A grooming algorithm applied to jets in order to remove soft and collinear radiation [10]. The parameters β = 0, z_{cut} = 0.1 are used. The jet mass of a groomed jet is referred to as soft drop mass MSD. In the rest of the document, for simplicity, the label jet mass refers to the M_{SD}.
N-subjettiness (τ_{N<\sub>):} A substructure variable related to the probability of a jet consisting of up to N subjets [11]. The ratios of τ_{21} = τ_{2}/τ_{1} are used to discriminate jets originating from W bosons against QCD multijet.
PFMET The module of the negative vector sum of all PF candidates in an event. PF candidates clustered in AK4 CHS jets with p_{T￼}>15 GeV are replaced by the jet p_{T} obtained after jet energy corrections are applied. The sum includes PF candidates associated to PU vertices.
PUPPI MET The module of the negative vector sum of all PF candidates weighted by their PUPPI weight in an event. PF candidates clustered in AK4 PUPPI jets with p_{T}>15 GeV are replaced by the jet p_{T} obtained after jet energy corrections are applied.
Hadronic recoil ( u ) in Z+jets events Defined as − p_{T}^{miss} − p_{T}(Z) [1,12]. The recoil can be projected into a parallel (u_{∥}) and a perpendicular (u_{⊥}) component to the Z boson p_{T}.
Hadronic recoil response Mean of a Gaussian function fitted to the −u_{∥}/p_{T}(Z) distribution.
Hadronic recoil resolution Width of a Gaussian function fitted to the −(u_{∥} + p_{T}(Z)) distribution, after correcting the recoil to reach a unity response.
I | Attachment | History | Action | Size | Date | Who | Comment |
---|---|---|---|---|---|---|---|
Fig1a.pdf | r1 | manage | 61.1 K | 2021-02-17 - 22:31 | LaurentThomas | ||
png | Fig1a.png | r1 | manage | 61.8 K | 2021-02-17 - 22:31 | LaurentThomas | |
Fig1b.pdf | r1 | manage | 61.5 K | 2021-02-17 - 22:31 | LaurentThomas | ||
png | Fig1b.png | r1 | manage | 65.2 K | 2021-02-17 - 22:31 | LaurentThomas | |
Fig1c.pdf | r1 | manage | 61.6 K | 2021-02-17 - 22:31 | LaurentThomas | ||
png | Fig1c.png | r1 | manage | 66.2 K | 2021-02-17 - 22:31 | LaurentThomas | |
Fig1d.pdf | r1 | manage | 60.5 K | 2021-02-17 - 22:31 | LaurentThomas | ||
png | Fig1d.png | r1 | manage | 65.9 K | 2021-02-17 - 22:31 | LaurentThomas | |
Fig1e.pdf | r1 | manage | 60.2 K | 2021-02-17 - 22:31 | LaurentThomas | ||
png | Fig1e.png | r1 | manage | 64.3 K | 2021-02-17 - 22:31 | LaurentThomas | |
Fig1f.pdf | r1 | manage | 59.6 K | 2021-02-17 - 22:31 | LaurentThomas | ||
png | Fig1f.png | r1 | manage | 62.1 K | 2021-02-17 - 22:31 | LaurentThomas | |
Fig2a.pdf | r1 | manage | 54.1 K | 2021-02-17 - 22:30 | LaurentThomas | ||
png | Fig2a.png | r1 | manage | 108.0 K | 2021-02-17 - 22:30 | LaurentThomas | |
Fig2b.pdf | r1 | manage | 53.8 K | 2021-02-17 - 22:30 | LaurentThomas | ||
png | Fig2b.png | r1 | manage | 108.4 K | 2021-02-17 - 22:30 | LaurentThomas | |
Fig3a.pdf | r1 | manage | 54.0 K | 2021-02-17 - 22:30 | LaurentThomas | ||
png | Fig3a.png | r1 | manage | 107.6 K | 2021-02-17 - 22:30 | LaurentThomas | |
Fig3b.pdf | r1 | manage | 54.0 K | 2021-02-17 - 22:30 | LaurentThomas | ||
png | Fig3b.png | r1 | manage | 104.7 K | 2021-02-17 - 22:30 | LaurentThomas | |
Fig4a.pdf | r1 | manage | 55.4 K | 2021-02-17 - 22:30 | LaurentThomas | ||
png | Fig4a.png | r1 | manage | 47.3 K | 2021-02-17 - 22:30 | LaurentThomas | |
Fig4b.pdf | r1 | manage | 55.6 K | 2021-02-17 - 22:30 | LaurentThomas | ||
png | Fig4b.png | r1 | manage | 48.9 K | 2021-02-17 - 22:30 | LaurentThomas | |
Fig4c.pdf | r1 | manage | 55.5 K | 2021-02-17 - 22:30 | LaurentThomas | ||
png | Fig4c.png | r1 | manage | 49.0 K | 2021-02-17 - 22:30 | LaurentThomas | |
Fig4d.pdf | r1 | manage | 55.0 K | 2021-02-17 - 22:30 | LaurentThomas | ||
png | Fig4d.png | r1 | manage | 44.8 K | 2021-02-17 - 22:30 | LaurentThomas | |
Fig4e.pdf | r1 | manage | 54.9 K | 2021-02-17 - 22:30 | LaurentThomas | ||
png | Fig4e.png | r1 | manage | 44.6 K | 2021-02-17 - 22:30 | LaurentThomas | |
Fig4f.pdf | r1 | manage | 55.0 K | 2021-02-17 - 22:30 | LaurentThomas | ||
png | Fig4f.png | r1 | manage | 44.5 K | 2021-02-17 - 22:30 | LaurentThomas | |
Fig5a.pdf | r1 | manage | 62.5 K | 2021-02-17 - 22:29 | LaurentThomas | ||
png | Fig5a.png | r1 | manage | 112.5 K | 2021-02-17 - 22:29 | LaurentThomas | |
Fig5b.pdf | r1 | manage | 62.1 K | 2021-02-17 - 22:29 | LaurentThomas | ||
png | Fig5b.png | r1 | manage | 107.9 K | 2021-02-17 - 22:29 | LaurentThomas | |
Fig6a.pdf | r1 | manage | 33.4 K | 2021-02-17 - 22:29 | LaurentThomas | ||
png | Fig6a.png | r1 | manage | 106.6 K | 2021-02-17 - 22:29 | LaurentThomas | |
Fig6b.pdf | r1 | manage | 33.4 K | 2021-02-17 - 22:29 | LaurentThomas | ||
png | Fig6b.png | r1 | manage | 102.2 K | 2021-02-17 - 22:29 | LaurentThomas | |
Fig7.pdf | r1 | manage | 55.1 K | 2021-02-17 - 22:29 | LaurentThomas | ||
png | Fig7.png | r1 | manage | 109.5 K | 2021-02-17 - 22:29 | LaurentThomas | |
Fig8a.pdf | r1 | manage | 60.8 K | 2021-02-17 - 22:28 | LaurentThomas | ||
png | Fig8a.png | r1 | manage | 92.3 K | 2021-02-17 - 22:28 | LaurentThomas | |
Fig8b.pdf | r1 | manage | 56.9 K | 2021-02-17 - 22:28 | LaurentThomas | ||
png | Fig8b.png | r1 | manage | 90.5 K | 2021-02-17 - 22:28 | LaurentThomas | |
Fig9.pdf | r1 | manage | 68.4 K | 2021-02-17 - 22:28 | LaurentThomas | ||
png | Fig9.png | r1 | manage | 118.1 K | 2021-02-17 - 22:28 | LaurentThomas | |
png | PUPPI.png | r1 | manage | 247.3 K | 2021-02-17 - 22:28 | LaurentThomas | |
png | PUPPIv151.png | r1 | manage | 282.0 K | 2021-02-17 - 22:28 | LaurentThomas | |
png | PUPPIv152.png | r1 | manage | 258.8 K | 2021-02-17 - 22:28 | LaurentThomas | |
png | PUPPIv153.png | r1 | manage | 262.4 K | 2021-02-17 - 22:28 | LaurentThomas |