BESIII Data Overview

• BESIII data is a curated collection of high-statistics experimental datasets from the BEPCII tau–charm collider, crucial for precision spectroscopy and new physics searches.
• The datasets feature record-breaking event yields from resonances like J/ψ and ψ(3686) and include detailed scans across energy ranges for diverse physics analyses.
• Advanced detector performance, including excellent momentum, timing, and energy resolutions, is supported by robust data acquisition and quality monitoring systems.

BESIII data refers to the experimental datasets produced by the BESIII detector at the Beijing Electron–Positron Collider II (BEPCII), a high-luminosity $e^+e^-$ collider optimized for the $\tau$–charm energy regime ($\sqrt{s} = 2.0$–$4.95$ GeV). BESIII data underpins a substantial physics program spanning charmonium spectroscopy, open charm, hyperons, light hadron structure, exotics (XYZ states), precision $\mathcal{R}$-value measurements, fragmentation-function phenomenology, baryon form factors, and searches for physics beyond the Standard Model. BESIII has accumulated the world's largest datasets at threshold and in the continuum for $J/\psi$, $\psi(3686)$, and open-charm production, coupled with state-of-the-art trigger, reconstruction, and data quality pipelines (Huang, 2012, Gao, 22 Sep 2025).

1. Data Samples: Yields, Luminosities, and Energy Coverage

BESIII data acquisition is partitioned into on-resonance, threshold, scan, and continuum samples. The most significant samples are:

Resonance / Process	$\sqrt{s}$ (GeV)	Sample Size	Notes
$J/\psi$	3.097	$10^{10}$ events	World-record, $\mathcal{L} \sim 3$ fb$^{-1}$ (Gao, 22 Sep 2025, Li et al., 8 Sep 2025)
$\psi(3686)$	3.686	$2.7 \times 10^9$ events	$\mathcal{L} \sim 3.8$ fb$^{-1}$
$\psi(3770)$ ($D\bar D$ th.)	3.773	$7.9$ fb$^{-1}$ ($\sim 10^{7-8}$ $D\bar D$ pairs)	Later: $+12.3$ fb$^{-1}$ (Gao, 22 Sep 2025)
$D_s D_s^*$ scan	4.128–4.226	$7.3$ fb$^{-1}$	$D_s$ open-charm physics
XYZ/Continuum	3.80–4.70	$>20$ fb$^{-1}$	Dense scan at Y(4S), $Y(4660)$ (Wu, 2020, Mezzadri, 2019)

Additional samples include a 14-point $\mathcal{R}$-scan (2.23–3.67 GeV, total $\mathcal{L}>1$ fb$^{-1}$), high-statistics $\psi(3770)$ datasets for charm, and dedicated baryon-pair scans (e.g., $p\bar p$, $\Lambda\bar\Lambda$, $\Lambda_c\bar\Lambda_c$) (Prasad, 2023, Morales, 2017). Run periods extend from 2008 to the present, with incremental luminosity and phase-space coverage expansion.

2. Detector, Data Acquisition, and Event Reconstruction

The BESIII detector is a cylindrical 4$\pi$ apparatus offering:

• Momentum resolution: $\delta p/p \approx 0.5\%$ at 1 GeV/$c$ (MDC).
• TOF resolution: 80 ps (barrel), 110 ps (endcap).
• EMC energy resolution: $2.5\%$ at 1 GeV (barrel), $|\cos\theta|<0.83$; $5\%$ at 100 MeV (endcap).
• Muon identification: $>90\%$ for $p>0.5$ GeV/$c$, spatial resolution $\sim 2$ cm.
• Overall charged track acceptance: $93\%$ of $4\pi$.

Event acceptance, tracking, and PID efficiencies reach $98\%$ for charged tracks, TOF-based $K/\pi$ separation $>2\sigma$ up to $\sim 1$ GeV/$c$, and photon-detection efficiency near $95\%$ above 100 MeV (Morales, 2017, Li et al., 8 Sep 2025, Huang, 2012).

Data acquisition employs a two-level trigger (L1 hardware, L3 software event filter), followed by event buffering, distributed raw data storage, and a scalable online Data Quality Monitoring (DQM) system (Sun et al., 2011). BESIII DQM samples events (up to 360 Hz), reconstructs them with the full offline BOSS chain (excluding KalFitAlg for reduced latency), and populates a suite of detector, subdetector, and physics monitoring histograms with automatic alarm thresholds. All data streams are archived in the ROOT format, with subsets available as open-source .txt exports for public analysis (Liao et al., 19 Sep 2025).

3. Data Processing, Open Datasets, and Format

BESIII data is processed through a calibrated, multi-stage workflow:

• Raw data: Acquired post-trigger (RTRAW files); digitizes subdetector readouts as ROOT objects.
• REC: Includes raw + reconstructed track/hit/shower/counter info.
• Analysis ($\mathbf{.root}$): Selected flat TTrees focusing on physical observables (kinematics, PID).
• Text export ($\mathbf{.txt}$): For open data, events are written as blocks containing all reconstructed objects' parameters (track helix, charge, $\chi^2$, TOF, EMC, MUC, hit-level ADCs) (Liao et al., 19 Sep 2025).

A hierarchical metadata and workflow is provided, from run/event headers to detector objects. Monte Carlo samples include full GEANT4-based simulation with standard decays and detailed digitization. MC normalization follows

$$\sigma = \frac{N}{\mathcal{L} \, \varepsilon}\,, \qquad \mathcal{L} = \frac{N}{\sigma \, \varepsilon}$$

with event weights and normalization constants given per channel.

Public data (ScienceDB ID: cstr.cn/31253.11.sciencedb.21486) covers flagged resonance energies ($J/\psi$, $\psi(3686)$, $D\bar D$) and select physics processes (e.g., $Z_c(3900)$ production, QED, cosmic-ray triggers), totaling $O(10^2)$ events per channel, primarily for methodology and outreach (Liao et al., 19 Sep 2025).

4. Analysis Methodology and Statistical Treatment

Analysis strategies are tailored to the topology and final state:

• Tagging/double tagging: Leveraged for $D$, $D_s$ at threshold to enable near-background-free absolute branching ratio determinations (Weidenkaff, 2017).
• Binned and unbinned maximum-likelihood fits: Applied for yields, mass, and width extraction of hadronic resonances, employing relativistic Breit–Wigner forms, interference models, and background PDFs.
• Cross-section normalization: $$\sigma_B (\sqrt{s}) = \frac{N_f}{\mathcal{L} \cdot \varepsilon_f \cdot (1+\delta) \cdot \prod_i \mathcal{B}_i}$$ where $N_f$ is the corrected signal yield, $\varepsilon_f$ is the total efficiency, $(1+\delta)$ is the radiative/ISR factor, and the $\mathcal{B}_i$ are daughter branching ratios (Wu, 2020).
• Systematic uncertainties: Dominated by luminosity ($1$–$1.5\%$), tracking/PID ($1$–$2\%$ per track), photon efficiency, fit models, and background parameterizations. Final state– or channel–specific variations are reported in the literature.
• Statistical methods: Both frequentist and likelihood-based approaches are used, with upper limits (90% C.L.) incorporating systematic errors as per individual analyses (Gao, 22 Sep 2025, Prasad, 2023).

5. Major Physics Programs and Data Exploitation

The comprehensive BESIII dataset enables:

1. Precision spectroscopy of charmonium and open charm:
   • Absolute measurements of $m(\eta_c)$, $m(h_c)$, decay widths, transition rates, and radiative line shapes (including $E_\gamma^7$ weight in hindered M1 transitions) (Dong, 2011, Huang, 2012).
   • Determination of decay constants $f_{D^+}$, $f_{D_s}$, semileptonic form factors $f_+(q^2)$, and CKM elements $|V_{cd}|$, $|V_{cs}|$ with double-tagging and LQCD validation (Weidenkaff, 2017, Li, 2012).
2. Light hadron and baryon spectroscopy:
   • Identification and characterization of glueball candidates ($f_0(1710)$), isospin-violating processes, light baryon exotics, and anomalous decays, employing partial-wave analysis on world-record $J/\psi$ samples (Liu, 2015).
3. Exotic and XYZ states:
   • Energy scans and ISR studies resolving $Y(4220), Y(4360), Y(4660)$, $X(3872)$, $Z_c(3900)$, and $Z_c(4020)$ structures with well-defined mass and width (Wu, 2020, Nerling, 2018, Mezzadri, 2019).
4. $\mathcal{R}$-value, fragmentation, and form-factor studies:
   • $\mathcal{R}$-scan with 2.6–3.0% uncertainty, crucial for $a_\mu$ and $\alpha(M_Z^2)$ (Prasad, 2023).
   • Global NNLO QCD fits for $\pi^0, K_S^0, \eta$ fragmentation including higher-twist corrections, exploiting BESIII's low-energy SIA points (Li et al., 2024).
   • Baryon electromagnetic form factor extractions for $p, \Lambda, \Lambda_c$ and SU(3) partners, covering near-threshold and high-$q^2$ physics (Morales, 2017).
5. Searches for BSM processes:
   • Large datasets for rare decay and forbidden processes ($K^0_S \to$ invisible, $D^0$ FCNC, LNV), as well as dark photon and axion-like particle searches (Gao, 22 Sep 2025).

6. Online/Offline Data Quality Monitoring and Performance Metrics

Real-time integrity of BESIII data is ensured by the DQM system (Sun et al., 2011):

• Distributed reconstruction and analysis client system, operating in parallel to DAQ.
• Sampling rate up to 360 Hz; representative for physics triggers (40–60 Hz for Bhabha, dimuon, hadron events).
• Pushes histograms and 2D event displays to OHP, OHD, and web dashboards.
• Performance: 2–5 min average histogram latency, $>$99% uptime, and rapid error detection (hardware, misconfigurations, timing).
• Derived metrics (MDC spatial/momentum resolution, TOF/EMC time/energy) are archived per run for calibration/trending.

User-defined plug-in algorithms enable per-channel or subdetector monitoring, event tagging, physics benchmark tracking, and statistical validation via $\chi^2$-tests and efficiency estimators.

7. Open Access, Reproducibility, and Usage Guidelines

Open BESIII event data, including both data and MC, are released under CC-BY license for public research and education (Liao et al., 19 Sep 2025). Analysis routines are based on ROOT/pyROOT, Python/NumPy/matplotlib, and C++ BOSS for complete chain access (IHEP firewall-limited). Detailed event parsing and analysis are documented, with example cuts ($\chi^2$/ndf, vertexing), PID likelihood requirements, invariant mass reconstruction, and MC efficiency correction protocols provided.

Users are required to cite Z.J. Li et al., Front. Phys. (Beijing) 19 (2024) no.6, 64201 (for open dataset use) and relevant BESIII collaboration works for physics results.

---

BESIII data, by virtue of its unmatched statistics, detector performance, and rigorous data management, forms the global reference in $\tau$–charm physics, hadron structure, fragmentation phenomena, and BSM searches (Gao, 22 Sep 2025, Wu, 2020, Sun et al., 2011, Li et al., 8 Sep 2025, Liao et al., 19 Sep 2025).