CEPC Research Report: Higgs Physics Analysis

The measurement potential of Higgs decay to an invisible channel on CEPC is investigated[9]. In SM, the Higgs boson can only decay invisibly via $H\to ZZ^{*}\to\nu\nu\nu\nu$ or dark matter, so any evidence of invisible Higgs decay that exceeds BR ($H\to inv$.) will immediately point to a phenomenon that is beyond the standard model (BSM). In this analysis, the upper limit of BR ($H\to inv$) is estimated for three channels, including two leptonic channels and one hadronic channel, under the assumption predicted by SM. With the SM ZH production rate, the upper limit of BR ($H\to inv$) could reach 0.24% at the 95% confidence level, as shown in Figure3. The impact of the BMR on the ZH ($Z\to qq,H\to inv$.) signal strength accuracy has been studied as well, and the obtained curve shows that the analysis result would be one of the indicators for the CEPC detector optimization study.

In the $H\to\tau\tau$ analysis, the signal strength measurement is studied[10]. Using the full simulation analysis, the CEPC is expected to measure the signal strength to a relative accuracy of 0.8%. Multiple analysis technologies and dedicated $\tau$ finding algorithms are developed. A precise reconstruction of the impact parameter is essential for the $\tau$ events identification, and the PFA oriented detector design and reconstruction are critical for this analysis. The physics requirement on the mass resolution of the Higgs boson with hadronic decay final states is also discussed, as shown in Figure 4. It shows that the CEPC baseline design and reconstruction fulfill the accuracy requirement of the $H\to\tau\tau$ signal strength.

Jet reconstruction is critical for the precision measurement of Higgs boson properties and the electroweak observables at the CEPC[11]. The jet energy and angular responses of benchmark 2- and 4-jet processes are analyzed with fully simulated samples with the CEPC baseline detector geometry. As shown in Figure 5, the relative resolution of 3.5% and 1% are observed on the jet energy and angular measurement for jets in the detector barrel region ($\left|\cos\theta\right|<0.6$) with energy greater than 60 GeV. Meanwhile, the jet energy/angular scale can be controlled within 0.5/0.01%. The differential dependences of the jet response on the jet direction and energy are extracted. The impact on the jet responses induced by different jet clustering algorithms and matching criteria are also analyzed, which yields a relative difference of up to 8%.

Identifying the leptons inside jets is critical for the measurements of Higgs boson and the flavor physics program at the CEPC[12]. Using the CEPC baseline software, the identification performance of leptons generated inside a jet is analyzed. The jet leptons are identified with typical efficiency and mis-identification rate of 98% and 1% for energy higher than 2 GeV, with the $Z\to bb$ process at 91.2 GeV center-of-mass energy. At the benchmark channel of the CEPC flavor program of $B_{c}\to\tau\nu$ with $\tau\to e\nu\nu$, the electrons are identified with inclusive efficiency times purity of 97%, providing sufficient signal selection for the physics potential study. Compared to the single leptons, these jet leptons identification efficiency degrades about 1-3%, and the mis-identification rate increases by less than 1% at the same working point, as shown in Figure 6. The dependence between the performance of the shower reconstruction and the lepton identification is quantified.

A global analysis is proposed to improve the precision of a set of branching fractions of same resonance significantly, by taking all the cross talks among all processes into account and benefiting from statistical character of the multinomial distribution or binnomial distribution in the simplest case. It was first used in $\tau$ decay branching fractions by the ALEPH collaboration [13] and a good demonstration can also be found in [14] by the BESIII collaboration. Usually only one signal and its background processes are studied in a branching fraction measurement, which is a typical classification problem of signal and background processes. If more than one decay modes from same resonance are measured simultaneously, taking the Higgs decay as an example, therefore two types of ”backgrounds” must be studied, one is the backgrounds from other Standard Model (SM) processes, the other is the contamination among all the decay modes of the Higgs (cross talks).

The main ingredients of the global analysis are determination of the efficiency matrices, matrix manipulation, and minimization of $\chi^{2}$. The first one could be accomplished with ML technique elegantly [15, 16]. The last two are easily to be realized with some mature and widely used software tools.

The improvement of the precision of the Higgs decay branching fractions brought by the global analysis approach could be explained with a few points. First, the global analysis makes use of the full covariance of all counts, which means more information is used to determine the fitting parameters. Second, it deploys the law of multinomial distribution in the analysis. Therefore, the statistical uncertainties are significantly smaller than those of individual analysis, because they are calculated with produced numbers of signals according to the multinomial distribution and then are transformed into the ones of observed signals. Third, the global analysis introduces an unitary constraint of the total branching fractions. In conclusion, the global analysis approach adapts more information from data, statistical principle, and prior of physics.

The preliminary results based on fast simulation show that the precision could be improved by a factor of 2 in average [17].