Belle II super-B factory experiment takes shape at KEK

Since CERN’s LHC switched on in the autumn of 2008, no new particle colliders have been built. SuperKEKB, under construction at the KEK laboratory in Tsukuba, Japan, is soon to change that. In contrast to the LHC, which is a proton–proton collider focused on producing the highest energies possible, SuperKEKB is an electron–positron collider that will operate at the intensity frontier to produce enormous quantities of B mesons.

At the intensity frontier, physicists search for signatures of new particles or processes by measuring rare or forbidden reactions, or finding deviations from Standard Model (SM) predictions. The “mass reach” for new-particle searches can be as high as 100 TeV/c², provided the couplings of the particles are large, which is well beyond the reach of direct searches at current colliders. The flavour sector provides a particularly powerful way to address the many deficiencies of the SM: at the cosmological scale, the puzzle of the baryon–antibaryon asymmetry remains unexplained by known sources of CP violation; the SM does not explain why there should be only three generations of elementary fermions or why there is an observed hierarchy in the fermion masses; the theory falls short on accounting for the small neutrino mass, and it is also not clear whether there is only a single Higgs boson.

SuperKEKB follows in the footsteps of its predecessor KEKB, which recorded more than 1000 fb^–1 (one inverse attobarn, ab^–1) of data and achieved a world record for instantaneous luminosity of 2.1 × 10³⁴ cm^–2 s^–1. The goals for SuperKEKB are even more ambitious. Its design luminosity is 8 × 10³⁵ cm^–2 s^–1, 40 times that of previous B-factory experiments, and the machine will operate in “factory” mode with the aim of recording an unprecedented data sample of 50 ab^–1.

The trillions of electron–positron collisions provided by SuperKEKB will be recorded by an upgraded detector called Belle II, which must be able to cope with the much larger beam-related backgrounds resulting from the high-luminosity environment. Belle II, which is the first “super-B factory” experiment, is designed to provide better or comparable performance to that of the previous Belle experiment at KEKB or BaBar at SLAC in Stanford, California. With the SM of weak interactions now well established, Belle II will focus on the search for new physics beyond the SM.

SuperKEKB was formally approved in October 2010, began construction in November 2011 and achieved its “first turns” in February this year (CERN Courier April 2016 p11). By the time of  completion of the initial accelerator commissioning before Belle-II roll-in (so-called “Phase 1”), the machine was storing a current of 1000 mA in its low-energy positron ring (LER) and 870 mA in the high-energy electron ring (HER). As currently scheduled, SuperKEKB will produce its first collisions in late 2017 (Phase 2), and the first physics run with the full detector in place will take place in late 2018 (Phase 3). The experiment will operate until the late 2020s.

B-physics background

The Belle experiment took data at the KEKB accelerator between 1999 and 2010. At roughly the same time, the BaBar experiment operated at SLAC’s PEP-II accelerator. In 2001, these two “B factories” established the first signals of CP violation, therefore revealing matter–antimatter asymmetries, in the B-meson sector. They also provided the experimental foundation for the 2008 Nobel Prize in Physics, which was awarded to theorists Makoto Kobayashi and Toshihide Maskawa for their explanation through complex phases in weak interactions.

In addition to the observation of large CP violation in the low-background “golden” B → J/ψ K_S-type decay modes, these B-factory experiments allowed many important measurements of weak interactions involving bottom and charm quarks as well as τ leptons. The B factories also discovered an unexpected crop of new strongly interacting particles known as the X, Y and Z states. Since 2008, a third major B factory, LHCb, entered the game. One of the four main LHC detectors, LHCb has made a large number of new measurements of B and B_s mesons and B baryons produced in proton–proton collisions. The experiment has tightly constrained new physics phases in the mixing-induced weak decays of B_s mesons, confirmed Belle’s discovery of the four-quark state Z(4430), and discovered the first two clear pentaquark states. Together with LHCb, Belle II is expected to be equally prolific and may discover signals of new physics in the coming decade.

Asymmetric collisions

The accelerator technology underpinning B factories is quite different from that of high-energy hadron colliders. For the coherent production of quantum-mechanically entangled pairs of B and B mesons, measurements of time-dependent CP asymmetries require that we know the difference in the decay times between the two B mesons. With equal energy beams, the B mesons travel only tens of microns from their production point and cannot experimentally be distinguished in silicon vertex detectors. To allow the B factory experiments to observe the time difference or spatial separation of the B vertices, the beams have asymmetric energies, and the centre of mass system is therefore boosted along the axis of the detector. For example, at PEP-II, 9 GeV electron and 3.1 GeV positron beams were used, while at KEKB the beam energies were 8 GeV and 3.5 GeV.

Charged particles within a beam undergo thermal motion just like gas molecules: they scatter to generate off-momentum particles at a rate given by the density and the temperature of the beam. Such off-momentum particles reduce the beam lifetime, increase beam sizes and generate detector background. To maximise the beam lifetime and reduce intra-beam scattering, SuperKEKB will collide 7 and 4 GeV electron and positron beams, respectively.

Two strategies were employed at the B factories to separate the incoming and outgoing beams: PEP-II used magnetic separation in a strong dipole magnet near the interaction point, while KEKB used a crossing angle of 22 mrad. SuperKEKB will extend the approach of KEKB with a crossing angle of 83 mrad, with separate beamlines for the two rings and no shared magnets between them. While the beam currents will be somewhat higher at SuperKEKB than they were at KEKB, the most dramatic improvement in luminosity is the result of very flat low-emittance “cool beams” and much stronger focusing at the interaction point. Specifically, SuperKEKB uses the nano-beam scheme inspired by the design of Italian accelerator physicist Pantaleo Raimondi, which promises to reduce the vertical beam size at the interaction point to around 50 nm – 20 times smaller than at KEKB.

Although the former TRISTAN (and KEKB) tunnels were reused for the SuperKEKB facility, many of the other accelerator components are new or upgraded from KEKB. For example, the 3 km-circumference vacuum chamber of the LER is new and is equipped with an antechamber and titanium-nitride coating to fight against the problem of photoelectrons. This process, in which low-energy electrons generated as photoelectrons or by ionisation of the residual gas in the beam pipe are attracted by the positively charged beam to form a cloud around the beam, was a scourge for the B factories and is also a major problem for the LHC. Many of the LER magnets are new, while a significant number of the HER magnets were rearranged to achieve a lower emittance, powered by newly designed high-precision power supplies at the ppm level. The RF system has been rearranged to double the beam current with a new digital-control system, and many beam diagnostics and control systems were rebuilt from scratch.

During Phase 1 commissioning, after many iterations the LER optics were corrected to achieve design emittance. To achieve low-emittance positron beams, a new damping ring has been constructed that will be brought into operation in 2017. To meet the charge and emittance requirements of SuperKEKB, the linac injector complex has been upgraded and includes a new low-emittance electron gun. Key components of the accelerator – including the beam pipe, superconducting magnets, beam feedback and diagnostics – were developed in collaboration with international partners in Italy (INFN Frascati), the US (BNL), and Russia (BINP), and further joint work, which will also involve CERN, is expected.

During Phase 1, intensive efforts were made to tune the machine to minimise the vertical emittances in both rings. This was done via measurements and corrections using orbit-response matrices. The estimated vertical emittances were below 10 pm in both rings, which is close to the design values. There were discrepancies, however, with the beam sizes measured by X-ray size monitors, especially in the HER, which is under investigation.

The early days of Belle and BaBar were plagued by problems, with beam-related backgrounds resulting from the then unprecedented beam currents and strong beam focusing. In the case of Belle, the first silicon vertex detector was destroyed by an unexpected synchrotron radiation “fan” produced by an electron beam passing through a steering magnet. Fortunately, the Belle team was able to build a new replacement detector quickly and move on to compete in the race with BaBar to measure CP asymmetries in the B sector. As a result of these past experiences, we have adopted a rather conservative commissioning strategy for the SuperKEKB/Belle-II facility. This year, during the earliest Phase 1 of operation, a special-purpose device called BEAST II consisting of seven types of background measurement devices was installed at the interaction point to characterise the expected Belle-II background.

At the beginning of next year, the Belle-II outer detector will be “rolled in” to the beamline and all components except the vertex detectors will be installed. The complex quadrupole superconducting final-focusing magnets are among the most challenging parts of the accelerator. In autumn 2017, the final-focusing magnets will be integrated with Belle II and the first runs of Phase 2 will commence. A new suite of background detectors will be installed, including a cartridge containing samples of the Belle-II vertex detectors. The first goal of the Phase-2 run is to achieve a luminosity above 10³⁴cm^–2 s^–1 and to verify that the backgrounds are low enough for the vertex detector to be installed.

Belle reborn

With Belle II expected to face beam-related backgrounds 20 times higher than at Belle, the detector has been reborn to achieve the experiment’s main physics goals – namely, to measure rare or forbidden decays of B and D mesons and the τ lepton with better accuracy and sensitivity than before. While Belle II reuses Belle’s spectrometer magnet, many state-of-the-art technologies have been included in the detector upgrade. A new vertex-detector system comprising a two-layer pixel detector (PXD) based on “DEPFET” technology and a four-layer double-sided silicon-strip detector (SVD) will be installed. With the beam-pipe radius of SuperKEKB having been reduced to 10 mm, the first PXD layer can be placed just 14 mm from the interaction point to improve the vertex resolution significantly. The outermost SVD layer is located at a larger radius than the equivalent system at Belle, resulting in higher reconstruction efficiency for K_s mesons, which is important for many CP-violation measurements.

A new central drift chamber (CDC) has been built with smaller cell sizes to be more robust against the higher level of beam background hits. The new CDC has a larger outer radius (1111.4 mm as opposed to 863 mm in Belle) and 56 compared to 50 measurement layers, resulting in improved momentum resolution. Combined with the vertex detectors, Belle II has improved D^* meson reconstruction and hence better full-reconstruction efficiency for B mesons, which often include D^*s among their weak-interaction decay products.

Because good particle identification is vital for successfully identifying rare processes in the presence of very large background (for example, the measurement of B → X_d γ must contend with B → X_s γ background processes that are an order-of-magnitude larger), two newly developed ring-imaging Cherenkov detectors have been introduced at Belle II. The first, the time-of-propagation (TOP) counter, is installed in the barrel region and consists of a finely polished and optically flat quartz radiator and an array of pixelated micro-channel-plate photomultiplier tubes that can measure the propagation time of internally reflected Cherenkov photons with a resolution of around 50 ps. The second, the aerogel ring-imaging Cherenkov counter (A-RICH), is located in Belle II’s forward endcap region and will detect Cherenkov photons produced in an aerogel radiator with hybrid avalanche photodiode sensors.

The electromagnetic calorimeter (ECL) reuses Belle’s thallium-doped cesium-iodide crystals. New waveform-sampling read-out electronics have been implemented to resolve overlapping signals such that π⁰ and γ reconstruction is not degraded, even in the high-background environment. The flux return of the Belle-II solenoid magnet, which surrounds the ECL, is instrumented to detect K_L mesons and muons (KLM). All of the endcap KLM layers and the innermost two layers of the barrel KLM were replaced with new scintillator-based detectors read out by solid-state photomultipliers. Signals from all of the Belle-II sub-detector components are read out through a common optical-data-transfer system and backend modules. GRID computing distributed over KEK-Asia-Australia-Europe-North America will be used to process the large data volumes produced at Belle II by high-luminosity collisions, which, like LHCb, are expected to be in the region of 1.8 GB/s.

Construction of the Belle-II experiment is in full swing, with fabrication and installation of sub-detectors progressing from the outer to the inner regions. A recent milestone was the completion of the TOP installation in June, while installation of the CDC, A-RICH and endcap ECL will follow soon. The Belle-II detector will be rolled into the SuperKEKB beamline in early 2017 and beam collisions will start later in the year, marking Phase 2. After verifying the background conditions in beam collisions, Phase 3 will see the installation of the vertex-detector system, after which the first physics run can begin towards the end of 2018.

Unique data set

As a next-generation B factory, Belle II will serve as our most powerful probe yet of new physics in the flavour sector, and may discover new strongly interacting particles such as tetraquarks, molecules or perhaps even hybrid mesons. Collisions at SuperKEKB will be tuned to centre-of-mass energies corresponding to the masses of the ϒ resonances, with most data to be collected at the Υ(4S) resonance. This is just above the threshold for producing quantum-correlated B-meson pairs with no fragmentation particles, which are optimal for measuring weak-interaction decays of B mesons.

SuperKEKB is both a super-B factory and a τ-charm factory: it will produce a total of 50 billion b b, c c and τ⁺ τ^– pairs over a period of eight years, and a team of more than 650 collaborators from 23 countries is already preparing to analyse this unique data set. The key open questions to be addressed include the search for new CP-violating phases in the quark sector, lepton-flavour violation and left–right asymmetries (see panel opposite).

Rare charged B decays to leptonic final states are the flagship measurements of the Belle-II research programme. The leptonic decay B → τν occurs in the SM via a W-annihilation diagram with an expected branching fraction of 0.82^+0.05_–0.03 × 10⁻⁴, which would be modified if a non-standard particle such as a charged Higgs interferes with the W. Since the final state contains multiple neutrinos, it is measurable only in an electron–positron collider experiment where the centre-of-mass energy is precisely known. Belle II should reach a precision of 3% on this measurement, and observe the channel B → μν for tests of lepton-flavour universality.

Perhaps the most interesting search at Belle II will be the analogous semi-leptonic decays, B → D^*τν and B → Dτν, which are similarly sensitive to charged Higgs bosons. Recently, the combined measurements of these processes from Babar, Belle and LHCb have pointed to a curious 4σ deviation of the decay rates compared to the SM prediction (see figure X). Since no such deviation is seen in B → τν, making it difficult to resolve the nature of the potential underlying new physics, the Belle-II data set will be required to settle the issue.

Another 4σ anomaly persists in B → K^* l⁺l^– flavour-changing neutral-current loop processes observed by LHCb, which may be explained by the actions of new gauge bosons. By allowing the study of closely related processes, Belle II will be able to confirm if this really is a sign of new physics and not an artifact of theoretical predictions. More precisely calculable inclusive transitions b → sγ and b → s l⁺l^– will be compared to the exclusive ones measured by LHCb. The ultimate data set will also give access to B → K^*νν and Kνν, which are experimentally challenging channels but also the most precise theoretically.

Beyond the Standard Model

There are many reasons to choose Belle II to address these and other puzzles with the SM, and in general the experiment will complement the physics reach of LHCb. The lower-background environment at Belle compared to LHCb allows researchers to reconstruct final states containing neutral particles, for instance, and to design efficient triggers for the analysis of τ particles. With asymmetric beam energies, the Lorentz boost of the electron–positron system is ideal for measurements of lifetimes, mixing parameters and CP violation.

The B factories established the existence of matter–antimatter asymmetries in the b-quark sector, in addition to the CP violation that was discovered 52 years earlier in the s-quark sector. The B factories established that a single irreducible complex phase in the weak interaction is sufficient to explain all CP-violating effects observed to date. This completed the SM description of the weak-interaction couplings of quarks.To move beyond this picture, two super-B factories were initially proposed: one at Tor Vegata near Frascati in Italy, and one at KEK in Japan. Although the former facility was not funded, there was a synergy and competition in the two designs. The super-B factory at KEK follows the legacy of the B factories, with Belle II and LHCb both vying to establish the first solid existence of new physics beyond the SM.