Résumé

Fermilab : des préparatifs intensifs

Fin juillet, le premier faisceau de protons depuis avril 2012 a circulé dans l’injecteur principal de Fermilab, marquant la fin de la période d’interruption du complexe. Le long arrêt a permis d’apporter à tous les accélérateurs des modifications importantes visant à accroître l’intensité des faisceaux de protons pour augmenter le potentiel scientifique des expériences du laboratoire. Ainsi, l’accélérateur Cockcroft-Walton a été remplacé par un quadripôle radiofréquence, de nouvelles améliorations ont été apportées au système RF du Booster, et une reconfiguration de l’anneau de recyclage a été mise en œuvre, ce qui a nécessité de nouvelles lignes de faisceau d’injection et d’extraction.

When a beam of protons passed through Fermilab’s Main Injector at the end of July, it marked the first operation of the accelerator complex since April 2012. The intervening long shutdown had seen significant changes to all of the accelerators to increase the proton-beam intensity that they can deliver and so maximize the scientific reach of Fermilab’s experiments. In August, acceleration of protons to 120 GeV succeeded at the first attempt – a real accomplishment after all of the upgrades that were made – and in September the Main Injector was already delivering 250 kW of proton-beam power. The goal is to reach 700 kW in the next couple of years.

With the end of the Tevatron collider programme in 2011, Fermilab increased its focus on studying neutrinos and rare subatomic processes while continuing its active role in the CMS experiment at CERN (CERN Courier October 2011 p20). Accelerator-based neutrino experiments, in particular, require intense proton beams. In the spring of 2012, Fermilab’s accelerator complex produced the most intense high-energy beam of neutrinos in the world, delivering a peak power of 350 kW by routinely sending 3.8 × 1013 protons/pulse at 120 GeV every 2.067 s to the MINOS and MINERvA neutrino experiments. It also delivered 15 kW of beam power at 8 GeV, sending 4.4 × 1012 protons/pulse every 0.4 s to the MiniBooNE neutrino experiment.

Higher intensities

This level of beam intensity was pushing the capabilities of the Linac, the Booster and the Main Injector. During the shutdown, Fermilab reconfigured its accelerator complex (see figure 1) and upgraded its machines to prepare them for the new NOvA, MicroBooNE and LBNE experiments, which will demand more muon neutrinos. In addition, the planned Muon g-2 and Mu2e experiments will require proton beams for muon production. With the higher beam intensities it is important to reduce beam losses, so the recent accelerator upgrades have also greatly improved beam quality and mitigated beam losses.

Before the shutdown, four machines were involved in delivering protons for neutrino production: the Cockcroft–Walton pre-accelerator, the linear accelerator, the Booster accelerator and the Main Injector. During the past 15 years, the proton requests for the Linac and Booster have gone up by more than an order of magnitude – first in support of MiniBooNE, which received beam from the Booster, and then in support of MINOS, which received beam from the Main Injector. Past upgrades to the accelerator complex ensured that those requests were met. However, during the next 10 years another factor of three is required to meet the goals of the new neutrino experiments. The latest upgrades are a major step towards meeting these goals.

For the first 40 years of the laboratory’s existence, the initial stage of the Fermilab accelerator chain was a caesium-ion source and a Cockcroft–Walton accelerator, which produced a 750 keV H beam. In August 2012, these were replaced with a new ion source, a radiofrequency quadrupole (RFQ) and an Einzel lens. The RFQ accomplishes transverse focusing, bunching and acceleration in a single compact device, significantly smaller than the room-sized Cockcroft–Walton accelerator. Now the 750 keV beam is already bunched, which improves capture in the following Linac (a drift-tube linear accelerator). The Einzel lens is used as a beam chopper: the transmission of the ions can be turned on and off by varying the voltage on the lens. Since the ion source and RFQ are a continuous-wave system, beam chopping is important to develop notches in the beam to allow for the rise times of the Booster extraction kicker. Chopping at the lowest possible energy minimizes the power loss in other areas of the complex.

The Booster, which receives 400 MeV H ions from the Linac, uses charge-exchange injection to strip the electrons from the ions and maximize beam current. It then accelerates the protons to 8 GeV. For the first 30 years of Booster operation, the demand for proton pulses was often less than 1 Hz and never higher than about 2 Hz. With the advent of MiniBooNE in 2002 and MINOS in 2005, demand for protons rose dramatically. As figure 2 shows, in 2003 – the first year of full MiniBooNE operation – 1.6 × 1020 protons travelled through the Booster. This number was greater than the total for the previous 10 years.

Booster upgrades

A series of upgrades during the past 10 years enabled this factor of 10 increase in proton throughput. The upgrades improved both the physical infrastructure (e.g. cooling water and transformer power) and accelerator physics (aperture and orbit control).

While the Booster magnet systems resonate at 15 Hz – the maximum number of cycles the machine can deliver – many of the other systems have not had sufficient power or cooling to operate at this frequency. Previous upgrades have pushed the Booster’s performance to about 7.5 Hz but the goal of the current upgrades is to bring the 40-year-old Linac and Booster up to full 15 Hz operation.

Understanding the aperture, orbit, beam tune and beam losses is increasingly important as the beam frequency rises. Beam losses directly result in component activation, which makes maintenance and repair more difficult because of radiation exposure to workers. Upgrades to instrumentation (beam-position monitors and dampers), orbit control (new ramped multipole correctors) and loss control (collimation systems) have led to a decrease in total power loss of a factor of two, even with the factor of 10 increase in total beam throughput.

Two ongoing upgrades to the RF systems continued during the recent shutdown. One concerns the replacement of the 20 RF power systems, exchanging the vacuum-tube-based modulators and power amplifiers from the 1970s with a solid-state system. This upgrade was geared towards improving reliability and reducing maintenance. The solid-state designs have been in use in the Main Injector for 15 years and have proved to be reliable. The tube-based power amplifiers were mounted on the RF cavities in the Booster tunnel, a location that exposed maintenance technicians to radiation. The new systems reduce the number of components in the tunnel, therefore reducing radiation exposure and downtime because they can be serviced without entering the accelerator tunnel. The second upgrade is a refurbishment of the cavities, with a focus on the cooling and the ferrite tuners. As operations continue, the refurbishment is done serially so that the Booster always has a minimum number of operational RF cavities. Working on these 40-plus-year-old cavities that have been activated by radiation is a labour-intensive process.

The Main Injector and Recycler

The upgrades to the Main Injector and the reconfiguration of the Recycler storage ring have been driven by the NOvA experiment, which will explore the neutrino-mass hierarchy and investigate the possibility of CP violation in the neutrino sector (CERN Courier June 2013 p9). With the goal of 3.6 × 1021 protons on target and 14 kt of detector mass, a significant region of the phase space for these parameters can be explored. For the six-year duration of the experiment, this requires the Main Injector to deliver 6 × 1021 protons/year. The best previous operation was 3.25 × 1021 protons/year. A doubling of the integrated number of protons is required to meet the goals of the NOvA experiment.

In 2012, just before the shutdown, the Main Injector was delivering 3.8 × 1013 protons every 2.067 s to the target for the Neutrinos at the Main Injector (NuMI) facility. This intensity was accomplished by injecting nine batches at 8 GeV from the Booster into the Main Injector, ramping up the Main Injector magnets while accelerating the protons to 120 GeV, sending them to the NuMI target, and ramping the magnets back down to 8 GeV levels – then repeating the process. The injection process took 8/15 of a second (0.533 s) and the ramping up and down of the magnets took 1.533 s.

A key goal of the shutdown was to reduce the time of the injection process. To achieve this, Fermilab reconfigured the Recycler, which is an 8 GeV, permanent-magnet storage ring located in the same tunnel as the Main Injector. The machine has the same 3.3 km circumference as the Main Injector. During the Tevatron collider era, it was used for the storage and cooling of antiprotons, achieving a record accumulation of 5 × 1012 antiprotons with a lifetime in excess of 1000 hours.

In future, the Recycler will be used to slip-stack protons from the Booster and transfer them into the Main Injector. By filling the Recycler with 12 batches (4.9 × 1013 protons) from the Booster while the Main Injector is ramping, the injection time can be cut from 0.533 s to 11 μs. Once completed, the upgrades to the magnet power and RF systems will speed up the Main Injector cycle to 1.33 s – a vast improvement compared with the 2.067 s achieved before the shutdown. When the Booster is ready to operate at 15 Hz, the total beam power on target will be 700 kW.

To use the Recycler for slip-stacking required a reconfiguration of the accelerator complex. A new injection beamline from the Booster to the Recycler had to be built (figure 3), since previously the only way to get protons into the Recycler was via the Main Injector. In addition, a new extraction beamline from the Recycler to the Main Injector was needed, as the aperture of the previous line was designed for the transfer of low-emittance, low-intensity antiproton beams. New 53 MHz RF cavities for the Recycler were installed to capture the protons from the Booster, slip-stack them and then transfer them to the Main Injector. New instrumentation had to be installed and all of the devices for cooling antiproton beams – both stochastic and electron cooling systems – and for beam transfer had to be removed.

Figure 4 shows the new injection line from the Booster (figure 5) to the Recycler, together with the upgraded injection line to the Main Injector, the transfer line for the Booster Neutrino Beam programme, and the Main Injector and Recycler rings. During the shutdown, personnel removed more than 100 magnets, all of the stochastic cooling equipment, vacuum components from four transfer lines and antiproton-specific diagnostic equipment. More than 150 magnets, 4 RF cavities and about 500 m of beam pipe for the new transfer lines were installed. Approximately 300 km of cable was pulled to support upgraded beam-position monitoring systems, new vacuum installations, new kicker systems, other new instrumentation and new powered elements. Approximately 450 tonnes of material was moved in or out of the complex at the same time.

The NuMI target

To prepare for a 700 kW beam, the target station for the NuMI facility needed upgrades to handle the increased power. A new target design was developed and fabricated in collaboration with the Institute for High Energy Physics, Protvino, and the Rutherford Appleton Laboratory, UK. A new focusing horn was installed to steer higher-energy neutrinos to the NOvA experiment (figure 6). The horn features a thinner conductor to minimize ohmic heating at the increased pulsing rate. The water-cooling capacity for the target, the focusing horns and the beam absorber were also increased.

With the completion of the shutdown, commissioning of the accelerator complex is underway. Operations have begun using the Main Injector, achieving 250 kW on target for the NuMI beamline and delivering beam to the Fermilab Test Beam Facility. The reconfigured Recycler has circulated protons for the first time and work is underway towards full integration of the machine into Main Injector operations. The neutrino experiments are taking data and the SeaQuest experiment will receive proton beam soon. Intensity and beam power are inceasing in all of the machines and the full 700 kW beam power in the Main Injector should be accomplished in 2015.