PIP-II’s international engagement is the secret of success

27 September 2021

Fermilab’s Proton Improvement Plan II (PIP-II) is relying on international collaborations to shape the future of accelerator-based particle physics in the US. Lia Merminga and Eduard Pozdeyev provide an insider take.

Rendering of PIP-II superconducting linac

The Proton Improvement Plan II (PIP-II) is an essential upgrade – and ambitious reimagining – of the Fermilab accelerator complex. An all-new, leading-edge superconducting linear accelerator, combined with a comprehensive overhaul of the laboratory’s existing circular accelerators, will deliver multimegawatt proton beam power and, in turn, enable the world’s most intense beam of neutrinos for the international Long Baseline Neutrino Facility (LBNF) and Deep Underground Neutrino Experiment (DUNE). While positioning Fermilab at the forefront of accelerator-based neutrino research, PIP-II will also provide the “engine room” for a diverse – and scalable – experimental programme in US particle physics for decades to come. Put simply, PIP-II will be the highest-energy and highest-power continuous-wave (CW) proton linac ever built, capable of delivering both pulsed and continuous particle beams.

Another unique aspect of PIP-II is that it is the first US Department of Energy (DOE)-funded particle accelerator that will be built with significant international participation. With major “in-kind” contributions from institutions in India, Italy, the UK, France and Poland, the project’s international partners bring wide-ranging expertise and know-how in core accelerator technologies along with an established track-record in big-physics initiatives. What’s more, PIP-II is not going to be the last DOE project to benefit from international collaboration – there will be more to come – so a near-term priority is to provide a successful template that others can follow. 

Deconstructing neutrino physics

Neutrino detector modules

Operationally, LBNF/DUNE is a global research endeavour comprising three main parts: the experiment itself (DUNE); the facility that produces the neutrino beam plus associated infrastructure to support the experiment (LBNF); and the PIP-II upgrade to the Fermilab accelerator complex, which will power the neutrino beam. 

At Fermilab, PIP-II will accelerate protons and smash them into an ultrapure graphite target. The resulting beam of neutrinos will travel through the DUNE near detector on the Fermilab site, then through 1300 km of earth (no tunnel required), and finally through the DUNE far detector at Sanford Lab in South Dakota (see figure). Data from neutrino interactions collected by the experiment’s detectors will be analysed by a network of more than 1000 DUNE collaborators around the world.

In this way, DUNE will enable a comprehensive programme of precision neutrino-oscillation measurements using νμ and νμ beams from Fermilab. Key areas of activity will include tests of leptonic charge-parity conservation; determining the neutrino mass ordering; measuring the angle θ23 in the Pontecorvo–Maki–Nakagawa–Sakata mixing matrix; and probing the three-neutrino paradigm. Furthermore, DUNE will search for proton decay in several decay modes and potentially detect and measure the νe flux from any supernovae that take place in our galaxy. 

To provide unprecedented detail in the reconstruction of neutrino events, the DUNE experiment will exploit liquid-argon time-projection-chamber (LArTPC) detectors on a massive scale (technology itself that was first deployed at scale in 2010 for the ICARUS detector as part of the CERN Neutrinos to Gran Sasso facility). The LArTPC implementation for DUNE is currently being developed in two prototype detectors at CERN via the CERN Neutrino Platform, an initiative inaugurated in 2014 following the recommendations of the 2013 European Strategy for Particle Physics to provide a focal point for Europe’s contributions to global neutrino research. 

In addition to the prototype DUNE detectors, the CERN Neutrino Platform is contributing to the long-baseline Tokai-to-Kamioka (T2K) and future Hyper-Kamiokande experiments in Japan. Construction of the underground caverns for DUNE and Hyper-Kamiokande is under way, with both experiments chasing similar physics goals and offering valuable scientific complementarity when they come online towards the end of the decade. 

A key driver of change was the recommendation of the 2014 US Particle Physics Project Prioritization Panel (P5) that the US host a world-leading international programme in neutrino physics. “Its centrepiece,” the P5 report asserts, “would be a next-generation long-baseline neutrino facility (LBNF). LBNF would combine a high-intensity neutrino beam and a large-volume precision detector [DUNE] sited underground a long distance away to make accurate measurements of the oscillated neutrino properties… A powerful, wideband neutrino beam would be realised with Fermilab’s PIP-II upgrade project, which provides very high intensities in the Fermilab accelerator complex.”

Fast forward to December 2020 and full DOE approval of the PIP-II baseline plan, at a total project cost of $978m and with completion scheduled for 2028. Initial site preparation actually started in March 2019, while construction of the cryoplant building got under way in July 2020. Commissioning of PIP-II is planned for the second half of this decade, with the first delivery of neutrino beam to LBNF/DUNE in the late 2020s (see “Deconstructing neutrino physics” panel). With the help of Fermilab’s network of international partners, a highly capable, state-of-the-art accelerator will soon be probing new frontiers in neutrino physics and, more broadly, redefining the roadmap for US high-energy physics.  

Then, now, next

If that’s the future, what of the back-story? Fermilab’s particle-accelerator complex originally powered the Tevatron, the first machine to break the TeV energy barrier and the world’s most powerful accelerator before CERN’s Large Hadron Collider (LHC) came online a decade ago. The Tevatron was shut down in 2011 after three illustrious decades at the forefront of particle physics, with notable high-points including discovery of the top quark in 1995 and direct discovery of the tau neutrino in 2000. 

A powerful, wideband neutrino beam would be realised with Fermilab’s PIP-II upgrade project

Today, about 4000 scientists from more than 50 countries rely on Fermilab’s accelerators, detectors and computing facilities to support their cutting-edge research. The laboratory comprises four interlinking accelerators and storage rings: a 400 MeV room-temperature linac; an 8 GeV Booster synchrotron; an 8 GeV fixed-energy storage ring called the Recycler; and a 60–120 GeV Main Injector synchrotron housed in the same tunnel with the Recycler. The Main Injector generates more than 800 kW of proton beam power, in turn yielding the world’s most intense beams of neutrinos for Fermilab’s flagship NOvA experiment (with the far detector located in Ash River, Minnesota), while supporting a multitude of other research programmes exploring fundamental particles and forces down to the smallest scales.

A leading-edge SRF proton linac

The roll-out of PIP-II will make the Fermilab complex more powerful again. Replacing the 50-year-old linear accelerator with a high-intensity, superconducting radio­frequency (SRF) linac will enable Fermilab to deliver 1.2 MW of proton beam power to the LBNF target, providing a platform for scale-up to multimegawatt levels and the capability for high-power operation across multiple particle-physics experiments simultaneously. 

Deconstructed, the PIP-II linac is an 800 MeV, 2 mA H machine consisting of a room-temperature front-end (up to 2.1 MeV) followed by an SRF section designed to operate in CW mode. The CW operation, and the requirements it places on the SRF systems, present some unprecedented challenges in terms of machine design. 

Precision engineering

The H source (capable of 15 mA beam current) is followed by a low-energy beam transport (LEBT) section and a radio­frequency quadrupole (RFQ) that operates at a frequency of 162.5 MHz and is capable of 10 mA CW operation. The RFQ bunches, focuses and accelerates the beam from 30 keV to 2.1 MeV. Subsequently, the PIP-II MEBT includes a bunch-by-bunch chopping system that removes undesired bunches of arbitrary patterns from the CW beam exiting the RFQ. This is one of several innovative features of the PIP-II linac design that enables not only direct injection into the Booster RF bucket – thereby mitigating beam losses at injection – but also delivery of tailored bunch patterns for other experiments. The chopper system itself comprises a pair of wideband kickers and a 20 kW beam absorber.

In terms of the beam physics, the H ions are non-relativistic at 2.1 MeV and their velocity changes rapidly with acceleration along the linac. To achieve efficient acceleration to 800 MeV, the PIP-II linac employs several families of accelerating cavities optimised for specific velocity regimes – i.e. five different types of SRF cavities at three RF frequencies. Although this arrangement ensures efficient acceleration, it also increases the technical complexity of the project owing to the unique challenges associated with the design, fabrication and commissioning of a portfolio of accelerating systems.

Mapped versus increasing energy, the PIP-II linac consists of a half-wave resonator (HWR) operating at 162.5 MHz at optimal beta-value of 0.112; two types of single-spoke resonators (SSR1, SSR2) at 325 MHz and optimal betas equal to 0.222 and 0.472, respectively; and two types of elliptical cavities with low and high beta at 650 MHz (LB650, HB650) and optimal betas equal to 0.65 and 0.971. The HWR cryomodule has been built by the DOE’s Argonne National Laboratory (Lemont, Illinois), while an SSR1 prototype cryomodule was constructed by Fermilab, with a cavity provided by India’s Department of Atomic Energy. Both cryomodules have now been tested successfully with beam by the PIP-II accelerator physics team. 

Innovation yields acceleration

Each of the five accelerating systems comes with unique technical challenges and requires dedicated development to validate performance requirements. In particular, the CW RF mode of operation necessitates SRF cavities with high-quality factors at high gradient, thereby minimising the cryogenic load. For the SSR2, LB650 and HB650 cavities, the Qo and accelerating gradient specifications are: 0.82 × 1010 and 11.4 MV/m; 2.4 × 1010 and 16.8 MV/m; 3.3 × 1010 and 18.7 MV/m, respectively – figures of merit that are all beyond the current state-of-the-art. Nitrogen doping will enable the elliptical cavities to reach this level of performance, while the SSR2 cavities will undergo a rotational-buffered chemical polishing treatment. 

PIP-II prioritises international partnerships

International engagement

PIP-II is the first DOE-funded particle accelerator to be built with significant international participation, leveraging in-kind contributions of equipment, personnel and expertise from a network of partners across six countries. It’s a similar working model to that favoured by European laboratories like CERN, the European X-ray Free Electron Laser (XFEL) and the European Spallation Source (ESS) – all of which have shared their experiences with Fermilab to inform the PIP-II partnership programme. 


Partners: Argonne National Laboratory; Fermilab (lead partner); Lawrence Berkeley National Laboratory; Thomas Jefferson National Accelerator Facility

Key inputs: HWR, RFQ and resonance control systems


Partners: Bhabha Atomic Research Centre (BARC); Inter-University Accelerator Centre (IUAC); Raja Ramanna Centre for Advanced Technology (RRCAT); Variable Energy Cyclotron Centre (VECC)

Key inputs: room-temperature and superconducting magnets, SRF cavities, cryomodules, RF amplifiers


Partner: Italian Institute for Nuclear Physics (INFN)

Key inputs: SRF cavities (LB650) 


Partner: Science and Technology Facilities Council as part of UK Research and Innovation (STFC UKRI)

Key inputs: SRF cryomodules (HB650)


Partners: French Alternative Energies and Atomic Energy Commission (CEA); French National Centre for Scientific Research/National Institute of Nuclear and Particle Physics (CNRS/IN2P3) 

Key inputs: cryomodules (LB650) and SRF cavity testing (SSR2)


Partners: Wrocław University of Science and Technology; Warsaw University of Technology; Lodz University of Technology

Key inputs: cryogenic distribution systems and high-performance electronics (e.g. low-level RF and RF protection instrumentation).

A further design challenge is to ensure that the cavity resonance is as narrow as possible – something that is necessary to minimise RF power requirements when operating in CW mode. However, a narrow-bandwidth cavity is prone to detuning owing to small acoustic disturbances (so-called microphonic noise), with adverse effects on the required phase, amplitude stability and ultimately RF power consumption. The maximum detuning requirement for PIP-II is 20 Hz – achieved via a mix of passive approaches (e.g. cryomodule design, decoupling cavities from sources of vibration and more rigid cavity design) and active intervention (e.g. adaptive detuning control algorithms). 

Another issue in the pulsed RF regime is Lorentz force cavity detuning, in which the thin walls of the SRF cavities are deformed by forces from electromagnetic fields inside the cavity. This phenomenon can be especially severe in the SSR2 and LB650 cavities – where detuning may be approximately 10 times larger than the cavity bandwidth – though initial operation of PIP-II in CW RF and pulsed beam mode will help to mitigate any detuning effects.

The management of risk 

Given the scale and complexity of the linac development programme, the Fermilab project team has constructed the PIP-II Injector Test facility (also known as PIP2IT) as a systems engineering testbed for PIP-II’s advanced technologies. Completed last year, PIP2IT is a near-full-scale prototype of the linac’s room-temperature front-end, which accelerates protons up to 2.1 MeV, and the first two PIP-II cryomodules (HWR and SSR1) that then take the beam up to about 20 MeV. 

The testbed is all about risk management: on the one hand, validating design choices and demonstrating that core enabling technologies will meet PIP-II performance goals in an operational setting; on the other, ensuring seamless integration of the in-kind contributions (including SRF cavities, magnets and RF amplifiers) from PIP-II’s network of international partners (see “PIP-II prioritises international partnerships”). Beam commissioning in PIP2IT was completed earlier this year, with notable successes versus a number of essential beam manipulations and technology validations including: the PIP-II design beam parameters; the bunch-by-bunch chopping pattern required for injection into the Booster; and acceleration of beam to 17.2 MeV in the first two PIP-II cryomodules. Significant progress was also registered with successful testing of the SRF/cryomodule technologies, first operation of the laser-wire profile monitor, and the application of machine-learning algorithms to align the orbit through the cryomodules. 

There’s no duplication of effort here either. Post-commissioning, after completion of full system and design validation, the PIP2IT accelerator will be disassembled, moved and reinstalled in the PIP-II facility as the SRF linac’s upstream front-end. The testbed location, meanwhile, is being transformed into the PIP-II Cryomodule Test Facility, where most of the cryomodules will be tested with full RF power before being installed in the tunnel. 

Notwithstanding construction of the new SRF linac, PIP-II also involves fundamental upgrades to Fermilab’s existing circular accelerators – the Booster, Recycler Ring and Main Injector – to enable the complex to achieve at least 1.2 MW of proton beam power while providing a scalable platform towards multi-MW capability. More specifically, the path to 1.2 MW from Fermilab’s Main Injector, over the energy range 60 to 120 GeV, requires a number of deliverables to come together: increase of the Fermilab Booster beam intensity by roughly 50% compared to current operation (i.e. an increase in the number of protons extracted per Booster cycle from 4.3 × 1012 to 6.3 × 1012); reduction of the Main Injector cycle from 1.33 to 1.2 s; and an increase of the Booster repetition rate from 15 to 20 Hz. 

PIP-II: a flexible, versatile design

If PIP-II is primarily about providing a platform for doubling the beam power from Fermilab’s Main Injector, the project is also designed with scalability in mind to enable future upgrades versus a broad spectrum of scientific opportunities. 

At 2 mA average beam current at 800 MeV, the PIP-II linac is capable of delivering up to 1.6 MW of CW beam power. The LBNF/DUNE experiment requires approximately 17 kW of that power budget – which constitutes 1% of the available beam power – though that requirement will scale to twice as much over the long term. The rest of the beam can be delivered to other user programmes by combining RF separators and fast switching magnets, yielding a high-power beam with flexible bunch patterns for diverse experiments simultaneously. 

One key outcome of PIP-II is an upgrade path for a 10-fold increase in beam power delivered to the Mu2e experiment. The resulting Mu2e-II project will measure how muons decay to electrons with the aim of observing forbidden processes that could point to physics beyond the Standard Model. PIP-II will also underpin a low-energy muon research programme, using not only its large beam power but also its extremely flexible bunch structure to support two different classes of slow-muon experiments – i.e. those involving continuous beams and those involving pulsed beams.

Along the energy coordinate, meanwhile, the PIP-II linac tunnel currently includes space and infrastructure for two more HB650 cryomodules – additions that would increase the beam energy above 1 GeV. There is considerable interest, for example, in using the excess protons at about 1 GeV that PIP-II could provide when operated in continuous mode. Coupling to a proton storage ring to drive a MW-class proton beam dump facility would support new lines of enquiry in high-energy physics, including the search for accelerator-produced dark matter. Extending the tunnel (and adding even more cryomodules) would offer an upgrade path to further increase the beam energy to roughly 2 GeV.

Versatility is a given. Although the PIP-II linac accelerates H ions, it’s worth noting that most of the accelerator components are also suitable for acceleration of protons without modifications or changes to their polarity or phase.

Right now, beam losses in the Booster – which occur during injection, transition and extraction – prevent the intensity increase and limit the performance of the accelerator complex to roughly 900 kW. The PIP-II SRF linac injection into the Booster mitigates high-intensity effects and reduces losses on two fronts: first, the higher injection energy (800 MeV vs 400 MeV) will mitigate space-charge forces at higher beam intensities; second, the high-quality, lower-emittance beam will allow “beam painting” at injection in all three degrees of freedom, further reducing space-charge forces and beam losses at high intensity. Other upgrades are also in the works to further reduce and control losses, with some of them to be made available early, several years before PIP-II commissioning, to benefit the NOvA experiment. 

In PIP-II, the 8 GeV Booster beam will be injected into the Fermilab Recycler ring – equipped with new 53 MHz RF cavities capable of larger beam current – where 12 Booster transfer batches are accumulated and slip-stacked. Next, the Recycler beam will enter Fermilab’s Main Injector – equipped with double the number of power amplifiers and vacuum tubes – which accelerates this intense beam anywhere from 60 to 120 GeV, delivering at least 1.2 MW of beam power at 120 GeV. Further, the Booster upgrade to 20 Hz will support an 8 GeV science programme, including Fermilab’s muon-to-electron conversion experiment (Mu2e) and studies of short-baseline neutrinos (see “PIP-II: a flexible, versatile design”). 

International collaboration  

Over the next decade, the PIP-II roadmap is clear. Phase one of the project will see the front-end of the Fermilab accelerator complex replaced with an 800 MeV SRF linac while performing necessary upgrades to the existing rings. Completion will see PIP-II deliver an initial beam power of 1.2 MW on the LBNF target, though the longer-term objective is to upgrade to 2.4 MW through replacement of the Booster synchrotron.

Operationally, its worth reiterating that PIP-II is very much a collective endeavour

Operationally, it’s worth reiterating that PIP-II is very much a collective endeavour – in fact, the first US accelerator to be built with the help of a network of international partnerships. In this way, PIP-II is very much a trail-blazer, with the excellence and sustained commitment of the project’s international partners essential for the construction – and ultimately the successful delivery – of this next-generation accelerator complex by the end of the decade. 

bright-rec iop pub iop-science physcis connect