Novel multi-beam front end for LANSCE accelerator facility

The LANSCE accelerator facility has been in operation for 50 years performing important scientific support for national security. The unique feature of the LANSCE accelerator facility is multi-beam operation, delivering beams to five experimental areas. To reduce long-term operational risks and to realize future beam performance goals in support of the laboratory missions, we are developing a novel high-brightness Front End injector. The proposed injector includes two independent low-energy transports for H+ and H- beams merging beams at the entrance of a single RFQ. These beamlines also perform preliminary beam bunching before RFQ. The challenge of the present project is associated with the simultaneous acceleration of protons and H- ions with multiple beam flavors in a single RFQ, which has never been done before. The proposed injector must provide better than existing beam parameters while beam intensity is supposed to be increased by a factor of two and injection energy is reduced from 750 keV to 100 keV. The paper discusses the details of the design and presents injector parameters.


Introduction
LANSCE linear accelerator consists of a 201.25 MHz Drift Tube Linac (DTL) accelerating particles from 0.75 MeV to 100 MeV and 805 MHz Coupled-Coupled Linac (CCL), accelerating particles from 100 MeV to 800 MeV [1].This accelerator facility simultaneously delivers various beams to multiple targets.A 100 MeV proton beam is delivered to the Isotope Production Facility (IPF), while 800 MeV H -beams are distributed to four experimental areas: the Lujan Neutron Scattering Center, the Weapons Neutron Research facility (WNR), the Proton Radiography facility (pRad), and the Ultra-Cold Neutron facility (UCN).To reduce long-term operational risks and to realize future beam performance goals in support of the laboratory missions, we developed a novel Front End including a high-brightness Radio-Frequency Quadrupole (RFQ) based injector [2,3].The layout of the proposed injector is illustrated in Fig. 1 and the parameters are presented in Table 1.

Front-end configuration with multi-beam RFQ
The proposed novel RFQ-based injector for LANSCE operation includes two independent low-energy transports merging beams at the entrance of a single RFQ, which accelerates simultaneously both protons and H -ions with multiple flavors of the beams (see Fig. 1).The beamlines are designed to perform pre-bunching in front of the accelerator section with the subsequent simultaneous acceleration of two different beams in a single RFQ.Each beamline contains a chopper, pre-buncher, lowfrequency buncher (in the H -only), deflector, solenoids, and quadrupoles accompanied by emittance measurement stations to match the beam to key beamline parameters.The accelerator operates at a 120 Hz repetition rate.The time pattern of the accelerator is presented in Ref. [2].It includes 20 Hz of acceleration of the H -Lujan beam and 100 Hz of simultaneous acceleration of WNR/IPF beams.Beams delivered to the UCN and pRad facilities "steal" their cycles from the WNR beam.Various time structures of the beams are achieved with a combination of the slow-wave chopper and bunching cavities.While H -(WNR) beam exists only as a sequence of single bunches, other beams are trains of bunches with various time structures.The level of bunching of beams is determined by the requirement to combine high beam capture and low emittance growth in the RFQ [3].

RFQ shielding considerations
One of the design issues of the project is the estimation of shielding requirements for the injector.Beam losses in the RFQ can result in neutron and gamma radiation.The radiation dose rates from accelerator operations are determined by the average beam current.Simulation of the proposed RFQ [3] shows that the typical total beam losses are at the level of 10%-15% of injected beam current.Most of the losses happen at the beginning of RFQ acceleration at beam energy within 100 keV -1 MeV.Such losses do not contribute to neutron production, because the threshold of neutron production in 63 Cu is 2.167 MeV.
According to simulations, the transverse particle losses in the RFQ, which hit the RFQ electrodes at the final energy of 3 MeV, are at the level of 1 out of 10 5 modeled particles.The LANSCE accelerator facility is designed for operation with an average current 1.25 mA, with the possible doubling of beam current after upgrades to the 805 MHz RF power system for the CCL.Therefore, the maximal RFQ losses are estimated as 25 nA of 3 MeV protons incident on a copper target.Based on data from Ref. IOP Publishing doi:10.1088/1742-6596/2687/5/0520063 [4], the expected doses at this level of beam loss are 0.25 mrem/h (2.5 µSv/h) neutron and 0.0025 mrem/h (0.025 µSv/h) photon.Reference [4], however, does not provide copper target impurity data or distance from the target to the detectors.The contribution from electrons stripped from 3 MeV H -ions is expected to be negligible.Figure 2 illustrates the spectrum of ions that are not captured into RFQ acceleration and are propagating in the RFQ structure.Such particles are expected to be lost downstream of the RFQ, mostly in the Medium Energy Beam Transport, and in the DTL.For 3 MeV, normal 1 W/m losses in the MEBT are expected to result in dose rates, smaller than 5 mrem/h (<50 µSv/h).These levels are far below the 100 mrem/h (1 mSv/h) threshold above which LANL requirements would drive the need for fences or radiation interlocks [5].However, a full power beam loss in MEBT could result is dose rates in excess of 10 rem/h (0.1 Sv/h).Though local shielding and interlocked area radiation monitors could be adequate controls for consequences from abnormal operations, it is likely to put the RFQ and MEBT in the shielded DTL tunnel within the Personnel Access Control System (PACS) boundary [6].

Chopper requirements
The chopper system of the future Front End must provide the same time structure as that in the existing LANSCE facility.One of the main issues in the design of choppers is the minimization of chopper rise/fall time [7].It is required for reducing the beam loss due to insufficient chopping.Usually, it can be achieved by utilizing traveling wave structures driven by fast pulsers or using low capacitive loads driven by fast switching devices [8].The possible chopper design is a traveling-wave structure based on a meander line with a dielectric substrate, similar to those developed for fast choppers at SNS and CERN [9].The transient time in traveling wave structures is affected by the edge field of the chopper.Even the ideal slow-wave structure with instantaneously switching voltage has a transient time , where is the chopper aperture [7].The RFQ provides a strongly bunched beam at the exit.The time interval between bunches in the RF field at 201.25 MHz is around 5 ns.It is expected that the single WNR bunch will be accompanied by satellite WNR bunches formed between a sequence of IPF bunches (see Fig. 3).Satellite WNR bunches must be removed with a Medium Energy Beam Transport chopper.It is expected to be done using two chopper pulses: before and after the primary WNR bunch.The length of the chopper pulses is 5-10 ns, but the time interval of 5 ns between pulses is fixed to allow propagation of an undisturbed WNR bunch.The required time structure of the chopper can be achieved using a dual-helix beam deflector with 200-Ω strip lines [10], which use a custom pulse generator with less than 4-ns rise and fall times.The traveling-wave current structures developed for this energy range demonstrated 2-ns rise/fall times, or even below that [9].Table 2 contains the expected parameters of LEBT and MEBT slow-wave choppers [11].A significant challenge for this work is to develop fast voltage pulse generators for these choppers.

Drift Tube Linac
After the RFQ, beams are accelerated in the Drift Tube Linac from 3 MeV up to 100 MeV.The DTL structure in this conceptual design is selected to be the classical Alvarez structure, which is simple and efficient in this ion energy range, especially with permanent-magnet quadrupoles inside its drift tubes.We will consider efficient alternatives, e.g., H-mode structures [12,13] in our analysis of alternatives.The proposed DTL is divided into 6 tanks each with a length of 7.5 m, average accelerating gradient = 2.5 MV/m, and synchronous phase for all DTL tanks.Relatively short tanks of the new DTL, compared to that of our existing DTL tanks (~18…19 m), are required for sufficient separation of the fundamental RF mode of 201.25 MHz from higher order modes.Additional parameters are the aperture radius = 1 cm, outer diameter of drift tubes = 12 cm, and the ratio = 0.15 -0.31,where is the gap between drift tubes.The parameters of this DTL are summarized in Table 3.A few cell shapes along the DTL are illustrated in Fig. 4.
The value of the Kilpatrick field limit at 201.25 MHz is = 14.77MV/m [14].The ratio of the maximal surface electric field in these DTL tanks to the Kilpatrick field limit, varies from 1.5 in Tank 1 to 1.8 in Tank 6 [15].The ratios up to ~ 2 are considered safe for pulsed linacs with pulse lengths below 1 ms [14].The effective shunt impedance per unit length, , changes from well above 50 MΩ/m in Tank 1 to slightly above 30 MΩ/m in Tank 6.These values are noticeably higher than those in the existing LANSCE DTL, contributing to reduced RF power requirements for the new design.
The required RF power to each tank will be delivered by the recently upgraded 201.25 MHz RF power stations based on Diacrode ® power amplifier tubes at the LANSCE DTL [16].In the present DTL, the output of two power amplifiers is combined to drive each of the three long tanks.A smaller tetrode power amplifier drives the 5 MeV Tank 1.This system will be reconfigured so that the tetrode power station will drive the new RFQ and the six Diacrode ® amplifiers will separately drive each of the DTL tanks.Each of the Diacrode ® power amplifiers can provide 1.7 MW peak power at 14% RF duty factor.

Summary
The new multi-beam Front End for the LANSCE accelerator facility is proposed.The upgrade concept consists of two independent 100 keV beamlines for H + and H -beams merged in front of 3 MeV RFQ, µ s 14th International Particle Accelerator Conference Journal of Physics: Conference Series 2687 (2024) 052006

Figure 2 :
Figure 2: Energy spectrum of uncaptured particles in RFQ.

4 Figure 3 :
Figure 3: Time structure of WNR and IPF bunches after RFQ, and required MEBT chopper time pattern.

Figure 4 :
Figure 4: Electric field lines calculated by DTL fish in a few DTL cells.

Table 1 :
Parameters of the Proposed LANSCE Injector

Table 2 :
Parameters of LANSCE Injector Choppers

Table 3 :
Parameters of Novel LANSCE Drift Tube Linac