LNL Annual Report Accelerator Operation and Development

showed a 200 µm longitudinal bending (banana shape). Moreover the module 2 presented an important displacement in one transversal direction that entailed the opening of eight more tuner holes for frequency recovery. Relaxation of stress induced during machining was identifi ed as the main effect in comparison with the deformations due to the brazing heat treatment.

Consequently machining procedure changed. For the remaining modules a full annealing at 600 °C preceded fi nishing. In addition fi nishing was as soft as possible in order to minimize induced stress. The success of the new procedure was evident after the fi rst brazing of RFQ3 (Dec. 2007). Banana shape and transversal movements generated by brazing heat treatment turned out to maintain nominal tolerances (±20 µm). Almost the same results were obtained after fi rst brazing of RFQ4 (May 2008), RFQ5 (Jul. 2008) and RFQ6 (Oct. 2008) (see fi gure 1).

Some minor problems characterized the second brazing of RFQ3 (Jul. 2008), the fi rst brazing of RFQ4 and RFQ6 and the preparation for the second brazing of RFQ4.

RFQ3 experienced a vacuum leak after vertical brazing. Leak was caused by a bad nickelization process of the head SS fl anges. A repairing brazing was attempted before sum-mer without success. This forced to remove the head fl anges and to repeat the vertical brazing with new nickelized fl anges (Feb. 2009). The new repairing brazing was partially suc-cessful. The brazing of one of the head fl anges was perfect, but a vacuum leak was found in the other one. The problem was identifi ed in the graphite supporting the module inside the brazing furnace. To solve the problem, it is necessary to re-machine the fl ange and to re-braze in a better confi gura-tion (with ceramic spacers instead of graphite plate).

Regarding RFQ4, the water channel plugs experienced some problem of brazing alloy fl ow. Plugs with not visible brazing alloy were removed and will be re-brazed in the ver-tical brazing. Moreover during preparation of the module for the second brazing, the SS collars of the vacuum ports were measured out of tolerances. New collars were machined and sent to CERN for thermal treatment. The treatment revealed an error in the material choice. Material used for four col-lars was AISI321 (with Titanium addition) instead of LN316.

INTRODUCTION

The high intensity RFQ under construction at LNL developed within TRASCO project [1] for ADS study, will be devoted to the application of Boron Neutron Capture Therapy (BNCT) for the treatment of skin melanoma [2].

The RFQ consists of three electromagnetic segments 2.4 meters long resonantly coupled via two coupling cells in order to reduce sensitivity to machining errors. Each segment consists of two 1.2 meters long modules, which are the basic construction units.

In order to accelerate the 30 mA CW proton beam up to 5 MeV, the maximum amount of power expected to be delivered to the RFQ is about 995 kW (148 kW beam power + 847 kW dissipated power). The possibility of a ten percent voltage increase respect to the nominal is considered in the dissipated power value. The power is generated by one klystron and supplied to the RFQ by a WR2300 waveguide system. Due to the fact that each RF coupler is rated for a maximum of 140 kW [3], the RF power is split in eight ways via magic-Tee dividers.

MECHANICAL FABRICATION AND BRAZING

Each module, built in Oxygen Free Electronic (OFE) copper, is made of four main parts. The head fl anges between modules and the rectangular vacuum fl anges are made of SS (LN316). To reduce the number of brazing joints, the longitudinal cooling passages are deep-hole drilled from one side and closed with brazed plugs on the fl at surfaces of the RFQ module (opposite to the coupling or end cells). Moreover, the vacuum grids with their cooling channels are directly machined on the copper bulk.

Two brazing steps occur. In the fi rst the four main parts are brazed in horizontal position in a horizontal vacuum furnace, as well as the OFE plugs for the cooling channels. The brazing alloy is Ag68CuPd-807/810 (57.3% at., 38,4% at., 4.3% at.) and the brazing temperature is 815 ºC. After fi rst brazing, housings for the head fl anges and the fl at end surfaces (where the cooling channel plugs are located) are machined. In the second brazing cycle the head SS fl anges, the inlet and outlet cooling water SS tubes, the SS supports for vacuum ports fl anges and the SS fl anges for couplers are brazed in vertical position in a vertical vacuum furnace. The brazing alloy is Ag72Cu-779 (40% at.) and the brazing temperature is 785 ºC.

Maximum temperature of the OFE copper stabilization cycle before fi nishing turned out to be extremely important to avoid deformation after brazing. For the fi rst two modules this temperature was 250 ºC. After brazing both the modules

Fabrication and Testing of SPES-BNCT RFQ

E. Fagotti1,2, M. Comunian2, F. Grespan3,4, S.J. Mathot4, A. Palmieri2, A. Pisent2, C. Roncolato2

1 Consorzio RFX Padova, Italy, 2 INFN Laboratori Nazionali di Legnaro, Italy, 3 Univ. degli Studi di Milano, Italy, 4 CERN, Switzerland

FIG. 1. RFQ3,4,5,6 in different phases of the brazing procedure.

213

LNL Annual Report Accelerator Operation and Development

This was unacceptable for the brazing process and the four collars were re-machined.

RFQ6 experienced an important longitudinal movement (about 170 microns) and a not expected transversal move-ment (about 60 microns). Fortunately, transversal movement maintained the distance between lateral electrodes and so, no frequency shift was found with rf measurements. The reason for these displacements is still under investigation. Accord-ing to the schedule, RFQ5 and RFQ6 will be brazed before the end of April.

RF MEASUREMENTS ON THE 1ST RFQ SEGMENT (RFQ1+RFQ2)

The fi rst RFQ segment was mounted at LNL for low power RF and vacuum tests (fi gure 2).

In order to allow a complete RF characterization, the segment was equipped on low and high energy sides with two aluminum Tunable End Cells (fi gure 3) equipped with variable Dipole Stabilizing Rods (DSR).

Prior to the vertical assembling of the two modules, the beam axis of each RFQ was characterized, and the neces-sary reference planes and pins were machined.

FIG. 2. The 1st RFQ segment mounted at LNL.

FIG. 3. The Tunable End Cell.

FIG. 4. Perturbative quadrupole (black) and dipole (red and green) components before tuning.

FIG. 5. Perturbative quadrupole (black) and dipole (red and green) components after tuning.

[1] SPES Technical Design Report, Eds. G. Prete and A. Covello INFN-LNL 223 (2008).[2] A. Pisent, et al., Progress on the accelerator based SPES-BNCT project at INFN Legnaro. J. Phys.: Conf. Ser. 41, 391-399, doi: 10.1088/1742-6596/41/1/043.[3] A. Palmieri, et al., LINAC02, Gyeonju, Rep. of Korea, p.722.

The RF measurements were performed in steps: fi rst the optimum length and transverse position of DSRs that guaranteed the maximum dipole-free region (about ±7 MHz) around the operating mode were determined: the optimum DSR length is equal to 145 mm. Then, the boundary conditions for the TE210 mode were tuned by inserting the octagon in the RFQ volume.

Finally the bead pull measurements and the tuning operations (based on an on-purpose developed algorithms making use on the RFQ modal expansion) were performed. The effectiveness of the tuning algorithms is shown in the following two fi gures, where the voltage perturbations due to upper quadrupole modes and the two kinds of dipole modes U(z)/U

0 and U

qd1,qd2(z)/U

0 are shown before and after tuning

(fi gures 4 and 5).

214