Operation and Maintenance Activities of the LNL Vacuum Furnace
for Heat Treatments and Brazing Tests of the IFMIF/EVEDA RFQ Components
C. Roncolato, F. Poletto, F. Stivanello, P. Favaron.
INFN, Laboratori Nazionali di Legnaro, Legnaro (Padova), Italy.
INTRODUCTION for air 6000 lt/s) capable of lowering the pressure down to
the high vacuum range 10-6÷10-7 mbar. The primary
Each one of the 18 modules of the RFQ for the pumping system has been replaced with a new one, be-
IFMIF/EVEDA project is composed of 4 electrodes made cause the old pump could not reach the specified ultimate
in OFHC copper that assembled form the RF cavity. In pressure. The new pumping system installed can now reach
addition there are 80 parts made in 316LN stainless steel as the decade of 10-3 mbar in less than one hour. The ultimate
end flanges, cooling channels connections, and side ports pressure measured after 2 hours is 1 10-3 mbar. As a result,
for tuners and vacuum pumping. All these mechanical on one hand the Cryo-Pump cross-over point is lowered
parts require strong, uniform, leakproof joints (vacuum and thus pumps can be regenerated later, on the other hand
tightness of 10-9 mbar•lt/s, and water circuits able to work the regeneration is carried out more efficiently.
up to 3 bar) to produce structurally sound assemblies.
Moreover, the RFQ cavity involves very narrow tolerances
on the vane tips relative positions respect to the beam axis
of ±50 mm. The brazed joints are not only expected to
preserve electrical conduction between the pieces on the
internal part of the cavity surface, but also to guarantee
thermal flow on the whole section. The large number of
connections to be done with high reproducibility, the
capability for precision production tolerances, the ability to
join dissimilar metal, are some of the reasons for choosing
the brazing technique. A brazed joint is strong and
generally produces less thermally induced distortion,
because the entire part is brought up to the same brazing
temperature.
VACUUM MAINTENANCE
Fig. 1. Charts of the furnace pressure and temperature for an
When clean metals are heated to brazing temperature, annealing cycle.
their surface may form oxides if the atmosphere around the
part contains oxygen. Oxidized metal surfaces are usually Another maintenance activity regarded the two Cryo-
difficult to wet with most brazing filler metals. Brazing in Pumps system. The absorber cartridges for pumping light
vacuum atmosphere prevents the oxidation, and certain gases were changed and the first stage grid cleaned. The
oxides of base metal (chromium oxides of the stainless two Cryo-Compressors were checked and the helium filter
steel part) will dissociate if the dew point is kept low. It exchanged. Unfortunately after the maintenance one pump
can be seen that chromium oxides is stable at 500 °C when did not reach the base temperature of 12÷14 K at the
the water pressure is above 10-4 mbar [1]. Although at second stage. At first it was thought that temperature
higher temperature the water content can also be higher, reading was wrong, thus the low temperature thermometers
the oxide prevention should be prevented through all the of the Cryo-Pumps were replaced with new ones, while the
heating curve. Generally it is considered as safety value a old ones moved on the first stage. During this operation
total pressure of 5 10-5 mbar. one sensor turned out to be incorrectly cabled. After this
Before the brazing test, much work has been done on the operation the quickly temperature fluctuations, previously
vacuum system of the furnace. In facts the first run, made observed when the furnace temperature was rising, disap-
with small pieces of copper to test the annealing cycle, peared. Despite of this, the expected Cryo-Pump’s perfor-
showed a level of vacuum close to the specifications. mance could be reached only after the substitution of the
Besides, the pumps could not withstand long time cycles. Cryo-Head. It turned out that some small lead spheres
The vacuum system of the furnace has two stages: a which are normally inside the moving part escaped from
primary pumping system composed of a multi-claw dry the displacer and flattened on the inner surface of the
pump with a root stage for roughing the pressure up to cylinder, spoiling the helium gasket. As result the cryo-
10-3 mbar, and a couple of Cryo-Pumps (pumping speed genic power was reduced, because less amount of gas
LNL Annual Report 294 Laboratory Infrastructures
passed through the porous plug for the Joule Thompson Finally a small version of the RFQ prototype was brazed
cooling effect. and successfully leak tested by the end of 2010 [4].
CHEMICAL CLEANING
A second important factor for a reliable joint is the
surface preparation of the parts. Oil, dirt, and oxides must
be fully removed from the base and filler metals before
brazing, because only then can uniform capillary attraction
be obtained. Typically the copper pieces coming from the
machine shop were either dusty or shining. In facts when
the part is cut from the copper blocks with the Electric
Discharge Machine its surface is rough and covered with
residues of the cutting wire, while when the surface is Fig. 2. Photo of semimanufactured copper pieces loaded on
milled it is smooth and oily. In addition the copper molybdenum grids of the furnace before the heat treatment of
structure has some deep holes for the cooling channels that annealing.
are an additional source of grease and particles.
At the beginning the copper pieces were cleaned with Table 1. The list of the operations of 2010.
industrial detergents and afterwards passivated with citric
Date Description Program
acid. However, the dirty surface needed a more aggressive
13/11/09 Copper Annealing Test 800°C x 6 h
cleaning solution. Some pieces for brazing tests were
degreased and gently etched with Ammonium Persulfate. 04/12/09 Copper Annealing Mod. Prot. I 800°C x 6 h
In spite of the satisfactory results, this compound could not 27/01/10 SS316LN Annealing 980°C x 1 h
be used later since the larger pieces could be treated only 03/02/10 Baking of Furnace 300°C x 17 h
in stainless steel ultrasonic tanks that are also etched. 09/02/10 SS316LN Annealing 925°C x 2 h
A tolerable cleaning solution based on phosphoric acid 28/05/10 Baking of Furnace 1050°C x 4 h
was used. Firstly copper pieces stays in this bath at pH 1 at
09/06/10 Copper Annealing Mod. Prot. II 830°C x 6 h
75°C, then is rinsed with water, and afterwards dipped in a
18/06/10 Copper Annealing Mod. Prot. II 830°C x 6 h
passivating bath at pH 10 at the same temperature. Both
chemical treatments last 10-15 min. After a final rinsing 26/07/10 Brazing Test C-Pieces and Blocks 820°C x 1 h
with demineralised water, the piece is dried with nitrogen. 15/10/10 Tempering Test 1050°C x 8 h
19/10/10 Solution Treating Nim. 90 springs 1050°C x 8 h
OPERATIONS IN 2010 20/10/10 Aging Nimonic 90 springs 700°C x 16 h
22/10/10 Test of Nimonic 90 springs 860°C x 2 h
As shown in table 1 many runs were done before brazing 27/10/10 Brazing MiniPrototype 800°C x 3 h
of a small version of the RFQ. After a first run for testing
the annealing cycle, some copper and stainless steel parts
were treated. Annealing before brazing is required for FUTURE WORKS
relaxation of the mechanical stress induced during
The maintenance activities in 2010 concerned mainly the
mechanical working. These parts were assembled in the
vacuum system, but additional work should be carried out
first prototype RFQ module brazed at CERN [2].
as soon as possible. Firstly an absolute calibration of the
A long stop from March to June was done for the
thermocouples is mandatory for a fine control of the
installation of the primary pumping system and a complete
temperature during brazing. Secondly, the planarity of the
new set of 12 Type K thermocouple.
Molybdenum support should be increase to minimize the
Afterwards, additional annealing cycles were carried out
distortions during thermal cycle. Last but not least, the
for the copper pieces that compose the second prototype
PLC will be upgraded because obsolete; since there are no
RFQ model and for a couple of C-piece (400 mm in
available spare parts on the market even a minor problem
length) that were vertically brazed in the test of July with
could compromise the time schedule of the project.
the aid of some heat-resisting spring in Nimonic 90.
Although the positive results of the brazing, the springs
were found to be plastically deformed by creep. The
problem was solved by properly treating the alloy [3]. In [1] M.M. Schwartz, Brazing, ASM Int., Metals Park, Ohio
facts the Nimonic 90 can be aged for precipitation (USA), p.42.
hardening. A second Nickel phase can be segregate at the [2] A. Pepato et al., LINAC10 Conf., MOPEC060.
grain boundaries for hindering dislocation motion. Thus in [3] AA.VV., The Heat Treating Source Book, American Society
October, after a test of tempering to check the quench for Metals, Metals Park, Ohio (USA) 1986.
speed, the springs were treated and tested. [4] A. Pepato et al., this Annual Report.
LNL Annual Report 295 Laboratory Infrastructures