Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
VERTICAL ELECTRO-POLISHING AT CEA SACLAY: COMMISSIONING OF A NEW SET-UP AND MODELING OF THE PROCESS APPLIED TO
DIFFERENT CAVITIES
F. Éozénou , S. Chel , Z. Wang , Y. Gasser , J-P. Poupeau , C. Servouin#a a ab a a a
aCEA-Saclay, DSM/Irfu/SACM-91191 Gif-Sur-Yvette – France bENSAM ParisTech – 75013 Paris – France
Abstract
Reproducible operation at high performances of sc cavities is required for superconducting linacs, at least in their high energy section in the case of accelerators for proton beams. High beta elliptical cavities are thus of concern, and, to achieve required performances for such cavities, surface preparation including electro-polishing (EP) is recommended.
Moreover, for large dimension cavities we consider the EP treatment in vertical configuration (abbreviated as VEP) to be more appropriate. For this reasons, a VEP set-up has been designed at Saclay for the electro-polishing of muticells elliptical cavities.
Chosen equipment will make it possible to use a wide range of parameters (voltage, flowrate, temperature, nitrogen inerting) with R&D purpose in mind. Optimization will be done using modeling with COMSOL software for different cavities. As examples, we present some results for the 704 MHz high-beta SPL cavity and the 1300 MHz ILC cavity and show the influence of cathode shape on both acid flow and electric field distribution during the process. At last, importance of the size of the cavity will be commented.
INTRODUCTION Electro-polishing (abbreviated as EP) is believed to be
the most desirable treatment for SRF cavities [1]. EP is an anodic electrochemical treatment carried out in concentrated hydrofluoric - sulphuric (HF-H2SO4) acids. Generally, the cavity is electro-polished following a process developed by KEK [1]: A voltage is applied between the cavity and a cylindrical aluminium cathode (set in its centre) while it is placed in horizontal position, rotating, and half-filled with the circulating acid. This process makes it possible to reach high gradients on 9cell Tesla shape cavities and has been chosen for the surface treatment of cavities for the XFEL linac. However, it induces some drawbacks: the rotating seals are more easily prone to leaks, the cavity must be switched full of acid for draining and the footprint of the set-up is rather large. Furthermore, the removal rate in the cavity depends
on the location of the cell [2]. To overcome these drawbacks, some laboratories are investigating an alternate process where the cavity is electro-polished in vertical position [3]. Cornell University has proved that VEP with static acid could make it possible to reach high gradients [4], and as VEP is suitable to the treatment of large elliptical cavities, we have developed at CEA-Saclay a VEP set-up sized for surface treatment of the largest high beta and high gradient cavity we have designed, that is to say a 704 MHz cavity for the Superconducting Proton Linac. All other “smaller” cavities, as ILC cavities, fit into the set-up too.
In this paper, this VEP set-up is described as well as results concerning the modeling of VEP for both high beta SPL [5] and ILC cavities. The acid flow and electrical field distribution expected during VEP will be discussed.
CHARACTERISTICS OF THE SET-UP
Generalities This set-up is designed to electro-polish a cavity in
presence of circulating acid electrolyte. The expected flowrate range is between 5L/min and 40L/min. A constant voltage EP process has been chosen for the treatment of the cavities. The available electric power for the electro-polishing is 30kW (20V – 1500A). The set-up is divided into two distinct ventilated cabinets: The acid storage area and the main treatment cabinet (Figure 1). Insertion of the cathode into the cavity and assembly of connecting pieces are done on a dedicated table close to the set up. The cavity is then transferred in the main cabinet with a specific handling tool.
The acid storage tank is located in a pit connected to the liquid exhaust storage area of the laboratory. The storage capacity is 300L. This capacity has been calculated in order to maintain a niobium concentration in the electrolyte below 10g/L after a bulk EP for SPL cavities.
The cavity is filled up from the bottom by a membrane pump located in the pit. Once the acid reaches the connection tube on the top of the cavity, the acid flows to
Proceedings of SRF2011, Chicago, IL USA TUPO062
07 Cavity preparation and production 549
the tank by gravity. The circulation of acid is uninterrupted when voltage is on. Temperature of the acid bath is stabilized thanks to circulation of cold water through a Teflon heat exchanger inserted inside the acid tank, as shown on the flowsheet of the process on Figure 2.
At the end of an electro-polishing sequence, the acid is drained back to the tank and the cavity is rinsed with Ultra-Pure water. A predetermined sequence of filling (from the bottom)/draining of the cavities is followed by a continuous circulating sequence. Rinsing stops automatically as soon as water conductivity is low enough.
All the process is automated and run through a touchscreen (See Figure 2).
Figure 1: VEP set-up at CEA Saclay. 1: Main cabinet for cavity treatment. 2: Acid tank located in a pit. 3: Table for assembly prior to VEP.
Figure 2: Flowsheet of the process displayed on the touchscreen.
Safety Safety aspects have been carefully taken into
consideration to design the set-up. Nitrogen is blown on
the top of the cavity and on the top of the acid tank to dilute the hydrogen generated during the process and prevent any risk of explosion. Furthermore, the program of the automat includes specific procedures if sensors (temperature, pressure, hydrogen, flows, acid level) should detect any failure:
- high temperature detected: the process is stopped and acid is drained,
- insufficient nitrogen flowrate: the current is stopped,
- extraction failure: The nitrogen flowrate is stopped as well as the generator,
- high level of acid detected on the top of the cavity: the pump stops.
Fundamental parameters (temperature, flowrate, voltage) evolution is monitored on the touchscreen of the automat, as well as failure occurrence. A USB port allows an easy exportation of data.
Materials The acid tank is coated with Teflon and piping/valves
are made of PFA. PVDF has been chosen for some pieces requiring higher mechanical resistance (connections for the cavity, inserts for sensors).
MODELING OF VERTICAL EP
2D Axi-symmetry Model with COMSOL Multiphysics
The described set-up makes it possible to use a wide range for operating parameters. (acid flowrate between 5 and 40L/min, voltage between 0 and 20V, nitrogen flowrate between 0 and 50L/min, etc.). It was designed with R&D purpose in mind, with the view of tracking the optimal parameters for VEP process. A complementary approach consists in modeling the process using, COMSOL Multiphysics software which allows evaluating some parameters not easily accessible by experiments. Previously, studies (fluids, electric field, concentrations, thermal properties) have been carried out with different softwares. [6,7,8] In this paper, fluid distribution and electric field have been modeled for SPL and ILC cavities with a cylindrical cathode (30 mm diameter) shape. In a second step, alternative cathode shapes have been numerically investigated for possible improvement of the process.
Contrarily to horizontal EP, VEP is a symmetrical process. Gravity force, represented by Fz on Figure 3. does not depend on radial component. It is thus possible to create a 2D model with symmetry according to the cavity z-axis (see Figure 3) which makes the calculation easier compared to a 3D model.
Fluid Dynamics Modeling For the considered set-up, acid is introduced in the cavity through eight holes equidistant from the cathode. To
1
2
3
TUPO062 Proceedings of SRF2011, Chicago, IL USA
550 07 Cavity preparation and production
pv
FWw
Alm
(fBc
E
hbd
e
pel
tu
simplify the profile, has beevelocity in the
Figure 3: 2D We have conswith a paraboli
Low fluid vAs a consequlaminar flow model the fluid
is the dyna(m/s) ; p the force, only theBoundary condcathode nor at
Electric FielFor the elect
has been chosbeen applied distribution is some parts oexplained in fo
FLU
Contrarily topaper, a holloenables to injelocated in froshown that s
uu (2
model, a crowen considered,center of the c
axi-symmetry sidered a crownic profile.
velocities (<0.8ence, Navier-Sof incompres
d dynamics in t
amic viscosity
pressure (N/me gravity is conditions chosenthe cavity surf
ld Modelingtric field modeen for the cath
to the cavconsidered the
of the cathodollowing parag
UID DYNAHORIZO
o the process owed cathode ect the acid innt of equatorsuch a configu
. u
Gpuu ).
wn of acid, as described in
crown is called
used for modn of fluid ente
8m/s) were useStockes equatissible fluid wthe cavity:
(Pa.s) ; u the
m2) ; G (N/m3
nsidered here. n: the fluid neitface.
g ling, a potentiahode and a 15Vvity. Only prere. To improvdes have beengraphs.
MICS: VEPONTAL EP
investigated is used for honto the cavitys. Previous muration is ina
0
G
with parabolin Figure 3. Th
d vmax.
deling of VEPering the cavity
d in the modelions, related to
were chosen to
(1)
(2) e fluid velocity), the resultan
ther slips at th
al equal to zeroV potential harimary currene homogeneityn insulated a
P VS
in the presenorizontal EP. Iy through hole
models [6] havappropriate fo
c he
P. y
l. o o
y nt
he
o as nt y, as
nt It es ve or
homogenoincreases inhomogeof field fcavities [2the VEP oResults oare shownthe centrashown in
Figure 4vmax=0.2line (1 to and the ca
This si
distributiothe cavitydistinct continuatisecond bu(10% of modify thrate of nimake it ppurpose inin our cobstacles surface,
Four almore homeach caselocations
Fluid v
a)
ous fluid distfrom the cente
eneity might bflatness observ2]. Similar invof 9cell cavitieobtained with an on Figure 4.al cell, the flu4c).
: a) Fluid m2m/s. b) Zoom
5), the distribavity is similar
imulation highon does not dey. The graph parts: the fion of the beaump characteri
the maximumhe geometry tiobium. It has possible to impn mind, differ
configuration. to guide the
lternative cathmogeneous flue, fluid velociof the central c
elocity m/s
tribution. In er of the cavitybe responsible ved after hori
vestigation has s with a cylinda rod cathode Along each lin
uid profile is s
modeling in thm of the centrabution profile br.
hlights that inepend on the po
highlights a ffluid flows pam pipe (see stic of the flow
m flow). It isto improve thebeen proved t
prove the procrent cathode sh
They are aie acid flow
hode shapes wid velocity in ity was studiecell of the cavi
fact, fluid vey to extremities
for the degradizontal EP of been carried o
drical cathode.and vmax=0.
ne 1-5 considesimilar to the
he ILC cavital cell. c) Forbetween the ca
n VEP case, osition of the cfluid profile inprincipally inFigure 4c.) an
w in the cell is s then necessae expected remthat shaped ca
cess [7, 9].Withapes were moimed at genetowards the c
were investigatecells (Figure
ed at five difity.
elocity s. This dation
f 9cell out for
2 m/s ered in
shape
ty for r each athode
fluid cell in n two n the nd the lower
ary to moval athode th this odeled erating cavity
ed for 5). In
fferent
c)
b)
Proceedings of SRF2011, Chicago, IL USA TUPO062
07 Cavity preparation and production 551
tdecd
FV
ccrb
vtv
Figure 5: Sinside the ceFrom the an- The sh
velocity- Uniform
using shThese t
tested during
INFLUE
The experimthe Tesla-shapdedicated to tweffect of poliscavity whosdimensions, is
Figure 6: 704 VEP set-up.
Similar equa
chosen to mcathode. Fluirepresented inbeam pipe is lo
We notice tvery low. Thethat the velocivelocity at the
85
Shapes tested tells. alysis of relate
hape D allowy close to the cmity of fluid vhape C. two shapes areg VEP.
ENCE OF TCAV
ments have highpe 9cell cavitiwo types of cshing parametese drawing, shown in Figu
MHz SPL cav
ations and boumodel fluid dy
d distributionn Figure 7, witocated on top. that fluid velo fluid distributity in the cell icathode surfac
212.9
Riris = 6
to improve flu
ed data, we havs a 39% incell.
velocity is imp
e promising ca
THE SIZE OVITY hlighted fluid ies. As the seavity, it is viters on the lar
including ure 6.
vity to be elec
undary conditiynamics withn in such coth vmax=0.8m
ocity at the cation along the is negligible cce.
4
4.7 Req
uid distribution
ve concluded: crease of fluid
proved by 33%
andidates to b
OF THE
distribution foet-up is mainlytal to anticipatrger 5Cell SPL
characteristi
ctro-polished in
ions have been a cylindricaonfiguration i
m/s. The shorte
avity surface ired line showompared to th
144
q = 191
n
d
%
e
or y te L c
n
n al is er
is ws he
Figure 7:b) fluid ve
These r
cavities. Ibecause o
- low- loc
It is th
geometriealternative(Figure 8)at the extlarge prot
a)
Figure 8:b) fluid cathode mcell.
The fl
(called F)available velocity imultiplied
In thattreatmentmodels dduring pr
F
a)
a) Fluid distribelocity profile
results show tIn fact, stagna
of: w convection acal saturation o
hen vital to aes. Similarly e shape was a). We have contremities to altuberances.
b)
a) Fluid distribvelocity prof
makes it possib
luid distributio is shown on for the fluid
in the proximd by 5.5) and int way, shaped. We also hav
do not take inrocess. Hydrog
Fluid velocity m/
)
bution in the 5along the red l
that EP is lessant acid in the
at the cathode sof the acid in ni
anticipate morto the ILC
also designed nsidered a beamllow the use o
bution in the 5file along theble to increase
on achieved Figure 8. Th
d circulation mity to the cath
n cells (multipld cavities are ve to keep in mnto account thgen forming a
/s
5cell SPL cavitline.
s adapted for e cell is undes
surface iobium.
re adapted caC cavity case
for the SPL cm pipe withoutof the cathode
5cell SPL cavite red line. Sfluid velocity
with this cahe narrowed seincreases the hode (flow rolied by 5). desirable for
mind that prese forming of at the cathode
b)
ty and
larger sirable
athode e, an cavity t taper e with
ty and haped in the
athode ection
fluid oughly
VEP sented gases
e, and
TUPO062 Proceedings of SRF2011, Chicago, IL USA
552 07 Cavity preparation and production
pg
uiroc
hoicofm
Faa
pirta
ifcc
potentially, oxgreat impact on
STUDYModels des
uniformity ofimproved by removal rate dof ions. Thus, cells is desirab
The electrichas been modeof 15V betweein the model.curves in a cobserved betwfield Ei: 825V/more than forty
Figure 9: Potena cylindrical canode, V=15V
Benefits of
[7, 9]. Thanksprevious paragis more homremoval rate athe parts of additional imp
In order to investigated cafor both 9cconfigurationsconfigurations
- cylindri- cylindri- shaped
insulatio- shaped
V=
xygen at the cn fluid mixing
Y OF THE Escribed in pref the acid fl
using a shaduring EP also a uniform elec
ble for uniform field during Eeled in the celen the cavity a. Figure 9 becell. Close to ween the lines/m. At the equay: 20V/m.
ntial distributiocathode. At th
V.
shaped cathods to the protugraphs, the ave
mogenous, favoalong the cell. I
the cathode provement of th
evaluate quanathodes, the elecell and 5c (as an ex used for the 9ical cathode (shical cathode wicathodes (sha
on cathode with in
=0
cavity surface.
ELECTRICevious paragralow inside caped cathodedepends on ele
ctric field durinm removal of the
EP with a cylinlls of ILC caviand the cathodeelow shows th
the iris, shors, generating aator, the field E
on in a Tesla she Cathode, V
des have alreaduberances inveerage cavity-caorable for a In a second stein front of t
he uniformity. ntitatively the ectric field has
cell cavities, xample, Figurcell cavity): hape A), withoith insulation aapes B, C, D,
nsulation
V=1
High field at
Low equat
e might have
C FIELD aphs show thaells might b. Furthermoreectro-migrationng EP along the surface. ndrical cathodities. A voltage has been usedhe isopotentiart intervals ara high electriEe is divided by
shape cell usingV=0 and at th
dy been studieestigated in thathode distancmore uniform
ep, insulation othe iris allow
benefits of ths been modeled
in differenre 10 show
out insulation at the iris
E, F) withou
5V
iris
field at tor
a
at e
e, n
he
de ge d al re c y
g he
d he
e m of
ws
he d
nt ws
ut
Figure numericV=15Vcathode Results
cathode) a
Table 1: EILC cavity
ShapeABC D E A BCD E
Similar
5cell caviTested cat
10: Alternacally tested fo
V, at the cathoe are in red (A,
s (Voltage of are summed up
Electric field ay for different
Insulation NoNoNo No No Yes YesYesYes Yes
r investigation ity with a cylithodes are show
ative insulateor the Tesla caode. V=0. Ins, B, C) and in l
15V between p in the recapit
at the iris and acathode configEi (V/m) Ee
825 816 913
1138 600 200 96225 128 240
has been carrindrical cathodwn below (Fig
d cathode savity. At the csulated parts olight blue (D an
the cavity antulative Table1
at the equator gurations e (V/m) Ei/E
20 4118 4533 2743 2635 1750 4.15 6.30 7.35 3.32 7.
ried out for thede and the shagure 11.).
shapes cavity: of the nd E).
nd the .
of the
Ee .2.3.7 .5 .1 0 457 5
e SPL ape F.
Proceedings of SRF2011, Chicago, IL USA TUPO062
07 Cavity preparation and production 553
T
mtir
ahapcMatcef
CFToA
Figure 11:numerically V=15V, at cathode are The obtained
Table 2: ElectrSPL cavity for
Shape Insu
A F A YF Y
For both IL
make it possibthe cells. Howinsulated cathremoval rate.
OUTLThe Electro
adapted for lahas been comachieved withprocess. Howechosen in oModeling wappropriate catake into accocathode and telectro-polishefor full underst
ACWe acknow
Community-RFP7 programmThis work hasof the “ConseiASTRE progra
Alternative tested for thethe cathode. in red.
d results are pr
ric field at ther different catho
ulation Ei (VNo 292No 409Yes 80Yes 111
LC and SPL ble to improvewever, we havhode will be
LOOK AND-Polishing in Varge elliptical mmissioned ath circulating ever, proper prder to achi
with COMSOathodes designount parameterthe cavity sured, results wiltanding of the
CKNOWLEwledge the
Research Infrasme (EuCARD been carried oil General de lam.
insulated cae SPL cavity. V=0. Insulate
resented in Tab
iris and at theode configurat
V/m) Ee (V/m2.7 8.1 9.5 21.7 0 5.7 1.7 19.0
cavities, desie electric fieldve to keep inresponsible fo
D CONCLUVertical configcavities. A de
t Saclay. The acid through arameters migieve satisfactoOL allows
ns. The studiedrs such as gas face. Once thl be comparedprocess.
EDGEMENsupport of
structure ActivD, contract num
out with the finl’Essone” in th
athode shapeAt the cavity
ed parts of th
ble2 below.
e equator of thtions.
m) Ei/Ee 36.1 18.9 14.1 5.9
igned cathoded uniformity inn mind that anor a decreased
USION guration is weledicated set-up
VEP will ban automated
ght be carefullyory treatment
investigatingd cases did noforming at th
he first cavitied with models
NTS the Europeanvity under thmber 227579)nancial supporhe frame of th
es y: he
he
es n n d
ll p e d y t. g
ot he es s,
n he ). rt
he
[1] K. Sa217 1
[2] E.KaWork2007
[3] R.L. InternIthac
[4] Z.A.InternBerlin
[5] J. Plobeta=confe
[6] M. BInternBeijin
[7] N. StInternBeijin
[8] C.E.WorkGerm
[9] A. WWorkGerm
REFaito et al., Par1998. ko et al., in Prokshop on RF S, WEP10, pp. 4
Geng, et alnational Worka, NY, 2005, TConway et
national Workn, Germany, 2ouin et al., “O=1 cavity forerence. Bruchon et anational Workng, China, 200teinhau-Kuehlnational Workng, China, 200Reece in Proc
kshop on Rmany, 2009. TH
Wu et al., in Prokshop on R
many, 2009, TH
FERENCESrticle Accelera
oceedings of thSuperconductiv453-457. l., in Proceedkshop on RF THP04. al., in Proceekshop on RF 009, TUPPO00
Optimized RF dr pulsed pro
al., in Proceekshop on RF 07, TUP51, pp.
et al., in Prockshop on RF 07, TUP33, pp.ceedings of thRF Supercon
HPPO061, pp. 7ceedings of theRF Supercon
HPPO064, pp. 7
S ators, Vol. 60,
he 13th Internavity, Beijing, C
dings of the Superconduc
edings of the Superconduc
04, pp. 176-17design of 704 oton drivers”,
edings of the Superconduc
247-250. ceedings of the
Superconduc 204-206.
he 14th Internanductivity, B742-745. e 14th Internatnductivity, B755-759.
, 193-
ational China,
12th ctivity,
14th ctivity, 9. MHz
, this
13th ctivity,
e 13th ctivity,
ational Berlin,
tional Berlin,
TUPO062 Proceedings of SRF2011, Chicago, IL USA
554 07 Cavity preparation and production