A cglccHdc

wamDdAe

p

ar

ct

pc

wfl

cc

mqpac

AN EXPE

Abstract Cyclic therm

can be the lgradients for rlimits could bcopper, coppeconstructing pHere we predetermining thcyclic thermal

Cyclic thermwas originallyaccelerators atmotivated expDesign of fudamping of waAdding dampenhances surfastresses. In tpotentially limstudies of braccelerators shrates and pulstructures aimcan only go stotally or partia

EXThe experime

permits testingcavity itself, wshape concentrwhere the tesfields on the clike and has ssurfaces excepcavity includecircular apertu The mechani

sample holder,mode convertequality factor parameters, theamplitude of calculate the p ____________________________

*Work supported AC03-76SF005

ERIMENT

SLAM. Aich

mal stresses prolimiting factorroom temperatube pushed higher alloys, or partial or comesent an exphe potential of

fatigue due to

INTRODmal stresses proy considered t extremely h

perimental studfuture linear akefields cause

ping features tface rf magnetthese structur

mit the reliablereakdown ratehowed a direcsed heating[5]

med at reducinso far. The onally change the

XPERIMENental setup utg of the mate

which is made orates the magn

st material is cavity walls anubstantially rept that of the es one rf coupure in the top. ical design of , and a picture er is shown in

is ~22,800. Fe simulation fi

the high populsed temper________________

by the Departmen15

TAL INV

L. LaurentAC National

heler, S. HeikY. Hig

oduced by rf r on the attaure linear acceher by using sother conduct

mplete acceleraperimental stuthese materialrf magnetic fie

DUCTION oduced by rf a limiting facigh frequenciedies of the phcolliders dem

ed by higher orto these structic fields and res, thermal e operating gres in high gct correlation ]. Geometrica

ng surface rf mnly alternative e structure mat

NTAL SETUtilizes a spec

erials without of OFHC coppnetic field on oplaced. There

nd the mode uteduced magnet

material undpling slot throu

f the cavity, a of the cavity c

n Figure 1. TheFrom the lowield profile, anower input prature rise pattnt of Energy under

ESTIGAT

t, S. Tantawil Laboratory,

kkinen, W. Wgashi, KEK, T

pulsed heatingainable reliablelerators. Thespecial types oting metals inator structuresudy aimed als for toleratingelds.

pulsed heatingctor for lineaes [1-2]. Thihenomenon [3]manded strongrder modes [4]

ctures typicallyhence thermastresses couldradient. Recengradient lineabetween thes

l variations omagnetic fieldafter that is to

terial.

UP ial cavity thadamage to th

per. The cavityone flat surface is no electritilized is TE013tic fields on al

der test[6]. Thugh a centered

picture of thconnected to the cavity loaded

w power cavitynd the measuredulse, one cantern across thContract DE-

TION ON

, V. Dolgash, Menlo Park

Wuensch, CERTsukuba, Iba

g e e

of n s. at g

g ar is ]. g ]. y al d

nt ar e

of ds o

at e y e c

3-ll e d

e e d y d n e

flat plate m

Figure 1attached tsource thr

is added the sampl The high50MW, 1circulatoruse a longabout 4 dthe klystrexperimencavity at a

The pulSLAC Nawith CERto differethe matercopper zicrystal cosilver, and The firElectronicwas inititemperatuThermal temperatuFatigue crepeated microscopincrease i70oC run extrusionsduring thetested to

CAVITY

hev, C. Nantisk, CA 94025RN, Genevaaraki, Japan

made of the ma

(a)

: (a) Pulsed to cavity end rough a TE10 (W

to guarantee ale. h power tests a1.42 GHz, X-b available at th

g stainless steeldB to provide 8ron and the cants is approxima pulse repetiti

RF TElsed heating ational Accele

RN and KEK.nt machining rials used in trconium, coppopper, electropd silver plated st two sampc grade (OFEially processe

ure of 70oC fofatigue crack

ure and was preracking resultscyclic stresse

pic cracks willin the number showed many

s and propagae first run. A 2a temperature

Y PULSED

sta, 5 U.S.A. , Switzerland

aterial under te

)

heating cavitcap and (b) c

WR90) to TE01

a constant tem

are performed band klystron. his frequency al waveguide w8 dB round triavity. The pulmately 1.5μm’on rate of 60 H

EST RESULexperiments w

erator LaboratoThe samples tand heat treatthis study incper chromium, plated coppercopper.

ples tested wC10100) copped to a peaor approximateks were alreaedominate alons when a matees. Above a l begin to formof rf cycles d

y newly formedation of sites 2nd OFE Coppe of 110oC. T

D HEATIN

d

est. A feedback

(b)

ty with 3” saconnected to p1 mode convert

mperature profi

using a SLACBecause there

and power levewith an attenuat

ip isolation belse length for s at the input

Hz.

LTS were conductory in collabotested were extment processeluded OFE coHIP copper,

r, Glidcop®, c

were Oxygen per. The first saak pulsed heely 4x106 rf pdy evident ang grain bounderial is subject

certain threm on the surfacduring a 2nd sd fatigue crackthat were ini

per (Cu) samplThe surface da

NG*

k loop

ample power ter.

file on

C XL4 e is no el, we tion of tween these

to the

ted at oration xposed es and opper, single

copper

Free ample eating

pulses. t this

daries. ted to shold,

ce. An imilar

ks and itiated le was amage

MOP076 Proceedings of Linear Accelerator Conference LINAC2010, Tsukuba, Japan

232

03 Technology

3B Room Temperature RF

ictdIattbt

p

Fc

4

bca4igiii Zmroo pparabvp

a

w

included bothcracks. Transgthe crystal anddislocations tIntergranular falong grain btypes of fatiguthermal fatiguband extrusiontested in this 110oC was signpulsed heating

(

Figure 2. (a) Scopper samplesubsurface dam40μms below t

Metallograph110oC to inveboundary sepachannels that wappearance. T40μm’s belowintergranular grains boundaris heavily disinitiation and intergranular f The next tw

Zirconium (Cumuch less damreaching a temon the CuZr observed on th The remaini

processed witpulsed heatinapproximatelyrate of 60 Hz.are listed in orby a Rockwellvalues (HR15Tphotograph. Tsignificantly mare shown on samples with were minimall The Single-P

sample was c1000oC for 15

h intergranulargranular fatigued are a stochathat lead to fatigue cracks

boundaries. Anue cracks is shue stresses pron in this samp

study. The snificantly mor

g sample tested

(a)

SEM images she tested to 110o

mage along grthe surface.

hy was conducstigate subsurfaration and damwidened near

The channels ew the surface.

damage. Interies are sites wsturbed. This

growth and fatigue cracks hwo samples teuZr (0.13-0.2%mage than the

mperature of 10sample appro

he Cu sample teing samples th approximateng temperaturey one week of

These samplerder of their ml Superficial HT) are located iThe pulsed hemore pronounce

the top row hardness valuely impacted by Point Diamonchemically etcminutes. The h

r and transgre cracks occur

astic process thslip lines anare microfract

n SEM image own in Figureduced a very

ple and in all surface damage severe compto 70oC.

(b)

howing surfaceoC. (b) Metalloain boundaries

cted on the saface damage (Fmage were obthe surface cr

extended from These results

ernal flaws, inwhere locally th

may facilitatebe the reason

have been founested were C

%)). At 70oC the Cu sample. 00oC that the soached the levested to 70oC. tested in thi

ely 107 rf pue of 110oC. runtime at a p

es are shown imaterial hardneHardness Testerin the top righteating surfaceed on the softeof Figure 3. Ies >70 HR15Tcyclic thermal

d Turned (SPched and hydheat treatment

ranular fatigur on the face ohat begins withnd slip bandstures that occu

showing bothe 2a. The cycli

noticeable slipof the sample

ge observed aared to the firs

e extrusions onography showeds extending 20

ample tested toFig. 2b). Grain

bserved to formeating a funnedepths of 20

suggested onlynclusions, and

he crystal lattice fatigue crackn why mainly

nd. C15000 Coppe

his sample hadIt wasn't unti

surface damagvel of damag

is study werlses at a peakThis required

pulse repetitionin Figure 3 andess as measuredr. The hardnest corner of eache damage waer materials thaIn contrast, thT (bottom rowl fatigue. PDT) OFE Cudrogen fired a

resulted in this

ue of h s. ur h c p

es at st

n d

0-

o n

m el 0-y d e k y

er d il e e

re k d n d d

ss h

as at e

w)

u at s

Figure 3:110oC sho

sample (Fto the preof indiviimpacted distance (Back Scadegradatiothe crystagrains hadRef. [7], wThe grain(lighter rorientatio A [100] test did sufficient temperatuthe extrusthan in thwas foundheating tecyclic temapproximsurface ha Two HIPwas chemheating dobserved the softer

Figure 4: with neighdamage.

: Pulsed heatiown in order of

Fig. 4) having aevious two copdual grains rmore than oth

(i.e. same pulsattered Diffracon arising fromallographic oried the least amwould correspn orientation regions) wouln. single crystal not suggest tto overcome

ures. On the sinsions and fractue previous samd to increase aemperature. A mperature rise ately 60K has ardening. P Cu samples mically etcheddamage and between the elarge grain sam

(a) (a) SPDT Cu

hboring grains

ing samples tf their material

a much larger pper samples revealed that her grains witsed heating temction has showm pulsed heatientation of the

mount of surfacond to a [100]with the mo

ld correspond

copper samplethat grain oriee surface fatigngle crystal saures were muc

mples tested. Thalmost proportiradial distributalso showed

to be overcom

were tested bd. Both had there was no

etched and nonmples, the dam

u sample and having varyin

that were testl hardness.

grain size comtested. Examinsome grains

thin the same mperature). Elewn [7] that suing is dependee grains. The dce damage and] crystal orientost severe da

to a [111]

e was tested buentation alonegue issues at ample the lengch smaller (< 1he material harionally to the ption of hardnesthat a thresho

me in order to in

but only one sasignificant p

o major diffen-etched samp

mage on some g

(b) (b) HIP Cu sa

ng degrees of su

ted to

mpared nation

were radial ectron urface ent on darker

d from tation. amage

grain

ut this e was

these gths of 0 um) rdness pulsed ss and old of nitiate

ample pulsed erence ple. In grains

ample urface

Proceedings of Linear Accelerator Conference LINAC2010, Tsukuba, Japan MOP076

03 Technology

3B Room Temperature RF 233

had a matted appearance while other grains had a weave-like appearance. The weave pattern is created by the crossing of transgranular fractures that terminate on grain boundaries (Fig. 5a). The grain with the matted surface is created by extrusions occurring through submicron size pits that form on the face of the crystal (Fig. 5b). It is speculated that the observed differences in types of grain damage may be due to a difference in crystallographic orientation. The outer edge and lower temperature region of the pulsed heating ring revealed that grain boundaries are typically the first location that is targeted.

(a) (b)

Figure 5: (a) Transgranular extrusions oriented in different directions and terminating along grain boundaries creating a weave-like pattern and (b) extrusions through micropits. There were two Cu chromium (C18200) samples that were tested and one was subjected to a 988oC vacuum braze cycle. The softer heat treated sample had significantly more surface damage than the one without heat treatment and most of the damage on the harder sample occurred along grain boundaries. A cold worked copper zirconium (CuZr) and an annealed copper zirconium sample were also tested. Remarkably, at a pulsed heating temperature of 110oC, there was no clear evidence of surface damage on the cold worked sample. This sample was retested to 150oC which resulted in a faintly visible pulsed heating ring. The softer annealed sample had significantly more surface damage in contrast to the harder cold worked CuZr sample. From the experiments that contained both soft and hard materials, the harder non-annealed materials substantially outperformed the softer materials. Although this may have been a likely outcome, it is problematic in the development of high power rf components that typically require a high temperature brazing cycle. Two copper silver (C10700) samples were tested and neither one of them were subjected to heat treatment. The hardness of the CuAg sample was similar to the CuCr sample and the experimental results were similar showing that the surface damage was dominant along grain boundaries. A 300μm silver plated Cu sample was also tested but the silver plating was severely damaged due to cracking and spalling. The sample with the highest measured hardness value was Glidcop® Al-15 (C15715). Small striation lines of pulsed heating damage were observed around the pulsed heating ring that had the same orientation irrespective of location. The cause is unclear but may have been due to

material defects, or surface imperfections created in the diamond fly-cutting process that was used. However, this process was also used on other samples and the striation pattern was not observed in any other experiment. Except for these small strips, the surface damage on the Glidcop® sample was minimal.

CONCLUSIONS This study has shown the possibility of pushing the gradient limits due to cyclic thermal fatigue by the use of copper zirconium and copper chromium alloys. Intergranular and transgranular extrusions were observed to varying degrees on all the samples tested to 110oC except for the cold worked CuZr sample which wasn't observed until after the 150oC test run. At low cyclic temperatures, grain boundaries were observed to be the initial sites impacted by pulsed heating. At higher cyclic temperatures (110oC), metallography revealed subsurface intergranular damage that extended 20-40μm’s below the surface on heat treated copper. A pulsed heating surface damage dependence on crystallographic orientation was determined using Electron Back Scattered Diffraction technology. There were two types of surface damage observed on the grain surfaces. Transgranular surface extrusions oriented in different directions and terminating on grain boundaries which created a weave pattern appearance and the second type of damage observed was due to extrusions through micro-pits. It is speculated that the differences between these two may be due to a difference in crystallographic orientation. In general, non-heat treated samples which had a higher material hardness and smaller grain size significantly outperformed the heat treated samples. This may be due to the small grain boundaries limiting the propagation of surface extrusions which terminated on grain boundaries.

ACKNOWLEDGEMENTS The authors would like to thank: S. Beebe, J. Eichner, A. Haase, E. Jongewaard, J. Lewandowski, D. Martin, C. Pearson, R. Talley, B. Vandersyl, D. Yeremian, and C. Yoneda for their help and support.

REFERENCES [1] I.H. Wilson, CLIC-Note-52, CERN, Geneva (1987). [2] W. Wuensch, in Proc. of EPAC08 (2008). [3] D.P. Pritzkau, R. Siemann, Phys. Rev. ST Accel.

Beams 5, 112002 (2002). [4] R. Jones, C. Adolphsen, J. Wang, and L. Zenghai,

Phys. Rev. ST Accel. Beams 9, 102001 (2006). [5] V. Dolgashev, S. Tantawi, Y. Higashi, and T. Higo,

in Proc. of EPAC 2008, Genoa, Italy (2008). [6] C. Nantista, S. Tantawi, J. Weisend, R. Siemann,

V. Dolgashev, L. Campisi, in Proc. of PAC2005. Knoxville, TN (2005).

[7] M. Aicheler, S. Sgobba, G. Arnau-Izquierdo, M. Taborelli, S. Calatroni, H. Neupert, and W. Wuensch, submitted to International Journal of Fatigue (2010).

MOP076 Proceedings of Linear Accelerator Conference LINAC2010, Tsukuba, Japan

234

03 Technology

3B Room Temperature RF