Construction and first operation of a pilot CW superconducting electron accelerator

Construction and first operation of a pilot CW superconducting electron accelerator

Nuclear Instruments and Methods in Physics Research 224 (1984) 5-16 North-Holland, Amsterdam 5 CONSTRUCTION AND FIRST OPERATION OF A PILOT CW SUPERC...

899KB Sizes 0 Downloads 17 Views

Nuclear Instruments and Methods in Physics Research 224 (1984) 5-16 North-Holland, Amsterdam

5

CONSTRUCTION AND FIRST OPERATION OF A PILOT CW SUPERCONDUCTING ELECTRON ACCELERATOR * T. G R U N D E Y , H. H E I N R I C H S , U . K L E I N **, G. M O L L E R , G. N I S S E N ÷ a n d H. P I E L Fachbereieh Physik der Universiti~t-Gesamthochschule Wuppertal, 5600 Wuppertal, Germany

H. G E N Z , H.-D. G R A F , M. J A N K E , A. R I C H T E R , M. S C H A N Z , E. S P A M E R a n d O. T I T Z E Institut ff~r Kernphysik der Technischen Hochschule Darmstadt, 6100 Darmstadt, Germany

Received 29 September 1983 and in revised form 31 January 1984

We report on the first operation of a small superconducting linear accelerator which serves as a pilot project for the Darmstadt superconducting recyclotron. A five cell niobium accelerating structure operated at 3 GHz was used to accelerate a chopped 200 keV electron beam to an energy of 850 keV. The Q value of the structure and its accelerating field (at 1.8 K) remained at 4x 10 9 and 5.7 MV/m respectivelyduring one year. Design, fabrication and test of the accelerating structure are discussed. A description of the pilot accelerator including its gun, chopper, and cryostat is given.

1. Introduction The successful experiments with superconducting (sc) niobium cavities in the late 1960s [1] have in a short time led to the application of rf superconductivity in high energy physics and nuclear physics. The first large scale instruments based on this new technology were the Stanford superconducting accelerator [2], the Illinois microtron [3], the CERN-Karlsruhe particle separator [4], and the heavy ion accelerator Atlas at Argonne [5]. These devices have been in operation for several years and have demonstrated that sc cavities can be operated reliably in linear accelerators. It was shown that the drastic reduction of the rf surface resistance in sc cavities could be achieved even in complex resonators. The early expectations, however, to reach the very high electric accelerating or deflecting fields promised by the elementary theory of superconductors in radio-frequency fields were not fulfilled. Limitations in the accelerating field showed up typically at values slightly above 2 MV/m. In recent years the phenomena leading to these limitations have been investigated experimentally and considerable understanding has been gained. The resonant multiplication of electrons (multipacting) in rf * Work supported by Deutsche Forschungsgemeinschaft. ** Now at INTERATOM, D-5060 Bergisch-Gladbach 1, Germany. + Now at Hewlett Packard, D-7030 BOblingen,Germany. 0167-5087/84/$03.00 © Elsevier Science Publishers B.V. (North-Holland Physics Publishing Division)

cavities was analysed [6] and cavity shapes were found which suppress this effect [7]. Improved diagnostic methods, especially the temperature mapping of cavities immersed in subcooled helium [8], have shown that small defects on the cavity surface are responsible for the observed quench phenomena [9]. Field induced emission of electrons at sites of high electric fields can be reduced by helium ion sputtering [10,11] and appears to be a limiting factor only at peak electric surface fields above 15 MV/m. These advances in the understanding of field limitations in sc accelerating structures have led us to propose the construction of a superconducting recyciotron [12], which will replace the Darmstadt normal conducting linear accelerator DALINAC [13]. This continuous wave (cw) electron accelerator is designed for an electron beam of 130 MeV with an energy spread of about 10 -4 and a cw current of 20 #A. From the point of view of sc structures the most important design parameter is the accelerating field of 5 MV/m. The funding of this project was granted in the fall of 1981. In order to test the superconducting accelerating structures, the most crucial components of the planned 130 MeV superconducting recyclotron, and to gain experience with superconducting technology a pilot accelerator has been set up. The pilot accelerator is described in sect. 2 and its operation and performance in sect. 3. The conclusion (sect. 4) summarizes the major results and gives an outlook

6

Grundev et a/. / Pdot

7"

cw

o G

LP

Ch

PS

L

Ch RF Chopper Covlhes FC Faraday Cup G 250kV Electron Gun HeC 18K He Cryostot L Magnetic Lens M : 9 0 " Bending Magnet P Pump Q Quadr upo~es S Chopping Slit SD E~ectron S p e c t r o m e t e r X Crystal Spectrometer SI Slhcon Detector

P

HeC

X

=O Sp i

i

FC

lm

Fig. 1. Floor plan of the pilot cw superconducting accelerator.

2. Superconducting accelerator 2.1. General layout

The main components of the superconducting cw accelerator are shown 'in fig. 1. The electron gun * produces a continuous electron beam with a maximum kinetic energy of 250 keV. This beam is chopped at the accelerator frequency (3 GHz) as described in section 2.4 below. The helium cryostat is designed to house a 5-cell accelerating structure (capture section) plus one 20-cell accelerating structure (standard section for the 130 MeV recyclotron). Behind the cryostat Mattscattered electrons from a thin (15 t~g/cm2) 12C foil can be detected by a cooled silicon surface barrier detector to determine the energy of the accelerated electron beam. Experimental set-ups like a high resolution crystal spectrometer for X-ray detection and a magnetic electron spectrometer for the investigation of Matt- and Meller scattering allow the beam to be used not only for acceleration tests but also for atomic physics experiments. Fig. 2 shows a photograph of parts of the pilot accelerator. 2.2. Accelerating structure 2.2.1. Design

The operating frequency ( f ) of the superconducting structure can be chosen in a range of about 0.5 to 10 GHz. Earlier arguments to choose a frequency as high as possible for the benefit of higher gradients or to go to low frequencies because of the f2-dependence of the BCS surface resistance are no longer supported by experiments with extended resonators which produce accelerating voltages of more than 1 MV. Accelerating fields in excess of 5 M V / m have been reached in the above frequency range. The experimentally reached rf losses per MV and per meter of accelerating structure

* On loan from lnstitut for Kernphysik, Universit~t Mainz.

superconducting electron accelerator

show a minimum around 3 GHz (for a recent review see ref. [14]). This experimental evidence and the vast experience with normal conducting S-band linacs support the choice of 3 GHz as operating frequency. Because of its high critical temperature and promising fields, reached in single cell cavities [14], Nb3Sn would be the best choice for the material of the superconducting structure [15]. However, the technology of fabrication of extended structures is at present not developed enough to base an accelerator design on this material. Therefore, we have chosen niobium which at 3 GHz enforces - due to the temperature dependence of the surface resistance of Nb an operating temperature of 2 K. At this temperature a quality factor Q of the structure of 3 × 109 can be achieved. The use of a spherical or elliptical cavity design is an efficient way to suppress electron multipacting. We have chosen the spherical design (fig. 3). The chemical surface treatments with highly reactive chemicals, the high temperature annealing above 1100°C and the mechanical manipulations to finish a structure after welding, ask for a length of the structure not very much longer than one meter. A very short structure would have the consequences of many rf coupling ports, an expensive rf system and a poor filling factor. We consider a length of one meter for the standard accelerating unit a good compromise. For preacceleration of the low energy beam one shorter structure is necessary as a capture section. Superconducting cavities are limited in their performance by the maximum surface magnetic field (defect induced quenching) a n d / o r by the electric surface field (field-emission limitation). Therefore, a small ratio of the surface electromagnetic field to the accelerating field is required. This favours the ~- mode operation of the standing wave accelerating structures. The ~r mode furthermore has the advantage of the highest specific shunt impedance r / Q per unit length. Its main disadvantage compared to the alternative choice of a 7r/2 mode is its sensitivity against tuning errors which lead to different field excitations in the individual cells. To reduce this sensitivity a strong cell to cell coupling is required, which can be achieved by a large iris aperture. The accuracy in the manufacturing process of a single cell asks for a cell to cell coupling factor of 4.1% resulting in an iris diameter of 35 mm. This large opening is beneficial to ease the coupling to the higher order modes excited by the beam. The most dangerous of these unwanted cavity modes is the TM~10-deflecting mode. Its frequency is close to 4.0 GHz. The diameter of the cut-off tube was chosen to allow the propagation of TEl1 waves with frequencies above 3.5 GHz. A sufficient and simple loading of the deflecting modes can therefore be accomplished by probes in the cut-off tubes. Because of their large diameter the length of the cut-off tubes has to be 14 cm to reduce the rf losses of the fundamental mode at

T. Grundey et al. / Pilot cw superconducting electron accelerator

7

Fig. 2. Photograph of the 5 m long cryostat housing the superconducting accelerating structure. The electron beam enters from the left and is deflected to the right by a 90 o bending magnet after acceleration reaching the electron spectrometer (labeled Sp in fig. 1). The helium is filled through the tower sitting on top of the cryostat. The gas recovery and pumping line connected to this tower is also visible.

normal conducting parts to a negligible amount. One disadvantage of a large iris diameter is the lowering of the frequency of the TE1H structure mode. It has to be

F.

avoided that the pass band of this mode crosses the pass band of the fundamental accelerating mode. In our design the lowest frequency of the TEl11 pass band is at 3.38 GHz. To compensate the influence of the cut-off tubes the diameter of the end cells has to be reduced to get a uniform field distribution in the structure. This leads to different sets of dimensions for the two cell types (table

1). The small superconducting accelerator described in this paper contains a 5-cell accelerating structure. This cavity is shown in fig. 4 and its rf parameters are given in table 2. For economy reasons it is fabricated with the same deep drawing tools as the standard 20 cell unit. This means that it is optimized for the acceleration of fl = 1 electrons. Calculations show that such a cavity can be used to accelerate electrons with an initial energy of 200 keV (see sect. 3.2). 2.2.2. Fabrication

Fig. 3. Shape of a single cell. The dimensions are given in table 1.

The fabrication of the cells is done in several steps. First, a disk ( ~ 150 mm) of 2 m m N b sheet material

8

7~ Grundey et al. / Pilot cw superconducting electron accelerator

Table 1 Nominal cell dimensions (cm) at operating temperature

Standard cell End cell

L

D

a

R

r

h

5.0 5.0

9.006 8.934

3.5 3.5

2.2 2.2

0.3 0.3

0.253 0.217

4

5

2

1

3

6

mm 0

50

6

2

5

1 0

Fig. 4. Five cell accelerating structure with rf probes. The different parts are labeled as follows: niobium cavity (1), stainless steel coupler housing (2), rf input probe with Qext = 1.4x 109 (3), rf output probe with Qext = 8 X 1011 (4), spare ports for higher order mode couplers (5), iadium seals (6) and entrance flange (7).

Table 2 Characteristic rf parameters for a 3 GHz 5-cell structure Normalized shunt impedance r/Q Geometry factor G Field enhancement factor Ep/Eac c Field enhancement factor Hp/Eac c Cell to cell coupling factor

20 12/cm 293 I2 3.0 4.2 m T / ( M V / m ) 4.1%

(KBI Corp.) is formed by deep drawing to an ashtraylike cup. T h e n the cups are stress-annealed at 1100 o C. Next the iris aperture is deep drawn. After a short chemical polishing (10/xm) the cups are welded together by a n electron b e a m a n d the resonant frequencies of the individual cells are measured. Our present experience resulting from m e a s u r e m e n t s on more t h a n 30 cells shows that a b o u t 90% of all cells are within a n interval of 3.5 M H z a r o u n d the design frequency a p p r o p r i a t e for this fabricational step. This places a n u p p e r limit for the variations of the m a i n diameter D of + 4 5 / x m . Individual chemical polishing ( 5 0 - 1 0 0 /tm) reduces the frequency deviations to less than 300 kHz. Before the final welding the five cell structure is set up to measure its field flatness. A n almost u n i f o r m excitation (field deviation < 2%) of all cells is achieved by a n appropriate p e r m u t a t i o n of the cells. Deep drawing, machining and welding was p e r f o r m e d at INTER.ATOM. 2.2.3. Final treatment and tests without beam After welding, the five cell structure was chemically

polished a n d tested in a vertical b a t h cryostat following conventional procedures. A t e m p e r a t u r e m a p p i n g system was used to detect regions of increased rf losses in the structure. Fig. 5 shows the experimental set-up. The final cavity t r e a t m e n t necessary to reach the design values is summarized in table 3. In test 1 an accelerating field of 3.9 M V / m was achieved and the q u e n c h location detected with fixed resistors was located close to the weld of iris 4. O p e r a t i n g the structure in the other pass b a n d modes showed a n o t h e r q u e n c h location at iris 5. Therefore, all iris welds which could be reached by a simple grinding tool (irises 1,2,5 a n d 6) were ground, followed by a short chemical polishing a n d rinsing with demineralised dust-free water. The cavity performance improved. A first temperature m a p was taken a n d is s h o w n in fig. 6. It shows two significant features. First of all three spots of increased rf losses are clearly visible. All of them are located at irises. The one at iris 6 (resistor 40) is responsible for the observed quench. Second, it is seen that the m a x i m u m temperature increases above the u n i f o r m b a c k g r o u n d are found in the iris region. This can be due to two effects. First of all the cooling of the iris region in subcooled helium is worse than close to the equator. Equal heat fluxes at e q u a t o r a n d iris will therefore result in higher temperature increases at the iris. On the other hand, dielectric losses could at least in part be responsible. The separation of these two m e c h a n i s m s needs further investigation. To reduce possible surface layers of material with dielectric losses the structure was fired at 1500 o C in an

T. Grundey et aL / Pilot cw superconducting electron accelerator ! I I

-~[

1L

_

--_!!.

View

A- B

Cj_.

View C - D

Fig. 5. Experimental set-up for tests in a vertical bath cryostat. It is equipped with a rotating system of 40 sliding carbon resistors. Each single cell is surrounded by 8 resistors. The resistors number 1, 8 and 9, 16 and 17, 24 and 25, 32 and 33 and number 40 sit at the irises numbers 1 through 6, respectively.

CP(5/tm), Ac, Me, wet mounting Grinding of irises 1, 2, 5 and 6 CP(20/xm). H 2 0 , wet mounting UHV-annealing at 1500 ° C for 6 h, dry mounting H 2 0 , wet mounting

CP(20/~m) and UHV-annealing at 1500 o C, dry mounting

Structure pressurized with dust free N 2, equipped with straight through valves

1 2

5

6

4

3

Cavity treatment

Test

4.0 x 109

5.2 x 109

1.6 x l0 ~

3.0 x l09

1.0 x 109 3.4 x 109

Q0 at E~'~ff~ T = 1.8 K

5.7

6.2

4.8

3.8

3.9 4.9

E~,~ ~ (MV/m)

Table 3 Final treatment of accelerating structure. CP = Chemical polishing with a solution of HF, H N O 3, H3PO4(I : 1 : 1 ). Ac = Rinsing with commercial dust-free acetone. Me = Rinsing with commercial dust-free methanol. H 2 0 = Rinsing with demineralized, dust-free water (conductivity 0.1 ~tS, filter size 0.22/~m).

Temperature m a p (fig. 6b) quench at iris 6 but different location compared to test 2. Temperature map (fig. 6c) quench at iris 5, increase of quench field from 3.3 M V / m to 4.8 M V / m within a few minutes of rf processing. Observation of slight field emission at m a x i m u m field. Quench field increased from 5.4 to 6.2 M V / m by rf processing. Quench location shifted from iris 4 to iris 3 (temperature m a p fig. 6d). Observation of slight field emission at m a x i m u m field. Mounted in the horizontal cryostat at Wuppertal no temperature mapping.

Fixed carbon resistors used, quench at iris 4. First temperature map (fig. 6a) quench at iris 6.

Remarks

r~

T. Grundey et al. / Pilot cw superconducting electron accelerator

11

&T

-24ink

AT

t~Tmax=402ink

Eocc = 3 . 6 M V / m

0'" 40

0 360"

1

10 N~

o

20 :.~or o~ ce~St er

36o*~

lO

er o~ ces~St°c

AT I ~Tmax = 4 7 4 mR

Ea¢¢= 4 . 6 M V / m

&T E ace = 5 . 0 M V I m

&Tmox=375mK

0 a

O*

o

10

20

0'~

3O re~ist°r

360

1

Nom~De ~-o~ ce~Sl°r

Fig. 6. Temperature maps taken during test 2 (fig. 6a, top left), test 3 (fig. 6b, top right), test 4 (fig. 6c, bottom left), and test 5 (fig. 6d, bottom right). Every cell was equipped with 8 resistors, giving a total of 40. The positions of the resistors can be seen in fig. 5. Resistor number 1 is on top of the cavity. The temperatures were measured in subcooled helium in angular steps of/t~ = 10 o.

ultra-high vacuum furnace. 1 5 0 0 ° C is the maximum temperature at which this structure is not deformed. Higher processing temperatures are possible with an appropriate mechanical support system. During mounting into the vacuum system a dust particle must have fallen into the resonator, which clearly shows up in the temperature map (fig. 6). A water rinsing was applied to the cavity and the accelerating field of test 2 was reproduc~l. For the first time, however, slight non-resonant electron loading (field emission) was observed and the quench field, initially at 3.3 M V / m , could be increased to 4.8 M V / m by several minutes of rf processing. The attempt of test 3 to improve the cavity performance by a high temperature treatment was repeated prior to test 5 and this time the design values of the accelerating structure could be exceeded. The structure was then pressurized with dry nitrogen and equipped with

straight-through valves-for its later accelerator application. Prior to its shipment it was tested in the horizontal accelerator cryostat at Wuppertal and showed no significant reduction in its performance. It was transported to Darmstadt under vacuum, mounted in the accelerator cryostat and a first test of its superconducting features did not show any changes compared to the results achieved at Wuppertal. 2.3. Cryostat

The helium cryostat of the pilot accelerator has been designed at Wuppertal and was built in the workshop of the institute at Darmstadt. A cross section of part of the cryostat with a 5-cell accelerating structure installed is shown in fig. 7. The outer vacuum jacket is fabricated from stainless steel and has a length of 5 m and a diameter of 0.4 m. Inside, a liquid nitrogen cooled

=

l

...[_~

---T

~i¸71¸ii i i ~ ~ "

@ 4,.-

~

Fig. 7. Cross section of half of the 5 m long helium cryostat with the 5-cell accelerating structure• The different parts arc labeled: ~acuum jacket ( 1 ), LNz-cooled radiati¢m ,~hield (2). LHe vessel (3), niobium cavity (4) and cut-off tubes with rf couplers (5).

®

®

@

i!

e~

.2

3~ e~

e~

4.

T. Grundey et al. / Pilot cw superconducting electron accelerator

radiation shield (aluminium) is provided to keep heat losses low. The liquid helium vessel (stainless steel) has a usable length of 3.5 m and an inner diameter of 0.20 m. Both the radiation shield and the helium vessel are wrapped by 20 layers of aluminized mylar foil (superinsulation) and 5 layers of glassfiber cloth to reduce radiation losses. In the range of 1.8-4.2 K the standby losses of the cryostat are almost independent of temperature and amount to 2.1 W corresponding to a liquid helium evaportion rate of 3 1/h. The whole beam line from entrance to exit of the cryostat is shielded magnetically by two layers of a 0.29 mm thick Cryoperm foil. Since there is not yet a closed loop refrigeration system the cryostat was operated with liquid helium from dewars. After precooling the cryostat to liquid nitrogen temperature 190 liters of liquid helium are required to fill the helium vessel. Temperature reduction to 1.8 K is achieved by lowering the helium vapour pressure to 17 mbar by a combination of a roots pump (800 m3/h) and a forepump (160 m3/h). During this process another 100 liters of liquid helium are transferred to the cryostat to maintain the helium level. Finally the cryostat contains about 100 liters of helium at 1.8 K which allows an acceleration test of about 6 - 8 h depending on the rf fields in the accelerating structure used during the test. 2.4. Injection

The electron gun operates at injection voltages up to 250 kV, however, most of the acceleration tests have been performed at 200 kV. Measurements of the beam profile using wire scanners at two different positions yielded a radial emittance of er < 5 mm • mrad for beam currents in the range from 20 # A to 300 ttA. The continuous electron beam from the gun is chopped into bunches by two transverse deflecting rf cavities and a chopping orifice. The operation principle is illustrated in fig. 8. The cavities oscillate in the TMl20 and TM210 mode, respectively, at the accelerator

13

frequency. The beam is deflected by the magnetic field in the cavities, the first cavity deflecting vertically, the second cavity horizontally with a 90 o phase shift with respect to the first one. This causes the beam to move on an elliptical path around the chopping orifice positioned 1.7 m downstream the beamline. The position of the ellipse is shifted horizontally by a set of steering coils (magnetic bias) to a position where one branch crosses the orifice to produce one bunch per rf cycle. Excitation of the cavities by 100 W and 10 W, respectively, yields an ellipse size of 24 mm by 8 mm. The beam is focussed to a diameter of 2 mm at the chopping orifice which is 4 mm in diameter. The resulting bunch has a width of 18 ° (fwhm) and a leading and trailing edge of 9 o referred to one rf cycle. 2.5. Electronics

A schematic diagram of the rf set-up used to excite the chopper cavities and the superconducting 5-cell accelerating structure is shown in fig. 9. A voltage controlled oscillator (VCO) is phaselocked to the resonant frequency of the superconducting structure. Its output power is split, the main part is used to drive the accelerating structure (indicated schematically) through a 20 W T W T amplifier. The PIN modulator incorporated in this path allows pulsed operation in order to determine the rf properties of the superconducting structure from incident, reflected and transmitted power monitor signals (labeled Pinc, Pren and Pout, respectively, in fig. 9). The phase-locked loop is formed by two directional couplers, insulators, attenuator, amplifier, phaseshifter and a double balanced mixer (DBM). A small fraction of the output power from the VCO is used to drive the chopper cavities through a 200 W klystron amplifier. Constant output power in this branch is maintained by the PIN-attenuator, whereas the relative phase between chopper and accelerating structure is adjusted by the phaseshifter in front of the amplifier.

3. Accelerator operation and results 3.1. Overall performance

if

Electron B e a m

RF

TM~,o ~

-

Cavit)es

Fig. 8. Principle of chopper operation.

J

Chopping Oriflce

The superconducting 5-cell accelerating structure used in the accelerating tests has been mounted in the cryostat for more than one year now. In this period of time six tests have been performed to improve the beam transport system and the beam position monitors and to compare experimental beam energies with the results of calculations. After each test the cryostat was warmed up to room temperature and taken apart to allow for minor changes in the beamline, like installation of correction coils, a position monitor, etc. The accelerating structure was always kept under vacuum to prevent a contamina-

14

7". (/rundev el aL / Pilot cw superconducting ele~ trun a{ ~el~'ral,)r

4CHOPPER

kSLSkAA3

CAVlT ES

5 CELL STRUCTURE

Fig, 9. Schematic diagram of rf set-up.

tion of the niobium surface. The unloaded Q of about 4 × 109 and the maximum effective accelerating field ( Ea~~ = 5.7 M V / m ) at 1.8 K remained unchanged in all tests. The effective accelerating field is defined as the maximum voltage gain of a synchronous electron passing the accelerating structure and its cut-off tubes, divided by the length of the structure without cut-off tubes (25 cm in our case). A relativistic electron (/3 = 1) would experience an energy gain of 1.25 MeV at Ea~~ = 5 MV/m. 3.2. E f f e c t i v e a c c e l e r a t i n g f i e l d s a n d M o t t s c a t t e r i n g s p e c tra

The beam from the electron gun is in the velocity range of/3 = 0.7-0.74 for injection voltages between 200 kV and 250 kV, respectively, whereas the 5-cell accelerator section is a /3 = 1 structure. Therefore, calculations of the kinetic energy of the electrons in this structure for various combinations of injection energy, injection phase and accelerating field have been performed using field distributions as determined from coumputer codes L A L A [16] and S U P E R - F I S H [17]. In fig. 10 the result of such a calculation for an accelerating field of 5 M V / m and injection energies of 200 keVand 250 keV at

optimum injection phase is shown. Two results should be noticed: 1) The fact that the accelerating field penetrates into the cut-off tubes on both sides of the structure decreases the injection energy by almost 100 keV, which results in a less effective acceleration. 2) The phase slip of the electrons which is increased by the aforementioned effect causes a low efficiency towards the end of the structure (in the region between 15 cm and 25 cm). Increasing the injection energy improves this situation (as can be seen by comparing the full and the dashed curve in fig. 10) but injection energies in excess of 250 keV result in an insufficient bunching by the preaccelerator section. So far, acceleration tests have been performed at injection energies up to 200 keV. Four typical spectra of Mott scattered electrons from a 15 ~tg/cm 2 ~2C foil taken with a cooled Si-detector are displayed in fig. 11. The spectrum at the top was taken with the direct beam from the electron gun with no acceleration field present. Applying accelerating fields of 3.2, 4.3 and 5.7 M V / m yields energies of the Mott scattered electrons of 350, 565 and 845 keV, respectively. A 2°9Bi source with electron energies between 480 keVand 1050 keV was

T. Grundey et al. / Pilot cw superconducting electron accelerator i

5CELL

i

i

1.5

i

EN'ERGY'vs. RF'-FIEL'D

UNIT

Eo=250keV

--Eo=2OOkeV

' /

E

g

/

i.o

//

)(.9

,i/ X\ill/ ~\\\

v >~

~

/

~,o~,~,,~

o

W Z w U

r \j I

I

L

~/

t~ 0.5

/

Z

500

//

/////~9

l-

tW u

'

- - - Eo= 200 keV -Eo=225keV - - EO= 250 key

1000 --

15

/

/

/

/"

,/

.

/

o/

//

/

// P"

3

4

/

/

./ /'/

r

/

0

~

2 ACCELERATING

0 I

-5

I

0

5

10

Accelerotor

15

20

25

3i0

Length (cm)

I

I

5 CELL

I

I

UNIT

400

Eo= 2 0 0 keV

200

ACC. FIELD O M V / m

..J o

4o

l

i

I

Z < -'r U 20 u3 F-

0

3.2MV/m

I

I

6

I

7

(MV/m)

used for the energy calibration of the Si-detector. In fig. 12 calculated energies for the electron beam after being accelerated by the 5-cell structure are shown as a function of the effective accelerating field for three injection energies. The experimentally determined energies are extracted from the spectra of Mott scattered electrons. The corresponding values for the effective accelerating field were calculated from rf power measurements. At the injection energy of E 0 = 200 keV there is satisfactory agreement with the calculated values. Note that the maximum accelerating field of 5.7 M V / m is the highest field ever achieved with a superconducting multicell structure operated in an accelerator.

1

4.3MV/m 10

0

I

5.7MV/m

80

40

I!

i f

I

0

FIELD

I

5

4. Conclusion [

ou

0

I

Fig. 12. Comparison of calculated beam energies at the end of the 5ocell preaccelerator for three injection energies with measured values determined from the Mott-scattering spectra. All measurements were performed at 200 keV injection energy.

Fig. 10. Calculated beam energy in a 5-cell preaccelerating structure for two injection energies. The effective accelerating field is 5 M V / m in both cases. The vertical dashed lines indicate the location of the irises. 600

I

200

400

600

800

1000

K I N E T I C E N E R G Y (keV)

Fig. 11. Energy spectra of Mott-scattered electrons on 12C for different accelerating fields. Note that the maximum obtained accelerating field is 5.7 MV/m.

A superconducting 5-cell accelerating structure of niobium was fabricated using procedures applicable to mass production. A Q value of 4 × 109 and an accelerating field of 5.7 M V could be achieved. Both values exceed the design parameters for the Darmstadt recyclotron. This cavity was operated in a pilot accelerator to accelerate an electron beam to an energy of 850 keV. The structure operated reliably and kept its performance over one year. It was not influenced by the numerous rearrangements of parts of the accelerator. Also all other components of the accelerator including the cryostat performed according to our expectations. The next step in the project is the installation of an additional 20-cell structure behind the 5-cell preaccelerator and their combined operation. This 1 m long

16

"F. Grundey et al, / Pilot cw superconducting electron accelerator

structure is presently being tested at Wuppertal. Concerning the 130 MeV superconducting recyclotron we finally remark that the accelerator building is almost ready to take major parts of the accelerator and its associated equipment. We thank Prof. Ehrenberg (Mainz) for his kind loan of the electron gun and Prof. Citron (Karlsruhe) for his help with the Cryoperm foil. Prof. G. Weber and H. Lagerpusch from the Solid State Physics Institute provided us generously with liquid helium. We are also grateful to J. Foh, G. Fleck, G. Kaster, W. L6w, J. Mushak, H. Pl6sser, H. Schmidt and S. Simrock for their help at various stages of the work.

References [11 J.P. Turneaure and Ira Weissman, J. Appl. Phys. 39 (1968) 4417. [2] C.M. Lyneis, M.S. McAshan, R.E. Rand, H.A. Schwettman, T.I. Smith and J.P. Turneaure, 1EEE Trans. Nucl. Sci. NS-28 (1981) 3445. [31 P. Axel, L.S. Cardman, H.-D. Gr~f, A.O. Hanson, R.A. Hoffswell, D. Jamnik, D.C. Sutton, R.H. Taylor and L.M. Young, IEEE Trans. Nucl. Sci. NS-26 (1979) 3143. [4] A. Citron, G. Dammertz, M. Grundner, L. Husson, R. Lehm and H. Lengeler, Nucl. Instr. and Meth. 164 (1979) 31. [5] L.M. Bollinger, Proc. National Accelerator Conf. (1983) IEEE Trans. Nucl. Sci NS-30 (1983) 2065.

[6] C.M. Lyneis, H.A. Schwettman and J.P. "l'urneaure, Appl. Phys. kerr. 31 (1977) 541. [71 U. Klein and D. Proch, Proc. Conf. on Future possibilities for electron accelerators, eds., J.S. McCarthy and R.R. W hitney, Charlottesville, Va. USA (1979 } p. N l - 17 [8] H. Piel, Proc. Workshop on Rf superconductivity, cd.. M. Kuntze, KfK-Report 3019, Kernforschungszentrum Karlsruhe (1980) p. 85. [9] Ph. Bernard, G. Cavallari, E. Chiaveri, E. Haebel, H. Heinrichs. H. Lengeler, E. Picasso, V. Piciarelli, .I. Tiickmantel and H. Piel, N ucl. Instr. and Meth. 190 ( 1981 ) 257. [10] H.A. Schwettman, J.P. Turneaure and R.F. Bates, J. Appl. Phys. 45 (1974) 914. [11] H. Padamsee, J. T0ckmantel and W. Weingarten, IFEE Trans. Magnet. MAG-19 (1983) 1308. [12] H. Heinrichs, U. Klein, G. MUller, H. Piel, D. Proch, W. Weingarten, H. Genz, H.-D. Gr~.f, T. Grundey, A. Richter. E. Spamer, Lecture Notes in Physics 108 11979) 176. [13] H.D. Gr~.f, H. Miska, E. Spamer, O. Titze and Th. Walcher. Nucl. Instr. and Meth. 153 (1978) 9; D. Sch0ml, J. Foh, H.D. Gr~if, H. Miska, R. Schneider, E. Spamer, tt. Theissen, O. Titze and Th. Walcher, ibid., p. 29. J. Foh, R. Frey, R. Schneider, D. Scht~ll, A. Schwierczinski, H Theissen and O. Titze, ibid., p. 43. [14] M. Tigner and H. Padamsee. CLNS-82/553, Cornell (1982), to be published in ATP Conf. Proc. of SLAC Summer Accelerator School (1982). [15] B. Hillenbrand, H. Martens, H. Pfister, K. Schnitzke and Y. Uzel, IEEE Trans. MAG-13 (1977) 49l. [16] H.C. lloyt, D.D. Simmons and W.F. Rich, Rev. Sci. Instr. 37 (1966) 755. [17] K. Halbach and R.F. Holsinger, Part. Accel. 7 (1976) 213,