A multipole wiggler magnet for DORIS

Nuclear Instruments and Methods 208 (1983) 163-166 North-Holland Publishing Company

A MULTIPOLE

WIGGLER

MAGNET

163

FOR DORIS

P. G I S R T L E R HAS YLAB at Deutsches Elektronen-Synchrotron DESY, 2000 Hamburg 52, Germany A. JACKSON * Daresbury Laboratory, Warrington WA4 4AD, England

To increase the photon flux of synchrotron radiation from the VUV up to the X-ray region ( - 1 ,&), a multipole wiggler device is currently under development at HASYLAB for installation in the storage ring DORIS. The magnetic array consists of Rare Earth Cobalt (REC) permanent magnet blocks with four blocks per period. The magnetic field of 0.45 T gives a critical wavelength of - 3.4 A at 3.5 GeV. For a 10 period device, the gain in photon flux with respect to a bending magnet will be a factor of - 40 at 1200 A (3.5 GeV) and - 12 at 1.5 A (5 GeV).

1. Introduction

HASYLAB ,

Currently there is great interest amongst users of s y n c h r o t r o n radiation at - 1.5 A to o b t a i n higher photon fluxes at their samples. One way to achieve this is to use long curved mirrors to collect a large angular spread of b e a m from o r d i n a r y b e n d i n g magnets a n d focus it o n t o the sample. A n o t h e r m e t h o d which is being persued at several s y n c h r o t r o n radiation laboratories is to use a multipole wiggler magnet which produces radiation into a small opening angle with an intensity which increases linearly with the n u m b e r of poles. Such a device is currently u n d e r d e v e l o p m e n t at H A S Y L A B for installation in the storage ring DORIS. This p a p e r describes the magnet, the radiation it produces a n d discusses the effects it will have o n the operation of DORIS.

2. The multipole wiggler magnet The magnet will be installed in the straight section of D O R I S which includes the electron injection elements. This site was chosen because: a) there is sufficient free space available to install a magnet > 1 m long, a n d b) the radiation p r o d u c e d by the m a g n e t will shine into a n area of H A S Y L A B available for d e v e l o p m e n t (see fig. 1). In order to service experiments requiring short wavelengths ( _< 10 ,~) the field in the m a g n e t should be m a d e as high as possible. However, space restrictions in the

* DESY/HASYLAB summer visitor 1982. 0 1 6 7 - 5 0 8 7 / 8 3 / 0 0 0 0 - 0 0 0 0 / $ 0 3 . 0 0 © 1983 N o r t h - H o l l a n d

i i

wiggler

ex#erim

station

I i

DORIS (quadrant I )

.

.

.

.

.

.

.

Fig. 1. A schematic overview of the HASYLAB-Hall and the first quadrant of the DORIS-storage ring showing the site chosen for the wiggler and the position of the wiggler beam line in the experimental hall. chosen area make the use of electromagnets impractical a n d so a solution utilising rare earth cobalt (REC) p e r m a n e n t magnet arrays has been sought. A m a g n e t c o m p r i s i n g the c o n v e n t i o n a l four b l o c k s / p e r i o d array produces a sinusoidal field with a peak value B 0 given by [1]: B 0 = 1.80 B r ( l - e 2rrL/X°) e -~h/x°, where B r is h is L is )~o is

the the the the

r e m a n e n t field of the magnet material, gap, block height, a n d magnet period.

II. W I G G L E R S / U N D U L A T O R S / O T H E R DEVICES

P. Giirtler, A. Jackson / Multipole wiggler magnet

164 s~ W

/ I

/'"

/-"1 I

e-Ih i J- t I

I

ho

Fig. 2. A schematic drawing of the permanent magnet blocks in the wiggler. One period is shown with the magnetic orientation of the blocks.

Fig. 2 shows a schematic view of a magnet period. Therefore in order to maximise B 0 it is necessary to select a material with high remanent field and design the magnet with a small gap and large block height, The value of X0 is a compromise between the two exponential terms. There are however constraints on the design, the most important being on the gap and on the volume of R E C material. The minimum aperture is determined from the vertical beta f u n c t i o n , / 3 , of D O R I S II [2] and the restricting aperture of the vacuum vessel in the main dipole magnets (50 mm). This gives an aperture at the wiggler magnet of 37.5 m m for the standard synchrotron radiation optic for H A S Y L A B and 32.0 m m for the colliding beam optic. It is proposed that the magnet be built outside the vacuum chamber (for ease of removal and field manipulation), so an allowance must be made for c h a m b e r walls and tolerances, bringing the minimum gap to 42 ram. Fig. 3 shows the fields which can be achieved from a

~ ~= /,2 mira

I

I

Br =095T

0.6

I~- 0.t, mo

0.2

I 40

I

l 80

I ~

I

I

L=~

(=Xo)

magnet array with four b l o c k s / p e r i o d and a gap of 42 mm, as a function of period, X0, and block height. The curves assume a remanent field of 0.95 T which is readily available in blocks of samarium cobalt. Also shown in this figure are curves of constant K, the deflection parameter, which defines the opening angle of the radiation through c~ = K / y . Note that when K = the electron energy in GeV, the full opening angle -- 1 mrad. Thus in order to achieve the maximum magnetic field one should choose the maximum possible period and maintain L - )t 0. However, for a magnet with fixed overall length the number of periods must decrease as )t o increases defeating the very purpose of the magnet. Also increasing the block height makes the required volume of R E C prohibitive. The compromise reached for the D O R I S wiggler is a magnet with ?t o = 120 mm, L = X0/4 (i.e. square crosssection blocks), and with a width of 100 m m to give an acceptable field uniformity at the centre of the aperture. This gives a sinusoidal field with a peak value of 0.45 T, and with 10 periods gives the intensity spectra shown in fig. 4. These spectra were calculated by dividing a full wiggler period into 22 segments and assuming each segment to have a uniform field given by B , = B 0cos

22

,

n = 1,22.

The spectrum from each segment was then calculated using the H A S Y L A B code " S P E C T R A " and these were summed to obtain the final curves. N o interference (or undulator) effects have been assumed. This assumption

1013

I

I

I

T

I Mulhpole

1

I

Wiggler 110 Per,o~s )

L X/~ n

10~2

o~

/

/

5 eev

. . . .

10" I

I 120

I

I 160

0

X o (mm)

Fig. 3. The magnetic field B0 as a function of the periodic length X0 with a gap height of 42 mm and a remanent field of 0.95 T. The shaded area is obtainable with different block heights L varying from Xo/4 to oe (which is nearly reached for L ~ X0). Also drawn are curves of constant K which means a constant deflection angle at a specific energy. The cross at Xo = 120 mm indicates the condition we have chosen for the wiggler.

101°

01

1

10 X (A)

i

i

100

1000

Fig. 4. The intensity of the radiation of the wiggler at 3.5 and 5 GeV. The spectra from a bending magnet are shown for comparison. The bars indicate the harmonics of the "undulator wavelength" where peaks due to interference effects could arise.

P. Giirtler, A. Jackson / Multipole wiggler magnet

165

Table 1 The intensities in photons/s-mA-mrad in 0.1% bandwidth and the gain obtained with respect to 1 mrad from a bending magnet at several wavelength (energies). 2, [J,]

1250 125 12 2.5 1.5 1.0

E [keV]

0.01 0.1 1 5 8 12

Intensity x 10 lo, at 3.5 GeV

at 5 GeV

Gain b at 3.5 GeV

at 5 GeV

90 180 250 100 40 15

75 150 250 190 130 80

45 43 33 14 7 4

38 36 29 17 12 8

a Intensity in photons/s- mA. mrad in 0.1% bandwidth. b Gain with respect to 1 mrad from a bending magnet.

is reasonable in the i m m e d i a t e area of interest ( _< 10 A), however at the f u n d a m e n t a l u n d u l a t o r wavelength (172 A at 3.5 GeV) a n d h a r m o n i c s up to the order - 10, the spectrum will be e n h a n c e d by interference effects. The positions of these h a r m o n i c s are m a r k e d on the graph. Spectra from the D O R I S dipole magnets are shown for comparison. The immediate gains, assuming that an experiment c a n n o t collect more radiation t h a n is given by the opening angle, ( = K / V ) , are given in table 1, for some wavelengths of particular interest at H A S Y L A B . The total power emitted by a 100 m A electron b e a m passing t h r o u g h the m a g n e t is: 189 W into 1.4 m r a d (horizontal) at 3.5 GeV; 386 W into 1.0 m r a d (horizontal) at 5.0 GeV. By opening the m a g n e t gap it will be possible to operate the wiggler at a lower peak field. This has two immediate benefits for users in the V U V region of the spectrum. a) It will increase the p h o t o n flux still further, a n d b) it will eliminate the u n w a n t e d higher energy radiation. It is clear that the wiggler is a very versatile source of s y n c h r o t r o n radiation indeed.

3. The effects of the multipole wiggler on D O R I S The effects on D O R I S can be divided into three parts: a) Since the wiggler comprises a series of parallel ended magnets it will produce vertical focusing a n d so modify the /3z function, the vertical b e t a t r o n tune, Q , and any vertical closed orbit distortions. b) The magnet increases the electron energy l o s s / t u r n , so changing the synchronous phase angle a n d the radiation d a m p i n g times, which in turn alter the electron b e a m size. c) Like all other magnets it is a potential source of field

error, the most i m p o r t a n t arising through f B d l ~: 0, giving rise to radial closed orbit changes. All these effects have been investigated and shown to be negligible to first order of m a c h i n e operation. A t 3.5 GeV, where the effects are most pronounced, the relev a n t p a r a m e t e r changes are: C h a n g e in vertical b e t a t r o n tune: AQz = 1.0 x 10 -3 for H A S Y L A B optic, = 0.9 x 10 -3 for colliding b e a m optic. Increase in radiation loss: A U = 1.89 k e V / t u r n ,

A U / U = 5 X 10 3.

Residual dipole field:

f Bdl=

1.7 x 10 -3 T - m,

which gives a net deflection of Ax' = 0.15 × 10 -3 rad.

4. Conclusions It has been shown that a R E C multipole wiggler m a g n e t can be inserted into D O R I S without u n d u e disturbance. The radiation spectrum it produces will benefit users who work with wavelengths > 1 ~,. Table 1 tabulates these gains for specific regions of interest. W i t h a m e c h a n i s m which permits the magnet gap to be increased the spectrum can be specifically tailored for V U V experiments a n d indeed, in this m o d e the m a g n e t m a y begin to show interference effects which will enh a n c e the spectrum still further. F o r future d e v e l o p m e n t it m a y be possible to increase the field in the wiggler by concentrating the field with ferrous pole pieces thus extending the useful range of the device into the s u b - A n g s t r o m region of the spect r u m [3]. Presently the engineering problems associated with the design of the wiggler are being pursued. II. W I G G L E R S / U N D U L A T O R S / O T H E R DEVICES

166

P. Gi~rtler, A. Jackson / Multipole wiggler magnet

This work has drawn heavily on the report, ref. 1, and we should like to thank the authors for their assistance during the course of the initial design work. Thanks are also due to the staff of HASYLAB, in particular Dr. W. Graeff, the author of the program SPECTRA, without whose help the production of the all important spectral distribution would have proven very time consuming, and also Prof. Kunz for many useful discussions. We especially acknowledge the help of Dr. H. Nesemann and Dr. K. Wille from the D O R I S machine group. One of us (A.J.) would also like to thank Prof. Kunz and Dr. Koch for giving him the opportunity to work

on this project and also for the hospitality of all staff during his stay at HASYLAB.

References [1] M.W. Fan, M.W. Poole and R.P. Walker, Some aspects of the design of plane periodic permanent magnets for use in undulators and free electron lasers, DL/SCI/TM30A. [2] K. Wille, DESY-Report 81-047 and private communication. [3] K. Halbach, Design of focussing and guide structures for charged particle beams using REPM, Proc. 5th Int. Workshop on REPM, Roanoke (1981).