A HIGH RESOLUTION WIRE SCANNER FOR MICRON-SIZE PROFILE MEASUREMENTS AT THE SLC

Transcription

1 c ;- SLAC-PUB-4605 LBL UM-HE Apri 1988 P/A) A HIGH RESOLUTION WIRE SCANNER FOR MICRON-SIZE PROFILE MEASUREMENTS AT THE SLC R. FULTON, J. HAGGERTY, R. JARED, R. JONES, J. KADYK Lawrence Berkeey Laboratory, Berkeey, CA C. FIELD, W. KOZANECKI Stanford Linear Acceerator Center, Stanford University, Stanford, CA W. KOSKA University of Michigan, Ann Arbor, MI ABSTRACT _ Fine conductive fibers have been used to measure transverse beam dimensions of a few microns at the Stanford Linear Coider (SLC). The beam profie is obtained by scanning a fiber across the beam in steps as sma as 1 micron, and recording the secondary emission signa at each step, using a charge sensitive ampifier. We first outine the mechanica construction and the anaogue eectronics of the wire scanner. We then describe its performance in test beams and in actua operation. The paper coses with a brief discussion of performance imitations of such a beam profie monitor..- _T. --c- Submitted to Nucear Instruments and Methods _- *Work supported by the Department of Energy, contracts DE-AC03-76SF00515, DE-AC03-76SF00098 and DE-AC02-76ER01112.

2 c.,c- I. Introduction The tuneup of the Stanford Linear Coider (SLC) [] presents a chaenge to measure beam sizes on a scae not previousy encountered in high energy physics. The goa is to achieve bunches of eectrons and positrons with approx- imate rms dimensions of 2 pm transversey and 1 mm ongitudinay. Coisions of beam bunches with such sma dimensions have never before been attempted, and commony used techniques (such as phosphor screen monitors) have a much too coarse resoution _. _- The device described here is a wire scanner that uses a very thin carbon fiber. This fiber is moved in uniform steps across the beam path whie measuring the eectrica signa generated by secondary emission [2]. Wire diameters of approximatey 7 and 25 pm have been used in SLC beams. Athough this technique has been used at severa high energy acceerators [3-81, we are not aware of its previous use at such a sma scae of dimensions. Both the beam profie and the beam position are accuratey determined by digitizing the wire position in steps as sma as 1 pm, whie recording an eectrica signa whose ampitude is proportiona to the amount of beam striking the wire. Other detection methods have been considered [3-41, such as the measurement of scattered beam partices or change in resistance due to temperature rise in the fiber, but-at these beam energies and repetition rates-secondary emission appeared to be the simpest method. Some new questions arise regarding the viabiity of the wire scanner in very intense beams. The SLC design cas for 5 x Oo eectrons or positrons in a beam bunch of the very sma dimensions given above. The heating which resuts from impact of such intense bunches may cause damage to or break the fiber. Carbon was chosen as the fiber materia because the energy oss per unit ength due to a crossing beam partice is reativey sma, the meting (or subi- - --*- mation) point is quite high, and the mechanica properties are otherwise good. The resuts described here were obtained in the tuneup phase of the SLC with the beam intensity we beow its design vaue, so viabiity of the fibers has not been put to a significant test.

3 ,c- II. Mechanica Design Figure 1 shows the SLC ayout in pan view, with the wire scanner ocations denoted. On each SLC machine puse, eectron and positron beam bunches trave the fu ength of the Stanford inac, about 3000 m, gaining an energy of about 50 GeV per partice. The design intensity is 5 x Oo partices per bunch, but in the tuneup phase, intensities of about 3 x 10 were typica. At the end of the inac the transverse radius of each bunch is about 90 pm, and it is here that the eectron and positron bunches are separated by a magnet; then each is transported about 1300 m around an arc section to the fina focus section. In the fina focus the beam size is reduced to its minimum vaue at the coision, or Interaction Point (IP), of the eectron and positron bunches. - Here, in a vacuum of about o- Torr, the fina focus wire scanners must operate. At the IP, three wire scanners were used, each having a measuring head with two carbon fibers of different diameters: about 7 pm and 25 pm. Each wire scanner measured profies in one of three transverse directions: X (horizonta), Y (vertica) and V (at 45 to X and Y). T wo of the three measuring heads (X and Y) coud be inserted into the beam simutaneousy, by virtue of a 1 mm offset in position aong the beam direction (2). A d rawing showing the arrangement of the three wire scanners on their support pate is shown in fig. 2. A photograph of a measuring head is shown in fig. 3, with wires instaed. The carbon fibers, which are very britte, were drawn around a curved surface of radius 4.5 mm on the ceramic hoder to minimize the possibiity of breakage. They are not mounted in a singe pane norma to the beam axis, but are offset in 2 by about 350 pm, and spaced ateray (as seen by the beam) by 200 pm. The surface of the ceramic hoder has been rendered sighty conductive by rubbing graphite on it; this ~ _- prevents buidup of charge on the surface. Lead-out wires were kept as short and -4 _ as neary perpendicuar to the beam direction as possibe, to minimize the beam induction signa. These eads were insuated with ceramic beads, since commony used (e.g., organic) insuators are unsuitabe in a high vacuum environment and are aso quite vunerabe to radiation damage. 3

4 c,c- The measuring head is attached to a support shaft, and eads from each end of the fiber are brought to feed-through insuators in a housing mounted on the actuating stage of a commerciay avaiabe inear positioning device [9]; the stage is moved by a stepping motor in 1 pm steps. The vacuum sea is made by ceramic insuated feed-throughs (BNC) weded into the wa of the housing, which is in turn couped to the vacuum pipe using a beows of nomina 3 inch ength, aowing a tota actuator motion of about 30 mm. Lead weights counterbaance the 5 kg vacuum oading on the beows. The maximum residua oading of 3.2 kg, due ony to the beows spring constant, is we within the manufacturer s 5 kg rating. Microswitches define the IN and RETRACTED positions, and interock the stepping motor controers. The norma running - speed is 0.5 mm/set, corresponding to a 500 Hz stepping rate. About 45 seconds are required to insert the device to a PARK position cose to the beam from which the actua scan is initiated. A typica beam scan takes about seconds. The entire system is competey under computer contro and may be operated from the Main Contro Room by the SLC operators, or from any of severa sateite ocations. Another version of the wire scanner has been buit for the Reverse Bend (RB) ocations in the SLC arcs (fig. 1). At each of these ocations the beams are nominay circuar with about 100 pm radius. Focusing on the wire in either the X or the Y pane is expected to reduce the size to about 12 pm, and from this size the beam emittance can be determined. Moreover, verification can be made of the aignment of the beam optica axes with the X and Y panes, as we as the absence of any X-Y correation. The design aowing this additiona measurement capabiity, not present in the fina focus unit, is iustrated in fig. 4. G The measuring head in this case supports the two wires of 25 pm diameter with a-- a caibrated spacing (- 200 pm). Whie in the case of the fina focus wire _- scanners the norma operation consists of mechanicay scanning the wire past the beam, the measurement with the RB wire scanners is performed by defecting the beam in stepwise increments past the wire, using sma steering magnets. The 4

5 stepping motor drive is used to insert and retract the pair of wires, and to rotate them about the beam axis after the measuring head reaches the fuy inserted position, as indicated by an eectrica contact. The gear and ratchet mechanism which accompishes this is seen in fig. 4. The caibrated wire spacing aows a direct caibration of the defection magnet sensitivity, setting the scae for profie measurements. The wire rotation aows the beam waist measurement to be made in the directions of the X or Y axes, or any intermediate direction if X-Y correations are present. - III. Eectrica and Eectronic Design - A. Preampifier and CAMAC Modue Signas from each of the scanner wires are taken via 3 m engths of RG223 cabe (doube-shieded) to a preampifier which is competey surrounded by an eectrostatic encosure. In addition, the preampifier is contained inside a heavy ead shied which protects it from ionizing radiation. The circuit bock diagram -. is shown in fig. 5. The input signa from the wire is buffered by a charge sensitive preampifier with provisions for pacing a bias votage on the wire and injecting a charge for the purpose of caibration. The signa from the preampifier is ampified and shaped with a peaking time of about 3 ps. The output is sent via 125 m of shieded muticonductor twisted pair cabe to a CAMAC modue which serves as an interface to the rest of the digitization and contro modues, having connections for: caibration, d.c. wire bias, stepper motor contro, encoder readout and anaogue wire signa output. This signa is digitized by a CAMAC anaogue-to-digita converter (ADC). _._T. r.-- 5

6 B. Signa*-Response Fundamenta to the approach is the response of the eectronics to the two types of input signa as iustrated in fig. 6. As the beam approaches the wire, a charge is induced that resuts in a current fow to the preampifier; this occurs whether or not the beam strikes the wire. As the beam eaves the vicinity of the wire, the induced charge is removed, resuting in a current equa in magnitude, but opposite in direction to the first impuse. The separation in time of the two fows is of the order of a few picoseconds. The resuting shape (a differentia of a Gaussian) due to the doubet input is shown in fig. 6. There is an additiona signa when the beam does strike the wire: the energy oss process in the wire resuts in the emission of secondary eectrons. The resuting charge impuse is ampified to produce the Gaussian shape shown. To maintain the inear response of the preampifier, essentia to this approach, a 1 nf shunt capacitor was added to attenuate the votage extremes of the doubet input. In operation the anaogueto-digita converter sampes the wave form at the crossover region of the doubet response. As is shown in the figures, this sampe time wi correspond to a maximum in the desired response to the secondary emission of eectrons. The wire signa digitization is done in a LeCroy 2249W ADC, gated by programmabe gate deay and width units. The gate timing is provided by the SLC timing system. The d.c. response is a current of about 12 PA to the ADC per pc of charge at the preampifier input, or with a 0.5 ps gate width, 6 pc output to the ADC (channe 24). As an exampe, if the effective coefficient for secondary emission is about 2010, and if 1 x 10 eectrons (or positrons) struck the carbon fiber, then a 20 pc charge woud be digitized by the ADC (in channe 80). -. _ C. Position Readout Positiona readout is provided by an incrementa encoder on the stepper motor drive [9]. This encoder digitizes the optica signas produced as a circuar grating attached to the motor drive rotates in front of a set of photoces. The encoder is sensitive to the direction of motion, and the entire system has a nom- 6

7 i,c- ina accuracy of one step, or 1 pm. In addition; there is a inear potentiometer. connected to the stage motion, having an accuracy of roughy 20 pm. This is primariy to provide a means of detecting possibe mafunctions in the digita readout. Both of these readouts shoud agree with the number of steps requested by the command program. D. Caibration and Fiber Bias The eectronic gain of the preampifier can be determined by injection of a caibrated votage puse into the CAMAC modue. There is aso provision for introducing a negative d.c. bias votage on the wire at this modue, a technique which has been used to enhance the signa according to severa pubished reports [4,5,7]. The basis of the effect is beieved to be that some fraction of the very ow energy eectrons knocked out of the wire may return to the wire due to oca eectric fieds, uness repeed by a sma negative bias on the wire (usuay - 30 vots is sufficient). As wi be described beow, no effect due to the bias was found in tests with our wire scanner. - IV. Tests: Probems and Soutions A. Bench Test of Beam Induction Efect As mentioned earier, the response to a fast unipoar impuse (i.e., asting < 3,~s)~is a bipoar puse of - 3 /.LS zero-crossing time. The induction due to the fied of the passing beam bunch, which has about a 1 mm ength (rms), creates just such an impuse on the scan wire, asting ony about 10 ps. For a bunch cose to _._T. the wire the ampitude of the induction signa can be many times the secondary emission signa. &.re to its transient behavior, however, no signa from this effect _- shoud contribute to the measurement of the d.c. eve at the samping (crossover) time at 3,xs. Nevertheess, there was concern that some residua effects might perturb the desired signa (e.g., due to non-inearities in the eectronics system), 7

8 i,c- since the-induction ampitude coud be much arger. Measurements were made. of the beam induction pickup using a simuation of the beam bunch by a fast (- 300 ps) puse on a thin wire strung through the actua wire scanner housing. The actua preampifier system was aso used. The induction puse was virtuay unmeasurabe at simuated beam eves of - Oo e-/puse. When the scanning wire was oriented parae to the beam, however, a distinctive induction signa was observed, at east 10 times arger than for the norma wire orientation. This is as expected, and ed to carefu pacement of signa ead-out wires as neary perpendicuar to the beam axis as possibe. Another feature is aso advantageous: The induction signa is argey composed of very high frequency components, and that part generated at the fiber tends to be fitered out by the capacitances associated with the preampifier input and the natura resistivity of the carbon fibers (- 9 KR for the 2.6 cm ength of 7 pm fiber from the eectrica connection to the center of the fiber). B. Beam Tests - _- Initia testing was performed in a 46 GeV eectron beam at the end of the inac, with intensities of about 5 x 10 e-/puse and with 5 puses/set. The drive system was controed by a program running under BASIC on an IBM PC. Severa crucia facts became apparent during these tests. Most importanty, this test area was not we shieded from stray partice fuxes, and aso contained arge amounts of eectromagnetic radiation background, particuary cose to the beam pipe. It was soon apparent that the preampifier needed to be paced away from the beam pipe (- 2.5 m) and inside a ead-brick house to shied against partice background. In addition, it was found necessary to have doubeshieded coax signa eads to protect against EM radiation. The combination of - fir- these improvements reduced the background eve by a factor of It was aso discovered that the signa cabe must have soid-core insuation, as partice backgrounds produced gas ionization in the air-core cabe used initiay. This was verified by appying a bias votage of about 100 vots, first with positive poarity, 8

9 i,; then negative, directy to the air core cabe. There was a arge background signa. observed with either poarity, presumed to be due to coection of positive ions or eectrons resuting from the ionization of the air core of the cabe. In the case of soid insuation no such effect was observed. After these improvements a profie was easiy measured using copper sensing wires, having diameters of 500 pm and 150 pm, comparabe with the measured beam size of pm rms (fig. 7). V. Experience Using Wire Scanners in the Tuneup of SIX The wire scanner has been used to measure X and Y profies of both e- and e+ beams at the IP. Exampes of typica profie pots are shown in figs Each point represents the measurement of a singe beam puse. The beam profie is obtained from a sequence of such measurements as the wire is scanned across the beam path. The vertica scae represents the wire signa response to a singe SLC beam puse, in ADC channe number, and the wire position is potted in pm on the horizonta axis. The pm transverse spacing between the two wires is apparent in fig. 8, which shows the response from both wires on the same beam scan. In this case, the scan was done in the X direction, and the resuting profie width is about 15 pm rms. It wi be noticed that the peak response of the arge wire is about eight times that of the sma wire, whie the ratio of beam intensities subtended by the wire diameters is ony about 3: for a Gaussian shape. This effect is currenty being studied by investigation of the detais of the SEM mechanism. Figure 9 is an exampe of a measurement of the X and Y profies on the sma wire of an eectron beam, taken ess than one minute apart, showing rms sizes 8.3 and 6.5 pm, respectivey. The sizes quoted represent the widths of Gaussian distributions fitted to the raw data. They therefore are _ the convoution of the actua beam size, the puse-to-puse beam position jitter (if any), and the resoution effects due to the finite wire size. Figure 10 shows X and Y scans of positron and eectron beams which were brought to the IP simutaneousy, but which were not coiding at the time these scans were taken. 9

10 i This fi-gure aso ceary iustrates how the wireirscanning technique can be used. to bring the two beams into coision. VI. Sense Wire Response A. Secondary Emission The principa mechanism of secondary emission produces soft eectrons from distant coisions (i.e., arge impact parameters) with the passing projectie partice. A beam partice produces a secondary eectron in a few percent of the coisions, based upon previous reports [3-71. In the present device, we observe secondary emission efficiencies which depend upon wire diameter and aso upon beam partice, e- or e+. It is about 1% for a e- beam with the 25 pm diameter fiber. Based upon ony one instance where sufficient data was avaiabe, it appears -that the efficiency for e+ is significanty higher than for e-. A difference in response between e- and e+ may we be expected from the effects of the eectric fied of the beam bunch upon the emitted secondary eectrons. Another effect, observed by others but not seen by us is that of signa enhancement by pacing a sma (- 30 vot) negative d.c. bias on the wire: as much as a factor of two signa increase has been reported [5]. Th is is expained as a repusion from the wire of some very ow energy secondary eectrons which woud otherwise return, due to a negative ambient potentia in the neighborhood. In the present case, it seems ikey that the intense eectric fied from the SLC bunches (up to severa MV/cm) dominates the oca potentia so there is itte infuence from a reativey sma d.c. bias votage. - 10

11 B. Heating of the Carbon Fibers In thin fibers, part of the ionization energy oss escapes in the form of deta rays [8]. The greater part, however, remains to heat the fiber and very simpe considerations show that the carbon subimation point woud easiy be exceeded in a singe SLC puse fuy focussed onto the fiber with design intensity and spot size. At ower intensities, therma shock may cause faiure, and the threshod for this is very uncertain. During the SLC deveopment period discussed here, the beam intensity was aways ess than 5 x 10 partices per puse, the transverse rms dimensions were greater than 5 microns, and a fibers survived. Under these conditions the maximum temperature of the fiber is cacuated to be ess than 300 degrees C. Between SLC puses the fibers coo by conduction primariy, with some radiation cooing at higher temperatures. This shoud be true up to the fu 120 Hz repetition rate, but for this report the SLC operating frequency was ony 5 Hz. - VII. Summary A wire scanner making use of secondary emission from very thin carbon fibers has been used successfuy in the measurement of beam profies at the Stanford Linear Coider. Beams as sma as 5 pm rms have been measured. After proper shieding techniques were impemented, the instrument has proven to be quite reiabe and was used routiney in the tuneup of the fina focus section of then coider. Acknowedgments Many peope have contributed generousy their effort and interest to make the wire scanner a successfuy operating instrument. We wish to acknowedge in particuar Nan Phinney, Tony Gromme and Joanne Bogart for their work on the software, and Dan Wright, for his hep on vacuum probems and other 11

12 instaation work. The ideas, assistance and support of Gerry Fischer, Roger. Erickson, Ted Fieguth and Phiip Bambade were very vauabe, and are gratefuy acknowedged, as we as other ideas received from others who had aso worked on wire scanners, especiay R. Jung and L. Evans at CERN, and J. Seeman at SLAC. References [] SLC Design Handbook, December 1984, Stanford Linear Acceerator Cen- ter, Stanford, Caifornia [2] J. S. Aen, Phys. Rev. 55, 336 (1939); A. G. Hi et a., Phys Rev. 55, (1939). [3] J. Bosser et a., Nuc. Inst. & Meth., A235, (1985). [4] R. Jung and R. J. Cochester, CERN/LEP-BI/85-160, [5] J. Kider and C. Hojvat, Fermiab-Pub-85/176, _. - [6] W. T. Weng et a., IEEE Transactions on Nuc. Sci., NS-30, 2331 (1983). [7] R. Chehab et a., LAL/RT/85-05 (ORSAY), [8] J. Bosser et a., CERN SPS/86-26 (MS), Our estimate of energy carried away by deta rays is significanty smaer than the 70% quoted in this reference. [9] The stage, stepping motor and digita encoder were suppied as a unit by Kinger Scientific Corporation, Richmond Hi, New York The mode used was: UTOO-75PP. C -.L 12

13 Figure Captioos Fig. 1. Layout of Stanford Linear Coider (SLC). Wire scanners were buit for the Interaction Point (IP), and the North and South Reverse Bend (RB) ocations. Fig. 2. Schematic of the arrangement of the three wire scanners mounted at the IP ocation. Some components are iustrated for the X-motion scanner. Fig. 3. Photograph of a measuring head used at the IP. Two measuring wires, 7 pm and 25 pm in diameter, are ceary visibe. Fig. 4. Photograph of an RB Wire Scanner assemby. Evident are the components iustrated in Fig. 2, as we as the ratchet and gear mechanism used to rotate the measuring head whie in the inserted position. Fig. 5. Bock diagram of eectronics for wire scanners. F.ig. 6. Response of eectronics to beam-induced signa and to the secondary emission signa. L Fig. 7. First successfu measurements of a beam profie performed at the end of the inac. (a) Profie taken with a 500 pm diameter wire. (b) Profie taken with pm diameter wire. Both wires were made of copper, and the beam size was approximatey 500 pm rms, as determined by a scan using a phosphor screen. Fig. 8. Beam profie measurement in X direction at the IP. (a) Profie taken with - 25 pm diameter fiber; (b) Profie with 7 pm diameter fiber. Both fibers are carbon fiaments. The pm transverse spacing between the fibers is evident. The beam size is about 15 pm rms. Fig. 9. Profie scans in X and Y directions of the beam taken ess than one minute ~- -- apart using the smaer fiber. The beam size was 8.3 pm in X and 6.5 pm _- in Y, rms. Fig. 10. X and Y scans whie both eectron and positron beams were present at the IP, but not coiding. These scans were taken with the arge fiber. 13

14 i,c- Interaction / Point (IP). Positron production target II L Existing inac -..._. Positron - return ine C 4 /B - Damping rings Existing inac +--- Eectron gun Fig. 1 XBL

15 Stepper mot01 1 I 1 I I I X moton we scanner I Mrcroswitches / (stage Imit) Y motion we scanner Fig. 2 C

16 Fig. 3

17

18 PREAMPLIFIER AMPLIFIER n CALIBRATION - I XBL Fig. 5 _ ^T. ~-. -e- _-

19 BEAM INDUCED 1 INPUT SIGNAL r AMPLIFIER RESPONSE TO BEAM INDUCED---- INPUT SIGNAL +--- SECONDARY EMISSION INPUT SIGNAL 1 1 AMPLIFIER RESPONSE TO SECONDARY - EMISSION INPUT I SIGNAL I I I ADC SAMPLE XBL Fig. 6 C

20 25. on ai 0) z 6 10 (a> It I -r I I I * o X scan position (mm) i i 31.5 (b) T -r X scan position (mm) Fig i.5 XBL C

21 I I I I. 80 L!? g 60 s c 0 0 $ 40 X scan position, microns XBL x a 8 01 I I i o X scan position, microns Fig. 8 XBL

22 25. cn 5 E 20 $ a 1oc 5+ e-e..j t eee eee I X scan position, microns I 3 25r---=J I f I I I I I Y scan position, microns - -Fe-- Fig. 9 XBL

23 ,c- I I I I I I I I e (a). cn 5 E ee e- eee pepee 1 e ee I ' X scan position, microns -.._ g 200- s e+ I UN z _- - -Fe-- Y 0 9 oo- e- ee e - e e@e *eeeee * eeeeeep I 4 o- I OO- -- scan position, microns Fig. 10 XBL