- Email: [email protected]
A rudimentary electron energy analyzer for accelerator diagnostics
Nuclear Instruments and Methods in Physics Research A 453 (2000) 507}513
A rudimentary electron energy analyzer for accelerator diagnostics R.A. Rosenberg*, K.C. Harkay Advanced Photon Source, Argonne National Laboratory, 9700 South Cass Avenue, Argonne, IL 60439, USA Received 6 March 2000; accepted 22 March 2000
Abstract We have constructed a compact, planar retarding "eld analyzer for the diagnostics of low-energy, background electrons in a high-energy particle accelerator. Bench measurements of the analyzer have been made to characterize it, and the results are reasonable in light of models of this type of analyzer. Comparisons to results obtained using a beam-position monitor (BPM) show the advantages of this analyzer for electron diagnostics. Sample results from analyzers installed at the Advanced Photon Source storage ring at Argonne National Laboratory and the Proton Storage Ring at Los Alamos National Laboratory show how the analyzers can be used for studying the intensity, energy, and time structure of electrons in an accelerator environment. 2000 Elsevier Science B.V. All rights reserved. PACS: 29.30.Aj; 29.30.Dn; 29.30.Ep Keywords: Electron analyzer; Accelerator; Diagnostics
1. Introduction Electrons in accelerators are ubiquitous. They can be produced directly by irradiation of vacuum chamber surfaces by X-rays, ions, and other particles or indirectly by bombardment of electrons, which leads to production of secondary electrons. Under many circumstances, electrons are not detrimental to accelerator performance; however, there are operating conditions that can lead to ampli"cation of the electrons. If the number density of the electrons becomes su$ciently large, they can lead to degradation of the particle beam either by direct * Corresponding author. Tel.: #1-630-252-6112; fax: #1-630252-8742. E-mail address: [email protected] (R.A. Rosenberg).
interaction or through electron-stimulated desorption of gases. The dense pockets of electrons are termed as an electron cloud (EC), and the resulting corruption of the beam is termed as an electron cloud instability (ECI). The ECI was described and modeled numerically by K. Ohmi after experimental evidence for it was found at the KEK Photon Factory [1,2]. The ECI is essentially nonresonant in nature, unlike the well-known beam-induced multipacting e!ect, "rst seen 23 years ago at the CERN Intersecting Storage Ring (ISR) [3]. The e!ective wake"eld extends only a few bunches, but for a large bunch number, the oscillation amplitude in the tail of the train grows exponentially [2]. The experimental evidence for ECI is as yet circumstantial, as direct observations of the electrons have not been made. The most
0168-9002/00/$ - see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 9 0 0 2 ( 0 0 ) 0 0 4 7 2 - 1
508
R.A. Rosenberg, K.C. Harkay / Nuclear Instruments and Methods in Physics Research A 453 (2000) 507}513
convincing pieces of evidence are the similarities between theoretical predictions and experiments performed at the Photon Factory, the Beijing electron}positron collider (BEP-C), and the Cornell Electron Storage Ring (CESR) that include positron vs. electron behavior, growth rate, e!ect of bunch spacing, e!ect of bunch current on spectrum, and beam energy. In order to directly measure the properties of the electron cloud, a special vacuum chamber was built and installed into the Advanced Photon Source (APS) storage ring. It contains ten rudimentary electron energy analyzers that were used to examine the intensity and electron energy distribution of the electrons in the storage ring under di!erent operating conditions. Subsequently, they were also employed at the Proton Storage Ring (PSR) at Los Alamos National Laboratory to diagnose the electrons produced at this facility. The results of these studies are published elsewhere [4,5]. It is the purpose of this paper to give a description of the detectors and to present data aimed at characterizing them.
2. Design The simplest possible electron energy analyzer is the retarding "eld analyzer, or RFA [6]. In its simplest con"guration, it consists of a retarding grid followed by a collector. It is desirable to use spherically symmetric grids if possible, but geometric constraints forced us to opt for planar grids. The theoretical resolution of this type of analyzer was given by DiStefano and Pierce [7,8]. For a parallel beam of monoenergetic electrons with energy ; traveling in a direction perpendicular to the plane of the grids, the transmission of the beam should be a step function as shown by the solid line in Fig. 1(a). However, if the divergence of the electrons is taken into account, then the dashed line in Fig. 1(a) results. The equation of this line is I"I (1!e;/; ), where I is the overall current of emitted electrons, ; is the retarding potential, and ; is the energy of the electrons. This equation was derived assuming the electrons originate from a point source on a surface parallel to the grids with a Lambertian angular distribution, P(a) da"2sin a cos a da,
Fig. 1. (a) Theoretical transmission of a planar retarding "eld analyzer. Solid line * ideal case for a parallel, nondivergent monoenergetic beam of energy ; . Dashed line * transmission assuming the electrons originate from a point source from a parallel surface with a Lambertian angular distribution, P(a) da"2sin a cos a da, where a is the angle between the electrons and the surface normal [7]. Also shown is the transmission curve for a cos a distribution (dotted line). In (b) is shown the di!erentiated signal for the three cases.
where a is the angle between the electrons and the surface normal. Also shown in Fig. 1(a) is the curve for a cos a distribution (I"I (1e;/; ) ) (dotted line). The "rst derivative of these curves is shown in Fig. 1(b). A schematic of the electron detector is shown in Fig. 2(a). It consists of two 70-lines/in. (90% transmission) copper grids and a collector. The "rst grid is grounded to present a uniform "eld to the incoming electrons. The second grid is biased at a retarding potential (E ) such that only electrons with kinetic energies greater than E are transmitted to the collector. The collector is graphite-coated to
R.A. Rosenberg, K.C. Harkay / Nuclear Instruments and Methods in Physics Research A 453 (2000) 507}513
509
opposite the antechamber channel as its geometry will allow. Mu-metal was wrapped around the tube to shield the detector from magnetic "elds. The location is shifted somewhat from where standard BPMs are mounted; 653 compared with 753 for a regular BPM. For comparison, a BPM is mounted opposite a detector at the same 653 at a few locations. The mounted detectors and BPM are shown in cross section in Fig. 2(b). The collector surface area is larger than a standard BPM button, but the aperture, as de"ned by three ellipsoidal vacuum slots cut into the chamber wall, is about the same: about 1 cm. Slots were chosen for RF heating and machine coupling impedance considerations. The ten electron detectors and three standard BPMs were mounted on a standard aperture (85;42 mm), 5-m-long, straight vacuum chamber.
3. Bench measurements
Fig. 2. (a) Schematic diagram of the analyzer. (b) Cross-sectional view of a vacuum chamber showing the orientation of the detector at a location where two detectors are mounted (top) and where a detector and BPM are mounted (bottom).
lower the secondary electron yield (SEY) and biased at 45 V with a battery to increase the collection e$ciency. The assembled detector was mounted on a 2.75-in. Con#at #ange with two BNC feedthroughs that were used to provide the retarding voltage and measure the collector current. The retarding voltage was applied to all ten detectors simultaneously via a programmable power supply, while the collector currents were read sequentially using a multiplexer tied into a picoammeter. Data acquisition was performed using a personal computer. The electron detectors are mounted through a 2.75-in. #ange into a 1.5-in. OD tube on a standard-aperture vacuum chamber as close to directly
All electron detectors were calibrated on the bench using an &1 lA electron beam from a lowenergy electron gun. The gun was mounted directly opposite the detector so the electron beam was parallel to the axis of the detector and approximately nondivergent. Fig. 3(a) shows the detector response and the derivative for a typical detector. The average FWHM for all detectors is &4% (*E/E). Due to the idealized nature of the electron beam from the electron gun, the detector response approximates the step function shown in Fig. 1(a), and the derivative approaches the delta function shown in Fig. 1(b). At retarding voltages greater than zero, the retarding grid e$ciently collects and focuses the scattered electrons onto the collector. This accounts for the increased signal at positive energies. However, as the voltage on the retarding grid approaches that of the battery (45 V), electrons are de#ected away from the collector, thereby decreasing the signal. The maximum observed signal is close to the theoretical value of 1 lA;(0.9)"0.8 lA, where the (0.9) factor stems from the 90% transmission of the two grids. In order to gain insight into the detector behavior under more realistic conditions, bench measurements were also made of electrons scattered from a surface. In this case, a monoenergetic beam of
510
R.A. Rosenberg, K.C. Harkay / Nuclear Instruments and Methods in Physics Research A 453 (2000) 507}513
Fig. 3. (a) Transmission curves for a monoenergetic electron beams directed along the axis of the analyzer for energies of 53 and 105 eV. The di!erentiated signal is also shown. (b) Transmission curves of monoenergetic electrons (365, 1000 eV) scattered from an Al target. The inset shows the di!erentiated signal of the 365 eV beam near the transmission threshold. (c) Signal produced from a BPM irradiated by 60 eV (solid line) and 80 eV (dashed line) electrons as a function of bias voltage applied to the BPM.
electrons was directed at an Al target at an angle of &303 to the surface normal. The analyzer was positioned to detect electrons scattered at &303. Representative data are shown in Fig. 3(b) for 1000 and 365 eV electrons. Except near 0 V, the data
monotonically increases in a fashion reminiscent of the theoretical curves represented by the dashed or dotted lines in Fig. 1(a). It is important to note that the electrons scatter from the surface elastically and inelastically, so the data shown in Fig. 3(b) contain a large contribution from low-energy electrons. The inset in Fig. 3(b) shows the di!erentiated signal resulting from scattering of the 365 eV electrons in the vicinity of the elastically scattered beam. Although there is considerable broadening, this data shows a step-like increase in the signal in a fashion similar to the dashed or dotted lines in Fig. 1(b). Most accelerators are equipped with button-like BPMs for diagnostic purposes and initially we tried to use these devices for in situ ECI studies. These preliminary data, however, proved di$cult to interpret. To characterize these BPMs, bench measurements were made with the low-energy electron gun directly irradiating the stainless-steel BPM surface. In this case, the electrometer was #oated at the retarding voltage so that the BPM current could be read as a function of applied retarding voltage. Results are shown in Fig. 3(c) for two di!erent energy electron beams, 60 and 80 eV. There is a broad, barely discernible increase in the signal at the beam energy that decreases until the retarding voltage becomes positive. The signal then rises steadily because the secondary electron collection e$ciency increases. These data reveal the general problems with trying to use a BPM, or any other single-component detector for electron diagnostics. The voltage applied to the BPM surface also decelerates or accelerates the incident electrons, thereby changing their energy. The signal read by the electrometer depends not only on the total current but also on the electron energy, because the SEY coef"cient of the surface is a strong function of the energy. Electrons liberated from the BPM surface decrease the measured current. So as voltage is applied to the BPM, the electron energy and SEY coe$cient are changing, making analysis of such data an overwhelmingly complex problem. In addition, as the positive voltage applied to the detector increases, the collection length increases, thereby attracting more distant electrons. This e!ect will depend on the density and energy distribution of the electrons in the vicinity of the BPM so getting even qualitative information is problematic at best.
R.A. Rosenberg, K.C. Harkay / Nuclear Instruments and Methods in Physics Research A 453 (2000) 507}513
To further characterize the RFA detector we used it to detect 5 kV electrons scattered from a Cu target and compared the signal to that obtained using a commercial, energy-dispersive, hemispherical electron energy analyzer (VG 100AX). Fig. 4(a) shows the spectrum obtained from the VG 100AX, while Fig. 4(b) shows that obtained from the RFA. Both were situated&1.5 in. from the target surface. The data in Fig. 4(a) show a typical Auger electron spectra from an atmospherically contaminated Cu surface: a steadily rising background with structure
511
attributable to C, O, and Cu Auger transitions. The relatively crude RFA shows no structure, only the steadily rising signal that is consistent with the behavior expected on the basis of Fig. 1(a) (dashed line). Since the RFA is an integrating analyzer, we attempted to simulate its behavior by reverse (high energy to low energy) integration of the data in Fig. 4(a). After scaling the results, we obtain the data shown in Fig. 4(c) (dashed line) and compare it to the RFA data. Integration results in the loss of the C structure at &265 eV and causes the VG100 AX data to mimic that of the RFA. Conversely, di!erentiation of the RFA data does not yield any meaningful structure, since any peaks would be totally broadened by the instrumental response.
4. Accelerator measurements
Fig. 4. (a) Electron energy spectrum from Cu target irradiated by a 5 kV electron beam obtained using a commercial, VG 100AX analyzer. (b) Spectrum obtained from the same target using the retarding "eld analyzer. (c) Comparison of a magni"ed spectrum from (b) (solid line) with the reverse integrated spectrum from (a) (dashed line).
Upon installation in the APS storage ring the detectors were used to measure electron data while varying the time structure and other properties of the stored beam. Initial measurements were made with positron beams while later measurements were made with electrons. Preliminary results were published previously [5], while a detailed analysis is forthcoming [4]. For the purposes of this paper we will only present some representative data that show the utility of the RFA analyzer for diagnosis of the electrons. Shown in Fig. 5(a) are two curves from one of the detectors under di!erent operating conditions using the positron beam. The solid line is data taken under nonmultipacting conditions (ten bunches with a spacing of 128 RF wavelengths (2.8 ns), 2 mA/bunch), while the dashed curve is data taken with stored beam in a resonant, multipacting state (ten bunches with a spacing of 7 RF wavelengths, 2 mA/bunch). Comparison of the two curves reveals that the multipacting condition produces signi"cantly more electrons (factor of 9), and they have higher energies. Further insight into the energy enhancement produced by multipacting is revealed by the processed data of Fig. 5(b). Here, the scaled nonmultipacting data have been subtracted from the multipacting data (solid line) and also di!erentiated (dotted line). This type of behavior is expected based on current models of the ECI.
512
R.A. Rosenberg, K.C. Harkay / Nuclear Instruments and Methods in Physics Research A 453 (2000) 507}513
Fig. 6. Time-dependent signal from an analyzer installed in the PSR at Los Alamos, with a 5 V bias on the retarding grid (solid line), and the signal from a current monitor (dashed line) [9].
dashed curve is the signal from a wall current monitor. The analysis of these data will be published elsewhere, but it is worth noting that reasonable signal levels are obtained and timing measurements in the nanosecond regime are possible.
5. Conclusions
Fig. 5. (a) Electron energy spectra obtained from an analyzer installed in the APS storage ring: (solid line) ten positron bunches spaced by 7;2.8 ns bunches, (dashed line) ten positron bunches spaced by 128;2.8 ns bunches; (b) (solid line) subtraction of the scaled 128;2.8 ns curve from the 7;2.8 ns curve, (dashed line) di!erentiation of the subtracted data.
Recently, the detectors have also been utilized for studies of instabilities at the Proton Storage Ring (PSR) at Los Alamos. In this case, the signal of interest was the time dependence of the electrons. Instead of the 45 V battery, an ampli"er was inserted immediately following the collector and the data were displayed on an oscilloscope. Some typical data are shown in Fig. 6 [9]. The solid curve is the RFA signal with a 5 V retardation, while the
When we designed these detectors the primary concern was to be able to maximize our detection e$ciency, since it was not possible to predict what the signal-to-noise levels would be. If there is su$cient intensity it would be possible to improve the detector by inserting a pair of grounded aperture plates in front of the retarding grid or a variation thereof [10]. These would help to de"ne a parallel beam that would simulate the ideal case shown by the solid line in Fig. 1. In addition, a channeltron or set of microchannel plates could be inserted before the collector to amplify the signal. However, these options could add to the complexity of the device and the data acquisition, which may not be warranted for a given situation. In summary, we have constructed a compact, planar RFA for accelerator diagnostics of electrons. Bench measurements to characterize the analyzer have been performed and the results are
R.A. Rosenberg, K.C. Harkay / Nuclear Instruments and Methods in Physics Research A 453 (2000) 507}513
reasonable in light of models of this type of analyzer. Comparison to results obtained using a BPM show the advantages of this analyzer for electron diagnostics. Sample results from analyzers used at the APS and PSR show how the analyzers can be used for studying the intensity, energy, and time structure of electrons in an accelerator environment.
Acknowledgements We would like to thank Mike McDowell, Je! Warren, Joe Gagliano, and George Goeppner (ANL) for experimental assistance and Andrew Browman, Dan Fitzgerald and Bob Macek (LANL) for use of their data prior to publication. This work was performed at the Advanced Photon Source and supported by the U.S. Department of
513
Energy, O$ce of Basic Energy Sciences under Contract No. W-31-109-ENG-38. References [1] M. Izawa, Y. Sato, T. Toyomasu, Phys. Rev. Lett. 74 (1995) 5044. [2] K. Ohmi, Phys. Rev. Lett. 75 (1995) 1526. [3] O. Grobner, in: The 10th International Conference on High Energy Accelerators, Protvino, 1977. [4] K.C. Harkay, R.A. Rosenberg, in preparation. [5] K.C. Harkay, R.A. Rosenberg, in: Proceedings of the 1999 Particle Accelerator Conference IEEE, New York, 1999, p. 1641. [6] J.L. Erskine, in: F.B. Dunning, R.G. Hulet (Eds.), Atomic, Molecular, and Optical Physics: Charged Particles, Vol. 29A, Academic Press, San Diego, 1995, p. 209. [7] T.H. DiStefano, D.T. Pierce, Rev. Sci. Instr. 41 (1970) 180. [8] J.A. Simpson, Rev. Sci. Instr. 32 (1961) 1283. [9] R. Macek, A. Browman, D. Fitzgerald, unpublished data, 1999. [10] C.L. Enloe, Rev. Sci. Instr. 65 (1994) 507.
A rudimentary electron energy analyzer for accelerator diagnostics R.A. Rosenberg*, K.C. Harkay Advanced Photon Source, Argonne National Laboratory, 9700 South Cass Avenue, Argonne, IL 60439, USA Received 6 March 2000; accepted 22 March 2000
Abstract We have constructed a compact, planar retarding "eld analyzer for the diagnostics of low-energy, background electrons in a high-energy particle accelerator. Bench measurements of the analyzer have been made to characterize it, and the results are reasonable in light of models of this type of analyzer. Comparisons to results obtained using a beam-position monitor (BPM) show the advantages of this analyzer for electron diagnostics. Sample results from analyzers installed at the Advanced Photon Source storage ring at Argonne National Laboratory and the Proton Storage Ring at Los Alamos National Laboratory show how the analyzers can be used for studying the intensity, energy, and time structure of electrons in an accelerator environment. 2000 Elsevier Science B.V. All rights reserved. PACS: 29.30.Aj; 29.30.Dn; 29.30.Ep Keywords: Electron analyzer; Accelerator; Diagnostics
1. Introduction Electrons in accelerators are ubiquitous. They can be produced directly by irradiation of vacuum chamber surfaces by X-rays, ions, and other particles or indirectly by bombardment of electrons, which leads to production of secondary electrons. Under many circumstances, electrons are not detrimental to accelerator performance; however, there are operating conditions that can lead to ampli"cation of the electrons. If the number density of the electrons becomes su$ciently large, they can lead to degradation of the particle beam either by direct * Corresponding author. Tel.: #1-630-252-6112; fax: #1-630252-8742. E-mail address: [email protected] (R.A. Rosenberg).
interaction or through electron-stimulated desorption of gases. The dense pockets of electrons are termed as an electron cloud (EC), and the resulting corruption of the beam is termed as an electron cloud instability (ECI). The ECI was described and modeled numerically by K. Ohmi after experimental evidence for it was found at the KEK Photon Factory [1,2]. The ECI is essentially nonresonant in nature, unlike the well-known beam-induced multipacting e!ect, "rst seen 23 years ago at the CERN Intersecting Storage Ring (ISR) [3]. The e!ective wake"eld extends only a few bunches, but for a large bunch number, the oscillation amplitude in the tail of the train grows exponentially [2]. The experimental evidence for ECI is as yet circumstantial, as direct observations of the electrons have not been made. The most
0168-9002/00/$ - see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 9 0 0 2 ( 0 0 ) 0 0 4 7 2 - 1
508
R.A. Rosenberg, K.C. Harkay / Nuclear Instruments and Methods in Physics Research A 453 (2000) 507}513
convincing pieces of evidence are the similarities between theoretical predictions and experiments performed at the Photon Factory, the Beijing electron}positron collider (BEP-C), and the Cornell Electron Storage Ring (CESR) that include positron vs. electron behavior, growth rate, e!ect of bunch spacing, e!ect of bunch current on spectrum, and beam energy. In order to directly measure the properties of the electron cloud, a special vacuum chamber was built and installed into the Advanced Photon Source (APS) storage ring. It contains ten rudimentary electron energy analyzers that were used to examine the intensity and electron energy distribution of the electrons in the storage ring under di!erent operating conditions. Subsequently, they were also employed at the Proton Storage Ring (PSR) at Los Alamos National Laboratory to diagnose the electrons produced at this facility. The results of these studies are published elsewhere [4,5]. It is the purpose of this paper to give a description of the detectors and to present data aimed at characterizing them.
2. Design The simplest possible electron energy analyzer is the retarding "eld analyzer, or RFA [6]. In its simplest con"guration, it consists of a retarding grid followed by a collector. It is desirable to use spherically symmetric grids if possible, but geometric constraints forced us to opt for planar grids. The theoretical resolution of this type of analyzer was given by DiStefano and Pierce [7,8]. For a parallel beam of monoenergetic electrons with energy ; traveling in a direction perpendicular to the plane of the grids, the transmission of the beam should be a step function as shown by the solid line in Fig. 1(a). However, if the divergence of the electrons is taken into account, then the dashed line in Fig. 1(a) results. The equation of this line is I"I (1!e;/; ), where I is the overall current of emitted electrons, ; is the retarding potential, and ; is the energy of the electrons. This equation was derived assuming the electrons originate from a point source on a surface parallel to the grids with a Lambertian angular distribution, P(a) da"2sin a cos a da,
Fig. 1. (a) Theoretical transmission of a planar retarding "eld analyzer. Solid line * ideal case for a parallel, nondivergent monoenergetic beam of energy ; . Dashed line * transmission assuming the electrons originate from a point source from a parallel surface with a Lambertian angular distribution, P(a) da"2sin a cos a da, where a is the angle between the electrons and the surface normal [7]. Also shown is the transmission curve for a cos a distribution (dotted line). In (b) is shown the di!erentiated signal for the three cases.
where a is the angle between the electrons and the surface normal. Also shown in Fig. 1(a) is the curve for a cos a distribution (I"I (1e;/; ) ) (dotted line). The "rst derivative of these curves is shown in Fig. 1(b). A schematic of the electron detector is shown in Fig. 2(a). It consists of two 70-lines/in. (90% transmission) copper grids and a collector. The "rst grid is grounded to present a uniform "eld to the incoming electrons. The second grid is biased at a retarding potential (E ) such that only electrons with kinetic energies greater than E are transmitted to the collector. The collector is graphite-coated to
R.A. Rosenberg, K.C. Harkay / Nuclear Instruments and Methods in Physics Research A 453 (2000) 507}513
509
opposite the antechamber channel as its geometry will allow. Mu-metal was wrapped around the tube to shield the detector from magnetic "elds. The location is shifted somewhat from where standard BPMs are mounted; 653 compared with 753 for a regular BPM. For comparison, a BPM is mounted opposite a detector at the same 653 at a few locations. The mounted detectors and BPM are shown in cross section in Fig. 2(b). The collector surface area is larger than a standard BPM button, but the aperture, as de"ned by three ellipsoidal vacuum slots cut into the chamber wall, is about the same: about 1 cm. Slots were chosen for RF heating and machine coupling impedance considerations. The ten electron detectors and three standard BPMs were mounted on a standard aperture (85;42 mm), 5-m-long, straight vacuum chamber.
3. Bench measurements
Fig. 2. (a) Schematic diagram of the analyzer. (b) Cross-sectional view of a vacuum chamber showing the orientation of the detector at a location where two detectors are mounted (top) and where a detector and BPM are mounted (bottom).
lower the secondary electron yield (SEY) and biased at 45 V with a battery to increase the collection e$ciency. The assembled detector was mounted on a 2.75-in. Con#at #ange with two BNC feedthroughs that were used to provide the retarding voltage and measure the collector current. The retarding voltage was applied to all ten detectors simultaneously via a programmable power supply, while the collector currents were read sequentially using a multiplexer tied into a picoammeter. Data acquisition was performed using a personal computer. The electron detectors are mounted through a 2.75-in. #ange into a 1.5-in. OD tube on a standard-aperture vacuum chamber as close to directly
All electron detectors were calibrated on the bench using an &1 lA electron beam from a lowenergy electron gun. The gun was mounted directly opposite the detector so the electron beam was parallel to the axis of the detector and approximately nondivergent. Fig. 3(a) shows the detector response and the derivative for a typical detector. The average FWHM for all detectors is &4% (*E/E). Due to the idealized nature of the electron beam from the electron gun, the detector response approximates the step function shown in Fig. 1(a), and the derivative approaches the delta function shown in Fig. 1(b). At retarding voltages greater than zero, the retarding grid e$ciently collects and focuses the scattered electrons onto the collector. This accounts for the increased signal at positive energies. However, as the voltage on the retarding grid approaches that of the battery (45 V), electrons are de#ected away from the collector, thereby decreasing the signal. The maximum observed signal is close to the theoretical value of 1 lA;(0.9)"0.8 lA, where the (0.9) factor stems from the 90% transmission of the two grids. In order to gain insight into the detector behavior under more realistic conditions, bench measurements were also made of electrons scattered from a surface. In this case, a monoenergetic beam of
510
R.A. Rosenberg, K.C. Harkay / Nuclear Instruments and Methods in Physics Research A 453 (2000) 507}513
Fig. 3. (a) Transmission curves for a monoenergetic electron beams directed along the axis of the analyzer for energies of 53 and 105 eV. The di!erentiated signal is also shown. (b) Transmission curves of monoenergetic electrons (365, 1000 eV) scattered from an Al target. The inset shows the di!erentiated signal of the 365 eV beam near the transmission threshold. (c) Signal produced from a BPM irradiated by 60 eV (solid line) and 80 eV (dashed line) electrons as a function of bias voltage applied to the BPM.
electrons was directed at an Al target at an angle of &303 to the surface normal. The analyzer was positioned to detect electrons scattered at &303. Representative data are shown in Fig. 3(b) for 1000 and 365 eV electrons. Except near 0 V, the data
monotonically increases in a fashion reminiscent of the theoretical curves represented by the dashed or dotted lines in Fig. 1(a). It is important to note that the electrons scatter from the surface elastically and inelastically, so the data shown in Fig. 3(b) contain a large contribution from low-energy electrons. The inset in Fig. 3(b) shows the di!erentiated signal resulting from scattering of the 365 eV electrons in the vicinity of the elastically scattered beam. Although there is considerable broadening, this data shows a step-like increase in the signal in a fashion similar to the dashed or dotted lines in Fig. 1(b). Most accelerators are equipped with button-like BPMs for diagnostic purposes and initially we tried to use these devices for in situ ECI studies. These preliminary data, however, proved di$cult to interpret. To characterize these BPMs, bench measurements were made with the low-energy electron gun directly irradiating the stainless-steel BPM surface. In this case, the electrometer was #oated at the retarding voltage so that the BPM current could be read as a function of applied retarding voltage. Results are shown in Fig. 3(c) for two di!erent energy electron beams, 60 and 80 eV. There is a broad, barely discernible increase in the signal at the beam energy that decreases until the retarding voltage becomes positive. The signal then rises steadily because the secondary electron collection e$ciency increases. These data reveal the general problems with trying to use a BPM, or any other single-component detector for electron diagnostics. The voltage applied to the BPM surface also decelerates or accelerates the incident electrons, thereby changing their energy. The signal read by the electrometer depends not only on the total current but also on the electron energy, because the SEY coef"cient of the surface is a strong function of the energy. Electrons liberated from the BPM surface decrease the measured current. So as voltage is applied to the BPM, the electron energy and SEY coe$cient are changing, making analysis of such data an overwhelmingly complex problem. In addition, as the positive voltage applied to the detector increases, the collection length increases, thereby attracting more distant electrons. This e!ect will depend on the density and energy distribution of the electrons in the vicinity of the BPM so getting even qualitative information is problematic at best.
R.A. Rosenberg, K.C. Harkay / Nuclear Instruments and Methods in Physics Research A 453 (2000) 507}513
To further characterize the RFA detector we used it to detect 5 kV electrons scattered from a Cu target and compared the signal to that obtained using a commercial, energy-dispersive, hemispherical electron energy analyzer (VG 100AX). Fig. 4(a) shows the spectrum obtained from the VG 100AX, while Fig. 4(b) shows that obtained from the RFA. Both were situated&1.5 in. from the target surface. The data in Fig. 4(a) show a typical Auger electron spectra from an atmospherically contaminated Cu surface: a steadily rising background with structure
511
attributable to C, O, and Cu Auger transitions. The relatively crude RFA shows no structure, only the steadily rising signal that is consistent with the behavior expected on the basis of Fig. 1(a) (dashed line). Since the RFA is an integrating analyzer, we attempted to simulate its behavior by reverse (high energy to low energy) integration of the data in Fig. 4(a). After scaling the results, we obtain the data shown in Fig. 4(c) (dashed line) and compare it to the RFA data. Integration results in the loss of the C structure at &265 eV and causes the VG100 AX data to mimic that of the RFA. Conversely, di!erentiation of the RFA data does not yield any meaningful structure, since any peaks would be totally broadened by the instrumental response.
4. Accelerator measurements
Fig. 4. (a) Electron energy spectrum from Cu target irradiated by a 5 kV electron beam obtained using a commercial, VG 100AX analyzer. (b) Spectrum obtained from the same target using the retarding "eld analyzer. (c) Comparison of a magni"ed spectrum from (b) (solid line) with the reverse integrated spectrum from (a) (dashed line).
Upon installation in the APS storage ring the detectors were used to measure electron data while varying the time structure and other properties of the stored beam. Initial measurements were made with positron beams while later measurements were made with electrons. Preliminary results were published previously [5], while a detailed analysis is forthcoming [4]. For the purposes of this paper we will only present some representative data that show the utility of the RFA analyzer for diagnosis of the electrons. Shown in Fig. 5(a) are two curves from one of the detectors under di!erent operating conditions using the positron beam. The solid line is data taken under nonmultipacting conditions (ten bunches with a spacing of 128 RF wavelengths (2.8 ns), 2 mA/bunch), while the dashed curve is data taken with stored beam in a resonant, multipacting state (ten bunches with a spacing of 7 RF wavelengths, 2 mA/bunch). Comparison of the two curves reveals that the multipacting condition produces signi"cantly more electrons (factor of 9), and they have higher energies. Further insight into the energy enhancement produced by multipacting is revealed by the processed data of Fig. 5(b). Here, the scaled nonmultipacting data have been subtracted from the multipacting data (solid line) and also di!erentiated (dotted line). This type of behavior is expected based on current models of the ECI.
512
R.A. Rosenberg, K.C. Harkay / Nuclear Instruments and Methods in Physics Research A 453 (2000) 507}513
Fig. 6. Time-dependent signal from an analyzer installed in the PSR at Los Alamos, with a 5 V bias on the retarding grid (solid line), and the signal from a current monitor (dashed line) [9].
dashed curve is the signal from a wall current monitor. The analysis of these data will be published elsewhere, but it is worth noting that reasonable signal levels are obtained and timing measurements in the nanosecond regime are possible.
5. Conclusions
Fig. 5. (a) Electron energy spectra obtained from an analyzer installed in the APS storage ring: (solid line) ten positron bunches spaced by 7;2.8 ns bunches, (dashed line) ten positron bunches spaced by 128;2.8 ns bunches; (b) (solid line) subtraction of the scaled 128;2.8 ns curve from the 7;2.8 ns curve, (dashed line) di!erentiation of the subtracted data.
Recently, the detectors have also been utilized for studies of instabilities at the Proton Storage Ring (PSR) at Los Alamos. In this case, the signal of interest was the time dependence of the electrons. Instead of the 45 V battery, an ampli"er was inserted immediately following the collector and the data were displayed on an oscilloscope. Some typical data are shown in Fig. 6 [9]. The solid curve is the RFA signal with a 5 V retardation, while the
When we designed these detectors the primary concern was to be able to maximize our detection e$ciency, since it was not possible to predict what the signal-to-noise levels would be. If there is su$cient intensity it would be possible to improve the detector by inserting a pair of grounded aperture plates in front of the retarding grid or a variation thereof [10]. These would help to de"ne a parallel beam that would simulate the ideal case shown by the solid line in Fig. 1. In addition, a channeltron or set of microchannel plates could be inserted before the collector to amplify the signal. However, these options could add to the complexity of the device and the data acquisition, which may not be warranted for a given situation. In summary, we have constructed a compact, planar RFA for accelerator diagnostics of electrons. Bench measurements to characterize the analyzer have been performed and the results are
R.A. Rosenberg, K.C. Harkay / Nuclear Instruments and Methods in Physics Research A 453 (2000) 507}513
reasonable in light of models of this type of analyzer. Comparison to results obtained using a BPM show the advantages of this analyzer for electron diagnostics. Sample results from analyzers used at the APS and PSR show how the analyzers can be used for studying the intensity, energy, and time structure of electrons in an accelerator environment.
Acknowledgements We would like to thank Mike McDowell, Je! Warren, Joe Gagliano, and George Goeppner (ANL) for experimental assistance and Andrew Browman, Dan Fitzgerald and Bob Macek (LANL) for use of their data prior to publication. This work was performed at the Advanced Photon Source and supported by the U.S. Department of
513
Energy, O$ce of Basic Energy Sciences under Contract No. W-31-109-ENG-38. References [1] M. Izawa, Y. Sato, T. Toyomasu, Phys. Rev. Lett. 74 (1995) 5044. [2] K. Ohmi, Phys. Rev. Lett. 75 (1995) 1526. [3] O. Grobner, in: The 10th International Conference on High Energy Accelerators, Protvino, 1977. [4] K.C. Harkay, R.A. Rosenberg, in preparation. [5] K.C. Harkay, R.A. Rosenberg, in: Proceedings of the 1999 Particle Accelerator Conference IEEE, New York, 1999, p. 1641. [6] J.L. Erskine, in: F.B. Dunning, R.G. Hulet (Eds.), Atomic, Molecular, and Optical Physics: Charged Particles, Vol. 29A, Academic Press, San Diego, 1995, p. 209. [7] T.H. DiStefano, D.T. Pierce, Rev. Sci. Instr. 41 (1970) 180. [8] J.A. Simpson, Rev. Sci. Instr. 32 (1961) 1283. [9] R. Macek, A. Browman, D. Fitzgerald, unpublished data, 1999. [10] C.L. Enloe, Rev. Sci. Instr. 65 (1994) 507.