Spin-dependent scattering of polarized protons from a polarized 3He internal gas target

Nuclear Instruments

and Methods in Physics Research A 354 (1995) 437-457

ELSEVIER

Spin-dependent

NUCLLAR INSTRUMENTS 8 METHODS IN PHYSICS RESEAftCW SectionA

scattering of polarized protons from a polarized 3He internal gas target

C. Bloch a, J. Doskow a, C.D. Goodman a, W.W. Jacobs a, M. Leuschner a, H.O. Meyer a, B. von Przewoski a, T. Rinckel a, G. Savopulos a, A. Smith a, J. Sowinski a, F. Sperisen a, W.K. Pitts b, D. DeSchepper ‘, R. Ent ‘,‘, J-O. Hansen ‘, J. Kelsey ‘, W. Korsch ‘,‘, L.H. Kramer ‘, K. Lee ‘, N.C.R. Makins ‘, R.G. Milner ‘, * , S.F. Pate ‘, C. Tschal’ar ‘, T.P. Welch c*3,D. Marchlenski d, E. Sugarbaker d, W. Lorenzon e, P.V. Pancella f, J.F.J. van den Brand g, H.J. Bulten g, C.E. Jones g74,M.A. Miller g, J. Neal g, 0. Unal g, Z-L. Zhou g aIndiana University Cyclotron Facility, Bloomington, Indiana 47405, USA b University of Louisville, Louisville, Kentucky 40292, USA Center and Laboratory for Nuclear Science, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA d The Ohio State University, Columbus, Ohio 43210, USA e TRIUMF, British Columbia, Canada V6T _&I3 f Western Michigan Universiry, Kalamazoo, Michigan 49007, USA g University of Wisconsin-Madison, Madison, Wisconsin 53706, USA

’ MIT-Bates Linear Accelerator

Received 1 August 1994

Abstract We describe the first ex eriment to use a polarized internal gas target and polarized beam in a storage ring. A laser P . optically pumped polarized He mtemal gas target has been used with circulating beams of 197-414 MeV polarized protons to carry out an extensive set of measurements of spin dependent scattering. A large acceptance non-magnetic detector system consisting of wire-chambers, scintillators and microstrip detectors was used to detect protons, neutrons, deuterons, and 3He nuclei from the beam-target interaction. It is demonstrated that these techniques result in low backgrounds ( < 1%) due to scattering from species other than the polarized target gas and allow detection of low energy recoiling nuclei. Specific issues such as interfacing the experiment to the storage ring and monitoring the luminosity and polarizations are discussed in detail.

1. Introduction

[2-61. We report here the technical

The use of a polarized gas target internal to a storage ring is recognized as having many advantages in the measurement of spin-dependent scattering observables [l]. Indeed major efforts are underway at many laboratories worldwide to undertake measurements using this technique

* Corresponding author, tel. + 1 617 258 5439, E-mail: [email protected]. ’ Present address: CEBAP, Newport News, VA 23606, USA. ’ Present address: Caltech, Pasadena, CA 9112.5, USA. 3 Present address: Oregon State University, Corvallis, OR 97331-6507, USA. 4 Present address: 60439, USA.

Argonne

National

Laboratory,

Argonne

IL

0168-9002/95/$09.50 0 1995 Elsevier Science B.V. All rights reserved SSDI 0168-9002(94)01057-9

details of measurements carried out using the Indiana University Cyclotron Facility (IUCF) Cooler Ring with a laser optically pumped polarized 3He target and a polarized proton beam. The first demonstration of these techniques based on initial measurements was reported previously [7]. The experiment was designed to study the spin structure of the 3He ground state, which is of great current interest. The polarized 3He nucleus has the feature that it is expected to be a good approximation to a polarized neutron. Thus, it should be possible to determine the charge 181 and spin [9] distributions of the neutron from measurement of elastic and deep inelastic form-factors via electron scattering from polarized 3He. Recently, the electric and magnetic form-factors of the neutron have been determined for the first time [lO,ll] using spin dependent electron scattering from polarized 3He. Further quasielastic

438

C. Bloch et al. /Nucl. Instr. and Meth. in Phys. Res. A 354 (1995) 437-457

measurements are planned [4,12,13] at several intermediate energy electron facilities. In addition, the first deep inelastic measurements of spin dependent scattering from polarized 3He have been carried out [14] and further experiments are planned at SLAC [15] and DESY [5]. Crucial to the extraction of precise information on the neutron is an understanding of the spin structure of the 3He ground state wave function [16]. This is best determined using medium energy quasielastic spin-dependent scattering from polarized 3He. In addition, it is important to obtain a quantitative understanding of the quasielastic reaction mechanism. An initial series of experiments to measure quasielastic 3He(p, 2p) and (p, pn) scattering was carried out at TRIUMF [17-191. However, these data surprisingly showed a large deviation from the picture of polarized 3He as an effective polarized neutron. The experiment described here (CE-25) was carried out to extend the TRIUMF measurements in both kinematic range and incident proton energy. Further, CE-25 was designed to measure the various asymmetries in quasielastic spin-dependent scattering as well as to discriminate between two-body and three-body breakup channels of the 3He nucleus. In this paper the experimental techniques used in carrying out the measurements are described in detail. The results of the measurements will be presented elsewhere. In Section 2 the design of the experiment is outlined. Section 3 discusses some aspects of interfacing the experiment to the storage ring. In Section 4 the operation of the laser driven polarized 3He internal gas target is described. A description of the detectors is provided in Section 5 and the electronics and data acquisition procedures are detailed in Section 6. Section 7 contains information on the methods used for monitoring the luminosity and polarizations. The performance of the apparatus is discussed in Section 8.

-

-.-. Mm1

R

I

\_

,’

“\\/

Fig. 2. (a) A schematic layout of the experiment. The AE detectors are labeled SL and SR where L(R) indicates left (right). Multi-wire proportional counters are labeled X(Y) for horizontal (vertical) position information and l(2) for front (back). The neutron detector arrays are labeled nDET and the backing plastics labeled BP. The Helmholtz coils for the polarized target are labeled HC. (b) An expanded view of the target region. The target cell has 3 large area silicon detectors labeled SDL/SDR on each side for recoil detection. A silicon detector and plastic scintillator labeled Monl are placed just downstream of the target on each side and used for luminosity monitoring.

have momenta p1 and pZ. The spin-dependent differential cross section when both beam and target spins are oriented normal (n) to the scattering plane can be written as [20] d%

dp,dp, 2. The design of the experiment

&

= This experiment was designed to measure the spin-dependent cross section for both 3He(p, 2p) and 3He(p, pn> quasielastic scattering over a large kinematic range. The reaction mechanism is shown in Fig. 1 where a proton is incident with momentum pint and the outgoing nucleons

\/

mA

m*-l~TA-l

Fig. 1. Quasielastic nucleon knockout by an incident proton with momentum piof resulting in outgoing nucleons with momenta p1 and pz.

i

I

[l+(ri.P,)A,,o+(ir.P,)A,,

(1) where fi is a unit vector in the direction of pine Xp,; P, and P, are the polarization vectors of the beam and target; (d%Jdp,dp,) is the unpolarized coincidence cross section; A,, is the spin-correlation parameter, and A,, are the beam and target analyzing powers, and A,, respectively. CE-25 measured the quantities AmO, Am”, for both 3He(p, 2p) and ‘He(p, pn). and A,, Fig. 2 shows a schematic layout of the experiment. The apparatus was designed to detect two nucleons in coincidence, one on each side of the beam, over a range of angles from approximately 21” to 67” with the energy and angle measured to reconstruct the momentum vector of each nucleon. This configuration was chosen to match free scattering kinematics where the spin observables are large.

C. Bloch et al. /Nucl. Instr. and Meth. in Phys. Res. A 354 (1995) 437-457 Each main detector arm consisted of wire chambers, thin plastic scintillators and large position-sensitive scintillator bars. The detectors also triggered on scattered deuterons and pions thus allowing detection of various other reactions. Data were taken at three incident proton energies: 197, 300 and 414 MeV. This energy range, together with the wide range of momentum transfers covered by the large detector acceptance, allowed for a thorough investigation of the reaction mechanism. Because the target was ultra-thin, low energy recoil particles, e.g. sub MeV 3He, could exit the target with sufficient energy to be detected. Such particles were detected by large area silicon detectors located near the target. In this way 3He(p, p3He1 elastic scattering was measured. These detectors also allowed a separation of ‘He(p, 2p)pn and 3He(p, 2p)d final states by detecting slow recoiling deuterons. Further details on the implementation of these techniques are provided below in Section 5. The experiment was performed using a polarized proton beam stored in the IUCF electron-cooled storage ring (the Cooler) [21]. This ring is a light ion synchrotron which can accelerate a beam of particles of charge 4 and atomic number A to about 500 MeV q2/A. The increase in the emittance of the stored beam due to the presence of internal targets is compensated by electron cooling, resulting in long beam lifetimes. The stored polarized protons scattered from a polarized ‘He internal gas target in the A section of the Cooler at a typical luminosity of 6 X 10z8 cm-‘s-’ (see Fig. 3). This location was chosen because of the beam’s minimal Bfunction values (1.44 m horizontally in the bend plane and 1.96 m vertically) and zero dispersion. A minimum pfunction implies a minimum transverse size to the stored

c

' CE-01

Fig. 3. A schematic layout of the IUCF Cooler Ring. The stored beam is injected in Section I and the electron cooling occurs in Section C. Section A contains the apparatus for the CE-25 experiment. Section G contains the CE-01 apparatus where additional ‘Htp, 2p) measurements were made to calibrate the luminosity monitors of the main experiment.

439

beam and consequently allows a smaller diameter target cell and thicker target. The beam polarization was reversed on every injection into the ring and the target polarization was reversed every 180 s independent of the status of the beam. At regular intervals during the CE-25 experiment, molecular hydrogen gas was flowed in the target so that ’ H(p, 2p) elastic measurements could be carried out. These elastic measurements served to calibrate the time-of-flight system, monitored the stability of the electronics, allowed a determination of the kinematic resolutions, and provided information on the background rate. When used with a large acceptance detector, this beam-target configuration had several advantages over conventional polarized targets. The target contained only 3He atoms which were constantly replenished, and the beam did not have to pass through any foils to reach the target. There was no radiation damage to the target and the magnetic holding field was low.

3. Interface of the experiment to the storage ring 3.1. Impact on storage ring With external conventional target experiments the accelerator and experiment are essentially unconnected. In this experiment, where the target was internal to the storage ring, there were several new considerations. The apertures, vacuum properties and magnetic field of the target, in addition to the presence of gas target atoms in the path of the stored beam, all directly influenced operation of the storage ring. The transverse dimensions of the target cell were carefully chosen based on several considerations. The thickness of the target increases rapidly as the transverse dimensions of the cell are decreased. The effectiveness of the flow limiters also improves rapidly with decreased radius. However, sufficiently small apertures can intersect the beam halo and produce backgrounds. In addition, small apertures can lower the geometrical acceptance of the accelerator thus impacting the beam lifetime and making it more difficult to inject beam into the ring. The target cell and pumping aperture dimensions were chosen to be approximately equal to the machine design acceptance 357~ mm mrad. The best performance seen in previous experiments was half this value [22]. By choosing the target acceptance to be larger than the operating machine acceptance it was straightforward to tune the stored beam through the target. In addition, the scattering of beam halo from the target cell was minimized. The presence of target gas reduced the stored beam lifetime. This directly influenced the time required for data taking and refilling the ring. With no target gas in the cell the beam lifetime was typically 2000 s and depended on the tune of the Cooler. With the target gas flowing this

440

C. Bloch et al. /NucI. Instr. and Meth. in Phys. Res. A 354 (1995) 437-457

value was reduced to 500-900 s. It took about 100-200 s to fill the ring to 100 pA. The polarized target required a holding magnetic field of about 1 mT both to orient the 3He atoms by optical pumping and to define the direction of polarization. The field was present over about 1 m along the beam line resulting in a maximum field integral of 1 mT m. This was sufficiently large to produce a non-negligible deviation in the trajectory of the stored beam. Without correction, the stored beam lifetime and the rate of background events in the detectors would be different for the two field directions. Further, any change in the magnetic holding field would disrupt beam injection. To correct for the effect of the target holding field two additional magnets were added to the straight section immediately following the target. These magnets were controlled using a shunt across the power supply which provided current to the target magnetic field coils. This combination of three magnets then allowed for a local motion of the beam without affecting the closed orbit elsewhere in the ring. With this configuration deflections of the beam in the target cell were less than +O.l mm, and operation of the ring could be made independent of the field in the target. 3.2. Synchronization

with the cooler

Experiments with the Cooler have a macroscopic time structure that consists of cycles. This is shown schematically in Fig. 4 where the stored current in the ring as measured by a parametric current transformer (PCT) [23] is plotted as a function of time. The polarized protons are injected, accelerated to the desired energy, and cooled. The data taking then commences as the intensity of the stored beam decays because of loss mechanisms. When the stored beam has decayed to about one half of its initial intensity the beam is kicked from the ring. This constitutes one cycle. The duration was chosen to optimize the integrated luminosity based on the fill time and lifetime of the stored beam. The Cooler was filled to 100 FA of polarized beam using the method of RF-stacking [24]. The beam was then accelerated to the final energy and cooled. At this point a

start pulse was sent from the Cooler controls computer to the experiment. The start pulse was used to trigger a programmable timing and sequencing module (Jorway). The Jorway issued the command to ramp the wire chamber voltages from the quiescent level at injection to their plateau voltage, to clear CAMAC modules at the beginning of each cycle and to enable data taking. After 600-900 s, the data taking was stopped and the wire chambers were reset to the quiescent voltage levels. The residual beam was kicked from the Cooler, the ring magnets were reset and the ring was then filled with beam of the opposite spin state. 3.3. The polarized proton beam In this experiment the polarized proton beam was injected into the Cooler at approximately 200 MeV. The stored beam was then ramped to the desired energy for data taking. Ramping polarized beams in a storage ring is complicated by the presence of depolarizing spin resonances. Such a resonance is excited in the Cooler near 200 MeV. This resonance has the form v,,,=Gy=n+mv,,,

(2)

where vspin is the spin tune defined as the number of times the spin of a beam particle precesses as it goes around the ring, 1; is the betatron tune or the number of times the beam oscillates vertically as it goes around the ring, G is the anomalous magnetic moment of the beam particles, and y is the relativistic kinematic factor. The resonance occurs when n and m are integers. In the IUCF Cooler the resonance of interest for protons occurs when II = 7 and m = - 1. The depolarizing effects of this resonance have been measured [25]. The width of the resonance was found to be narrow and it can be easily avoided by tuning such that the resonance occurs below the injection energy. This is accomplished by changing the strengths of the Cooler quadrupole magnets. In addition to the intrinsic resonance near 200 MeV, depolarization can also be caused by coupling of the vertical and horizontal betatron tunes. Skew quadrupoles were adjusted to minimize this coupling. Using the known circumference and the measured radiofrequency associated with the stored beam in the Cooler the incident beam energies were determined to be 197.5, 299.6, and 414.3 MeV with an uncertainty of kO.1 MeV.

4. The polarized He-3 internal gas target

011 0

”

200

”

400

”

600

”

800

’

1’

1000

Time (sec.)

Fig. 4. A schematic figure of the time structure of the experiment.

The polarized 3He internal gas target consisted of a laser optically pumped source of atoms directed into a thin walled target cell. A large differential vacuum pumping system interfaced the target cell to the storage ring vacuum, The polarization of the atoms in the pumping cell was monitored by an optical technique.

The 3He source polarized the atoms by metastability exchange optical pumping [26]. If a weak electric discharge is maintained in a low pressure 3He gas, a small fraction of the atoms (= 10m6) will be in the long-lived 23S, metastable state. Circularly polarized pumping light incident upon the sample along a weak applied magnetic field excite transitions between the 3S, and 3Po states. Angular momentum is thus transferred from the pumping light to the metastable atoms, and the metastable atoms become polarized. Transfer of polarization to the ground state atoms is achieved through metastability exchange collisions. The target [27] operated by flowing ‘He atoms through a glass pumping cell of volume 400 cm3 at a rate of 1.2 X 10” atoms/s. The 3He gas had an input pressure of 5 Torr and traversed an input precision capillary with a conductance of 0.9 cm3/s. The polarized gas in the pump-

ing cell had an average pressure of 0.5 Torr and exited through a second precision capillary with a conductance of 8 cm3/s to an aluminum target cell, which is described in Section 4.3 below. Fig. 5 shows the schematic layout of the target apparatus. This configuration resulted in a target thickness of 1.5 X lOI4 atoms/cm2. The target was polarized along the holding magnetic field provided by a 1 m diameter Helmholtz pair. The target assembly was mounted on a single aluminum flange, which was attached to an aluminum ultra-high vacuum chamber. The pumping cell was located in a well in the target flange to minimize the effects of depolarizing field gradients. The target chamber was designed to be an all metal bakeable ultra high vacuum system. It was manufactured from a single forged block of 6061-T6 aluminum. Nonmagnetic metal was used because of the depolarizing effects of non-uniform magnetic fields in the vicinity of the target pumping cell. The seals used were aluminum conflat from Ulvac Ltd. and Viton rubber seals. The

m

HELUHOLTZ COILS

Fig. 5. A schematic diagram of the CE-25 polarized 3He internal gas target.

-0.6’

’

500

1

j

’

’

1500

1000

’

1 2000

seconds

Fig. 6. Target polarization as a function of time.

complete vacuum chamber assembly was coated with titanium nitride to harden the sealing surfaces. An infrared laser system was used as a source of 1.083 pm photons. A 4 mm diameter X 79 mm long crystal of La,,,Nd,,,MgAl,,O,, (LNA) was the lasing medium in a Lasermetrics 9650 YAG cavity. The rod had 60 cm radius concave ends to minimize the effects of thermal lensing. The LNA gain curve has peaks at 1.054 and 1.084 pm. An uncoated etalon of thickness 0.5 mm was used for fine tuning to the 23S, - 23Po 1.083 pm helium transition. Laser tuning was facilitated by passing the light that leaked out one of the end mirrors through a 3He discharge cell and observing the fluorescence signal in a photomultiplier tube at 90”. A 99.25% reflecting mirror at the other end of the cavity was used as an output coupler. Typical output power was 4 W at the helium transition. The output laser light was linearly polarized by passage through a polarizing beamsplitter cube and then circularly polarized by passage through a Pockels cell which acted as a quarter-wave plate. The average polarization in the pumping cell was 0.46 with a variation of f0.02 and was measured using detection of the 667 nm line in the 3He discharge [25], as described in Section 4.2 below. The target polarization was reversed every 180 s by reversing the circular polarization of the laser light. The target polarization is shown over a time period of 2000 s in Fig. 6 The target field direction was typically reversed every four cycles to minimize systematic errors. The target operated reliably at an efficiency close to 100% over a period of about one year. This was in large part due to the stability of the LNA laser and the gas feed system and the ability to continuously obtain on-line a large amount of diagnostic and control information. 4.2. Polarization

monitor

The nuclear polarization of the 3He atoms in the target pumping cell was determined from the circular polariza-

tion of light emitted in atomic transitions excited by the gas discharge. This is possible because initially unpolarized excited atomic states acquire a certain degree of polarization during their lifetime due to the hyperfine interaction, so that light emitted in the decay of these states is circularly polarized. Because collisions between atoms may cause atomic depolarization, the degree of circular polarization is pressure-dependent, and a cross-calibration with a sample of known nuclear polarization has to be erformed [28]. In the present target apparatus, the 667 nm !? He line was used for polarization monitoring. The polarimeter viewed the pumping cell at an angle of 13.5”. Light entering the polarimeter first passed through an Oriel 539-65 narrow-band filter, which strongly attenuated noise from background light, and then through a CVI 2WPM-667-10-4 quarter-wave plate, rotating with a frequency v = 20 Hz followed by a Polaroid fixed linear polarizer. An Amperex XP2023B photomultiplier tube detected the transmitted light. Any circular polarization component in the incident light caused a 2v-periodic intensity variation at the phototube, whereas unpolarized or linearly polarized light caused a constant background. The photomultiplier signal (together with a reference signal of frequency w) was fed into a lockin amplifier (EG&G Model 5209) whose output was proportional to the AC component of the input signal. The DC component of the phototube signal was measured by a low-pass/DC amplifier combination. Both lockin and DC amplifier output signals, V,, and V,, were digitized by CAMAC analog-to-digital converters once every second. The pumping cell polarization was computed as

wherefp,,,,is the pressure-dependent calibration constant, corrected to account for the angle of the polarimeter with respect to the magnetic field axis (13.5”) and for the effect of the magnetic holding field [28]. For the present experiment, jp,,is = 7.55. 4.3. Target cell The exit capillary of the 3He pumping cell was connected to a glass flange which was in turn attached to the top of the target chamber. The lower end of the capillary extended into the chamber and was coupled to the target cell with a short Teflon sleeve. The target cell, which was 400 mm long, 16.6 mm high and 13.1 mm wide, was constructed from 0.2 mm thick aluminum as shown in Fig. 7. The sides of the cell, which served as exit windows for reaction products including low energy recoil particles, were 1.7 pm aluminized mylar sheets attached to the target cell with Torr Seal low vapor pressure resin [29]. The cell dimensions, which were chosen to meet the designed storage ring acceptances at the target location,

443

C. Bloch et al. /Nucl. Instr. and Meth. in Phys. Res. A 354 (1995) 437-457 provided 357r mm mrad acceptance in the vertical direction and 301~ mm mrad in the horizontal plane. The scattering chamber was centered along the beam axis using a transit and crosshairs placed in the end flanges. The target cell, transport tube and pumping cell were attached rapidly to the top flange, which was bolted to the chamber. Pins were placed in two of the bolt holes so that the top flange could be removed and replaced at a reproducible position. Initial alignment of the target cell was done using apertures placed in the two ends of the cell. Measurements of the position of target cell relative to the top flange allowed the cell to be removed and replaced with an alignment accuracy of less than f 1 mm in both horizontal and vertical directions. 4.4. Differential vacuum system Significantly lower vacuums are required in storage rings than when conventional nuclear physics techniques are employed. In the IUCF cooler ring the base vacuum is typically 10m9 Torr. A comparable vacuum in the absence of target gas was attained in the target chamber. This allowed for routine operation of the storage ring for other experiments. When the target was in operation a gas flow of 1.2 X 10” ‘He atoms/s was pumped away in the target region and differential pumping stages were used to attain the storage ring vacuum in as short a distance as possible. The 3He gas leaving the cell was pumped differentially along the beam axis. Removal of this background gas was important because it contributed to the stored beam loss rate, and thus adversely affected the average luminosity. Therefore, the design goal of the differential pumping system was to minimize the thickness of the background gas. In the present experiment this was accomplished with three turbomolecular and two cryogenic pumps in an arrangement shown in Fig. 8. Turbo pumps were located on the scattering chamber (Balzers TPH 2200, nominal He pumping speed: 3200 l/s) and on each neighboring pumping stage (Balzers TPH 1500, nominal He pumping speed: 1500 l/s). On the two outer stages, which connected the experimental area to the rest of the ring, cryo pumps (keybold RPK 3OOOS12, nominal He pumping speed: 1300 l/s) were employed.

1300,k

I-

. FL

1500 1:s turbo

3200 l/S turbo

Fig. 8. A schematic layout of the differential vacuum system. The flow limiters between pumping stages are labeled FL. The target cell is labeled TC and the Helmoholtz coils HC.

Cryo pumps have a limited capacity to store helium which made frequent regeneration necessary (typically once or twice a day in this experiment). For this reason, the cryo pumps were equipped with an internal heater which could be activated remotely to regenerate the pumps within a few minutes. The differential pumping stages were separated by gas flow limiters which, like the target cell, were located outside the ring acceptance, both vertically and horizontally. Their position and length were optimized, within the constraints imposed by particle detection requirements, such that the thickness of the 3He gas outside the target cell was a minimum, and amounted to only 2% of the target thickness. The two limiters upstream of the target cell were 11.3 mm wide by 15.3 mm high. Downstream, the first limiter was 16.6 by 17.6 mm’ and the second 26.8 by 23.3 mm2 in width and height, respectively. The typical pressures at a 3He flow rate of 1.2 X 1017 atoms/s, measured by ionization gauges in the differential pumping stages, were (from upstream) 6, 140, 1500, 210 and 20 x 1O-9 Torr.

5. The detectors 5.1. Overview

Fig. 7. A schematic layout of the target cell.

The detectors consisted of three main subsystems designed to simultaneously record data from a number of different reactions: main detector arms to the left and the right of the target, recoil detectors inside the target vacuum chamber on each side of the target cell, and forward angle detectors. The quasielastic events, 3He(p, 2p), 3He(p, pn), and 3He(p, pd), were defined by a charged particle in one of the main detector arms and a charged or neutral particle (a neutron) in the other (see Section 6.2). The internal recoil detectors were used to detect low energy (sub MeV)

C. Bloch et al. / Nucl. Instr. and Meth. in Phys. Res. A 354 (1995) 437-457

444

SL

_

SR _

BPL n

Im

Fig. 9. A side view of the main detector arm. The labels are as defined in the caption to Fig. 2.

particles such as 3He and 2H in coincidence with either the main detectors or the forward angle detectors. The forward angle-recoil coincidence events originated from fast protons due to 3He(p, p3He) elastic scattering in coincidence with the ‘He recoil particles. These detectors were left-right symmetric to allow for separation of the luminosity from spin dependent quantities. 5.2. The main detector arms The main detector arms consisted of a AE-E scintillator pair with two planes of x-y wire chambers positioned between them. A pair of thick backing scintillators were located behind the E scintillators. This is shown schematically in Fig. 9. The AE detectors were 3 mm thick pieces of BC400 scintillator [33]. They were coupled to Burle 8575 photomultiplier tubes with 70 cm long tapered light guides designed to emphasize direct rather than reflected photons. In addition, the AE detectors were configured to give optimum timing resolution while minimizing the effect of the target holding field on the PMTs. Each E detector consisted of a stack of six NE102 scintillator bars 102 cm long, 10 cm high and 15 cm thick. An Amperex 2240 photomultiplier tube was attached to each end with a hyperbolic light guide to optimize the timing characteristics of the bars. Calibration of the pulse height gain, timing, and position within each E detector stack was accomplished with minimally ionizing cosmic rays [30]. Spacers were required between the bars to avoid stressing the light guides, resulting in a 65 cm high stack on the left and a 63 cm high stack on the right. The scintillator arrays covered an angular acceptance of 36” < 0, < 67” and 21” < 0a < 47” combined with - 10” < $I < + 10”. The neutron detection efficiency at 197 MeV was approximately 15%. Multi-wire proportional chambers (MWPC) [31] were used to reconstruct charged particle tracks and thus to determine the event vertex. They were arranged in two

pairs: a large and a small set, located on each of the two major left-right detector arms. The first detector of each pair was positioned with its wires in a vertical direction (XnR, XnL chambers) and the second detector with its wires running horizontally (YnR, YnL chambers). The small forward chambers (XIL, XlR) had nominal wire spacings of 2 mm and 2.5 mm, respectively; the rear chamber (X2L, X2R) wire spacings were about twice as large, giving an intrinsic angular resolution of about 0.2”. The LeCroy PCOSII readout system, employing model 7700 MWPC chamber boards (locally modified for noise reduction), and separate model 2700 CAMAC-based controllers, was used to record the MWPC hit patterns. Alternate odd and even wires were read out on separate chamber boards to minimize cross talk. The wire chamber information was not incorporated into the hardware trigger but was read out for every event generated by the main detector arms. Constraints on the MWPC hit patterns were incorporated in the replay software. The efficiency of the wire chambers was measured to be 99% with minimum ionizing particles. Although the present experiment was conducted in the Cooler A region, which is located approximately half way around the ring from the injection point (see Fig. 3), considerable background was present during injection of the polarized beam into the ring. For this reason the chamber voltages (nominally 3.1 kV for the small, and 2.8 kV for the large chambers) were reduced by 50% during the fill, ramp, and initial cooling phase of the Cooler cycle. This was accomplished by a logic level supplied to the bin gate of commercially available high voltage modules (with a slight internal modification to achieve the 50% voltage reduction) which then held the MWPC at full voltage for a period longer than the data acquisition time during each cycle. The chambers used an argon-ethane gas mixture (50%/50%) to which n-propyl alcohol vapor was added to reduce sparking. The MWPC trip circuit was integrated into the data “run gate” in order to guard against the effects of occasional sparking during the acquisition period. Two scintillator planes were placed behind each E detector stack to provide additional energy information on high energy particles that did not stop in the E detectors. The first plane was 7.0 cm thick with two photomultiplier tubes attached at each end. The second was 2.54 cm thick with three PMT’s attached along the upper edge. The total scintillator thickness was sufficient to stop the highest energy deuterons, thus allowing the separation of protons and deuterons at all energies. 5.3. Recoil detectors For recoil detection, three silicon microstrip detectors were mounted on each side of the target cell, 3 cm from the beam axis and slightly downstream of the target center. These were type 1 detectors from Micron Semiconductor

C. Bloch et al. /Nucl.

Instr. and Meth. in Phys. Rex A 354 (1995) 437-457

[32] with an active area 4 cm high and 6 cm long. One was 500 pm thick, the others 300 pm. Although each detector was segmented into 7 strips, in this experiment the strips were read out in parallel and treated as a single large detector. To reduce noise due to the turbomolecular pumps on the target vacuum system, the detectors were mounted inside a Faraday cage connected to the electronic ground. The other main noise source was the infrared laser light that leaked down into the target chamber through the 3He capillary. To block this light, an additional 0.2 p,m of aluminum was evaporated onto the mylar target cell windows. In addition, a small opaque disk was placed on top of the pumping cell to prevent laser light from travelling directly down the capillary. The effective solid angle subtended by the recoil detectors for the entire target length, weighted by the triangular target density distribution, amounted to 3.1 sr. The microstrip detectors were calibrated continuously with a 2zsTh alpha source mounted inside the Faraday cage. This source had resolvable lines at 5.38, 5.69, 6.29, 6.78, and 8.78 MeV. The recoil detector energy resolution was typically 200 keV (FWHM) with the laser beam shuttered. With the laser on the target, the light noise limited the resolution to about 300 keV and the energy threshold to 0.5 MeV. 5.4. The luminosity monitor detectors The luminosity is the product of the beam current and target thickness. This was monitored measuring 3He(p, p3He) elastic scattering at forward angles outside the acceptance of the main detector arms. The forward scattered protons were detected in separately mounted plastic scintillators and the coincident 3He nuclei were detected in the recoil detectors. There were two left-right symmetric sets of plastic monitor scintillators, one inside (Monl) and one outside (Mon2) the target vacuum chamber. The Monl detectors were 6.35 cm high, 4.45 cm wide, 1.27 cm thick, and centered at 27.8 cm on a line 16.1” from the target center. The angular acceptance of the scintillators was centered near the forward maximum in the analyzing power of the 3He(p, p3He) reaction at all incident beam energies. The Mon2 scintillators were 10.2 cm wide, 25.4 cm tall, and 5.1 cm thick, were centered at 115.1 cm on a line 18.0” from the center of the cell. The scintillator material for the Monl detectors located in the Cooler vacuum was Bicron BC-400 plastic scintillator [33]. A polystyrene light guide, chosen due to its lower outgassing rate compared to plexiglass, brought the light from the scintillator to an optical viewport. The scintillator and light guide were loosely wrapped in one layer of 12.7 pm thick aluminum foil, which had a slit to allow easy escape of the gas load from the plastic. The additional gas load and reduced bakeout temperature due to the Monl detectors did not significantly affect the quality of the ring vacuum.

445

6. Electronics and data acquisition 6.1. Data acquisition and readout Online data were processed using a network of 4 computers plus one VME crate for CAMAC readout. A schematic block diagram of this arrangement is shown in Fig. 10. The main data acquisition and control computer of the system was linked to the other computers via an Ethernet network. Four CAMAC crates were read out through the VME crate which sent buffered data by Ethernet to the main computer, where the data were written to disk (later archived to 8 mm tape) and sorted into histograms that could then be used for online analysis. The control computer for the polarized target transmitted updates of the target status in five minute intervals to the main computer. 6.2. Trigger The utilization of a windowless internal target with a cooled proton beam provided a relatively background free environment. To take advantage of the low background and the large kinematic acceptance provided by the various detectors, the electronic trigger system was designed with the flexibility to accommodate all possible coincidence combinations. To initiate the “fast” trigger logic a coincidence between the AE and E detectors on one side was required, indicating the passage of a charged particle. A coincidence was then sought between this first level trigger and a signal from one of the detectors on the opposite side of the beam. The first level trigger was also used to start the conversion of the analog scintillator signals in fast encoding ADC and TDC modules. In the absence of a double-arm coincidence, a delayed version of the first level trigger provided a fast clear to the ADC and TDC modules. The

Fig. 10. Schematic block diagram of the electronics acquisition system used in the experiment.

and data

446

C. Bloch et al. / Nucl. Instr. and Meth. in Phys. Res. A 354 (199.5)437-457

total electronic processing time (dead time) for each event was about 2.6 ps. A valid (p, pn) trigger was identified as a coincidence between a first level trigger on one side of the beam with an E detector on the opposite side. The (p, 2p) events were identified in software by determining that the AE and wire chambers had fired on both sides. In addition to events involving a coincidence between the two main AE-E detector arms, the trigger electronics also responded to coincidence between the first level trigger and any of the silicon microstrip recoil detectors on the other side of the beam. Thus, events where one charged particle (a proton) is detected on one side, and one heavy recoil particle is detected on the other, were recorded in the data stream. The most important event of this type was 3He(p, p3He) elastic scattering. A third (“slow”) trigger system was implemented to handle events in the luminosity monitors. Since both the fast and the slow trigger can be generated by a coincidence involving the microstrip detectors, it was necessary that a trigger in either system inhibit the generation of a trigger in the other, so that the recoil detector ADC’s and TDC’s did not receive multiple starts. Other than this mutual inhibit, the fast and slow triggers operated independently. A fourth trigger was generated every 0.1 s to initiate scaler reads. These scalers monitored rates at various stages of the electronics. For the fast trigger, valid information from the detector ADCs and TDCs and wire chambers were read out with zero and overflow suppression. Detector hit patterns, along with status bits indicating the beam and target polarization and the target holding field direction, were latched into a coincidence register for later offline decoding and filtering. Due to the time structure of the cooler beam it was possible to make a direct measurement of the quasielastic accidental rate. The time separation between the cooler beam pulses was typically about 100 ns, and so it was possible to delay a copy of the first level trigger into the next beam burst, and then look for coincidences with the undelayed first level trigger from the other side of the beam. The accidental coincidence rate measured in this manner was typically about 0.1% of the total quasielastic trigger rate at the luminosities attained during the experiment. 6.3. Target diagnostic data acquisition and remote control To make available diagnostic information on the target system and to record important experimental data, a number of target parameters were monitored in parallel to polarization data acquisition. The diagnostic data were continuously monitored for immediate recognition of potential problems and inserted into the data stream at regular intervals (every 5 min). A modular system of converter and adaptor units was developed to allow for easy interchange and addition of

analog signals to the diagnostic data stream. Each signal of interest was converted to a voltage in the range + 10 V and digitized in one of two CAMAC 32-channel scanning digital voltmeters (DSP Model 2032). The physical channel assignments at the voltmeters were translated into logical signal numbers via a lookup table in software. Additional diagnostic information was acquired via GPIB and from a digital input register in order to monitor the status of several instruments (lockin amplifier, Gaussmeter, rf oscillator) and the target control electronics. Remote control of the target apparatus was facilitated by standard CAMAC analog and digital output. Of special importance for the performance of the target during data taking was the ability to control the gas flow rate and the polarization of the laser optical pumping light. The gas flow was regulated with a Balzers UDV135 thermovalve, whose setpoint could be programmed via CAMAC. Similarly, the laser polarization could be altered by varying the voltage on the LiNO, Pockels cell (INRAD 204-080). A high-voltage amplifier (Lasermetrics Model 8403) was used as a driver for the Pockels cell.

7. Luminosity and spin monitoring

7.1. Introduction The integrated, relative luminosity must be known for each combination of spin states in order to extract asymmetries. Two independent techniques were used for this in the CE-25 experiment. One method was to independently monitor the circulating beam current using the parametric current transformer (PCT) [23] and to monitor in addition the target thickness using a precision pressure transducer. The second method of determining luminosity was based upon detection of 3He(p, p3He) elastic scattering in a set of scintillators at forward angles (the luminosity monitors), as described previously in Section 5.4. The formalism for extracting the luminosity and beam polarization from the monitor results is outlined in Appendix A, and the numerical results are listed in Appendix B. 7.2. Analysis of the luminosity monitor data The elastic scattering events were isolated using the correlation between the pulse height in the plastic scintillatars, the calibrated pulse height in the silicon detectors, and the relative time difference between the silicon detectors and the plastic scintillators. The forward protons passed through the scintillators at all energies, giving a proton energy loss and a timing start signal for the event. Software conditions for these events gated a spectrum of the pulse height in the silicon detector. A typical spectrum is plotted in Fig. 12 for the 197 MeV sorting conditions. The background at about 1 MeV is due to 3He(p, 2p) events; this peak was also present when the target cell was filled

C. Bloch et al. / Nuel. Instr. and Meth. in Phys. Res. A 354 (1995) 437-457

447 WC 2

earn

I

I

: :

Reconstructrd

Slllcon

Track

Detectors

Fig. 13. A schematic

layout of the CE-01 detector [34]

CE25 Trlgger Circuit Fig. 11. Schematic diagram of the trigger system used in the experiment. Octal logic modules (0.L.M.) were used to generate the event trigger.

with H, indicate

gas. The dashed the

summing

summing

limit

was

lines at 3.2 and 15.0 MeV

region Ec3He)

for

the

peak

= 7.50-23.00

sums. MeV

The

at 414

MeV incident proton energy. The small amount of real 3He(p, 2p) background that extends above 3.20 MeV has no effect upon the final results, since the polarization dependence of the monitor is calibrated for a given set of sorting conditions. The calibration was insensitive to small changes in the sorting conditions within statistical uncertainty.

7.3. Analysis

of data

from the CE-01 detector

It was necessary to calibrate the effective analyzing power and spin correlation parameter of the luminosity monitor. This then allowed the separation of beam and target analyzing powers from the luminosity. The details are described in Appendix A.

The CE-01 detectors and an unpolarized hydrogen target were used to determine the luminosity and beam polarization during the monitor calibration. Its operation was the same as in previous analyzing power measurements at these energies [34]. The detectors comprise a combination of scintillators and multi-wire counters optimized for the detection of forward going charged particles. A schematic view of the apparatus is shown in Fig. 13. Both the forward scattered and recoil protons were detected in coincidence. The low energy recoil protons were detected in two 4 X 6 cm’, 600 pm thick silicon detectors mounted in the innermost chamber of the target. The detectors were mounted at 5.7 cm from the nominal beam axis. The inner chamber of the target was filled with H, gas to form a long, diffuse target. The pressure was adjusted to roughly balance the count rates in the CE-01 and CE-25 detectors. The wire chambers covered a laboratory angular range of 5” to 18” with limited azimuthal coverage at the larger angles. The analysis procedure was similar to that in earlier polarization measurements with the detector [34]. The coincidence requirement resulted in a large fraction of the data being good ’ H(p, 2~) scattering events. Fig. 14 shows

300

250

EOO

150

100

50

0

50

3He Km&c

10 0

Energy

15 0 (tdev)

Fig. 12. 3He pulse height spectrum for the luminosity monitor at an incident proton energy of 197 MeV. The dashed lines indicate the summing limits.

200

400

ml

800 1000 1200 Pulse Height

1400

le.00

1800

2000

Fig. 14. The correlation between forward scattering angle and energy deposited in the silicon detector for the CE-01 detector system.

448

C. Bloch et al. /Nucl. Instr. and Meth. in Phys. Res. A 354 (1995) 437-457

the kinematic between the scattering angle in the chambers and pulse height in the detector. This has a requirement that event vertex within f cm of target center. that there almost no away from kinematic locus ‘H(p, 2p) tering. The ‘H(p, 2p) were required have a which intersected vertex, hit wire chamber the opposite from the silicon detector, intersected the thick E In addition, was required the target be flowing at least s, allowing H, gas to stabilize. typical computer was only few percent was indeof the state of beam. The streams of CE-01 and detectors shared common signal formed hardware failures. final data sorted by into histograms the laboratory angle. The and right were inspected determine a range of at 197 consistent with azimuthal acceptance the detector. of the conditions was from the values, and statistically significant in the of spin luminosities was 7.4. The

analysis

The of the was also from measurements the relative thickness and current. The thickness was from a of the in the cell, which acquired as of each There was solid connection the pumping to the target cell, thus the thickness was related to measured value the input to the cell. The current was with a current transformer [23]. This is a transformer, capable measuring beam of less 100 p,A of beam structure and variations in beam’s position size, but a noise accuracy. To precise measurements the PCT is necessary monitor drifts the zero level. To the gain the PCT known flux generated with fixed current through the on a At least 10 s, pA calibration per cycle passed through wire to the reAt the of each the beam dumped about s before start of next cycle, a measurement the DC The PCT voltage was to a frequency and on a Hz timescale, the beam was reconstructed software. A histogram of MeV beam sity is in Fig. The PCT data were fit to a functional form which was the sum of a smooth decay, the calibration pulses, and a background term with a linear time dependence. The fit

Deviation

from Centroid

(u)

Fig. 1.5. Histogram of the ratio of the PCT to luminosity monitor normalizations at 197 MeV. The binning unit (T is the average of the statistical uncertainties for all 134 comparisons.

was made for each cycle. The width of the calibration pulse and the slope of its edges were also parameters of the fit. The time dependence of the DC offset was found by simultaneously fitting all cycles in a given run. Only runs which were good fits to the functional form were accepted - which was the case in the great majority of the data runs. The smooth decay part of the fit was then used to determine the target spin-sorted charge. The luminosity was generated from the product of the extracted spin-sorted charge and the target pressure. 7.5, Comparison of the two methods The comparison of the luminosity monitor and PCI normalizations is a necessary part of the measurement. The PCT normalization was divided by the luminosity monitor results for the 197 MeV data, and the statistical uncertainty of this ratio was given by the propagated statistical uncertainties. The distribution plot in Fig. 15 shows the scatter of the ratio for 134 separate comparisons. The data are binned in units of the (simple) average statistical uncertainty of the ratio. The Gaussian shape of the distribution, with width = (T, shows a good agreement between the two methods. The results of the experiment do not depend upon the choice of normalization method.

8. Performance The polarized internal gas target results in relatively low luminosity which is well matched to a large acceptance detector. In addition, the backgrounds from the internal target technique are low. In this section the performance of the apparatus is described. In particular, those experimental quantities which are central to the quality of the measurement are discussed in detail.

C. Bloch et al. /Nucl. Instr. and Meth. in Phys. Res. A 354 (1995) 437-457

449

8.1. Vertex reconstruction and background One of the principle advantages of the polarized internal target technique is the possibility to perform scattering experiments using an isotopically and chemically pure target species. High purity polarized 3He gas atoms are injected into the storage cell at the center. The gas flows out the open ends of the cell resulting in a density distribution that is approximately triangular and centered on the cell. Fig. 16 shows the vertex distribution along the beam axis (z) from wire chamber raytracing of two final-state protons from the 3He(p, 2p) reaction. The distribution is in good agreement with the expected shape and indicates that backgrounds are small. In order to make quantitative estimates of the background in the quasielastic scattering measurements, e.g. from either the walls of the cell or unpolarized background gas outside the cell, it was necessary to take data on scattering from hydrogen gas. Such measurements cannot simply be made with an empty cell since the beam phase space properties are determined from a complicated interplay of beam heating effects due to scattering in the target and damping effects from the electron cooling. Fig. 17 shows the distribution of the sum of the two outgoing proton angles for 197 MeV. The distribution is peaked around eL + H, = 87” (FWHM N 1”) as expected for equal mass elastic collisions in relativistic kinematics at this energy. At a level of lo-’ from the maximum there is a broad distribution of events that do not have the characteristic angle correlation of free scattering. We note that this spectrum without any cuts is already much cleaner than that which would be obtained with alternative techniques. 8.2. Time of flight calibration The A E-E detectors were used in coincidence to measure the time-of-flight of charged particles. The time-offlight and the length of the charged particle’s flight path were determined using the reconstructed trajectory mea-

2000

,

I

I

I

0

20

I

1500 -

1000

-

500 -

0 -40

-20

40

Z (cm) Fig. 16. Reconstructed vertex position along the beam axis from wire chamber ray-tracing of 3He(p, 2~) reaction products.

0, + 0. (degrees)

Fig, 17. The distribution of ‘H(p, 2~) elastic scattering events at 197 MeV as a function of the angle between the protons in the final state.

sured with the wire chambers. This information was then used to calculate the particle’s speed which was then corrected for energy loss in the detectors, the surrounding air and the target chamber window to provide the particle’s momentum immediately after the reaction. In the case of neutrons, which were defined as particles which triggered the E but not the AE detectors nor the wire chambers, the time-of-flight was calculated from the relative timing between the E-bar that stopped the neutron and the reaction time calculated by projecting back along the trajectory of the proton detected in the opposite arm. The time-of-flight (TOF) calibrations were based on the elastic ‘H(p, 2~) measurements which were routinely carried out. The TOF calibration was started by the left or right AE-detector signal, depending on the side the “fast” proton was scattered. There were three stop signals: the AE on the side of the “slow” proton and either of the E detectors. This provided a complete set of timing signals to determine all the required kinematic quantities, since the geometry of the detector system was known and the azimuthal and polar scattering angles were determined from the wire chamber information. It was important to correct for the energy loss of the proton as it passed through the different materials in the detector. The calibrations were obtained by comparing scatter plots of data (e.g. TOF vs eP or kinetic energy vs eP> with the theoretical expectation. Each E scintillator bar was calibrated independently. Only elastic ‘H(p, 2p) events, which had a vertex position along the beam axis within + 1 cm from the center of the target, were used. This corresponded to the vertex resolution of the wire chambers when multiple scattering is included. Fig. 18 shows calibrated TOF plotted against scattering angle. At 197 MeV the TOF resolution was 1.7 ns (0). The solid curve is an energy loss corrected calculation. This method allowed the TOF offset to be determined to better than 50.1 ns or about + 1% for

450

C. Bloch et al. /Nucl, Instr. and Meth. in Phys. Res. A 354 (1995) 437-457

knock-out reaction 3He(p, pd) whereas the lower band corresponds to quasielastic (p, 2p) scattering. 8.3. Angie calibration

lo.o

theta (deg)

Fig. 18. Calibrated TOF vs scattering angle compared calculated value for 197 MeV incident proton energy.

to the

a TOF of 10 ns and an incident proton energy of 197 MeV. At 414 MeV the accuracy decreased to about L-0.2 ns. The run-to-run fluctuations were of comparable magnitude so that the total error in the TOF calibrations was about + (2-3)%. Time-of-flight information was also useful for particle identification. Fig. 19 shows the correlation between timeof-flight and energy deposited in the E scintillators. The upper band corresponds to deuterons from the direct

‘He(p,Zp)

200

1

175

:

150

r

125

r

100

1

75

r

50

Y

25

I

and

(p,pd)

The angular calibrations of the detectors were carried out using elastic ‘H(p, 2p) scattering. In this reaction, the angle between the two outgoing protons is determined solely by relativistic kinematics and the beam energy as shown in Fig. 17. In addition, since the scattering involves identical particles, the analyzing power must be zero at 90” in the center of mass frame. The absolute angles of the detector arms were determined to f0.6” using a two-step procedure. First, the nominal angle between the detectors was adjusted to give the correct kinematic opening angle between the outgoing protons. Then the angle between the beam and the detector pair was adjusted to give the proper zero crossing angle in the ‘H(p, 2p) analyzing power for both arms. The measured analyzing powers at 414 and 197 MeV are shown in Fig. 20. The main uncertainty in this process was in determining the luminosity for the two beam spin states, estimated to be less than *3% arising from the PCT. This uncertainty primarily accounts for the quoted angle uncertainty. Fig. 21 shows the distribution of ‘H(p, 2p) elastic scattering events in azimuthal angle at

particle

ID at 200

MeV

o-

Pulse

Height

(MeV)

vs Time-of-Flight

Fig. 19. The correlation between TOF and energy deposited in the E detector for quasielastic applied for particle identification.

(ns)

events at 197 MeV. The lines indicate the cuts

451

C. Bloch et al. /Nucl. lnscr. and Meth. in Phys. Rex A 354 (1995) 437-457

-0.2 -0.4 I 20

I

I

I

30

I I I I I, 40

50

60

I 70

I

8

t~~‘~‘~‘~‘~“~l 0

lab

= right arm

20

A 5 left arm = SAID phase shift

30

40

50 8

60

70

lab

Fig. 20. The beam analyzing power in ‘H(p, 2p) elastic scattering as a function of laboratory scattering angle at 414 and 197 MeV. The curve is the SAID phase shift fit SM89 [36].

197 MeV. This demonstrates momentum vectors. 8.4.

Low energy

the coplanarity

of the two

hadrons

8.5. Kinematic

Another important advantage of the internal target technique is the possibility to detect low-energy hadrons. The target thickness is extremely small (several ng/cm”) and the walls of the target cell can be made ultra-thin: 1.7 pm of mylar were used in the present experiment. Elastic “He(p, p3He) scattering events at 197 MeV incident energy were identified by measuring the proton in coincidence with the 3He nucleus detected in the Si-strip detectors located inside the vacuum chamber. Fig. 22 shows the correlation between the proton angle and ‘He energies. The locus for elastic “He(p, p3He) scattering is clearly visible. The threshold for detecting low-energy hadrons was set by the noise from the turbomolecular pumps and the pickup of laser light by the detectors. Heavily ionizing “He recoil nuclei were detected with energies as low as 500 keV. In Fig. 22 the two bands correspond to recoil

150

170

an -

particles passing through the 300 and 500 pm thick silicon detectors.

180

190

200

210

220

aL (degrees)

Fig. 21, The distribution of ’ H(p, 2p) elastic scattering events as a function of the azimuthal angle between the protons in the final state at 197 MeV.

quantities

A central aspect of the CE-25 experiment was the ability to analyze the data in terms of the initial 4-momenturn of the quasielastically scattered nucleon. Thus, the quasielastic coincidence data were described in terms of two kinematic variables [35]: the missing momentum defined as pm =pinc -p, -pz and the missing energy defined as E, = T, - T, - T2 - qecoi,, where qecoil is the kinetic energy of the recoiling system. The missing momentum p, and missing energy E, are the most important kinematic quantities for extraction of information on the ground state spin structure of polarized 3He. The large acceptance of the detector configuration used in this experiment allowed simultaneous data taking over a large kinematic range. Fig. 23 shows this range in terms of missing energy and momentum at 197 MeV incident proton energy. Approximately 2 X 10” 3He(p, 2~) events and 2.5 X 10’ ‘He(p, pn) quasielastic events were acquired at each incident proton energy. The resolution in these quantities for the CE-25 experiment was determined using the reaction ‘H(p, 2~1, where in the limit of perfect resolution one should find E, =p, = 0. We find AE, = 1.5 MeV and Ap,, = 20 MeV/c, APmr = 14 MeV/c, and Ap,,,; = 26 MeV/c (FWHM). In Fig. 24 we show the measured distributions for E, and p,. It is clear that this missing energy resolution does not allow the separation of the 3He(p, 2p)‘H two-body breakup channel, at E, = 5.5 MeV, from the 3He(p,2p)n channel which opens up at E, = 7.7 MeV. The two-body breakup channel was isolated in two different ways. First, the deuteron knockout reaction ‘He(p, pd) provides a direct measurement of the two-body breakup. That this reaction can be cleanly isolated is demonstrated in Fig. 19. Secondly, we have measured the

C. Bloch et al. /Nucl. Instr. and Meth. in Phys. Res. A 354 (1995) 437-457

452

3He(p,p3He) 35

at 200

I

.

A

.

. . . .

30

25

MeV

. ..

-

I-

20

-

15

-

10

-

. .

l

.

.

nA

.

..

.

_m

. I.

/

. _I

.

‘He(p,Zp)

200

MeV

100 lo2

80 60 40 20

1%

0 -20 -40 -60 -80 _ iO0

E, (MeV)

vs

pm (Mev/cj

Fig. 22. (top) The correlation between the proton angle and the ‘He energies indicate the cuts applied for particle identification. Fig. 23. (bottom) The distribution

of ‘He(p, 2p) quasielastic

in ‘He(p, p3He) elastic scattering

events in terms of missing energy and momentum

at 197 MeV. The lines

at 197 MeV incident energy.

C. Bloch et al. / Nucl. Instr. and Meth. in Phys. Rex A 354 (I 995) 437-457 20000

-

14OQO

12000 15000

10000

-

8000

-

6000

-

4000

-

2000

-

12500 10000 7500

:

5000

:

-80

-60

-40

-20

0 Em

20

40

60

80

0 -100

,()I,,,’ -80

-60

-40

-20

0

20

40

60

80

pm2 (MeV)

(MeV)

Fig. 24. The reconstructed missing energy and missing momentum distributions in ’ H(p, 2~) elastic scattering at 197 MeV.

recoil deuterons after proton knockout from 3He. Determining the energy of the recoiling nuclear system in the three-fold coincidence measurement 3He(p, ppd) provides a direct measurement of the missing momentum. In Fig. 25 we show the energy deposited in the recoil detector vs. the magnitude of the missing momentum p,. There is a clear locus which we identify as events v&e a deuteron is carrying the full missing momentum thus identifying the 3Hetp, 2p)‘H two-body breakup channel.

‘He(p,Zp)

8.6. Systematic uncertainties The dominant contributions to the uncertainty in the determination of asymmetries were due to: the beam polarization, the target polarization, and the determination of the spin-sorted luminosity. 8.6.1. Uncertainty in the beam polarization The beam polarization was determined by comparing the ‘H(p, p) elastic scattering measurements carried out

200

MeV

I 1C2

8

10

6

0

50

:oil

Energy

100

(MeV)

150

vs Missing

200

Momentur

Fig. 25. Correlation between the recoil detector energy and the missing momentum at 197 MeV incident proton energy

454

C. Bloch et al. /Nucl. Instr. and Meth. in Phys. Res. A 354 (1995) 437-457

with CE-01 apparatus and a phase shift fit to previous data (see Appendix A). The fractional uncertainty (APJP,,) in the knowledge of the beam polarization was f 1.5%. 8.6.2. Uncertainty in the target polarization The polarization of the 3He atoms was determined in the pumping cell by the optical monitor technique as described in Section 4.2 above. There is a fractional uncertainty of f 2% in the calibration of this technique. In addition, there was a noise on the detection system of f 3.5%. The polarized 3He atoms travel from the pumping cell to the target cell. All estimates of depolarizing effects produce less than 1% fractional decrease in the polarization observed in the pumping cell. These estimates were supported by a comparison of the measured target analyzing power for 3He(p, p3He) elastic scattering at 45 MeV with previous data [7]. Thus, we have assumed that the polarization in the target cell is identical to that measured in the pumping cell and have assigned a fractional uncertainty (AP,/P,) of *4% to the knowledge of the target polarization. 8.6.3. Uncertainty in the luminosity The relative luminosity between spin states was determined using two techniques. The method for determining the luminosity from the PCI and target thickness is described in Section 7.4 and the monitor detector method is discussed in Section 7.2 and appendix A. From the data on ratios of PCI to monitor detector luminosities we conclude that the random error on luminosity determination was f0.7% and the systematic error was f 1.3%. The actual extracted spin observables for 3He(p, 2~1 and 3He(p, pn> differed by at most 0.02 between the two luminosity determinations. Subsets of the data taken with different field directions were consistent within statistical uncertainties. Simultaneous with the quasielastic measurements we have carried out measurements of the beam and target analyzing powers and the spin correlation parameter in elastic 3He(p, p3He) scattering at each energy. The 197 MeV data have been compared in detail with previous data from TRIUMF. The beam and target analyzing power data are in good agreement. From these direct comparisons of the data for internal and external consistency we conclude that the systematic error due to the luminosity deterrnination was f 0.02 for each of the measured spin observables.

9. Summary A laser optically pumped polarized 3He internal gas target has been successfully used in a medium energy storage ring with a polarized beam to carry out an extensive series of spin-dependent scattering measurements. The unique advantages of internal targets such as purity, high polarization and the ability to detect low energy recoil

nuclei have been demonstrated. A large acceptance detector has been used to simultaneously measure the scattering asymmetries in elastic and quasielastic scattering. The techniques developed in this experiment for luminosity and stored beam polarization monitoring are of general import for future experiments using polarized beams and polarized internal targets. The target thickness of 1.5 X 1014 atoms/cm’ was limited by the rate at which data could be acquired without substantial deadtime corrections. For other experimental configurations, e.g. medium energy electron storage rings, the target thickness could be at least one order of magnitude thicker using a higher flow rate and cryogenic cooling of the target cell [39]. The CE-25 experiment has demonstrated that the use of polarized internal target and beam with a large acceptance detector allows the study of the spin dependent reaction over a large kinematic range with high statistical and systematic precision. Similar experiments planned with electron beams [4] should provide measurement of electromagnetic polarization asymmetries over a broad range of momentum transfer.

Acknowledgements This work is supported in part by the National Science Foundation under Grants no. PHY-9015957 (IUCF), PHY9019983 (Madison), PHY-9018242 (Ohio State) and the NSF Research Opportunity Award Program (Louisville and W. Michigan) and by the Department of Energy under Contract no. W-31-109-ENG-38 (ANL) and DE-AC0276ER03069 (MIT). RGM acknowledges a Presidential Young Investigator Award from the NSF.

Appendix A

Formalism for the calibration and operation of the luminosity monitors The reaction rates measured in the CE-25 luminosity monitor depend upon beam and target polarization since the elastic scattering reaction is polarization dependent. This polarization dependence must be calibrated for an accurate determination of the luminosity from the measured yields. The reaction rate is,sensitive to the effects of the beam analyzing power A,,, target analyzing power A-, and spin correlation coefficient A-. The analyzing powers are left-right asymmetric, and their effects can be minimized in a left-right symmetric detector. The spin correlation coefficient A,, results in a left-right symmetric change in the measured yields, and its effects cannot be eliminated through a proper choice of monitor detector geometry. The effect of the AOOnnspin correlation

455

C. Bloch et al. / Nucl. Instr. and Meth. in Phys. Res. A 354 (1995) 437-457

coefficient must be determined in a separate experimental calibration with precise, independent measurements of the beam current, beam polarization, and target polarization. The luminosity and beam polarization were measured with the CE-01 detector using an unpolarized internal hydrogen gas target during this separate calibration process. The CE-01 detector was located in another section of the ring, and was operated simultaneously with the 3He target and CE-25 luminosity monitor during the calibration. The 3He target polarization was determined from the optical signal, as described in Section 4.2 above. The calibration of the A wnv, A,,, and A,, effects were determined from the measured CE-25 monitor yields and the CE-01-derived luminosity and beam polarization. The formalism for the response of the monitors is developed below in detail for the + + combination of beam and target polarization with a simplifying assumption of equal magnitude polarization in each of the two spin projections. The left and right peak sums are L ++ =

ciintA++

41

+~~,A'oono

+f't4,,,.

'YR(~

The luminosity A ++ and the integrated cross section a,,, = /da o( f3) d R will affect both sides equally. The (Ye and ~a coefficients explicitly include those effects which may differ for the two sides, such as the detector solid angle and efficiency. The effective polarization observables such as A’,, result from weighting the observable with the cross section over the solid angle of the detector. The effective beam analyzing power A’,,, for example, is defined as

(A.3) and R,,

h++gint{(aL+

EOO~

+

Efalse(

‘hn~

+

ECNXI~)

The gint(aL + a,> term will affect all spin state combinations equally and can be ignored. In the calibration process the individual A’s are known from the CE-01 detector, and the polarization dependence of the monitor can be determined. The left/right differences can be used to extract the polarization of the beam and target. The asymmetry of the + + combination of beam and target polarization is L ++ -R++ E++=

L+++

R,,

pFSA’,,*

=

+ PTA&~,

1 +PTPBAoo,~~

+ %,x(1

+ PTP~A,,,)

+ %~PB&Io~~~

’

+PTA’OOO,)

which can be rewritten in terms of the measured asymmetries as

1 + %o””

+

The individual asymmetries then rearranged as

for the spin combinations

= %0”0 + %o” + %se(l E+-{l

I1 =

=

are

+ %I”” ).

- Eoonn+ %lse(~oOnll - ~000”)~

= %o”O - %Gi3”+ E-+

64.9)

loono + EOOO” )I

E*+ { 1 + QO”” + %e(

-

EOO””

-_E

oono

is "R)(1+PTPBA60nn)

Efadl + %d Efalse( %o”O + %30” 1

%onO+ %o” + EC+=

E--I1

L+++ R++=

1 ’

(A.7)

-P,Ab,,,-P,Ab,,,+P,P,Ab,,,).

(A.2)

The sum of L,,

L+++R++

ai”,( &L -I- (yR)

(‘4.8)

and aintA++

1 A ++=

+f'd't4w,,)~ (A.11

R ++ =

which can be rewritten to find the luminosity

+ -

+

loonn

%o”O

-

+

Efad

%M” +

Efalse Cl+

~oono

Efalse(l

EfalkEoono %xQ”

+

Efalse (1

%ll”n)~ +

~OOO”~l

-

%o”“h

-

~000”)~

+

%I”“).

(A.10)

f(a~-~Yto(~~AbOn~f~~AbOan)}. (A4

The origins and effects of these corrections are more easily seen by rewriting the equation in terms of the measured polarization asymmetries lOOnO= P~A’~,,~, eOOOn= pT A’-, and Ed, = pepTAftin. The effects of those false asymmetries due to (Ye # (~a are best included as L,-R, Efalse = -&+&I

=p

(Y,_-LY~

(A.9

(YL+aR

In this notation the sum becomes L+++R++=

h++aint(aL'

(yR)

X{l+~~~~+~false(~,,,+~,,j}, (A-6)

The appropriate combinations of these corrected asymmetries yields the polarizations. The combination E++ + E+_ - l_+- E__, with each term corrected by the appropriate 1 * Eh + Ef**J f Em0 f Eooon>, gives an excellent approximation to 4~~~ since the target polarization and false asymmetry terms cancel. The separate plus and minus polarizations can be found from combinations such as l++ + E+_. The false asymmetry contributions to this sum do not cancel, however, and it is necessary to know efalse to determine the separate plus and minus polarizations of the beam. Any possible changes in efalse can be detected by monitoring the measured asymmetries E,,, E+_, E_+, and E__ for internal consistency throughout the experiment. Changes in efalsc would appear as a global shift of all 4 of these measured quantities. However, no significant

456

C. Bloch et al. /Nucl. Instr. and Meth. in Phys. Res. A 354 (1995) 437-457

shift of the measured asymmetries was observed. Similar relations can be formed for the target asymmetry. These relations are normalized to the optical signal from the pumping cell, and serve only as a measure of the internal consistency of the experiment.

Appendix B

Results of the calibration The average beam polarization measured with the CE-01 detector was pn = 0.722 + 0.010 during the 197 MeV calibration, determined using the cross-ratio method for the polarization asymmetry and the WA0 phase shift solution of Amdt [36]. This phase shift solution is a local fit to data from O-350 MeV, and its immediate predecessor (VL35) is known to be a very accurate description of analyzing power data in this angle and energy range [34]. The use of the cross-ratio technique results in a precise determination of the average polarization [37]. The individual polarization states can be determined if the false asymmetry is known. The false asymmetry lfalse of the CE-01 detector was era,= = - 0.068 f 0.006, measured with an “unpolarized” beam formed by turning off the rf hyperfine transition units in the atomic beam source. There is a small residual polarization of about -0.04 in this beam due to the solenoidal field of the electron ionizer, and its contribution to the false asymmetry has been subtracted. The polarizations of the individual states were pB = 0.771 + 0.028 and pB = -0.675 + 0.027. The average target polarization was 0.457, determined from the 667 nm optical signal. The average beam and target analyzing power asymmetries of the CE-25 luminosity monitor were eOOn,,= 0.477 f0.003 and lOOoo= 0.111 = 0.005. These asymmetries were determined using the cross-ratio method with either the beam or target unpolarized. The spin correlation asymmetry was ~~~ = 0.116 + 0.006. The false asymmetry, required for corrections to the luminosity, was lfalse = -0.076 f 0.018 for the luminosity monitors. The above errors are statistical only. Dividing by the appropriate polarization combination yields the systematic uncertainties of the extracted 3He(p, 2p) and 3He(p, pn) polarization FiA,, = +0.004, and observables: 6A,, = f0.013, SA,, = kO.017. The average beam polarization was 0.693 f 0.012 at 414 MeV, including the statistical uncertainty of the preThis result is generated using the C400 diction for A,,. solution of Amdt. This solution is a local energy solution which includes data from 375 to 425 MeV. The quoted uncertainty of this solution (6(A,,) = 2 X 10e3) is much less than the spread of available phase shift solutions. The C400 solution was judged to be the best representation of the data since the Amdt local energy solutions best repro-

duced absolute polarization measurements in this energy range [38]. The false asymmetry was -0.112 f 0.008, and the polarizations of the individual states were 0.724 + 0.029 and -0.645 f 0.029. The average target polarization was 0.458. The average beam and target analyzing power asymmetries of the luminosity monitor were Ed,, = 0.422 + 0.006 and lOOOn = 0.117 + 0.005, determined from cross-ratio asymmetries with unpolarized target and unpolarized beam respectively. The spin correlation asymmetry was es,,,,” = 0.010 f 0.006. The false asymmetry, required for corrections to the luminosity, was lfalse = -0.046 k 0.019 for the luminosity monitor. The above errors are statistical only. Dividing by the appropriate polarization combination yields the systematic uncertainties of the extracted 3He(p, 2p) and 3He(p, pn) polarization observables: iSA,, = +O.Oll, 6A,, = kO.004, and 6A,, = _t 0.017.

References

[II

K.-H. Altoff and W. Meyer, eds., Proc. 9th Int. High-Energy Spin Physics, Bonn, Germany, 1990, Vol. 1. (Springer, Berlin, 1990). El R. Gilman et al., Phys. Rev. Lett. 65, (1990) 1733. [31 NIKHEF proposal 91-12, spokesmen J.F.J. van den Brand, C. de Jager. [41 Bates Large Acceptance Spectrometer Toroid (BLAST) proposal, September 1991. unpub151HERMES proposal to DESY, DESY-PRC-90-01, lished 1990. b1 M.A. Ross et al., Nucl. Instr. and Meth. A 344 (1994) 307. [71 K. Lee et al., Phys. Rev. L&t. 70 (1993) 738. [81 B. Blankleider and R.M. Woloshyn, Phys. Rev. C 29(1984) 538. [91 R.G. Mimer, Proc. Workshop on Polarized 3He Beams and Targets, Princeton, 1984, AJP Conf. Proc. no. 131, p. 186. 1101M. Meyerhoff et al., Phys. Lett B 327 (1994) 201. 1111H. Gao et al., Caltech preprint OAP-729, Phys. Rev. C 50 (1994) R 546. I121 Bates proposal 88-25, spokesmen A. Bernstein, T. Chupp, R. Milner. [I31 CEBAF proposal 89-020, 1989, spokesman R.D. McKeown. D41 P.L. Anthony et al., Phys. Rev. Lett. 71 (1993) 959. [151 SLAC proposal E154, spokesman E. Hughes. [161 J.L. Friar et al., Phys. Rev. C 37 (1988) 2852. [I71 A. Rahav et al., Phys. Lett. 275 (1992) 259. iI81 A. Rahav et al., Phys. Rev. C 46 (1992) 1167. [191 E.J. Brash et al., Phys. Rev. C 47 (1993) 2064. DOI J. Bystricky et al., J. Phys. (Paris) 39 (1978) 1. ml R.E. Pollock, Ann. Rev. Nucl. Part. Sci. 41 (1991) 357. m.1 M.A. Ross et al., Nucl. Instr. and Meth. A 326 (1993) 424. [23l K. Unser, IEEE Trans. Nucl. Sci 28 (1981) 2344. [241 R.E. Pollock, Proc. 19th Symp. on Cooler Rings and their Applications, eds. T. Katayama and A. Noda, (World Scientific, Singapore, 1991) p. 4. P51 J. Sowinski, in: Symmetry and Spin in the Standard Model, eds. B.A. Campbell, L.G. Greeniaus, A.N. Kamal, F.C. Khanna (World Scientific, Singapore, 1992) p. 360.

C. Bloch et al. /Nucl. Instr. and Meth. in Phys. Rex A 354 (1995) 437-457 [26] F.D. Colegrove, L.D. &hearer 132 (1963) 2561. [27] K. Lee et al., Nucl. Instr. and [28] W. Lorenzon, T.R. Gentile, Phys. Rev. A 47 (1993) 468. 1291 Varian Associates, Vacuum [30] k Wang, [31] S.E. Vigdor 1321 Type “I”, BN15 8UN. 1331 The Bicron OH 44065.

and G.K. Walters,

Phys. Rev.

Meth. A 333 (1993) 194. H. Gao and R.D. McKeown, Products

Division,

Lexington,

Ph.D. Thesis, Ohio University (1993). et al., Phys. Rev. C 46 (1992) 410. Micron Semiconductor, Ltd., Lancing, Sussex England. Corporation, 12345 Kinsman Road, Newbury,

451

[34] W.K. Pitts, W. Haeberli, J.S. Price, H.O. Meyer, S.F. Pate, R.E. Pollock, B. von Przewoski, T. Rinckel, J. Sowinski, F. Sperisen and P.V. Pan&la, Phys. Rev. C 45 (1992) Rl. [35] S. Frullani and J. Mougey, Adv. Nucl. Phys. 14 (1984) 1. [36] R.A. Amdt, John S. Hyslop III; and L. David Roper, Phys. Rev. D 35 (1987) 128; and program SAID (Scattering Analysis Interactive Dial-up) Virginia Polytechnic Institute, Blacksburg, VA 24061. [37] R.C. Hanna, Proc. 2nd Int. Conf. on Polarization Phenomena of Nucleons, eds. P. Huber and H. Schopper (Birkhauser, Basel, 1966) p. 280. [38] C. Amsler et al., J. Phys. G4 (1978) 1047. [39] L.D. Kramer et al., to be published.