Measurement of the Gravitational Acceleration of the Antiproton: An Experimental Overview

An ambitious experiment Is being developed to measure the force on the antiproton due to the gravitational field of the earth. The technique consists in obtaining antlprotons of the lowest energy possible from the LEAR facility at CERN, decelerating them further In an external beam line, trapping and cooling them to ultralow energy, and measuring their gravitational acceleration by time-of-fllght methods. The experiment has been granted and initial development ef!orts are described. tested with H- ions before shipment to CERN for the antiproton experiment.


Introduction
An adventurous and difficult experiment to measure the gravitational Interaction of antimatter (an antiproton) with matter (the earth) is being planned, The effect of earth's gravity on other fundamental particles has been measured (3), but no measurement of the gravitational Interaction of antimatter with either matter or antimatter has ever been made.
Recent theory (1 ,2,3) suggest that the antimatter-matter gravitational interaction could be slgnific My different than the matter-matter Interaction with which we are so familiar. One typical estimate Is that the antimatter-matter interaction could be a factor of 3 stronger (4). In keeping with the nature of this conference I will not detail the theoretical background, but I shall first outllne the basic oxperlment as planneu by our scientific collaboration developing the project (5), then discuss some of the more challen~ing aspects of the work, and finally comment on our current progress.
2. The Basic Exporlment ,4 simple sci~ematlc of the experiment Is shown In figure 1. Antlprotons In a thermal (statistical) distribution In an electromagnetic trap are released to traverse upwards through a 1-m shielded drift tube to a detector, most likely a multichannel plate detector (MCP). The time of flight (TOF) of the particles from release to detection Is recorded. At some slow velocity, because of the gravitational attraction, the particles will fall back and not reach the detector. This cutoff time, TC=(2L/g)lU, about 0.45 seconds,~allows the detorml ation of g If the drift length L Is known.

The Real Exporlmeu[
Transforming the basic experiment Into reality lfitrOdLWs numerous and challenging complications.
a. For a 1-m drift tube, tho release energy of the particle with the cutoff TOF is aboM 10-7 eV (5 m/s); gravity Is weakl Supplying sufficient numbers of antlprotons to map out the TOF spectrum In the region of the cutoff Is a problem, If the antiproton cluster In the launch trap has been cooled to Ilquld helium temperature, 4 K, the mean energy of the particles Is about 10-9OV (1 rneV), and few particles have a low enough energy to be useful. Monte Carlo studies Indicate that roughly 107 particles wIII have to be launched from a trap at 4 K to measure g to 1'/oi In order to ob~aln sufficient numbers of antlprotons, we plan to perform the experiment at the Low Energy Antlproton Ring (LEAR) faclllty at CERN In Geneva, Switzerland, where our proposal (6) has been accepted as experiment PS-200, Details of this experiment ha')e been discussed previously (7)(8)(9)(10)(11).
b. In order not to compe!e with gravltatlon, electromagnetic forces must be SeptSmber 25, 1986 extremely small in the drift tube region: less than 10-7 V/m and 1 part in 10-5variation in the magnetic guide field needed for such low energy oarticles.
c. The drift length L will not be known well because of elec!ric field penetration into the drift tube. An H-ion has the same electric charge and very nearly the same ineflial mass and magnetic moment as an antiproton. An essential part of the experiment will be to repeat the measurements using H-ions for a calibration and comparison. Thus H-ion sources will be needed: at 20 keV for tests and development, at 2 MeV for "checkout of the deceleration and final experiment setup at Los Alamos, and at 2 MeV from LEAR for beam alignment and tuning. Dand 0-beams will provide different mass ions to test and help understand the time-of-flight experiment. d. Extremely low vacuums will be needed, on the order of 10-14 Torr, to avoid antiproton annihilation and H-neutralization by the residual gas atoms.
e. The antiprotons created at the CERN facility at about 3.6 GeV/c (2.7 GeV) must be decelerated, (trapped) and cooled by 16orders of magnitude in energy to obtain 10-7eV pafiicles, and this must be done without undue losses.

Fllllng the Antiproton Reservoir
A more complete schematic of the experiment is shown in figure 2, The antiprotons produced at CERN are collected, decelerated to 600 MeV/c and injected into LEAR for further deceleration and cooling, We hope to receive from LEAR a 200 ns blvst of about 10aor more 2 MeV antiprotons. We will then decelerate this bunch to 20 i"~"d using a radiofrequency quadruple (RFQ). A design study of the use of the RFQ as a d?ce/ef8tor has been given by J.H. Billen et al,(l 2). Transverse currents in the RFCl pole structure provide alternating electric polarities resulting in a transverse electric quadrupoie field. This field produces focusing in one transverse plane and defocusing in the other. The fields reverse direction one half an RF period later to produce a net strong focusing effect. The pole tips are machined with an oscillatory variation in radius to produce longitudinal accderating (or decelerating) fields as well as transverse focusing fields. The strong focusing in an RFQ permits the preservation of good beam quality during deceleration. The RFQ freouenhy offers the advantage of reduced size over an electostatlc system. A current design (13) gives an RFQ le~lgth of less than 2 meters. An RF buncher 8 m upstream is necessaty to create the initial 200 MHz structure that the RFCt requires.
Beam tuning and steering to insure that the single LEAR antiproton pulse will properly strike the first trep is critical, and development of such an ability for the 2 MeV and 20 keV beam lines Is underway.
The necessary timing ior the LEAR beam bunch, chopped H-beam, trap electrode voltage changes will be ha~ldlod by~small computer; which wIII also analyze the experimental time-of-flight data as Itis taken.
The 20 keV pafilcles are then further decelerated to 5 keV by electrostatic forces, and then caught in an electomagrwtlc Penning trap (9,14,15).
Here an Important transition takes place from the panicles having a directed enargy of a beam to a thermal distribution of a cluster held In the center of a trap. Particle confinement In a Penning tlap is achieved by a combination of static electric and magnetic fieldc, with the electic field being that of an axially symmetric quadruple, the magnetic field being uniform and In Pie axial direction (the symmetry axis being the beam direction, ) Cooling the thermal particle bunch Is possible by several means. Stochastic cooling Is being studied by collaborators at the University of Genoa and Piss In Italy, electron cooling at Rice University, and r~sistive coollng at Texas A&M Unlverslty and Los Alamos.
Information on trapping and cooling 3 September 25, 1!386 processes is reported by Kenefick(16) and Church (17) at this conference.
The catching trap must be long in order to contain a large fraction of the antiproton burst. This length interferes with the harmonicity of the trap and thus its ability to cool the thermal distribution. An example of a design of a catching trap that also has some cooling ability is shown in figure 3. Here the shaping of the electric field with additional electrodes will allow efficient catching of the bunch and still retain enough resonant harmonic,ity to allow resistive cooling to be used to bring the temperature down to about 10-100 eV on the order of I hour or less. A 6-T magnet provides the axial field necessary for trap operation and for guiding the entering and exiting particles. "Closing the door to catch the incoming particles is not a routine matter. 10 to 20 kilovolts in 10's of ns must be applied with careful timing to the front electrode of the catching trap, A programmed high-voltag~level shifter is under development which will provide the various potential changes in the trap electrode.
-The now-small particle bunch is transfered to one or more smaller highl -compensated harmonic traps for additional and fast cooling (to a few Kelvin, 1 -10-eV) and launching. Even after such cooling, the particles of interest are far down on the slow part of the velocity distribution at 10-7 eV.
Successful development of a source of a large number of cold thermal antiprotons may open the door to a number of interesting experiments using antiptrotons or antihydrogen in addition to the gravity experiment.

Vacuums and Cryostat Technology
The annihilation cross section rises rapidly as the velocity of the antiproton falls (18). Vacuums of 10"10to 10'12 Torr should suffice in the first catching trap. Such vacuums require painstaking but known techniques. In the launch-drift region where the energies are very low, the vacuums must be 10-14 Torr or better for adequate antiproton survival, This will necessitate complete enclosure at a temperature of 4 K or colder (19). The design of a system to provide all the voltage and signal loads, es well GSa beam "trap door' (see Fig. 1), all to fit In the bore of a superconducting magnet, will be a stimulating challenge.

The Gravity Experhnent Itaolf
Launching and dehscting such slow particles In an understandable way will be difficult. The pif)neerlng experiment by Witteborn and Fairbank (20) with electrons has and wIII be a great help in investigating some of the possible systematic effects and errors, Table I lists some of the effects that must be considered. I shall discuss several of these effects. See Ref.
(3) for more details. The patch effect occurs because the surface of even the best of conductors Is 4 September 25, 1986 not at a uniform potential because it consists of many crystal faces (patches) that can have differing work functions (21). The rms axial electric field produced by a random patch distribution on the surface of a typical conductor is es~imated to be considerably larger than the gravitational equivalent (22). However, we expect to coat (23) the inner surface with an amorphous but still conducting material to reduce this field significantly. Furthermore, it has been discovered (22) that the patch field is strongly reduced at liquid-helium temperatures, although the reason for this suppression is not understood.
A temperature gradient along the drift tube will cause a potential variation (Thompson emf) of a few pV/K. In our experiment we will need a temperature difference less than 10-2 K. Immersing the entire tube in liquid helium cannot accomplish this, because the change in boiling point with pressure causes a temperature gradient of 0.3 Wm. However, it is possible (20) to reduce the gradient to <10-5 K/m by having the tube in contact with the helium bath at only one location (see Fig. 1 "Thermal Suppoti Link").
The gravitational force on an antiproton at the earth's surface is equal to the electrostatic force between two antiprotons 12 cm apart. Thus the interaction of neighbor particles In the launch might seriously deplete the number of very low-energy particles. However, In a conducting tube the antlprotons are partially shielded from each other. One finds that the effect Is small compared to gravity if the antiprotons are separated by at leas'. two to three times the tube's radius. The effect of such CoulomlJ forces on the velocity distribution of the launched particles is currently being studied by computer simulation. First results indicate that the number of particles per launch should be In the range of 10-100 to avoid a serious problem. This que~tlon wIII be studied experimentally In tests with H-ions,

Present Status
We have constructed at Los Alamos a test beam line to begin developing the apparatus for the antiproton gravity experiment. A 20 keV H-beam, chopped to resemble an antlproton bunch is sent through a horizontal section consisting of valves, cold traps, four-way slits, vacuum pumps, magnetic steerers, and electrostatic lenses, The beam is then turned Into the vertical direction by a 90°m agnet mounted In a large vertical support stand, Above the 90°magnet, the beam Is steered and focusc d into a Penning trap situated in the 6-T field of a superconducting soienold magnet.
Our ongoing tests with this apparatus wIII allow us to study vacuum Isolation and bakeout procedures, ultimate vacuum capability with a room-temperature Ion trap, trap pulslng to capture protons or H-ions, the features needed In subsequent trap designs, and the type of H-source to be taken to LEAR. To date, we have succeeded In passing a 10 VA beam of 20 keV H-Ions through the trap's 3 mm diameter apertures In a 6-T magnetic field. We have also demonstrated simple trapping of NNN Ions for Tll seconds by a fast raise of the voltages on the appropriate traps caps, followed by release and detection with an MCP. This Is an impofiant step: capture of In-flight Ions has only recerltly been demonstrated (24), In the future, an Improved system with a new 2G keV H" source, buncher and RFQ decelerator wIII be constructed and Installed at Los Alamos to receive a 2 MeV H-beam from the Los A!amos Vertical Van-de-Graaff.
Research w~~hthis system will lead to a final choice of traps and drltl tube assembly, which in turn will be 5 September 25, 1986 proof tested with H-ions before shipment to CERN for the antiproton experiment.

Ack~~owledgements
Besides expressing appreciation to my colleagues (5) for discussions and help, I would like to thank D. C. Lizon, R. Martinez, R. R. Showalter, and C. B, Webb for their expert help in constructing and operating the present test system at the Los Alamos Ion Beam Facility.
This work was supporled by the U.S. Department of Energy under Contract No.

Figure Captions.
1. Schematic of an apparatus design for the time-of-flight antiproton gravity experiment. Ten eV antiprotons coming from below are trapped and cooled in the final Penning "iaunch" trap to a temperature of roughly 10 K. The antiprotons would be extracted, a few at a time, to CM up the shield tube; and, if not pulled back by gravity, accelerated to strike the detector. A rough scale is given by the 1-m tall drift lube.
2. A possible schematic diagram for the Antiproton Gravity Experiment. The layout is a plan view except for the section after the electrostatic mirror which is a side view. The diagram is not to scale. The region inside ?!N dotted lines represents a "thermal source" of low to very-low energy antipmtcms that would be available for a variety of experiments.
3. Schematic cross section of a design for a first stage catching trap. The device is azimuthally symmetric about the horizontal axis. The entrance electrode is elongated to accommodate the beam burst Isngth after some electrostatic deceleration. The Penning trap proper centers at the torus of circular cross section, and has a extended array of electrodes to provide a long trap that still has sufficient harmonicity for cooling. The entire trap length in this design is 50 cm. 3. Schematic cross section of a design for a first stage catching trap. The device is azimuthally symmetric about the horizontal axis. The entrance electrode is elongated to accommodate the beam burst length after some electrostatic deceleration. The Penning trap proper centers at the torus of circular cross section, and has a extended array of electrodes tc provide a long trap that still has sufficient harrnonicity for cooling. The entire trap length in this design is 50 cm.