Biasing Wire Scanners and Halo Scrapers for
Measuring 6.7-MeV Proton-Beam Halo+
J. Douglas Gilpatrick *, Michael Gruchalla t, James Kamperschroer ^[,
James O'Hara*
* Los Alamos National Laboratory, MS H808, LANL, Los Alamos, NM, 87545
f Honeywell FM&TINM, PO Box 5250, Albuquerque, NM, 87185
f General Atomics, Los Alamos, NM, 87544
Abstract. Wire scanners and halo scrapers (WS/HS) were used to acquire projected beam
distributions over a very wide dynamic range in order to determine the extent and study the
formation of beam halo at the Low Energy Demonstration Accelerator (LEDA). To detect
beam distributions over a large dynamic range, it was necessary to understand the effects of
WS/HS biasing for optimizing wire and scraper signal amplitudes. Both wire scanners and halo
scrapers were biased with both positive and negative potentials to +/- 200 V. WS/HS depletedcharge data were acquired at these different potentials and the amount of signed charge leaving
or accumulating on the wire or scraper was measured. This paper will show these data, and will
offer a discussion of an optimal biasing potential for these types of projected beam profile
measurement devices.
HALO INSTRUMENTATION
At the LEDA, a 100-mA, 6.7-MeV beam is injected into a 52-quadrupole-magnet
lattice (see Fig. 1). Within this 11-m FODO lattice, there are nine wire scanner/halo
scraper (WS/HS) stations, five pairs of steering magnets and beam position monitors,
five loss monitors, three pulsed-beam current monitors, and two image-current
monitors for monitoring beam energy [1]. The WS/HS instrument's purpose is to
measure the beam's transverse projected distribution [2].
These measured
distributions must have sufficient detail to understand beam halo resulting from
upstream lattice mismatches [3,4]. The first WS/HS station, located after the fourth
quadrupole magnet, verifies the beam's transverse characteristics after the RFQ exit.
A cluster of four WS/HS located after magnets #20, #22, #24, and #26 provides phase
space information after the beam has debunched. After magnets #45, #47, #49, and
#51 reside the final four WS/HS stations. These four WS/HS acquire projected beam
distributions under both matched and mismatched conditions. These conditions are
generated by adjusting the first four quadrupole-magnetic fields so that the RFQ
output beam is matched or mismatched in a known fashion to the rest of the lattice.
Because the halo takes many lattice periods to fully develop, this final cluster of
WS/HS are positioned to be most sensitive to halo generation.
Work supported by the US Department of Energy.
CP648, Beam Instrumentation Workshop 2002: Tenth Workshop, edited by G. A. Smith and T. Russo
© 2002 American Institute of Physics 0-7354-0103-9/02/$19.00
297
FIGURE 1. The 11-m, 52-magnet FODO lattice includes nine WS/HS stations that measure the beam's
transverse projected distributions.
As the RFQ output beam is mismatched to the lattice, the WS/HS actually observe
a variety of distortions to a properly matched Gaussian-like distribution [3,4]. These
distortions appear as distribution tails or backgrounds. It is the size, shape, and extent
of these tails that represent specific types of halo. However, not every lattice WS/HS
observes the halo generated in phase space because the resultant distribution tails may
be hidden from the projection's view. Therefore, multiple WS/HS are used to observe
the various distribution tails.
WS/HS DESCRIPTION
Each station consists of a horizontal and vertical actuator assembly (see Fig. 2) that
can move a 33-|im-carbon monofilament and two graphite/copper scraper subassemblies [5]. The carbon wire and scrapers are connected to the same movable
frame. Attached to this movable frame is a linear encoder that provides the wire and
scraper edges' relative position to within a typical rms error of 5 jim, and an additional
linear potentiometer provides an absolute approximate position for LEDA's runpermit systems. A stepper motor coupled to a ball lead screw is used to drive the
moveable frame. A motor brake and microswitches limit the frame's movement.
298
FIGURE 2. The WS/HS assembly contains a movable frame on which a 0.03-mm carbon wire resides
between two water-cooled graphite scrapers.
The carbon wire, which senses the beam's core, is cooled by thermal radiation. If
the beam macropulse is too long, the wire temperature continues above 1800 K
resulting in the onset of thermionic emission [6]. Thermionic emission causes an
inaccurate appearance to the distribution by exaggerating the core's current density.
To eliminate these effects for the halo experiment, the maximum pulse length and
repetition rate is limited to approximately 30 jis and 1 Hz, respectively.
The halo scrapers are composed of a 1.5-mm thick graphite plate brazed to a watercooled 1.5-mm thick copper plate. Since 6.7-MeV protons average range in carbon is
approximate 0.3 mm, the beam is completely stopped within the graphite plate.
Cooling via conduction lowers the average temperature of the scraper sub-assembly
and allows the scraper to be cooled more rapidly than the wire. The lower average
temperature and faster cooling allows the scraper to be driven in as far as 2 rms widths
from the beam distribution peak without the peak temperature increasing above 1800
K.
The movement and positioning of each wire and scraper pair is controlled by a
motion control system that contains a stepper motor, stepper motor controller, a linear
encoder, and an electronic driver amplifier [7]. The controller's digital PID loop
controls the speed and accuracy at which the assembly is moved and placed.
The target position, as defined by the WS/HS operator, is relayed from the EPICS
control screen via a database process variable to a National Instruments Lab VIEW
Virtual Instrument (VI). The VI also calibrates the relative position of the linear
encoders based on the measured position of the limit switches, and provides some
error feedback information [7]. The total error between the target wire position and
the actual wire position attained is within a total 4% range of a typical 1-mm rmswidth beam.
As the wire is moved through the beam, it senses the projected beam core
distribution. A small portion of the beam's energy is imparted to the wire causing
secondary electron emission to occur. The secondary electrons leaving the wire are
replaced by negative charge flowing from the electronics. This current flow for both
axes is connected through a negative bias battery to an electronic lossy integrator
circuit and followed by an amplification stage.
The integrator capacitance and amplifier gain are set to allow a very wide range of
values of accumulated charge [8]. Data are acquired by digitizing the accumulated
299
charge through the lossy integrator at two different times within the beam pulse. This
charge difference, acquired by subtracting the two values of charge, provides a low
noise method of relative beam charge acquisition. The wire and scraper accumulated
charge signals are digitized using 12- and 14-bit digitizers, respectively. The analog
noise floor has been measured to be 0.03 pC, a noise level slightly lower than the
scraper digital LSB noise level of 0.15 pC using the highest gain settings within the
detection electronics.
The front-end electronic circuitry, mounted on a daughter printed circuit board, is
connected to a motherboard that has all of the necessary interface electronics to
communicate with EPICS via a controller module within the same electronics crate. A
software state machine sequence was written within EPICS to control and operate
WS/HS instrumentation [9]. The state machine instructs the VI to move the wire and
scraper to a specific location, acquire synchronous distribution data from either the
wire or scraper, trigger the IDL routine to normalize the acquired charge with a nearby
toroidal current measurement, graph the normalized data, and write the distribution to
a file. The sequence also instructs IDL to calculate the first through fourth moments,
fit a Gaussian distribution to the wire scanner data, and calculate the point at which the
beam distribution disappears into the background noise.
To plot the complete beam distribution for each axis, the wire scanner and two
scraper data sets must be joined [10]. To accomplish this joining, several analysis
tasks are performed on the wire and scraper data including,
(1) scraper data are spatially differentiated and averaged,
(2) wire and scraper data are acquired with sufficient spatial overlap, and
(3) differentiated scraper data are normalized to the wire beam core data.
The scraper data need only be normalized in the relative charge axis since the
distances between each wire and scraper edge are known to within 0.25-mm. In
addition, the first four moments and the point at which the beam distribution
disappears into the noise are also calculated for the combined distribution data.
WIRE AND SCRAPER PHYSICS
For several different wire scanners and halo scrapers, an emission current curve
was generated as a function of device bias potential. This was done to verify the
acquisition goals of detecting only secondary electrons for the WS wire and 6.7-MeV
protons for the halo scraper. The procedure included acquiring an integrated charge
waveform at each bias potential. The emission currents for each point on the
following graphs were acquired by fitting a straight line to the integrated charge data
and calculated the fitted slope or current. Emitted current data were acquired during
the last 10 jis and at least 10 jis after the 30-|is-length beam pulse.
Wire Scanner Bias
Several interesting details show up in these wire scanner bias curves. As the wire
bias is positively increased from ~+6 V to > +200 V, the wire secondary electron
emission is nearly completely inhibited and the net current flowing on to the wire
300
reduces to very near zero (as shown in both wire scanners #22 and #51 of Fig. 3). At
approximately +150 V, the secondary emission is completely inhibited. As the bias
potential is reduced from ~-6 V to < -200 V, the net current varies from being
relatively stable initially to a slight reduced trend. For the purposes of the halo
experiment, the stable or 0-slope area in the -6 V to -12 V region is what was chosen
as the optimal operating point and is the region in which the wire emission current
does not change with bias. Furthermore, it appears that the wire collects positive ions
with < -25 V bias potentials after the beam pulse, as shown by the red crosses in Fig.
3. This ion collection additionally limits the amount of negative bias that is applied to
the wire for proper secondary emission operation. As the negatively biased wire's
potential is further reduced, the small positive slope of the "during pulse" wire data
appears to reinforce the premise that slow ions are being collected.
One rather interesting area of the graph is that of the emitted wire current near the 0
V bias potential. It is peaked at 0 V approximately 15% higher than at either +6 V or
-10 V bias potential. One proposed explanation of this 0-bias elevated wire current is
the interception of electrons and ions from protons ionizing residual background gas both of these ionized species creating further secondary emission. As the wire is
biased negatively, low energy intercepted electrons are rejected causing a reduction in
secondary emission, and as the wire is biased positively, the intercepted ions are
rejected causing a reduction in secondary emission. If the intercepted electron-ion
pairs are the mechanism for the elevated net current, the wire should be biased to not
include this additional current component since this is not directly due to 6.7-MeV
beam impingement. Since we did not have clear proof as to what the real cause of the
elevated current is, it was decided to bias the wire away from this effect, i.e., a -12 V
bias.
160
WS #51
OJ
* Oping Futse
n Ate Pulse
OJ
0.4
' 03
60S!
01
40-1
20 !
0j
m I
40
0.0
m
m
-04
90
-2m
-1(30
O
100
1)Q
300
Bias Potential (Y)
FIGURE 3. The two above graphs of wire scanner #22X and WS #51 show the wire scanner emission
current as a function of wire potential. The goal of proper biasing is to optimize the emission current
such that the detection current is only due to secondary electron emission. This goal was addressed by
biasing the wire with approximately -12V where the curves slope is approximately zero.
Halo Scraper Bias
The scraper bias data, Fig. 4, show similar results as the wire data. Increasing the
scraper bias from ~+ 6 V to > +100 V, reduced the amount of electrons leaving the
scraper. At +20 V to +40 V, the total amount of current detected levels out to a
301
minimum and is composed only of deposited protons and is constant with respect to
increasing bias. For the purposes of the halo experiment, a scraper bias of +25 V was
chosen. Also, note that with this +25 V bias, the data show that no after pulse current
flowing. One interpretation of the after pulse current was that it is composed primarily
of slow low-energy positive ions. This interpretation seems to hold based both on the
temporal waveforms observed during data acquisition and the fact that after pulse
current is a positive non-zero value for negative bias potentials.
With 0 V applied to the scraper, the scraper net current is also elevated. As with
the wire, it is not clear what is causing this increase in emission current so it was
decided to bias the scraper outside this region, i.e. a +25 V bias.
160 i
HS£24
*
$ During Pulse
0
n After Pulse
00
80
O
^o
:Q
40
o
o
0°
%
§ 0,6
o
*
;
Q- 6 J ,
25
50
75
100
125
150
:
s
^ 0.2
11P5
o During Pulse
^
3is After Pulse
&
°° \,
g 04
° °1
<4Q
)
'
" ggrff^lilii*.
0.0
-3 0
0
Bias Potential (V)
*
30
0
*
o
60
*>
90
<
12
Bias Potential (V)
FIGURE 4. The two above graphs of halo scraper #20 and WS #24 show the scraper emission current
as a function of scraper potential. The goal of proper biasing is to optimize the emission current such
that the detection current is only due to deposited protons. This goal was addressed by biasing the
scraper with approximately +25 V where the curve slope is approximately zero.
Secondary Emission Yield
As was shown in the previous section, the WS wire is biased negatively to optimize
secondary emission (SE) yield, where these SE yield is defined as the ratio of the
emitted secondary electron current and the proton beam current intercepted by the
wire. TABLE 1 shows the measured values of a representative sample of the lattice
and HEBT WS wires. All of the wires in the halo lattice WS are configured with a 33jim, carbon monofilament. The HEBT WS is configured with a 100-jim SiC wire. All
of the SE currents were acquired with the nominal -12 V bias so that SE was
optimized and with the wire placed in the core of the proton beam. In both cases, the
6.7-MeV protons did not stop in the wire but deposited sufficient beam energy into the
wire to cause SE.
One particular model for emission of secondary electrons resulting from energetic
ions impinging on various materials is described by E. J. Sternglass. In his 1957
paper, Sternglass defined the secondary emission yield, F, as
e dx
(1)
where P is the probability that an electron will escape (approximately -0.5), d is the
average depth from which electrons escape the material (in this case a small
302
monolayer of - 1 nanometer), £is the average amount of kinetic energy lost by an ion
or proton per ionization in material (~ 25 eV), and dEldx is the stopping power of the
proton beam for the wire material (163 MeV/cm for SiC and 162.9 MeV/cm for C)
[11, 12]. The resulting calculated yield based on the Sternglass model for both C and
SiC is approximately 33%.
However, the above model does not include the geometric factor of a round wire.
In this case, a round wire will have areas on both sides of the wire that have a longer
distance to deposit the proton beam's energy in the 1-nm annular "electron escape"
region of the wire. This "form factor," as Sternglass describes it in his paper, is a
value proportional to a sec(9) function where 0 is the angle between two rays, an
incoming proton trajectory ray and a ray between the wire center and the impact point
of the proton on the wire perimeter. The "form factor" has been calculated to be -3.8
and -4.2, for the 33-|im and 0.1-mm round wires, respectively.
Additionally, Sternglass also mentions a SE yield temperature dependency that can
lower the emission efficiency by as much as 50%. If both the temperature dependency
effect and the "form factor" are treated as multipliers to the initial theoretical yield, a
final yield is calculated to be -63% and -70% for the C and SiC wire, respectively.
The acquired experimental data, as shown in Table 1, are within a few 10s of percent
and approximately agrees with the calculated yields as suggested by Sternglass's
model.
TABLE 1. Secondary Emission Yield: SiC and C Wires with 6.7-MeV Proton Impingement
Wire Scanner
X/YBeam
X/Y
X/YRms
X/YS.E. Current
Number
Width (mm)
(mA)
Current (mA)
Yields (%)
22
75/76
0.86/0.67
0.63/0.6
55/41
24
76/76
0.78/0.88
0.55/0.61
42/54
47
76/76
0.7/0.75
0.6/0.6
42/47
51
11111
0.8/0.77
0.65/0.61
51/46
HEBT
75/75
5/8.2
0.30/0.24
51/66
SUMMARY
This paper has described the general operation of the WS/HS combination profile
measurement used in the LEDA. In order to operate the WS and HS properly, a series
of biasing measurements was performed. The wire scanner and halo scraper overall
acted as expected at various biasing potentials and the resulting V-I curves show that
the wire and scraper are optimally biased at -12 V and +25 V, respectively.
However, an additional effect near 0 V potential of an elevated emission was
unexpected and not well understood. A proposed explanation for this elevated
emission current was suggested but certainly not proven. The amount of secondary
emission yield was measured to be approximately the expected amount based on the
Sternglass model of secondary emission of electrons as a energetic ion impinges on a
specific material (in this case, 6.7-MeV protons on C or SiC).
REFERENCES
1. Gilpatrick, J. D., et al., "Experience with the Low Energy Demonstration Accelerator (LEDA) Halo
Experiment Beam Instrumentation," Proceedings of the 2001 Particle Accelerator Conference, June
18-22, 2001, pp.2311-2313.
2. Gilpatrick, J. D., et al., " Beam-Profile Instrumentation for Beam-Halo Measurement: Overall
Description and Operation," Proceedings of the 2001 Particle Accelerator Conference, June 18-22,
2001, pp. 525-527.
3. Wangler, T. P., et al., "Experimental Study of Proton-Beam Halo Induced by Beam Mismatch in
LEDA," Proceedings of the 2001 Particle Accelerator Conference, June 18-22, 2001, pp. 29232925.
4. Colestock, P. L., et al., "Measurements of Halo Generation for a Proton Beam in a FODO Channel,"
Proceedings of the 2001 Particle Accelerator Conference, June 18-22, 2001, pp. 170-172.
5. Valdiviez, R., et al., "The Final Mechanical Design, Fabrication, and Commissioning of a Wire
Scanner and Scraper Assembly for Halo-Formation Measurements in a Proton Beam," Proceedings
of the 2001 Particle Accelerator Conference, June 18-22, 2001, pp. 1324-1326.
6. Valdiviez, R., et al., "The High-Heat Flux Testing of an Interceptive Device for an Intense Proton
Beam", Proceedings of the 2001 Particle Accelerator Conference, June 18-22, 2001, pp. 1321-1323.
7. Barr, D., et al., "Design and Experience with the WS/HS Assembly Movement Using LabVIEW
Vis, National Instrument Motion Controllers, and Compumotor Electronic Drive Units and Motors,"
Proceedings of the 2001 Particle Accelerator Conference, June 18-22, 2001, pp. 794-796.
8. Gruchalla, M., et al., "Beam Profile Wire-Scanner/Halo-Scraper Sensor Analog Interface
Electronics," Proceedings of the 2001 Particle Accelerator Conference, June 18-22, 2001, pp. 23142316.
9. Day, L., et al., "Automated Control and Real-Time Data Processing of Wire Scanner/Halo Scraper
Measurements," Proceedings of the 2001 Particle Accelerator Conference, June 18-22, 2001, pp.
1309-1311.
10. Kamperschroer, J., et al., "Analysis of Data from the LEDA Wire Scanner/Halo Scraper,"
Proceedings of the 2001 Particle Accelerator Conference, June 18-22, 2001, pp. 1306-1308.
11. Sternglass, E. J., "Theory of Secondary Electron Emission by High-Speed Ions," The Physical
Review, Second Series, Vol. 108, No. 1, October 1, 1957, pp. 1-12.
12. Borovsky, J. E. and Suszcynsky, D. M., "Experimental Investigation of the z2 Scaling Law of FastIon-Produced Secondary Electron Emission," The Physical Review A, Vol. 43, No. 3, February 1,
1991, pp. 1416-1432.
304