kgAPSs05 - University of Richmond

Download Report

Transcript kgAPSs05 - University of Richmond

INTRODUCTION
The purpose of the Thomas Jefferson National Accelerator Facility (JLab) is to understand the fundamental
properties of matter in terms of quarks and gluons. We describe here how data is collected at Jefferson Lab
and how we determine the electron fiducial volume of one of the end station detectors called CLAS (CEBAF
Large Acceptance Spectrometer). We do this by focusing on data where the efficiency of the detector is well
understood.
Fiducial Cuts for the CLAS E5
Data Set
Most of the fits we made were high quality, but a few required
intervention. Here we discuss how we identified and fixed
those fits.
Stage 1: First Generation Fits
s1_p08_plot44
s6_p08_plot34
K. Greenholt (G.P. Gilfoyle)
CEBAF
The Continuous Electron Beam Accelerating Facility (CEBAF) at JLab in Newport News,
Virginia, is used to study the properties of quark matter. CEBAF is capable of producing electron
beams of 2-6 GeV. The accelerator is about 7/8 of a mile around and is 25 feet underground.
The electron beam is accelerated through the straight sections and magnets are used to make the
beam travel around the bends [See Fig. 1].
Fig. 1 JLab
Accelerator
and Halls A,
B, and C
Hall C
Department of Physics
University of Richmond, Virginia
WHAT’S THE CHALLENGE?
(2) Minimized 2: Based on a numerical fit, is the 2 small
enough? The left-hand panel shows a fit with a high . Again
in the right-hand panel we show the same data with an
improved  achieved with a better starting point for the fit.
s5_p08_plot35
(3) Reasonable Fit Parameter Uncertainties: We observed
uncertainties (2) on some fit parameters that were orders of
magnitude smaller than expected (~10-3). This would cause the
point to be weighted more, distorting the second generation fits.
The fit uncertainties were more reasonable after using the
results from the original fit in the left-hand panel as a new
starting point. The right hand panel shows the final results.
In regions of the azimuthal electron scattering angle near the current-carrying
coils that produce the CLAS magnetic field the efficiency, or acceptance, of
CLAS is not well known. To prevent the inclusion of these events in our sample,
we generate constraints [fiducial cuts] on electron scattering angles to exclude the
regions of the magnetic field near the coils.
Stage 2: Second Generation Fits
Hall B
Hall A
An electron beam can travel around the accelerator up to five times near the speed of light. The
beam is sent to one of three halls where the beam collides with a target and the debris is
measured. These data were collected in Hall B with CLAS [Fig. 2].
(1) Direct Observation: How well does the fiducial graph fit
the actual data plot? In other words, are we cutting out good
data, or including events that should be excluded? The lefthand plot shows a fit where MINUIT failed to find the correct
low edge. The right-hand panel shows the same data with an
improved fit. This was done by choosing a better starting
position.
We seek to define a function which effectively limits the data analyzed to focus
on regions of CLAS in which the acceptance of the detector is well understood.
The left-hand panel shows the  versus fit for the original first and second
generation fits. The right-hand panel shows the data with the improved fit
obtained by excluding low-statistics points. This fit defines the regions of
CLAS in which we can well understand the acceptance.
Goal: To generate electron fiducial cuts on data from the CLAS
detector, so as to focus on regions of CLAS in which the acceptance
of the detector is well understood.
What is Acceptance?
Acceptance (often referred to as efficiency) is the ratio of the events measured in the detector versus the
actual events produced in the nuclear reaction. In plain English, “how much of the good stuff do we
actually catch in the detector?”
Stable Acceptance:
focus on the flat, smooth
regions
The red line is what we might
expect to measure from an ideal
or perfect detector; the black
line is what we actually
measure.
CLAS
CLAS is located in Hall B and is used to detect pions, electrons, protons and other subatomic particles. The
detector is able to detect most particles created in a nuclear reaction, because of its unique nearly-full-solidangle structure. There are six different layers of CLAS [see Fig. 2] which produce electrical signals,
providing us with information on velocity, momentum, and energy, and allow us to identify different
subatomic particles.
Fig. 2 CLAS
Event
Display(CED),
displays signals
received from each
layer of CLAS.
The drift chambers make up the first three layers, and determine the paths of different particles. The next
layer is the Cerenkov counters which separate electrons from pions. The following layer is made of the
time of flight scintillators to determine time of flight and hence velocity. The calorimeters, used to measure
the energy of the particles, make up the final layer. Also in CLAS is a toroidal magnet that causes charged
particles to bend as they pass through the drift chambers. This bending is used to determine momentum.
This magnetic interaction is of particular interest to us, as we attempt to define the fiducial volume of the
detector, because it affects the regions of stable efficiency.
Where do we Start?
Fig 3. Data plot from CLAS
showing  versus  for the
electron. Note: six sector
orientation.
Fig. 4. Fiducial cut in terms of
events plotted against  angle,
showing the region of stable
efficiency in the distribution
for the electrons in the labeled
momentum and bin.
Stage 3: Third Generation Fits
Upper Edges
PROCEDURES:
Stage 1: First Generation Fit We plot the number of events versus the  angle for a particular momentum
bin and  angle bin. We then use a CERN program called Minuit to fit a trapezoidal curve to the data points.
The fiducial cut is defined as the edge of the plateau in Fig. 4.
Stage 2: Second Generation Fit We fit the upper and lower sector edges defined by the first generation fits,
and plots them against the electron angle. We then use Minuit to fit another curve to these data points.
While often this fit is symmetric, the procedure does not require symmetry.
Stage 3: Third Generation Fit We plot the results generated by the first and second generation fits against
the momentum of the electron (ascertained when the particle passes through the toroidal magnet), and fit these
data with a polynomial function.
References:
'Fiducial Cuts for electrons in the CLAS/E2 data at 4.4 GeV', D. Protopopescu, F. W. Hersman, M. Holtrop,
UNH, S. Stepanyan, CNU, and CLAS/E2 run group, CLAS-Note 2000-007, November 27, 2000.
Upper Edges
We plot the different fit
parameters against the
momentum of the electron
and fit the curves with
polynomials. We note that
there should be some
symmetry between the upper
and lower edges. Fig. 5
(sector 1) and Fig. 6 (sector 3)
show this behavior.
Lower Edges
Fig. 5.
Fig. 6.
Lower Edges
Stage 4: Conclusions
1)
We have fitted more than 10,560 distributions of  for the E5 data sets at 2.56GeV normal torus polarity and reversed torus
polarity (first generation fits).
2)
Some distributions required new starting positions to find acceptable fits.
3)
We fitted the edges measured in the first generation fits successfully and excluded low statistics regions.
4)
We observed smooth dependence on electron momentum for all fit parameters.