1:
2: %\section{Data Sample and Analysis}
3: The data sample consists of $2.9\times 10^8$ and $3.7\times 10^8$
4: pulses at beam energies of $45.0$ and $48.3$ GeV respectively,
5: collected over three data runs in 2002 and 2003.
6: Roughly 60\% of the data were accumulated in the 2003 run,
7: which featured a novel ``superlattice'' photocathode \cite{superlattice}
8: with $\approx 90\%$ beam polarization.
9:
10: Data were collected at 120 Hz, with $\sim 1$ Hz of pulses blanked to
11: measure baseline signals. Alternate triggers fall into two 60~Hz
12: fixed-phase ``time slots''. Within these time slots, right-left pulse
13: pairs are formed for independent asymmetry analyses. The helicity
14: sequence in each pulse pair was chosen pseudo-randomly. In addition to
15: the fast helicity flips, the sign of the electron polarization was
16: passively reversed in two independent ways. First, the state of a
17: half-wave plate in the laser line was toggled each day, guarding
18: against helicity-correlated electronics crosstalk. Second, spin
19: precession in the $24.5^\circ$ bend after acceleration created
20: opposite helicity orientations at $45$ and $48$ GeV beam energies.
21: For each of the 2002 runs the beam energy was changed once, while the
22: change was made
23: roughly every four days during the 2003 run.
24: Roughly equal statistics were accumulated with
25: opposite signs of the measured asymmetry,
26: suppressing many classes of systematic effects.
27:
28: %\subsection{Raw Asymmetry}
29: %analysis chain description
30:
31: We select pulses with beam charge greater than $10^{11}$
32: electrons and require that the beam crosses each beam position monitor (BPM)
33: within 1~mm of its geometric center.
34: We also require that the beam position and
35: charge for each pulse be within 6 standard deviations from the running
36: mean value. Typically, the beam charge per pulse varies
37: between $(4-6)\times 10^{11}$ electrons with $0.3\%$ pulse-to-pulse
38: jitter, and the beam position is within $100~\mu$m of each
39: BPM center, with jitter on the order of $50~\mu$m.
40: In order to avoid
41: helicity-dependent biases, we reject several pulses before and after a
42: pulse which fails a cut. Other than the demand that
43: the beam charge asymmetry measured by two independent monitors agree to within
44: $10^{-3}$, no helicity-dependent cuts are made.
45:
46: The right-left asymmetry in the integrated detector response for each pulse
47: pair is computed by normalizing to incident charge and
48: then correcting for beam fluctuations.
49: To first
50: order, six correlated parameters describe the
51: beam trajectory: charge, energy, and horizontal and vertical
52: position and angle. Each parameter is measured by two
53: independent monitors, such that device resolution and systematic
54: effects can be studied.
55:
56: Two methods are used to calibrate
57: the detector sensitivity to each beam parameter and remove
58: beam-induced random and systematic effects from the raw asymmetry.
59: One method uses a calibration subset (4\%) of the pulses,
60: where each beam parameter is modulated periodically
61: around its average
62: value by an amount large compared to nominal beam fluctuations.
63: The other method applies an unbinned least squares linear regression
64: to the pulses used for physics analysis.
65: They yield statistically consistent results to within $3$~ppb.
66: Final results are obtained with the latter, statistically more powerful
67: technique.
68:
69: %
70: % second order effects
71: %
72: Additional bias to the measured asymmetry may arise from
73: asymmetries in higher order moments of beam distributions, such as
74: temporal variations of the beam position or
75: energy within a 270~ns beam pulse, coupled to the intra-pulse
76: variation of position or energy asymmetries. Such higher order biases
77: are small for the Inner and Middle \Moller\ detector rings, but are
78: observed to be significant for the Outer ring.
79:
80: During the physics runs, great care was taken to minimize
81: residual time structure of the beam position at the target, keeping it
82: typically within 1 mm. In order to measure the possible
83: bias due to such effects, six BPMs were instrumented with
84: 23 additional readout channels before the 2003 run. Thus, in addition to
85: the average beam parameters for each pulse, BPM signals for charge,
86: energy, positions, and angles are each digitized in four independent time
87: slices (three slices for energy). Corrections due to intra-pulse
88: variation of beam asymmetries are computed by linearly regressing
89: \Moller\ asymmetries against the beam asymmetries in time slices.
90:
91: For the 2003 data, the regression analysis limits the possible
92: contribution to the detector asymmetry due to the
93: intra-pulse variations to 3 ppb. Since time slice
94: data were not available for 2002 data, the Outer \Moller\ ring channels
95: are only used in 2003 data sample.
96: For the 2002 datasets,
97: these channels set a limit on the maximum possible bias
98: in the two innermost \Moller\ rings (containing the bulk of the
99: statistical weight) of less than 5 ppb.
100:
101: After linear regression, the integrated responses of all the
102: selected rings are averaged to form the raw asymmetry
103: $A_{\mathrm{raw}}$. Near-perfect azimuthal symmetry reduces the
104: sensitivity to beam fluctuations and right-left beam asymmetries. The
105: $A_\mathrm{raw}$ pulse-pair distribution has an average RMS of
106: $215$~ppm for the 2002 data and $185$~ppm for the 2003 data.
107: The cumulative beam asymmetry
108: correction is $-9.7\pm 1.4$~ppb. A correction due to an azimuthal
109: modulation of $A_{\mathrm{raw}}$ \cite{transverse} from a small
110: non-zero transverse component of the beam polarization is found to be $-3.8\pm
111: 1.5$~ppb.
112:
113: The average electron beam polarization is
114: $P_b=0.89\pm 0.04$, measured every few days by a
115: polarimeter using \Moller\ scattering of beam electrons
116: off a magnetized foil. The linearity of the calorimeter response
117: is determined to be $\epsilon=0.99\pm 0.01$ from special calibration runs.
118:
119: The physics asymmetry $A_{\rm phys}$ is formed from $A_{\rm raw}$ by
120: correcting for background
121: contributions, detector linearity and beam polarization:
122: \[
123: A_\mathrm{phys} = \frac{1}{P_b \epsilon}
124: \frac{A_\mathrm{raw} - \sum_i \Delta A_i}{1-\sum_i f_i} \, .
125: \]
126: Asymmetry corrections $\Delta A_i$ and dilutions $f_i$ for various
127: background sources are
128: listed in \tabref{tab:systematics}.
129: %
130: %correction and dilution table
131: %
132: \begin{table}[b]
133: \caption{Corrections $\Delta A_i$ and dilutions $f_i$
134: to $A_{\rm raw}$ and associated systematic uncertainties.
135: }
136: \label{tab:systematics}
137: \begin{center}
138: \begin{tabular}{|l|r|c|}
139: \hline
140: Source & $\Delta A$ (ppb) & $f$ \\
141: \hline
142: Beam (first order) & $-10\pm 1$& \\
143: Beam (higher order) & $0\pm 3$& \\
144: Transverse polarization & $-4\pm 2$ & \\
145: $e^-+p\rightarrow e^-+p(+\gamma)$ & $-7\pm 1$ & $0.056\pm 0.007$ \\
146: $e^-(\gamma)+p\rightarrow e^-+X$ & $-22\pm 4$ & $0.009\pm 0.001$ \\
147: $\gamma+e^-\rightarrow e^-+\gamma$ & $0\pm 1$ & $0.005\pm 0.002$ \\
148: High energy photons & $3\pm 3$ & $0.004\pm 0.002$\\
149: Synchrotron photons & $0\pm 1$ & $0.002\pm 0.001$\\
150: Pions & $1\pm 1$ & $0.001\pm 0.001$\\
151: \hline
152: \end{tabular}
153: \end{center}
154: \end{table}
155:
156:
157: