1. Hello everyone, I'm Diancheng Wang from Univerisity of Virginia. Today I will be talking about the PVDIS experiment at JLab 6GeV. I'll start with a brief introduction of the Physics, followed by an overview of the experimental setup. And then I'll go through the systematic uncertainties rather quickly. And at last I'll show some preliminary results. 2. In this experiment, we measured the parity violating asymmetry of a polarized electron beam scattering off an unpolarized deuterium target in the deep inelastic scattering region. Under parity transformation, the helicity state of the electron changes sign. As we all know, parity is violated in weak interation. This causes a cross section difference between the scattered electrons of different helicity states. The observable is this parity violating asymmetry here. The weak interation by itself is very small compared to electro-magnetic interactions. However the interference between this two gives a rise to the weak interaction so that this parity violating asymmetry is measurable. Under the framwork of standard model, the PVDIS asymmetry from a deuterium target is expressed in this formula. The Q2 is four momentum transform. The R's are ratioes of the parton distribution functions of different quarks. The big Y contains kinematics information. And the C1 and C2's are weak neutral coupling constants as define here, which are basically as fundermental as the weak mixing angle theta W. So by measuring this asymmetry to a high precision, we can extract the coupling constants, which provides us a test on the standard model. One thing that's unique about PVDIS is that it's the only one to measure C2's among current EW experiments.... 3 As shown in this slide. A list of SM test experiments are shown here. All of them are quite famous. The Qweak experiment at Jefferson Lab has just completed its production recently. As you can see, these different experiments probe different parts of the Lagrangian, and PVDIS is the only one accessing the C2 couplings. 4 This slide here shows our current knowledge on the weak coupling constants. The left figure is for the C1 couplings. There are many expeiments that measured C1, and the best constrain will be provided by the Atomc Parity Violating (Cs) experiment and the Q-weak experiment. The star here shows the standard model value. Compared with the C1 couplings, our knowledge on the C2's is rather limited. There are only a couple of experiments and their precisions are pretty poor. The main motive of this experiment is to improve our knowledge on the C2 couplings. 5 The experiment was carried out at Jefferson Lab, which is a linear accelerator providing highly polarized electron beam with beam energy upto 6GeV. Actually the 6GeV era of JLab has just ended and it will be upgraded to 12GeV in the coming years. The high luminorsity at JLab makes it perfect for precision measurements. There are three experiment halls, A, B and C (actually another hall D is currently under construction). and this experiment was done in Hall A. 6 Here is an oveview of the experimental setup in Hall A for this experiment. The beam comes in from the accelerator. There are two polarimetries along the beam line, the Compton and Moller. Also there are various beam monitors to monitor beam current and position. Then the beam hits the targe. The scattered electrons are detected by two hig resolution spectrometers, the Left HRS and the right HRS. They are very similar to each other. Here is a side view of the HRS. The electrons are bent by a combination of magnets, which are usually referred to as the Optics. Then they reach the detector hut where all detectors are located. There are two layeres of Verticle Drift Chambers for tracking purpose, and other detectors including scintillators, Gas Cherenkov, and Lead Glass counters for particle identification. We measured the asymmetry at two different kinematics points, as shown here. 7 In order to count the electrons precisly, we built a scaler-based fast counting DAQ which does the particle identificatin on the hardware level. Here is a sketch of the trigger logic of our DAQ system. A good electron trigger would require coincidence of all the detectors. Also, in order to cleanly seperate the electrons from pions, two hardware cuts are applied using discriminators, one cuting on the preshower energy deposit, the other cutting on the sum of prshower and shower energy deposit, as explained in this spectrum here. 8 And this is a plot from real data showing the perfomance of this DAQ system. As you can see, all pions are cleanly cut out. One issue about this DAQ system is that it induces extra systematics, namely the Deadtime and the PID Efficiency. And we specifially designed two identical DAQ path with known discriminator width for deadtime study. 9 Here are two pictures showing the DAQ in the detector hut. 10 Next I will go through the systematic uncertainties in this experiment as listed here. I'll cover all of them,,, but rather quickly. 11 The first and also the largest systematic comes from the Beam Polarization. The Beam Polarization is measured by two independent polarimetries, the Moller and the Compton. The Moller results are shown here. Moller measurement is invasive, we need to stop the production whenever we take a Moller measurement. So we took one measuerment every several days. The uncertainty from Moller is 2%. Compton is non-invasive, so we take measurements contineously as the experiment was running. But the Compton only worked for the later half of the experiment. Compton results are shown here in this plot, together with several moller measurements. The uncertainty for Compton is also about 2%, which mainly from the analyzing power. When there are both Moller and Compton results, we combined the two and get a smaller uncertainty. 12 13 The PID performance of the DAQ is studied in detail. Here are the plots for the Electron detection efficiency and the Pion rejection factor of the lead glass PID, plotted versus the vetical dimension of the acceptance. We do the same analysis for the Gac Cherenkov. And overall, we achieved 95% of Electron efficiency and 10 to the 4th of pion rejection factor. The PID will affect the measured asymmetry only if it varies over the acceptance or if there are "holes". As you can see, the efficiency is quite stable over the acceptance, as a result, the correction is pretty small. 14 The next systematic comes from the Q2 measurement, which mainly depends on the optics calibration. The Optics is calibrated using a multiple-foil target and a multiple-hole collimator, which is usually referred to as "the seive". On the right side are the data plots after calibration. The sharp peaks here corresponds to the foils here. And here you can see distinctive holes clearly. After calibration, the uncertainty due to Q2 measurement is less than 1%. 15,16,17 18 Another systematic is the false asymmetry. There are mainly two kinds of false asymmetries for this experiment, both related to the beam. One is the charge asymmetry, caused by unequal beam intensity for different helicity states. During the experiment, we used a feedback mechanism to minimize the charge asymmetry. The beam intensity is measured using beam current monitors. And the information is then fed back to the accelerator side and beam intensity is then adjusted accordingly. With this feedback mechanism, charge asymmetry converges to zero much faster, as shown here. In just a couple of hours, it converges to sub-ppm level. 19 Another false asymmetry comes from the beam motion. If the beam moves in a way that's helicity correlated, it will cause false asymmetry. We have two methods to correct the beam modulation. The Dithering method, which intentionally moves the beam around, and study how the asymmetry changes as a result. The And regression method use the natural motion (or jittering) of the beam. This plot here shows how the dithering works. There are five beam monitors at different locations, providing detailed information of the beam's position. And this plot here shows the sensitivities of the asymmetry wit respect to different monitors. The correction due to beam modulation turned out to be small. 20 Various backgrounds are considered and corrected. One is the transverse asymmetry. We took some data with transverse spin, and the results are shown in this table. Also the uncertainty to our final asymmetry is shown. Another background is the positron from pair production. We took some positron runs during the experiment, and the positron asymmetry results are consistent with zero, although with a large erro bar. Also, despite that we have very large pion rejection factor, there's still some pion contamination. The pion asymmetries are observed to be non-zero, as shown here. Besides, the target cell is made of aluminum, so there's a little background from the endcap. We actually took some aluminum runs, but the statistics was too low. So we just used the SM calculated values to estimated the correction due to aluminum endcap. All these background corrections are very small except for the transverse asymmtry. 21 And next I'll show some of the results from this experiment. 22 We did a blinded data analysis while extracting the raw asymmetries, and two independent analysis were carried out as cross checks, so as to make sure no human errors were made. Here I'm showing the statical quality of the data. Plotted are the histograms of raw asymmetry within a 66 mili second helicity pair. Black is data and the red is a gaussian fit. A pure gaussian distribution means that the data is purely statistical. 23 This is a talbe summarizing all the uncertainties. We are dominated by the statistical uncertainties, expecially for kine #2. The systematics mainly come from beam polarization, radiative correction, Q2, transverse asymmetry and the deadtime. Others are very small. 24 And here are the finalized asymmtry results. From these results, we can extract the C2 coupling constants. One thing that complicates the interpretation of the asymmetry results is the higher-twist effect, which by itself is a very interesting subject and not all well understood. So we currently have two ways of interpreting our results. 25 One way is to extract the C2's using the high Q2 data point, and assuming that the higher-twist effect is negligible. The preliminary result is shown here, in good agreement with the standard model. And Delta(2C2u - C2d) is 0.052, which is a factor of five improvement over previous data. 25 Another way is using both data points combined and assume the higher-twist effect affects the asymmetry in this way, characterized by the beta HT parameter. The C2q-Beta HT Correlation is shown in this plot, the red ellipse. The result suggests that higher-twist effect is small. 26 summary 27