Designing a Current Source for Toroid Integrated Intensity System Calibration. MELANIE DAY (Rochester Institute of Technology Rochester, NY 14623) M.A IBRAHIM (Fermi National Accelerator Laboratory, Batavia, IL, 60510)

At Fermilab, the toroid integrated intensity system is used to measure beam intensity. This measurement is especially important when lowest possible error is critical, as in the MINIBooNE neutrino oscillation experiment. In order to reduce the error in toroid measurement, a calibration module was commissioned by the MINIBooNE collaborators through Columbia University. However, because of certain dependencies, it was necessary to design a current source to replace the existing pulse generating circuitry of the module. Several current source designs were examined, with special consideration given to the specific needs of the toroid integrated intensity system. Models done in WinSpice simulated the current output for each design. The simulation was then used to analyze the difference between an ideal source and the actual output of each model due to its dependence on change in the terminating resistance of the circuit. In simulation, the final design gave an optimal error margin of approximately +/-1mA for every .1ohm change in terminating resistance, or a +/-.5mV change in the voltage pulse input to the toroid integrator. Simulation suggests this value could be modified by changing the terminating resistor from the current 50ohm load to a higher value resistance. Extensive testing of the model has not yet been done. Future alterations may be necessary to deal with effects that cannot be seen in the simulation. Also, other circuitry in the pulse generator will need to be altered in the future to deal with the new requirements of the current source.

Determining the rate of boron 12 decay in the MiniBooNE neutrino detector. RUTH TONER (Yale University New Haven, CT 06520) MORGAN WASCKO (Fermi National Accelerator Laboratory, Batavia, IL, 60510)

In addition to its main goal of investigating neutrino oscillation, the MiniBooNE collaboration conducts several physics projects involving the search for exotic particles with energies as low as 10 MeV. It is therefore essential to understand all sources of background at this low energy, including that of the beta decay of boron 12. This background appears when cosmic ray muons stop in the detector and are captured by mineral oil carbon 12 nuclei. The most important step in calculating the cross section of this decay reaction in the detector was accurately determining its rate in the detector. The MiniBooNE exotics pre-report predicts that the rate in the MiniBooNE detector should be approximately 11 Hz. Sample data was taken from the detector strobe stream, which had been recorded by the data acquisition system (DAQ) with an external trigger independent of the neutrino beam. A series of filters and pre-cuts were applied to the data, turning the DAQ raw data into a usable set of events with detailed information about energy, event location and time, etc. The total boron 12 decay rate was taken to be the total number of boron 12 events observed divided by the integrated time of observation. The total time for the rate calculation was taken to be the time during which the DAQ recorded good data - i.e., the number of events passing the filters multiplied by their constant time duration. The total number of boron 12 events was taken to be those events passing the precuts and a second series of cuts applied in ROOT, including a rough energy cut only passing those events below 15 MeV. By knowing the rate and the approximate number of detector carbon atoms, the frequency per carbon atom was calculated to be 1.76 x 10-30 Hz. This number can later be used to get a cross section, once the flux of muons stopping in the detector is better understood. Although the order of magnitude seems reasonable, the decay rate number used in this calculation (33.43 Hz) seemed somewhat high. The study therefore recommend a more complicated series of fitting functions to apply weighted cuts and determine a more accurate count for the boron 12 events.

Event Display tools for the Cryogenic Dark Matter Search (CDMS). REGINA CAPUTO (Colorado School of Mines Golden, CO 80401) DAN BAUER (Fermi National Accelerator Laboratory, Batavia, IL, 60510)

Dark matter is an attractive solution to the problem of missing matter in the universe as observed in galactic rotation curves and other large body observations. Cold non-baryonic Dark Matter (CDM) emerges as the leading candidate among the other types of dark matter because it has properties required for the formation of the current universe. The Cryogenic Dark Matter Search (CDMS) is using cutting edge technology to aid in the direct detection of Weakly Interacting Massive Particles (WIMPs), a type of CDM. This paper discusses the process by which WIMPs, as well as other incoming particles, are detected by the CDMS experiment and how the data is collected and characterized. Once data is collected, however, accessing the raw data by members of the collaboration is important. This paper also focuses on event display software that allows easy access to raw data. A problem occurs when raw data is stored where there are restrictions to accessing the files. This problem can be solved by incorporating two versions of event display software into general use. The first, written in ROOT, can graph raw data and apply a curve fitting so that, after integrating the curve, the energy of the particle can be determined. The second version, written in CGI for the Internet, can allow access to graphed raw data saved in a site accessible to the collaboration.

Optimizing The B _s → μ ⁺ μ ^- Decay Search At CDF Using An Artificial Neural Network. BARRY AMOS (Saint Mary's College of California Moraga, CA 94565) CHENG-JU LIN (Fermi National Accelerator Laboratory, Batavia, IL, 60510)

In the standard model (SM), the Flavor Changing Neutral Current B _s → μ ⁺ μ ^- decay is highly suppressed. The SM expectation for this branching fraction is B(B _s → μ ⁺ μ ^- ) = 3.42 +/- 0.54x10^-9, which is about an order of magnitude smaller than the current experimental sensitivity at the Tevatron. However, new physics contributions can significantly enhance this branching fraction. An observation of this decay at the Tevatron would be unambiguous evidence for physics beyond the SM. However, identifying Bs candidates is an experimental challenge due to the small branching fraction and large background contamination in the data. In a previous search, a multivariate likelihood technique (LH) was used to discriminate between signal (B _s → μ ⁺ μ ^- events) and background (non B _s → μ ⁺ μ ^- events). In this analysis, we study the feasibility of using a Neural Network to improve signal and background separation using of data collected by the CDF II detector. To quantify the improvements, we train the Neural Network using the same three input discriminating variables and data sample as the ones used in the LH analysis. The first input variable is the proper decay time of the Bs candidate. The second input variable is the opening angle between the Bs candidate momentum and decay axis, where the decay axis is defined as the vector pointing from the Bs production to the decay vertex. The last input variable is the Bs candidate isolation, which is defined as the ratio of the Bs momentum to the scalar sum of the momentum of other adjacent tracks around the Bs candidate. Using independent samples for testing and training the Neural Network, we were able to optimize the number of training cycles and structure of the network. To train the network we use a standard-backpropagation algorithm. The structure of the network consists of eight nodes in the first hidden layer, six nodes in the second, and a single output node. Our results for the Neural Network show about a 10% increase in signal acceptance with the same background rejection compared to the LH. With future plans of using more input variables to further optimize the Neural Network, we expect to improve significantly the signal acceptance and background rejection.

Optimizing the Handling of Large Datasets. CHRISTOPHER DORAN (DePauw University Greencastle, IN 46135) FRANK CHLEBANA (Fermi National Accelerator Laboratory, Batavia, IL, 60510)

Collision Detector at Fermilab (CDF) Run II has recently surpassed 1 fb^-1 of integrated luminosity and it is hoped that it will reach at least 6 fb^-1 before the Tevatron is decommissioned. The number of events is proportional to the integrated luminosity, and if the goal of 6 fb^-1 is reached, the result will be close to 500 TB of raw data. The large amount of data presents a challenge for offline data analysis. The current inclusive-jet dataset, sampled from 500 pb^-1 of raw data, requires more than 5 hours to analyze. Steps must be taken to increase the speed and convenience of analyses as the data sample continues to grow. By condensing the data such that it contains only the information necessary for the particular analysis, and/or by running the analysis in parallel on multiple machines, one can increase the speed of analysis significantly. Condensing also makes it easier to store and access the data locally, increasing data access speed and convenience. Results show that the user is able to increase speed by several factors. These strategies will make the growing data sample more manageable, and they may be applied to any analysis of large datasets.

Quality Assurance Testing of HDI Circuitry for CMS FPix Detector. DANIEL ROKUSEK (University of Illinois at Urbana-Champaign Urbana, IL 61801) GREG SELLBERG (Fermi National Accelerator Laboratory, Batavia, IL, 60510)

The Large Hadron Collider (LHC) at CERN will be used to a variety of aspects of high energy physics. One such aspect is the breaking of symmetry in the electroweak force. It has been theorized that the discovery of the Higgs particle could uncover the nature of this symmetry breaking. The Compact Muon Solenoid (CMS) will be used to search for the Higgs, beginning with the Forward Pixel (FPix) detector located at the center of CMS. FPix is a robust particle tracker, featuring silicon strip and pixel detectors. Before the components of FPix can be assembled, it is of the utmost importance that each individual piece of circuitry be tested to verify proper performance. The work presented here consists of quality assurance (QA) testing of the high density interconnect (HDI) circuitry for FPix. Prior to testing the HDIs, software was written to automate the probing process. National Instruments' LabVIEW 7.0 was used to write a custom probing software package. This package uses the NI GPIB (IEEE 488) interface to control a Rucker & Kolles 680A Semi-Automatic probe station and an Agilent 4284A precision LCR meter. Features of the software package include a user interface to enter input data of HDI surface feature locations generated from the HDI Gerber file, perform probing tasks and analysis, and generate output data. Special care was taken to design a HDI alignment sub-routine to allow a user without previous knowledge the HDI circuitry, software, or probe station to perform the QA tests. Results from the testing of the initial batch of HDIs show shorting errors. Errors were manually inspected and verified, and the information was passed onto the HDI design team and to the manufacturer for analysis. Once the errors are remedied, full production of HDIs can be approved and the HDIs will be tested with this software package. Upgrades to the software package could begin with the replacement of the semi-automatic RK680A probe station with an Electroglas 2001X probe station. The incorporation of additional probes and motion stages to the 2001X probe station will result in a fully-automatic flying-head probe station, and will significantly decrease the HDI probing time.

Simulation of Neutron Therapy at the Neutron Therapy Facility Using Geant4. AMANDA WEINMANN (Saint Mary's University of Minnesota Winona, MN 55987) ERIK RAMBERG (Fermi National Accelerator Laboratory, Batavia, IL, 60510)

Neutron therapy is a type of radiation therapy used to treat malignant tumors. Unlike conventional radiation therapy that uses photons, neutron therapy destroys the cells by interacting with atomic nuclei via several, varying mechanisms. It is important to be able to predict how the neutrons will interact and lose energy in tissue in order to develop unique and effective treatment plans for patients. For this reason, this project sought to create a simulation that could model the energy deposition of neutrons in a patient at the Neutron Therapy Facility (NTF). A simulation toolkit named Generation and Tracking 4 (Geant4) was used. Geant4 is designed to model particles passing through matter, and it is capable of simulating many user-defined details, such as particle generation, physics processes, and geometries. The geometries used by NTF are imported from Computed Tomography (CT) images that meet a medical standard known as Digital Imaging and Communications in Medicine (DICOM). The Hounsfield numbers of the DICOM images are converted into linearly-correlated tissue densities and constructed into Geant4 geometries. Once the geometry is developed, aspects of the neutron beam must be characterized. Other reference materials discuss specifics of the NTF beam. Currently, the simulation models the initial beam spot, energy spectrum, and downstream flux of the neutrons. The beam spot is simulated as the size of the neutron beam that is emitted from 66 MeV protons striking a beryllium target; the neutrons' energies are reproduced from a plot of the NTF neutron energy spectrum; and the downstream flux is simulated as a uniform and flat spatial distribution at the exit of a variable collimator. This simulation is distinctive because, just as neutron therapy is relatively rare compared to photon therapy, most radiation therapy simulations are intended to model photon radiation. It is also different from other techniques because it does not require that ionization profiles be defined for each type of tissue or material. The Geant4 simulation is a useful technique for NTF, but future work must be conducted to describe other aspects of the beam and obtain acceptable accuracy levels.