S13CharacterizationofGammaShieldfortheK100DilutionRefrigerator

Characterization of a Gamma Lead Shield using a Dilution Refrigerator for the SCDMS Test Facility

Alex Codoreanu and Nafiz Jaidye

University of Minnesota

Methods of Experimental Physics Spring 2013

Abstract

A Gamma Lead Shield was constructed for a test cryostat in order to limit the influence of background triggered events in the characterization of Ge detectors for the Super Cryogenic Dark Matter Search collaboration. The detectors are calibrated using the emission lines of and sources, thus the area of interest for our study is [0,380] keV. A rejection distribution was created as a function of incoming gamma particle energy by removing the influence of the sources and normalizing for the same exposure time for a no shield and with shield spectra. This rejection distribution’s fit is:

This rejection function is a useful tool in simulating the influence of the gamma-background in future testing facilities.

Introduction

The Super Cryogenic Dark Matter Search, S.C.D.M.S., collaboration aims to make the first direct interaction detection of Dark Matter particles. Using the Wilkinson Microwave Anisotropy Probe, WMAP, results it was found that Dark Matter and Dark Energy dominate the energy density in our universe. The current distribution is 5% regular matter, 25% Dark Matter and 70% Dark Energy [1]. Galaxies themselves form inside the DM halos and their gravitational influence is visible in the galactic rotation curves as seen below in figure 1[2]. The term “Dark Matter” was coined because DM is not visible through light and only interacts gravitationally and possibly through the weak force. Studies of gravitational lensing of distant galaxies have also provided information on the size of DM clusters/halos but a measurement of the rest mass energy of the individual constituent particles has not been made [2]. A few candidates have been proposed for the nature of DM which includes both baryonic and non-baryonic matter [2]. One candidate is a Weakly Interacting Massive Particle, W.I.M.P. These non-baryonic particles are expected to only interact via the weak force with regular matter and it is this property that S.C.D.M.S. is investigating [2].

Theory

The S.C.D.M.S. collaboration uses Ge detectors to detect possible interactions of incoming particles with target nuclei. These incoming particles can consist of high-energy photons which can release outer shell electrons through the photoelectric effect or neutrons which release outer shell electrons through nuclear recoil [2]. The R&D facility uses Am-241 and Ba-133 as known gamma sources in order to characterize and calibrate the detectors. The testing facility also has a natural gamma background [5] which arises out of the radioactive decay of and several isotopes as seen below in figure 2. Moreover, cosmic rays are another contributor to this gamma background.

An electric field is applied across the detector in order to collect the electron-hole pairs resulting from an interaction event. As the original ejected electron travels through the Ge detector lattice, it creates a cascade effect which releases other electrons and the total collection of collected electrons is the ionization signal. The phonon channel information is directly proportional to the total change in temperature due to the propagation of this cascade effect. Phonon channel information will not be used in this analysis as only gamma sources will be used.

In order to isolate source photons from background photons a gamma shield was constructed out of Pb. The absorption as a function of energy and thickness of the shield

where n = Z density, = energy dependent cross section, x = distance traveled in the material. The energy dependent cross section was calculated using the National Institute of Standards and Technology’s website . Accounting only for absorption due to the photoelectric effect and using the Beer-Lambert law, the probability for a γ-ray to penetrate a material of thickness x and density ρ is:

The expected spectrum, assuming a lead cube with no seams and total absorption can be seen below in figure 3. As can be seen in figure 3, a half-inch of lead provides an attenuation of approximately 100 if no other factors, such as isotropy of the design, are taken into consideration. Based on the weight requirement that the total structure weigh less than 8000 pounds along with this attenuation consideration, it was decided that the structure would be designed as a .5” shield.

Our goal was to find a rejection ratio of incoming Gamma particles as a function of their energy for the Gamma-Lead shield.

Experimental Setup

The detectors are attached to a tower support structure which is then installed inside of the K100 Cryostat. The detector’s ionization channels are electrodes which are embedded inside of the Ge detector. These are connected to a trans-impedance (charge –to-voltage) amplifier, whose diagram is in figure 4[6].

Figure 4. Simplified schematic circuit diagram of the ionization channel readout for a single charge channel. Ccoupling (300pF) blocks low frequency noise from the detector side of the circuit. Rf is composed of 4 inline 10MΩ resistors and the parasitic capacitance of the circuit is modeled as a capacitor to ground (Cp). The innate capacitance of the resistors is Cf(≈1pF) from [6].

The ‘output’ signal is then passed to a Superconducting Quantum Interface Device (S.Q.U.I.D.) paired with a Field Effect Transistor card (S.Q.U.E.T.) connected to a SQUID line which leads to an outer EBOX. This EBOX acts as a Faraday cage to further isolate the signal from outside influences. The Data Acquisition System (DAQ) is connected to the EBOX via Detector Control and Readout cards (DCRC’s). The DAQ, via the DCRC’s provides the trigger information and collects the charge and phonon data from the detectors. The DAQ first collects 500 random triggered events in order to create a smart filter algorithm and then triggers once a set threshold voltage has been overcome. The analysis package also performs an initial fit to a predefined function defined with a 1μs rise time and 40μs decay which can be seen below in figure 5.

Figure 5. Left panel: single event pulse Right Panel: A two event pulse as triggered by the first pulse. The second event piles up on the signal resulting in a large value when fitting to the expected single pulse information [7]

It then performs a FOURIER transformation of the pulse signal and records the data. The software then creates a ROOT tree with the respective data run information in which the applied bias voltage,, total charge collected, event number and event time for each event are recorded.

Figure 6. Left Panel: Photograph of G101b detector, Side 2 Right Panel: Cross-section view of the electric field lines resulting from the applied bias voltages.[2]

The detectors collect charge information on both sides of the detector, side 1 and side 2 and the field lines can be seen below in figure (6)

Shield Structure

The Gamma Lead Shield, in combination with the neutron polyethylene shield has been constructed for minimal background count in the lab. Currently, an interior top shield is being constructed. The shields are mounted on a 2 cart system and the total weight of the set-up is close to 6500 lbs. The carts have been designed to safely handle up to 8000 lbs and the 3D drawing and the installed shield can be seen in figure 7 below.

Figure 7. 3D mock up and final installation of the Lead Gamma Shield.

Data Collection

Data was collected with the shield structure at different biases: +/- 4V, -/+ 4V, +/- 2.5V, -/+ 2.5V, +/- 3V and -/+ 3V where the bias values represent the applied voltages to the charge channels on side 1 and side 2, respectively. Two internal gamma sources, 210Pb and 241Am, were installed on side 2 which will be referred to as source side. Side 1 will be referred to as opposite source side. The 210Pb source was installed with the goal of completing a low energy study which focused on surface event rejection for the G101b detector design. The 241Am source is used as a quality of data indicator as its’ 60 keV emission line has a penetrating depth of ~1mm thus a quality data set will only have a 60 keV spectral spike on source side only. A 133Ba was also used in order to calibrate our energy scale as will be discussed in the Data Analysis section.

Shortly into the run we experienced a breakdown of the D.C.R.C.’s charge channel amplifiers which impeded our ability to continue taking data without the shield at the same applied bias voltages. A He tight vacuum seal also broke down in our circulation pump and the decision was made to abort the run until it could be fixed. The lead time for the seal was 5 weeks, thus we focused on previous collected data without the shield, with the same detector and with the same applied bias voltages in order to complete our study. After an investigation into the previously collected sets we found a +/- 3V and a +/- 4V collection of series taken without the shield which could be used. Further investigation led us to focus on the +/- 4V for the final investigation into the rejection function. We encountered the troubles before completing the +/- 4V collection with the shield and we only had data with the 210Pb and 241Am sources but the no shield data was taken with the 241Am and 133Ba sources. The detectors were designed as to have a linear relationship between collected charge and incoming energy of photons for the same bias voltage thus the energy scale calibration from the no shield data set can be applied to the with shield data. The final data sets used were:

- Run 36/With Shield, series ending in 0120, total exposure time of 400 minutes

- Run 28/No Shield, series ending in 0337, total exposure time of 210 minutes

Analysis and Results

The goal of the analysis procedure is to isolate data points characteristic of a single event in which the total ionization energy has been collected. In order to ensure total collection of ionization energy, bulk events must be isolated. The bulk refers to the cylindrical volume defined by the area seen below in figure 8. The detector has been manufactured with a radius of 44mm and a radius of 50mm. This design, effectively uses the volume defined by only as a statistical veto tool by comparing the associated collection with the associated collection for each event.

Figure 8. Qinner and Qouter charge channel

In order to isolate the Q(i) dominant events we define,

and select those events which meet the 0.9 < Qpartition < 1.00 criteria. Qi and Qo are scalar values which are related to the total energy deposited across the detector area, and respectively. Each increment in and corresponds to as defined by the F.E.T.’s resolution. The collected data and the applied cut can be seen below in figure 9.

Figure 9. Left Panel: plot for NO SHIELD data set Right Panel: plot for WITH SHIELD data set

The final spectra with associated errors can be seen below in figure 16.

Figure 16. Left Pane

l: Final Bin Spectra NO SHIELD, with errors Right Panel: Final Bin Spectra, WITH SHIELD, with errors

The rejection function is simply the best fit to the distribution of and the error associated with each value in this distribution is

The best fit in our range of interest, defined by the location of the is seen below in figure 17.

Figure 17. Final Rejection Ratio distribution with associated errors and best line fit.

The final fit of the rejection ratio distribution is

with reduced chi^2 = 0.921 and p-factor = .847.

Conclusion

The fit to the rejection ratio distribution function provides a statistical tool needed in simulations of future detector designs and their response to the lab’s background. Another application of this fit to the rejection ratio distribution can also provide the underpinning to anticipating whether or not the shield will continue to be effective with a different background profile as will be the case when the lab will switch locations. Further improvements and verification of this statistically found rejection function can be made by ensuring that both No_Shield and With_Shield data sets are collected with identical sources thus simplifying the analysis.

References

[1] N, Jarosik, “Seven – year Wilkinson microwave anisotropy probe (WMAP) observations: Skymaps, systematic errors, and basic results.” APJS, 192, 14 (2011).

[2] J. M. Kiveni, “A Search for WIMP Dark Matter using an Optimized Chi-Square Technique on the Final Data from the Cryogenic Dark Matter Search Program (CDMS II)”, http://surface.syr.edu/phy_etd/129/

[3] “DARK MATTER”, http://bustard.phys.nd.edu/Phys171/lectures/dm.html , Feb. 26th 2013

[4] G. Steigman and M.S. Turner, Nucl. Phys. B253, 375 (1985)

[5] H. Chagani and A. Kennedy, “Brief Feasibility Study of Gamma Shield for K100”.

[6] R. Radpour, “DESIGN, CONSTRUCTION, AND ASSESMENT OF A NEUTRON SHIELD FOR CDMS TEST FACILITIES”, Aug 2011.

[7] “Characterization of neutron shielding for the K100 cryostat for the CDMS experiment”, Sean Vig, University of Minnesota Twin Cities, May 5th 2011.

Acknowledgements

Many thanks to Hassan Chagani and Prf. Greg Pawloski for their instruction and direction with the analysis procedure. We would also like to acknowledge Roxanne Radpour and Allison Kennedy for their help with data collection.

The final fit of the rejection ratio distribution is

Figure 17. Final Rejection Ratio distribution with associated errors and best line fit.

with

reduced

and p-factor = .847.