NSS-MIC 2006 Conference Record Template



Timing Measurements of a New Micro-Channel Plate Photomultiplier Incorporating a Novel Equal-Time Anode Structure

Karen Byrum, John Anderson, Gary Drake, Camden Ertley, Henry Frisch, Harold Sanders, Fukun Tang, Jerry Va’vra

(Abstract–We have measured the timing properties on two new 2-inch micro-channel plate photomultiplier tubes from Photonis; one tube with 25 micron pores and the other tube having 10 micron pores. The 10 micron pore tube uses a novel charge-collection scheme at the anode to provide equal arrival time of the signal independent of the hit position of incident light on the face of the tube. We have performed these measurements using a Hamamatsu PLP-10 picosecond laser and a commercial Camac readout electronics system. We present these results, and compare timing properties to a standard Burle/Photonis 85011 micro-channel plate photomultiplier tube having 25 micron pores.

INTRODUCTION

O

ur goal is to build a Time-of-flight (TOF) detector that has resolution on order of a picosecond (psec), which corresponds to a distance of 300 microns at the speed of light. The goal of a psec is set by the application in High Energy Physics for particle identification by velocity measurement [1] in future detectors, such as the 4th Detector [2] at the International Linear Collider (ILC) [3], in muon collider experiments [4], possible future precision flavor experiments at the Tevatron [5], and other HEP applications; high-resolution applications also exist in many fields, including radiology, astrophysics and nuclear physics. While spatial resolution has improved by more than an order of magnitude with the advent of silicon strip detectors, time resolution has been stagnant for 30 years or more. As a result, the ability to measure the mass of particles produced in very high-energy collisions, such as those now at the Tevatron, and soon to be at the Large Hadron Collider (LHC) [6], has remained out of our reach beyond momenta of a few GeV. Typical timing resolutions in current state-of-the-art TOF systems today are on order 100 psec; this resolution is set by the characteristic transverse size of an inch. By using the Cherenkov light, with no bounces, directly on a device with characteristic size in the sub-millimeter range, we can access much better time resolutions.

Micro-channel plate (MCP) photomultiplier tubes (PMTs) have been shown to have excellent timing properties. We are interested in exploiting and optimizing these properties to develop a technique for measuring transit time of particles on the order of 1 picosecond resolution.

To achieve our goal of developing a large-area time of arrival system with psec timing resolution, we are focusing on a multi-prong approach which includes 3-D modeling of the photo-optical devices, developing a laser facility at Argonne National Laboratory to fully characterize and compare the performance of these devices (which is the focus of this paper), to design and construct the ultra-fast electronics to readout these devices, to develop a complete end-to-end simulation, and ultimately to perform studies in the Fermilab Test Beam Facility. In this paper, we are reporting on timing measurements performed in the Argonne laser teststand using several Burle/Photonis MCPs, one which incorporates a novel charge-collection scheme at the anode.

Photodetector

Micro-channel-plate photomultiplier tubes have been shown to have excellent timing properties. The best results that we are aware of are from Jerry Va’vra at SLAC and from Takayoshi Ohshima of University Nagoya. Va’vra measured an upper limit of the MCPPMT resolution of σMCPPMT~5ps using two Burle/Photonis 10micron MCP hole diameter tubes and a PiLAS red laser diode at Npe ~ 50 [7]. Ohshima has reached a σMCPPMT~6.2ps in a test beam using two identical Hamamatsu R3809U-50-11X MCPPMTs in a beam with a quartz radiator (Npe ~ 50) [8].

Optimization of the MCPPMT for fast timing performance is a crucial step to achieve our goal of order 1 psec timing resolution. For good timing performance, it is desirable to use the smallest pore size possible. We started with a tube having 25 micron pores, and recently obtained one with 10 micron pore size. We use the detection of Cherenkov light generated in the face window of an MCP. The charge collection at the anode is a primary area of development, since the timing skew over a 2 inch square tube can be a few 100 picoseconds depending on where on the window the light is emitted. We have developed a new charge-collection scheme at the anode of the tube called the “equal time anode” [9]. The 10 micron tube has a 32 by 32 array of anode pads, and these are collected into 4 quadrants each of which is a 4 x 4 array. The 16 pads for each quadrant are routed using equal length traces. Each quadrant is summed into a SMA connector, which goes to the oscilloscope. Custom electronics that would sit directly on the back of the MCPPMT module to receive the summed signals on a low-inductance path is being developed.

For the measurements reported in this paper, we have studied 3 MCPPMTs in the Argonne laser test stand. These include a Burle/Photonis 85011-501 25um pore, 64 anode tube and two other MCP tubes which were developed by Burle/Photonis with special features. The second MCPPMT studied, which we call Mark-0, is shown in Fig 2a. Mark-0 is a 25um pore, 64 anode tube that has an improved ground plane to correct for the current return path [10]. All timing measurements of the Mark-0 were done with the center four pixels tied together. The third MCPPMT, which we call Mark-1, is a 10um pore, 1024 anode tube. This tube, shown in Fig 2b, also has the improved ground plane and the equal time anode clamped with a conducting elastomer connector to the MCP [11]. (This clamping was temporary and used for testing; the next step will be to directly fasten the PC board to the MCP). Each quadrant of the equal time anode collects charge over 1in2 and has an equal path length for each anode.

[pic] [pic]

Fig 2. a) The left hand figure shows the “Mark-0”. The right figure is Mark-1. See text for description.

Argonne Laser Facility

The Argonne laser teststand was developed specifically to perform precision timing measurements. The optical bench setup is shown in Fig 6 and a schematic of our setup is shown in Fig 7. The light source, a Hamamatsu PLP-10 picosecond laser light pulser [12] provides light with either a 405nm or 635nm diode head. The PLP-10 laser pulse to pulse jitter is less than 10psec, as specified by the manufacturer. The light was optically transmitted onto an optical splitter, focusing lenses, diaphragms and neutral density filters to focus the beam onto the face(s) of the MCPPMT(s). The MCP was attached to an xyz stager unit; we also have the capability a attaching a second MCP to a xy stager.

The DAQ consisted of a CAMAC control unit, a WEINER CC-USB [13], which was used to initiate the start pulse generated by the Hamamatsu PLP-10. The signals were processed through an Ortec 566 time to analog converter (TAC) [14] to measure the amount of time between a start pulse and when the output of the MCPPMT happened. The laser was focused onto a 500 μm hole and a neutral density filter. The attenuated laser was split and focused on the face of the MCPPMT. A sync pulse, created by the laser controller, was used as a start pulse for the TAC. The anode of the MCPPMT was connected to an Ortec 9327 [15] constant fraction discriminator (CFD). The output from the CFD was used as the stop pulse for the TAC. The signal from the TAC was measured by an Ortec AD114 analog to digital converter (ADC) [16]. The data was saved on a computer and then analyzed with ROOT [17].

The numbers of photo-electrons (Npe) were adjusted by changing the amount of filters between the laser and the MCPPMT. The Npe was calculated by estimating the amount of charge in a single pulse from the MCPPMT and then calculated using the known gain of the MCPPMT.

The equipment was calibrated using a Pulse Generator from Impeccable Instruments with pulse to pulse jitter of less than 3ps. The pulser produced a start and stop pulse. The stop pulse was delayed using a length of cable and readout with a TAC. The intrinsic jitter of the system (excluding the Laser) was found to be σPulser+AMP/CFD+TAC+ADC~6.2ps. The timing jitter of the Pulser + TAC + ADC was found to be σPulser++TAC+ADC~3.3ps.

[pic]

Fig 6. The Argonne laser teststand optical bench. See text for description.

[pic]

Fig 7. Schematic of Argonne laser teststand. See text for description.

Results

In this paper, we focus on the first timing resolutions we have achieved using the PLP-10 laser and one MCPPMT. The best timing resolution was obtained with the Mark-0 or Mark-1 with the Mark-1 giving a slightly better value. The timing resolution of the Mark-0 as a function of the Npe is shown in Fig 9. For this measurement, the laser light was 635nm and the MCPPMT HV was 1900V. An upper limit on the timing resolution for the MCP device was measured at σMARK-0 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download