Home  |  Past Issues  |  Editorial Board  | RHIC

 

 

About the Author

Leo Greiner is a research engineer at Lawrence Berkeley National Laboratory who has been working on the Heavy Flavor Tracker at STAR since 2003.

A MAPS Vertex Detector for STAR – HFT Pixel Development

By Leo Greiner

The STAR Heavy Flavor Tracker upgrade group is working to extend the capabilities of the STAR detector in the heavy flavor domain by providing a tracking system that will allow for very high resolution vertex measurements. This upgrade is designed to allow for the direct topological reconstruction of D and B mesons through the identification of decay vertices displaced from the primary interaction vertex by 100 – 150 microns. The Pixel Detector is composed of two layers of high resolution Monolithic Active Pixel Sensors (MAPS) and is the core of the vertex detector upgrades for STAR.

Figure 1: D⁰ decay typical of the topology that the STAR HFT upgrade is designed to detect.

Achieving the required sub 40 micron resolution is quite challenging on many fronts. Multiple coulomb scattering constrains the detector to a design limit of < 0.5% radiation length per layer in the sensitive region. Contained in this small amount of material we need to have the MAPS sensors, a readout cable and a highly rigid support structure to maintain the position of the pixels, within an internal detector reference system, to 10 microns. In this stringent environment, the mechanical design requirements interact strongly with the sensor and readout electronics design and sensors thinned to 50 microns, air cooling, a 500 micron thick Be beam pipe, and aluminum rather than copper conductor readout cables become necessary. The expected final system design is an array of 33 sensor ladders with ten 2 cm x 2 cm sensors per ladder and parallel independent ladder readout systems. Each sensor is a 640 x 604 array of pixels giving the Pixel Detector system a total pixel count of > 135M. The eta coverage of the Pixel Detector is the same as the outer tracking detectors with -1 < eta < 1. We are pursuing an incremental approach to reaching the final design by constructing prototype generations of sensors and readout electronics.

Figure 2: End view of the Pixel Detector for STAR. There are two effective layers of MAPS sensors mounted to support ladders. The inner radius R1 = 5.0 cm. The outer staggered arrangement has R2 = 6.5 cm and R3 = 7.5 cm. The detector active length is ~ 20 cm and is composed of 33 identical ladders of 10 sensors each. The spoke structure shown is a beam pipe support.

As part of our incremental approach to solving the design challenges presented, we have been working with an existing prototype Mimostar2 sensor. The Mimostar2 is a MAPS developed by our collaborators at IPHC in Strasbourg, France with an array of 128 x 128 pixels. We have constructed a three sensors prototype detector in a telescope configuration, and concentrated our current design efforts on readout and integration with the existing STAR trigger and DAQ system. One of the main challenges facing this type of detector is that of data sparsification. The raw data rate for our 135M pixel detector is 50.7 GB / s. We implemented a 50 MHz on-the-fly data reduction system for test in the RHIC run 7. In this system, the sensor serial data output is resorted to give a raster scan through the pixel array. This allows us to identify clusters by doing threshold pattern recognition in an FPGA over a synchronous window that examines a sequential 3 x 3 geometric array of pixels. When a cluster is identified by these threshold criteria, only the address is read out. This allowed us, with hit densities comparable to what is expected in the final system, to reduce the data rate by a factor of ~1000 in the telescope readout system. This data reduction rate should directly extend to the final detector design. We also successfully integrated our system with the STAR trigger system and delivered our data to a STAR standard DAQ PC. In our analysis, we have used the hits from our sparsified data set to do tracking through the three layers of this telescope and have imaged the interaction diamond at STAR. We have also measured the charged particle density and the noise environment, both critical parameters for future designs. Further analysis of the data taken is currently in progress.

Figure 3: A frame of data from one Mimostar2 sensor in our prototype telescope. The hits from interaction at STAR clearly stand out from the noise.

This successful test of a prototype system is an important step and a proof of principle for design ideas that will allow us to deploy a full Pixel vertex detector system at STAR in the future.

More information about the STAR upgrades is available here.

The Relativistic Heavy Ion Collider at Brookhaven National Laboratory is a world-class scientific research facility primarily funded by the U.S. Department of Energy Office of Science. Hundreds of physicists from around the world use RHIC to study what the universe may have looked like in the first few moments after its creation. What physicists learn from these collisions may help us understand more about why the physical world works the way it does, from the smallest subatomic particles, to the largest stars.