PERFORMANCE OF THE CMS ECAL LASER MONITORING SOURCE IN THE TEST BEAM

Transcription

1 PERFORMANCE OF THE CMS ECAL LASER MONITORING SOURCE IN THE TEST BEAM A. BORNHEIM CALTECH 2 E. California Blvd., Pasadena, CA 925, USA bornheim@hep.caltech.edu On behalf of the CMS ECAL Collaboration. The CMS detector at LHC will be equipped with a high precision lead tungsten electromagnetic calorimeter, the CMS ECAL. In 23, the laser source for the monitoring of the CMS ECAL was used in its final design for several months in a beam test of a CMS ECAL module at CERN. The laser source consists of two different laser systems providing laser pulses at four different wave lengths plus one back-up system. We report on the experience with this device and present results on the performance and stability of the system over a time period of several months. The achieved performance is discussed in view of the design goals for the CMS ECAL.. Introduction One of the main physics goals of the CMS detector at LHC is the search for the Higgs boson. In the mass range around 25 GeV the decay H γγ is of particular interest because of its clean final state. This mode however requires a very good energy resolution for the electromagnetic calorimeter to take advantage of the expected narrow decay width of the Higgs boson in this mass range. To fulfil this requirement, the CMS experiment features a high resolution crystal calorimeter utilizing 8 lead tungstate (PWO) crystals. Extensive RD efforts lead to the development of radiation hard lead tungstate crystals 2. These crystals do experience only a very slight, dose-rate dependent decrease of their transparency under irradiation which recovers to a large extend in irradiaton free periods. In CMS, the variation of the crystal transparency under irradiation and the corresponding decrease in light output of the crystals will be monitored with a high precision light monitoring system. In this paper we describe the experience with the final design laser source for this monitoring system in the test beam at CERN, operating on a close to final design CMS ECAL barrel module.

2 2 2. The CMS ECAL monitoring system at the CERN H4 test beam area. The CMS ECAL monitoring system consists of a laser based light source (LS) 4, shown in Figure, and a light distribution system (LDS) which distributes the light to the individual PWO crystals, which are read out by avalanche photo diodes (APD), and to reference PN diodes. Ext. Trigger NET - RS232 Ti:S Controller Digital Delay Monochromator Osc. Digitizing PC Diagnostic Ti: Sapphire Main FS ( 44 or 495 nm ) Nd:YLF (527 nm) Diagnostic Bus Extender Bus Extender 3-5 m RS232 Ti:S Controller Ti: Sapphire ( 44 or 495 nm ) Ext. Trigger Digital Delay Nd:YLF (527 nm) Diagnostic Monochromator Main FS Osc. Digitizing FC 2 x Optical Switch o o o o o o CAMAC Multi-Channel ADC o o o o o o x 8 Monitoring Optical Switch Box o o o o o o Level Two Fanout Figure. Schematic view of the laser source for the CMS ECAL monitoring system. The light source has to meet a number of performance parameters to ensure a precise monitoring of the radiation damage in the PWO crystals. Laboratory test have shown that the radiation induced absorption can be measured with best linearity and adequat sensitivity at a wavelength of 44 nm (blue) 3, close to the maximum of the PWO emission spectrum. This wavelength has therefore been chosen as the primary monitoring wavelength. As a systematic cross check light at 495 nm (green) can be used. The radiation induced loss of transparency is mostly restricted to wavelength below 7 nm. Thus, light at longer wavelength can be used to monitor the entire chain from the LDS to the crystals, the APDs and the electronics. For this purpose light at 7 nm (red) and 8 nm (infrared) is provided. The laser light is pulsed with a pulse width of less than 4 ns to match the shaping time of the electronics. The light intensity has to be around mj per pulse to have sufficient light injected into the LDS to probe a large portion of dynamic range of the ECAL up to the equivalent of TeV, taking into account all losses of the LDS. In CMS sets of up to 7

3 3 crystals, corresponding to the structural sub-unit of the barrel part of the ECAL called supermodule, will be illuminated by individual laser pulses simultaneously. The pulse to pulse energy fluctuation should be less than % to minimize possible non-linear effects in the ratio of the signal from the APDs and PN diodes. The time jitter of each pulse is less than 3 ns to ensure precise timing with respect to the LHC bunch crossings, spaced 25 ns apart. It was decided to use a two stage laser system for this task. A YLF:Nd pump laser which drives a Ti:S laser, providing blue and green light and an independent second system, also a YLF:Nd pumped Ti:S laser, which provides red and infrared light. The wavelength for each system can be selected with a movable birefringent filter inside the Ti:S laser. The selection of red and blue system is done via an optical 2x switch. The choice between the subsets of ECAL modules to be illuminated will be done with an x8 optical switch. This entire system is currently installed in its final design at the H4 test beam area at CERN 6. The experimental area at CERN is equipped with a temperature stabilized experimental hall which houses a movable table onto which ECAL supermodules can be installed and positioned in the beam with a precision of better than mm. The air temperature in the hall is stable to approximately ±.5 C o. In 23 a final design bare supermodule with close to final design electronics was tested for several month in the beam. The crystals and the APDs of the supermodule are temperature stabilized with a precision water cooling system to approximately. C o. The laser systems as well as the x2 and the x8 optical switches are housed in a separate barrack to ensure a stable and clean environment. The laser pulses are sent via an optical fibre to the ECAL supermodule. The entire setup, including the fibre length from the laser barracks to the ECAL module, mimics the environmental conditions envisioned for the final installation in the CMS underground cavern. 3. The Performance of the Laser Source in the 23 Test Beam. In the 23, a variety of tests were performed on the ECAL module such as energy and position resolution determination 5, long term stability of the ECAL and irradiation test using electron and pion beams to induce radiation damage. Throughout the entire period, the laser system was used to monitor the ECAL module and measure the radiation damage, in total more of 2 hours of operation. During this period the laser system exceeded its design performance by a significant margin. To quantify the

4 4 short term performance we quote the spread of the pulse energy, pulse width and pulse timing distribution over a period of 3 minutes, the time it will take to scan all crystals on the full ECAL in CMS during LHC operation. We quote a medium term stability over a period of 25 hours, about two times the duration of a LHC run, which will last about 2 hours and is referred to as one LHC cycle. For the long term stability we quote the mean pulse energy degradation per day. This long term stability is mainly driven by the degradation of the pump lamp of the YLF:Nd laser which has a lifetime of hours. A significant increase in the long term pulse degradation can also indicate damages in the optical components of the laser system. The short term stability is driven by the laser system design, the medium term stability by the laser barrack environment. The performance achieved in 23 are summarized in Table. Table. Stability of the CMS ECAL laser source in the 23 beam test. Stability Pls. Hgth. Pls. Hgth. Pls. Wdth. Pls. Jter. Puls. Degr. [3min] [25h] [25h] [25h] 44 nm.5 % 3.2 % 2.7 % 2.8 ns.4 % day 495 nm 7.6 % 6.8 % 7 nm 8.2 % 5.7 % 8 nm 2.8 % 3.2% 2.6 %.2 ns.6 % day nm 8 nm nm 7 nm nm 8 nm. 44 nm 8 nm Normalized Pulse Energy.7.3 Normalized Pulse Energy 25 3 Pulse Width [ ns ] Pulse Jitter [ ns ] Figure 2. Typical spectra for the normalized pulse width for all four wavelength (left and centre left), the pulse width (centre right) and the pulse jitter for 44 nm and 8 nm (right). As can be seen from the values in the table and the spectra in Figure 2 the performance in terms of the pulse stability is very similar for the 44 nm and the 8 nm setting of the corresponding laser systems. For the pulse width and the pulse jitter, the 44 nm laser performs slightly better than the 8 nm system while the 495 nm and the 7 nm, both corresponding to a higher harmonic mode of the respective laser cavities, are slightly less performant because they are off the emission spectrum peak of Ti:S. During the beam test monitoring data has been recorded with the 44 nm and the

5 5 8 nm laser throughout the entire test period, while data with 495 nm and 7 nm has only been recorded for a limited period of time to study systematic effects. In Figure 3 the long term trend of the mean laser pulse energy is shown, illustrating the slow decrease of the 44 nm system as expected due to the pump lamp degradation. The 8 nm system shows a much slower decrease because it is operated at a lower pump current. This is of particular interest for studying the stability of the ECAL. Int(NORM).5 Int(NORM) nm.85 8 nm T [hours] T [hours] Figure 3. The long term stability of the 44 nm and the 8 nm laser system over a period of 4 hours. 4. Examples of Monitoring Performance of the CMS ECAL. As shown above the performance of the laser source exceeded the design performance by a significant margin. This allowed a detailed testing of the higher level monitoring system and the monitoring procedure of the radiation induced transparency loss of the crystals itself. In Figure: 4 we show a comparison of the mean laser pulse height at 8 nm recorded by the laser source internal monitor, compared to the mean response of one ECAL channel over a period of 2 hours. Both data sets have not been corrected off line in any way. As can be seen from the plot the two curves track each other within a few tenth of a percent which shows that the entire light distribution system and the ECAL module itself, including the lower level light distribution system inside the laser barracks, is very stable over many LHC cycles. This offers the possibility to use the laser internal monitoring system as a cross check independent of the entire ECAL installation. This is of particular interest since the laser system will not be located in the main CMS cavern but in the adjacent electronics cavern, thus having very different environmental conditions, in particular much lower radiation levels. It should be pointed out that in the CMS ECAL only the ratio between the signals of the APDs and the PN reference diodes are used to track the radiation damage. This ratio is stable to better than. %.

6 6 rel. Intensity ECAL Laser Runs (averaged, Xtal 67) Laser Monitor 8 nm % T [ hours ] Figure 4. Comparison of the light source monitoring and the mean response as seen by the ECAL, averaged over time bins. The plots are normalized to each other. The long term fluctuations of the absolute laser pulse energy is on the order of %, indicated by the arrow at T=45 hours. The LS monitoring and the absolute ECAL response track each other to within a few tenth of a percent. One of the main goals of the test beam was to prove the feasibility of the radiation damage monitoring with the desired precision. This was done by irradiating crystals in a high energy electron beam with dose rates similar to the ones expected in CMS. During the irradiation the crystals are calibrated with the electron beam in regular time intervals, followed by laser monitoring runs. The loss in the electron beam signal response can then be correlated with the loss in the signal response to laser shots, which is shown in Figure 5. Figure 5. Correlation between the relative loss of the signal response to a fixed energy electron beam and the laser monitoring signal. The slope α as shown in the plot is used to characterize the correlation. Figure 6. The distribution of the parameter α for 9 PWO crystals measured in the 22 beam test. The spread of this parameter was found to be 6.3 %. This result was confirmed in 23. The plot exhibits a clear linear relationship. The relative loss in the laser response can thus be used to correct the observed reduction of the electron signal response. It is not foreseen to measure the parameter α, which characterizes the correction function, for each crystal of the CMS ECAL. As can be seen in Figure 6 the spread of α was measured to be 6.3 % in

7 7 a sample of 9 crystals, which is sufficient for the average signal loss for CMS-like radiation environment which is expected to be around 5 %. The change in the ECAL response during normal LHC operation might however be smaller than this since the recovery time of the crystals is longer than one LHC cycle. 5. Summary. The final design laser source for the CMS ECAL monitoring system has been operated in the test beam at CERN for more than 2 hours in 23. It has exceeded its design performance in terms of stability by a significant margin. The precise tracking of radiation induced transparency loss in a PWO crystal calorimeter with a laser based monitoring system has been proven to work in the test beam. Further beam tests in 24 and possibly 26 will be performed to fine tune the details of the monitoring procedure. The laser system presented here will be transferred to the CMS underground cavern towards the end of 26 to be ready for CMS startup in 27. Acknowledgements I would like to thank the entire ECAL collaboration for their outstanding effort, in particular R.-Y. Zhu for guidance and fruitful collaboration, and D. Bailleux, L. Zhang and K. Zhu for their superb work on the laser system. I also thank the organizers of CALOR 24 for an interesting and pleasant conference. References. J. Fay, Status of the CMS Electromagnetic Calorimeter, these proceedings. 2. CMS collaboration, The Electromagnetic Calorimeter Project Technical Design Report, CERN/LHCC (997); F. Cavallari, Performance of the PWO Crystals of the CMS ECAL, Proc. of the th Int. Conf. on Cal. in Part. Phys. 22; further references in 3 3. X. Qu, L. Zhang,R.-Y, Zhu, Radiation Induced Color Centers and Light Monitoring for Lead Tungstate Crystals, IEEE Trans. Nucl. Sci. NS-47 (2). 4. L. Zhang, K. Zhu, R.-Y. Zhu, D. Liu, Monitoring Light Source for CMS Lead Tungstate Crystal Calorimeter at LHC, IEEE Trans. Nucl. Sci. NS-48 (2) I. van Vulpen, Pulse Reconstruction in the CMS ECAL, these proceedings. 6. D. Bailleux, A. Bornheim, L. Zhang, K. Zhu, R.Y. Zhu, ECAL Monitoring Light Source at H4, CMS Internal Note,IN-23/45.