Overview of the Silicon Detectors Operation at Cryogenic Environment*

,

(Ioffe Institute, 26 Politekhnicheskaya str., St. Petersburg 194021 Russian Federation)

Abstract：The increased luminosity and maximal particle energy in the upgraded Large Hadron Collider (High-Luminosity LHC) requires more precise data on the intensity of radiation field in the vicinity of the magnet coils of the collider. For that, the radiation sensors must be placed in the proximity of the coils, which minimizes the influence of debris on the beam loss. Silicon sensors were chosen as one of the candidates and the related R&D was started seven years ago jointly by the CERN BE-BI-BL group, CERN-RD39 collaboration and the Ioffe institute. The sensors developed at the Ioffe institute and in situ tested at cryogenic temperatures 1.9 K and 4.3 K surveyed irradiation by 23 GeV protons showing reproducible characteristics and giving new findings in the detector physics.

Key words：silicon detectors ；cryogenic temperatures；beam loss；in situ test

1 Introduction

One of the most important elements of modern accelerators is a magnetic system that precisely maintains a needed trajectory of the particles in the ion conductor. The existing accelerators produce bunches of particles with the energy up to 14 TeV as in High-Luminosity LHC, and therefore their magnetic systems operate in superconductive mode, which makes it possible to create magnetic fields sufficient for confinement of particles in orbits with a few kilometer radius and spatial accuracy of colliding events in the interaction points within ten micrometers. Safe operation of superconducting magnets is the most important complex task solved by on-line monitoring. One of the factors capable of creating failure in the accelerator operation is heating of the superconducting magnet coils by particles that escape from the beam (beam loss). Fig.1[1]shows an example of a problematic loss situation when an accidental beam loss on the beam screen can lead to almost direct energy deposition in the superconducting coil, local overheating and explosion of the magnet.

The LHC machine uses a monitoring system for the radiation field at its critical points shown in Fig.2[2]. In the existing system, the sensors are mounted on the outer surface of the magnet at a distance of 80 cm from its coil. Obviously, background radiation, constantly existing in the tunnel of the accelerator, limits the sensitivity threshold of monitors. The LHC upgrade program assumes the development and testing of silicon BLMs[3-4]installed in the proximity of the magnet coils.Below is a brief review of the latest results of studies of the radiation degradation of silicon monitors made in the scope of the development program for the LHC machine at CERN.

2 Experimental samples and measurement technique

In the investigations of Si detectors developed as BLMs the timeline and the sequence of the measurements atT=1.9～4.2 K were the following.

(1) 2012: characterization of nonirradiated Si detectors at 4.2 K (the test denoted as T0)[5];

(2) 2012: the firstinsituirradiation test (T1) aimed at getting the first data on detector radiation degradation at 1.9 K, and performed with detectors processed on wafers with a standard thickness (d) of 300 μm and different silicon resistivities[6-7];

(3) 2014: the secondinsituirradiation test (T2) aimed at the first comparison between the radiation degradations atT=4.1 K of detectors with standard and reduced (100 μm) thicknesses[8];

(4) 2015: the thirdinsituirradiation test (T3) atT=4.1 K focused at getting statistically significant data on the signal degradation in the modules of Si detectors with 300 and 100 μm thicknesses[9].

The silicon samples used in all the tests were planar p+-n-n+pad detectors designed and manufactured jointly by the Ioffe Institute, St. Petersburg, and the Research Institute of Material Science and Technology, Zelenograd, both Russia. The detectors were processed mainly on wafers with a resistivity of 10～15 kΩcm.

Insituirradiation tests were carried out at the CERN PS irradiation facility. The detectors were irradiated by a proton beam with an energy of 23 GeV. Monitoring of the beam position and the intensity was performed along the entire duration of the tests by the Beam Position Monitors (BMP) incorporated in the PS facility. The detector modules were mounted in a cassette installed inside a cryostat with superfluid or liquid helium. In addition to BMPs, two silicon beam telescope modules were used to control alignment of the cassette.

The experimental study of the detectors included the following measurements:① TheI-Vcharacteristics；② The current pulse response of the detector arising from the nonequilibrium charge generated by a spill；③ The collected chargeQcdetermined by integrating the current response over a spill duration(T2 and T3).

The parameters of 26 detectors studied in T3 and grouped in eight detector modules are listed in Tab.1.

Tab.1 Parameters of detector modules studied in the test T3

A compact arrangement with a minimum gap between the detectors in the modules resulted in diminishing the fluctuations of the proton beam intensity and of the accumulated fluence. Two of the modules labeled as “TeleIN” and “TeleOUT” also contained four samples each and were exploited as beam telescopes located near the entrance and exit planes of the cassette with installed modules. The reference detectors denoted as Ref1 and Ref2 were identical to those in the module MM1. These detectors were needed to compare the data recorded by different DAQ systems. Fig.3(a) shows a schematic cross-section of the detector irradiation typical of modules MM1～MM4. The cassette with the installed detector modules studied in T3 is presented in Fig.3(b).

Two versions of the DAQ system were used in T3. The first was the CERN-DAQ dedicated for application with standard BLM equipment. The second,Ioffe-DAQ, was the system elaborated at the Ioffe Institute, which included a 16-channel sampling unit with an incorporated ADC operated at the sampling frequency of 250 Hz and linked with the detectors from the modules MM1～MM4. The signal from the individual detector was amplified, digitized in the nonstop regime, and after being triggered, recorded by a PC. This system recorded the signals automatically all over the T3 duration and manually whiles the voltage scans were carried out.Monitoring of the beam parameters by BMPs showed that its intensity and position were not stable in time. The BPM data filtered from abnormal signals were employed to build a 3D image of the beam. With an account of geometrical and positional parameters of the detectors, the number of protons delivered by the spill and passed through the detectors was obtained. This number was used to recalculate the detector signal recorded by the DAQ systems into the charge generated by a single proton (mip) and collected in the detector. The error of the signal measurements in T3 was about10%.

3 Results of the in situ irradiation test T3

The results ofinsituirradiation test T3 and its analysis were presented in[10-12].In Fig.4 the signal vs. fluence (F) dependences are shown. The data gave statistical signal distribution for all detectors installed in the modules MM1～MM4 ((a)～(d), respectively) and operated at reverse biasVrev= -300 V. Here and in Fig.5 and Fig.6 the signal expressed as the charge collected per mip was determined by integrating the individual current responses from a single spill recorded by the Ioffe Institute DAQ system over the whole spill duration and normalized in respect of the corresponding number of protons passed through the detectors.

At a fixedVrevof -300 V the signal spread in the 300 μm thick detectors (Fig.4(a) and (b)) was higher than that in the 100 μm thick ones. The maximum spread was observed for the samples from the module MM2 processed on the 500 Ωcm silicon. It indicated presumably the problems with the reproducibility of the electric field distribution inside the bulk. On the other hand, the signal spread was minimal in the 100 μm thick detectors (Fig. 4(c) and (d)), especially in the module MM3. The specific feature of the curves shown in Fig. 4 is that for a majority of the detectors (13 from 16) signal reduction differed from the usual monotonic degradation with irradiation. The maximum signals of the detectors from MM3 and MM4 were shifted toF= 2×1014p/cm2while for the six detectors from MM1 and MM2 only a marginal signal increase was observed at this fluence. A similar shift was observed also in theQc(V) dependences for detectors irradiated within the fluence range 3.8×1013～5.5×1015p/cm2and operated atVrev.

The advantage in terms of charge collected in thinner detectors can be seen in Fig.5, where the voltage scan datas on the CCE were compared directly for 100 and 300 μm thick detectors at reverse and forward (Vforw) bias voltages. The values were calculated taking into account that the ratio of nonequilibrium charges generated by MIPs in the samples with 300 and 100 μm thicknesses,Qo(300)/Qo(100), equals three. At the lowestFof 3.8×1013p/cm2and absVrev≥150 VCCE(100) is about 50% higher thanCCE(300), whereas atF=1×1015p/cm2the ratioCCE(100)/CCE(300) increases and reaches 3 and 5～6 atVrevandVforw, respectively. A higher efficiency of thinned detectors is due to two factors: ① the mean electric field is higher than in the 300 μm thick ones, and ② even atF～ 1×1015p/cm2the carrier drift length is comparable with the detector thickness, which leads to lower trapping-related charge loss. It results in a lower rate of radiation degradation in thinner detectors, which agrees with the calculated data[10].

Finally, Fig.6 shows the silicon detector modules installed on the end of the cold mass containing superconductive coils of the magnets, for their operation as BLMs[10].

4 Summary

The results of theinsituirradiation tests showed that Si detectors can operate at 1.9 K after irradiation toF=1.1×1016p/cm2required for their application as BLMs. Operation in the forward bias mode was advantageous up toF～2×1015p/cm2. Thinned (100 μm) detectors provided a significantly higherCCE, a lower rate of signal degradation, and minimum spread of the sensitivity.It should be noted that the explanations of some features observed in theQ(F) dependences are not yet fully understood. Meanwhile, the results clearly illustrate that silicon p-i-n detectors can operate even at 1.9 K with high level of reproducibility of their characteristics. As a consequence of our work, silicon detector modules were installed on the end of the vessel containing superconductive coil of a magnet immersed in superfluid helium for their operation as BLMs in real experimental conditions[10-12].