Induced radioactivity in ATLAS cavern measured by MPX detector network

An application of the use of MPX detectors for the measurement of induced radioactivity in the Large Hadron Collider (LHC) experiment environment is presented. The sixteen ATLAS-MPX detectors measured the associated photon ambient dose equivalent rate (H*(10) rate) in ATLAS due to the LHC operation. It is presented that in 2010 the average H*(10) rate determined with all MPX detectors varies around 0.015 μSv/h in the ATLAS cavern. In 2013, roughly after 50 days from the end of high luminosity physics runs, the highest H*(10) rate measured with the MPX detectors has reached 14.9 μSv/h (in inner detector region). All the other dedicated MPX detectors measured H*(10) rate which reached the value up to 1.3 μSv/h, depending on the MPX detector.


Introduction
A network of sixteen MPX pixel detectors (ATLAS-MPX network) based on the Medipix2 chip [1] was installed within the ATLAS experiment at CERN, see page 50 in [2]. Their location is illustrated in figure 1 with their coordinates given in table 1. A detailed description of the ATLAS-MPX detectors network and operation can be found in [3] and [4].
The present paper reports an example of application of MPX detectors. It consists of the measurement of the photon ambient dose equivalent rate (H*(10) rate) performed with the MPX detectors during the period before the start of LHC collisions in 2010 and after the LHC shutdown in 2013. The number of interacting photons and electrons in the detectors was determined from the amount of measured clusters/tracks produced by corresponding radiation quanta in the pixelated sensor. The cluster rate was recalculated into H*(10) rate utilizing a conversion coefficient determined from an MPX detector of reference exposed to gamma ray sources ( 60 Co and 137 Cs) of known H*(10) rate. The value of the conversion coefficient was confirmed by two independent methods: by MCNPX™ Monte Carlo simulation modelling the experimental set-up of reference and by comparing the MPX detector response to natural radiation background in Prague with the reading of a calibrated dosimeter.

MPX detectors
The utilized MPX device consists of a 300 µm thick silicon sensor matrix of 256 × 256 cells (matrix element 55 × 55 µm 2 ) bump-bonded to a pixelated read-out chip. Each MPX device is fitted in a duralumin box with an entrance window located above the sensor. The network detectors were operated in tracking mode with low threshold (8-10 keV). The signature of particles interacting -1 - in the silicon layer is visualized in the form of a cluster of adjoining activated pixels (a track), similarly to nuclear emulsion with different size and shape depending on the type of particles, their energies, incidence angles, and the nature of their interactions in the sensor. The remotely settable integration time window (depending on the luminosity and the MPX position in ATLAS) was set short enough to avoid overlap of tracks created by the individual particles. The integration times were in the range from 10 ms to 30 s. Data collected in the tracking mode were analyzed with a pattern recognition algorithm [3] and subsequently stored for further evaluation.

The method
The MPX detector does not measure directly the energy deposited by the particles. However, based on the cluster shape analysis it can provide the composition of the radiation field, i.e., the particle types and their approximate energy ranges [3]. Then, it can measure the radiation field composed of natural radiation background and LHC induced radioactivity (due to decay of radioactive nuclei generated in the ATLAS environment during LHC collisions), see pages 51-52 in [2]. After an LHC collision period, this radiation field primarily consists of photons and electrons with prevailing energy between few hundreds of keV and few MeV. Figure 2 shows examples of measurement of such radiation fields with MPX detectors between 6 and 9 December 2012. The interpretation of the measurement of these induced radiation fields is simpler than that of the complex radiation field generated during LHC collisions [3,4]. Using a calibration with standardized photon beams in that energy range, it was possible to calculate the H*(10) rate from the measured cluster rate.
-2 - Table 1. MPX detector locations and positions with respect to the central interaction point. X, Y and Z axes correspond to the standard ATLAS coordinate system, R = (X 2 + Y 2 ) 1/2 is the distance from the beam axis at position Z. The approximate orientation of the devices is given with respect to the beam axis (Z-axis) with the sensitive surface of the detectors facing the interaction point. Z determines beam direction, positive sign corresponds to side A, negative to side C, Y -positive sign corresponds to up direction, negative to down, X -positive corresponds to orientation toward the center of the LHC ring, negative toward U.S.A. 15 (outside the ring).

MPX detector calibration in reference photon beams
It has been experimentally confirmed that the MPX detector response to photons and charged particles is consistent from detector to detector within 20% [3]. Therefore, only one reference device was exploited for the H*(10) rate calibration. For the calibration, the standardized 137 Cs and 60 Co photon beams (divergent collimated fields of point-like sources) available in the Czech Metrology Institute in Prague were used. The energies of photons emitted in the process of the radionuclides disintegration are observed to correspond roughly to the mean energies of natural radiation background and expected LHC induced radioactivity in the ATLAS environment. During the calibration procedure, the detector was positioned perpendicularly to the 60 Co and 137 Cs photon beam axis, 300 cm from the source, and irradiated from the front side and then from the back side. To keep calibration conditions close to the ATLAS radiation conditions, the detector was calibrated without its duralumin box so the electrons created in the air and surroundings could contribute to the measured signal as well, in addition to electrons created directly inside the sensor -3 -

JINST 14 P03010
chip. To obtain the calibration coefficient for 60 Co and 137 Cs, the reference H*(10) rate of the corresponding photon beam at the MPX detector calibration position was divided by the rate of all clusters registered on the whole area of the sensor.
Monte Carlo simulations were performed in order to cross-check the measurement results. The geometry of the experiment, including the shape and composition of MPX devices [3,4], was modeled in the MCNPX™ code [5]. The simulations were performed for both photon beams and for a detector without a duralumin box (as in the calibration experiment), with a duralumin box (as in ATLAS), and with a duralumin box without an entrance window in the box (with the aim to estimate differences on detector sensitivity with entrance window oriented either to open space or to wall of box fixation). Table 2 presents the comparison between measured and simulated values of the conversion coefficient. It can be concluded from table 2 that the value of the calibration coefficient is slightly dependent on both the orientation of the detector and photon energy. However, the differences between measured and simulated coefficients as well as their variation are within 15%. Therefore, we chose to set the calibration coefficient converting the cluster rate into H*(10) rate to the value of 2.0 × 10 −10 Sv/cluster. Assuming that the induced radiation field in ATLAS is close to be isotropic and taking into account the MPX detector energy-angular detection efficiency for photons [3], one can expect that the angular dependence of the conversion coefficient is also well within 15%. The value is valid for the MPX detectors in tracking mode of operation at low threshold, and for the cluster rate obtained from the whole area of the sensor.
The difference observed between simulated and measured conversion coefficients could be caused by an over-simplification of the model, which neglected scattering objects, like irradiation hall walls and floor or a detector holder. Also, the discrepancy could be caused by the difference between the real photon fluence spectra and the ones utilized in the simulations. Taking into -4 -account the over-simplifications, the agreement between simulation and measurement is satisfactory. Therefore, one can conclude that the simulation confirmed the value of the conversion coefficient determined by the experiment.
It can be also concluded that the conversion coefficient does not significantly depend on the cover of the detector. The difference between the conversion coefficient for a detector without a box, inside a box, or inside a box but without an entrance window is lower than 10%.

MPX detectors calibration in the environment of natural radiation background
To certify the calibration, the simultaneous measurement of the H*(10) rate of natural radiation background was performed inside the Institute of Experimental and Applied Physics (IEAP) in Prague during August 2013 with MPX03, MPX08, MPX10, MPX12, and MPX16 (taken out of the ATLAS cavern for this purpose) and with two certified electronic dosimeters UltraRadiac™ (Canberra). Table 3 summarizes the comparison between the H*(10) rates of natural radiation background as measured by the MPX detectors and the UltraRadiac™ dosimeters. A very good agreement was found between all MPX detectors and both UltraRadiac™ dosimeters. The H*(10) rates agree within 3%, which is at the level of statistical uncertainty on the measurements and fluctuations of the natural radiation background. The comparative experiment proved that the calibration coefficient of 2.0 × 10 −10 Sv/cluster is valid also for the natural radiation background at ground level.

LHC induced radioactivity
The similarity between the photon calibration field and the natural radiation background with LHC induced radioactivity field has to be demonstrated for justifying the use of the value of the conversion coefficient stated above. Figure 3 presents an example of MPX frames acquired under exposure of a) the MPX detector of reference to 60 Co photon beam, b) the MPX03 to 137 Cs source during the Tile Calorimeter inter-calibration in ATLAS on 3 August 2010 [3,6], c) the MPX03 to natural radiation background at Prague on 8 August 2013, and d) the MPX03 to LHC induced radioactivity on 17 February 2013. The frame length (acquisition time) was adjusted according to a count rate to -5 -minimize the number of overlapping tracks in frame. For the calibration with 60 Co, as an example, the frame length was set to 0.1 ms at dose rate 0.1 Sv/h and 1 s at 10 µSv/h. The structure and the distribution of lengths of clusters in all frames are quite similar.

Figure 2. Example of LHC induced radioactivity as measured with MPX detectors between 6 and 9
December 2012. Cluster rates in each MPX detector are multiplied by the value N for data scaling. MPX14 data do not represent induced radioactivity but referred to as HETP -High Energy Transfer Particles in our previous papers ( [3,4]) generated during either beam collisions or some beam operations, as this detector was set at high threshold [3], being insensitive to LETP (Low Energy Transfer Particles [3,4] Table 4 summarizes the fractions of different types of clusters in calibration data and radiation background measurements. A measurement with an 241 Am source emitting 59.5 keV photons is included in the table to show the ratio of clusters that resulted in low energy photon field. Comparing the data in table 4, it can be said that the LHC induced radioactivity field in ATLAS after 50 days of "cooling", after the end of high luminosity physics runs, is similar to the natural radiation background. It is also similar to 137 Cs photon field but it includes slightly more dots (clusters of one pixel), i.e., more low energy photons (tens of keV) are detected. One can conclude that the spectral composition of LHC induced radioactivity in ATLAS is rather close to that of natural radiation background which justifies the use of the value of the calibration coefficient stated above for determination of H*(10) rate caused by LHC induced radioactivity. Table 5 presents the H*(10) rates caused by the natural radiation background and LHC induced radioactivity in the ATLAS cavern, as measured by the MPX detectors in February 2013, roughly after 50 days of "cooling". For comparison, data from the beginning of February 2010, before the -6 - LHC collisions started, are given as well. The ATLAS recorded total integrated luminosity in 2012 was 21.79 fb −1 [7].

Results and discussion
The H*(10) rates measured by MPX detectors in 2010 are consistent between each other and they correspond to expected level of natural radiation background in ATLAS cavern located 92 m below the ground (cavern floor level; the cavern height is 35 m [8]). The H*(10) rates are about one order of magnitude lower than those at the ground. The higher H*(10) rate in MPX10 and MPX11 compared to other MPX detectors is most likely caused by natural radioactivity in the concrete because these detectors were mounted onto a concrete wall. Other slight variations in the detector -7 - Table 4. Distribution of diverse types of clusters measured in different radiation fields dominated by photons and electrons, as obtained from cluster pattern recognition analysis [3]. Dots: cluster of one pixel; Small blob: cluster of 2-4 pixels; Curly track: cluster with a shape of a curly line of more than 4 pixels in length. There have been several Monte Carlo studies [9,10] and another direct measurement of the induced radioactivity in ATLAS [11]. Table 5 reports the H*(10) rates as measured by several MPX detectors in February 2013, approximately after 50 days of "cooling". However, a sensible comparison with the results of [12] was not possible as both types of detectors were located in different positions and the amount and composition of material in between and around them was different. The locations of MPX detectors in ATLAS system are reported in table 1.
As an example, the decay of the LHC induced radioactivity with short half-life within a few days is presented in figure 4. The MPX03 cluster rate per frame and the corresponding H*(10) rate is shown for a time period of 5 days starting just after the end of LHC collisions on 6 December 2012.

Conclusions
The induced radioactivity and associated photon ambient dose equivalent rate (H*(10) rate) in ATLAS due to the Large Hadron Collider (LHC) operation was obtained from measurements with ATLAS-MPX silicon pixel detectors (MPX detectors). It was demonstrated that the ATLAS-MPX network, consisting of 16 MPX detectors, delivers on-line information about radiation level across the whole ATLAS detector during and after LHC collision periods. In 2013, roughly after 50 days of "cooling", the highest H*(10) rate measured with MPX detectors dedicated to radioactivity measurement inside the ATLAS detector was obtained with MPX01 located between the JM plug shielding and the inner detector reaching 14.9 µSv/h. For all the other dedicated MPX detectors, the H*(10) rate reached a value up to 1.3 µSv/h, depending on the detector position.
The currently installed ATLAS-TPX network [13] (upgraded version of ATLAS-MPX network) has also the capability, as the MPX detector network, to determine induced radioactivity in a broad -8 - range of decay half-lives, from minutes to years (table 5, reference [3]). The capability of MPX and now TPX network can be utilized for instance to predict growth and decay of the LHC induced radioactivity based on the measured data a) to estimate its contribution to ATLAS radiation field during LHC collisions and b) to predict the H*(10) rate any time ahead according the expected LHC schedule.