Summer Student project report Mika Väänänen September 1, 2017 Abstract In this report I give a brief overview of my activities during the summer student project. I worked on the scintillating fibre (SciFi) tracker, which will replace LHCb s current IT/OT tracker. More specifically, I worked with quality assurance of production phase fibres and calibrated an instrument used to determine radiation doses received by the fibre samples. 1 Introduction During the Long Shutdown 2 in 2019-2020, the particle tracking system in LHCb will be replaced to handle the higher luminosity and higher readout frequency. The Level-0 trigger will be removed and events will be read out at the full collision frequency of 40MHz. LHCb will also take data up to an integrated luminosity of 50 fb 1. The current downstream tracker is a combination of two technologies. The section of the tracker closer to the beam pipe (inner tracker) is a silicon microstrip tracker and the section further from the pipe (outer tracker) is a straw-tube drift tracker. The whole downstream tracker will be replaced with a scintillating fibre tracker [2], composed of 2.5 meter long fibres as the active elements of the tracker. 1.1 Scintillators Scintillators are materials that produce light when exposed to ionizing radiation. Scintillators are divided into two categories, organic and inorganic scintillators. Organic scintillators are based on excitation of the constituent molecular energy levels, whereas inorganic crystalline scintillators produce light through excitation and de-excitation of energy levels in the crystal band structure. Inorganic scintillators are made of dense materials like lead and tungsten and organized into a crystal structure. Inorganic scintillators have high light yield (photons per MeV deposited) and are often used in, for example, calorimetry and medical imagining. Organic scintillators are carbon-based materials, usually containing a benzene ring which acts as actual scintillator. The light yield and radiation hardness are usually worse than in inorganic scintillators, but organic scintillators are much more affordable, so they can be used in large-scale projects like the SciFi tracker. 1
2 LHCb tracking detectors The LHCb particle tracking system is formed by five tracking stations: the Tracker Turiciensis (TT), located between the interaction point and the dipole magnet and three downstream tracker stations T1-T3 between the magnet and RICH2 and the Vertex Locator VELO, located before the magnet and TT. During the Long Shutdown 2 the downstream tracker will be replaced by a SciFi tracker, and thus this section mostly discusses T1-T3 tracking stations and their replacement. 2.1 SciFi Tracker The IT/OT structure of each tracking station will be replaced by a scintillating fibre tracker. The geometry of the tracking stations will be similar to the current tracker: there will be three stations, each consisting of four detection layers of which two are rotated by a ±5 stereo angle to allow y-coordinate measurement of the passing particles. The detection layers are made of fibre mats which in turn are made of six layers of 250 µm diameter fibres with a 275 µm pitch. The fibres are of type SCSF-78, produced by Kuraray. Figure 1: Schematical view of LHCb after the upgrade. The SciFi tracker is in the center, between the magnet and RICH2. SciFi tracking stations T1 and T2 will consist of 10 modules and T3 of 12 modules, resulting in total 128 modules. A single module is 5 meters in height and 0.52 meters in width, resulting in final dimensions of 5 m by 6 m in the x-y plane of the detector. The modules are made of eight six-layer fibre mats 2.5 meters in length. The mats will have mirrors at the joining ends to maximize light output and silicon photomultipliers at the outer ends that serve as readout 2
detectors. The tracker will have a single hit spatial resolution of 100 µm or less. To accommodate the beampipe structure, all tracker layers will have a cutout in the middle. The SciFi tracker geometry is described in figure 2. Figure 2: Geometry of the SciFi tracker. The stereo angle layers are visible in blue and green. 3 Project activities On average about every two weeks, the SciFi quality assurance lab receives a batch of 24 fibre spools, 12.5 km each. A total of 11000 kilometres of fibre will be produced. The fibres are tested for six characteristics: attenuation length, light yield, fibre diameter, cladding integrity, radiation hardness and minimal bending radius. My main project was calibrating an optical densitometer, which is used to determine the radiation dose received by the fibre samples during the radiation hardness test. A piece of X-ray film is irradiated along with the fibre and the optical density (i.e. darkness) of the film is linearly proportional to the received dose. I also designed a new measurement setup to measure the minimal bending radius of the fibres and carried out other fibre QA procedures as described in the following. 3
3.1 SciFi quality assurance Attenuation length describes the attenuation of light travelling in the fibre. A sample of each spool is measured by exciting the sample with UV light at different points and measuring the light output at the end of the fibre for each excitation point. The results are then fitted with an exponential: I(x) I 0 = e x Λ, (1) where I(x) is the intensity of light as a function of distance, I 0 is the intensity at zero distance, x is the distance from the detecotr and Λ is the attenuation length. Light yield is measured by exciting the fibre samples with 1 MeV electrons at different points to trigger the whole scintillation process, measuring the light output for each excitation point, fitting with an exponential to the measured light outputs (due to the attenuation of light being exponential, described in eq. 1) and extrapolating to zero distance. Each spool is sampled. The diameter of each fibre spool is also measured. The whole spool is run through a custom built scanner, which records the diameter and the locations of diameter defects. The scanner also includes a tool to shrink moderate diameter defects. Also, cladding defects are recorded. Radiation hardness refers to the fibres resistance to radiation. Three spools per batch are tested. A fibre sample is irradiated to around 1 kgy of X-ray radiation and the attenuation lengths before and after the irradiation are compared. The dose rate is measured attaching a dosimetric Gafchromic HD-810 film to the fibre sample for the first 10 minutes of the measurement. The film is taken out and the fibre sample is irradiated for 35 minutes more, resulting in a 45 minute total irradiation time. According to the manufacturer, the dose of the film as a function of the film s absorbance is D = 125 0.72 (A A 0), (2) where D is the dose, A the absorbance and A 0 the absorbance before irradiation ( background optical density ). The factor 0.72 is a correction due to the radiation type. [1] If the fibre is bent with a radius smaller than the minimal bending radius and kept at the radius for a few days, it will start to produce microcracks and increase light losses. The manufacturer has given the minimal bending radii for fibres of other diameters, but not for the 250 µm fibre. For example, for a 1 mm fibre the minimal bending radius is 50-100mm and for a 500µm fibre 25-50 mm, so for the 250 µm the minimal bending radius is expected to be around 13-25 mm. 3.2 Densitometer calibration Before this summer, a commercial densitometer borrowed from CERN radiation protection was used to analyze the absorbance of the X-ray dosimetric film. This was not optimal, since several groups were using the same densitometer and thus it is only available for a day or two at a time and radiation hardness measurements had to be scheduled accordingly. The densitometer works by 4
shining a red light of known intensity through the film and then measuring the transmitted intensity at the other side. The optical density or absorbance A is i Φe = log10 T, (3) A = log10 Φte Φi where T = ( Φte ) 1 is the transmittance of the film, Φie the incident intensity e and Φte the transmitted intensity. A new, custom made densitometer built at the CERN workshop works in a very similar way (see figures 3 and 4). The instrument consists of two halves: bottom half containing a red LED (NTE30034, peak wavelength 660nm) that works as the light source and the top half containing a PIN diode (BPX65) to read out the light intensity. If a film is placed in the middle, the intensity of the light on the PIN diode, and thus the produced current, drops according to the transmittance of the film. Figure 3: On the left: new densitometer. The gold colored connector on top is for the PIN diode. The LED connector at the bottom is not visible. On the right: sketch of the new densitometer showing most important components. 5
Figure 4: Inside view of the densitometer. On the left is the bottom half with the hole of the LED in the middle. On the right is the top half, with the PIN diode visible in the center. My task was to calibrate the new instrument and create a user interface using LabVIEW. The calibration was done measuring the transmittance of a number of films irradiated with different doses with the commercial densitometer. The films were then placed in the new densitometer and the current produced by the PIN diode, dependent on the transmittance of the films, was recorded. Since the currents measured were in the order of hundreds of na, even small fluctuations in the intensity produced by the LED created large errors in the calibration. Thus, instead of relating the current directly to the optical density, the ratio of the photocurrent at a certain transmittance and so called zerodensity current was used: I(T ) R=, (4) I(1) where I(T ) is the current as a function of transmittance T and I(1) is the photocurrent without a film in the densitometer (transmittance practically one, or optical density practically zero, thus the name zero-density current). This proved to be a reliable measure giving accurate results over the whole optical density range used. The relation was found to be quadratic: T (R) = ar2 + br + c (5) The coefficients a, b and c were found by fitting a polynomial to the data. Three datasets were collected with different voltages of the densitometer LED. The fits, which can be seen in figures 6, 7 and 8, were found to be very similar between all three datasets, proving that the LED and the PIN diode worked in a non-saturated regime. The final calibration fit uses the average of the parameters over the three measurements, and the relation between transmittance and current ratio becomes T (R) = 0.445R2 + 0.581R + 0.0007 The results of all calibrations are presented in table 1. 6 (6)
a b c 1.5V 0.464 0.542 0.006 1.6V 0.440 0.566 0.004 1.7V 0.430 0.596-0.008 Table 1: Fit parameters for all different measurements. Figure 5: Photocurrent as a function of the luminous power incident on the photodiode. Efficiency of 40% is assumed for the LED. The incident power should be in the order of milliwatts to achieve linear response. The T-R relation was originally expected to be linear. This deviation from the original expectation may be caused by a few factors. One possible explanation is that the photodiode is operating in a non-linear regime. The incident intensity on the diode when measuring a film is very low (total power of the order 0.05 mw), and so it may be close to the diode noise floor. Some tests with a more precise current sourcewere done to determine the linear regime of the PIN diode, i.e. whether it was saturated during the calibration and whether the diode was working in the non-linear regime. The LED was supplied with different currents ranging from 0.09 ma (minimum current needed to run the LED) to 20 ma, the absolute maximum rated current, at a constant voltage of 2 volts and the photocurrent was measured. The PIN diode didn t show any signs of saturating even at the highest LED power. At the lower end of the current range, the photodiode showed some non-linear behavior, as shown in figure 5, which may explain the non-linear response of the densitometer. 7
Since the densitometer has so far been using a Keithley sourcemeter limited to 2 ma output current, the possibility of finding a more suitable power supply is being investigated. To rule out the poor calibration of the commercial densitometer, the transmittances of the films were measured with a dedicated instrument designed to measure transmittance over a wide wavelength band. The transmittances seem to be in accordance with those measured with the commercial densitometer. Figure 6: Calibration done with V LED = 1.5V 8
Figure 7: Calibration done with V LED = 1.6V Figure 8: Calibration done with V LED = 1.7V 9
The LabVIEW program serving as the user interface measures the current ratio R, calculates the transmittance using the fit described in equation 6 and then converts the transmittance to optical density according to equation 3. The interface is displayed in figure 9. The LabVIEW program calculates the dose rate received by the dosimetric film, and from this and the total irradiation time, the total dose received by the fibre. Figure 9: LabVIEW interface to the densitometer. Zero-density current and photocurrent are I(1) and I(T ) in equation 4, respectively. Transmittance and optical density are calculated from these according to equations 3 and 6. Irradiation time is the irradiation time of the film, total irradiation time is the total irradiation time of the fibre sample and background optical density is A 0 of equation 2. 10
4 Calibration results A A The accuracy of the measurement, calculated with = refarefmeas, where Aref is the optical density measured by the commercial densitometer and Ameas the optical density measured with the new instrument, is on average 5 percent. The error of the calibration measurement can be estimated as the standard deviation of the measured optical density between measurements with different LED voltages. The standard deviation was between 0.001 and 0.1, for Aref = 0 and Aref = 1.575 measured with the commercial densitometer, respectively. Standard deviations for each fit parameter are a = 0.014, b = 0.022 and c = 0.006. The negative sign of the fit parameter c in the 1.7V measurement explains the large standard deviation of the parameter. 5 Bending test The bending test utilizes the light yield setup. A glass tube bent to a circle with a diameter of 30 mm is placed between the fibre excitation point and light sensor, bending the fibre to the same radius. The light yield of the fibre is measured once an hour. Since the radiation dose, due to the electrons used to excite the fibre, received by the fibre is negligible (under 1 mgy), only source of increasing light losses are expected to be the microcracks produced by the bending. The setup is depicted in figure 10. Figure 10: Bending test setup. The red line marks the path of the fibre. Since the microcracks appear over the course of a few days, it is not practical to do the measurements on weekdays since the light yield setup is needed to keep 11
the quality assurance process running smoothly. Thus the test has to be done over the weekend and it has to be automatized. I wrote a script in Visual Basic, based on code written previously by Laura Gavardi, to automatically start a measurement and save the accumulated data once an hour. After the measurement is done, the light output can be calculated for each point in time and its evolution over time determined. Preliminary analysis suggests that the minimal bending radius is below 15 mm, but more detailed analysis will be done later. 6 Summary During this summer, I succeeded in calibrating the new densitometer. The instrument gives accurate, reproducible results. I also managed to create a user interface, which expands the features of the densitometer: now the user doesn t have to calculate the dose received by a film by hand, but the program does the calculation, thus removing one manual step from the X-ray radiation hardness measurement. Also, I was able to design a new measurement setup for the bending test using the already existing light yield setup and automatize the measurement with Visual Basic. Even though not yet completely analyzed, preliminary results suggest that the minimal bending radius of the fibres is at least smaller than 15 mm, as expected. References [1] GAFCHROMIC R HD-810 Radiochromic Dosimetry Film. [2] The LHCb Collaboration. LHCb Tracker Upgrade Technical Design Report. Technical report, CERN, 2014. 12