CMS Note Mailing address: CMS CERN, CH-1211 GENEVA 23, Switzerland

Transcription

1 Available on CMS information server CMS NOTE CMS NOTE 27/3 The Compact Muon Solenoid Experiment CMS Note Mailing address: CMS CERN, CH-2 GENEVA 23, Switzerland 5 JAN 27 Radiation-hardness measurements of high OH content quartz fibres irradiated with 24 GeV protons up to.25 Grad. K. Cankoçak a), N. M. Bakırcı b), S. Çerçi b), E. Gülmez c), J.P. Merlo d), Y. Onel d), F. Özok e), I. Schmidt d), N. Sönmez f) a) Mugla University, Department of Physics, Mugla, Turkey b) Cukurova University, Department of Physics, Adana, Turkey c) Bogazici University, Department of Physics, Istanbul, Turkey d) University of Iowa, Department of Physics and Astronomy, Iowa City, USA e) Istanbul University, Department of Physics, Istanbul, Turkey f) Ege University, Department of Physics, Izmir, Turkey Abstract We investigated the darkening of two high OH content quartz fibres irradiated with 24 GeV protons at the Cern PS facility IRRAD. The two tested fibres have a.6 mm quartz core diameter, one with hard plastic cladding (qp) and the other with quartz cladding (qq). These fibres were exposed at about.25 Gigarad in 3 weeks. The fibres became opaque below 38 nm, and in the range nm. The darkening under irradiation and damage recovery after irradiation as a function of dose and time are similar to what we observed with electrons. The typical attenuation at 455 nm are.44 ±.22 and 2.2 ±.5 db/m at Mrad for qp and qq fibres, respectively. The maximum damage recovery is also observed near this wavelength.

2 Introduction For a few years, in association with Turkish Particle Physics groups, the University of Iowa group investigated the radiation hardness of different types of fibres for the Hadronic Forward calorimeter (HF) of the Compact Muon Solenoid experiment (CMS) at LHC. The results of darkening measurements of quartz fibres of.3 and.4 mm core diameter, irradiated with 5 MeV electrons up to 5 Mrad have already been published []. To complete our knowledge of radiation damage in quartz fibres and to investigate their possible use at the Super LHC (SLHC) we initiated a high level irradiation with 24 GeV protons at CERN. About 6 protons/cm 2 were sent onto.2 m of quartz fibres, corresponding to a dose of.25 Grad (i.e. about 2 years of HF operation at LHC in the hottest tower). The fibre radiation damage induced by protons exhibits the same well known behaviour as with electrons: high light attenuation below 38 nm and in the band nm. Moderate attenuation in the band 4-52 nm and practically no attenuation above 7 nm was observed. The damage varies exponentially with dose; fast in the first hours and slow after. Above.6 Grad we observe a new phenomenon: the radiation damage is not recoverable in the range nm and below 38 nm. The two fibres tested,.6 mm quartz core diameter, one with hard plastic cladding (qp) and the other with quartz cladding (qq), are supplied by the US firm Polymicro Tech. Inc. (PT). 2 Experimental set-up and data acquisition 2. Experimental set-up The fibres were installed, with their light support, on a remote controlled table (Fig. ) at the CERN facility IRRAD [2]. Figure : Experimental set-up Three loops of qq and qp fibers (4 cm long) are wrapped together and placed with a slope of 2.5% relative to the beam. The total irradiated length (L) is about.2 m. No material was inserted in front of the fibres as in the case of electrons []. The 24 GeV proton beam at IRRAD was delivered as 3-4 sub-cycles in a (9.6 ± 3.)s long supercycle [2]. This beam delivers p/cm 2 /burst in a spot of 2x2 cm 2. The subcycle sharing changed during the irradiation. The number of protons delivered was recorded and the dose calibration was done using thin Al samples. We regularly checked the proton beam stability in position, width and time. The optical setup uses Ocean Optics components. The two-channel spectrometer (SD 2) is the same as in electron measurements []. At the beginning of irradiation we used a triggerable deuterium halogen lamp (d+h) and at the end a Xe lamp. These lamps deliver a wider spectrum than the range of the spectrometer (35-85 nm). The attenuation measurements were performed in-situ, the irradiated fibres linked to the spectrometer and source with 25 m long fibers. The light was injected into a Y fibre splitter and then into the qq and qp fibres. 2.2 Data acquisition The intensity of the light source (d+h) was quite low for the SD2 that required an integration time of s inducing a noticeable background (Fig. 2). In the range of nm and with a dose of about one Grad, the 2

3 signal strength in the SD2 was still significant. The qp channel was less sensitive than the qq one. The rate of light flashes and data acquisition changed from one per cycle (9.6s) at the beginning of irradiation to one per 4 cyc. (3min) and one per 8cyc. (26min) at the end according to the damage versus dose or time (see section 3). () (2) ADC counts c(λ) (3) (4) ADC counts Figure 2: Evolution of the transmitted light spectra from a.25 m of qq fibre as a function of dose ( to.25 Grad). 2.3 Dose estimate In the 2.5% tilted fibres, protons crossed 2.4 cm of quartz, well before the hadronic shower maximum, and one can only consider the minimum energy loss by protons. Then the dose D(MeV/g) depends simply on the minimum energy loss in quartz de/dx =.7 MeV/(g/cm2) and the total number of protons/cm 2 N. D(MeV/g) = N(dE/dx) () For 6 protons/cm 2, D =.7 6 MeV/g = 272 Mrad. At IRRAD with protons/cm 2, the total dose of the proton irradiation is.25 Grad. The electron irradiations at LIL were done up to 5 Mrad []. 3

4 3 Analysis and results To measure the attenuation in the fibre samples we had to take data outside the spills of protons not to include the Cherenkov light generated by protons. But the PS cycling structure was changing during the run and some runs included Cherenkov light, these runs were discarded. Then, the data are histogrammed in nm wavelength bins centered at λ i. Runs are arranged in groups covering a few minutes of irradiation at the beginning and hours after Mrad dose. 3. Analysis method We first checked that in absence of input light the spectrometer background is independent of λ, dose and time. We added all the background distributions corresponding to each λ bin to get an overall background distribution (Fig.3), which we fit with a Moyal distribution. The position b of the maximum in ADC counts is taken as background in each λ bin, for a given lamp and spectrometer channel at a certain dose and time. At a given dose D in the bin λ i the raw signal distribution S(λ i, D) in ADC counts starts at the cusp value c(λ i, D) (Fig.4) and exhibits a Landau shape due to the background entanglement. The distributions above c(λ i, D), i.e. [S c](λ i, D), are similar for each i bin and characterizes the background entanglement. Summing over the λ i bins one gets the total distribution in ADC counts S T (D) = i [S c](λ i, D) (2) Fitting S T (D) with a Moyal (Landau) function, gives the maximum (µ(d)) of the S T (D) distribution (Fig.4). We assume that the signal transmission intensity (I(λ i, D)) can be expressed as a function of c, µ and b as I(λ i, D) = c(λ i, D) + µ(d) b (3) 3.2 Results We calculate the ratios I(λ, D)/I(λ, ) which decrease as a function of dose as shown in Figs. 5 to 8 giving the typical behaviour at 455 and 65 nm for qq and qp fibres. The transmitted light attenuation in the fibre is well represented by the following function [3]. A(λ, D) = α(λ)[d/d s ] β(λ) (4) α and β parameters for qq and qp fibres are determined by fitting the ratios as a function of wavelength and dose. I(λ, D)/I(λ, ) = exp[ (L/4.343)α(λ)(D/D s ) β(λ) ] = exp[ A(λ, D)L/4.343] (5) Choosing a scale factor D s = Mrad, and L being expressed in meters, α is the attenuation at Mrad, in db/m. The results for qq and qp fibres are shown in (Figs. 9 and ) for all wavelengths and in Table for 455 and 65 nm. For the qp fibre the results are the mean of two successive nm bins to increase the statistics. Statistical errors are larger than those in electron irradiations because of the lower light intensity and some beam unstability. Comparing to the former analysis of electron irradiation data we fixed the systematic errors to be α=.5 independent of lambda and β β =.5 in the range nm and. below 4 nm or above 6 nm. Table : Values of α, β parameters at 455 and 65 nm for qq and qp fibres of. 6 and.3 mm core diameter irradiated with protons. Fibre α( nm) β( nm) α(65 62 nm) β(65 62 nm) qq (.6) p 2.2 ±.6.53 ± ±.5.66 ±. qp (.6) p.44 ±.6.44 ± ± ±.2 qp (.3) e.54 ±.2.33 ±. 6.3 ±.6.6 ±. 4

5 The values of α and β parameters for qq and qp fibres irradiated with protons are in good agreement with the results for.3-.4 mm core diameter irradiated with electrons, as long as they are drawn by Polymicro Techn. with the same type of quartz. The fits up to.25 Grad near 45 nm show that the qq fibre has an higher attenuation than qp fibre. The fit of proton data up to 5 Mrad like the electron data [] this difference disappears Background distribution (qq) 75 5 Overall signal distribution (qq) ADC values ADC values Figure 3: Overall background distribution in ADC counts and the Moyal function fit defining the background b as the position of the maximum for qq fibre. Figure 4: Overall background distribution S T (D) and the Moyal function fit defining the additional value µ(d) as position of the maximum for qq fibre. 4 Radiation damage recovery 4. Fit the recovery data The radiation damage recovery was measured in qq and qp fibres after a dose accumulation of.4 Grad (24 hours of irradiation). The qq fibre recovery was also measured after.25 Grad (65 hours). The qp fiber broke before the end of maximum irradiation. We are not sure of the cause of the breakage. Nevertheless the steep decrease of the transmitted light in qp fibre above Grad (Fig.7) favors the hypothesis of cladding deterioration. The time evolution of defects in quartz can be described by exponential functions or by their approximate form at low values of the exponent. [ ( t ) η ] Q(λ, t) = Q(λ, ) exp τ(λ) For small damage recovery effect, τ(λ) gets large values and it is impossible to fit the data. For this reason we use Γ(λ) as second parameter Γ(λ) = ( ) η(λ) (7) τ(λ) We choose to fit the recovery data, attenuation versus time, with the following function (as was done in ref. [] and [3]) ( t ) η(λ) A D (λ, t) = A(λ, D)/( + [γ(λ) ]) (8) t irr (6) 5

6 I(D) / I() at λ= 455 nm qq fibre I(D) / I() at λ= 65 nm qq fibre D/D s D/D s Figure 5: I(D)/I() for qq fibre versus dose at 455 nm. Figure 6: I(D)/I() for qq fibre versus dose at 65 nm. qp fibre qp fibre I(D) / I() at λ= 455 nm I(D) / I() at λ= 65 nm D/D s D/D s Figure 7: I(D)/I() for qp fibre versus dose at 455 nm. The fit is done up to Grad. Figure 8: I(D)/I() for qp fibre versus dose at 65 nm. 6

7 α (db/m) β Figure 9: α parameters versus λ for qq and qp fibres, corresponding to the fit of irradiation data up to.25 Grad. Figure : β parameters versus λ for qq and qp fibres, corresponding to the fit of irradiation data up to.25 Grad. Where A D (λ, t) is the value of the light attenuation at the post exposure time t (in hour) after an irradiation at a dose of D in a time t irr. The parameter γ(λ) being related to τ(λ) as, τ(λ) = t irr /(γ(λ)) /η(λ) (9) Figures -2 show the decrease of the attenuation versus time, expressing the radiation damage recovery for qq and qp fibres at nm after 4 Mrad irradiation. Figures 3-4 show the corresponding increase of the transmitted signal. Fitting the recovery data with function (8) one gets the parameters γ and η versus λ shown in Figures 5-6 for the qq and qp fibres. In Table 2 the γ and η values at 455 nm and 65 nm for proton irradiation are presented and compared with the electron irradiation results []. The τ values are recalculated according to Eq. (9). One notices a strong recovery maximum near 45 nm with γ values close to and a steep decrease below 4 nm and above 5 nm. The qq data exhibits the shrinkage of the peak between 4 Mrad and.25 Grad corresponding to the absence of recovery below 38 nm and above 6 nm. The quartz becomes opaque to these wavelengths at high dose. Table 2: Values of γ and η parameters at 455 and 65 nm for qq and qp fibres of. 6 mm core diameter (proton irradiation) and.3 mm core diameter (e irradiation at LIL). At 455 nm the recovery lifetime τ is calculated according Eq.(9), at 65 nm τ is meaningless regarding the errors. Fibre γ(455) η(455) τ(455) in h γ(65) η(65) qq (24 h).93 ±.25.3 ±.8 28 ± 57.8 ±..25 ±.7 qq (65 h).79 ±..22 ±. 635 ± 3.3 ±..57 ±.45 qp (24 h).89 ±.8.3 ±.6 23 ± 47.8 ±.2.49 ±.33 qp (e 45 h).6 ±.2.26 ±. 4 ± 3.77 ±.4.9 ±.5 7

8 ..5 A D (t)/a(d) versus time (hour) for qq, λ= 455 nm..5 A D (t)/a(d) versus time (hour) for qp, λ= 455 nm Figure : Decrease of the fibre light attenuation versus the post irradiation time for the qq fibre at nm after 4 Mrad irradiation. Figure 2: Decrease of the fibre light attenuation versus the post irradiation time for the qp fibre at nm after 4 Mrad irradiation..5 I(t)/I() versus time (hour) for qq, λ= 455 nm.25 I(t)/I() versus time (hour) for qp, λ= 455 nm Figure 3: Increase of the transmitted signal versus the post irradiation time for qq fibre at nm after 4 Mrad irradiation. Figure 4: Increase of the transmitted signal versus the post irradiation time for qp fibre at nm after 4 Mrad irradiation. 8

9 .4.6 γ η λ λ Figure 5: Recovery parameters γ versus λ for qq and qp fibres versus wavelength, after 25 hours and 65 hours of irradiation. Figure 6: Recovery parameters η versus λ for qq and qp fibres versus wavelength, after 25 hours and 65 hours of irradiation. 4.2 The damage recovery at 45 nm In an iron-fibre calorimeter like CMS-HF the Cerenkov light is detected by PMTs sensitive in the 4-5 nm range. Near 45 nm the damage recovery is maximum and gamma is close to The knowledge of the evolution of light transmission in fibres after a break in data taking is important for the calibration measurements [4]. From Equations 3-8 one can easily derive the increase of transmitted signal I D (t) after an irradiation to a total dose D in time t irr. I D (t)/i(d) = exp[a(d)(l/4.343))(γt η /(t η irr + γtη ))] () As an example we present in the Table 3 the values of the ratios I D (t)/i(d) for three post irradiation times t for a 2 m long fibre irradiated at Mrad in t irr = 48 hours with γ=.9, η =.25 and A =.6 db/m. Table 3: Values of I D (λ, t)/i(λ, D) for three post irradiation times after the irradiation of 2 m of fibre at Mrad in t irr = 48 h, at 455 nm. t(h) I D (t)/i(d) One can see from Table 3 what kind of correction might be applied to a calibration performed at a post data taking time t to get the real calibration at the data taking time. The difficulty to determine accurately the damage recovery parameters leads us to propose to monitor in situ the light transmission in the quartz fibre of the CMS-HF calorimeter during and after data taking. 5 Summary Radiation damage due to protons and damage recovery are in agreement within errors with our previous results of electron irradiations with the same type of quartz used by Polymicro Technology Inc.. The two types of irradiation were quite different: - 24 GeV protons compared to.5 GeV electrons, in different geometries 9

10 -.25 Grad with protons compared to 5 Mrad with electrons - Different set-up and different fibre sizes As in electron irradiations we did not observe significant differences in proton irradiation between qq and qp fibre up to Grad. Near 45 nm the observed difference appears in the fits at high dose and above Grad there is a steep decrease of the transmitted signal in qp fibre (Figure 7). For doses above.5 Grad there is no recovery of damage below 38 nm and in the range of nm, the quartz becomes opaque. The PMTs detecting Cherenkov light are sensitive to the blue light (about 45 nm) emitted from the fibres where the effects are maximum. For a 2 m qp fibre with the parameters listed in Tables and 2, one expects: -a decrease of 3% in light transmission resulting from a dose accumulation of 2 Mrad -an increase in light transmission by 22% ten days after an irradiation to Mrad in 2 days. With a quartz-iron calorimeter like CMS-HF, the calibration [4] is done after the data taking. Also one should know the change of response of HF fibres between the end of data taking and the time when the calibration takes place. As mentioned in ref. [5] at 47 nm there is a luminescence band with a rather long and not well measured lifetime of 2- ms in quartz. This could explain the quartz radiation effects near 45 nm: relatively low darkening under irradiation and almost full recovery of the irradiation damage. But this quartz luminescence band is still not very well known and merits to be investigated more seriously to improve our knowledge of the quartz material and to understand the consequences in a quartz-metal detector as the HF-CMS calorimeter. Acknowledgments We thank Maurice Glaser responsible of the IRRAD facility who helped us in the installation of our set-up and calibrated the dose received, the PS operators and technicians providing the proton beam, the SPS-PS coordinator and the CMS test beam coordinator. We thank I.Dumanoglu for helpful discussions. We would also like to thank the University of Iowa for their continuing support for this research. This work was supported by the US Depart. of Energy (DE-FG2-9ER 4664) and NSF (NSF-INT ) and the Scientific and Technical Research Council of Turkey, TUBITAK, Turkish Atomic Energy Board (TAEK) and Bogazici University Research Fund (Grant no: 6B32). References [] I.Dumanoglu et al., Nucl. Inst. and Meth. in Phys. Res. A49 (22) [2] M. Glaser et al., Nucl. Inst. and Meth. in Phys. Res. A426 (999) [3] D.L. Griscom et al., Phys. Rev. Lett. 7 (993) 9. [4] CMS-HF collaboration, CMS NOTE 26/44 [5] L. Skuja, Journ. Non-Cryst. Solids 239 (998) 6.