In order to extract information about the charge in the electron bunch by evaluating the intensity of the signal from the scintillating screen, the respective screen together with the imaging system has to be calibrated with a well characterized electron source.
One existent calibration method that can be applied to this problem is the absolute cali- bration ofImage Plates(IP) [118, 119] done by Tanaka et al. [120] and Zeil et al. [121]. These photostimulable phosphor plates store energy in metastable color centers (lumines- cence centers) that is proportional to the energy deposition per area. Since the energy loss for electrons above ∼ 10 MeV is nearly constant, the deposited energy corresponds to the charge per area. The stored energy can be read out by a visible-wavelength laser. IPs are very sensitive and can detect even low charge electron bunches, but the read-out takes long and needs a separate scanner, which renders them impractical for permanent use in an electron acceleration experiment. But as their properties are well characterized with conventional accelerators, they can be employed to calibrate a setup consisting of a scintillating screen, imaging optics and a camera. The fact that the calibration is only valid for one geometry and imaging system is a major disadvantage of this method. In order to obtain a more general calibration for scintillating screens, we introduced an additional constant light source (CLS) as a luminosity reference [115]. Simultaneously, with the scintillator signal the CLS was imaged. By comparing the integrated CLS signal in the final experimental setup at the laser-wakefield accelerator (Fig. V.3) to the signal in
dipole magnet scintillating screen mirror camera objective CLS
Figure V.3.:Setup for spectrum and charge acquisition. CLS: constant light source
the calibration setup, the scaling factor for the scintillator-signal-to-charge conversion can be extracted. Hence, the calibration can be easily transferred to arbitrary imaging systems and geometries, without any additional tools as e.g. anImage Plate scanner.
Our CLS are small glass tubes, filled with gaseous tritium and coated with a luminous substance. The electrons emitted by the radioactive tritium permanently activate the lu- minous substance (radio luminescence), such that a constant photon emission is guaran- teed, although of course the half-life of tritium has to be taken into account. These tritium tubes are commercially available from mb-microtec[122] and can be easily attached to the scintillating screen. Our CLS-tube is 12 mm long and 2 mm in diameter. If the screen and the CLS together are calibrated at a conventional electron accelerator, the CLS gives a scaling factor for the light intensity versus stimulating charge depending on the imaging setup in the experiment.
The absolute charge calibration is detailed in ’Absolute charge calibration of scintillat- ing screens for relativistic electron detection’ Buck, Zeil, Popp et al.[115]. It was per- formed at the ELBE linear accelerator at Forschungszentrum Dresden-Rossendorf, de- livering electron bunch trains of variable length. The charge of a single electron bunch can be tuned up to 50 pC. For higher charges up to 100 nC several pulses are summed up with a 154 ns delay between the single pulses, which have a duration of 2 ps. The decay time of the scintillating screens is around one millisecond. Therefore, in order to simulate a higher bunch charge, several bunches can be accumulated. The electron energy was 40 MeV for all measurements. Reference charge measurements were conducted with Faraday Cups and Integrating Current Transformers (ICT), as they are routinely used at conventional accelerators.
Peak charge density (pC/mm2)
Total charge (pC)
Scintilla
tion Signal (photons/sr)
N scint / N CL S 100 101 102 103 104 105 10-2 10-1 100 101 102 103 104 103 102 101 100 1014 1013 1012 1011 1010
single bunches multiple bunches
ref
(a) Scintillator signal vs. deposited charge for various commercial screen types,NCLS is the signal of the constant light source recorded with a exposure time of 20 ms. In this workCAWO OG16 was used. The data point marked ”ref” is taken from an independent calibration done by Glinec et al. [114] for the
KODAKLanex Fine Screen.
Ρ
ICT (pC/mm2)Ρ
scint (pC/mm 2 ) 0 20 40 60 80 100 20 40 60 80 100(b) Measured charge density in the peak from the scintillation screenCAWO OG 16 vs the charge density calculated from an ICT measurement. Dashed line: linear fit (below saturation), solid line: Birks saturation law [123]
Figure V.4.:Charge calibration of different scintillating screens, original plots and data evalua- tion by A. Buck, for further details see [115]
experiments was calibrated. Figure V.4(a) shows the photon signal of the scintillating screen depending on the electron bunch charge. Both the integrated charge is shown as well as the charge per area. In order to extract the charge density a Gaussian was fitted to
the spatial intensity distribution. In addition to the absolute scintillator signal, it is given scaled with the CLS-signal. The important conversion for our experiment is: 1 pC will result in a signal on theCAWO OG 16 screen that is 4.86 times higher than the spatially integrated signal of the CLS. This holds for images taken with 20 ms exposure time. For longterm calibration a drop in luminosity of the CLS has to be taken into account. This is due to a combination of tritium decay and degradation of the luminous substance. In good approximation an exponential decay with 5 years decay time can be assumed (calibration date: October 2009).
The sensitivity of the CAWO OG 16 is relatively high, which is advantageous for low- charge electron bunches. The trade-offis the simultaneously reduced spatial resolution. The data point marked ”ref” in figure V.4(a) is taken from an independent calibration done by Glinec et al. [114] for one charge density and fits well to our measurement (both done for theKODAK Lanex Fine Screen).
Figure V.4(b) shows the measured charge density extracted from the peak of the scintillat- ing screen signal (CAWO OG 16) depending on the deposited charge as measured by an ICT. For low charge densities the signal is identical and can be fitted linearly. With higher charge densities the screen signal deviates from the linear fit, the scintillating screen satu- rates. The fitting curve there is based on Birk’s saturation law for scintillators [123]. The charge density that saturates the screen was determined to be 32.9±6.6 pC/mm2. At this
pointρscinthas dropped to 95% of the linear fit. The dispersed charge density (behind the
spectrometer magnet) produced in the experiments of this work is still well below that limit.