Experimental setup for the measurement of optical properties in the vacuum ultraviolet region

This work presents an experimental setup designed for measuring optical properties in the vacuum ultraviolet (VUV) region. Originally developed for the characterization of ICARUS photomultiplier tubes, the setup has been recently employed to measure the conversion efficiency of para-Terphenyl coatings, acting as wavelength shifters, as a function of thickness. The VUV region is of particular interest in particle physics experiments utilizing cryogenic noble gases as scintillating media. The setup includes a deuterium light source, a mirror system, an intensity monitor, and a sample chamber to host sample and photodetector. An example measurement of para-Terphenyl conversion efficiency — as wavelength shifter — with varying thickness is presented, demonstrating the versatility and effectiveness of the presented system.


Introduction
The investigation of optical properties in the vacuum ultraviolet (VUV) region holds significant importance across various scientific disciplines.In the context of particle physics, this significance is further heightened by the increasing utilization of cryogenic noble gases, as Argon and Xenon, as scintillating materials.These gases play a crucial role in experiments related to dark matter and neutrinos, exemplified by projects like ICARUS, MicroBooNE, DUNE, Xenon, LUX, DarkSide, and numerous others.
The appeal of Argon and Xenon lies in their high light yield, excellent purity, extended electron lifetime, and the capability for pulse shape discrimination.However, a trade-off emerges due to their emission spectra, centered at 128 nm and 178 nm, respectively, falling within the Vacuum Ultraviolet (VUV) region.Traditional optical materials exhibit low transmittance or reflectance in this spectral range.
This paper introduces an experimental setup tailored for the measurement of optical properties in the VUV region.Originally designed for characterizing ICARUS photomultiplier tubes [1], the setup has recently been employed to assess the conversion efficiency of para-Terphenyl coatingsserving as wavelength shifter -as a function of the coating thickness.

Setup description
The experimental setup, schematically shown in figure 1, can be divided into four main components: 1. Light source and monochromator.A Hamamatsu L2D2 Deuterium Lamp serves as the light source, emitting a spectrum primarily peaked at 121 nm and 161 nm.The 121 nm line is particularly relevant as it closely mimics the scintillation emission of liquid Argon, which peaks at 128 nm.A focusing elbow directs the light through a slit for intensity regulation, then to a McPhearson 234/302 monochromator with a diffraction grating featuring 120 ridges/mm.Spectral width is adjustable using a second slit, and the light passes through a collimator, reducing the beam diameter to approximately 8 mm.
-1 -2.Intermediate chamber.This chamber incorporates an aluminum mirror system at its center, and the mirror can be rotated around its vertical axis.This allows the beam to be directed towards different targets without the need to open the apparatus.This design enables alternating illumination of the device under test and the light intensity monitoring device.
3. Intensity monitor.To ensure measurement stability, the setup includes photodiodes (two for redundancy), than can be illuminated by properly orienting the mirror in the intermediate chamber.Photodiode currents are measured both before and after the device under test.This approach compensates for aging effects in the deuterium lamp, noticeable even at short time scales.We foresee to upgrade the intensity monitor with a calibrated photodiode, so that the primary photon flux can be determined and, thus, unlocking the absolute scale calibration for optical characterizations.The entire system is connected to a pumping stage capable of achieving a vacuum level of 10 -6 mbar.This involves both a primary pump and a turbomolecular pump, in addition to a pressure gauge monitor.Vacuum environment (or an inert atmosphere) is necessary because VUV light is strongly absorbed by oxygen, as the name suggests.-2 -

Mesurement example: para-Terphenyl conversion efficiency
Para-Terphenyl (C 6 H 5 C 6 H 4 C 6 H 5 ) (pTP), which absorbs light below 300 nm and re-emits it between 300 and 400 nm [2], is being explored as a potential wavelength shifter for converting light emitted during liquid Argon scintillation.It is currently under investigation as a coating for X-Arapuca, intended for installation in the DUNE Far Detectors [3,4].We started by coating a set of fused silica windows with different amounts of pTP.Then each sample has been measured by placing it in front of a PMT, with pTP coating on the front side of the fused silica substrate, and illuminating it with 121 nm light.VUV light emulates Argon scintillation light and is partially converted by the pTP coathing.Part of the re-emitted light then goes through the fused silica substrate and is eventually detected by the PMT while, on the contrary, the not-converted fraction of 121 nm light is lost or absorbed by the fused silica substrate.This ensures that the PMT signal is only due to converted light and thus proportional to pTP conversion efficiency.

Para-Terphenyl coatings preparation
To investigate the conversion efficiency of pTP as a function of its thickness, a series of 11 samples made of Fused Silica 1 ′′ -diameter windows coated with varying amounts of pTP were prepared.The samples were prepared using a custom-made evaporator located in Pavia Physics Department and each sample underwent a distinct evaporation process.Fused Silica windows were preliminarily weighted with sub-microgram precision, then coated and weighted again.The mass increment was used to extablish the coating thickness (as a mass per area unit).Different coating thicknesses were achieved simply by varing the amount of pTP put in the evaporator crucible.Indeed we found a good linearity between these two quantities (see figure 2).For the three thinnest samples it was not possible to direcly measure the coating thickness, so we simply used this linearity to deduce it from the pTP mass in the crucible.
The evaporation process works as follows.The coated windows were placed in the evaporator with the surface to be avaporated exposed to the crucible.The evaporator is composed by a vacuum chamber were a vacuum of the level of 10 -6 mbar can be reached.Once the desired vacuum level is reached the crucible temperature is gradually increased using an heater to reach the evaporation point of pTP (approximately 235 • C).Then the temperature is kept stable for a few minutes, to ensure complete evaporation, and the heater is then turned off, letting the crucible temperature to passively decrease.After a couple of hours the system is vented and opened to remove the sample.

PMT gain characterization
For converted light detection we used a 1 ′′ PMT, ET Enterprises model 9326FLB with MgF 2 window, with quantum efficiency close to 25% for light in the 300 nm to 400 nm range.
In the preliminary configuration, the PMT was set with all dynodes short-circuited and establishing a potential difference of 50 V between the photocathode and all the dynodes.In this configuration the cathode current can be used to measure the light flux incident on the PMT window without gain (G=1), thus allowing to work with higher luminosities without saturation effects.In a second moment we decided to work in an intermediate range, with a gain of the order of 10 3 .Subsequently the standard configuration was used, employing a voltage divider, and the PMT gain was studied as a function of PMT voltage supply.
-3 - The actual gain measurement was done at a power supply voltage of 900 V, a level where the PMT gain was sufficiently high to produce clearly distinguishable single photoelectron signals due to the PMT dark rate.These signals were used to construct the single photoelectron charge distribution.Data were acquired with the oscilloscope Teledyne LeCroy WaveRunner 610zi.We fitted the single photoelectron charge distribution with two gaussians (one for the single PE peak and one for the noise peak) obtaining a charge value of 26.8 pV•s corresponding to a gain of 3.3 × 10 6 at 900 V.
In a subsequent phase, the gain variation was measured by monitoring the PMT current, for voltages ranging from 900 V down to 400 V in steps of 50 V, under constant illumination.When we found that, after decreasing the voltage by one step, the output current became too low to consider noise contribution entirely negligible, we repeated the measurement at the previous step with an increased amount of light and we resumed the voltage scan with increased light.In these cases we calculated the correction factor as the ratio of currents obtained at the same voltage but with the PMT exposed to different light intensities.Such correction factor was then applied to rectify all the subsequent measurements.This approach facilitated the reconstruction of the gain trend as a function of high voltage (HV), and the gain curve we obtained is reported in figure 3.

Conversion efficiency measurement and results
In order to evaluate the conversion efficiency of pTP, each one of the samples described in section 3.1 was individually measured.We used an assembly that allows to fix the sample directly in front, with 1 mm gap, of the PMT which in turn was hanging from the sample chamber top flange.Despite the PMT being sensitive to the primary 121 nm light, the fused silica sample substrate acts as a filter for that light, ensuring that all the PMT signal is due to converted light.This was verified with a preliminary measurement in wich we used a sample window with no pTP, and we measured a current compatible with zero.
The PMT was negatively biased at 470 V, providing a gain of 3.8 × 10 3 and a picoammeter Keithley 2450 was used to measure the PMT anode current.In order to compensate the deuterium lamp intensity fluctuations (and ageing process) we paired all the PMT current measurements with -4 - four measurements of the current coming from the reference photodiodes.This was possible by redirecting the light beam thanks to the rotatable mirror at the center of the intermediate chamber.For each sample we took current values following this scheme: photodiode-1, photodiode-2, PMT (i.e. the actual measurement), photodiode-2 and photodiode-1.The PMT anode current value obtained for each sample was then divided by the average current measured with one of the two diodes, for the same sample.In this way we obtained two datasets, depending on the considered photodiode, and we found the two datasets to be consistent within the uncertainties.This quantity is proportional to the fraction of light converted by the PMT and provided the relative efficiency curve (see figure 4).It can be seen that for coatings thicker than 200 μg/cm 2 some sort of plateau is reached.This confirms that -5 -

JINST 19 C04036
the design thickness, of 550 μg/cm 2 , foreseen for DUNE X-Arapuca windows is in the plateau region with broad margin so that a certain variability of the thickness obtained during the mass production will not impact the detection efficiency homegeneity along the X-Arapuca surface.
The reproducibility of these measurements was evaluated by repeating the measurement of one of the samples, obtaining a relative difference of less than 1%.For this reason we think that the spread of the values in the plateau region is due to non-uniformities in the pTP depositions.

4 .
Sample chamber.The sample chamber contains the device under test and is connected to the intermediate chamber through a KF-40 flange.Three types of interchangeable sample chambers are currently availabe: two of them are simple cylindrical chambers where both the photodetector (a PhotoMultiplier Tube, or PMT) and the sample hang from the top, while the third one is a Model 121 goniometric chamber by McPhearson.The latter, in particular, allows the rotation of both the sample and the PMT, enabling the variation of both incidence and reflection/transmission angles within a single measurement cycle.

Figure 1 .
Figure 1.Scheme of the setup described in this work, here in the configuration adopted for photomultiplier tubes characterization.

Figure 2 .
Figure 2. pTP coating thickness as a function of the amount of pTP that was put in the evaporator crucible.The three leftmost data points are for reference only, since it was not possible to direcly measure their coating thickness.

Figure 3 .
Figure 3.The photomultiplier tube gain, measured as described in section 3.2, spanning over three orders of magnitude.

Figure 4 .
Figure 4.The para-Terphenyl conversion efficiency, defined as the anode current divided by the average current measured by the photodiodes, for different coating thicknesses.For very thin coatings, below 100 μg/cm 2 , efficiency rises with thickness; then, starting from 100-200 μg/cm 2 , a plateau region is visible.