1. Introduction

Recently CsI photocathodes with electron multiplier structures like a Thick Gas Electron Multiplier (THGEM) have gained a lot of attention due to their applicability in vacuum and gas operated imaging detectors in the UV spectral range [1,2]. UV photon detectors like gaseous photomultipliers (GPM) can operate in two different photocathode configurations.

 

Fig. 1 Schematic representation of the working principle of a gaseous photomultiplier with semitransparent and reflective photocathode.

Download Full Size | PDF

Photoelectron yield from the photocathode surface has a significant role in deciding the UV photon detection efficiency of the GPM. If photoelectron yield is smaller, the number of photoelectrons reaching the electron multiplier will be smaller leading to reduced detection efficiency of the GPM. Since CsI is hygroscopic in nature, moisture plays an important role in the decay of the photoelectron emission properties [3,4]. Even short-term exposure of about half an hour to humid air during transfer and mounting of the photocathode may become crucial in terms of photoelectron yield. The reason behind this is the strong attractive force between the polar water molecules and the ionic CsI crystals. Since water has high vapor pressure, the rate of adsorption of water molecules is much higher than the desorption rate. Hence water molecules accumulate over the surface of the film. Since the water molecules are polar, once one layer is formed, it attracts more molecules and a film formation occurs in which the CsI dissolves. The damage to the photocathode becomes more pronounced due to the solubility of CsI in water [3]. The photoemission property of the photocathode is affected differently for different thicknesses under exposure to humid air. For thicker films the degradation in the photoelectron yield is not very prominent for 30-60 minutes of exposure to ambient atmosphere [2]. But for thin films (20-50nm), the film continuity is affected even for a few minutes of exposure to humidity [4] leading to degraded photoelectron yield.

In this work, we report a simple technique of vacuum treatment of the CsI film (exposed to humid air for short duration) for the improvement of photoemission properties of the film. The microstructural changes in the film due to vacuum treatment are also analyzed under scanning electron microscope (SEM) to correlate this effect with the microstructure of the film.

2. Materials and methods

2.1 Preparation of CsI film

As discussed before, GPMs can operate under two different photocathode mounting configurations. In the present work for ST mode operation, a 2 mm thick quartz plate coated with a 30 nm CsI film with a thin under layer of chromium (4-5 nm) was used as the photocathode, while in Ref mode, a thicker film (300 nm) coated over the top electrode of the THGEM made from copper cladded FR4 was used as photocathode. A thin layer of Cr/Au was sputtered below the CsI layer for Ref mode application. This prevents the direct contact of the copper material with CsI, due to which the copper might dissociate [5]. In this work, the samples were prepared using a thermal evaporation technique. Ultra pure cesium iodide (99.995% purity) was used for deposition. The film deposition was done under a vacuum of 10⁻⁵ Torr. The melting point of CsI at 10⁻⁵ Torr is 621°C. CsI material (in the form of crystals) was loaded in the evaporation chamber in a molybdenum boat which can withstand high temperature. The deposition rate was maintained at 1-2 nm/sec. The thickness and the rate of deposition were monitored by a quartz crystal monitor. After deposition, the samples were annealed under high vacuum at 70°C for 4 hours to enhance the photoemission from the photocathode [5]. Throughout the remainder of this paper, we will refer to films with 30 nm thickness as thin films and films with 300 nm thickness as thick films.

2.2 Measurement of photocurrent

The photoemission property of the photocathode can be estimated by measuring the photocurrent from the photocathode. The measurements on thin films were carried out in ST configuration and for thick films were performed in Ref configuration. The setups used for these measurements are shown in Fig. 2(a) and Fig. 2(b). The photocurrent was measured using an electrometer (Keithley 6517A). In our measurements, we used a low pressure Hg vapour lamp emitting UVC radiation in the wavelength range of 100-280 nm. The quantum efficiency of CsI varies between 70% and 0.1% in the wavelength range between 120 and 220 nm [6]. This overlaps with the wavelength range emitted by the lamp. Before starting the experiment, we had monitored the UV lamp stability by observing photocurrent using setup 2(a) for a period of 3 hours. Initially the fluctuation in the photocurrent (and hence photon flux) was 10-15% and after a period of about 45 minutes, a much more stable value of 3-5% fluctuation was observed. In order to ensure a stable photon flux, all our measurements were carried out after the UV lamp was ON for 45 minutes.

 

Fig. 2 Schematic of the setup used for photocurrent measurement in (a) ST photocathode configuration, (b) Ref photocathode configuration.

Download Full Size | PDF

The initial photocurrent before any exposure to humidity was measured for both thin and thick films. The as-deposited films (both thick and thin) were directly transferred from the evaporation chamber to the test chamber for photocurrent measurement. Care had been taken not to expose the films for more than 1-2 minutes during unloading and loading. The photocurrent measurements were carried out at a vacuum of 10⁻⁵ mbar which was achieved in about 5 minutes after opening the high vacuum valve. For thin film the initial photocurrent was around 1.1 nA whereas for thick film it was around 175 nA. After this, both thick and thin film samples were kept exposed to moist air (relative humidity~70%) for an hour. Air with relative humidity around 70-80% is considered to be high humidity air [3]. Hence one hour exposure to this air should be sufficient to degrade the photoemission property significantly. If the exposure is for much longer duration (2 hrs or more) irreversible damage of the film is reported [3]. Hence we had restricted our study to one hour exposure only. After exposure to humid air for 1 hour, the photocurrent measurement was carried out during constant evacuation and also in the presence of gas mixture for both thick and thin films. In situ photocurrent measurement during the process of evacuation was carried out. The measurement was done for two different electric fields (E_D) at the photocathode surface. Since GPMs work in gas medium [1], photocurrent measurement was also carried out in the gas medium to find out the actual effect of vacuum treatment for the application. The gas mixture used was P10 at 900 mbar pressure. The current for a particular duration of evacuation was measured first under high vacuum (~10⁻⁶ mbar). Immediately after this, gas was filled and readings were taken. The gas was evacuated out after each measurement, and the same procedure was repeated for the next set of reading.

3. Results and discussion

3.1 Effect of vacuum treatment on photocurrent

As shown in Fig. 3 it was observed that for thin CsI films, the photocurrent increases with the duration of evacuation for both the electric fields, and it becomes constant after 90-100 minutes of evacuation. It was observed that after 1hour exposure to humid air, the photocurrent under vacuum decreased to around 67% of the value obtained for the as-deposited film. Vacuum treatment of the film for 1.5 hours restores the photocurrent up to around 86% of that obtained with freshly deposited film. Thus a fraction of the lost sensitivity of the film due to exposure to humidity is regained after vacuum treatment. The increase in photocurrent under evacuation can be explained as follows. CsI is a negative electron affinity material, so water molecules on its surface can give it positive electron affinity which reduces the photoelectron yield. During evacuation, the moisture adsorbed at the film surface gets desorbed. As the film thickness is small, there is not much penetration of moisture to the bulk of the film. The adsorbed moisture over the CsI film absorbs the incident UV [3], thus degrading the photocurrent emission from the photocathode surface. Due to evacuation the adsorbed moisture starts getting removed from the surface which enhances the photoelectron emission. When all the moisture content is removed from the film, the photocurrent reaches saturation and there is no more increase in current with evacuation time. The photocurrent was much smaller in gas medium than in vacuum due to the well-known backscattering of photoelectrons to the photocathode [7].

 

Fig. 3 Effect of evacuation time on photocurrent for thin film in vacuum as well as in gas medium.

Download Full Size | PDF

Figure 4 shows the results obtained for thick CsI films under constant evacuation. It was observed that the photocurrent increases slowly up to about 45 minutes of evacuation and after that it starts increasing at a much faster rate with time, and it continues to increase even after 3.5 hour of evacuation. However, the rate of increase of photocurrent slows down after 3.5 hours. This could be due to slow removal of water molecules from the bulk of the film. The increase in current was much higher than that for thin CsI films. The increase in current was seen both under vacuum and also in gas medium. It was observed that in vacuum the photocurrent decreased to less than 5% of the value obtained with the freshly deposited film after 1hour exposure to humid air. Evacuation for 3.5 hours restored the photocurrent to around 90% of the value measured with the as-deposited film. Continuous evacuation for both thin and thick CsI films showed similar improvement without filling gas in between as compared to multiple cycles of evacuation with gas filling in between. In order to understand the effect of evacuation on the microstructure of the film, the surface morphologies of different thick film samples kept under evacuation for different durations of time were studied.

 

Fig. 4 Effect of evacuation time on photocurrent for thick film in vacuum as well as in gas medium.

Download Full Size | PDF

3.2 Surface morphology study with duration of evacuation

For surface morphology study, we prepared three samples of thick CsI film. These samples were kept exposed to humid air (relative humidity~70%) for 1 hour and then loaded to the evacuation chamber. The three different samples were kept under high vacuum (~10⁻⁶ torr) for 30 minutes, 45 minutes and 1 hour respectively. Care had been taken to avoid the contact of air with the vacuum treated samples during transfer from the evacuation chamber to the SEM chamber. The samples were exposed to air for a maximum of 1-2 minutes during loading and unloading. To obtain a clear understanding of the progression of morphology of the CsI film from the original state through vacuum treatment, we had taken SEM images at different stages. Figure 5(a) shows the SEM image of the as-deposited film. It can be seen that the film is densely packed with grains having an average size of 300 nm. Figure 5(b) shows the image of a film exposed to humid air. Here the grain size increased slightly (900 nm) due to coalescence of neighboring grains. This behavior has also been reported earlier [3]. But once exposed to humid air, a layer of water molecules is produced over the film. This reduces the photoelectron yield from the surface. Figure 5(c), Fig. 5(d) and Fig. 5(e) are the images of the samples treated under vacuum for 30 minutes, 45 minutes and 1 hour respectively. It can be observed that with the increase in evacuation time, the film recrystallizes to form bigger size grains. The grain sizes initially do not vary much for 30-45 minutes of evacuation and then suddenly increase from 1.2 µm for 45 minutes of evacuation to 5.4 µm for 1 hour evacuation. The observation is similar to photocurrent variation with evacuation time, where up to 45 minutes the current increases slowly and thereafter starts increasing at a much faster rate. Now the increase in photocurrent with evacuation time as observed in Fig. 4 can be explained as follows. Initially with evacuation, the water vapor adsorbed at the surface gets removed. Thus there is not much change in the grain size between 30 and 45 minutes of evacuation. This removal of adsorbed water molecules from the surface of the film increases the photoelectron emission to a certain extent. With longer duration of evacuation, as the grain size increases the photocurrent also starts increasing drastically. As pointed out in [8], smaller grains have a detrimental effect on the photoelectron yield due to increased number of grain boundaries. These grain boundaries trap photoelectrons and prevent them from leaving the film. As the grain size increases, the number of grain boundaries decrease. This increases the probability of photoelectrons coming out of the film. Thus with the evacuation for one hour, the photocurrent increases due to increased grain size. Also due to slow removal of water molecules from the bulk of the film, the photocurrent increases at a slow rate even after 3.5 hours of evacuation.

 

Fig. 5 SEM images of thick CsI film (a) as deposited film, (b) after 1 hour exposure to humid air, (c) after 30 minutes of evacuation, (d) after 45 minutes of evacuation (e) after 1 hour of evacuation.

Download Full Size | PDF

4. Conclusion

CsI is used as a photocathode material for UV photon detectors. The detection efficiency of the detector strongly depends on the photoemission property of the photocathode. CsI is hygroscopic in nature. Thus moisture in air degrades the photoelectron emission from the photocathode surface due to hydrolysis of the material surface. Vacuum treatment of CsI films exposed to humid air for short duration can be used to overcome this problem. Constant evacuation for long times (90-100 minutes) removes the adsorbed moisture from the thin CsI films thus improving the photoelectron yield. Hence the first effect of water on CsI photocathode is to increase the electron affinity of the surface, lowering the photoelectron yield in a manner that can be reversed by removal of water. Thicker films undergo recrystallization to form bigger grains due to longer evacuation. Larger grains improve the photoelectron yield due to reduced number of grain boundaries which increases the escape probability of the photoelectron from the film. The improvement in photocurrent holds good for the working gas also. Thus for GPMs a prolonged evacuation before gas filling might lead to better detection efficiency.