HASPIDE: a project for the development of hydrogenated amorphous silicon radiation sensors on a flexible substrate

Hydrogenated amorphous silicon (a-Si:H) is a material with a very good radiation hardness and with the possibility of deposition on flexible substrates like Polyimide (PI). Exploiting these properties, the HASPIDE (Hydrogenated Amorphous Silicon PIxels DEtectors) project has the goal of developing a-Si:H detectors on flexible substrates for beam dosimetry and profile monitoring, neutron detection and space experiments. The detectors for this experiment will be developed in two different structures: the n-i-p diode structure, which has been used up to now for the construction of the planar a-Si:H detectors, and the recently developed charge selective contact structure. In the latter the doped layers (n or p) are replaced with charge selective materials namely electron-selective conductive metal-oxides (TiO2 or Al:ZnO) and hole-selective conductive metal oxides (MoO x ). In this paper preliminary data on the capabilities of these detectors to measure X-ray and electron fluxes will be presented. In particular, the linearity, the sensitivity, the stability and dark current in various conditions will be discussed.


Introduction
Hydrogenated amorphous silicon (a-Si:H) is a form of amorphous silicon obtained form the addition of hydrogen atoms in an amorphous silicon structure.This fabrication process improves the electrical characteristics of this material used as radiation detector.
Unlike the more commonly used crystalline silicon, a-Si:H inherits from amorphous silicon the characteristic of not having an ordered atomic structure but instead displaying a disordered arrangement of atoms.Hydrogen is introduced into the material to saturate the so-called "dangling bonds" or unpaired bonds that occur in the disordered structure of amorphous silicon.These dangling bonds can cause defects and charge traps in the material, reducing the charge collection efficiency for this material used as radiation sensor.
The hydrogenation reduces the density of defects in the material, improving its structural quality.Furthermore, hydrogen helps regulating the width of the band-gap of amorphous silicon.Another peculiar property of this type of material is the high radiation resistance.
These attributes makes a-Si:H a useful material in various applications, including solar cells, radiation detectors, and electronic devices [1].
One of the most commonly employed techniques for fabricating thin film a-Si:H is plasmaenhanced chemical vapor deposition (PECVD).In order to achieve the proper hydrogen content, the film is grown from a mixture of Silane (SiH4) and hydrogen at temperatures ranging from 250 to 300 • C. Due to the low deposition temperature of a-Si:H, it can be easily deposited on plastic materials.This allows the sensor to exhibit strong adaptability to curved surfaces and notable flexibility [2].
The plastic materials most frequently used are based on Polyimide (PI) like Kapton.This material also has the advantage of being considered tissue-equivalent (TEM) [3], and in addition to its high chemical stability and radiation hardness, it allows for use in medical and dosimetric applications.
The HASPIDE project [4] focuses on developing a-Si:H sensors deposited on PI.The envisioned applications include beam monitoring for high-energy physics and TRansmission Detectors (TRD) for electron and proton clinical accelerators, X-ray beam dose profiling for medical and industrial purposes, detectors for solar flare events in space missions [5], and neutron detection for industrial, nuclear safeguard, and homeland security applications [4].

Sensors and setup description
The radiation sensors within the HASPIDE project are available in two distinct configurations: n-i-p diodes and charge-selective contact devices (CSC).N-i-p diodes consist of a thin layer (tens of nm) of p-doped a-Si:H, a thicker layer (1-10 μm) of intrinsic a-Si:H, and a thin layer of n-type doped a-Si:H.
On the other hand, charge-selective contact devices [6] follow a three-layer structure, including a thin layer of metal oxides where hole mobility exceeds electron mobility (MoO  or WO), a thick layer of intrinsic a-Si:H, and a thin layer of a metal oxide where electron mobility is greater than hole mobility (such as TiO 2 or Aluminum-doped Zinc Oxide).
The electric contact of the sensors is provided by a layer of ITO (Indium Tin Oxide) deposited by sputtering.
The sensors used for the characterization described in this paper include 6 pads, differing in their active area: one with 5×5 mm 2 surface and five surrounding pads having 2×2 mm 2 surface.The thickness of the sensors described in this work are 2.5 μm or 5 μm depending on the tested production batch.
Sensors deposited on Kapton are fixed to a dedicated PCB, onto which bonds for the electrical contacts of the sensor will then added.
As illustrated in figure 1, it can be noted that the sensors are positioned within a dedicated opening within the PCB, on an additional layer of Kapton (which serves as support).This configuration not only allows for complete electrical isolation of the sensor but also provides a minimal material layer for possible dosimetry measurements behind the sensor.The Kapton tape, in fact, can be considered almost transparent and also "tissue-equivalent" for the X-ray spectrum used in these measurements.Its effect can thus be easily identified (and removed if necessary) during the analysis.
The bonds ensuring the electrical contact of the readout pads are soldered with tin soldering on the PCB and with silver-based glue on the sensor under test.The use of glue is necessary as ultrasonic soldering on a-Si:H is not advisable due to adherence problems on the ITO layer.
The various contacts are then grouped in a dedicated connector, one for each pad to be measured.The bias is provided to the pad while the common contact on the sensor film base (composed of n doped a-Si:H or TiO 2 ) is grounded.
-2 - The measurement setup includes a PCB board for the connection to SMU (Source Measuring Unit, Keithley 2410) that is used also as mechanical support.This assembly is located inside a climatic chamber designed to maintain stable temperature conditions.The radiation source is a X-ray tube (Newton Scientific XRSource 50 KV 10 W and a Semat 200 KV) or a 90 Sr electron source.
For the measurements presented, each data acquisition run has been divided into 2 steps.The first step involves IV characterization to obtain preliminary information about the dark current.This helps in selecting the correct bias voltage.The second step is essentially the characterization of the sensor's performance when irradiated by selected dose rates (X-ray radiation) or known electron flux from 90 Sr (31 KBq).
It is important to note that the measurements are conducted while keeping the sensor's position fixed, even during dark current measurement.This is done to ensure that X-ray characterization (or with other sources) is performed at a defined and consistently fixed distance (the setup is shown in figure 2a); figure 2b shows the calibration curve dose-rate in function of the tube current.Due to the long time-scale for a-Si:H electrical stabilization, it was necessary to set an initial stabilization time of 600 s for both IV measurements and characterizations with sources.
For the same reason, an additional stabilization time of 180 s was added into the measurement procedure at each voltage change.
The X-ray irradiation is performed at fixed voltage, and consequently, no additional stabilization was foreseen, unlike in the case of IV measurements.Both stabilization times were defined empirically and iteratively to achieve a stable dark current condition considered acceptable.
an example of the IV characterization of two sample sensors from the entire tested population is shown.As observed, the sensors generally exhibit high homogeneity in current measurement, given that they come from the same production batch.This enables the straightforward and relatively quick identification of faulty sensors within the tested population.In this case, the characterization procedure involves applying a maximum of 2 V/μm in reverse polarization (−5 V).To verify the rectifying properties of the sensors the IV measurement extended to 0.4 V/μm in direct polarization (1 V).It was decided not to exceed this value as higher forward polarization that could risk the damage of the sensor.The maximum expected current in reverse polarization is approximately 60 nA.
The current increase above 0.8 V/μm (−2 V) helps locating the working point at which the sensor is used during most tests.For some samples (2.5 μm thick sensors), an additional measurement was conducted at voltages up to 16 V/μm (−40 V).This type of measurement was carried out to assess the voltage range in which the sensor can be considered in a saturation condition.In figure 4, it can be seen that even after applying 16 V/μm, the sensors still do not exceed the current value of 100 nA.
-4 - In figure 5, the characterization of a sensor under X-ray exposure is presented.As evident, the sensitivity maintains linearity with the X-ray dose rate to which the sensor is irradiated.This behavior persists even under intense exposition to light prior to the measurement, but in this case a reduction in the slope is observed due to the effect of light exposure.It has been observed that the sensors can recover their maximum sensitivity after a 12-hour annealing session at 100 • C.This consistent behavior across various production batches led to the decision to implement a preliminary annealing session before each characterization session.The annealing procedure also helps to shorten the stabilization time of the signal as it can be observed in figure 5a and figure 5b.
After each measurement, the sensors are stored in individual specialized black boxes to prevent potential damage from light exposure or handling.

JINST 19 C04025
In figure 6, the current generated by the radiation from the 90 Sr source, with the contribution due to the dark current subtracted, is presented.Up to approximately 8 V/μm, there is minimal difference between the signal in irradiation condition and the dark current value.However, above 8 V/μm, an exponential trend is observable, suggesting the presence of a charge amplification process.The cause of the signal amplification may be attributed to an avalanche-like effect.The phenomenon is somehow unexpected, and the results are still preliminary.More in-depth studies will be conducted to clarify the presence of an avalanche-like effect.

Conclusion
The assembly procedures, electrical characterization of the HASPIDE sensors prototype have been presented.Additionally, measurements were conducted to characterize the sensor performance under X-ray radiation and electrons from a 90 Sr source.It was observed that the responses of the detectors to irradiation at different X-ray fluxes is linear in the range of the observations.Leakage current measurements were also performed and subtracted to the observed photocurrent.Effects from annealing on light exposed detectors were also shown demonstrating a substantial increment in sensitivity.Results from the irradiation with 90 Sr source at various bias voltages has been shown.The data exhibit an exponential increase of the signal with voltage possibly due to an internal amplification mechanism to be further investigated.

Figure 1 .
Figure 1.Example of two sample sensors (N-i-p on the left and CSC on the right) mounted on PCB.

Figure 2 .
Figure 2. a) Setup installed under the X-ray tube.b) Calibration curve dose-rate as a function of the tube current.

Figure 3 .
Figure 3. Example of IV characterization for two 5 μm thickness sensors.

Figure 4 .
Figure 4. Example of IV characterization for two sensors (2.5 μm of thickness).The first one (in blue) up to 5 V/μm, and the second one (green) up to 16 V/μm.

Figure 5 .
Figure 5. Example of X-ray characterization of a sensor.Figure a) shows real-time data acquisition of the sensor signal at various X-ray dose rates.The sensor was not subjected to annealing.In figure b) the same real-time data acquisition was repeated after an annealing session.The X-ray source activation is sequenced (On-Off in the graph) to optimize dead times due to the electrical stabilization of the sensor.In figure c), the sensitivity of a selected group of sensors before and after the annealing procedure is illustrated.

Figure 6 .
Figure 6.Exponential trend of the sensor signal (electron from 90 Sr) as a function of the applied electric field.The dark current was already subtracted from the signal.