In recent years, in studies of the interaction of X-rays with electronics, it has been concluded that X-rays can cause damage to electronic devices. Ideally, an X-ray detector should have a high sensitivity that captures every incident X-ray photon without any additional noise and without any degradation of performance during the trial. However, it is unlikely that such an ideal detector exists in reality.

Current image sensors can do as much as possible to achieve the above two points of high sensitivity, low noise, by taking certain measures to protect against irradiation to try to achieve the performance does not decay. Here we have to mention the seed player, the CMOS image sensor, which has a very high sensitivity and introduces very little additional electronic noise to the signal.CMOS image sensors are claimed to be irradiation tolerant up to 100 krad (10 kGy). However, 1Gy = 1 J-kg-1, which means that for every 1g the chip is exposed to about 1 Joule of energy before it fails. Thankfully, in real-world scenarios, where the intensity of X-rays is limited, the accumulation of irradiation dose is a very slow process.

Mechanisms of damage to chips by X-ray irradiation

Simply throw out a conclusion - integrated circuits fail after absorbing ionising radiation energy on the order of 1 joule. In reality irradiation damage is a gradual process, with many changes accumulating within an electronic device before it fails. For X-photons with energies below 1 MeV, their interaction with matter is mainly excitation (photoelectric effect) and scattering (Compton effect) of electrons outside the nucleus. The interaction with electrons in matter does not affect the lattice structure of the material or the degree of atomic ordering of the crystal, but only produces a large number of free electrons and positively charged ions or holes.

In other words, if the irradiated material is a conductor, the resulting positive and negative charges will quickly compound, resulting in a rapid return to electrical neutrality within the material. If the material absorbing the X-ray energy is an insulator, when the electrons are excited or scattered away, the positive and negative charges will not be able to compound quickly, leaving a permanent positive charge accumulation area within the material.

Integrated circuits rely on insulating media to separate conductors, semiconductor materials, in order to form the electric field inside the device.CCD and CMOS devices use silicon dioxide (SiO2) insulating media to separate the electrodes from the silicon substrate below. If X-rays interact with the SiO2 insulating layer and cause a positive charge to build up inside, this can cause the charge migration characteristics of the device to change and the threshold voltage of the transistor to slowly drift until the device is always on or completely off. Digital integrated circuits or analogue circuits of a particular design can tolerate a certain degree of threshold voltage drift and maintain normal operation. But when the transistor loses its ability to turn on and off due to charge build-up, the device will eventually fail completely.

In CMOS image sensors, the most significant effect of irradiation is dark current variation.

The dark current from the photodiode increases before the transistor in the sensor fails to function, which is an indication of the gradual accumulation of charge in the insulating medium. x-ray detectors for CMOS sensors have a very low starting dark current (on the order of 25 pA/cm2), so that even if they are used for a considerable period of time in irradiated environments, an increase in the dark current does not have any significant effect on the image quality. The mechanism by which the dark current increases as the irradiation dose accumulates is the accumulation of positive charges in the oxide insulating layer described above. The charge distribution at the boundary of the depletion region in contact with the photodiode is a significant perturbation of the internal electric field of the CMOS device. The accumulated positive charge increases the leakage current in the diode PN junction until the device fails completely when the diode leakage rate exceeds the charge readout rate.

Detector lifetime hours: The total lifetime of a radiation imaging detector depends primarily on the environment in which it is used.

For example, what is the energy spectrum of the radiation to which the detector is exposed in the application? What is its dose rate? What is the maximum X-photon energy? Is the radiation source additionally filtered? Are the X-ray exposures continuous, or do they follow a shot-by-shot pulse? How much time is the detector in exposure use during a typical day, week, or month? What type of subject is placed in front of the detector - does the detector routinely receive unobstructed X-rays, or is it primarily used to image thick metal plates shielded from most of the radiation? Is the detector designed to be radiation hardened, or does it inherently degrade rapidly under irradiation?

Finally, what are the criteria for "failure" of the detector?

All of these questions affect the length of a detector's life in a given environment.

Of course, many of these different conditions of use described above can be summarised by counting the total dose absorbed by the detector. However, different irradiation energy spectra will differ. For example, a detector used at a tube voltage of 50 kVp will have different results after receiving a certain dose than the same irradiation dose at 150 kVp. In addition, it is difficult to measure or calculate the actual energy absorbed in crystalline silicon. In practice it is often the case that the dose rate is assessed by measuring the exposure at the air, or rather at the incident surface of the detector. This type of measurement is adaptable to different applications as long as the ray energy spectrum does not change significantly.

You may have noticed that no matter what type of detector manufacturer, when giving the detector lifetime, the unit used is kGy, not 10 years. We need to understand the customer's usage scenario first and then assess a time accordingly. So when a customer asks us directly how many years the detector will last, all we can say is that it depends and we need more information.

Test conditions

In order to determine the extent of degradation of detector performance with increasing radiation dose, two different configurations of their detectors, the standard version and the version reinforced with radiation protection, were tested by foreign manufacturers under different conditions. Table 1 shows the different test conditions. The standard version of the detector was tested at 25, 45, 100 and 160 kVp ray energy spectra. The radiation hardened version (EV version) had essentially no performance degradation at 25 kVp and was therefore tested only at 45, 100 and 160 kVp.

Table 1 Test conditions

Standard test

The main radiation effect that limits the lifetime of a detector is the increase in its dark current with cumulative exposure . None of the detectors tested failed in any other way before dark current saturated the detector at room temperature . Of course, the saturation point can be delayed by cooling the sensor or reducing the integration time.    

The selected test standards are the detector dark field signal (measured in ADU) and the detector dark current (measured in electrons/second) with an integration time of 500 ms . The latter can be calculated from the difference between the dark field signals of two images taken at different integration times; for example, the difference between the dark field images at two integration times of 500 ms and 1500 ms is divided by the difference in integration time. The average gain (signal level under fixed exposure conditions), signal-to-noise ratio (under the same fixed exposure level) and spatial resolution (line-to-card, MTF measured by edge method) of some detectors were also measured.

Test Results

The test results are summarized in the charts shown in Figures 1 to 3. Since the dark field signal and dark current are strongly temperature dependent (dark current doubles approximately for every 8°C increase), the data were normalized to an average detector temperature of 25°C. Since these parameters do not change much with X-ray energy, typical response curves for flat-field signal and signal-to-noise ratio are shown. The MTF remains constant at both energy and total dose.