Direct Detection

In direct detection sensor technology, X-ray photons are directly converted into electric charge in the solid-state sensor (Fig. 1). In contrast, classical indirect detection sensors use scintillators with fiber optics or lenses to transform X-ray photons into light which is then converted into electric charge in the optical solid-state sensor.

Direct detection sensor technology prevents a number of image artifacts from being introduced by scintillators and fiber optics, such as fixed patterns, gain variations and distortion.

Direct detection sensors are fully depleted monolithic sensors and are typically fabricated from high-resistivity solid-state sensor materials such as silicon. Ohmic and diode type pixelated sensors with various sensor thicknesses are available with ohmic contacts or diode structures in each single pixel. With respect to quantum efficiency, a typical 450 µm thick diode-type silicon sensor yields maximum performance for experiments in the energy range of 3 to 15 keV. However, the sensor can be used for energies of up to 40 keV and thicker sensors are available for high-energy applications.

Fig. 1: Principle of direct detection of X-ray photons in a solid-state sensor.

Hybrid Pixel

In hybrid pixel technology, a pixelated solid-state sensor is connected to a CMOS readout ASIC by bump or wire bonding technology (Fig. 2). The sensor, which is typically a one- or two-dimensional array of pn-diodes processed in high-resistivity silicon, is connected to an array of readout channels designed in CMOS technology. Each sensor element is typically wired to its corresponding readout channel through a micro-bump bond, e.g., a microscopic indium ball with a diameter of about 18 µm. This specific micro-bump bonding process has been refined at the Paul Scherrer Institut. The great advantage of this approach is that well-proven standard technologies can be independently employed for both the sensor and the CMOS readout ASIC, thus guaranteeing the highest level of quality. Both of these technologies can be optimized separately since the best substrate for X-ray sensors and for high-speed and low-noise CMOS circuits are very different. Moreover, the pixel’s small size and interconnection results in a very low input- node capacitance, which has the beneficial effect of reducing the noise and power consumption of the pixel readout electronics.

Fig. 2: Principle of a hybrid pixel with a sensor bump bonded to a CMOS readout ASIC.

Single-Photon Counting

In single-photon counting technology, every photon is detected individually and counted in the discriminating detector (Fig. 3). The required signal processing of the electric charge generated in the sensor is included in each readout channel of the CMOS readout ASIC. The readout channel comprises an amplifier, which amplifies the electric charge pulse generated in the sensor by an incoming X-ray photon, a discriminator, which generates a digital pulse signal if the incoming charge pulse exceeds an adjustable predefined threshold, and a digital counter, which counts the number of generated digital pulses. In Fig. 3, the first diagram shows the sequence of incoming X-ray photons with their corresponding energy. The second diagram shows the analog voltage pulses at the output of the amplifier and the predefined threshold voltage of the successive comparator that acts as the discriminator. The fourth diagram shows the digital voltage pulse sequence that appears at the output of the comparator and that feeds the counter in order to determine the number of incoming X-ray photons with energy higher than the predefined threshold energy. This represents a completely digital detection and storage scheme and achieves a noiseless determination and readout of the number of detected X-ray photons per pixel.

Fig. 3: Principle of single-photon counting detectors vs. integrating detectors.

In contrast to single-photon counting technology, classical integrating detectors such as charge-coupled devices or CMOS active pixel sensors accumulate the electric charge generated in the sensor as shown in the third diagram of Fig. 3. An additional high-resolution analog-to-digital converter (ADC) is required to convert the integrated charge into a digital output value after the exposure time in order to determine the intensity of the incoming X-ray photons. In addition to the primary signal, the integration process accumulates undesired signal contributions such as leakage current, noise, background and interferences. This, together with the effective resolution of the ADC, greatly affects the overall performance of the integrating detector and limits the achievable signal-to-noise ratio and dynamic range.

The single-photon counting principle features zero dark signal and zero readout noise and achieves excellent signal-to-noise ratio. In addition, the energy threshold provides energy resolution and a single energy threshold can be used for low-energy suppression. Due to the completely digital detection and storage scheme, single-photon counting achieves short readout time and high frame rates. High-performance single-photon counting detectors typically include counter sizes in the range of 20 bits, dynamic ranges of up to 1,000,000, readout times of less than 1 ms and frame rates of several 100 Hz. The maximum count rates, typically about 106 photons per second in a single pixel, allow the handling of the high flux of modern synchrotron light sources.

Modular Detector

In modular detector technology, the complete active area and pixel array of a large-area detector consists of multiple identical modules of a predetermined size. The detector module is the fundamental unit of the multi-module detector and includes a multi-chip module with a single monolithic sensor and a number of CMOS readout ASICs, a mounting bracket and control electronics. The detector modules are mounted onto a high-precision mechanical frame to create multi-module detectors with up to 60 modules or more.

For example, PILATUS detector modules include a multi-chip module with a single sensor and an 8 x 2 array of CMOS readout ASICs assembled by bump-bonding technology (Fig. 4).

Fig. 4: Multi-chip module with 16 CMOS readout ASICs bump-bonded to a single sensor.

Each sensor is a continuous 487 x 195 array of 94,965 pixels without dead areas and covers an active area of 83.8 mm x 33.5 mm. The multi-chip module is wire-bonded to the mounting bracket with the control electronics and forms the PILATUS detector module (Fig. 5). Multiple PILATUS detector modules are assembled in a multi-module setup to form large-area PILATUS detectors (Fig. 6).

Fig. 5: PILATUS detector module.
Fig. 6: PILATUS detector modules assembled in a multi-module setup in a large-area PILATUS detector.

As another example, MYTHEN detector modules include a single sensor and a linear array of 8 CMOS readout ASICs assembled by wire bonding technology. The sensor is a linear array of 1280 microstrips on a 50 um pitch with a total active area of 8 mm x 64 mm. Multiple MYTHEN detector modules can be combined and operated with a single detector control system to form a large multi-detector system (Fig. 7).

Fig. 7: MYTHEN detector system consisting of a detector control system and one multi-chip detector module (without housing).