Designing image sensors for the ESA’s Euclid mission demonstrates the importance of customization for specialty uses and harsh environments. Adapting commercial electronic parts for space applications usually consists of little more than the addition of screening tests and environmental qualifications. But customizing image sensors from the design through to manufacture can optimize the performance of these devices for both the application and the space environment. Euclid is a European Space Agency (ESA) mission to map the dark universe, and it has been selected for launch in 2020. The Euclid space telescope is a product of a long-standing inquiry: “How did the universe originate, and what is it made of?” Its primary goal is to help advance understanding of “the nature of dark energy and dark matter by accurate measurement of the accelerated expansion of the universe through different independent methods.” The data generated by the space telescope will enable the testing of current theoretical models, including modifications to general relativity. The ESA’s Euclid satellite will map the dark universe; it will launch in 2020. Photo courtesy of European Space Agency The Euclid satellite will sit in the second sun-Earth Lagrangian point, approximately 1.5 million km away from the Earth. Euclid’s visible and near-IR imaging channels (VIS) instrument will survey 15,000 square degrees of the extragalactic sky for photons in the 550- to 900-nm-wavelength range; e2v is supplying 36 CCD image sensors to form the focal plane. The images will be used to measure the shapes of more than 2 billion galaxies accurately enough to assess the distortion created by gravitational fields resulting from dark matter, creating a dark-matter map of the universe. This CCD will be based on one that was developed for NASA’s Solar Dynamic Observatory, innovating to produce a unique instrument specifically optimized for Euclid’s goals as well as the challenges of the environment it will face. One important aspect of CCD performance is the charge transfer efficiency (CTE). This is the measurement of how well a CCD can move charge from its collection point (pixel) to the output of the sensor so users can obtain the data. The charge packet is transferred from one pixel to the next by changing voltages in sequence. The VIS focal plane is required to detect a wide range of light sources, right down to the very faint, equivalent to only tens of electrons (very small amounts of charge). Each time the charge packet is transferred from one pixel to the next, it can lose some of its charge. Euclid will include 36 CCD image sensors that will form the focal plane. Here, e2v’s CCD273-84 is shown with handling jig and protective cover (a), and with the protective cover removed (b). The CTE is the proportion of charge successfully transferred each time. Poor CTE can lead to the loss of important information regarding the intensity of the source (particularly relevant when detecting faint stars) and charge trailing, resulting in a blurred image. Therefore, the CTE is a critical characteristic when looking at small distortions of distant galaxies. The main cause of poor CTE is traps, which capture electrons and then release them after a period of time. Traps can be created as a result of chip design, processing of the silicon and lattice defects, or impurities in the silicon. Advances have led to high-quality silicon with very few defects or impurities, but the nature of the Euclid environment can counter this. Space offers many challenges that terrestrial systems do not encounter: The sun emits large amounts of charged particles, but here on Earth, we are protected by a magnetic field. In space, these particles can penetrate spacecraft and cause damage. Although the presence of the Earth between the Euclid spacecraft and the sun will offer some protection, these charged particles can cause ionization in the oxides within image sensors or alter the silicon structure, creating a significant number of additional traps. To reduce these effects (and those of CTE degradation of the image sensors over time), several techniques must be incorporated specifically for the Euclid mission. The scientists’ requirements for the CCD mean that the channel that conducts the charge does not have to be as large as that of our standard designs. Narrowing this channel can reduce the number of radiation-induced traps that the charge packets encounter during transfer, so that the image sensor continues to have high performance throughout the mission. In addition, a charge-injection structure has been added to the CCD, which allows the user to inject a defined signal level into the image area of the CCD before an image is taken. This artificial background level fills some of the traps to prevent them from trapping the charge from the actual image, and therefore the essential data remains intact. Ionization of the oxide structures in the CCD can result in changes in the capacitances, which in turn alter the size and shape of the charge packets. The Euclid CCDs also have thinner oxide structures to lower the amount of charge that can build up and reduce these changes resulting from radiation damage. The thinner oxide allows for lower-voltage operation, reducing power consumption. This is an added benefit, but there are sacrifices to consider within the design. A few rows of pixels are lost to allow for the charge-injection structure, and the narrow channel means that less signal can be transferred out of the device. These are factors that are weighed specifically for each project. It is not only sensor design to be considered; packaging of the image sensor must be tailor-made for the mission. The Euclid CCDs are designed to operate at temperatures far below -100 °C and to survive shocks and vibrations associated with a launch burning several tons of rocket fuel. The materials for the package must be strong and durable, and must not produce any substances that may cause damage to the spacecraft. Because of the wide-ranging temperature exposure, the materials for the Euclid CCD were specifically chosen to have similar thermal-expansion characteristics to avoid significant stress on the device. Each image sensor will be integrated into a large focal plane within the VIS instrument. This focal plane must be extremely flat, and each CCD sensor must be as close to its neighbor as possible to reduce any “dead space” where the array is not sensitive to the incident light. To achieve this, the Euclid image sensor was designed with precision ground shims for submillimeter height accuracy and a closely fitted package to allow the edge of each sensor to be within less than a couple of millimeters of the adjacent one. The package that supports the CCD sensor is much thicker than the sensor itself, but has pockets machined to reduce mass, essential for space applications – but it retains significant structural strength to resist mechanical deformation from cooling to cryogenic temperatures. The Euclid CCD is robust to radiation damage, sub -100 °C temperatures and mechanical shock, and can be tightly packed into an array of 49 such devices designed to accurately map 2 billion distant galaxies to unprecedented precision. This CCD will give the ESA the opportunity to test the theories proposed by physicists around the world, advancing our understanding of the universe. Meet the author Andy Grey is a technical authority in e2v’s High Performance Imaging Solutions division in Chelmsford, England; email: andy.grey@e2v.com