All surfaces in our surrounding radiate heat as an electromagnetic wave - similar to light, but invisible to our eyes. The ability to sense and quantify thermal radiation opens up a vast field of application scenarios such as person detection, thermal analysis of buildings or equipment, and health monitoring.Especially the latter has become a matter of utmost importance with the sudden outbreak of the SARS-CoV-2 pandemic in the beginning of 2020. In such an event, the most important task is to closely observe the spreading of the infection and to initiate adequate measures for social distancing.In many cultures, monitoring the health of citizens is a legally delicate issue, and prescribing confinement to an individual would raise difficult questions when it comes to compensate for lost wages. However, having precise instruments at hand to monitor and control the pandemic would allow governments to provide individual support rather than bringing the whole economy to a shutdown.The widespread availability of medical grade temperature sensors could contribute valuable information to a multi-parametric pandemic monitoring. Many infections are accompanied by an elevated body temperature, and often fever is not only one of the most obvious symptoms but also an early stage indicator. Hence, there is a renewed interest in ultra-small medical grade IR sensors which can be integrated into mobile consumer products such as wearables and smart phones.In order to produce IR sensors for a global mass market, a thorough consideration of all aspects on the system level is necessary to optimize cost, size and performance for the targeted application. Infrared detectors for temperature measurement have not yet reached high-volume applications for two main reasons: (1) The optical design necessitates far-infrared transparent materials like ZnSe, CaF or Ge, which are expensive and not always robust for the target application. (2) The state-of-the-art assembly and packaging mostly uses large metal housings (“TO cans”) for their vacuum capability. The size and price of such housings definitely prevents their integration into extremely compact, mobile devices.We are presenting a technological approach that allows effective scaling of production volumes into high quantities by using a fully wafer-based optical assembly and integration using spherical silicon lens preforms. The complete vacuum packaging and optical alignment are performed as a CMOS post-process in our MEMS cleanroom, using industrial state-of-the-art equipment for 200 mm wafers.Considering a spot thermometer where light just needs to be concentrated, convex lenses can either be manufactured as Fresnel lens or as a spherical lens. Both types can be produced only laboriously with standard Silicon technology using dry etching processes and generally multiple lithography cycles when it comes to larger apertures. For a small F number, the Fresnel lens would rather generate stray light than concentrate incoming radiation on the detector element. Additionally, it is not sufficient to just produce a lens - obtaining a precise spherical shape with defined radius of curvature with good uniformity across the wafer is challenging with etching processes. Expensive equipment would be needed for 100% optical test coverage and any lens failure or parametric deviation would equally bring down the CMOS yield when using wafer-to-wafer bonding.In our paper, a detailed review of the IR packaging solutions, lens materials and their fabrication process will be given. A novel technology platform being developed at Fraunhofer ISIT together with the end user will be presented. At the core of the fabrication process, preformed Si spheres with a defined curvature are used that are subsequently grinded to form multiple plano-convex lenses on the wafer level. The process is a new element in our glass forming process portfolio that already enabled many technological innovations in wafer-level optics. Using spherical preforms results in excellent optical properties for IR spot detectors and eliminates the need for expensive 100% optical testing. Although certain limitations with respect to the pitch and the availability of intermediate sphere sizes exist, a clever design (e.g. focal length adjustment, placement of IR pixels etc.) allows covering a commercially very relevant spot on the IR detector market. Figure 1
Read full abstract