Lens-free microscopy on a chip
According to the National Institutes of Health, the emphasis of health care is shifting toward prevention and early disease detection, as well as management of chronic conditions.1 In that context, point-of-care (PoC) testing can enable a more patient-centered approach to health-care delivery by providing immediate results in nonlaboratory settings. However, there has yet to be developed an adequately compact, affordable, and easy-to-use PoC diagnostic device that can—for example—provide instant blood analysis results. The future of PoC diagnostics lies in novel silicon technologies. These take advantage of device scaling combined with enhanced sensing functionality and reduced cost per unit due to monolithic integration and mass production.2
A microscope-on-chip would offer an essential building block for many life science and industrial applications. However, only a fully integrated solution can benefit from the advantages of silicon scaling. Lens-free in-line holographic microscopy (LHM) is an imaging technique that requires no lenses or high-precision mechanical components3 and is therefore an ideal candidate for further miniaturization.
Our goal is to build a fully integrated, mass-producible LHM module that can be customized for different applications. Examples include an affordable benchtop microscope (see Figure 1) or a compact blood analysis device.4, 5 However, achieving this requires the development and combining of several different technologies: suitable image sensors, compact yet high-quality light sources, and systems with sufficient computational power for LHM image reconstruction.

On-chip light source integration is one of the main challenges to realizing a microscope-on-chip. The basic functionality of the source in our LHM system is to deliver a controlled wavefront of coherent light onto the sample and the image sensor. Uniformity of the wavefront shape (phase) and intensity, as well as sufficient spatial and temporal coherence, are the main requirements that have a direct impact on the imaging capabilities of the microscope. The on-chip integration brings further challenges, as all building blocks need to be part of a single miniaturized module. Semiconductor lasers in the visible or near-IR range (532 or 635–660nm) implemented in the form of edge-emitting laser diodes or vertical-cavity surface-emitting lasers are suitable sources of coherent light for life science applications. In our solution, we plan to use flip-chip bonding (with soldering) to integrate the laser with the rest of the system.6 The main challenge lies in increasing the light coupling efficiency between the laser and the photonic in-coupling grating of the light delivery network.
A light distribution network, based on single-mode photonic waveguides and processed with lithographic precision, can deliver light to different subsystems in a controlled way with minimal losses.7 Fractal distribution networks8 are suitable for applications that require high uniformity of illumination and lower out-coupler density. Alternatively, we can implement high-density arrays using evanescently coupled distribution.7 Here, achieving uniform illumination over large areas is more difficult due to the inherent exponential intensity decay. The main challenges lie in increasing the processing uniformity to reduce optical losses, but also in design and development of novel distribution architectures for high-density phased arrays that will enable implementations requiring an extremely compact large-field-of-view LHM.
To couple light out of the distribution network, we are able to design photonic structures to create almost any complex wave-front by modulating the phase and intensity of light.9 For example, we can generate a submicron focused spot with Gaussian intensity distribution and high numerical aperture just a few micrometers above the chip surface, creating the perfect illumination element for our point source-LHM system. For the large-field-of-view LHM application, a collimated beam of light emitted from the chip surface at a predefined angle and in phase with all the other beams creates a scalable large-area light source.
Our first prototype devices, fabricated using silicon nitride photonic process technology, show promising results in terms of phase and intensity control, as well as overall efficiency. Figure 2 shows an example of such a device. The next-generation photonic prototypes for microscope-on-chip are designed to optimize and extend the light-steering capabilities. Fabrication and characterization of these components are currently ongoing.

Part of this work was supported by the European Research Council under the Consolidator Grant SCALPEL (617312) to Liesbet Lagae.
Richard Stahl is a senior researcher in the Integrated Imaging team. His main research interests are holography, diffractive optics, and their application in display and imaging technologies.