When you think of any sort of optical imaging arrangement, you naturally assume there’s a lens in the path to focus the light onto the photon-sensitive surface, whether it’s old-fashioned film, CCD, CMOS, or other. The need for the lens adds size, weight, assembly issues, and cost to devices, but it seems to be an unavoidable necessity.
But maybe that is not going to be the case in the future. A team at the California Institute of Technology (CalTech), led by Ali Hajimiri, Bren Professor of Electrical Engineering and Medical Engineering in the Division of Engineering and Applied Science, has built a demonstration lensless imaging device using standard silicon-based photodetectors in an 8 × 8 array, Figure 1 . By “stepping” through these in a carefully timed sequence, they are able to focus and capture an image. Admittedly, the 64-pixel resolution of this array is crude, but it’s a starting point for the proof of concept, Figure 2 . (The team hopes to next scale up to larger receivers with higher resolution and sensitivity.)
The 8 x 8 OPA chip, placed on a penny for scale, is a first-pass effort at a lensless imaging arrangement. (Credit: Caltech/Hajimiri Lab)
The OPA chip within its larger housing and in its test and evaluation arrangement. (Credit: Caltech)
In simplest terms, the method they use is a variation on the phased-array technique, which we generally associate with sonar/ultrasound, radar, and even MIMO RF antennas. In a phased-array configuration, the array is electronically steered by precisely controlling activation timing of each element and thus their phase relationship. Here, the elements are not RF antennas or ultrasound transducers, but an ultra-thin optical phased array (OPA). The timing of the OPA scan must be controlled with femtosecond accuracy.
As with any phased-array system, it is possible to dynamically steer the imaging angle, and also adjust the field of view from narrow to wide, by adjusting which elements are activated along with their phase (timing) relationship, something which cannot be done with a fixed-in-place conventional lens. Further, the all-silicon/no-lens approach has the potential to be a much-thinner imaging implementation, which would allow thinner phones. While it’s not clear if that’s a priority application, it would also enable rethinking of how and where imaging subsystems are placed in many other products.
Making this lens-less camera takes much more than just the imaging sensors and timing control. It requires analog mixers, amplifiers, filters, and more. Although the full paper ” An 8X8 Heterodyne Lens-less OPA Camera” is behind a paywall, the researchers’ paper on its predecessor, a linear, 32-element array “A One-Dimensional Heterodyne Lens-Free OPA Camera” reveals details about the circuitry needed. That design used heterodyne architecture with adjustable phase shifters in the local oscillator (LO) path to improve the sensitivity of the receiver, reject leakage pickups by the receiver photodiode, and allow for easier amplification, Figure 3 . The earlier device achieved a steering range in excess of 60o with no blind spots, and resolved a beam width of 0.74o in 30o field of view.
(a) The heterodyne receiver diagram, (b) the block diagram of the OPA receiver, and (c) die photo of the fabricated 1.2 mm × 1.0 mm, 32-element linear lensless array. (Credit: Caltech/Hajimiri Lab)
Will all-silicon, multi-sensor arrays and associated analog circuitry replace today’s lens-based imaging subsystems? I certainly don’t know, and I doubt anyone does. When it comes to technology advances and their adoption, it’s important to “never say never”, however contradictory as that may sound. After all, it wasn’t that long ago that the expert opinion was that CMOS-based sensors would not replace CCD-based ones for true high-performance imaging, at least for “the foreseeable future.” We did now know that the foreseeable future was only close by, as CMOS sensor technology is used in all but the most demanding or specialty applications.
What’s your view on the practicality and possibilities for this lens-less, phased-array imaging scheme?