CMOS lock-in optical sensor for parallel detection in pump-probe systems Roger A. Light* a , Richard J. Smith b , Nicholas S. Johnston a , Michael G. Somekh a and Mark C. Pitter a a IBIOS, University of Nottingham, University Park, Nottingham, NG7 2RD, UK; b Applied Optics, University of Nottingham, University Park, Nottingham, NG7 2RD, UK ABSTRACT In pump-probe type experiments the signal of interest is often a very small fraction of the overall light intensity reaching the detector. This is beyond the capabilities of conventional cameras due to the necessarily high light intensity at the detector and its limited dynamic range. To overcome these problems, phase-sensitive or lock-in detection with a single photodiode is generally used. In phase-sensitive detection, the pump beam is modulated and the probe beam is captured with a photodiode connected to a lock-in amplifier running from the same reference. This provides very narrowband detection and moves the signal away from low frequency noise. We have developed a linear array detector that can perform shot-noise limited lock-in detection in 256 parallel channels. Each pixel has four independent wells to allow phase-sensitive detection. The depth of each well is massively increased and can be controlled on a per-pixel basis allowing the gain of the sensor to be matched to the incident light intensity, improving noise performance. The array reduces the number of dimensions that need to be sequentially scanned and so greatly speeds up acquisition. Results demonstrating spectral parallelism in pump-probe experiments are presented where the a.c. amplitude to background ratio approaches 1 part in one million. Keywords: CMOS image sensor, active pixel sensor, modulated light, phase sensitive detection 1. INTRODUCTION Scientific imaging often poses a set of problems that are not well catered for by conventional commercial cameras. Two main problem areas exist: where the overall light level is very low and where there is a high light level at the detector, but only a small proportion of this is the signal of interest. Conventional cameras are designed for imaging the every day world, where contrast is relatively high and the dynamic range requirements of the camera are modest. In most cases the noise performance of the device is much less important than the size and number of pixels. In low light experiments even the brightest of signals is small so great care must be taken to reduce noise. The dominant noise sources here are dark current and read noise. In astronomy, detectors are typically cryogenically cooled to reduce the dark current to a minimum, and are read out very slowly to reduce the read noise 1,2 . Another example of this type of imaging is found in fluorescence microscopy in the biology field, where the signal of interest is produced with fluorescent markers placed selectively in a sample. The fluorescence caused by the markers is very dim. Conventional cameras are not suited to this type of work, but there are alternatives such as the electron multiplying CCD which is popular because of its superior low light performance. We are concerned with experiments with high light levels where the signal of interest is a small proportion of the background. To make these kinds of measurements, a high dynamic range is required. There is a lot of interest in making high dynamic range cameras using techniques such as taking multiple exposures 3,4 , using multiple integration wells per pixel for low and high light conditions 5 , using pixels with logarithmic response 6 or with combined linear and logarithmic response 7 . Although these cameras tend to offer impressive dynamic range results, the design intent of the pixels is to cope with a large range of light intensities in a single frame. They aren’t intended to cope with small changes in int ensity across the whole of their working range and as such their optimizations aren’t useful here. This problem of very low contrast is common in pump-probe type experiments 8 and is traditionally solved by using a single photodetector connected to a lock-in amplifier. The optical input to the experiment is modulated by some means *roger.light@nottingham.ac.uk; phone 44 115 8468848; fax 44 115 9515616; nottingham.ac.uk/ibios