Raman spectroscopy is a hugely informative tool with a plethora of applications from biomedicine to analytical chemistry. Potentially, the technique could improve liver transplantation success rates through investigating Raman signals associated with metabolic changes prior to transplant rejection. However, studying biological systems is challenging since background fluorescence dominates the weak Raman signal. Thus, there is a need to improve signal-to-noise and Raman-tofluorescence ratios and drive down spectral acquisition times. Pulsed lasers combined with time-resolving single photon avalanche diode (SPAD) detection systems have been shown to enhance Raman and fluorescence discrimination. We report significant advances in time-correlated single photon counting (TCSPC) Raman spectroscopy using a laser exhibiting up to 200 W peak power and 40 MHz repetition rates in combination with a 512 spectral channel, 16.5 gigaevent/s throughput SPAD histogramming line sensor. Using a diamond sample, we report 0.4 MHz Raman count rates, millisecond spectral acquisition times, and signal-to-noise ratios of over 200. We demonstrate simultaneous, singleexposure acquisition of Raman and fluorescence signals in sesame oil. Time-based Raman-fluorescence discrimination techniques are subject to fluorescence signal tail influences from previous pulses, and data obtained with laser periods of 25 ns and 50 ns are presented. We achieved optimised Raman-to-fluorescence ratios through adjustment of histogram bin positions in 63 ps increments. Achieving high count rates while discriminating fluorescence from Raman signals unlocks the potential of combined Raman/fluorescence lifetime spectroscopy for biomedical imaging applications.
Time-resolved separation of Raman scattering from background fluorescence is demonstrated using a recently developed 512 pixel, 16.5 giga-events CMOS SPAD line sensor 1 . The system is being developed with the aim of evaluating the suitability of liver tissue for transplant surgery, as these samples pose particular challenges to commercial Raman systems due to their high fluorescence emission across a wide spectral excitation range. Compared to previous work 2 , the spectrometer operates without the use of time-gating, relying entirely on inpixel time-correlated single photon counting (TCSPC). Indeed, by employing the unique features of the sensor, such as on-chip histogramming and zoomable time resolution from 50ps to 6.4ns, the system is able to deliver both Raman and time-resolved fluorescence decay data. Time-resolved separation of Raman and fluorescence signals allows the spectrometer to be operated in the visible range (using a 532-nm pulsed laser), thus providing enhanced Raman scattering intensity compared with the use of a near-infrared laser, since scattering emission is proportional to λ-4 3 . The system is calibrated using a Neon calibration source and benchmarked using samples of pure distilled-water fluorescein, paracetamol and sesame oil in comparison with results from continuous wave excitation in a Renishaw InVia spectrometer. The Raman band of water at a Raman shift of 3000-3800 cm-1 is chosen to evaluate sensor performance because of its low intensity and its characteristic spectral profile which is readily compared with the literature 4 .
KEYWORDS: Fluorescence resonance energy transfer, Sensors, Luminescence, Single photon, Time resolved spectroscopy, CMOS sensors, Spectroscopy, Molecules, Molecular energy transfer, Time correlated photon counting
We demonstrate a new 512x16 single photon avalanche diode (SPAD) based line sensor with per-pixel TCSPC histogramming for time-resolved, time-zoomable, FRET spectroscopy. The line sensor can operate in single photon counting (SPC) mode as well as time-correlated single photon counting (TCSPC) and per-pixel histogramming modes. TCSPC has been the preferred method for fluorescence lifetime measurements due to its collection of full decays as a histogram of arrival times. However, TCSPC is slow due to only capturing one photon per exposure and large timestamp data transfer requirements for offline histogramming. On-chip histogramming improves the data rate by allowing multiple SPAD pulses (up to one pulse per laser period) to be processed in each exposure cycle, along with secondly reducing the I/O bottleneck as only the final histogram is transferred. This can enable 50x higher acquisition rates (up to 10 billion counts per second), along with time-zoomable histogramming operation from 1.6ns to 205ns with 50ps resolution. A broad spectral range can be interrogated with the sensor (450-900nm). Overall, these sensors provide a unique combination of light sensing capabilities for use in high speed, sensitive, optical instrumentation in the time/wavelength domain. We test the sensor performance by observation of fluorescence resonance energy transfer (FRET) between FAM and TAMRA and between EGFP and RFP FRET standards.
KEYWORDS: Sensors, Spectrometers, Signal to noise ratio, Luminescence, CMOS sensors, Single photon, Spectroscopy, Imaging spectroscopy, Interference (communication), Time resolved spectroscopy
Time-resolved spectroscopy in the presence of noise is challenging. We have developed a new 512 pixel line sensor with 16 single-photon-avalanche (SPAD) detectors per pixel and ultrafast in-pixel time-correlated single photon counting (TCSPC) histogramming for such applications. SPADs are near shot noise limited detectors but we are still faced with the problem of high dark count rate (DCR) SPADs. The noisiest SPADs can be switched off to optimise signal-to-noiseratios (SNR) at the expense of longer acquisition/exposure times than would be possible if more SPADs were exploited. Here we present detailed noise characterization of our array. We build a DCR map for the sensor and demonstrate the effect of switching off the noisiest SPADs in each pixel. 24% percent of SPADs in the array are measured to have DCR in excess of 1kHz, while the best SPAD selection per pixel reduces DCR to 53+/-7Hz across the entire array. We demonstrate that selection of the lowest DCR SPAD in each pixel leads to the emergence of sparse spatial sampling noise in the sensor.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.