As a bread-and-butter diagnostic tool in biomedicine, optical microscopy has fueled spectacular progress in unraveling the complexity, diversity and the physiological processes of biological tissues and cells. While most of the current research in optical microscopy has overwhelmingly emphasized improving the spatial resolution, improvement in temporal resolution enabling studies of dynamical processes, especially non-repetitive transient phenomena, is of equal importance.
However, conventional charge-coupled devices (CCDs) and their complementary metal oxide semiconductor (CMOS) counterparts, the major workhorse in microscopy, are incapable of capturing fast dynamical processes with high sensitivity and resolution. This is mainly due to the fundamental trade-off between sensitivity and frame rate - at high frame rates, fewer photons are collected during each frame - a predicament that affects virtually all optical imaging systems.
Our goal is to develop a new type of ultrafast and sensitive imaging system built on the recent demonstration of an entirely new imaging modality by the principle investigator (PI) and co-workers that has established the world record for frame rate and sensitivity for continuous real-time imaging, called Serial Time-Encoded Amplified Microscopy (STEAM). This approach employs optical image amplification and other innovations to overcome the aforementioned trade-off between imaging sensitivity and speed. It captures the images with a single-pixel detector, eliminating the need for the CCD/CMOS imager and its associated speed limitations. The STEAM prototype developed here, in its native form, could play a significant role in non-invasive in-vivo reflectance confocal imaging, which has been employed in a myriad of clinical applications, e.g. to monitor fast response to therapeutic treatment such as laser surgery.
Get your grade
or your money back
using our Essay Writing Service!
Building on this STEAM prototype, we will also explore the potential of incorporating STEAM with quantitative phase microscopy (QPM), which can provide quantitative cellular information at nanometer scale without the need for any cell preparation or using exogenous labels. This combined system thus possesses high-speed and high-content (quantitative) imaging capabilities, which pave the way to achieving fast, high-throughput, sensitive, quantitative and label-free cell imaging and characterization in biomedicine, e.g. flow cytometry.
High speed optical microscopy has been an important but challenging area to study dynamical processes, especially non-repetitive transient phenomena. Examples are the spatiotemporal study of biochemical waves in cells/tissues [1-2], and high-throughput cell characterization (e.g. flow cytometry [3-4]). However, conventional CCD/CMOS imagers, the major workhorse in microscopy [5-7], are incapable of capturing fast dynamical processes with high sensitivity and resolution. This is mainly due to the fundamental trade-off between sensitivity and frame rate - at high frame rates, fewer photons are collected during each frame - a predicament that affects virtually all optical imaging systems. This project aims to develop a new high-speed imaging tool that is able to overcome this limitation. The proposed imaging system is built on the recent demonstration of an entirely new imaging modality by the PI and co-workers that has established the world record for frame rate (> millions frames per second) and sensitivity (with image amplification in the optical domain) for continuous real-time imaging, called Serial Time-Encoded Amplified Microscopy (STEAM) [8-9]. It encodes the image onto the spectrum of a laser pulse and then converts the spectrum into an optically-amplified time-domain serial data. This allows us to capture images with a single-pixel detector, eliminating the need for the CCD/CMOS imagers and their associated trade-off between imaging sensitivity and speed. Specifically, we aim to:
Develop the amplified dispersive Fourier transform (ADFT) technique in near infra-red (NIR) -Amplified dispersive Fourier transform (ADFT) is a powerful technique which was originally developed for ultrafast real-time spectroscopic measurements with spectral acquisition rate of >MHz, a speed not achievable with conventional spectrometers [10-12]. It is the central element enabling ultrafast imaging in STEAM. ADFT has been demonstrated mainly in the wavelength range of ~1500 nm, we here, in contrast, plan to develop a new ADFT system operating in the shorter wavelength range (~700-900nm). It is well-known to be the preferable working regime, compared with the 1500-nm range, for biomedical applications because of the considerably less optical loss (and the associated sample heat-up/damage) due to water absorption [13-14] and the higher achievable diffraction-limited spatial resolution in microscopy .
Develop the STEAM systems in NIR - Another key feature in STEAM is the mapping of the image into the spectrum of the optical pulse, i.e. spectrally-encoded imaging [15-19]. The spectrum, which is encoded with the image, can thus be measured by ADFT, allowing unprecedented imaging speed. We will design and build the spectrally-encoded imaging system in NIR range. As a result, combining ADFT with this spectrally-encoded imaging technique, we will develop a new STEAM system operating in NIR range with improved image quality in contrast to the prior proof-of-principle demonstration [8-9], and with high-speed (up to MHz) and sensitive imaging (with image amplification in the optical domain) capabilities.
Always on Time
Marked to Standard
Explore the potential applications of STEAM - The above prototype of STEAM, in its native form, is essentially a type of reflectance confocal microscopy (RCM) which has been extensively used in numerous non-invasive, label-free, in-vivo bio-imaging applications such as diagnosis of skin tumors [20-22]. However, its imaging speed has to compromise the weak detection sensitivity, which is limited by the weak back-scattered light (< few %) from the biological tissue. Hence, STEAM can offer orders-of-magnitude improvement in imaging speed and sensitivity in these applications. We will also investigate the potential of STEAM in other more compelling applications. One example is the combination of STEAM and quantitative phase microscopy (QPM). QPM has been proven to be a powerful technique to provide quantitative evaluation of cellular morphology and function at nanometer scale without the need for any cell preparation or using exogenous labels [23-27]. Hence, such a combined system could impact imaging flow cytometry applications, which demand rapid, high-throughput and quantitative characterization of different cells, e.g. circulating tumor cells .
LONG TERM IMPACTS
STEAM is an entirely new imaging modality which overcomes the fundamental trade-off between imaging sensitivity and speed in the conventional imagers. In the long term, it will clearly have a transformative impact on having better understanding of cellular dynamics in biological systems and in clinical cell analysis in biomedicine, e.g. flow cytometry, by increasing the throughput and reducing the associated costs. Utilizing STEAM in a wide range of applications could lead multidisciplinary collaboration opportunities (among the field of engineering, physics, chemistry and medicine) in local and international universities/industries. The study of our new architecture of STEAM will also show a strong justification to pursue patent applications. All these are indispensable in supporting Hong Kong's migration to a knowledge-based economy.
Recent breakthroughs in optical microscopy have resulted in dramatic advances in science and medicine. While most of the current research has overwhelmingly emphasized improving the spatial resolution [29-32], improvement in temporal resolution is of paramount importance. High-speed imaging enables studies of dynamical processes, especially non-repetitive transient phenomena. It is becoming increasingly important in microscopy because over the Î¼m-scale, even slow moving phenomena require high temporal resolution. One example is the spatiotemporal study of biochemical waves in cells and tissues, which requires imaging with a Î¼s- to ns-response time and is crucial for studying cell signaling and drug trafficking [1-2]. Another key application is flow cytometry [3-4], where high-speed cameras are required to provide high-throughput cell characterization. The following sections describe the existing techniques for high-speed imaging and other tools related to our proposed imaging system, STEAM:
Existing imagers - The CCD or CMOS imager is by far the most widely deployed optical imaging technology. The high-end versions can operate at tens or even hundreds of kHz by reducing the number of pixels that are downloaded from the sensor arrays. In particular, a CMOS imager (a faster, but less sensitive version of the CCD) operating at up to 675 kHz frame rate , and an ultrafast camera operating at 1 MHz frame rate have been achieved . However, traditional camera technologies, such as these, suffer from the fundamental trade-off between sensitivity and speed - during a short integration time fewer photons are collected leading to loss of sensitivity [5-7]. Another drawback of fast CCD/CMOS imagers is that they require cooling to reduce thermal noise - a solution that comes at the expense of the added complexity and cost of refrigeration. Because they lack any means to optically amplify the image, high-speed CMOS imagers also require high-intensity illumination to ensure adequate signal-to-noise ratio. However, high-intensity illumination causes damage to, or modification of, the object being imaged, especially in microscopy where the light is focused on a miniature field of view. This is a fundamental problem that renders high-speed CMOS imagers unsuitable for microscopy - a fact that explains why such imagers are marketed mostly towards industrial and non-microscopy scientific applications. Another type of high-speed image sensor is the framing streak camera that has been employed for diagnostics in laser fusion, plasma radiation and combustion. This device operates in burst mode only  and requires synchronization of the camera with the event to be captured, rendering streak cameras also unable to capture unknown or random events. This, along with the high cost of the camera, limits its use in practical applications.
This Essay is
a Student's Work
This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.Examples of our work
Image intensifiers - Image intensifiers are complex devices developed for sensitive imaging by amplifying the photo-generated electrons, as opposed to photons in our STEAM imager. They have a low image acquisition rate up to ~10 kHz in continuous mode [36, 37] - performance that is adequate for its intended use in night-vision cameras because they only need to operate at the video rate. Photon amplification is fundamentally superior as amplification is performed before the signal is contaminated with thermal noise. As a case study, in fiber optic communication, optical amplification is preferred over the use of avalanche photo-diodes (APDs) because it provides superior signal-to-noise ratio .
Time-resolved pump-probe measurements - In scientific applications, high-speed imaging is often achieved by the time-resolved pump-probe technique with a temporal resolution down to picoseconds [39-40]. Although this technique can capture the dynamics of fast events, but it can only do so if the event is repetitive. Because they do not operate in real time, they are unable to capture non-repetitive random events that occur only once or do not occur at regular intervals such as rogue events . Detection of such events requires an imaging technology with fast, continuous, and real-time capability.
Spectrally-encoded imaging - An alternative confocal microscopy technique called spectrally encoded confocal microscopy (SECM) was recently developed to provide a simplified laser scanning mechanism and to enable in vivo imaging through a miniaturized fiber-based probe [16-17]. It uses a spatial disperser, e.g. prism, diffraction grating, which transforms incident broadband light into a 1D spatial spectral pattern resembling a spectral shower. The 1D spatial information of the sample is then encoded into the spectrum of the back-reflected spectral shower. By scanning this 1D spectral shower in an orthogonal direction, a 2D image can be captured. However, these previous demonstrations are incapable of capturing ultrafast dynamical processes with a temporal resolution better than ïs to ns. They are primarily limited by the speed of galvanometric scanner, the speed of the CCD-based spectrometer or the wavelength sweeping speed of the swept laser source [16-19]. The frame rate demonstrated so far is only up to few tens of frames per second. In addition, using the CCD-based detection still suffers the aforementioned trade-off between speed and sensitivity.
Extending such a 1-D SECM technique, the PI and co-workers recently developed a technique which can map the spectrum of the broadband source into a 2D space by using a 2D spatial disperser that consists of a diffraction grating and a virtually-imaged phased array (VIPA) [42-45]. In this way, mechanical scanning for imaging acquisition can be eliminated. This technique also enables simultaneous imaging and laser surgery, called spectrally encoded confocal microscopy and microsurgery (SECOMM) .
WORK DONE BY THE INVESTIGATORS
Amplified Dispersive Fourier Transform (ADFT)
The key element enabling high-speed imaging in STEAM is Amplified Dispersive Fourier transform (ADFT) [10-12]. It is a powerful technique in which spectral information of an optical pulse is mapped into the time domain using chromatic dispersion (also called temporal dispersion) in the dispersive element. (Note that in the context of STEAM, the spectral information is encoded with the spatial information of the sample.) During the process, the pulse is also optically amplified (Fig. 1). The advantages of ADFT are threefold: (i) its pulsed nature provides extremely rapid spectral acquisition capability that can span >100 nm in a single shot. (ii) Its time-domain operation uses a single photodetector and a high-speed electronic digitizer to acquire extremely high resolution spectra in real time. The spectral acquisition rate is essentially given by the repetition rate of the laser pulse train (typically on the order of 10 MHz). (iii) The optical amplification in ADFT overcomes the fundamental trade-off between optical loss and temporal dispersion, hence circumventing the trade-off between sensitivity and speed.
The PI and co-workers have demonstrated ADFT by employing (i) a dispersive fiber as the dispersive element because of its high dispersion and relatively low loss; and (ii) Raman amplification within the DCF - as the mean of optical amplification [8-12]. This established the foundation of ADFT, which led to many applications e.g. absorption  and Raman  spectroscopy, optical frequency-domain reflectometry , displacement sensing and barcode reading , and also microscopy, i.e. STEAM - the main theme of this project.
Fig. 1 Working principle of ADFT.
Serial Time-Encoded Amplified Microscopy (STEAM)
The PI and co-workers recently developed an entirely new imaging modality called Serial Time-Encoded Amplified Microscopy (STEAM) [8-9]. It overcomes the limitations of CCD/CMOS imagers and offers two orders of magnitude higher speed than these imagers (frame rate of 6 MHz). It also overcomes the loss of sensitivity at high speed using novel optical image amplification. STEAM operates continuously and can capture ultrafast events without any knowledge of the timing of their occurrence. The key feature of STEAM is the mapping of a 2D image into a serial time-domain waveform by a two-step approach (Fig. 2): (i) We first encode the 2D spatial information of an object onto the spectrum of a broadband pulse by using a 2D spatial disperser [9, 15]. (ii) A dispersive fiber is then used to perform ADFT - mapping the spectrum into time. The optical spectrum, which is encoded with the image, now appears as a serial sequence in time. To simultaneously amplify the image, the dispersive fiber is pumped with secondary light sources to implement Raman amplification directly within fiber. This powerful approach allows us to perform wavelength-to-time conversion with net signal gain of ~15 dB . Without the optical amplification, the image would not be visible because the signal would lie below the thermal noise of the photodetector, as can be clearly seen in Fig. 3(a)  . Optical image amplification allows high speed imaging at low light levels. Due to this serialization, the image can be detected with a single-pixel photodiode and captured, not by a CCD/CMOS camera, but instead, with any real-time digitizer. The main attributes of the new imager are hence the image amplification in the optical domain and the elimination of the CCD/CMOS - when combined, they enable continuous real-time operation with speed and sensitivity not achievable with CCD/CMOS imagers. As a proof-of-concept demonstration, we used STEAM to capture ultrafast microfluidic flow in real-time continuously (Fig. 3(b)). The flow of microspheres from the right to
Fig. 2 Schematic of the generic STEAM system.
the left is clearly observed (at a speed of 2.4 m/s). This is the first time that such ultrafast microfluidic flow has been observed in real time with such a fine temporal resolution.
RESEARCH PLAN AND METHODOLOGY
This project aims to develop a new STEAM system which operates in the near-infrared (NIR) wavelength range of ~ 700-900 nm, in contrast to the prior proof-of-principle demonstration operating in ~1500 nm [8-9]. Working in this NIR regime offers a more practical tool for biomedical imaging-based diagnosis. This is because it suffers considerably less optical loss due to water absorption and the associated sample heat-up/damage in biological tissue/cell imaging [13-14]. In addition, shorter wavelengths can achieve higher diffraction-limited spatial resolution, improving the image quality . Another advantage is the availability of ultra-broadband pulse laser sources in this wavelength range (e.g. Titanium:Sapphire laser , super-continuum sources ).The last point relates to the performance of ADFT, and hence STEAM because a broadband spectrum (>100 nm) is a prerequisite enabling ADFT. A wider spectrum increases the number of sampling points, i.e., wider field-of-view in STEAM . As described, STEAM consists of two main sub-systems: spectrally-encoded imager and the ADFT system. The project will encompass the design & development of these two parts which can operate efficiently in NIR.
Develop ADFT in NIR
Here we plan to develop a new ADFT system operating at shorter NIR wavelengths (700-900 nm). The key elements are the dispersive element and the optical amplifier:
Dispersive element in ADFT - A highly dispersive element operating in NIR is essential in this project. In ADFT, the dispersion relates to the spectral resolution (i.e. spatial resolution in STEAM). It can be understood by recognizing that the electrical bandwidth of the digitizer limits the spectral resolution, a relation given by ï¤ï¬=0.35(Dfdig)âˆ’1 [11-12], where ï¤ï¬ is the spectral resolution, D is the total temporal dispersion, and fdig is the digitizer bandwidth. The fundamental trade-off in ADFT is self-evident: the product of ï¤ï¬D is fixed by the digitizer bandwidth. To increase the spectral resolution (reduce ï¤ï¬), one is forced to increase D. Dispersive fiber by far is the most viable candidate to achieve dispersive Fourier transform because of its compactness, and the ability to engineer its dispersive properties, such as dispersion compensation fibers (DCFs) and photonic crystal fibers (PCFs). There are dispersive fibers available in NIR, e.g. a dispersive fiber has a dispersion of ~0.15 ns/nm-km at 800 nm (Nufern ) and a PCF has a dispersion of ~0.2 ns/nm-km at 700-900 nm (Crystal Fibre ).
We have done a preliminary study on dispersive Fourier transform operation at ~800 nm (Fig. 4(a)). We employed a dispersive fiber with dispersion of ~500ps/nm at 800 nm. The input spectrum (shaped by the bandpass filter) of the individual pulse is mapped into the time domain by the dispersive Fourier transform and then is captured in real-time by an oscilloscope (Fig. 4(b)). Although no optical amplification was performed in this case, which results in a weaker detected signal, the good agreement between the spectra captured by the oscilloscope and by the spectrometer as shown in Fig. 4(b) clearly validates the dispersive Fourier transform operation in ~800nm. Adding optical amplification can thus greatly enhance the sensitivity.
Optical amplifier in ADFT - It is desirable to have high dispersion in order to achieve high spectral resolution. However, this comes at the expense of increased optical loss and reduced detection sensitivity. This dispersion-loss trade-off is a fundamental connection described by the Kramers-Kronig relations . Optical amplification is thus essential to circumvent this trade-off and is playing an important role in ADFT and hence, STEAM.
There are a number of viable amplification schemes, e.g. Raman amplification, as demonstrated in the PI's preliminary works [8-9, 12]. Because of its intrinsic properties, such as its widely tunable gain spectrum and its naturally broadband gain spectrum allowed by the amorphous nature in optical glass fiber, Raman amplification is generally favorable for ADFT. The gain bandwidth can be further broadened by using multi-wavelength pump lasers, and, surprisingly but fortuitously, extremely broadband gain spectra can be realized using incoherent pump sources . This is highly desirable because a large optical bandwidth, as we mentioned earlier, results in a large number of resolvable points in STEAM.
Another feasible amplification approach is fiber optical parametric amplification (FOPA). The advantages of FOPA are that the gain spectrum is widely tunable, and the gain spectral shape can be flexibly designed by dispersion engineering of the fibers. By using the dispersion-engineered fibers, FOPA is well-known to exhibit broadband optical gain under the phase-matching condition [52-56]. For examples, it has been demonstrated to generate optical gain of > 30dB in the wavelength range of 800-900nm using PCFs ; octave-spanning (600 nm-1200 nm) parametric gain of >10-20 dB was also demonstrated using PCFs .
The above preliminary results of the dispersive Fourier transform, together with the successful demonstration of both Raman amplification and FOPA, show great promise to realize the ADFT system at NIR. More thorough studies will be carried out to fully incorporate Raman amplification or FOPA into the new ADFT system and further improve its performance. We will also perform a theoretical study to investigate the fundamental limit of different schemes.
Develop the STEAM systems in NIR
The architecture of the new STEAM system in NIR range is similar to the previous proof-of-principle demonstration operating in ~1500 nm  (Fig. 2). The key differences are (a) the requirements of the ADFT part, which has already been discussed earlier and (b) the spatial disperser (i.e. diffraction grating and/or VIPA) which should be able to perform spectrally-encoded imaging efficiently in NIR. We will first build a simpler version of STEAM, i.e. 1-D STEAM, to evaluate its basic functions in NIR. Note that 1-D STEAM has also successfully been demonstrated in ~1500 nm by the PI and co-workers for ultrafast (world record) bar-code reading and displacement sensing . In 1-D STEAM, a 1-D spatial disperser (e.g. prism and diffraction grating, which are the off-the-shelf optical components) is the key component for spectrally-encoded imaging. The design and development of the spectrally-encoded imaging part will be carried out based on the PI's recent theoretical analysis of STEAM [12, 57]. The performance of STEAM can also be further improved based on this analysis as well. For instance, the frame rate can be further increased by employing a parallel multi-channel architecture with the so-called virtual time gating technique .
One possible experiment to verify the performance of STEAM is to monitor microfluidic transport, similar to ref. . But here, the 1-D STEAM imager operates in the line scan mode. Consecutive line scans each at a different horizontal plane (in the x direction) are obtained as the microparticle passes the beam. The 2D transversal image of the particle (x and y directions) is then reconstructed by combining the 1D line scans by image processing (Fig. 5). An immediate application of such configuration is rapid cell counting and identification in flow-cytometry . This configuration also serves as the basis of developing the phase-contrast STEAM system, which is able to perform high-speed quantitative evaluation of biological cells in imaging flow cytometry (see the next section). To realize 2-D STEAM, we can simply replace the 1-D spatial disperser with the 2-D spatial disperser (i.e. diffraction grating and VIPA), and leave the ADFT system intact.
Explore the potential applications of STEAM
STEAM, in its native form, is essentially a type of reflectance confocal microscopy (RCM). It is a non-invasive and label-free imaging technique revealing the cell morphology and tissue architecture in-vivo using back-scattering from different cellular/tissue components to provide image contrast [20-22]. Due to its confocal nature, RCM offers a significant improvement in image quality over conventional brightfield/ darkfield microscopy. It has been extensively used in many clinical applications, e.g. diagnosis of skin tumors , and characterization of human breast tissue . However, its imaging speed has to compromise the weak detection sensitivity, which is limited by the weak back-scattered light (< few %) from the biological tissue [20-22]. Thus, combining RCM with STEAM, i.e. RC-STEAM, can offer orders-of-magnitude improvement in imaging speed and sensitivity. We will investigate the utilization of RC-STEAM in areas such as monitoring fast response to therapeutic treatment, e.g. laser surgery [60-62]. In this regard, the Faculty Core Imaging Facility in the Faculty of Medicine at HKU will provide support to this project on practical biomedical imaging studies, e.g. to provide comprehensive evaluation of the STEAM system using their facilities, to explore the potential of integrating STEAM with the existing microscopes.
Moreover, many biological entities, e.g. live red/white blood cells, are optically transparent and cannot be imaged without stains or labeling, which generally requires modification of the cellular structures. It is less favorable for characterization of live cells in their native physiological states. By obtaining the optical phase change of the transparent cells, quantitative phase microscopy (QPM) was recently developed for stain/label-free cell imaging. It can provide not only high image contrast of these cells, but also quantitative evaluation of cellular information (e.g. cell volume, mass, refractive index) at nanometer scale [23-27]. However, its image acquisition rate is mainly limited by the speed of the recording device, i.e. CCD/CMOS imagers, which possess the same speed-sensitivity trade-off mentioned before. Hence, combining STEAM with QPM, called quantitative phase STEAM (QP-STEAM), paves the way to achieve fast, high-throughput, sensitive, quantitative and label-free cell imaging and characterization in biomedicine, e.g. monitoring cellular dynamic during laser surgery , and flow cytometry [3-4].
Built on the 1-D STEAM prototype developed in (2), a possible scheme realizing QP-STEAM is to incorporate an interferometer with the prototype. Fig. 6 shows a preliminary study on QP-STEAM operating at ~1500 nm based on a Michelson interferometer. By performing 1D spectral-shower line scans (see Fig. 5), the system obtain the phase images of the PDMS microbeads (~ 20-50 ïm), which flow along the microfluidic channel at a high speed of ~ 5 m/s. This preliminary study suggests that the QP-STEAM in principle can perform quantitative imaging at a speed (i.e. ~100,000 cells/sec) not achievable with the current state-of-the-art "imaging" flow cytometry (1,000 cells/sec) . Further investigation will be carried out to develop a practical QP-STEAM prototype, operating in NIR in which the image resolution can be further improved. The coordinator of the Faculty Core Imaging Facility, Prof. George Tsao, will also provide support of essential infrastructures for this part of the project, e.g. the existing flow cytometers.
RESEARCH SCHEDULE AND TECHNICAL CONCERNS
(also see Table 1 in the "Additional figures and tables" page)
Year 1 (I)
Investigate different types of dispersive fibers and identify the best candidate (~3 months).
There are a number of commercially available dispersive fibers operating at ~700 - 900 nm, supplied by e.g. Nufern , Crystal Fibre . (risk: low)
Establish simulation capability for optical system design of STEAM (~ 2 months).
We will employ an industrial-standard simulation tool, called ZEMAX , to design the STEAM system. It is the tool used for designing the previous STEAM system . (risk: low)
Perform theoretical modeling of ADFT. (~3 months)
Based on our recent work on ADFT analysis , the modeling work will focus on modeling the ADFT performance based on the design parameters used in this project and will also further generalize the existing model. (risk: low)
Evaluate the dispersive properties of the dispersive fiber (~ 3 months)
Many well-known methods exist to evaluate the fibers' dispersive properties . (risk: low)
Develop the ADFT experimental setup and evaluate its performance (~ 6 months).
The PI is experienced in Raman amplification (as demonstrated in both fiber and silicon [8,9,12, 66-68]). The Co-I has a strong track record in FOPA research [69-73] and will work closely with the PI in ADFT design and development. (risk: low - intermediate)
Year 2 (II)
Design the STEAM system based on the established simulation capabilities (~ 4 months).
Same as Task IIB (risk: low).
Develop the 1-D and 2-D spectrally-encoded imaging experimental setup (~ 6 months)
To build the spectrally-encoded imaging setup, the off-the-shelf optical components are required to operate efficiently in NIR. By a proper system design (Task IIA), this part is relatively less challenging. (risk: low)
Setup the basic prototype of 1-D STEAM. (~ 2 months)
Combining ADFT and spectrally-encoded imaging together is straightforward and does not require significant modification of the individual set-up. (risk: low)
Explore the potential applications of STEAM (all year long).
The PI got the support from the Faculty Core Imaging Facility in the Faculty of Medicine at HKU. The facility has been actively involved in a myriad of optical bio-imaging research areas. Their input can benefit this part of the project significantly. (risk: low)
Year 3 (III)
Continue Task IID, i.e. implement 1-D STEAM as well as 2-D STEAM (~ 3 months).
Same as Task IIC.
Evaluate the STEAM system performance (~6 months).
The pros and cons of the 1D and 2D STEAM prototypes will be identified. Even if the 2D version does not show satisfactory performance, we can always resort to the 1D STEAM system which is less challenging in terms of implementation and it indeed has a great potential in many applications, e.g. flow cytometry. (risk: low - intermediate)
Develop the QP-STEAM experimental setup and evaluate its performance (~6 months).
QPM based on interferometric configuration has already been successfully demonstrated. We need to explore the possibility of incorporating such configuration with the 1-D STEAM prototype. (risk: intermediate)