Open Access
Add to BibSonomyAbstract
Ultrahigh resolution optical coherence tomography (OCT) is demonstrated at 800 nm and 1300 nm using continuum generation in a single photonic crystal fiber with a parabolic dispersion profile and two closely spaced zero dispersion wavelengths. Both wavelengths are generated simultaneously by pumping the fiber with ~78 mW average power at 1064 nm in a 52 MHz, 85 fs pulse train from a compact Nd:Glass oscillator. Continuum processes result in a double peak spectrum with > 110 nm and 30 mW average power at 800 nm and > 150 nm and 48 mW at 1300 nm. OCT imaging with < 5 μm resolution in tissue at 1300 nm and < 3 μm resolution at 800 nm is demonstrated. Numerical modeling of propagation was used to predict the spectrum and can be used for further optimization to generate smooth, broad spectra for OCT applications.
©2006 Optical Society of America
1. Introduction
Optical coherence tomography (OCT) enables high-resolution cross-sectional imaging of tissue microstructure without the need for tissue excision and processing [1] and has therefore been investigated for minimally invasive imaging in several areas of biology and medicine [2]. OCT performs optical imaging in scattering tissue using low-coherence interferometry, and the axial range resolution scales inversely with the optical bandwidth, Δλ, of the light source as λo 2/Δλ, where λo is the center wavelength of the source. Traditional OCT imaging systems are based on superluminescent diode light sources, which enable imaging resolutions of 10 – 15 μm. Broadband modelocked Ti:Al2O3 lasers have been demonstrated to extend the axial resolution to 1–5 μm in tissue [3, 4], leading to the term ultrahigh resolution OCT. These sources directly provide broad bandwidth and high power and are well suited for real time biomedical imaging, particularly in the relatively transparent tissues of the eye [5]. The high cost and complexity of femtosecond Ti:Al2O3 sources has driven the development of lower cost alternatives for ultrahigh resolution OCT imaging, including low-threshold Ti:Al2O31 lasers capable of using low-power pump lasers [6, 7], continuum generation in highly nonlinear fibers [8], diode-pumped Cr:LiCAF lasers [9], thermal light sources [10], superluminescent Ti:Al2O3 sources [11, 12], and multiplexed superluminescent diode sources [13]. As a result of these developments as well as commercialization of suitable Ti:Al2O3 laser sources, ultrahigh resolution OCT imaging in the 800 nm wavelength range is becoming widely available to the research community.
In vivo OCT imaging in scattering tissues other than the eye is typically performed at 1300 nm to take advantage of reduced scattering and enhanced penetration at the longer near-infrared wavelength [14]. Suitably broadband ultrafast lasers for ultrahigh resolution are not common at this wavelength and research has therefore focused on spectral broadening techniques using nonlinear mechanisms in optical fibers. A Cr:forsterite femtosecond laser with the bandwidth of the output pulse broadened in a dispersion shifted fiber has been used to achieve < 5 μm axial resolution [15, 16]. The realization of supercontinuum generation in photonic crystal fibers (PCF’s) [17] has enabled ultrahigh resolution OCT at wavelengths throughout the visible and near infrared [18–20]. OCT was performed at 1300 nm with resolutions as high as 2.5 μm using a portion of the broadband continuum produced by a Ti:Al2O3 pump laser and a PCF with zero dispersion at 800 nm [18]. Broadband continuum from highly nonlinear fibers has also been used to demonstrate ultrahigh resolution OCT at 800 nm in the wavelength range around 1000 nm as an alternative to 800 nm and 1300 nm [21–24]. This wavelength range has generated considerable interest for imaging in scattering tissues due to the optimal characteristics of water dispersion at 1000 nm [25]. Similarly, ultrahigh resolution OCT imaging using continuum generation at 1500 nm has recently been demonstrated [26].
Using continuum generation, reliable commercially available femtosecond lasers can be used to access different wavelengths throughout the near-infrared region. However, higher exposure limits for non-ocular tissues necessitate higher power sources for OCT imaging at wavelengths other than 800 nm. Controlling the continuum generation to simultaneously achieve high power and broad bandwidth with small spectral modulation and low amplitude noise is of primary interest. Broadening by self-phase modulation in the normal dispersion region of highly nonlinear fibers has proven effective in achieving these goals [27]. A commercially available, turnkey Nd:Glass femtosecond laser combined with self-phase modulation in a high numerical aperture fiber was recently demonstrated for real time ultrahigh resolution imaging at 1000 nm with < 5 μm axial resolution and > 90 mW output power [23]. Using a portable Cr:Forsterite and spectral broadening in a dispersion shifted fiber, axial resolution of < 5 μm has been achieved for in vivo endoscopic imaging [16]. Compact Cr:forsterite lasers are not commercially available, which will limit this technique for use at 1300 nm to investigators with significant femtosecond laser experience. Pumping near the zero dispersion wavelength of a PCF can generate very broad spectra extending beyond 1300 nm, but can result in unacceptably strong spectral modulation and noise amplification [28–30]. Furthermore, spectral filtering of a wavelength region away from the pump wavelength is inefficient and results in low source output power. A compact erbium fiber laser combined with a highly nonlinear fiber has been demonstrated for ~2 μm axial resolution at 1300 nm but had only 4 mW of output power due to inefficient spectral conversion [31]. A continuous-wave, all-fiber Raman light source based on a microstructured fiber was capable of > 300 mW output power from a 10 W pump source, but the continuum suffered from excess intensity noise [32].
Photonic crystal fiber technology is extremely powerful because it enables customization of the fiber dispersion profile, which can be used to control the nonlinear processes responsible for continuum generation. Recently, a novel photonic crystal fiber with two closely spaced zero dispersion wavelengths centered around 860 nm was reported for the generation of stable, compressible double peak spectra using a modelocked Ti:Al2O3 laser broadened by a combination of self-phase modulation and four-wave mixing [33]. Unlike supercontinuum produced when a standard PCF is pumped near the zero dispersion wavelength, the continuum from the new PCF design was relatively insensitive to parameters of the input Ti:Al2O3 pulse resulting in stable, low-noise spectra. In addition, the pump energy was efficiently depleted during the continuum to generate high brightness spectral peaks at higher and lower wavelengths, with the location of the peaks set by the zero dispersion wavelengths of the fiber.
In this paper, we investigate this novel PCF design concept for a broadband, stable continuum light source for ultrahigh resolution OCT imaging. By injecting pulses at 1064 nm from a Nd:Glass laser into a photonic crystal fiber with two closely spaced zero dispersion points centered around 1050 nm, we numerically and experimentally demonstrate that a stable continuum can be generated with high brightness double peaks centered at 800 nm and 1300 nm, two primary wavelengths of interest to the OCT imaging community. The experimentally measured continuum achieves > 110 nm bandwidth and 30 mW power in the 800 nm spectral peak and > 150 nm and 48 mW in the 1300 nm peak, thereby enabling ultrahigh resolution imaging with < 5 μm in tissue at both wavelengths. In addition, we investigate via numerical simulation the conditions for achieving a smooth, modulation free spectrum for optimal OCT imaging. Combined with previous results from [23], this work suggests that a single commercially available oscillator combined with various highly nonlinear fibers can be used to generate stable, high brightness spectra for ultrahigh resolution OCT imaging throughout the near infrared wavelength region.
2. Numerical simulations
This work is based on a commercially available photonic crystal fiber with a parabolic dispersion profile and two closely spaced zero dispersion wavelengths (Crystal Fibre A/S, NL-1050-ZERO-2). Zero dispersion wavelengths are nominally present at ~1022 nm and ~1075 nm with the parabolic maximum of ~0.6 ps/nm/km in the anomalous region at 1050 nm and monotonically decreasing dispersion at wavelengths below and above the zero dispersion points. The fiber is non-polarization maintaining and uses a microstructured cladding with air holes to guide light in a pure silica core. It has high nonlinearity of ~0.37 (Wkm)^-1 with a small mode field diameter of ~ 2.2 μm and high numerical aperture of 0.37.
Numerical simulations were conducted to study continuum generation in the novel PCF for OCT imaging applications. The details of the simulation procedure have been described previously [34, 35]. In summary, pulse propagation in the fiber was modeled by solving the generalized nonlinear Schrodinger equation using the split-step Fourier method. Dispersion terms up to 10th order were included as well as absorption losses from Rayleigh scattering, infrared absorption, and OH ion absorption. Dispersion parameters were estimated from the measured dispersion profile, with values outside the available wavelength range obtained by extrapolation. Third order effects including self-phase modulation (SPM) and four-wave mixing (FWM) are included in the simulations as well as the effect of self-steepening (SS). Stimulated Raman scattering (SRS) has been ignored. To most closely match the experimental setup, a sech^2 pulse with 85 fs full width at half maximum duration and 18 kW peak power at the Nd:Glass center wavelength of 1064 nm was used for the simulations.
2.1 Fiber length dependence of the continuum
Figure 1 presents numerical results for the length dependence of the continuum, with spectra displayed on a linear scale. As demonstrated for a similar fiber design previously [33], this continuum rapidly evolves into a double peak structure with nearly complete depletion of the pump wavelength. These results predict that the main spectral peaks can be localized to the important OCT imaging wavelength regions near 800 nm and 1300 nm. The input spectrum is rapidly broadened within 10 cm to a full spectral width of more than one octave between 700 nm and 1500 nm. Additional propagation in the fiber leads to further depletion of the region between the two spectral peaks as well as filling in and smoothing of the main peaks. Slight modulation appears on the inner facing edges of the spectral peaks between 0.5 and 1.0 m, and propagation after 0.5 m results in little change in the spectrum.
Several theoretical and experimental studies have been performed to elucidate mechanisms of supercontinuum generation in typical photonic crystal fibers when pumping in the anomalous dispersion region near the zero-dispersion wavelength [36–40]. When using picosecond or nanosecond pulses, stimulated Raman scattering and parametric four-wave mixing are dominant [40], while the use of femtosecond pulses leads to spectral broadening through fission of higher-order solitons into red-shifted fundamental solitons and blue-shifted nonsolitonic radiation [41]. By contrast, the mechanism for supercontinuum generation in PCF’s with two closely spaced zero dispersion wavelengths has been attributed by Hilligsoe, et al. to a combination of self-phase modulation and phase-matched four-wave mixing, with relative suppression of the Raman scattering, self-steepening, and soliton fission mechanisms [33]. These authors report that self-phase modulation is the dominant broadening mechanism and provides seed wavelengths for four-wave mixing, which leads to efficient depletion of wavelengths between the zero dispersion points. Additional non-degenerate four wave mixing also contributes to extending the depleted region between the main peaks beyond the region between the zero dispersion wavelengths [33]. Subsequent work has questioned the role of four-wave mixing in the spectral broadening and has demonstrated a role of generation of dispersive waves followed by soliton self frequency shift (SSFS) if the separation between the zero dispersion wavelengths is large enough [42]. Additional work on the fundamental mechanisms in such fibers is necessary to fully understand the discrepancies in these studies.
In our study, the fiber chosen has a very narrow anomalous dispersion region between the zero dispersion wavelengths, which makes soliton effects unlikely due to the rapid shift of energy outside of the zero dispersion wavelengths to the normal dispersion region [42]. Comparison of the characteristic lengths of dispersion, LD, and nonlinearity, LNL, for peak power Po, pulse width To, and dispersion β 2 at the pump wavelength reinforces that pulse evolution is initially dominated by high nonlinearity since LD/LNL ≡ (γPoTo 2)/|β 2| ≈ 104 ≫1 [35]. Small group-velocity dispersion in the region of the pump wavelength and high fiber nonlinearity allows self-phase modulation to dominate dispersive broadening over the first several centimeters resulting in fast spectral broadening. The main spectral peaks lie in the normal dispersion region and dispersive broadening subsequently leads to a reduction in peak power. With continued propagation, nearly complete depletion of the region between the main peaks results, potentially due to phase-matched four-wave mixing [33]. Further theoretical studies are underway to confirm the precise mechanisms of continuum generation in this fiber.
Fig. 1. Simulated fiber length dependence of the continuum spectrum generated by 85 fs, 18 kW peak power pulses at 1060 nm. Nearly complete depletion of the pump wavelength is observed along with creation of two high brightness main peaks centered near 800 nm and 1300 nm.
2.2 Pump wavelength dependence of the continuum
To further understand how to control spectral shape for OCT applications, we simulated the effect of varying the pump wavelength with respect to the zero dispersion wavelengths. Previous results suggested that the spectral extent is not very sensitive to the position of the pump wavelength and that the relative power between the two main lobes can be adjusted by choice of the pump wavelength position [33]. Our simulations present somewhat different results using the photonic crystal fiber designed for 1050 nm. This fiber has extremely small anomalous dispersion in the region between the zero dispersion wavelengths, which may complicate the wavelength dependence by enhancing additional nonlinear mechanisms. Figure 2 presents spectra resulting from simulations of 85 fs, 18 kW peak power pulses propagated through a 2 m length of the photonic crystal fiber at various center wavelengths around the depletion region. At each pump wavelength, the majority of the spectral power remains in the long wavelength peak, and for pump wavelengths near the lower zero dispersion wavelength, the continuum can extend farther into the infrared. The short wavelength edge of the spectrum remains relatively fixed with respect to changing pump wavelength. The asymmetry in the spectrum seems to depend on higher order dispersion terms. In addition, the depletion region between the main spectral lobes increases in span with pump wavelengths in the normal dispersion ranges. This could be due to enhanced four-wave mixing processes resulting from stronger seed energy deposited by self-phase modulation around the zero dispersion wavelengths [33].
For OCT imaging, spectral shape is a critical determinant of image quality since non-Gaussian spectra with large modulation can lead to large coherence wings on the interference point spread function [43]. Our simulations indicate that smooth, twin peak spectra can be achieved when the wavelength of the pump pulse is located between the two zero dispersion wavelengths.
Fig. 2. Simulated pump wavelength dependence for optimal continuum generation. Input pulses of 85 fs, 18 kW peak power illustrate differences in continuum as a function of pump wavelength.
Fig. 3. Time-frequency characteristics of the continuum for input pump pulse center wavelength of 1120 nm. The temporal profile (a) and spectrogram (b) illustrate how temporal overlap of spectral components can lead to temporal and spectral structure by cross-phase modulation.
The increased fine spectral modulation for pump wavelengths in the normal dispersion region results from cross phase modulation due to temporally overlapped spectral components. Figure 3 displays the temporal pulse profile and spectrogram corresponding to the 1120 nm pulse with spectral profile shown in Fig. 2. The short wavelength spectral components trail the long wavelength components for the most part, but temporal overlap between wavelengths on the edges of the depletion region leads to cross-phase modulation. This overlap is minimized for pump wavelengths located between the zero dispersion points. Further investigation will be necessary to completely understand the mechanisms behind the wavelength dependence of this continuum. Since choosing the optimal pump wavelength is largely equivalent to optimizing the fiber dispersion profile for a fixed pump wavelength, custom photonic crystal design and fabrication should enable ultrahigh resolution OCT with several commercially available femtosecond oscillators.
3. Experimental measurements of the continuum
The experimental setup used to measure the continuum spectrum is shown in Fig. 4. A diode pumped femtosecond Nd:Glass laser (High Q Laser Production) generates a 52 MHz train of 85 fs pulses with 165 mW average power at 1064 nm center wavelength. The compact, prismless oscillator (53 cm × 20 cm) is pumped by two 1 W laser diodes and is soliton modelocked using a semiconductor saturable absorber mirror (SESAM) and intracavity dispersion compensating chirped mirrors. The ends of a 1 m length of PCF are flat-cleaved using a standard cleaver and mounted on high-precision 3-axis stages for coupling into and out of the fiber. The laser output is coupled directly into a 1 m length of photonic crystal fiber using a 3.3 mm focal length aspheric lens and is recollimated at the output using a 3.1 mm aspheric lens. An isolator was tested and found to be unnecessary because the oscillator was relatively insensitive to feedback from the fiber end facet reflection. The fiber is non-polarization maintaining, and adjustment of the input via a half-wave plate also did not appreciably alter the continuum spectrum.
To enable OCT imaging at both 800 nm and 1300 nm, the spectrum had to be filtered with high purity to prevent interference from the unused wavelength region. The 800 nm continuum was filtered using 5 bounces on dielectric mirrors with broadband antireflection coatings centered at 800 nm. These mirrors have typical reflectances of > 99% across the wavelength range around 800 nm while transmitting with low loss the wavelengths at 1300 nm, thereby providing 60–70 dB of spectral separation of the two continuum main lobes. Rejection of residual 800 nm light from the 1300 nm continuum was achieved with a polished silicon absorber. After spectral filtering, the 800 nm and 1300 nm light was separately coupled into standard single mode fibers designed for those wavelengths using 5 mm focal length aspheric lenses.
Fig. 4. Experimental setup for the continuum generation light source at 800 nm and 1300 nm for ultrahigh resolution OCT imaging. Spectral filtering is performed using broadband dielectric mirrors and a silicon absorber to ensure removal of the opposite wavelength for OCT imaging. TM, turning mirror; DM, dielectric mirror.
The measured continuum spectrum is presented in Fig. 5. The pump spectrum is also shown in Fig. 5(a). The continuum spectrum shown is a concatenation of spectra measured separately with 0.5 nm resolution after filtering the 800 nm and 1300 nm spectra. Separate measurement of the spectral regions enabled more precise representation of spectral shape and relative amplitude, since the power in both wavelength regions could be individually measured and fiber coupling to the optical spectrum analyzer could be optimized to minimize chromatic focusing aberrations. The effect of filtering on the combined spectrum was minimized by concatenating the spectra at 1000 nm in the depleted region between the two main peaks. The silicon absorption edge and the dielectric mirror reflectivity both drop sharply around this wavelength. The linear spectrum in Fig. 5(a) agrees reasonably well with previously presented simulation results. Bandwidths of 156 nm and 116 nm are achieved at 1300 nm and 800 nm, respectively. The simulated spectrum shown in Fig. 1 for 1060 nm pump wavelength extends further into the infrared than the measured spectrum. This may be due to leakage losses in the PCF at the long wavelengths [33] or perhaps to deviation from the specified dispersion profile. As shown in Fig. 2, the extent of the long wavelength side of the continuum also depends on the pump wavelength relative to the two zero dispersion points. Power in the spectra recorded before coupling to the single mode fibers measured ~ 30 mW at 800 nm and ~ 48 mW at 1300 nm. For average incident pump power of 165 mW, the measured output power corresponds to ~ 48% coupling efficiency. At 800 nm, 8–11 mW was available in the single mode HI-780 fiber, while at 1300 nm 25–30 mW was coupled into SMF-28. As predicted in the simulation, energy is efficiently depleted from the pump wavelength and transferred to the two spectral bands around 800 nm and 1300 nm. The spectrum is shown on a log scale in Fig. 5(b). The region between the two spectral main lobes after filtering is shown to be > 20 dB below the main spectral lobe amplitude.
Fig. 5. Experimental measurement of continuum spectrum on (a) linear and (b) log scales. The pump laser spectrum is also shown in (a). These spectra are created from concatenation of individually measured spectra in the 800 nm and 1300 nm wavelength regions.
The modulation seen on the spectrum in Fig. 5(a) is greater than expected based on numerical simulations for an injected pulse at 1064 nm. There are several possible reasons for this. First, there may be some error in the approximation of the higher order dispersion terms from the available measured data of the fiber dispersion profile. The form of the continuum is strongly dependent on the dispersion characteristics of the PCF. The observation that the short wavelength component is centered closer to 800nm in the experimental measurements compared to the numerical results may also indicate deviation from the manufacturer’s dispersion specification. In addition, the dispersion profile may vary slightly over the length of the fiber, since it depends on physical air hole structure on the micron scale. We observed that, while the general form of the continuum remained fixed, the spectral modulation varied slightly when the fiber was recleaved and when the fiber ends were switched such that the propagation direction was reversed. The spectral shape of the continuum was also sensitive to very small adjustments of the XYZ coupling stage without significant coupled power variation suggesting that the precise way in which the mode is launched into the fiber must be considered. It is possible that light propagating in the microstructured cladding may cause modulation on the spectrum. To minimize these effects, steps will be taken in future experiments to control the coupling with even higher precision stages.
Occasionally, we observed severe degradation in coupled power into the PCF. This was the result of increased scattering from dust or moisture at the tip and subsequent fiber end facet damage. This effect was seen at both the input and output coupling ends and is a well known observation with photonic crystal fibers, since the fine air-silica microstructure is extremely sensitive to ambient environmental conditions. We were able to minimize this effect by covering the bare end facet, but in future work the use of commercially available options for fiber end facet protection will likely improve long term power stability. Photonic crystal fibers are now readily available with integrated collimating lenses or with hermetically sealed end facets.
Excess amplitude noise has been measured previously on continuum spectra measured from photonic crystal fibers as the result of temporal instability and nonlinearly amplified quantum noise [28]. Prior work with PCF’s using two closely spaced zero dispersion wavelengths, however, has measured relatively low noise continuum because of the suppression of several higher order nonlinear mechanisms including soliton fission [33]. To verify that this holds for the continuum presented in Fig. 5, the intensity noise in the filtered wavelength regions of 800 nm and 1300 nm was measured separately using an RF spectrum analyzer.
Fig. 6. Experimental measurement of RF noise spectra for the filtered wavelength regions around (a) 800 nm and (b) 1300 nm. The noise spectra are compared to the RF spectrum for the Nd:Glass oscillator alone.
Figure 6 shows the recorded RF noise spectra acquired using a fast photodiode. The measurements were made at the first harmonic of the laser repetition rate and normalized to the carrier power. Detected first harmonic carrier power was 1.13 dBm for the oscillator alone and -22.97 dBm and -17.0 dBm for the wavelength regions of 800 nm and 1300 nm, respectively. The noise decreases nearly exponentially at frequencies above 10^5 Hz and is consistent with previous measurements of the amplitude noise spectrum for continuum generated predominantly by self-phase modulation in a high numerical aperture fiber [23]. The noise spectra largely tracks that of the oscillator alone implying that OCT imaging can be performed with high sensitivity due to the lack of excess amplitude noise.
4. OCT imaging results and discussion
4.1 OCT system setup
The continuum spectrum presented in Fig. 5 was used with time-domain, ultrahigh resolution OCT systems to demonstrate imaging in biological tissue. Figure 7 shows the generalized system diagram, which was implemented at both 800 nm and 1300 nm. Light from the filtered continuum was coupled into a dual-balanced interferometer consisting of two 50/50 splitters. At the final splitter, light is divided between a reference beam and a sample beam. Optical path length scanning of the reference beam is performed using a retroreflecting galvanometer, while the sample beam is focused by a 1:1 microscope to a small spot size approximately equal to the fiber mode field diameter, ~ 6 μm at 800 nm and ~ 9 μm at 1300 nm. Dispersion compensating glass is used in the reference arm to match the dispersion of the sample arm optics. Sample scanning is achieved using mechanical translation with a computer controlled stage.
Fig. 7. OCT system diagram. Similar systems were implemented at both 800 nm and 1300 nm for these experiments. DCG, dispersion compensating glass; TIA, transimpedance amplifier; A/D, analog to digital converter.
Light returning from the sample and reference arm interferes at the balanced detectors, which subtract common-mode noise while adding the interference terms. The electronic OCT signal is amplified, filtered, and logarithmically demodulated before being sampled by a computer for display. Different couplers, sample arm optics, and photodiode detectors are used for the 800 nm and 1300 nm wavelengths, but the galvanometer, scanning stage, and filtering/demodulating electronics are common for both systems. Filtering is performed at 205 kHz center frequency with 53 kHz bandwidth. For axial scan depth of 2 mm at both wavelengths, this corresponds to 38 Hz, 133 mm/s scanning at 1300 nm and 25.4 Hz, 84.5 mm/s scanning at 800 nm. These imaging speeds are slower than usual for most time-domain OCT systems, but are sufficient for evaluating the imaging performance of the continuum light source.
4.2 System performance evaluation
System performance characterization is presented for both wavelengths in Fig. 8. System resolutions and point spread functions were characterized by measuring reflections from a mirror with neutral density filters in the sample arm. Minimum neutral density attenuation was chosen to prevent saturation of the transimpedance amplifier, which occurred for > 2.5 OD. Figures 8(a) and (b) compares the point spread function on a linear scale for both wavelengths. At 800 nm, 3.0 μm full width at half maximum (FWHM) is achieved, which for index of refraction in tissue of 1.38 will provide ~ 2.2 μm in tissue. At 1300 nm, 6.5 μm FWHM provides ~ 4.7 μm in tissue. Both point spread functions are symmetric and correlate well to the Fourier transform interference bandwidths presented in Fig. 8(e) and (f), indicating that dispersion mismatch is minimized. The interference bandwidths indicate that some spectral shaping in the OCT system results in reduction of optical bandwidth on the short wavelength edges of the spectra. The couplers used in this study have been previously demonstrated to support broad optical bandwidth [4, 18]. Spectral shaping likely results from wavelength dependent focusing aberration in the optics as well as some multimode leakage in long optical fibers bringing light to the OCT system setup. This can be improved with further attention to system design.
Figures 8(c) and (d) present the logarithmically demodulated output signals for 800 nm and 1300 nm, respectively. The traces have been normalized and scaled to reflect the 50 dB sample arm attenuation (2.5 OD filter, double-pass) used to measure them. At both wavelengths, the sidelobe coherence artifacts on the point spread functions are present at ~ 25–30 dB. These results correlate well to the expected point spread function obtained by Fourier transforming the input optical bandwidths to the systems indicating that the source of the coherence artifact is the spectral modulation. Sidelobe levels of 40–50 dB are desirable for OCT imaging, and further work to investigate spectral smoothing in the PCF will be important for performance improvement. Sensitivities were recorded as the minimum detectable reflection levels below the sample power for the systems. Measured values were ~ -103 dB for 1.0 mW sample exposure at 800 nm and ~ -107 dB for 3.0 mW sample exposure at 1300 nm. Reference arm optical power on the detectors was set to achieve shot noise limited sensitivity with maximum dynamic range.
4.3 OCT imaging results
Figure 9 presents representative OCT imaging results in biological tissue at 800 nm and 1300 nm. Imaging was performed of formalin fixed hamster cheek pouch, which provides an excellent and durable tissue model for evaluation and comparison of light source performance. OCT imaging was performed with axial resolutions in tissue of 2.2 μm and 4.7 μm at 800 nm and 1300 nm, respectively. Acquired images were ~ 1 mm × 1.5 mm (transverse × axial) in dimension with 1000 × 2000 pixels. Sample power was 3.7 mW at 1300 nm and 1.0 mW at 800 nm.
Examination of the 800 nm and 1300 nm images in Fig. 9 illustrates that fine features in the fixed tissue specimen can be visualized with high resolution. Some blurring of the images results from the sidelobe coherence artifact, but this can be improved with further optimization of the continuum light source. Enhanced penetration depth at 1300 nm is evident in the images. The OCT system at 1300 nm has a larger depth of field by a factor of ~ 2 corresponding to the larger fiber spot size at this wavelength. In addition, there is reduced scattering and enhanced penetration of longer wavelengths in tissue. This has been an important reason for imaging in the 1300 nm wavelength range in highly scattering tissues. In the formalin fixed tissue sample, there is some shrinkage compared to the unfixed in vivo images which have been presented previously [18], but the layered structure is preserved. The epithelium (e), connective tissue (c), and muscular layers (m) are visible with high contrast.
Fig. 8. Ultrahigh resolution OCT performance evaluation. Linear (a,b) and logarithmic (c,d) point spread functions are shown for both 800 nm (a,c) and 1300 nm (b,d). The Fourier transform of the point spread functions in (a) and (b) are shown in blue in (e) and (f) to indicate the interference bandwidth. They are overlapped with the input source spectra shown in red.
The OCT imaging results presented here suggest that ultrahigh resolution OCT can be achieved at both 800 nm and 1300 nm with a single compact, stable, and user-friendly source with sufficient output power for in vivo imaging applications. While it is unlikely that continuum generation sources will find widespread use for ultrahigh resolution retinal imaging at 800 nm due to the availability of broadband superluminescent diode (SLD) sources [13], they offer a very promising alternative to SLD’s for ultrahigh resolution imaging in other applications where more power and/or longer wavelengths are required. The 1300 nm source demonstrated here, for example, may with further development provide a widely available replacement for the Cr:Forsterite modelocked laser source currently used in ultrahigh resolution endoscopic OCT [16]. It is feasible that the light source and fiber can be integrated into a single package to provide a completely turn-key operation, high power ultrahigh resolution OCT source at 1300 nm. Currently, the performance of this light source is not equal to the Cr:Forsterite source due to coherence artifact from spectral modulation, but based on the numerical simulations presented in section 2 and on the work of others [33, 42, 44], it appears that generation of smoother continuum will be realizable with further optimization of such fibers. Dispersion management solutions using combinations of fibers with different dispersion profiles [45] or tapering of photonic crystal fibers are particularly promising [44]. It should be noted that, while these sources were demonstrated for time-domain ultrahigh resolution OCT imaging, they can also be implemented with spectrometer based spectral domain OCT systems for high speed, ultrahigh resolution OCT at 800 nm and 1300 nm [46–49].
The paradigm of using theoretical analysis and numerical simulation to custom design photonic crystal fibers that achieve a desired spectral distribution will be extremely powerful for applications such as OCT that are critically dependent on the precise spectral characteristics. The novel PCF design with two closely spaced zero dispersion wavelengths offers a solution to the problems of excess noise and low power efficiency that have hindered the use of supercontinuum generation for ultrahigh resolution OCT in the 1300 nm wavelength range. Furthermore, the ability to tune the wavelengths of the main peaks by setting the zero dispersion wavelengths opens up the possibility that investigators can achieve ultrahigh resolution imaging at any wavelength throughout the near infrared by using a single pump laser source with a specific fiber. This would greatly expand the capabilities of research groups by enabling multiple expensive laser systems to be replaced with a single system and several optical fibers.
Fig. 9. Ultrahigh resolution OCT images of formalin fixed hamster cheek pouch at (a) 800 nm and (b) 1300 nm. Enhanced image penetration is noticeable at 1300 nm and important features such as the epithelium (e), connective tissue bands (c), and muscular layers (m) can be distinguished.
4. Conclusion
We have developed and demonstrated a light source for ultrahigh resolution OCT imaging at 800 nm and 1300 nm based on a novel photonic crystal fiber with two closely spaced zero dispersion wavelengths. The source was developed entirely with commercially available components, including the photonic crystal fiber and the turnkey Nd:Glass femtosecond oscillator. High sensitivity imaging was achieved with < 5 μm axial resolution in tissue at both wavelengths, although coherence artifacts from spectrum modulation degraded the performance somewhat. To address this, we have studied using numerical simulations the conditions for generating smooth spectral peaks in the PCF. Our results suggest that optimization of the continuum can be achieved by tuning the pump dispersion wavelength relative to the zero dispersion wavelengths of the fiber. This work provides preliminary data that should motivate additional development and optimization of similar photonic crystal based light sources for ultrahigh resolution OCT.
Acknowledgments
This research is supported in part by National Institutes of Health RO1-CA75289-07 and RO1–EY11289–20, National Science Foundation ECS–01–19452 and BES–0522845, Air Force Office of Scientific Research Medical Free Electron Laser Program F49620–01–1–0186 and F49620–01–01–0084, the Poduska Family Foundation Fund for Innovative Research in Cancer, and the philantrophy of Mr. G. Andlinger. A.D. Aguirre acknowledges graduate fellowship support from the Whitaker Foundation. N. Nishizawa is visiting from the Department of Quantum Engineering, Nagoya University, Japan, and is supported by the Telecommunications Advancement Foundation (TAF).
References and links
1. D. Huang, E. A. Swanson, C. P. Lin, J. S. Schuman, W. G. Stinson, W. Chang, M. R. Hee, T. Flotte, K. Gregory, C. A. Puliafito, and J. G. Fujimoto, “Optical coherence tomography,” Science 254, 1178–1181 (1991). [CrossRef] [PubMed]
2. J. G. Fujimoto, “Optical coherence tomography for ultrahigh resolution in vivo imaging,” Nat Biotechnol 21, 1361–1367 (2003). [CrossRef] [PubMed]
3. B. Bouma, G. J. Tearney, S. A. Boppart, M. R. Hee, M. E. Brezinski, and J. G. Fujimoto, “High-resolution optical coherence tomographic imaging using a mode-locked Ti:Al2O3 laser source,” Optics Letters 20, 1486–1488 (1995). [CrossRef] [PubMed]
4. W. Drexler, U. Morgner, F. X. Kartner, C. Pitris, S. A. Boppart, X. D. Li, E. P. Ippen, and J. G. Fujimoto, “In vivo ultrahigh-resolution optical coherence tomography,” Optics Letters 24, 1221–1223 (1999). [CrossRef]
5. W. Drexler, U. Morgner, R. K. Ghanta, F. X. Körtner, J. S. Schuman, and J. G. Fujimoto, “Ultrahigh-resolution ophthalmic optical coherence tomography,” Nat Med 7, 502–507 (2001). [CrossRef] [PubMed]
6. A. M. Kowalevicz, T. R. Schibli, F. X. Kartner, and J. G. Fujimoto, “Ultralow-threshold Kerr-lens mode-locked Ti:Al2O3 laser,” Optics Letters 27, 2037–2039 (2002). [CrossRef]
7. A. Unterhuber, B. Povazay, B. Hermann, H. Sattmann, W. Drexler, V. Yakovlev, G. Tempea, C. Schubert, E. M. Anger, P. K. Ahnelt, M. Stur, J. E. Morgan, A. Cowey, G. Jung, T. Le, and A. Stingl, “Compact, low-cost Ti:Al2O3 laser for in vivo ultrahigh-resolution optical coherence tomography,” Optics letters 28, 905–907 (2003). [CrossRef] [PubMed]
8. D. L. Marks, A. L. Oldenburg, J. J. Reynolds, and S. A. Boppart, “Study of an ultrahigh-numerical-aperture fiber continuum generation source for optical coherence tomography,” Optics Letters 27, 2010–2012 (2002). [CrossRef]
9. P. C. Wagenblast, T. H. Ko, J. G. Fujimoto, F. X. Kaertner, and U. Morgner, “Ultrahigh-resolution optical coherence tomography with a diode-pumped broadband Cr3+: LiCAF laser,” Optics Express 12, 3257–3263 (2004), http://www.opticsinfobase.org/abstract.cfm?id=80534. [CrossRef] [PubMed]
10. L. Vabre, A. Dubois, and A. C. Boccara, “Thermal-light full-field optical coherence tomography,” Optics Letters 27, 530–532 (2002). [CrossRef]
11. A. M. Kowalevicz, T. Ko, I. Hartl, J. G. Fujimoto, M. Pollnau, and R. P. Salathe, “Ultrahigh resolution optical coherence tomography using a superluminescent light source,” Optics Express 10, 349–353 (2002), http://www.opticsinfobase.org/abstract.cfm?id=68496. [PubMed]
12. C. Grivas, T. C. May-Smith, D. P. Shepherd, R. W. Eason, M. Pollnau, and M. Jelinek, “Broadband single-transverse-mode fluorescence sources based on ribs fabricated in pulsed laser deposited Ti : sapphire waveguides,” Appl Phys a-Mater 79, 1195–1198 (2004).
13. T. H. Ko, D. C. Adler, J. G. Fujimoto, D. Mamedov, V. Prokhorov, V. Shidlovski, and S. Yakubovich, “Ultrahigh resolution optical coherence tomography imaging with a broadband superluminescent diode light source,” Optics Express 12, 2112–2119 (2004), http://www.opticsinfobase.org/abstract.cfm?id=79925. [CrossRef] [PubMed]
14. J. M. Schmitt, A. Knuttel, M. Yadlowsky, and M. A. Eckhaus, “Optical-coherence tomography of a dense tissue: statistics of attenuation and backscattering,” Physics in Medicine and Biology 39, 1705–1720 (1994). [CrossRef] [PubMed]
15. B. E. Bouma, G. J. Tearney, I. P. Bilinsky, B. Golubovic, and J. G. Fujimoto, “Self-phase-modulated Kerr-lens mode-locked Cr:forsterite laser source for optical coherence tomography,” Optics Letters 21, 1839–1841 (1996). [CrossRef] [PubMed]
16. P. R. Herz, Y. Chen, A. D. Aguirre, J. C. Fujimoto, H. Mashimo, J. Schmitt, A. Koski, J. Goodnow, and C. Petersen, “Ultrahigh resolution optical biopsy with endoscopic optical coherence tomography,” Optics Express 12 (2004), http://www.opticsinfobase.org/abstract.cfm?id=80642. [CrossRef] [PubMed]
17. J. K. Ranka, R. S. Windeler, and A. J. Stentz, “Visible continuum generation in air-silica microstructure optical fibers with anomalous dispersion at 800 nm,” Optics Letters 25, 25–27 (2000). [CrossRef]
18. I. Hartl, X. D. Li, C. Chudoba, R. K. Hganta, T. H. Ko, J. G. Fujimoto, J. K. Ranka, and R. S. Windeler, “Ultrahigh-resolution optical coherence tomography using continuum generation in an air-silica microstructure optical fiber,” Optics Letters 26, 608–610 (2001). [CrossRef]
19. B. Povazay, K. Bizheva, A. Unterhuber, B. Hermann, H. Sattmann, A. F. Fercher, W. Drexler, A. Apolonski, W. J. Wadsworth, J. C. Knight, P. S. J. Russell, M. Vetterlein, and E. Scherzer, “Submicrometer axial resolution optical coherence tomography,” Optics Letters 27, 1800–1802 (2002). [CrossRef]
20. W. Drexler, “Ultrahigh-resolution optical coherence tomography,” Journal of Biomedical Optics 9, 47–74 (2004). [CrossRef] [PubMed]
21. B. Povazay, K. Bizheva, B. Hermann, A. Unterhuber, H. Sattmann, A. F. Fercher, W. Drexler, C. Schubert, P. K. Ahnelt, M. Mei, R. Holzwarth, W. J. Wadsworth, J. C. Knight, and P. S. Russel, “Enhanced visualization of choroidal vessels using ultrahigh resolution ophthalmic OCT at 1050 nm,” Optics Express 11, 1980–1986 (2003), http://www.opticsinfobase.org/abstract.cfm?id=74008. [CrossRef] [PubMed]
22. Y. Wang, Y. Zhao, J. S. Nelson, Z. Chen, and R. S. Windeler, “Ultrahigh-resolution optical coherence tomography by broadband continuum generation from a photonic crystal fiber,” Optics Letters 28, 182–184 (2003). [CrossRef] [PubMed]
23. S. Bourquin, A. D. Aguirre, I. Hartl, P. Hsiung, T. H. Ko, J. G. Fujimoto, T. A. Birks, W. J. Wadsworth, U. Bunting, and D. Kopf, “Ultrahigh resolution real time OCT imaging using a compact femtosecond Nd : Glass laser and nonlinear fiber,” Optics Express 11, 3290–3297 (2003), http://www.opticsinfobase.org/abstract.cfm?id=78091. [CrossRef] [PubMed]
24. H. Lim, Y. Jiang, Y. Wang, Y. C. Huang, Z. Chen, and F. W. Wise, “Ultrahigh-resolution optical coherence tomography with a fiber laser source at 1 microm,” Opt Lett 30, 1171–1173 (2005). [CrossRef] [PubMed]
25. Y. Wang, J. S. Nelson, Z. Chen, B. J. Reiser, R. S. Chuck, and R. S. Windeler, “Optimal wavelength for ultrahigh-resolution optical coherence tomography,” Optics Express 11, 1411–1417 (2003), http://www.opticsinfobase.org/abstract.cfm?id=72649. [CrossRef] [PubMed]
26. N. Nishizawa, Y. Chen, P. Hsiung, E. P. Ippen, and J. G. Fujimoto, “Real-time, ultrahigh-resolution, optical coherence tomography with an all-fiber, femtosecond fiber laser continuum at 1.5 mu m,” Optics Letters 29, 2846–2848 (2004). [CrossRef]
27. T. Hori, N. Nishizawa, T. Goto, and M. Yoshida, “Wideband and nonmechanical sonogram measurement by use of an electronically controlled, wavelength-tunable, femtosecond soliton pulse,” J Opt Soc Am B 20, 2410–2417 (2003). [CrossRef]
28. K. L. Corwin, N. R. Newbury, J. M. Dudley, S. Coen, S. A. Diddams, K. Weber, and R. S. Windeler, “Fundamental noise limitations to supercontinuum generation in microstructure fiber,” Physical Review Letters 90, 113904/1–4 (2003). [CrossRef] [PubMed]
29. N. R. Newbury, B. R. Washburn, K. L. Corwin, and R. S. Windeler, “Noise amplification during supercontinuum generation in microstructure fiber,” Optics Letters 28, 944–946 (2003). [CrossRef] [PubMed]
30. A. Apolonski, B. Povazay, A. Unterhuber, W. Drexler, W. J. Wadsworth, J. C. Knight, and P. S. Russell, “Spectral shaping of supercontinuum in a cobweb photonic-crystal fiber with sub-20-fs pulses,” J Opt Soc Am B 19, 2165–2170 (2002). [CrossRef]
31. K. Bizheva, B. Povazay, B. Hermann, H. Sattmann, W. Drexler, M. Mei, R. Holzwarth, T. Hoelzenbein, V. Wacheck, and H. Pehamherger, “Compact, broad-bandwidth fiber laser for sub-2- mu m axial resolution optical coherence tomography in. The 1300-nm wavelength region,” Optics Letters 28, 707–709 (2003). [CrossRef] [PubMed]
32. P. L. Hsiung, Y. Chen, T. H. Ko, J. G. Fujimoto, C. J. S. de Matos, S. V. Popov, J. R. Taylor, and V. P. Gapontsev, “Optical coherence tomography using a continuous-wave, high-power, Raman continuum light source,” Optics Express 12, 5287–5295 (2004), http://www.opticsinfobase.org/abstract.cfm?id=81626. [CrossRef] [PubMed]
33. K. M. Hilligsoe, T. V. Andersen, H. N. Paulsen, C. K. Nielsen, K. Molmer, S. Keiding, Y. Kristiansen, K. P. Hansen, and J. J. Larsen, “Supercontinuum generation in a photonic crystal fiber with two zero dispersion wavelengths,” Optics Express 12, 1045–1054 (2004), http://www.opticsinfobase.org/abstract.cfm?id=79252. [CrossRef] [PubMed]
34. T. Hori, N. Nishizawa, T. Goto, and M. Yoshida, “Experimental and numerical analysis of widely broadened supercontinuum generation in highly nonlinear dispersion-shifted fiber with a femtosecond pulse,” J Opt Soc Am B 21, 1969–1980 (2004). [CrossRef]
35. G. P. Agrawal, Nonlinear Fiber Optics (Academic Press, London, 2001).
36. A. V. Husakou and J. Herrmann, “Supercontinuum generation of higher-order solitons by fission in photonic crystal fibers,” Physical Review Letters 8720, art. no.-203901 (2001).
37. A. L. Gaeta, “Nonlinear propagation and continuum generation in microstructured optical fibers,” Optics Letters 27, 924–926 (2002). [CrossRef]
38. B. R. Washburn, S. E. Ralph, and R. S. Windeler, “Ultrashort pulse propagation in air-silica microstructure fiber,” Optics Express 10, 575–580 (2002), http://www.opticsinfobase.org/abstract.cfm?id=69331. [PubMed]
39. G. Genty, M. Lehtonen, H. Ludvigsen, J. Broeng, and M. Kaivola, “Spectral broadening of femtosecond pulses into continuum radiation in microstructured fibers,” Optics Express 10, 1083–1098 (2002), http://www.opticsinfobase.org/abstract.cfm?id=70205. [PubMed]
40. S. Coen, A. H. L. Chau, R. Leonhardt, J. D. Harvey, J. C. Knight, W. J. Wadsworth, and P. S. J. Russell, “Supercontinuum generation by stimulated Raman scattering and parametric four-wave mixing in photonic crystal fibers,” J Opt Soc Am B 19, 753–764 (2002). [CrossRef]
41. A. V. Husakou and J. Herrmann, “Supercontinuum generation, four-wave mixing, and fission of higher-order solitons in photonic-crystal fibers,” J Opt Soc Am B 19, 2171–2182 (2002). [CrossRef]
42. M. H. Frosz, P. Falk, and O. Bang, “The role of the second zero-dispersion wavelength in generation of supercontinua and bright-bright soliton-pairs across the zero-dispersion wavelength,” Optics Express 13, 6181–6192 (2005), http://www.opticsinfobase.org/abstract.cfm?id=85282. [CrossRef] [PubMed]
43. S. R. Chinn and E. A. Swanson, “Blindness Limitations in Optical Coherence Domain Reflectometry,” Electronics Letters 29, 2025–2027 (1993). [CrossRef]
44. P. Falk, M. H. Frosz, and O. Bang, “Supercontinuum generation in a photonic crystal fiber with two zero-dispersion wavelengths tapered to normal dispersion at all wavelengths,” Optics Express 13, 7535–7540 (2005), http://www.opticsinfobase.org/abstract.cfm?id=85486. [CrossRef] [PubMed]
45. T. Hori, J. Takayanagi, N. Nishizawa, and T. Goto, “Flatly broadened, wideband and low noise supercontinuum generation in highly nonlinear hybrid fiber,” Optics Express 12, 317–324 (2004), http://www.opticsinfobase.org/abstract.cfm?id=78593. [CrossRef] [PubMed]
46. M. Wojtkowski, T. Bajraszewski, P. Targowski, and A. Kowalczyk, “Real-time in vivo imaging by highspeed spectral optical coherence tomography,” Optics Letters 28, 1745–1747 (2003). [CrossRef] [PubMed]
47. M. Wojtkowski, V. J. Srinivasan, T. H. Ko, J. G. Fujimoto, A. Kowalczyk, and J. S. Duker, “Ultrahigh-resolution, high-speed, Fourier domain optical coherence tomography and methods for dispersion compensation,” Optics Express 12, 2404–2422 (2004), http://www.opticsinfobase.org/abstract.cfm?id=80147. [CrossRef] [PubMed]
48. N. Nassif, B. Cense, B. H. Park, S. H. Yun, T. C. Chen, B. E. Bouma, G. J. Tearney, and J. F. de Boer, “In vivo human retinal imaging by ultrahigh-speed spectral domain optical coherence tomography,” Opt Lett 29, 480–482 (2004). [CrossRef] [PubMed]
49. S. H. Yun, G. J. Tearney, B. E. Bouma, B. H. Park, and J. F. de Boer, “High-speed spectral-domain optical coherence tomography at 1.3 mu m wavelength,” Optics Express 11, 3598–3604 (2003), http://www.opticsinfobase.org/abstract.cfm?id=78225. [CrossRef] [PubMed]
