1. Introduction

Second harmonic generation (SHG) is a process where two photons simultaneously interact with non-centrosymmetrical structures without absorption, producing radiation at exactly half of the excitation wavelength [1–3]. SHG allows for the unique ability to observe fine structural details of connective tissues such as collagen, tendon and muscle in animals and humans without the aid of contrast agents [4, 5].

The properties of SHG offer several advantages for live cell or tissue imaging. Unlike fluorescence imaging, SHG does not involve excitation of endogenous florescent molecules. Thus, it does not suffer from phototoxicity effects or photo bleaching, which limits the usefulness of fluorescence imaging [6, 7]. The excitation of SHG imaging uses near-infrared wavelengths, which allows for a large penetration depth, making it a prime candidate for analysis of thick tissue specimens with the enhanced axial depth discrimination. SHG signals have well defined polarization, so SHG polarization anisotropy can be used to determine the absolute orientation and the degree of organization of proteins in tissues [7–9]. The SHG excitation is a nonlinear process and the signal intensity is directly proportional to the square of the excitation intensity. This provides SHG imaging with intrinsic confocality and deep sectioning in complex tissues.

SHG microscopy is presently being introduced as a tool for basic biomedical research to noninvasively visualize and obtain high-resolution images of unstained collagen structures [1–5, 10]. However, to date almost all research in SHG imaging has been carried out using a conventional bench top microscope system. Such a system does not provide the adjustment flexibility needed for in vivo imaging, where specimens need to be mounted on the microscopy stage, limits the system’s use. A biopsy is required for SHG imaging which will introduce intrusion to animal or human and variation of collagen structures because of excision from a live body. An endomicroscope system having a long flexible fiber and a miniaturized probe is necessary for translating SHG imaging into in vivo imaging [11–15].

In this letter, we demonstrate SHG imaging using our recently developed compact handheld nonlinear endomicroscope which is capable of in vivo 3D imaging [16]. Experimental results show that the core of a double clad fiber (DCF) which the endomicroscope uses has a “polarization maintaining” feature in particular directions. Using this “polarization marinating” feature, the nonlinear endomicroscope can observe clear 3D collagen structure of a mouse tail tendon within a depth of 88 µm. High resolution SHG imaging from the endomicroscope has shown more micro-structure information of biological tissue than two-photon fluorescence imaging.

2. The nonlinear endomicroscopic system for SHG imaging

The SHG endomicroscope is schematically depicted in Fig. 1 , where a passive mode-locked Ti:Sapphire laser generates an 80 MHz 100 fs optical pulse train with wavelength tuneable between 730 nm and 870 nm. 800 nm pulses can introduce high SHG signal from mouse tail tendon and are used for SHG imaging. The pulses propagate through a pair of gratings which provide a negative frequency chirp. The pulses pass through a half wave plate (λ/2 plate) and a Glen Thompson polarizer being used to control the polarization direction of the optical pulses by rotating the direction of the λ/2 plate and the Glen Thompson polarizer accordingly. The optical pulses are then coupled into the core of a 2.5 m long DCF by an objective lens (Olympus MA20X/0.4), where the chromatic dispersion of the DCF is compensated by the pre-chirp in the pulses. The pulse width at the output of the endomicroscope is 1.5 ps.

 

Fig. 1 Schematic representation of experimental setup for SHG imaging using the designed nonlinear endomicroscope.

Download Full Size | PDF

The DCF, with a core diameter of 3.6 µm and an inner cladding diameter of 105 µm, is used for the delivery of the excitation laser beam to the specimen and the collection of the SHG signal back for detection. The other end of the DCF is placed inside the endomicroscope probe. The detail structure of the probe can be found in the previous report [17, 18]. In the probe, a multiple element lens focuses the excitation laser beam delivered by the DCF into the specimen. The numerical aperture (NA) of the lens is 0.35. The lens well corrects the aberration within the field of view and realizes diffraction limited resolution [16–18], which reaches better than 1 μm and 10 μm of lateral and axial resolutions. A micro-scanner inside the probe scans the DCF line by line and moves the lens with the DCF backward and forward for 3D imaging. The excited SHG signal is collected through the lens and the cladding of the DCF, separated from the excitation laser beam by a dichroic mirror and coupled to another DCF by another objective lens (Olympus MA20X/0.4) which sends the SHG signal into the SHG endomicroscope detection box. Inside the detection box, a photomultiplier tube (PMT) is employed to detect the SHG signal after it is filtered by a 400/9 nm bandpass filter to ensure that only the SHG signal is detected. The endomicroscope’s electronics and the host CPU synchronize the endomicroscope system and display SHG images with a speed of 2 frames per second.

3. Degree of polarization of the excitation laser beam delivered by the nonlinear endomicroscope

The DCF that the nonlinear endomicroscope uses has a circular core and it should not be polarization maintaining. As a linearly polarized excitation laser beam is coupled into the core of the DCF, we found that the degree of polarization, defined as γ = (I_max - I_min)/(I_max + I_min) by measuring the maximum (I_max) and minimum (I_min) intensity of the output laser beam, changes by varying the polarization direction of the incident linearly polarized laser beam. Figure 2 shows the measured degree of polarization of the output laser beam at wavelength 800 nm from the fiber probe as a function of the incident polarization direction. The degree of polarization γ of the excitation laser beam from the output of the probe is changed between 0.35 and 0.96 for different incident polarization directions. The perfect circular structure of the core of the DCF should not have such features. However, minor deviations in the symmetry of the DCF during fabrication causes the 2.5 m DCF “polarization maintaining” along two directions: 0° and 90°. Since the DCF is only 2.5 m long and free running, polarization maintaining of the DCF is stable during imaging.

 

Fig. 2 Degree of polarization of the excitation laser beam and signal to noise ratio of SHG imaging as a function of the polarization direction of the input linearly polarized laser beam

Download Full Size | PDF

The degree of polarization of the excitation laser beam from the nonlinear endomicroscope directly affects the signal to noise ratio of SHG imaging. The SHG intensity and the background noise of SHG imaging of a mouse tail tendon fiber by the nonlinear endomicroscope are measured. The signal to noise ratio (SNR), defined as the ratio of the SHG intensity to the background noise, is also displayed in Fig. 2. The SNR of SHG imaging has a similar trend as the degree of polarization γ of the excitation laser from the endomicroscope. SHG imaging has a high SNR as the output laser beam is close to be linearly polarized, and the SNR of SHG imaging reduces while γ of the output laser beam decreases.

4. Three-dimensional SHG imaging by nonlinear endomicroscopy

The nonlinear endomicroscope permits 3D anatomical and functional imaging of organs with a large field of view and high resolution [9]. Figure 3 is 3D SHG images of a mouse tail tendon obtained by the nonlinear endomicroscope with axial distance of 8 µm, where polarization direction of incident laser beam is 0°, aligning with the “polarization maintaining” direction of the DCF. The size of the images is 475 µm × 475 µm. Figure 3 shows clear structural details of collagen in the mouse tail tendon at different depth. This nonlinear endomicroscope can observe the collagen structure of a mouse tail tendon 88 µm under the surface.

 

Fig. 3 SHG images of a mouse tail tendon with an axial step of 8 µm. The image size is 475 µm × 475 µm. The excitation wavelength is 800 nm. The excitation power is 30 mW on the sample. Arrows are the polarization directions of the excitation laser beam

Download Full Size | PDF

5. SHG imaging and two-photon excited fluorescence imaging by nonlinear endomicroscopy

The nonlinear endomicroscope can observe two-channel of images – a SHG image and a two-photon excited fluorescence (TPF) image, by injecting a mouse with fluorescein and switching between the filter for SHG imaging and the filter for TPF imaging, which is a 3 mm thick BG18 glass filter (Schott Pty Ltd). Figure 4 reveals high resolution two-channel images of a mouse tail tendon.

 

Fig. 4 Two channel images of a mouse tail tendon by the nonlinear endomicroscope. (a), (d) SHG images. (b), (e) TPF images. (c), (f) Combination of two-channel images. Red: TPF image. Green: SHG image. The size of the images: 150 µm × 150 µm. Arrows are the polarization directions of the excitation laser beam

Download Full Size | PDF

The structure of the SHG images and the TPF images is different. The TPF images show the distribution of fluorescein in the mouse tail tendon absorbed from the bloodstream. On the other hand, SHG images display non-centrosymmetrical orientation and structures of proteins in the mouse tail tendon. Therefore, SHG offers the nonlinear endomicroscope with another channel of image which helps to understand orientation and degree of organisation of proteins and collagen fiber in the tissues.

5. Conclusions

The nonlinear endomicroscope can be used for in vivo SHG imaging. The experimental results show that the degree of polarization of the excitation laser beam from the output of the endomicroscopy probe changes with different incident polarization directions. The circular DCF that the nonlinear endomicroscope uses shows a “polarization maintaining” feature at particular directions due to minor deviations in the symmetry of the DCF during fabrication. As a result, the SNR of the SHG images obtained by the endomicroscope changes with different incident polarization directions. With the polarization direction of the incident laser beam aligning with the “polarization maintaining” direction of the DCF, the nonlinear endomicroscope shows clear 3D collagen structures of a mouse tail tendon. Comparing with two-photon excited fluorescence imaging by the nonlinear endomicroscope, the high resolution SHG imaging offers further information about the orientation and degree of organisation of proteins and collagen fiber in biological objects.