1. Introduction

In recent years, picosecond and femtosecond optical pulses have been used in biophotonic applications such as the inducing of nonlinear multiphoton bioimaging and coherent anti-Stokes Raman scattering (CARS) [1–3]. As ultrashort optical pulse sources, mode-locked solid-state lasers are generally used for two-photon bioimaging [1,4–7]. However, these lasers are large, costly, and not maintenance free. Furthermore, controlling optical pulse features such as repetition rate and electronic synchronization is not easy. Therefore, for real world biophotonic applications it is necessary to develop a turnkey ultrashort optical pulse source that is simple, compact, and low cost.

In this paper we describe a simple scheme for kilowatt-peak-power picosecond optical pulse generation at a wavelength of 0.77 μm using a gain-switched laser diode (LD), and its successful application for nonlinear bioimaging using two-photon fluorescence microscopy.

2. High-peak-power optical pulse generation

The configuration for high-peak-power optical pulse generation and two-photon microscopy is shown in Fig. 1. Picosecond optical pulses are generated by the gain-switching of an InGaAsP distributed-feedback-Bragg structure laser-diode (DFB-LD). The operation wavelength is 1548 nm under gain-switching [8]. We developed an electronic pulse generator, which could produce 5 V amplitude and 100 ps duration electronic pulses at a flexible repetition rate up to 200 MHz. Under the excitation with these electronic pulses, the DFB-LD directly generated optical pulses less than 7 ps duration without any optical pulse compression techniques.

 

Fig. 1. Experimental configuration for generating high-peak-power picosecond optical pulses with a semiconductor laser. The two-photon fluorescence microscope is also indicated in the dashed-line-surrounding area.

Download Full Size | PDF

As a preamplifier of optical pulses, we used a low-average-power (10 mW saturation power) erbium-doped fiber amplifier (EDFA). After preamplification, optical pulses were filtered out with a 1 nm bandwidth optical filter to remove the wideband spontaneous emission noise.

However, as the main optical amplifier, we employed a low-nonlinear-effect EDFA to avoid serious, unintentional spectral distortions by self-phase-modulation (SPM) inside the EDFA [9]. This main EDFA was specially designed for the present experiment [10], and the active fiber length was shortened to 150 mm. By this main optical amplifier, the maximum average output power of approximately 20 mW was possible. Note that with 10 mW average optical power at pulse repetition rate of 1 MHz, even if pulse duration is 10 ps, the optical peak power reaches 1 kW. At this power level, we observed a slight spectral broadening by SPM. This broadening was, however, approximately 1/10 (at -10 dB level from the peak) of the SPM broadening in a conventional EDFA whose active fiber is several meters long. Figure 2 indicates the optical spectra at 1 kW peak power after the present main EDFA (a), and after a conventional EDFA (b). Suppression of the SPM spectral broadening can assure a conversion efficiency increase in the second-harmonic generation (SHG).

 

Fig. 2. Optical spectra at 1 kW peak power optical pulses from the low-nonlinear-effect EDFA (a) and a conventional EDFA (b).

Download Full Size | PDF

It should be noted that in the present configuration, there were not any high-average-power optical devices. The mechanism by which we can obtain high-peak-power optical pulses is attributed to a long upper-laser-level lifetime of several milliseconds in the EDFA. If the repetition rate of the incident optical pulses is decreased, the energy stored in the EDFA during the pulse interval time is increased, and each optical pulse can receive higher saturation energy after amplification. If we decrease the pulse repetition rate less than a megahertz range, amplified-spontaneous-emission (ASE) power will exceed the average power of amplified optical pulses. Therefore, we operated the DFB-LD mainly at a repetition rate of more than 1 MHz.

3. High-efficiency wavelength conversion

After the main EDFA, 1548 nm optical pulses were converted to second-harmonic (SH) light by a bulk periodically-poled MgO-doped LiNbO₃ (PPMgLN) device [11], and then several milliwatt average-power SH output was obtained.

We mainly operated the DFB-LD to generate 7 ps optical pulses. With this pulse duration, the average optical output power from main optical amplifier at 1 MHz was approximately 10 mW, thus the optical peak power reached 1.4 kW. After SHG wavelength conversion by the PPMgLN device, we obtained 5.5 mW average optical output power after a 780 nm optical filter; the SHG power conversion efficiency was as high as 50%. In the following experiment, the peak power and average power were detected just after the 780 nm optical filter. Since the temporal width of SH optical pulses was measured to be 5 ps, the SH-pulse peak power reached 1.1 kW. This peak power will be high enough for two-photon bioimaging.

4. Two-photon imaging

After SHG wavelength conversion by the PPMgLN device, we directed the laser beam into a fluorescence microscope (Olympus IX71) that has been modified for two-photon imaging. The focus of the laser beam is scanned on a specimen by a two perpendicular-axis pair of galvanometer mirrors through a pupil-transfer-lens into the side port of the inverted microscope. To change the focal point depth in the specimen, a stepping motor was used. Finally, the beam is focused by a 60 × and 1.2 NA water-immersion objective lens onto a specimen. Two-photon fluorescence was collected by a photomultiplier tube (PMT).

The two-photon fluorescent intensity I increases with the duration of the excitation pulse τ, the square of the peak power P_peak, and with the reciprocal of optical pulse period T.

The factor k depends on the two-photon absorption cross section of the fluorophore at the laser wavelength, and on the spatial energy distribution in the focus. Because the optical pulse width is fixed in the experiment, the two-photon fluorescence intensity depends on the pulse period and square of peak power. The peak-power of the laser is represented as $P_{\mathit{peak}}=\frac{T}{\tau }{P}_{\mathit{av}}$, where P_av is the average power, then Eq. (1) is rewritten to be

Therefore, even with low average power and pulse duration up to a few picoseconds, we can improve the two-photon fluorescence intensity by decreasing the pulse repetition rate. At repetition rates of 1 MHz, 2 MHz, and 20 MHz, the peak power was approximately 1.1 kW, 0.84 kW, and 0.06 kW, respectively. Then, the transmission of the laser light through the optics of the two-photon microscope onto a specimen was found to be 24%.

At different pulse repetition rates of 2 MHz and 20 MHz, the feasibility of two-photon imaging is shown in Fig. 3. The diameter of polystyrene microspheres is about 2.0 μm and their single-photon absorption peak is 505 nm, 535 nm, and 580 nm, respectively.

 

Fig. 3. Two-photon images of polystyrene microspheres taken with 774 nm optical pulses at repetition rate of 2 MHz (a) and 20 MHz (b).

Download Full Size | PDF

When the repetition rate was 2 MHz, the average power is 6.2 mW (1.5 mW on the specimen). While the repetition rate was increased up to 20 MHz, the average power is 8.3-mW (2.0 mW on the specimen). Based on the relation of Eq. (2), the fluorescence light intensity for 2 MHz is expected to be more than five times higher than that for 20 MHz. This is clearly shown in Fig. 3. The fluorescence of 2 MHz provides an image much brighter than that of 20 MHz.

 

Fig. 4. Two-photon excited fluorescence intensity image of actin filaments in PtK2 cell stained with Alexa Fluor 488. The optical peak power was 1.1 kW at repetition rate of 1 MHz.

Download Full Size | PDF

Since our optical pulse source can induce high resolution two-photon imaging, we next attempted two-photon imaging of a biological sample. As a bioimaging specimen, we used epithelial cells (PtK2) of a rat kangaroo kidney, which were stained with Alexa Fluor 488. The single-photon absorption peak of Alexa Fluor 488 is 495 nm, and the emission maximum occurs at a wavelength of 520 nm. As illustrated in Fig. 4, we have obtained a very clear image of actin filaments in PtK2 cells with 1.1 kW peak-power optical pulses at repetition rate of 1 MHz. These results, shown in Figs. 3 and 4, clearly indicate the usefulness of the present picosecond optical source for nonlinear bioimaging.

5. Conclusions

In conclusion, we have generated kilowatt peak power, 1.55 μm picosecond optical pulses using a gain-switched semiconductor laser and low-average-power fiber amplifiers. Using a bulk PPMgLN crystal, SH optical pulses of 774 nm wavelength and 5 ps duration were generated with the maximal conversion efficiency of 50%; the SHG optical pulse peak power exceeded 1 kW at repetition rate of 1 MHz. Subsequently, employing these SHG optical pulses we successfully demonstrated high-resolution and high-contrast two-photon bioimaging for actin filaments in PtK2 cells labeled with Alexa Fluor 488.

We should emphasize that there are not any large-size high-average-power optical devices in the experiment. Our present results will stimulate many applications of stable and compact semiconductor-laser-based optical pulse sources in biophotonics.