1. Introduction

Phase-shifted fibre Bragg grating (PSFBG), which is showing promise as an adaptable high finesse transmission filter or switching element for future dense WDM optical communication systems [1,2,3], is one of the most successful types of bandpass filter. Its transmission spectra can be characterized by dips at wavelengths that satisfy the resonance condition. PSFBG has also found a wide range of applications in discrete feedback Bragg (DFB) laser [4], and sensors [5] for temperature, strain, vibration, and chemical measurements.

A PSFBG can be inscribed into photosensitive fiber by use of the phase-shift phase mask technique [6] or the UV post-processing technique [7]. However, high quality phase masks are very expensive and UV post-processing requires extra long exposed time.

External perturbations techniques, such as using heat [8] or tension [9], have also been reported. But the implementations of such schemes are mostly complex and costly. Furthermore, the temporary phase shift will be removed from the FBG in the absence of external disturbance.

In this letter, on the other hand, we proposed a new technique using CO₂ laser pulses to obtain the PSFBGs. As neither UV exposure nor hydrogen loading is required, the CO₂ laser has been widely used to produce the long period grating (LPG) on single mode fiber and even on photonic crystal fiber [10,11]. The effect of CO₂ laser irradiation on the refractive index change in optical fibers has been investigated by Bok Hyeon Kim, et al [12]. Now, we discover that the phase shifted bandpass filter can be written on the fiber Bragg grating (FBG) with 10.6μm freespace wavelength CO₂ laser pulses. This new proposed technique based on the residual stress relaxation effect [12] has the advantage of effectively controlling the spectral characteristics of the fiber grating structure, introducing the permanent phase shift, fast and compact fabrication, etc.

2. Experiment setup

A schematic diagram of the experimental setup is shown in Fig. 1. The setup includes a CO₂ laser with galvanometer that directs and focuses the beam diameter to an ~240μm Gaussian focal spot on the fiber through a lens and mirror, two fiber holders, and two CCD cameras. Each CCD camera is mounted at a 45° angle to capture the fiber image and provide visual verification that the energy is generated. One side of the fiber is fixed on the left fiber holder, whose movement can be controlled with translation stages attached to it. A fixed weight of 3.3g is attached to the other side of fiber over the right fiber holder to provide constant tension and to keep the fiber straight during the experiment. An optical spectrum analyzer with a light source is used to trace the transmission spectrum profiles of the gratings in situ.

 

Fig. 1. Schematic diagram of the experimental setup

Download Full Size | PDF

The fiber Bragg grating, used in our experiment, is inscribed previously by UV light into the photosensitive fiber. The reflectivity is about 20dB and the grating length is 12mm. Then, this grating is put onto the fiber holders and exposed by the CO₂ laser pulses. The output power from each CO₂ laser pulse is almost constant during the experiment, and no physical deformation is observed after the irradiation, which ensures that this kind of PSFBG is compact and easy to package. Fig. 2 shows the optical micrograph of a PSFBG.

 

Fig. 2. Optical micrograph of a PSFBG

Download Full Size | PDF

3. Results

Theoretically it has been shown that the peak shifts either with the amount of phase-shift at the centre location or with the location of the introduced phase [13]. The amount of the phase shift depends on the total fluence of the CO₂ pulses exposure that determines the amount of induced refractive index change through the residual stress relaxation effect.

As the fiber grating is moved to each new position by the translation stage, the CO₂ laser is turned on and the light beam is incident on the fiber grating where stress relaxation is created. In this work, the PSFBGs of different passband number are fabricated sequentially.

3.1 Single passband filter

The CO₂ laser pulses shoot on the central position of FBG. The single passband is observed at a wavelength for which the phase shift introduced by CO₂ laser pulses compensates for the phase mismatch between transmitted waves. The Fig. 3 gives the transmission spectrum of original FBG and the spectrum with the single bandpass peak in the central position of FBG’s stopband. It is seen that the reflectivity of PSFBG has decreased from the one of original FBG.

 

Fig. 3. The transmission spectrums of FBG and PSFBG

Download Full Size | PDF

3.2 Two passbands filter

For the beam size of the CO₂ pulses is about 240μm, the advantage is employed to localize the exposure position on the fiber grating. In this case, two segments, located in 5mm, 7mm of the longitudinal grating position, are shot by equal number of CO₂ laser pulses. The Fig. 4 gives the experiment and simulation results of this PSFBG. There are two bandpass peaks existed in the stopband of FBG with the decreasing reflectivity. Here, a transfer matrix approach is adopted to calculate the transmission spectrum of the PSBG. A simple phase shift between input and output waves can be obtained with transfer matrix $M=\left[\begin{array}{cc}{e}^{-j\phi }& 0\\ 0& e^{j\phi }\end{array}\right]$ . For the grating sections, the elements of the transfer matrix N can be derived by direction solution of the coupled mode equations [14]. The phase shift amount is chosen to π in the simulation, with Λ=538.74μm, L=12mm, and κ=250m^-1.

 

Fig. 4. The transmission spectrum of two bandpass peaks

Download Full Size | PDF

3.3 Three passbands filter

Based on the work of 3.2, the central segment, which located in 6mm of the longitudinal grating position, is also be exposed by the equal pluses number. This time, there are three bandpass peaks appeared in the stopband of FBG. The channel space of these three bandpass peaks is about 0.07nm, and the bandwidth of bandpass peak is about 0.02nm In this case, the reflectivity of the FBG’s stopband has decreased to 7dB. The Fig. 5 shows the corresponding experiment and simulation results.

 

Fig. 5. The transmission spectrum of three bandpass peaks

Download Full Size | PDF

4. Discussion

In the work of 3.1, with the pulse number increasing, the phase change is introduced and the resonant wavelength is shifted from the long wavelength to short wavelength within the stopband of FBG. This phenomenon is illustrated in Fig. 6, which gives the transmission spectrums of PSFBG in two different time slot (the red curve represents the time which is before the black dash one).

The above PSFBG can be erased through high temperature, thus the high temperature associated to CO₂ laser exposure may change the grating refractive index modulation, thus decrease the average refractive index. But the effect of the drawing force during the fiber drawing process, which influences the extent and distribution of the residual stress in the fiber, on refractive index change and subsequent refractive index change on the CO₂ laser irradiation also exists.

For the temperature erasure, the first several pulses may have completely erased the average refractive index change caused by UV written. However, the movement of the resonance, which is from the long to the short wavelength side, still exists with the pulses increasing, indicating a continuous reduction in the refractive index of the material in the processed region. This is consistent with the premise that irradiating by CO₂ laser light is reversing the positive change of index induced by UV light, which moves the resonance from short to long wavelengths. So the results demonstrate that the residual stress relaxation is the main mechanism for the refractive index change in the optical FBGs by the CO₂ laser irradiation.

In the work of 3.2, we even can tune the bandpass profile in the stopband of FBG. The Fig. 7 below shows that the constrast ratio of bandpass can be changed through controlling the CO₂ pulses number. The red curve here is recorded before the black dash one with the exposed pulses increasing.

 

Fig. 6. The phase peak shifts in PSFBG

Download Full Size | PDF

 

Fig. 7. The bandpass profile control of PSFBG.

Download Full Size | PDF

The difference in the experiment and simulation results shown in Fig. 4 and Fig. 5 is the contrast ratio of bandpass peak. The reason maybe lies on that optical phase shift introduced by residual stress relaxation is not exactly π, or the resolution of measurement equipment.

5. Conclusion

In a summary, in this work, we propose a new, yet simple, method for introducing bandpass peaks inside the stopband of a FBG, by controlling CO₂ laser pulses irradiation position and accumulated power. As a result of this new technique, not only can the bandpass peak be made tunable, but multi-bandpass peaks can also be opened simultaneously.