1. Introduction

Active elements, such as Er, Yb, Tm, Bi, Ho, Pr, Cr and so on [1–11], have been doped in glass [2], or silica-based [3–9], or polymer fibers [10,11], or waveguides [12,13], for realizing the optical amplifiers and lasers for optical fiber communication and sensing. The comprehensive knowledge of both the spectral emission and absorption characteristics is very important for evaluating and utilizing these active fibers and waveguides for various applications [1,14,15].

For spectral emission measurement, usually the co- and counter-pumping schemes are employed. The counter-pumping is preferred since that, in the co-pumping scheme, the residual pump power could be significant and produces a background noise mixed with the emission spectra to be tested. However when the emission spectra are broad, e.g. over 200 nm [6] or even wider as those of the Bi/Er co-doped fiber shown in this paper, it is difficult to find a very broadband WDM coupler needed in the counter-pumping scheme. In addition, the selection of the active fiber length is always an issue because the emission spectra could be changed with fiber lengths due to the spectral absorption [1]. So a short section of active fiber is normally used to minimize the effect in order to observe the original emission spectra.

For spectral absorption measurement, the cutback method is usually employed [1,15]. It involves cutting the fiber and its measurement accuracy, determined by the repeatability of power transmitted or coupled before and after the fiber cutting, can be easily affected by the uncertainty in cutting, re-splicing or realigning fibers - especially research-purpose active optical fibers with non-conventional materials and small core geometries.

The direct side pumping scheme we propose here could overcome the abovementioned problems. This scheme is free from fiber cutting, intrinsically low pump background, and simultaneous accurate measurement of emission and absorption. Moreover it can be a simple probe method with appropriate spatial resolution that provides local emission and absorption information in the range of millimeters to centimeters - determined by the pump beam size. To test our scheme, we measured the emission and absorption of a commercial L-band Er-doped fiber against the known results from the supplier. We also successfully demonstrated the scheme’s application over a very broad spectral range of a Bi and Er co-doped fiber we made ourselves.

2. Side pumping scheme

2.1 Experimental setup

The proposed side pumping scheme is shown in Fig. 1 . A test sample can be simply prepared by sandwiching a section of the active optical fiber to be tested (Fiber-Under-Test, FUT) between two half standard SMF connectors. Then the emission and absorption of the FUT can be tested by side pumping a short section of FUT and by connecting one to two Optical Spectrum Analyzers (OSAs). In the experiment, the whole FUT is hold by two fiber holders sat on two movable stages and the FUT is in between, which could realize the parallel shift of FUT easily. The pump laser light is compressed as a line shape and focused perpendicularly on FUT based on a cylinder lens system to realize the side pump. Then the FUT begins to emit the light spectra. So the emission spectra is got because a part of them would be attracted by the fiber and guided to the spectrometer.

 

Fig. 1 (a) The configuration of the emission and absorption measurement of active optical fiber, (b) the coordinate system of the active fiber.

Download Full Size | PDF

2.2 Data process algorithm

The FUT absorption in the emission band can be realized based on the spectra that obtained from two measurements at two different positions $Z_{\text{A}}$and $Z_{\text{B}}$, and their coordinate system is given as Fig. 1(b). The FUT is along Z axis and their left and right connection points are 0 and $Z_{\text{0}}$. The spectra, ${\text{I}}_{\text{1}}(\lambda )$ and ${\text{I}}_{\text{2}}(\lambda )$, are got from the channel 1 and channel 2 when the pump is in the position $Z_{\text{A}}$. They can be expressed as

3. Experiments and discussions

We measure a FUT (EDL001) of 15cm in length based on the methods proposed above. The pump laser is a multimode laser diode of 810nm in central wavelength and 600mW in total power. The laser light is compressed as a line shape, smaller than ~125um wide and ~7mm long, and projected on the FUT (EDL001). Two channels spectrometer with a 5nm resolution, based on two synchronized spectral analyzers (Aglent 86143B), are used to collect the emission spectra. In order to observe the emission spectra with the minimum absorption effect, we pumped the FUT nearby the two connection points. The two spectra, when the pump is projected near the left connection point, are shown in Fig. 2(a) . Their profiles are obviously different because the spectrum at channel 2 experienced the absorption of the ~13cm long Er doped fiber and the other at channel 1 didn’t. In order to observe the emission spectrum without the FUT re-absorption, we pump the fiber at the right connection point and the emission spectra is recorded by the nearby OSA (channel 2), shown in Fig. 2(b).

 

Fig. 2 The side pumping experimental results of Er doped fiber (EDL001). (a) two channel spectra when pumped at Z_A = 2cm, (b) an emission spectrum sample without FUT’s re-absorption and an absorption spectrum based on one channel method. (c) the peak power vibration of the emission spectra when pumped by a multimode laser diode of ~810nm in wavelength, (d) the peak power vibration of the emission spectra when pumped by a single mode laser of ~244nm in wavelength,(e) the calculated peak absorption based on the multimode laser diode and one channel method, (f) the calculated peak absorption based on the single mode laser and one channel method, (g) the calculated absorption spectra based on two channel method when Z_B-Z_A = 6cm, (h) the calculated absorption spectra based on two channel method when Z_B-Z_A = 12cm.

Download Full Size | PDF

3.1 One channel results

Based on the Eq. (5), we can calculate the absorption by only one channel spectra. We measure ten times for each pump position when Z_B-Z_A = 12cm. The spectra vibration could be found and its maximum value is 2dB, shown in Fig. 2(c), which comes from the power and pattern change of our multimode laser source mainly. It could be improved by changing a more stable single mode laser. A frequency double Argon laser of 244nm in wavelength is used here. Its output power is ~150mW and the power vibration is within 2%. The corresponding emission spectra are stable and its vibration is small than 0.8dB, shown in Fig. 2(d). For the case of the multimode laser diode, We use the ten-times-average spectra to calculate the absorption and the results are shown in Fig. 2(g), which approach the data provided by the company, the peak absorption~33dB/m near 1530nm. Here the low pass filter based on FFT is used in order to remove the high frequency noise. The signal and noise ratio is good at the strong emission band and become bad at the weaker emission band. The calculated absorption spectra are vibrated within a range for different measures. Their peak absorption values vary from 28.5dB to 33.2dB based on the multimode laser diode and its standard deviation (Std) is 1.423, shown in Fig. 2(e). Based on the stable 244nm laser, the peak absorption vibration could be reduced to ~2dB (31.1~33.1dB) and the Std is 0.823,shown in Fig. 2(f).

3.2 Two channel results

We also use the two channel method to calculate the FUT absorption. This method is based on the four spectra from two channels, described in Eq. (6). It is free from the spectra vibration in principle. So we use the multimode laser diode of 810nm in wavelength to prove that. The experimental results agree well with the data provided by the supplier, the peak absorption~33dB/m near 1530nm and ~10dB/m near 1480nm, shown in Fig. 2(h). The Std of the absorption spectra in the range from 1460nm to 1600nm are smaller than 1 and have the smallest value ~0.215 at ~1530nm, shown in Fig. 2(k), which is much smaller than the one channel measurements in Fig. 2(e) and Fig. 2(f). The Std has the smallest value at the strongest-emission wavelength and become larger at the weak-emission wavelength. It is reasonable because the emission spectra are used as the measured signals and the stronger signals give the more accurate results. Here, the absorption accuracy is reasonable at the most part of spectral emission range, shown in Fig. 2(h). The difference between the measurements in Fig. 2(g) and Fig. 2(h) is that the side pump spaces, the values of (Z_B-Z_A) in Fig. 1(b), are 6cm and 12cm, respectively. The measurement with a smaller (Z_B-Z_A) value shows a larger Std, shown in Fig. 2(i), which is easy to understand from Eq. (6). The similar results are because of the FUT’s uniformity. They also show that the side pump measurements can be easier to realize the spatial resolution of a small section of FUT without cutting and re-splicing fiber, compared with the traditional cutback method, and could realize the localized measurements.

At last, we show the measurement of our lab-made Bi/Er co-doped, Al activated silica optical fiber sample based on the same setup. Our fiber is fabricated by in situ doping of [Er₂O₃] ~0.01, [Al₂O₃] ~0.15, [Bi₂O₃] ~0.16, [P₂O₅] ~0.94, and [GeO₂] ~12.9 mol %, respectively. The core of our non-standard fiber sample is off axis of about 5um, which is a big trouble for the normal cutback scheme because the coupling with other components is quite different every time. While it is convenient to use the proposed side pumping based scheme. The emission and absorption spectra can be readily obtained, as shown in Fig. 3(a) and Fig. 3(b), respectively. The bandwidth of the emission spectrum is as wide as over 300nm. There are two obvious emission and absorption bands at 1400 nm for ES1→GS of Bi center [7] and 1530 nm for ⁴I_13/2→⁴I_15/2 of Er ions [1]. The stability of the peak absorption at ~1390nm is shown in Fig. 3(c).

 

Fig. 3 (a) The emission spectra of Bi/Er co-doped fiber, (b) the calculated absorption spectra based on two channel method, and (c) the measurement stability of the peak absorption at around 1390nm.

Download Full Size | PDF

4. Discussions

The proposed simple and non-destructive scheme could be used as a method to measure the uniformity of the doping concentration along the active fiber just by measuring the spectra at a few different pumping positions. That is needed and expected to be applied in the active waveguide area as well.

It needs to be noted that the absorption measurement based on the proposed scheme is limited to the emission band. For the very weak emission, the lock-in amplifier technique is expected to apply to obtain the spectra, instead of using OSAs simply. It could be used as a calibration method to calibrate the results of the cutback measurement when the absorption spectrum in the non-emission band is needed.

It needs to be mentioned that the higher pump power is needed for the proposed side pump scheme compared to the case that pump based on a WDM coupler. However, the very cheap, lower than 100 dollars for our case, multi transverse and longitudinal mode laser diodes can be found easily, which is seldom used in SMF system because of the coupling difficulty.

It needs to be noted as well that the resonance pump should be applied by using the appropriate laser to realize the fully inversion if the emission cross section of active fiber is needed. Based on 810nm laser diode and our side-pumping scheme, we did not realize the fully inversion for FUT (EDL001) and we observed the Bi emission saturation of our home-made active fiber.

5. Conclusion

The test of spectral emission and absorption of active optical fiber are realized based on a simple and non-deconstructive side-pumping technique and its feasibility is proved by the measurement results of a commercialized Er doped optical fiber. The results of Our Non-perfect Bi/Er co-doped optical fiber, which has a 5um off-axis in the fiber core and an over 300nm width emission band, show its robustness. It is expected to become a standard measurement scheme for variable types of active optical fibers.