1. Introduction

Significant scientific effort has been devoted to photodarkening [1] (PD) mitigation or complete bleaching in silica optical fibers. The main applied methods have been: non - isothermal bleaching [2,3], codoping with Al, P [4] or Ce [5], mitigation by using peculiar fiber design [6] or different active ion inversions [7], O₂ [8] or H₂ loading [9] and photobleaching (PB) with UV - VIS irradiation [10–12]. All approaches cited above have been utilized in order to examine the bleaching as well as the PD mechanism since their final goal is to offer a photodarkening free fiber.

Several articles on PB have been published. The evidence of partial PB with 543 nm irradiation and 1 mW pump power given in [10] showed that the resonant 977 nm Yb³⁺ absorption peak increases accordingly. Supported by the fact that Yb²⁺ exist in aluminosilicate glasses, the author suggested Yb²⁺ → Yb³⁺ as the main PB mechanism.

The total bleaching of the Yb-doped, large mode area (LMA) optical fiber by 355 nm and 450 mW light irradiation at 5 kHz repetition rate was also reported [11]. Several PD ↔ PB consecutive processes showed no difference between pristine and treated fibers which indicated that no residual PD loss was present.

The PB by 407 nm laser diode with cladding – irradiation of 120 W/cm² intensity in Yb doped aluminosilicate fibers was demonstrated [12] and defined as the one photon process. On one hand, a semi – empirical model for PD and PB description was developed and confirmed that BL in Yb doped LMA fibers is one photon process [13]. On the other hand, the bleaching of Tb doped fibers is assumed as the 2 photon process [4] where residual PD loss in the consecutive PD ↔ PB process should be taken into account.

Since the photodarkening phenomenon can be described as the transformation of atomic precursors to color centers (CC), it is important to investigate PD the other way around, through the annihilation of CC i.e. from the viewpoint of PB. In the present study, PB of the aluminosilicate, Yb³⁺ doped fibers with different dopant concentrations under various 633 nm irradiation intensities was extensively examined. This paper can be divided into three different parts due to the characteristics of the initial PD process and dopant concentration. In the first part, bleaching of the optical fibers with a different CC number and dopant concentrations was compared. In the second part, dopant concentration was fixed while the CC number was changed and finally, in the third part dopant concentration was different while the number of CC was approximately the same.

2. Experimental

The set up employed for measuring PD loss and bleached loss amount was analogue to the one described in [14]. Single mode, aluminosilicate fibers with 1.1, 1.35, 1.8 wt% Yb³⁺ dopant concentrations were fabricated by means of the modified chemical vapor deposition (MCVD) technique. The fiber diameter (6.6, 6.6, 7.8 µm) and Al³⁺ content (3.2, 3.2, 1.8 wt %) were varying in the sequence as the dopant concentration was increasing. Fiber samples were from 3 to 6 cm long in order to provide uniform PD loss over the length and ensure an adequate PB process.

During the first step, each sample was irradiated with 976 nm (1.27 eV) light (power ~200 mW) and characteristic photodarkening curves and values were obtained in 24h period [14,15]. The second step considered bleaching with an HeNe laser at 633 nm (1.96 eV) emission wavelength and powers in 0.07 – 3.9 mW range. Estimated bleaching power determination error was 5%.

3. Bleaching of photodarkening for various 633 nm light powers

In the first step, we irradiated pristine testing fibers with 633 nm light and no transmission degradation on a long time scale was observed. In the first set of experiments, the PD losses were induced by 976 nm photons. Irradiation with different powers of 633 nm light was carried out afterwards in order to measure the bleached loss amplitude. An example of bleaching curves obtained for 1.1 wt% doped fiber is shown in Fig. 1 .

 

Fig. 1 Bleaching of 1.1 wt% Yb³⁺ doped fiber for different HeNe laser (633 nm) powers. Initial PD excess loss was 170 ± 5dB/m.

Download Full Size | PDF

Experimental curves were fitted on the stretched exponential equation [7] and corresponding parameters were shown. Bleached loss was presented as the absolute value and of course, that meant the decrease of photodarkening loss for a certain percentage as shown in Fig. 2 . We noticed that the PB equilibrium state was increasing as the HeNe laser power was increasing. It was also observed that stretching parameter β had the range from 0.5 to about 0.8 and was not strongly correlated with the bleaching intensity and the ytterbium concentration.

 

Fig. 2 Percentage of the photodarkening bleached loss as the function of laser power at 633 nm.

Download Full Size | PDF

In Fig. 2, the percentage of the bleached photodarkening as the function of bleaching power for all three tested fibers was shown. The percentage was calculated as the ratio of the bleached loss amplitude and the equilibrium PD loss.

Figure 2 showed that a) the increase of bleaching power led to a rather quick increase of bleached loss percentage and the very small change for the subsequent power increase which seems to suggest a saturation effect, b) for the maximum bleaching power of 3.9 mW (~100 W/mm²) at 633 nm, unbleachable loss was demonstrated, c) the minimum HeNe power of 0.07 mW showed 8% of bleaching which suggest that µW power level should not drastically influence PD measurements, d) the bleached loss percentage showed similar behavior regardless the dopant concentration.

The value of 1.8 wt% at 3.9 mW was somewhat lower than expected and should be further investigated. However, such discrepancy was also observed in the case of 407 nm irradiation [12].

4. Bleaching of initially different photodarkening loss at fixed Yb³⁺ concentration and vice versa

To investigate the role of doping level on the bleaching mechanism, two experiments at fixed bleaching power of 0.8 mW were performed. In the first experiment, the PD process of one selected fiber (1.8 wt% Yb³⁺) was stopped at different moments of its evolution i.e. before reaching the equilibrium value as labeled in Fig. 3 . In this way, we kept the same dopant concentration while changing the number of generated color centers which were bleached afterwards. In the second experiment, the bleaching process in two fibers with different dopant concentrations started from approximately the same PD loss amplitude value (Fig. 4 ). In a contrast with the previous case, here we have a similar number of CC defects, but in the fibers with different dopant concentrations.

 

Fig. 3 (a) Bleaching with 0.8 mW light at 633 nm on the same fiber doped with 1.8 wt% Yb³⁺. Each curve presents bleaching which started from different PD loss as it is assigned in the label. (b) The bleaching amplitude as the function of initial PD loss amplitude.

Download Full Size | PDF

 

Fig. 4 (a) Bleaching of 1.8 wt% and 1.1 wt% Yb³⁺ fibers with 0.8 mW of 633 nm light irradiation. (b) Bleaching of 1.8 wt % and 1.35 wt% Yb³⁺ fibers with 0.8 mW light. The initial PD loss level was similar.

Download Full Size | PDF

With reference to Figs. 3 and 4, the following conclusions can be made. Figure 3 demonstrated that 0.8 mW could bleach only ~80% of total PD loss in the range from 170 - 507 dB/m. In the case of the maximum photodarkening level of 652 dB/m the percentage was lower. The final, maximum value of bleached loss increased linearly as the initial PD loss amplitude increased, up to certain value of saturation (in our case 507 dB/m PD) which was probably determined by the BL light power (Figs. 3(a) and 3(b)). Hence, fixed bleaching power of 0.8 mW was not able to bleach more CC after a certain number even if their number would be increased in the previous PD process. This may suggest that during the last period of PD process (saturation part of the PD curve [14]) the most of the created defects required photons with higher energy in order to be bleached. On the contrary, in the beginning about 80% of defects can be bleached with 633 nm photons. We therefore suggest that the type of generated defects is different during PD process. This conclusion extracted from PB adds support to a coherent picture provided by thermal bleaching [2,16].

In Figs. 4 (a) and 4(b), the bleaching curves of fibers with different dopant concentrations, but approximately the same initial PD loss value showed different bleaching dynamics and a final bleaching amplitude. It again implies that if PD process is stopped at its early stage about 80% of defects can be bleached while the percentage decrease if PD reaches the final, near equilibrium stage. Note that the 1.1 wt% Yb³⁺ doped fiber (Fig. 4(a)) was bleached after complete PD process while in the case of 1.8 wt% Yb³⁺ fiber, PD process was stopped at ~30% of the final PD loss value. These assumptions and the indication of the PB behavior should be taken with caution since different Al³⁺ content could influence the results [17].

The experiments showed residual unbleachable losses in aluminosilicate single mode fibers when irradiated with 633 nm light with power up to 3.9 mW. This fact leads to following assumptions regarding CC nature: a) Yb³⁺ clustering prevented certain defects being bleached, b) certain defects could not be bleached with a red photon irradiation due to different mechanism of CC creation which provided different qualities and CC energy level. Since the similar bleaching behavior is observed for different dopant concentrations and therefore different cluster number, the first option is not likely. These defects can be bleached by higher energy photons [11] as stated in the introduction part while their nature is still unsolved probably because of overlapping of several different mechanisms.

It was suggested [18] that the band from 400 – 700 nm is associated with Al – oxygen hole centers (Al-OHC) as the result of UV (196 nm) or IR irradiation. Since tested fibers were codoped with Al³⁺, we assumed that PB with 633 nm photons was related with dopant ions and the pointed band. According to that, it is supposed that 976 nm IR photons created CC related with Al-OHCs which were afterwards partially bleached with 633 nm light. The adequate energy of 633 nm light should rearranged bonds among Yb ions and its neighboring ligands enabling the bleaching process. The loss reduction may be related with local charge rearrangement described as Yb²⁺ → Yb³⁺ and proved by the Al-OHCs band amplitude change in the future experiments.

5. Conclusion

The importance of photobleaching process examination lays in understanding of not only photobleaching, but also of photodarkening process and color - center formation mechanism. The bleaching process for various dopant concentrations, photobleaching powers or initial photodarkening loss was compared in order to describe the mechanism and process characteristics.

The research brought several novelties. It was shown that the percentage of bleached photodarkening loss is similar for different concentrations at given bleaching power in the fibers where photodarkening loss was induced to its maximum. Furthermore, photobleaching was related with the number of photodarkening induced color centers while the dopant concentration was fixed. Photobleaching was dopant concentration dependent while the number of color centers was approximately the same. Unbleachable loss was observed for the 633 nm light at 0.07 – 3.9 mW power range propagating inside the core of aluminosilicate fibers doped with Yb³⁺ ions from 1.1 – 1.8 wt%.

In addition, different type of defects generated at distinct stages of the PD process evolution may be suggested.