AMPLIFYING FIBER AND OPTICAL AMPLIFIER

Abstract

An amplification fiber includes a core which is doped with an erbium ion and a cladding which surrounds the core and has a refractive index lower than a refractive index of the core, and a relative refractive index difference Δn.sub.51 between the core and the cladding is not more than a smaller one of values of a relative refractive index difference Δn.sub.1 expressed as a predetermined expression related to a radius a of the core and a relative refractive index difference Δn.sub.2 expressed as a predetermined expression related to the radius a of the core.

Claims

1. An amplification fiber comprising: a core which is doped with an erbium ion; and a cladding which surrounds the core and has a refractive index lower than a refractive index of the core, wherein a relative refractive index difference between the core and the cladding is not more than a smaller one of values of a first relative refractive index difference expressed as the following Expression (1) and a second relative refractive index difference expressed as the following Expression (2): [Math. 1]
Δn.sub.1=75.64−280.33a+443.18a.sup.2372.88a.sup.3+175.04a.sup.4−43.384a.sup.5+4.44a.sup.6   (1) [Math. 2]
Δn.sub.2=13.34a.sup.−1.98   (2) wherein, in (1) and (2) described above, a denotes a radius [μm] of the core, Δn.sub.1 denotes the first relative refractive index difference [%], and Δn.sub.2 denotes the second relative refractive index difference [%].

2. The amplification fiber according to claim 1, wherein a plurality of the cores are provided to be spaced apart from each other, each of relative refractive index differences between the plurality of the cores and the cladding is not more than the smaller one of the values of the first relative refractive index difference and the second relative refractive index difference, and the radius of the core satisfies Expression (5): [Math. 3]
a≥2.3   (5)

3. The amplification fiber according to claim 1, wherein the relative refractive index difference between the core and the cladding is not more than 2.8% in an area of the radius of the core in which the smaller one of the values of the first relative refractive index difference and the second relative refractive index difference is more than 2.8%.

4. An amplification fiber comprising: a core which is doped with an erbium ion; and a cladding which surrounds the core and has a refractive index lower than a refractive index of the core, wherein a relative refractive index difference between the core and the cladding is not more than a smaller one of values of a third relative refractive index difference expressed as the following Expression (3) and a fourth relative refractive index difference expressed as the following Expression (4): [Math. 4]
Δn.sub.3=45.12−134.20a+160.96a.sup.2−87.78a.sup.3+18.30a.sup.4   (3) [Math. 5]
Δn.sub.4=13.34a.sup.−1.98   (4) wherein, in (3) and (4) described above, a denotes a radius [μm] of the core, Δn.sub.3 denotes the third relative refractive index difference [%], and Δn.sub.4 denotes the fourth relative refractive index difference [%].

5. The amplification fiber according to claim 4, wherein a plurality of the cores are provided to be spaced apart from each other, each of relative refractive index differences between the plurality of the cores and the cladding is not more than the smaller one of the values of the third relative refractive index difference and the fourth relative refractive index difference, and the radius of the core satisfies Expression (5): [Math. 6]
a≥2.3   (5)

6. The amplification fiber according to claim 4, wherein the relative refractive index difference between the core and the cladding is not more than 2.8% in an area of the radius of the core in which the smaller one of the values of the third relative refractive index difference and the fourth relative refractive index difference is more than 2.8%.

7. The amplification fiber according to claim 2, wherein the cladding includes: a first cladding which surrounds the plurality of the cores and has a refractive index lower than the refractive index of the core; and a second cladding which surrounds the first cladding and has a refractive index lower than the refractive index of the first cladding.

8. The amplification fiber according to claim 1, which is wound around a bobbin and is bonded to the bobbin.

9. An optical amplifier comprising: an amplification fiber comprising: a core which is doped with an erbium ion; and a cladding which surrounds the core and has a refractive index lower than a refractive index of the core, wherein a relative refractive index difference between the core and the cladding is not more than a smaller one of values of a first relative refractive index difference expressed as the following Expression (1) and a second relative refractive index difference expressed as the following Expression (2): [Math. 1]
Δn.sub.1=75.64−280.33a+443.18a.sup.2372.88a.sup.3+175.04a.sup.443.384a.sup.5+4.44a.sup.6   (1) [Math. 2]
Δn.sub.2=13.34a.sup.−1.98   (2) wherein, in (1) and (2) described above, a denotes a radius [μm] of the core, Δn.sub.1 denotes the first relative refractive index difference [%], and Am denotes the second relative refractive index difference [%]; a semiconductor laser light source which emits excitation light input to the amplification fiber; and a feedback circuit which generates current supplied to the semiconductor laser light source from gain detected according to powers of signal light input to the amplification fiber and signal light output from the amplification fiber.

10. The optical amplifier according to claim 9, wherein a plurality of the cores are provided to be spaced apart from each other, each of relative refractive index differences between the plurality of the cores and the cladding is not more than the smaller one of the values of the first relative refractive index difference and the second relative refractive index difference, and the radius of the core satisfies Expression (5): [Math. 3]
a≥2.3   (5)

11. The optical amplifier according to claim 9, wherein the relative refractive index difference between the core and the cladding is not more than 2.8% in an area of the radius of the core in which the smaller one of the values of the first relative refractive index difference and the second relative refractive index difference is more than 2.8%.

Description

BRIEF DESCRIPTION OF DRAWINGS

[0033] FIG. 1 is a cross-sectional view of an amplification fiber of a first embodiment of the present invention.

[0034] FIG. 2 is a graph showing a relationship between a radius of a core and a relative refractive index difference for indicating an area satisfied by the core and the relative refractive index difference of the amplification fiber shown in FIG. 1.

[0035] FIG. 3 is another graph showing the relationship between the radius of the core and the relative refractive index difference for indicating the area satisfied by the core and the relative refractive index difference of the amplification fiber shown in FIG. 1.

[0036] FIG. 4 is a schematic view of ac optical amplifier of the first embodiment of the present invention.

[0037] FIG. 5 is a graph showing gain change with respect to elapsed time in the optical amplifier shown in FIG. 4.

[0038] FIG. 6 is a cross-sectional view of an amplification fiber of a second embodiment of the present invention.

[0039] FIG. 7 is a cross-sectional view showing a modification of the amplification fiber shown in FIG. 6.

[0040] FIG. 8 is a schematic view of an optical amplifier including the amplification fiber shown in FIG. 6 or 7.

[0041] FIG. 9 is a graph showing a relationship between. a radius of a core and a power consumption ratio of the optical amplifier shown in FIG. 8 conventional EDFA.

[0042] FIG. 10 is a graph showing a relationship between a radius of a core and a relative refractive index difference for indicating an area satisfied by the core and the relative refractive index difference of the amplification fiber shown in FIG. 6 or 7.

[0043] FIG. 11 is another graph showing the relationship between the radius of the core and the relative refractive index difference for indicating the area satisfied by the core and the relative refractive index difference of the amplification fiber shown in FIG. 6 or 7.

[0044] FIG. 12 is a graph showing gain change with respect to elapsed time in the optical amplifier shown in FIG. 8.

[0045] FIG. 13 is a graph showing gain change with respect to elapsed time in a conventional optical amplifier.

DESCRIPTION OF EMBODIMENTS

[0046] Hereinbelow, embodiments of an amplification fiber and an optical amplifier of the present invention will be described with reference to the drawings. Note that, in the present description and the drawings, configurations having the same function are designated by the same reference numeral, and the repeated description thereof will be omitted.

First Embodiment

[0047] An amplification fiber 51 of a first embodiment of the present invention is used as, e.g., an amplification medium of an optical fiber amplifier. As shown in FIG. 1, the amplification fiber 51 includes one core 61 and a cladding 65 which surrounds the core 61, and is what is called a single-core erbium-doped fiber. The core 61 is doped with an erbium ion.

[0048] A relative refractive index difference Δn.sub.51 between the core 61 and the cladding 65 is limited to a shaded area shown in FIG. 2 or 3. A graph in FIG. 2 shows a relationship between the radius of the core and the relative refractive index difference Δn.sub.51 [%] which are numerically calculated on the assumption of The case where the amplification fiber 51 is applied to amplification of C-band signal light. A graph in FIG. 3 shows the relationship between the radius of the core and the relative refractive index difference which are numerically calculated on the assumption of the case where the amplification fiber 51 is applied to amplification of L-band signal light. The above-mentioned C band denotes a frequency band of not less than 1530 nm and not more than 1565 nm, and the L band denotes a frequency band of not less than 1565 nm and not more than 1625 nm. Note that, when a radius a is less than 1 μm, a fiber length required for the amplification of the erbium-doped fiber becomes extremely long. From this, the erbium-doped fiber having the radius a of less than 1 μm is not used substantially, and hence the case where the radius a is not more than 1 μm is not shown in the graph in each of FIGS. 2 and 3.

[0049] In order to suppress change of gain to less than 0.2 dB with respect to change of an input signal light power in the amplification fiber 51, a gain change speed immediately after the change of the input signal light power needs to be less than 0.2 dB/μs. This is because the band of each of a feedback control circuit and an LD drive circuit is about 1 MHz. The gain change speed differs depending on the radius a [μm] of the core 61 and the relative refractive index difference Δn.sub.51. When the amplification fiber 51 is applied to the amplification of the C-band signal light, as shown in FIG. 2, as the radius a increases from 1 μm, a relative refractive index difference (first relative refractive index difference) Δn.sub.1 at which the gain change speed is 0.2 dB/μs is an area smaller than a boundary line 1-1 expressed as the following Expression (1):

[Math. 7]

[0050]
Δn.sub.1=75.64−280.33a+443.18a.sup.2−372.88a.sup.3+175.04a.sup.4−43.38a.sup.5+4.44a.sup.6   (1).

[0051] Similarly, when the amplification fiber 51 is applied to the amplification of the L-band signal light, as shown in FIG. 3, as the radius a increases from 1 μm, a boundary line 1-2 of a relative refractive index difference (third relative refractive index difference) Δn.sub.3 at which the gain change speed is 0.2 dB/μs is an area smaller than the boundary line 1-1 expressed as the following Expression (3):

[Math. 8]

[0052]
Δn.sub.3=45.12−134.20a+160.96a.sup.2−87.78a.sup.3+18.30a.sup.4   (3).

[0053] In an area in which the relative refractive index difference Δn.sub.51 is smaller than the relative refractive index difference Δn.sub.1 in the C band, and in an area in which the relative refractive index difference Δn.sub.51 is smaller than a relative refractive index difference Δn.sub.2 described later in the L band, the gain change speed is about 0.2 dB/μs. In the case where the relative refractive index difference Δn.sub.51 is fixed and the radius a is increased, when the radius a becomes larger than a predetermined. radius, the permissible mode of the core 61 becomes a higher-order mode, and the amplification. efficiency of the amplification of The signal light in a single mode is reduced due to mode coupling. The predetermined radius mentioned above is changed according to the relative refractive index difference Δn.sub.51. However, in fusion splicing between the amplification fiber 51 and a pigtail fiber (not shown) constituting the optical amplifier, by performing The fusion splicing such that the center of the core 61 matches the center of the, core of the pigtail fiber, excitation of higher-order mode light is suppressed at an input end of The amplification fiber 51.

[0054] The amplification fiber 51 is wound around a bobbin into a coil shape, and is bonded and fixed to the bobbin with an adhesive or the like. With this, mechanical agitation such as vibration of the amplification fiber 51 caused by, e.g., ventilation is averted. Thus, by preventing the signal light in the single mode from being converted into the signal light in the higher-order mode during propagation in the amplification fiber 51, it is possible to handle the amplification fiber 51 as a single mode fiber (SMF) effectively when a cutoff wavelength is not more than 2 μm.

[0055] When the amplification fiber 51 is applied to the amplification of the C-band signal light in an area in which the radius a is larger than the predetermined radius described above, as shown in FIG. 2, a boundary line 2-1 of the relative refractive index difference (second relative refractive index difference) Δn.sub.2 at which the cutoff wavelength of the amplification fiber 51 is 2 μm is expressed as the following Expression (2):

[Math. 9]

[0056]
Δn.sub.2=13.34a.sup.−1.98   (2)

[0057] Similarly, when the amplification fiber 51 is applied to the amplification of the L-band signal light in the area in which the radius a is larger than the predetermined radius described above, as shown in FIG. 3, a boundary line 2-2 of a relative refractive index difference (fourth relative refractive index difference) Δn.sub.4 at which the cutoff wavelength of the amplification fiber 51 is 2 μm is expressed as the following Expression (4):

[Math. 10]

[0058]
Δn.sub.4=13.34a.sup.−1.98   (4)

[0059] As can be seen from FIG. 2, when the amplification fiber 51 is applied to the amplification of the C-band signal light, the relative refractive index difference Δn.sub.51 is not more than the smaller one of the values of the relative refractive index difference Δn.sub.1 and the relative refractive index difference ≢n.sub.2. Similarly, as can be seen from FIG. 3, when the amplification fiber 51 is applied to the amplification of the L-band signal light, the relative refractive index difference Δn.sub.51 is not more than the smaller one of the values of the relative refractive index difference Δn.sub.3 and the relative refractive index difference Δn.sub.4. Note that, in the case where the values of the relative refractive index difference Δn.sub.2 and the relative refractive index. difference Δn.sub.2 are equal to each other, and in the case where the values of the relative refractive index difference Δn.sub.3 and the relative refractive index. difference Δn.sub.4 are equal to each other, the above-described “smaller one of the values” denotes the values which are equal to each other.

[0060] According to the amplification fiber 51 described. above, the relative refractive index difference Δn.sub.51 is limited to the shaded area shown in each of FIGS. 2 and 3, and a correlation between the radius a and the relative refractive index difference Δn.sub.51 is thereby limited to an area in which a gain change value satisfies a predetermined reference value at the time of the change of the input signal light power. Herein, the predetermined reference value means a reference value which does not degrade transmission characteristics of an optical network, and is, e.g., less than 0.2 dB. The correlation between the radius a and the relative refractive index difference Δn.sub.51 is limited as described above, whereby, the gain change can be suppressed to less than 0.2 dB when gain control based on feedback control is performed to cope with the change of the input signal light power caused by the change of the number of wavelengths of WDM signal light.

[0061] Note that, in FIG. 3, the boundary line 1-2 and the boundary line 2-2 intersect each other in an area of Δn>2.8%. However, even when quartz glass, or tellurite glass or bismuth oxide glass which has a refractive index higher than that of quartz glass is used, it is difficult to implement Δn.sub.51 >2.8% with a glass fiber. From this, in light of practicality of the amplification fiber 51, in the area of the radius a in which the boundary line 1-2 and the boundary line 2-2 intersect each other with Δn>2.8%, the relative refractive index difference Δn.sub.51 is not more than 2.8%. In FIGS. 2 and 3, conditions for Δn=2.8% to be satisfied are shown as boundary lines 3-1 and 3-2 of a relative refractive index Δn.sub.5.

[0062] An optical amplifier 101 shown in FIG. 4 includes, in addition to the amplification fiber 51 described above, LD light sources (semiconductor laser light sources) 12-1 and 12-2, multiplexer/demultiplexers 13-1 and 13-2, optical branching devices 14-1 and 14-2, photodetectors 15-1 and 15-2, a gain equalizer 16, and a feedback circuit 17. Each of the LD light sources 12-1 and 12-2 emits excitation light input to the amplification fiber 51. Each of the multiplexer/demultiplexers 13-1 and 13-2 includes an optical isolator which is riot shown, and multiplexes or demultiplexes signal light and. excitation light. Each of the optical branching devices 14-1 and 14-2 causes part of input light to branch. Each of the photodetectors 15-1 and. 15-2 converts the light caused to branch by each of the optical branching devices 14-1 and 14-2 to current or voltage according to the power of the light. The gain equalizer 16 equalizes wavelength dependence of gain of the amplification fiber 51. The feedback circuit 17 calculates the gain by using detection results in the photodetectors 15-1 and 15-2, and generates adjustment current (current) supplied. to the LD light sources 12-1 and 12-2 from the calculation result.

[0063] The feedback circuit 17 adjusts drive current to each of the LD light sources 12-1 and 12-2 such that the gain calculated from detected values of the photodetectors 15-1 and 15-2 is constant when the input signal light power is changed, and the gain control in the optical amplifier 101 is thereby performed.

[0064] A solid line in FIG. 5 indicates the state of the gain control of one wavelength which has remained when the number of wavelengths of the signal light has been changed from 40 to 1 in the optical amplifier 101 based on the assumption that the radius a is 2 μm, the relative refractive index difference Δn.sub.51 is 0.8%, and the signal light is in the C band. A broken line in FIG. 5 indicates, as a reference, the state of the gain control of the optical amplifier 101 which uses an erbium-doped fiber having the radius a of 2 μm and the relative refractive index difference Δn of 1.5% instead of the amplification fiber 51, and has other conditions similar to those of the state indicated by the solid line. The condition that the radius a is 2 μm and the relative refractive index difference Δn.sub.51 is 0.8% is in the shaded area in FIG. 2, but the condition that the radius a is 2 μm and the relative refractive index difference Ln.sub.51 is 1.5% is outside the shaded area in FIG. 2. As indicated by the solid line in FIG. 5, in the case where the amplification fiber 51 having the radius a of 2 μm and the relative refractive index difference Δn.sub.51 0.8% is used, the maximum value of the gain change is lower than 0.2 dB. On the other hand, in the case where the erbium-doped fiber having the radius a of 2 μm and the relative refractive index difference Δn of 1.5% is used, it can be seen that the gain change reaches up to about 0.35 dB. From this result, according to the amplification fiber 51, it is possible to determine that, when the gain control based on the feedback control is performed to cope with the change of the input signal light power caused by the change of the number of wavelengths of the WDM: signal light, the gain. change can be suppressed to less than 0.2 dB.

[0065] In addition, in the configuration in which the signal light was in the C band in the optical amplifier 101, the excitation light power required for a gain of 25 dB in the case where the amplification fiber 51 which satisfied the condition. in the shaded area in FIG. 2 and had the radius a of 3.5 μm and the relative refractive index difference Δn.sub.51 of 0.8% was used was compared with the excitation light power required for a gain of 25 dB in the case where the erbium-doped fiber which had the radius a of 3.5 μm. and the relative refractive index difference Δn of 1.5% was used. As a result, in the amplification fiber 51 having the radius a of 3.5 μm and the relative refractive index difference Δn.sub.51 of 0.8%, the excitation light power which was nine times the excitation light power of the erbium-doped. fiber having the radius a of 3.5 μm and the relative refractive index difference Δn of 1.5% was required, and the cutoff wavelength exceeded 2 μm.

[0066] Further, in the case where the amplification fiber 51 having the radius a of 3.5 μm and the relative refractive index difference Δn.sub.51 of 0.8% was simply wound around a bobbin, the gain temporally fluctuated from a normal value by about (+ −) 0.8 dB under an environment of ventilation of 1 m/s. The above (+ −) represents plus and minus signs. In contrast to this, in the case where the above-described amplification fiber 51 was wound around the bobbin and was bonded and fixed to the bobbin, it was determined that the gain did not change from the normal value.

Second Embodiment

[0067] As shown. in FIG. 6, an amplification fiber 52 of a second embodiment of the present invention includes seven cores 61 which are disposed at regular intervals, and a cladding 65 which surrounds the cores 61. The cladding 65 includes a first cladding (cladding) 62 which surrounds the seven cores 61 without any gap, and a second cladding 63 which surrounds the first cladding 62.

[0068] The refractive index of the first cladding 62 is lower than the refractive index of the core 61. The first cladding 62 propagates the excitation light, and is able to collectively excite erbium ions with which all of the cores 61 are doped with the excitation light. The refractive index of the second cladding 63 is lower than the refractive index of the first cladding 62. The amplification fiber 52 is a multi-core erbium-doped fiber having a double cladding structure. Each of relative refractive index differences Δn.sub.52 and Δn.sub.53 of the amplification fibers 52 and 53 denotes the relative refractive index difference between the core 61 and the first cladding 62.

[0069] The amplification fiber 53 shown in FIG. 7 is a modification of the amplification fiber 52, and includes nineteen cores 61 which are disposed at regular intervals, and a cladding 65 which surrounds the cores 61.

[0070] An optical amplifier 102 shown in FIG. 8 includes, in addition to the amplification fiber 52 or the amplification fiber 53, LD light sources 12-1 and 12-2, multiplexer/demultiplexers 13-1 and 13-2, optical branching devices 14-1 and 14-2, photodetectors 15-1 and 15-2, a gain equalizer 16, and a feedback circuit 17. Note that the feedback circuit 17 calculates the gain by using detection results in the photodetectors 15-1 and 15-2, and generates only the adjustment current supplied to the LD light source 12-2 from the calculation result, as will be described later.

[0071] The optical amplifier 108 further includes a fan-in component 18-1 and a fan-out component 18-2. The fan-in component 18-1 is provided between the multiplexer-demultiplexer 13-1 and the optical branching device 14-1, and performs conversion from a single-core fiber to a multi-core fiber. The fan-out component 18-2 is disposed between. an end portion of the amplification fiber 52 or the amplification fiber 53 on a side opposite to the multiplexer/demultiplexer 13-1 and the multiplexer/demultiplexer 13-2, and performs conversion from the multi-core fiber to the single-core fiber. The LD light source 12-1 is a multi-mode LD. The excitation light of the LID light source 12-1 is coupled. to the first cladding 62 of the amplification fiber 52 or the amplification fiber 53 via the multiplexer/demultiplexer 13-1, and is used as cladding pumping light. On the other hand, the LD light source 12-2 is a single-mode LD. The excitation light of the LID light source 12-2 is multiplexed with the signal light by the multiplexer/demultiplexer 13-2, is input to the core 61 of the amplification fiber 52 or the amplification fiber 53, and is used as core excitation light. The feedback circuit 17 adjusts only the drive current to the LD light source 12-2 such that the gain calculated from the detected values of the photodetectors 15-1 and 15-2 is constant when the input signal light power is changed, and the gain control is thereby performed. At this point, the drive current of the LD light source 12-1 is kept constant.

[0072] The multi-core EDFA which uses cladding pumping requires only one light source for the cladding pumping, and hence a low power consumption effect of being able to reduce the power consumption of the optical amplifier is achieved. However, when the radius of the core of the erbium-doped fiber is smaller than a predetermined radius, the low power consumption effect cannot be achieved. This is because, when the radius of the core is small, the erbium ion with which the core is doped cannot absorb the excitation light propagating in the first cladding adequately, the excitation light power which is not used in amplification is increased, and the power consumption of the optical ampler is thereby increased. The radius of the core of a boundary for determining the presence or absence of the low power consumption effect is 2.3 μm. FIG. 9 shows evidence that the boundary of the presence or absence of the low power consumption effect is the radius a=2.3 μm.

[0073] As shown in FIG. 9, a power consumption ratio to the conventional EDFA is plotted with respect to the core radius. The power consumption ratio to the conventional EDFA is calculated by dividing the power consumption of the amplification fiber 52 or the amplification fiber 53 serving as the multi-core EDFA which uses the cladding pumping by a value obtained by the power consumption of the single-core EDFA having equivalent amplification characteristics×the number of cores of the multi-core EDFA. A solid line in 9 indicates the result of calculation of the power consumption of the amplification fiber 52. A broken line in FIG. 9 indicates the result of calculation of the power consumption of the amplification fiber 53. The sold line and the broken line in FIG. 9 substantially coincide with each other. The power consumption ratio of each of the amplification fibers 52 and 53 to the conventional EDFA is less than 1 when the radius a is not less than 2.3 μm. When the radius a is not less than 2.3 μm, the amplification fiber 52 or the amplification fiber 53 achieves the low power consumption effect.

[0074] Consequently, in the amplification fiber 52 or the amplification fiber 53, the limited area of the radius a and the relative refractive index difference Δn.sub.52 or Δn.sub.53 in which the gain change speed of 0.2 dB/μs is obtained corresponds to an area obtained by limiting the shaded area in each of FIGS. 2 and 3 with a limitation in which the radius a is not less than 2.3 μm, as shown in FIGS. 10 and 11. That is, FIG. 10 shows the limited area of the radius a and the relative refractive index difference Δn.sub.52 or Δn.sub.53 in the case where the amplification fiber 52 or the amplification fiber 53 is applied to the amplification of the C-band signal light. FIG. 11 shows the limited area of the radius a and the relative refractive index difference Δn.sub.52 or Δn.sub.53 in the case where the amplification fiber 52 or the amplification fiber 53 is applied to the amplification of the L-band signal light. As shown in FIG. 11, in the case where the amplification fiber 52 or the amplification fiber 53 is applied to the amplification of the L-band signal light, the boundary line 1-2 is in an area in which the radius a is less than 2.3 μm, and does not appear in the shaded limited area.

[0075] A solid line in FIG. 12 indicates the state of the Gain control of one wavelength which has remained when the number of wavelengths of the signal light has changed from 40 to 1 in the case where the signal light is in the C band in the optical amplifier 102 which uses the amplification fiber 52 or 53 having the radius a of 2.4 μm and the relative refractive index difference Δn.sub.52 or Δn.sub.53 of 1.7%. A broken line in FIG. 12 indicates, as a reference, the state of the gain control in the case where the erbium-doped fiber having the radius a of 2.4 μm and the relative refractive index difference Δn of 2.0 is used instead of the amplification fiber 52 or 53. The condition that the radius a is 2.4 μm and the relative refractive index difference Δn.sub.52 or Δn.sub.53 is 1.7% is in a shaded area shown in FIG. 10. However, the condition that the radius a is 2.4 μm and the relative refractive index difference Δn.sub.52 or Δn.sub.53 is 2.0% is outside the shaded area shown in FIG. 10. As indicated by the solid line in FIG. 12, in the case where the amplification fiber 52 or 53 is used, the maximum value of the gain change is lower than 0.2 dB. On the other hand, as indicated by the broken line in FIG. 12, in the case where the erbium-doped fiber having the radius a of 2.4 μm and the relative refractive index difference Δn.sub.52 or Δn.sub.53 of 2.0% is used, the maximum value of the gain change is about 0.3 dB and is higher than 0.2 dB. From this result, according to the amplification fibers 52 and 53, it can be seen that, when the gain control based on the feedback control is performed to cope with the change of the input signal light power caused by the change of the number of wavelengths of the WDM signal light, the gain change can be suppressed to less than 0.2 dB.

[0076] In the configuration of the optical amplifier 102, the core excitation light power required for the gain change of 25 dB in the case where the amplification fiber 52 or 53 having the radius a of 3.5 μm and the relative refractive index difference Δn.sub.52 or Δn.sub.53 of 0.8% in the shaded area shown in FIG. 10 was used was compared with the core excitation light power required for the gain change of 25 dB in the case where the erbium-doped fiber having the radius a of 3.5 μm and the relative refractive index difference Δn of 1.5% was used. As a result, in the erbium-doped fiber having the radius a of 3.5 μm and the relative refractive index difference Δn of 1.5%, the core excitation light power which was seven times the core excitation light power of the amplification fiber 52 or 53 having the radius a of 3.5 μm and the relative refractive index difference Δn.sub.52 or Δn.sub.53 of 0.8% was required, and the cutoff wavelength exceeded 2 μm. Further, in the case where the amplification fiber 52 or 53 having the radius a of 3.5 μm and the relative refractive index difference Δn.sub.52 or Δn.sub.53 of 0.8% was simply wound around. the bobbin, the gain temporally fluctuated from the normal value by about ±0.8 dB under an environment of ventilation of 1 m/s. In contrast to this, in the case where the above-described amplification fiber 52 or 53 was wound around the bobbin, was bonded and fixed to the bobbin, and was used, the gain did not change from the normal value. Similarly to the amplification fiber 51 of the first embodiment, it was determined that it was possible to suppress the change of the gain in the case where the amplification fiber 52 or 53 was wound around the bobbin and was bonded and fixed to the bobbin.

[0077] As described above, according to the amplification fibers 51, 52, and 53 and the optical amplifiers 101 and 102 according to the first embodiment and the second embodiment, even when the input signal light power is sharply changed, the gain is prevented from being sharply changed, and the change of the gain can be suppressed to not more than 0.2 dB.

[0078] While the preferred embodiments of the present invention have been described, the present invention is not limited to the above-described embodiments. The configuration of the present invention can be changed as long as the objects and effects described above can be achieved. In addition, specific structures and shapes when the present invention is implemented may be other structures and shapes as long as the objects and effects of the present invention can be achieved.

REFERENCE SIGNS LIST

[0079] 51, 52, 53 Amplification fiber

[0080] 61 Core

[0081] 62 First cladding (cladding)

[0082] 65 Cladding

[0083] 101, 102 Optical amplifier

[0084] a Radios

[0085] Δn.sub.51, Δn.sub.52, Δn.sub.52 Relative refractive index difference