Highly Stable Switchable Emissions of an Erbium-Doped Fiber Ring Laser Using Cascaded MZIs Based on CHCF

Abstract

A stable, single, and dual-wavelength erbium-doped fiber laser (EDFL), based on two Mach–Zehnder interferometers (MZIs), arranged in a cascade configuration, was proposed for experimental purposes. Both MZIs were assembled by splicing a capillary hollow-core fiber (CHCF) section between two multimode fibers (MMFs) segments. The novelty of this single and dual-wavelength EDFL is that the switchable operation of the laser is achieved by thermally tuning the interference pattern of one MZI and not by adjusting the polarization state inside the fiber ring cavity. The maximum measured value of SNR was 58.9 dB for the single and dual-wavelength laser emissions. Moreover, the stable output power exhibited by this EDFL, in terms of minimal power and wavelength fluctuations, at 0.05 dB and 10 pm, was detected during the single and dual-wavelength operation. It is worth noticing that switching is achieved at exact wavelength locations with a separation of 1.8 nm and not randomly, as reported by other works. These features make this switchable EDFL an appealing candidate for application in optical fiber communication systems and fiber sensing.

1. Introduction

In the last two decades, multiwavelength erbium-doped fiber lasers (EDFLs) have become reliable light sources for different applications such as wavelength division multiplexing [1,2], optical fiber sensing [3,4], and microwave photonics [5,6], to mention but a few. As is well known, homogeneous line broadening produces high mode competition, thus causing the generation of unstable emissions in multiwavelength EDFLs [7,8,9]. One possible solution to this problem is to use fiber wavelength filters; most of them work similarly to Sagnac [10], Fabry–Perot [11], and Mach–Zehnder [12] interferometers, though other filters based on different fiber structures can also be used [13,14,15,16,17,18,19,20,21,22]. Among these devices, the Mach–Zehnder interferometers (MZIs) have attracted significant attention due to their simple structure and the ease with which they can be manufactured; therefore, different optical fibers and fiber structures have been used to implement these interferometers. For instance, Y. Lv et al. [23] proposed a Mach–Zehnder interferometer (MZI) that was based on a triple-core photonic crystal fiber in order to implement a single, dual, and triple-wavelength EDFL. In this case, the total length of the fiber structure, including the length of the multimode fiber (MMF) that acts as a modal coupler, was 5.9 cm. They reported a side-mode suppression ratio (SMSR) of 50, 45, and 37 dB for single, dual, and triple-wavelength emissions, respectively, and the measured 3 dB linewidth was 0.026 nm for all cases. The maximum power fluctuation for single, dual, and triple-wavelength emissions were 0.44, 0.77, and 0.96 dB, respectively. Moreover, Y. Qi et al. [24] proposed a single, dual, and triple-wavelength EDFL, wherein the switchable operation of this laser depended on an MZI that was constructed using the core-offset technique where 57 cm of few-mode fiber was spliced between two single-mode fibers (SMF). They reported a maximum signal–noise ratio (SNR) of 54, 43, and 49 dB for single, dual, and triple-wavelength emission, respectively, and the measured 3 dB linewidth was 0.02 nm for all cases. The maximum power fluctuation for the single, dual, and triple-wavelength emissions were 1.14, 2, and 4.33 dB, respectively. Additionally, J. A. Martin-Vela et al. [12] proposed an MZI based on the core-offset technique where the sensing SMF, with 5 cm of length, was covered with Al that was 30 nm thick, in order to construct a single, dual, and triple-wavelength EDFL. They reported a maximum SMSR of 55, 50, and 50 dB for single, dual, and triple-wavelength emissions, respectively, and the measured 3 dB linewidth of the single-wavelength emission was 0.02 nm (the other linewidths were not reported). The maximum peak power fluctuation for the single, dual, and triple-wavelength emissions were 0.2, 1, and 5 dB, respectively. In the three examples mentioned above, the multiwavelength lasing operation was achieved by adjusting the polarization controller (PC) that was inside the laser cavity. Changing the polarization alters the gain and loss inside the cavity, and it modifies the interference pattern of each MZI that was used in these examples. It is worth mentioning that the number of wavelengths that are capable of lasing is determined by the peaks of maximum transmission in the fringe pattern of the MZI. For example, if a single-wavelength emission switches to a dual-wavelength emission by adjusting the PC, this dual operation could be due to the occurrence of two consecutive peaks in the fringe pattern (i.e., the wavelength spacing between these laser emissions are equal to the free spectrum range (FSR) of the interference pattern); however, if the laser switched to single-wavelength emissions, and then came back to using a dual operation, it is possible that the lasing peaks would not be consecutive. This problem can be solved by adjusting the polarization state to obtain the same result, but it requires more time. In our view, adjusting the polarization state until a specific lasing peak appears is time-consuming and it exhibits low control of the multiwavelength laser emissions in real-time.

On the other hand, several authors have proposed using two wavelength filters in a cascade configuration inside an EDFL, in a ring configuration, in order to obtain multiwavelength laser emissions with narrower linewidths, better stability, and higher SMSR [25,26,27,28,29,30]. For example, Z. Tang et al. [25] manufactured two MZIs that were set in an EDFL to implement a single and dual-wavelength laser. The first MZI was constructed by splicing a 11.2 cm segment of two core photonic crystal fibers between two SMF segments, and its FSR was approximately 25 nm. The second interferometer was constructed using two conventional fiber couplers to assemble an MZI, and the path difference of the two arms was 2.1 mm, with an FSR of 0.78 nm. We should mention that the last interferometer is usually large, and the arms’ lengths are approximately two meters. They reported an SMSR of 45 and 42 dB for single and dual-wavelength emissions, respectively, and the measured 3 dB linewidths of the single and dual-wavelength emissions were 0.026 and 0.03 nm, respectively. The maximum power fluctuation for single and dual-wavelength emissions was 0.47, 2, and 0.8 dB, respectively. Additionally, L. Zhang et al. [28] proposed using two MZIs to implement a single and dual-wavelength EDFL. The first MZI was manufactured by tapering a seven-core fiber (waist of 8.2 µm, waist length of 8 mm) with an FSR of 13.93 nm. As in the previous example, the second MZI was constructed using two conventional fiber couplers to generate an FSR of 0.37 nm. For single and dual-wavelength emissions, they reported that the values of the SMSR, linewidth, and maximum peak power fluctuation were 45 dB, 0.0171 nm, and 0.155 dB, respectively. It is worth noting that by adjusting the PC inside the cavity, multiwavelength emissions were obtained, as was also the case in the configuration where they used one interferometer. Another drawback shown in these examples is the assembly of MZI using conventional couplers, as it makes the laser more complex and bulkier.

In this paper, a stable single and dual-wavelength EDFL, based on two cascaded MZIs, was proposed and experimentally investigated. The all-fiber MZIs were constructed by splicing a section of a capillary hollow core fiber (CHCF) between two segments of MMF. The maximum length of these MZIs was 5 mm, making these interferometers attractive due to their easy construction, small size, low cost, and sturdiness. Moreover, the MZI with the longer FSR was used as the wavelength selector, whereas the MZI with the smaller FSR exhibits narrower linewidth emissions and determines the separation of switchable steps. The novelty of this single and dual-wavelength EDFL is that the switchable operation of the laser is achieved by thermally tuning the interference pattern of one MZI and not by adjusting the polarization state inside the fiber ring cavity, as in the examples mentioned above. This scheme provides a simple way to switch wavelengths over well-known wavelength locations (peak emissions), thus eliminating the randomness obtained via polarization tuning. The maximum measured value of SNR was 58.9 dB for the single and dual-wavelength laser emissions, and the linewidth for all the peaks was 30 pm. Moreover, no power and wavelength fluctuations were detected during the single and dual-wavelength operation, thus making the laser emissions of this switchable EDFL highly stable.

2. Materials and Methods

A fiber ring laser cavity was assembled, as shown in Figure 1, using a pump laser diode (BL976-PAG700, Thorlabs, Newton, NJ, USA) that coupled light to a 980/1550 wavelength division multiplexer (WDM) fiber coupler. A 2 m long segment of EDF (Er80-8/125, Thorlabs, Newton, NJ, USA) was used as the gain medium, and it was pumped with the light of the laser diode at a wavelength of 976 nm. A PC and an isolator were added to the laser cavity to obtain a better SNR and to observe the unidirectional operation, respectively. To monitor the laser output, an optical fiber coupler (90/10) was set into the cavity; the 10% port was connected to the OSA, and the 90% port was used to close the cavity for continuous laser generation. Moreover, to develop a switchable EDFL, two MZIs that worked similarly to filters were placed into the laser cavity. Their fabrication process and working principle are described in the next section.

Fabrication Process and Working Principle

The structure of these filters consisted of a spliced segment of CHCF (whose inner and outer diameters are 62.5 and 125 µm, respectively) between two pieces of 1 mm of MMF (105/125 µm), and this fiber structure was sandwiched between two conventional SMF (see Figure 1). The CHCF was composed of SiO₂, and its coating was composed of UV-cured acrylate (not commercial fiber). It should be noted that the coatings of the CHCF and MMF were removed before starting the construction process, and the SMF was handled as a standard splice. It is crucial to note that MMF segments work similarly to mode couplers, sending light toward the hollow core and the ring cladding of the CHCF. The equation that describes the output signal of this MZI is [31]

$${I=I}_1{+I}_2+\sqrt{I_1{I}_2}cos\left({2\pi \Delta n}_{eff}L/\lambda \right)$$

(1)

where I₁ and I₂ are the intensities of the light in the hollow core and the ring cladding, respectively. L, λ, and Δn_eff are the length of the MZI, the wavelength of the light, and the effective refractive index difference between the modes that travel in the CHCF (a detailed explanation can be found in [31]), respectively. Moreover, the FSR of the output signal of this MZI can be calculated using $FSR={\lambda }^2/\mathsf{\Delta }{n}_{eff}L$, and one may observe that it is feasible to change the FSR of the interference pattern by modifying the length of the CHCF. We constructed four MZIs with different FSR values by taking advantage of this property and our capability of cutting CHCF segments with high precision. Implementing a switchable laser requires two MZIs set into a fiber laser cavity in a cascade configuration. One of these devices (MZI₁) had an FSR that was higher than 10 nm, and the other (MZI₂) had an FSR that was lower than 2 nm. For the MZI₁, three MZIs were constructed, and their FSRs are 17.82, 13.29, and 10.98 nm (see Figure 2d), which correspond to the estimated lengths of 0.3, 0.4, and 0.5 mm, respectively. The MZIs with lengths of 0.3, 0.4, and 0.5 will be named MZI_1a, MZI_1b, and MZI_1c, respectively. For the MZI₂, one MZI was constructed, and its FSR was 1.8 nm (see Figure 2d), corresponding to an estimated length of 3 mm. A picture of the MZIs that were used in this experiment is shown in Figure 3.

Table 1. Characteristics of the MZIs.

	Experimental FSR (nm)	Calculated Length * (mm)	Designed CHCF Length (mm)	Error (%)	Contrast (dB)
MZI1a	17.8	0.304	0.3	1.33	42.2
MZI1b	13.3	0.407	0.4	1.75	19.7
MZI1c	11.0	0.493	0.5	1.40	19.0
MZI2	1.8	2.942	3	1.93	15.4

* This value was calculated using the experimental FSR and the equation (${FSR=\lambda }^2/{\Delta n}_{\mathit{eff}}L$).

shows the essential features of the MZIs.

It should be noted that we previously reported [32] on a tunable erbium-doped fiber ring laser using an MZI based on CHCF. In that work, we demonstrate that the spectral response of the MZI undergoes a redshift as the temperature increases, with a slope of 33 pm/°C. This thermal tunability allowed us to use two MZIs in a fiber ring laser cavity, using the MZI₁ as a wavelength selector (larger FSR); conversely, MZI₂ was used as a reference device in a steady state condition. In other words, the interference pattern of the MZI₁ tunes over the spectrum of the MZI₂ to choose a peak (at a specific wavelength), or peaks (multi-wavelengths), that will emit laser light. It should be noted that the MZI₂ determines the switchable step (FSR = 1.8 nm), as well as the narrower laser linewidth when the two MZIs are connected in a cascade configuration. Moreover, all the MZIs that were used in this experiment are polarization-independent elements; this means that the amplitude and phase of their output fringing patterns do not change with the input polarization state of light.

The spectral responses of the cascade combination of MZI_1a, MZI_1b, and MZI_1c, with the MZI₂, are shown in Figure 4. These transmission spectra were taken using the amplified spontaneous emission (ASE) of the Erbium-doped fiber (in an open cavity) when the pump power was fixed at 250 mW. In these three cases, the interference pattern of the MZI₂ was modulated by the signal generated by MZI_1a, MZI_1c, and MZI_1c, respectively (see black curves in Figure 3). For a better understanding, the peaks that show the maximum transmission in the interference pattern were produced by two cascaded MZIs, which were marked to indicate that they were highly likely to produce a laser emission. The wavelengths of the peaks were measured, and the results are shown in

Table 2. Measured transmission spectra of the MZI₂ connected in series with the (a) MZI_1a, (b) MZI_1b, (c) MZI_1c, and (d) MZI₂ (cascade configuration).

Cascaded MZIs	Peak	Wavelength (nm)
MZI2 and MZI1a	P11	1563.7
	P12	1565.5
MZI2 and MZI1b	P21	1562.7
	P22	1564.5
	P23	1566.3
MZI2 and MZI1c	P31	1560.0
	P32	1561.8
	P33	1563.6
	P34	1565.4

.

3. Results

An experimental study was carried out to analyze the switchable operation of an EDFL using two MZIs connected in a series configuration, see Figure 1. Three cases were investigated; this includes the combination of the MZI₂ with the MZI_1a, MZI_1b, and MZI_1c, respectively. It should be mentioned that changing the order of the MZI₂ and MZI₁ (in the three cases) does not affect the performance of the laser, as we observed in the experiments. The MZI₂ was set on a hot plate (Echotherm, Model IC20) at a constant temperature, 20 °C, whereas the MZI₁ was on a hot plate (Thermo Scientific™, model HP88854100, Waltham, MA, USA) that underwent different temperatures; this produced the interference pattern of the MZI₁ which suffered a redshift. In the following sections, the results of the three cases are shown.

3.1. MZI₂ and MZI_1a in a Cascade Configuration

The transmission spectrum of the cascaded MZI₂ and MZI_1a is shown in Figure 4a. This indicated that the laser emission of this EDFL would probably occur at P₁₁, since this peak shows the highest transmission. When these two cascade interferometers were set in the laser cavity (see Figure 1) both interferometers were at 20 °C, the polarization controller was properly adjusted, and the pump power was fixed at 250 mW to obtain the maximum SNR at room temperature. All measurements were acquired using this pump power. In this scenario, we observed a laser emission of 1563.7 nm (P₁₁), with an SNR of 58.9 dB, as shown in Figure 5a. The temperature of the MZI_1a was then increased in increments of 1 °C, from 20 °C to 29 °C, and it was observed that the single-wavelength laser emission remained fixed at 1563.7 nm (P₁₁) within this temperature range. Nevertheless, when the MZI_1a temperature reached 30 °C, another laser emission appeared at 1565.5 nm (P₁₂) with an SNR of 58.6 dB. This dual-wavelength laser emission was achieved because the peak of the MZI_1a, with a large FSR, was tuned to the point where peaks P₁₁ and P₁₂, from the MZI₂, which had small FSRs, exhibited the same intensity after a round trip. In such scenarios, those peaks are enhanced, and they eventually achieve laser emissions. Since this condition was achieved due to the shape of the tuning peak from MZI_1a, the dual-wavelength emission was observed within a narrower temperature range, as shown in Figure 5. This dual-wavelength laser emission continued until the MZI_1a reached a temperature of 40 °C, and at that point, at P₁₂, the EDFL switched to single-wavelength laser emissions. The laser emission at P₁₂ remained fixed when the temperature of the MZI_1a varied between 40 to 160 °C, as can be seen in Figure 5a. The linewidth of the P₁₁ and P₁₂ was 0.03 nm. Furthermore, the wavelength stability of these laser emissions against temperature changes was measured and it is presented in Figure 5b. In the case of single-wavelength laser emissions, the peaks P₁₁ and P₁₂ did not vary their amplitudes and positions as the temperature changed from 20 to 29 °C and from 40 to 160 °C, respectively. In the case of dual-wavelength laser emissions, the peaks P₁₁ and P₁₂ did not modify their amplitudes and positions as the temperature changed from 30 to 39 °C. This means that the MZI_1a can undergo temperature fluctuations of up 9 °C, and the laser emission (single/dual) will not suffer changes in terms of their amplitudes or positions (wavelengths). This also indicates that for real applications, the temperature controller used to set the temperature on the MZI_1a does not require high accuracy. Similar behaviors are observed in the other two cases (MZI_1b and MZI_1c) (see the following sections for more information).

3.2. MZI₂ and MZI_1b in a Cascade Configuration

Since the peak (P₂₁) shows the highest transmission, as is shown in Figure 4b, the laser emission is likely to occur at that peak. In the laser cavity, these two interferometers were 20 °C and connected in a series configuration (see Figure 1). Laser emission was detected at 1562.7 nm (P₂₁), with an SNR of 50.3 dB, as shown in Figure 6a, when the pump power was fixed at 250 mW. As with the previous case, a similar analysis was carried out by increasing the temperature of the MZI_1b in increments of 1 °C, from 20 to 160 °C (see Figure 6a,b). The linewidth of the P₂₁, P₁₂, and P₂₃ was 0.03 nm. The overall measurement results are summarized in

Table 3. Summarized results of the MZI₂ connected and cascaded with MZI₁.

Cascaded MZIs	Temperature (°C)	Single-Wavelength Laser Emission (nm)	Dual-Wavelength Laser Emission	SNR (dB)
MZI2 and MZI1a	20–29	P11 (1563.7)		58.9
	30–39		P11 and P12	58.9 and 58.6
	40–160	P12 (1565.5)		58.6
MZI2 and MZI1b	20–79	P21 (1562.7)		50.3
	80–89		P21 and P22	50.3 and 50.9
	90–149	P22 (1564.5)		50.9
	150–159		P22 and P23	50.9 and 50.8
	160	P23 (1566.3)		50.8
MZI2 and MZI1c	20–29	P31 (1560.0)		54.5
	30–39		P31 and P32	54.5 and 54.7
	40–89	P32 (1561.8)		54.7
	90–99		P32 and P33	54.7 and 54.2
	100–149	P33 (1563.6)		54.2
	150–159		P33 and P34	54.2 and 55.1
	160	P34 (1565.4)		55.1

.

3.3. MZI₂ and MZI_1c in a Cascade Configuration

The transmission spectrum of the cascaded MZI₂ and MZI_1c is shown in Figure 4c. The peak that shows the highest transmission is likely to produce the laser emission of the EDFL; in this case, P₃₁. In the laser cavity, these two interferometers were connected in a series configuration (see Figure 1). The pump power was fixed at 250 mW and both interferometers were at 20 °C. Laser emission was observed at 1560.0 nm (P₃₁), with an SNR of 54.5 dB, as shown in Figure 7a. An analysis similar to that of the previous cases was performed by increasing the temperature of the MZI_1c in increments of 1 °C, from 20 to 160 °C (see Figure 7a,b). The linewidth of P₃₁, P₃₂, P₃₃, and P₃₄ was 0.03 nm. The general results of the measurement are shown in Table 3. It is worth noting that as the FSR of MZI₁ was reduced, the wavelength switching range increased, as can be observed from Figure 5, Figure 6 and Figure 7. First, it is necessary to say that the temperature sensitivity of the three MZI₁ is around 33 pm/°C; therefore, the wavelength shift as a function of temperature for these filters is the same. From Figure 2a–c, one may observe that the MZI_1a, with the larger FSR, also exhibits a larger plateau at its peak, compared with the MZI_1b and MZI_1c. Since the plateau is larger, we need a more significant wavelength shift to observe laser switching, which translates to a small number of laser switching transitions for a given temperature range. On the other hand, in the case of MZI_1c, which has a smaller FSR and a sharper peak (smaller plateau), the number of laser switching transitions increases significantly for the same temperature range; therefore, a smaller FSR (i.e., sharper spectrum) provides more laser transitions.

4. Discussion

The experimental results regarding the switchable operation of the EDFL for the three possible combinations of the MZI₂ with the MZI_1a, MZI_1b, and MZI_1c, respectively, are summarized in Table 3. Regarding single-wavelength laser emissions, they were detected twice in the first combination of MZI₂ and MZI_1a, at peaks P₁₁ and P₁₂; in the second combination of MZI₂ and MZI_1b, they were identified three times, at peaks P₂₁, P₂₂, and P₂₃, and in the third combination of MZI₂ and MZI_1c, it was detected four times, at peaks P₃₁, P₃₂, P₃₃, and P₃₄. For the case of dual-wavelength laser emissions, it was detected once in the first combination of MZI₂ and MZI_1a, at peaks P₁₁ and P₁₂. In the second combination of MZI₂ and MZI_1b, it was observed twice, at P₂₁ and P₂₂, and P₂₂ and P₂₃. In the third combination of MZI₂ and MZI_1c, it was identified three times, at P₃₁ and P₃₂, P₃₂ and P₃₃, and P₃₃ and P₃₄. Regarding dual-wavelength switchable operations, they were detected by decreasing the FSR of the MZI₁ and noting the number of times that the dual operation increased in the same temperature range. Similar behavior was observed in the single-wavelength switchable operation (see Table 3). The output power of the laser emissions is another interesting feature of this EDFL that needs to be analyzed. The output power of the laser emissions decreased as the laser switched from single to dual operation, and the power increased when the laser switched from dual to single operation. This is because the power is redistributed on the peaks. When MIZ_1a was used in the laser, a maximum power change of 2.2 dBm was measured when switching from single to dual laser emissions, and a maximum power change of 2.7 dBm was measured when switching from dual to single laser emissions. In the case of the MZI_1b, maximum power changes of 3.2 and 2.7 dBm were measured when switching from single to dual laser emissions, and from dual to single laser emissions, respectively. In the case of the MZI_1c, the maximum power changes when switching from single to dual laser emissions and from dual to single laser emissions were the same, at 3.7 dBm.

It should be noted that the most significant characteristic of the proposed EDFL is that it can be switched from lasing single to dual wavelengths by increasing the temperature of the MZI₁. This performance offers an advantage over the switchable laser based on polarization [10,11,12,13,23], not only because of its simple straight-line operation but also because of its stability against temperature fluctuations. Switchable lasers based on polarization have a problem with temperature fluctuations, and these fluctuations produce changes in the polarization state; therefore, small adjustments need to be applied to the polarization controller to keep the wavelength and amplitude of the laser emission. However, the laser emissions of this switchable laser exhibited fluctuations in wavelength and power when the temperature of the MZI₁ was changed within a certain temperature range. These experimental results would appear to indicate that a highly accurate temperature controller is not required to regulate the temperature of the MZI₁, since temperature oscillations of up to 9 °C do not change the wavelength or amplitude of the single or dual wavelengths, and in some cases, this temperature value can be even larger (see Table 3). This can be achieved due to the small size of both MZIs. It is worth noting that the level linearity and wavelength linearity of the OSA are 0.05 dB and ±10 pm, respectively; this means that our equipment is not able to detect power or wavelength fluctuations that are lower than the values mentioned above. It is important to highlight two vital features of this proposed switchable EDFL, which both come from the use of the MZI₂. The first one is related to the linewidth of 0.03 nm for all the laser emissions, although the other interferometers (MZI_1a, MZI_1b, and MZI_1c) have different FSR values. The second is the fact that the wavelength spacing between dual-wavelength laser emissions is always 1.8 nm, which is the FSR of the MZI₂.

Table 4. Laser parameter comparison using recently published works.

No.	Structure	Switchable Lines	SMSR/NSR	Linewidth
1 [26]	MZI (bi-tapered) and Sagnac (Coupler)	5	NA/>14.98 dB	NA
2 [10]	Sagnac (coupler) and Filter (PMF)	3	NA/23 dB	<600 pm
3 [25]	MZI (coupler) and Filter (TCPCF)	2	>45 dB/NA	26 pm
4 [28]	MZI (coupler) and MZI (TSCF)	1	45 dB/NA	17.1 pm
5 [27]	Filter (SE-HSOFF) and Sagnac (coupler)	4	NA/>30 dB	NA
6 [29]	MZI (Coupler) and Sagnac (Coupler)	5	48.7 dB/NA	60 pm
7 [30]	NOLM (Coupler) and Filter (Lyot)	23	>25 dB/NA	500 pm
8 [13]	Filter (FLCSCF)	6	NA/>50 dB	20 pm
This work	MZI (CHCF) and MZI (CHCF)	2	NA/58.9 dB	30 pm

was expanded upon to compare the performance of our laser with that of similar switchable lasers, using recently published studies. It is worth noting that in most of these works (from 1 to 7), a fiber interferometer was constructed using conventional fiber couplers, thus causing the system to become more complex and bulkier. The laser linewidth of our switchable laser is narrower than 2, 6, and 7, but wider than 3, 4, and 8. Regarding the SNR, its value is higher than all the lasers in Table 4. These striking results demonstrate the good performance of our switchable laser.

5. Conclusions

This paper has proposed and experimentally demonstrated the use of a single and dual-wavelength EDFL, incorporating two MZIs, in a series configuration. The all-fiber MZIs were constructed by splicing a section of a capillary hollow core fiber (CHCF) between two segments of MMF. These MZIs exhibit good characteristics, such as ease of construction, small size, low cost, and robustness when being handled. Moreover, their interference patterns have contrast well and do not change with the input polarization’s state of light (polarization-independent element). Moreover, an MZI with a longer FSR, was used as a wavelength selector, whereas an MZI with a shorter FSR made the linewidths narrower and determined the size of the switchable step. The novelty of this single and dual-wavelength EDFL is that the switchable operation of the laser is achieved by thermally tuning the interference pattern of the MZI, and not by adjusting the polarization state inside the fiber ring cavity. The maximum measured value of SNR was 58.9 dB for the single and dual-wavelength laser emissions, and the linewidth of all the peaks was 30 pm. Furthermore, no power or wavelength fluctuations were detected during the single and dual-wavelength operation, making the laser emissions of this switchable EDFL highly stable. The good performance of this switchable EDFL makes it a candidate for potential application in optical fiber communications systems and fiber sensing.