Browse

• Abstract

SECTION 1

## INTRODUCTION

Erbium-doped fiber (EDF) has emerged as one of the strong candidates to be employed as the gain medium in a fiber laser. Consequently, by exciting its Erbium ions, it can provide amplification of signals at wavelength attractive for optical communication, which is around 1550 nm without introducing any effects of gain narrowing [1], [2]. Nevertheless, EDF ring lasers usually suffer from homogeneous gain broadening, mode hopping, mode competition, and multimode oscillation [3]. In addition, the issue of multimode output that rises from fiber ring laser due to mode hopping, longer cavity length, and very narrow longitudinal mode spacing restricts the fiber lasers from obtaining single-longitudinal-mode (SLM) operation, which results in formation of noises in frequency domain [4]. In this regard, SLM output in EDF-based fiber lasers have become an important criteria for multitude of operations including fiber optic sensors, modern instrumentation, wavelength-division-multiplexing (WDM) communications, and microwave photonics system [5]. In the past, many works pertaining to SLM generation have been demonstrated [6], [7], [8], [9], [10], [11], [12], [13], [14], [15]; however, most employ highly complex techniques such as multiple ring cavity structures [16], tunable ring resonators [17], external light injection [18], unidirectional loop mirrors [19], acousto-optic tunable filters [20], and fiber Fabry–Perot filter (FFPF) [21]. Simpler configurations have also been demonstrated using unpumped EDFs as a saturable absorber [22], though at the cost of performance due to high insertion losses and, therefore, requiring high pump powers. Nowadays, graphene has been attaining much interest and attention in both photonics and optoelectronics application due to its outstanding and unique features, including high mobility, good optical transparency, and ultrawideband tunability, which arises from the zero bandgap energy, which itself is a consequence of the linear dispersion of Dirac electrons [23]. There are numerous journal papers that have reported recently on graphene for various potential significance, for instance, a graphene-based saturable absorber for mode-locked fiber laser [24], [25], [26]. In this paper, a stable and inexpensive tunable SLM C-band fiber laser is proposed and demonstrated. The fiber laser uses a short length of highly doped EDF as the active gain medium and multilayer graphene acting as a saturable absorber. SLM lasing in a simple cavity design with the tuning range between 1547.88 and 1559.88 nm is obtained. The deposition process of graphene on the surface end of the fiber ferrule is done by using a novel technique employing index matching gel (IMG) to adhere the graphene flakes onto the fiber ferrule. In addition, the characteristics of output power, wavelength, signal-to-noise ratio (SNR), and radio frequency (RF) spectrum are performed and analyzed.

SECTION 2

## DEPOSITION OF GRAPHENE

IMG is used to adhere graphene flakes onto the surface of the fiber ferrule as an alternative approach. The graphene flakes used are suspended in a solution as obtained from Graphene Research Ltd. The IMG used has an index of refraction that closely approximates to that of the silica optical fiber, which has a value of 1.463. Normally, the IMG is used to reduce Fresnel reflection at the end surfaces of the fiber connectors. The other benefit of IMG is that it has a very high transparency, low evaporation, excellent adhesion, and good mechanical shear stability. In this paper, the IMG is used in a unique manner, whereby it is used to provide adhesion of the graphene flakes onto the ferrule.

In the deposition process, the IMG is firstly spread thoroughly onto the surface end of the fiber ferrule. Then, the fiber ferrule with the attached IMG is immersed into the graphene solution. After this, the attached graphene as well as the IMG is dried at room temperature. Fig. 1 shows the schematic illustration of the fiber ferrule with the deposited graphene on its surface end. The thick layer of graphene appears finely dispersed around the surface of the fiber ferrule without excessive clustering of the graphene.

Fig. 1. Schematic representation of the fiber ferrule deposited with graphene using index matching gel (IMG: Index Matching Gel; FF: Fiber Ferrule; G: Graphene).

The result of the deposition method is microscopically examined by Raman spectroscopy (Renishaw) as to measure the Raman spectrum of the sample. The Raman spectrum is acquired by laser excitation at 532 nm (2.33 eV) with an exposure time of 10 s using a 1800-lines/mm grating. The incident power and the depth of field are set to be 5 mW and 1 $\mu\hbox{m}$, respectively. The detector used in this Raman spectroscopy is a charge-coupled device camera. Using a 100× objective lens with a numerical aperture NA of 0.85, we obtain a spot size of 0.5 $\mu\hbox{m}$. The spot size is defined as the diameter of the laser spot on the sample.

The Raman spectrum of the deposited graphene is shown in Fig. 2, which exhibits the intensity peaks at Raman shift of approximately 1350, 1580, and 2700 $\hbox{cm}^{-}1$. The most intense features in Raman spectrum of graphene are the two prominent peaks, one located around the Raman shift of 1580 $\hbox{cm}^{-1}$, commonly called the G peak, and the other one located around the Raman shift of 2700 $\hbox{cm}^{-1}$, namely 2-D peak [27], [28]. The peak profile of the Raman spectrum obtained in Fig. 2 matches the specified peak profile of graphene in the Raman spectrum, as stated above, and as reported previously in other works [27], [28], [29]. This indicates that graphene is well deposited on the fiber ferrule in this work.

Fig. 2. Raman spectrum of the thick layer graphene deposited on the fiber ferrule.

In defected graphene, another peak will be observed around the Raman shift of 1350 $\hbox{cm}^{-1}$, historically named D peak, and its relative signal strength (compared with the G peak) depends strongly on the amount of disorder in the graphitic material, which originates from the graphene edge [28]. The width of the 2-D peak can be used to determine the number of graphene layers (whereby the width is larger as the number of graphene layers increase). This effect reflects the change in the electron bands through a double resonant Raman process based on the electronic structure and the phonon dispersion [27], [28]. Besides that, another way to distinguish the single layer graphene from multilayer graphene is by calculating the intensity ratio of G peak over 2-D peak. Single-layer graphene is indicated by the low intensity ratio of G/2-D, which is generally lower than 0.5, whereas multilayer graphene is identified by a high intensity ratio of G/2-D, which is larger than or close to 1 [29]. In our work, the calculated intensity ratio of G/2-D from Fig. 3 exceeds the value of 1 (1.04499), thus signifying that the deposited graphene is multilayer. The estimated thickness of the graphene layer, based on the above value, is approximately about 0.712 nm.

Fig. 3. Spot image of the deposited graphene under Raman spectroscopy.

From the comprehensive explanation of Fig. 2, it can be concluded that the multilayer graphene is successfully and properly deposited on the fiber ferrule. The noise in the Raman spectrum observed from the figure is most probably attributed by the Raman shift of the IMG. In Fig. 3, the spot image of the deposited thick layer graphene viewed under Raman spectroscopy is presented.

The advantage of this method is its low complexity and cost effectiveness, as no other optical, chemical, or electrical methods are needed. The IMG also has the additional advantage of having minimal optical loss. This has been proven experimentally by checking the power loss using a TLS, a patchcord with IMG applied onto its ferrule end, and an optical power meter. Moreover, by using this method, a very thick graphene layer can be deposited at a time. The basic operation of graphene as the saturable absorber is described in the other section of this paper.

SECTION 3

## EXPERIMENTAL SETUP

Fig. 4 shows the experimental setup of the proposed SLM tunable C-band fiber laser, which consists of a 1-m EDF (LIEKKI™ Er80-8/125), which is a highly doped large mode area erbium fiber with core absorption coefficients of 41 and 80 $\hbox{dBm}^{-1}$ at 980 and 1530 nm, respectively. This type of fiber is ideal for medium peak power pulse amplification as it has low splice loss, high doping, and a large core, with a mode field diameter of 9.5 $\mu\hbox{m}$ at 1550 nm, as well as core numerical aperture of 0.21. Its high erbium concentration reduces the required fiber length considerably while providing strong gain and reduced nonlinear effects like Four-Wave Mixing, Stimulated Raman Scattering, and Stimulated Brillouin Scattering. The length of the highly doped EDF is chosen such that it produces the optimum gain in the fiber laser, as well as allowing the discharge of excess pump power to ensure that the EDF is totally in saturated condition.

Fig. 4. Experimental setup for tunable highly Er-doped fiber ring laser.

The fiber is pumped by a 980-nm laser diode at 143 mW through a 980/1550 WDM, and the other end is connected to Port 1 of an optical circulator (OC), whereas Port 2 of the OC is connected to the tunable fiber Bragg grating (TFBG) to provide the tuned reflected wavelength. The TFBG provides the tuning mechanism giving a tuning range of more than 10 nm by applying mechanical stress (extension or compression) to the FBG, which resulted in the shift of the Bragg resonance wavelength. This is done by bending a piece of Perspex with a low Young modulus in upward or downward direction with the FBG glued onto it. The details about this TFBG design and basic operation are described in [6]. Port 3 of the OC in this experimental setup is then connected to a 95:5 fused coupler with the 95% port connecting back to the 1550-nm port of the WDM, thus creating a ring cavity. The 5% port of the fused coupler is connected to an optical spectrum analyzer (OSA), which serves as the output of the fiber laser. In between the 95% port of the coupler and the 1550 port of the WDM, there is a ferrule with commercial graphene layer, which acts as the saturable absorber to generate SLM output in this fiber laser. The basic principle of the graphene-based saturable absorber is described in the section below.

SECTION 4

## GRAPHENE AS SATURABLE ABSORBER

The photonic properties of graphene are significant. Graphene, which is composed of a single layer from 3-D graphitic crystalis, is one of the carbon allotropes. It consists of one-atom-thick planar sheets of $\hbox{sp}^{2}$-bonded carbon atoms that are densely packed in a honeycomb crystal lattice. Its 2-D sheet, which differs from Carbon Nanotubes (CNT), reveals its uniqueness. The CNT, which is commonly used as saturable absorber, and the principles are clearly explained in [23], [30], and [31]. In achieving SLM laser operation, the saturable absorber is used to suppress the multilongitudinal mode and noises in a fiber laser. In the case of low intensity light incident, photons are highly absorbed, and the electrons in the valence band are excited up to conduction band of the saturable absorber material. In the case of high intensity light incident, some photons are not absorbed due to the occupation of electrons in the conduction band, which are excited by the photons from the low intensity light. Therefore, only high intensity light can pass through the saturable absorber with very low loss and vice versa. In principle, the optical absorption of graphene layers is proportional to the number of layers, each absorbing ${\rm A} \approx 1 - {\rm T} \approx \pi \approx 2.3\%$ over the visible spectrum [23]. Hence, it can be concluded that the thicker the graphene deposited, the more optical absorption it exhibits.

SECTION 5

## RESULTS AND DISCUSSION

Fig. 5 shows the experimental tunability of the SLM C-band fiber laser using a highly doped 1-m-long EDF taken from an Ando OSA (AQ6317) with a spectral resolution of 0.02 nm. The indicated tuning range of this fiber laser spans from 1547 to 1560 nm, and this is not limited as the tuning range can exceed above 1560 nm and below 1547 nm, respectively. By using a differential micrometer head, the tuning resolution can be further improved [6].

Fig. 5. Output spectra versus wavelengths in the tuning range of 1547.88 to 1559.88 nm.

The lasing wavelengths are observed to be of a SLM and can be tuned continuously over the desired wavelength region. No mode hoping is observed as the wavelengths are tuned, and this can be attributed to the single wavelength lasing allowed by the FBG.

As opposed to other earlier works [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], the graphene-based saturable absorber mechanism used in this experiment experiences lower insertion loss, which leads to higher and stable output power. From the figure, there is a very low output power variation over the entire wavelength range, with peak amplitudes ranging from −5.65 to −7.18 dBm. The maximum peak amplitude, which is about −5.6 dBm, is at 1553.88 nm, varying slightly among all the other wavelengths in the figure. This is plotted in Fig. 6, which shows the variation of the peak amplitude power against the tuned SLM wavelengths. In addition, the SNR, which is an important characteristic for a fiber laser, is also shown in Fig. 6. From the plot of the SNR against the tuned SLM wavelength, it can be seen that the distribution of SNR is obtained between 66.0 and 68.3 dB. An excellent SNR profile is observed with the highest SNR value of 68.25 dB at 1553.88 nm, and the corresponding output power is −5.65 dBm. The low deviation of the SNR values across the whole wavelength range in the figure indicates that the quality of the proposed fiber laser is of very high, as opposed to that in the earlier work [6].

Fig. 6. Output power and SNR versus wavelength.

To realize and investigate the stabilities of output power and output wavelength in the proposed fiber laser, a short-term stability measurement is carried out, and the result is shown in Fig. 7. The observation time is over 60 min. at 1552.09 nm with an output power of −5.71 dBm initially. The power and wavelength variations are observed to be less than 0.04 dB and 0.12 nm, respectively. This proved that the output stability of the proposed fiber laser is well maintained over time.

Fig. 7. Output stability measurements of the fiber laser over 60 min observation time.

The verification of the SLM is measured using a high-speed photodetector (HP 53440B, 6 GHz) and an RF spectrum analyzer (Anritsu 2683A), which are generally used to show SLM behavior. The result is shown in Fig. 8. From the figure, it is observed that there is no beat detected in the RF spectrum. In a fiber laser, the beating can occur when there is more than one mode oscillating in the cavity. Thus, SLM laser operation can be indicated by zero beating, as observed in the RF spectrum.

Fig. 8. RF spectrum of the output laser.

In this paper, apart from Fig. 8, further verification of SLM operation is reaffirmed using a delayed self-heterodyne RF spectrum technique. The basic setup consists of a 3 dB (1 × 2) coupler with one port connecting to a 500-m-long single-mode fiber (SMF), which functions as the delay line, and the other port is connected to the Acousto Optic Modulator (AOM), is then recombined using a 3 dB (2 × 1) coupler [6]. In this basic setup, the input coupler divides the signal from the fiber laser into two portions of the same power, with one portion propagating into the 500 m long SMF, while the other portion propagates into the AOM, which operates at 80 MHz. Both signals are then recombined at the output coupler [32]. The measured line-width from the RF spectrum is shown in Fig. 7, and this equals 206.25 kHz, which proves that the output of the fiber laser operates in SLM. The proposed setup using a short length of highly doped erbium fiber in a ring cavity configuration with graphene as saturable absorber is capable to produce SLM operation. To the best of the authors' knowledge, this is the first report of graphene-based saturable absorber providing SLM operation in a tunable fiber laser. The RF spectrum of delayed self-heterodyne signal is shown in Fig. 9, while the linewidth measurement for 13 tuning wavelengths is shown in Fig. 10. Only small linewidth variations are observed for different wavelengths in Fig. 10.

Fig. 9. RF spectrum of delayed self-heterodyne signal.
Fig. 10. Linewidth measurement versus wavelength.
SECTION 6

## CONCLUSION

As a summary, the proposed graphene-based saturable absorber for tunable EDF laser with an SLM operation using a short length of 1 m of highly doped EDF as the gain medium has been demonstrated. The graphene flakes are deposited mechanically as multiple layers using IMG on the fiber ferrule, which is verified by intensity peaks at Raman shifts of approximately 1350, 1580, and 2700 $\hbox{cm}^{-1}$. Being the gain medium in the ring cavity laser, a 1-m EDF has core absorption coefficients of 41 and 80 $\hbox{dBm}^{-1}$ at 980 and 1530 nm, respectively, and is used in conjunction with the saturable absorber and is used to suppress the multilongitudinal mode and noises in a fiber laser, thus giving SLM operation. Verification of SLM operation is done by detecting the RF spectrum, which in this case shows no frequency beating and using the delayed self-heterodyne technique. The measured linewidth is 206.25 kHz and clearly indicates SLM operation. Tunability of the fiber laser is obtained by the inclusion of a TFBG, which gives a wavelength output of between 1547.88 and 1559.88 nm with an average peak power of −6.48 dBm. The tuning can be further enhanced by having an improved TBFG design. The measured SNR is between 66.0 and 68.3 dB and reaches its maximum value of 68.25 dB at 1553.88 nm with corresponding output power of −5.65 dBm. The result confirms that the graphene-based saturable absorber by the simple deposition method using IMG is applicable to give a good performance of the SLM tunable fiber laser.

## Footnotes

This work was supported by the University of Malaya under HIR Grant (Terahertz, UM.C/HIR/MOHE/SC/01) and by MOHE. Corresponding author: H. Ahmad (e-mail: harith@um.edu.my).

## References

No Data Available

## Cited By

No Data Available

None

## Multimedia

No Data Available
This paper appears in:
No Data Available
Issue Date:
No Data Available
On page(s):
No Data Available
ISSN:
None
INSPEC Accession Number:
None
Digital Object Identifier:
None
Date of Current Version:
No Data Available
Date of Original Publication:
No Data Available