Effect of laser linewidth and fiber length on self-pulsing dynamics and output stabilization of...

Effect of laser linewidth and fiber length on self-pulsingdynamics and output stabilization of single-mode

Yb-doped double-clad fiber laser

Brahma N. Upadhyaya,1,* Antony Kuruvilla,1 Usha Chakravarty,1

Mangalpady R. Shenoy,2 Krishna Thyagarajan,2

and Shrikant M. Oak1

1Solid State Laser Division, Raja Ramanna Centre for Advanced Technology, Indore 452013, India2Department of Physics, Indian Institute of Technology Delhi, New Delhi 110016, India

*Corresponding author: [email protected]

Received 4 January 2010; revised 22 March 2010; accepted 22 March 2010;posted 23 March 2010 (Doc. ID 121626); published 14 April 2010

We report an experimental study of the effect of fiber length and laser linewidth on self-pulsing dynamicsand output stabilization of a single-mode Yb-doped double-clad CW fiber laser. It is found that initiationof self-pulsing under low-level pumping conditions is due to relaxation oscillations and saturable absorp-tion in the weakly pumped region of the doped fiber, irrespective of the fiber length and the laser line-width. However, with an increase in pump power, depending on fiber length and laser linewidth, thepulses initiated due to relaxation oscillation get amplified, and result in short-duration giant pulsesdue to either stimulated Brillouin scattering (SBS) or stimulated Raman scattering (SRS). In the caseof fiber lasers that employ a broadband mirror and wherein the fiber length is sufficient to reach the SRSthreshold, the giant self-pulses are generated by SRS, whereas in the case of fiber lasers using a fiberBragg grating, characterized by narrowband reflection and with sufficient fiber length to reach the SBSthreshold, the giant self-pulses are generated by SBS. Output stabilization and, hence, elimination ofself-pulsations can be achieved either by suppressing the relaxation oscillations with the addition ofan appropriate length of a passive fiber to sufficiently increase the cavity photon lifetime, or by increasingthe pump power to achieve gain uniformity along the doped fiber such that relaxation oscillations andreabsorption effects are suppressed. © 2010 Optical Society of America

OCIS codes: 140.0140, 140.3510, 140.3615.

1. Introduction

The output power from single transverse mode Yb-doped double-clad CW fiber lasers has been scaledto the kilowatt level recently, and such high-powerfiber lasers are being used in various industrialand medical applications as a potential replacementfor conventional bulk solid-state lasers [1,2]. How-ever, achieving truly CW operation of high-powerfiber lasers is not trivial due to self-pulsations thatoccur in most of the rare-earth-doped fiber lasers

[3–6]. These high peak power short-pulse self-pulsa-tions may result in pulse energies enough to causecatastrophic damage of the doped fiber and can makethe fiber unusable. There are several reports on in-vestigations of self-pulsations and techniques to sup-press self-pulsation in rare-earth-doped fiber lasersunder different experimental conditions of pumppower, resonator, and mirror configurations [7–28].The various mechanisms that may cause self-pulsingin fiber lasers include external perturbations, suchas pump noise [19], reabsorption by an unpumpedpart of the fiber [3,17,18], scattering processes, suchas Rayleigh scattering and stimulated Brillouin scat-tering (SBS) [12,13], and formation of ion pairs in

0003-6935/10/122316-10$15.00/0© 2010 Optical Society of America

2316 APPLIED OPTICS / Vol. 49, No. 12 / 20 April 2010

highly doped fibers [14,15]. However, further inves-tigations are required to clearly pinpoint the me-chanisms responsible for self-pulsing instability inreal laser configurations so as to achieve higher out-put power and truly CW operation of the laser, freefrom amplitude modulations.In this paper, we have carried out a detailed study

on the self-pulsing dynamics of a single-mode Yb-doped double-clad fiber laser using long and shortfiber lengths, and with wide and narrow laser oscil-lation bandwidths, which is realized by employingbroadband mirrors and narrowband fiber Bragggrating (FBG) reflectors. It has been found that self-pulsing is initiated by relaxation oscillations and sa-turable absorption along weakly pumped portions ofthe doped fiber, irrespective of the fiber length or thelaser linewidth. However, with an increase in pumppower, depending on fiber length and laser linewidth,relaxation oscillation pulses get amplified and giverise to high-peak-power giant pulses due to SBS orstimulated Raman scattering (SRS), which maycause catastrophic damage to the doped fiber. In thecase where a broadband mirror is used and thelength of the fiber is sufficient to reach the SRSthreshold, short-duration giant self-pulses are gener-ated by SRS, and in the case where a FBG mirrorwith narrow linewidth is used with sufficient fiberlength to reach the SBS threshold, short-durationgiant self-pulses are generated by SBS. It has alsobeen found that self-pulsing suppression can beachieved by suppression of relaxation oscillations byadding an appropriate length of a passive fiber to suf-ficiently increase photon lifetime in the resonator, orby increasing the pump power and gain uniformity toa level such that relaxation oscillations and reab-sorption effects are suppressed. To the best of ourknowledge, this is the first report of the study ofthe effect of broadband and narrowband FBG cavityreflectors on self-pulsing dynamics and output stabi-lization observed in a single experiment.

2. Experiment

Figures 1(a)–1(d) show the experimental setups forthe study of the self-pulsing dynamics of a Yb-dopedCW fiber laser. Four different configurations werestudied. Figures 1(a) and 1(c) show setups for thestudy with broadband bulk mirrors and narrowbandFBG mirrors, respectively, while Figs. 1(b) and 1(d)show setups for the study with provision for insertingan additional length of passive fiber. In each case, aYb-doped double-clad fiber having a mode field dia-meter (MFD) of 5:5 μm at 1060nm with numericalaperture (NA) of 0.15, and an inner-clad diameterof 125 μm with a NA of 0.46 has been used. ThisYb-doped fiber had an octagonal inner-clad geometryand clad-pump absorption of 1:7dB=m at 975nm. A20W fiber-coupled laser diode having a center wave-length of 975nm at 20 °C has been used to pump an8m length of the Yb-doped fiber from one end. Thisresulted in absorption of about 13:6dB, or 95.6%, ofpump power. The pump laser output, having a

200 μm core diameter fiber pigtail with 0:22NA,was collimated using a lens of 40mm focal lengthand then focused using another lens of 25mm focallength to have a focused spot diameter of about125 μm on the input end of the doped fiber. The dopedfiber was perpendicularly cleaved at both the ends tosustain higher damage thresholds. Two dichroic mir-rors were used in this configuration: one of them hashigh transmission at 975nm and high reflection inthe broad range of 1040−1120nm for normal inci-dence; the other mirror has similar characteristicsfor a 45° angle of incidence. In each case, pumpingwas carried out from the left end of the doped fiberand the laser would be in the “backward-pumpingconfiguration,” in which the cleaved pump inputend of the doped fiber with ∼4% Fresnel reflectionacts as the output coupler; feedback is provided bythe highly reflecting broadband bulk mirror or the99.6% FBG mirror at the other end. The FBG mirrorused in the experiment had a FWHM linewidth of0:2nm with center peak at 1093nm and the Bragggrating fiber had a MFD of 6:2 μm at 1060nm with0:14NA. To study the effect of an increase in the cav-ity photon lifetime, the cavity length was increasedby splicing a passive fiber of length 80mwith a modefield diameter of 6:2 μm at 1060nm and a NA of 0.14to the Yb-doped fiber. Further, to study the effect ofdoped fiber length on self-pulsing dynamics, anotherYb-doped double-clad fiber of 18m length havingcore diameter of 10 μm with NA of 0.075, and an in-ner-clad diameter of 400 μm with a NA of 0.46, wasused in the experimental setup of Fig. 1. This Yb-doped fiber also had an octagonal inner-clad geome-try and clad-pump absorption of 0:8dB=m at 975nm,which resulted in absorption of 14:4dB, or 97.4% ofpump power.

Figure 2 shows variation of the fiber laser outputpower as a function of the input pump power for eachcase. In the case of broadband mirror without aspliced passive fiber, a maximum output power of2:58W was obtained at an input pump power of12:36W, with a slope efficiency of ∼22:7%. Whenan 80m length of passive fiber was spliced, the maxi-mum output power was reduced to 1:76W due tosplice loss. In the case of the FBG mirror, the max-imum output power was 1:68W, and when 80m ofpassive fiber was spliced, the maximum outputpower was reduced to 0:92W due to splice loss at twolocations. Threshold pump power for lasing wasabout 1:3W. The laser output varies linearly with in-put pump power in each case. Output slope efficiencywas relatively small due to low pump coupling effi-ciency. The spectrum of the laser output was ob-served using an optical spectrum analyzer (AgilentTechnologies) with a resolution of 0:06nm. The laseroutput spectrum was peaked at 1093nm with aFWHM of ∼6nm in the case of the broadband mirrorand of ∼0:2nm in the case of the FBG mirror. Thelaser output was in the single transverse mode withdiffraction-limited beam quality, and was emitted ina full cone angle of 300mrad in the case of the 8m

20 April 2010 / Vol. 49, No. 12 / APPLIED OPTICS 2317

length of the Yb-doped fiber and of 150mrad in thecase of the 18m length of the Yb-doped fiber, definedby the NA of the core of the doped fibers. The randomself-pulsing of the fiber laser output was recordedusing a 1GHz photoreceiver and a 1GHz, 2GSa=sdigital storage oscilloscope.

3. Results and Discussion

Figures 3(a)–3(d) show the self-pulses in the output,corresponding to the case of broadband mirror with-out addition of passive fiber. It can be seen that the

self-pulses are initiated with relaxation oscillations,which are separated by about 6 μs, and the peak am-plitude of these self-pulses increases as the pumppower increases from 2.27 to 11:78W. The scale onthe y axis (a.u.) corresponds to the photodetector out-put that is proportional to the incident power; the re-lative strengths of the outputs are on the same scalein all measurements for a particular configuration.Figure 7(a) shows the output spectrum correspond-ing to these pulses. The output spectrum is peakedat 1093nm with an FWHM linewidth of ∼6nm.

Fig. 1. (Color online) Experimental setup for study of self-pulsing dynamics in Yb-doped fiber laser using (a) a broadband mirror, (b) abroadband mirror and a spliced passive fiber, (c) a FBG mirror, and (d) a FBG mirror and a spliced passive fiber.


Because of the short length of the resonator, thethreshold for onset of nonlinear SBS and SRS effectsare very high and, hence, there is no accompaniedsignal corresponding to SRS or SBS in the outputspectrum. When an 80m length of the passive fiberis added to the 8m length of active fiber, self-pulsesdisappear from the output over the entire range ofpump power up to 14:4W. Figures 4(a) and 4(b) showoutput dynamics at input pump powers of 3.31 and12:36W, respectively; Fig. 7(b) shows the corre-sponding output spectrum. The output spectrum getsshifted slightly toward the lower wavelength side,with the peak at 1088nm and FWHM linewidth of∼6nm. It can be observed that output dynamics getsmodified with the addition of a long length of passivefiber and relaxation oscillation spikes are totallyeliminated. It is clear that, in spite of a lower thresh-old, SBS and SRS are not initiated due to eliminationof the relaxation oscillations. Thus, with the additionof an extra length of the passive fiber, the onset ofrelaxation oscillation was eliminated and, hence,the formation of high-peak-power self-pulses wasalso eliminated. Although, with the addition of extralength of passive fiber, the threshold for SBS and

Fig. 2. (Color online) Variation of fiber laser output from an 8mlong Yb-doped double-clad fiber as a function of the input pumppower for the cases of a broadband mirror, a broadband mirrorwith a spliced passive fiber, a FBG mirror, and a FBG mirror witha spliced passive fiber in the resonator.

Fig. 3. Self-pulses in the case of an 8m long Yb-doped double-clad fiber, using a broadband mirror without a passive fiber, at input pumppowers of (a) 2.27, (b) 5.40, (c) 7.56, and (d) 11:78W.


SRS decreased significantly, due to absence of initialrelaxation oscillation pulses, the signal power wasstill below nonlinear thresholds.

Figs. 5(a)–5(d) show self-pulses when the broad-band mirror was replaced by a narrowband FBG re-flector. It can be seen that the self-pulses again start

Fig. 4. Self-pulses in the case of an 8m long Yb-doped double-clad fiber, using a broadband mirror with a spliced passive fiber, at inputpump powers of (a) 3.31 and (b) 12:36W.

Fig. 5. Self-pulses in the case of an 8m long Yb-doped double-clad fiber, using a FBGmirror without a passive fiber, at input pump powersof (a) 2.27, (b) 5.40, (c) 11.8, and (d) 14:4W.


with initial relaxation oscillations, and the peak am-plitude of these self-pulses increases with an increasein pump power up to 11:8W.With further increase ofthepumppower, self-pulsesdisappear fromtheoutputat an input pump power of 14:4W [Fig. 5(d)]; Fig. 7(c)showsthecorrespondingoutput spectrum.Theoutputspectrum is peaked at 1093nm with an FWHM line-width of 0:2nm. Figures 6(a)–6(d) show self-pulseswhen an 80m length of passive fiber is spliced be-tween the active fiber and the FBG. In this case also,the peak power of self-pulses increases with an in-crease in pump power, and reaches amaximum valueof 800 a:u: at an input pump power of 4:88W; with afurther increase of the pump power, the self-pulsesdisappear fromthe output [Fig. 6(d)] at an input pumppower of 9:67W. Figure 7(d) shows the correspondingoutput spectrum; the output spectrum is peaked at1093nmwithaFWHMlinewidthof0:2nm.Oncarefulobservation of the output spectrum, we find that it isbroadened by about 0:7nm toward the longer wave-length side at the input pump power of 4:88W, i.e.,when thepeakpower of the self-pulses is large enoughto reach the SBS threshold.With a further increase in

pumppower, these self-pulses disappear from the out-put, and broadening of the output spectrum is alsoeliminated. This shows that broadening of the outputis due to high-peak-power pulses and can be attribu-ted to initiation of nonlinear SBS process.

Figures 8(a)–8(d) show self-pulsation in the case ofa fiber laser with an 18m long Yb-doped double-cladfiber. In this case also, the self-pulses start fromrelaxation oscillations and its peak amplitude in-creases with an increase in pump power. Figure 9shows the corresponding output spectrum with theincrease in pump power. The spectrum is peakedat 1089nm with an FWHM linewidth of ∼6nm. Itcan be seen that, when the peak power of theseself-pulses is enough to reach the SRS threshold,an SRS peak at 1146nm that is shifted from the sig-nal peak by about 56nm appears in the output spec-trum. To analyze the nonlinear effects of SBS andSRS, we look for thresholds for the onset of SRSand SBS, which are given by [29,30]

ðP0crÞSRS ≈

16Aeff

gRLeff; ð1Þ

Fig. 6. Self-pulses in the case of an 8m long Yb-doped double-clad fiber, using a FBG mirror with a spliced passive fiber, at input pumppowers of (a) 1.25, (b) 2.27, (c) 4.88, and (d) 9:67W.


ðP0crÞSBS ≈

21Aeff

gRLeff

�1þ Δvs

ΔvB

�; ð2Þ

where Aeff is the effective core area and Leff is theeffective fiber length given by

Leff ¼1αs

½1 − expð−αsLÞ�: ð3Þ

The Brillouin and Raman gain coefficients in silicaare gB ¼ 5 × 10−11 m=W and gR ¼ 1 × 10−13 m=W, re-spectively [29,30].ΔvB ≈ 30MHz is the Brillouin gainbandwidth and αs ≈ 5 × 10−3 m−1 is the scattering lossat the signal wavelength. The calculated values ofSBS and SRS thresholds for different configurationsused in the experiment are listed in Table 1. Fromthis table, we see that, in the case of fiber laser withthe broadband mirror, due to large linewidth of thesignal, the SRS threshold is much lower than theSBS threshold and high-peak power self-pulses canbe generated through SRS. However, in the case ofFBG mirrors, the SBS threshold is lower and cancause generation of high-peak-power self-pulses.As soon as the pulse peak power exceeds the thresh-old for nonlinear effects, short narrow pulses of highpeak power appear in the output. Under feedback,

these pulses pass through the population-invertedportion of the fiber, and thereby extract most ofthe stored energy to emit high-peak-power pulses.The appearance of short narrow pulses due to non-linear effects is also confirmed by the presence ofeither a SBS or a SRS signal in the output spectrum.When a long fiber length with a broadband mirrorwas used, a SRS peak shifted by about 56nm appearsin the output [Fig. 9], and when a narrow linewidthFBG mirror was used, broadening of the signal spec-trum due to SBS appears in the output [Fig. 7(d)]. Asmall broadening of signal spectrum due to SBS to-ward the longer side of the spectrum is in concur-rence with the report from Salhi et al. [12], wherethey also observed evidence of SBS in the generationof self-pulses. Although, in the case of a FBG mirrorwith a spliced passive fiber, the calculated value ofthe threshold for SRS is lower than the SBS thresh-old (Table 1), no SRS peak was observed. This may bebecause, under experimental conditions, the SRSthreshold might be higher than the SBS threshold,because the calculated figures are based on approx-imate analytical expressions.

It is evident from these results that the root causesfor the onset of self-pulsing are relaxation oscilla-tions at low pump powers and saturable absorption

Fig. 7. (Color online) Output spectrum in the case of an 8m long Yb-doped double-clad fiber with variation in input pump power for thecases of (a) a broadband mirror, (b) a broadband mirror and a passive fiber, (c) a FBG, and (d) a FBG and passive fiber.


along the weakly pumped portion of the fiber,whereas nonlinear SBS and SRS effects are respon-sible for increasing the peak power of these relaxa-tion oscillation pulses. In fiber lasers with single-endor double-end pumping (with wide or narrow laseroscillation bandwidths), most of the pump power isabsorbed near the pump end of the doped fiber and,hence, there is always a weakly pumped region—atthe farther end of the fiber length in the case ofsingle-end pumping, and at the middle portion ofthe fiber length in the case of double-end pumping.The presence of a weakly pumped region results insignal reabsorption along the fiber length, andconsequent occurrence of self-pulsation due to the“saturable absorption effect,” as in passively Q-switched lasers. However, due to the “distributednature” of saturable absorption of the signal alongthe fiber length, self-pulsing is random, with varia-tions in the frequency of self-pulses and the separa-tion between them in the time domain. With anincrease in pump power, the frequency of theseself-pulses increases, which is evident from the ob-served sequence of self-pulses under different condi-tions, and is a characteristic of passively Q-switched

lasers. With a further increase in pump power, gainuniformity increases, with a corresponding reductionin saturable absorption of the signal along the fiber

Fig. 8. Self-pulses in the case of an 18m long Yb-doped double-clad fiber, using a broadbandmirror without a passive fiber, at input pumppowers of (a) 2.77, (b) 4.36, (c) 7.56, and (d) 10:25W.

Fig. 9. (Color online) Output spectrum in the case of an 18m longYb-doped double-clad fiber with variation in input pump powerwhen the broadband mirror is used for feedback.


(due to absence of a weakly pumped region); thiseliminates self-pulsation at higher pump inputs.Passively mode-locked pulses with a characteristicseparation equal to the round-trip time may also ap-pear in the self-pulse envelope, depending on the sig-nal intensity and excited state absorption of thesignal in the doped fiber, as in bulk “passively Q-switched with simultaneous mode-locked lasers.”Such mode-locked pulses were not observed in ourexperiments.Since the lifetime of photons in the cavity plays an

important role in the dynamics of relaxation oscilla-tions and, consequently, the self-pulsing dynamics,we consider the average lifetime of photons τc in theresonator, which is given by the round-trip time ofthe photons in the resonator tr divided by the frac-tional loss ε per round trip [31]:

τc ¼trε ¼ 2nL

cð− lnR1R2 þ 2αlÞ ; ð4Þ

where L is the length of the resonator, l is the lengthof the active medium, c is the speed of light in va-cuum, n is the refractive index of the medium, R1and R2 are the reflectivities of the resonator mirrors,and α is the scattering loss. When the total length ofthe resonator is increased by splicing a passive fiberto the doped fiber, the lifetime of the photons in thecavity increases. In our case, R1 ¼ 1, R2 ¼ 0:04, and,with α ¼ 5 × 10−3 m−1 in Eq. (4), the photon lifetimefor the 8m length of fiber is about 23ns, and that forthe 88m length of the fiber is about 256ns. Thus,with an increase in fiber length, there is a significantincrease in the photon lifetime, which can modify re-laxation oscillation dynamics. Now we see that, inthe case of the FBG mirror without a spliced fiber,self-pulses disappear beyond 14:4W, whereas, withthe spliced fiber, self-pulses disappear beyond9:67W. We can further observe that, in the case ofbroadband mirror, self-pulses are completely elimi-nated when an 80m passive fiber is spliced, while,in the case of the FBG mirror, self-pulses still appearin a narrow range of pump power as compared to thatwithout the passive fiber. This may be due to a de-crease in the photon lifetime because of an increasein loss in the resonator with one more splice and the

narrow linewidth of the FBG mirror. Thus, a longerlength of passive fiber may be needed to completelyeliminate self-pulsing in the case of the FBG mirror,also. This is in agreement with the results reportedby Guan and Marciante [25], where they also ob-served a decrease in the pump power range forself-pulsing with an increase in the length of splicedpassive fiber, and finally a complete elimination ofself-pulsing was observed when a length of 2329mof passive fiber was used. Further, Hideur et al. [3]also reported a reduction in self-pulsing when theyused a good cavity (R1 ¼ 100%, R2 ¼ 80%), as com-pared to that with a bad cavity (R1 ¼ 100%,R2 ¼ 4%). This can also be explained by an increasein the photon lifetime when a good cavity configura-tion is used as compared to that with a bad cavityconfiguration and, hence, a corresponding decreasein output modulations due to relaxation oscillations.

When the length of the active fiber is short, self-pulses are initiated by relaxation oscillations atlow pump power, and as soon as the pumping ratebecomes sufficient to replenish the population inver-sion decay below the threshold level (due to lowerlifetime of photons in the cavity), self-pulses areeliminated. When the fiber length is long, self-pulsesare still initiated by relaxation oscillations; however,the saturable absorption effect in the weakly pumpedportion of the fiber also plays a role. When the peakpower of these self-pulses due to the saturable ab-sorption effect increases to reach the SBS or theSRS threshold, giant peak-power self-pulses appearin the output. Thus, complete elimination of self-pulsing requires either removal of relaxation oscilla-tions by increasing the photon lifetime, or increasingof the pump power to replenish gain and minimizethe saturable absorption effect by having gain unifor-mity along the fiber length.

4. Conclusion

We have carried out an experimental study to showthe effect of fiber length and laser linewidth onself-pulsing dynamics and output stabilization of asingle-mode Yb-doped double-clad CW fiber laser.It is found that onset of self-pulsing under low-levelpumping conditions is due to relaxation oscillationsand the saturable absorption effect in the weakly

Table 1. Calculated Values of Signal Threshold for Stimulated Brillouin Scattering and Stimulated Raman Scattering under Different ResonatorConfigurations

S. No. Resonator Configuration Signal Wavelength Laser Linewidth SBS Threshold SRS Threshold

1. 8m Yb-doped fiber, without spliced passive fiber,and broadband mirror

1093nm 6nm 52:7kW 400W

2. 8m Yb-doped fiber, with spliced 80m passive fiber,and broadband mirror

1088nm 6nm 5:85kW 44W

3. 8m Yb-doped fiber, without spliced passive fiber,and FBG mirror

1093nm 0:2nm 1:76kW 400W

4. 8m Yb-doped fiber, with spliced 80m passive fiber,and FBG mirror

1093nm 0:2nm 193W 44W

5. 18m Yb-doped fiber, without spliced passive fiber,and broadband mirror

1093nm 6nm 96:2kW 730W


pumped region of the doped fiber, irrespective of fiberlength and laser linewidth. However, with an in-crease in pump power, depending on the fiber lengthand the laser linewidth, the relaxation oscillationpulses get amplified and give rise to short-durationgiant pulses due to SBS or SRS. In the case where abroadband mirror is used and the length of the fiberis sufficient to reach the SRS threshold, the giantself-pulses are generated by SRS; on the other hand,where a narrow linewidth FBG mirror is used withsufficient fiber length to reach the SBS threshold, thegiant self-pulses are generated by SBS. Output sta-bilization and, hence, elimination of self-pulsations,can be achieved by suppression of relaxation oscilla-tions either by adding an appropriate length of a pas-sive fiber to sufficiently increase the photon lifetimein the resonator, or by increasing the pump powerand gain uniformity along the fiber to a level suchthat relaxation oscillations and reabsorption effectsare suppressed.

References1. Y. Jeong, J. K. Sahu, D. N. Payne, and J. Nilsson, “Ytterbium-

doped large-core fiber laser with 1:36kW continuous-waveoutput power,” Opt. Express 12, 6088–6092 (2004).

2. http//www.ipgphotonics.com.3. A. Hideur, T. Chartier, C. Ozkul, and F. Sanchez, “Dynamics

and stabilization of a high power side-pumped Yb-doped dou-ble-clad fiber laser,” Opt. Commun. 186, 311–317 (2000).

4. P. Glas, M. Naumann, A. Schirrmacher, L. Daweritz, andR. Hey, “Self pulsing versus self locking in a cw pumped neo-dymium doped double clad fiber laser,” Opt. Commun. 161,345–358 (1999).

5. E. Lacot, F. Stoeckel, and M. Chenevier, “Dynamics of an er-bium-doped fiber laser,” Phys. Rev. A 49, 3997–4008 (1994).

6. S. D. Jackson and T. A. King, “Dynamics of the output of heav-ily Tm-doped double-clad silica fiber lasers,” J. Opt. Soc. Am. B16, 2178–2188 (1999).

7. B. Ortac, A. Hideur, T. Chartier, M. Brunel, G. Martel, M.Salhi, and F. Sanchez, “Influence of cavity losses on stimu-lated Brillouin scattering in a self-pulsing side-pumped ytter-bium-doped double-clad fiber laser,” Opt. Commun. 215,389–395 (2003).

8. A. Martinez-Rios, I. Torres-Gomez, G. Anzueto-Sanchez, andR. Selvas-Aguilar, “Self-pulsing in a double-clad ytterbium fi-ber laser induced by high scattering loss,” Opt. Commun. 281,663–667 (2008).

9. R. Rangel-Rojo and M. Mohebi, “Study of the onset of self-pulsing behaviour in an Er-doped fiber laser,” Opt. Commun.137, 98–102 (1997).

10. Y. H. Tsang, T. A. King, D. K. Ko, and J. Lee, “Output dynamicsand stabilization of a multi-mode double-clad Yb-doped silicafiber laser,” Opt. Commun. 259, 236–241 (2006).

11. F. Brunet, Y. Taillon, P. Galarneau, and S. LaRochelle, “A sim-ple model describing both self-mode locking and sustainedself-pulsing in ytterbium-doped ring fiber lasers,” J. Light-wave Technol. 23, 2131–2138 (2005).

12. M. Salhi, A. Hideur, T. Chartier, M. Brunel, G. Martel,C. Ozkul, and F. Sanchez, “Evidence of Brillouin scatteringin an ytterbium-doped double-clad fiber laser,” Opt. Lett.27, 1294–1296 (2002).

13. A. A. Fotiadi, P. Mégret, and M. Blondel, “Dynamics of aself-Q-switched fiber laser with a Rayleigh-stimulatedBrillouin scattering ring mirror,” Opt. Lett. 29, 1078–1080(2004).

14. H. Takara, S. Kawanishi, and M. Saruwatari, “Stabilizationof a mode-locked Er-doped fiber laser by suppressing the re-laxation oscillation frequency component,” Electron. Lett. 31,292–293 (1995).

15. F. Sanchez, P. Le Boudec, P.-L. François, and G. Stephan,“Effects of ion pairs on the dynamics of erbium-doped fiber la-sers,” Phys. Rev. A 48, 2220–2229 (1993).

16. A. V. Kir’yanov, Y. O. Barmenkov, and I. L. Martinez, “Coop-erative luminescence and absorption in ytterbium-doped sili-ca fiber and the fiber nonlinear transmission coefficient atλ ¼ 980nm with a regard to the ytterbium ion-pairs’ effect,”Opt. Express 14, 3981–3992 (2006).

17. S. D. Jackson, “Direct evidence for laser re-absorption as in-itial cause for self-pulsing in three-level fiber lasers,”Electron.Lett. 38, 1640–1642 (2002).

18. D. Marcuse, “Pulsing behavior of a three-level laser withsaturable absorber,” IEEE J. Quantum Electron. 29,2390–2396 (1993).

19. A. Hideur, T. Chartier, and C. Özkul, “All-fiber tunableytterbium-doped double-clad fiber ring laser,” Opt. Lett. 26,1054–1057 (2001).

20. J. Li, K. Ueda, M. Musha, and A. Shirakawa, “Residual pumplight as a probe of self-pulsing instability in an ytterbium-doped fiber laser,” Opt. Lett. 31, 1450–1452 (2006).

21. Y. O. Barmenkov and A. V. Kir’yanov, “Pump noise as thesource of self-modulation and self-pulsing in erbium fiberlaser,” Opt. Express 12, 3171–3177 (2004).

22. L. Luo and P. L. Chu, “Suppression of self-pulsing inan erbium-doped fiber laser,” Opt. Lett. 22, 1174–1176(1997).

23. W. H. Loh, and J. P. de Sandro, “Suppression of self-pulsingbehavior in erbium-doped fiber lasers with resonant pumping:experimental results,” Opt. Lett. 21, 1475–1477 (1996).

24. V. Mizrahi, D. J. DiGiovanni, R. M. Atkins, S. G. Grubb,Y.-K. Park, and J.-M. P. Delavaux, “Stable single-mode erbiumfiber-grating laser for digital communication,” J. LightwaveTechnol. 11, 2021–2025 (1993).

25. W. Guan and J. R. Marciante, “Complete elimination of self-pulsations in dual-clad ytterbium-doped fiber lasers at allpumping levels,” Opt. Lett. 34, 815–817 (2009).

26. H. Chen, G. Zhu, N. K. Dutta, and K. Dreyer, “Suppression ofself-pulsing behavior in erbium-doped fiber lasers with asemiconductor optical amplifier,” Appl. Opt. 41, 3511–3516(2002).

27. A. Suzuki, Y. Takahashi, M. Yoshida, and M. Nakazawa, “Anultra low noise and narrow linewidth λ=4-shifted DFBEr-doped fiber laser with a ring cavity configuration,” IEEEPhotonics Technol. Lett. 19, 1463–1465 (2007).

28. B. N. Upadhyaya, U. Chakravarty, A. Kuruvilla, A. K. Nath,M. R. Shenoy, and K. Thyagarajan, “Effect of steady-stateconditions on self-pulsing characteristics of Yb-doped cw fiberlasers,” Opt. Commun. 281, 146–153 (2008).

29. R. G. Smith, “Optical power handling capacity of low lossoptical fibers as determined by stimulated Raman andBrillouin scattering,” Appl. Opt. 11, 2489–2494 (1972).

30. G. P. Agrawal, Nonlinear Fiber Optics, 2nd ed. (Academic,1995).

31. W. Koechner, Solid State Laser Engineering, 5th ed.(Springer, 1999).


http//www.ipgphotonics.com



Effect of laser linewidth and fiber length on self-pulsing dynamics and output stabilization of...

Documents

Transcript of Effect of laser linewidth and fiber length on self-pulsing dynamics and output stabilization of...