All-fiber wavelength-tunable picosecond nonlinear reflectivity measurement setup for characterization of semiconductor saturable absorber mirrors

All-fiber wavelength-tunable picosecond nonlinear reflectivity measurement setup for characterization of semiconductor saturable absorber mirrors

Optical Fiber Technology 31 (2016) 74–82 Contents lists available at ScienceDirect Optical Fiber Technology www.elsevier.com/locate/yofte Regular A...

3MB Sizes 0 Downloads 49 Views

Optical Fiber Technology 31 (2016) 74–82

Contents lists available at ScienceDirect

Optical Fiber Technology www.elsevier.com/locate/yofte

Regular Articles

All-fiber wavelength-tunable picosecond nonlinear reflectivity measurement setup for characterization of semiconductor saturable absorber mirrors K. Viskontas ⇑, N. Rusteika Center for Physical Sciences and Technology, Savanoriu ave. 231, LT-02300 Vilnius, Lithuania Ekspla UAB, Savanoriu ave. 237, LT-02300 Vilnius, Lithuania

a r t i c l e

i n f o

Article history: Received 11 February 2016 Revised 13 June 2016 Accepted 15 June 2016

Keywords: Ultra-short pulse Saturable absorber Nonlinear optical properties Mode-locked fiber laser

a b s t r a c t Semiconductor saturable absorber mirror (SESAM) is the key component for many passively mode-locked ultrafast laser sources. Particular set of nonlinear parameters is required to achieve self-starting modelocking or avoid undesirable q-switch mode-locking for the ultra-short pulse laser. In this paper, we introduce a novel all-fiber wavelength-tunable picosecond pulse duration setup for the measurement of nonlinear properties of saturable absorber mirrors at around 1 lm center wavelength. The main advantage of an all-fiber configuration is the simplicity of measuring the fiber-integrated or fiberpigtailed saturable absorbers. A tunable picosecond fiber laser enables to investigate the nonlinear parameters at different wavelengths in ultrafast regime. To verify the capability of the setup, nonlinear parameters for different SESAMs with low and high modulation depth were measured. In the operating wavelength range 1020–1074 nm, <1% absolute nonlinear reflectivity accuracy was demonstrated. Achieved fluence range was from 100 nJ/cm2 to 2 mJ/cm2 with corresponding intensity from 10 kW/cm2 to 300 MW/cm2. Ó 2016 Elsevier Inc. All rights reserved.

1. Introduction Ultrashort pulse laser is an important research tool adapted in many scientific laboratories as well as industrial settings [1–3]. Mode-locking technique is exclusively applied to achieve ultrashort (<10 ps) pulse duration from the lasers. The peak-powerdependent transmission element with a sufficiently fast response can lock the longitudinal modes of the laser’s resonator, which enables ultrashort pulse generation [4–7]. There are two main methods to obtain mode-locking operation – active and passive. Active technique usually involves an electro-optic or acoustooptic device which periodically modulates the losses of the resonator. By adjusting the frequency of the modulator close to the pulse repetition rate of the resonator it is possible to attenuate the trailing edges of the pulse after each round trip and shorten the pulse duration [8]. But when the pulse duration becomes sufficiently short, other limiting effects become dominant (e.g. chromatic dispersion or limited gain bandwidth) which restricts further pulse shortening [8]. Therefore, with this technique it is ⇑ Corresponding author at: Center for Physical Sciences and Technology, Savanoriu ave. 231, LT-02300 Vilnius, Lithuania. E-mail address: [email protected] (K. Viskontas). http://dx.doi.org/10.1016/j.yofte.2016.06.005 1068-5200/Ó 2016 Elsevier Inc. All rights reserved.

not possible to reach the pulse duration shorter then few picoseconds without an external pulse compression [9]. Also, the stable synchronization of actively mode-locked laser is very tricky to achieve [10]. Passive mode-locking utilizes saturable absorber element which has a peak-intensity dependent loss and no external modulator signal is required. With this technique it is possible to get the shortest [11], highest energy pulses [12] and highest average power [13] directly from the laser oscillator. Simplicity, robustness and low-noise performance of passive saturable absorber are the reasons that currently the absolute majority of commercial ultrafast lasers are built using passive mode-locking. When initially introduced in 1966 [14] passive mode-locking technique suffered from lack of proper materials and technology to fabricate saturable absorber with required properties. Saturable absorbers based on organic dyes were employed in ultrafast lasers for more than 20 years. But a passively mode-locked laser with an expensive, toxic and maintenance intensive absorber was as difficult to work with as an actively mode-locked laser. In 1980s the progress in the high-brightness pump diodes caused these pump sources to replace inefficient, bulky and expensive flash-lamps. The effort to have a robust and simple passive component in a laser resonator became even more important to further simplify and reduce the cost of the diode-pumped ultrafast lasers. The

K. Viskontas, N. Rusteika / Optical Fiber Technology 31 (2016) 74–82

breakthrough came with the invention of SESAM [15] and Kerrlens mode-locking (KLM) [16]. KLM is a technique based on an artificial saturable absorber when intensity of the short pulse in a laser gain medium is so high that it induces the self-focusing of a laser beam and by incorporating an aperture it is possible to modulate the losses of the resonator. This nonlinear optical self-focusing process is known as Kerr effect. Although, the KLM technique is adopted in many scientific laboratories and suitable to generate shortest broadly tunable ultra-short pulses [17], it has some major shortcomings. Firstly, Kerr mode-locking is usually realised near the stability boundary of a resonator and therefore is quite sensitive to environmental changes (temperature, mechanical perturbations). Secondly, with KLM it is difficult to achieve a reliable self-starting mode-locking and often some additional starting mechanism is required [18]. Reliable self-starting passive mode-locking was achieved by introducing SESAM [4–6,15]. This type of absorbers is widely used for more than 25 years [19]. The improved performance of the passively mode-locked solid-state lasers and further development of the SESAM devices expanded the application area of ultrafast lasers allowing them to be used not only in a scientific environment, but also in the industry. SESAM is made of different layers of semiconductor material stacked with sub-nanometer accuracy. By controlling the growth parameters of the semiconductor material and properly choosing the cavity design it is possible to tune all the relevant parameters (e.g. recovery time, saturation fluence, modulation depth, absorption wavelength) of a SESAM [6,20,21]. These parameters are particularly important to satisfy the stability and self-starting conditions of solid-state laser mode-locking [22]. Typically an absorbing part of a SESAM device is made of a very thin (<10 nm) semiconductor layer embedded in another semiconductor. Such structure is called a quantum well (QW) [23]. This thin layer is essential for the nonlinear optical properties of the SESAM. A semiconductor QW is a two dimensional structure in which quantum confinement plays a major role by enhancing effect of saturation of the absorption [24]. It is also possible to use other types of the low-dimensional structures instead of a semiconductor QWs. Although there are a lot of promising new nano-materials with good nonlinear optical properties (e.g. quantum dot absorbers [25], carbon nanotubes [26] and graphene [27]), a lot of work remains to be done to fabricate reliable and reproducible saturable absorbers of this type. In contrast, fabrication of SESAM devices with the molecular beam epitaxy (MBE) is highly precise and robust procedure [6]. Due to the rapid development of optical communication industry, especially of optical fiber and semiconductor laser diode technologies, fiber lasers emerged and began to expand into the field of ultrashort pulse lasers. These lasers have a number of desired qualities: ultrafast fiber lasers can be made environmentally stable, maintenance free, low cost and compact size [28]. Moreover, a small quantum defect in the Yb-doped fiber leads to high power efficiency and reduced thermal effects in the high output power lasers. Recent advances in Yb doped ultrafast fiber lasers also benefitted from SESAM technology [5]. Compared to Kerr type modelocking it has similar advantages as in solid state lasers: easy self-starting operation, flexibility of obtainable laser parameters (pulse duration, power etc.) and excellent environmental stability. The major obstacle remains degradation of semiconductor material in the SESAM’s structure due to high intensity of the incident radiation (hundreds MW/cm2 [29]. This problem is more severe in fiber lasers than in solid state lasers [30]. First, the nonlinearity which is necessary to achieve mode-locking in typical ultrashort pulse fiber lasers is of the order of 10% compared to 1% in solid state lasers [31]. This put constrains on the design of the SESAM and implies that non-saturable losses are also quite high. Second, material degradation of the SESAM in solid state laser is quite easily miti-

75

gated by moving an incident light spot (which is typically ten to hundred microns in diameter) to a different location on the SESAM chip. This in principal can be done in fiber lasers too. However the concept of monolithic fiber laser with a robust no-alignment operation is compromised when there is a need for movable SESAM. To complicate the problem further, there seems to be almost no scientific activity in researching for long lifetime SESAM architectures suitable for fiber lasers. Therefore in order to design and manufacture the SESAMs with required nonlinear properties comprehensive characterization techniques must be developed for experimental evaluation of the grown structures. In this work we set to develop set-up for spectral characterization of nonlinear properties of SESAMs and other saturable absorbers at around 1um wavelength, corresponding to usual window of operation of Yb-doped fiber lasers. 2. Methodology To reveal the most important parameters of the saturable absorber, nonlinear reflectivity measurement technique, where absorber reflectivity is measured as a function of incident optical intensity is often used [32]. Two regimes of saturation of absorption can be distinguished. First, when incident pulse duration is much longer than carrier relaxation time in the SESAM structure (typically 1–10 ps for mode-locking), saturation process can be well described as a function of intensity (measured in W/cm2). However such regime is rarely achieved in practical ultrafast lasers. Second, when incident pulse duration is much shorter compared to the carrier relaxation time, is well described in terms on incident energy fluence [33] (measured in J/cm2). In this case the saturation level does not depend on the pulse duration as the absorber remains saturated long time after the ultrashort pulse has passed. Using this approximation it is possible to define saturation fluence (Fsat), modulation depth (DR), low intensity reflectivity (R0), saturated reflectivity Rsat and nonsaturable losses (DRns). Fsat is the pulse fluence at which reflectivity of the absorber increases from the initial low intensity reflectivity R0 to 1/e level of the fully saturated nonlinear reflectivity Rsat. The modulation depth DR is the nonlinear change of the reflectivity. The nonsaturable losses DRns is the difference between the 100% reflectivity and the reflectivity Rsat. When measuring at the higher fluences, it is often necessary to include the inverse absorption coefficient F2. The F2 depends on two-photon absorption (TPA) and other nonlinear effects in the absorber material. For SESAMs, the chosen design may also affect an inverse absorption coefficient [34]. Fig. 1 depicts all the most important nonlinear parameters of typical saturable absorber. Logarithmic scale for the incident energy fluence is used as the nonlinear saturation takes place in a large range of fluence values (3 orders of magnitude). Generally, for a flat-top-shaped beam profile, the non-linear reflectivity of saturable absorber with respect to a pulse energy fluence (Fp) can be approximated by [29]:

RðF p Þ ¼ 1  DRns  DR

1  eðF p =F sat Þ  F p =F 2 : F p =F sat

ð1Þ

All static non-linear response parameters of saturable absorber can be found by fitting the experimental data to the Eq. (1). Some additional approximation must be used for beams with non-uniform intensity distribution. For the Gaussian-shaped beam, the equivalent incident fluence can be calculated:

Fp ¼

Ep ; Aeff

ð2Þ

here Ep is the pulse energy, Aeff is the effective area calculated from beam radius at the 1/e2 intensity level from the peak.

76

K. Viskontas, N. Rusteika / Optical Fiber Technology 31 (2016) 74–82

3. Experimental setup

Fig. 1. Nonlinear reflectivity of a typical SESAM device. The nonlinear reflectivity measured (data points) with respect to pulse fluence and fitted (solid line) to an Eq. (1). Using the fitting function it is possible to define all static nonlinear SESAM parameters: saturation fluence (Fsat), modulation depth (DR), low intensity reflectivity (R0), saturated reflectivity (Rsat) and non-saturable losses (DRns).

A figure of merit to describe the saturation level of the absorber is the saturation parameter S. It is defined as a ratio of the incident fluence to the saturation fluence of the SESAM:



Fp : F sat

ð3Þ

This parameter allows to compare the saturation level of the absorbers with different saturation fluences. It is particularly important in the degradation studies of saturable absorbers, where the saturation fluence of the absorber is modified during the exposure. Therefore, the saturation parameter also changes in time. Typically, the SESAMs in mode-locked fiber lasers are saturated in the range of S = 10–30 and in solid state lasers to S < 10 [33]. Nonlinear optical parameters play an important role in ultrashort pulse formation and stabilization process [6,26,33]. For instance, the saturation fluence of a saturable absorber needs to be sufficiently low with respect to the pulse fluence to initiate self-start mode-locking and avoid Q-switched mode-locking instabilities [22]. Other parameters usually depend on the laser type. For solid state lasers low modulation depth (1%) and low nonsaturable losses (1%) are required to reduce thermal heating in a high power lasers and avoid Q-switching [29]. For fiber lasers modulation depth of the absorber has to be much higher (of the order of 10%) due to strong amplified spontaneous emission (ASE) and typically large resonator losses [6] to ensure a self-starting and stable operation of the oscillator. For SESAMs, higher modulation depth is usually achieved by increasing the number of quantum wells (layers of the absorbing material). The lattice of QWs is not exactly matched to the GaAs lattice, which leads to higher nonsaturable losses and excessive mechanical strain within the device. This can make the SESAM more susceptible to faster degradation and optical damage [35]. Several methods to measure nonlinear reflectivity were suggested [36,37]. Most of these methods were realized using lockin amplifiers and solid state lasers. More recently, nonlinear reflectivity measurement system was introduced which used a fiber laser as an ultrashort optical pulse source [37]. All-in-fiber setup is a convenient tool to investigate passive saturable absorbers for fiber lasers as these devices are usually integrated with the optical fiber. Environmental stability and durability of an all-fiber setup also enables a possibility to track changes of absorber’s nonlinear parameters over a long period of time.

The experimental setup for characterization of saturable absorbers with main optical elements is shown in Fig. 2. The main blocks of the setup were a fiber oscillator, a preamplifier, a variable attenuator, a signal detection chain with two photodiodes and a saturable absorber (sample). Setup is compatible with absorbers in different configurations: fiber integrated, butt-coupled or conventional free space (bulk) type. For the reference, high reflectivity mirror was used to calibrate the system, e.g. to eliminate the nonlinearities of the photodiodes and to identify the absolute value of the reflectivity. Ultrashort optical pulses for the setup were generated from a SESAM mode-locked picosecond fiber oscillator with tunable center wavelength. The main components of the oscillator were a chirped fiber Bragg grating (CFBG) for dispersion control, an ytterbium-doped polarization maintaining (PM) single-mode fiber (YDFO) as a gain medium, a fiber coupled tunable wavelength filter (TWF), a micro-optical fiber pigtailed beam splitter (PBSO) for pulse extraction from the resonator and a SESAM for the initiation of mode-locking. The chip of the SESAM was butt-coupled to a FC/ PC optical connector and acted as the end-mirror of the resonator. We used the InGaAs quantum well (QW) SESAM with a high modulation depth (DR > 25%) in a broad spectral range. High modulation depth was particularly important to satisfy the self-start mode-locking conditions in a wavelength tunable fiber oscillator with the high losses (>10 dB) in the resonator. The oscillator was pumped by a laser diode (LD) with the center wavelength of 976 nm. Spectral characteristics of CFBG, SESAM and TWF components which define tunability of the picosecond oscillator are presented in Fig. 3. When pumped at a constant 145 mW power the oscillator generated nearly transform limited optical pulses with the duration in the range of 5–9 ps with a repetition rate of around 32 MHz, depending on the center wavelength. A tuning range of the oscillator was 1020–1074 nm with spectral bandwidth at a fixed wavelength around 0.2 nm. The main limitation of the tuning range in this realization was due to a limited bandwidth of the CFBG. However, it would be difficult to extend this range considerably because amplification bandwidth of Yb doped fiber would likely cut-off the spectrum below 1010 nm and above 1100 nm [38,39]. The output spectra from the oscillator across the tuning range are shown in Fig. 4(a) along with the repetition rate of the laser. The spectral bandwidth varies in the range 0.17–0.32 nm. The shape of the output pulse spectra remains Gaussian across the entire tuning range. The repetition rate of the laser decreases slightly when the center wavelength is increased. This effect can be attributed to the chirp of the fiber Bragg grating as the different wavelengths are reflected at the different positions of the grating changing the effective length of the resonator. A good match of the measured pulse repetition rate to the calculated one using fixed dispersion value of CFBG (20 ps/nm, provided by the manufacturer) verify this statement (Fig. 4(a)). Anomalous total cavity dispersion along with Kerr nonlinearity induced SPM allows to achieve soliton mode-locking regime [40]. The CFBG induced dispersion also defines the duration of a quasi-soliton pulse [41]. The autocorrelation traces and a close up spectrum of the picosecond pulses measured at 1030 nm wavelength are shown in Fig. 4(b). Kelly sidebands in the spectrum indicate quasi-soliton regime of the oscillator. At the mode-lock threshold the Gaussian-shaped pulses are almost transform limited with the time bandwidth product TBP  0.43 and marginal Kelly sidebands. The increase of pump power results in higher output pulse energy which increases the amplitude of Kelly sidebands and TBP from 0.43 to 0.49.

K. Viskontas, N. Rusteika / Optical Fiber Technology 31 (2016) 74–82

77

Fig. 2. All-fiber setup for measurement of nonlinear reflectivity of saturable absorber mirror. The oscillator and preamplifier are highlighted by a grey rectangular box. The output of the laser diode was split by fiber beam splitter (BS): 30% to the oscillator and 70% to the preamplifier. PF is a laser diode protection filter. Outside the grey rectangular box is the nonlinear reflectivity measurement system: VFOT – variable fiber optical attenuator, PBS – fiber micro-optic beam splitter (90% to the sample and 10% to the reference channel), Ph1 – photodiode for the reference beam, Ph2 – photodiode for the beam reflected from the sample. Two types of sample coupling are depicted inside the red ellipse: butt-coupling to the fiber and free-space coupling. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3. Reflectivity spectra (left scale) of a CFBG (black solid line) and a SESAM (blue solid line/blue dashed line). Transmittance spectrum (right scale) of a TWF (red solid line) at the 1040 nm wavelength position. Dotted lines indicate a tuning range of an oscillator. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

The pulse duration of the soliton mode-locked fiber oscillator depends on the total dispersion and nonlinearity of the resonator [42]. The pulse duration at different wavelengths measured with autocorrelator just above the mode-locking threshold is shown in Fig. 5(a) along with the output pulse energy of the oscillator. The

close to threshold pulse duration varies from 6.6 ps to 9.6 ps and increases almost linearly from 1020 nm to 1074 nm wavelength. The pulse duration is an important parameter for the measurements of nonlinearity and should be taken into account when measuring the nonlinear reflectivity of an absorber as it influences the inverse absorption coefficient F2 [34] which may decrease the modulation depth. At the mode-locking threshold the output pulse energy also linearly increases from 18 pJ to 28 pJ changing the center wavelength (Fig. 5(a)). However, to overcome the CFBG induced losses, the pump power must be increased sharply at the edges of operating spectral range (Fig. 5(b)). For the setup self-starting and stable mode-locking of the oscillator with a single pulse in the resonator is preferred. This regime was achieved with the pump power fixed at 145 mW in all tuning range. To verify the pulse stability of the tunable oscillator the radio frequency (RF) spectrum [43] was measured near the spectral edges of CFBG reflectivity at 1030 nm and 1070 nm wavelengths. The RF spectra, measured with the resolution of 10 Hz and the span of 30 kHz are shown in Fig. 6. At 1070 nm wavelength the RF spectra give a signal-to-noise ratio of >70 dB which is slightly higher than >65 dB ratio for 1030 nm wavelength. High signal-to-noise ratio show good mode-locking stability and low amplitude of pulse fluctuations [43]. At a fixed 145 mW pump power the oscillator produced up to 2 mW of average output power at 32 MHz with the maximum pulse energy in a range of few tens of pJ. For the nonlinear reflectivity setup this pulse energy was too low to completely saturate the absorber and therefore the preamplifier section was necessary. Picosecond pulses from the fiber oscillator were coupled with pump light by a wavelength division multiplexer/optical isolator hybrid (WIDM) and then amplified in ytterbium-doped fiber. The isolator integrated in a WIDM was necessary to avoid the output

78

K. Viskontas, N. Rusteika / Optical Fiber Technology 31 (2016) 74–82

0.025

33.2

1020 nm ~ 1074 nm

0.020

32.8

0.015

32.4

0.010 32.0

0.005

Frequency (MHz)

Optical Intensity (a.u.)

Frequency Theoretical frequency (right scale)

31.6

0.000 1020

1030

1040

1050

1060

1070

1080

Wavelength (nm)

(a)

(b)

Fig. 4. Properties of a tunable oscillator at a fixed pump power of 145 mW. (a) Color solid lines (left scale): the spectra of the continuously tunable mode-locked fiber oscillator; data points (right scale): the frequency of an oscillator at the different wavelengths; red solid line: theoretical frequency, calculated using the fixed dispersion value (20 ps/nm) of CFBG. (b): the autocorrelation traces of pulses from the oscillator working at 1030 nm center wavelength. Color lines indicate measurements at the self-starting threshold (cyan), stable mode-locking threshold (magenta) and fixed pump power 145 mW (black). Inset is the spectra of the corresponding pulses. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 5. (a) Measured pulse energy (black data points) and pulse duration (blue data points) of a mode-locked fiber oscillator for the different center wavelengths at the modelock threshold; (b) pump power at mode-lock threshold. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

power instabilities, arising from the back reflected light from the sample. The oscillator output was amplified to 50 mW of average power using 323 mW of pump power. Maximal pulse energy of 1.5 nJ was reached at 1040 nm (Fig. 7(a)). At this pulse energy, achievable energy fluence was 4000 lJ/cm2 assuming a 6.6 lm beam diameter (equal to mode field diameter of a single mode fiber) on the sample. For comparison, typical saturation fluence of the SESAM is 10–100 lJ/cm2. To verify the reliability of the nonlinear reflectivity measurement system the long-term optical output power measurements of the oscillator and preamplifier were performed at 1064 nm center wavelength. The stability of a preamplifier is shown in Fig. 7(b). The long-term (>14 h) peak to peak fluctuation of an output power of the preamplifier was 0.4% and the short-term (<20 min) peak to peak power fluctuations were only 0.06%. For the setup the short-

term stability was more important as the measurement of nonlinear reflectivity could be done in few minutes. The output pulses from the preamplifier were used for the nonlinear reflectivity measurements. In order to change the optical power (fluence) on the sample, the variable fiber optical attenuator (VFOT) was connected to the preamplifier output by PC/APC connector (see Fig. 2). Finally a micro-optical fiber pigtailed beam splitter with a 90/10 splitting ratio was spliced in between a VFOT and sample. The splitter provided three outputs – to the sample (90% port3), to the reference detector (10% port1) and to the reflection channel detector (port2). To detect the signals at the reflection and reference channels two photodiode power sensors (Thorlabs S130C) with a dual-channel power meter console (Thorlabs PM320E) were used. In order to determine the absolute value of sample reflectivity, fiber pigtailed mirror with the insertion loss

K. Viskontas, N. Rusteika / Optical Fiber Technology 31 (2016) 74–82

Fig. 6. Fundamental RF spectra of an oscillator output power at the fixed 145 mW pump power for two different wavelengths.

of 0.8 dB was used as a reference to calibrate the system for the nonlinearity of photodiodes.

4. Measurement of nonlinear properties of SESAMs In the following experiments the samples were coupled with two different methods – simple butt-coupling (direct optical contact with a fiber) or free-space coupling using collimating and focusing optics. In both cases the beam size on the sample was 6.6 lm which corresponds to mode field diameter of the optical fiber. For a fixed 323 mW pump power maximum energy fluence and intensity delivered on the sample at different wavelengths are shown in Fig. 8. To reveal the most important advantages and disadvantages of the nonlinear reflectivity measurement system, SESAMs with low and high modulation depth and with two different coupling methods were measured. First, a SESAM with a low modulation depth (2.5%) and narrowband exciton absorption at 1073 nm was cou-

79

Fig. 8. Pulse fluences (black data points) and pulse intensity (blue data points) on a sample at the different center wavelengths and fixed pump power (323 mW) produced from the preamplifier. The fluence was determined from an Eq. (2). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

pled into the set-up with free-space optics. This type of semiconductor absorber mirror was the best choice to verify the capabilities of a setup to measure the small changes in reflectivity at different wavelengths. The reflectivity of both a SESAM and a calibration mirror with respect to the incident fluence are shown in Fig. 9(a). Although the standard deviation of the mirror reflectivity across the fluence range was relatively high 0.5%, the measurements had good repeatability indicating systematic nature of the scatter. The repeatability of a reflectivity measurement was better than 0.06% over the whole spectral range. By dividing the measured reflectivity of a saturable absorber from the reflectivity of the calibration mirror, resulting data could be fitted to the Eq. (1) with the RMS error only 0.04%. Using this fit, the determined modulation depth was DR = 2.0%, nonsaturable losses DRns = 15.5%, saturation fluence Fsat = 35 lJ/cm2 and inverse absorption parameter F2 = 9300 lJ/cm2. Measurements at other wavelengths were also performed with low modulation depth SESAM. The excitonic peak at 1070 nm

Fig. 7. (a) Measured pulse energy (black data points) and pulse duration (blue data points) of a fiber preamplifier at the different center wavelengths and fixed pump power for the oscillator (145 mW) and preamplifier (323 mW). (b) The long-term (<14 h) measurement of a preamplifier output power at the fixed 323 mW pump power. Dashed lines define the interval of peak to peak power fluctuation. Inset (b): the short-term (<20 min) measurement of a preamplifier output power at the same 323 mW pump power. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

80

K. Viskontas, N. Rusteika / Optical Fiber Technology 31 (2016) 74–82

Fig. 9. (a) Measured nonlinear reflectivity of a SESAM with low DR (black – data points, red curve – fit to an Eq. (1) and a fiber pigtailed dielectric mirror (magenta data points) with respect to the incident fluence. (b) Measured nonlinear reflectivity at the three different wavelengths – 1069 nm, 1064 nm and 1040 nm. Inset (b) depicts the linear reflectivity spectrum of the same SESAM. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

wavelength can be clearly seen from the low intensity reflectivity spectrum in the inset of Fig. 9(b). Measurements at three different wavelengths – 1069 nm (at the excitonic peak), 1064 nm (on the blue slope) and 1040 nm (out of excitonic absorption), showed the stark change of nonlinear parameters (see Fig. 9(b)). The highest modulation depth DR = 2.9% was for the 1069 nm wavelength, indicating highest nonlinear absorption near the excitonic maximum position. The saturation fluence at this wavelength was 30 lJ/cm2. At the 1040 nm wavelength the modulation depth slightly decreased to DR = 2.3%, but the saturation fluence increased fourfold to Fsat = 126 lJ/cm2 clearly showing the effect of the excitonic absorption. Usually, during the SESAM engineering, e.g. after the ion-bombardment or thermal annealing, the position of excitonic peak may change [20]. Therefore, the wavelength tunable all-fiber setup could be a useful tool to investigate the SESAM during the post-growth processing.

Later, the high modulation depth SESAM was butt-coupled to the optical fiber and its nonlinear reflectivity was measured at the different wavelengths (see Fig. 10(a)). The saturation fluence of the device only slightly varied (37–51 lJ/cm2) in all spectral range, showing the absence of excitonic absorption feature. On the other hand, modulation depth changed from 16% to 6% at 1044 nm and 1064 nm respectively, indicating resonant behavior of this SESAM structure. To have full spectral information, reflectivity spectra were measured at low (0.01 lJ/cm2) and high (1200 lJ/cm2) incident fluence (Fig. 10(b)). The measurement revealed the modulation depth of a SESAM within the broad spectral region with clear maximum (resonance) at 1040 nm. One of the advantages of all in fiber nonlinear reflectivity measurement setup compared to the free space system is the ability to measure the nonlinear reflectivity of the saturable absorber before and after using it in the laser resonator for mode-locking. A typical

Fig. 10. (a) Measured nonlinear reflectivity data for a SESAM with high DR at four different wavelengths – 1030 nm (black), 1044 nm (magenta), 1053 nm (cyan) and 1064 nm (dark yellow). Solid color curves: fit to Eq. (1). (b) Reflectivity spectrum of a SESAM, measured at a low fluence (light grey) and high fluence (dark grey). The corresponding modulation depth (black) was derived by subtracting the low fluence spectrum from high fluence spectrum. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

K. Viskontas, N. Rusteika / Optical Fiber Technology 31 (2016) 74–82

beam diameter on the SESAM is of the order of few to few tens of micrometers, therefore it is quite difficult to perform optical investigation of the same spot afterwards. Example of measuring SESAM’s nonlinear properties after its long term (5000 h) operation in the fiber laser resonator by using fiber pigtailed measurement system described in this work was provided in reference [29]. In that study, by directly measuring nonlinear properties of the same spot on the SESAM before and after laser operation, two modes of optical degradation were determined – catastrophic optical damage, mostly affecting linear reflectivity of the absorber, and slow reduction of nonlinear reflectivity. These two modes were not distinguished before when only linear SESAM reflectivity as an indication of the degradation was measured [28]. More detailed study investigating change of the nonlinear properties of the SESAM after different duration of the operation in the laser using setup described in this work is planned in the future. 5. Summary and conclusions To summarise, we constructed an all-fiber nonlinear reflectivity setup with the wavelength tunable picosecond fiber laser in the range 1020–1073 nm. The measurement accuracy was sufficient for investigating low (2%) modulation depth SESAM. Spectrally dependent nonlinear properties were determined for this structure and the presence on sharp excitonic absorption was confirmed in agreement with linear reflectivity measurements. Wavelength dependent saturation fluence was measured to change from 35 lJ/cm2 at excitonic peak to 126 lJ/cm2. In conclusion, the presented all-fiber tunable nonlinear reflectivity system is an accurate, stable and powerful tool to investigate saturable absorbers with different nonlinear properties. Acknowledgments

[12]

[13]

[14]

[15]

[16]

[17]

[18] [19]

[20]

[21]

[22]

[23] [24]

[25]

The work described here was financially supported by Research Council of Lithuania under contract number LAT-10/2016.

[26]

References

[27]

[1] A.H. Zewail, Femtochemistry: atomic-scale dynamics of the chemical bond using ultrafast lasers (Nobel lecture), Angew. Chem. Int. Ed. 39 (2000) 2586– 2631. 10.1002/1521-3773(20000804)39:15<2586::AID-ANIE2586>3.0.CO;2O. [2] A. Stolow, A.E. Bragg, D.M. Neumark, Femtosecond time-resolved photoelectron spectroscopy, Chem. Rev. 104 (2004) 1719–1757, http://dx. doi.org/10.1021/cr020683w. [3] J. Adamonis, N. Rusteika, R. Danilevicˇius, A. Krotkus, A compact terahertz burst emission system driven with 1 lm fiber laser, Opt. Commun. 293 (2013) 61– 64, http://dx.doi.org/10.1016/j.optcom.2012.11.100. [4] U. Keller, Recent developments in compact ultrafast lasers, Nature 424 (2003) 831–838, http://dx.doi.org/10.1038/nature01938. [5] O. Okhotnikov, A. Grudinin, M. Pessa, Ultra-fast fibre laser systems based on SESAM technology: new horizons and applications, New J. Phys. 6 (2004) 1–22, http://dx.doi.org/10.1088/1367-2630/6/1/177. [6] U. Keller, K.J. Weingarten, F.X. Kartner, D. Kopf, B. Braun, I.D. Jung, et al., Semiconductor saturable absorber mirrors (SESAM’s) for femtosecond to nanosecond pulse generation in solid-state lasers, IEEE J. Sel. Top. Quantum Electron. 2 (1996) 435–451, http://dx.doi.org/10.1109/2944.571743. [7] G. Steinmeyer, D. Sutter, L. Gallmann, N. Matuschek, U. Keller, Frontiers in ultrashort pulse generation: pushing the limits in linear and nonlinear optics, Science 286 (1999) 1507–1512, http://dx.doi.org/ 10.1126/science.286.5444.1507. [8] D. Kuizenga, A. Siegman, FM and AM mode locking of the homogeneous laser – Part I: theory, IEEE J. Quantum Electron. 6 (1970) 694–708, http://dx.doi.org/ 10.1109/JQE.1970.1076343. [9] U. Keller, K.D. Li, B.T. Khuri-Yakub, D.M. Bloom, K.J. Weingarten, D.C. Gerstenberger, High-frequency acousto-optic modelocker for picosecond pulse generation, Opt. Lett. 15 (1990) 45–47. [10] F. Zhou, G.P. Malcolm, A.I. Ferguson, 1-GHz repetition-rate frequencymodulation mode-locked neodymium lasers at 1.3 microm, Opt. Lett. 16 (1991) 1101–1103, http://dx.doi.org/10.1364/OL.16.001101. [11] R. Ell, U. Morgner, F.X. Kãârtner, J.G. Fujimoto, E.P. Ippen, V. Scheuer, et al., Generation of 5-fs pulses and octave-spanning spectra directly from a Ti:

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

81

sapphire laser, Opt. Lett. 26 (2001) 373–375, http://dx.doi.org/10.1364/ OL.26.000373. J. Neuhaus, D. Bauer, J. Zhang, A. Killi, J. Kleinbauer, M. Kumkar, et al., Subpicosecond thin-disk laser oscillator with pulse energies of up to 25.9 microjoules by use of an active multipass geometry, Opt. Express 16 (2008) 20530–20539, http://dx.doi.org/10.1364/OE.16.020530. C.R.E. Baer, C. Kränkel, C.J. Saraceno, O.H. Heckl, M. Golling, R. Peters, et al., Femtosecond thin-disk laser with 141 W of average power, Opt. Lett. 35 (2010) 2302–2304, http://dx.doi.org/10.1364/OL.35.002302. A.J. Demaria, D.A. Stetser, H. Heynau, Self mode-locking of lasers with saturable absorbers, Appl. Phys. Lett. 8 (1966) 174–176, http://dx.doi.org/ 10.1063/1.1754541. U. Keller, D.A. Miller, G.D. Boyd, T.H. Chiu, J.F. Ferguson, M.T. Asom, Solid-state low-loss intracavity saturable absorber for Nd:YLF lasers: an antiresonant semiconductor Fabry-Perot saturable absorber, Opt. Lett. 17 (1992) 505–507, http://dx.doi.org/10.1364/OL.17.000505. D.E. Spence, P.N. Kean, W. Sibbett, 60-fsec pulse generation from a self-modelocked Ti:sapphire laser, Opt. Lett. 16 (1991) 42–44, http://dx.doi.org/10.1364/ OL.16.000042. D.H. Sutter, G. Steinmeyer, L. Gallmann, N. Matuschek, F. Morier-Genoud, U. Keller, et al., Semiconductor saturable-absorber mirror assisted Kerr-lens mode-locked Ti:sapphire laser producing pulses in the two-cycle regime, Opt. Lett. 24 (1999) 631–633, http://dx.doi.org/10.1364/OL.24.000631. S. Chen, J. Wang, Self-starting issues of passive self-focusing mode locking, Opt. Lett. 16 (1991) 1689–1691, http://dx.doi.org/10.1364/OL.16.001689. U. Keller, Ultrafast solid-state laser oscillators: a success story for the last 20 years with no end in sight, Appl. Phys. B Lasers Opt. 100 (2010) 15–28, http://dx.doi.org/10.1007/s00340-010-4045-3. M.J. Lederer, V. Kolev, B. Luther-davies, H.H. Tan, C. Jagadish, Ion-implanted InGaAs single quantum well semiconductor saturable absorber mirrors for passive mode-locking, J. Phys. D Appl. Phys. 34 (2001) 2455–2464, http://dx. doi.org/10.1088/0022-3727/34/16/309. F. Schättiger, D. Bauer, J. Demsar, T. Dekorsy, J. Kleinbauer, D.H. Sutter, et al., Characterization of InGaAs and InGaAsN semiconductor saturable absorber mirrors for high-power mode-locked thin-disk lasers, Appl. Phys. B 106 (2011) 605–612, http://dx.doi.org/10.1007/s00340-011-4697-7. C. Hönninger, R. Paschotta, F. Morier-Genoud, M. Moser, U. Keller, Q-switching stability limits of continuous-wave passive mode locking, J. Opt. Soc. Am. B 16 (1999) 46, http://dx.doi.org/10.1364/JOSAB.16.000046. M. Jaros, Wave mechanics applied to semiconductor heterostructures, J. Mod. Opt. 38 (1991) 1211–1212, http://dx.doi.org/10.1080/09500349114551351. S. Schmitt-Rink, D.S. Chemla, D.A.B. Miller, Linear and nonlinear optical properties of semiconductor quantum wells, Adv. Phys. 38 (1989) 89–188, http://dx.doi.org/10.1080/00018738900101102. D.J.H.C. Maas, A.-R. Bellancourt, M. Hoffmann, B. Rudin, Y. Barbarin, M. Golling, et al., Growth parameter optimization for fast quantum dot SESAMs, Opt. Express 16 (2008) 18646–18656, http://dx.doi.org/10.1364/OE.16.018646. S. Kivistö, T. Hakulinen, A. Kaskela, B. Aitchison, D.P. Brown, A.G. Nasibulin, et al., Carbon nanotube films for ultrafast broadband technology, Opt. Express 17 (2009) 2358–2363, http://dx.doi.org/10.1364/OE.17.002358. Z. Sun, T. Hasan, F. Torrisi, D. Popa, G. Privitera, F. Wang, et al., Graphene modelocked ultrafast laser, ACS Nano 4 (2010) 803–810, http://dx.doi.org/10.1021/ nn901703e. M.E. Fermann, I. Hartl, Ultrafast fiber laser technology, IEEE J. Sel. Top. Quantum Electron. 15 (2009) 191–206, http://dx.doi.org/10.1109/ JSTQE.2008.2010246. C.J. Saraceno, C. Schriber, M. Mangold, M. Hoffmann, O.H. Heckl, C.R.E. Baer, et al., SESAMs for high-power oscillators: design guidelines and damage thresholds, IEEE J. Sel. Top. Quantum Electron. 18 (2012) 29–41, http://dx.doi. org/10.1109/JSTQE.2010.2092753. K. Viskontas, K. Regelskis, N. Rusteika, Slow and fast optical degradation of the SESAM for fiber laser mode-locking at 1 lm, Lith. J. Phys. 54 (2014) 127–135, http://dx.doi.org/10.3952/physics.v54i3.2951. R. Herda, O.G. Okhotnikov, Dispersion compensation-free fiber laser modelocked and stabilized by high-contrast saturable absorber mirror, IEEE J. Quantum Electron. 40 (2004) 893–899, http://dx.doi.org/10.1109/ JQE.2004.830194. M. Haiml, R. Grange, U. Keller, Optical characterization of semiconductor saturable absorbers, Appl. Phys. B 79 (2004) 331–339, http://dx.doi.org/ 10.1007/s00340-004-1535-1. R. Paschotta, U. Keller, Passive mode locking with slow saturable absorbers, Appl. Phys. B Lasers Opt. 73 (2001) 653–662, http://dx.doi.org/10.1007/ s003400100726. R. Grange, M. Haiml, R. Paschotta, G.J. Spühler, L. Krainer, M. Golling, et al., New regime of inverse saturable absorption for self-stabilizing passively mode-locked lasers, Appl. Phys. B Lasers Opt. 80 (2005) 151–158, http://dx.doi. org/10.1007/s00340-004-1622-3. K.L. Kavanagh, R.S. Goldman, C. Lavoie, B. Leduc, T. Pinnington, T. Tiedje, et al., In situ detection of misfit dislocations by light scattering, J. Cryst. Growth 174 (1997) 550–557, http://dx.doi.org/10.1016/S0022-0248(97)00032-8. D.J. Maas, B. Rudin, A.-R. Bellancourt, D. Iwaniuk, S.V. Marchese, T. Südmeyer, et al., High precision optical characterization of semiconductor saturable absorber mirrors, Opt. Express 16 (2008) 7571–7579, http://dx.doi.org/ 10.1364/OE.16.007571. L. Orsila, A. Härkönen, J. Hyyti, M. Guina, G. Steinmeyer, Ultrahigh precision nonlinear reflectivity measurement system for saturable absorber mirrors

82

K. Viskontas, N. Rusteika / Optical Fiber Technology 31 (2016) 74–82

with self-referenced fluence characterization, Opt. Lett. 39 (2014) 4384, http:// dx.doi.org/10.1364/OL.39.004384. [38] D.C. Hanna, R.M. Percival, I.R. Perry, R.G. Smart, P.J. Suni, A.C. Tropper, An Ytterbium-doped monomode fibre laser: broadly tunable operation from 1010 lm to 1162 lm and Three-level operation at 974 nm, J. Mod. Opt. 37 (1990) 37–41, http://dx.doi.org/10.1080/09500349014550601. [39] R. Paschotta, J. Nilsson, A.C. Tropper, D.C. Hanna, Ytterbium-doped fiber amplifiers, IEEE J. Quantum Electron. 33 (1997) 1049–1056, http://dx.doi.org/ 10.1109/3.594865. [40] L.F. Mollenauer, R.H. Stolen, J.P. Gordon, Experimental observation of picosecond pulse narrowing and solitons in optical fibers, Phys. Rev. Lett. 45 (1980) 1095–1098, http://dx.doi.org/10.1103/PhysRevLett. 45.1095.

[41] O. Katz, Y. Sintov, Y. Nafcha, Y. Glick, Passively mode-locked ytterbium fiber laser utilizing chirped-fiber-Bragg-gratings for dispersion control, Opt. Commun. 269 (2007) 156–165, http://dx.doi.org/10.1016/j. optcom.2006.07.034. [42] J.D. Kafka, T. Baer, D.W. Hall, Mode-locked erbium-doped fiber laser with soliton pulse shaping, Opt. Lett. 14 (1989) 1269–1271, http://dx.doi.org/ 10.1364/OL.14.001269. [43] D. von der Linde, Characterization of the noise in continuously operating mode-locked lasers, Appl. Phys. B 39 (1986) 201–217, http://dx.doi.org/ 10.1007/BF00697487.