Open Access
Add to BibSonomy
Abstract
Dual-comb lasers are a new class of ultrafast lasers that enable fast, accurate and sensitive measurements without any mechanical delay lines. Here, we demonstrate a 2-µm laser called MIXSEL (Modelocked Integrated eXternal-cavity Surface Emitting Laser), based on an optically pumped passively modelocked semiconductor thin disk laser. Using III-V semiconductor molecular beam epitaxy, we achieve a center wavelength in the shortwave infrared (SWIR) range by integrating InGaSb quantum well gain and saturable absorber layers onto a highly reflective mirror. The cavity setup consists of a linear straight configuration with the semiconductor MIXSEL chip at one end and an output coupler a few centimeters away, resulting in an optical comb spacing between 1 and 10 GHz. This gigahertz pulse repetition rate is ideal for ambient pressure gas spectroscopy and dual-comb measurements without requiring additional stabilization. In single-comb operation, we generate 1.5-ps pulses with an average output power of 28 mW, a pulse repetition rate of 4 GHz at a center wavelength of 2.035 µm. For dual-comb operation, we spatially multiplex the cavity using an inverted bisprism operated in transmission, achieving an adjustable pulse repetition rate difference estimated up to 4.4 MHz. The resulting heterodyne beat reveals a low-noise down-converted microwave frequency comb, facilitating coherent averaging.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Modelocked laser-based optical frequency combs [1–3] offer high-precision dual-comb applications in the fields of spectroscopy, ranging, safety, pharmaceuticals, and healthcare [4–7]. These dual-comb lasers enable fast, accurate and sensitive time-resolved measurements and Fourier transform spectroscopy, eliminating the need for mechanical delay lines that can limit measurement speed and scan range. However, conventional optical dual-comb systems, relying on two modelocked lasers with four active stabilization loops, face challenges in terms of cost, complexity and size, hindering the industrial realization of dual-comb applications. To overcome these limitations, we have developed a solution using polarization [8,9] and spatial [10] multiplexing within a single laser cavity, eliminating the need for stabilization in diode-pumped solid-state or semiconductor lasers. We validated dual-comb spectroscopy using such a free-running single-cavity dual-comb laser by measuring weak water absorption at 968 nm – at that time a paradigm shift in dual-comb spectroscopy [11].
For many applications there is a need for a dual-comb source in the shortwave infrared (SWIR) and mid-infrared (mid-IR) regime. In this paper we present the first demonstration of a MIXSEL (Modelocked Integrated eXternal-cavity Surface Emitting Lasers) capable of both single and a dual-comb operation at a 2-µm center wavelength. Initially demonstrated in the near-IR spectrum [12], the MIXSEL combines an optically pumped semiconductor laser with cost-effective multi-mode diode lasers for optical pumping and power scaling.
Compared to diode-pumped solid-state lasers, the MIXSEL offers a simplified approach to passive modelocking by integrating both gain medium and saturable absorber into a unified, high-reflective mirror chip. This design forms a compact linear laser cavity with an output coupler positioned a few centimeters away from the MIXSEL chip to facilitate gigahertz comb spacing. A MIXSEL advantage over edge-emitting semiconductor diode laser is low-noise passive modelocking enabled by diode pumping, a high-Q cavity, and a laser beam aligned perpendicular to the quantum well gain and absorber layers reducing the interaction with the carrier dynamics [13]. The MIXSEL surpasses diode-pumped solid-state lasers in wavelength flexibility, stable gigahertz pulse repetition rates without Q-switching instabilities, and a simpler cavity structure. Moreover, many applications require pulse repetition rates between 1 and 10 gigahertz. For a 3-GHz pulse repetition, a cavity length of approximately 5 cm in air is needed, posing challenges for further integration.
The 2-µm MIXSEL is realized using III-V semiconductor GaSb (gallium-antimonide) molecular beam epitaxy (MBE), incorporating 15 InGaSb quantum wells (QWs) for gain, a single InGaSb QW saturable absorber, and AlAsSb/GaSb and AlAsSb/AlGaAsSb distributed Bragg reflectors for high reflectivity at the lasing and the pump wavelengths, respectively. In single-comb operation, we achieved 1.5-ps pulses with an average output power of 28 mW at a pulse repetition rate of 4 GHz and a center wavelength of 2.035 µm. Operating the laser close to room temperature still allows for a similar optical-to-optical efficiency as for the near-IR MIXSELs. Unlike previous dual-comb MIXSELs, our approach for this 2-µm result utilizes an inverted biprism within the MIXSEL cavity to generate two spatially multiplexed combs. The pulse repetition rate difference $\Delta {f_{\textrm{rep}}}$ between the two modelocked optical combs is adjustable up to 4.4 MHz and obtained within a 4-GHz cavity length. This laser design is well-suited for dual-comb applications, as the sufficiently high repetition rate ${f_{\textrm{rep}}}$ prevents aliasing at fast measurement times of $\frac{1}{{\Delta {f_{\textrm{rep}}}}}$ with low noise. We demonstrate for a $\Delta {f_{\textrm{rep}}}$ = 595 kHz direct resolution of microwave comb lines and the potential for coherent averaging of interferograms, as previously shown with more complex diode-pumped solid-state lasers in the near-IR [14]. We take advantage of semiconductor bandgap engineering to provide versatility in the operation wavelength within the SWIR regime by employing type-I InGaSb quantum wells. Furthermore, the incorporation of type-II quantum wells holds promise for extending this laser technology into the mid-IR regime.
Since the initial demonstration of a near-IR MIXSEL using GaAs-based MBE in 2007 [11], significant progress has been made in generating shorter pulses [15,16], achieving higher power [17] and scaling the pulse repetition rate up to 100 GHz [18,19]. A major milestone was the demonstration of the first dual-comb MIXSEL in 2015, employing polarization multiplexing [8]. This single-cavity dual-comb MIXSEL produced a pair of modelocked pulse trains, enabling applications such as molecular spectroscopy [13,20] and LiDAR applications [21] in free-running operation. So far, all previous MIXSEL results have used GaAs-based semiconductor materials, resulting in near-IR emission around 1 µm. However, many dual-comb laser applications, especially molecular absorption spectroscopy could benefit from emission wavelengths above 2 µm. GaSb has emerged as the most suitable semiconductor material system for this wavelength range, with the successful demonstration of vertical external cavity surface emitting laser (VECSEL) and semiconductor saturable absorber mirror (SESAM) technology [22–24]. Recent advancements include upside-down processing, backside cooling, power scaling of InGaSb VECSELs [25,26] and passive modelocking with a SESAM at 2 µm [27,28]. In our work presented here, we integrate the SESAM and VECSEL into a single semiconductor wafer, significantly simplifying the cavity design and creating a monolithically grown semiconductor MIXSEL chip of only 10 µm thickness.
In recent years, significant advancements have been made in the development of single-cavity dual-comb lasers based on diode-pumped solid-state and semiconductor lasers in various configurations [6–10,29–31]. These lasers have found applications in pump-probe measurements [32–34], ultrasonics [35], LiDAR [21,36] and spectroscopy [13,14,20]. One key advantage of these lasers is that both modelocked pulse trains share the same cavity elements, resulting in reduced uncorrelated noise for the pulse repetition rate difference $\Delta {f_{\textrm{rep}}}$ [14]. Several single-cavity multiplexing techniques have been implemented in many different lasers based on polarization [8,9], wavelength [37,38] and circulation-direction [39,40], and spatial multiplexing techniques in both reflection [10,31] and transmission [14]. In comparison to other comb sources such as ion-doped solid-state lasers, fiber lasers [41,42], quantum cascade [43] lasers or micro resonator combs [44,45], the MIXSEL offers a simple architecture for the 1-10 GHz comb spacing regime, delivering high power per comb line, and low noise.
2. Results and discussion
2.1 GaSb-based MIXSEL chip design
Figure 1 illustrates the design and characterization of the 2-µm MIXSEL chip, grown in the FIRST Lab at ETH Zurich. The chip is optically pumped and comprises a 19.5-pair AlAs08Sb92/GaSb distributed Bragg reflector (DBR) designed for the lasing wavelength. Within the chip on top of this DBR there is a thick layer of GaSb with an embedded In27Ga73Sb saturable absorber QW of 9.3 nm thickness to achieve fast recombination dynamics [24]. A pump-DBR, consisting of 10.5 pairs of AlAs08Sb92/Al15Ga85AsSb follows the saturable absorber layers to reflect the incident pump and prevent saturation of the absorber QW. The resonant periodic gain region is grown on top of the pump-DBR and comprises 5 × 3 In27Ga73Sb QWs with a thickness of 8.5 nm each. The QWs are barrier pumped, with the majority of the pump light absorbed outside the QW, as illustrated by the band diagram in Fig. 1 inset. The Al15Ga85AsSb ensures that carriers are confined around each QW which increases carrier capture rates and photoluminescence (PL). The Al15Ga85AsSb layers do not absorb the pump but provide small refractive index contrast to the surrounding GaSb layers, which limits internal reflections from the active region interfaces and therefore improves group delay dispersion (GDD). The top section of the chip consists of non-pump-absorbing semiconductor layers (grown monolithically) and a plasma-enhanced chemical vapor deposition (PECVD) grown Si3N4 dielectric layer (grown separately after flip-chip processing). The refractive index of Si3N4 is matched to the refractive index of GaSb via ${n_{\textrm{Si}3\textrm{N}4}} \approx \sqrt {{n_{\textrm{GaSb}}}} $ so that it serves as antireflective layer. This section ensures a flat, close-to-zero GDD for the laser wavelength and increases the field-intensity enhancement in the structure. The thickness of the total structure is slightly more than 10 µm, with all bulk materials (i.e., layer thickness >10 nm) grown lattice matched to avoid strain-relaxation. The design accounts for different field-intensity enhancement in the gain and absorber QWs to achieve gain saturation at a higher fluence than absorber saturation [46,47] (i.e. 0.85 for the gain QWs and 1.22 for the absorber QW).
Fig. 1. Design and characterization of the InGaSb MIXSEL chip: a) Layer structure of the MIXSEL chip, depicting the refractive index across the epitaxial structure's thickness. The electric field intensities of the laser mode (red line) and pump mode (green line). The various sections are labeled as distributed Bragg reflector (DBR) and AR (antireflection). The right axis displays the simulated temperature for typical laser operation parameters. A cross-sectional SEM image of the upside-down grown structure is shown at the bottom. The bottom-right panel depicts the zoomed-in band edges of a quantum well (QW), with green arrow indicating pump absorption and red arrow indicating laser emission. b) Linear reflectance and photoluminescence (PL) of the processed chip. c) Measured group delay dispersion (GDD) without the dielectric Si3N4 top layer (blue). This measurement facilitates the correction of initial design (red) for growth errors (green dashed) and consequently achieves an overall flattened GDD in the presence of growth errors and with the dielectric top coating (purple, dotted).
To compensate for internal heating-induced wavelength shifts, the QW thickness is adjusted in the design. One-dimensional heat simulations at a typical pump power reveal a temperature difference of approximately 10 °C between gain and absorber QWs. Using the Varshni formula [48], we estimate the corresponding wavelength shift and incorporate it into the QW thickness variation in the design. The layer thickness is validated by comparing a scanning electron microscopy image of the cleaved facet to the design (Fig. 1(a), bottom).
Figure 1(b) illustrates the measured stop band of the high reflecting mirror, exhibiting a pronounced dip on the blue side corresponding to the absorption of the 15 gain QWs. The pump-DBR stopband around 1470 nm is also visible, with a peak reflectance of slightly below 70% and an estimated pump absorption of approximately 35% through the gain section. Photoluminescence (PL) can be observed, matching the strong absorption dip. During lasing operation, the emission wavelength of the QW undergoes a redshift of 20 nm due to pump-induced heating, aligning it with the center of the stopband. Note that the integrated absorber QW cannot be characterized inside the MIXSEL structure and therefore requires analysis in separate test growth runs without the gain section. For the integrated absorber, a modulation depth of 2.2% and a saturation fluence of 1.2 µJ/cm2 are expected based on previous SESAM structures [23].
Figure 1(c) displays the measured GDD of the MIXSEL chip before applying the dielectric Si3N4 top layer. Even with precise calibration, thickness variations during MBE growth of a single layer can reach up to 2%. This variation can lead to shifts in the DBR stopband, the quantum wells, and the pump-DBR relative to one another. While many of these characteristics are quite resilient to minor fluctuations, and thus do not significantly impact the overall results, the group delay dispersion (GDD) is sensitive to even minor growth deviations. Therefore, it requires meticulous attention during the MBE growth process.
By evaluating the spectral positions of the resonances, growth errors during MBE can be estimated and compensated for by adjusting the thickness of the dielectric layer. After correcting for the growth deviations in the design, the simulated GDD resonances (yellow solid line) closely match the measured ones (blue solid line). Note that in semiconductor lasers, the quasi-soliton modelocking mechanism requires a slightly positive and close-to-zero GDD for optimal and shortest pulse formation [49]. Following the application of the dielectric top layer, an overall flat and close-to-zero GDD (purple dotted line) is expected, enabling the generation of short pulses from the MIXSEL.
2.2 2-µm MIXSEL cavity configurations
Figure 2(a) shows the straight, linear MIXSEL cavity, measuring 3.9 cm in length, designed for single-comb operation. Optical pumping is achieved by using a 1450-nm, 15-W, fiber-coupled diode laser bar with an M2 of 94, that is focused onto the MIXSEL chip at an external angle of 45°. Two cylindrical lenses are employed for beam shaping, resulting in a round spot with a diameter 1/e2 of 360 µm. The curved output coupler with a transmission of 1.8% has a radius of curvature of 50 mm. The estimated intracavity mode size 1/e2 diameter on the MIXSEL chip is 240 µm. Fine-tuning of the GDD and achieving p-polarized output are accomplished by using YAG Brewster windows of different thickness, introducing –134 fs2/mm of round-trip GDD. Figure 2(b) displays a photograph of the single-comb MIXSEL cavity with the 1-mm YAG Brewster window. The MIXSEL chip, soldered on a copper piece, is mounted on a water-cooled, Peltier-stabilized heat sink for efficient thermal management.
Fig. 2. MIXSEL cavities: a) Schematics of the single-comb MIXSEL cavity, comprising the MIXSEL chip, Brewster window, and output coupler (OC) with radius of curvature (Roc) and intensity transmission (Toc). b) Photograph of the single-comb MIXSEL with the key elements. c) Schematics of the dual-comb MIXSEL cavity, where the pump is split to simultaneously invert two modes on the chip. An inverted biprism functions as a Brewster window and a spatial multiplexer. d) Photograph of the dual-comb MIXSEL with the key elements. e) Propagation of the two modes within the cavity showing side and top-down view. The shaded area represents the 1/e2 intensity in side view, according to the beam waist w in the final subpanel. Two resonant modes with an approximate separation of 2 mm on the MIXSEL chip propagate on top of each other in the top-down view.
For dual-comb operation, Fig. 2(c) and 2(d) present the same cavity configuration with the implementation of a special Brewster window, shaped as an inverted biprism. This inverted biprism, constructed from CaF2 substrate with an apex angle of inverted 6°, enables multiplexing (Fig. 2(e)). This special geometry deflects the intracavity modes passing through the top and bottom parts of the biprism differently and allows for control of the pulse repetition rate difference $\Delta {f_{\textrm{rep}}}$. By adjusting the transverse position of the biprism, $\Delta {f_{\textrm{rep}}}$ can be tuned between 0 and 4.4 MHz. The pump is equally split using a fiber-splitter and then focused on different positions on the chip, with a spot diameter of 210 µm, to avoid thermal cross-talk. The radius of curvature of the output coupler remains the same, but the output coupling rate is reduced to 1% compared to the single-comb MIXSEL configuration. Figure 2(d) depicts the biprism, while Fig. 2(e) provides detailed analysis of the biprism’s impact on the resonant cavity modes in a side view and a top-down view. The modes propagate on top of each other, with a spatial separation of approximately 2 mm on the chip. In the side view, the laser cavity mode size is indicated by the shaded area and shows that both modes are well separated on the chip, minimizing thermal cross-talk of the two modes. The top-down-view the biprism at Brewster’s angle, ensuring minimal losses and p-polarized output for both beams. The small lateral shift to the side needs to be taken into account when placing the output coupler. The modesize propagation demonstrates that both laser modes have nearly round shapes and are matched compared to the pump spot size on the MIXSEL chip.
2.3 Modelocking diagnostics of single-comb MIXSEL
Figure 3 shows the modelocking characterization of the single-comb operation. All the data was obtained at a chip heat-sink temperature of 15.7 °C, using a pump power of approximately 3 W and yielding output power of 28 mW. A 2-mm YAG Brewster window adds –269 fs2 of intracavity GDD.
Fig. 3. Modelocking diagnostics of the single-comb MIXSEL: a) Measured optical spectrum, exhibiting a center wavelength of 2035nm and a full width at half maximum (FWHM) of 4.1 nm. b) Measured autocorrelation with a sech2 fit, resulting in a pulse duration of 1.46 ps. The inset displays a wide scanning measurement. c) Microwave spectrum of the pulse repetition rate peak at 4 GHz, sampled with a resolution bandwidth (RBW) of 3 kHz. The inset exhibits a larger span of the microwave spectrum (RBW 100 kHz), indicating the presence of higher harmonics of the repetition rate with equal spectral power.
The measured spectrum, with a full width at half maximum (FWHM) of 4.1 nm centered at 2035nm, is displayed in Fig. 3(a). Using intensity autocorrelation, we determined a pulse duration of 1.5 ps (Fig. 3(b)), leading to a time-bandwidth product of 0.451. This value is 1.4 times greater than that of the ideal sech2 pulse shape. The absence of side pulses is verified through an extended autocorrelation scan over ±150 ps, as shown in the inset of Fig. 3(b). Although shorter pulses were not achievable with this first demonstration, future devices with tighter tolerances and flatter GDD are anticipated to produce shorter pulses. Additionally, further efforts to minimize two-photon absorption are expected to be required to generate shorter pulses. Figure 3(c) shows the measured microwave spectrum indicating clean fundamental modelocking at 4 GHz with equal power of the higher harmonics in the inset. The first harmonic exhibits a high signal-to-noise ratio of 60 dB, measured with a resolution bandwidth (RBW) of 3 kHz.
The MIXSEL also operates across various temperatures and pump powers. At lower pump powers, we initially observe continuous wave lasing at threshold levels. As pump power is incrementally increased, the laser naturally transitions into self-starting modelocking. Ultimately too much increase in pump power leads to a multi-pulsing regime, a phenomenon previously observed before in the near-IR regime [50]. The stable modelocking regime can be adjusted by changing the output coupler.
In our experiments, output couplers with transmission rates ranging from 0.5% to 1.8% were employed during single comb operation, within which stable modelocking was achievable. This stability is attributed to the high internal gain of our chip [25]. Particularly, the 1.8% output coupler yielded the best modelocking results, as it maintained the lowest intracavity power. Maintaining lower intracavity power in the current configuration was important to prevent gain saturation [49] and to minimize inverse saturable absorption effects. Additionally, we can fine-tune the intracavity GDD by adjusting the heat-sink temperature, altering the pump power, and employing Brewster windows of varying thicknesses [27]. The optical-to-optical efficiency of the MIXSEL, even near room temperature, remains comparable with that of modelocked GaAs-based semiconductor lasers operating in the near-IR [19].
2.4 Dual-comb operation and the potential for coherent averaging
After splitting the pump and replacing the Brewster window by the inverted biprism, dual-comb modelocked operation is easily achieved, resulting in the emission of two vertically separated, p-polarized pulse trains. Figure 4 shows the characterization of dual modelocking under the following conditions: chip heat-sink temperature of 7.6 °C, pump power of 5 W per comb, and an average output power of 45 mW (49 mW) for the upper (lower) comb as shown in the side view in Fig. 2(e). The 3-mm thick, CaF2 inverted biprism introduces approximately -140 fs2 of intracavity, round-trip GDD. The spectra in Fig. 4(a), centered at 2038nm (2039nm) with a width of 6.8 nm (6.6 nm), exhibits reasonable overlap at their main peak. Despite slight asymmetry, the spectra appear smooth at the highest resolution of 0.1 nm, indicating clean modelocking. Both combs demonstrate good beam quality (Fig. 4(b)), with the lower comb exhibiting slightly better performance due to better processing quality on that part of the chip. Autocorrelation measurements reveal picosecond pulse durations of 1.87 ps (1.16 ps). The slight difference is likely attributed to growth variation during MBE and the gradient towards the edge of the dielectric layer caused by the PECVD coating technique. More homogenous growth and processing within a commercial effort should give more symmetric dual-comb operation similar to diode-pumped solid-state lasers. Combining spectral and pulse length information yields a time-bandwidth product of 2.9 (1.8) times the ideal sech2 pulse shape, indicating a satisfactory GDD balance for this initial demonstration at 2.04 µm. The microwave spectrum in Fig. 4(c) demonstrates clean modelocking with a high signal-to-noise ratio exceeding 70 dB and minimal broadening of the 4.4-GHz peak at a RBW of 300 Hz. From the microwave spectrum, the difference in repetition rates $\Delta {f_{\textrm{rep}}}$ is 595 kHz, which can be adjusted by vertically translating the inverted biprism. The two resolved harmonics remain stable for both combs, as shown in the inset of Fig. 4(c), measured with a RBW of 100 kHz.
Fig. 4. Modelocking diagnostics of the dual-comb MIXSEL, with the upper (lower) comb shown in red (blue). a) Measured optical spectra, featuring a center wavelength of 2038nm (2039nm) and a full-width half maximum (FWHM) of 6.8 nm (6.6 nm). b) Two measured autocorrelations with sech2 fits with a pulse duration of 1.87 ps (1.16 ps) and the two combs’ beam profiles in the inset. c) Measured microwave spectra of the pulse repetition rate peak at 4.4 GHz, sampled with a resolution bandwidth (RBW) of 300 Hz. The difference in repetition rates is set to 595 kHz. The inset exhibits a larger span of the microwave spectrum (RBW 100 kHz), revealing the presence of a higher harmonic of the repetition rate with equal spectral power.
To study the relative coherence properties of the two combs, their heterodyne beating is examined. The two combs are combined on a non-polarizing beam splitter, and the beating is detected on a fast photodiode. Interferogram traces are recorded for 200 ms with a sampling rate of 5 GHz, with one interferogram displayed in Fig. 5(a). Figure 5(b) provides a zoomed-out view of the interferograms, where the spacing between interferograms corresponds to the inverse of laser repetition frequency difference $\frac{1}{{\Delta {f_{\textrm{rep}}}}}$. Figure 5(c) shows the Fourier transform of a single interferogram, providing the down converted microwave spectrum originating from the overlapping part of the optical spectra.
Fig. 5. Heterodyne beat of the dual-comb outputs recorded on a fast photodiode. Time traces undergo digital filtering using a 500 MHz low-pass filter. a) Single interferogram resulting from the beat of the two pulse trains. b) Zoomed-out view of the interferogram traces. c) Fast Fourier transform (FT) of a single interferogram, revealing the microwave spectrum generated by the beat of the two combs. d) FT of the interferogram traces resulting into the microwave comb. The inset provides a magnified view of the individual comb lines. e) Evolution of the first and second-order differences in the interferogram phase over the entire 200 ms recording time. Notably, the second-order difference of the phase remains bounded between ± π.
A microwave comb (Fig. 5(d)) for which the individual comb lines are fully resolved is crucial for dual-comb spectroscopy without the need for additional stabilization. Several replicas of this microwave spectrum are observed around the laser harmonics. The analysis focuses on the replica between zero frequency (DC) and half of the pulse repetition frequency, with the lowest frequency chosen to avoid overlap with the strongly fluctuating DC components. For 15 interferogram periods the comb structure is clearly conserved with a contrast between the comb lines and the background 36 dB.
Analyzing the full 200-ms data trace, we observed that the microwave comb lines undergo a gradual shift and broaden over extended measurement durations. However, the interferograms’ phase can be tracked, enabling the application of coherent averaging algorithms [51–53,14]. This can be seen in Fig. 5(e), where the continuously tracked interferograms’ phase exhibits fluctuations, yet its second-order difference remains strictly bounded between ${\pm} \pi $. This signifies that the phase variations occur at a slow enough rate for coherent averaging algorithms. In the future, this holds promise for spectroscopy applications without the need for laser stabilization to an optical reference.
The gigahertz repetition rate of our laser allows for the sampling of substantial bandwidths without aliasing. In our current configuration, with a pulse repetition rate of approximately 4.4 GHz and a difference in repetition rate of $\Delta {f_{\textrm{rep}}}$= 595 kHz, we can potentially sample bandwidth up to 225 nm at our present center wavelength of approximately 2 µm [4]. This capability opens up opportunities for high-speed acquisition, and the compact footprint renders the dual-comb MIXSEL particularly advantageous for field applications.
3. Conclusions
In conclusion, we have successfully demonstrated a dual-comb laser source operating at a center wavelength above 2 µm with a repetition rate of approximately 4 GHz, making it suitable for dual-comb spectroscopy. Our laser source, known as MIXSEL, comprises a semiconductor gain and absorber region fully integrated inside a highly reflective semiconductor chip, along with an external output coupler and a multiplexing device. The MIXSEL chip, fabricated using III-V semiconductor molecular beam epitaxy (MBE), incorporates InGaSb gain and absorber quantum wells (QWs) within a single monolithic structure. The laser emits fundamentally modelocked pulses with a duration of 1.5 ps and an average power of 28 mW. Through spatial multiplexing using an inverted biprism, the laser cavity is multiplexed with a repetition rate difference of $\Delta {f_{\textrm{rep}}}$= 595 kHz. This configuration enables the emission of two spatially separate, p-polarized pulse trains with an average output power of 45-49 mW at a repetition rate of 4.4 GHz.
The heterodyne beat of the two pulse trains can be Fourier transformed to investigate the microwave comb. Remarkably, our free-running dual-comb MIXSEL produces resolvable comb lines from a few interferograms without requiring any stabilization. The two distinct repetition rates facilitate sampling of large bandwidths without aliasing while maintaining fast acquisition times. Furthermore, the phase of the heterodyne beat interferograms can be tracked, allowing the application of coherent averaging algorithms [14].
While the recently published dual-comb Cr:ZnS laser [31] excels in output power and pulse duration, the MIXSEL stands out for its wavelength flexibility through semiconductor bandgap engineering. Additionally, its compact laser architecture delivers multi-gigahertz repetition rates in dual-comb operation, facilitating comb-line-resolved microwave spectra.
The versatility of the semiconductor platform utilized in the MIXSEL enables bandgap engineering and freely adjustable operation wavelength. In the current GaSb material system, we can extend the operation wavelength from 1.7 µm up to 3 µm, covering the SWIR spectral region and potentially reaching into the mid-IR range, which is well-suited for trace gas spectroscopy. Our first proof-of-principle demonstration as reported in this publication lays the foundation for widespread application of dual-comb SWIR MIXSEL technology as free-running sources.
3.1 Methods
3.1.1 Design and chip characterization
The chip was grown using MBE following the method described in Ref. [23] with a growth uncertainty of 2% or less. It was then processed in an upside-down configuration, as outlined in Ref. [25]. The layer structure was analyzed by cleaving a fresh facet and examining it using a Zeiss ULTRA 55 plus scanning electron microscope (SEM). Linear reflectance measurements were performed using a Varian Cary 5e spectrophotometer, while photoluminescence (PL) was measured using a Bruker Vertex 80 v spectrometer equipped with a PL II unit. The PL II unit utilized a 1064 nm excitation laser with a power of up to 100 mW. Group delay dispersion (GDD) was measured using a custom-built setup. The layer structures were simulated using a custom-made transfer matrix simulation tool, which utilized complex refractive indices to calculate reflection, absorption, field intensities, and GDD.
3.1.2 Heat simulations
Heat simulations are based on a one-dimensional heat conduction model, where each layer is assigned a specific heat conductivity based on its material properties. This approach is justified because the pump-spot diameter exceeds 200 µm, while the total structure thickness is approximately 10 µm. The stationary heat equation, subject to appropriate boundary conditions, is applied, where z = 0 represents the heatsink and z = zmax corresponds to the air interface
This equation can be analytically solved using integrals
3.1.3 Laser characterization
The optical spectra of the laser are measured using a Yokogawa AQ6376 spectrometer. Pulse durations are measured using a Femtochrome FR-103WS intensity autocorrelator. Microwave spectra are recorded using an EOT ET-5000F InGaAs 12.5-GHz fast photodiode, coupled with a 25-dB gain preamplifier (Agilent Technologies 87405C) and a microwave spectrum analyzer (Agilent Technologies E4405B). The heterodyne beat note is analyzed using the same fast photodiode connected to a Teledyne LeCroy WAVEPRO 254HD oscilloscope (5 GHz), employing a 2.85 GHz low pass filter (Mini Circuits VLF-2850+). Beam profiles are measured using a DataRay Wincam IR-BB.
3.1.4 Multiplexing device
The inverted biprism was constructed using a 3° calcium fluoride (CaF2) wedge obtained from ISP optics. To achieve an inverted apex angle of 6°, the wedge was segmented and the opposite thin facets were securely bonded together using UV-curing adhesive.
Funding
H2020 European Research Council (787097).
Acknowledgements
We thank Dr. Chris Phillips for developing and providing the data analysis software used to generate Fig. 5(e). FIRST clean room facility at ETH Zurich.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. H. R. Telle, G. Steinmeyer, A. E. Dunlop, J. Stenger, D. H. Sutter, and U. Keller, “Carrier-envelope offset phase control: A novel concept for absolute optical frequency measurement and ultrashort pulse generation,” Appl. Phys. B: Lasers Opt. 69(4), 327–332 (1999). [CrossRef]
2. A. Apolonski, A. Poppe, G. Tempea, C. Spielmann, T. Udem, R. Holzwarth, T. W. Hänsch, and F. Krausz, “Controlling the Phase Evolution of Few-Cycle Light Pulses,” Phys. Rev. Lett. 85(4), 740–743 (2000). [CrossRef]
3. D. J. Jones, S. A. Diddams, J. K. Ranka, A. Stentz, R. S. Windeler, J. L. Hall, and S. T. Cundiff, “Carrier-Envelope Phase Control of Femtosecond Mode-Locked Lasers and Direct Optical Frequency Synthesis,” Science 288(5466), 635–639 (2000). [CrossRef]
4. I. Coddington, N. Newbury, and W. Swann, “Dual-comb spectroscopy,” Optica 3(4), 414 (2016). [CrossRef]
5. T. Fortier and E. Baumann, “20 years of developments in optical frequency comb technology and applications,” Commun. Phys. 2(1), 153 (2019). [CrossRef]
6. R. Liao, H. Tian, W. Liu, R. Li, Y. Song, and M. Hu, “Dual-comb generation from a single laser source: Principles and spectroscopic applications towards mid-IR - A review,” J. Phys. Photonics 2(4), 042006 (2020). [CrossRef]
7. C. Zhang, F. Qu, P. Ou, H. Sun, S. He, and B. Fu, “Recent Advances and Outlook in Single-Cavity Dual Comb Lasers,” Photonics 10(2), 221 (2023). [CrossRef]
8. S. M. Link, A. Klenner, M. Mangold, M. Golling, B. W. Tilma, and U. Keller, “Dual-comb modelocked lasers,” Opt. Express 23(5), 5521–5531 (2015). [CrossRef]
9. B. Willenberg, J. Pupeikis, L. M. Krüger, F. Koch, C. R. Phillips, and U. Keller, “Femtosecond dual-comb Yb:CaF 2 laser from a single free-running polarization-multiplexed cavity for optical sampling applications,” Opt. Express 28(20), 30275 (2020). [CrossRef]
10. J. Pupeikis, B. Willenberg, S. L. Camenzind, A. Benayad, P. Camy, C. R. Phillips, and U. Keller, “Spatially multiplexed single-cavity dual-comb laser,” Optica 9(7), 713 (2022). [CrossRef]
11. S. M. Link, D. J. H. C. Maas, D. Waldburger, and U. Keller, “Dual-comb spectroscopy of water vapor with a free-running semiconductor disk laser,” Science 356(6343), 1164–1168 (2017). [CrossRef]
12. D. J. H. C. Maas, A. R. Bellancourt, B. Rudin, M. Golling, H. J. Unold, T. Südmeyer, and U. Keller, “Vertical integration of ultrafast semiconductor lasers,” Appl. Phys. B 88(4), 493–497 (2007). [CrossRef]
13. M. Mangold, S. M. Link, A. Klenner, C. A. Zaugg, M. Golling, B. W. Tilma, and U. Keller, “Amplitude Noise and Timing Jitter Characterization of a High-Power Mode-Locked Integrated External-Cavity Surface Emitting Laser,” IEEE Photonics J. 6(1), 1–9 (2014). [CrossRef]
14. C. R. Phillips, B. Willenberg, A. Nussbaum-Lapping, F. Callegari, S. L. Camenzind, J. Pupeikis, and U. Keller, “Coherently averaged dual-comb spectroscopy with a low-noise and high-power free-running gigahertz dual-comb laser,” Opt. Express 31(5), 7103 (2023). [CrossRef]
15. M. Mangold, V. J. Wittwer, C. A. Zaugg, S. M. Link, M. Golling, B. W. Tilma, and U. Keller, “Femtosecond pulses from a modelocked integrated external-cavity surface emitting laser (MIXSEL),” Opt. Express 21(21), 24904 (2013). [CrossRef]
16. C. G. E. Alfieri, D. Waldburger, J. Nürnberg, M. Golling, and U. Keller, “Sub-150-fs pulses from an optically pumped broadband modelocked integrated external-cavity surface emitting laser,” Opt. Lett. 44(1), 25 (2019). [CrossRef]
17. B. Rudin, V. J. Wittwer, D. J. H. C. Maas, M. Hoffmann, O. D. Sieber, Y. Barbarin, M. Golling, T. Südmeyer, and U. Keller, “High-power MIXSEL: an integrated ultrafast semiconductor laser with 64 W average power,” Opt. Express 18(26), 27582 (2010). [CrossRef]
18. M. Mangold, C. A. Zaugg, S. M. Link, M. Golling, B. W. Tilma, and U. Keller, “Pulse repetition rate scaling from 5 to 100 GHz with a high-power semiconductor disk laser,” Opt. Express 22(5), 6099 (2014). [CrossRef]
19. B. W. Tilma, M. Mangold, C. A. Zaugg, S. M. Link, D. Waldburger, A. Klenner, A. S. Mayer, E. Gini, M. Golling, and U. Keller, “Recent advances in ultrafast semiconductor disk lasers,” Light: Sci. Appl. 4(7), e310–310 (2015). [CrossRef]
20. J. Nürnberg, C. G. E. Alfieri, Z. Chen, D. Waldburger, N. Picqué, and U. Keller, “An unstabilized femtosecond semiconductor laser for dual-comb spectroscopy of acetylene,” Opt. Express 27(3), 3190 (2019). [CrossRef]
21. J. Nürnberg, B. Willenberg, C. R. Phillips, and U. Keller, “Dual-comb ranging with frequency combs from single cavity free-running laser oscillators,” Opt. Express 29(16), 24910 (2021). [CrossRef]
22. M. Guina, A. Rantamäki, and A. Härkönen, “Optically pumped VECSELs: Review of technology and progress,” J. Phys. D: Appl. Phys. 50(38), 383001 (2017). [CrossRef]
23. BÖ Alaydin, M. Gaulke, J. Heidrich, M. Golling, A. Barh, and U. Keller, “Bandgap engineering, monolithic growth, and operation parameters of GaSb-based SESAMs in the 2–2.4 µm range,” Opt. Mater. Express 12(6), 2382 (2022). [CrossRef]
24. J. Heidrich, M. Gaulke, B. O. Alaydin, M. Golling, A. Barh, and U. Keller, “Full optical SESAM characterization methods in the 1.9 to 3-µm wavelength regime,” Opt. Express 29(5), 6647–6656 (2021). [CrossRef]
25. M. Gaulke, J. Heidrich, B. Özgür Alaydin, M. Golling, A. Barh, and U. Keller, “High average output power from a backside-cooled 2-µm InGaSb VECSEL with full gain characterization,” Opt. Express 29(24), 40360 (2021). [CrossRef]
26. N. Huwyler, M. Gaulke, J. Heidrich, M. Golling, A. Barh, and U. Keller, “3-W output power from a 2-µm InGaSb VECSEL using a hybrid metal-semiconductor Bragg reflector,” Opt. Mater. Express 13(3), 833 (2023). [CrossRef]
27. J. Heidrich, M. Gaulke, M. Golling, B. O. Alaydin, A. Barh, and U. Keller, “324-fs Pulses From a SESAM Modelocked Backside-Cooled 2-µm VECSEL,” IEEE Photon. Technol. Lett. 34(6), 337–340 (2022). [CrossRef]
28. A. Härkönen, C. Grebing, J. Paajaste, R. Koskinen, J.-P. Alanko, S. Suomalainen, G. Steinmeyer, and M. Guina, “Modelocked GaSb disk laser producing 384 fs pulses at 2 µm wavelength,” Electron. Lett. 47(7), 454–456 (2011). [CrossRef]
29. C. P. Bauer, S. L. Camenzind, J. Pupeikis, B. Willenberg, C. R. Phillips, and U. Keller, “Dual-comb optical parametric oscillator in the mid-infrared based on a single free-running cavity,” Opt. Express 30(11), 19904 (2022). [CrossRef]
30. S. L. Camenzind, T. Sevim, B. Willenberg, J. Pupeikis, A. Nussbaum-Lapping, C. R. Phillips, and U. Keller, “Free-running Yb:KYW dual-comb oscillator in a MOPA architecture,” Opt. Express 31(4), 6633 (2023). [CrossRef]
31. A. Barh, A. Nussbaum-Lapping, J. Heidrich, M. Gaulke, M. Golling, C. R. Phillips, and U. Keller, “Single-cavity dual-modelocked 2.36-µm laser,” Opt. Express 31(4), 6475 (2023). [CrossRef]
32. A. Nussbaum-Lapping, C. R. Phillips, B. Willenberg, J. Pupeikis, and U. Keller, “Absolute SESAM characterization via polarization-resolved non-collinear equivalent time sampling,” Appl. Phys. B 128(2), 24 (2022). [CrossRef]
33. J. Pupeikis, W. Hu, B. Willenberg, M. Mehendale, G. A. Antonelli, C. R. Phillips, and U. Keller, “Efficient pump-probe sampling with a single-cavity dual-comb laser: Application in ultrafast photoacoustics,” Photoacoustics 29, 100439 (2023). [CrossRef]
34. T. Flöry, V. Stummer, J. Pupeikis, B. Willenberg, A. Nussbaum-Lapping, E. Kaksis, F. V. A. Camargo, M. Barkauskas, C. R. Phillips, U. Keller, G. Cerullo, A. Pugžlys, and A. Baltuška, “Rapid-Scan Nonlinear Time-Resolved Spectroscopy over Arbitrary Delay Intervals,” Ultrafast Science 3, 0027 (2023). [CrossRef]
35. J. Pupeikis, B. Willenberg, F. Bruno, M. Hettich, A. Nussbaum-Lapping, M. Golling, C. P. Bauer, S. L. Camenzind, A. Benayad, P. Camy, B. Audoin, C. R. Phillips, and U. Keller, “Picosecond ultrasonics with a free-running dual-comb laser,” Opt. Express 29(22), 35735 (2021). [CrossRef]
36. S. L. Camenzind, J. F. Fricke, J. Kellner, B. Willenberg, J. Pupeikis, C. R. Phillips, and U. Keller, “Dynamic and precise long-distance ranging using a free-running dual-comb laser,” Opt. Express 30(21), 37245 (2022). [CrossRef]
37. X. Zhao, G. Hu, B. Zhao, C. Li, Y. Pan, Y. Liu, T. Yasui, and Z. Zheng, “Picometer-resolution dual-comb spectroscopy with a free-running fiber laser,” Opt. Express 24(19), 21833–21845 (2016). [CrossRef]
38. R. Liao, Y. Song, W. Liu, H. Shi, L. Chai, and M. Hu, “Dual-comb spectroscopy with a single free-running thulium-doped fiber laser,” Opt. Express 26(8), 11046–11054 (2018). [CrossRef]
39. S. Mehravar, R. A. Norwood, N. Peyghambarian, and K. Kieu, “Real-time dual-comb spectroscopy with a free-running bidirectionally mode-locked fiber laser,” Appl. Phys. Lett. 108(23), 231104 (2016). [CrossRef]
40. T. Ideguchi, T. Nakamura, Y. Kobayashi, and K. Goda, “Kerr-lens mode-locked bidirectional dual-comb ring laser for broadband dual-comb spectroscopy,” Optica 3(7), 748–753 (2016). [CrossRef]
41. I. Coddington, W. C. Swann, and N. R. Newbury, “Coherent Multiheterodyne Spectroscopy Using Stabilized Optical Frequency Combs,” Phys. Rev. Lett. 100(1), 013902 (2008). [CrossRef]
42. P. Giaccari, J.-D. Deschênes, P. Saucier, J. Genest, and P. Tremblay, “Active Fourier-transform spectroscopy combining the direct RF beating of two fiber-based mode-locked lasers with a novel referencing method,” Opt. Express 16(6), 4347–4365 (2008). [CrossRef]
43. J. Faist, G. Villares, G. Scalari, M. Rösch, C. Bonzon, A. Hugi, and M. Beck, “Quantum Cascade Laser Frequency Combs,” Nanophotonics 5(2), 272–291 (2016). [CrossRef]
44. T. J. Kippenberg, A. L. Gaeta, M. Lipson, and M. L. Gorodetsky, “Dissipative Kerr solitons in optical microresonators,” Science 361(6402), eaan8083 (2018). [CrossRef]
45. A. L. Gaeta, M. Lipson, and T. J. Kippenberg, “Photonic-chip-based frequency combs,” Nat. Photonics 13(3), 158–169 (2019). [CrossRef]
46. C. G. E. Alfieri, D. Waldburger, E. Gini, G. Eisenstein, M. Golling, S. M. Link, and U. Keller, “Optical efficiency and gain dynamics of modelocked semiconductor disk lasers,” Opt. Express 25(6), 6402–6420 (2017). [CrossRef]
47. Y. Barbarin, A. R. Bellancourt, D. J. H. C. Maas, M. Shafiei, M. Hoffmann, M. Golling, T. Südmeyer, and U. Keller, “Low saturation fluence antiresonant quantum dot SESAMs for MIXSEL integration,” Opt. Express 17(12), 9704 (2009). [CrossRef]
48. Y. P. Varshni, “Temperature dependence of the energy gap in semiconductors,” Physica 34(1), 149–154 (1967). [CrossRef]
49. O. D. Sieber, M. Hoffmann, V. J. Wittwer, M. Mangold, M. Golling, B. W. Tilma, T. Südmeyer, and U. Keller, “Experimentally verified pulse formation model for high-power femtosecond VECSELs,” Appl. Phys. B 113(1), 133–145 (2013). [CrossRef]
50. D. Waldburger, C. G. E. Alfieri, S. M. Link, S. Meinecke, L. C. Jaurigue, K. Lüdge, and U. Keller, “Multipulse instabilities of a femtosecond SESAM-modelocked VECSEL,” Opt. Express 26(17), 21872–21886 (2018). [CrossRef]
51. J. Roy, J.-D. Deschênes, S. Potvin, and J. Genest, “Continuous real-time correction and averaging for frequency comb interferometry,” Opt. Express 20(20), 21932–21939 (2012). [CrossRef]
52. L. A. Sterczewski, J. Westberg, and G. Wysocki, “Computational coherent averaging for free-running dual-comb spectroscopy,” Opt. Express 27(17), 23875–23893 (2019). [CrossRef]
53. N. B. Hébert, V. Michaud-Belleau, J.-D. Deschênes, and J. Genest, “Self-Correction Limits in Dual-Comb Interferometry,” IEEE J. Quantum Electron. 55(4), 1–11 (2019). [CrossRef]
