1 Introduction

Laser emission at multiple wavelengths is a potential tool in various applications such as optical sensing [1], microwave photonics [2], absorption spectroscopy [3], wavelength division multiplexed communication systems [4], and in wavelength-modulated Raman spectroscopy (WMRS) [5].

Distributed Bragg reflector (DBR)-based diode laser arrays offer narrow laser emission at various wavelengths from a compact device [6]. However, each array has its own aperture and thus requires a method for spatial combination of the individual arrays/wavelengths for most applications [7].

Alternatively, monolithically combined ridge waveguide (RW)-based DBR lasers offer emission at multiple wavelength emitted from a single aperture. This is, e.g. realized in the so-called Y-branch DBR-RW laser where two branches/arms are combined to offer two wavelengths from a single aperture [8]. These light sources are suitable for beat signal generation [9, 10], shifted excitation Raman spectroscopy [11], terahertz frequency generation [12, 13] and in tunable high-power master oscillator power amplifier (MOPA) systems [14].

In this work, we investigate the further development of these light sources by considering four-arm DBR-RW laser sources emitting around 970 nm. These light sources consist of four individual RWs with corresponding DBR gratings, which are combined and emitted from relatively compact (4 × 0.8 mm2) devices.

In particular, two of such devices with different intersection structures are considered: one where all four arms are combined at a single intersection point and a second device with two individual intersections. The motivation of this work is to investigate the effect of different curvature of the inner vs outer arms and the coupling between those. By understanding and optimizing the shape and intersections between multiple RWs, we believe one could develop compact multi-wavelength laser sources.

2 Setup

The investigated lasers are both implemented on an asymmetric super-large optical-cavity (ASLOC) vertical structure developed using metal–organic vapour phase epitaxy (MOVPE). An InGaAs single quantum well is utilized as the active layer, which is embedded between AlGaAs waveguide and cladding layers. The RW structures are made using projection lithography with an I-line wafer stepper. The DBRs are seventh order, deeply etched surface gratings realized using E-beam lithography. The structure emits TE (transverse electric)-polarized light which has a vertical far-field angle of 26.7°, measured at full width half maximum (FWHM).

For reference purposes, the threshold current and the slope efficiency are determined for uncoated, unmounted 1-mm-long laser diodes with a stripe width of 100 µm. These measurements were taken in pulsed mode with a 1-µs current pulse length at a repetition rate of 1 kHz. A threshold current Ith = 158 mA, a slope efficiency S = 0.60 W/A, and a differential efficiency of ηD = 0.94 are determined.

Using these excitation conditions on devices with different resonator lengths, the following material data were determined: internal quantum efficiency ηi ≈ 0.95, internal losses αi = 1.2 cm− 1, transparency current density JTR = 72.5 A/cm2, threshold current density for devices with infinite length J∞= 78.2 A/cm2, modal gain coefficient Γg0 = 15.9 cm− 1 and the characteristic temperature of the threshold current T0 = 133 K for a 1-mm-long device. The high internal quantum efficiency and the relatively low internal losses make the developed structure well optimized for laser processing.

The contact scheme of the processed four-arm DBR-RW lasers is shown in Fig. 1. As can be seen, each waveguide has its own DBR grating, and the four waveguides are joined (common section) before exiting from the front section. Note that the different sections (the individual arms, common- and front sections) have isolated gold surface contacts to control the current injections separately.

Fig. 1
figure 1

Contact scheme of the investigated lasers

Full size image

Each waveguide consists of three parts: an angled part, a bend part and a front “straight” part. The bend sections follow the S-bend formula [15]:

\(Y\left( z \right)=\frac{w}{2}\left[ {1 - {\text{Cos}}\left( {\frac{{\pi z}}{L}} \right)} \right],\)where Y is the lateral position of the waveguide at position z, winner = 20 µm and wouter = 60 µm, are half the lateral offset, for a bending length of Linner = 2000 µm and Louter = 1000 µm, for the inner and outer arms, respectively. The bend losses are reduced by implementing a large bend radius according to the etch depth and effective index step [16]. Note that the S-bend formula provides two bends to achieve a lateral shift, while in this investigation only a single bend is used. The inner two arms are angled by ± 1.8°, while the outer are angled by ± 3.4°. These values are found to provide a smooth transition between the angled and the bend sections. Note that the DBR gratings are angled with respect to the angle of the arms. The transition between the angled and bend parts is indicated in Fig. 1.

The motivation behind this particular structure is to use bend structures with low curvature, to ensure combination of the individual arms while minimize bending losses. The angled RW sections ensure further displacement between the individual arms at the back side of the devices, providing more space for the processing of the individual DBR gratings.

For this investigation, two lasers with different common section structures are investigated. The first laser has a single intersection point for all four waveguides, while the second laser has two separate points, see Fig. 2a, b, respectively. In the latter case, the inner and outer arms’ intersection points are displaced by 150 µm from one another. The individual arms of laser A are denoted as shown in Fig. 2 as A1, A2, A3 and A4 and likewise for Laser B: B1, B2, B3 and B4.

Fig. 2
figure 2

Close-ups of the common/front section of a laser A with a single connection point and b Laser B with two connection points

Full size image

The facets of the investigated lasers are cleaned by applying atomic hydrogen and a sealing process with ZnSe before being reflectivity coated [17]. Both lasers have a front facet reflectivity of Rf = 5% and an anti-reflection (AR) coating on the rear facets, with a reflectivity about the order of Rr ≈ 5 × 10− 4.

Both lasers are mounted p-side up on CuW sub-mounts, and each is mounted on top of an AlN heat spreader. These subassemblies are then mounted on 25 mm × 25 mm conduction cooled package laser mounts, allowing efficient heat removal.

3 Experimental results

In this section, the experimental results of the lasers with the two different designs are investigated and compared with one another.

3.1 Output power characteristics

The power–voltage–current (PUI) curves of the two lasers when operated at a heat sink temperature of T = 25 °C are shown side-by-side in Fig. 3. For this investigation, the common section is not operated, while the front section is injected with Ifront = 35 mA, and the injection current through each arm is varied between Iinj = [0, 500] mA, with step size of ΔIinj = 5 mA. Note that the small Ifront current value is chosen to keep the front section transparent and to avoid absorption in the front section.

Fig. 3
figure 3

Power–voltage–current of a laser A and b laser B

Full size image

The four arms of each laser are characterized and the threshold current Ith, the maximum measured output power Pmax (@ Iinj = 500 mA), and the slope efficiency S are summarized in Table 1. The slope efficiencies are obtained through linear fits between Iinj = [100, 300] mA, a region above threshold and below the role-over effect. Table 1 shows a clear difference in performance when comparing the outer (A1 and A4) and the inner arms (A2 and A3) of laser A. This is seen in Ith, Pmax, S and also in Fig. 3a.

Table 1 Laser characteristics of the investigated lasers at a heat sink temperature of T = 25°C
Full size table

The laser characteristics of the four arms of laser B are likewise summarized in Table 1. In this case, it is seen that the four arms have similar performance, with a smaller discrepancy between the outer arms (B1 and B4) and the inner arms (B2 and B3). This is also clear in the power curves in Fig. 3b.

The observed behaviour in Fig. 3a indicates comparative losses between the inner (or outer) arms, but different losses between the inner and outer arms. This effect is less evident in the power characteristics in Fig. 3b, where comparative performance of the four arms is observed.

Note that the Ith values are obtained by first significant data points. By neglecting saturation effects and looking at the estimated zero crossing current values, similar Ith values are expected for A1 and A4, and likewise for B1 and B4.

The observed laser characteristics suggest different losses in the inner and outer arms in the case of laser A, and comparable performance in the case of the four arms of laser B. This is explained by the different common section designs, and the comparable values for the inner and outer arms of laser B suggest that the curvature of the individual arms has a minor effect on the opto-electrical performance.

3.2 Wavelength characteristics

Next, the spectral behaviour of the two lasers is investigated. Figure 4 shows the normalized emission wavelength of the individual arms, as function of the injection current in a false colour contour diagram. These measurements are taken using a double echelle monochromator DEMON from LTB Lasertechnik Berlin, with a spectral resolution of 17 pm at 976 nm.

Fig. 4
figure 4

False colour contour plots of the emission wavelength of the studied lasers, as a function of the injected current

Full size image

Laser A emits at wavelengths of about λA1 = 978 nm, λA2 = 975 nm, λA3 = 973 nm and λA4 = 971 nm, respectively. As can be seen in Fig. 4, all eight arms of the two lasers maintain single-mode operation over the investigated injection current range. The discrepancy in the threshold currents between the inner and outer arms of laser A (see Fig. 3a) can also be observed in the spectral emission. Laser B emits at similar wavelengths of laser A, and similar threshold values for all four arms are observed, in agreement with the power curve (see Fig. 3b).

The individual optical spectra of the investigated lasers at Iinj = 200 mA are shown in Fig. 5, and the emission wavelength values are summarized in Table 1. From this measurement, the emission linewidth of all eight measurements is estimated to be smaller than ΔλFWHM < 17 pm, limited by the resolution of the spectrometer. A side mode suppression ratio (SMSR) of about 30 dB is observed in all eight measurements, limited by the dynamic range of the monochromator.

Fig. 5
figure 5

Normalized individual optical spectra of the investigated lasers at Iinj = 200 mA

Full size image

The spectral distance between the four arms was designed to be ~ 2.2 nm. In addition, the DBR gratings were designed to provide similarly wavelengths from both lasers (A1 and B1, A2 and B2), i.e. the observed shift in wavelength in Fig. 5 is due to manufacturing tolerances.

Note that the chosen injection current of Iinj = 200 mA provides an output power of about P = 100 mW, which we intend to apply in future applications of these lasers. Nonetheless, an emission linewidth smaller than ΔλFWHM < 17 pm is maintained over the entire investigated injection current range as indicated in Fig. 4.

3.3 Spatial characteristics

The spatial characteristics of both lasers are obtained according to the moving slit method [20]. The normalized lateral intensity distributions of the near field (@ Iinj = 200 mA) are shown side-by-side in Fig. 6.

Fig. 6
figure 6

Normalized near-field profiles of the investigated lasers at Iinj = 200 mA

Full size image

The inner arms of laser A (A2 and A3) show two defined central lobes, with a number of low-intensity side lobes. In the case of A2, these neighbouring peaks are mainly concentrated on the rhs (positive x-position), and vice versa for A3. In the case of the outer arms, a defined central lobe is once again seen, however, with strong side lobes that are concentrated toward the inner arms. The near fields of both the inner and outer arms have mirror symmetry, and show the effect of having stronger side lobes towards the aperture (front section).

In the case of laser B, the near fields of the inner arms show two broad neighbouring peaks with a defined central lobe in between. As seen in the case of laser A, the side lobes toward the aperture are larger than those toward the neighbouring outer arm. The outer arms of laser B show similar behaviour to the ones of laser A, having defined central lobes with side lobes mainly concentrated toward the inner arms (aperture).

The near-field widths of both lasers A and B are summarized in Table 2. In the case of laser A, comparable near-field values are observed for the inner arms (~ 5 µm) and for the outer arms (~ 27 µm). In the case of laser B, the same behaviour is observed, however, with a smaller discrepancy between the inner (~ 6 µm) and outer arms (~ 9 µm).

Table 2 Beam quality parameters of the investigated lasers at Iinj = 200 mA
Full size table

By considering the near-field distributions, the power in central lobe PCL of each arm is obtained and summarized in Table 2. Once again, laser A have comparable values for the inner (or outer) arms, but with a large discrepancy from 32 to 78% between those. In the case of laser B, the PCL values of the inner and outer arms only vary between 45 and 53%.

The measured far-field intensity profiles of the two lasers are shown in Fig. 7. The profiles of both lasers consist of multiple high-intensity peaks, showing mirror symmetry between the inner and outer arms. Using the far-field angles together with the near-field values, the propagation values M2 (@ 1/e2) in Table 2 are obtained.

Fig. 7
figure 7

Normalized far-field profiles of the investigated lasers at Iinj = 200 mA

Full size image

In the case of laser A, the inner arms have comparable M2 values (~ 1.5), and likewise for the outer arms (~ 8.2). The same effect is seen in the case of laser B, however, with smaller discrepancy in the M2 values, ranging between 1.6 for the inner and 2.8 for the outer arms, respectively.

At higher injection currents, the beam quality deteriorates for both lasers. However, the overall observed behaviour is maintained, e.g. the large discrepancy between the inner and outer arms for laser A and the more comparable behaviour observed for laser B (Table 3).

Table 3 Propagation beam quality parameters of the investigated lasers at Iinj = 200 mA
Full size table

The presented spatial characteristic results suggest that the common section plays a crucial effect on the spatial characteristics, whereas the curvature of the individual arms has a smaller effect.

4 Conclusions

In this study, two different four-arm DBR-RW lasers are investigated: laser A having a single intersection and laser B having two separate intersection points between the individual arms. Through this investigation, it is clear that the two structures exhibit different laser characteristics. This is observed in the output power characteristics, where laser A has distinct performance when comparing the inner and outer arms. In particular, the outer arms had smaller slope efficiencies and lower output power values in comparison to the inner arms. On the other hand, laser B showed similar slope efficiencies and output power values for all four arms.

Similar behaviour is observed in the spatial characteristics, where laser A had higher M2 values (~ 8.2) for the outer arms than for the inner arms (~ 1.5), and likewise smaller PCL values (32%) for the outer arms when compared to the inner arms (78%). The four arms of laser B, on the other hand, had comparable beam qualities, where the M2 values only ranged between 1.6 and 2.8 and the PCL ranged between 45 and 53%.

The spectral characteristics of the two lasers did, however, not show this kind of discrepancy. Single-mode operation and a spectral width smaller than 17 pm (resolution limit) were maintained over the investigated injection current range for both lasers.

The spatial characteristics of both lasers, especially the far-field profiles, indicate non-diffraction limited beams, which is the “cost” of having four monolithically combined laser arms. This is believed to be caused by the joining sections, where in the case of laser A, the RW suddenly becomes four times as wide, and in the case of laser B becomes twice as wide at the first joining point, and about three times as wide at the second joining point. In both cases, this change allows higher order (spatial) modes to propagate, as seen in the near-field measurements. The comparable results of the four arms of laser B suggest that the curvature of the individual arms plays a minor role on the spatial characteristics.

While the non-diffraction limited beams are a clear drawback in these light sources, we target applications where spatial filtering takes place, either by coupling the light into single-mode fibres or in MOPA systems. This is achievable due to the relatively high output powers, concentrated within the central lobe.

Finally, it is clear that laser B is preferred over laser A in almost every aspect. Future work should include waveguide simulations of the intersection between the RW to understand the experimental behaviour shown in this study, and to further improve the spatial characteristics of multi-arm DBR-RW lasers.