Open Access
Add to BibSonomyAbstract
Parasitic lateral lasing in certain optically pumped semiconductor disc lasers drains the gain of the vertical mode and thus causes power scaling degradation and premature rollover in surface emitting operation. We have observed this effect in both multiple quantum wells (MQW) (GaInAs/GaAs) and double heterostructures (DHS) (GaInP/GaAs/GaInP) under pulsed excitation even when the gain chip lateral dimensions are much larger than the diameter of the pump laser. Lateral lasing occurs persistently between cleaved facets at a band-tail wavelength much longer than the peak of the gain. We show that the effect of bandgap renormalization due to Coulomb screening explains this phenomena. Exploiting the simple analytical plasma theory of bulk semiconductors (Banyai & Koch, 1986), we can account for such an effect in double heterostructures.
© 2015 Optical Society of America
1. Introduction
Vertical External Cavity Surface Emitting Lasers (VECSELs), also known as Optically Pumped Semiconductor Lasers (OPSLs), have been generating a great amount of interest due to their superior performance in beam quality, power scaling capability and wavelength flexibility [1–3].
In order to power scale the VECSEL device, the pump spot size needs to be increased. The large pump area will increase the gain length in the lateral plane amplifying the spontaneous emission. This effect was mentioned in the very early development of Vertical Cavity Surface Emitting Lasers (VCSELs) [4] as well as other types of disk lasers [5]. Detection of such lateral spontaneous emission has been used to determine the non-radiative recombination coefficient in VCSELs [6] as well as for VCSEL optical power monitoring [7]. With sufficient feedback from the edges of the device, lateral lasing can be observed, which in turn hampers the lasing operation in the vertical direction [8–11]. Lateral lasing can be suppressed either by choosing a large enough chip size or by facet modification [8, 11]. Even if lateral lasing is suppressed, amplified spontaneous emission (ASE) will still be a significant factor raising threshold and reducing efficiency of vertical lasing [12]. One of the useful applications of the interplay between vertical and lateral lasing was demonstrated in the semiconductor optical amplifier reported by Francis et al. [13], wherein gain clamping upon lasing in the vertical direction linearized optical amplification in the lateral direction. Hessenius et al. [8] also observed an interesting yet puzzling phenomenon in which lateral lasing happens even when the pump spot size is significantly smaller than the lateral dimension of the sample. This implies that, the total optical gain within the small pumped region has to be larger than the total optical absorption in the much longer un-pumped region of the sample. They attribute this phenomenon to the thermal shift of the bandgap: As the device is pumped, the active region begins to heat, leading to the red-shifting of the bandgap in the pumped region compared to the surrounding area. This, they reasoned, results in the un-pumped region to be transparent to the spontaneous emission from the active region.
In this paper, however, we show that even after the heating effect is eliminated by using a pulsed optical excitation with very low duty cycle, the same lateral lasing effect is still observed in a lifted-off VECSEL structure bonded to a crystal substrate. Such pulsed lateral lasing occurs in both multiple-quantum well (MQW) as well as double heterostructure (DHS) gain chips, and cannot be explained by the thermal model. We propose a new model based on the effect of bandgap renormalization due to Coulomb screening. We show that theoretical results based on a simple plasma theory is in good agreement with the experimental data for GaAs bulk DHS gain chips.
2. Observation of lateral lasing in VECSELs
A typical VECSEL chip consists of an active region and a Distributed Bragg Reflector (DBR), which has tens of pairs of lattice-matched alternating index layers. The high thermal resistance associated with the DBR layer hampers the thermal management of the gain chip and ultimately limits the power of VECSEL. A novel design, called DBR-free VECSEL, was proposed and demonstrated recently [14, 15], in which the semiconductor gain chip without DBR was lifted off from the substrate and Van Der Waals bonded to a transparent crystal. Two external mirrors complete the cavity. For gain-chip testing and screening, various samples were bonded to glass or sapphire substrates and pumped with low-duty cycle pulse lasers to avoid any thermal degradation. Samples studied included GaAs/GaInP DHS and GaInAs MQW gain chips. We encountered lateral lasing in many of these samples under high enough excitation.
The description of the experiment is as follows. The MQW gain chip is grown by metalorganic chemical vapor deposition (MOCVD) and consists of 10 GaInAs quantum wells in a resonant periodic gain structure barriered by GaAs layers. The structure is passivated by AlGaAs layers both at the top and at the bottom. Another thin layer of AlGaAs with high Al content is grown between the substrate and the structure as a release layer for lift off. The photoluminescence of the multi quantum wells peaks at 1001nm. The optical pump source is a miniature diode-pumped Q-switched Er:YAG laser (1535 nm) (Photop Technologies, Model DPQL-1535-C-0040-005N-03) frequency doubled to 767nm using KTP crystal. It delivers ~3.5ns (FWHM) pulses having ~0.6 μJ energy at 1 kHz repetition rate.
Figure 1(a) shows the top-view image of the MQW sample during the experiment; it clearly shows the scattered (laser) light from the chip’s four edges due to lateral lasing along both directions. Figure 1(b) is the spectrum collected from the edge of the sample. The lasing wavelength (shown in Fig. 1(b)) is about 1030nm, much longer than the peak of the gain (λ∼1005 nm) as designed for the vertical lasing operation.
Fig. 1 (a) Quantum well sample (2mm × 2mm square) bonded to substrate showing lateral lasing scattered from chip edges, (b) Spectral measurement from the edge of the sample shows the lateral lasing light emission.
In addition to MQWs, we also investigated GaInP/GaAs/GaInP double heterostructures. These samples were investigated prior to lift-off (i.e. on the epitaxial GaAs substrate) as well as after lift-off and bonding onto a ZnS or glass substrate. Lateral lasing was observed in both cases. Figure 2 shows the results for a 600μm diameter GaInP/GaAs/GaInP (0.75μm/0.75μm/0.75μm thick) DHS bonded to a ZnS substrate (See inset of Fig. 2(a)). The lasing operation is indicated by the observation of a distinct threshold [Fig. 2(a)] at 0.33 μJ, corresponding to a threshold carrier density of, which is obtained by solving the rate equation assuming radiative recombination coefficient B = . Carrier diffusion can be neglected since the estimated diffusion time (28ns) is much longer than the pulse width. Similar to MQW structures, the lasing wavelength at λ∼890 nm is much longer than the peak of the gain (or luminescence) at 860nm [Fig. 2(b)]. The instant temperature rise of the chip at threshold carrier density is estimated to be ~10K, which will only cause the peak of the gain to red-shift by ~2nm. The thermal model thus cannot explain the lateral lasing effect caused by pulsed excitation.
Fig. 2 (a) Lateral lasing threshold for a double heterostructure sample (600µm in diameter) bonded to substrate showing lateral lasing scattered from chip edge (b) Spectral measurement showing the lateral lasing emission.
To confirm that the lasing operation indeed happen in the lateral direction, we investigated samples from the same gain chip prior to lifting off from the GaAs substrate, making it impossible to lase vertically. Similar lasing action was still observed indicating that the lateral lasing action is rather a universal phenomenon in optically (pulse-) pumped semiconductor disk lasers.
As was also reported in [8], it is quite striking that lateral lasing is persistent even when the dimension of the pump area is significantly smaller than the lateral dimension of the sample. As shown in Fig. 3, the square MQW sample is 2mm wide and long, while the pump spot size is only about 50 µm by calculation. However, the use of pulsed excitation in our experiment assures negligible heating within the active region, therefore the thermal model in [8] cannot explain the lateral lasing phenomenon observed in our samples. In the following, we present a plausible mechanism based on bandgap renormalization due to Coulomb screening in the presence of a high concentration of optically injected electron-hole (e-h) density.
Fig. 3 Dimension of the pump area is significantly smaller than the lateral dimension of the sample.
3. The effect of coulomb screening
Coulomb screening is the damping of electric fields caused by the presence of mobile carriers. In semiconductors, when the excited carrier density is high, the effect of Coulomb screening will cause bandgap renormalization. Following the theory of Banyai and Koch [16, 17], the renormalized bandgap is given by:
whereis the unrenormalized semiconductor band-gap energy, is the (exciton) Rydberg energy,, with κ denoting the inverse of Coulomb screening length as determined by:and is the exciton Bohr radius. Here is the reduced mass of the electron-hole pair. The common physical constants used are: = permittivity of free space, = reduced Planck constant, kB = Boltzmann constant, and e = electronic charge. At a given temperature T, the distribution of electrons and holes in the conduction and valence bands are given by the Fermi-Dirac functions:wheredenotes the quasi-Fermi levels. Parabolic bands are assumed with dispersion where is the effective mass for electrons and holes. The variation of chemical potentials with the injected electron-hole density (N) is therefore given by:where F1/2 is the Fermi integral.Using Eq. (1), we can calculate the renormalized band-gap energy for GaAs at room temperature as a function of the excited carrier density, as shown in Fig. 4.
The absorption coefficient is given by:
The contributions from bound and free excitons as given by the plasma theory of Banyai and Koch, are [17]:
where n is the main quantum number for excitons, is the exciton binding energy and is a broadened delta-function having a width Γ that represents phenomenological carrier-phonon interaction. For T∼300K, we assume Γ∼ER [15]. From theory and experiments, α0∼104 cm−1 [18, 19].Figure 5(a) shows the calculated absorption (gain) spectra for the active region for different carrier densities and for the un-pumped passive region of the GaInP/GaAs/GaInP double heterostructure sample at 300K.
Fig. 5 (a) Calculated absorption (gain) spectra for the active region for carrier densities of a) 2 × 1018cm−3, b) 4 × 1018 cm−3, c) 6.5 × 1018cm−3, d) 9 × 1018cm−3, e) 1 × 1019cm−3 and for the un-pumped passive region of the GaInP/GaAs/GaInP double heterostructure sample at 300K. (b) Simulated lateral gain spectrum for different carrier densities. Net gain needs to be greater than 1 to reach lasing threshold condition (above the dashed line).
As can be seen, the renormalized band-gap energy in the active region red-shifts the gain spectrum. When the carrier density is sufficiently high, there is a spectral window in the long-wavelength end of the gain spectrum where the passive (unpumped) region of the sample is nearly transparent. With sufficient feedback from the facets of the sample, lateral lasing will occur.
Based on the absorption and gain spectrum, we can calculate the lateral emission integrated gain spectrum for different carrier densities [Fig. 5(b)] and estimate the lasing wavelength. We calculate the net gain given by R2exp(γD-α(L-D)) where R is the Fresnel reflectivity of the facets, γ is the calculated gain ( = -α from Eq. (5)) in the pumped region and α is the absorption in the passive (N = 0) region. The results are shown in Fig. 5(b) for different excitation carrier densities. The lateral dimension of the wafer (L = 600μm) and diameter of the pumped region (D = 50μm) are shown in the inset of the Fig. 5(b). The result predicts a lasing wavelength of ~887nm when the carrier density is above for the double heterostructure sample, which is in great agreement with our observation as shown in Fig. 2(b).
4. Conclusion
In conclusion, we observe the lateral lasing effect in both optically pumped quantum well and double heterostructure semiconductors and show that the phenomenon can be explained by effect of bandgap renormalization due to Coulomb screening.
References
1. M. Kuznetsov, F. Hakimi, R. Sprague, and A. Mooradian, “Design and characteristics of high-power (>0.5-W CW) diode-pumped vertical-external-cavity surface-emitting semiconductor lasers with circular TEM00 beams,” IEEE J. Sel. Top. Quantum Electron. 5(3), 561–573 (1999). [CrossRef]
2. O. G. Okhotnikov, Semiconductor Disk Lasers: Physics and Technology (Wiley, 2010).
3. B. Heinen, T.-L. Wang, M. Sparenberg, A. Weber, B. Kunert, J. Hader, S. Koch, J. Moloney, M. Koch, and W. Stolz, “106 W continuous-wave output power from vertical-external-cavity surface-emitting laser,” Electron. Lett. 48(9), 516–517 (2012). [CrossRef]
4. K. Iga, F. Koyama, and S. Kinoshita, “Surface emitting semiconductor lasers,” IEEE J. Quantum Electron. 24(9), 1845–1855 (1988). [CrossRef]
5. J. A. Glaze, S. Guch, and J. B. Trenholme, “Parasitic suppression in large aperture Nd:glass disk laser amplifiers,” Appl. Opt. 13(12), 2808–2811 (1974). [CrossRef] [PubMed]
6. J. Shin and Y. H. Lee, “Determination of nonradiative recombination coefficients of vertical‐cavity surface‐emitting lasers from lateral spontaneous emission,” Appl. Phys. Lett. 67(3), 314–316 (1995). [CrossRef]
7. C. Bringer, V. Bardinal, E. Daran, T. Camps, Y. G. Boucher, G. Almuneau, O. Gauthier-Lafaye, P. Dubreuil, J.-B. Doucet, and C. Fontaine, “Detection of lateral spontaneous emission for VCSEL monitoring,” Proc. SPIE 5453, 209–216 (2004). [CrossRef]
8. C. Hessenius, M. Fallahi, J. Moloney, and R. Bedford, “Lateral lasing and ASE reduction in VECSELs,” Proc. SPIE 7919, 791909 (2011). [CrossRef]
9. R. G. Bedford, M. Kolesik, J. L. A. Chilla, M. K. Reed, T. R. Nelson, and J. V. Moloney, “Power-limiting mechanisms in VECSELs,” Proc. SPIE 5814, 199–208 (2005). [CrossRef]
10. S. Kaspar, M. Rattunde, T. Topper, R. Moser, S. Adler, C. Manz, K. Kohler, and J. Wagner, “Recent advances in 2-μm GaSb-base semiconductor disk laser—power scaling, narrow-linewidth and short-pulse operation,” IEEE J. Sel. Top. Quantum Electron. 19(4), 1501908 (2013). [CrossRef]
11. T. Töpper, M. Rattunde, S. Kaspar, R. Moser, C. Manz, K. Köhler, and J. Wagner, “High-power 2.0 μm semiconductor disk laser—influence of lateral lasing,” Appl. Phys. Lett. 100(19), 192107 (2012). [CrossRef]
12. J. Khurgin, “Theoretical and experimental investigation of amplified spontaneous emission in electron-beam-pumped semiconductor lasers,” IEEE J. Quantum Electron. 23(2), 194–204 (1987). [CrossRef]
13. D. Francis, S. DiJaili and J. Walker, “A single-chip linear optical amplifier,” in Optical Fiber Communication Conference and Exhibit (OSA, 2001), paper PD13. [CrossRef]
14. M. Sheik-Bahae, “Evanescent-wave subcavity amplifiers and lasers,” United States Patent 11,845,367 (2007).
15. Z. Yang, A. R. Albrecht, J. G. Cederberg, and M. Sheik-Bahae, “DBR-free optically pumped semiconductor disk lasers,” Proc. SPIE 9349, 934905 (2015). [CrossRef]
16. L. Bányai and S. W. Koch, “A simple theory for the effects of plasma screening on the optical-spectra of highly excited semiconductors,” Z. Phys. B 63(3), 283–291 (1986). [CrossRef]
17. H. Haug and S. Koch, Quantum Theory of the Optical and Electronic Properties of Semiconductors (World Scientific, 1994).
18. H. C. Casey, D. D. Sell, and K. W. Wecht, “Concentration dependence of the absorption coefficient for n− and p−type GaAs between 1.3 and 1.6 eV,” J. Appl. Phys. 46(1), 250–257 (1975). [CrossRef]
19. H. C. Casey and F. Stern, “Concentration‐dependent absorption and spontaneous emission of heavily doped GaAs,” J. Appl. Phys. 47(2), 631–643 (1976). [CrossRef]
