Experimental results and laser characteristics discussion

The sample was cleaved to form a 1 mm long cavity and optically pumped by a 1064nm (~1.2eV) pulsed fiber laser shaped via two cylindrical lenses, resulting in a beam size of ~ 200m (width) x 3mm (length). The long axis of the pump beam was aligned in parallel with the cavity direction. The pumping light penetrated the thick InAlAs (~1.5eV) cladding layer and was absorbed by the active region. Photo-generated carriers then down converts the pumping light into mid-IR light through the radiative recombination in the “W” QWs. The light output from the cleaved mirror went through a long-pass filter and two focusing lenses.

It was then dispersed through a monochromator and detected by a thermal electric cooled InGaAsSb photodetector. The room temperature lasing behavior is shown in Fig. 5.4. The L-L curve was measured under pulsed operation with 20 ns pumping pulses at a 100 Hz repetition rate. The lasing phenomenon is confirmed by the dramatically increased output power as well as the abruptly shrunk spectrum after a threshold power density (Pth) of ~40 kW/cm². Using separate measurements of the transmission and reflection of samples with and without the

Position along growth direction (m) Near field

Fig. 5.3 The simulated near field TE0 mode and wave guide index profile. The inset figure is far field pattern versus vertical angle.

71

region. The transmission spectrum and the reflection spectrum are shown in Fig. 5.5. Taking into account of the surface reflection and the percentage of the power absorbed, the absorbed threshold power density per well is ~0.467kW/cm². The spectra at 0.98P_th and 1.02P_th are shown in the inset of Fig. 5.4. The lasing wavelength is at 2.56 µm corresponding to the transition between the lowest E1 state and the first quantized heavy hole state (HH1).

Fig. 5.4 The L-L curve of a 1mm long “W” type laser at room temperature.

The inset figure is the lasing spectra at ~0.98Pth (Intensity 50x) and ~1.02Pth.

72

We have measured L-L curves of lasers with different cavity length (L) to extract internal loss (i)at room temperature. The obtained external quantum efficiency (d') can be described as the actual external quantum efficiency (d) multiplied by a correction factor C accounted for the imperfect light collection. Although C is unknown, the plot of 1/d' versus L can still be used to derive i based on the relation d'=C*d=C*i /(1+i/ln(1/R)*L), with i referring to internal quantum efficiency and R referring to the facet reflectivity (here ~0.27). The derived i is around 10.7cm^-1(in below Fig 5.6.).

800 900 1000 1100 1200 1300 1400 0

800 900 1000 1100 1200 1300 1400 0

Fig. 5.5 Transmission, reflection, and absorption spectra of (a) n+ InP substrate only indicated with 17.3% absorption at 1064nm pumping wavelength, and (b) “W” laser structure on n+ InP substrate indicated with 42.2% absorption at 1064nm. ~25% absorption of active region is estimated, based on the calculation of 42.2%-17.3%.

73

In order to further characterize our laser, we have measured the lasing performance at different temperatures. The temperature dependent L-L curves are plotted in Fig. 5.7. The extracted Pth as a function of temperature is shown in the inset (in log scale). It is found that the slope efficiency hardly changes for temperatures below 250K. The slowly increased Pth

from 78K to 250K has a exponentially fitted characteristic temperature (T₀) of 487.8K.

However, when the temperature is increased above 250K, the slope efficiency decreases and the threshold power increases at a much faster rate. The characteristic temperature is 41.8K

Fig. 5.6 (a) The L-L curves measured with different cavity length, and (b) the plot of inverse external quantum efficiency (d') versus cavity length (L) to extract internal loss (i).

74

when the temperature goes beyond 250K. It is known that the monomolecular (Shockley-Read-Hall) recombination and the radiative recombination are dominating processes at low temperatures. But they could not cause this T₀ change behavior because the Shockley-Read-Hall coefficient varies with ~T^1/2 and the radiative coefficient gradually decays with T. According to the theoretical analysis of ref.[53], the integrated PL intensity should decay with T^-2 as temperature goes up if the defect related Shockley-Read recombination dominates the whole process. Our results, however, is quite different. The comparison between our measured result and the theoretical T^-2 curve is shown in Fig. 5.8.

The intensity decays at a much slower rate than what is predicted by the Schottky-Read-Hall process. The leakages due to escaped carriers should also not play a role here since both the E1 level in the conduction band and the HH1 level in the valence band are more than 300meV deep compared to the surrounding potential barriers. The most probable reason for the rapid increase of the threshold power when temperature goes beyond ~250K is the Auger

Fig. 5.7 L -L curves measured at different temperatures. The inset shows the plot of Pth as a function of temperature.

75

The amount of the Auger effect for our laser is analyzed in the following. We first calculated the laser’s modal gain as a function of carrier density for multiple “W” quantum wells using the 8–band k.p method including a Lorenzian line broadening function (=meV) and a confinement factor (17.6%) based on previous wave guide simulation. The result of the modal gain spectra with 2D carrier density varied from 1 x10¹²cm^-2 to 3 x10¹²cm^-2 are shown in Fig. 5.9. The theoretically obtained 2D transparency carrier density was ~1.1x10¹²cm^-2. Based on the measured internal loss, 1mm long cavity mirror loss and the calculated modal gain, we estimated a 2D threshold density around 1.5x10¹²cm^-2 per well at room temperature.

We can estimate the carrier life time (h) at threshold using the relation, th ≈ Nthℏωp/Pthfab(1-R), where Nth, ℏωp, fab, and R refer to the 2D threshold current density, the photon energy of the pump beam, the absorbed fraction, and the surface reflectivity, respectively. The obtained th is in the range of ~0.6 ns. This will lead to an Auger coefficient,

0 50 100 150 200 250 300 0.01

0.1 1

Our "W" QWsintegrated PL intensity

T-2 curve (theory)for Shockley Read behavior

In tensity (n orma li zed)

Temperature (K)

Fig. 5.8 The comparison between our “W” QWs integrated intensity changed along with temperature and the T^-2 curve for the case of Shockley Read process dominated.

76

th(Nth/d)²), of 1.67 x10^-27 cm⁶/s at room temperature, where d of 15nm is used here for the thickness of one period “W” QW. This relatively high Auger coefficient explains why T0

goes down at high temperatures. The value obtained is comparable to those in InAs/InGaSb type-II quantum wells on GaSb substrate, which also show similar T0 at room temperatures [57, 58].

The Auger processes often start to become dominated in a higher carrier density which is estimated around 5x10¹¹ cm^-2 for the “W” laser discussed above according to the threshold value at 250K. Not all “W” lasers we grown posse such a large T0 in low temperature and an obvious T0 change behavior as the operation temperature is raised. The temperature dependent L-L curves and the laser T0 value versus operation temperature form the “W” laser sample named Rn1024, are shown in Fig. 5.10(a) and Fig. 5.10(b). The “W” structure layer thicknesses are the same as the precious one, however with Sb fraction around 0.65 in GaAsSb and total with 30 stacks of “W”QWs in the active region. The highest operation

0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 -200

-100 0 100 200

c~17.6%

for mutiple QWs

Modal gain

c g

cm -1

Energy (eV)

carrier 2D density (cm-2) from 1e12.. to 3e12 step 0.25e12

Total loss

Fig. 5.9 Modal gain calculations based on the 8-band k.p theory with 2D carrier density varied from 1 x10¹²cm^-2 to 3 x10¹²cm^-2. The simulated confinement factor is ~17.6% and the total optical loss is indicated by the dash line.

77

temperature is up to 87^oC with lasing wavelength at 2.42 m shown in the inset of Fig.

5.10(b). We can see that the T0 value in low temperature is 113.6K, which is smaller than the previous one, and is 41.7K around RT, which is similar to the previous sample. The T₀ change behavior is smoother and still around 250K. We found that the T0 value changes to 22.2K as the operation temperature is raised above 340K. These phenomena, different T₀ behavior µm with a threshold pumping power density of ~40kW/cm². The laser shows a characteristic temperature (T0) of 487.8K when it is operated below 250K and a T0 of 41.8K near room temperature. The small T0 at room temperature is considered due to the dominated Auger processes. An Auger coefficient of 1.67 x10^-27 cm⁶/s was estimated. More studies are needed to improve the laser performance. We believe that the “W” structure on InP substrates are promising for the fabrication of mid-IR optoelectronics devices.

0 50 100 150 200 250 300 350 400

Fig. 5.10(a) Temperature dependent L -L curves for the “W” laser named Rn1024, and (b) the T0 valueversus the laser operation temperature. The inset shows the lasing spectrum at the highest operation temperature with a lasing peak at 2.42m.