Band-tail Model Research Articles

Photoluminescence Study on InGaN/GaN Quantum Well Structure Grown on (112̄0) Sapphire Substrate

An investigation of temperature-dependent photoluminescence is carried out on InGaN/GaN multiple-quantum-well (MQW) structures with 2.5 nm well thickness grown on (0001) and (112̄0) orientation sapphire substrates. Based on a band-tail model, the exciton localization effect is investigated with regard to the substrate orientation. On the (0001) sapphire substrate, the emission energy decreases monotonically with increasing temperature from 10 K to room temperature. However, on the (112̄0) sapphire substrate, the emission energy shows a temperature-induced blue shift at temperatures above 70 K, which is a typical characteristic of the occurrence of the exciton localization effect in an InGaN/GaN quantum well structure. Therefore, our result indicates that the exciton localization effect can be enhanced by growing InGaN/GaN MQW structures on (112̄0) sapphire substrates. Since the exciton localization effect is generally accepted to be advantageous for enhancing the emission quantum efficiency of the InGaN/GaN MQW, the results presented in this paper should be seriously considered.

Influence of the quantum-well thickness on the radiative recombination of InGaN/GaN quantum well structures

Temperature-dependent photoluminescence (PL) measurements are performed on In0.23Ga0.77N/GaN single-quantum-well structures with different well thickness. Based on a band-tail model, the exciton localization effect is studied. The exciton localization effect is enhanced by increasing quantum-well thickness up to 2.5 nm. If the quantum-well thickness is further increased to above 2.5 nm, the exciton localization effect becomes weak. Finally, when the quantum-well thickness is increased to 5 nm, the exciton localization effect cannot be observed. In addition, the PL intensity decreases monotonically with increasing the quantum-well thickness. In connection with an excitation-power dependent PL measurement, the result of the quantum-well thickness dependent PL intensity can be attributed to quantum confined Stark effect, which becomes particularly strong in the wide quantum-well structure. Based on our optical investigation, the presented article indicates that the emission mechanism is dominated by the exciton localization effect in the thin quantum-well structures, while the quantum confined Stark effect dominates the radiative recombination in the wide quantum-well structures. Since understanding the emission mechanism is very important for further improving the performance of an InGaN/GaN-based optical device, the presented results in this article should be highly emphasized.

Effect of silicon doping on the optical and transport properties of InGaN/GaN multiple-quantum-well structures

Temperature-dependent photoluminescence and transport measurements were performed on the In0.13Ga0.87N:Si/GaN:Si multiple-quantum-well (MQW) structures with different doping levels. By fitting the temperature-dependent emission energy of these samples using the band tail model, an obvious localization effect is observed in lightly doped MQW structures. Correspondingly, the electron mobilities in these structures are significantly higher than those of undoped and heavily doped MQW structures. Furthermore, when the localization effect is stronger, the mobility is higher.

Band-tail model and temperature-induced blue-shift in photoluminescence spectra of InxGa1-xN grown on sapphire

A band-tail model of inhomogeneously broadened radiative recombination is presented and applied to interpret experimental data on photoluminescence of various bulk and quantum-well epitaxial InGaN/GaN structures grown by MOCVD. The temperature dependence of the spectral peak position is analyzed according to the model, explaining the anomalous temperature-induced blue spectral shift. Significant differences are observed between epilayers grown on sapphire substrates and on GaN substrates prepared by the sublimation method. No apparent evidence of band tails in homoepitaxial structures indicates their higher crystalline quality.

Radiative and nonradiative recombination in polymerlikea−C:Hfilms

Localized electronic states corresponding to \ensuremath{\pi} electrons in hydrogenated amorphous carbon have been investigated using photoluminescence (PL) and optical-absorption spectroscopies. Carbon films contain \ensuremath{\pi}-bonded ``grains'' with different configurations which spatially confine photogenerated electron-hole pairs. A model of PL in $a\ensuremath{-}\mathrm{C}:\mathrm{H},$ including excitonlike radiative recombination and electron tunneling to nonradiative acceptor centers at constant energy, is proposed to explain the following results. The broad PL emission spectra are composed of three bands (at 2.20--2.40 eV, 2.65 eV, and 2.95 eV, independent of the excitation energy) arising from different structural units containing $\mathrm{C}\mathrm{=}\mathrm{C}$ double bonds. Resonance features have been evidenced in the PL excitation spectra and attributed to excitation of confined electron-hole pairs similar to confined exciton optical transitions. Excitonic behavior is consistent with the redshift of peak (b) excitation resonance observed as a function of increasing gap (decreasing \ensuremath{\varepsilon}). Relative values of the radiative rates deduced from PL efficiency indicate that the exciton radius has the same order of magnitude for all three emissive centers. Nonradiative recombination cannot be explained by multiphonon emission processes. The dependence of PL efficiency on the density of acceptor sites (deduced from $\ensuremath{\pi}\ensuremath{-}{\ensuremath{\pi}}^{*}$ optical transitions) can rather be described by electron tunneling from a confined exciton toward nonradiative localized ${\ensuremath{\pi}}^{*}$ states arising from distorted \ensuremath{\pi}-bonded sites. This mechanism accounts for the anticorrelation of the PL efficiency with the optical gap, the decrease of the decay time as a function of emission energy, and the relative enhancement of the blue emission (2.95-eV peak) as a function of decreasing optical gap. In contrast, the band-tail model proposed earlier leads to unrealistically low values of the electron Bohr radius.

Experimental and theoretical photoluminescence study of heavily carbon doped GaAs grown by low-pressure metalorganic chemical vapor deposition

Carbon (C) doped GaAs epilayers grown by low-pressure metalorganic chemical vapor deposition have been analyzed by photoluminescence as functions of hole concentration and temperature. We have obtained an empirical relation of the effective band gap shrinkage for C doped GaAs as a function of hole concentration, ΔEg=−2.17×10−8p1/3. The temperature dependent band gap energy of C doped GaAs with a hole concentration of 9.0×1019 cm−3 has been analyzed. The resulting band gap at 0 K is (1.422±0.005) eV. The photoluminescence spectra have been analyzed on the bases of the Halperin and Lax bandtail model and an effective energy dependent optical transition matrix element. We have shown that the intensity distribution position of the photoluminescence spectrum of the conduction band to the heavy hole valence band transition is clearly separated from the intensity distribution of the conduction band to the light hole valence band transition. The conduction band to the light hole valence band transition is very prominent in the photoluminescence emission spectra of heavily C doped GaAs, and particularly at low temperatures. We have suggested that the shoulder peak which appears on the higher energy side of photoluminescence spectra is mainly due to the conduction band to the light hole valence band transition. At room temperature, the shoulder peak disappears and the full width at half maximum of the photoluminescence spectra increases; this is in good agreement with experimental results.

Activation energy of source-drain current in hydrogenated and unhydrogenated polysilicon thin-film transistors

The activation energy of the drain current in polysilicon thin-film transistors (TFTs) and the effects of hydrogenation on this energy are discussed. The activation energy data are fitted using different models of the density of states in the material. It is shown that a model which assumes a distribution of brand tail states and localized deep states can account for the activation energy data of unhydrogenated polysilicon TFTs. However, the activation energy data on hydrogenated TFTs cannot be explained with the band tail model. Instead, a simple model of deep states localized at the grain boundary can fit this data quite accurately. Also, it is shown that there is a characteristic kink in the activation energy data of the hydrogenated TFTs which is a signature of the location of the deep states relative to the valence band edge. Analysis indicates that these deep states are located approximately 0.36 eV from the valence band edge. This value is consistent with that obtained from absorption measurements using photothermal deflection spectroscopy.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

The effect of band-tails on the design of GaAs/AlGaAs bipolar transistor carrier-injected optical modulator/switch

The variations in the absorption coefficient and the refractive index in the waveguide region of a double-heterostructure N-AlGaAs/p-GaAs/N-AlGaAs bipolar transistor carrier-injected optical modulator/switch are calculated theoretically using a band-tail model. The refractive index change is calculated by a Kramers-Kronig analysis of the absorption curve, which changes in the vicinity of the absorption edge primarily because of the effect of band-filling, band-gap shrinkage, and plasma dispersion associated with the presence of free carriers. The performance of bipolar transistor carrier-injected optical modulator/switch is analyzed on the basis of the calculated results, including the alpha-parameter and the maximum modulation depth. It is shown that small frequency chirping and high modulation depth with low absorption loss can be achieved by properly selecting the optimum operating conditions. >

EFFECT OF ANNEALING ON THE BANDTAIL STATES IN a-SixC1-x:H FILMS

Photoluminescence (PL) spectra of annealed a-SixC1-x:H films were studied. The-widths of the exponential bandtails were determined by means of the temperature dependences of PL peak energy positions. The experimental results shows that PL peak energy positions shift towards the lower energy with annealing temperature increasing. The bandtail widths have been calculated according to the exponential bandtail model, the calculation shows that the bandtails contract with increase of annealing temperature. A discussion about the results has been given.

Dispersive band tail relaxation in amorphous semiconductors

Within the novel non-equilibrium theory of incoherent motion we study the current and energy relaxation of excited charge carriers as observed in transient photoconductivity experiments. The theory is applied to an exponential band tail model commonly used for a-As 2Se 3 and a-Si:H. The typical temperature dependence of the decay parameter can be explained in the whole temperature regime and the mean energy loss as a function of initial energy is calculated.

Study of recombination of photocarriers in a-As 2Se 3 and As 2Te 3 by photoconductivity and photo-absorption measurements

Time-dependent photoconductivity and photo-absorption measurements, the latter using probe beams of 0.1 eV and 1.4 eV, were analyzed on the basis of a saturated band tail model. By a best fit procedure of theory to experiment band structure and recombination parameters for a-As 2Se 3 were obtained.

Automatic frequency control in a semiconductor laser and an optical amplifier

Automatic frequency control (AFC) in an injection locked or resonant type amplifier in an AlGaAs semiconductor laser was achieved through using the terminal voltage change induced by light injection. Signal-to-noise ratio in the control signal of 10 dB was obtained when the input optical power was -47 dBm and the optical gain Was 51 dB. The AFC was maintained for 3 h with an 0.3-percent output power fluctuation for 2°C ambient temperature change and 65-MHz frequency stability. Step response showed that the system response time was 1.5 s. Sensitivity to input optical power deteriorates at -49 dBm, with a 53-dB locking gain, because of frequency deviation caused by temperature modulation. The second derivative of the induced voltage and it's relation to the optical frequency is constant at 5 \times 10^{-10} [V/(MHz)2] for all input power levels in a buried-heterostructure (BH)-AlGaAs laser. Terminal voltage change induced by light injection is calculated by simple rate equations with a Gaussian-Halperin-Lax (GHL) bandtail model. Good agreement with experimental results was seen.

Noise in an AlGaAs Semiconductor Laser Amplifier

The noise characteristics in a Fabry-Perot (FP) cavity type semiconductor laser amplifier, biased at just below its oscillation thresh-old current, have been studied theoretically and experimentally. Quantum mechanical multimode rate equations containing a Langevin shot noise source and an input signal term were numerically solved for an exponential band-tail model with no k-selection rule. Noise power calculated using this rate equation was compared with a simpler photon statistic master equation method. The experimental results on noise power for an AlGaAs laser amplifier are in reasonable agreement with the two different theoretical predictions. Dominant noise powers in a semiconductor laser amplifier are beat noise powers between signal and spontaneous emission, and between spontaneous emission components. Noise characteristics in a Fabry-Perot cavity type laser amplifier can be improved both by the reduction of the facet mirror reflectivities and by use of an asymmetric cavity configuration with low-input and high-output mirror reflectivities. Two beat noise powers are expressed in simple analytic form by introducing an equivalent noise bandwidth and an excess noise coefficient as figures of merit in an optical amplifier.

Noise in an AlGaAs semiconductor laser amplifier

The noise characteristics in a Fabry-Perot (FP) cavity type semiconductor laser amplifier, biased at just below its oscillation threshold current, have been studied theoretically and experimentally. Quantum mechanical multimode rate equations containing a Langevin shot noise source and an input signal term were numerically solved for an exponential band-tail model with no <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">k</tex> -selection rule. Noise power calculated using this rate equation was compared with a simpler photon statistic master equation method. The experimental results on noise power for an AlGaAs laser amplifier are in reasonable agreement with the two different theoretical predictions. Dominant noise powers in a semiconductor laser amplifier are beat noise powers between signal and spontaneous emission, and between spontaneous emission components. Noise characteristics in a Fabry-Perot cavity type laser amplifier can be improved both by the reduction of the facet mirror reflectivities and by use of an asymmetric cavity configuration with low-input and high-output mirror reflectivities. Two beat noise powers are expressed in simple analytic form by introducing an equivalent noise bandwidth and an excess noise coefficient as figures of merit in an optical amplifier.

Gain, frequency bandwidth, and saturation output power of AlGaAs DH laser amplifiers

Signal gain, frequency bandwidth, and saturation output power, which are important parameters in determining preamplifier and linear repeater system performance, are studied theoretically and experimentally for AlGaAs Fabry-Perot cavity type laser amplifiers. Multimode rate equations with an input signal term are solved for the exponential bandtail model with no k -selection rule. Experimental results are in reasonable agreement with theoretical predictions. The maximum signal gain, frequency bandwidth, and the tolerance of the optical center frequency control are determined by facet mirror reflectivity and cavity length. The unsaturated signal gain increases with the pumping rate and the maximum signal gain as high as 21-28 dB is obtained near oscillation threshold for various cavity structures. The frequency bandwidth, at which signal gain is decreased by 3 dB from the low frequency value, is 1 GHz at 21 dB signal gain for a 900μm cavity length amplifier. The saturation output power, at which the signal gain is decreased by 3 dB from the unsaturated value, is -10 to -5 dBm. The saturation output power will be improved by 18 dB in a traveling wave type amplifier.

The impurity density-of-states tails in neutral semiconductors: application to the diffusion current in GaSb p-n junctions

A simplified approach of the Halperin and Lax model of the density-of-states tails in heavily doped semiconductors is presented: the trial wavefunctions are assumed to be hydrogen-like ones. Analytical expressions of the density-of-states tails are obtained, in very good agreement with the computed Halperin and Lax results.The connection of the density-of-states tail with the unperturbed parabolic band is then achieved in the same way as previously done by Hwang, and numerical values are calculated for various doping levels in GaSb. The influence of the band tails on the electrical I-V characteristic, and on the optical luminescence spectra, is calculated. Good agreement is shown with experimental results on GaSb p-n junctions. Thus valuable information about the location and spectral properties of the recombinations in p-n junctions can be deduced from this simplified band-tail model.

Optical Absorption Edge of Compensated Germanium

The optical absorption of heavily doped and closely compensated Ge is calculated using a Halperin-Lax band-tail model and an ad hoc matrix element. The only adjustable parameter is an effective temperature ${T}_{\mathrm{ion}}$ for the contribution of the position correlations of the ions to the screening; ${T}_{\mathrm{ion}}$ is estimated to be about 7000 \ifmmode^\circ\else\textdegree\fi{}K. The results are in reasonable agreement with experiment; direct transitions at $k=0$ make the dominant contribution to the edge.

Band-Tail Model for Optical Absorption and for the Mobility Edge in Amorphous Silicon

Band-Tail Model for Optical Absorption and for the Mobility Edge in Amorphous Silicon