Cathodoluminescence Intensity Research Articles

Rapid identification of rare earth element bearing minerals in ores by cathodoluminescence method

In mineral ores, the identification of monazite, bastnäsite, and xenotime, which are currently exploited rare earth element bearing minerals, is important for the development of the mines that reserve rare earth elements. A conventional method for their identification, an electron probe microanalysis, is time-consuming. Here a method to rapidly identify these minerals is presented by acquiring cathodoluminescence (CL) images and spectra. Monazite, xenotime, and bastnäsite emitted red, blue-green, and red luminescence in their CL images, respectively. We detected CL peaks related to some rare earth elements, such as Ce, Tb, Pr, Sm, and Nd for monazite; Gd, Ce, Tb, Er, Tm, Dy, Sm, and Nd for xenotime; and Pr, Sm, and Nd for bastnäsite using our custom CL spectrometer. Monazite, bastnäsite, and xenotime are distinguishable from other minerals that coexist with these three minerals based on their CL colors, intensities, peak wavelengths, and peak intensity ratios. The analysis time to acquire the CL images and spectra was within 100 s. Therefore, the acquisition of CL images and spectra contributes to the development of the rare earth element reserving mines and the sustainable supply thereof.

Cathodoluminescence study of electric field induced migration of defects in single crystal m-plane ZnO

Internal electric fields can have a significant effect on the behavior of charged defects, dopants, and impurities in operating electronic devices that can adversely impact on their long-term performance and reliability. In this paper, we investigate the redistribution of charged centers in single crystal m-plane ZnO under the action of a DC electric field at 873 K using in-plane and in-depth spatially resolved cathodoluminescence (CL) spectroscopy. The CL intensities of the ultra-violet near band edge (NBE) emission at 3.28 eV and green luminescence (GL) at 2.39 eV were observed to both uniformly increase on the anode side of the electrode gap. Conversely, toward the cathode, the NBE and GL steadily decrease and increase, respectively. The GL quenched after hydrogen donor doping, confirming that the emission is related to acceptor-like centers. Based on the electro-migration and hydrogen doping results, the GL is attributed to radiative recombination involving ZniandVZn pairs. The intensity of an orange luminescence centered at 2.01 eV was unaffected by the electric field and is assigned to substitutional Li acceptors.

Imaging hole-density inhomogeneity in arsenic-doped CdTe thin films by scanning capacitance microscopy

We report nanometer-scale hole-density imaging in As-doped CdTe films by scanning capacitance microscopy (SCM). The hole-density profiling is validated by contrasting a CdTe cross-section sample made by molecular-beam epitaxy with systematically increased As doping in a staircase pattern over the range of 1015–1018/cm3 within a single film. In polycrystalline films, the carrier distribution is significantly nonuniform, with inhomogeneity ranging from several hundred nm to several μm and hole density varying by one order of magnitude (low 1016 to low 1017/cm3). This nonuniformity is distributed randomly, independent of the grain structure and grain boundaries (GBs). Kelvin probe force microscopy (KPFM) maps the surface potential and is correlated to cathodoluminescence (CL) and SCM by imaging over identical areas. Higher potential and lower CL intensity are found on GBs but not by SCM contrast; this indicates positive GB charging and recombination, but not GB-specific hole-density changes. The overall KPFM potential image is in rough agreement with the SCM carrier distribution and indicates band-edge potential fluctuations. Nonuniform carrier concentration, potential fluctuations, and defect recombination can all cause voltage and performance loss in As-doped CdTe solar technology.

Carbon isotopic evolution of aqueous fluids in CM chondrites: Clues from in-situ isotope analyses within calcite grains in Yamato-791198

We report the variability in carbon and oxygen isotopic compositions, chemical compositions, and cathodoluminescence intensities within calcite grains in the Yamato-791198 CM chondrite measured by secondary ion mass spectrometry. To understand the change in carbon isotopic compositions during calcite formation, the carbon isotope-analyses were performed on a series of crystal growth bands of each calcite grain. The crystal growth of calcite grains was inferred from comprehensive analyses of oxygen isotopes, chemical compositions, and cathodoluminescence characteristics.The observed δ18O variations within individual grains are as large as 13‰. The oxygen-isotope data plot on a single straight line with a slope of 0.61 ± 0.06 (2σ) in an oxygen three-isotope diagram. This slope steeper than that of the terrestrial fractionation line indicates that the oxygen isotopic compositions of aqueous fluids evolved from higher δ18O and Δ17O to lower δ18O and Δ17O compositions due to the oxygen-isotope exchange between water and anhydrous silicates in the parent body. Thus, calcite crystals grew from higher Δ17O to lower Δ17O areas. The crystal growth inferred from oxygen isotopic compositions is corroborated by the morphology and cathodoluminescence characteristics of the calcite grains. The minor element concentrations of the calcite grains did not increase/decrease monotonically during calcite formation.The δ13C variations within individual grains are no more than 4‰ except for one grain. The intra-grain δ13C variations observed here are much smaller than inter-grain δ13C variations of ∼80‰ previously reported. These observations indicate that the carbon isotopic compositions of dissolved carbon species did not change during calcite formation and that they were locally heterogeneous which reflects variable proportions of carbon reservoirs with different isotopic compositions.

Ore-forming mechanism and physicochemical evolution of Gutaishan Au deposit, South China: Perspective from quartz geochemistry and fluid inclusions

The Gutaishan is an important gold deposit of the Jiangnan Orogen in South China. The deposit is characterized by multiphase alteration/mineralization, including barren quartz (pre-ore Stage I) quartz + dolomite + pyrite + arsenopyrite (Stage II), quartz + dolomite + native gold + pyrite + arsenopyrite + stibnite (Stage III), and quartz + dolomite ± pyrite (Stage IV).Cathodoluminescence (CL) observations show that each of Stage II and III quartz comprises two sub-stages (Stage II: QII1 and QII2; Stage III: QIII1 and QIII2), and Stage IV quartz (QIV) contains no sub-stages. QIII1 has distinct core-rim texture, with the CL intensity decreasing from core to rim. QII1, QII2, QIII2 and QIV are texturally more homogeneous. Variations of aluminium concentration correlate with the CL intensity and reveal the determinated of pH in fluids during Stage II and III, with marked decline to Stage IV. Fluid inclusions (FIs) trapped in quartz are liquid-rich and CO2-rich, suggesting a CO2-H2O-NaCl fluid system. Fluid inclusion microthermometry data suggest that the ore-forming fluids were under conditions of medium-low salinity (Stage II: 1.02–11.46 wt% NaCl equiv.; Stage III: 0.83–9.84 wt% NaCl equiv.; Stage IV: 4.39–8.10 wt% NaCl equiv.) and temperature (Stage II: 190–285 °C; Stage II: 170–270 °C; Stage II: and 163–240 °C), consistent with the low Ti content (<10 ppm) in the quartz. Fluid boiling, coupled with decrease of temperature, may have controlled gold precipitation at Gutaishan, based on the coexistence of liquid-rich and CO2-rich FIs with similar homogenization temperatures and large salinity variations. Integrating regional tectonic evolution, previous H–O–S–Pb data, and close temporal and spatial relationship with the Baimashan granite rocks, we propose that the Gutaishan gold mineralization may have been genetically linked to the Mesozoic magmatism related to the Baimashan granite rocks formed in an extensional setting during subduction of the Paleo-Pacific plate.

Kinetic properties of YAG: Eu3+ emission upon electron beam excitation

Luminescent properties of yttrium-aluminum garnet activated by different amount of europium (YAG:Eu3+) single crystals were studied by local cathodoluminescence (CL) technique. Samples were grown by flux method. CL spectra were obtained in the range from 400 to 800 nm with spectral resolution of 0.1 nm. The rise and decay kinetics of various CL bands were investigated. The concentration dependencies of CL intensity and decay times of transitions were plotted. The concentration quenching for 5D1-7F2 Eu3+ transition was observed at about 3.5 at.% of Eu by Y, and for 5D2-7F1 and 5D3-7F3 transitions – at about 2.5–3 at.% of Eu by Y. For transition 5D0-7F1 no concentration quenching was observed. The high-energy excitation capture efficiency by different energy levels of Eu3+ in YAG was estimated based on the dependence of the CL rise rate of these levels on the electron beam current density. It was found that 5D1 energy level excitation capture efficiency is about three times higher than that of the 5D0, 5D2 and 5D3 energy levels.

Luminescence behavior of semipolar (101¯1) InGaN/GaN “bow-tie” structures on patterned Si substrates

In this work, we report on the innovative growth of semipolar “bow-tie”-shaped GaN structures containing InGaN/GaN multiple quantum wells (MQWs) and their structural and luminescence characterization. We investigate the impact of growth on patterned (113) Si substrates, which results in the bow-tie cross section with upper surfaces having the (101¯1) orientation. Room temperature cathodoluminescence (CL) hyperspectral imaging reveals two types of extended defects: black spots appearing in intensity images of the GaN near band edge emission and dark lines running parallel in the direction of the Si stripes in MQW intensity images. Electron channeling contrast imaging (ECCI) identifies the black spots as threading dislocations propagating to the inclined (101¯1) surfaces. Line defects in ECCI, propagating in the [12¯10] direction parallel to the Si stripes, are attributed to misfit dislocations (MDs) introduced by glide in the basal (0001) planes at the interfaces of the MQW structure. Identification of these line defects as MDs within the MQWs is only possible because they are revealed as dark lines in the MQW CL intensity images, but not in the GaN intensity images. Low temperature CL spectra exhibit additional emission lines at energies below the GaN bound exciton emission line. These emission lines only appear at the edge or the center of the structures where two (0001) growth fronts meet and coalesce (join of the bow-tie). They are most likely related to basal-plane or prismatic stacking faults or partial dislocations at the GaN/Si interface and the coalescence region.

Влияние отжига в газообразном цинке на люминесценцию в видимом и среднем ИК диапазонах ZnSe : Fe-=SUP=-2+-=/SUP=-

The study results of the effect of annealing in a zinc vapor atmosphere on the defect and impurity composition of iron diffusion-doped ZnSe:Te plates on the characteristics of cathodoluminescence (CL) in the IR range are described. The formation of precipitates at the boundaries of twin faults was found. It is shown that in structures with a surface iron concentration of more than 1 at%, a significant increase in the CL intensity in the IR range is observed as a result of annealing in a zinc vapor atmosphere. A partial recovery of interband luminescence in regions with a high iron concentration as a result of annealing was found.

Two-dimensional analysis of the nonuniform quantum yields of multiple quantum wells for AlGaN-based deep-ultraviolet LEDs grown on AlN templates with dense macrosteps using cathodoluminescence spectroscopy

AlGaN-based deep-ultraviolet light-emitting diodes (LEDs) incorporating uneven multiple quantum wells (MQWs) with inclined and terrace zones, which were fabricated on an AlN template with dense macrosteps, have exhibited a high internal quantum efficiency (IQE). To investigate the microscopic structure of uneven MQWs, cathodoluminescence (CL) mapping characterization was carried out, and the maps of the CL intensity at 300 K relative to that at 38 K were obtained for uneven MQWs that targeted 265 and 285 nm LEDs. At an electron beam current of less than 1.0 nA, the signals from inclined and terrace zones of the uneven MQWs were confirmed to satisfy the nonsaturated excitation condition at 300 K. Nonradiative recombination (NR) was insufficiently frozen even at 38 K, specifically on the terraces in the 265 nm MQW, suggesting high concentrations of NR centers due to point defects (PDs). In contrast, NR in the 285 nm MQW at 38 K was closer to freeze-out. The concentration of PDs in the 285 nm MQW was likely to be lower than that in the 265 nm MQW. Finally, the ratios of the CL intensity at 300 K to those at 38 K were mapped, demonstrating an approach to creating an approximate map of IQE. The values in the CL intensity ratio maps are discussed by considering the analytical error factors. The results support the model of localized current injection through Ga-rich stripe zones in the n-AlGaN cladding layer.

The influence of iron concentration on the cathodoluminescence kinetics in the mid-IR range in ZnSe:Fe crystals

Integral intensity and kinetics dependences of cathodoluminescence of ZnSe:Fe crystals in mid-IR spectral range was studied. Concentration of iron varied from 0.01 to 14 wt.%. The experiments were carried out in 77 - 300 K temperature range. It is established that the luminescence pulse rise time is longer than excitation pulse at low iron concentration at 77 K temperature. The increasing ones lead to shortening of the luminescence pulse rise time, decreasing of the luminescence lifetime and increasing of the luminescence intensity in mid-IR. The annealed ZnSe:Fe concentration series in vapour Zn are reproduced the same trends. The nature of observed experimental effects is discussed.

Ultraviolet cathodoluminescence of pure zinc aluminate ZnAl2O4

The features of the intrinsic cathodoluminescence of pure zinc aluminate ZnAl2O4 having a radiation maximum in the range 250-270 nm are studied. The kinetics of the afterglow of a sample excited by short (∼ 2 ns) pulses is analyzed. Analysis and possible explanation of the influence of the type and composition of the atmosphere, in which zinc aluminate was annealed, is given. Possibilities for a multiple increase in the cathodoluminescence intensity of zinc aluminate for use in bactericidal lamps are demonstrated.

Determination of the Carrier Diffusion Length inGaNfrom Cathodoluminescence Maps Around Threading Dislocations: Fallacies and Opportunities

We investigate, both theoretically and experimentally, the drift, diffusion, and recombination of excitons in the strain field of an edge threading dislocation intersecting the GaN{0001} surface. We calculate and measure hyperspectral cathodoluminescence maps around the dislocation outcrop for temperatures between 10 to 200 K. Contrary to common belief, the cathodoluminescence intensity contrast is only weakly affected by exciton diffusion, but is caused primarily by exciton dissociation in the piezoelectric field at the dislocation outcrop. Hence, the extension of the dark spots around dislocations in the luminescence maps cannot be used to determine the exciton diffusion length. However, the cathodoluminescence energy contrast, reflecting the local bandgap variation in the dislocation strain field, does sensitively depend on the exciton diffusion length and hence enables its experimental determination.

Cathodoluminescence degradation of Y2O3:Dy3+ nanophosphor for field emission displays

Cathodoluminescence (CL) degradation of Y2O3:Dy3+ nanophosphors prepared by the solution combustion method was explored for feasible applications in low voltage field emission displays (FEDs). Oxide materials are excellent candidates for FED fabrication due to their high melting points, chemical and radiation stability with long lifetimes, high color purity, and being environmentally friendly. Auger electron spectroscopy (AES) and x-ray photoelectron spectroscopy (XPS) were used to monitor changes in the surface chemical composition and correlation fit with CL degradation. AES and CL spectroscopy (2 keV energy electrons and with a beam current of 15 μA) measurements were done in high vacuum (1.5 × 10−8 Torr) and oxygen pressures of 1 × 10−7 and 5 × 10−7 Torr. The Y2O3:Dy3+ nanophosphor showed strong yellow (572 nm) and relatively weaker blue (492 nm) CL emissions. These CL emissions increased as carbon (C) was depleted from the surface, and then it slightly decreased at a high electron dose in both the vacuum and oxygen atmospheres, for electron doses up to about 690 C/cm2. The C was depleted from the surface due to electron stimulated reactions. No significant change in the chemical state of Y 3d was observed with XPS high resolution spectra for the postirradiation sample. The change in the CL intensity was, therefore, attributed to the depletion of C from the surface.

Rapid and Simple Identification of Free Magnesia in Steelmaking Slag Used for Road Construction Using Cathodoluminescence

Reuse of steelmaking slags is important for the effective use of natural resources. Free magnesia (f-MgO) in steelmaking slag may cause serious problems because of a hydration reaction followed by expansion when it is reused for road construction. We present a promising method to identify f-MgO that causes volume expansion rapidly by investigating cathodoluminescence (CL) images and spectra of a steelmaking slag sample. f-MgO emitted red–orange luminescence from a peak at 755 nm. The mineral phases, 3CaO·SiO2 and 2CaO·SiO2, emitted red and yellow luminescence from peaks at 720 and 590 nm, respectively. No luminescence of FeO and 2CaO·Fe2O3 was detected. f-MgO changed its composition in the slag sample that was immersed in hot (70 °C) water for a week. f-MgO that was responsible for the volume expansion (combined content of FeO and MnO below 30 mass pct) retained a red–orange luminescence, whereas the other f-MgO lost luminescence. The CL intensity of the f-MgO that retained luminescence was more than 10 times larger than that of 3CaO·SiO2 and 2CaO·SiO2. Therefore, we can distinguish f-MgO that causes volume expansion by detecting the intense red–orange luminescence from the peak at 755 nm in the CL image within a few seconds.

Cathodoluminescence and trace elements of scheelite: Constraints on ore-forming processes of the Dabaoshan porphyry Mo-W deposit, South China

Scheelite, a good proxy to record the complex history of hydrothermal evolution, has been investigated in the Dabaoshan porphyry Mo-W deposit, South China. Based on cathodoluminescence (CL) textures, mineral paragenesis, and chemical characterization, three generations of scheelite were identified (Sch1–Sch3). Sch1 is disseminated in porphyry rocks, or occasionally occurs in early quartz-molybdenite veins. It is CL-dark and contains high concentrations of V, Sr, Mo, Nb, Ta, and rare earth elements (REE), low U/Th ratios (0.37 mean) and light REE (LREE)-enriched patterns with negative Eu anomalies. Sch2 and Sch3 mainly occur in late quartz-pyrite-scheelite ± chalcopyrite veins. Sch2 is CL-bright and commonly rims Sch1, and/or is overgrown by late oscillatory Sch3. Compared with Sch1, Sch2 shows depletion in V, Sr, Mo, Nb, and Ta, but slight enrichment in U and Th, and is characterized by middle REE (MREE)-enriched patterns with negative Eu anomalies. Sch3 grains show reddish-brown color under transmitted light and exhibit oscillatory zonation under CL. Sch3 has elevated concentrations of Mn, U (up to 4484 ppm), and Th (up to 18.8 ppm) contents, and high U/Th ratios (1175 mean). Chemical mapping reveals a negative correlation between U concentrations and CL intensity, which is attributed to the radiation damage caused by the radioactive decay of U. Based on the REE patterns, Sch3 is subdivided into Sch3a and Sch3b. Sch3a shows MREE-enriched patterns with slight negative to positive Eu anomalies, while Sch3b is LREE-enriched and shows gradual depletion in MREE and HREE with increasing positive Eu anomalies in consecutive growth zones.The Dabaoshan porphyry and all generations of scheelites show similar Y/Ho ratios, indicating that ore-forming fluids were mainly derived from the Dabaoshan porphyry. Our results reveal that REE chiefly substitute into scheelite by the mechanism Ca2+ = REE3+ + Na+ while U substitutes for W into scheelite. Sch1 containing the highest concentrations of Mo, V, Nb, and Ta precipitates from the magmatic fluids at high temperature. It inherits the LREE-enriched patterns and negative Eu anomalies from the causative porphyry. From early to late generations, the gradual depletion of Mo, Nb, Ta, and V in scheelite is ascribed to the precipitation of molybdenite, rutile, and scheelite, whereas the increasing concentrations of U and Th in scheelite indicates the gradual enrichment of U and Th in ore-forming fluids. The increasing U/Th ratios in scheelites may result from not only the differential solubility between U and Th in fluids, but also the preferential incorporation of U relative to Th in scheelite. Compared with Sch1, Sch2 and Sch3a are depleted in LREE, which is interpreted to reflect the precipitation of apatite. Sch3b with various REE patterns within single grains is interpreted as precipitation from a closed-system accompanied by intermittent fluid replenishment. This study demonstrates that REE patterns of scheelite are controlled by multiple factors, including substitution mechanism, fluid composition, hydrothermal alteration, precipitation of REE-rich minerals, and hydrothermal dynamics during scheelite growth. The systematic variation of trace elements in scheelite suggests that scheelite chemistry can be used to monitor the evolution of ore-forming fluids in the Dabaoshan deposit.

Chemistry of quartz – A new insight into the origin of the Orlovka Ta-Li deposit, Eastern Transbaikalia, Russia

The Ta-Li deposit of Orlovka in Eastern Transbaikalia, Russia, is one of the best known rare-metal deposits in cupola-shaped plutons of Li, F-rich rare-metal granites. Quartz from all granite varieties from the deposit was studied using automated mineralogy (TIMA), cathodoluminescence (CL), and laser-ablation inductively-coupled mass spectrometry (LA-ICP MS) in order to find new constraints on the origin of the deposit. Subhedral to euhedral quartz crystals, strongly zoned in CL, relatively enriched in Ti and Li (up to 95 and 33 ppm respectively), and free of mineral inclusions were found in all varieties of biotite granites forming the deeper part of the pluton. Crystals of “snow-ball” quartz with zonally arranged inclusions of albite, with low intensity of CL, and slightly enriched in Al and Ge (up to 350 and 4.8 ppm, respectively) are common in all types of Li-mica granites forming the cupola-shaped upper part of the pluton. The distinction between the two principal types of quartz is generally consistent with the already proposed model of two comagmatic intrusions: deeper-seated biotite granites underwent a complicated magmatic evolution with repeated episodes of crystallization and resorption of quartz, only a limited degree of fractionation, and a small – if any – reaction with fluid. Li-mica granites, taking position within the cupola, strongly fractionated upwards. The upward transfer of Li and rare metals was, moreover, supported by fluid unmixed in the middle level of the cupola.

Influence of sol–gel-derived ZnO:Al coating on luminescent properties of Y2O3:Eu3+ phosphor

This study deposited transparent and conductive Al-doped ZnO (AZO) thin coating on the surface of Y2O3:Eu3+ phosphor particle via a sol–gel method and analyzed the influence of soaking pretreatment, sol concentration, stabilizers, and coating temperature on the microstructure and luminescent properties of AZO-coated Y2O3:Eu3+ phosphors with the aid of SEM, TEM, cathodoluminescence (CL), and photoluminescence (PL) measurements. A water soaking pretreatment was found to effectively improve the coverage and uniformity of the AZO coating, resulting in a continuous and coherent coating. This result is attributed to the presence of adsorbed water and hydroxyl groups on the Y2O3:Eu3+ particle surface after water soaking. Monoethanolamine as a stabilizer agent and higher sol concentration led to a thicker coating on the phosphors. At 800 V or lower, the CL intensity of AZO coated Y2O3:Eu3+ phosphors is significantly greater than that of the uncoated particles and increases with coating thickness. In addition, the PL emission intensity also rises with increasing coating thickness and is even higher than bare phosphors. Further increases in coating thickness beyond the optimum value led to a decrease in PL intensity. This study thus proposes that AZO coatings with sufficient thickness and surface coverage bring about in a higher electron/UV light transport and less surface defects, resulting in an improvement in emission intensity.

Investigation of a Contamination Film Formed by the Electron Beam Irradiation

In the study of materials on electron probe devices in the field of action of the electron beam, a contaminating hydrocarbon film is formed, which affects the experimental results. In this paper, we have studied the influence of a contamination film on carbon-film-coated dielectrics on the intensity of cathodoluminescence and characteristic X-ray lines. The absorption coefficient of the contamination film in the visible and UV ranges has been determined. Filming mechanisms at different parameters of the electron beam have been discussed.

Influence of Mg‐to‐Ce Concentration Ratio on Cathodoluminescence in LuAG and LuGAGG Single‐Crystalline Films

The cathodoluminescence (CL) characterizations of Lu3Al5O12:Ce,Mg (LuAG:Ce,Mg) and of multicomponent (Lu0.25Gd0.74)3(Ga2.48Al2.52)O12:Ce,Mg (LuGAGG:Ce,Mg) single‐crystalline garnet films are supported by the optical absorption spectroscopy. Using this approach, the influence of the Mg‐to‐Ce concentration ratio on the CL spectra and the CL decay is analyzed. The crucial role of stable Ce4+ centers in CL is shown in Mg‐rich film studies. Although the CL intensity is somewhat reduced, drastic improvements in timing performance are demonstrated, including the CL decay time as low as 4.2 ns and the afterglow as low as 0.015% at 500 ns after the e‐beam excitation cutoff. These results predetermine the utilization of the LuAG:Ce,Mg and LuGAGG:Ce,Mg single‐crystalline films in e‐beam detection systems, where especially fast scintillator response is crucial.

Temperature-dependent cathodoluminescence mapping of InGaN epitaxial layers with different In compositions

We performed temperature-dependent cathodoluminescence (CL) mapping of InGaN epitaxial layers, for which the presence of potential barriers around threading dislocations was previously confirmed by photoluminescence (PL) mapping using scanning near-field optical microscopy. The dark spot contrast in the panchromatic CL intensity images decreased with increasing temperature owing to the activation of nonradiative recombination processes. The change in the distribution of monochromatic CL (MCL) intensity with increasing temperature was ascribed to a carrier diffusion process in which thermally activated carriers diffuse to the potential minima. The change in the MCL intensity images of the InGaN epitaxial layer with higher In composition at higher temperatures was remarkable, indicating that the carriers cannot readily move to the potential minima owing to the higher potential barrier. This tendency is consistent with the temperature dependence of the intensity ratio of the two emission components previously observed in macroscopic PL measurements.