ETE:S a In this way, ompared relative threshold ratios for 13 and 26MQW structures by determining the quality factor, Q, and cor factor,I, using three-dimensional finite-difference time-domain(3D-FDTD) calculations because other parameters remain approximately constant. The modes with the highest Q(Fig 4b)exhibited a characteristic central intensity node and substantial overlap with MQW gain medium. The specifie values of Q and T for these modes obtained from the calculations for the 26MQW nanowire structure,(Q=2, 485 and T=0. 240), were larger than those values for the 13MQW nanowire,(Q=656 10- and I=0.122), thus indicating that the threshold for lasing should be lower in the 26MQW structures. This conclusion was further analysed by calculating the light input-output curves(see I 10-2 Supplementary Information, Methods)as shown in Fig. 4c.These data show that Lth is about 14 times lower for the 26MQW versus 13MQW nanowire structures at a lasing wavelength of 450 nm. Finally, the experimentally determined lasing threshold for a series of 13MQW and 26MQW nanowire structures(Fig. 4d)shows 100 that Lth is 4-10 times lower for the 26 versus 13 MQW structures across the 400-500 nm wavelength region, in good agreement with our numerical calculations. Variation in Lth within the same MQW structure type reflects expected differences of individual nanowire cavities. The thresholds for the In gan MQW nanowire lasers are comparable to the planar InGan MQW lasers pumped at similar excitation energies", although they are higher than the best reported value for homogeneous GaN nanowires. We attribute this last difference primarily to smaller confinement factors in the present MQW structures but believe that future nanowire structure optimization can reduce threshold values(see the Supplementary Information). Significantly, we note that the 26MQW nanowire Pump power density (W cm-2) lasing threshold exhibits little wavelength dependence, in contrast to typical In Gan MQW planar structures, where lasing thresholds exhibit an exponential increase with emission wavelength. Thi has been attributed to the deterioration of material quality with increasing In composition, which we believe is alleviated in ▲△▲ our case as the MQW nanowire structures are dislocation-free ted controlled synthesis the first MQW core/shell nanowire heterostructures based on well-defined Ill-nitride materials. TEM studies have demonstrated that the triangular single-crystalline Gan nanowire cores nable epitaxial and dislocation-free growth of highly uniform (In GaN/GaN)n quantum wells with n 3, 13 and 26 and In gan well thicknesses of 1-3 nm. Optical studies of individual MQW nanowire structures further demonstrated lasing with Wavelength (nm) In gan quantum-well composition-dependent emission from 365 to 494 nm. Calculations exploring this new nanowire laser architecture have shown that laser threshold is reduced for Figure 4 MQw na ode along the nanowire axis. Cross-sectional top view of a Maw nanowire taken increasing n-value owing to higher Q and T, and this prediction was confirmed by experimental studies showing a 4-10-fold along the A-A position indicated in the end-on view at the right The MOW shell and threshold reduction in 26 versus 13 MQW structures. There are GaN core are indicated in yellow and grey, respectively. The red curve represents the also areas that could benefit from further effort, including coupled mode and the dashed-dotted line indicates the middle of the nanowire udinal cavity. b, FDTD simulation of the dominant laser mode in 13MOW and synthesis and structural analyses designed to push the limit of In 26MOw nanowire structures. white lines indicate the nanowire profile and the composition in the quantum wells and to define differences in core/shell interface. The scale bar is 100 nm. c, Calculated ight-in versus light-out growth for the unique nanowire versus planar geometry. Finally, we believe it will be interesting to implement an extra level of curve for 13 MoW (red) and 2bMaW ( ue) nanowire structures. d, Measured lasing complexity with p-type algaN/GaN shells, as this could enable the nanowires. Solid symbols correspond to experimental data and lines are guides for realization of free-standing nanowire injection nanolasers the eyes. NANOWIRE SYNTHESIS Q is the mqw nanowire structures were synthesized on an r-plane sapphire substrate de, vg is the in a metal-organic chemical vapour deposition reactor(Thomas Swan Scientific Equipment Ltd) using trimethylgallium(TMG), trimethylindium naturematerialsvol7iSepTembeR2008iwww.nature.com/naturematerials @2008 Macmillan Publishers Limited. All rights reserved.LETTERS 2 4 6 8 1,000 2 4 6 8 Threshold (kW cm–2) 400 420 440 460 480 Wavelength (nm) Gain medium A A‘ Output power (a.u.) Pump power density (kW cm–2) 1 10–1 10–2 10–4 10–3 0.01 0.1 1.0 10.0 100.0 100 1,000 10,000 a b c d Figure 4 MQW nanowire lasing threshold. a, Schematic diagram of an optical mode along the nanowire axis. Cross-sectional top view of a MQW nanowire taken along the A-A’ position indicated in the end-on view at the right. The MQW shell and GaN core are indicated in yellow and grey, respectively. The red curve represents the optical mode and the dashed–dotted line indicates the middle of the nanowire longitudinal cavity. b, FDTD simulation of the dominant laser mode in 13MQW and 26MQW nanowire structures. White lines indicate the nanowire profile and the core/shell interface. The scale bar is 100 nm. c, Calculated light-in versus light-out curve for 13MQW (red) and 26MQW (blue) nanowire structures. d, Measured lasing thresholds as a function of wavelength for 13MQW (red) and 26MQW (blue) nanowires. Solid symbols correspond to experimental data and lines are guides for the eyes. density, Γ1 is the optical confinement factor per well, Q is the quality factor, ω is the frequency of the resonant mode, vg is the group velocity and ηi is the absorption ratio in quantum wells. In this way, we have compared relative threshold ratios for 13 and 26MQW nanowire structures by determining the quality factor, Q, and confinement factor, Γ, using three-dimensional finite-difference time-domain (3D-FDTD) calculations because other parameters remain approximately constant. The modes with the highest Q (Fig. 4b) exhibited a characteristic central intensity node and substantial overlap with MQW gain medium. The specific values of Q and Γ for these modes obtained from the calculations for the 26MQW nanowire structure, (Q = 2,485 and Γ = 0.240), were larger than those values for the 13MQW nanowire, (Q = 656 and Γ = 0.122), thus indicating that the threshold for lasing should be lower in the 26MQW structures. This conclusion was further analysed by calculating the light input–output curves (see Supplementary Information, Methods) as shown in Fig. 4c. These data show that Lth is about 14 times lower for the 26MQW versus 13MQW nanowire structures at a lasing wavelength of 450 nm. Finally, the experimentally determined lasing threshold for a series of 13MQW and 26MQW nanowire structures (Fig. 4d) shows that Lth is 4–10 times lower for the 26 versus 13 MQW structures across the 400–500 nm wavelength region, in good agreement with our numerical calculations. Variation in Lth within the same MQW structure type reflects expected differences of individual nanowire cavities. The thresholds for the InGaN MQW nanowire lasers are comparable to the planar InGaN MQW lasers pumped at similar excitation energies29, although they are higher than the best reported value11 for homogeneous GaN nanowires. We attribute this last difference primarily to smaller confinement factors in the present MQW structures but believe that future nanowire structure optimization can reduce threshold values (see the Supplementary Information). Significantly, we note that the 26MQW nanowire lasing threshold exhibits little wavelength dependence, in contrast to typical InGaN MQW planar structures, where lasing thresholds exhibit an exponential increase with emission wavelength31. This has been attributed to the deterioration of material quality with increasing In composition31, which we believe is alleviated in our case as the MQW nanowire structures are dislocation-free single crystals. In summary, we have reported controlled synthesis of the first MQW core/shell nanowire heterostructures based on well-defined III-nitride materials. TEM studies have demonstrated that the triangular single-crystalline GaN nanowire cores enable epitaxial and dislocation-free growth of highly uniform (InGaN/GaN)n quantum wells with n = 3, 13 and 26 and InGaN well thicknesses of 1–3 nm. Optical studies of individual MQW nanowire structures further demonstrated lasing with InGaN quantum-well composition-dependent emission from 365 to 494 nm. Calculations exploring this new nanowire laser architecture have shown that laser threshold is reduced for increasing n-value owing to higher Q and Γ, and this prediction was confirmed by experimental studies showing a 4–10-fold threshold reduction in 26 versus 13 MQW structures. There are also areas that could benefit from further effort, including coupled synthesis and structural analyses designed to push the limit of In composition in the quantum wells and to define differences in growth for the unique nanowire versus planar geometry. Finally, we believe it will be interesting to implement an extra level of complexity with p-type AlGaN/GaN shells, as this could enable the realization of free-standing nanowire injection nanolasers. METHODS NANOWIRE SYNTHESIS MQW nanowire structures were synthesized on an r-plane sapphire substrate in a metal–organic chemical vapour deposition reactor (Thomas Swan Scientific Equipment Ltd) using trimethylgallium (TMG), trimethylindium nature materials VOL 7 SEPTEMBER 2008 www.nature.com/naturematerials 705 © 2008 Macmillan Publishers Limited. All rights reserved