A Postsynthesis Decomposition Strategy for Group III–Nitride Quantum Wires

Published: August 22, 2011

r 2011 American Chemical Society 4559 dx.doi.org/10.1021/cg200809k | Cryst. Growth Des. 2011, 11, 4559–4564

ARTICLE

pubs.acs.org/crystal

A Postsynthesis Decomposition Strategy for Group III�NitrideQuantum WiresLance Brockway,† Chandrashekhar Pendyala,‡ Jacek Jasinski,§ Mahendra K. Sunkara,‡ andSreeram Vaddiraju*,†

†Department of Chemical Engineering, Texas A&M University, 3122 TAMU, College Station, Texas 77843, United States‡106 Ernst Hall, Department of Chemical Engineering, and §Conn Center for Renewable Energy Research, University of Louisville,Louisville, Kentucky 40292, United States

bS Supporting Information

’ INTRODUCTION

Vapor-phase synthesis of nanowires was mainly accomplishedin the past using catalyst-mediated techniques,1�7 oxide-assistedmethods,8�11 or direct synthesis schemes.12�17 Typically, theseprocesses suffer from two distinct disadvantages. Either thediameters of the nanowires obtained were greater than20 nm,1�17 or the nanowires obtained were contaminated withthe catalyst metal.18,19 Typically, nanowire diameters need to beless than 20 nm to observe any quantum confinement effects inmany of the III�V semiconductors. For example, diameters lessthan 11 nm are necessary to observe quantum confinementeffects in GaN.20�22 Of the many reports in the literature onnanowire synthesis, only a few indicated quantum confinementeffects in nanowires. Reports of quantum confinement effects ingroup III�nitride nanowires do not currently exist. Hence, it isimportant to develop methods for obtaining nanowires withdiameters less than 20 nm in a controllable manner and to studythe effect of size on their electronic and optical properties.

As mentioned above, only a few reports in the literaturediscussed the syntheses of nanowires with diameters less than20 nm. A few examples include the synthesis of 7.8 nm diametersilicon and 5 nm diameter germanium nanowires by the laserablation method using gold and iron as the catalysts by Lieberand co-workers,23 the synthesis of 15 nm diameter siliconnanowires using TiSi2 followed by surface oxidation to reducethe diameter of the silicon core by Kamins and co-workers,24 andthe synthesis of InP rods with diameters as small as 3 nm indiameter by thermal decomposition of an InP-containing meta-lorganic precursor by Nozik and co-workers.25 Further, Korgeland co-workers synthesized silicon nanowires as small as 4 nm26

and germaniumnanowires as small as 7 nm27,28 by the supercritical

fluid�liquid�solid (SFLS) method using gold as the catalyst. A0.3 eV shift in the bandgap of germanium was observed at thesesmall diameters. Solution-phase synthesis of 7 nm diameter and9 nm diameter CdSe nanowires using gold as the catalyst was alsodemonstrated by Kuno and co-workers.29 In a series of papers,Buhro and co-workers have also reported the synthesis of InP,30

CdSe,31 ZnSe-ZnTe,32 and InAs33 nanowires using the solu-tion�liquid�solid (SLS) method. Recently, SnO2 nanorods withdiameters as small as 2 nm have also been synthesized.34 Ingeneral, it is observed that the vapor�liquid�solid (VLS)mechanism rarely yields wires with diameters less than 20 nm,while the SLS mechanism typically yields nanowires with dia-meters in the range of 4�10 nm.35 Unfortunately, the SLSmechanism has not resulted in high-quality nitride nanowires.35

Because of the inability of the current methods in synthesizingnanowires with sub-10 nm diameters (or quantum wires),especially in III�nitride systems, a postsynthesis approach forreducing the diameter of nanowires would be ideal for obtainingquantum-sized nanowires in a reliable manner, irrespective of thetechnique used for their synthesis.

Two postsynthesis strategies useful for reducing the diameterof nanowires are thermal decomposition and chemical etching.In the case of III�nitride nanowires, it is possible to devisestrategies based on thermal decomposition easily. For example, itis well-known that GaN decomposes above temperatures of850 �C. Consequently, it is important to know whether thedecomposition of presynthesized GaN nanowires leads to a

Received: June 27, 2011Revised: August 18, 2011

ABSTRACT:A simple postsynthesis strategy for obtaining sub-5 nm nanowires by uniformlyreducing the diameters of presynthesized nanowires is reported. Specific experimentsperformed using GaN as an example indicated that the thermal decomposition of 25 nmdiameter nanowires led to the formation of 5 nm diameter quantum wires, and this furtherresulted in a 0.5 eV bandgap blue-shift. The mechanism and kinetics of decomposition seemto be an exclusive nanoscale phenomenon, dependent on the initial size of nanowires.

4560 dx.doi.org/10.1021/cg200809k |Cryst. Growth Des. 2011, 11, 4559–4564

Crystal Growth & Design ARTICLE

reduction in their diameters. Second, it is crucial to know if thisdiameter reduction is uniform all along the lengths of thenanowires. Finally, it is imperative to understand the parametersaffecting the decomposition to obtain nanowires with diametersless than 10 nm in a uniform and controlled manner.

Therefore, the primary aim of this Article is to study whetherpostsynthesis decomposition could be employed as a strategy forreducing the diameter of nanowires. Specifically, the decomposi-tion of GaN nanowires is studied and discussed in detail in thisArticle. Our results indicated that the decomposition of GaNnanowires leads to a uniform reduction in their diameter andresults in the formation of quantum wires. It is also the aim of thisArticle to understand the mechanism behind the uniform andcontrolled decomposition of GaN nanowires, along with thefactors affecting the rates of decomposition of GaN. On the basisof the results, a layer-by-layer decomposition model is proposedto explain the uniform thinning of the GaN nanowire diametersall along their lengths. This layer-by-layer decomposition is adirect result of the ballistic diffusion of gallium adatoms, observedexclusively on nanoscale surfaces. Ballistic diffusion is intended tomean that the diffusion lengths are very large, on the order of tensof micrometers. A final aim of this Article is to understand thevariation in the optical bandgap of the nanowires with diameter.The bandgap of GaN nanowires was observed to increase withdecreasing diameter. A blue-shift in the bandgap of GaN, asmuchas 0.5 eV, was observed. To our knowledge, no such studiesindicating the use of postsynthesis decomposition as a strategyfor uniformly reducing the diameter of nanowires currently existin the literature.

’EXPERIMENTAL SECTION

The synthesis and postsynthesis decomposition of GaN nanowireswere performed in a custom-built chemical vapor deposition (CVD)chamber. This stainless steel vacuum chamber, capable of operating atpressures as low as 1 mTorr, is connected to mass flow controllers(1479A type mass flow controllers from MKS Instruments Inc.) for thecontrolled flow of both hydrogen and ammonia into the chamber and apressure transducer (KJLC 317 type from Kurt J. Lesker Co.) for themeasurement of pressure. The CVD chamber is also equipped with aceramic substrate heater that can be heated to temperatures as high as1400 �C. This setup was also described in detail previously.16 Addition-ally, a schematic representing the experimental setup employed for thesynthesis of GaN nanowires is shown in Figure S1 in the SupportingInformation.The synthesis of GaN nanowires was performed in a manner similar

to that described by Li et al.,14 using the reactive vapor transportapproach13,14,16,17 with GaN powder and NH3 as the sources for galliumand nitrogen, respectively. The GaN source was placed in a heatedcrucible, and the substrate was physically placed on top of this crucible.The distance between the source and crucible was adjusted to 2mm. Thetemperature of the substrate was measured using a backside thermo-couple placed on the side of the substrate not facing the GaN powdersource. All of the experiments were performed under an ammonia flowof 75 sccm, at a substrate temperature of 850 �C, and at a pressure of 100mTorr. The typical duration of the synthesis experiments was 20 min.Experiments for the decomposition of GaN nanowires were per-

formed by placing the presynthesized GaN nanowires on the heater andheating them to an elevated temperature of 900 �C. These experimentswere performed both in the presence of only ammonia and only hydrogenin the gas phase for varying durations ranging from 1 to 10 min. Theas-obtained nanowires and the quantum wires obtained by their decom-position were characterized using scanning electron microscopy (SEM),

X-ray diffraction (XRD), transmission electron microscopy (TEM), andultraviolet�visible absorption (UV�vis) spectroscopy. As-obtained GaNnanowires and quantum wires were used for the SEM and UV�visanalysis, without any sample preparation. Samples for TEM analysis wereprepared by simply scraping the nanowires onto holey carbon-coatedcopper TEM grids.

’RESULTS AND DISCUSSION

The reactive vapor transport of gallium in the presence ofdecomposed ammonia proceeded as expected and led to theformation of GaN nanowires.14,16 The as-obtained nanowires(Figure 1a) had an average diameter of 23 nm and are tens ofmicrometers long (additional images of the synthesized GaNnanowires are provided in Figure S2 of the Supporting Infor-mation). XRD analysis of the GaN nanowires sample (Figure 1b)indicated that the obtained nanowires have a wurtzite crystalstructure with lattice parameters of a = 3.19 Å and c = 5.17 Å.7

Decomposition of the as-obtained GaN nanowires was per-formed in the presence of ammonia in a sequential manner fordurations of 3, 6, and 10 min. After each decomposition experi-mental step, the nanowires were characterized using SEM andTEM to understand the effect of high temperature decomposi-tion on their morphology and size. Representative TEM micro-graphs of the 23 nm diameter as-synthesized nanowires, the6.8 nm diameter nanowires resulting from their decompositionfor 6min, and the 4.8 nmdiameter nanowires resulting from theirdecomposition for 10 min in the presence of NH3 are respec-tively shown in Figure 2a�c. From Figures 2a�c, it can also beobserved that the reduction in the diameters of the nanowires isuniform all along their lengths.

HR-TEM analysis of the nanowires indicated that the as-obtained nanowires are single-crystalline in nature (Figure 2d).The lattice spacing in the growth direction was observed to be2.76 Å and indicated that they preferentially grew in the [1010]growth direction (Figure 2d).14 HR-TEM analysis also indicatedthat the growth direction of the nanowires remained unalteredduring the decomposition process and remained [1010](Figure 2e). No change in the growth direction of the wires isindicative of the fact that the nanowires undergo a partialdecomposition, leading to a uniform reduction in the wirediameters. If this process occurred via complete decomposition,followed by regrowth of the nanowires, then there would be achange in the growth direction of the nanowires from [1010] to[0001].14

Figure 1. (a) Scanning electron micrograph of the as-synthesized GaNnanowires obtained by the reactive vapor transport approach. Theaverage diameter of the obtained nanowires is observed to be 23 nm.(b) XRD analysis of the obtained GaN nanowires showing their wurtzitecrystal structure. Some peaks from excess gallium are also observed inthe samples and are indicated by an “*”.



A plot showing the variation in the mean diameter of thenanowires with decomposition time is shown in Figure 3. It isessential to mention here that for each sample, the diameters of25�50 nanowires were taken as a sample data set and measuredfor the determination of their mean diameter. The error in themeasurement is no more than (0.5 nm. The diameter distribu-tions of the as-synthesized nanowires and nanowires decom-posed for 3, 6, and 10 min are shown in the inset to Figure 3. Themean diameter of the nanowires was found to decrease exponen-tially with decomposition time as shown in Figure 3 and eq 1.

D ¼ Doe�kt ð1Þ

From eq 1, the decomposition rate (nm/min) can be determinedto be

dDdt

¼ � kDoe�kt ¼ � kD ð2Þ

where D is the mean diameter of the nanowires (nm) at time t(min), Do is the initial diameter of the nanowires, and k is adecomposition rate constant. This exponential rate of decom-position is proportional to the circumference (πD) due tomonolayer step creation being the rate-determining step. Thedecomposition rate constant is dependent on the experimentalconditions employed: decomposition temperature, pressure, andgas-phase composition. From the plot in Figure 3, the value of k isdetermined to be 0.156 min�1 under our experimental condi-tions. This implies that a duration of 4.44 min is required forhalving the diameter of the nanowires. The rate of decompositiondepends on the initial diameter of the nanowire and is observedto have amaximumof 3.28 nm/min and an average of 1.66 nm/min.The maximum and average decomposition rates correspond

to 13.12 and 6.64 monolayers per minute, respectively, assuminga thickness of 0.25 nm for one monolayer of GaN.36 The rate ofdecomposition of GaN nanowires is observed to be higher thanthat of bulk GaN crystals. In a previous study, Grandjean et al.studied the decomposition of bulk GaN in the presence of anammonia flux.37 Using the correlation presented, the decom-position of bulk GaN under our experimental conditions (850 �Cand 100 mTorr) was determined to be 0.36 nm/min. Assumingagain a thickness of 0.25 nm per monolayer for GaN,36 the bulkdecomposition of GaN was estimated to be 1.44 monolayers perminute. From the above analysis, the decomposition rate of bulkGaN is estimated to be approximately one-tenth of the maximumdecomposition rate and one-fifth of the average decompositionrate of GaN nanowires observed in our experiments. It was alsoobserved that the diameter distribution of the wires was reducedas the decomposition proceeded (Figure 3), resulting in a muchmore narrow distribution in the decomposed wires than wasfound in the as-synthesized nanowires.

Experiments performed to understand the effect of gas-phasechemistry on the decomposition of GaN indicated that the rate ofdecomposition of GaN nanowires is higher in the presence ofhydrogen than in the presence of ammonia. These experimentswere performed at a temperature of 900 �C, a pressure of 100mTorr, and under a hydrogen flow rate of 100 sccm. In thepresence of hydrogen, the complete decomposition of GaNnanowires (23 nm in diameter) into gallium was observed aftera duration of only 6 min (see Figure S3 in the SupportingInformation). It is essential to recall here that in the presence ofammonia 4.8 nm thin quantum wires were observed after 10 minof decomposition.

To understand whether the uniform reduction in the dia-meters of the nanowires is exclusive to nanoscale, the decom-position of bulk compound semiconductor wafers was alsostudied. Because of the lack of single-crystal GaN wafers, theseexperiments were performed using GaAs wafers. As expected, thedecomposition of a GaAs wafer led to the formation of etch pitsalong its surface (see Figure S4 in the Supporting Information).Similar behavior was previously observed and reported in GaNcrystals by Pisch et al.38

Figure 2. Transmission electron micrographs of (a) as-obtained GaNnanowires, (b) nanowires observed after decomposition in NH3 for6 min, and (c) nanowires observed after decomposition in NH3 for10 min. A comparison of the images clearly shows that decompositionled to a uniform reduction in the diameters of the nanowires, without anyother changes to their morphologies. The fuzzy structure surroundingthe nanowires is from the amorphous carbon part of the holey carbon-coated copper TEM grids. (d) High magnification TEM micrograph ofan as-obtained GaN nanowire showing that the lattice spacing in thegrowth direction is 2.76 Å, and that the growth direction of thenanowires is [1010]. (e) TEM micrograph of a 3.2 nm thick quantumwire resulting from the decomposition of a 23 nm thick nanowire. Thelattice spacing indicated that the growth direction of the nanowire didnot change during the decomposition process and remained [1010].

Figure 3. A plot showing the variation in the mean diameter of the GaNnanowires with decomposition time in the presence of ammonia at atemperature of 900 �C. The mean nanowire diameter is observed todecrease exponentially with time. The error in the measurement isestimated to be (0.5 nm. The histogram inset shows the diameterdistribution of the GaN nanowires at various times.



In principle, the partial decomposition of GaN nanowires mayfollow three routes: (a) layer-by-layer decomposition if thediffusion of both decomposed Ga and N species can be fast overthe length scales of nanowires, (b) creation of multiple etch stepsand coalescence to remove a monolayer eventually leading torough surfaces of the decomposed nanowires, and (c) creation ofetch pits in nanowires ultimately leading to the breakdown ofnanowires into individual crystals. Themode of decomposition isdependent on the gallium adatom diffusion lengths on thesurfaces of GaN. If the diffusion lengths of gallium adatoms areballistic, layer-by-layer decomposition on the surfaces of thenanowires occurs and leads to a uniform reduction in theirdiameters without changing their growth direction (Figure 4aand b). If the diffusion lengths of adatoms are smaller, thengallium droplets are formed all along the lengths of the nano-wires. The accelerated etching of GaN in the vicinity of thedroplet breaks the nanowires into multiple crystals (Figure 4c).This leads to the formation of polycrystalline GaN.

The diffusion lengths of gallium adatoms on the surfaces ofbulk GaN crystals are well-known to be in the range of10�80 nm.39 Similarly, the diffusion lengths of gallium on GaAscrystal surfaces are also in the 10�100 nm range.40 It is alsodocumented that gallium droplets aid in accelerating the decom-position in their vicinity.38 Hence, in the case of bulk crystals, theprimary result of decomposition is the formation of etch pits all

along the surfaces.These etch pits eventually break bulk crystalsinto a polycrystalline crust. Similar decomposition phenomenaoccur in thick nanowires (Figure 4c). In the case of thinner (sub-25 nm diameter) nanowires, ballistic diffusion of metal adatomson the surfaces can occur with diffusion lengths on the order oftens ofmicrometers.14,41 Therefore, the formation of galliumdropletson the surfaces of nanowires does not occur unless the desorption ofgallium adatoms proceeds slowly to allow its accumulation on thenanowire surface. Thus, layer-by-layer decomposition leads tothe uniform shrinking of the nanowires diameters.

In the case of layer-by-layer decomposition of nanowires, themonolayer etch rate can depend upon either the surface area ofthe nanowire or the circumference of the nanowire. In the case ofslower kinetics, in the presence of ammonia, the rate of mono-layer etching is limited by etch step creation, which is propor-tional to the number of edge sites (or the circumference � πD).In this case, the decomposition is initiated at the edge of thenanowire, and the step propagation occurs much faster than thenext etch step creation as seen in Figure 4a. For this mechanismto be operative, the gallium adatoms, resulting from the decom-position of GaN, diffuse away from the step edge and desorbquickly from the surface of the nanowire. Under these conditions,one can model the rate of monolayer etching as that proportionalto the monolayer etch step creation, the rate-determining step. Inother words, the rate of monolayer etching is proportional to thediameter (or the circumference) of the nanowires and a rateconstant as shown in eq 2. Because the rate of decomposition isdirectly proportional to the diameter of the wire, larger nanowires(with larger circumferences) decompose faster than smallernanowires. Consequently, the diameter distribution is expectedto be narrowed as the decomposition proceeds. The above modelis consistentwith the experimental observations shown in Figure 3.

In the presence of hydrogen, the activation energy of nitrogendesorption can be reduced leading to multiple etch step creationall along the lengths of nanowires instead of primarily at the edges asillustrated in Figure 4b. In this case, the time required for thecomplete decomposition of a GaN surface layer depends primarilyon the etch step propagation and coalescence time and is conse-quently not limited by the etch step creation time. Therefore, therate of decomposition of GaN nanowires is higher in the presenceof hydrogen than in the presence of ammonia. In sharp contrast, thediffusion lengths of adatoms on bulk GaN crystal/thick GaNnanowires surfaces are smaller (10�80 nm).39 These lower diffu-sion lengths of gallium adatoms lead to their accumulation and theformation of droplets all along the surface of the nanowire/crystal.The accelerated etching of GaN38 in the vicinity of the dropletessentially breaks a nanowire into multiple crystals (Figure 4c).

Decomposition of GaN nanowires was also performed inside aTEM (FEI Technai F20) equipped with a heating stage. Theseexperiments were preformed in a vacuum at 10�7 Torr, withoutthe supply of a nitrogen-containing gas, such as ammonia. Thedecomposition of GaN nanowires in the absence of ammoniawas observed to be similar to that expected in bulk GaN. Duringthis process, voids begin to form in nanowires as the resultof nitrogen evaporation and nitrogen vacancies accumulation(as indicated by the white arrows in Figure S5a). Simultaneously, thegallium resulting from the decomposition of GaN accumulates intodroplets. The preferential decomposition of GaN in the vicinity ofthe gallium droplets leads to the formation of holes (Figure S5b). Itis documented that in the presence of a nitrogen-containing gas,the activation energy of nitrogen desorption is increased;37

consequently, the amount of free gallium on the surface of the

Figure 4. A schematic illustrating the proposed mechanism of decom-position in thin GaN nanowires (diameters less than 25 nm) in thepresence of (a) NH3 and (b) H2. (c) Mechanism of decompositionexpected in thick GaN nanowires/bulk GaN crystals. (a) In thin GaNnanowires, the decomposition starts at the edge of the nanowires in thepresence of ammonia. This leads to the creation of a step. Aided by theballistic diffusion of gallium adatoms, the decomposition propagatesuntil this layer is completely decomposed. (b) The initiation of thedecomposition at multiple locations all along the surface is expected toenhance the decomposition rate in the presence of hydrogen relative tothat of ammonia due to an absence of a nitrogen overpressure. (c) Inbulk GaN crystals/thick GaN nanowires, the lower diffusion length ofgallium adatoms leads to the formation of gallium droplets on thesurfaces. Accelerated etching in the vicinity of these droplets ultimatelyleads to the formation of polycrystalline GaN (white circles indicategallium adatoms, while gray circles indicate gallium droplets).



GaN nanowires is insufficient to accumulate into droplets.Therefore, controlled decomposition in the presence of ammo-nia is essential for uniformly thinning GaN nanowires. Incontrast, the lower nitrogen desorption activation energy inthe absence of ammonia leads to the accumulation of gallium in theform of droplets, and hence the formation of etch pits in GaNnanowires. Additional experimentation on the in situ decompositionof GaN nanowires inside a TEM is currently being performedand will be reported later.

To estimate the effect of diameter on the bandgap of GaN,UV�vis absorption spectroscopic analysis of the as-synthesizedGaN nanowires and the quantum wires resulting from theirdecomposition was performed. As-obtained nanowires andquantum wires on quartz substrates were directly employed forthis purpose, without any additional sample preparation. A Taucplot was also generated from the transmittance (T) versus energy(hν) data obtained using UV�vis spectroscopy (Figure 5).42,43

The Tauc relation is given by αhν = B(hν� Eg)X, where α is the

absorption coefficient, h is Planck’s constant, ν is the frequency ofthe incident wave, andB is a constant. The factor x is 0.5 for directbandgap semiconductors (such as GaN) and 2 for indirectbandgap semiconductors. The absorption coefficient (α) wascalculated from the transmittance (T) versus energy (hν) datausing the procedure described previously.42,44,45

The Tauc plot of GaN nanowires and quantumwires, showingthe variation of (αhν)2 with energy (hν), is shown in Figure 5.The optical bandgap can be determined by the x-intercept of atangent drawn through the linear portion of the plot. Thistechnique confirmed that the bandgap of the as-synthesized23 nm GaN wires is 3.4 eV.20�22 The Tauc plot also indicatedthat the bandgaps of GaN nanowires with average diameters of13.5, 6.8, and 4.8 nm are, respectively, 3.4, 3.7, and 3.9 eV. A plotshowing the variation of the bandgap of GaN with nanowirediameter is shown in the inset to Figure 5. As expected, thereduction in the diameter of GaN nanowires led to an increase in

the bandgap.20 A maximum blue-shift of 0.5 eV in the bandgap ofGaN was observed in nanowires with an average diameter of4.8 nm. The bandgaps estimated from the Tauc plots wereobserved to be close to those predicted by theoreticalestimates.22 For example, theoretical estimates predict a blue-shift of 0.5 eV is expected in GaN crystals with sizes of 4 nm.20,22

The average diameter of our nanowires observed from SEM andTEM analysis exhibiting a bandgap of 3.9 eV was observed to be4.8 nm, close to that predicted by theory.

’CONCLUSION

A simple postsynthesis strategy for uniformly reducing thediameters of presynthesized nanowires and obtaining quantumwires is presented. The decomposition occurred in a layer-by-layer fashion in nanowires with sizes less than 25 nm and led tothe formation of quantumwires with diameters as small as 4.8 nmon average. This layer-by-layer decomposition is a result of theballistic diffusion of gallium adatoms (resulting from the decom-position of GaN) on GaN nanowire surfaces. Layer-by-layerdecomposition is exclusive to GaN nanowires with diameters lessthan 50 nm and unlike that of bulk GaN crystals where decom-position leads to the formation of etch pits all along the surfaces.The rate of decomposition was also observed to be dependent onthe gas-phase chemistry, lower in the presence of ammonia andhigher in the presence of hydrogen. A reduction in the diameter ofthe GaN nanowires resulted in an increase in their opticalbandgap. A blue-shift in the bandgap of 0.5 eV is observed inthe obtainedGaNquantumwires with amean diameter of 4.8 nm.This shift is attributed to quantum confinement resulting from thediameter reduction of the nanowires. This postsynthesis decom-position strategy for reducing the diameters of nanowires isexpected to be applicable to other III�V compound semicon-ductor systems.

’ASSOCIATED CONTENT

bS Supporting Information. A schematic of the experimen-tal setup employed for the synthesis of GaN nanowires, addi-tional micrographs of GaN nanowires, micrographs of GaNnanowires decomposed in the presence of hydrogen, and micro-graphs of nanowires at various stages of in situ decompositionobserved inside a TEM. This material is available free of chargevia the Internet at http://pubs.acs.org.

’AUTHOR INFORMATION

Corresponding Author*Phone: (979) 862-1615. Fax: (979) 845-6446. E-mail: [email protected].

’ACKNOWLEDGMENT

We thank ConnCenter for Renewable Energy at the Universityof Louisville for help in characterizing the GaN nanowires andquantumwires. L.B. and S.V. thank the NSF/DOE thermoelectricpartnership program for financial support (CBET 1048702).

’REFERENCES

(1) Agarwal, R.; Lieber, C. M. Semiconductor nanowires: optics andoptoelectronics. Appl. Phys. A: Mater. Sci. Process. 2006, 85, 209–215.

(2) Chen, C. C.; Yeh, C. C. Large-scale catalytic synthesis ofcrystalline gallium nitride nanowires. Adv. Mater. 2000, 12, 738.

Figure 5. Tauc plots of (A) as-obtained GaN nanowires, and nanowiresobserved after decomposition in NH3 for (B) 3 min, (C) 6 min, and (D)10 min showing a 0.5 eV blue-shift in the bandgap as the diameter wasdecreased from 23 to 4.8 nm. The inset shows the variation in thebandgap of GaN nanowires as a function of their average diameters.



(3) Duan, X. F.; Lieber, C. M. Laser-assisted catalytic growth ofsingle crystal GaN nanowires. J. Am. Chem. Soc. 2000, 122, 188–189.(4) Duan, X. F.; Lieber, C. M. General synthesis of compound

semiconductor nanowires. Adv. Mater. 2000, 12, 298–302.(5) Duan, X. F.; Wang, J. F.; Lieber, C. M. Synthesis and optical

properties of gallium arsenide nanowires. Appl. Phys. Lett. 2000, 76,1116–1118.(6) Wang, D.; Qian, F.; Yang, C.; Zhong, Z. H.; Lieber, C. M.

Rational growth of branched and hyperbranched nanowire structures.Nano Lett. 2004, 4, 871–874.(7) Zhong, Z. H.; Qian, F.; Wang, D. L.; Lieber, C. M. Synthesis of

p-type gallium nitride nanowires for electronic and photonic nanode-vices. Nano Lett. 2003, 3, 343–346.(8) Goldberger, J.; He, R. R.; Zhang, Y. F.; Lee, S. W.; Yan, H. Q.;

Choi, H. J.; Yang, P. D. Single-crystal gallium nitride nanotubes. Nature2003, 422, 599–602.(9) Shi, W. S.; Zheng, Y. F.; Wang, N.; Lee, C. S.; Lee, S. T. A general

synthetic route to III�V compound semiconductor nanowires. Adv.Mater. 2001, 13, 591.(10) Shi, W. S.; Zheng, Y. F.; Wang, N.; Lee, C. S.; Lee, S. T.

Microstructures of gallium nitride nanowires synthesized by oxide-assisted method. Chem. Phys. Lett. 2001, 345, 377–380.(11) Shi, W. S.; Zheng, Y. F.; Wang, N.; Lee, C. S.; Lee, S. T. Oxide-

assisted growth and optical characterization of gallium-arsenide nano-wires. Appl. Phys. Lett. 2001, 78, 3304–3306.(12) Chandrasekaran, H.; Sunkara, M. K. Growth of gallium nitride

textured films and nanowires on polycrystalline substrates at sub-atmo-spheric pressures. In Gan and Related Alloys-2001; Northrup, J. E.,Neugebauer, J., Look, D. C., Chichibu, S. F., Riechert, H., Eds.; MaterialsResearch Society: Warrendale, 2002; Vol. 693, pp 159�164.(13) Chin, A. H.; Vaddiraju, S.; Maslov, A. V.; Ning, C. Z.; Sunkara,

M. K.; Meyyappan, M. Near-infrared semiconductor subwavelength-wire lasers. Appl. Phys. Lett. 2006, 88.(14) Li, H. W.; Chin, A. H.; Sunkara, M. K. Direction-dependent

homoepitaxial growth of GaN nanowires. Adv. Mater. 2006, 18, 216.(15) Stach, E. A.; Pauzauskie, P. J.; Kuykendall, T.; Goldberger, J.;

He, R. R.; Yang, P. D. Watching GaN nanowires grow. Nano Lett. 2003,3, 867–869.(16) Vaddiraju, S.;Mohite, A.; Chin, A.;Meyyappan,M.; Sumanasekera,

G.; Alphenaar, B. W.; Sunkara, M. K. Mechanisms of 1D crystal growth inreactive vapor transport: Indium nitride nanowires. Nano Lett. 2005, 5,1625–1631.(17) Vaddiraju, S.; Sunkara, M. K.; Chin, A. H.; Ning, C. Z.;

Dholakia, G. R. Synthesis of group III antimonide nanowires. J. Phys.Chem. C 2007, 111, 7339–7347.(18) Allen, J. E.; Hemesath, E. R.; Perea, D. E.; Lensch-Falk, J. L.; Li,

Z. Y.; Yin, F.; Gass, M. H.; Wang, P.; Bleloch, A. L.; Palmer, R. E.;Lauhon, L. J. High-resolution detection of Au catalyst atoms in Sinanowires. Nat. Nanotechnol. 2008, 3, 168–173.(19) Hannon, J. B.; Kodambaka, S.; Ross, F. M.; Tromp, R. M. The

influence of the surface migration of gold on the growth of siliconnanowires. Nature 2006, 440, 69–71.(20) Brus, L. Electronic wave-functions in semiconductor clusters -

experiment and theory. J. Phys. Chem. 1986, 90, 2555–2560.(21) Preschilla, N.;Major, S.; Kumar, N.; Samajdar, I.; Srinivasa, R. S.

Nanocrystalline gallium nitride thin films. Appl. Phys. Lett. 2000, 77,1861–1863.(22) Yoon, J. W.; Sasaki, T.; Roh, C. H.; Shim, S. H.; Shim, K. B.;

Koshizaki, N. Quantum confinement effect of nanocrystalline GaN filmsprepared by pulsed-laser ablation under various Ar pressures. Thin SolidFilms 2005, 471, 273–276.(23) Morales, A. M.; Lieber, C. M. A laser ablation method for the

synthesis of crystalline semiconductor nanowires. Science 1998, 279,208.(24) Guichard, A. R.; Barsic, D. N.; Sharma, S.; Kamins, T. I.;

Brongersma, M. L. Tunable light emission from quantum-confinedexcitons in TiSi2-catalyzed silicon nanowires. Nano Lett. 2006, 6,2140–2144.

(25) Ahrenkiel, S. P.; Micic, O. I.; Miedaner, A.; Curtis, C. J.;Nedeljkovic, J. M.; Nozik, A. J. Synthesis and characterization ofcolloidal InP quantum rods. Nano Lett. 2003, 3, 833–837.

(26) Holmes, J. D.; Johnston, K. P.; Doty, R. C.; Korgel, B. A.Control of thickness and orientation of solution-grown silicon nano-wires. Science 2000, 287, 1471.

(27) Hanrath, T.; Korgel, B. A. Quantum confinement effects ingermanium nanowires studied by electron energy loss spectroscopy(EELS). InPhysical Chemistry of Interfaces andNanomaterials Iii; Hartland,G. V., Zhu, X. Y., Eds.; SPIE-The International Society for OpticalEngineering: Bellingham, Washington USA, 2004; Vol. 5513, pp 40�47.

(28) Hanrath, T.; Korgel, B. A. A comprehensive study of electronenergy losses in Ge nanowires. Nano Lett. 2004, 4, 1455–1461.

(29) Grebinski, J. W.; Hull, K. L.; Zhang, J.; Kosel, T. H.; Kuno, M.Solution-based straight and branched CdSe nanowires. Chem. Mater.2004, 16, 5260–5272.

(30) Yu, H.; Li, J.; Loomis, R. A.; Wang, L.-W.; Buhro, W. E. Two-versus three-dimensional quantum confinement in indium phosphidewires and dots. Nat. Mater. 2003, 2, 517–520.

(31) Yu, H.; Li, J.; Loomis, R. A.; Gibbons, P. C.;Wang; Buhro,W. E.Cadmium selenide quantum wires and the transition from 3D to 2Dconfinement. J. Am. Chem. Soc. 2003, 125, 16168–16169.

(32) Dong, A. G.; Wang, F. D.; Daulton, T. L.; Buhro, W. E.Solution-liquid-solid (SLS) growth of ZnSe�ZnTe quantum wireshaving axial heterojunctions. Nano Lett. 2007, 7, 1308–1313.

(33) Wang, F. D.; Yu, H.; Jeong, S. H.; Pietryga, J. M.; Hollingsworth,J. A.; Gibbons, P. C.; Buhro, W. E. The scaling of the effective band gaps inindium�arsenide quantumdots andwires.ACSNano2008, 2, 1903–1913.

(34) Xi, G.; Ye, J. Ultrathin SnO2nanorods: Template- and surfactant-free solution phase synthesis, growth mechanism, optical, gas-sensing, andsurface adsorption properties. Inorg. Chem. 2010, 49, 2302–2309.

(35) Wang, F.; Dong, A.; Sun, J.; Tang, R.; Yu, H.; Buhro, W. E.Solution�liquid�solid growth of semiconductor nanowires. Inorg.Chem. 2006, 45, 7511–7521.

(36) Erenburg, S. B.; et al. Quantum dots microstructure by XAFSspectroscopy: GaN/AlN system depending on preparation conditions.J. Phys.: Conf. Ser. 2006, 41, 261.

(37) Grandjean, N.; Massies, J.; Semond, F.; Karpov, S. Y.; Talalaev,R. A. GaN evaporation in molecular beam epitaxy environment (vol 74,pg 1854, 1999). Appl. Phys. Lett. 1999, 75, 3035–3035.

(38) Pisch, A.; Schmid-Fetzer, R. In situ decomposition study ofGaN thin films. J. Cryst. Growth 1998, 187, 329–332.

(39) Kukushkin, S. A.; Bessolov, V. N.; Osipov, A. V.; Luk’yanov,A. V. Mechanism and kinetics of early growth stages of a GaN film. Phys.Solid State 2002, 44, 1399–1405.

(40) Shitara, T.; Nishinaga, T. Surface-diffusion length of galliumduring MBE growth on the various misoriented GaAs(001) substrates.Jpn. J. Appl. Phys., Part 1 1989, 28, 1212–1216.

(41) Jensen, L. E.; Bjork, M. T.; Jeppesen, S.; Persson, A. I.; Ohlsson,B. J.; Samuelson, L. Role of surface diffusion in chemical beam epitaxy ofInAs nanowires. Nano Lett. 2004, 4, 1961–1964.

(42) Chakrapani, V.; Thangala, J.; Sunkara, M. K. WO3 and W2Nnanowire arrays for photoelectrochemical hydrogen production. Int.J. Hydrogen Energy 2009, 34, 9050–9059.

(43) Chen, Z. B.; Jaramillo, T. F.; Deutsch, T. G.; Kleiman-Shwarsctein,A.; Forman, A. J.; Gaillard, N.; Garland, R.; Takanabe, K.; Heske, C.;Sunkara, M.; McFarland, E. W.; Domen, K.; Miller, E. L.; Turner, J. A.;Dinh, H. N. Accelerating materials development for photoelectrochemicalhydrogen production: Standards for methods, definitions, and reportingprotocols. J. Mater. Res. 2010, 25, 3–16.

(44) Hiramatsu, H.; Seo, W. S.; Koumoto, K. Electrical and opticalproperties of radio-frequency-sputtered thin films of (ZnO)(5)In2O3.Chem. Mater. 1998, 10, 3033–3039.

(45) Ong, C. H.; Gong, H. Effects of aluminum on the properties ofp-type Cu-Al-O transparent oxide semiconductor prepared by reactiveco-sputtering. Thin Solid Films 2003, 445, 299–303.

A Postsynthesis Decomposition Strategy for Group III–Nitride Quantum Wires

Documents

Transcript of A Postsynthesis Decomposition Strategy for Group III–Nitride Quantum Wires