Growth Condition-Oriented Defect Engineering for Changes in Au–Zno Contact Behavior from Schottky to Ohmic and Vice Versa

nanomaterials

Communication Growth Condition-Oriented Defect Engineering for Changes in Au–ZnO Contact Behavior from Schottky to Ohmic and Vice Versa

Abu ul Hassan Sarwar Rana and Hyun-Seok Kim *

Division of Electronics and Electrical Engineering, Dongguk University-Seoul, Seoul 04620, Korea; [email protected] * Correspondence: [email protected]; Tel.: +82-2-2260-3996; Fax: +82-2-2277-8735

Received: 14 November 2018; Accepted: 26 November 2018; Published: 27 November 2018

Abstract: ZnO has the built-in characteristics of both ionic and covalent compound semiconductors, which makes the metal–ZnO carrier transport mechanism quite intricate. The growth mechanism- centric change in ZnO defect density and carrier concentration also makes the contact formation and behavior unpredictable. This study investigates the uncertainty in Au–ZnO contact behavior for application-oriented research and the development on ZnO nanostructures. Herein, we explain the phenomenon for how Au–ZnO contact could be rectifying or non-rectifying. Growth method- dependent defect engineering was exploited to explain the change in Schottky barrier heights at the Au–ZnO interface, and the change in device characteristics from Schottky to Ohmic and vice versa. The ZnO nanorods were fabricated via aqueous chemical growth (ACG) and microwave-assisted growth (MAG) methods. For further investigations, one ACG sample was doped with Ga, and another was subjected to oxygen plasma treatment (OPT). The ACG and Ga-doped ACG samples showed a quasi-Ohmic and Ohmic behavior, respectively, because of a high surface and subsurface level donor defect-centric Schottky barrier pinning at the Au–ZnO interface. However, the ACG-OPT and MAG samples showed a more pronounced Schottky contact because of the presence of low defect-centric carrier concentration via MAG, and the removal of the surface accumulation layer via the OPT process.

Keywords: ZnO; metal-semiconductor contact; crystal defects; nanorod; microwave; oxygen plasma treatment

1. Introduction ZnO has been emerged as an enticing semiconductor material in the research and development on wide bandgap optoelectronic and nanoelectronic applications [1–4]. At ambient conditions, it has a high exciton binding energy of 60 meV and a direct bandgap of 3.37 eV, which make it an alluring candidate for the semiconductor industry. Its worth is further highlighted because of the formation of different polymorphic shapes, including ZnO nanorods (ZNRs), nanoﬂowers, nanotubes, nanoparticles, nanotetrapods, nanostars, and many more, by judiciously controlling the growth conditions [5–8]. Depending upon the formation of multiple nanostructured shapes, it has potential applications in the realm of nanoelectronics, optoelectronics, nanoscale lasers, toxic gas sensors, and ultraviolet photodetectors [9–15]. The metal contact formation of any semiconductor is one of the most important and inevitable processes for device fabrication. For instance, Ohmic contacts to ZnO are important for various electronic and optoelectronic applications, such as solar cells, ﬁeld effect transistors, and light-emitting diodes. Similarly, Schottky contacts to ZnO are important for applications in the realm of gas and

Nanomaterials 2018, 8, 980; doi:10.3390/nano8120980 www.mdpi.com/journal/nanomaterials Nanomaterials 2018, 8, 980 2 of 11 chemical sensors and photodetectors, where the Schottky barrier at the interface plays a significant role in determining the device efficiency [16,17]. As per the Schottky–Mott model, metals with high work functions are prime candidates for making Schottky contacts to semiconductors [18]. However, for ZnO, it has already been argued that the metal should have low oxygen affinity, to prevent oxide layer formation at the metal–ZnO interface, which leaves us with the choice of Pt, Pd, and Au being inert to oxidation [19]. The theoretical Au–ZnO contact should be a Schottky contact according to the classical Schottky model: ϕSB= ϕAu−χZnO (1) where ϕSB is the Schottky barrier height, ϕAu (5.1 eV) is the Au work function, and χZnO (4.2 eV) is the ZnO electron affinity. The theoretical ϕSB calculated via Equation (1) is 0.9 eV. However, Au–ZnO contacts show disparities to the results calculated via the classical Schottky model, because the experimental calculations not only show different Schottky barrier heights, but also an Ohmic contact to ZnO because of its intricate growth chemistry and defect formation. Herein, the practical reasons to the above mentioned Au–ZnO contact unpredictability are addressed via experimentation and reasoning. Au is a mandatory contact to ZnO, because of its non-oxidizing nature. Au is an inert metal, where some of the Ohmic metals, such as Ag, form an oxide layer on the metal–semiconductor surface. Hence, Au is preferred upon Ag for contact formation. However, Au makes a Schottky contact, and some devices use Ohmic contacts for their operation. This study propounds how Au could be used as both Schottky and Ohmic contacts for ZnO device characteristics, and why the Au–ZnO contact behavior is unpredictable at times. The prevailing Au–ZnO contact disparities were explained under the ambit of complex ZnO growth chemistry, crystal quality, and defect engineering. We found that the unpredictability in Au–ZnO contact formation lies in the growth method-dependent defect engineering, which is still a matter of debate in research and development on ZnO. Four different ZNR samples, grown with aqueous chemical growth (ACG), Ga-doped ACG, oxygen plasma-treated (OPT) ACG, and microwave-assisted growth (MAG), were considered. The defect engineering was monitored by studying the ZNR crystalline, structural, and optical characteristics. Also, the defect engineering-centric Schottky barrier heights and Au–ZnO contact behaviors were predicted by studying the current–voltage (I-V) and Hall-effect measurement results. It was found that, depending upon the defect engineering-centric change in majority carrier concentration, the ACG and Ga-doped ACG samples showed a quasi-Ohmic and complete Ohmic contact to ZnO, respectively, while the OPT-ACG and MAG samples showed Schottky contacts to ZnO. We believe that the propounded results would be very beneficial for the application-oriented research and development on ZnO.

2. Materials and Methods Flexible polyethylene terephthalate (PET) was used as a substrate for ZNR growth and device testing. The substrates were cleaned with acetone, isopropyl alcohol, and deionized (DI) water, and then dried with nitrogen gas. The device patterns were defined by spin coating positive photoresist (AZ 5214 E) and photolithography. The bottom electrode was deposited by evaporating 150 nm Au thin film via e-beam metal evaporation and lift-off processes. The seed layer-assisted ZNRs were directly grown onto the bottom electrode. ZNRs were fabricated via the most facile and widely accepted ACG and MAG methods, where the details of the methods are already explained in our previous reports [20,21]. A thin film seed layer deposition is necessary for vertical ZNR growth [22]. The homogeneous seed solution was made by mixing 22 mM zinc acetate dihydrate (Zn(CH3COO)2·2H2O) (MW 219.51 g/mol) in n-propanol (C3H8O) (MW 60.10 g/mol). The seed solution was spin coated twice on the PET substrate surface, which were then annealed at 70 ◦C for 2 min for the first coating and at 100 ◦C for 2 hr for the second coating. The growth solution was made by stirring 50 mM zinc nitrate hexahydrate (Zn(NO3)2·6H2O) (MW 297.48 g/mol) and 50 mM methenamine (C6H12N4) (MW 140.186 g/mol) in DI water. After one hour of continuous stirring, the Au-coated seeded PET substrates were immersed in the solution autoclaves, which were Nanomaterials 2018, 8, 980 3 of 11

Nanomaterials 2018, 8, x FOR PEER REVIEW 3 of 11 then placed on a hotplate at 75 ◦C for 12 hr for ACG growth, and in a domestic 850 watt (2.7 GHz) microwaveoven for the oven MAG for process. the MAG One process. of the ACG One samples of the ACG was samplesdoped with was 5% doped Ga content. with 5% The Ga details content. of Thethe detailsdoping ofprocess the doping are explained process arein our explained previous in reports our previous [23,24]. reports In short, [23 gallium,24]. In short,nitrate galliumhydrate nitrate(Ga(N hydrateO3)3·xH2 (Ga(NOO) (MW 3255.7)3·xH42 O)anhydrous (MW 255.74 basis) anhydrous was mixed basis) with was the mixed 50 mM with zinc the nitride 50 mM hexahydrate zinc nitride hexahydrateand methenamine and methenamine growth solution growth before solution putting before the solution putting autoclave the solution onto autoclave the hot plate. onto To the probe hot plate.the effects To probe of surface the effects and subsurface of surface cleaning, and subsurface another cleaning, ACG sample another was ACGsubjected sample to OPT was with subjected a flow torate OPT of 100 with sccm a ﬂow at 100 rate W. of Afterward, 100 sccm at the 100 150 W. nm Afterward, top Au contact the 150 was nm dire topctly Au evaporated contact was via directly e-beam evaporatedand lift-off viaprocesses e-beam on and the lift-off ACG, processes Ga-doped on ACG, the ACG, OPT Ga-doped-ACG, and ACG, MAG OPT-ACG, ZNRs without and MAG any further ZNRs withoutsurface any processing. further surface The final processing. device structures The ﬁnal device for all structures the aforementioned for all the aforementioned devices were the devices same, werewhich the issame, presented which in isSchematic presented Figure in Schematic 1. Figure1.

FigureFigure 1. 1.Step-by-step Step-by-step schematic schematic representation representation of of the the device device fabrication. fabrication.

ToTo look look at at the the detailed detailed morphology, morphology, the the ZNRs ZNRs were were characterized characterized with with scanning scanning electron electron microscopymicroscopy (SEM:(SEM:Hitachi HitachiS-4800, S-4800,Hitachi, Hitachi,Tokyo, Tokyo, Japan). Japan).The The structural structural and and crystalline crystalline properties properties werewere probed probed withwith X-rayX-ray diffraction ( (XRD:XRD: Rigaku Rigaku Ultima Ultima IV IV,, Rigaku Rigaku,, Tokyo Tokyo,, Japan Japan)) operating operating at ata Cu a CuKα Kradiationα radiation of λ of =λ 0.15418= 0.15418 nm. nm. The Thedefect defect centers centers and andoptical optical characteristics characteristics were were analyzed analyzed with withphotoluminescence photoluminescence (PL: (PL: Accent Accent RPM RPM 2000 2000,, Accent Accent Optical Optical Technologies Technologies Inc. Inc.,, Bend Bend,, OR,OR, USA)USA) spectroscopyspectroscopy withwith aa PL PL range range of of 300–700 300–700 nm nm at at room room temperature. temperature. The The ZNR ZNR carrier carrier concentrations concentrations werewere found found with with a Hall-effect a Hall-effect measurement measurement system system (ECOPIA: (ECOPIA: AHT55T5, AHT55T5 Ecopia,Corporation, Ecopia Corporation Anyang,, Gyeonggi-do,Anyang, Gyeonggi South- Korea).do, South The KoreaI-V characteristics). The I-V characteristics were measured were with measured Keithley with2410 measurement Keithley 2410 setupmeasurement (Keithley Instruments, setup (Keithley Solon, Instruments OH, USA) at, ambient Solon, OH, conditions. USA) Theat photolithographic ambient conditions. patterns The werephotolithographic deﬁned with Karlpatterns Suss were MA defined 6 mask with aligner Karl (SUSS Suss MA MicroTec, 6 mask Munich,aligner (SUSS Germany), MicroTec, and Munich, the Au contactsGermany) were, and deposited the Au contacts via e-beam were metal deposited evaporator via e-beam (Korea metal Vacuum: evaporator KVE-T5560, (Korea Korea Vacuum: Vacuum KVE- Tech.,T5560 Gimpo,, Korea Gyeonggi-do,Vacuum Tech. South, Gimpo Korea), Gyeonggi at high-do vacuum, South conditions.Korea) at high vacuum conditions.

3.3. ResultsResults andand DiscussionDiscussion

3.1.3.1. ZnOZnO SurfaceSurface MorphologyMorphology FigureFigure2 2shows shows that that the the ACG, ACG, Ga-doped Ga-doped ACG, ACG, OPT-ACG, OPT-ACG, and and MAG MAG ZNRs ZNRs have have almost almost the the same same sizessizes and and dimensions, dimensions, which which were were deliberately deliberately kept kept constant constant to standardize to standardize the surface-to-volume the surface-to-volume ratio ofratio the samplesof the samples grown grown with different with different techniques. techniques. In this particularIn this particular backdrop, backdrop, the morphology-induced the morphology- surface-to-volumeinduced surface-to ratio-volume standardization ratio standardization is very important, is very soimportant that the, realso that essence the ofreal Au–ZnO essence contact of Au– behaviorZnO contact can be behavior probed can because be probed of defect because engineering-induced of defect engineering change-induced in Schottky change barrier in Schottkyheights, ratherbarrier than heights the, change rather than in morphology. the change in All morphology. of the ZNRs All had of thethe sameZNRs diameter had the same of ~200 diameter nm, and of ~200 the µ ZNRnm, and lengths the wereZNR ﬁxedlengths to ~1.5were fixedm. The to MAG~1.5 µm. solution The MAG replacement solution method replacement was used method to control was used the MAGto control ZNR the morphology, MAG ZNRwhere morphology, the 50 mMwhere solution the 50 wasmM replacedsolution was three replaced times to three obtain times the to desired obtain resultsthe desired [21]. results [21].

Nanomaterials 2018, 8, 980 4 of 11 Nanomaterials 2018, 8, x FOR PEER REVIEW 4 of 11

FigureFigure 2.2. SEMSEM imagesimages ofof ZNRsZNRs growngrown withwith ((aa)) ACG,ACG, ((bb)) Ga-dopedGa-doped ACG,ACG, ((cc)) OPT-ACG,OPT-ACG, andand ((dd)) MAGMAG methods.methods. TheThe scalescale barsbars inin thethe toptop andand bottombottom imagesimages correspondcorrespond toto 33 andand 11 microns,microns, respectively.respectively.

3.2.3.2. ZNRZNR StructuralStructural andand CrystallineCrystalline PropertiesProperties XRDXRD analysisanalysis waswas exploitedexploited toto studystudy thethe crystallinecrystalline andand structuralstructural characteristics of the ZNRs. FigureFigure3 3 shows shows the the XRD XRD responses responses of of all all the the samples. samples. It It is is confirmed confirmed that that all all the the samples samples showed showed the the typicaltypical hexagonalhexagonal wurtzite wurtzite ZnO ZnO character character with with all all the the peaks peaks along along the the pertinent pertinent ZnO ZnO phase. phase. In all In of all the of samples,the samples, the tallest the peaktallest along peak the along 002 directionthe 002 confirms direction the confirms ZNR vertical the ZorientationNR vertical perpendicular orientation toperpendicular the substrate to surface. the substrate The XRD surface. phase-dependent The XRD phase best-dependent ZNR orientation best ZNR is orientation found in MAG is found ZNRs, in whereMAG theZNRs 100, where and 101 the peak 100 intensitiesand 101 peak were intensities the minimum. were However,the minimum. the worst However, ZNR orientationthe worst ZNR was foundorientation in Ga-doped was found ACG in ZNRsGa-doped sample. ACG The ZNRs crystalline sample. quality The crystalline and defect quality structure and weredefect confirmed structure bywere measuring confirmed the by full measuring width half the maximum full width (FWHM) half maximum of the 002(FWHM) peaks, of which the 002 was peaks, further which used was to calculatefurther used the crystalliteto calculate size the of crystallite the samples size via of the Scherrersamples via formula the Scherrer [25]: formula [25]:

DD==((0.890.89λ휆))//(B퐵 cos푐표푠휃θ)),, ( (2)2) where B is the FWHM of the 002 peak in radians, λ is the X-ray wavelength, and θ is the 002 peak where B is the FWHM of the 002 peak in radians, λ is the X-ray wavelength, and θ is the 002 peak diffraction angle. Besides the crystallite size, the crystalline qualities of the samples were further diffraction angle. Besides the crystallite size, the crystalline qualities of the samples were further checked by measuring the strain along the 002 phase. The strain (ε) was calculated by the relation: checked by measuring the strain along the 002 phase. The strain (ε) was calculated by the relation: 퐶− 퐶 (3) ε = 표 × 100, 퐶표 C − C ε = o × 100, (3) where co is the lattice constant of stress-free bulkCo ZnO (5.205 Å ). For stain calculation, the lattice constants (c) of the fabricated samples were measured by: where co is the lattice constant of stress-free bulk ZnO (5.205 Å). For stain calculation, the lattice 휆 푐 = , constants (c) of the fabricated samples were measured푠𝑖푛휃 by: (4) where θ corresponds the XRD 002 peak, and λ is theλ X-ray wavelength. The details of the structural c = , (4) characteristics of all the samples are provided in Tablesinθ 1. It is evident that the DMAG > DOPT-ACG > DGa- doped ACG > DACG, which confirms the deterioration in the crystalline structure in ACG and Ga-doped where θ corresponds the XRD 002 peak, and λ is the X-ray wavelength. The details of the structural ACG samples. Similarly, MAG and OPT-ACG samples were less prone to strain than ACG and Ga- characteristics of all the samples are provided in Table1. It is evident that the D > D doped ACG samples. The negative sign in strain values implies that all of the samplesMAG encounteredOPT-ACG > D > D , which conﬁrms the deterioration in the crystalline structure in ACG and compressiveGa-doped ACG strain ratherACG than tensile. The yield stress (σ) in the samples could be calculated by the Ga-doped ACG samples. Similarly, MAG and OPT-ACG samples were less prone to strain than Hall–Petch equation: ACG and Ga-doped ACG samples. The negative sign in strain values implies that all of the samples 푘 encountered compressive strain rather thanσ tensile.= 휎 + The, yield stress (σ) in the samples could be 0 √D (5) calculated by the Hall–Petch equation: where 휎0 and 푘 are constants to the material that makek the factor σy inversely proportional to the σ = σ0 + √ , (5) crystallite size D. Hence, the least stress was expected Dof the large crystallite size MAG sample, and the highest stress was encountered in least crystallite size ACG sample. Furthermore, the XRD results were in accordance to the PL results provided in the next section.

Nanomaterials 2018, 8, 980 5 of 11

where σ0 and k are constants to the material that make the factor σy inversely proportional to the crystallite size D. Hence, the least stress was expected of the large crystallite size MAG sample, and the highest stress was encountered in least crystallite size ACG sample. Furthermore, the XRD results wereNanomaterials in accordance 2018, 8, x FOR to the PEER PL REVIEW results provided in the next section. 5 of 11

Figure 3. XRD patterns of ZNRs grown with ( a)) ACG, ( b) Ga Ga-doped-doped ACG, (c)) OPT OPT-ACG,-ACG, and ( d) MAG methods.

Table 1.1. Structural and opticaloptical parametersparameters ofof allall thethe samples.samples.

Sample θ (Deg.) FWHFθ (Rad.)FWHF DD (nm) c c (Å)ε ε (%) R Sample R ACG 17.28(Deg.) 0.0048 (Rad.) (nm) 30.44 (Å) 5.184(%) −0.403 03.33 Ga-doped ACGACG 17.26 17.28 0.0047 0.0048 30.44 31.13 5.184 5.185−0.403 −03.330.384 25.00 OPT-ACGGa-doped 17.24 ACG 17.26 0.0033 0.0047 31.13 44.19 5.185 5.202−0.384 −25.000.057 47.12 MAG 17.23 0.0024 62.27 5.203 −0.038 58.15 OPT-ACG 17.24 0.0033 44.19 5.202 −0.057 47.12 MAG 17.23 0.0024 62.27 5.203 −0.038 58.15 3.3. ZNR Photoluminescent Optical Characteristics 3.3. ZNRThe crystallinePhotoluminescent structure, Optical optical Characteristics characteristics, and the presence of multiple defects were further classified with PL spectroscopy. Figure4 shows the room temperature PL response of all the samples The crystalline structure, optical characteristics, and the presence of multiple defects were in the range of 300 to 700 nm. All the samples showed a high intensity peak in the UV region because further classified with PL spectroscopy. Figure 4 shows the room temperature PL response of all the of the exciton recombination process. However, the comparative intensity of the UV peaks and the samples in the range of 300 to 700 nm. All the samples showed a high intensity peak in the UV region UV-to-visible peak intensity ratio could be further exploited to test the crystalline structure, purity, because of the exciton recombination process. However, the comparative intensity of the UV peaks and smooth optical character of the ZNRs. The highest UV peaks were shown by the OPT-ACG and and the UV-to-visible peak intensity ratio could be further exploited to test the crystalline structure, MAG samples; however, the smallest UV peak was acquired from the ACG sample. purity, and smooth optical character of the ZNRs. The highest UV peaks were shown by the OPT- ACGThe and defectMAG engineeringsamples; however, was verified the smallest by studying UV peak the was PL acquired responses from in the ACG visible sample. region. It is evident that, depending upon the growth conditions, the PL peak intensities are different in the visible The defect engineering was verified by studying the PL responses in the visible region. It is region.evident The that, heighest depending peak upon intensity the growthwas shown conditions, by the ACG the PL sample, peak which intensities confirms are different the presence in the of visible region. The heighest peak intensity was shown by the ACG sample, which confirms the presence of many defects in the crystal structure corresponding to the visible band. However, the peak intensity is reduced for the Ga-doped ACG sample, which confirms that Ga tends to fix many subsurface defects in the ZnO crystal [23]. On the contrary, the OPT-ACG and MAG samples show almost a flatband condition in the visible region, which confirms that the OPT process removes many surface and subsurface defects corresponding to the visible region, and the downward band bending

Nanomaterials 2018, 8, 980 6 of 11 many defects in the crystal structure corresponding to the visible band. However, the peak intensity is reduced for the Ga-doped ACG sample, which confirms that Ga tends to fix many subsurface defects in the ZnO crystal [23]. On the contrary, the OPT-ACG and MAG samples show almost a flatband conditionNanomaterials in 2018 the, 8, visiblex FOR PEER region, REVIEW which confirms that the OPT process removes many surface6 andof 11 subsurface defects corresponding to the visible region, and the downward band bending in Figure4c isin theFigure evidence 4c is the to the evidence reported to fact.the reported The highest fact. UV The peak highest intensity UV peak and theintensity least visible and the peak least intensity visible ispeak shown intensity by the is MAGshown sample, by the MAG which sample, testifies which the perfect testifies ZNR the perfect crystal structureZNR crystal with structure best optical with characteristics.best optical characteristics. The UV-to-visible The UV band-to-visible ratio (R)band is anratio indicator (R) is an of indicator the ZNR of crystal the ZNR quality, crystal where quality, the greaterwhere the value greater of R isvalue analogous of R is analog to a betterous crystallineto a better crystalline structure. Asstructure. per the As values per the provided values inprovided Table1, itin is Table evident 1, it thatis evident RMAG that> ROPT-ACG RMAG > R>OPT RGa-doped-ACG > RGa ACG-doped> ACG RACG > R,ACG which, which is in is accordance in accordance to the to restthe rest of the of structuralthe structural characteristics characteristics in Table in Table1. The 1. PLThe optical PL optical and and XRD XRD structural structural characteristics characteristics established established the basethe base to explain to explain the disparitiesthe disparities in Au–ZnO in Au–ZnO contact contact behavior behavior presented presented in the in nextthe next section. section.

FigureFigure 4. PLPL spectra spectra of ofZNR ZNRss grown grown with with (a) ACG, (a) ACG, (b) Ga (b)-doped Ga-doped ACG, ACG, (c) OPT (c)-ACG, OPT-ACG, and (d and) MAG (d) MAGmethods. methods.

3.4.3.4. Growth Method-DependentMethod-Dependent Defect Engineering ZnO crystal quality and defect formation play a crucial role in Au–ZnO Au–ZnO Schottky Schottky barrier barrier engineering. ZnO ZnO has has built-in built-in properties properties of both of both covalent covalent and ionic and ionic semiconductors, semiconductors because, because the carrier the transportcarrier transport and defect and formationdefect formation are quite are complex quite complex in ZnO. in The ZnO. defect The generationdefect generation plays a plays crucial a crucial role in modulatingrole in modulating the ZnO the carrier ZnO concentration,carrier concentration because, because most of most the ZnO of the defects ZnO aredefects donor are defects donor [defects26–28]. The[26– topic28]. The is still topic very is still controversial, very controv andersial this is, and a moot this pointis a moot in research point in and research development, and development, but oxygen vacanciesbut oxygen are vacancies still considered are still to considered be one common to be defect one common that acts defect as an electron that acts donor as an to electron ZnO. The donor theory to forZnO. ambient The theory hydrogen for ambient being hydrogen used as a being shallow used donor as a shallow to ZnO donor was ﬁrst to ZnO propounded was first propounded by Mollwo, by Mollwo, and substantiated by Van de Walle [29,30]. It is believed that hydrogen reacts with ZnO surface oxygen and forms OH− ions at ambient growth conditions by donating an electron to the surface accumulation layer [31–33]. The process could be explained by the equation [34]: H + O2− OH− + e− (6)

where O2− is the surface-adsorbed oxygen via the chemisorption process, which is desorbed by H, and the depletion region is relaxed. In contact behavior studies, it is believed that the extrinsic factors,

Nanomaterials 2018, 8, 980 7 of 11 and substantiated by Van de Walle [29,30]. It is believed that hydrogen reacts with ZnO surface oxygen and forms OH− ions at ambient growth conditions by donating an electron to the surface accumulation layer [31–33]. The process could be explained by the equation [34]:

H + O2− → OH− + e− (6) where O2− is the surface-adsorbed oxygen via the chemisorption process, which is desorbed by H, and the depletion region is relaxed. In contact behavior studies, it is believed that the extrinsic factors, such as surface preparation and growth conditions, have a strong impact on the reduction of the Schottky barrier lowering. However, we believe that the intrinsic factors, such as defect formation, are directly dependent upon the extrinsic factors such as growth conditions. For instance, the thermal energy provided to the ZnO atoms to sit on proper crystallographic planes was not sufﬁcient in low-temperature ZNR growth via the ACG method, and the process leads to the generation of many defects, and the vacant cites have been supplanted by donor impurities such as hydrogen [35]. Secondly, the existence of the sample in the growth solution for a long time (12 hr in the ACG method) paves the way for impurity adsorption on the ZNR surface. On the contrary, the OPT process removes the adsorbed impurities on the ZNR surface, and high-power MAG provides the energy required by the atoms to nucleate from the proper crystal sites, and both processes lead to the least generation of defects. Hence, the extrinsic factors directly affect the generation of defects that modulate the fermi level position and Schottky barrier pinning. Because of defect density-induced Schottky barrier pinning, the effective barrier height as well as width at the Au–ZnO interface would be changed, which supports Ohmic/Schottky behaviors to ZnO. Furthermore, some defects are introduced into the ZnO bandgap to support electron hopping through the barrier.

3.5. Au–ZnO Contact Behavior Depending upon the base provided in Section 3.4, the Au–ZnO contact behavior could be divided into multiple sections and is provided in the next section.

3.5.1. Au–ZnO Ohmic Contact to ZNRs Figure5 shows the electrical characteristics of ACG and Ga-doped ACG samples. It is evident that the ACG sample shows a quasi-Ohmic response, while Ga-doped ACG sample shows a more pronounced Ohmic response to the Au contact. The reason for this unusual response to Au is the presence of many surface and subsurface defects in ACG ZNRs, as explained earlier in Section 3.4. Also, further impetus was provided by the formation of an accumulation layer via hydrogen adsorption, which denounces the formation of the Schottky barrier at the ZNR surface via the chemisorption process. This leads to the Schottky barrier pinning, where the barrier height and width are decreased at the interface. The relation between the tunneling energy “E” and the carrier concentration “N” is governed by the equation [36]: √ qh N E = √ (7) 4π mεoεs where h is the Plank constant, m is the effective mass, and εs is the ZnO dielectric√ constant. It is evident in Equation (7) that the tunneling energy E is directly proportional to the factor N. The average carrier concentrations in the ACG and Ga-doped samples were measured via a Hall-effect measurement setup, which was of the order of 1018 and 1021 cm−3, respectively. Hence, the carrier transport in ACG and Ga-doped ACG samples was governed by thermionic emission by reducing the Schottky barrier height, ﬁeld emission through the thin barrier, and thermionic ﬁeld emission near the top of the barrier. At high voltages, the deep-level defects could be further supported by electron hopping within the forbidden band. To increase the carrier concentration, the ACG sample was further doped with Ga. The doped sample showed a more pronounced Ohmic contact to Au with a highest intensity current, as shown in Figure5b. The Fermi level position in bulk ZnO was predicted to be 0.3 eV under Nanomaterials 2018, 8, 980 8 of 11 the conduction band [37]. However, we believe that the Fermi level position that was very close to the conduction band and diffusion into the conduction band in ACG and Ga-doped ACG samples, respectively, was the reason for these Au–ZnO quasi-Ohmic (ACG) and Ohmic contacts (Ga-doped ACG). The Schottky barrier height is governed by: qϕSB = qVB + Ec − Ef (8) where the quantity (Ec − Ef) < 0 for ACG and Ga-doped ACG samples at the interface and the effective valueNanomaterials of ϕSB 2018was, 8, decreased. x FOR PEER REVIEW 8 of 11

FigureFigure 5.5. II--VV characteristicscharacteristics ofof ((aa)) ACGACG ZNRsZNRs andand ((bb)) Ga-dopedGa-doped ACGACG ZNRs.ZNRs.

3.5.2.3.5.2. Au–ZnOAu–ZnO SchottkySchottky ContactContact toto ZNRsZNRs ContraryContrary toto thethe ACGACG andand Ga-dopedGa-doped ACGACG samples,samples, thethe OPT-ACGOPT-ACG andand MAGMAG samplessamples showedshowed Au–ZnOAu–ZnO SchottkySchottky contacts. contacts. Figure 6 6a,ba,b showshow the the II--VV characteristics of thethe OPT-ACGOPT-ACG and and MAG MAG samples,samples, respectively.respectively. Both Both samples samples show show Schottky-type Schottky-typeI- VI-Vcharacteristics characteristics because because of of the the decrease decrease in thein the carrier carrier concentration-induced concentration-induced increase increase in Schottky in Schottky barrier barrier height height and and the barrierthe barrier broadening broadenin atg the at interface.the interface. The The OPT OPT process process was was used used as a ZNRas a ZNR surface surface cleanser, cleanser, where wh itere removed it removed the hydrogen the hydrogen from ZNRfrom surface,ZNR surface which, which is a source is a source of donor of donor electrons. electrons. Furthermore, Furthermore, the accumulationthe accumulation layer layer formed formed on theon the ZNR ZNR surface surface near near the the Au–ZnO Au–ZnO interface interface was was also also removed removed in in the the OPT-ACG OPT-ACG and and MAG MAG samples, samples, whichwhich provided provided further further impetus impetus to the to chemisorption the chemisorption process process and strengthened and strengthened the interface the Schottkyinterface barriersSchottky in barriers both samples. in both Znsamples. vacancies Zn vacancies act as deep act acceptors, as deep acceptors and oxygen, and vacancies oxygen vacancies are the primary are the sourcesprimary of source deeps donors of deep in donors ZnO. The in ZnO. OPT processThe OPT did process not affect did thenot Zn-acceptoraffect the Zn vacancies,-acceptor vacancies but ﬁlled, thebut potential filled the sites potential for donor sites oxygen for donor vacancies, oxygen reducing vacancies, the overall reducing n-type the carrier overall concentration n-type carrier in theconcentration ZNR channel. in the It alsoZNR removed channel. theIt also hydrocarbon removed the contamination hydrocarbon from contamination the surface, from which the aided surface in, thewhich defect aided formation in the defect process. formation As explained process. in As Section explained 3.4, the in Section immaculate 3.4, the nature immaculate of MAG nature method of grewMAG cleanmethod and grew low-defect clean and ZNRs low with-defect the ZNRs least carrierwith the concentration. least carrier concentration. A reductionin A visiblereduction band in intensityvisible band in PL intens spectraity (see in Figure PL spectra4c,d) is(see the Fevidenceigure 4c to,d) the is above-mentioned the evidence to phenomenon.the above-mentioned Hence, – thephenomenon. MAG and OPTHence, processes the MAG did and not OPT only processes eliminate did the OHnot onlyion eliminate inserted to the the OH surface– ion inserted accumulation to the layer,surface but accumulation also reduced layerthe deep, but level also defects reduced to eliminate the deep the level effects defects of processes to eliminate such the as tunneling, effects of depletion-widthprocesses such thinning,as tunneling, and hopping. depletion As-width per Equation thinning, (7), and the ho tunnelingpping. As energy per EEquation is the minimum (7), the intunneling OPT-ACG energy and MAG E is the samples, minimum hence, in theOPT carrier-ACG transportand MAG is samples, only governed hence, by the thermionic carrier transport emission is at high voltage levels. Also, the factor E − E was greater than zero because of the Fermi level shifting only governed by thermionic emissionc at highf voltage levels. Also, the factor Ec − Ef was greater than belowzero because the conduction of the Fermi band, level which shifting further below strengthened the conduction the Schottky band, barriers which further (ϕSB) as strengthened per Equation the (8). Schottky barriers (φSB) as per Equation (8).

Figure 6. I-V characteristics of (a) OPT-ACG ZNRs and (b) MAG ZNRs.

Nanomaterials 2018, 8, x FOR PEER REVIEW 8 of 11

Figure 5. I-V characteristics of (a) ACG ZNRs and (b) Ga-doped ACG ZNRs.

3.5.2. Au–ZnO Schottky Contact to ZNRs Contrary to the ACG and Ga-doped ACG samples, the OPT-ACG and MAG samples showed Au–ZnO Schottky contacts. Figure 6a,b show the I-V characteristics of the OPT-ACG and MAG samples, respectively. Both samples show Schottky-type I-V characteristics because of the decrease in the carrier concentration-induced increase in Schottky barrier height and the barrier broadening at the interface. The OPT process was used as a ZNR surface cleanser, where it removed the hydrogen from ZNR surface, which is a source of donor electrons. Furthermore, the accumulation layer formed on the ZNR surface near the Au–ZnO interface was also removed in the OPT-ACG and MAG samples, which provided further impetus to the chemisorption process and strengthened the interface Schottky barriers in both samples. Zn vacancies act as deep acceptors, and oxygen vacancies are the primary sources of deep donors in ZnO. The OPT process did not affect the Zn-acceptor vacancies, but filled the potential sites for donor oxygen vacancies, reducing the overall n-type carrier concentration in the ZNR channel. It also removed the hydrocarbon contamination from the surface, which aided in the defect formation process. As explained in Section 3.4, the immaculate nature of MAG method grew clean and low-defect ZNRs with the least carrier concentration. A reduction in visible band intensity in PL spectra (see Figure 4c,d) is the evidence to the above-mentioned phenomenon. Hence, the MAG and OPT processes did not only eliminate the OH– ion inserted to the surface accumulation layer, but also reduced the deep level defects to eliminate the effects of processes such as tunneling, depletion-width thinning, and hopping. As per Equation (7), the tunneling energy E is the minimum in OPT-ACG and MAG samples, hence, the carrier transport is only governed by thermionic emission at high voltage levels. Also, the factor Ec − Ef was greater than Nanomaterialszero because2018 of, 8 ,the 980 Fermi level shifting below the conduction band, which further strengthened9 of the 11 Schottky barriers (φSB) as per Equation (8).

FigureFigure 6. 6.I -IV-V characteristicscharacteristics ofof ((aa)) OPT-ACGOPT-ACG ZNRsZNRs andand ( b(b)) MAG MAG ZNRs. ZNRs.

3.5.3. Au Deposition Process Details

At times, the metallization process also incorporates many defects into the ZnO crystal, which further supports field emission and thermionic field emission processes. The metallization chemistry and phenomenon is complicated, and is still a matter of debate in research and development on ZnO metallization schemes. The metallization process has the probability of dislodging many oxygen atoms from the crystal structure which move towards the metal surface via diffusion. The out-diffusion of oxygen atoms leaves behind oxygen vacancies that are donor defects to ZnO, and further supports tunneling near the metal-semiconductor surface. However, we believe that it was not the case in our devices’ metallization process, because at first, the metal was evaporated via a state of the art e-beam metal evaporation process under high-vacuum conditions. Secondly, Au is one of the inert metals that does not incorporate defects at the Au–ZnO interface via oxygen vacancies’ out-diffusion, and supports barrier pinning. Also, it does not react with ambient oxygen to form an insulating oxide layer at the Au–ZnO interface to support barrier width broadening. Hence, the propounded results are purely on the basis of defect engineering rather than any complementary processes.

4. Conclusions In this study, the disparities in Au–ZnO contact behavior were probed via experimentation and results. The growth method-induced defect engineering was cited as the main reason for the said disparities. Two antithetical growth methods, namely ACG and MAG, were used for ZNR growth. The ACG samples were further doped with Ga and subjected to OPT to probe into the subsurface and surface-level defect engineering-centric change in carrier concentration, and its effects on Au–ZnO Schottky barrier heights. Also, the ZNR morphology and dimensions were standardized to study the change in Au–ZnO contact behavior because of defects/carriers, rather than the change in surface-to-volume ratio. The XRD and PL results qualiﬁed that the MAG sample had the best crystalline quality and optical character. It was further argued that the OPT process removed many surface and subsurface level defects in the ZnO crystal, which improved the ZNR crystalline structure and optical characteristics. Despite having a theoretical Schottky barrier at the Au–ZnO interface, the ACG and Ga-doped ACG samples showed a quasi-Ohmic and Ohmic behaviors, respectively. This was because of the defects and doping-induced increase in carrier concentration in the ZNR channel and the Schottky barrier pinning, which facilitated tunneling through the thin barrier. Contrarily, the defect scanty OPT-ACG and MAG samples further strengthened the Schottky barriers at the Au–ZnO interface and showed Schottky contacts to ZnO. Hence, the reason for the change in Au–ZnO contact behavior lies in the method-dependent defect engineering, which needs to be modulated for the speciﬁed use of contact in application-oriented research and development on Au–ZnO. Nanomaterials 2018, 8, 980 10 of 11

Author Contributions: A.u.H.S.R. fabricated the devices, designed and conducted all the experiments, and analyzed the data. H.-S.K. planned and supervised the project. Both authors contributed to discussing the results and writing the manuscript. Funding: This work was supported by the Ministry of Trade, Industry and Energy (MOTIE, Korea) under the Technology Innovation Programs (No. 10063682 and No. 10073122). Conﬂicts of Interest: There are no conﬂict of interest to declare.

References

1. Bagnall, D.M.; Chen, Y.F.; Zhu, Z.; Yao, T.; Koyama, S.; Shen, M.Y.; Goto, T. Optically pumped lasing of ZnO at room temperature. Appl. Phys. Lett. 1997, 70, 2230–2232. [CrossRef] 2. Keis, K.; Vayssieres, L.; Rensmo, H.; Lindquist, S.E.; Hagfeldt, A. Photoelectrochemical properties of nano-to microstructured ZnO electrodes. J. Electrochem. Soc. 2001, 148, A149–A155. [CrossRef] 3. Amin, M.; Manzoor, U.; Islam, M.; Bhatti, A.S.; Shah, N.A. Synthesis of ZnO nanostructures for low temperature CO and UV sensing. Sensors 2012, 12, 13842–13851. [CrossRef][PubMed] 4. Kong, X.Y.; Wang, Z.L. Spontaneous polarization-induced nanohelixes, nanosprings, and nanorings of piezoelectric nanobelts. Nano Lett. 2003, 3, 1625–1631. [CrossRef]

5. Rana, A.S.; Chang, S.B.; Kim, H.S. NH4OH-Oriented and pH-dependent growth of ZnO nanostructures via microwave-assisted growth method. J. Nanosci. Nanotechnol. 2018, 18, 2125–2127. [CrossRef][PubMed] 6. Somvanshi, D.; Jit, S. Analysis of temperature-dependent electrical characteristics of n-ZnO nanowires (NWs)/p-Si heterojunction diodes. IEEE Trans. Nanotechnol. 2014, 13, 62–69. [CrossRef] 7. Rana, A.S.; Kang, M.; Jeong, E.S.; Kim, H.S. Transition between ZnO nanorods and ZnO nanotubes with their antithetical properties. J. Nanosci. Nanotechnol. 2016, 16, 10772–10776. [CrossRef] 8. Vinod, R.; Sajan, P.; Achary, S.R.; Tomas, C.M.; Munoz-Sanjose, V.; Bushiri, M.J. Enhanced UV emission from ZnO nanoﬂowers synthesized by the hydrothermal process. J. Phys. D Appl. Phys. 2012, 45, 425103. [CrossRef]

9. Hjiri, M.; El Mir, L.; Leonardi, S.G.; Donato, N.; Neri, G. CO and NO2 selective monitoring by ZnO-based sensors. Nanomaterials 2013, 3, 357–369. [CrossRef][PubMed] 10. Konenkamp, R.; Word, R.C.; Schlegel, C. Vertical nanowire light-emitting diode. Appl. Phys. Lett. 2004, 85, 6004–6006. [CrossRef] 11. Goldberger, J.; Sirbuly, D.J.; Law, M.; Yang, P. ZnO nanowire transistors. J. Phys. Chem. B 2005, 109, 9–14. [CrossRef][PubMed] 12. Huang, M.H.; Mao, S.; Feick, H.; Yan, H.; Wu, Y.; Kind, H.; Weber, E.; Russo, R.; Yang, P. Room-z temperature ultraviolet nanowire nanolasers. Science 2001, 292, 1897–1899. [CrossRef][PubMed] 13. Suehiro, J.; Nakagawa, N.; Hidaka, S.I.; Ueda, M.; Imasaka, K.; Higashihata, M.; Okada, T.; Hara, M. Dielectrophoretic fabrication and characterization of a ZnO nanowire-based UV photosensor. Nanotechnology 2006, 17, 2567. [CrossRef][PubMed] 14. Zhang, F.; Chao, D.; Cui, H.; Zhang, W.; Zhang, W. Electronic structure and magnetism of Mn-Doped ZnO nanowires. Nanomaterials 2015, 5, 885–894. [CrossRef][PubMed] 15. Law, M.; Greene, L.E.; Johnson, J.C.; Saykally, R.; Yang, P. Nanowire dye-sensitized solar cells. Nat. Mater. 2005, 4, 455–459. [CrossRef][PubMed] 16. Liang, S.; Sheng, H.; Liu, Y.; Huo, Z.; Lu, Y.; Shen, H. ZnO Schottky ultraviolet photodetectors. J. Cryst. Growth 2001, 225, 110–113. [CrossRef] 17. Rana, A.S.; Lee, J.Y.; Hong, Y.P.; Kim, H.S. Transient current response for ZnO nanorod-based doubly transparent UV sensor fabricated on ﬂexible substrate. Phys. Status Solidi RRL 2018, 12, 1800001. [CrossRef] 18. Mott, N.F. Note on the contact between a metal and an insulator or semi-conductor. Proc. Cambridge Philos. Soc. 1938, 34, 568. [CrossRef] 19. Rabadanov, R.A.; Guseikhanov, M.K.; Aliev, I.S.; Semiletov, S.A. Properties of metal-zinc oxide contacts. Soviet Phys. J. 1981, 24, 548–551. [CrossRef] 20. Rana, A.S.; Ko, K.; Hong, S.; Kang, M.; Kim, H.S. Fabrication and characterization of ZnO nanorods on multiple substrates. J. Nanosci. Nanotechnol. 2015, 15, 8375–8380. [CrossRef][PubMed] Nanomaterials 2018, 8, 980 11 of 11

21. Rana, A.S.; Kang, M.; Kim, H.S. Microwave-assisted facile and ultrafast growth of ZnO nanostructures and proposition of alternative microwave-assisted methods to address growth stoppage. Sci. Rep. 2016, 6, 24870. [CrossRef][PubMed] 22. Rana, A.S.; Chang, S.B.; Chae, H.U.; Kim, H.S. Structural, optical, electrical and morphological properties of different concentration sol-gel ZnO seeds and consanguineous ZnO nanostructured growth dependence on seeds. J. Alloy. Comp. 2017, 729, 571–582. [CrossRef]

23. Rana, A.S.; Kim, H.S. NH4OH Treatment for an optimum morphological trade-off to hydrothermal Ga-doped n-ZnO/p-Si heterostructure characteristics. Materials 2017, 11, 37. [CrossRef][PubMed] 24. Rana, A.S.; Shahid, A.; Lee, J.Y.; Kim, H.S. High-Power microwave-assisted Ga doping, an effective method to tailor n-ZnO/p-Si heterostructure optoelectronic characteristics. Phys. Status Solidi A 2018, 215, 1700763. [CrossRef] 25. Cullity, B.D.; Rstock, S. Elements of X-ray Diffraction; Prentice Hall: Upper Saddle River, NJ, USA, 2001. 26. Lany, S.; Zunger, A. Anion vacancies as a source of persistent photoconductivity in II-VI and chalcopyrite semiconductors. Phys. Rev. B 2005, 72, 035215. [CrossRef] 27. Look, D.C.; Hemsky, J.W.; Sizelove, J.R. Residual native shallow donor in ZnO. Phys. Rev. Lett. 1999, 82, 2552. [CrossRef] 28. Janotti, A.; Van deWalle, C.G. Hydrogen multicentre bonds. Nat. Mater. 2007, 6, 44–47. [CrossRef][PubMed] 29. Thomas, D.G.; Lander, J.J. Hydrogen as a donor in zinc oxide. J. Chem. Phys. 1956, 25, 1136–1142. [CrossRef] 30. Van de Walle, C.G. Hydrogen as a cause of doping in zinc oxide. Phys. Rev. Lett. 2000, 85, 1012. [CrossRef] [PubMed] 31. Moormann, H.; Kohl, D.; Heiland, G. Variations of work function and surface conductivity on clean cleaved zinc oxide surfaces by annealing and by hydrogen adsorption. Surf. Sci. 1980, 100, 302–314. [CrossRef] 32. Heiland, G.; Kunstmann, P. Polar surfaces of zinc oxide crystals. Surf. Sci. 1969, 13, 72–84. [CrossRef] 33. Nakagawa, M.; Mitsudo, H. Anomalous temperature dependence of the electrical conductivity of zinc oxide thin films. Surf. Sci. 1986, 175, 157–176. [CrossRef] 34. Henrich, V.E.; Cox, P.A. The Surface Science of Metal Oxides; Cambridge University Press: Cambridge, UK, 1994; p. 297. 35. Rana, A.S.; Lee, J.Y.; Shahid, A.; Kim, H.S. Growth method-dependent and defect density-oriented structural, optical, conductive, and physical properties of solution-grown ZnO nanostructures. Nanomaterials 2017, 7, 266. [CrossRef][PubMed] 36. Crowell, C.R.; Rideout, V.L. Normalized thermionic-field (TF) emission in metal-semiconductor (Schottky) barriers. Solid State Electron. 1969, 12, 89–105. [CrossRef] 37. Jacobi, K.; Zwicker, G.; Gutmann, A. Work function, electron affinity and band bending of zinc oxide surfaces. Surf. Sci. 1984, 141, 109–125. [CrossRef]

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).