CS

Xa

b

a

ARRA

KZcp

1

tbrpsfl

tsidtm([

IC

h1

Journal of Manufacturing Processes 32 (2018) 438–444

Contents lists available at ScienceDirect

Journal of Manufacturing Processes

j ourna l ho me pa g e: www.elsev ier .com/ locate /manpro

haracterization and Modeling of Catalyst-free Carbon-Assistedynthesis of ZnO Nanowires

iangcheng Kong a, Chuang Wei a, Yong Zhu b, Paul Cohen a, Jingyan Dong a,∗

Edward P. Fitts Department of Industrial and Systems Engineering, USADepartment of Mechanical and Aerospace Engineering, North Carolina State University, Raleigh, NC, USA

r t i c l e i n f o

rticle history:eceived 15 November 2017eceived in revised form 21 February 2018ccepted 13 March 2018

eywords:nO nanowire synthesishemical vapor deposition (CVD)

a b s t r a c t

ZnO nanowires have been widely studied due to their unique material properties and many potentialapplications in electronic and optoelectronic devices. Many growth processes have been developed tosynthesize ZnO nanowires. It is critically important to develop predictive process models so as to maxi-mize the output of the nanowire synthesis. Here we report a method to characterize, quantify, and modela catalyst-free carbon-assisted ZnO nanowire growth process. Two key factors were identified for thesynthesis conditions, which are reaction temperature and flow rate. Based on a factorial design method,we conducted experiments with different combinations of the two factors to study their effects on the

rocess modeling process output (i.e. density of the nanowires), which was evaluated by a scanning electron microscope(SEM). The experimental results were analyzed using ANOVA test, and then a semi-empirical model wasbuilt to correlate the ZnO nanowire output with synthesis conditions. This model was able to describethe ZnO nanowire density with respect to synthesis conditions, which can provide guideline for synthesisparameters selection and process optimization.

© 2018 Published by Elsevier Ltd on behalf of The Society of Manufacturing Engineers.

. Introduction

ZnO nanowire is an important semiconducting and piezoelec-ric nanomaterial [1] with a large exciton binding energy, a wideandgap and excellent mechanical properties [2,3]. They haveeceived significant attention in recent years because of theirotential applications in electronic and optoelectronic devices,uch as solar cells [4,5], field emission devices [6,7], transistorsor transparent and flexible electronics [8,9], photodetectors [10],ight-emitting diodes [11] and piezo nanogenerators [12–14].

The methods to synthesize ZnO nanowires can be classified intowo categories: vapor phase synthesis and solution phase synthe-is. A number of vapor phase techniques have been developedncluding vapor liquid solid (VLS) growth [15–18], chemical vaporeposition (CVD) [19–22], metal organic chemical vapor deposi-ion (MOCVD) [23–26], physical vapor deposition (PVD) [27–30],

olecular beam epitaxy (MBE) [31–34], pulsed laser depositionPLD) [35–38], and metal organic vapor phase epitaxy (MOVPE)39–41]. Among these methods, VLS is the simplest vapor phase

∗ Corresponding author at: 414-C Daniels Hall, Campus box 7906, Department ofndustrial and Systems Engineering, North Carolina State University, Raleigh, Northarolina, 27695-7906, USA.

E-mail address: jdong@ncsu.edu (J. Dong).

ttps://doi.org/10.1016/j.jmapro.2018.03.018526-6125/© 2018 Published by Elsevier Ltd on behalf of The Society of Manufacturing E

technique to synthesize ZnO nanowires, especially in large scale[42]. During the VLS growth, the quality and growth behavior ofZnO nanowires are affected by many factors including oxygen par-tial pressure, thickness of the catalyst layer and chamber pressure.Catalyst-free metal-organic chemical vapor deposition (MOCVD)has been applied to produce high-purity ZnO nanowire as theimpurities introduced by catalysts have been eliminated. Also, thegrowth temperature is lower than the growth temperature in VLSprocesses [43]. Solution phase synthesis has many advantages overvapor phase synthesis, including low temperature, low cost, goodscalability. Moreover, many substrates can be chosen for solu-tion phase synthesis. Solution phase synthesis methods includinghydrothermal method [44–50], microemulsion [51], and ethanolbase methods [52]. Due to oxygen vacancies, ZnO nanowires syn-thesized by hydrothermal methods have more crystalline defects[53], which allow nanowires exhibit visible light photo catalysiseven without doping with transition metal. Statistical design ofexperiments have been applied to the hydrothermal depositionprocess to improve growth performance and quality of the ZnOnanowires [54]. Besides ZnO nanowires, kinetics modeling and sta-tistical modeling have been studied for many other nanomaterials

[55–57].

In this work, we developed a method to characterize and modela catalyst-free carbon-assisted ZnO nanowire growth process. Tosynthesize ZnO nanowires using the CVD method, ZnO patterns

ngineers.

X. Kong et al. / Journal of Manufacturing Processes 32 (2018) 438–444 439

system

wdtctctuVsrtbqgt

2

aaatnat

tp(wwn1tw4rs

mqsnd1

Fig. 1. (a) Chemical vapor deposition

ere printed onto a silicon substrate using a novel electrohy-rodynamic (EHD) printing process. We found that two factors,emperature and flow rate, significantly affect the ZnO growth pro-ess. To correlate the growth process conditions with the processhroughput, we conducted experiments with different synthesisonditions base on blocked factorial design method, and quantita-ively characterized the process throughput (i.e. nanowire density)sing scanning electron microscope (SEM) images. The Analysis Ofariance (ANOVA) tests were performed to evaluate the effects andignificance of the process parameters (i.e. temperature and flowate) on the process throughput. A modified two-dimensional ellip-ical Gaussian model was developed to describe the relationshipetween the process parameters and process throughput, so as touantitatively predict the output (i.e. throughput) of ZnO nanowirerowth process. The results from the model agreed very well withhe experimental results.

. Experimental methods

A CVD system was developed for the growth of ZnO nanowires,s shown in Fig. 1, which includes a temperature-controlled furnacend a gas flow control system. The furnace including a quartz tubend a temperature controlled heating element was used to heat uphe sample and chemical reactants to the desired temperature foranowire growth. A quartz tube with the inside diameter of 22 mmnd the length of 100 cm was placed inside the high-temperatureube furnace.

To selectively grow ZnO nanowires on the substrate, ZnO pat-erns were printed onto the silicon substrate by the EHD printingrocess. The ZnO ink was obtained by mixing 1 weight percentwt%) of ZnO nanoparticles and 5 wt% of polyvinylpyrrolidone (PVP)ith ethanol, followed by ultrasonication for two hours. The PVPas used as the surfactant to prevent the aggregation of the ZnO

anoparticles in the ink. During the EHD printing, a voltage of500 V between the nozzle and the substrate was applied to ini-iate the printing process. The printed patterns on the substrateere dried in air and then transferred to an oven for 3 hours at

50 ◦C to decompose the PVP coating on ZnO nanoparticles. Theesulting substrate with the ZnO patterns was used as the collectorubstrate for ZnO nanowire growth process.

In the growth process, ZnO power and graphite powder wereixed with 1:1 ratio as reactants. The reactant was placed in the

uartz boat located in the center of the furnace. The collector sub-

trate (i.e. a silicon die with the printed ZnO patterns) was locatedear the tube outlet. An Argon gas flow was supplied to the tubeuring growth process, with mass flow rates between 1 sccm and00 sccm. The pressure of furnace and the growth time were set

(b) Furnace with quartz tube inside.

to be 100 Torr and 30 minutes respectively. After the temperatureof quartz boat reaches reaction temperature, the ZnO is reduced bygraphite, and gaseous products will be produced including Zn vaporand CO. As the printed ZnO patterns contain huge amount of ZnOnanoparticles with large specific surface area, which will serve asnucleus sites for nanowire growth. Once the gaseous products aretransported to cooler area where the collector is located by Argongas flow, the Zn vapor is deposited on the surface of the nucleussites and re-oxidized, resulting in the growth of ZnO nanowires.

3. Results and discussion

High-resolution ZnO patterns were successfully printed ontothe silicon substrate using EHD printing process [58–60] of liquidink with ZnO nanoparticles. The printed ZnO pattern was shown inFig. 2(a) with the smallest line width about 25 �m. With the high-resolution ZnO patterns, ZnO nanowires were selectively grown onthe substrate using CVD process, as shown in Fig. 2(b). With theseselectively fabricated patterns, it is easier to quantitatively measurethe density of the nanowires and to study the effect of the processparameters (such as temperature gradient) to the nanowire growthprocess.

Based on our observations from pre-experiments, a few syn-thesis parameters possibly affect the throughput of ZnO nanowiregrowth process, including furnace temperature, Argon flow rate,and the location on the substrate along the temperature gradientdirection. Significant difference of the nanowire growth results atdifferent process conditions and locations was clearly observed,which result in different nanowires density (i.e. productivity) onthe substrate as shown in Fig. 3.

To control and optimize the synthesis process for achievinghigh throughput of ZnO nanowires, we need to understand therelationship between the produced ZnO nanowires and the corre-sponding synthesis parameters. To analyze and model the process,we first quantitatively characterize the throughput of the pro-duced nanowires. The substrate after growth process was imagedby Scan Electronic Microscope (SEM). The density of the producednanowires is used to evaluate the throughput of the nanowiregrowth process. As it is time consuming to enumerate all the syn-thesis conditions experimentally, we applied Blocked FractionalFactorial Designs method to reduce unnecessary experiments.Based on our pre-experiments, furnace temperature is set to rangefrom 920 ◦C to 1000 ◦C, flow rate varies from 60 to 100 sccm, and

substrate dimension along the temperature gradient direction isabout 14 mm.

We design experiments to cover all the synthesis conditions,including the entire temperature and flow rate range on the whole

440 X. Kong et al. / Journal of Manufacturing Processes 32 (2018) 438–444

Fig. 2. (a) Electrohydrodynamic (EHD) printed ZnO pattern on the substrate. (b) Selectively grown ZnO nanowires.

Fig. 3. SEM image for grown ZnO nanowire under different fabrication conditions (a) Temperature = 960 ◦C, flow rate = 60 sccm, and distance = 6 mm. (b) Temperature = 960 ◦C,flow rate = 80 sccm and distance = 6 mm. (c) Temperature = 960 ◦C, flow rate = 100 sccm and distance = 6 mm.

Table 1ZnO density (number/�m2) under different temperatures and flow rates.

Temperature/◦C

920 940 960 980 1000

Flow rate/ sccm 50 0 0 0 0 060 0 0 0.03 0.04 070 0 0 0.06 0.06 080 0 0.02 0.09 0.11 0

sarfmafaoo

scdeapetJ

Table 2ANOVA test results.

Source Temperature Flow rate Location on Temperature*Flow

90 0 0 0.04 0.04 0100 0 0 0.04 0.03 0

ubstrate. Since we are more interested in the effect of temperaturend flow rate on the growth process, we set temperature and flowate as variables, which are changed for each experiment. As the dif-erent spots on the substrate along the axial direction of the furnace

ay affect the growth results of ZnO nanowires due to the temper-ture distribution, we set the location of the substrate as the blockactor. Thus, for each synthesis experiment (with a fixed temper-ture and flow rate), we take measurements at different locationsn substrate, from 2 mm to 14 mm to reflect the effect of locationsn the growth process.

As shown in Fig. 3, the throughput of the growth process istrongly affected by the synthesis conditions. We quantitativelyharacterize the throughput of the fabricated nanowires by its areaensity to provide a numerical comparison of results from differ-nt growth conditions. The area density of ZnO nanowires is defineds the number of ZnO nanowires per unit area in the region with

rinted ZnO pattern. To estimate the density of ZnO nanowires atach location, the total number of nanowires in a given area inhe SEM image was counted using an open source software, Image. The number of nanowires were manually counted disregarding

substrate rate

Prob>|t| <.0001 0.0027 <.0001 0.0514

their various shapes and lengths. To study the effect of temperatureand flow rate on the throughput of the overall nanowire growthprocess, the average density on a substrate is derived by calcu-lating the mean value of density from different locations on thesubstrate under the same temperature and flow rate. The resultsare summarized in Table 1.

To evaluate the significance of temperature and flow rate on thethroughput of the nanowire growth process, we applied AnalysisOf Variance (ANOVA) to determine the significant factors in deter-mining the throughput of ZnO nanowire growth. From Table 2, thep-value for temperature, flow rate and location on substrate are allsmaller than 0.005, which means, all of these three variables arestatistically significant. As mentioned earlier, we set distance asblock factor, so we will not take distance as significant factor in theanalysis of the overall performance on each substrate. The signifi-cant analysis results match well with the experimental observation.As can be observed from Fig. 4, under each specific value of flowrate, the plot has similar bell-shape with the peak value occurs ataround 970–980 ◦C. When the reaction temperature is below thisvalue, the fabricated ZnO nanowire density will increase with theincrease of temperature, while when temperature is above the opti-

mal value, the ZnO nanowire density decreases with the increasedtemperature.

Based on the statistical analysis, the furnace temperature andflow rate were identified to be the significant factors in determin-

X. Kong et al. / Journal of Manufacturing Processes 32 (2018) 438–444 441

ire de

idtabrbstwiiaf

Wrutpbaam

D

swveflFtFt

Fig. 4. Relationship between ZnO nanow

ng ZnO nanowire growth process. To predict the ZnO density underifferent given synthesis conditions, a quantitative model is neededo describe the relationship between the process input parametersnd process throughput. Based on the experimental results, theell-shape trends can be observed for the nanowire density withespect to the temperature and flow rate. However, the relationshipetween the temperature and the resulting nanowire density is notymmetrical. The nanowire density increases relatively slowly ashe reaction temperature gradually approaches to the peak value,hile the density drops rapidly as the reaction temperature further

ncreases from the peak value. Given such asymmetrical behav-or, the relationship between ZnO density and synthesis factorsre described by a modified two-dimensional elliptical Gaussianunction, as shown in Eq. (1):

Density = c1 exp{

−c5(T − Ts)c6[c2(T − T0)2

+c3(Flowrate − F0)2 + c4 ∗ T ∗ Flowrate]}

(1)

here T0 and F0 are center of peak for the temperature and the flowate. c1 to c4 are coefficients of the model. The term c5(T − Ts)c6 issed to provide the skew response with respect to the reactionemperature. When the Ts is smaller than T0, the term c5(T − Ts)c6

rovides less decay to the density function at smaller temperature,ut much larger decay at higher temperature ranges, which give thesymmetrical function around the center of peak for the temper-ture. The model parameters were calculated by the least squareethod using data in Table 1. The final model is shown in Eq. (2):

ensity = 0.228 ∗ exp

{(−3.02 ∗ 10−10(T − 760)4 ∗ [0.0056(T − 975)2

+0.0039(Flowrate − 82.4)2 + 1.4 ∗ 10−5 ∗ T ∗ Flowrate]

}

(2)

From the resulting model, the peak value of the nanowire den-ity occurs around temperature of 975 ◦C and flow rate of 82.4 sccm,hich agrees well with our experimental observation. To pro-

ide more detailed comparison between the developed model andxperimental results, the effect of the reaction temperature andow rate on the nanowire growth process is plotted on Fig. 5.

ig. 5(a) shows the resulting nanowire density at different reactionemperature for each specific flow rate used in the experiments.ig. 5(b) shows the resulting nanowire density with respect tohe flow rate for each specific reaction temperature used in the

nsity vs Temperature at each flow rate.

experiments. Clearly the regression model successfully predicts thetrends of the experimental results.

To better understand the process and verify the results of theregression model, we selected a representative flow rate withchanging temperatures and a representative temperature withchanging flow rates to compare the effect of these process con-ditions on the nanowire growth process. Fig. 6 shows the resultsof the nanowire growth at a flow rate of 80 sccm with the reactiontemperature changing from a low value of 910 ◦C to a high value of1000 ◦C. The SEM pictures at the center of the substrate are plottedfor a few selected conditions. As can be clearly observed, the regres-sion model well describes the trends of the experimental data. TheSEM images also verify the difference in the nanowire density atdifferent process conditions (i.e. furnace temperature in the Fig. 6).The optimal reaction temperature at this condition is around 980 ◦C,and ZnO nanowire density follows the trend of a skewed bell shapewith respect to the temperature. When the temperature is too low(under 920 ◦C), there is almost no observable nanowire growth.When the temperature gradually increases from 920 ◦C to 980 ◦C,more and more ZnO more nanowires can be counted on the sub-strate. The largest nanowire density was observed at 980 ◦C at thisspecific flow rate. When the temperature becomes even higher, arapid drop of nanowire density can be observed with almost nonanowire growth.

The theoretical explanation for the effect of the temperature onthe ZnO growth process is very complex. In general, the tempera-ture affects the ZnO nanowire growth process in the following twoways: determining the amount of condensed vapor, and affectingthe vapor surface diffusion length. At the low end of temperaturerange (temperature under 930 ◦C), the vapor has little energy, andthe atoms cannot diffuse much as they are in low energy level,which forms a film on substrate and prohibits the nanowire growth.At intermediate temperature zone, as the temperature is highenough, there are sufficient vapors to grow ZnO nanowires and highenough energy for a long atom diffusion length to ensure nanowiregrowth. With even higher temperature (temperature above 990 C),very small amount of vapor condensed on the substrate, result-ing in small even no ZnO nanowire growth. The observed trends

are comparable to the previously reported results [54,61], althoughthe specific process conditions are different due to different processand system configuration.

442 X. Kong et al. / Journal of Manufacturing Processes 32 (2018) 438–444

Fig. 5. (a) Plot of both the modeled and measured value of ZnO density at different temperature for each flow rate with different line color representing model predictionat different flow rate, dots are measured nanowire density. b) Plot of both the modeled and experimental value of ZnO density at different flow rate for each temperature.Different color representing model prediction at different temperature.

F rent teo ZnO

rlprcmt

ig. 6. Plot of modeled and experimental results for ZnO nanowire density at diffef the substrate are displayed. The red dots are the experimental data for measured

Fig. 7 shows the results of the nanowire growth process at a fixedeaction temperature of 960 ◦C with the flow rate changing from aow value of 50 sccm to a high value of 100 flow rate. The SEMictures at the center of the substrate are plotted for a few flow

ate to provide an intuitive comparison among different processonditions. As can be observed, the prediction from the regressionodel fits well to the experimental data (red dots in Fig. 5). Again

he ZnO nanowire density follows the trend of a slightly skewed

mperature with flow rate of 80 sccm, the corresponding SEM images at the centernanowire density and the blue line is the result from the model.

bell shape with respect to the flow rate. The peak nanowire densitywas achieved at 80 sccm, which agrees with our observation fromthe SEM images. With a small flow rate of 60 sccm, there are onlya small number of ZnO nanowires produced. With the increasing

of flow rate, the ZnO nanowires density become larger, the peaknanowire density was obtained under 80 sccm with plenty of ZnOnanowires produced. If the flow rate is further increased, the ZnOnanowire density begin to decrease.

X. Kong et al. / Journal of Manufacturing Processes 32 (2018) 438–444 443

F ent flot r meas

otr(lrigoarrt

4

mZwtwmewcttZagwemt

[

[

[

ig. 7. Plot of modeled and experimental values for ZnO nanowire density at differhe center of the substrate are displayed. The red dots are the experimental data fo

The observed trends of the flow rate can be explained by its effectn the vapor transportation process, as it will change the concen-ration Zn vapor and other product, resulting in different oxygen/Znatios to affect the process of Zn vapor oxidation. With low flow ratee.g. 60 sccm), the oxidation of Zn vapor is possibly inhibited by theimitation of available oxygen and Zn vapor inside the tube. As aesult, very few ZnO nanowires will be fabricated. When flow ratencreases (e.g. 80 sccm), the concentration of Zn vapor and oxy-en/Zn ratios increases, which will provide sufficient oxygen toxidize Zn droplets, so ZnO nanowire synthesis gets boosted with

large amount of ZnO nanowires produced. With a very high flowate, Zn vapor will be quickly blown out of the quartz tube, and theate of the nucleation of ZnO nanowire will be reduced, thereforehe reaction speed decreases.

. Conclusion

In this paper, we report a method to characterize, quantify, andodel the growth of ZnO nanowires. The density of the synthesized

nO nanowires was used to characterize the process throughput,hich was measured using SEM images, so as to compare the

hroughputs under different synthesis conditions. The experimentsere performed based on the Blocked Fractional Factorial Designsethod to cover a broad nanowire growth conditions. From the

xperimental observation and the ANOVA test, two key factorsere identified for this process – temperature and flow rate. We

onducted experiments with a variety of combinations of these fac-ors to study the effects of the process conditions on the processhroughput, and then a regression model was built to correlate thenO nanowire throughput with synthesis conditions (i.e. temper-ture and flow rate). This model was able to describe the effect ofrowth process conditions on the resulting ZnO nanowire densities,

hich provides a guideline for the selection of the synthesis param-

ters and process optimization. The characterization and modelingethod reported here can be readily extended to growth of other

ypes of nanomaterials.

[

w rate with the reaction temperature of 960 ◦C, the corresponding SEM images atured ZnO nanowire density and the blue line is the result from the model.

Acknowledgement

This work was supported in part by the National Science Foun-dation under Grant Award CMMI-1129817.

References

[1] Wang ZL. Zinc oxide nanostructures: growth, properties and applications. J PhysCondens Matter 2004;16:R829–58.

[2] Xu F, Qin Q, Mishra A, Gu Y, Zhu Y. Mechanical properties of ZnO nanowiresunder different loading modes. Nano Res 2010;3:271–80.

[3] Cheng G, Miao C, Qin Q, Li J, Xu F, Haftbaradaran H, Dickey EC, Gao H, ZhuY. Large Anelasticity and Associated Energy Dissipation in Single-CrystallineNanowires. Nat Nanotechnol 2015;10:687–91.

[4] Law M, Greene LE, Johnson JC, Saykally R, Yang P. Nanowire Dye-SensitizedSolar Cells. Nat Mater 2005;4:455–9.

[5] Jean J, Chang S, Brown PR, Cheng JJ, Rekemeyer PH, Bawendi MG, Gradecak S,Bulovic V. ZnO Nanowire Arrays for Enhanced Photocurrent in PbS QuantumDot Solar Cells. Adv Mater 2013;25:2790–6.

[6] Lee CJ, Lee TJ, Lyu SC, Zhang Y, Ruh H, Lee HJ. Field Emission from Well-Aligned Zinc Oxide Nanowires Grown at Low Temperature. Appl Phys Lett2002;81:3648.

[7] Yang Q, Liu Y, Pan C, Chen J, Wen X, Wang ZL. Largely Enhanced Efficiency in ZnONanowire/p-Polymer Hybridized Inorganic/organic Ultraviolet Light-EmittingDiode by Piezo-Phototronic Effect. Nano Lett 2013;13:607–13.

[8] Ju S, Facchetti A, Xuan Y, Liu J, Ishikawa F, Ye P, Zhou C, Marks TJ, Janes DB. Fab-rication of Fully Transparent Nanowire Transistors for Transparent and FlexibleElectronics. Nat Nanotechnol 2007;2:378–84.

[9] Nasr B, Wang D, Kruk R, Rösner H, Hahn H, Dasgupta S. High-Speed, Low-Voltage, and Environmentally Stable Operation of Electrochemically Gated ZincOxide Nanowire Field-Effect Transistors. Adv Funct Mater 2012;23:1750–8.

10] Liu X, Gu L, Zhang Q, Wu J, Long Y, Fan Z. All-Printable Band-Edge Modu-lated ZnO Nanowire Photodetectors with Ultra-High Detectivity. Nat Commun2014;5.

11] Jeon PJ, Kim JS, Lim JY, Cho Y, Pezeshki A, Lee HS, Yu S, Min S-W, ImS. Low Power Consumption Complementary Inverters with N-moS2and P-wSe2Dichalcogenide Nanosheets on Glass for Logic and Light-Emitting DiodeCircuits. ACS Appl Mater Interfaces 2015;7:22333–40.

12] Xue X, Nie Y, He B, Xing L, Zhang Y, Wang ZL. Surface Free-Carrier ScreeningEffect on the Output of a ZnO Nanowire Nanogenerator and Its Potential as a

Self-Powered Active Gas Sensor. Nanotechnology 2013;24:225501.

13] Zhao Y, Deng P, Nie Y, Wang P, Zhang Y, Xing L, Xue X. Biomolecule-Adsorption-Dependent Piezoelectric Output of ZnO Nanowire Nanogeneratorand Its Application as Self-Powered Active Biosensor. Biosens Bioelectron2014;57:269–75.

4 cturin

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[of Silver Nanoparticles by Using a Focused Nanocolloid Jet. Appl Phys Lett2007;90:081905.

[61] Tateno K, Zhang G, Gotoh H, Sogawa T. VLS Growth of Alternating InAsP/InPHeterostructure Nanowires for Multiple-Quantum-Dot Structures. Nano Lett

44 X. Kong et al. / Journal of Manufa

14] Hsu C-L, Chen K-C. Improving Piezoelectric Nanogenerator Comprises ZnONanowires by Bending the Flexible PET Substrate at Low Vibration Frequency.J Phys Chem C 2012;116:9351–5.

15] Li SY, Lee CY, Tseng TY. Copper-Catalyzed ZnO Nanowires on Silicon (100)Grown by Vapor–liquid–solid Process. J Cryst Growth 2003;247:357–62.

16] Li SY, Lin P, Lee CY, Tseng TY. Field Emission and Photofluorescent Char-acteristics of Zinc Oxide Nanowires Synthesized by a Metal CatalyzedVapor-Liquid-Solid Process. J Appl Phys 2004;95:3711–6.

17] Wang X, Summers CJ, Wang ZL. Large-Scale Hexagonal-Patterned Growth ofAligned ZnO Nanorods for Nano-Optoelectronics and Nanosensor Arrays. NanoLett 2004;4:423–6.

18] Yang P, Yan H, Mao S, Russo R, Johnson J, Saykally R, Morris N, Pham J, He R,Choi HJ. Controlled growth of ZnO nanowires and their optical properties. AdvFunct Mater 2002;12(5):323.

19] Xiang B, Wang P, Zhang X, Dayeh SA, Aplin DPR, Soci C, Yu D, Wang D. RationalSynthesis of P-Type Zinc Oxide Nanowire Arrays Using Simple Chemical VaporDeposition. Nano Lett 2007;7:323–8.

20] Chang P-C, Fan Z, Wang D, Tseng W-Y, Chiou W-A, Hong J, Lu JG. ZnO NanowiresSynthesized by Vapor Trapping CVD Method. Chem Mater 2004;16:5133–7.

21] Bae SY, Seo HW, Park J. Vertically Aligned Sulfur-Doped ZnO Nanowires Syn-thesized via Chemical Vapor Deposition. J Phys Chem B 2004;35.

22] Wang D, Dai H. Low-Temperature Synthesis of Single-Crystal GermaniumNanowires by Chemical Vapor Deposition. Angew Chem 2003;34.

23] Zhang BP, Binh NT, Segawa Y, Wakatsuki K, Usami N. Optical Properties ofZnO Rods Formed by Metalorganic Chemical Vapor Deposition. Appl Phys Lett2003;83:1635.

24] Kim S-W, Fujita S. ZnO Nanowires with High Aspect Ratios Grown by Met-alorganic Chemical Vapor Deposition Using Gold Nanoparticles. Appl Phys Lett2005;86:153119.

25] Kim K-S, Kim HW. Synthesis of ZnO Nanorod on Bare Si Substrate Using MetalOrganic Chemical Vapor Deposition. Phys B 2003;328:368–71.

26] Yang JL, An SJ, Park WI, Yi G-C, Choi W. Photocatalysis Using ZnO Thin Films andNanoneedles Grown by Metal-Organic Chemical Vapor Deposition. Adv Mater2004;16:1661–4.

27] Kong YC, Yu DP, Zhang B, Fang W, Feng SQ. Ultraviolet-Emitting ZnONanowires Synthesized by a Physical Vapor Deposition Approach. Appl PhysLett 2001;78:407–9.

28] Wang L, Zhang X, Zhao S, Zhou G, Zhou Y, Qi J. Synthesis of Well-Aligned ZnONanowires by Simple Physical Vapor Deposition on c-Oriented ZnO Thin Filmswithout Catalysts or Additives. Appl Phys Lett 2005;86:024108.

29] Lyu SC, Zhang Y, Lee CJ, Ruh H, Lee HJ. Low-Temperature Growth of ZnONanowire Array by a Simple Physical Vapor-Deposition Method. Chem Mater2003;15:3294–9.

30] Greene LE, Law M, Tan DH, Montano M, Goldberger J, Somorjai G, Yang P. Gen-eral Route to Vertical ZnO Nanowire Arrays Using Textured ZnO Seeds. NanoLett 2005;5:1231–6.

31] Heo YW, Varadarajan V, Kaufman M, Kim K, Norton DP, Ren F, Fleming PH.Site-Specific Growth of Zno Nanorods Using Catalysis-Driven Molecular-BeamEpitaxy. Appl Phys Lett 2002;81:3046.

32] Kong XY. Single-Crystal Nanorings Formed by Epitaxial Self-Coiling of PolarNanobelts. Science 2004;303:1348–51.

33] Park WI, Kim DH, Jung S-W, Yi G-C. Metalorganic Vapor-Phase Epitaxial Growthof Vertically Well-Aligned ZnO Nanorods. Appl Phys Lett 2002;80:4232.

34] Tien LC, Norton DP, Pearton SJ, Wang H-T, Ren F. Nucleation Control for ZnONanorods Grown by Catalyst-Driven Molecular Beam Epitaxy. Appl Surf Sci2007;253:4620–5.

35] Okada T, Agung BH, Nakata Y. ZnO Nano-Rods Synthesized by Nano-Particle-Assisted Pulsed-Laser Deposition. Appl Phys A 2004;79:1417–9.

36] Tien LC, Pearton SJ, Norton DP, Ren F. Synthesis and Microstructure of Verti-cally Aligned ZnO Nanowires Grown by High-Pressure-Assisted Pulsed-LaserDeposition. J Mater Sci 2008;43:6925–32.

37] Guo RQ, Nishimura J, Matsumoto M, Nakamura D, Okada T. Catalyst-Free Syn-thesis of Vertically-Aligned ZnO Nanowires by Nanoparticle-Assisted PulsedLaser Deposition. Appl Phys A 2008;93.

38] Morales AM. A Laser Ablation Method for the Synthesis of Crystalline Semicon-ductor Nanowires. Science 1998;279:208–11.

g Processes 32 (2018) 438–444

39] Park WI, Kim DH, Jung S-W, Yi G-C. Metalorganic Vapor-Phase Epitaxial Growthof Vertically Well-Aligned ZnO Nanorods. Appl Phys Lett 2002;80:4232.

40] Song T, Choung JW, Park J-G, Park WI, Rogers JA, Paik U. Surface Polarity andShape-Controlled Synthesis of ZnO Nanostructures on GaN Thin Films Based onCatalyst-Free Metalorganic Vapor Phase Epitaxy. Adv Mater 2008;20:4464–9.

41] Fan Z, Lu JG. Zinc Oxide Nanostructures: Synthesis and Properties. J NanosciNanotechnol 2005;5:1561–73.

42] Suh D-I, Byeon CC, Lee C-L. Synthesis and Optical Characterization of Verti-cally Grown ZnO Nanowires in High Crystallinity through Vapor–liquid–solidGrowth Mechanism. Appl Surf Sci 2010;257:1454–6.

43] Huang MH, Wu Y, Feick H, Tran N, Weber E, Yang P. Catalytic Growth of ZincOxide Nanowires by Vapor Transport. Adv Mater 2001;13:113–6.

44] Li Q, Kumar V, Li Y, Zhang H, Marks TJ, Chang RPH. Fabrication of ZnO Nanorodsand Nanotubes in Aqueous Solutions. Chem Mater 2005;17:1001–6.

45] Ko SH, Lee D, Kang HW, Nam KH, Yeo JY, Hong SJ, Grigoropoulos CP, Sung HJ.Nanoforest of Hydrothermally Grown Hierarchical ZnO Nanowires for a HighEfficiency Dye-Sensitized Solar Cell. Nano Lett 2011;11:666–71.

46] Sun Y, George Ndifor-Angwafor N, Jason Riley D, Ashfold MNR. Synthesis andPhotoluminescence of Ultra-Thin ZnO Nanowire/nanotube Arrays Formed byHydrothermal Growth. Chem Phys Lett 2006;431:352–7.

47] Wang JX, Sun XW, Yang Y, Huang H, Lee YC, Tan OK, Vayssieres L. Hydrother-mally Grown Oriented ZnO Nanorod Arrays for Gas Sensing Applications.Nanotechnology 2006;17:4995–8.

48] Wang J, Gao L. Hydrothermal Synthesis and Photoluminescence Properties ofZnO Nanowires. Solid State Commun 2004.

49] Qiu J, Li X, He W, Park S-J, Kim H-K, Hwang Y-H, Lee J-H, Kim Y-D. The GrowthMechanism and Optical Properties of Ultralong ZnO Nanorod Arrays with aHigh Aspect Ratio by a Preheating Hydrothermal Method. Nanotechnology2009;20:155603.

50] Baruah S, Mahmood MA, Myint MTZ, Bora T, Dutta J. Enhanced Visible LightPhotocatalysis through Fast Crystallization of Zinc Oxide Nanorods. Beilstein JNanotechnol 2010;1:14–20.

51] Lim SK, Hwang SH, Kim S. Microemulsion Synthesis and Characterization ofAluminum Doped ZnO Nanorods. Cryst Res Technol 2010;45:771–5.

52] Wu C, Shen L, Yu H, Huang Q, Zhang YC. Synthesis of Sn-Doped ZnO Nanorodsand Their Photocatalytic Properties. Mater Res Bull 2011;46:1107–12.

53] Athauda TJ, Ozer RR. Hydrothermal Growth of ZnO Nanostructures on NylonFabrics. MRS Proc 2012;1512.

54] Xu S, Adiga N, Ba S, Dasgupta T, Wu CFJ, Wang ZL. Optimizing and Improvingthe Growth Quality of ZnO Nanowire Arrays Guided by Statistical Design ofExperiments. ACS Nano 2009;3:1803–12.

55] Puretzky AA, Geohegan DB, Jesse S, Ivanov IN, Eres G. In situ measurementsand modeling of carbon nanotube array growth kinetics during chemical vapordeposition. Appl Phys A 2005;81(2):223–40.

56] Yousefi AT, Tanaka H, Bagheri S, Elfghi F, Mahmood MR, Ikeda S. Vectorial crys-tal growth of oriented vertically aligned carbon nanotubes using statisticalanalysis. Cryst Growth Des 2015;15(7):3457–63.

57] Oliver CR, Polsen ES, Meshot ER, Tawfick S, Park SJ, Bedewy M, Hart AJ. Statis-tical analysis of variation in laboratory growth of carbon nanotube forests andrecommendations for improved consistency. ACS Nano 2013;7(4):3565–80.

58] Wei C, Dong J. Direct Fabrication of High-Resolution Three-Dimensional Poly-meric Scaffolds Using Electrohydrodynamic Hot Jet Plotting. J MicromechMicroeng 2013;23:025017.

59] Wei C, Qin H, Ramírez-Iglesias NA, Chiu C-P, Lee Y, Dong J. High-ResolutionAc-Pulse Modulated Electrohydrodynamic Jet Printing on Highly InsulatingSubstrates. J Micromech Microeng 2014;24:045010.

60] Lee D-Y, Shin Y-S, Park S-E, Yu T-U, Hwang J. Electrohydrodynamic Printing

2012;12:2888–93.