Interfacial energetics of NaCl–organic composite layer at an OLED anode

This article has been downloaded from IOPscience. Please scroll down to see the full text article.

2012 J. Phys. D: Appl. Phys. 45 455304

(http://iopscience.iop.org/0022-3727/45/45/455304)

Download details:

IP Address: 121.50.21.21

The article was downloaded on 17/10/2012 at 03:57

Please note that terms and conditions apply.

View the table of contents for this issue, or go to the journal homepage for more

Home Search Collections Journals About Contact us My IOPscience

IOP PUBLISHING JOURNAL OF PHYSICS D: APPLIED PHYSICS

J. Phys. D: Appl. Phys. 45 (2012) 455304 (7pp) doi:10.1088/0022-3727/45/45/455304

Interfacial energetics of NaCl–organiccomposite layer at an OLED anodeJeongho Kim1,4,5, Yeonjin Yi2, Jeong Won Kim3,5, Seok Hwan Noh1 andHeon Kang4

1 LG Electronics, Seoul, 153-801, Republic of Korea2 Department of Physics, Yonsei University, Seoul 120-749, Republic of Korea3 Korea Research Institute of Standards and Science, Daejeon 305-340, Republic of Korea4 Department of Chemistry, Seoul National University, Seoul 151-747, Republic of Korea

E-mail: kim0011@snu.ac.kr (J Kim) and jeongwonk@kriss.re.kr (J W Kim)

Received 7 August 2012, in final form 11 September 2012Published 16 October 2012Online at stacks.iop.org/JPhysD/45/455304

AbstractAlthough low work function alkaline halides are widely used as a cathode interlayer fororganic light-emitting diodes (OLEDs), NaCl–organic composites are shown to be an efficientanodic buffer. Here we suggest a mechanistic origin of the improved OLED performance uponthe use of a NaCl-containing organic buffer layer between an indium tin oxide (ITO) anodeand N ,N ′-bis(naphthalene-1-yl)-N ,N ′-bis(phenyl)benzidine (NPB), based on the studies withultraviolet photoelectron spectroscopy and atomic force microscopy. While a pure NaClinterlayer has a high hole-injection barrier (1.40 eV), the NPB : NaCl composite layer exhibitsa substantially lower barrier (0.84 eV), which is comparable to the value at a bare ITO/NPBinterface. Furthermore, the wettability of the composite onto ITO is enhanced due tosignificant adhesive interactions of NaCl with both ITO and NPB, leading to effective electricalcontacts. The two key factors, i.e. the plausible hole-injection barrier and better wettability ofthe NPB : NaCl composite, contribute to the improved hole injection efficiency and lifetime.

(Some figures may appear in colour only in the online journal)

1. Introduction

Interface engineering between electrodes and organic layersin organic light-emitting diodes (OLEDs) has been one ofthe main technological approaches because charge injectionand device stability are strongly dependent on their interfaceformation [1, 2]. This is why much effort has been madeto improve the interface properties and understand themechanism governing the charge injection and transport. Ingeneral, to reduce charge injection barriers, the anodic sideconsists of high work function materials and the cathodicside of low work function materials. In addition, surface orinterface free energy is another important factor to be countedin terms of interfacial compatibility between inorganic andorganic materials.

Various approaches have been tried to modify both thework function and surface energy of indium tin oxide (ITO), themost common transparent anode. Physical surface treatments

5 Authors to whom any correspondence should be addressed.

such as oxygen plasma and ultraviolet ozone (UVO) have beenwidely used to enhance hole injection [3, 4]. These treatmentsare effective in removing residual surface contaminants andinducing oxygen-rich surface at ITO, resulting in an increasein the work function via surface band bending of ITO orinterfacial dipole [5, 6]. However, the surface treatments arenot enough to obtain interfacial compatibility between the ITOand organic hole transport layer (HTL) because the plasmaor UVO treatment makes the ITO surface more hydrophilic[7, 8], possibly exhibiting adverse interfacial properties tomost organic HTL materials with high hydrophobicity [9].Another or an additional important approach for modifyingthe anodic interfacial region is to introduce metals or metaloxides with high work function [10–15], organic materials suchas a star-burst amine [16] and fluorocarbon plasma polymers[17–19], and even insulators [20–23] between the ITO andorganic HTL. Hole injection enhancement has variously beenascribed to mechanisms such as energy level realignmentvia interface dipoles or surface band bending [14, 24–26],

0022-3727/12/455304+07$33.00 1 © 2012 IOP Publishing Ltd Printed in the UK & the USA

J. Phys. D: Appl. Phys. 45 (2012) 455304 J Kim et al

interfacial stability induced by mechanical adhesion [10, 21–23], or tunnelling probability due to a reduced effectivebarrier [20].

On the other hand, metal halides which are usuallyapplied at cathode/organic interfaces, have also been triedfor better anode/organic interface properties [27, 28]. Theapplication of metal halides on the anodic side can allowsimpler fabrication of highly efficient OLEDs by introducingthe same metal halide for both electron and hole injectionenhancement. Furthermore, metal halides are usually cheapand easy to handle, compared with other organic or metaloxide interlayer materials such as V2O5 [14], MoO3 [15] andbuckminsterfullerene [29, 30]. LiF, the most common electroninjection material on the cathode surface, was demonstratedto be effective in hole injection enhancement on hydrogenplasma treated ITO anodes but not on the ones with morepractical treatments such as oxygen plasma or UVO, based onthe tunnelling model [31]. A composite form of metal halidessuch as LiF and MgF2 with hole transport materials improvesthe thermal stability of OLEDs [28, 32]. Furthermore, wehave recently shown that a NaCl-incorporated interlayer at theITO and HTL interface improves not only the anodic interfacestability but also the hole injection efficiency with implicationof multiple mechanistic origins for improved performance[33]. However, the reasons for hole injection and devicelifespan enhancement are not clear yet despite the practicaladvantages of metal halides as an anodic buffer.

To tackle the underlying mechanism of OLED perfor-mance enhancement using NaCl composites, the interfacialelectronic and thermodynamic properties are studied. Here, weprepared ITO/NaCl-containing buffer/N ,N ′-bis(naphthalene-1-yl)-N ,N ′-bis(phenyl)benzidine (NPB) interfaces, and mea-sured ultraviolet photoelectron spectroscopy (UPS) and thewettability of the thin films by morphological study withatomic force microscopy (AFM). First, the hole injection bar-rier at the interface of the ITO and NPB : NaCl composite layerwas comparable to that at the ITO/NPB interface, showing thatNPB at the interface is kept both chemically and electronicallyintact even in the presence of NaCl. Second, the wettability ofNPB in the composite onto ITO was enhanced by the significantadhesive interactions of NaCl with both ITO and NPB, leadingto effective electrical contacts across the anode side. This wassupported by the calculation of interface free energies, workof adhesions and spreading coefficients. These studies revealthe importance of concomitant consideration of both electroniccharge injection barrier and morphological adhesion propertyfor a rational design of organic/inorganic interfaces in organicelectronic devices.

2. Experiment

Both the growth of thin solid layers and the UPS measurementswere performed in a ultrahigh vacuum system (base pressureof 10−9 Torr) composed of three interconnected chambersequipped for surface pretreatment, film growth and analysis.An ITO-coated glass substrate with a sheet resistance of50 � cm−2 was used and UVO treatment was employedto clean the substrate in the surface pretreatment chamber.

Figure 1. UPS spectra collected near the Fermi level and secondarycutoff region during the step-by-step layer deposition of NPB onITO.

NPB and NaCl films were deposited onto the ITO substrateby resistive heating in the film growth chamber. Thedeposition rate was 0.01 nm s−1 for both NPB and NaCllayers. Mixed films of NPB : NaCl (1 : 3 wt%) were preparedby coevaporation of NaCl and NPB. UPS spectra acquisitionwas performed in the analysis chamber using the unfilteredHe I (21.2 eV) radiation lines of a gas discharge lamp and ahemispherical electron energy analyser (VG ESCALAB 220isystem). At each step of interface formation, the valenceband and the onset of photoemission were recorded. Theonset or secondary electron cutoff, which represented thevacuum level, was measured with a bias of −10 V on thesample to clear the vacuum level of the detector. Vacuumlevel shifts and HOMO level changes induced by the NPBor NPB : NaCl composite overlayers were thus measured.Surface morphology measurements were performed usingAFM (dimension 3100).

3. Results and discussion

The evolutions of secondary electron cutoff and valence regionspectra for NPB/ITO with increasing coverage are shownin figure 1. The shape of the UPS spectrum agrees wellwith the previous report along the measured whole energyrange, showing the characteristic features of the highestoccupied molecular orbital (HOMO) corresponding to thecentral biphenyl unit of NPB [3, 34]. All spectra wereplotted with respect to the Fermi level (EF). The normalizedsecondary cutoff edges of NPB are shown on the left panel offigure 1. The vacuum levels of the samples were determinedby linear extrapolation of secondary electron cutoffs on thehigh-binding-energy side of the UPS spectra (15–19 eV). The

2

J. Phys. D: Appl. Phys. 45 (2012) 455304 J Kim et al

vacuum level shifts abruptly towards a higher binding energyfor the first sub-monolayer deposition (nominal thickness of0.2 nm) by about 0.2 eV, thus indicating the presence of aninterface dipole as previously reported [35, 36]. As the NPBlayer becomes thicker after the second step of deposition, thevacuum levels shift gradually towards higher binding energies.The final secondary cutoff edge is assigned to the 4 nmthickness NPB spectrum. The right panel of figure 1 showsHOMO onsets for NPB. The HOMO onset was determinedvia a linear extrapolation of the leading edge (closer to theFermi level) in the spectrum and is found at 0.84 eV for NPB(0.2 nm)/ITO below EF, followed by a gradual higher bindingenergy shift with a further increase in NPB thickness.

To investigate the electronic effect of the NaCl-containinginterlayer between ITO and NPB, NaCl was deposited onthe ITO surface in the forms of both pure NaCl layers andNPB : NaCl mixed layers. Figure 2(a) shows the secondaryelectron cutoffs and HOMO regions of UPS spectra for thesamples containing a pure NaCl interlayer between NPB andITO. The secondary cutoff shifts abruptly towards higherbinding energies at low coverages under 0.5 nm thicknessby 0.26 eV, and then saturates at a thickness of 1.0 nm. Assoon as the NPB is deposited on the pre-deposited NaClsurface, the secondary cutoff shifts abruptly again in the samedirection by about 0.3 eV, and then moves gradually towardshigher binding energies with further deposition of NPB. Thus,interface dipoles are expected to be present on both sides ofthe NaCl interlayer between NPB and ITO. In the right panelof figure 2(a), the NPB HOMO features are clearly observedon the NaCl surface, showing that the HOMO onset at thelow coverage of 0.2 nm thickness is 1.40 eV and then shiftsslowly towards higher binding energies with the increase inNPB coverage. Features above the HOMO level or in theenergy gap are replica of the valence state due to the weak He I

β-line from the unfiltered He I discharge lamp. Apart fromthese, no new state appears at the NPB/NaCl interface.

Unlike the pure NaCl interlayer, the mixed film type ofNPB : NaCl shows a relatively pronounced gradual shift ofthe secondary cutoff towards higher binding energies duringthe deposition of the composite by 0.59 eV rather than suddenshift(s), as shown in the left part of figure 2(b). The HOMOfeatures of NPB in the NPB : NaCl mixed phase are clearlyobserved even though the intensities are relatively weak atlower coverages of the composite, shown in the right panelof figure 2(b). The shapes and widths of the NPB HOMOfeatures in the mixed phase are found to be similar to thoseof the HOMO features of pure NPB overlayer in figure 2(a),implying that NPB in the mixed phase with NaCl is keptchemically intact. Even in the composite layer, NPB and NaCldo not induce any new UPS feature and maintain their originalchemical properties. Although the HOMO cutoff positionof NPB : NaCl with 0.2 nm thickness is hard to be assignedbecause of the weak signal from NPB, we could extract theposition to be 0.84 eV by considering the lower binding energyshift of the observed HOMO maximum by 0.16 eV from thatof NPB : NaCl with 0.5 nm thickness, as marked in the rightpanel of figure 2(b).

Work functions or vacuum levels, and HOMO onsetswith different NPB coverages were extracted and summarized

Figure 2. UPS spectra collected near the Fermi level and secondarycutoff region during the step-by-step layer deposition of (a) NaCland (b) NPB : NaCl, followed by NPB, on ITO.

by the schematic energy level diagrams for NPB/ITO,NPB/NaCl/ITO, and NPB/NPB : NaCl/ITO interfaces, asshown in figure 3. With respect to the Fermi level, thechange in the secondary cutoff corresponds directly to the workfunction or vacuum level change. For all the samples, bothsurface work functions and HOMO onsets shift downwardswith the increase in NPB or NPB : NaCl thickness. Thus,the polarization effect on the inorganic substrate (ITO) couldbe considered to be negligible. Otherwise, a difference inthe HOMO and vacuum level shifts from the interface to thefilm surface would be observed by the substrate polarization

3

J. Phys. D: Appl. Phys. 45 (2012) 455304 J Kim et al

Figure 3. Energy level diagram of (a) NPB/ITO, (b) NPB/NaCl/ITO and (c) NPB/NPB : NaCl/ITO. Evac, EF, eD and Φh correspond to thevacuum level, Fermi level, interface dipole and hole injection barrier, respectively.

effect upon photoionization [37]. Thus, the overall downwardshifts are attributed mainly to the space charge redistribution[36, 38], compensating for the interface dipole formation. Thepure NaCl interlayer raises the hole injection barrier fromITO to NPB by 0.56 eV compared with the direct contactof ITO/NPB. The specific mechanisms related to the chargeinjection efficiency might be extremely complicated, but it isclear that the higher hole injection barrier contributes to thehigher turn-on voltage of OLEDs with a pure NaCl thin filmbuffer in the previous report [27]. Meanwhile, the injectionbarrier of NPB : NaCl on ITO is comparable to that of pureNPB on ITO, implying that the NPB maintains its intrinsicinjection property even in the composite phase on the ITOsurface. Furthermore, the pure NaCl interlayer between ITOand NPB shows two stepwise decreases in the vacuum levelby the interface dipoles formed on both sides of the interlayer,whereas the NPB : NaCl composite layer shows a gradualdecrease throughout the mixed phase. The total vacuum levelchange from ITO to the NPB : NaCl/NPB interface is 0.59 eVin figure 3(c), which is similar to that of 0.60 eV from ITOto the NaCl/NPB interface in figure 3(b). Thus, the amountof vacuum-level lowering by NaCl observed in the pure NaClinterlayer between NPB and ITO via interface dipoles is likelyexpressed as a large band bending in the mixed phase of theNPB : NaCl composite interlayer, even though the origin ofthe large band bending is not yet clearly revealed. Due tothe large band bending in the mixed phase, the HOMO levelis relatively far from the Fermi level at a thickness of 4 nm(1.28 eV) compared with that of pure NPB at the same nominalthickness (1.09 eV).

Based on the above energy level alignment diagrams, theschematic energy level changes for pure NPB and NPB : NaClcomposite layers on ITO under a bias condition are comparedin figures 4(a) and (b), respectively. When a forward biasis applied, the NPB : NaCl layer creates a larger electric fielddue to its large resistivity induced by the insulating propertyof NaCl compared with pure NPB. In fact, the thickness andcomposition of the composite interlayer could be optimizedfor the best device performance. As the layer thickness goesabove several nm, the current density of the device rapidlybecomes worse [33]. Thus, a larger energy level gradient in

Figure 4. Schematic band diagram: (a) without and (b) withNPB : NaCl buffer layer. Energy level changes with applied bias areindicated. The shaded area denotes the approximate tunnellingbarrier under the bias voltage.

the mixed buffer layer reduces the tunnelling barrier, whichis approximately displayed by the shaded area in figure 4(b).Compared with the pure NPB layer in figure 4(a), this reducedeffective tunnelling barrier may lead to an easier hole injectioninto the HTL through the NPB : NaCl buffer layer. Therefore,the hole injection enhancement could be partially attributed tothe tunnelling mechanism [1, 39, 40].

Since the wettability of the charge injection or transportlayer onto the electrode is another factor determining thecharge injection efficiency, the surface morphologies of thepure NPB and NPB : NaCl composite (1 : 7 wt%) films werecompared by AFM images. In figures 5(a) and (b), the pureNPB film forms large aggregates showing numerous dewettingspots on the revealed ITO grain pattern. The root mean square

4

J. Phys. D: Appl. Phys. 45 (2012) 455304 J Kim et al

Figure 5. AFM images (2 × 2 µm2) of 4 nm thickness of (a) NPB and (b) NPB : NaCl, and 8 nm thickness of (c) NPB and (d) NPB : NaCl.The thin films were deposited onto the oxygen plasma pretreated ITO surface. The root mean square roughness values are 5.13 nm, 4.17 nm,6.06 nm and 6.86 nm for (a), (b), (c) and (d), respectively.

(rms) roughness value is 5.13 nm at the nominal thicknessof 4 nm. However, the NPB : NaCl composite film does notshow NPB aggregates with a lower rms roughness of 4.17 nm,implying that the mixed film is more wettable on the ITOsurface. With the increase in the film thickness, the surfaceroughness of the composite film increases more rapidly andsurpasses that of pure NPB. The rms roughness values at 8 nmthickness for pure NPB and NPB : NaCl composite films are6.06 nm and 6.86 nm, respectively, as shown in figures 5(c)and (d). Therefore, NaCl significantly affects the film growthmode in the NaCl : NPB mixed film. The film wetting andgrowth behaviour can be described by work of adhesion andsurface free energy at each interface. The interfacial freeenergy (γ ), work of adhesion (WA) and spreading coefficient(S) were calculated, according to equations (1), (2) and (3),respectively [41, 42]:

γA : B = (γ1/2A − γ

1/2B )2 (1)

WAA : B = γA + γB − γA : B (2)

SA : B = γB − γA − γA : B (3)

Table 1. Calculated values of interfacial free energy (γA : B), work ofadhesion (W A

A : B) and spreading coefficient (SA : B) in mJ m−2. Eachsurface free energy of NPB, NaCl and ITO is referenced from theliterature, 31.1 mJ m−2, 227 mJ m−2 and 70.1 mJ m−2, respectively[43–45].

A : B γA : B W AA : B SA : B

NPB : ITO 7.82 93.4 31.2NaCl : ITO 44.9 252 −202NPB : NaCl 90.1 168 106

where γA : B is the interfacial free energy of the A/B interface,γA is the surface free energy of A, WA

A : B is the workof adhesion between A and B, and SA : B is the spreadingcoefficient of A on surface B. The surface free energies ofNPB, NaCl and ITO were referenced from the reported values,31.1 mJ m−2, 227 mJ m−2 and 70.1 mJ m−2, respectively [43–45]. Table 1 exhibits the summarized values of γA : B,WA

A : B and SA : B where A : B is NPB : ITO, NaCl : ITO orNPB : NaCl. Although the interface free energy of NPB : ITOis less than that of NaCl : ITO, NaCl shows a higher work

5

J. Phys. D: Appl. Phys. 45 (2012) 455304 J Kim et al

of adhesion (252 mJ m−2) onto the ITO surface than that ofNPB (93.4 mJ m−2) due to the high surface free energy ofNaCl, but its spreading coefficient is less and even negative(−202 mJ m−2). This means that NaCl can strongly adhereto the ITO surface but is not so wettable on it—reminiscentof Volmer–Weber (V–W) film growth mode in which three-dimensional crystallites nucleate immediately upon contact[46]. As indeed shown in the above AFM results, NaClin the NPB : NaCl mixed film makes the composite filmgrowth mode close to the V–W type, preventing the NPBaggregation observed in pure NPB films. Furthermore, NPBis expected to be a good adhesive to NaCl from the highWA of NPB : NaCl (168 mJ m/2), as additionally confirmed byAFM images showing well-intermixed uniform morphologiesin figures 5(b) and (d). NaCl, thus anchoring to the ITO surfacewith its relatively high WA, seems to facilitate the contactbetween ITO electrode and NPB by suppressing the cohesiveinteractions between NPB molecules and simultaneouslyinducing moderate adhesive interactions between NPB andNaCl. Therefore, such a morphological effect caused byNaCl incorporation into the NPB film contributes to theprevious report of hole injection enhancement by bettermechanical adhesion, leading to ‘effective’ electrical contactsand prolonged lifetime.

4. Conclusions

We investigated the role of a NaCl thin film and a NPB : NaClcomposite layer as anodic buffer between an ITO anode anda NPB transport layer using UPS and AFM measurements.The hole injection barrier is increased with the insertion ofthe pure NaCl layer compared with the interface without abuffer layer. The NPB : NaCl composite, however, showsa nearly identical injection barrier to that of the NPB/ITOinterface. Furthermore, NaCl in the NPB : NaCl mixed filmassists the mechanical adhesion and the resultant electricalcontact between ITO and NPB via better wettability ofthe mixed film on the ITO surface. As a whole thetunnelling through the highly resistive composite layer witha comparable injection barrier height and enhanced electricalcontact are associated with the improved performance of theOLEDs using a NPB : NaCl composite anodic buffer. Toachieve high-end OLEDs, interface control in terms of bothelectronic charge injection barrier and mechanical adhesion isnecessary.

Acknowledgments

This work was supported by Brain Korea 21 (BK21) projectof the Ministry of Education, Science and Technology, byresearch projects of the National Research Foundation of Korea(Grants 2011-0026100 and 2012-0002303), and by the KoreaResearch Council of Fundamental Science and Technology(KRCF) through Basic Research Project managed by the KoreaResearch Institute of Standards and Science (KRISS).

References

[1] Parker I D 1994 J. Appl. Phys. 75 1656–66[2] Malliaras G G and Scott J C 1998 J. Appl. Phys. 83 5399–403[3] Ding X M, Hung L M, Cheng L F, Deng Z B, Hou X Y,

Lee C S and Lee S T 2000 Appl. Phys. Lett. 76 2704–6[4] Sugiyama K, Ishii H, Ouchi Y and Seki K 2000 J. Appl. Phys.

87 295–8[5] Milliron D J, Hill I G, Shen C, Kahn A and Schwartz J 2000

J. Appl. Phys. 87 572–6[6] Kim S Y, Lee J-L, Kim K-B and Tak Y-H 2004 J. Appl. Phys.

95 2560–3[7] Huang Z H, Zeng X T, Sun X Y, Kang E T, Fuh J Y H and

Lu L 2008 Org. Electron. 9 51–62[8] Kim J S, Friend R H and Cacialli F 1999 J. Appl. Phys.

86 2774–8[9] Cui J, Huang Q, Veinot J C G, Yan H, Wang Q,

Hutchison G R, Richter A G, Evmenenko G, Dutta P andMarks T J 2002 Langmuir 18 9958–70

[10] Shen Y, Jacobs D B, Malliaras G G, Koley G, Spencer M Gand Ioannidis A 2001 Adv. Mater. 13 1234–8

[11] Chan I M and Hong F C 2004 Thin Solid Films450 304–11

[12] Moon J-M, Bae J-H, Jeong J-A, Jeong S-W, Park N-J,Kim H-K, Kang J-W, Kim J-J and Yi M-S 2007 Appl. Phys.Lett. 90 163516

[13] Hu W, Manabe K, Furukawa T and Matsumura M 2002 Appl.Phys. Lett. 80 2640–1

[14] Zhang H M and Choy W C H 2008 IEEE Trans. ElectronDevices 55 2517–20

[15] You H, Dai Y, Zhang Z and Ma D 2007 J. Appl. Phys.101 026105

[16] Chen S-F and Wang C-W 2004 Appl. Phys. Lett.85 765–7

[17] Tong S W, Lee C S, Lifshitz Y, Gao D Q and Lee S T 2004Appl. Phys. Lett. 84 4032–4

[18] Hung L S, Zheng L R and Mason M G 2001 Appl. Phys. Lett.78 673–5

[19] Hsiao C-C, Chang C-H, Lu H-H and Chen S-A 2007 Org.Electron. 8 343–8

[20] Qiu C, Xie Z, Chen H, Wong M and Kwok H S 2003 J. Appl.Phys. 93 3253–8

[21] Poon C O, Wong F L, Tong S W, Zhang R Q, Lee C S andLee S T 2003 Appl. Phys. Lett. 83 1038–40

[22] Deng Z B, Ding X M, Lee S T and Gambling W A 1999 Appl.Phys. Lett. 74 2227–9

[23] Gyoutoku A, Hara S, Komatsu T, Shirinashihama M,Iwanaga H and Sakanoue K 1997 Synth. Met. 91 73–5

[24] Kim S Y, Hong K and Lee J-L 2007 Appl. Phys. Lett.90 183508

[25] Lee H, Cho S W, Han K, Jeon P E, Whang C-N, Jeong K,Cho K and Yi Y 2008 Appl. Phys. Lett. 93 043308

[26] Tong S W, Lau K M, Sun H Y, Fung M K, Lee C S, Lifshitz Yand Lee S T 2006 Appl. Surf. Sci. 252 3806–11

[27] Shi S W, Ma D G and Peng J B 2007 Eur. Phys. J. Appl. Phys.40 141–4

[28] Tokito S and Taga Y 1995 Appl. Phys. Lett. 66 673–5[29] Yuan Y Y, Han S, Grozea D and Lu Z H 2006 Appl. Phys. Lett.

88 093503[30] Yuan Y, Grozea D and Lu Z H 2005 Appl. Phys. Lett.

86 143509[31] Zhao J M, Zhang S T, Wang X J, Zhan Y Q, Wang X Z,

Zhong G Y, Wang Z J, Ding X M, Huang W and Hou X Y2004 Appl. Phys. Lett. 84 2913–5

[32] Grozea D, Turak A, Yuan Y, Han S, Lu Z H and Kim W Y2007 J. Appl. Phys. 101 033522

[33] Kim J, Kim M, Kim J W, Yi Y and Kang H 2010 J. Appl. Phys.108 103703

6

J. Phys. D: Appl. Phys. 45 (2012) 455304 J Kim et al

[34] Zhang R Q, Lee C S and Lee S T 2000 J. Chem. Phys.112 8614–20

[35] He P, Wang S D, Wong W K, Lee C S and Lee S T 2001 Appl.Phys. Lett. 79 1561–3

[36] Ishii H, Sugiyama K, Ito E and Seki K 1999 Adv. Mater.11 605–25

[37] Gao W and Kahn A 2003 J. Appl. Phys. 94 359–66[38] Yan L, Mason M G, Tang C W and Gao Y 2001 Appl. Surf.

Sci. 175–176 412–8[39] Kim Y-E, Park H and Kim J-J 1996 Appl. Phys. Lett.

69 599–601[40] Zhang S T et al 2004 Appl. Phys. Lett. 84 425-7

[41] Murakami T, Kuroda S-i and Osawa Z 1998 J. ColloidInterface Sci. 202 37–44

[42] Adamson A W 1990 Physical Chemistry of Surfaces(New York: Wiley-Interscience)

[43] Kim J K, Park J W, Yang H, Choi M, Choi J H and Suh K Y2006 Nanotechnology 17 940

[44] Somorjai G A 1994 Introduction to Surface Chemistry andCatalysis (New York: Wiley-Interscience)

[45] You Z Z and Dong J Y 2006 J. Colloid Interface Sci.300 697–703

[46] Zangwill A 1988 Physics at Surfaces (Cambridge: CambridgeUniversity Press)

7