Patterning of self-assembled monolayers based on differences in molecular conductance

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

2009 Nanotechnology 20 245306

(http://iopscience.iop.org/0957-4484/20/24/245306)

Download details:

IP Address: 138.251.105.135

The article was downloaded on 23/06/2009 at 11:49

Please note that terms and conditions apply.

The Table of Contents and more related content is available

HOME | SEARCH | PACS & MSC | JOURNALS | ABOUT | CONTACT US

IOP PUBLISHING NANOTECHNOLOGY

Nanotechnology 20 (2009) 245306 (6pp) doi:10.1088/0957-4484/20/24/245306

Patterning of self-assembled monolayersbased on differences in molecularconductanceCai Shen and Manfred Buck

EaStCHEM School of Chemistry, University of St Andrews, St Andrews KY16 9ST, UK

E-mail: mb45@st-andrews.ac.uk

Received 25 February 2009, in final form 13 April 2009Published 26 May 2009Online at stacks.iop.org/Nano/20/245306

AbstractScanning tunneling microscopy (STM) is used for replacement patterning of self-assembledmonolayers (SAMs) of thiols on a sub-10 nm scale. Contrasting other schemes of scanningprobe patterning of SAMs, the exchange of molecules relies on differences in conductance and,thus, occurs under tunneling conditions where the resolution of the tip is maintained. Exchangetakes place at the boundary between different thiols but only when the tip moves from areas oflower to higher conductance. In combination with SAMs which exhibit excellent structuralquality, patterns with a contour definition of ±1 molecule, lines as thin as 2.5 nm and islandswith an area of less than 20 nm2 are straightforwardly produced. It is suggested that the shearforce exerted onto the molecules with the lower conductance triggers displacement of the onewith higher conductance.

(Some figures in this article are in colour only in the electronic version)

1. Introduction

With self-assembled monolayers (SAMs) constituting anessential element of nanotechnology [1–4], their patterningon an ever-shrinking length scale is a topic crucial foraddressing both fundamental aspects in nanoscience andtechnological applications. To access the sub-100 nmrange a number of patterning techniques based on electrons,photons, mechanical forces or electrochemical processeshave been developed such as electron beam lithography(EBL) [5, 6], extreme-UV interference lithography (EUV-IL) [7] and scanning probe microscopies (SPM), with thelatter comprising near-field optical microscopy (SNOM) [8],scanning tunneling microscopy (STM) [9–19] and atomicforce microscopy (AFM) [20–25]. Dip pen nanolithography(DPN) [26–28] as another AFM-derived technique whichis not based on replacement like nanografting [20, 21] butuses localized transfer of material also allows rather routineaccess to feature sizes below 100 nm, even in a parallelizedfashion [29, 30].

Regardless of the techniques applied to generate patternedSAMs, values of typically 10–20 nm, i.e. still significantlylarger than molecular dimensions, have been reported indemonstrations of resolution. Furthermore, a closer look

at structures reveals that feature dimensions are not exactlydefined but exhibit significant variations up to severalnanometers, a problem making reproducible and accuratestudies on the lower end of the nanometer scale difficult.To tackle this challenge it is important to address not onlythe patterning technique but also consider the relationshipbetween the structure of a SAM and its behavior towardspatterning. With regard to the former, STM as the techniquewith the highest resolution seems an obvious choice forthe generation of ultrasmall patterns and it has indeedbeen exploited for nanoscale patterning by either applyingsufficiently high tunneling bias, typically of the order ofseveral volts [11, 17, 18, 31], or high current setpointwhich moves the tip close to the surface [19, 31, 32].In the case of voltage-induced modification electrochemicalprocesses have been identified to be important in a liquidenvironment [10, 12, 18] whereas in vacuum/air field emissionof electrons and/or current-induced bond breaking have beensuggested as mechanisms [17, 21]. In the case of hightunneling current the tip penetrates the SAM, thus resultingin the mechanical displacement of molecules similar tonanoshaving/nanografting by AFM [20, 21]. While SAMmodification by STM was demonstrated already quite sometime ago [11, 16, 31] this approach has, however, been

0957-4484/09/245306+06$30.00 © 2009 IOP Publishing Ltd Printed in the UK1

Nanotechnology 20 (2009) 245306 C Shen and M Buck

relatively limited [9–17] as it is plagued by a lack ofreproducibility due to the required strong interactions betweenSTM tip and SAM which easily result in tip modificationand quick deterioration of resolution. Therefore, STM doesnot seem to offer any advantage over AFM techniques withtheir somewhat lower resolution but the use of a more robusttip. As far as the system is concerned, there has been nodetailed investigations of how structure and energetics of aSAM enter into the patterning process. This, however, can beexpected to be important since size distribution and orientationof domains, inhomogeneities in molecular packing and, thus,intermolecular and substrate–molecule interactions will affectthe precision to which a SAM can be modified on the envisagedlength scale.

In this paper we report a patterning mechanism whichdiffers from previously described ones and demonstrate that,through a properly designed SAM, patterning by STM canbe performed reliably and at high accuracy. Contrastingexisting grafting schemes, the approach taken here exploits thedifference in tunneling resistance between matrix and guestmolecules. As a consequence, conditions under which thematrix SAM is modified are more gentle compared to voltage-or current-induced patterning. Major changes of the STMtip are, thus, avoided and molecular resolution is maintainedthroughout the experiment.

2. Experimental details

Octadecanethiol (95%, Fluka), dodecanethiol (98%, Fluka),hexadecane (99%, Aldrich) and ethanol (AnalR, BDH) wereused as received. 2-(4′-methyl-biphenyl-4-yl)-ethanethiol(BP2) was synthesized as described previously [33]. Micasubstrates with an epitaxial Au(111) layer 300 nm thick werepurchased from Georg Albert PVD, Heidelberg, Germany.Substrates were flame annealed before immersion intosolutions of 1 mM BP2 in ethanol at 345 K overnight (15 h).After immersion, samples were rinsed with pure ethanol andblown dry with nitrogen. The thus-obtained low temperaturephase of BP2 was converted into the δ phase (vide infra)by annealing [34] of the sample in a sealed, nitrogen-purgedcontainer (418 K for 10 h).

For imaging a PicoPlus microscope (molecular imaging)was used. Tips were mechanically cut from a 0.25 mm Pt/Irwire (80:20, Goodfellow). STM images were recorded inconstant current mode and are all presented as acquired. STM-controlled replacement experiments were carried out at roomtemperature in a home-built Teflon cell using a 10 mM solutionof an alkane thiol (CH3(CH2)n−1SH, MCn) in hexadecane.Replacement requires an initial defect in the BP2 for whicheither a random defect already present in the BP2 SAM or anSTM-generated defect was used. The former can be controlledthrough sample preparation, e.g. cleanliness of the substrate.The latter was created by gentle pulsing (2.5–3.0 V for 50 ms)of the STM tip.

3. Results

The matrix SAM consists of 2-(4′-methyl-biphenyl-4-yl)-ethane thiol (CH3C6H4–C6H4–(CH2)2SH, BP2) which be-

longs to a class of molecules whose characteristics is the com-bination of a rigid aromatic moiety with an aliphatic spaceras described in detail previously [33–39]. The length of thespacer was chosen according to the concept of competitivedesign [38, 39] which has been shown to yield polymorphicSAMs with phases of very high structural quality [34, 38, 39].For the present experiments the δ phase of BP2 SAMs was usedwhich exhibits a rectangular (2

√3 × 2) unit cell [34]. Since a

defect-free δ phase of the BP2 SAMs is stable against spon-taneous exchange not only by mere exposure to a thiol solu-tion [34] but also upon scanning in a thiol-containing solution,an initial small defect is required for STM-induced patterningto occur, as will become clear later. Usually such a defect wascreated by gentle pulsing of the STM tip but also intrinsic de-fects [40], which are occasionally present on the sample de-pending on the pretreatment of the substrate, could be used.

The process of STM-induced replacement is illustratedin figure 1(a) where replacement of BP2 by octadecane thiol(MC18) is depicted in a sequence of STM images usinga single set of tunneling parameters, i.e. imaging andmodification are done simultaneously. Starting from twocircular defects, repeated scanning of the same area results inan exchange seen as an increase in patch size. Three points arenoteworthy here. Firstly, in pronounced contrast to the usualmodification parameters, the present ones (0.5 V/50 pA) arerather gentle and the tip does not penetrate the BP2 SAM [41].Furthermore, it does not depend on bias sign. Secondly,the presence of the thiol in the solvent is required since inpure hexadecane no alteration of the BP2 SAM is seen evenfor extended scanning under tunneling conditions where thetip is significantly closer to the substrate. Thirdly, the wayreplacement progresses is also quite different from normalnanografting. Instead of exchange occurring all along a scanline, displacement occurs by translation of the BP2/alkane thiolboundary. Evidence that the growing patches consist of thealkane thiol is provided in figure 1(b) which shows molecularlyresolved images of an MC12 patch surrounded by the originalBP2 SAM. The patch shows domains of well-ordered MC12molecules with, as revealed by the Fourier transform (upperleft inset), a hexagonal packing of the alkane thiol moleculeswhich contrasts with the rectangular geometry for BP2. Thehighly crystalline arrangement of the alkane thiols within a fewminutes of the experiment is in agreement with earlier reportsthat the kinetics of SAM formation proceeds significantlyfaster in confined geometry than in an unrestricted one [42, 43].With respect to the tunneling contrast we note that in theconstant current images areas of alkane thiols appear deeper,irrespective of the length of the molecules which was variedfrom MC6 to MC18, i.e. compared to BP2 from geometricallyshorter to longer molecules. This is caused by the differencein tunneling resistance which is higher for the saturatedhydrocarbon chains of alkane thiols compared to BP2 bearingaromatic moieties.

This gentle modification protocol can be exploited forSAM patterning at very high resolution and with goodreproducibility as illustrated by figure 2. Features such aslines and islands with sub-5 nm width can be easily written.Equally important as the average achievable feature size is

2

Nanotechnology 20 (2009) 245306 C Shen and M Buck

Figure 1. (a) Sequence of STM images showing displacement ofBP2 by MC18. Starting from two initial defects (I) the darkerappearing areas of MC18 expand upon continuous scanning (II–IV).(b) Molecularly resolved image of an MC12 patch produced byreplacement of BP2. Insets show Fourier transform of areas of BP2(lower right) and MC12 (upper left). Imaging parameters are0.5V/50 pA (a) and 0.6 V/30 pA (b).

the precision at which contours can be defined. From thehigh resolution images (figures 2(a)–(d)) one can see that the

Figure 2. (a) Left: STM image of a 3 nm line of MC12 patternedonto a SAM of BP2 in MC12/hexadecane solution. Line wasgenerated by the tip scanning only along the 〈112〉 direction for 30 swith a scan speed of 11.4 lines s−1. Parameters of replacement: 6 pAand 1.2 V. Shown height profile is marked by a solid line. Right:cartoon illustrating the structure of BP2 at the boundary betweenBP2 and MC12. (b) Series of lines about 5 nm wide, each generatedby scanning at a scanning speed of 5.7 lines s−1 for 1 s using 0.1 nAand 0.1 V. ((c), (d)) STM images of BP2 islands surrounded byMC12. Line about 3 nm wide (c) and island of 20 nm2 (d).Parameters used for patterning are 0.05 nA, 0.5 V (c) and 0.1 nA, 0.1V (d). (e) Line–dot line patterns with gaps of about 2 nm. Parametersused for patterning are 0.05 nA and 0.5 V.

boundaries between areas of BP2 and MC12 are very welldefined within a precision of ±1 molecule. A very interestingfeature which is particularly obvious in figures 2(a) and (c) isthe BP2/MC12 boundary along the 〈112〉 direction. It exhibitsa corrugation which nicely reflects the structure of the BP2SAM with its (2

√3 × 2) unit cell (see the model in figure 2(a))

for which a herringbone packing of the biphenyl moieties hasbeen suggested [34]. We note at this point that the height

3

Nanotechnology 20 (2009) 245306 C Shen and M Buck

difference between the areas of BP2 and MC12 as detailed bythe height profile in figure 2(a) is only coincidentally close toa monatomic step of the gold surface. A different combinationof molecules results in a different height due to differences intunneling resistance of the molecules and molecule–substratecontacts.

A closer look at the modification parameters revealsthat displacement occurs at a current even as low as 6 pA(figure 2(a)). In combination with a tip potential of 1.2 Vsuch low current values are usually chosen for non-perturbingimaging and, as already mentioned above, the tip is well abovethe BP2 SAM under these conditions. This demonstrates,on the one hand, that only a little interference between tipand SAM is required to trigger displacement. In combinationwith the significantly different parameters used to generate thestructure of figure 2(b) one can, on the other hand, concludethat the achievable resolution, i.e. level of precision at whicha BP2/MCn boundary can be generated, does not seem todepend sensitively on tunneling current and bias. However,what is decisively affected by the tunneling parameters is therate at which the BP2/MCn boundary moves. Bringing the tipcloser to the surface increases the writing speed significantlyas inferred from comparison of figures 2(a) and (b) which took30 s (equivalent to 342 scan lines) and 1 s (equivalent to about6 scan lines), respectively.

Another important point is that the nanostructures arestable in the solution-based environment under which thepatterning is performed and, thus, further modification ofpatterns or individual nanostructures is possible. This isdemonstrated in figure 3(a) where individual nanoislands ofBP2 are eliminated one by one from an array of islands.Control over the size of a nanostructure is exemplified infigure 3(b) where an island about 70 nm2 in size is reducedin steps to less than 30 nm2. It is noteworthy that the molecularresolution is maintained throughout the whole process.

As already indicated, the displacement scheme presentedhere differs in several aspects from established ones. Withcurrents as small as 6 pA and/or voltages well below theenergy required to break an S–Au bond, current-induced bondbreaking [44, 45] can be rather safely excluded. Also, field-induced desorption cannot be the reason as thiol exchangeoccurs at potentials which are much lower than the onerequired for field-induced effects. Furthermore, the polarityof the bias does not affect displacement. An electrochemicalprocess is excluded by the low bias.

This leaves us with a mechanical effect even thoughits mechanism is not immediately obvious since tunnelingparameters are such that the tip is not in mechanicalcontact with the BP2 SAM. However, the fact thatdisplacement progresses via movement of a well-definedboundary separating areas of BP2 and MCn (see figure 1(a))indicates that the transition region plays the decisive role. Themost obvious difference between the molecules is the tunnelingdistance which, for a given set of tunneling parameters, issignificantly smaller for MCn compared to BP2. How thisaffects the displacement is revealed by the experiment shown infigure 4 which starts from a line of an alkane thiol SAM writtenonto a BP2 SAM. As illustrated by the scheme (figure 4(a)),

the STM tip performs a rectangular loop and during onecycle it crosses the line of alkane thiol twice but in oppositedirections. Cycles in clockwise direction produces the patternshown in figure 4(c). The clear asymmetry in the patternproves that displacement occurs only when going from high(MC12) to low (BP2) tunneling resistance, i.e. when thetip retracts. We propose that the mechanism is based onthe shear force the tip exerts onto the thiol with the lowerconductance. When the tip moves from molecules of high tothose of low conductance the shear force points away fromthe former ones (see figure 4(d)), whereas upon the oppositemovement, i.e. low to high conductance, a pushing forceis exerted onto the molecules of high conductance whichweakens the Au–S bond sufficiently to be displaced by theother thiol. While this, at first glance, appears very similar tonanografting there is a fundamental difference as the tip doesnot displace the high conductance thiol directly but requiresthe low conductance thiol as mediator. Furthermore, the effectrelies on the combination of conductivity and length of themolecules, i.e. their conductance. The mechanism works fordisplacing molecules which are shorter but have a conductancesufficiently low to give rise to a shear force above the thresholdrequired for replacement. Correspondingly, a molecule canhave a higher conductivity as long as the length of the moleculeis sufficient to bring the conductance below the value requiredto reach the threshold shear force. The latter is corroborated byexperiments using a terphenyl thiol replacing BP2 (not shown).Assuming that the conductivity is determined by the aromaticunits and, thus, is very similar for both types of molecules, thelower conductance of the terphenyl thiol due to the length ofthe molecule again causes displacement.

Looking for other factors which might explain theexperimental observations, pulling forces due to tip-induceddipolar interactions which could weaken the bonding of BP2molecules to the substrate have also to be considered. While acontribution from such forces cannot be completely excludedit is considered small. This can be concluded, on the one hand,from experiments using the terphenyl thiol where the tip ismore retracted compared to BP2. Further corroboration comes,on the other hand, from experiments varying the chain length ofthe alkane thiols. In this case the distance between tip and BP2does not vary much or even increases slightly with increasinglength of the alkane chain. Thus, the dipolar force does notincrease and its effect on replacement should not be dependenton the chain length which is in pronounced contrast to theexperiment where variation from MC6 to MC18 for a givenset of tunneling parameters for the BP2 SAM causes a changefrom no replacement to a very quick one. As a final remarkwe note that the exact replacement parameters are not onlydetermined by the conductance but are also influenced by thefactors determining the energy of the SAM matrix such as thestrength of the Au–thiol bond or intermolecular interactions.For SAMs such as BP2 which, due to the competitive design,are in an energetically higher state than alkane thiol SAMsor BPn homologues with an odd number of CH2 moieties,displacement is easier and thus should occur at lower shearforces compared to more stable SAMs.

4

Nanotechnology 20 (2009) 245306 C Shen and M Buck

Figure 3. (a) Series of STM images acquired in a solution of MC12/hexadecane which shows a one-by-one elimination of BP2 islands.Modification and imaging parameters are 0.1 nA/0.1 V and 5 pA, 1.2 V, respectively. (b) Sequence of three STM images showing the shapingof a BP2 island at molecular resolution. Patterning and imaging parameters are −0.5 V/30 pA and −1.2 V/5 pA, respectively.

Figure 4. (a) Schematic illustration of patterning across an existingline. The STM tip follows a rectangular path in a clockwisedirection, thus crossing the line twice but in opposite directions. ((b),(c)) STM images of MC12 in a BP2 matrix with line only (c) andafter 25 cycles with the STM tip (50 pA, 0.5 V) moving at0.1 μm s−1. (d) Cartoon of the nanolithography process.

4. Conclusion

Sub-10 nm scale structures in SAMs can be created in areproducible way by STM under hexadecane solution in the

presence of a second type of molecule. The presented methodproduces chemically well-defined phases as replacement iscomplete and, thus, no mixed phases are generated. Whilethe conductance-based patterning method works also for othercombinations of molecules, the advantage of using SAMsbased on the architecture of molecules such as BP2 is theirsuperb structural quality, i.e. extended domains and extremelylow defect density, in combination with a low activation barrierrequired for displacement. These features, together with thescanning probe operating at molecular resolution, offer anapproach for the generation of precisely defined structures atthe ultrashort length scale.

Acknowledgments

This work has been financially supported by EPSRC. CSgratefully acknowledges support through an EaStCHEMstudentship.

References

[1] Smith R K, Lewis P A and Weiss P S 2004 Prog. Surf. Sci. 75 1[2] Love J C, Estroff L A, Kriebel J K, Nuzzo R G and

Whitesides G M 2005 Chem. Rev. 105 1103[3] Kramer S, Fuierer R R and Gorman C B 2003 Chem. Rev.

103 4367[4] Tan Y H, Liu M, Nolting B, Go J G, Gervay-Hague J and

Liu G-y 2008 ACS Nano 2 2374[5] Sondag-Huethorst J A M, van Helleputte H R J and

Fokkink L G J 1994 Appl. Phys. Lett. 64 285

5

Nanotechnology 20 (2009) 245306 C Shen and M Buck

[6] Golzhauser A, Eck W, Geyer W, Stadler V, Weimann T,Hinze P and Grunze M 2001 Adv. Mater. 13 806

[7] Turchanin A, Schnietz M, El-Desawy M, Solak H H,David C and Golzhauser A 2007 Small 3 2114

[8] Leggett G J 2006 Chem. Soc. Rev. 35 1150[9] Williams J A and Gorman C B 2007 Langmuir 23 3103

[10] Schoer J K, Zamborini F P and Crooks R M 1996 J. Phys.Chem. 100 11086

[11] Schoer J K, Ross C B, Crooks R M, Corbitt T S andHampdensmith M J 1994 Langmuir 10 615

[12] Schoer J K and Crooks R M 1997 Langmuir 13 2323[13] Perkins F K, Dobisz E A, Brandow S L, Calvert J M,

Kosakowski J E and Marrian C R K 1996 Appl. Phys. Lett.68 550

[14] Osada T, Zhu N, Zhang Y F and Komeda T 2008 J. Phys.Chem. C 112 3835

[15] Mizutani W, Ishida T and Tokumoto H 1998 Langmuir14 7197

[16] Marrian C R K, Perkins F K, Brandow S L, Koloski T S,Dobisz E A and Calvert J M 1994 Appl. Phys. Lett. 64 390

[17] Delamarche E, Hoole A C F, Michel B, Wilkes S, Despont M,Welland M E and Biebuyck H 1997 J. Phys. Chem. B101 9263

[18] Gorman C B, Carroll R L, He Y F, Tian F and Fuierer R 2000Langmuir 16 6312

[19] Kim J, Uchida H, Honda N, Hashizume N, Hashimoto Y,Kim H, Nishimura K and Inoue M 2003 Japan. J. Appl.Phys. 1 42 4770

[20] Liu M, Amro N A and Liu G Y 2008 Annu. Rev. Phys. Chem.59 367

[21] Liu G-Y, Xu S and Qian Y 2000 Acc. Chem. Res. 33 457[22] Zhao J W and Uosaki K 2001 Langmuir 17 7784[23] Jang J W, Sanedrin R G, Maspoch D, Hwang S, Fujigaya T,

Jeon Y M, Vega R A, Chen X D and Mirkin C A 2008 NanoLett. 8 1451

[24] Liang J, Rosa L G and Scoles G 2007 J. Phys. Chem. C111 17275

[25] Liu J F, Von Ehr J R, Baur C, Stallcup R, Randall J andBray K 2004 Appl. Phys. Lett. 84 1359

[26] Salaita K, Wang Y H and Mirkin C A 2007 Nat. Nanotechnol.2 145

[27] Mirkin C A 2007 ACS Nano 1 79[28] Ginger D S, Zhang H and Mirkin C A 2004 Angew. Chem. Int.

Edn 43 30[29] Salaita K, Wang Y H, Fragala J, Vega R A, Liu C and

Mirkin C A 2006 Angew. Chem. Int. Edn 45 7220[30] Bullen D, Chung S W, Wang X F, Zou J, Mirkin C A and

Liu C 2004 Appl. Phys. Lett. 84 789[31] Kim Y T and Bard A J 1992 Langmuir 8 1096[32] Keel J M, Yin J, Guo Q and Palmer R E 2002 J. Chem. Phys.

116 7151[33] Rong H T, Frey S, Yang Y J, Zharnikov M, Buck M, Wuhn M,

Woll C and Helmchen G 2001 Langmuir 17 1582[34] Cyganik P, Buck M, Strunskus T, Shaporenko A, Witte G,

Zharnikov M and Woll C 2007 J. Phys. Chem. C 111 16909[35] Azzam W, Cyganik P, Witte G, Buck M and Woll C 2003

Langmuir 19 8262[36] Cyganik P, Buck M, Azzam W and Woll C 2004 J. Phys.

Chem. B 108 4989[37] Cyganik P, Buck M, Wilton-Ely J D E T and Woll C 2005

J. Phys. Chem. B 109 10902[38] Cyganik P and Buck M 2004 J. Am. Chem. Soc. 126 5960[39] Cyganik P, Buck M, Strunskus T, Shaporenko A,

Wilton-Ely J D E T, Zharnikov M and Woll C 2006 J. Am.Chem. Soc. 128 13868

[40] Silien C and Buck M 2008 J. Phys. Chem. C 112 3881[41] Su G J J, Aguilar-Sanchez R, Li Z H, Pobelov I, Homberger M,

Simon U and Wandlowski T 2007 ChemPhysChem 8 1037[42] Xu S, Laibinis P E and Liu G Y 1998 J. Am. Chem. Soc.

120 9356[43] Yu J J, Tan Y H, Li X, Kuo P K and Liu G Y 2006 J. Am.

Chem. Soc. 128 11574[44] Schulze G, Franke K J and Pascual J I 2008 New J. Phys.

10 065005[45] Hla S W, Bartels L, Meyer G and Rieder K H 2000 Phys. Rev.

Lett. 85 2777

6