BlO6RAPH YEthan D. Minot reeeit€di:;his PhD in $$lA from 1

Cornell Univenity' in20M.Hecompletedpost- l.doctoral work at DelftUniversity of Technology krbefore joining the L'physia faculty at Oregon State University.Dr Minot's research focuses on the elec-tronic and optical properties of carbon nan-otubes. Many of his experiments rely onscanning probe techniques such as atomicforce microscopy and scanning photo-current microscopy.

ABSTRACTUsing CNT network sensors as our workingexample, we review AFM-based techniqueswhich are used to study and engineer nano-electronic devices. We have used dc-EFM

and ac-EFM to identify the locations andresistances of individual CNTs that are elec-trically connected in parallel. Next, 5GM andtm-SGM were used to reveal the semicon-ducting response of each CNT. With theinformation available in these scans a singleCNT with desirable properties was singledout for further experimentation. Theunwanted CNTs were electrically cut with a

biased AFM probe to leave a device con-taining a single CNT. An atomic-sized tran-sistor with chemical functionality was engi-neered in the remaining CNT using a voltagepulse from the AFM probe. This nano-transistor was then demonstrated to be a

single molecule sensor sensitive to EDC in anaqueous environment. This multi-step mea-surement and manipulation process illus-trates the power of AFM-based techniquesto map out and control the properties ofna noelectron ic devices.

KEYWORDSscanning probe microscopy, electric forcemicroscopy, scanning gate microscopy, car-bon nanotubes, nanoengineering, nano-electronics

AUTHOR DETAILSDr Ethan D. Minot, Department of Physics,

Oregon State University, Corvallis,Oregon 97331-6507, USA

Tel: +1 541 737 9671Ema i | : m inote@science.oregonstate.edu

Microscopy and Analysis 26(3):7-'10 (Atrlt), ZOtZ

Scanning Probe Microscope Techniques for the

Engineering of Nanoelectronic DevicesLandon Prisbrey,l li-Yong Park,2 Kerstin Blank,3 Anir Mosha7a and Ethan D. Minotl1. Department of Phpics, Oregon State University, Corvallis, OR, USA. 2. Department of Physics and Division of EnergySystens Research, Ajou University, Suwon, Korea. 3. lnstitute for Molecules and Material, Radboud University Nijmegen,The Netherlands. 4. Asylum Research, Santa Barbara, CA, USA

INTRODUCTIONResearch in nanoscale systems relies heavily onscanning probe microscopy. Many of thesenanoscale systems have unique properties thatare amenable to scanning probe techniquesand, therefore, a wide variety of specializedscanning probe techniques have been devel-oped. For example, in the field of magneticmemory research, magnetized probes are used

to map magnetization patterns [1], while fer-roelectric materials are studied by locating theelectric fields from polarized domains [2]. Inmany cases the scanning probe microscope is

used both as a manipulation tool as well as animaging tool. lt is possible, for example, to flipdiscrete magnetic or ferroelectric domains andthen image the resulting changes [2, 3].

Nanoscale Electronic Devices

A growing toolkit of scanning probemicroscopy-based techniques is enabling newways to build and investigate nanoscale elec-tronic devices. In this article we review keytechniques that allow for mapping the distrib-ution of electrical resistances, measuring localfield-effect sensitivity, and engineering newelectrical characteristics into nanoscaledevices.

Here we review several advanced techniquesto characterize and manipulate nanoelec-tronic devices using an atomic force micro-scope (AFM). Starting from a carbon nanotube(CNT) network device that is fabricated by con-ventiona I photol ithog ra phy (m icrometer-sca le

resolution), individual carbon nanotubes canbe characterized, unwanted carbon nan-otubes can be cut, and an atomic-sized tran-sistor with single molecule detection capabili-ties can be created. While we focus on CNT

devices, many of the same techniques are

being used for nanowire [a] and graphene [5]devices.

These measurement and manipulationprocesses were performed with the MFP-3D

AFM (Asylum Research, Santa Barbara, CA)with the Probe Station Option (Figure 1) andillustrate the recent progress in AFM-basedtechniques for nanoelectronics research.

We first discuss how electric forcemicroscopy (EFM) and scanning gatemicroscopy (SGM)techniques are used to char-acterize the electrical response of nanoelec-tronic devices. The information contained in

EFM and 5GM images is extremely valuable forrelating the local electrical behavior of a

device to the global device properties. ln thespecific example of our CNT network device,we use the knowledge of local resistance andlocal semiconducting behavior to choose the'best' CNT out of a mixture of parallel CNTs.

ELECTRIC FORCE AND SCANNINGGATE MICROSCOPYTo acquire EFM and SGM signals, a conductingAFM probe hovers above an electrically con-tacted device (Figure 2a). During a 'hoverpass', the microscope maintains a constantseparation between the conductive AFM tipand the surface (typical separations are 20 to200 nm). To achieve this constant separationdistance, the microscope alternates betweenstandard topographical line scans and hoverpass line scans. A typical electrical contact con-figuration involves a current amplifier andthree voltage sources (V,6, V6, and V6o).

Depending on the particular imaging tech-nique, different voltage signals will be appliedand the AFM tip may or may not be piezoelec-trically driven Gee Table 1).

To demonstrate the utility of EFM and SGM,

tigure 1:

Experimental setup lhe Probe Shfron attaches lo the MFP-3D AFM scanner and allows easy electrrcal probing of sample properties, electrrcalbiasing,

andothermeasurementswhtlerhesamphuberngscannedwiththeAFM.Avarieflof electrrca/connectronscan bemadeandconbinedwithvari.

ousinaging modes,

MICROSCOPY AND ANALYSIS SPM ISSUE APRIL 2012

we used dc-EFM [6], ac-EFM [7], SGM [8] andtip modulated SGM (tm-SGM) [9] to image acarbon nanotube (CNT) network device (Fig-

ure 2). The device was fabricated by growingCNTs on a SiOr/Sisubstrate (300 nm oxide) [10],and then patterning electrodes with pho-

tolithography. The images in Figures 2b and3a are a combination of a dc-EFM data (color

scale) and AFM topography. Many individualCNTs are seen bridging the gap between thesource and drain electrodes. We are interestedin learning which CNTs are metallic, which are

semiconducting, which are well connected tothe metal electrodes and which containnatural defects.

Electric Force Microscopy Tech niquesThe electrostatic forces which give imagingcontrast in dc-EFM [6] (Figures 2b and 3a) are

due to a voltage difference between the tipand the grounded electrodes. During each

hover pass, the AFM cantilever was piezoelec-

trically driven at a fixed frequency (the funda-mental resonance of the cantilever). lf theAFM tip is above a grounded conductori a ver-tical gradient in electrostatic force causes a

shift in the resonance frequency of the can-

tilever. This resonant frequency shift modu-lates both the phase and amplitude of the can-

tilever motion. The dc-EFM image shown inFigure 3a allows us to quickly locate all con-ducting objects in the large (50 pm) imagingarea and requires minimal setup.

The vertical gradient in electrostatic force,which gives rise to the dc-EFM signal, dependson the tip-sample capacitance and the tip-sam-ple voltage difference, When the tip hovers

above a large-area electrode, the vertical gra-

dient in tip-sample capacitance is large. When

the tip hovers over a small CNL the verticalgradient in tip-sample capacitance is small.Thus, the EFM signal from CNTs is weaker thanfrom the electrodes. There has been steadyprogress in mapping out the vertical gradientin tip-sample capacitance to allow more quan-

titative imaging of the potential differencebetween the tip and the sample 111 ,121.

Figure 3b shows an ac-EFM [7] image inwhich different voltage signals are sent to thesource and drain electrodes, respectively. This

type of ac-EFM scan is used to visualize thevoltage profile along the length of electrically-biased CNTs. A similar measurement can be

done with dc-EFM, however; the ac-EFM signalis cleaner (more discussion below). CNTs

labeled 1-4 in Figure 3b exhibit a sharp voltagedrop where they meet the inner electrode (the

length of the CNT appears uniform black). This

sudden voltage drop indicates a high contactresistance. The CNT labeled 5 exhibits a sharpvoltage drop midway along its length. The

location of the voltage drop corresponds to a

sharp bend in the CNT, showing that the bendis a point of high resistance. CNTs labeled 6-10

show a gradual voltage gradient along theirlength, indicating that these CNTs are wellconnected to both electrodes, and are freefrom major electrical defects. These CNTs withsmall contact resistance will be subsequently

singled out for additional engineering andexperiments.

MICROSCOPY AND ANALYSIS SPM ISSUE APRIL 2012

23"

=LLt!

0

Figure2:

(a) Schematicillustrating a carbon nanotubefted-effecttransistor and 'horlerpas'6ee texf) microscopy geometnes. Not to sca/e.

(b) AFM topognphy colored with EFM phase (V* = 8 V; height = 20 nn),

Table 1:

Key diff erences betwmn modes us ed f or rmaging nanoe/rtrontc derltces.

Vra Piezoelectrically Driven? 'Hover Pass'Signal

dc-EFM DC yes phase and amplitude

ac-EFM AC @-/.rn,ir"r", no amplitude

SGM DC no 1to

tm-SGM DC yes /ro, AC @./cantit"u"..

Figure j:'Hover pass' microscopy data of a typrcalcarbon nanolube field-effntranststor. (a)EFM M,r=20 nV, V os= 0 V, Vrr= 5V, height = 20 nm). Color

sca/eshowsphase of cantileuerresponse, (b)AlternailngcurrentEfM (V,a= 100 mV @63 kHz,Von=0V,V,r=2V, height=20 nm) Co/orsca/eshows

anplitude of canf/ever response. (c) SGM (V ,o= 29 ftM, V r= 0 V, V nr= 5 V, height = 20 nm), Co/or scah shor,rn the change in current A l,o through

thedeyrce. (d)Tip-moduhtedSGM(Vil=20mV,Vw=0V,Vrr=5V,herght=20nm).Colorsca/eshor,rnthechangeincurrentatthecantietler

fundanental f requenq, A 1,0.,,

To obtain the ac-EFM signal in Figure 3b, anAC bias was applied to the outer electrode(frequency matching the cantilever's mechan-ical resonance) while the inner electrode wasgrounded and the AFM tip was biased to 2 V.

During the hover pass the cantilever is notpiezoelectrically driven, instead cantileveroscillations are driven by capacitive couplingto the AC biased electrode. These oscillationsare detected by monitoring the cantileverdeflection signal using a lock-in amplifier. Theoscillation amplitude is plotted as a function oftip position to create the ac-EFM image, Can-tilever oscillations are strongest when theprobe is nearthe biased electrode, The ac-EFM

signal is'cleaner' than dc-EFM signals becausestray DC signals (such as charge on the insulat-ing substrate)are invisible in ac-EFM.

Scanning Gate Microscopy TechniquesWe now turn our attention to scanning gatetechniques, which allow researchers to furthercharacterize the electronic properties of indi-vidual CNTs. Scanning gate techniques identifythe local semiconducting response of a

nanoscale conductor, revealing the presence

of localized semiconducting 'hotspots' andbottlenecks for electron transport [8].

Figure 3c shows an SGM image, where theAFM probe acts as a roaming gate electrode tomodulate the doping in the underlying elec-

tronics. The 5GM image (Figure 3c)shows a sin-gle feature, indicating that the global semi-conductor response of the device is dominatedby a single CNT (labeled 10 in Figure 3b). OtherCNTs are present in the imaging area, but theSGM signals from these are significantlysmaller. Methods to improve sensitivity arediscussed below.

The imaging contrast in Figure 3c corre-sponds to changes in global device currentwhen the biased probe locally changes thedoping level in the underlying electronics. Asmall source-drain voltage (4a) is appliedacross the device to drive global current. By

monitoring changes in current as a function oftip position, the local field-effect sensitivitycan be mapped.

A variation of 5GM which offers greater sen-

sitivity is tip-modulated SGM (tm-SGM) [9].Compared to conventional SGM, tm-SGMoffers higher spatial resolution and is capableof resolving weaker signals. Figure 3d shows a

tm-SGM image. This scan reveals five CNTs

with a semiconducting response (labeled 6-10

in Figure 3b). One CNT contains a region ofhighly localized semiconducting response,

indicating the presence of a natural defect(labeled 8 in Figure 3b).

In tm-SGM the AFM cantilever is biased andpiezoelectrically driven at its resonant fre-quency, resulting in an AC electric field whichmodulates the local doping level in the under-lying electronics. The local field-effectresponse of the sample is mapped as a func-tion of tip position by isolating the componentof the conductance signal at the cantilever res-

onant frequency using a lock-in amplifier.

Interpreting SGM and EFM DataThe detailed characterization outlined in

coa

600

500

40time (sec)

tigure4:

Elxtrtcalcuttingof unwantedcarfunnanotubes. (a)Elettrrcforcemrcroscopermageof adevtce.CuttingpathindicatdbydottdanowU,=25mV,

Vw=0V,V,,r=8V, height=20 nm), (b)finetraceof cutilngevent(/*=25 nV,Vr=0V,Vrr=-$V, contactforce=6-9 nN).

n +r r*15 msU-I I-IJ.3 V(b)

l-

Eo

80

60

40

20

0

-_\it^\

10x\/

-4-2 0 2von U)

4

q3 nm height 0

:0 A/.//o -1

:0 a/sd//o -1

Figure 5:

Defect engrnenn g in a carbon nanotube fteld-effn transistor. (a) khematk rllustrating defn enginertng gnmetry. b) Tranststor charactenstra

before(duhedblack)andafter(sohdrdJdefeooeatron.Therdcurvehasben mu/trp/rd by l0 ffr=25ptl). @ AFMtopqnphyof aCNT devtce.

A defxt wu enginerd at the /ocatrb n indkatd by an arrow. (d) SGM daa of the derlrce sh own in (c) taken before and (e) after defect engrnenng

Ud=25tnV,V*=0V,V,r=SV,height=20nm).Theco/orsca/e (Aldll)rndrcateschangesrn devtcecurrentrelatrrletothe fuselnevalue,l,

ln (d) and (e) the values of I o are 280 nA and 7 nA, rapfrruely.

MICROSCOPY AND ANALYSIS SPM ISSUE APRIL 2012

Figure 3 allows us to pick the best CNT out ofthe ten CNTs for building a nanotransistor withchemical functionality. Using dc-EFM (Figure

3a), we identified the locations of all CNTs in a

large area scan, providing a road map for engi-neering. With ac-EFM (Figure 3b)we identifieddesirable CNTs with low contact resistance (no

sudden drop in voltage at metal-CNT contact)

and an absence of natural defects (no sudden

drop in voltage along length of CNT). Finally,

using SGM methods (Figure 3 c,d), we identi-fied CNTs with a semiconducting response.

The most desirable CNT for building a nan-

otransistor is one with low contact resistance,

no natural defects, and a uniform semicon-ducting response along its length. The CNT

labeled 10 in Figure 3b satisfies these criteria,and is singled out for further experimentation.

Atomic Force Microscopy EngineeringWith initial characterization complete, wemove to AFM manipulation techniques. Once

a desired CNT is identified, we use AFM engi-neering to prune unwanted CNTs [13], thenengineer an atomic-sized transistor elementinto one remaining CNT [14]. These techniquescombine AFM lithography modes with controlof the tip-sample voltage. Mechanical manip-

ulation by AFM has also been used by many

authors [15, 16], but is not discussed here.To cut unwanted CNTs, a Pt-coated AFM

probe is biased to 4ip = -10 V engaged withthe surface, and then dragged through theCNTs in contact mode (contact force -5-9 nN).

Previously obtained dc-EFM and ac-EFM scans,

such as in Figure 3 a,b, provide a convenientroad map of the locations of the unwantedCNTs. Figure 4 shows a time trace of globaldevice conductance during electrical cutting.Sharp decreases in device conductance areobserved as the CNTs are severed. Afterwards,ac-EFM and SGM scans may be used to verifythat a single CNT forms the only electricalcontact between the electrodes.

A single CNT can be converted into an

atomic-sized transistor by further AFM-basedengineering. Such nanotransistors approachthe ultimate limit of miniaturization. To engi-neer an atomic-sized transistoL a conductingAFM probe is brought into gentle contact withthe CNT and a -3 V 15 ms square wave pulse is

applied to the tip (Figure 5a). During the pulse,

a chemical defect is incorporated into the side-

wall of the CNI and the device's electricalproperties change significantly as a result (Fig-

ure 5b) 114, 17, 201. Typical defects addbetween 10 kO to 1 MO to the overall resis-

tance of the circuit, and change the field-effect sensitivity of the device.

Figure 5c shows a single-CNT device pro-

duced by the electrical 'nicking' methoddescribed above. SGM measurements showthat the pristine CNT is uniformly gate sensi-

tive (Figure 5d). Figure 5e shows an identicalmeasurement following defect engineeringwhere gate sensitivity is localized to the regionaround the defect. The SGM scan reveals thatthe defect acts as a gate sensitive bottleneckfor transport, and the remainder of the CNT

serves only as contact electrodes to the minia-ture transistor.

A unique attribute of point-defect nan-

otransistors is their sensitivity to electrostaticpotential within a very small detection vol-ume. As such, they are ideally suited for single-molecule sensing applications. Recently, point-defect nanotransistor sensors have been used

to study biochemical reactions at the singlemolecule level [18, 19]

Figure 5 demonstrates the use of an AFM-engineered point-defect nanotransistor as asingle-molecule sensor. The reaction betweenN-(3-d i methylam i nopropyl)-N'-ethylcarbod i-

imide (EDC) and a carboxyl group is used as a

model reaction. EDC is frequently used to acti-vate carboxyl groups in bioconjugation chem-istry, and reacts reversibly with a carboxylpoint defect on the CNT sidewall. The point-defect nanotransistor was immersed in an

electrolyte solution containing EDC (Figure

6a). Figures 5 b,c show the conductance datafor a point-defect CNT before and after theaddition of 20 pM EDC, respectively. Discrete

switching events are observed in the presence

of EDC.

This two-state telegraph noise is believed toreflect the fluctuating electrostatic environ-ment as individual EDC molecules react withthe carboxyl group on the CNT sidewall [18].

coNcLUstoNsUsing CNT network sensors as our workingexample, we have reviewed AFM-based tech-niques which are used to study and engineernanoelectronic devices. We have used dc-EFM

and ac-EFM to identify the locations and resis-

tances of individual CNTs that are electricallyconnected in parallel. Next, SGM and tm-SGMwere used to reveal the semiconductingresponse of each CNT. With the informationavailable in these scans a single CNT with desir-able properties was singled out for furtherexperimentation. The unwanted CNTs wereelectrically cut with a biased AFM probe toleave a device containing a single CNT. Anatomic-sized transistor with chemical func-tionalitywas engineered in the remaining CNT

using a voltage pulse from the AFM probe.

This nanotransistor was then demonstrated to

MI5rngh-moku/e de ter.tton of

EDC,

(a) Schemafrc rllustrattng

detectron geometry, The hq-

urd potentral rs contro//ed

wtth a ltquid-gate e/ectrode

btasedtoV,n.

(b) l*trace of a nanotranri-

tor in MES buffer (pH 4.5, V d

= 25 nV, V b= 250 mV) ,

(c) l,o trace of the same

dev tce f ollowing the addttr on

of 20 ttMEDC(pH4.5,Vil=

25 mV,V,n=100 mV).

be a single molecule sensor sensitive to EDC in

an aqueous environment. This multi-step mea-

surement and manipulation process illustratesthe power of AFM-based techniques to mapout and control the properties of nano-electronic devices.

ACKNOWLEDGEMENTSWe thank Kristina Prisbrey for assistance withFigures 2, 5, and 6, and Bo Criss for assistance

with Figure 3. We thank Sophie Ripp for initialexperiments with EDC using suspended CNTs.

We thank Paul Schuele and Sharp Labs ofAmerica for assistance with device fabrication.This work is funded by the Human FrontierScience Program.

REFERENCES1. Pukin,S. etal Science 320190-194, 2008.

2, Gruverman, A, l. Vacuum Sci, & Technology B: Micro-

e/ectronics and Na nometer Structures l 4:602, 1 996.

3, Gibson, G. A,, Smyth, J. F., Schula,5., Kem, D, P.IEEE

Tranncilons on Magnetrcs 27:5187-5189, l99l ,

4. Zhou, X,, Dayeh, S, A,, Wang, D,, Yu, E. T, Applied Physks

1etten90233118,2007.

5, lalilian,R. eta/, Nanotech nology 22295705, 2011,

6. Martin, Y,, Abraham, D, W,, Wickramasinghe, H. K. Applied

Physks Letten 521 1 03, 1 988,

7, Bachtold, A., Fuhrer, M.5,, Plyasunov,S, et a/, Phys. Rev, letf,

I9.

10,

11.

12,

13,

14,

t0.

a1U.

18,

19,

20,

84:5082,2000,

Tans, S, J,, Dekker, C. Nature 404:834-835,2000.

Wilson, N, R,, Cobden, D, H, NanoLettenS:2161-2165,2008.

Kong, L et al, Nature 395:878-881,1998.

Lei, C, H, et al tVanotechn ology 15:627-534,2004,

Yan, M,, Bernstein, G, H, Surtace and lnbrtace Analysis

39:354-358,2007,

Puk, l,-Y, et al, Appl, Phys, Lett,80'M64448,2002,

Park, l.-Y, Appl. Phys. Lett, 90:0231 1 2-3, 2007,

lunno, T,, Deppert, K,, Montelius, 1,, Samuelson, L. Applied

Phlnrcs leffers 5 53527, 1 995,

Postma, H. W, C,, Teepen, T,, Yao, 2,, Grifoni, M., Dekker, C.

Science 293:7 5-7 9, 2001,

Kim, D,-H,, Koo, l,-Y,, Km, J.J, Phys. Reu B 68113406, 2003,

Goldsnith, B. R,, Coroneus, L G,, Kane, A. A,, Weiss, G, A,,

Collins, P, 6. l/ano Letten 81 89-l 94, 2008,

Sorgenfrei,S, et al tVature tVa no, 6126-132,2011 ,

Pr'tsbrey, L. et al, L Phys, Chem. C 1151961-1965,2012,

@2012 John Wiley & Sons, Ltd

(c)180

cl

Jf801

0

(b)

10 10time (ms)time (ms)

10 MICROSCOPY AND ANALYSIS SPM ISSUE APRIL 2012