University of Alberta National Institute for Nanotechnology Edmonton, Alberta, Canada

No Slide Title

University of AlbertaNational Institute for NanotechnologyEdmonton, Alberta, Canada

Molecules in Circuits, a New Type of Microelectronics?

Richard L. McCreery

#1Organic Electronics:Sony organic LED display

charge transport by hopping across 100 nm 10 m distancescarriers are radical ions and/or polarons, often low stabilityWhat happens when charge transfer distance is decreased to 1-25 nm, and electric fields may exceed 106 V/cm ? #26000 rpmcommercial photoresist (novolac resin)Soft Baking90OCClean Si WaferSpin-CoatingGas Flow95% N2;5% H2(1000OC, 1 hour)Pyrolysis

50 mOptionalLithographyOur substrate: Pyrolyzed Photoresist Films (PPF)PPF (roughness < 0.4 nm rms)

50 m

Kim, Song, Kinoshita, Madou, and White, J. Electrochem. Soc., 1998, 145, 2315Ranganathan, McCreery, Majji, and Madou, J. Electrochem. Soc., 2000. 147, 277282Ranganathan, McCreery, Anal. Chem, 2001, 73, 893-900.#3Delamar, M.; Hitmi, R.; Pinson, J.; Saveant, J. M.; J. Am. Chem. Soc. 1992, 114, 5883. Liu, McCreery, J. Am. Chem. Soc., 1995, 117, 11254. Belanger, Pinson, Chem. Soc. Reviews 2011, 40, 3995 .PPF electrodeeCH3CN+RN2+RNH2HNO2 surface bond stable to > 500 oC high coverage very low in pinholes prone to multilayer formation. Surface modification via diazonium reduction:sp2 carbonRRN2+ Re-#4modified electrode Ox Redsp2 carbonAdv. Mater. 2009, 21, 4304 J. Phys. Chem. C 2010, 114, 15806J. Am. Chem. Soc. 2011, 133, 191681-6 nmVCuFIB/TEM:

silicon (for contrast)CuSi10 nmBTB

PPFe-CAuAue-carbon

Packaged: #5

Pyrolyzed Photoresist Film (PPF) microfabrication:100 mm

1. lithography2. pyrolysisPPF structure and conductivity similar to glassy carbon roughness < 0.5 nm rms by AFM

#6

PPF Echip 4 waferMagnification = 1molecules attached to PPF by electrochemical reduction of diazonium ions

PPF leadsJunctionMag. = 10

Cu/Au depositedthrough shadow maskMag. = 10next slide#7

cleave through junction region, then SEM of edgeTEMMag. = 60

SiO2SiPPFSiO2Cu/Au1 mjunction regionMag. = 30,000

Cu/AuSiO2PPFMolecules(not resolved)100 nmMag. = 300,000

5 nmsilicon (for contrast)FIB/TEM:Mag. = 5,000,000CuAusp2 carbon#

near Jasper, Alberta#9

log scaleexponential above ~0.2 VV

frequency independent, 0.01 to 105 Hz symmetric can scan > 109 cycles without change survived 150 oC for > 40 hrsACS Applied Materials & Interfaces 2010, 2, 369380 x 80m junctionNNNOO V1 mA0.5 Vno molecule-2-1012-1.0-0.50.00.51.0J(A/cm2)NAB(nitroazobenzene)NAB

32 junctionslooks likeMarcus kineticsIs it?#10

1000/T, K-1Arrhenius:

T = 5 K37 meV0.03 meVstrong thickness dependence

d = 2.2 nm2.8 nm3.3 nm4.5 nm5.2 nmmolecular layerthicknessJ. Phys. Chem C, 2010, 114, 15806not activated for T400 junctions)More molecules:NO EFFECT!

Sayed, Fereiro, Yan, McCreery, Bergren; PNAS (2012) 109, 11498.

#13

PPFNO2BrCCHEnergy levels of PPF/molecule from Ultraviolet Photoelectron Spectroscopy:

* method of Kim, Choi, Zhu, Frisbie, JACS 2011, 133, 19864*hv=21.2 eVEFermieEF - EHOMOeWFmean for 8aromatics=1.3 0.2 eValiphatic=2.0 0.1 eVvacuumWF=4.404.534.884.97#14Energy (eV vs. Vacuum)-5-6-7-4-3-2-8PPFBTBABPhBrNABNPNC8H17Free molecule HOMOs:1.3 0.2 eV(aromatics)NC8H172.0 0.1 eVMolecules bonded to PPF:UPS and Simmons barriersare consistent, and correlatewith observed values0.692.99Sayed, Fereiro, Yan, RLM, Bergren, PNAS, 109, 11498 (2012)#15Why?PPF(unmodified)WF ~ 4.6+ eVelectron donating0.7 eVNNNOOWF ~ 5.1+ eV4.6 eVelectron withdrawingHOMO ~6.6*2.0NNNOOSSWF ~ 4.4+ eV4.6 eV+1.2 eV1.3 eVHOMO ~5.3*SSFree molecules:leveling effect of strong coupling need to consider the system, not just isolated components* B3LYP 6-31G(d)+ experimental#16

90o dihedralHOMO-1 orbital*

Strongcoupling:graphene electrodemolecule

60o dihedral

37o dihedral

20o dihedral

0o dihedraldihedral anglebetween G9and NABSubstituents on moleculeaffect both HOMO and Fermi levels in strong coupling limit*Gaussian 03, B3LYP/6-31G(d)Where does the electrodestop and the molecule begin?Cu#17Something special about molecular tunnel junctions:

electron transit time 15 fsec (15 x 10-15 seconds)maximum frequency > 10,000 GHzwhatever we can do with tunnel junctionsshould be VERY fast5 nm+1 V0 Vemolecules#18a problem with tunneling:

J (A/cm2) = K exp(-d)

so far, we are really doing barrier electronicswe need to seriously reduce if molecular electronics is to be broadly practical = 9 nm-1 (alkane)

= 3 nm-1 (aromatic)

= 1 nm-1 (??) = 0.03 nm-1 (??)00.10.20.30.40.50.60.70.80.91012345distance, nm9310.03probability#19

= 8.7 nm-1 metalliccontactsEfermie (electron tunnellingbarrier)h (holetunnellingbarrier)what happens beyond tunnelling?

SSBTBd#204.5 22 nm multilayerPPF (sp2 carbon)Aue-beam carbonVSSSSSSSSSSSSSSSSSSSSSSSSSSSScollaboration with: Maria Luisa Della Rocca, Pascal Martin, Philippe Lafarge, Jean-Christophe Lacroix, University of Paris

BTB (bis-thienyl benzene) in all-carbon junctionmetalliccontacts-4.8-5.3-1.5 eVEnergy vs. vacuum, eVSS#21

300 K4.55.08.0132216

4.55.08.010.513227.016300 Kthickness, nm:tunnelling? unlikely across 22 nmnote curvature, notsimple exponentialconventional wisdom: activated hopping by redox exchange for d > 5-10 nm#22

4.5 nm8.010.522 15 nm, T>200 K Arrhenius slope=160 meV (>200 K)0.3 meV (5-50 K)(bulk polythiophene:130 280 meV)#24

300 K4.5 nm8.010.522

< 10K4.5 nm22 10.58.0the usual suspects: characteristic plot:

Fowler Nordheim (field emission) linear lnJ vs 1/E (E= electric field)variable range hopping linear lnJ vs T-1/2 or T-1/4Schottky emission (i.e. thermionic) linear lnJ vs 1/Tredox hopping linear ln J vs 1/Tspace charge limited conduction linear J vs V2

Poole-Frenkel transport between traps linear ln (J/E) vs E1/2 R2 = 0.9989(note: controlled by electric field, not thickness. certainly NOT tunneling)

#25

(trap-Poole-Frenkel transport between coulombic traps:trape-V=0V < 0'trape-trap depth decreases with increasing electric fieldV 106NAB, AB, etc#31