Electron Transport in Strongly Coupled Molecular Electronic Junctions

Transcription

1 Electron Transport in Strongly Coupled Molecular Electronic Junctions Richard McCreery, Adam Bergren, Sergio Jimenez Bryan Szeto, Jie Ru, Andriy Kovalenko, Stan Stoyanov University of Alberta National Institute for Nanotechnology

2 National Institute for Nanotechnology University of Alberta, Edmonton, Alberta, Canada Founded 2002 Building dedicated 2006 $$$: National Research Council NSERC (Canada) Alberta Ingenuity Fund Canada Fund for Innovation

3 Molecular electronics molecular junctions: Note: molecules become circuit elements critical dimension is 1-10 nm Today s question: How are electrons transported through molecules??

4 Electrochemical reduction: Molecular junction: E vs. NHE -1.0 ionic double layer molecule in solution E= E=0 V distance from electrode LUMO HOMO energy relative to vacuum, ev distance metallic contacts E fermi metal electrode two electrodes, no double layer, no solution

5 Two common electron transport models: off-resonant resonant LUMO is close (~within kt) to Fermi level energy relative to vacuum, ev e - distance metallic contacts LUMO e.g. tunneling, Schottky, field emission - E fermi HOMO e - metallic contacts e.g. resonant tunneling, orbital mediated tunneling - E fermi HOMO

6 The scientific question: How are electrons transported through 1-5 nm thick molecular layers? Outline: fabrication of molecular junctions characterization electronic properties transport mechanism

7 Things we do differently from everyone else: V slow electron beam deposition of Cu top contact, often covalently bonded to molecule conjugated, partially ordered mono- or multilayer, 1-5 nm thick, molecules in parallel Au Cu covalent C-C surface bond, stable to > 500 o C very flat (< 0.5 nm rms) graphitic carbon substrate [Pyrolyzed next slide Photoresist Film, PPF, essentially metallic, with ρ=0.006 Ω-cm] sp 2 carbon Phys. Chem. Chem. Phys, 2006, 8, 2572 J. Chem. Phys. 2007, 126, J. Phys. Cond. Matter, 20, (2008)

8 SEM TEM 1 µm molecular layer (not resolved) EM mount metal polypyrrole carbon 20 nm

9 Microfabricated E-chips PPF Echip 4 wafer PPF Echip Clip electrode used for electrochemistry cut after molecule deposition PPF leads Junction next slide Bryan Szeto Jie Ru

10 500 µm current amplifier Au strip 1 mm Junction area: 2.5 x 2.5 µm to 400 x 400 µm V sense V - +

11 Cross section of a PPF/NAB/Cu/Au junction (SEM) Left side Right side SiN Cu/Au Cu/Au SiN PPF SiO 2 Molecules (not resolved) Molecules (not resolved) PPF SiO 2 Si Si

12 Schematic of junction structure: (not to scale) Au 30nm Cu 120nm SiN 70nm SiO 2 50nm PPF 1µm SiO2 SiN Cu/Au molecules PPF SiO 2 Si SiN SiO2 to scale: Cu/Au SiN/SiO 2 > 100 nm PPF 1-5 nm molecular layer is really thin compared to metals, does it survive metal deposition??

13 backside Raman of buried interface: Au N =N NO2 Cu ~4 nm NAB ~10 nm PPF ~50% transparent (~ 4 nm thick multilayer) Quartz substrate (0.13 mm) nm laser excitation at 45 Collect Raman scattered light normal to substrate

14 Intensity(counts/30s) Raman intensity (30 sec, 19 mw) Quartz/PPF/NAB/Cu/Au (after metal deposition) =N N NO Quartz/PPF/NAB, no Cu or Au Raman Shift (cm -1 ) Adam Bergren Amr Mahmoud (Monday, 4:20 PM)

15 FTIR of buried interface: NAB on Au/Ti - log (R/R o ) Au/Ti NAB Ti Au wavenumber, cm -1 Ti is primer layer for NAB bonding to Au after 100 nm Au deposition: - log (R/R o ) Au IR transparent Si wavenumber, cm -1 IR beam Adam Bergren Amr Mahmoud (Monday, 4:20 PM)

16 Electronic behavior: V sense current amplifier Labview V - + carbon/molecule = fluorene 12 nm 20 nm nm 2 μm Au Cu V carbon on SiO 2

17 20 Start with something familiar: PPF/SiO 2 /Cu current density, J, A/cm Cu PPF PPF/Cu slope= 1.6 Ω Control FL-SiO2 V 10 nm Cu SiO 2 V V, carbon relative to Au J, A/cm slope= 420,000 Ω breakdown ~ 3 MV/cm

18 How about a molecule instead of SiO 2? nitrobiphenyl multilayer 4.6 nm thick: 3 Cu NO 2 J, A/cm NBP film much more conducting than SiO nm PPF NO2 NO Au SiO 2 V, carbon relative to Au or Cu

19 Nitrobiphenyl junctions of differing thickness: V 7 A/cm 2 ~ 10 6 e - /sec/molecule nm 2.8 nm 4.6 nm Au NBP 2 1 J, A/cm NBP4.5 FL-SiO2 rsd 5-15% 2 NBP 2.8 nm NBP 1.6 yield > 90% typical error bar V J. Phys. Chem. B, 2005, 109, 11163

20 3 2 Same data, on a log scale: NBP(1.6 nm) NBP(2.8 nm) ln J, A/cm V V, V ln(j, 0.1 V) No obvious shape change from 1.6 to 4.6 nm thickness Symmetric with minimal hysteresis Repeatable > 10 8 cycles NBP(4.6 nm)

21 -7 β = 0.22 Å -1 NO2 d Literature: β alkanes (echem or junctions) J = B e -βd 0.8 Å -1 aromatic (echem, 1999) 0.22 conjugated (echem, SAM) 0.3 to 0.6 Ln I (V=0.1 V), A polyene* (2005) 0.22 oligothiophene* (2008) 0.1 oligoporphyrin* (2008) 0.04 oligophenyleneimine + (2008) 0.3 * single molecule junction + ~100 molecule junction d= thickness, nm β = 0.21 Å -1 β is smaller for aromatic structures (i.e. conjugated molecules are better conductors )

22 Is this behavior molecular? (1.5 nm) NO2 (2.0 nm) (1.4 nm) j (A cm 2 ) no molecule strong effect of structure and thickness on conduction (1.9 nm multilayer) NO2 (3.7 nm multilayer) V (PPF relative to Au) Adam Bergren, J Phys. Cond. Matt. 2008, 20,

23 Various transport mechanisms: Weakly Temperature dependent: Coherent tunneling, "superexchange" Incoherent, diffusive tunneling Field emission (Fowler Nordheim) distance dependence: exp(- β d) exp(- β d) (V/d) exp(- a d) Strongly Temperature dependent ( activated ): Thermionic (Schottky) emission exp (-c d 1/2 ) Poole-Frenkel effect ( coulombic traps") hopping, including redox exchange (Marcus-Levich) exp (-c d 1/2 ) d -1

24 A good probe of mechanism: Temperature dependence activated 250 K K 325 K 100, 120, 150 K J, A/cm temperature independent -1.0 j (0.2 V) V J. Phys. Cond. Matter, 20, (2008)

25 ln j 0.2 V (A cm 2 ) K 20 K Arrhenius plots 10 K BP(1.4) FL(1.8) NAB(3.3) AB(3.2) (1.4) T 1 (K 1 ) 5 K (1.8 nm) N= N NO 2 (3.3 nm) AJB 25

26 mev (AB) 40 mev (NAB) 200 K 100 K T 1 (K 1 ) 26 µev 56 µev 50 µev N =N NO 2

27 j (A/cm 2 ) T = 5 K (T independent, K) N =N NO (4.5 nm multilayer) V 0.01 need to explain: apparent ohmic contact T-independent, despite: 4.5 nm thick (too thick for tunneling) j (A/cm 2 ) V

28 All common off-resonance tunneling mechanisms fail: J (A/cm2) PPF/NAB (2.6nm )/Cu Simmons with image charge NAB experimental (2.6 nm) Applied Voltage (V) Simmons, φ = 0.85 ev (i.e. tunneling through a rectangular barrier) Field emission (Fowler Nordheim)

29 Tunneling with reduced electron mass (modified Simmons equation): J ( 2m *) q qv 2 2m * qv qv 2 qv = φ exp φ d φ + exp φ + d 2 2 4π ηd 2 η 2 2 η 2 m* = effective electron mass, where mass of charge carrier = m* x 9.1 x kg NAB(3.3), experiment Simmons Model d= 3.3 nm NAB(4.5), experiment Simmons Model d= 4.5 nm Φ = 1.1 ev m* = 0.4 m e Φ = 1.1 ev m* = 0.4 m e Experimental data collected at 5 or 6 K off resonant tunneling just doesn t work Simmons, J. G. Journal of Applied Physics (1963), 34,

30 An alternative approach: we expect these levels to be broadened, by: electronic coupling to substrate intermolecular interactions variable bonding geometry uncertainty (i.e. lifetime) broadening Some evidence for broadening: ev vs vacuum PPF(-4.9)* NAB LUMO (-3.0 ev, from DFT) Cu (-4.7)* HOMO (-6.6, DFT) NO2 N =N Optical Absorbance NAB in hexane (X.02) NAB bonded to carbon NO2 *experimental, from Kelvin probe N =N wavelength, nm Appl. Spectros. 2007, 61,

31 DFT with periodic boundary conditions for graphene E, ev NAB NAB-graphene -2 LUMO -4-6 HOMO Stan Stoyanov Kirill Koshelev Andriy Kovalenko NINT

32 HOMO and LUMO energies vary with torsion angle

33 Modeling of both contacts and molecule E, ev -2 C_NAB_Au C_NAB_Cu strong interaction of both Cu and graphene with NAB LUMO -4 HOMO -6 Koshelev

34 Zero bias: positive bias: a range of orbital energies -2 energy relative to vacuum, ev LUMOs HOMOs E fermi next slide distance filled states in metal

35 once + bias creates empty metal orbitals, electrons can leave HOMO V=0 + bias e - + HOMO fills again from negative electrode, effectively hole transport e - - Note that more HOMOs become accessible for higher bias, causing upward curvature HOMOs

36 N =N NO2 90 sech distribution σ = 0.31 ev E f E HOMO = 1.7 ev N chan = 10 5 more HOMOs becoming accessible with increased bias nm J (A/cm2) gaussian distribution σ = 0.52 ev E f E HOMO = 1.7 ev N chan = Applied Voltage (V) observed σ = half width of orbital energy distribution E f E HOMO = Fermi level to orbital offset N chan = total number of active channels Sergio Jimenez Adam Bergren

37 N =N NO2 Ln (J) PPF/NAB (2.6 nm)/cu sech observed gaussian Applied Voltage (V) Main parameters of the model: E f E HOMO HOMO linewidth (σ) and number of channels (N) molecular layer thickness (d) for the electrochemists: NOT Marcus/Butler-Volmer; similarity due to distribution of orbital energies rather than thermal fluctuations

38 Important notes: broadening caused by coupling and local environment, not thermal fluctuations main parameters are distribution width (σ), energy offset (E F E HOMO ), and thickness σ LUMO energy overlap of metal and molecule orbitals may (and probably does) occur at zero bias E f E F -E HOMO depending on offset between molecular orbitals and Fermi level, we can greatly vary conductance HOMO The punch line: strong interactions between molecule and contacts result in resonant electron transport rather than classical tunneling

39 Adam Bergren (NRC) Sergio Jimenez (visit. prof.) Andriy Kovalenko Stan Stoyanov Kirill Koshelev Jie Ru (Uof Alberta) Bryan Szeto NINT Also : Rory Chisholm (2:00 PM Monday) Mark McDermott