Molecules in Circuits, a New Type of Microelectronics?

Transcription

1 1 Molecules in Circuits, a New Type of Microelectronics? Richard L. McCreery University of Alberta National Institute for Nanotechnology Edmonton, Alberta, Canada

2 Organic Electronics : ony organic LED display charge transport by hopping across 100 nm 10 μm distances carriers are radical ions and/or polarons, often low stability 2 What happens when charge transfer distance is decreased to 1-25 nm, and electric fields may exceed 10 6 V/cm?

3 Our substrate: Pyrolyzed Photoresist Films (PPF) 20 µm commercial photoresist (novolac resin) oft Baking 90 O C Clean i Wafer 6000 rpm pin-coating Optional Lithography 50 µm PPF (roughness < 0.4 nm rms) Pyrolysis Gas Flow 50 µm (1000 O C, 1 hour) 95% N 2 ;5% H 2 3 Kim, ong, Kinoshita, Madou, and White, J. Electrochem. oc., 1998, 145, 2315 Ranganathan, McCreery, Majji, and Madou, J. Electrochem. oc., , Ranganathan, McCreery, Anal. Chem, 2001, 73,

4 urface modification via diazonium reduction: R NH 2 HNO 2 e - R N e CH 3 CN R + N 2 R R PPF electrode sp 2 carbon surface bond stable to > 500 o C high coverage very low in pinholes prone to multilayer formation. 4 Delamar, M.; Hitmi, R.; Pinson, J.; aveant, J. M.; J. Am. Chem. oc. 1992, 114, Liu, McCreery, J. Am. Chem. oc., 1995, 117, Belanger, Pinson, Chem. oc. Reviews 2011, 40, 3995.

5 Packaged: V FIB/TEM: modified electrode Au Ox Red Cu e-carbon 10 nm Cu silicon (for contrast) i 1-6 nm sp 2 carbon Au e-c BTB PPF 5 Adv. Mater. 2009, 21, 4304 J. Phys. Chem. C 2010, 114, J. Am. Chem. oc. 2011, 133, 19168

6 6 Pyrolyzed Photoresist Film (PPF) microfabrication: 1. lithography 2. pyrolysis 100 mm PPF structure and conductivity similar to glassy carbon roughness < 0.5 nm rms by AFM

7 7 PPF Echip 4 wafer molecules attached to PPF by electrochemical reduction of diazonium ions Cu/Au deposited through shadow mask next slide PPF leads Junction Magnification = 1 Mag. = 10 Mag. = 10

8 Mag. = 60 cleave through junction region, then EM of edge FIB/TEM: Mag. = 5,000,000 TEM silicon (for contrast) 5 nm Au Cu sp 2 carbon Cu/Au io 2 PPF io 2 Cu/Au 100 nm io 2 i PPF Molecules (not resolved) 8 1 µm junction region Mag. = 30,000 Mag. = 300,000

9 9 near Jasper, Alberta

10 J(A/cm 2 ) J(A/cm 2 ) 80 x 80 μm junction 2 no molecule O N O 1 1 ma 32 junctions N N NAB (nitroazobenzene) 0-1 NAB 0.5 V V frequency independent, 0.01 to 10 5 Hz symmetric can scan > 10 9 cycles without change survived 150 o C for > 40 hrs looks like 10 Marcus kinetics -2 Is it? 10-3 log scale 10-4 exponential above ~0.2 V AC Applied Materials & Interfaces 2010, 2,

11 J (A cm -2 ) j (A/cm 2 ) 15 Arrhenius: 10 5 T = 5 K 37 mev mev V molecular layer thickness 2.2 nm 2.8 nm 3.3 nm 5.2 nm 4.5 nm V (V) d = 2.2 nm 2.8 nm 3.3 nm 4.5 nm 5.2 nm 1000/T, K -1 not activated for T<200 K strong thickness dependence NOT activated electron transfer, so what is it? J. Phys. Chem C, 2010, 114, 15806

12 ln J (0.1) Tunneling: molecules, insulator, etc sp 2 carbon e Cu tunneling barriers: ϕ electron LUMO electron density vacuum d eˉ E Fermi J (A/cm 2 ) = K exp(-βd) ϕ hole HOMO 4 2 metal, β ~ 0 12 vacuum: β ~ 25 nm alkane, β = 8.6 nm -1 NAB: β = 2.5 nm d, nm now, vary HOMO and LUMO energies

13 13 More molecules: BTB O N O 2.4 ev range of HOMOs, 2.3 ev range of LUMOs NO EFFECT! N N also: biphenyl, nitrophenyl, ethynyl benzene, anthraquinone (9 molecules, >400 junctions) NH N N β = 8.7 nm -1 ayed, Fereiro, Yan, McCreery, Bergren; PNA (2012) 109,

14 Intensity (a.u.) Energy levels of PPF/molecule from Ultraviolet Photoelectron pectroscopy: vacuum e e Br WF hv=21.2 ev WF= CH C PPF NO 2 E F - E HOMO E Fermi WF, ev * E F - E HOMO Binding Energy (ev) 16.1 PPF (bare) 4.53 biphenyl PhCCH PhBr PhNO NAB mean for 8 aromatics= 1.3 ± 0.2 ev C aliphatic= 2.0 ± 0.1 ev 14 * method of Kim, Choi, Zhu, Frisbie, JAC 2011, 133, 19864

15 Energy (ev vs. Vacuum) -2-3 Free molecule HOMOs: Molecules bonded to PPF: BTB 1.3 ± 0.2 ev (aromatics) -6-7 PPF AB PhBr NAB ± 0.1 ev NC 8 H 17-8 NP NC 8 H 17 UP and immons barriers are consistent, and correlate with observed β values 15 ayed, Fereiro, Yan, RLM, Bergren, PNA, 109, (2012)

16 Why? leveling effect of strong coupling need to consider the system, not just isolated components WF ~ ev WF ~ ev 4.6 ev ev Free molecules: 0.7 ev HOMO ~5.3* 2.0 electron donating 4.6 ev PPF (unmodified) N N HOMO ~6.6* O N O 1.3 ev WF ~ ev electron withdrawing N N O N O 16 * B3LYP 6-31G(d) + experimental

17 trong coupling: graphene electrode HOMO-1 orbital* 0 o dihedral o o dihedral dihedral angle between G9 and NAB molecule ubstituents on molecule affect both HOMO and Fermi levels in strong coupling limit Cu 17 Where does the electrode stop and the molecule begin? *Gaussian 03, B3LYP/6-31G(d)

18 j (A/cm 2 ) omething special about molecular tunnel junctions: nm V eˉ molecules electron transit time 15 fsec (15 x seconds) maximum frequency > 10,000 GHz 0 V +1 V whatever we can do with tunnel junctions should be VERY fast 18

19 19 probability a problem with tunneling: J (A/cm 2 ) = K exp(-βd) β = 9 nm -1 (alkane) β = 3 nm -1 (aromatic) β = 0.03 nm -1 (??) β = 1 nm -1 (??) distance, nm we need to seriously reduce β if molecular electronics is to be broadly practical so far, we are really doing barrier electronics

20 20 BTB ϕ e (electron tunnelling barrier) ϕ h metallic contacts E fermi (hole tunnelling barrier) d β = 8.7 nm -1 what happens beyond tunnelling?

21 BTB (bis-thienyl benzene) in all-carbon junction Energy vs. vacuum, ev Au e-beam carbon -1.5 ev V nm multilayer metallic contacts PPF (sp 2 carbon) 21 collaboration with: Maria Luisa Della Rocca, Pascal Martin, Philippe Lafarge, Jean-Christophe Lacroix, University of Paris

22 22 thickness, nm: tunnelling? unlikely across 22 nm 300 K note curvature, not simple exponential 22 conventional wisdom: activated hopping by redox exchange for d > 5-10 nm 300 K

23 nm J, A/cm K V 4.5 nm 8.0 weakly activated, if at all, not consistent with redox exchange <10 K 23

24 ln J (1V) What about β, the attenuation coefficient? 300 K 10 β = 3.0 ± 0.3 nm -1 (N=8) coherent tunnelling for d< 8 nm 5 0 β = 1.0 ± 0.2 nm -1 (N=19) so what is this? not activated, β~ 1 nm -1-5 activated hopping, d> 15 nm, T>200 K -10 β = 0.1 ± 0.1 nm -1 (N=6) at 300K -15 6K 100K 200K Arrhenius slope= 160 mev (>200 K) 0.3 mev (5-50 K) 250K 300K d, nm (bulk polythiophene: mev) 24 Haijun Yan, et al. PNA, 110, 5326 (2013)

25 25 the 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 -1/4 chottky emission (i.e. thermionic) linear lnj vs 1/T redox hopping linear ln J vs 1/T space charge limited conduction linear J vs V 2 (note: controlled by electric field, R 2 = Poole-Frenkel transport between traps linear ln (J/E) vs E 1/2 300 K not thickness. certainly NOT tunneling) < 10K 4.5 nm 4.5 nm

26 Poole-Frenkel transport between coulombic traps : V=0 V < 0 V << 0 trap e - e - ' trap e - ' trap trap depth decreases with increasing electric field (ϕ trap - q Arrhenius slope (ln J vs 1/T) = ϕ trap -( E 1/2 πε ) 1/2 26 note: a field-dependent activation barrier

27 27 Arrhenius slope, mev q Arrhenius slope (ln J vs 1/T) = ϕ trap -( E 1/2 πε ) 1/2 trap nm nm 10.5 nm K E 1/2 E 1/2 (V/cm) 1/2 activation barrier 0 at high field

28 uppose the trap is actually an occupied molecular orbital (e.g. HOMO): BTB subunits, connected but not polarons ~0.35 V trap (i.e. HOMO) V=0 Molecular field ionization Note: field ionization in gas phase occurs at ~10 7 V/cm; our field is ~10 6 V/cm, but trap is much shallower - A bit like dry electrochemistry, but without a double layer and not activated 28 V<0 +

29 Richard Feynman, 1959 There's Plenty of Room at the Bottom β = 2.7 nm -1 β = 8.7 nm the bottom for aromatic molecular junctions? barrier electronics and not much control over barrier height

30 tunnel junctions transport distance, nm bulk semiconductors nm coherent tunnelling conjugation length?? field ionization scattering length?? resonant transport hopping, redox exchange There is plenty of room in the middle!! (low mobility, reactive species) 30 Phys. Chem. Chem. Phys., 2013,15, 1065

31 31 A preview of some new molecules: β 3 nm -1 ϕ e NAB, AB, etc 0-5 β 0.3 nm -1, E act < 30 mev metallic ϕ contacts h E fermi -10 BTB factor of >10 6 β 0.12 nm -1, E act < 50 mev -15 β 0, but thermal d, nm β 1 nm -1