M. Meyyappan Director, Center for Nanotechnology NASA Ames Research Center Moffett Field, CA

Transcription

1 M. Meyyappan Director, Center for Nanotechnology NASA Ames Research Center Moffett Field, CA web:

2 CNT is a tubular form of carbon with diameter as small as 1 nm. Length: few nm to microns. CNT is configurationally equivalent to a two dimensional graphene sheet rolled into a tube. CNT exhibits extraordinary mechanical properties: Young s modulus over 1 Tera Pascal, as stiff as diamond, and tensile strength ~ 200 GPa. CNT can be metallic or semiconducting, depending on chirality.

3 The strongest and most flexible molecular material because of C-C covalent bonding and seamless hexagonal network architecture Young s modulus of over 1 TPa vs 70 GPa for Aluminum, 700 GPA for C-fiber - strength to weight ratio 500 time > for Al; similar improvements over steel and titanium; one order of magnitude improvement over graphite/epoxy Maximum strain 10-30% much higher than any material Thermal conductivity ~ 3000 W/mK in the axial direction with small values in the radial direction

4 Electrical conductivity six orders of magnitude higher than copper Can be metallic or semiconducting depending on chirality - tunable bandgap - electronic properties can be tailored through application of external magnetic field, application of mechanical deformation Very high current carrying capacity Excellent field emitter; high aspect ratio and small tip radius of curvature are ideal for field emission Can be functionalized

5 High strength composites Cables, tethers, beams Multifunctional materials Functionalize and use as polymer back bone - plastics with enhanced properties like blow molded steel Heat exchangers, radiators, thermal barriers, cryotanks Radiation shielding Filter membranes, supports Body armor, space suits Challenges - Control of properties, characterization - Dispersion of CNT homogeneously in host materials - Large scale production - Application development

6 CNT quantum wire interconnects Diodes and transistors for computing Capacitors Data Storage Field emitters for instrumentation Flat panel displays THz oscillators Challenges Control of diameter, chirality Doping, contacts Novel architectures (not CMOS based!) Development of inexpensive manufacturing processes

7 CNT based microscopy: AFM, STM Nanotube sensors: force, pressure, chemical Biosensors for Astrobiology Molecular gears, motors, actuators Batteries, Fuel Cells: H 2, Li storage Challenges Controlled growth Functionalization with probe molecules, robustness Integration, signal processing Fabrication techniques Nanoscale reactors, ion channels Biomedical - in vivo real time crew health monitoring - Lab on a chip - Drug delivery - DNA sequencing - Artificial muscles, bone replacement, bionic eye, ear...

8 CNT has been grown by laser ablation (pioneered at Rice) and carbon arc process (NEC, Japan) - early 90s. - SWNT, high purity, purification methods CVD is ideal for patterned growth (electronics, sensor applications) - Well known technique from microelectronics - Hydrocarbon feedstock - Growth needs catalyst (transition metal) - Multiwall tubes at deg. C. - Numerous parameters influence CNT growth

9

10 Adsorption and decomposition of feedstock on the surface of the catalyst particle Diffusion of carbon atoms into the particle from the supersaturated surface Carbon precipitates into a crystalline tubular form Particle remains on the surface and nanotube continues to lengthen - base growth mechanism Growth stops when graphitic overcoat occurs on the growth front - catalytic poisoning C x H y -H 2 C x H y MC y M = Fe, Ni, Co, Pt, Rh, Pd and others -H 2 -H 2 C x H y C x H y Tip Growth Typically occurs when there are very weak metal-surface interactions MC y Base Growth Occurs when the metal-surface interactions are strong

11 Solution based techniques are time consuming; also hard to confine catalysts in ultrasmall patterns. Physical techniques (sputtering, laser deposition ) are much quicker, amenable for patterning.

12 Alloying a catalyst with a non-catalytic metal increases the number of reactive sites through surface clusters, as shown in a thermodynamic and kinetic study (Nolan et al, JPC, 1998). No need for any other chemical pre-treatment or preparation by ion bombardment to create particles. Allows tuning the final electrical conductivity of the structure The underlayer can play the role of a barrier between an incompatible catalyst and the substrate; an example of this is Fe on HOPG In some cases, excess and unused catalyst lifts off and catalyses a second layer of nanotubes - sort of an anomalous layered growth pattern. Use of an underlayer prevents this from happening.

13 Catalyst surface characterized by AFM (with SWNT tip) and STM. AFM image of as-sputtered 10 nm iron catalyst (area shown is 150 nm x 150 nm). Also, the same surface after heating to 750 C (and cooled) showing Fe particles rearranging into clusters. STM image of a nickel catalyst showing nanoscale particles These results are consistent with high resolution TEM showing particles as small as 2 nm.

14 - Surface masked by a 400 mesh TEM grid - Methane, 900 C, 10 nm Al/1.0 nm Fe/0.2 nm Mo

15 2 mw laser power, 1 µm focus spot Characteristic narrow band at 1590 cm -1 Signature band at 1730 cm -1 at SWNTs Diameter distribution 1.14 nm to 2 nm; consistent with TEM results High metallic % of NTs

16 - Surface masked by a 400 mesh TEM grid; 20 nm Al/ 10 nm Fe; nanotubes grown for 10 minutes Grown using ethylene at 750 o C

17 MWNT towers have been grown on a wide variety of substrates: Si, quartz, mica, HOPG Copper substrate - failure (due to alloying of Cu with other metals?) Very strong attachment of the towers to the silicon substrate (scotch tape test); HOPG surface exhibited the weakest attachment Catalyst/underlayer formulations are similar for SWNT and MWNT growth; but the growth conditions are different C for SWNT vs. 750 C for MWNT - Methane for SWNT vs. ethylene for MWNT Theoretical analysis by Karzow and Ding (Phys. Rev. B, 1999) shows that higher temperatures (900 C) and low supply of carbon favor SWNTs; the opposite favors MWNTs

18 Plasma CVD offers a variation of the well-established thermal CVD, offers additional control parameters such as power, substrate bias to exert influence on growth if possible Conventional wisdom of plasma processing does NOT seem to apply - Typically PECVD enables a lower temperature operation than thermal CVD since electrons are energetic and lead to the production of reactive species. - CNT growth is catalyst promoted and requires 600 C for catalyst activation. So, traditional cold plasma with very low substrate temperatures is not possible. So, is plasma just another way to produce species? - Maybe - If anything, excessive production of C-bearing species may lead to only MWNTs and filaments, not SWNTs - One recognized advantage is vertical alignment of nanotubes due to the electric field.

19 Inductively coupled plasmas are the simplest type of plasmas; very efficient in sustaining the plasma; reactor easy to build and simple to operate Quartz chamber 10 cm in diameter with a window for sample introduction Inductive coil on the upper electrode MHz independent capacitive power on the bottom electrode Heating stage for the bottom electrode Operating conditions CH 4 /H 2 : 5-20% Total flow : 100 sccm Pressure : 1-20 Torr Inductive power : W Bottom electrode power : W

20 Effect of changing the capacitive power from a) 20, b) 30, c) 40, and d) 50 W power Transformation from NTs to filaments occurs at W At > 50 W, always filaments a b c d Conditions: 3 Torr, 100 W Inductive power, 800 C, 20 sccm methane + 80 sccm H 2 10 nm Fe with 10 nm Al.

21 Filaments consist of stacked-cone arrangement of graphite basal plane sheets; grow with particles at the tip; hydrogen is believed to satisfy the valences at cone edges in filaments The orientation angle θ (between graphite basal planes and tube axis) increases with increasing hydrogen concentration When θ = 0, MWNTs. These are filaments with no graphite edges, requiring no valence-satisfying species such as hydrogen Nolan et al provide evidence (all thermal CVD) that material produced with CO disproportionation (without any H 2 ) was MWNTs; addition of H 2 produced filaments; as % of H 2, θ up to 30 Better to call these structures as MWNFs (filaments) instead of graphitic carbon fibers (GCFs) or vapor grown carbon fibers (VGCFs) both of which denote solid cylinders Nolan et al, JPC, B, 1998

22 Conditions: 70 W substrate power Effect of changing the inductive power from a) 0, b) 50, c) 100 and d) 200 W power There is very little or no change associated with changing the inductive power; capacitive power at 70 W is conducive to filament production. a b c d

23 Plasma grown MWNTs Grown with low capacitive power or Ar Base growth Some amorphous carbon is present on the NTs

24 Plasma grown filaments Grown with higher capacitive power Tip growth Tubes are free of amorphous carbon

25 MWNF D band centered at 1350 cm -1 Tangential G band at 1590 cm -1 Shoulder peaked around 1616 cm -1 only in filaments

26 Previous results showed that for 10 nm Fe/ 10 Al, at standard conditions + 70 W capacitive power, only MWNFs resulted. If an additional layer of 10 nm Ir or 10 nm Mo is added, then at 70 W MWNTs are formed. Higher capacitive power is needed for growth transition to filaments.

27 A 2 x 2 Π Emission spectroscopy of plasma, CH (0, 0) band near 430 nm; atomic H Balmer series (α peak at 680 nm and β peak at 486 nm); several H 2 peaks (strongest at 464 nm) MWNT growth is accompanied by a low peak intensity of atomic hydrogen. At high intensities filaments As power to substrate, filaments result; may be due to increased dissociation of H 2 from increased n e, T e at fixed inductive power and pressure Increased dilution with argon reduces concentration, as evident from the intensity, which coincides with MWNT production

28 Neural tree with 14 symmetric Y-junctions Branching and switching of signals at each junction similar to what happens in biological neural network Neural tree can be trained to perform complex switching and computing functions Not restricted to only electronic signals; possible to use acoustic, chemical or thermal signals 500 nm

29 Atomic Force Microscopy is a powerful technique for imaging, nanomanipulation, as platform for sensor work, nanolithography... Conventional silicon or tungsten tips wear out quickly. CNT tip is robust, offers amazing resolution. Simulated Mars dust H. Dai

30 Transition metal catalyst is deposited from liquid phase or sputtered on the tip of the cantilever Carbon nanotube is grown in thermal CVD or plasma reactor

31 280 nm line/space. Array of polymeric resist on a silicon substrate.

32 Non-destructive characterization for photolithography and photoresist processing - Interference Lithography SEM AFM with MWNT tip ~ 5 J/cm 2 ~ 2 J/cm 2

33 2 nm thick Au on Mica Si 3 N 4 on Silicon substrate 5 nm thick Ir on Mica

34 Red Dune Sand (Mars Analog) Optical image AFM image using carbon nanotube tip

35 DNA PROTEIN

36 p OH O O O O O O- O O OH OH OH Self-assembled Monolayer on Au-coated Tip OH OH-Monolayers on Patterned Gold Lines OH OH Si-Tip OH OH OH Phase Images Collected in ph 10 Buffer Good Phase Contrast No Contrast

37 Shortening Process by Electric Field Etching Produces COOH Terminated SWNT Tip OH O OH O - O HO OH O O O - OH OH Good Phase Contrast: Attraction Between Surface and SWNT Tip. Topography Image Phase Image Phase Image X-section

38 Our interest is to develop sensors for astrobiology to study origins of life. CNT, though inert, can be functionalized at the tip with a probe molecule. Current study uses AFM as an experimental platform. The technology is also being used in collaboration with NCI to develop sensors for cancer diagnostics - Identified probe molecule that will serve as signature of leukemia cells, to be attached to CNT - Current flow due to hybridization will be through CNT electrode to an IC chip. - Prototype biosensors catheter development High specificity Direct, fast response High sensitivity Single molecule and cell signal capture and detection

39 Electrode Electrode G T C A G C G A G T C A G C G A C A G T C G C T Immobilization Hybridization Transducing into electronic signal by electrochemical mechanism Transducing mechanism: By redox of DNA bases By redox of metal chelate indicators or other intercalators By indicator-free mechanisms based on conducting polymers (polypyrrole etc.)

40 Top View Each individual array electrode is electronically addressable dia ~ 10 to 100 nm dnn ~ 500 nm to 2000 nm Side View 10 to 100 µm Immobilized with PNA or DNA Probes

41 Design A Design B - Probes functionalized to the end of the CNT arrays - Employing metal chelate or other intercalating indicators - Probes attached to conducting polymers wrapped around CNTs - Hybridization transducing into EC signal through the change in the internal properties of conducting polymer itself Both need well-defined CNT arrays with specific density and purity.

42 Carbon Nanotube Nanoelectrode Ensemble End of the CNT Open space CNT SiO 2 Top view of the original MWCNT arrays prepared by PECVD Top view of the MWCNT arrays filled with SiO 2 by CVD of TEOS followed by mechanical polishing

43 Electrical Properties of CNT in the Nanoelectrode Ensemble Topography Deflection Current Sensing Current (na) Voltage Bias (V) Big Bundle Current (na) SiO Voltage Bias (V) Current (na) Voltage Bias (V) Individual Nanotube

44 Functionalization of DNA to CNT Solid phase synthesis - Array of Nanotubes as solid support and DNA in aqueous phase Use water soluble carbodiimide reagents H 3 C N C CH 3 N N + Cl - H CH 3 EDC: 1-ethyl-3-(3-dimethylaminopropyl) Carbodiimide Hydrochloride O HO N SO 3 Na O Sulfo-NHS: N-Hydroxysulfosuccinimide DNA O OH H 3 C N H C O CH 3 N N + O H Cl - CH 3 NaO 3 S O O N O O O N EDC sulfo-nhs H 2 N DNA CH 3 sulfo-nhs H 3 C N H O C H N N + Cl - H CH 3

45 Coupling to Cy3-labeled DNA EDC/Sulfo-NHS DNA* H 2 O/RT Naked nanotubes 1. EDC/Sulfo-NHS H 2 O Washed 2. DNA* 60 o C/1 hr. Nanotubes with spin-on film *DNA = H 2 N(CH 2 ) 6 -ACACGAGTCAGCGCAGCCATCGC-Cy3

46 Control Experiments Spin-on Glass: Same DNA coupling reaction condition and wash procedure to spin-on glass film (without any nanotube) sample exhibit no fluorescent signal. DNA not covalently attached to spin-on glass EDC and Sulfo-NHS: DNA coupling reaction without EDC and sulfo-nhs show relatively very low intensity fluorescent signal. Open ends of nanotubes contain other chemical functional groups, such as phenol and aldehyde, which can also react DNA with DNA without coupling reagents O N

47 Hybridization Test - Same assay widely use in DNA biochip EDC/Sulfo-NHS PNA* H 2 O Washed 60 o C/1 hr. Washes Hybridization c-dna-(cy3-labeled) # *PNA = H 2 N(CH 2 ) 6 -GCCGATGCACC # c-dna = CGGTACGTGG-Cy3

48 *SWNT-based logic device Inverter demonstration (Appl. Phys. Lett., Nov. 2001) by Chongwu Zhou (USC) and Jie Han (NASA Ames) V0 Vout n-type p-type Carbon nanotube Vin V DD =2.9 V V DD p VDD I DS (na) 100 V 80 DS =10 mv 60 p-mosfet V g (V) V out (V) V in (V) V i n 2.0 n 0 V V ou t 2.5 I DS (na) V DS =10 mv n-mosfet V g (V) 10

49 Carbon nanotube transistor gas sensors DNA electronics sensors Membrane Nanotube gas A + B dsdna current DNA hybridization S D S D voltage current A B voltage S dsdna D No DNA chip or sensor now can detect attomole (millions Of) DNA. The proposed DNA electronics sensors aim at 10 to 1000 DNA molecules detection for clinical in-situ DNA diagnostics

50 Molecular Hydrogen Sensor I-V curves before and after exposure to 2% Hydrogen in air: Response time: I (A) before flow 2% H 2 in air 2 minutes after 4 minutes after 6 minutes after 10 minutes after I (na) x V (V) -2 0 Positive bias: holes injected from Au or electrons injected from Pd; Negative bias: holes injected from Pd or electrons injected from Au time (s) 1200 Conclusion: For our Pd/molecule system, conduction is through the valence band. Chongwu Zhou, USC

51 When subjected to high E field, electrons near the Fermi level can overcome the energy barrier to escape to the vacuum level Fowler - Nordheim equation: Critical: low threshold E field, high current density, high emission site density (for high resolution displays) Common tips: Mo, Si, diamond Applications: - Cathode ray lighting elements - Flat panel displays - Gas discharge tubes in telecom networks - Electron guns in electron microscopy - Microwave amplifiers I = av 2 exp( bφ 1.5 /βv )

52 Needs - For displays, 1-10 ma/cm 2 - For microwave amplifiers, > 500 ma/cm 2 To obtain low threshold field - Low work function (φ) - Large field enhancement factor (β) depends on geometry of the emitter; β _ ~ 1/5r Threshold field values (in V/µm) for 10 ma/cm 2 - Mo Si P-type diamond Graphite Powder Carbon nanotubes (stable at 1 A/cm 2 )

53 Working full color flat panel displays and CRT-lighting elements have been demonstrated in Japan and Korea Display - Working anode, a glass substrate with phosphor coated ITO stripes - Anode and cathode perpedicular to each other to form pixels at the intersection - Phosphors such as Y 2 O 2 S: Eu (red), Zns: Cu, Al (green), ZnS: Ag, Cl (blue) display showing a uniform and stable image Lighting Element - Phophor screen printed on the inner surface of the glass and backed by a thin Al film (~100 nm) to give electrical conductivity - Lifetime testing of the lighting element shows a lifespan over 1000 hrs.

54 More & more components are packaged in smaller spaces where electromagnetic interference can become a problem - Ex: Digital electronics coupled with high power transmitters as in many microwave systems or even cellular phone systems Growing need for thin coatings which can help isolate critical components from other components of the system and external world Carbon nanofibers have been tested for EMI shielding; nanotubes have potential as well - Act as absorber/scatterer of radar and microwave radiation - High aspect ratio is advantageous - Efficiency is boosted by small diameter. Large d will have material too deep inside to affect the process. At sub-100 nm, all of the material participate in the absorption - Carbon fibers and nanotubes (< 2 g/cc) have better specific conductivity than metal fillers, sometimes used as radar absorbing materials.

55 Fully automated control of vehicles to enhance safety and mobility Lateral control steering to control position relative to the center of the traffic lane Longitudinal control speed & headway Original contender: Magnetic sensors together with magnetic highway markings for (lateral) + radar technology (for longitudinal) Cement paste with 0.5 vol% carbon filaments exhibits reflectivity at 1 GHz that is 29 db higher than transmittivity [ D.D.L. Chung, in Carbon Filaments and Nanotubes, NATO Science Series, Kluwer Academic 2001].

56 Electromagnetic technology is better than magnetic technology: Why? 1. Low material cost - Reflecting concrete is only 30% more than regular concrete - Still much less than concrete with magnetic strips or embedded magnets 2. Low labor cost 3. Low peripheral electronics cost (off-the-shelf oscillators and detectors) 4. Reflecting concrete exhibits better mechanical properties and lower drying shrinkage than conventional concrete; embedded magnets weaken concrete 5. Good reliability, less affected by weather as frequency, impedance and power selectivity provide tuning capability 6. High durability; demagnetization and marking detachments are not issues 7. Magnetic field from a magnetic marking can be shielded by electrical conductors (such as steel) between the marking and the vehicle, whereas electromagnetic field cannot be shielded easily.

57 Attaching chemical groups to the sidewall of CNTs to modify the properties as needed for applications - Chemical modification of the sidewall may improve the adhesion characteristics of CNTs in a host matrix to make composites - Chemical or biosensors Saturation of 2% the C atoms in SWNTs with C-Cl sufficient to change electronic band structures dramatically; done with reacting SWNTs with dichlorocarbon (Chen et al, Science, 282, 95 (1998). Fluorination of SWNTs with F 2 gas flow at C for 5 hrs. (Michelson et al, CPL, 296, 188 (1998) has been shown to attach F covalently to the sidewall Cold plasma approach to functionalization (Khare et al, NanoLett. 2, 73, (2002)