High Resolution Imaging of Nanoscale Structures by Scanning Probe Microscopy Techniques

Transcription

1 High Resolution Imaging of Nanoscale Structures by Scanning Probe Microscopy Techniques Prof. Marco Farina, Senior Member IEEE Dipartimento di Ingegneria dell Informazione Università Politecnica delle Marche

2 Our Team Andrea Di Donato (Assistant Professor) Giuseppe Venanzoni (Research Fellow) Davide Mencarelli (Research Fellow) Tamara Monti (PhD Student) Francesco Bigelli (PhD Student) Antonio Morini (Associate Professor)

3 Scanning Probe Microscopy (SPM) A quite recent class (1981) of microscopy techniques that has improved our understanding of surfaces and materials at sub-nanometric scale. In 1986 this work earned Nobel Prize to Gerd Binnig and Heinrich Rohrer in Zurich) Today IBM labs still holds records in resolution! Imaging the charge distribution within a single molecule F. Mohn et al, Nature Nanotechnology 7, (2012) Imaging of naphthalocyanine (left) and DFT-model (right)

4 Scanning Probe Microscopy (SPM) In all cases a probe is scanned in close proximity of the surface of the sample (or vice-versa) and, depending on the type of probe, variations of some physical parameter arising from the interaction between surface and probe are recorded Generally piezoelectric membranes are used to displace the sample (or the probe) at sub-nanometric scale: membranes are driven by some feedback system SPM techniques may achieve atomic resolution at room temperature without need for vacuum!

5 Scanning Probe Microscopy (SPM) In the Atomic Force Microscope (AFM) in contact mode, the device recovers the sample topography by a measurement of the deflection of a mechanical sharp tip, when the latter "touches" -in some sense- the sample surface. The tip is a few atoms at its edge The deflection is detected by means of a laser beam Images can be obtained by processing either the lateral or the normal deflection

6 Atomic Force Microscopy Credits animation: J.C. Bean, University of Virginia Virtual LAB

7 Atomic Force Microscopy In the semi-contact mode the tip oscillates nearby its mechanical resonance; the interaction with the sample modifies amplitude and phase of the mechanical oscillation (there is also a frequency shift) Useful for softer materials, such as polymers and bio-organic samples; phase allows to detect material inhomogeneity In the non-contact mode, still the tip oscillates, but it is farther Image: courtesy NT-MDT

8 Conducting Atomic Force Microscopy Also called spreading resistance microscopy: a bias is applied and the current recorded via a conducting tip. The current is proportional to the sample local resistivity It may be not as easy as it seems: -in air there is always a water meniscus, -there is chance to damage or contaminate the conductive coating during scans -the measurement depends on the contact area, and hence the landing conditions Image: courtesy NT-MDT

9 Techniques derived from AFM: Electric Force Microscopy (EFM) Using a conductive tip: there are many versions; e.g. in a common semicontact two pass technique the first pass recovers the topography. During the second pass the cantilever is driven at a given distance and following the surface profile, while it oscillates at resonant frequency and cantilever is biased. Capacitive tip-sample electric force (actually its derivative) leads to resonance frequency shift. Variations of the capacitance (Scanning Capacitance Microscopy SCM), or the surface potential distribution can be imaged by reporting variations in the oscillation amplitude Image: courtesy NT-MDT

10 Techniques derived from AFM: SCM Alternative (typical) implementation of SCM: Electrode (SCM probe) X-y scanner Input coupled line Transmission line resonator Output Coupled line Resonator Varactor Variations of the capacitance are detected as frequency shift of a microwave resonator In any case the above measurements are qualitative and differential (variations are shown) Microwave source (around 1 GHz) Detector

11 Other AFM-related Techniques Kelvin Probe Microscopy: measurement of contact potential difference between tip and sample; often a two-pass technique and a static potential is applied by a feedback system to keep the system in equilibrium. This potential maps the contact potential Magnetic Force Microscopy: Investigation of magnetic domains; several possible modes even in this case and many more Image: courtesy NT-MDT Actually a good deal of confusion arises in classifying the large number of possible modifications in the original SPM (e.g., contact, semi-contact, non-contact, tapping mode.)

12 ...AFM manipulation AFM tip can be used to move object at nano-scale and to perform some lithography (below some example at our Dept. [DIBET] old name: the whole text is 2mm wide; thickness 10nm)

13 A different approach: the Scanning Tunneling Microscopy (STM) Historically, the first SPM technique Exploits the tunneling current between a sharp tip and a conducting sample Sharp tips are obtained by simply cutting a conducting wire or by electrochemical etching Very sensitive as tunnel current is exponentially dependent from the distance! Virtually tunnel current occurs just at level of a single tip atom (or just a few); hence no problem related to the curvature radius of the tip (convolution) : atomic resolution! In AFM in fact the height image is a convolution between tip shape and surface geometry

14 the Scanning Tunneling Microscopy (STM) Credits: animation: J.C. Bean, University of Virginia Virtual LAB

15 Notes The convolution arising in AFM imaging is not necessarily a dramatic issue: by scanning a known profile, the tip profile can be estimated, and images can be processed by deconvolution STM on other hand can be used for other kind of measurements, such as the surface Density of States (DoS) by measuring the tunneling voltampère characteristic Image: courtesy NT-MDT

16 STM: imaging at atomic scale HOPG (graphite) surface as seen in our lab by the NT-MDT P-47 microscope Carbonium lattice

17 Microwave Imaging: basic principles Near-Field Scanning Microwave Microscopy (SMM): a first successful realization dates back to 70s[1] while the idea is credited to E.H. Synge [2], 1928 Microwaves are used to resolve sample details well below the Abbe's barrier, namely the wavelength limit; this is done by exploiting evanescent (near field) microwave fields interacting with probe and sample The general line of reasoning: A sharp tip fed with microwave signal, having curvature radius R 0 <<l, will generate evanescent waves with wave vector k in the order of 1/ R 0, hence rapidly decaying from the metal tip and giving resolution power in the order of R 0, in spite of the wavelength..actually the resolution can be much higher than R 0 (shared some principles with synthetic radar) Main issue: results are a convolution of effects due topography and local composition of the sample; difficult interpretation of results [1]E. A. Ash and G. Nicholls, "Super-resolution Aperture Scanning Microscope", Nature, vol. 237, pp , June 1972 [2] E.H. Synge, "A suggested method for extending microscopic resolution into the ultra-microscopic region", Phil. Mag , 1928

18 Microwave Imaging: basic principles Scanning Microwave Microscopy Aperture microscopes: a miniature waveguide is used to generate evanescent fields Pros: near field interaction may dominate (shielding) Cons: usually resolution limited (micrometric) Apertureless microscopes: a sharp tip excites quasi-singular fields Pros: simple, achieves sub-nanometric resolution Cons: tip radiates: strong coexistence of local and non-local interaction with sample makes somehow harder quantitative measurements and interpretation of data

19 Microwave Imaging: Aperture vs. apertureless Aperture Vs Apertureless waveguide coax aperture sample Parallel strip TL sample coax cantilever shield sample coax sample STM tip sample

20 Microwave Imaging: basic principles Scanning Microwave Microscopy Common question: How can we use microwaves (centimetric wavelengths) to image nanometeric features? Answer: we use electromagnetic fields to locally couple the probe and sample when they are very close (reactive interaction), not as radiated rays. The Abbe s barrier (wavelength diffraction limit) does not apply BUT: radiated fields still exist. Usually both local and non local interaction occur, making difficult data interpretation.

21 Microwave Imaging: basic principles Scanning Microwave Microscopy Generally SMM is associated to either AFM or STM. AFM or STM are used to control the tip-to-sample distance (for example: Agilent implementation uses AFM) There are important exceptions: in 90s group leaded by Weiss used STM tip/sample junction to generate microwave harmonics, using the microwave signal directly in the feedback controlling the tip distance. Important result: possibility to use STM also on insulating specimens In the most common implementation the SMM tip is part of a microwave resonant structure: they work at a fixed frequency and frequency shift is recorded by a PLL Pro: enhanced sensitivity, simpler quantitative measurements Cons: narrow-band; microwave spectroscopy not possible

22 Coax. Cable Microwave Imaging: Our Approach Our Software: Vector Network Analyzer VNA for Microwave signal source and reflection measurement: Ethernet Cable Conductive Pt/Ir tip Tunnel current - STM-SMM synchronization - Data processing: new algorithms for broadband processing. We process the complex reflection coefficient, not the resonant frequency shift GHz (PNA E8361A) - max Dynamic range 120 db XYZ piezo Sample SPM controller/feedback Voltage source STM/AFM feedback control and spatial resolution (Nt-MDT Solver P 47)

23 Microwave Imaging: Our Approach

24 Pros and Cons of STM Possible Atomic Resolution, quite easily, in ambient conditions Currently no longer restricted to conducting surfaces: very sensitive current amplifiers (<1pA) available, so that also measurement on biological samples is possible. possible to reduce parasitic interaction between tip and sample, owing to the typical shape of the tip. Minimum the piezo cross-talk effect. AFM tip STM tip STM is difficult over relatively large areas, or over dishomogeneous regions The STM information is never purely topographic Need to be careful in reducing interaction between microwave signal and STM electronics

25 Microwave Imaging: Our Approach Our implementation We have not inserted resonators: broadband (0-70GHz) Of course the tip is in any case a resonator, the mismatched cable is a resonator, the cavity is a resonator (generally broadband implementations involved aperture probe) Consequence: the sensitivity will be frequency dependent. Our broadband statement refers to data collection and manipulation rather than to a specific hardware implementation! Data acquisition: We use a Vector Network Analyzer in the STM-assisted system (usually VNAs in literature have been used in AFM-assisted systems). VNA allows unpaired dynamics and extremely broadband

26 Microwave Imaging: Our Approach Data Processing: Frequency In many frequency regions the sensitivity will not be sufficient, owing to the lack of resonances. How to improve sensitivity? The idea is straightforward: a sample imaged at different (eventually close) frequencies shows the same features but with different amount of noise; one can extract common features among images obtained at different frequencies This can be done by performing cross-correlation between images at different frequencies, or simply by normalizing and averaging images (M. Farina et al. IET Electronic Letters Jan 2010)

27 Example: HOPG Microwave Imaging: Our Approach to HOPG SMM before frequency processing (20.35GHz) SMM after frequency processing ( GHz) STM

28 Screen-shot of our software

29 A gallery of SMM results: Integrated circuit STM SMM

30 A gallery of SMM results: Example: Calibration grating chalcagenid glass, with gold surface and aluminum sublayer; the pattern height is 30 nm, period 278 nm STM 22.8GHz

31 A gallery of SMM results: C2C12 mouse muscle cell Example: Fixed C2C12 muscle cell STM (1pA, 8V) SMM (X band)

32 Example: Fixed C2C12 muscle cell (zoom) A gallery of SMM results: C2C12 mouse muscle cell (detail)

33 Example: Fixed C2C12 muscle cell (further zoom) A gallery of SMM results: C2C12 mouse muscle cell: detail

34 Quantitative TIME DOMAIN? measurement: calibration (IEEE MTT 2011, M. Farina et al.) What we measure is at the input of a box error, defining all effects not related to the sample (cable, tip body etc) Sample local admittance Raw Admittance Y (e) By assuming to know three different loads, we can evaluate the error box and remove it Note: here we assume that just one port connects the error to the sample: not trivial (multimode or multipath interaction is possible). Generally multipath interaction gives poor imaging.

35 TIME DOMAIN? Our Idea: the known loads The tip edge assumed to be as a sphere, and, if the tip-sample distance is known on a ground plane, the sphere capacitance becomes the known load! C 1 h 1 C 2 h 2 h 3 > h 2 > h 1 C 3

36 Capacitance (10-16 F) TIME DOMAIN? Comparison theory/experiment Height (nm) Tip capacitance against tip/ground distance (calculated square Vs theory circle)

37 TIME DOMAIN? Time Domain?

38 TIME DOMAIN? Time Domain? Question: How can we use microwaves (tents of picoseconds) to image nanometric features? 1nm at speed of light in vacuum femtoseconds... We can borrow some concepts of time-domain reflectometry and to Fourier-transform the recorded reflection coefficient Consider an open transmission line 1.5cm long. In vacuum at the speed of light a microwave signal is reflected back to the source in 100 picoseconds A Fourier-transform of a signal 1.5cm with 20GHz as upper frequency would give a pulse of 50picoseconds (actually worse for windowing): perfectly detectable

39 TIME DOMAIN? Time Domain? This transmission line could be the probe. Any added capacitance (tip-to-sample interaction) changes the effective length of the line (as known by Hertz...) The ability to resolve a small time-shift will depend on the system dynamic range, rather than on the upper frequency of the frequency acquisition (unfortunately the time delay depends on the frequency so that the pulse is also distorted)

40 Advantages? Time Domain:why? Probably easier understanding and interpretation of what is going on It is at zero cost (just a matter of post-process!) The signal in time is Real: features sometimes hidden either in the real or the imaginary parts (or mag and phase) of the frequency domain reflection coefficient are combined in a real signal and identified more easily Idea: reflections from the region of sample under the tip vertex (closer region) can be disentangled from reflections arising from the radiated waves, as the latter should reach the probe at different times. Appropriate selection of time should allow to disentangle local and non-local probe interactions

41 A model to understand: PORT P=1 Z=50 Ohm COAXP2 ID=CX2 Di=600 um Do=2620 um L=1.36e4 um K=2.12 A= F=0.1 GHz A Model Non-local tip-to-sample and tip-to-surround interaction Local tip to sample interaction CAP ID=C1 C=55.8 ff TLIN ID=TL1 Z0=792 Ohm EL=104 Deg F0=10 GHz TLIN TLIN ID=TL2 ID=TL3 Z0=2.754e4 Ohm Z0=2.294e4 Ohm EL=3.382 Deg EL=8.575 Deg F0=30 GHz F0=30 GHz TLIN ID=TL4 Z0=1.994e4 Ohm EL=3.835 Deg F0=50 GHz TLIN TLIN ID=TL5 ID=TL6 Z0=2.294e4 Ohm Z0=2.264e4 Ohm EL=4.195 Deg EL=347.5 Deg F0=60 GHz F0=70 GHz RES ID=R1 R=7.5e4 Ohm CAP ID=C2 C=0.7 ff RES ID=R3 R=657 Ohm RES ID=R2 R=20.5 Ohm RES ID=R4 R=51.3 Ohm RES ID=R5 R=46.2 Ohm RES ID=R6 R=42.2 Ohm RES ID=R7 R=1000 Ohm CAP ID=C3 C=8.8 ff CAP ID=C4 C=13.9 ff CAP ID=C5 C=11.4 ff CAP ID=C6 C=12.4 ff CAP ID=C7 C=18.1 ff CAP ID=C8 C=18.3 ff Parameters selected to fit the measured response; number of lines depends on the frequency band

42 Comparison with measured data A model to understand: experiment vs model

43 Time domain circuit simulation A proper selection of the time interval allows to disentangle local and non local interactions Local interaction dominates Non-local interaction dominates

44 Some result SMM time transform (no correction!) Original SMM in frequency 20.35GHz)...After all Time Domain transform involves a combination of spectral data...

45 Some result: Ti Domain HOPG Animation

46 Now STM/SMM in liquid (M. Farina et al. IEEE MWCL In press vol 22, issue )

47 Advantages? STM SMM in liquid The part of the tip immersed in water may change with piezo z-displacement: cross-talk Solutions: - shield going up to water - and/or time disentangling capability! See below

48 Advantages? STM SMM in liquid (HOPG) STM SMM time 1 SMM time 2

49 AFM assisted Advantages? SMM (M. Farina et al. Applied Physics Letters November 2012 in press) We are testing a new AFM assisted SMM: easier to land and to compare with topography However piezo cross-talk is in this case relevant, owing to parasitic tip to piezo capacitance

50 Some result: Interaction nanotubes-cells (co-work with Time University Domain Animation of Trieste; University of Chieti, Dr. Tiziana Pietrangelo) AFM Microwave

51 Some result: Time Domain Interaction Animation nanotubes-cells (smaller area around the nanotube) AFM Microwave