Advanced Materials Characterization Workshop

Scanning Electron Microscopy (SEM) Wacek Swiech Jim Mabon Vania Petrova Mike Marshall Ivan Petrov Supported by the U.S. Department of Energy under grants DEFG02-07-ER46453 and DEFG02-07-ER46471 © 2008 University of Illinois Board of Trustees. All rights reserved.

Parallel Acquisition

Sequentioal Acquisition

Basic Comparison to Optical Microscopy

Optical

SEM (secondary electron)

The higher resolution and depth of focus available with the SEM are clearly observed, electrons have small mass (m0).

2

de Broglie : λ = h p −1

λrelat (x102 nm)

v (x107 m/s)

500

5.48

1.33

eU = m0 v 2 / 2 ⇒

1,000

3.88

1.87

hence : λnon − rel = h (2m0 eU ) −1/ 2

U (V)

5,000

1.73

4.16

10,000

1.22

5.85

30,000

0.70

9.84 (≈c/3!)

p = m0 v = h (2m0 eU )1/ 2

λrelat = h [2m0 eU (1 + eU (2m0 c 2 )−1 )]−1/2 From Scanning Electron Microscopy and X-Ray Microanalysis, Joseph I. Goldstein et al. Plenum Press

What does the SEM do? • An instrument for observing and analyzing the surface microstructure of a bulk sample using a finely focused beam of energetic electrons • An electron-optical system is used to form the electron probe which is scanned across the surface of the sample (raster pattern). • Various signals are generated through the interaction of this beam with the sample. These signals are Many Applications: collected by appropriate detectors. Most widely employed microscopy • The signal amplitude obtained at technique other then optical microscopy each position in the raster pattern is Surface morphology (SE, BSE, FSE) assembled to form an image. Composition analysis (X-EDS, WDS) Crystallography (electron diffraction and channeling techniques) Optical properties (cathodoluminescence CL) Many other more specialized applications

3

Animation from A Guide to X-Ray Microanalysis, Oxford Microanalytical Instruments

Secondary and Backscattered Electrons Energy Distribution of Emitted Electrons

Secondary electrons are low energy electrons ejected from the specimen atoms by the energetic primary beam 4

Backscattered electrons are primary beam electrons scattered back out of the sample. Adapted From L. Reimer, Scanning Electron Microscopy, 2nd edition, Springer Verlag

Electron Beam / Specimen Interactions CL UV/Visible/IR Light

EDS/WDS

Incident Ib, E =0.5-30 keV, α = 0.2-1o Beam

Characteristic Xrays

Secondary electrons (SE)

Imaging

Auger electrons

Backscattered electrons (BSE)

Bremsstrahlung X-rays

Elastic Scattering Inelastic Scattering Micrometer-size Interaction Volume

Heat Specimen Current

Imaging

Isc

Isc = Ib− ηIb − δIb = Ib (1−(η+δ)) 5

Monte-Carlo simulations of electron scattering C

15 keV Al

PMMA @ 20kV Everhart et.al. (1972)

15 keV 1 μm

Ti

Ti

5 keV

15 keV

• Determine effective lateral or depth resolution for a particular signal in a defined sample • Simulate X-ray generation / X-ray spectra in a defined sample • Simulate image contrast 6

W

15 keV

Ti

25 keV Simulations are very useful for testing if a measurement is possible or interpreting results Animation from A Guide to X-Ray Microanalysis, Oxford Microanalytical Instruments Monte Carlo Calculations, CASINO

Sequential Image Acquisition in SEM

•

•

M = Ldisplay/Lspecimen

7

Animations from, The Oxford Guide to X-Ray Microanalysis, Oxford Instruments Microanalysis Group

The scan of the electron beam and the screen raster are synchronized with intensity proportional to the collected signal Magnification is given by the ratio of the length of the line on display device to length scanned on the real sample Figure from Scanning Electron Microscopy and X-Ray Microanalysis, Joseph I. Goldstein et al. Plenum Press

Typical SEM Column / Vacuum Conditions JEOL 6060LV

Courtesy JEOL USA 8

•

An SEM specimen chamber typically operates at high vacuum conditions: Vacuum better resolution

•

Longer Focal Lengths(W2) –> larger d2, smaller α2 -> better depth of field

•

Smaller Apertures –> smaller d2, smaller α2 -> better resolution & better depth of focus

•

Correction coils are used to correct asymmetries in lens (correct astigmatism)

16

Adapted from Scanning Electron Microscopy and X-Ray Microanalysis, Joseph I. Goldstein et al. Plenum Press

Lens Aberrations / Optimum Aperture Angle Spherical Aberration Aperture diffraction causes a fundamental limit to the achievable probe size

Optimum aperture angle determined by combined effect of spherical aberration and aperture diffraction

17

Adapted from Scanning Electron Microscopy and X-Ray Microanalysis, Joseph I. Goldstein et al. Plenum Press

Chromatic Aberration

Lens Aberrations / Astigmatism Astigmatism is caused imperfections in lens or other interference. Lens by Aberrations – the Astigmatism It can be corrected using additional elements called stigmators contained inside the objective lens

Magnetostatic quadrupole lens is basis of a stigmator

Octupole lens stigmator

18

Adapted from Scanning Electron Microscopy and X-Ray Microanalysis, Joseph I. Goldstein et al. Plenum Press

Secondary Electron Detector / Imaging • •

•

Faraday Cage (collector) is usually biased a few hundred volts positive (for collection efficiency) Scintillator is biased +10kV to accelerate electrons to sufficient energy to efficiently excite scintillating material Amplified output level is directly used to set brightness (offset) and contrast (gain) in corresponding pixel in image

Everhart-Thornley SE detector

Contrast from predominately angular dependence of secondary electron yield and edge effects. Reactive ion etching of Al/Si(001) 19

Secondary Electron Yield Dependence of SE yield with angle (local) of incidence with surface

Escape probability for SE’s as a function of depth of generation Æ image resolution for SE imaging will approach the probe size 20

From Scanning Electron Microscopy and X-Ray Microanalysis, Joseph I. Goldstein et al. Plenum Press

Edge Effect on SE Yield

21

Adapted From L. Reimer, Scanning Electron Microscopy, 2nd edition, Springer Verlag

Analogy to Oblique and Diffuse Optical Illumination • Secondary electron yield is strongly dependent on local angle of incidence with beam • Backscattered electrons are also directly and indirectly detected (image is not pure SE) • Together this, along with the high depth of focus of the SEM, gives the familiar SEM images with a good perceptive sense of surface topography 22

From Scanning Electron Microscopy and X-Ray Microanalysis, Joseph I. Goldstein et al. Plenum Press

Exceptionally High depth of Focus 10,000X original magnification

Electrodeposited Gold Dendritic Structure

100,000X original magnification

Carbon Nanotube

23

Extremely Wide Range of Magnifications 12X original magnification

Miniature Sensor Device Calorimeter

24

500,000X original magnification

Sputtered Au-Pd on Magnetic Tape

Backscattered Electron Detectors and Yield Objective lens pole piece

•

• • Typical 4 quadrant solid state BSE detector

•

Solid State 4 quadrant Backscattered Electron detector placed annular to bottom of objective lens Composition image – electronically sum signal from all 4 quadrants BSE topographic images – differencing various detector quadrants

Backscattered electron yield is a strongly dependent on sample mean atomic number

Graphic from, The Oxford Guide to X-Ray Microanalysis, Oxford Instruments Microanalysis Group

25

Backscattered Compositional Contrast Secondary Electron Image

Backscattered Electron Image

SnBi alloy • most useful on multi-phase samples • can be sensitive to < 0.01 average Z differences • flat-polished specimens essential 26

Non-Conductive Specimens - Charging

Isc = Ib− ηIb − δIb = Ib (1−(η+δ)) =0 Total emitted electron coefficient η+δ as a function of beam energy When η+δ=1 Æ charge balance

Upper cross-over energy, E2, for several materials

27

From Scanning Electron Microscopy and X-Ray Microanalysis, Joseph I. Goldstein et al. Plenum Press

Variable Pressure (VP / LV / Environmental) SEM • A different and simpler solution to specimen charging of un-coated nonconductive samples is to introduce a gas (air, etc.) into the specimen chamber. • The high energy electrons ionize the gas, thus positive ions are available to dynamically neutralize any charge on the sample. • Available in both Schottky and Thermionic (Tungsten) instruments

28

@ high vacuum

Uncoated Dysprosium Niobium Oxide Ceramic

@ 20 Pa air

JEOL JSM-7000F Analytical Scanning Electron Microscope EDS

EBSD WDS

CL

• Energy-Dispersive Spectroscopy (EDS) – solid state detector simultaneously measures all energies of X-ray photons • Wavelength Dispersive Spectroscopy (WDS) – sequentially measures intensity vs X-ray wavelength (energy). Superior energy resolution and detection limits (P/B ratio). • Electron Backscattered Diffraction (EBSD) – acquires electron diffraction information from surface of highly tilted bulk sample with lateral resolution of 10’s of nm • Cathodoluminesence (CL) – optical emission spectrometer and imaging system for 3001,700nm. Liquid He cooled stage module. 29

Characteristic X-ray Generation • A scattering event kicks out an electron from K,L,M, or N shell of atom in specimen • An electron from an outer shell falls to fill in the vacancy • Energy difference results in release of an X-ray of characteristic energy/wavelength or an Auger electron X-ray vs. Auger Generation

30

Graphic from, The Oxford Guide to X-Ray Microanalysis, Oxford Instruments Microanalysis Group

Atomic Number

Energies of Characteristic X-rays to 20 keV

From Scanning Electron Microscopy and X-Ray Microanalysis, Joseph I. Goldstein et al. Plenum Press

31

Energy Dispersive X-ray Detector: Si(Li)

Copper Rod (at Liq. N2 Temperature)

FET charge sensitive amplifier

Si(Li) Detector X-ray Window Magnetic Electron Trap Collimator

32

Graphics from A Guide to X-Ray Microanalysis, Oxford Microanalytical Instruments

Voltage

Mechanism of X-ray Energy Determination Voltage step X-ray event

Charge restore

Time

Animations from, The Oxford Guide to X-Ray Microanalysis, Oxford Instruments Microanalysis Group

•

• • • 33

X-ray loses energy through inelastic scattering events creating electron / hole pairs High voltage bias keeps generated pairs from re-combining Charge sensitive amplifier “counts” pairs generated by X-ray Spectrometer calibration effectively multiplies by energy/pair (3.8 eV) to determine X-ray energy

EDS Spectral Resolution and Count Rates Pulse processing time constants are used to adjust available count rate versus spectral peak resolution

• Long time constants are best for single acquisition analysis for best energy resolution. • Short time constants are best for fast acquisition of X-ray elemental maps (elemental distribution images) or line-scans (intensity or concentration profiles). • Compromise is often needed. 34

X-ray EDS Microanalysis in the SEM • •

Fast Parallel Detection Qualitative elemental analysis

Oxford Instr.: Link-ISIS software

– From beryllium up on periodic table – Sensitivities to