tutorial joakimreuteler fib-artifacts-and-how-to … · 'hod\hulqj /rz uhghsrvlwlrq ... d qp...

$: Tutorial JoakimReuteler FIB-artifacts-and-how-to … · 'HOD\HULQJ /RZ UHGHSRVLWLRQ ... D QP WRWDO LQFO QP DPRUSKRXV ... Tutorial_JoakimReuteler_FIB-artifacts-and-how-to-overcome-them_2017-05$
||

Tutorial

EuFN FIB workshop, Graz, July 4th/5th, 2017

Joakim Reuteler, ScopeM, ETH Zürich

FIB artifacts and how to overcome them

2016-07-05Joakim Reuteler 1

|| 2016-07-05Joakim Reuteler 2

Outline

1. Basics

2. Curtaining

3. Rippling

4. Crystal damage

5. Implantation

6. Mixing and demixing

Acknowledgements:

Samples, Material and slides provided by:

Livie Carter-Dorwling, Dr. Robin Schäublin, DinaKlimentyeva, Dr. Karsten Kunze, Yuan Xiao, Dr. JeffWheeler, Dr. Tomáš Hrnčíř, Marek Šikula, Dr. AnnaEvans, Dr. Brandon Bürgler, Dr. Henning Galinski,Tamara Popovic, Anna Hambitzer


Different types of FIBs and applications 1. Basics

Liquid Metal FIB

• Typically Ga source

• Alloy sources: Au, Bi, Si, ….

• Current range: 1 pA – 50 nA

Plasma FIB

• Typically Xe

• Current range: 100 pA – 1 uA

Applications

Cross section flat cut

TEM lamella thin & defect free

Pillar steep sidewalls, no overetch for mechanical testing

Nanostructuring no redeposition, smooth surfaces

Ga Spots on Si: 3 sec. (a) 27 nA no defocus,(b) with defocus, (c) 700 pA, (d) 40 pA

(a)(b) (c)

(d)

Xe Spots on Si: 3 sec. (a) 300 nA, (b) 10 nA, (c) 1 nA (d) 100 pAand 10 sec. (e) UHR setting

(a) (b)(c)

(d) (e)

Gas Field Ionization Source• He /Ne

Laser Cooled Ion Source• Cs, …


Scan strategy

Layer by layer

• Used for imaging

• Parallel floor

• Delayering

• Low redeposition

Parameters we control

• Dwell time

• Spacing

• Order

• Dose

Rings /following outlines

• Used for holes /pillars /tips

• Circle by circle: polishing

• Ring over ring: parallel

1. Basics

Line by line

• Used for cutting(cross sections /polishing)

• Fast removal of large volumes

• Tilted floor: shallow to deep

• Redeposition on floor and sides


Sputter yield and sputter rate

Sputter yield Y = atoms sputtered per incoming ion

Sputter rate S = volume removed per dose

1. Basics

θ

Incidention

sampleElement Sputter yield*

Y ( – )Sputter rate**

S (um3/nC)

C 1.69 0.18

Al 3.47 0.29

Si 2.78 0.24

Ti 2.28 0.46

Cr 5.15 0.1

Au 15.75 1.5

W 7.59 0.12

Sputtering by 30 kV Ga+ normal incidence Angle dependence of sputter yield **

* SRIM data from: “Introduction to focused ion beams” Lucille A. Giannuzziand Fred A. Stevie (Eds.) Springer, 2005

** “Nanofabrication using focused ion and electron beams” Ivo Utke, Stanislav Moshkalev, Phillipp Russell, Oxford University Press, 2012


Collision cascade model

Interaction of high energy ion with a solid: ion Incidence is normal to solid surface.

Ion solid interaction

• Surface atoms can leave(sputtering /milling)

• Liquid state after 1 ps *

• Relaxation and remainingdefects after 10 ps

• Implanted ions

• Amorphization

• Electrical damage zone

• 30 kV Ga+ on Si: ca. 30 nm

1. Basics

* “Introduction to focused ion beams” Lucille A. Giannuzziand Fred A. Stevie (Eds.) Springer, 2005


Milling at an edge

“Edge effect”

• Increased sputter yield (ca. 8x)

• Scan strategy matters because former pixel creates a side wall line by line milling is efficient

• Long dwell time strong side walls fast milling

Evolution of profile of line by line milling box (side view)

Interaction of high energy ion with a solidAt an edge, i.e. tangential irradiation)

1. Basics


Reminder on everyday tricks and tipps

General tricks

• Small currents for high precision

• Large currents for speed

• Defocus on high currents *

• Tilt sample into beam for perpendicular cut

• Short dwell time for high control of shape /depth

• Long dwell time for hard materials /speed

• Several passes for nanopatterning to avoid redeposition

• Work with a well tuned beam

• Observe and reflect

• Be open to new effects, they might be useful

• Literature

1. Basics

* Gierak, J. “Focused ion beam technology and ultimate applications”, Semicond. Sci. Technol. 24 (2009) 043001 (23pp)

Tilt sample into beam for perpendicular cuts


What is curtaining? 2. Curtaining

Curtaining (aka waterfall effect) • Milled face shows lines (more / less material was removed)• Irregular non flat surface: bad for quantitative image analysis, may obscure features

Porous ceramic on Ni: strong curtaining due to rough surface Porous cermet: mild curtaining due to pores

Samples by Dr. Anna Evans, Dr. Henning Galinski, NMW, ETH Zürich


Simple picture to explain curtaining

Sample topography leads to deflection of ions, thereby modulating the dose locally and thus the milling speed

Curtaining is created by spatial variation of the sputter rate of the sample and the modulation of the current density by forward scattering of the ions:

• Porous materials

• Rough surfaces

• Height steps (e.g. gates in semiconductors)

• Composites of hard and soft materials

2. Curtaining

* “Nanofabrication using focused ion and electron beams” Ivo Utke, Stanislav Moshkalev, Phillipp Russell, Oxford University Press, 2012

Ion beam gets deflected by tilted faces *


How to reduce /avoid curtaining? #1

Tricks

• Hide it:

- BSE signal

- Post processing

• Thick, uniform and dense protection cap to

smoothen surface, small currents may help

• Backside milling

- Works at an edge *

- For TEM lamellas **, ***

2. Curtaining

Samples from Dr. Henning Galinski and Dr. Brandon Bürgler, NMW, ETH Zürich

* “Nanofabrication using focused ion and electron beams” Ivo Utke, Stanislav Moshkalev, Phillipp Russell, Oxford University Press, 2012

** Simon-Najasek, M, et.al., “Advanced FIB sample preparation techniques for high resolution TEM investigations of HEMT structures", MicroelectronicsReliability 54 (2014) 1785-1789

*** Vieweg, BF, et. al., "TEM preparation method for site- and orientation-specific sectioning of individual anisotropic nanoparticles based on shadow-FIB geometry", Ultramicroscopy 113 (2012) 165-170



Tricks

• Stage rocking

• Rocking stage

Rocking stage from Tescan:Piezo moteors allow for additional tilt axis can monitor progress of removal of curtaining

2. Curtaining

Stage rocking: From (a) rotate stage 90° to get to (b) then tilt back 10° to get to (c), then blindly mill until curtaining changed direction, get back to (a) and polish again, stop before curtaining reappears

http://www.tescan.com/en-us/technology/accessories/rocking-stage



Tricks

• Infiltrate the sample:

low viscosity resin

vacuum – air cycling

Porous cermet infiltrated with epoxy resin under vacuum and cured under isostatic pressure.

2. Curtaining

Vacuum chamber for vacuum infiltration and two embedding resins: Epon and Körapox.

Cylinders with embedded samples for metallography (mechanical polishing).


Rippling vs Curtaining 3. Rippling

Ion rippling

• Self organization

• Instability of process

• Material dependent (PI, rocks, are bad Al and Si are fine)

• Ion dependent (size of ion vs atoms in solid)

• Current density dependent

• Mostly see it in Xe PFIBTypical ion rippling of Xe PFIB cut cross section on silicate rock*: small plateaus and then curtaining at each end

The geometry of the rippling artifact makes that stage rocking won’t help.

* Sample from: Dina Klimentyeva, Institute of Geochemistry and Petrology, ETH Zürich and Marja Lehtonen, Hugh O'Brien - Geological Survey of Finland (GTK)


How to get rid of rippling?

Tricks

• Thick and very dense deposition

Pt deposited at 10 kV in PFIB

• Masking technique

cover ROI with a thick piece of Si wafer

Dense and homogeneous platinum deposition created using 10 kV Xe beam.

3. Rippling

Cross section polished using Si mask

Sample from: Dina Klimentyeva, Institute of Geochemistry and Petrology, ETH Zürich and Marja Lehtonen, Hugh O'Brien -Geological Survey of Finland (GTK)

Tomáš Hrnčíř, Marek Šikula, Jozef Oboňa & Pascal Gounet: “How to achieve artefact-free FIB milling on polyimide packages”, ISTFA 2016 Conference proceedings, pg. 642-646

(a) Pick a Si mask using manipulator and SEM glue

(b) Place mask onto ROI

(c) Mill using a high current

(a)

(b) (c)


Masking is fantastic!

Massive ion rippling when milling SiC sample, microstructure completely obscured.

3. Rippling

Cross section polished using Si mask, microstructure clearly visible.

Tomáš Hrnčíř, Marek Šikula, Jozef Oboňa & Pascal Gounet: “How to achieve artefact-free FIB milling on polyimide packages”, ISTFA 2016 Conference proceedings, pg. 642-646

For more details see poster “Plasma FIB artefact-free milling using TRUE X-sectioning technique” presented by Miloš Hrabovský.


Types of damage to crystal structure 4. Crystal damage

Relevant parameters

• Energy and type of ion

• Incidence angle

• Type of solid

Amorphization

• Local melting and quenching of solid

• Irregular atom positions obscure HRTEM

Point defects

• Vacancies and interstitials are created

• Electronic structure is disturbed affects TEM Holography

• Accumulation of point defects leads to dislocation loops will show in diffraction contrast TEM

Damage zones in a crystalline solid C

A = amorphized

P = point defects


Amorphization 4. Crystal damage

Thickness of amorphization layer

(a) 160 nm total incl. 60 nm amorphous

(b) 90 nm total incl. 60 nm amorphous

(c) 40 nm total incl. 10 nm amorphous

(a) (b) (c)

* Kelley, R, "Xe+ FIB milling and measurement of amorphous silicon damage", Microsc. Microanal. 19 (Suppl2), (2013) 862-863

** Schaffer, M, et. al., "Sample preparation for atomic-resolution STEM at low voltages by FIB", Ultramicroscopy 114 (2012) 62-71

*** Süess, M, et. al., “Minimization of amorphous layer in Ar+ ion milling for UHR-TEM" Ultramicroscopy 111 (2011) 1224-1232

To reduce thickness of amorphous layer (*, **, ***)

• Low kV milling

• Larger ion

• Lower angles

Overview of amorphization layer thickness in Si, from *


Tungsten – flash polishing

100nm 100nm

Flash timing ca. 0.3s

Power Supply20 V, ca. 100 mA

NaOHsolution

• FIB creates damage to W crystal that in turn spoils TEM image• Flash polishing in NaOH removes damaged layer TEM reveals the oxide particles embedded in the W matrix.

TEM image of W-2Y2O3 ipreparedby FIB, thickness: 150 nm

TEM image after flash polishingthe FIB prepared lamella

‘Focused ion beam application on the investigation of tungsten-based materials for fusion application’L. Veleva, R. Schäublin, A.Ramar, Z. Oksiuta, N. BalucEuropean Microscopy Congress 2008, Volume 2: Materials Science, Eds. S. Richter and A. Schwedt, Springer-Verlag Publication, Berlin Heidelberg, 2008, p. 503-504

4. Crystal damage


Micropillars for mechanical testing 5. Implantation

Pillar morphology

• Presence of Ga in Al is visible by SEM

• XeF2 while Xe milling improves quality of pillar

2 µm

Ga XeXe-Ga

Displacement controlled microcompression setup by Alemnis

Jeff Wheeler, Yuan Xiao, LNM, ETH Zürich


Comparison between Ga and Xe prep’d 5. Implantation

Ga in Al

• Ga accumulates at grain boundaries weakening of grain boundaries

• Deformation mode is changed grains pop out !

Liquid Metal Embrittlement (LME)

• Ga is one of the worst, it affects:Al, Cd, Cu, Fe, Ag, Sn, Zn

Empirical rules for LME

1. Metals form alloy /intermetallic no embrittlement.

2. Metals are immiscible embrittlement possible!

Y. Xiao1,2, T. Al-Samman2, S. Korte-Kerzel2, R. Spolenak1, J.M. Wheeler1

1 ETH Zurich, Laboratory for Nanometallurgy2 RWTH Aachen, Institut für Metallkunde und Metallphysik

Comparison of tested micropillars


Ga vs Xe FIBed pillars 5. Implantation

Microcompression testing: ultra-fine grained aluminum cylinders prepared by Ga FIB and Xe FIB

Y. Xiao, et. al. , Investigation of the deformation behavior of aluminum micropillars produced by focused ion beam machining using Ga and Xe ions, Scripta Materialia, Volume 127, 15 January 2017, Pages 191-194, http://dx.doi.org/10.1016/j.scriptamat.2016.08.028

M.H. Kamdar. “Liquid Metal Embrittlement”. in: Briant CL, (Ed.). Embrittlement of Engineering Alloys. Elsevier Science, 2013.


Mixing of layers

STEM EDX and imaging of “B”

STEM EDX and imaging of patterned thinfilms “A”

Sample from Livie Dorwling-Carter (Biomedical Engineering), TEM STEM by Christian Zaubitzer (ScopeM)

Procedure

• Si wafer with Si3N4 passivation

• Evaporation of 3-layer stack either“A” Ti-Au-Ti (10-5-10 nm)“B” Cr-Au-Cr (10-5-10 nm)

• FIB patterning at 10 pA, several circles from out to inside

• TEM lamella across patterned area

• STEM imaging and EDX mapping

Mixing is much stronger for Ti than Cr!

6. Mixing & demixing

Material Sputter rate (um3/nC)

Al2O3 0.08

Cr 0.1

Ti 0.46

Au 1.5


Bleeding /preferential milling

Bright spots covering FIB polished InP

Explanation *

• P sputters faster than In

• Accumulation of In at surface

• Heat during sputtering allows droplet formation

Tricks hinder droplet formation

• Apply XeF2 etching gas**

• Cooling to low temperature*

• Small milling currents*

6. Mixing & demixing

HEMT sample from Tamara Popovic and Anna Hambitzer, MM-Wave Electronics Group, ETH Zürich

*Chew, N.G. and Cullis, A.G., "The preparation of transmission electron microscopy specimens from compound semiconductors by ion milling", Ultramicroscopy 23 (1987) 175-198

** “Nanofabrication using focused ion and electron beams” Ivo Utke, Stanislav Moshkalev, Phillipp Russell, Oxford University Press, 2012


Questions

tutorial joakimreuteler fib-artifacts-and-how-to … · 'hod\hulqj /rz uhghsrvlwlrq ... d qp...

Documents