SEM/FIB with TOF-SIMS:

Introduction and Application

Examples

James Whitby, Deborah Alberts and Johann Michler

Empa, Swiss Federal Laboratories for Materials Science and

Technology

Laboratory for Mechanics of Materials and Nanostructure

July 10th for TESCAN ‘webinar’

Outline

Who are Empa?

What is TOF-SIMS?

Motivation for FIB-SIMS

The Empa instruments

Example applications Depth profiling: masks and contamination from plasma asher

Imaging: lateral resolution standard

Imaging: grain boundaries in alloys

Quantification: boron in silicon

Isotopic labelling: origin of hydrogen in ALD films

Molecular identification?: PTFE

Real-world complications

2

Who are Empa?

Empa is a German acronym for the Swiss

Federal Laboratories for Materials Science and

Technology

Funded largely by the Swiss government

Administered similarly to ETHZ/EPFL

Budget of 100 million Swiss Francs

Plus 50 million from 3rd parties

Employs 1500 people at three locations,

publishes about 500 papers per year

In top ten for reputation of European R&D institutions

3

Bern

Zürich

Bellinzona

Lausanne

Basel

Chur

Empa sites in Switzerland

St. Gallen

Dübendorf

Thun

What do Empa do?

Empa pursues a range of applied research

programs connected to materials science:

Nanostructures, Health, Natural resources and

pollutants, Sustainable built environment, Materials

for energy technologies

Services for companies

R&D (textiles, sensors, buildings, materials)

Reports/tests/analyses (e.g. failed cable-car cables!)

Consulting (nanotoxology, life cycle analysis)

Training courses

6

Empa

Laboratory for Mechanics of Materials and

Nanostructures

Nanomechanics E.g. nano-indentation tests in SEM

Nanomaterial development Electrochemical

FEBID/FIBID

ALD

FIB/SEM, Raman, EBSD, EDX, GD-MS, FIB-SIMS, XRD…

Development of new instrumentation Characterization

FIB-TOFSIMS, GD-TOFMS

Mechanics In SEM hot and cold indentation stages

Miscellaneous Nano-grippers

Gas injection systems

Functionalized AFM tips

7

What is (TOF)-SIMS?

Secondary Ion Mass Spectrometry (SIMS) When material is sputtered by high energy particles (e.g.

focused ion beam) a small fraction is ionized

Mass analysis of sputtered ions gives composition of sample

Versatile but expensive analysis/imaging technique Geochemisty/Cosmochemistry

Semiconductor industry

Time-Of-Flight mass spectrometry Type of mass analysis with performance and price

between quadrupole and magnetic sector mass analyzers

Very fast measurements (100 kHz) Can use normal FIB scan speeds / dwell times

8

Secondary Ion Mass Spectrometry (SIMS)

Energy

analyzer

Mass

spectrometer

Detector

Mass Spectrum

Depth profile

Image

collision cascade

● Solid surface is bombarded by primary ions (1-30 keV

energy). Surface atoms and molecules overcome

surface binding energy, some are charged (+

or -)

2

Detector Detector

HV

Pulser

HV

Pulser

ToF chamber

Time of Flight

is measured

Ion transfer optics

Primary ions

Sample

FIB chamber

Orthogonal ToF-SIMS

Fiblys Workshop 2010, Brno

Flight time [us]

co

un

ts

Ga

Ca

Prototype FIB/SEM-TOFSIMS at Empa

Fiblys Workshop 2010, Brno

Sample stage

SIMS

optics

FIB

SEM

Commercial version (on Lyra3 GM)

12

Mass analyser (TOF)

Ion pump for mass

analyser

Motivation for FIB-SIMS

Demand for chemical imaging SIMS much faster than EDX, with better lateral resolution, limits

of detection and dynamic range Works for light elements, allows isotopic labelling

Allows depth profiling and 3D imaging

Compared to traditional dedicated SIMS, SEM/FIB allows excellent imaging and sample manipulation easy location of regions of interest

milling and slicing with FIB beam (dedicated software)

flexible stage design

Manipulation (FEBID/FIBID, micromanipulators, other accessories)

Given a FIB, then adding a mass analyzer to do SIMS has a cost comparable to EDX equipment

13

TOF-SIMS

Structure/Imaging

Composition/properties

organic inorganic

EDX

SNOM

Raman

TERS, SERS

AFM

Secondary electron image from

SEM/FIB

FIB-TOFSIMS development at Empa

Started project with Tescan (SEM/FIB instruments) and Tofwerk (time-of-flight mass analyzer) in 2008 EU research funding, part of larger project

FIBLYS project succesfully concluded in 2011

Tescan now offer SEM/FIB-TOFSIMS commercially Lyra3-GM

Further development of original prototype Improve sensitivity, integrate more with AFM, Raman

EU funding (part of larger UnivSEM project)

Independently of Tescan: Feasibility study of higher mass resolution version (Swiss

funding with Tofwerk) Much less signal and would be much more expensive, but x10 mass

resolving power. Useful?

15

The Empa FIB-SIMS instruments

Two SEM/FIB-SIMS instruments, both using orthogonal-extraction time-of-flight mass analyzers (Tofwerk AG) on Tescan SEM/FIB instruments (Vela XM and Lyra GM)

The original prototype, M/DM ~500, equivalent to the commercial version from Tescan/Tofwerk

A higher mass resolution version (M/DM <4000, but lower signal). Still in development! No plans for commercial version.

Orthogonal extraction time-of-flight analyzer No need to pulse primary ion beam (so no need for monoisotopic material to preserve mass

resolution)

no modification to FIB/SEM necessary

TOF has better detection limits than quadrupole for complex nanomaterials (many elements to monitor in limited amount of material)

Dynamic range adequate (>106)

Extraction ion optics have earthed tip, so stage collision alarm still works and there are no strong extraction fields that might distort electron beam images.

Both instruments use a gallium ion beam for FIB and SIMS Canion 31 from Orsay Physics

50 pA at 20 nm resolution, 1 pA at 7 nm resolution (30 keV)

Charge compensation using electron microscope beam

16

0.00001

0.0001

0.001

0.01

0.1

1

0 50 100 150 200 250 300 350

Counts

/Extr

actio

n

Frames

Al 27 Ga 69 In 115 As 75

InP

cap

laye

r, t

= 1

0 n

mA

l 0.5

3In

0.4

7A

s w

ind

ow

laye

r, t

= 8

0 n

m

5×

2 A

l 0.1

4G

a 0.1

8In 0

.68A

s Q

Ws,

t =

7 n

m5

×3

Al 0

.28G

a0

.26I

n 0.4

6As

stra

inco

mp

en

sati

on

laye

rs, t

= 1

0 n

m6

×In

Psp

ace

rs,

t (s

ub

cavi

ty)

= 3

×λ/

n

35

pa

irs

of

Al 0

.9G

a 0.1

As/

Ga

As

DB

R, t

= 0

.25

×λ/

n

Wa

fer

fusi

on

GaA

ssu

bstr

ate

Ga

0.2

1In

0.7

9As 0

.45P

0.55

etch

sto

p la

yer,

t =

40

nm

InP

sub

stra

te

InP

cap

laye

r, t

= 10

nm

Al 0

.53In

0.4

7A

s w

indo

wla

yer,

t =

80 n

m

5×2

Al 0

.14G

a 0.1

8In 0

.68A

s Q

Ws,

t =

7 n

m5

×3

Al 0

.28G

a 0.2

6In 0

.46A

s st

rain

com

pe

nsa

tio

nla

yers

, t =

10

nm

6×

InP

spa

cers

, t

(su

bca

vity

) =

3 ×

λ/n

35

pai

rso

f A

l 0.9

Ga 0

.1A

s/G

aAs

DB

R, t

= 0

.25

×λ/

n

Wa

fer

fusi

on

GaA

ssu

bstr

ate

Ga

0.2

1In

0.7

9As 0

.45P

0.55

etch

stop

laye

r, t

= 40

nm

InP

subs

trat

e

InP

cap

laye

r, t

= 10

nm

Al 0

.53In

0.4

7A

s w

indo

wla

yer,

t =

80 n

m

5×2

Al 0

.14G

a 0.1

8In 0

.68A

s Q

Ws,

t =

7 n

m5

×3

Al 0

.28G

a0

.26I

n 0.4

6As

stra

inco

mp

en

sati

on

laye

rs, t

= 1

0 n

m6

×In

Psp

ace

rs,

t (s

ub

cavi

ty)

= 3

×λ/

n

35

pai

rso

f A

l 0.9

Ga 0

.1A

s/G

aAs

DB

R, t

= 0

.25

×λ/

n

Wa

fer

fusi

on

GaA

ssu

bstr

ate

Ga

0.2

1In

0.7

9As 0

.45P

0.55

etch

stop

laye

r, t

= 40

nm

InP

subs

trat

e

GaA

s (1

00 n

m)

Al 0

.9G

a 0.1

As

(100

nm

)

InP

cap

laye

r, t

= 10

nm

Al 0

.53In

0.4

7A

s w

indo

wla

yer,

t =

80 n

m

5×2

Al 0

.14G

a 0.1

8In 0

.68A

s Q

Ws,

t =

7 n

m5

×3

Al 0

.28G

a0

.26I

n 0.4

6As

stra

inco

mp

en

sati

on

laye

rs, t

= 1

0 n

m6

×In

Psp

ace

rs,

t (s

ub

cavi

ty)

= 3

×λ/

n

35

pai

rso

f A

l 0.9

Ga 0

.1A

s/G

aAs

DB

R, t

= 0

.25

×λ/

n

Wa

fer

fusi

on

GaA

ssu

bstr

ate

Ga

0.2

1In

0.7

9As 0

.45P

0.55

etch

stop

laye

r, t

= 40

nm

InP

subs

trat

e

InP

cap

laye

r, t

= 10

nm

Al 0

.53In

0.4

7A

s w

indo

wla

yer,

t =

80 n

m

5×2

Al 0

.14G

a 0.1

8In 0

.68A

s Q

Ws,

t =

7 n

m5

×3

Al 0

.28G

a0

.26I

n 0.4

6As

stra

inco

mp

en

sati

on

laye

rs, t

= 1

0 n

m6

×In

Psp

ace

rs,

t (s

ub

cavi

ty)

= 3

×λ/

n

35

pai

rso

f A

l 0.9

Ga 0

.1A

s/G

aAs

DB

R, t

= 0

.25

×λ/

n

Wa

fer

fusi

on

GaA

ssu

bstr

ate

Ga

0.2

1In

0.7

9As 0

.45P

0.55

etch

stop

laye

r, t

= 40

nm

InP

subs

trat

e

InP

(10

0 n

m)

Al x

Ga 1

-xIn

yAs 1

-y (

44 n

m)

InP

(10

0 n

m)

InP

(10

nm

) A

l 0.5

3G

a0

.47A

s (8

0 n

m)

InP

(90

nm

)

35x 5x G

aA

s su

bst

rate

2.7 µm

10 µm

10 µm

VECSEL depth profile

VECSEL SAMPLE: Mass Spectrum

Principal elements

C

C

Multiple charged ions

Principal elements

Cluster ions

Contamination

Multiple charged ions

Co

un

ts/E

xtr

act

ion

m/q

APPLICATIONS

(1)

13

Depth profile in mask

Is there or is there not a thin layer of titanium at

depth at the marked location in the component?

Given information:

sample consists of gold on lithium niobate or lithium

tantalate

Presumed to have an adhesion layer (Pt or Ti?)

between gold and lithium niobate

19

20

Silver dag

connection to

electrically ground

metallic layer

Craters

from SIMS

analysis

SEM image sample in FIB-SIMS chamber showing

area indicated for depth profile measurements

Pen mark

21

TOFSIMS depth profiles for gold mask on lithium tantalate

• platinum layer at interface (plotted as GaPt+, gives clearer signal)

• titanium layer at interface (below platinum)

Empa’s customer was very happy to have the existence and nature

of the Pt/Ti adhesive layer confirmed!

Depth profile: Contamination from plasma

asher

Our plasma asher was found to be depositing aluminium oxide at all power and pressure conditions (red curve vs black curve).

The aluminium was shown to come from the powered electrode.

22

Al+ FIB-SIMS image from the BAM L200 certified reference material. Pairs of lines

spaced by as little as 34 nm can be distinguished (dimensions in image refer to the

period of the line pairs. The gap between lines is half the period e.g. 67.5/2 = 34 nm).

Beam spot size was calculated to be 40 nm (20-80% across step); image pixel size is

16 nm.

Lateral resolution

Limited by beam spot size

7 nm for Canion 31

Lateral primary ion scattering

effects are negligible

In practice, S/N and time to

acquire image are important

TESCAN specify 50 nm

Best lateral resolution yet at

Empa: ~40 nm

Imaging: grain boundaries

Magnesium alloy with hard-phase skeleton (Al,

Ca rich)

24

Mg Al Ca

Close up of same Mg alloy

Ten minutes data collection

25

Ca+

Ion-induced secondary electron

image after measurement Al+

Mg+ Secondary electron image after

measurement

Imaging: grain boundary in aluminium

99.5% aluminium alloy, heat treated

Magnesium has concentrated at grain

boundaries

NB different sputtering rates on different

crystallographic orientations

26

Mg+ Al+

Ion-induced secondary

electron image

Quantification:

Boron in silicon FIB-SIMS calibration

Detection limit < 10 ppm

Conditions: 14.7 nA, 10x10 microns, 512x512x121 pixels,30 keV

Oxygen flooding used: 6.0×10-5 mbar O2

27

Isotopic Labelling to Study Atomic Layer

Deposition (ALD)

ALD: Conformal, self-limiting, coating technique

e.g. for Transparent Conducting Oxide films, semiconductors

To make nominal ZnO / Al2O3 layers

Diethyl zinc / water

Triethyl aluminium / water

Impurities?

FIB-SIMS and GD-OES showed boron to sometimes be present in the diethyl zinc

Origin of hydrogen? Using FIB-SIMs and ERDA/RBS with deuterium

labelled water as a reagent:

Hydrogen in ALD material comes from water, and not from the ethyl- ligands.

28

ALD deposited ZnO/Al2O3

pairs of layers, 20 – 40 nm

thick.

1H+ 2D+

Molecular identification??

29

Despite low ion yields and problems with charge compensation,

molecular information can sometimes be obtained.

Real-world complications

SIMS! Secondary ion yields (sensitivity) vary by orders of magnitude across periodic table

Very matrix sensitive (orders of magnitude), makes quantification challenging Response usually not linear to concentration except at low concentrations (~ppm)

only Ga+ primary ions in typical FIB Relatively low secondary ion yields (sensitivity) compared to caesium or oxygen for

most materials limits useful spatial resolution

Not ideal for organic materials

Not ideal for samples with gallium (e.g. semiconductor lasers, CIGS photocells)

‘High’ vacuum chamber pressure (not UHV) 1×10-6 – 1×10-5 mbar

NB since our prototype, Tescan option for < 1×10-6 mbar on Lyra3 GM

affects secondary ion yields e.g. causes drift during day as chamber pumps down

Can be beneficial for sensitivity

gives background from residual gases

non-linear scaling of signal with current and scan area

30

Effect of deliberate oxygen flooding on secondary ion signal for cobalt in CoNiFe over typical working pressure range

N.B. still some effect even at lowest possible pressure of SEM/FIB!

Similar effects seen for water vapour

31

4

3

2

1

0

En

ha

nce

me

nt fa

cto

r w

ith

oxyg

en

22x10-62018161412108

Pressure / mbar

Roughening can reduce depth resolution in

some materials

polycrystalline copper metal oxides on silicon

Acknowledgements

Tofwerk AG

Fredrik Oestlund

Empa

J Baudot, Ivo Utke, M Bechelany (ALD samples)

Jeff Wheeler (magnesium and aluminium alloys)

Sebastian Schmidt (ICP-MS, GD-MS on B/Si)

EPFL/Beam Express

V Iakovlev, A Sirbu (VCSEL sample)

European Union:

FP7 projectS FIBLYS (2008-2011) and UnivSEM

(2012-2015)

33

Secondary ion image of

implanted gallium after

analysis of ALD material