E345: Techniques and Processes of failure analysis

Why every thing is built on ICs'

1) Reduction of the interconnection parasitic - an IC with multilevel metallization can substantially reduce the overall wiring length.

2) Full utilization of semiconductor wafer's - as devices can be closed-packed within an IC chip.

3) drastic cost reduction in processing cost - as wire bonding is error prone and time-consuming

Semiconductor materials

1) Insulator - very low conductivities

2) Conductors - high conductivities

3) Semiconductors - conductivity in-between insulators and conductors

Reasons for Failure analysis

Integrated chips (ICs) contain hundreds millions of transistors within the microprocessors

Be it expected/sudden, failures have severe business impacts. And need to understand and solve the issues asap

Finding cause of failure is most important as there's hundreds of thousands of transistors

IC manufacturing

Wafer fabrication

Lithography

Etching

Packaging

Wafer Fabrication Flow (Possible steps that can go wrong)

Thin film formation

Impurity Doping

Photolithography

Etching

Forms insulation and metal layers for semiconductors

Form regions with controlled conductivity by adding impurities into bulk regions

To define patterns of structure to be formed

Form patterned profile by removing unwanted material

Measured Attributes - Tools

Film or layer thickness - Secondary Electron Microscopy (SEM), Transmission Electron Microscope (TEM) and Atomic Force Microscopy (AFM)

Morphology & Roughness - Secondary Electron Microscopy (SEM) or Atomic Force Microscopy (AFM)

Film Stoichiometry & Depth Profile - Rutherford Backscattering Spectroscopy (RBS), Automated Electrical Sort (AES) and X-ray Photoelectron Spectroscopy(XPS)

Bulk Impurities (including atmospherics) - >0.5% ( Electrical Discharge System (EDS), Auger Electron Spectroscopy (AES), X-ray Photoelectron Spectroscopy (XPS), Hot Filament Source (HFS)) and <0.5% ( Secondary Ion Mass Spectrometry (SIMS), Glow Discharge Mass Spectrometry (GDMS))

Surface Composition

Elemental: X-ray Photoelectron Spectroscopy (XPS), Auger Electron Spectroscopy (AES)

Chemical: X-ray Photoelectron Spectroscopy (XPS)

Organic: Time-of-Flight Secondary Ion Mass Spectrometry (TOF-SIMS)

Surface Impurities:

Metallic: Total Reflection X-ray Fluorescence (TXRF), Time-of-Flight Secondary Ion Mass Spectrometry (TOF-SIMS), Surface Secondary Ion Mass Spectrometry (SurfaceSIMS), X-ray Photoelectron Spectroscopy (XPS), Auger Electron Spectroscopy (AES)

Organic: Time-of-Flight Secondary Ion Mass Spectrometry (TOF-SIMS)

IC Failure Analysis Techniques

Electrical Characterization Techniques: Four point probe, Hall measurement

To study the I-V, sheet resistance etc

Optical Techniques: Optical microscopy

For external or internal visual inspection

Thermal Imaging Techniques: Liquid Crystal, Infrared Thermography

To locate the hotspot (e.g. due to high current density)

Ion Beam technique - Focused Ion Beam (FIB), Secondary Ion Mass Spectrometer (SIMS)

X-ray Techniques: X-ray Diffraction (XRD), X-ray Photoelectron Spectroscopy (XPS)

Electron Beam Techniques: Secondary Electron Microscopy (SEM), Auger Spectroscopy, Energy Dispersive X-ray (EDX), Transmission Electron Microscopy (TEM)

Scanning Probe Technique: Scanning Tunnelling Microscopy (STM) and Atomic Force Microscopy (AFM)

Why Failure Analysis

Modern computer chips contains hundreds millions of transistors within its microprocessors.

Size of transistors are getting smaller to pack more transistor into a set area for better efficiency. i.e. compare a computer in 1900s to a computer in 2023

Failures are inevitable due to complexities and rapidly shrinking devices

Occurs due to several points

1) Product development

2) Qualification

3) yield learning

4) Reliability improvement

5) system manufacture

6) field applications

Main points for failure analysis

Sudden/anticipated failures could result in dire business impacts

Understanding failure and take corrective actions quickly

Goal: locate fault site

Challenging to locate a single failed transistor out of the hundreds of millions

Need a systematic and effective approach to locate the failure

Types of Failures

Short Failures

Open Failure

Cracks causes open circuit

Particle causes short circuit

Reasons for open circuit

Failure fault
( Reported fault )

Open-circuit (O/C)

Failure defect ( Analysis report )

Missing interconnects

Corrosion

Mechanical damage/scratch

Open bond

Open metallization

Metallization mircocracks

Local disruption of interconnects

Die cracking

Failure mechanism / Failure cause ( Interpretation )

Mask errors

1) Moisture ( internal gas; poor seal )

2) Contaminants + moisture

Poor processing
/ handling

1) Contaminant present on pad

2) Overpressure in bonding

3) Package stress

4) Al / Au intermetallic formation

5) Materials compatibility ( purple plague )

6) Fatigue failure

7) EM in metallization

8) Processing defects ( Step coverage; lithography )

9) Corrosion

10) Wrong passivation composition

Stress migration in as-deposited metallization

1) Electrical overstress ( EOS )

2) Electrostatic discharge (ESD)

Thermal / Mechanical stress

Failure Analysis Terminology

Failure mode: nature of device malfunction which resulted in the observed failure characteristics

e.g. open circuit, short-circuit, leakage

Failure Defects: nature of fault causing the device malfunction

e.g. micro-cracks, growth

Failure mechanism: physical phenomenon which produced the failure defects

e.g. electro-migration, corrosion

Root-cause: nature of the action which caused the failure mechanism to occur

e.g. poor design, high current density, stress concentration

Failure Analysis Methodology

Step 1: Verify there is a failure

Errors in production testing can cause good IC chips to be falsely identified as failures

Step 2: Characterize the symptoms of the failure

A thorough understanding of the IC failure condition enables the analyst to proceed efficiently and utilize info from similar failures

Step 3: Verify that these symptoms caused the observed failure

There maybe more than one failure mechanism affecting an IC, but only one of these mechanisms is usually responsible for the malfunction being analysed

Step 4; Determine the root-cause of the failure

Involves localization of the failure and determination of the failure mechanism and root cause

Step 5: Suggest corrective action

Corrective actions can be implemented by a team of design, process, reliability and failure analysis engineers to overcome the failure

Step 6: Document the results of the failure analysis

Documentation communicates the result of the analysis and forms a historical record that is invaluable for future failure analysis work

Failure analysis process flow

1) fault localization

identifying the location of fault from a sea of transistors / device

2) De-processing

reversal of wafer fabrication process; films are removed in reverse order of application using etching process

3) Defect localization and characterization

Pin-point the location of defects and gather as much info as possible about the defect and it's location since subsequent steps are irreversible

4) Inspection and physical characterization

Inspection of defects and further characterization is performed to gather info to find out the root-cause

Moore's Law - Device scaling

no. of transistors on a chip will doubled every 18 months

Have to reduce size of transistors to continue increase the number

Measured Attributes

Process and Device Evaluation

Throughout the wafer-fabrication process, many tests and measurements are needed to determine wafer quality and process performance

One process mistake can make wafer completely useless

One "Killer" defect can ruin a die

Good characterization

warns of process that goes out of control & device characterization

essential to analyse circuit performance and conformance for customer specifications

Characterization

Some measurements are direct and some are indirect

Direct measurements

electrical measurements of test wafer and actual devices

Physical parameters

Measures direct effect of some process

Layer thickness and widths, composition and others

Metrology an Inspection

General term applied to the measurement physical surface features

Separated into 3 main classifications by application

Critical Dimension (CD) measurement

Overlay measurements

Particle and pattern defect detection

Thin film parameter measurement i.e. resistivity, thickness and stress

Common measured attributes

surface morphology: Thin Film thickness

Defects: Rsistivity

Cause: Impurity / contaminants concentration

What we measure: Defects

solid materials contains large amounts of imperfections

Common Si wafer defects

point defects

Examples: lattice vacancies, self-interstitial atoms, substitutional impurity atoms, and interstitial impurity atoms

Linear & interfacial defects

Examples: Edge dislocation, screw dislocation, An extra portion of a plane of atoms, or half-plane, the edge of which terminates within the crystal

Planar defects

Examples: Gain boundaries, stacking faults and external surfaces

Volume bulk defects

Examples: Pores, cracks

Associated with one or two atomic positions, including vacancies, self-interstitials (host atoms that occupy interstitial sites) and impurities

Interstitial solid solutions from for relatively small impurity atoms that occupy interstitial sites among the host atoms

For substitutional solid solution, impurity substitute for hose atoms

Dislocation is a linear or one-dimensional defect around which some of the atoms are misaligned

Boundaries that have two dimensions and normally separates regions of the material that have different crystal structures and/or crystallographic orientations

Typical dimension & microscopy (Defects)

Microscopes with different resolutions are used to view defects of particular sizes

Subatomic particles: 1X10-14 m to 1X10-11m

Atom / ion diameter: 1X10^-11m to 1X10^-9m

Unit cell edge lengths: 1X10^-9m to 1X10^8m

Dislocations (width): 1X10^-8m to 1X10^-7m

Second phase particles: 1X10^-8m to 1X10^-4m

Grains: 1X10^-8m to 1X10^-2m

Macrostructural features (porosity, voids, cracks): 1X10^-5m to 1X10^-1 m

Typical dimension & microscopy (Microscope)

Scanning probe microscopes: 1X10^-11m to 1X10^-7m

Transmission electron microscopes: 1X10^-10m to 1X10^-4m

Scanning electron microscopes: 1X10^-9m to 1X10^-3m

Optical microscopes: 1X10^-7m to 1X10^-3m

Naked eye: 1X10^-4m to 1X10^-1m

What we measure - impurity

The electrical conductivity of the semiconducting materials is not as high as that of the metals

Electrical properties of these materials are extremely sensitive to the presence of even minute concentration of impurities

Intrinsic semiconductor are those in which the electrical behaviour is based on the electronic structure inherent in the pure material

When the electrical characteristics are dictated by impurity atoms, the semiconductor is said to be Extrinsic

Extrinsic Semiconductors

Extrinsic semiconductors (both N and P type) are produced from materials that are initially of extremely high purity

Controlled concentrations of specific donors or acceptors are then intentionally added

Termed as doping

In extrinsic semiconductors, large numbers of charge carriers (either electrons or holes, depending on the impurity type) are created at room temperature, by the available thermal energy

Increase of impurities for such carries increase the conductivity in doped material

Impurity Measurement Techniques

Undesired / unintended doping (such as contaminants or change of composition) would cause drastic charge in electrical performance of devices

What we measure - Resistivity

One of the most important electrical characteristic of a solid material is the ease with which it transmits an electric current. Ohm's law relates the current to the applied voltage as shows

V = IR

p = (RA) / I

R = resistance of material through which current is passing

I = Distance between the two points at which the voltage is measured

A = Cross-sectional area perpendicular to the direction of the current

p = Resistivity (Independent or sample geometry). The units for p are ohm-meters

Resistivity (p)

Important for starting material as well as for semiconductor devices because it contributes to the device series resistance, capacitance, threshold voltage and other paraeters

The wafer resistivity is usually modified locally during device processing by diffusion and / or ion implantation

Four point probe + Sheet resistance

Four point probe

Commonly used to measure semiconductor resistivity

Absolute measurement without recourse to calibrated standards

measures bulk or thin film specimens

Consists of four equally spaced metal tips with finite radius

High impedance current source is supplied though outer two probes

Voltmeter measures the voltage across the inner two probes

Sheet resistance

Resistance between two ends is

R = p( L / A ) = p( L / W t ) = ( p / t ) ( L / W ) ohms

L/W has no units, p/t unit is ohms. To distinguish between R and p/t is called sheet resistance (ohms/square)

R = Rsh ( L / W ) ohms

Pro of using sheet resistance: When build square features on the wafer, ever the size of the features ( for example 11, 22, 3*3) are different, the resistance will be the same of these square features.

Film thickness measurement

Total thickness of the sample

Thickness of variously doped semiconductor layers

Thickness of external layers of foreign materials such as dielectric or metal films deposited on a semiconductor surface

Optical Techniques

Thickness typically range from a few millimetres to a few nanometres and thus require several different techniques in order to provide reasonable sensitivity over such as extended range

widely used as:

applicable to opaque and transparent films

generally yields thickness values of high accuracy

measurements are rapidly performed, frequently non-destructive, and utilize relatively in-expensive equipment

Interferometry and ellipsometry are basic optical methods for determining film thickness

Both interferometry and ellipsometry, incident light impinges on the film surface and a reflected portion is measured

Interferometry

If light is light is irradiated onto a wafer some beams of light are reflected on top of it and some penetrate into the film. The latter will be reflected on bottom of this layer or penetrate into another layer beneath and so on

These multiple-reflection light waves boost and weaken each other according to their phase difference

The phase difference of each multiple-reflection light is determined by the light wavelength and optical path length (i.e. the distance that light moves back and forth in the thin film multiplied by the film's refractive index).

Optical path difference is related to film thickness

The colour of the film is affected by the thickness and its refractive index so the film colour is merely a rough guide to thickness

Colour chart (Not applicable if the film thickness < half wavelength of the shortest visible light

t = to ( no / nf )

to: Thickness read from the chart

no: Reflective index of chart material

nf: Refractive index of measured film

Full colour chart refer to CA2 WS L4 Q2

Ellipsometer

non-destructive and non-contact

requires neither sample preparation nor special measurement environment.

Used to determine dielectric properties (complex index of refraction) and thickness of thin transparent films (a few Armstrong to a few um)

Measured attributes

film thickness

refractive index

Spectroscopic ellipsometer combined with variable angle of incidence is used

alloy ratio

crystallinity

Interfacial roughness

What happens if the film is not transparent visible light?

Using longer wavelength

What happens if the film is highly absorbed to the light source?

e.g. infrared light, provided the layer which may not transparent to

cannot use ellipsometer for example: metal

Light

Property

Light can be described as an electromagnetic wave travelling through space. For purposes of ellipsometry, it is adequate to discuss the wave's electric field behaviour in space and time, also known as polarization

When the light has completely random orientation and phase, it's considered unpolarized

For ellipsometry, however we are interested in the kind of electric field follows a specific path and traces out a distinct shape at any point. This is known as polarized light

Polarization

When two orthogonal light waves are in-phase, the resulting light will be linearly polarized The relative amplitudes determine the resulting orientation

If the orthogonal waves are 90 degree out-of-phase and equal in amplitude, the resultant light is circularly polarized

The most common polarization is "elliptical", one that combines orthogonal wave of arbitrary amplitude and phase. This is where ellipsometer gets it name

Working Principle

Measures change in polarization state of light reflected from the surface of a sample

Arrangement of optical components between the source and detector defines the type of ellipsometer being used. 3 basic designs of ellipsometer are null, polarization modulation and rotating element ellipsometer, but all three are approximately the same work.

Refer to CA2 slide 32

Mechanical Techniques: Surface Profiler

Takes measurements electromechanically by moving the sample beneath a diamond-tipped stylus

High-precision stage moves a sample beneath the stylus according to a user programmed scan length, speed and stylus force

The stylus is mechanically coupled to LVDT ( Linear Variable Different Transformer )

Surface variations cause stylus to be translated vertically. Electrical signals corresponding to stylus movement are produced as the core position of the LVDT changes.

The LVDT scales an AC reference signal proportional to position change

After A/D conversion, the digitized signals from performing a single scan are stored in computer memory.

Thin film thickness can be measured by sensing the mechanical movement of a stylus as it traces the topography of a film-substrate step. The step can be made by masking part of the area during thin film deposition.

Factors limiting accuracy of stylus measurements:

Stylus penetration and scratching of films:

Occurs on very soft films

Substrate roughness:

Excessive noise is introduced into the measurement as a result and this creates uncertainty in the position of the step especially when the film thickness is rather small

Thin Film Stress

Stresses are generally present in thin films. Such residual or internal stresses directly affect a variety of phenomena including:

1) Adhesion

2) Generation of crystalline defects

3) Formation of surface such as hillocks

Stress tends to increase with thickness, promoting film peeling, stress is a prime limitation to growth of very thick films

Substrate deformation and distortion also necessarily arise from stress in the overlaying films. In IC's even the slightest bowing of silicon wafer presents significant problems

Common issues in aluminium lines, stress contributes to reliability failure such as electro-migration

Cracks

voids

Tensile Stress

Develops when growing film is shorter than oxide

Compressive stress

Compressive stress develops when film is longer than substrate

Compatibility, requires both film and substrate have same length

Film is constrained and stretches while substrate contracts

Sometimes, tensile stresses are large causing film fracture. Similarly, exceedingly high compressive stresses can cause film wrinkling and loss of adhesion

Simple Lens

Thin lens is a lens with a thickness that is negligible compared to the focal length of the lens

Properties of light

Visible light, used in light microscopy

Is a form of electromagnetic radiation contained in discrete units i.e photons

Determine amount of energy associated with proton at a specific wavelength is E = hv

Energy, frequency and wavelength

E = hc / λ

λ = the wavelength (m)

E = energy (eV)

h = Planck's constant ( 6.63 X10^-34 js = 4.14X10^-15 eVs )

c = speed of light

Refraction & Reflection

Refraction

Occurs when light bends

Reflection

Diffuse reflection

Occurs when light is reflected off rough surface

Mirror-like reflection

Occurs when light is reflected off a smooth surface

Due to change in speed of light as light passes through different mediums

Distance along the optical axis between the two surfaces of the lens

Can be +ve or -ve

+ve lens converge light

-ve lens diverge light

Diffraction

Spreading light that occurs when a beam of light contacts an object

In microscope

Light from illuminator is diffracted by the image plane

collected by the objective lens and focused in the image plane

Wave constructively and destructively interfere to form a contrast image

Constructive Interference

Amplitude of the final waves is a sum of amplitude of 2 identical wavs that are in phase

destructive interference

Waves may annihilate each other when they are out of phase with each other by λ / 2

Airy disk

Central diffraction spot that contains

84% of the light from point source

result of passing an extended Electromagnetic wave front through a small aperture.

Diameter of the disk:

D = 1.22λ / NA

D = diameter of the disk

λ = wavelength

NA = numerical aperture

Rayleigh Criterion & Resolution Limit

2 adjacent object points are resolved when the airy disk of 1 point coincides with the first diffraction minimum of the other point in the image

Rayleigh set the limit for the smallest resolvable distance between 2 points as D = 1.22λ / NA

Rayleigh Resolution Limit

Numerical Aperture, NA, of the lens is the light collecting ability of the lens entering the objective form a foxed object distance

Related to the refractive index of a material that represents the optical density between materials like space between objective lens and specimen. (optical density - the speed of propagation of light rays)

NA = n * sinθ

θ = half of collection angle of objective

NA = Numerical aperture

n = refractive index of the lens working medium (refractive index of air = 1)

To improve resolution, n or sineθ needs to increase

Resolution VS Magnification

Magnification = how much bigger a sample appears under the microscope than in real life

Resolution = ability to distinguish between 2 points on an image at the smallest separation

Function of microscope = Enhance resolution & enlarges view of object that cant be seen by the human eye

Enlargement of image not necessary means higher/increased resolution. Increase resolution = image size increase not resolution

Commonly used contrast technique

Differences in light intensity between the specimen and the adjacent background relative to the overall background

Contrast mechanism

Bright field: Produce a dark image against a brighter background

Dark field: Produce a bright image against a dark background

Phase contrast: Enhance contrast for object when there is small phase shifts in the light passing through a transparent speciment

Chromatic & Spherical Aberrations

Chromatic Aberrations:

Occurs as lens refracts light differently depending on the wavelength

refer to CA3 slide 20

Spherical Aberrations:

Solution: Use a compound lenses that have different colour dispersing properties

Due to spherical surfaces of lens

refer to CA3 slide 20

Solution: Use a combination of +ve and -ve lenses with different thickness

Electron Microscopy

Used due to shrinking size of transistors

Advance microscopy techniques with higher resolution (e.g. SEM, TEM, AFM) are needed.

Electron beam microscope is developed to offer a better spatial resolution to overcome the limitation of light microscope due to diffraction of light

Optical Microscopy vs SEM

Optical Microscope (OM):

The smallest details which can be distinguished in a light microscope are approximately 200nm but an unaided human eye can only detect details up to 200um. As a result any magnification > 1000X only makes details bigger by not finer details visible.

Known as 'Empty Magification'

Higher magnification, requires 3 or more stages of magnification that introduces extra aberrations to the systems

Scanning Electron Microscopy (SEM)

Capability on order of magnitude better than optical microscope due to the small wavelength of the probed beam

wavelength of visible light is between 400 to 700nm

wavelength of electrons about 3600 eV or 0.02nm

Magnification: 5X-1500X, Depth of Field: 0.5mm, Resolution: approxi 0.2um

Magnification: 10X-500KX, Depth of Field: up to cm, Resolution 1.5nm

Depth of Field & Depth of Focus

Depth of Field

Range of distance along the optical axis in which the specimen can move without the image appearing to lose sharpness

Depth of focus

Extend of the region around the image plane in which the image will appear to be sharp

Extent of a zone on the specimen which appears focus at the same time

To improve depth of field:

Decrease the angle of convergence

Increasing the working distance

Decreasing the aperture size

D = (4X10^5 W) / (A M) um

W = working distance

D = depth of field

d = beam diameter

P = pixel diameter

A = aperture

a = beam convergence angle

m = Magnification

How it works:
(Uses electrons instead of light to form an image)

Beam of electrons is produced at the top of the microscope by heating of a metallic filament

Electron beam follows a vertical path through the column of the microscope

Makes it through electromagnetic lenses which directs and focuses beam down towards sample

When electron beam hits sample, other electrons are ejected from sample and collected by detector

Converted signal is sent to a viewing screen, producing an image

Key components in SEM:

Electron optical column

Vacuum system

Signal detection & display

Electron - Solid Interaction

Electron beam strikes a sample, a large no. of signals are generated

Interaction of the primary beam with the sample creates an excitation volume in which the electrons are scattered (Elastic & inelastic scattering)

Elastic: Electrons may be reflected with no loss of energy, absorbed by the specimen, giving rise to secondary electrons with low energy, together with X-rays

Inelastic: Absorb and give rise to emission of visible light (an effect is cathodoluminescence)

e.g. Auger Electron, Backscattered Electron, Secondary Electron, Continuous & Characteristic X-rays

e.g. Auger Electron, Cathoduminescence

Sample Characteristics

Provides info on surface topography. Used in Voltage Contrast & Magnetic Contrast Imaging

BSE

Provides info on atomic no. topography & materal

Surface sensitive & provides info on chemical composition through thickness (EDX, WDX)

Characteristics X-rays

Provides info on chemical composition through thickness

Emitted photons provides electrical info & crystal defects info

SEM Signals - Sample Characteristics

This is sensitive to surface morphology or topography

Secondary Electrons (SE) detectors measure the SE emission intensity as electron beam raster across the sample, especially at the edges where more SE escapes from the sharp edges

Due to low energy (< 50eV), only secondary electrons (SE) from within few nm of surface can escape.

As Secondary Electrons (SE) moves, they will interact with other atomic electrons and loss energy

Due to loss in Kinetic Energy (KE), the avg distance that a SE travels is very small

Most SE are brought to rest within interaction volume

Those created near surface may escape

Topographical Images

Secondary Electrons (SE) & Backscattered Electrons (BSE) coefficient are minimum when the surface of the specimen is perpendicular to the electron beam as there is less possibilities of electrons being scattered out of the surface

Hence SE & BSE electron coefficient varies with tilt of the specimen

BSE is able to get topographical images but as the direction of BSE is more peaked at the forward direction, the yield of electron is also lower.

As electron beam tilts, more electrons are likely to scatter out of the specimen rather than penetrating further into it.

There's relatively lower yield of BSE collected as compared to SE

Rough surface observed with BSE imaging will have more shadows compared to SE imaging.