Emission Microscopy (EMMI) is an optical fault-localization technique used in semiconductor failure analysis to identify electrically active defects inside integrated circuits. Instead of physically opening individual circuit structures and searching for damage, emission microscopy detects extremely weak light generated while a semiconductor device is electrically operating or biased.
Abnormal photon emission can help identify where electrical activity associated with a failure is occurring. Typical EMMI applications include localization of:
→ Leakage current
→ Junction breakdown
→ Abnormal transistor operation
→ I/O failures
→ Latch-up
→ ESD-related failures
→ Gate-dielectric defects
→ Electrically active shorts
→ Dynamic circuit and timing problems
Once the emitting region has been identified, the failure analyst can correlate the location with the IC layout and perform more targeted physical analysis using techniques such as FIB, SEM or TEM.
What Is Emission Microscopy?
Emission microscopy detects photons generated by semiconductor devices during electrical operation. It is also commonly referred to as Photon Emission Microscopy (PEM).
The basic process is:
Electrically stimulate the IC
↓
A semiconductor structure emits photons
↓
A sensitive optical detector collects the emission
↓
Emission locations are mapped onto an image of the die
↓
The suspect circuit region is identified
The emitted light can be extremely weak, so measurements are normally performed in a dark optical environment using a highly sensitive detector. Unlike laser-based fault-localization techniques, photon emission does not require an external light source to stimulate the semiconductor. The electrical operation of the device itself produces the optical signal.
Why Do Semiconductor Devices Emit Light?
Semiconductor devices can emit photons when electrical carriers undergo certain energy transitions. Two important mechanisms are involved.
Recombination-Related Photon Emission
Electrons and holes can recombine and release energy. Under suitable conditions, part of this energy can be released as photons. This type of emission is associated with bipolar or recombination processes and can be related to p-n junction behavior.
Electric-Field-Assisted Photon Emission
Carriers moving through strong electrical fields can gain significant kinetic energy. When these energetic carriers scatter and lose energy, part of that energy can be released as photons. This mechanism is particularly important in semiconductor structures operating under strong electric fields.
Both mechanisms can provide useful information during semiconductor failure analysis.
EMMI and Leakage Current
Leakage localization is one of the most important applications of emission microscopy.
A healthy reverse-biased p-n junction may produce so little current that its photon emission falls below detector sensitivity. A defective or leaky junction can behave differently. If leakage current passes through a localized high-field region, the combination of current density, electric field and energetic carriers can produce a detectable photon-emission site.
A typical investigation may proceed as follows:
Excessive leakage measured electrically
↓
Device biased under the failing condition
↓
EMMI detects localized emission
↓
Emission correlated with IC layout
↓
Physical analysis targets the suspect structure
Localized leakage sites can sometimes produce detectable emission even when total current is relatively small, especially when the current is concentrated in a very small region.
Junction Breakdown and EMMI
Photon emission can become particularly strong near avalanche breakdown. In a reverse-biased junction, carriers accelerate through a high-field depletion region. As the electric field increases, some carriers gain enough energy to create additional electron-hole pairs.
This multiplication process generates large numbers of energetic carriers and can result in strong photon emission. EMMI can therefore be effective for detecting abnormal junction breakdown.
Care must be taken during analysis because uncontrolled avalanche conditions can permanently damage the device. The goal is to reproduce the failing condition with enough electrical stress to generate a useful optical signal without creating additional damage.
Not Every Breakdown Produces Strong Emission
Different breakdown mechanisms can produce different optical behavior. Avalanche breakdown typically involves carriers gaining substantial kinetic energy and can therefore generate strong photon emission. Tunneling-dominated breakdown may generate weaker emission.
A weak or absent emission site therefore does not automatically mean that electrical breakdown is not occurring. This is one reason EMMI should always be interpreted together with electrical measurements.
Photon Emission From MOSFETs
MOSFETs can generate photon emission under certain operating conditions. In normal static operation, a properly functioning CMOS circuit generally produces very little photon emission.
However, when a MOSFET operates under conditions involving strong electric fields, energetic carriers can generate detectable photons. Abnormal transistor behavior may therefore create an emission pattern that differs from a known-good circuit. This can help identify suspect transistor regions.
Static vs. Dynamic Emission Microscopy
Photon emission analysis can generally be divided into static emission and dynamic emission.
Static Emission Analysis
Static EMMI is particularly useful when a failure produces a continuous or relatively stable emission site. Examples include:
→ Leakage
→ Junction breakdown
→ Abnormal current paths
→ Latch-up
→ Some transistor defects
The device is biased into the failing condition and the photon emission is integrated over time to create an image.
Dynamic Emission Analysis
Digital CMOS behaves differently. When a CMOS gate remains at logic 0 or logic 1, there may be very little emission. During switching, however, the transistors temporarily enter conditions that can generate photons.
The result is a short optical event associated with the switching transition. This makes photon emission useful for investigating dynamic circuit behavior.
Time-Resolved Emission
Time-Resolved Emission (TRE) measures when photon-emission events occur relative to the electrical operation of the circuit. Instead of simply showing where emission occurred, the technique also provides timing information.
This can allow the analyst to investigate:
→ Switching activity
→ Signal propagation
→ Clock activity
→ Timing differences
→ Circuit delays
The emission signal does not necessarily reproduce the electrical waveform directly. Instead, photon peaks correspond to periods during which transistors are switching and producing detectable emission. This allows switching events to be followed through different parts of a circuit.
Picosecond Circuit Analysis
Advanced time-resolved emission methods can provide extremely high temporal resolution. This can make it possible to compare transistor switching delays and investigate timing-related failures.
Potential applications include:
→ Marginal timing failures
→ Clock-tree problems
→ Delay differences
→ Dynamic logic failures
→ Scan-related failures
→ First-silicon debugging
Time-resolved analysis is more specialized than conventional EMMI, but it demonstrates that photon emission can provide both location and timing information.
Frontside EMMI
Photon emission can sometimes be collected from the front side of the semiconductor die. This approach is easier on devices containing relatively few metal interconnect layers.
Modern ICs, however, may contain many layers of metal above the active transistor structures. These metal layers can block or shadow photons traveling toward the microscope.
As interconnect complexity increases, frontside emission analysis becomes more difficult. Two electrically similar transistors may even appear to generate different emission intensities because their optical paths through the metal stack are different. This is one reason backside analysis has become increasingly important.
Backside EMMI
Backside photon emission microscopy observes active circuitry through the silicon substrate. Silicon can transmit significant near-infrared light under suitable conditions. Photon emission generated near the transistor structures can therefore travel through the silicon and be detected from the backside.
A typical backside EMMI arrangement is:
Active transistor region
↓
Near-infrared photon emission
↓
Silicon substrate
↓
Backside microscope optics
↓
Sensitive detector
Backside analysis avoids much of the optical obstruction created by frontside metal interconnects. For highly integrated ICs, backside access can therefore provide a much clearer route to the active circuitry.
Advantages of Backside EMMI
Reduced Metal Shadowing
The emission does not need to travel through the complete interconnect stack.
Better Access to Dense ICs
Modern multilayer interconnect structures may make frontside observation difficult or impractical.
Better Device-to-Device Comparison
A more uniform backside optical path can make emission patterns easier to compare.
Compatibility With Flip-Chip Devices
In a flip-chip package, the active side of the die faces the package substrate. The backside may therefore be the natural optical access point.
Silicon Thickness and Backside EMMI
Although silicon transmits certain near-infrared wavelengths, transmission depends on several factors:
→ Wavelength
→ Silicon thickness
→ Substrate doping
Reducing silicon thickness can improve optical transmission and therefore increase photon-emission sensitivity. Backside preparation may include:
→ Package material removal
→ Silicon grinding
→ Mechanical polishing
→ Localized thinning
→ Surface polishing
However, unnecessary thinning should be avoided. Sample preparation introduces risk, and electrical functionality must normally be preserved so that the failure can still be reproduced during EMMI.
EMMI Detectors and Near-Infrared Light
Photon emission from silicon devices often extends into the near-infrared (NIR) range. Detector selection is therefore an important part of emission microscopy.
Important detector characteristics include:
→ Spectral sensitivity
→ Detector noise
→ Cooling
→ Integration time
→ Spatial resolution
→ Time resolution
A detector optimized for visible light may not provide sufficient sensitivity for the wavelengths generated by modern semiconductor devices.
Why Detector Cooling Matters
Photon emission from semiconductor failures can be extremely weak. At the same time, the detector itself generates electronic noise. If detector noise becomes comparable to the optical signal, a weak failure site can disappear into the background.
Cooling the detector reduces noise and improves sensitivity. This can be particularly important when analyzing low-voltage or low-current devices.
Low-Voltage ICs and EMMI
As semiconductor operating voltages decrease, photon-emission analysis becomes more challenging. Lower voltage generally means carriers gain less kinetic energy.
The resulting emission may become:
→ Weaker
→ Shifted toward longer infrared wavelengths
This creates two challenges. First, weaker emission requires greater detector sensitivity. Second, longer wavelengths reduce optical spatial resolution.
Advanced low-power semiconductor technologies therefore place increasing demands on photon-emission equipment.
Spatial Resolution of EMMI
The spatial resolution of photon emission microscopy is partly limited by the wavelength of the emitted light. Near-infrared wavelengths are longer than visible wavelengths and therefore provide lower optical resolution.
This becomes increasingly significant as transistor dimensions shrink. An emission site may therefore identify a small Region of Interest without identifying one individual nanoscale defect. Additional fault-localization or physical-analysis techniques may still be required.
EMMI and Gate Dielectric Failures
Certain defects involving thin gate dielectrics can create electrically active leakage paths. Examples include:
→ Localized dielectric leakage
→ Tunneling-related current
→ Gate dielectric breakdown
Under suitable electrical conditions, these failures may generate detectable photon emission. EMMI can therefore help identify the approximate region of a gate-related failure before destructive analysis begins.
However, emission microscopy generally cannot reveal the physical structure of the dielectric defect. Final confirmation may require additional physical analysis such as FIB and TEM.
EMMI and Latch-Up
Latch-up can produce strong photon emission. Latch-up involves activation of parasitic semiconductor structures within an integrated circuit. When these structures enter conduction, substantial current can flow. The resulting carrier activity may generate strong photon emission.
EMMI can therefore help determine:
→ Where latch-up begins
→ Which circuit region is involved
→ How the event develops
Because latch-up can also create significant local heating, photon emission may be combined with thermal fault-localization techniques.
EMMI and ESD-Related Failures
Electrostatic-discharge damage can create several electrically active failure conditions. These may include:
→ Junction damage
→ Leakage
→ Localized breakdown
→ Damaged I/O structures
→ Abnormal parasitic conduction
If the damaged structure continues to conduct electrically, high-field or recombination processes may produce detectable photon emission. EMMI can therefore help localize certain ESD-related failure sites.
However, an emission site alone does not prove that ESD caused the failure. The failure mechanism must still be established using electrical and physical evidence.
EMMI and I/O Failures
I/O circuitry commonly contains:
→ Protection structures
→ Large transistors
→ Junctions
→ Parasitic bipolar structures
These structures can generate distinctive emission behavior when abnormal electrical conditions occur. EMMI can therefore be useful for investigating:
→ I/O leakage
→ Junction breakdown
→ Protection-device failures
→ Latch-up
→ ESD-related damage
Once the emission location has been identified, IC layout information can help determine which structure is involved.
CAD Navigation With EMMI
An emission image becomes significantly more useful when it is correlated with the IC design layout.
The image can be overlaid with information such as:
→ Metal layers
→ Transistor locations
→ Circuit blocks
→ Nets
→ Standard cells
A small emission site can then be associated with a particular circuit structure. The investigation becomes:
Emission site
↓
CAD correlation
↓
Suspected device or net
↓
Electrical hypothesis
↓
Targeted physical analysis
This can significantly reduce the area that must be investigated destructively.
EMMI as a Fault-Localization Technique
It is important to distinguish fault localization from root-cause analysis. EMMI may identify the region where abnormal electrical activity is occurring, but it does not necessarily identify the exact physical defect or explain why that defect occurred.
A complete investigation may therefore proceed:
Electrical failure verified
↓
EMMI identifies emission site
↓
CAD identifies suspect circuit
↓
FIB exposes the structure
↓
SEM or TEM identifies physical defect
↓
Failure mechanism determined
↓
Root cause established
EMMI is especially valuable because it reduces the search area before destructive physical analysis begins.
EMMI vs. Thermal Imaging
EMMI and thermal imaging can both localize electrically active failures, but they measure different physical effects.
Emission Microscopy
Emission microscopy detects photons generated by semiconductor carrier processes and is particularly useful for:
→ Leakage junctions
→ Breakdown
→ Transistor activity
→ Latch-up
→ Dynamic switching
Thermal Imaging
Thermal imaging detects heat generated by electrical power dissipation and is particularly useful for:
→ Shorts
→ Resistive leakage
→ Hot spots
→ Localized power dissipation
Some failures produce both optical emission and heat. Using both techniques can therefore provide complementary information.
EMMI vs. OBIRCH
EMMI detects photons produced by the electrical operation of the semiconductor device. OBIRCH uses a different principle. A laser scans across the device and locally heats the structure, while the electrical response to this heating is monitored to identify resistance changes.
EMMI is particularly useful when:
→ The defect naturally emits photons
→ Junction leakage is suspected
→ Breakdown is present
→ Latch-up is involved
→ Dynamic transistor behavior is being investigated
OBIRCH can be particularly useful when:
→ A resistive short exists
→ An interconnect has abnormal resistance
→ A leakage path responds to local heating
→ The failure produces little or no detectable photon emission
For difficult failures, the two techniques can be complementary.
EMMI vs. FIB
EMMI and FIB perform very different functions. EMMI locates electrically active behavior, while FIB physically accesses the selected location.
This creates a powerful combination:
Leakage current
↓
EMMI localization
↓
CAD correlation
↓
FIB cross-section
↓
SEM / TEM analysis
Instead of using FIB to search a large area, the analyst can target the region identified by EMMI.
Advantages of EMMI
→ Contactless optical fault localization
→ Device can operate electrically during analysis
→ No external optical stimulus is required
→ Sensitive to several electrically active semiconductor mechanisms
→ Suitable for backside analysis
→ Can localize small leakage regions
→ Can be correlated with CAD layout
→ Can provide time-resolved information
→ Can guide FIB and physical analysis
One of its greatest advantages is its ability to reduce a complex integrated circuit to a manageable Region of Interest.
Limitations of EMMI
The Defect Must Produce Detectable Emission
Not every semiconductor failure mechanism generates photons.
Signal Intensity Can Be Extremely Low
Long integration times and highly sensitive detectors may be required.
Low Supply Voltage Reduces Emission
Advanced low-voltage technologies can produce weaker optical signals.
Infrared Wavelength Limits Spatial Resolution
An emission site may cover several neighboring semiconductor structures.
Silicon Thickness and Doping Affect Backside Transmission
Sample preparation may sometimes be required.
Metal Layers Can Block Frontside Emission
Backside analysis may therefore be preferable.
EMMI Does Not Automatically Reveal Root Cause
The emission site must still be correlated with electrical and physical evidence.
Common Semiconductor Failures Investigated With EMMI
| Failure / Behavior |
How EMMI Can Help |
| Leakage current |
Locate electrically active leakage site |
| Junction breakdown |
Detect emission from a high-field region |
| Avalanche breakdown |
Identify localized breakdown site |
| Gate dielectric leakage |
Localize electrically active gate-related region |
| Abnormal transistor operation |
Compare emission with expected circuit behavior |
| I/O failure |
Identify suspect I/O or protection structure |
| Latch-up |
Locate parasitic conduction region |
| ESD-related damage |
Locate electrically active damaged region |
| Dynamic logic failure |
Track switching-related emission |
| Timing failure |
Compare switching timing using time-resolved emission |
Typical EMMI Failure Analysis Workflow
1. Verify the Failure
Reproduce the electrical failure under controlled conditions. Record information such as:
→ Supply voltage
→ Current
→ Input state
→ Output state
→ Temperature
→ Failure signature
2. Determine the Required Optical Access
→ Frontside access
→ Backside access
The choice depends on package construction, metal density and device technology.
3. Prepare the Sample if Necessary
For backside analysis, package opening or silicon thinning may be required while preserving electrical functionality.
4. Bias the Device
Apply the electrical conditions required to reproduce the failure.
5. Acquire a Reference Image
Record the physical die or backside image for navigation.
6. Collect Photon Emission
Integrate the optical signal until emission sites become detectable.
7. Compare With Known-Good Behavior
→ Emission location
→ Emission intensity
→ Bias dependence
→ Dynamic timing
8. Correlate With CAD
Identify circuit structures beneath the emission location.
9. Form a Failure Hypothesis
Determine whether the emission is consistent with:
→ Leakage
→ Breakdown
→ Latch-up
→ Abnormal transistor operation
→ Another electrical failure mechanism
10. Perform Targeted Physical Analysis
Additional techniques may include:
→ FIB
→ SEM
→ TEM
→ Nanoprobing
The objective is to identify the actual physical defect responsible for the electrical failure.
Frequently Asked Questions About EMMI
What does EMMI mean in semiconductor failure analysis?
EMMI generally refers to Emission Microscopy, an optical technique used to detect weak photon emission produced by electrically operating semiconductor devices.
Is EMMI the same as Photon Emission Microscopy?
The terms are commonly used together in semiconductor failure analysis. Photon Emission Microscopy emphasizes the detection of photons generated by electrically active semiconductor structures.
What failures can EMMI detect?
→ Leakage
→ Junction breakdown
→ Latch-up
→ Some transistor failures
→ I/O defects
→ ESD-related damage
→ Dynamic switching problems
Can EMMI detect leakage current?
Yes. Localized semiconductor leakage is one of the most important applications of photon emission microscopy.
Does EMMI require the IC to be powered?
Normally, yes. The device must generally be electrically biased or operated under conditions that produce the emission being investigated.
Why is EMMI often performed from the backside?
Modern ICs contain many metal interconnect layers that can block frontside optical access. Near-infrared photons can travel through silicon under suitable conditions, making backside inspection useful for advanced devices.
Can EMMI detect ESD damage?
EMMI can localize some electrically active defects associated with ESD, including leakage, junction damage and I/O-related breakdown. Additional evidence is required to establish ESD as the root cause.
Can EMMI identify a gate oxide failure?
Certain gate-dielectric leakage and breakdown conditions may generate detectable photon emission. Physical confirmation may still require techniques such as FIB and TEM.
What is time-resolved emission?
Time-resolved emission measures when photon events occur relative to circuit operation. It can be used to investigate switching activity and timing behavior.
What is the difference between EMMI and OBIRCH?
EMMI detects photons produced by the semiconductor’s electrical operation. OBIRCH uses a scanning laser to locally heat the circuit and monitors resulting changes in electrical resistance.
Is EMMI destructive?
The emission measurement itself is generally non-destructive. However, sample preparation may be required to obtain optical access, particularly for backside analysis.
Does EMMI determine root cause?
Not by itself. EMMI primarily localizes electrically active behavior. Physical analysis is often required to determine the exact defect, failure mechanism and root cause.
EMMI as Part of Semiconductor Failure Analysis
The real strength of emission microscopy is its ability to transform an electrical failure into a physical location. Without fault localization, a failure analyst may face millions or billions of transistor structures.
EMMI can progressively narrow the investigation:
IC fails electrically
↓
Failure signature reproduced
↓
Photon emission detected
↓
Emission site localized
↓
CAD identifies suspect circuit
↓
FIB / SEM / TEM targets the location
↓
Physical defect identified
↓
Failure mechanism determined
↓
Root cause established
For failures that generate detectable photon emission, this can dramatically reduce the amount of physical analysis required.
EMMI therefore occupies an important position between electrical characterization and physical failure analysis.
Looking for an EMMI / Photon Emission Failure Analysis Lab?
If your semiconductor device has unexplained leakage, junction breakdown, I/O failure, latch-up, abnormal transistor behavior or another electrically active failure, an EMMI-capable failure analysis laboratory may be able to localize the problem before destructive physical analysis begins.
Typical requirements may include:
→ Emission Microscopy
→ Photon Emission Microscopy
→ Backside EMMI
→ Leakage localization
→ Junction breakdown analysis
→ I/O failure localization
→ Latch-up analysis
→ Time-resolved photon emission
→ CAD correlation
→ Follow-up FIB / SEM / TEM analysis
When requesting EMMI analysis, provide information about the device technology, package type, electrical failure signature, required bias conditions, failure reproducibility and any fault-localization work already completed.
The appropriate semiconductor failure analysis provider can then be selected according to the required EMMI equipment, backside preparation capability, device expertise and follow-up physical-analysis capabilities.
