At a glance
- Port-level insertion loss confirms a split ratio is wrong, but cannot show where inside the coupling region the light actually goes.
- Near-field optical probing maps light propagation and loss inside a working photonic integrated circuit at wafer or die level, non-destructively.
- Omni-SiPh LightView offers near-field, deep-trench and angled probes, edge and grating coupling, and electro-optical measurements such as IV, LIV and wavelength.
- Conventional fallbacks stay dedicated test structures, simulation comparison and destructive cross-sectioning, which reveals geometry without revealing optical behavior.
InZiv
Published:
When a directional coupler returns from fabrication with a split ratio that misses its design target and drifts with wavelength, input and output measurements can confirm the error without locating it, because they report only the light going in and the light coming out. The measurement class that can locate it is spatially resolved optical characterization inside the working circuit: near-field optical probes, positioned close enough to interact with the evanescent field (the part of the optical field that extends beyond a guiding structure and decays rapidly with distance), map how light propagates and where optical power is lost along the actual device. Omni-SiPh LightView from InZiv applies this approach to photonic integrated circuits (PICs, chips containing interconnected optical components) at wafer and die level, with edge and grating coupling and electro-optical measurements such as IV, LIV and wavelength, so an engineer can see the bend, splitter, coupler or waveguide section where light is lost. The conventional alternatives remain dedicated test structures, comparison against simulation and, in some cases, destructive cross-sectioning, which shows the geometry and nothing about how the light behaved. As of November 2026, InZiv's silicon photonics systems, Glide-SiPh PIC and Omni-SiPh LightView, are available to order.
Why can a photonic integrated circuit pass wafer spec and still miss its designed split ratio?
A photonic integrated circuit can pass wafer spec and still miss its designed split ratio because neither the simulation nor the process data shows what the light does inside the coupling region. In the typical situation InZiv describes for PIC and photonics design engineers, a taped-out directional coupler comes back with a measured split ratio that is off and shifts with wavelength. The simulation says the design is correct, the process data says the wafer is within spec, and the coupling region stays a black box.
Port measurements do not resolve the question. Standard test equipment measures only the light going in and coming out, so it can confirm that the ratio is wrong without showing where inside the device the power went. Process and yield engineers face a related problem: they cannot always tell whether variation comes from the device or from the optical coupling. InZiv lists coupling variation that makes device results hard to interpret, and a yield problem that port data cannot explain, among the situations in which teams reconsider their existing test setup.
Locating the deviation inside the coupling region is the job that in-circuit optical measurement is meant to do.
What does mapping the optical field inside a PIC coupling region actually mean?
The canonical term is in-circuit optical characterization: spatially resolved optical measurement taken inside an accessible photonic circuit. Mapping the optical field inside a photonic integrated circuit means measuring light at points between the input and output ports of the device, so a split ratio that disagrees with simulation can be traced to the structure that caused it. Omni-SiPh LightView performs this class of measurement at wafer and die level.
The terms below carry specific meanings in this context.
- Coupling region: the part of a directional coupler where power transfers between the two waveguides. Port data does not show what happens inside it.
- Split ratio: the share of input power that reaches each output port of a splitter or coupler.
- Near-field optics: optical measurement or interaction at very short distances, including access to evanescent fields near a device.
- Port-level versus in-device measurement: port-level testing reports insertion loss, meaning the optical power lost along the path, and output power for the whole path. In-device measurement reports the position along that path where the power went.
Which approaches are available today for observing what happens in the coupling region?
Several approaches are available today for investigating a coupling region, and each one answers a different question about the device. Three criteria help decide between them:
- Spatial localization. Can the method say where inside the circuit power is lost, or only how much arrives at the output port? This becomes decisive when simulation and process data both look correct.
- Device integrity. Whether the method damages the device under test.
- Measurement breadth. Optical mapping alone versus optical plus electro-optical data such as current-voltage behavior and wavelength response.
| Approach | What InZiv's framing says about it |
|---|---|
| Pass/fail measurement at the ports | Measures only the light going in and coming out; costs nothing, changes nothing in the flow and is good enough while volumes and yields are acceptable |
| Dedicated test structures and comparison against simulation | Steps in finding the cause of excess loss, often alongside several design or fab iterations |
| Destructive cross-sectioning | Shows the geometry and nothing about how the light behaved |
| Near-field optical mapping | Maps the light as it travels through the working circuit, without damaging it, and shows the component where light is lost |
Omni-SiPh LightView applies the last approach in that table.
How does Omni-SiPh LightView help close the gap between simulation and measurement?
Omni-SiPh LightView helps with one specific investigation: a taped-out photonic integrated circuit, such as a directional coupler, whose measured split ratio disagrees with simulation while the process data says the wafer is within spec. The scope here is in-circuit optical characterization of a working device for analysis; high-throughput pass/fail sorting at volume is the job of Glide-SiPh PIC.
The attributes that matter for a coupling-region discrepancy:
| Attribute | What Omni-SiPh LightView provides |
|---|---|
| Measurement principle | Near-field mapping of light propagation and optical loss inside a working circuit, so engineers see where light is lost and why |
| Measurement level | Wafer level and die level |
| Coupling types | Edge coupling and grating coupling |
| Probe options | Near-field, deep-trench and angled probes, plus custom probe designs for specific customer architectures |
| Electro-optical measurements | IV, LIV and wavelength |
| Device integrity | Measures without damaging the device |
The work it supports spans process development and new product introduction, including design validation, debugging yield problems and offline failure analysis when something goes wrong in production. Per InZiv, pinpointing the exact location of optical loss or design deviation on the actual device is how Omni-SiPh LightView is designed to reduce iteration cycles from weeks to hours, so a fail at wafer test can be traced to where the light is actually lost. Omni-SiPh LightView is available to order.
Where in the test flow should coupling-region measurement be inserted?
Coupling-region measurement is most useful where its result can still change a decision about the design or the process. That is the territory Omni-SiPh LightView serves: process development and new product introduction, with offline failure analysis when something goes wrong in production.
Across the wider test flow, InZiv's intended coverage spans test insertion points 1, 2, and 3: wafer testing before and after bonding, as well as testing of individual dies after wafer dicing.
| Insertion point | Stage, in words |
|---|---|
| 1 | Wafer testing before bonding |
| 2 | Wafer testing after bonding |
| 3 | Testing of individual dies after wafer dicing |
In a bonded stack, a photonic integrated circuit is joined to an electronic integrated circuit that drives or reads it. Electrical pads and optical interfaces can then sit on opposite faces, which is why double-sided probing, meaning access from both sides of the device, becomes part of the test conversation at this stage. InZiv's role there is described by its intended coverage of wafer testing before and after bonding.
Frequently Asked Questions
What can near-field mapping show that port measurements cannot?
Port measurements report only the light going into and coming out of a photonic integrated circuit, so a split ratio that disagrees with simulation stays unexplained. Omni-SiPh LightView uses near-field optical probes to map light propagation and optical loss inside a working circuit at wafer and die level. Engineers see the specific bend, splitter, coupler or waveguide section where light is lost.
Does probing inside the circuit damage the device?
No. Omni-SiPh LightView maps the light as it travels through the working circuit, at wafer or die level, without damaging it, in contrast to a destructive cross-section that shows geometry and nothing about how the light behaved. Probe options include near-field, deep-trench and angled probes, with custom probe designs for specific customer architectures. Per InZiv, the system is designed to reduce iteration cycles from weeks to hours.
Can the same system handle production pass/fail testing?
No. InZiv separates the two jobs: Omni-SiPh LightView serves process development, new product introduction, design validation, debugging yield problems and offline failure analysis, while Glide-SiPh PIC performs high-throughput pass/fail sorting at volume. Per InZiv's published Glide-SiPh PIC specifications, that tester covers 200 mm and 300 mm wafers, is FOUP compatible, and provides multichannel optical and electrical characterization, optical insertion loss per channel and wafer-level yield and uniformity mapping. Both are available to order.
Which test stages does InZiv's silicon photonics coverage include?
According to InZiv, its intended coverage spans test insertion points 1, 2 and 3: wafer testing before bonding, wafer testing after bonding, and testing of individual dies after wafer dicing. Testing of the finished co-packaged module sits outside InZiv's fit today, as do very low-volume laboratories that need only input and output power measurements.
How long does an evaluation and purchase usually take?
Per InZiv, the time from a demonstration to the purchase order runs anywhere from 2 to 6 months and often depends on the customer's own timelines, ending at the order rather than at delivery or installation. The sequence covers a technical discussion and review of device architecture and test requirements, sample-based demonstration or feasibility work, agreement on configuration and acceptance criteria, a commercial proposal, the order, then installation and acceptance. A pilot or on-site evaluation is optional.
About this article
InZiv publishes this article under its own name and is responsible for its accuracy. Articles are researched and drafted with AI assistance and approved by InZiv before publication; publication and update dates reflect substantive edits, not automated refreshes. Last updated: 2026-11-01