Access Official Technical PDF

Share Summary:LinkedInEmail

In optical lithography, intensity images do not tell the whole story. Light diffracted by photomask edges can carry phase variations that change how a feature prints as focus shifts. Phase imaging in optical lithography provides ways to recover and interpret this information from measurements, helping engineers connect mask structure and imaging behavior with through-focus performance.

These methods are relevant to mask metrology, lithography process-window analysis, and imaging-system characterization. Their value lies in extracting useful information from changes in measured intensity rather than relying on a single image at best focus.

Why photomask edges create phase effects

A photomask absorber is not infinitely thin. Its height and sidewalls affect how electromagnetic waves interact with feature boundaries. Diffraction at those boundaries can alter the phase of light in the aerial image, including near regions intended to remain dark. The result is not simply a change in image contrast: phase also influences how intensity evolves through focus.

Phase-shift masks use interference to improve printed-image contrast, but their thicker absorber structures can increase edge diffraction. Opaque MoSi on glass (OMOG) masks use thinner, more strongly attenuating absorber materials, which can mitigate some effects. They do not necessarily eliminate edge-related phase variations.

For lithography engineers, the distinction matters because mask-induced phase can affect critical-dimension behavior across focus. A model that captures only absorber transmission or bulk phase may miss the local phase gradients at feature edges that influence the shape of through-focus curves.

Phase imaging in optical lithography with focus stacks

A practical way to investigate phase is to acquire intensity images at multiple focus positions. A single in-focus image may conceal phase differences, particularly when a feature is strongly absorbing. Comparing images on either side of focus can reveal asymmetries that point to phase modulation in the aerial image.

One approach is the Transport of Intensity Equation (TIE), which relates the change in intensity along the direction of propagation to phase in the image plane. With suitable through-focus measurements, the equation can be solved to estimate phase. In photomask analysis, this allows phase variations associated with absorber topography to be studied using an aerial imaging system.

The method depends on the measurement conditions and the assumptions used in reconstruction. For example, oblique illumination can introduce pupil-related asymmetries in addition to mask effects. Separating these contributions may require estimating both the wafer-plane field and the imaging pupil rather than interpreting a recovered phase map as a direct image of the mask alone.

Connecting edge phase to the lithography process window

Through-focus behavior is commonly examined with Bossung plots, which show how critical dimension changes with focus at a given exposure condition. If the intensity response is symmetric around focus, the corresponding curve may also be symmetric. Phase gradients near mask edges can break that symmetry and produce a tilt in the curve.

This connection gives phase imaging a practical role in process analysis. Measuring edge-related phase can help explain why a feature’s critical dimension changes as focus shifts and support feature-specific process-window assessment. It can also inform mask modeling and optical proximity correction (OPC), where an accurate representation of mask behavior is important to predicting the aerial image.

Thin-mask models offer one way to represent three-dimensional absorber effects without modeling the full geometry in every calculation. In these models, effective complex-valued boundary layers at feature edges approximate the influence of the mask near-field. The boundary layer can represent both an effective edge displacement and a phase contribution. Measurements can help constrain such models and check whether they reproduce observed through-focus behavior.

Different absorber properties do not necessarily affect the image in the same way. The described simulations and measurements indicate that absorber height changes bulk phase and transmission, while sidewall angle can change printed critical dimension without producing the same change in phase or Bossung tilt. Distinguishing these effects can help engineers avoid attributing every CD shift to a single mask parameter.

Differential intensity imaging reveals optical dynamics

Instead of comparing only the raw images in a focus stack, differential intensity imaging examines intensity changes between focus positions. Subtracting nearby measurements highlights where energy moves as the image propagates. These differences can make polarization-dependent interactions between light and mask sidewalls easier to see, even when the original intensity images appear similar.

Such measurements can support several lithography tasks:

  • Assessing focus sensitivity: Relate image changes to critical-dimension variation across focus.
  • Comparing mask regions: Identify differences in optical response that may indicate non-uniformity.
  • Evaluating models: Test whether a simplified mask representation reproduces measured through-focus behavior.

At corners, the energy-flow pattern can include curl that conventional TIE reconstruction may not capture accurately. An iterative approach can estimate and correct for this component, reducing artifacts in recovered phase maps. This is a reminder that phase-retrieval results depend not only on data quality but also on whether the reconstruction method represents the field’s behavior adequately.

Using EUV speckle to characterize imaging systems

Phase-related measurements can also characterize the imaging system, not just the photomask. In extreme ultraviolet (EUV) lithography, weak scattering from surface roughness can create speckle in aerial images. Under a weak-scattering approximation, information about imaging-system aberrations is encoded in the speckle’s spatial-frequency content.

Changing the illumination angle provides another measurement dimension. Comparing speckle under on-axis and tilted illumination can reveal different components of pupil aberration. In this way, a rough or blank mask can act as an optical probe, helping estimate aspects of the imaging system’s transfer behavior from image measurements.

This approach differs from analyzing strongly absorbing DUV masks. There, the main objective is often to measure phase associated with mask features and their edges. With weakly scattering EUV speckle, the pattern can instead help reveal aberrations in the imaging path. In both cases, measured intensity changes provide information about an underlying optical field that is not directly recorded by a conventional intensity image.

Practical implications for engineering teams

For mask, process, and metrology teams, the core lesson is that aerial-image intensity should be interpreted in context. Focus stacks and differential measurements can expose phase effects that a best-focus image hides. Used alongside suitable reconstruction and modeling, these methods can support more informed analysis of edge diffraction, polarization dependence, critical-dimension sensitivity, and imaging-system aberrations.

They do not replace careful process characterization or rigorous modeling. Rather, they provide measurement-based ways to test assumptions, identify relevant optical effects, and improve the connection between mask behavior and printed-image response.

Conclusion

Phase imaging in optical lithography turns changes in intensity across focus into insight about mask and imaging-system behavior. For DUV photomasks, it can help reveal edge phase and its relationship to critical-dimension behavior and process windows. For EUV systems, speckle measurements can provide a route to assessing pupil aberrations. The practical benefit is a more complete picture of how optical fields evolve from mask to wafer—and a stronger basis for measurement-informed lithography models and decisions.

Access Official Technical PDF

About The Author