CD-SEM is a tool used for
         i) Measuring the lines and holes and other fine feature patterns produced on semiconductor wafers. [2] Those are obtained by deriving a critical dimension (CD) from a pair of parallel edges extracted from the images.
         ii) Measuring CD variation and roughness, with the invaluable abilities inherent to image-based systems for spot-checking design weak points of discrete features.
         iii) Measuring contour information.
         iv) Qualitative excursion information gained only from an image. [15]
         v) Hybrid metrology. [16]

Over the past many years, some have stated that CD-SEM is almost at the end of its usefulness. A couple of past works showed that the spatial resolution of CD-SEMs has stagnated in recent years, and is falling behind the requirements of Moore’s law. [17, 18] CDSEM measurements are limitated and thus causes repeatability to decline in some cases: [2]
         i) Application on 22 nm node and below is very challenging.
         ii) High aspect ratio (HAR) trenches.
         iii) Holes in multilayer stacks.
         iv) Noise. Shot noise is cause by secondary electrons given off as a result EB irradiation.
         v) Visual field misalignment.
         vi) Auto focus variation refers to CD value variation caused by slight differences in calculated focal points when capturing images.
         vii) Brightness correction variation is caused by the accuracy of internal automatic adjustments of detected signal amplification factor required for imaging various materials with different secondary electron emission efficiencies at a fixed brightness level.
         viii) On throughput.
         ix) The exposure to the EB (electron beam) causes ArF (argon fluoride) resist patterns to shrink and other damaging effects. [2]
         x) Is often not very reliable when dealing with complex 2D patterns since the size of the measurement window can be very small which often decreases the reliability of the measurements unless spending a lot of effort to set up a robust metrology recipe. [10] In contrast, an alternative is to use SEM contour technique to extract all the edges of the image, which in principle is more versatile, but is not used in traditional CD-SEM.

There are three components in total repeatability (3-σ_total):
         i) Short-term repeatability (3-σ_short), which is the CD value variability when the same pattern is measured continuously multiple times.
         ii) Long-term stability (3-σ_longe), which is the CD value variability when the same place is measured after a time interval (e.g. one week).
         iii) Tool difference (3-σ_tool), which is the CD value variability when the same place is measured by different tools.
where,         
          ----------------------------------- [2360]

Due to a continuous reduction in feature size, a reduction in resist film thickness (FT) is required to prevent large aspect ratios that lead to pattern collapse, especailly <30 nm resist FT is expected when advancing to high NA EUVL. This brings along associated challenges with: [9]
       i) Resist critical dimension scanning electron microscope (CDSEM) metrology.
       ii) Resist patterning performance.

It is expected that, for high NA EUVL, the challenges can be subdivided into: [9]
         i) Metrology challenges due to reduced CDSEM image quality. The design of experiments for the metrology can be used to address this challenge:
           i.a) Visual comparison of the CDSEM images of the various resist film thicknesses (FTs).
           i.b) Verification of the resist array height compared to the nominal resist FT by AFM (see Figure 2360a).
           i.c) Assessment of the quality of the CDSEM images with the signal-to-noise ratio (SNR).
           i.d) Determination of the CD 3σ precision of a full CDSEM image.
           i.e) Simulation can be used to further strengthen the experimental data.
        ii) Challenges related to resist patterning performance. These challenges can be addressed through:
           ii.a) Determination of the uLWR scaling through resist FT.
           ii.b) A CDSEM frame variation study to further elucidate that observed uLWR scaling.
           ii.c) Simulation can be used to further strengthen the experimental data.

(a)

(b)

Figure 2360b shows user interfaces of CDSEM systems.

(a)
	(c)
(b)	(d)
(e)
(f)
(g)
(h)