The semiconductor industry fabricates integrated circuits and microchips that are used in a myriad of applications
from communications to data processing and automotive sectors. To improve device performance and reduce cost, advanced
manufacturing techniques and new technologies are being continuously developed. Semiconductor fabrication requires
extreme level of cleanliness, as any contamination may result in device failure. For this reason, great care is taken
to identify contamination and defects, and eliminate their sources. Quick and accurate identification of these sources
coupled with feedback into the manufacturing process can help ensure a stable supply of products and increase
profitability through yield improvement.
Many of the advances in semiconductor devices result from the ever-shrinking "technology nodes," which
relate to the smallest features in the devices. Over just the last 20 years, feature sizes have been reduced by a
factor of ten, from >100 nm features to <10 nm features. While smaller means cheaper and faster, miniaturization
also means that particles and defects of smaller size can have large effects on the devices and must therefore be
controlled.
For example, when photomask blanks used for extreme ultraviolet lithography (EUV) processes are produced, the goal is
to reduce the number of particles larger than ~20 nm, on a 6-inch photomask blank, to zero. Initial contamination
inspection is done using light scattering to locate defects, but it cannot provide accurate information about defect
type (i.e. particle vs. indent). To further study the contamination, scanning electron microscopy coupled with energy
dispersive X-ray spectroscopy (SEM/EDS) could be used, but it is difficult to analyze such small-sized particles using
SEM. The other technique that can be used is transmission electron microscopy (TEM/EDS).
To perform TEM analysis efficiently, the type of contamination should be identified (particle or defect) and its
exact location confirmed. Atomic force microscopy (AFM) has become an indispensable tool used to easily identify very
small contaminates and indicate if the contaminate is a particle, a hole defect, or a bump, while providing the exact
defect location.
The Jupiter Discovery AFM from Oxford Instruments (Figure 1) has several features that set it apart from conventional
AFMs, thereby making it the instrument of choice for contamination detection and analysis.

Figure 1: Jupiter Discovery large-sample AFM
The Jupiter Discovery AFM uses Oxford Instruments' exclusive blueDrive photothermal tapping mode that ensures
that the quality of data is accurate and repeatable. Most importantly, the Ergo software has 3-point alignment and
coordinate correction functions that facilitate particle localization (Figure 2).
The Jupiter Discovery AFM can scan large areas of the sample (up to 100 µm) with high pixel count to detect
nanometer-sized particles or defects. At a conventional image pixel count of 512×512, each pixel in a 100 µm image
measures almost 200×200 nm square. However, at a higher pixel count of 4096×4096 pixels, each pixel represents an area
of only about 2.5×2.5 nm square.

Figure 2: Ergo software interface showing 3-point alignment using fiducial markers
Additionally, imaging on a Jupiter Discovery AFM is performed rapidly such that a high pixel count image, comparable
to the one shown in Figure 3C (20 µm image with pixel size of 20×20 nm), can be obtained in under 2 minutes which is
an important consideration for industrial productivity.
Furthermore, Ergo has automation and Autopilot functions. Automation enables the AFM user to collect data at multiple
sites (up to 1,000 sites) without requiring user interaction, whereas Autopilot is an advanced algorithm which
determines optimal imaging parameters automatically to make the data collection simple for every AFM user.
All these features, coupled with the low noise of the Jupiter Discovery AFM, means the detection of nanometer-sized
defects and particles can be done with speed and accuracy surpassing any other AFM.
An example of a particle contamination measurement is shown in Figure 3. The sample has been first analyzed using a
scattering technique and the general area where contamination was found was marked with optically visible indents. The
indents are visible in the AFM video camera view (Figure 3A) and are used as reference marks to determine the area to
be imaged by AFM. A survey scan of 30 µm was acquired and included the 3 indent reference points (Figure 3B). A
slightly smaller image of 20 µm with a much higher pixel count (Figure 3C) was acquired inside the area bordered by
the indents, in order to locate the contamination that was first detected by a light scattering technique. The image
clearly showed contamination, specifically a particle that, when imaged more closely (Figure 3D), appeared to be an
irregularly shaped particle 250 nm in size.

Figure 3: Localization and imaging of sample contamination using a Jupiter AFM. a) Optical view of the sample showing
indents (arrows), the AFM tip (red dot), and the outlines of the different scan areas for reference: 20 µm (blue), 30
µm (green), 100 µm (yellow). Scale bar is 20 µm. b) AFM image showing the 3 indents visible in optical image. c) AFM
image of the area where the contamination was detected. d) Zoom-in into the contamination area showing a particle of
~250 nm in diameter.
In summary, characterization of this sample with an Oxford Instruments Jupiter AFM provided information about the
type of contamination present, an irregularly shaped particle, as well as its exact location on the sample. The
particle's position coordinates can now be used to further analyze it using SEM/EDS or TEM/EDS techniques.
The Jupiter AFM has become an indispensable tool for defect inspection used to accurately identify and provide the
exact location of nanoscale contaminates with high speed and accuracy.