Bits, Bandwidth, and Beam Placement Accuracy: A Primer for the Discerning FIB/SEM User

Session

FIB Applications & EM Samples Prep Techniques in Physical Sciences

Authors

Mr. Michael Phaneuf (1), Dr. Ken Lagarec (1)

Affiliations

1. Fibics Incorporated

Keywords

SEM, FIB, bits, bandwidth, beam placement accuracy, DNL, INL, nanopatterning, calibration, artifacts, volume EM

Abstract text

The FIB/SEM user who intends to undertake sample preparation, 3D tomography (“Volume EM”) or nanopatterning can often achieve excellent results for “typical” workflows, but is sometimes faced with peculiar surprises: “why does my cross section look nice and smooth when I section 40μm wide, but it gets “bumpy” when I section 400μm wide?”  Similar effects are sometimes observed as the “write field” for nanopatterning is made larger, or if dwell times are made shorter, for a range of applications.  In 2016, the authors participated as part of a larger effort [1] to characterise the limitations of “patterning in an imperfect world” as they pertained to focused ion beam placement and control in applications relating to advanced circuit editing of integrated circuits.  In the current presentation, we will attempt to broaden the discussion to pertain to a wider range of applications that may be of interest to the audience, and to examine the impact of factors such as the number of bits in the digital-to-analog converters (DACs) used in scanning systems, introducing some more esoteric details such as the differential nonlinearity (DNL) and  integral nonlinearity (INL) of these DACs.  We will briefly cover key points regarding how beam deflection systems work in certain types of SEM and FIB systems, including the impact of magnification ranges when viewed from the perspective of the user, as well as from the perspective of the electronics, and discuss how the noise floor in these ranges can impact the end results.   

As users seek to scan more quickly, or more sparsely, in FIB and SEM systems, they may encounter the concept of bandwidth limits as they pertain to a charged particle microscope’s deflection system.  These limits will be examined from both a large signal and small signal standpoint, and from the perspective of “slew rate” and what that means in terms more familiar to the microscopist such as scan speed or dwell time and image pixel count.   Beam placement accuracy is a broad term that reflects whether the FIB or SEM beam gets to the location of interest “properly” and “sufficiently quickly” to allow the user to obtain an image that is an accurate representation of what is truly on the sample, or to accurately nanofabricate the desired structure.   As dwell time is reduced and/or the spacing between points is increased, it is important to understand which system calibrations are critical to obtaining the required performance, and how to recognize artifacts for what they are, as well as how to construct the experiment so as to improve robustness against the types of artifacts observed.

The various factors discussed above can combine in ways that negatively impact the outcome when users are pushing the performance of their FIB/SEM systems.  We will illustrate some of the problems we have encountered when undertaking among the more challenging applications during our careers to date, and where possible provide information and suggestions as to how one might overcome similar issues, or at least how we proceeded to mitigate issues when faced with these problems.

References

[1] “Patterning in an Imperfect World: Limitations of Focused Ion Beam Systems and their Effects on Advanced Applications at the 14nm Process Node”, Christopher M. Scheffler; Richard H. Livengood; Haripriya E. Prakasam; Michael W. Phaneuf; Ken Lagarec; ISTFA 2016: Conference Proceedings from the 42nd International Symposium for Testing and Failure Analysis.  pp. 382-390.   DOI:10.31399/asm.cp.istfa2016p0382