Keeping analytical tools in line with NTRS goals: An industry challenge
01/01/1999
TEST/METEROLOGY
Keeping analytical tolls in line with NTRS goals: An industry challenge
Thomas J. Shaffner, National Institute of Standards and Technology, Gaithersburg, Maryland
As we enter the gigabit DRAM era, stringent demands for higher-performance materials characterization are prematurely pushing sophisticated and costly metrology tools into the manufacturing environment. In a climate of limited resources, one must learn to identify which of these tools are in the critical path of IC development and manufacturing. The SIA National Technology Roadmap for Semiconductors (NTRS) provides a framework from which to start [1, 2]. As useful as these roadmaps are, the analytical laboratory manager must still plan his or her specific laboratory plan of action relative to cycle time, available resources, and critical needs in pursuit of overall fab productivity optimization, yield maximization, and alignment with the NTRS.
 |
Figure 1. Simplified IC process flow, a) without and b) with analysis laboratory support. TCAD is the acronym for technology for computer assisted design.
The role of the analytical laboratory
Reliable measurements provide essential feedback for new process development and fine-tuning of the product line. Materials and device characterization thus continue to evolve in a synergistic partnership with microcircuit fabrication. To understand more about the nature of this interaction, consider the simple feedback process flow outlined in Fig. 1a. Circuit designs are improved based on the electrical test performance of real devices, which in turn are fabricated according to a design layout. The objective is to increase product functionality, reliability, and throughput. The analytical laboratory is cost-effective because it provides hard data on which critical fab and design decisions can be based. In effect, it provides other avenues for the feedback, as illustrated in Fig. 1b.
Strengthening the links
The links between the design, analysis, and production stages are a critical part of this process and must be kept strong for the laboratory to remain cost-effective. For example, the scanning electron microscope (SEM) established strong links during its evaluation in the 1970s. It has since become an indispensable tool and has been integrated into the "test" function itself.
When weakness occurs in the links, due to poor in-house cycle time, for example, engineers may outsource the analysis, but more frequently they opt to do without. This is perhaps the more insidious effect because it forces the system to revert to the simple flow in Fig. 1a, and the regression typically goes unnoticed. This leads to uninformed decision-making and ultimate loss in production or ramp-up in a new fab.
Most analysts will agree that there is a qualitative correlation between cycle time (Tc) and the number of samples submitted (Ns) to the laboratory:
Ns (1/Tc (1)
Tc is influenced by the precision and accuracy required of the measurement, diagnostic findings, equipment downtime, the number of operating shifts, and a number of other factors.
 |
Figure 2. Characterization tool utility, showing evolution to the critical path of manufacturing. The acronyms are defined on p. 40.
Equation 1 simply reflects the fact that cycle time is a major driver for the links. In the extreme, some choose to acquire questionable data (wrong technique) simply because the right technique is shut down for instrument repair, or otherwise unavailable. Obviously, cycle time can be abused, leading to weakening of the links.
To gain more understanding, we can examine the historical trends of characterization tool utility (Fig. 2). New analytical tools evolve through the novelty, useful, and critical path phases. In the industrial environment, these are most relevant to research, process development, and manufacturing objectives, respectively.
Figure 2 indicates that OM, SEM, FIB, and TXRF are critical path tools. What they have in common are reasonable cycle time, sensible cost, and ranking as essential (critical path) test equipment. These are the important drivers that strengthen links between the analytical laboratory and IC process flow. It is helpful to examine how these are shaping the development and applications of evolving techniques.
Transmission electron microscopy
Although the SEM is in the critical path of the fab, it is no longer able to image accurately the thickness of the industry`s ultrathin gate dielectrics. Typically, these are 10-nm-thick oxynitride composites, with <5% (only one or two monolayers) uncertainties tolerated. These are designed to increase the dielectric constant as well as minimize dopant diffusion into the gate.
Today, transmission electron microscopy (TEM) is the only tool capable of delivering the required resolution, so it is clearly in the critical path [3]. TEMs are rather expensive instruments ($1-1.5 million, depending on accessories), so resources will ultimately limit how many may be installed and maintained. Difficulty in preparing the required 10 to 100-nm-thick cross-sections, however, is a serious problem that directly limits cycle time.
The challenge is to resolve the sample preparation problem as soon as possible to support the NTRS timetable. If this cannot be done, engineers may ultimately forgo the analysis altogether, or even worse, continue to incorporate unreliable thickness measurements from SEM images distorted by electron scattering at the interface boundaries.
Focused ion beam sectioning
Focused ion beam (FIB) milling offers an elegant way to prepare TEM samples, but more improvement is needed before this can become a fully critical path tool in the fab. The instrument is capable of carving a 10 to 100-nm-thick cross-section from any select area on a full wafer inside the cleanroom. Current issues are damage to the fragile section, Ga+ ion implantation from the probe, and mounting of the film onto a TEM grid [4]. FIB is likely to become a critical path enabling tool for select-spot microscope imaging.
SIMS-based methods
Secondary ion mass spectrometry (SIMS) is considered a critical path tool by some because it is presently the workhorse tool for profiling dopants and impurities in small junctions and contacts. Since it is often difficult to use, SIMS has not found a comfortable place in the fab, but there is pressure for this to happen [5] in spite of poor spatial resolution (~0.6 ?m) and difficult calibration issues.
Variations of the SIMS theme include time-of-flight SIMS (ToF-SIMS) and quadrupole SIMS (Quad-SIMS). They specialize in the measurement of surface organics and impurities in insulating films, respectively. Data acquisition in ToF-SIMS is straightforward, but interpretations can be quite difficult. Normally, this would preclude evolution of the tool into a critical path mode, but when restricted to repetitive measurement of one or two signature peaks, it could become a surface contamination monitoring tool of higher sensitivity than TXRF. ToF-SIMS instruments are now available with coordinate-driven stages and cleanroom compatibility.
Template for alignment
There are perhaps as many different opinions on how to rank characterization tools as there are operators of these tools, although most agree that low cost, short cycle time, and critical path status are important in the industrial environment. When a tool is necessary to qualify ramp-up of a new fab, it has clearly reached the critical path. Optical microscopes and SEMs are obvious examples, but the situation is less defined as we project which tools will be required in the future.
 |
The table below summarizes how these tool evaluations may evolve relative to cost, cycle time, and critical path classification. The latter is perhaps the most important, as we have seen with the TEM example. In effect, we give priority to the critical path column, paying special attention to those tools coded white. Resources are applied here to reduce cost and cycle time. Second priority goes to tools shaded gray, and last (or none) to those coded black.
The emergence of revolutionary new tools is much more difficult to predict or manage. So far, we have only considered evolutionary improvements in existing techniques, which can be achieved according to milestone-driven programs. Breakthrough technologies rely more on an infrastructure that fosters innovation and cross-disciplinary activities in the fundamental sciences. Many characterization methodologies are rooted in Nobel Prize-quality research, including x-ray diffraction, TEM, Raman spectroscopy, XPS, and more recently, scanning tunneling microscopy. Research into new science and engineering is required to ensure continued impact of the analytical laboratory on semiconductor development and manufacturing.
 |
References:
1. The National Technology Roadmap for Semiconductors, The Semiconductor Industry Association, San Jose, CA, 1997. (The next version will be an international effort, the International Technology Roadmap for Semiconductors or ITRS.)
2. A.C. Diebold, "Metrology Roadmap: A Supplement to the National Technology Roadmap for Semiconductors," SEMATECH Technology Transfer Document #94102578 A-TR, 1994.
3. J. Mardinly, D. Susnitzky, D. Venables, "Transmission Electron Microscopy: A Critical Analytical Tool for ULSI Technology," in Characterization and Metrology for ULSI, D.G. Seiler, et al., eds., American Institute of Physics Conference Proceedings 449, Springer-Verlag, NY, 1998.
4. L.A. Giannuzzi, J.L. Drown, S.R. Brown, R.B. Irwin, F.A. Stevie, "Focused Ion Beam Milling and Micromanipulation Lift-Out for Site Specific Cross-Section TEM Specimen Preparation," in Specimen Preparation for Transmission Electron Microscopy of Materials IV, R.M. Anderson, S.CD. Walck, eds., Materials Research Society, Warrendale, PA, pp. 19-27, 1997.
5. W.R. Runyan, T.J. Shaffner, "Secondary Ion Mass Spectrometry," in Semiconductor Measurements and Instrumentation, 2nd edition, McGraw-Hill, NY, Chapter 15, 1998.
THOMAS J. SHAFFNER is the Materials Technology Group Leader in the Semiconductor Electronics Division at the National Institute of Standards and Technology. Technology Building 225, Room A305, 100 Bureau Road, Stop 8121, Gaithersburg, MD 20899-8121; ph 301/975-8009, fax 301/948-4081, e-mail [email protected].