Beam Broadening Effects in STEM/EDS Measurement of Radiation-Induced Segregation in High-Purity 304L Stainless Steel

^1- University of Michigan, Ann Arbor, MI 48109-2104,
^2- Metals and Ceramics Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831,
^3- Materials Science Division, Argonne National Laboratory, Argonne, IL 60439,
^4- Argonne National Laboratory-West, Idaho Falls, ID 83403

• Introduction
  
• Approach
  
• Experimental
  
  • Material and Irradiation Conditions
  • Intrumentation
  • Measurement Techniques
  • Sample Prep
• Results
  
• Discussion
  
• Conclusions
  
• Acknowledgements
  
• References
  

Radiation-induced segregation (RIS) is the spatial redistribution of elements during irradiation at defect sinks such as grain boundaries and free surfaces. RIS has been studied in a wide variety of materials and has been identified as a potential contributor to irradiation-assisted stress corrosion cracking (IASCC) of various nuclear reactor core components. . RIS is frequently measured using scanning-transmission electron microscopy with energy dispersive spectroscopy (STEM/EDS). STEM/EDS measurements, however, are influenced by the finite width of the beam and the degree of beam spreading as it passes through the specimen. Accurate determination of the grain boundary chemistry is important in understanding the role of RIS in IASCC.

Radiation-induced segregation (RIS) profiles are typically exceedingly narrow (usually less than 10 nm full width at half-maximum). For such narrow profiles, the incident probe size is a significant portion of the profile. Hence, the accuracy of the grain boundary composition measurements using STEM/EDS are dependent upon the volume of material in which x-rays are generated. This is depicted below in Fig. 1. Thus, all STEM/EDS measurements of grain boundary composition underestimate the actual amount of segregation since the surrounding matrix material is sampled as well. The volume of excited material is determined by both the incident electron beam diameter and the primary electron beam energy. However,increasing the beam energy or reducing the probe size reduces the excited volume of material.

Fig. 1. Schematic drawing of beam broadening effect.

Reducing the amount of excited volume leads to STEM/EDS measurements which more accurately reflect the actual grain boundary composition. In this work, the effect of beam broadening on STEM/EDS segregation measurements is assessed in a 304L stainless steel sample irradiated with 3.2 MeV protons at 400¡C to doses of 0.1 and 3.0 dpa. The STEM/EDS measurements are also compared to measurements made using Auger electron spectroscopy (AES) and model results.

Grain boundary segregation was measured using STEM/EDS in three different instruments with varying primary beam energy and varying incident spot size. To ensure that variations between grain boundaries and samples were minimized, measurements were taken

• on the same TEM samples
• on the same grain boundaries
• at approximately the same thickness

STEM/EDS results from all instruments results are compared to results of AES analysis of samples from the same heat and irradiations.

• Ultra-high purity 304L (Heat #1327)
• Composition determined by electron-microprobe analysis

• Irradiated with 3.2 MeV protons using a 1.7 MV Tandetron
• Dose rate = 7 x 10^-6 dpa/s to doses of 0.1 and 3.0 dpa.
• The damage profile created by protons is uniform over first 35 microns of particle range (~ 40 microns)
• Irradiation Temperature = 400^oC

• Irradiated samples were mechanically thinned from the unirradiated face using SiC paper to a thickness of approximately 150 microns
• 3 mm TEM disks were then cut using a slurry core drill cutter to minimize mechanical deformation.
• Thinned disks were jet thinned using a 5% perchloric acid, 95% methanol bath until perforation.
• Final thinning to electron transparency and removal of oxide films was performed using ion milling with Ar ions.
• Prior to analysis in the VG HB603Z instrument, samples underwent plasma cleaning using a South Bay PC-150 Plasma Cleaner

• Samples were analyzed in the Philips EM400T/FEG and Philips CM200/FEG in a double-tilt, liquid nitrogen-cooled Be specimen holder. The samples were kept at ~-128^oC during analysis to eliminated surface contamination by the focused beam.
• Analysis in the VG HB603-Z was done in a double-tilt Be specimen holder at room temperature. Plasma cleaning done prior to analysis minimized surface contamination.
• Boundaries were carefully tilted to be edge-on with the incident beam.
• Strong, low-order diffracting conditions were avoided to minimize complications associated with channeling and bremsstrahlung peaks.
• In-hole spectra were subtracted from all grain boundary and profile measurements to account for uncollimated radiation from the instrument.
• Intensities were converted to atomic concentrations using k-factors determined from matrix measurements.
• The same sample and same boundaries were analyzed in all three instruments. AES measurements were taken on samples from the same heat and irradiation batch. Details of the AES measurement techique can be found in Ref.6.

Grain boundary Cr and Ni measurements from the 0.1 dpa sample were taken using the Philips EM400T/FEG and Philips CM200/FEG are shown. The results of these measurements are shown to the right in Fig. 2. RIS profiles from the same boundary taken with the same two instruments are shown in Fig. 3. In both cases, the CM200/FEG measures a higher degree of segregation.

Fig. 2: A comparison of grain boundary measurements made on the 0.1 dpa sample by the Philips EM400T/FEG and Philips CM200 FEG.

Fig. 3: A comparison of grain boundary profiles made on the same boundaries of the 0.1 dpa sample by the Philips EM400T/FEG and Philips CM200 FEG.

Grain boundary measurements from the 3.0 dpa sample from all three STEM/EDS instruments and results from AES analysis are shown in Fig. 4 below. AES measures the most segregation while the EM400T/FEG measures the least. RIS profiles of Cr and Ni are shown in Figs. 5 and 6, respectively (below and to the right). The VG-HB603-Z gave the deepest profile of the three STEM/EDS instruments.

As illustrated in Fig. 4, AES analysis measures more segregation than any of the STEM/EDS instruments. However, AES should have slightly better resolution than the STEM/EDS instruments, since the escape depth of an Auger electron is about 7-10 monolayers. Further, AES analysis was done on an intergranularly fractured sample and only facets which showed no evidence of ductile tearing were analyzed. This measurement technique may inadvertently select those boundaries with the highest amount of segregation as those boundaries may be more likely to fracture intergranularly.

Fig. 4: A comparison of grain boundary measurements made on the 3.0 dpa sample by the Philips EM400T/FEG, Philips CM200 FEG, VG HB603-Z, and measurements made by AES.

Clearly, instrument type and beam broadening shows a significant effect on the measured segregation in both samples. As expected, the HB-603Z operating at 300 kV measures more Cr and Ni segregation than the CM200/FEG and the EM400T/FEG since less matrix material is excited by the higher energy incident beam which is smaller. The CM200/FEG measures more segregation than the EM400T/FEG, again due to the higher energy incident beam which experiences less beam broadening.

The segregation profiles measured by the HB-603Z are deeper than those measured by the CM200/FEG and EM400T/FEG. Further, the profiles measured by the Philips CM200/FEG are deeper and show more detail than those measured by the acquired with the EM400T/FEG Again, as less volume is excited, more segregation is measured and the amount of detail in the profile is increased.

As mentioned earlier, actual measurements will always underestimate the amount of segregation at a boundary due to beam broadening. Model calculations were compared to the experimental measurements in order to estimate the amount that segregation is underestimated due to beam broadening. A modified inverse-Kirkendall model developed by Allen was used.

Both modeled and measured grain boundary composition measurements are compared to the right in Fig. 7 as a function of dose. At low doses, model and measurement differ widely, but are quite close at 3.0 dpa. Simulated and shown in Fig. 8 for 304L SS irradiated to 3.0 dpa. The measured profiles compare quite favorably with the model results.

In Fig. 7, the segregation evolution of both measured and modeled compositions differ by very large amounts at 0.1 dpa. This difference at low doses is attributed to incorrect values of several parameters used in the model rather than errors in the measurements. However, at higher doses, model and measured values compare quite well. Model and measured profiles shown in Fig. 8 indicate that with increased beam energy and reduced probe size, STEM/EDS measurements are approaching the actual grain boundary compositions.

• The amount of segregation measured and the shape of the segregation profiles are dependent upon the primary beam energy and incide probe size.
• The order of sensitivity corresponds to the operating voltage of the electron column instruments ( HB-603Z > CM200 > EM400T ).
• AES is more sensitive than any of the STEM/EDS instruments, but has limited lateral spatial resolution.
• Since different instruments with different operating characteristics measure different amounts of segregation on the same boundaries, care must be taken when comparing experimental data taken with different instruments.

This research was supported by the U.S. Department of Energy under grant DE-FG02-93ER-12310, by the Associated Western Universities-Northwest under U.S. DoE grant DE-FG02-89ER-7552; by the Division of Materials Sciences, U.S. DoE under contract DE-AC05-96OR22464 with Lockheed Martin Energy Research Corp.; through the SHaRE Program under contract DE-AC05-76OR000333 with ORAU and also by the Division of Materials Science, U.S. DoE under contract W-31-109-ENG-38W with Argonne National Laboratory, Division of Material Science. The work involved herein was also facilitated through the use of the DoE2000 Materials Microcharacterization Collaboratory (http://tpm.amc.anl.gov/MMC) funded by the U.S.DoE/DMS-BES, DMS-EE, and CTR-MICS programs.

1. G.S. Was and P. Andresen, J.Metals 44 (4) (1992) 8.

2. D.L. Damcott, J.M. Cookson, R.D. Carter, Jr., J.R. Martin, M. Atzmon, and G.S. Was, Radiat. Eff. Def. Solids, 118 (1991) 383.

3. N.J. Zaluzec, B. Kestel, and D. Henriks, to be published in Proceedings of Microscopy and Microanalysis, 1997, Eds. Dimlich, Alexander, McCarthy, and Bailey.

4. E.A. Kenik, Scripta Metall. 21 (1987) 811.

5. J.I. Goldstein, in Principles of Analytical Microscopy, eds. D.C. Joy, A.D. Romig and J.I. Goldstein (Plenum, New York, 1986).

6.T.R. Allen, J.M. Cookson, D.L. Damcott, and G.S. Was, Evolution of Irradiation Microsctructure, Proceedings of the 1997 Fall MRS Meeting, eds. I.M. Robertson, G.S. Was, L.W. Hobbs, and T.D. de la Rubia (MRS, Pittsburgh PA, 1997) Vol 439, p 557-563.

Return to MMC Home Page