Nuclear Instruments and Methods in Physics Research B 268, 3269 (2010)

AFM characterization of model nuclear fuel oxide multilayer structures modified by heavy ion beam irradiation

M.E. Hawley^*, D.J. Devlin, C.J. Reichhardt, K.E. Sickafus, I.O. Usov, J.A. Valdez, and Y.Q. Wang

Materials Science and Technology Division, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, USA

ARTICLE INFO
Article history:
Received 9 October 2009
Received in revised form 2 June 2010
Available online 9 June 2010
Keywords:
Metal oxide films
Radiation effects
Heavy ions
Atomic force microscopy
Surface modification

Abstract
This work explored a potential new model dispersion fuel form consisting of an actinide material embedded in a radiation tolerant matrix that captures fission products (FPs) and is easily separated chemically as waste from the fuel material. To understand the stability of this proposed dispersion fuel form design, an idealized model system composed of a multilayer film was studied. This system consisted of a tri-layer structure of an MgO layer sandwiched between two HfO₂ layers. HfO₂ served as a surrogate fissile material for UO₂ while MgO represented a stable, fissile product (FP) getter that is easily separated from the fissile material. This type of multilayer film structure allowed us to control the size of and spacing between each layer. The films were grown at room temperature by e-beam deposition on a Si(1 1 1) substrate and post-annealed annealing at a range of temperatures to crystallize the HfO₂ layers. The 550°C annealed sample was subsequently irradiated with 10 MeV Au 3+ ions at a range of fluences from 5 × 10¹³ to 3.74 × 10¹⁶ ions/cm² . Separate single layer constituent films and the substrate were also irradiated at 5 × 10¹⁵ and 8 × 10¹⁴ and 2 × 10¹⁶, respectively. After annealing and irradiation, the samples were characterized using atomic force imaging techniques to determine local changes in microstructure and mechanical properties. All samples annealed above 550 °C cracked. From the AFM results we observed both crack healing and significant modification of the surface at higher fluences.

1. Introduction
2. Experiment
3. Results
References

1. Introduction

New composite dispersion fuel designs consisting of an actinide material imbedded in a radiation tolerant matrix are being explored to fill three critical needs: containment of fission products (FP), easy separation of FPs from spent nuclear fuel, and greater energy production [1]. To understand the stability of this proposed fuel form design under radiation conditions, idealized model systems were studied. In this study, a tri-layer oxide film structure was chosen to represent this potential model fuel. This type of structure allowed us to control the size of each layer, which represented the size of each component and the spacing between repeat layers.

The tri-layer film structure used in this work consisted of a thick layer (880 nm) of a surrogate fissile material sandwiched between two thinner layers (50 nm) of a radiation stable metal oxide matrix material [2]. The matrix material's role is to capture and accumulate the FPs and be easy to separate chemically from the fissile material. In this case, HfO₂ was chosen to serve as surrogate fissile material for UO₂ . Although MgO is known to experience void swelling under irradiation [2], it was selected as the matrix material partly because it can be removed from the sample in acidic solutions where its rate of dissolution is dependent on factors such as acid concentration, MgO grain size, and temperature [3]. The choice of these relatively simple oxides was also to facilitate the modeling effort, not covered in this paper. The films were post-deposition annealed to crystallize the HfO₂ layers, which is amorphous when deposited at room temperature. The onset of the crystallization to the monoclinic phase of HfO₂ starts around 500 °C and is fully developed by about 800° [4, 5, 6]. (1 1 1)-oriented MgO grows on Si(1 1 1) at room temperature [7], however, grown on the HfO₂ layer or the SiO₂ native oxide layer the film is polycrystalline.

The stability of the tri-layer film structure was tested by bombarding the samples in a flux of heavy ions at relatively high energies (10 MeV) and fluences (up to 4 × 10¹⁶). At these energies both nuclear and electronic stopping is expected. Resulting changes in topography, crystal structure, defects generation, interface mixing, and layer thickness were characterized using a suite of tools including Atomic Force Microscopy (AFM, also referred to a Scanning Force or Scanning Probe Microscopy) results presented here. AFM contributes nanometer to micron scale three-dimensional information, which includes changes in grain size and structure, surface roughness, surface morphology, and local variations in nano mechanical properties. The latter is achieved by monitoring changes in the oscillatory behavior of the cantilever probe as it scans across and in contact with the sample surface.

2. Experiment

Tri-layer films consisting of a 50 nm base layer of HfO₂ , an 800 to 880 nm layer of MgO, and a 50 nm top layer of HfO₂ were grown at room temperature by e-beam deposition on 4 00 Si(1 1 1) substrates. The room temperature growth resulted in a polycrystalline layer of MgO separating two amorphous HfO₂ layers. These thicknesses were chosen to represent the desired relative separation and quantity of the surrogate fissile material and the radiation tolerant material needed to trap the FPs. The substrates were used as received with a native oxide layer of about 2 nm [8]. After deposition, the wafer was cleaved into smaller pieces, which were then annealed in air for 12 h at temperatures between 200 °C and 1000 °C to determine the affect of annealing temperature on the film's microstructure and the crystallization of the HfO₂ layers.

The 550 °C substrate piece was further divided into pieces for a fluence-dependent irradiation study. These pieces were irradiated with 10 MeV Au 3+ ions in the Ion Beam Materials Laboratory at Los Alamos National Laboratory at fluences ranging from 5 × 10¹³ to 5 × 10¹⁶ ions/cm² using a DanFysik 20 kV high current ion implanter [9]. Single-phase films of the two component materials, MgO and HfO₂, were irradiated at 5 × 10¹⁵ ions/cm² and pieces of the bare substrate were irradiated at 8 × 10¹⁴ and 2 × 10¹⁶ ions/cm² for comparison with the irradiated tri-layer samples.

3. Results

The as-grown tri-layer film was relatively smooth (RMS = 7.3 nm), with 20 to 30 nm grains randomly oriented on the surface. For the annealed films, no increase in RMS roughness was observed until 800 °C, where the beginning of faceting of some of the grains was seen in the capping HfO₂ layer. However, even at the lowest annealing temperature, the grains had begun to coalescence into larger aggregates. All samples annealed above 500 °C had interconnected cracks in a dried mud pattern and, at the highest temperature, pieces of the film delaminated from the substrate.

Figure 1: Plot of RMS surface roughness as a function of fluence for the 550 °C annealed films irradiated with 10 MeV Au 3+ ions.

After irradiation, the AFM results revealed a roughing of the surface (shown in Fig. 1), increase in crack density, crack healing, and finally significant modification as the fluence increased (Fig. 2). Initially at the lowest fluences the crack density increased and the crack profile inverted. At 1 × 10¹⁴ ions/cm², the edges of the cracks became rounded out-of-plane. At ~5 × 10¹⁴ ions/cm² not only was the crack density greater, the edge rounding increased and depressions in the surface containing a central peak appeared. When the fluence was increased to 1 × 10¹⁵ ions/cm² the cracks disappeared, at least at the surface. The round depressions or craters were deeper and the central peaks more prominent. Note that these depressions are large (8-12 μ m in diameter) and possess well-defined ridges. The craters at all fluences were non-uniformly distributed across the sample surfaces. Interestingly, the height of its central peak did not correlate with the depth of its crater, that is very tall peaks were observed in both deep and shallow craters. Likewise relatively short peaks were found in deep craters. Further increase in fluence resulted in an increased density and deepening of the craters, although their diameters remained approximately the same. Even though the average depth of the craters did increased with fluence, there was variation in depth between craters on the same sample. At the highest density of craters, the ridges that defined their perimeters were no longer distinguishable and the RMS roughness increased to 20 times that of the sample irradiated at the lowest fluence. It is also important to note that the depth of the craters at the three highest fluences exceeded the 50 nm thickness of the top HfO² layer. At the highest fluence some of the craters were over 400 nm deep, i.e. 8 times the thickness of that layer.

Figure 2: 40 × 40 μm AFM profile images of the 550 °C annealed tri-layer films surface as a function of fluence. Note: the vertical scale for the two highest fluences is 100 and 250 nm compared to 50 nm for the rest of the images.

At intermediate fluences, the central peaks in the craters often protrude above the surface, extending as much as 150 to 300 nm from the base of the crater. Their measured contact angles ranged from around 6 to 8 degrees up to nearly 30 degrees, depending on the peak height.

Phase imaging revealed local changes in nano mechanical properties as well. The peaks and the areas surrounding the craters appear "softer" in the phase image than the base of the crater suggesting compositional variations across the surface. This is consistent with a non-uniform loss of HfO₂ from the top layer observed in TEM cross sectional data (not presented here).

Figure 3: 40 × 40 μm AFM images showing a comparison of the surface modification of the 550C annealed tri-layer film, a single MgO layer, and a single HfO₂ layer after irradiation with 5 × 10¹⁵ ions/cm² 10 MeV Au 3+ ions. Note the differences in RMS roughness and z vertical scales. The single layer films are the same thickness as in the tri-layer film.

A comparison of the radiation tolerance toward surface modification of the individual components of the tri-layer is shown in Fig. 3. Although there was a change in the grain structure of the MgO film, the RMS roughness did not increase after irradiation. However, this was not the case for the HfO₂ film, where the RMS roughness increased at least 10-fold. However, even with this large increase in roughness after irradiation, no craters were formed; they were only observed on the tri-layer sample.

One final observation, a couple of AFM phase images and local surface potential measurements (not shown), which were used to try to determine if subsurface cracks persisted as the fluence was increase, did reveal their existence but not in which layer, MgO or the bottom HfO₂ layer, or if they disappeared at the highest fluences.

4. Discussion

Figure 4: Plots of the coefficient of linear thermal expansion as a function of temperature for MgO and HfO₂ [10], and Si [11]. The left black line marks the deposition T, while the right is approximately 550 °C.

In order to understand the origin of cracking in the films annealed above 500 °C (753 K) one has to identify the probable origins. The first factor is the different thermal responses of the components that make up the tri-layer film and their substrate, the largest component. Fig. 4 presents plots of the linear thermal expansion coefficients, α, for MgO, HfO₂, and silicon as a function of temperature in degrees K over the relevant temperature range used in the annealing process [10, 11]. The black line on the left indicates the deposition temperature. The one on the right indicates the temperature where cracking was first observed. Clear from this figure is the increasing disparity in the linear thermal expansion properties between MgO and HfO₂ and the silicon substrate. At the film growth temperature MgO and silicon have similar values for α. However, by 550 °C (823 K) the MgO value is double that of the substrate. The situation is even worse when you compare the value for HfO₂ against MgO at that temperature. The factor is almost 8 times greater for MgO than HfO₂ .

The second potential candidate for driving cracking is the change in volume of the HfO₂ layers, particularly the trapped bottom layer, during the transformation from an amorphous phase to the monoclinic crystalline phase. Crystallization of HfO₂ is observed to begin around 500 °C and continue up to 800 °C [4, 5, 6], the temperature at which crystallites were observed in the AFM images. In addition, non-uniform crystallization at 550 °C would lead to non-uniform strain in the films. Both factors could also induce cracking. The situation is further complicated if heating the polycrystalline MgO layer lead to further crystal growth.

The increase in number of cracks at the lower fluences could reduce strain and aid dewetting between the MgO and HfO₂ layers, especially if these materials have significantly different surface energies. Accurate measurement of the contact angle at the crack edges are needed to help resolve this issue. However, the inversion of the crack edge to form rounded, raised ridges increasing with fluence is consistent with a dewetting mechanism.

The sudden onset of crack healing, at least in the top HfO₂ layer, when the fluence was doubled from 5 × 10¹⁴ to 1 × 10¹⁵ ions/cm² and the formation of craters with a central peak is harder to explain but appears to suggest a complicated intersection between competing mechanisms, dewetting and a non-uniform strain field. The non-uniform distribution of craters and crater sizes is consistent with a variation in forces driven by non-uniform structure and strain distribution.

The large size of the craters, increase in surface roughness, and appearance of a central peak, which often extends above the surface, does not point to a simple mechanism, especially if one considers that the depth of craters is larger than the thickness of the original top HfO₂ layer and approaches a large fraction of the tri-layer thickness. These dimensions require the loss or movement of a significant proportion of both materials in the tri-layer. Phase images also suggest that the inside of the crater wall is different from the central peak and surrounding surface material and could be a different phase. The lack of correlation between the crater depth and the central peak height might also point to their formation being decoupled in some way.

Irradiated of single-phase films of the two materials did not in either case result in the formation of these unusual structures. These experiments revealed totally different damage resistance. Little change except in grain size was seen in the MgO film. In contrast the HfO₂ surface roughness increased 10 times. Since neither formed craters with central peaks, this observation further points to non-uniform strain and dewetting mechanisms in the tri-layer structure grown on silicon.

We believe this is the first observation of crack healing and the formation of large circular craters with central peaks in these materials. The author could find no reference documenting the appearance of radiation induced large, 8 to 12 μm diameter, craters with central peaks protruding above the surface. Modeling efforts are under way to understand the mechanisms behind both the crack healing and the formation of the surface structures observed by AFM techniques.

Acknowledgment

This work was supported by a Los Alamos National Laboratory, Laboratory Directed Research and Development (LDRD) Grant.

References

[1]: G.D. Jarvinen, D.W. Carroll, D.J. Devlin, United States Patent Application No. S-100, (2004) 559.
[2]: K.E. Sickafus, L. Minervini, R.W. Grimes, J.A. Valdez, M. Ishimaru, F. Li, K.J. McClellan, T. Hartmann, Science 289 (5480) (2000) 748.
[3]: P. Raschman, A. Fedorockoa, Chem. Eng. J. 117 (2006) 205.
[4]: S. Nam, S.W. Nam, J.H. Yoo, and D.-H. Ko, Mat. Res. Soc. Symp. Proc. 716 B4.25.1 (2002).
[5]: H. Wang, Y. Wang, J. Feng, C. Ye, B.Y. Wang, H.B. Wang, Q. Li, Y. Jang, A.P. Huang, Z.S. Xiao, Appl. Phys. A 93 (2008) 681.
[6]: G. He, M. Liu, L.Q. Zhu, M. Chang, Q. Fang, L.D. Zhang, Surf. Sci. 576 (2005) 67.
[7]: T.L. Chen, X.M. Li, X. Zhang, W.D. Yu, X.D. Gao, Proc. of SPIE 5774 (2004) 361.
[8]: J.H. Mazur, R. Gronsky, and J. Washburn, Solid-State Phys. 37, page unknown (1983).
[9]: I.O. Usov, J.A. Valdez, J. Won, M. Hawley, D.J. Devlin, R.M. Dicerson, B.P. Uberuaga, Y.Q. Wang, C.J. Reichhardt, G.D. Jarvinen, K.E. Sickafus, Nucl. Instrum. Meth. Phys. Res. B 267 (2009) 1918.
[10]: Y.S. Touloukian, R.K. Kirby, et al. (Eds.), Thermal Expansion - Metallic Elements and Alloys. Thermophysical Properties of Matter: The TPRC Data Series, IFI/Plenum, New York, 1975.
[11]: H. Watanabe, N. Yamada, M. Okaji, Int. J. Thermophysics 25 (1) (2004) 221.

File translated from T_EX by T_THgold, version 4.00.
Back to Home