SANS and neutron diffraction study of ausformed EUROFER97/2 ferritic/martensitic steel

(cid:45) Small(cid:45)angle neutron scattering (SANS) and neutron diffraction have been utilized for micro(cid:45)structural characterization of Eurofer97/2 heats submitted to thermo(cid:45)mechanical treatments, with the aim to investigate macroscopic material volumes and to complete the local information provided by scanning electron microscopy. The measurements were carried out at MLZ in Garching, utilizing for SANS also polarized neutron beam. The investigated samples had been submitted to austenitization and tempering at different temperatures, as well as to double austenitization and to (cid:179)(cid:68)(cid:88)(cid:86)(cid:73)(cid:82)(cid:85)(cid:80)(cid:76)(cid:81)(cid:74)(cid:180) , that is austenitization followed by hot rolling. For most of the examined treatments, nearly identical SANS cross(cid:45)sections were measured, close to the one of Eurofer97/1. In the ausformed sample the nuclear SANS cross(cid:45)section is also nearly identical to the other samples but the magnetic one is one order of magnitude higher. Furthermore, over a wide experimental interval its nuclear(cid:45)magnetic interference, measured by polarized SANS, is of opposite sign with respect to the non(cid:45)ausformed samples. Neutron diffraction measurements provided strong evidence that the origin of such effects is due to the presence of the non(cid:45)magnetic austenite phase in the ausformed sample, with an estimated volume fraction of 0.17 ± 0.02; within the experimental resolution, it disappears after subsequent tempering.


Introduction
Ferritic/martensitic steels currently appear as one of the most realistic options as structural materials for DEMO technology; however, high temperature mechanical properties, creep resistance in particular, need to be improved for optimizing in service behavior at 650°C or at higher temperatures [1,2]. Within this frame, a series of innovative Eurofer97/2 heats has been developed; the most favorable elemental compositions and preliminary thermo mechanical treatments have been systematically investigated [3 5]. Namely, with respect to Eurofer97/1 the C, W and N contents have been varied to different extents in the different heats, in order to reduce M23C6 precipitation, promote V rich precipitation, contribute in optimizing grain size [3]. Optimization of the preliminary treatment concerned austenitization and tempering temperatures, the effect of a double austenitization and the effect of a combined thermo mechanical treatment consisting in a short hot rolling stage at 650°C or 750°C after austenitization. This treatment, ausforming , is aimed to tune carbide precipitation in such a way to favor pinning of dislocations with beneficial effects on creep resistance [6 9]. The produced heats were mechanically characterized by fatigue and creep tests; their grain size micro structure was investigated by scanning electron microscopy (SEM). Preliminary X ray diffraction measurements were also carried out [10]. Promising results were obtained concerning creep resistance and impact properties, but no clear correlation could be established between the achieved improvements and the changes introduced in chemical composition and thermo mechanical treatments. Consequently, small angle neutron scattering (SANS) and neutron diffraction measurements were carried out on some of these samples in order to contribute in characterizing their micro structure and trying to understand the effective role of the parameters selected for each heat.
In fact, neutron beams can probe magnetic micro structural features and allow for an investigation of massive samples, with typical volumes as large as 0.1 cm 3 . Therefore, the obtained results can be more directly compared with those of mechanical testing and are in any case more representative of bulk properties of the investigated samples. These two techniques are also complementary to each other: SANS provides the distribution of defects, such as precipitates or micro voids, while neutron diffraction allows to indentify the crystallographic phases present after the different metallurgical treatments. The micro structure of the investigated materials is very complex and only a limited number of samples was investigated, namely only one ausformed sample. Therefore the results presented here below are not intended as a conclusive study, but as a first, necessary step to contribute in the metallurgical characterization of such Eurofer97/2 heats.

Material characterization
The most significant SANS and neutron diffraction effects were detected in the three Eurofer97/2 samples listed in Tab. I. A sample of Eurofer97/1 submitted to standard treatment was also measured, both to check the changes associated to the different elemental compositions of the new heats and to compare with previous SANS measurements of this same sample [11]. SEM observations of these Eurofer97/2 samples are included in Ref. [3], showing grain size distribution and nitride precipitates. The samples utilized for the SANS measurements were prepared in the shape of rectangular platelets, approximately 1 cm x 1 cm x 0.7 mm in size. Each of them was cut from the original ingot, then its surfaces were mechanically polished to avoid spurious SANS effects. After ausforming, a grain texture with preferential direction is produced in the heats as a consequence of the hot rolling stage. In preparing the first samples for the SANS measurements it was not possible to cut all of them with the same orientation with respect to the rolling direction, therefore some uncertainties are present in the 2D SANS data, as discussed in sec. 3 here below.
Several other samples were investigated by SANS, submitted to double austenitization and to creep at 650°C for 1650 h, under a load of 100 MPa: within the experimental uncertainties, no significant changes were observed among their respective SANS cross sections and also comparing them to Eurfer97/1. The same samples utilized for the SANS measurements were utilized also for the neutron diffraction ones.

Experimental techniques
General information on the SANS technique is available in Ref.s [12,13]; the scheme of one of the two utilized SANS instruments is shown in Fig. 1. For studying magnetic samples, an external magnetic field of at least 1 T must be applied to saturate their magnetization and measure the nuclear and the magnetic SANS components separately. In fact, for such samples the total SANS cross where is the azimutal angle on the detector plane. Parallel to the magnetic field ( = 0°) the nuclear SANS cross section is measured, perpendicular to it ( = 90°) the sum of the nuclear and magnetic ones is measured; they are usually determined by selecting on the detector plane angular sectors 15 ° wide around these two directions. Their ratio R(Q) is related to the composition of the micro structural defects, 2 ) ( difference in neutron scattering length density (nuclear and magnetic respectively) between the observed nuclear and magnetic defects and the matrix. between the scattering vector Q and the magnetic field H, the cross sections measured in the spin directions parallel (+) and anti parallel ( ) to the applied magnetic field with polarization 100% can be written as: For = 0° N 2 is obtained. For = 90° the nuclear plus magnetic SANS component, N 2 + M 2 , and the nuclear magnetic interference term, NM, are obtained respectively as follows: The diffractograms of the measured samples were normalized for incoming neutron flux, and corrected for angular coverage and detector efficiency. However, given the restricted available Q range, a full refinement of the obtained data was not possible.

Results and discussion
In most of the examined Eurofer97/2 samples, submitted solely to preliminary thermal treatments, the SANS cross sections are nearly coincident among each other and close to the one of Eurofer97/1, independently of austenitization or tempering temperature. An example is provided in Fig. 2 The Eurofer97/2 ausformed sample shows quite different features. First, its 2D SANS intensity distribution is significantly more anisotropic compared to Eurofer97/1, as shown in Fig. 3. The magnetic field (vertical in the plane of the figure) is due to the fact that the ausformed ingots are textured and the two respective samples were not cut with the same orientation with respect to the rolling direction. Furthermore, the nuclear SANS components of these three samples are nearly coincident (Fig. 4 a) within the uncertainty associated to their different orientation, but the nuclear plus magnetic SANS component of the ausformed one is one order of magnitude higher compared to Eurofer97/1 and to Eurofer97/2 tempered after ausforming (Fig. 4 b). This reflects in the R(Q) ratio of the ausformed sample, compared in Fig. 5 with the Eurofer97/1 sample: for Eurofer97/2 ausformed R(Q) is nearly 10 times higher up to 0.1 Å 1 approximately, then drops to same value range as Eurofer97/1.
The polarized SANS measurements were carried out on the ausformed sample and on the Eurofer97/2 sample submitted to double austenitization. Fig. 6 a b shows the nuclear magnetic interference term

Conclusions
Finally, it is stressed that the obtained results although preliminary constitute a unique contribution to understand the micro structural effect of ausforming, obtainable by no other experimental method.     (blue, circles), for Eurofer97/2 (red, squares) and for Eurofer97/2 (green, triangles).     Calibrated SANS cross sections parallel (a) and perpendicular (b) to magnetic field for Eurofer97/1 (blue, circles), for Eurofer97/2 (red, squares) and for Eurofer97/2 (green, triangles).