Snoek-Type and Zener Relaxation in Fe-Si-Al Alloys

Abstract

Carbon-containing Fe - Si and Fe - Si - Al alloys were studied with respect to the carbonrelated Snoek-type and Zener relaxation using different mechanical spectroscopy techniques. In all alloys the temperature-dependent profile of the Snoek peak, relative to that in pure iron, is modified on its high-temperature side by the substitutional atoms. At least two components, an Fe - C - Fe (which correspond to C atom jumps (diffusion) in areas where it is surrounded by Fe atoms only) and Fe - C - Me peaks, where Me = Si, Al, can be distinguished in the Snoek-peak profile. In both binary Fe - Al and Fe - Si and ternary Fe - Si - Al alloys, a higher annealing temperature prior to quenching leads to an increase in the Fe - C - Fe and a decrease in the Fe - C - Me component of the Snoek peak. Heating to 1173K and above often lowers the peak height due to thermal vacancies. Low-temperature (<670K) ageing of quenched Fe - Si - Al and Fe – Si specimens reduces both the Fe – C - Fe and Fe – C – Al / Si peaks. Ageing at T > 670 K changes the temperature- as well as the amplitude-dependent parts of internal friction due to a redistribution of carbon between solid solution and dislocations. Both the Snoek-type peak height and the dislocation mobility – as can be concluded from the slope of the amplitude-dependent internal friction – increase, and a new peak appears at temperatures higher than that of the Snoek peak, which probably is a Snoek-Köster peak resulting from the motion of weakly pinned dislocations. A Zener peak appears if the concentration of substitutional atoms is > 6 at. %. The Zener peak relaxation strength is much lower in ternary alloys than in the binary ones probably due to mutual compensation of elastic distortions in presence of Al and Si atoms which are bigger and smaller, respectively, than Fe atoms.

Interactions between solute atoms in Fe–Si–Al–C alloys as studied by mechanical spectroscopy

“…Damping Q −1 and resonance frequency f of torsional or flexural vibrations were measured in two different frequency ranges, during heating with 1 K/min between room temperature and 873 K, using an inverted torsion pendulum (0.5-3 Hz) at Tula State University and different vibrating-reed set-ups (150-3000 Hz) at Braunschweig University. We will focus here only on internal friction in the known temperature-frequency ranges of the Snoek-type and Zener relaxations; microstructural characterization, elastic modulus data, and full damping spectra including other effects, as well as more experimental details, can be found elsewhere [18,[21][22][23][24]. …”

“…The Zener relaxation had been studied previously mainly in binary Fe-Al and Fe-Si alloys [25,26], but first results on ternary Fe-Si-Al were also considered briefly [23,24]. Being located at ∼900 K for frequencies of 200-500 Hz (e.g., in Fe-Al [22]), close to the upper temperature limit of our vibrating-reed equipment, the Zener peak is more readily studied using low-frequency techniques.…”

“…Snoek-and Zener-type effects have been found in both Fe-Si(-C) and Fe-Al(-C) alloys, and in particular Fe-Al has been studied extensively under many different conditions (see review [17]). With increasing Al content in Fe [18], and similarly in some Fe-Si alloys [19,20], the carbon Snoek-type relaxation can be phenomenologically divided into a "pure iron" Snoek peak (abbreviated as S Fe [18] or Fe-C-Fe [21][22][23]) being gradually reduced without changing its position, and a broader "interstitial-substitutional" (S Fe-Me [18] or Fe-C-Me [21][22][23]; Me = Al, Si) peak 1 growing at its high-temperature side and then shifting further to higher temperatures. Both peaks coexist from less than 3 to about 12 at% Al, where the "pure iron" component is finally suppressed.…”

“…As these compositions are clearly in the "concentrated" range where a quantitative theoretical treatment is difficult, various parameter variations are necessary to figure out the main characteristics of the relaxation processes. Instead of quantitative peak deconvolution partially attempted elsewhere [23], we will focus here in more brevity on some basic questions of solute interaction on a simpler, qualitative level.…”

“…When we adopt here the notation Fe-C-Fe/Fe-C-Me from preceding work[21][22][23], this should not be misunderstood as a description of some "triplet" point defect. The relaxing point defect is in both cases the interstitial C atom, which behaves in one case as if it were surrounded only by Fe atoms, but in the other case shows a clear influence of mixed Fe/Me surroundings.…”