Eddy current methodology in the non-direct measurement of martensite during plastic deformation of SS316L

1 Introduction

Austenitic steels are widely used in various industrial applications due to their high mechanical strength, corrosion resistance, and the stability of their non-magnetic properties, particularly in electrical and energy sectors. In these applications, maintaining the stability of paramagnetic austenite is crucial, as the formation of a ferromagnetic martensitic phase can occur under operational loads, both mechanical and thermal. The stability of austenite is primarily influenced by its chemical composition and the content of stabilizing alloying elements.

Austenitic steels are susceptible to martensitic transformation under plastic deformation conditions [1], and this process is more intensive at low temperatures [2]. It has been reported that in the case of 304 steel, ε-martensite (hexagonal close-packed [HCP], paramagnetic) forms at the beginning of deformation and reaches a peak value at about 5% of tensile strain. After exceeding this value, the volume fraction of this phase decreases to almost zero at 20% of plastic strain. On the other hand, the volume fraction of α′-martensite (body-centered cubic [BCC], ferromagnetic) increased steadily with deformation. At higher strain values, the α′ phase was the only martensite phase present in the material volume [3]. The transformation is accompanied by a change in the magnetic properties of the steel, which allows the use of magnetic techniques to assess its effects [4].

Similar results were reported for the AISI 316 steel deformed during rolling and subsequent tension test [5]. These properties are of great importance in the case of parts manufactured using additive technologies, because during the printing process, ferrite is precipitated at the grain boundaries of austenitic steel, the share of which reaches 6% [6,7]. Since the ferrite volume fraction in the structure of austenitic steel can adversely affect fatigue strength [8,9] it is advisable to control its content. The formation of martensite can be detected and evaluated by X-ray diffraction (XRD) [10] neuron diffraction [11], electron backscatter diffraction (EBSD) [12], magnetic methods including the eddy current method [13], metallography, and other techniques [14]. Monitoring the ferrite content released during martensitic transformation is crucial for evaluating the corrosion resistance of steel. Corrosion studies have demonstrated that martensitic transformation, induced by cold forming processes and cyclic loading conditions, significantly affects the corrosion resistance of AISI 316 austenitic stainless steel [15,16]. The formation of martensite is highly dependent on factors such as chemical composition, austenitic grain size [17], deformation temperature [18], and other variables. One should highlight that eddy current method is particularly sensitive to changes in the electrical resistivity and magnetic permeability of materials, which are influenced by variations in chemical composition, hardness, mechanical strength, and the degree of cold work. Consequently, this technique provides an indirect measurement of the overall physical and chemical properties of a material. It has been shown to be sensitive enough to detect phase transformations in stainless steels [19].

Plastic deformation of conventional and additive manufacture (AM) specimens may significantly influence their magnetic properties due to deformation-induced martensitic transformation [20]. The metastable ASSs exhibit tendency toward diffusion-free phase transformation during plastic deformation, especially when temperatures tend to 0 K [21]. In this process, initial austenite with a face-centered cubic (γ) structure transforms into martensite ε with a HCP structure and subsequently into martensite α′ with a BCC structure. It is worth noting that the transformation can also occur directly and, during tension at room temperature, mainly martensite α′ is observed [21,22].

Therefore, this article presents an investigation into the stability of austenite in 316 steel samples, produced from both rolled sheets and via AM, under static tensile loading conditions.

2 Methodology

The tests were conducted on two types of flat specimens: conventional (R) 316L and AM 316L.

The conventional specimens were fabricated using electrical discharge machining from a commercial 316L stainless steel sheet (2 mm thickness) with the tensile axis parallel to the rolling direction (RD). The semi-products were supplied in cold-rolled and annealed conditions (Figure 1). In such a form, the materials are typically used for structural components working in cryogenic conditions, down to the temperature of superfluid helium, 1.9.

Figure 1

Ultrafuse 316L filament, created by BASF for producing stainless steel parts through 3D printing, was employed (as an example of the possibilities of FFF technology).

The AM process utilizes Ultrafuse 316L, a metal–polymer composite filament specifically engineered for fused filament fabrication (FFF). This filament comprises over 80% metal powder, bound by a unique polymer matrix that is fully removed during debinding and sintering. The result is a high-strength, corrosion-resistant part composed entirely of 316L stainless steel. The printing process was conducted using an Ultimaker S5 printer with the following optimized parameters: a glass print platform coated with Magigoo PRO metal adhesive, a 0.4 mm hardened steel nozzle, platform temperature of 110°C, and printing temperature of 230°C. No print cooling was applied. The layer thickness was set to 0.15 mm, with a print speed of 25 mm/s and a fill density of 100%.

Debinding followed the Badische Anilin-und Soda-Fabrik (BASF) process, using a Nabertherm NRA 40/02-CDB debinding furnace at 120°C with HNO₃ at a concentration of >98%. Nitric acid was fed at a typical rate of 30 L/h, along with nitrogen purging gas at 500 L/h, ensuring safe and consistent processing. The debinding phase concludes upon reaching a minimal mass loss of 10.5%.

The final stage, sintering, was conducted in a 100% pure, dry hydrogen atmosphere. The sintering cycle involved a controlled temperature ramp as follows:

• From room temperature to 600°C at a rate of 5 K/min, followed by a 1 h hold at 600°C.

• From 600 to 1,380°C at a rate of 5 K/min, followed by a 3 h hold at 1,380°C.

• Gradual furnace cooling to complete the process.

The initial microstructure (Figure 3) of the as-received materials in all cases consists of equiaxial grains with recrystallization twins present in some of the larger grains. The average grain size in the conventional 316L steels, equaling 19.2 µm, while FFF 316L has average grain diameter of 37 µm. All steels have a heterogeneous distribution of grain sizes.

Figure 2

AM 316L specimen: (a) printed part and (b) final part after sintering.

Figure 3

As-received microstructures of the investigated materials: inverse pole figure maps in the top row and grain size distributions (equivalent circle diameter) in the bottom row.

Each specimen was divided into 14 areas, 10 of which were in the gauge section (Figure 4). In each area, the ferrite content and electromagnetic parameters were measured.

Figure 4

Specimen cut from sheet metal R 316L (a) and printed AM 316L (b).

The following tests were conducted to measure the ferrite content formed during the plastic deformation of specimens:

• Manual measurement of the ferromagnetic martensitic phase content, using a DMP30 ferritoscope.

• Manual measurement of impedance parameters (phase angle and amplitude) using the lift off technique and NORTEC 600 eddy current flaw detector.

• Automatic measurement of the amplitude and resonance frequency of the eddy current signal using developed setup based on the WIROTEST device made by WIT Łukasiewicz (Figure 5).

• Profile of the ferrite phase content using XRD and XRDynamic 500 diffractometer.

Ferrite phase content, phase angle, amplitude, and resonance frequency were measured in all specimens using the above-mentioned techniques before and after static tensile tests performed up to 8% in plastic deformation. Subsequently, one series of specimens was subjected to five loading–unloading tensile cycles, to obtain 1% deformation in the first cycle, 2% in the second, 4% in the third, 6% in fourth, and 8% of plastic deformation in the last, fifth cycle. After subsequent loading cycles, the ferrite content and impedance parameters were measured in each measuring area. Measurements with the WIROTEST setup (Figure 5) and XRD tests were performed after loading. The results obtained with these techniques refer only to the state before and after plastic deformation to a certain level.

Figure 5

Automated system for measuring the amplitude and frequency of the eddy current signal by WIROTEST.

The feritscope, widely used due to its speed and non-destructive nature, provides reliable results when calibrated appropriately [23]. The feritscope measures ferromagnetic microstructures in materials, primarily ferrite but also α′ martensite, which has a ferromagnetic BCC structure. It cannot measure ε martensite, as this paramagnetic, HCP phase does not generate a detectable magnetic signal. ε martensite is transitional, reducing with plastic strain and converting to α′ martensite at higher strains [23,24].

The feritscope detects voltage proportional to magnetic properties, but differences between ferrite and martensite magnetic permeability necessitate conversion of the readings. Talonen et al. [23] established a widely cited proportionality coefficient of 1.7 for martensite content, supported by other researchers [24].

Talonen et al. [23] demonstrated a linear relationship between martensite content and feritscope readings up to 55%, with a bilinear approximation beyond this. The formula for martensite content (%) based on feritscope readings is

The martensite content was measured using the feritscope Fisher FMP 30, an instrument that detects all magnetic phases, including ferrite. The martensite content for ASS can be calculated from the feritscope reading according to the formula presented by Talonen et al.

Thus, in the manuscript the ferrite content measured by means of commercial NORTEC 600 flaw detector was verified by feritoscope and WIROTEST measurement, as well as by XRD analysis.

3 Results

At the beginning of experimental procedure, three specimens of 316L steel (produced via rolling – R and AM) were subjected to static tensile loading at 77 K, inducing a similar plastic strain in AM and R. For these specimens, the results are reported for the states before and after deformation (Figures 7 and 8). For one pair of samples (AM and R), ferrite content measurements and commercial flaw detection data were collected after each cycle of tensile loading (Figures 9 and 10). In the case of AM 316L some initial ferrite grains are present at austenite grain boundaries as a result of high temperature sintering process, while the R316L sample presents typical single-phase microstructure (Figure 3).

Figure 6

Profiles of ferrite content measured along the symmetry axis for the rolled (R) and printed (AM) specimens after plastic deformation at 77 K, obtained by XRD method.

Figure 7

Profiles of ferrite content (a), phase angle (b), and impedance amplitude (c) changes measured for the rolled (R) and printed (AM) specimens, before and after deformation at 77 K.

Figure 8

Resonance frequency (a) and eddy current voltage amplitude (b) obtained by WIROTEST in the scanning mode. Ferrite content profiles for five states of plastic deformation, without elastic stresses (measurement at force 0) (c).

Figure 9

Ferrite content profiles (a) and eddy current phase angle profiles for five states of plastic deformation measured during unloading (b) and loading (c).

Figure 6 presents the ferrite phase content measured using the XRD method for two specimens manufactured by different techniques. The graphs illustrate the profile of ferrite content percentages at successive measurement points along the sample axis. To achieve a substantial amount of ferrite, experiments were carried out at 77 K (liquid nitrogen temperature). At this temperature, the deformation-induced phase transformation in austenitic steels occurs with much greater intensity compared to similar deformation conducted at room temperature [21]. This allowed for a comparison of ferrite distribution between conventionally manufactured and additively manufactured 316L steel. One can observe increased ferrite phase content in the AM specimen, both in the non-plastically deformed area (grip part) and along the entire length of the gauge area (Figure 6).

Figure 10

Comparison of ferrite content profiles for two types of SS316L after deformation obtained by XRD (line) and feritoscope (points).

The result shown in Figure 8a shows the ferrite contents in previously specified areas of the specimens measured by using ferritoscope. Uniaxial tensile tests were carried out at 77 K. Similar to the XRD measurements, an increased ferrite phase content was observed in the printed specimens, both, in the initial state and after deformation. The similar trend was observed during phase angle measurements (Figure 7b) and impedance amplitude measurements (Figure 7c) executed by applying an eddy current probe at the established measurement points. These are the results of the sintering process of AM specimens (Figure 2).

Subsequently, measurements of the resonance frequency and the voltage amplitude of the eddy current signal induced by the WIROTEST setup developed by the authors were presented. Figure 8a shows the frequency values obtained in automatic mode for two rolled and two AM specimens after plastic deformation. Since the WIROTEST setup also measures the voltage amplitude of the induced eddy currents, Figure 9b presents the results of these measurements for two pairs of specimens after plastic deformation. The eddy current amplitude changes along the axis of plastically deformed specimens indicating the possibility of detecting deformed areas and related sensitivity to the manufacturing method and the corresponded ferrite content. Furthermore, in the case of 316R, a local increase of the amplitude value was observed. Plastic front propagation is observed in 316L, but only at 77 K. This peak is related to the presence of the martensitic transformation front [22]. The results obtained for five static load cycles are presented only for the rolled 316L steel specimen. They include measurements of the ferrite content and the phase angle value after each cycle, in the unloaded stage (force 0) and at the maximum load applied (force max). The aim of such interrupted test was to assess the effect of elastic deformation on the measurable parameters. Figure 8c shows the increase in ferrite content along the specimen axis in subsequent cycles, i.e., as a function of increasing strain, for the unloaded variant. Figure 9a presents the changes in ferrite content measured at the maximum force of subsequent load cycles. With the increase of the plastic deformation level, differences in ferrite content are visible for the unloaded and maximum load conditions (Figure 9b and c). It is so called Villari effect [25,26].

4 Discussion

All applied research techniques demonstrate sensitivity to changes in the content of ferrite which result from manufacturing processes and deformation-induced phase transformation. The results of tests conducted under cyclic loading indicate the potential for identifying elastic deformation in both phases: austenite as well as in ferrite (Figures 8 and 9). This implies that flaw detector measurements, commonly used for detecting potential cracks, can also be utilized to identify and monitor the evolution of deformation-induced phase transformations in austenitic stainless steels. In the case of impedance amplitude measurements using the WIROTEST setup, two specimens exhibited a peak value near the end of the measurement section (points 10 and 11). This is likely associated with the location of the martensitic transformation front. However, this effect was not consistently observed. The results obtained from techniques based on magnetic induction demonstrate sensitivity to changes in magnetic permeability, which are linked to the presence of the ferromagnetic martensitic phase.

One should highlight the suitability of eddy current (EC) methodology in ferrite content identification. The comparison between feritscope and XRD techniques presented in Figure 10 proved the high efficiency of presented methodology.

5 Conclusions

Eddy current methodology is a promising non-destructive technique for indirectly measuring martensite formation in 316L during plastic deformation, though it comes with certain limitations. 316L, an austenitic stainless steel, can undergo deformation-induced phase transformation from austenite to martensite during, which it alters its magnetic properties. Since eddy currents are sensitive to changes in the material’s electrical conductivity and magnetic permeability, they can indirectly detect the presence of martensite, which is ferromagnetic, unlike the non-magnetic austenitic phase. This makes EC a suitable method for tracking martensitic transformation in real-time during mechanical loading without interfering with the deformation process. The high sensitivity of EC to local changes in phase composition is an advantage, allowing the detection of even small volume fraction of martensite. Additionally, EC is fast, scalable, and can be applied to complex geometries, making it versatile for in situ industrial applications. However, its sensitivity to other factors, such as temperature variations, strain-induced microstructural changes unrelated to martensite formation, and surface condition, may complicate accurate phase quantification. Calibration with complementary techniques like XRD or magnetic Barkhausen noise analysis might be required to enhance the accuracy of martensite content determination. Thus, while EC is suitable for detecting trends in martensitic transformation during plastic deformation, precise quantification and isolation of martensite may necessitate careful control of experimental conditions and the integration of additional methodologies.