**1. Introduction**

The 9 wt.% Cr creep strength enhanced ferritic (CSEF) steels (e.g., T/P91, T/P92, T/P911, C/FB2, MARBN, NPM1, etc.) represent advanced structural materials for application in high-efficiency power engineering. However, for constructing complex power generation equipment, fusion welding technologies are needed for joining individual functional parts. In accordance with the numerous research studies and ex-service experience, e.g., [1–5], it has been generally accepted that the fusion welded joints of ferritic steels represent the most critical component locations with respect to their preferential degradation and potential failure. Besides the regions of base material (BM) and weld metal (WM) within the structures of all welded joints, thermal effect of fusion welding on the welded ferritic steels' BMs typically results in the creation of a relatively wide heat-affected zone (HAZ) consisting of several, continuously created microstructural sub-regions, i.e., often called the "HAZ microstructural gradient". Its occurrence within the welded joint represents the primary, welding-induced microstructure degradation zone, since the individual HAZ sub-regions, such as the coarse-grained HAZ (CG-HAZ), fine-grained HAZ (FG-HAZ), inter-critical HAZ (IC-HAZ), and subcritical HAZ (SC-HAZ), possess mutually various microstructures and mechanical properties [6–8].

Depending on several factors including the welding metallurgy-related material properties and outer loading and/or environmental conditions, the fusion weldments can be susceptible to some of several typical failures [9]. The "Type I" and "Type II" failures, originating from intercrystalline cracks in weld metals, are generally related to the so-called "hot cracking" phenomena, typically occurring in weld solidified microstructures with higher impurity content. However, the occurrence of these failures has been considerably suppressed in the ferritic steels' weldments thanks to the recently developed ferritic filler materials of high metallurgical purity [10]. The "Type III" failure typically occurs within the CG-HAZ close to the weld fusion zone (FZ) of low alloy ferritic steel weldments. This failure type has been often related to the so-called "reheat cracking" due to either residual stress relief during the PWHT or superabundant secondary precipitation hardening in FZ/CG-HAZ during high temperature creep exposure [11–13]. In a specific case of dissimilar weldments, the considered failure type (sometimes referred to as "Type IIIa" failure [14]) is related to premature creep cracking within the area of soft, carbon-depleted CG-HAZ, created as a result of the decarburization processes driven by the carbon activity gradient at the interface between the lower grade ferritic steel and the higher grade weld metal. Last but not least, depending on acting environmental conditions, the "Type III" failure may also be related to so-called "cold cracking" phenomena, i.e., hydrogen-induced cracking (HIC) or environmentally assisted cracking (EAC) [15,16]. This failure occurs due to exceeding the critical hydrogen concentration in locally hardened FZ/CG-HAZ areas with the highest degree of transformation (i.e., martensitic) hardening as a consequence of the welding thermal cycle. Under long-term creep conditions, the welded joints of ferritic heat-resistant steels are typically prone to the "Type IV" failure within their FG-/IC-HAZs because these regions exhibit the lowest creep strength within the whole weldment. This failure is generally related to severe degradation of transformation hardening mechanism and preferential coarsening of Fe2(W,Mo)-based Laves phase within the failure location [17,18]. The study by Albert et al. [19] showed that the "Type IV" failure is caused by preferential creep strain accumulation in the soft, fine-grained HAZ regions (FG-/IC-HAZs) due to the multiaxial stress state induced by microstructural heterogeneity throughout the weld-joint. A specific failure type is related to "cracking in over-tempered base material" which typically occurs within the softened region of SC-HAZ and is characterized by a highly ductile fracture [20,21]. This failure type is observed usually in welded joints after the high temperature tensile tests or after high-stress short-term creep tests [21,22]. The mechanism of microstructural and property degradation in SC-HAZ, i.e., within the over-tempered base material, is believed to arise from the coarsening of precipitates during the welding thermal cycle. After longer durations of low-stress creep tests, the failure commonly shifts from the over-tempered region to the "Type IV" failure region. However, unlike the short-term "over-tempered base metal cracking", the long-term "type IV cracking" is characterized by low-ductility creep failure [23].

In common industrial practice, the weldments of CSEF steels are necessarily subjected to conventional PWHT, i.e., the subcritical tempering below the steel Ac1 critical transformation temperature. The main aim of such PWHT is to relieve residual stresses and thermally stabilize the weld microstructure with secondary phase precipitates, typically the M23 C6 (M = Cr, Fe ... ) carbides and MX (M = V, Nb; X = C, N) carbo-nitrides. The direct consequence of performing the subcritical PWHT procedure is related to the decrease of unallowably high hardness in FZ/CG-HAZ and improvement of the overall weld fracture resistance [19,24]. However, it has been proved [19] that the occurrence of premature "Type IV" creep failure in ferritic steels' weldments cannot be avoided by any variation in the subcritical PWHT regime, since their HAZ microstructural gradients remain still preserved within subcritically tempered microstructures. On the other hand, several studies [25,26] suggested that the only way to enable the "Type IV" failure suppression in ferritic welds is associated with so-called "full heat treatment" which involves the weld renormalization (i.e., the weld complete reaustenitization and its subsequent cooling on still air), followed by conventional subcritical tempering.

Our previous investigations [27,28] were focused on investigation of the e ffects of both the conventional tempering and quenching-and-tempering PWHT procedures of T92/TP316H

martensitic/austenitic weldments on their microstructure and creep behavior. The results showed that the quenching-and-tempering PWHT led to "Type IV" failure elimination and thus notable creep life improvement as a result of significant homogenization of the T92 steel microstructure, i.e., complete suppression of the T92 HAZ microstructural gradient thanks to performed reaustenitization. Moreover, our separate study [29] on the T92 HAZ local mechanical properties of the T92/TP316H weldments indicated, that compared to the weldments subjected to only conventional PWHT, the T92 HAZ of quenched-and-tempered weldments exhibited lower hardness and higher impact toughness. The combined e ffects of quenching-and-tempering PWHT and subsequent electrochemical hydrogen charging on room-temperature tensile properties of the T92/TP316H weldments were investigated in [30]. It has been revealed that the applied electrochemical hydrogen charging did not a ffect the strength properties of the weldments significantly, but it resulted in quite serious deterioration of their deformation properties along with significant impact on their fracture behavior and final failure localization. The most critical region was found to be the interfacial weld region close to the T92 steel FZ.

Our present study represents a continuous research work to our aforementioned former studies. It deals with investigation of the e ffects of initial PWHT conditions and subsequent electrochemical hydrogenation on the resulting room-temperature tensile properties and fracture behavior of T92/T92 welded joints. Mutual correlations between varying microstructural characteristics induced by di fferent initial PWHT regimes and resulting mechanical properties of the weldments in either hydrogen-free or hydrogen-charged conditions are discussed.

### **2. Materials and Methods**

Four segments of industrially normalized and tempered T92 tubes (outer diameter 38 mm, wall thickness 5.6 mm, approx. tube segmen<sup>t</sup> length 130 mm) were circumferentially welded in the company SES a.s. Tlmaˇce, Slovakia. The welded joints were produced by gas tungsten arc welding (GTAW) technique using T92-based filler metal Thermanit MTS 616 to prepare two equivalent T92/T92 weldments. The T92/T92 welds geometry was the same as also used in our previous study about long-term ageing e ffects on room-temperature tensile behavior of quenched and tempered T92/TP316H dissimilar weldments [31]. Specifically, the 60◦ groove angle and 2–3 mm root gap was used. Welding parameters for the preparation of T92/T92 welded joints were the following ones: welding current 120–160 A, voltage 12–17 V and heat input 9–12 kJ/cm. The diameter of TIG electrode was 2.4 mm and the negative polarity on the electrode was used. Table 1 shows chemical compositions of the T92 steel base material (T92 BM) and T92 steel-based filler metal (T92 FM) Thermanit MTS 616.

**Table 1.** Chemical composition (wt.%) of T92 base material (T92 BM) and T92-based filler metal (T92 FM) used for fabrication of T92/T92 weldments.


The chemical compositions in Table 1 represent certified alloy compositions by the material producers Tenaris Dalmine (Dalmine—BG, Italy) and Voestalpine Böhler Welding (Düsseldorf, Germany), respectively.

The two prepared weldments were individually subjected to mutually di ffering post-welding heat treatment (PWHT) procedures. Figure 1 shows schematic illustration of both these PWHT procedures in context with the equilibrium phase diagram including isoplethal section for T92 BM, computed by thermodynamic software ThermoCalc (version S, Thermo-Calc Software AB, Solna, Sweden) using thermodynamic database TCFE6.

**Figure 1.** Calculated equilibrium phase diagram with schematic illustrations of individual post-welding heat treatment (PWHT) regimes applied in the present study for T92/T92 weldments (The T92 steel composition is indicated in the diagram by vertical dashed line at 0.11 wt.% C).

The first T92/T92 weldment was conventionally tempered at 760 ◦C (i.e., below the Ac1 temperature of T92 steel) for 60 min and then slowly cooled within the tempering furnace (see the PWHT-1 in Figure 1). On the other hand, the second T92/T92 weldment was subjected to its full renormalization consisting of the complete reaustenitization at 1060 ◦C (i.e., well above the Ac3 temperature of T92 steel) for 20 min and subsequently cooled on still air, followed by its conventional subcritical tempering (see the PWHT-2 in Figure 1). From both weldments, twelve cylindrical tensile test specimens of cross-weld (c-w) configuration with partly discontinuous M6 thread (due to the above specified tube wall thickness) within their head portions were machined. A schematic illustration of the tensile test specimen is shown in Figure 2.

**Figure 2.** The tensile test specimen for cross-weld tensile testing of T92/T92 weldments (All dimensions are in mm), gauge length diameter 4 mm, gauge length 38 mm.

Electrolytic hydrogenation, i.e., cathodic hydrogen charging of prepared cylindrical c-w tensile specimens was performed in electrolytic solution of 1M HCl with 0.1N N2H6SO4 at a current density of 300 <sup>A</sup>/m2. The hydrogenation was realized at room temperature for 24 h. This procedure has been optimized and used in our several former studies [15,30,32] which indicated full saturation of tensile specimens by hydrogen after 24 h of their electrolytic hydrogenation. Similar findings, supported by hydrogen concentration measurements indicated the same or even shorter hydrogenation time for achieving the hydrogen concentration saturation in electrochemically hydrogen-charged alloy steels, as reported in other studies, e.g., [33–35]. Yin et al. [36] indicated that the content of diffusible hydrogen tends to be the saturation state when the hydrogen charging time reaches 48 h. However, they showed that the difference in diffusible hydrogen concentration for 24 and 48 h of hydrogen charging was

already rather small (i.e., within experimental value scattering). A schematic illustration of the whole experimental setup is visualized in Figure 3.

**Figure 3.** Schematic illustration of electrolytic hydrogenation.

The room-temperature tensile tests were performed for individual welds' PWHT states in both hydrogen-free and hydrogen-charged conditions. The tensile testing was carried out using TIRATEST 2300 universal testing machine (TIRA GmbH, Schalkau, Germany) at a crosshead speed of 0.05 mm/min. Three tensile test specimens per each state (i.e., "PWHT-1", "PWHT-2", "PWHT-1 + hydrogen", and "PWHT-2 + hydrogen") were investigated. The hydrogen-charged samples were tested immediately after the electrolytic hydrogen charging. The evaluation of c-w tensile properties (i.e., yield stress "YS" estimated as 0.2% proof stress, ultimate tensile strength "UTS", total elongation at fracture "EL", and reduction of area at fracture "RA") involved the calculation of their average values and corresponding standard deviations.

Local mechanical properties of studied weldments were characterized by means of hardness measurements which were performed using a Vickers 432 SVD hardness tester (Wolpert Wilson Instruments, division of Instron Deutschland GmbH, Aachen, Germany) on plain surfaces of longitudinal sections of fractured tensile specimens. This procedure was also helpful for indication of local strain hardening effects within the studied weldments during the tensile tests. The referential, i.e., un-deformed samples corresponding to both initial PWHT states were also tested for hardness. All the hardness measurements were performed at 98 N loading for 10 s per measurement.

Microstructural analyses of the studied weldments were performed on the conventionally prepared metallographic samples (i.e., wet grinding on SiC papers with granularity from 500 to 1200 grit, cloth polishing with a diamond paste suspension of a particle size ranging from 1 to 0.25 μm and final etching in a solution consisting of 120 mL CH3COOH, 20 mL HCl, 3 g picric acid, and 144 mL CH3OH) using the light optical microscope OLYMPUS GX71 (OLYMPUS Europa Holding GmbH, Hamburg, Germany) and the scanning electron microscope (SEM) JEOL JSM-7000F (Jeol Ltd., Tokyo, Japan). Fractographic analyses were carried out using the SEM Tescan Vega-3 LMU (TESCAN Brno, s.r.o., Czech Republic).

### **3. Results and Discussion**
