2. 1 Self-loaded hot crack testing: Hook cracks on Tensile test specimens

The evaluation of the so called Hook cracks on tensile specimens (Figure 1) according to DIN 32525 was recommended to be accomplished using a self-loaded hot crack testing method in 1990 [10]. Hot crack investigations with this method were already performed in earlier decades [11, 12]. As described, with this simple method only distinctions can be made between lower or higher HCS of weld metals, i.e. the preliminary selection of weld metals. For detailed investigation this method is not useable, as solidification cracks and liquation cracks cannot be kept apart [12]. For the evaluation of the HCS of ductile weld metals Pohle [13] has suggested to use the reduction of area over the number of Hook cracks found on the surface of broken tensile specimens by using a magnifier (6-10 fold). The Testing program I is shown in Table 3: The tensile test specimens were made out of all-weld metal (AWM), acc. EN 1597-1 and weld metal of a Plate Deposit (PD). The weld metal of PD is partially hindered in shrinkage by the plate thickness and the underlying layers, and therefore a higher HCS can be expected than for AWM, see Figure 2. The fabrication of the weld metals was performed using MCW and FCW respectively. The cross sections of them are shown in Figure 3. Additional weld metals were produced with solid wire (SW), MMA-electrode and FCW on the market. The applied welding procedures were GMA-welding for FCW (shielding gas 82Ar/18CO2) and for MCW (50Ar/50He), with direct current electrode positive (DCEP). The SW Ni-base 70/20 was welded using pulsed current (50Ar/50He).

2. 2 Externally-loaded hot crack testing: PVR-Test

For the development and quantification of hot crack susceptible fully austenitic weld metals the controlled deformation crack test or controlled flat tension test, PVR-test for short, was developed and employed the first time more than two decades ago [14, 15, 16], see Figure 4. The PVR-test belongs besides the Modified Varestraint-Transvarestraint-Test (MVT) and Hot tension test (using the Gleeble^® physical simulator) and others, to the externally loaded hot crack testing methods [17]. The PVR-test was created to evaluate the effects of welding procedure with regard to hot cracking [2]. Only one single test specimen is required to determine the HCS of a base metal and weld metal of the applied welding process. The ordinary test procedure of the PVR-test in Figure 4 uses flat specimens with the dimensions 40x10x300mm, clamped into a special tension fixture, lengthened in a horizontal servo-hydraulic system of the test equipment. The welding arc, moving with constant speed, is superposed by a linearly increased tension speed vPVR in welding direction.

The standard PVR-test procedure is carried out with tension rates linearly increasing from zero to 60mm/min, using bead-on-plate TIG-welding with Argon shielding gas and two types of heat input per unit length of about 7 and 10 kJ/cm (Iweld=180/220A, Uweld=12/14V, vweld=19cm/min) [18]. The critical tension speed vcr or critical elongation speed (CES) is the test criteria for the PVR-test. The CES should correspond to the first hot crack, detected visually at a magnification of 40. It can be determined for each of the hot crack types: Solidification cracks SC, Liquation cracks LC and Ductility dip cracks DDC, see Figure 5.

The success to assess all hot cracking types depends on the superposition of the local thermal cycle beside the bead with the tension rate. In addition, different indices for the critical tension speed vcr are possible, dependent on the size and number of visible cracks. For example: vcr1 is the critical tension speed for the first microscopically visible hot cracking feature. vcr3 describes the first three hot cracks per 10mm of weld bead and vcr9 determines the first nine hot cracks per 10mm of weld bead. The modelling of the PVR-test has shown the correlation between test criterion, hot cracking theory by Prokhorov, and its applicability [19].

The PVR-test I was performed for to proof the HCS of MMA-electrode Ni-base type 70/20 ∅4mm (Alloy No. 2) and the basic FCW Ni-base type 70/15 (Alloy No. 18), using two different heat inputs from the TIG-arc for remelting the PVR-weld metal (6.8 and 9.7 kJ/cm). To evaluate the HCS of the optimized FCW (Alloy No. 12b) in comparison to FCW market products (Alloy No. 14, 15, 16) the PVR-test II was performed. The final PVR-test III was to compare it’s HCS to MMA-electrode ∅5mm (Alloy No.2’), Solid Wire (Alloy No.1), and similar base metal Alloy 600, see Table 4. The standard heat input in the PVR-test was 6.8 kJ/cm, except PVR-specimen N31 and N33 having 9.7 kJ/cm.

2. 3Evaluation of the slag basicity

The basicity of the solidified slag was determined as a parameter for to adjust the metallurgical system, utilizing EDX-analysis of slag micro-sections and the modified basicity index by Bauné et al. [20] in Eq. 1 below.

CaF +O Na MgO CaO + O K +5.0 ×(O Fe MnO +O Cr )[at.%]

3. 1 Analysis of all-weld metal

Except the Nb-contents of the consumables Alloy No. 3 and 4 MCW 70/20 and Alloy No. 11 FCW 70/20, all of them are within the tolerances of weld metal Ni-base 70/20 from MMA-electrode (prEN ISO 14172) and bare electrodes and rods (DIN 1736, AWS A5.14) respectively. Only the Fe-content of weld metal FCW 70/15 (Alloy No.18) of approx. 1% is below the min. standard content of 2%Fe in Ni-base 70/15 from MMA-electrode (acc. DIN 1736), see Table 5 and Table 2. For the optimization of the manufactured flux-cored wires Alloy No.12 FCW 70/20 was chosen as a basis (2% Nb in the filling). By increasing the Nb-content to about 2.5% and making some variations mainly in Fe-, Mn- and Si-content, the required value for the fracture elongation of 30% could be reached. Taking into consideration the requirements for the arc stability and slag viscosity of the FCW, the final product was found to be 12b FCW 70/20, see Table 6.

3. 2Mechanical values of weld metals

With increasing Nb-contents up to 3 % in the weld metals the fracture elongation A5 and the tensile strength Rm as well were increased strongly, for FCW and MCW respectively, see Table

7. At 3.5 % Nb the fracture elongation for weld metals of MCW becomes lower again. As the fracture elongation and tensile strength as well is increased with increasing Nb-content from 0 to 3% in AWM and PD a finer subgrain structure is expected. For weld metals with high C-content between 2.2% and 3.6% Nb no significant change in the fracture elongation can be observed. Concerning the Nb-content of the solid wire SW, the A5 and Rm values fit to the tendency of the LC MCW, see Figure 6. With increasing Nb-content from 1 to 3 % in AWM and PD manufactured with FCW Alloy No. 11, 12 and 13 a strong improvement of fracture elongation and tensile strength can be found, too. FCW Alloy No. 15 and 16 from the market are below the required 30% fracture elongation, whereas Alloy No. 18 FCW 70/15 seems to be susceptible to higher welding heat input (compare 18a: 13 kJ/cm – 18b: 6 kJ/cm). The MMA-electrode, for lower and higher heat input as well, give very good A5-values, see Figure 7. With some variations in the metallurgical system of Alloy No. 12 FCW (Table 6), the fracture elongation could be improved to about 40% (12b-12d), so the required standard value of A5 >30%, could easily be fulfilled, see Table 8.

3. 3Reduction of area and Hook cracks

The differences in the mechanical values (Table 7, Figure 6 and 7) were due to the appearance of Hook cracks, which were evaluated over the reduction of area RoA (magnification 8x), as suggested by Pohle [13], see Figure 8 and Figure 9. The AWM and PD of MCW with high-carbon content with 2.2%Nb and 0.09%C (Alloy No. 6) reached the highest RoA (47/39%) and the fewest Hook cracks (6/31) as well. High carbon MCW with 3.3/3.7%Nb and 0.09%C (Alloy No. 8/10) give comparable good results too. Among weld metals with low-carbon content MCW with 3.4/3.0 %Nb and 0.02%C (Alloy No. 9/7) give the best values for AWM. The Plate Deposits PD with 3/2% Nb have the highest RoA and lowest number of cracks (Alloy No. 7/5). The weld metals of Alloy No. 1, solid wire type 70/20, show superior ductility, i.e. RoA more that 50%, whereby only a few Hook cracks appeared, see Figure 8.

For FCW testing wires the RoA increases significantly with the Nb-content (1.0, 2.0, 3.1%), like the fracture elongation in Figure 7. The lowest number of Hook cracks appeared for all-weld metal with 3.1%Nb (Alloy No. 13), see Figure 9. The weld metals of Alloy No. 17 FCW type 70/15 show superior ductility, like the solid wire in Figure 8. Astonishingly the Plate Deposit PD of the basic FCW produced with ~8kJ/cm heat input (Alloy No. 18b) show a RoA of 51%, having 102 small Hook cracks. The same Plate Deposit welded with 13kJ/cm (Alloy No. 18a) show a distinctly lower RoA (28%), but lesser cracks, too (33). From there it can be expected that this weld metal is sensitive to overheating, but on the other hand it has high fracture toughness. The Alloy No. 2 of MMA-electrode shows also very good ductility and just a few Hook cracks. Besides it no deterioration of the values using ~10kJ/cm welding heat input (2a) compared to ~6 kJ/cm (2b) for Plate Deposit PD can be observed, see Figure 9. A satisfying optimization of the manufactured Alloy No. 12 FCW could be reached by setting the Nb-content to about 2.5% and the C-content in AWM to about 0.05%. Other important factors are the ratios of Nb/Si and Ni/Cr/Fe, as well as the Mn- and Fe-content [25, 26]. According to the investigation of Dupont [27] a higher C-content lowers significantly the solidification interval for comparable alloys. Herewith the number of Hook cracks could be decreased strongly from 40 of Alloy No.12 FCW to 4 cracks of Alloy No.12b FCW, see Table

8. The AWM of No.12a FCW was somehow out the tolerances of this metallurgical system, i.e. slightly lower/higher C-/Si-content, compare Table 6, what can be seen at the lower Basicity Index (BI=0.59), see Figure 12.

3. 4Fractography of all-weld metal from FCW

For the necessary development and optimization of the weld metals from Flux-cored testing wires (Alloy No. 11, 12, 13) concerning the Nb/C ratio, investigations of the fracture surface of the tensile test specimens were performed using a stereo microscope and SEM-fractographs. As the ductility and tensile strength as well were improved with increasing Nb-content (Figure 7, Table 7), it could be expected to find a finer subgrain structure, i.e. smaller dendrite spacings, what could be confirmed. All-weld metal of FCW with 1%Nb (Alloy No. 11) shows brittle fracture behaviour, i.e. due to the intercrystalline cracks leading to fracture, a typical fibrous structure can be seen in the stereo microscope, Figure 10 top left. In the SEM-fractograph plane structures with partially melted zones appear, Figure 10 top right. All-weld metals of FCW with 2 and 3%Nb (Alloy No. 12 and 13) show a microstructural ductile fracture and the bright precipitations found in the SEM were analyzed to be NbC, Figure 10 middle/down. All three weld metals show oxidic inclusions of different size and shape, marked with “Ox.” The all-weld metal of the optimized No. 12a and 12b FCW 70/20 show totally different fractographs, despite the similar chemical analysis in Table 6: 12a show partially melted zones and a crack surface topography as well. On the contrary 12b shows a distinctive dimple fracture, see Figure 11 top/down.

3. 5 Slag Basicity and O-content

The basicity of the solidified slag was determined by the modified basicity index by Bauné et al. [20, 28] in Eq. 1. Herewith, a good correlation could be achieved between the O-content and the impact values of the all-weld metal from the flux-cored testing wires Alloy No. 11, 12, 13, see Table 9.

By this means the possibility to investigate the influence of relatively small variations of metals (1, 2, 3% Nb) in the filling on the basicity and the O-content in the weld metal is shown. For the optimized FCW Alloy No. 12a-12d also a good correlation was found: Falling O-content in all-weld metal with higher BI of the solidified slag and decreasing Hook cracks, see Figure 12.

3. 6 PVR-Test results

The results of the PVR-test I in Figure 13 show a reduction in the critical elongation speed (CES), i.e. higher HCS, when using higher heat input of the TIG-arc for remelting the weld metal. To that effect MMA-electrode Ni-base 70/20 ∅4mm (Alloy No. 2) is more susceptible to Solidification cracks (1^stSC criterion), while the basic FCW Ni-base type 70/15 (Alloy No. 18) shows a stronger shift of the liquation cracks criteria (3LC/10mm and 9LC/10mm) to lower elongation speeds. The reason for this is the bigger size of the mushy zone using a heat input of 9.7 kJ/cm, compared to 6.8 kJ/cm [19]. Apart from that, the weld metals have comparable low HCS, like it was estimated in the RoA-Hook-cracks diagrams, see 2a/b-18a/b in Figure 9. Ductility Dip cracks DDC were not evaluated in these weld metals. Figure 14 show two examples of longitudinal micro-sections at the CES of 3LC/10mm (v3LC=12.4mm/min) and at the CES of the 1^stSC criterion (vSC=31.1mm/min) of PVR specimen N30, MMA-electrode Ni-base 70/20 ∅4mm (Alloy No. 2). The weld metal of the optimized FCW 70/20 (Alloy No. 12b) in the PVR-test II show comparable CES to the FCW 70/20 market products for each of the hot crack types (Figure 15); but they are quite lower than the CES of MMA-electrode ∅4mm (Alloy No. 2) and the basic FCW 70/15 (Alloy No. 18) in Figure 13. A further comparison in the PVR-test III shows that the CES of Alloy No. 12b is far away from that of the Solid wire (Alloy No. 1). It is just comparable to MMA-electrode ∅5mm (Alloy No. 2’), concerning the vSC / vMicro-SC criterion. Still these values are quite below the similar base metal Alloy 600, see Figure 16.

CONCLUSIONS

Nb-additions to weld metals Nibas-70/20 can reduce the hot crack susceptibility HCS strongly, because of the formation of NbC, what makes finer subgrains in the weld metal. By that also the fracture elongation and the tensile strength is improved.

The tested weld metals of Plate Deposit PD show mostly more Hook cracks than all-weld metal AWM. Thus it should be a useful method for pre-selecting a weld metal concerning its HCS. For the optimization of slag bearing consumables the determination of the Basicity Index of the solidified slag can be a support to find the right metallurgical system.

Higher heat input in the manufacture of the weld metals as well as for the remelting in the PVR-test increases the HCS strongly. The HCS of weld metals from FCW 70/20 in the PVR-test is distinctly higher than for weld metals of MMA-electrode and solid wire. The HCS of weld metals Ni-base 70/15 for the investigated products was lower than for 70/20.

In general the HCS of similar base metal Alloy 600 is still much lower than of the weld metals.

ACKNOWLEDGEMENTS

The authors would like to give thanks to the Materials Center Leoben (Kplus program of the Austrian government) for financing this investigation. Likewise to Prof. Herold, Mrs. Streitenberger and Mr. Pchennikov from the Institute of Joining and Beam Technology at TU Magdeburg for the performing and evaluation of the PVR-tests. Last but not least we appreciated the expert advices of Mr. Tösch and Mr. Klagges from Böhler Schweißtechnik Austria and of Mr. Heinemann, UTP-Schweißmaterial very much.

……………………………………………………………………………………………………

Figure 1: Hook cracks on the surface of a tensile test specimen (solidification and liquation cracks SC/LC widened by loading).

Figure 2: (a1) Joint preparation for all-weld metal (AWM): Arrows show the direction of free shrinkage during welding. (a2) Cross section of the finished AWM-joint. (b1) Finished Plate Deposit PD: The weld metal is partially hindered in shrinkage by the plate thickness (20mm) and the underlying layers. (b2) Cross section of the finished PD. Doted circles in (a2), (b2): Diameter of tensile specimen acc. DIN 32525 (thread M16).

Figure 3: Cross sections of testing wires Ni-base type 70/20 ∅1.2mm: (a) Metal-cored wire MCW No. 3-10 and (b) Flux-cored wire FCW No. 11-13 in Table 3.

Figure 4: PVR-test, scheme and test procedure to determine the critical tension speed vcr initiating the 1st hot crack transverse to the welding direction; TU Magdeburg.

Figure 5: Solidification cracks SC, Liquation cracks LC and Ductility dip cracks DDC on a PVR-specimen.

Figure 6: Fracture elongation A5 and Tensile strength Rm of all-weld metal AWM and Plate Deposit PD of Metal-cored wire MCW No. 3-10 and Solid wire SW 1, see Table 3, 5 and 7.

Figure 7: Fracture elongation A5 and Tensile strength Rm of all-weld metal AWM and weld metal of Plate Deposit PD, manufactured with Flux-cored testing wires (11-13), Flux-cored wires market products (14-16: type 70/20, 17-18a/b: type 70/15) and MMA-Electrode (2a/b), see Table 3, 5 and 7.

Figure 8: Number of Hook cracks against Reduction of Area of all-weld metal AWM and weld metal of Plate Deposit PD manufactured with Metal-cored testing wires MCW, Alloy No. 3-10 and solid wire SW, Alloy No. 1, see Table 3, 5 and 7.

Figure 9: Number of Hook cracks against Reduction of Area of all-weld metal AWM and weld metal of Plate Deposit PD manufactured with Flux-cored testing wires (Alloy No. 11-13), Flux-cored wires market products (Alloy No. 14-16: type 70/20, 17-18a/b: type 70/15) and MMA-Electrode (Alloy No. 2a/b), see Table 3, 5 and 7.

Table 1: Standard requirements for the mechanical values of Ni-base 70/20 and 70/15 all-weld metals.

Table 2: Standard all-weld metal composition of Ni-base 70/20 and 70/15, bare electrodes and similar base metal.

Table 3: Nb- and C-content of the investigated welding consumables and heat input for manufacturing weld metals Ni-base 70/20 and 70/15 (Testing program I).

Table 4: Parameter for manufacture of PVR-specimens with different welding consumables (PVR-Test I, II, III).

Table 5: Analysis of all-weld metals AWM manufactured with different welding consumables, see Table 3.

Table 6: Analysis of all-weld metal AWM manufactured with optimized FCW 12a – 12d based on 12 FCW 70/20.

Table 7: Mechanical values of all-weld metal AWM and weld metal of Plate Deposit PD (Testing program I).

Table 8: Mechanical properties, No. of Hook cracks and heat input for manufacturing weld metals Ni-base 70/20 of optimized FCW based on Testing wire No. 12 (Testing program II).

Table 9: Basicity Index BI of the slag, O-content and Charpy-Impact Values Av of all-weld metal AWM from the investigated welding consumables (Testing program I).