INTRODUCTION

Additive technologies, namely, 3D printing with metals, have been intensively developing over the past two decades. The methods of selective laser melting (SLM) of powders, laser cladding of powder and electric arc cladding of wire are widely spread [1, 2]. Among the aforementioned methods, wire arc additive manufacturing (WAAM) [3–7] has numerous advantages, namely, high productivity (up to 12 kg/h) and uniformity. Such production technology of metal items can be successfully implemented in the existing production complexes, including CNC machines.

The operation of structures, which work for a long time in the Arctic and Far North, is necessarily associated with the need to control the physical and mechanical characteristics of the material used during the entire life cycle of products. The use of metallic materials at low temperatures is usually accompanied by a drop in impact toughness [8], and the risk of brittle fracture increases many times.

The ability of metallic materials to resist the action of dynamic loads is known to be related to the size of structural fragments (grains) [9]. In this regard, it is necessary to establish a interrelation between the technological parameters of the WAAM process and the resulting structure and properties of the material, in particular, providing the required cold resistance.

The heat input during 3D printing, thermal interlayer cycles, build rate and protective environment should be taken as controlled WAAM parameters. The proposed parameters are necessary for the creation of modern systems for controlling the structure and quality of the produced materials.

Thus, the purpose of this work is to investigate the peculiarities of structure formation in a 09G2S alloy during additive electric arc growth, as well as to determine the parameters of 3D printing necessary to form the required structure and, consequently, the mechanical properties of the alloy.

EXPERIMENTAL

A widely used 09G2S steel, which is used for welded structures that can be operated at low temperatures (down to –40°C), was chosen as the material for the research (the chemical composition is given in Table 1). The material for research was fabricated using the WAAM method (0.8 mm wire) integrated into a CNC machine. The construction of the workpiece was carried out in a CO2 environment as a protective gas.

Table 1. Chemical composition of steel, wt %
Full size table

Voltage U, current I, wire feed speed V, gap between wire and cladded workpiece z, as well as CO2 flow rate were taken as variable parameters of the WAAM process. The gap z, feed speed V, and shielding gas flow rate were chosen as constant variables in the experiment. In this case, z = 11 mm, and V = 350 mm/min.

During the experiment, the energy Q required for deposition of a unit weld length was determined,

$$Q = \frac{{0.8IU}}{V},$$

where 0.8 is the coefficient of energy loss (State Standard GOST R ISO 857-1–2009). The rest bench used for research allows changing the Q parameter in the range of 150–1200 J/mm.

The samples were made according to the following sequence: printing a layer, stopping the process and cooling the workpiece to the required temperature, printing the next layer. Thus, 3D printing was performed under three thermal cycles Tc: complete cooling of the layer (below 50°C), layer temperature 350°C according to the recommendations [11], and layer temperature 850–900°C (above the Ac3 point). The schematic view of the influence of thermal effects from laying subsequent layers during 3D printing on a point in the workpiece is illustrated in Fig. 1.

Fig. 1.
figure 1

Schematic view of thermal action from formation of subsequent layers on the region to be analyzed.

Full size image

Samples for studying the microstructure in different zones of the workpiece were cut out from three sections (next to the substrate, in the central part and in the final upper layers). To identify the microstructure of the material under study, short term etching in a 4% aqueous solution of nitric acid was used. The structures were imaged using a KYENCE–VHX1000 metallographic optical microscope.

The average grain size in the structures of 09G2S steel was assessed according to State Standard GOST 5639–82. To evaluate the cold resistance of the resulting material, Charpy type samples were made, which were cut from the walls along and across the cladded layers. This is due to the fact that materials made by WAAM 3D printing are characterized by anisotropy of properties [10].

Then the samples were tested on a MK-30A pendulum machine to determine impact toughness according to State Standard GOST 9454–78 at reduced temperatures (–80…+20°C) [11, 12]. The developed cryochamber [13] was used for low temperature tests.

RESULTS AND DISCUSSION

Analysis of microstructures (Fig. 2) of 09G2S steel formed at different energy modes of WAAM technology shows that the size of ferrite grains and structural homogeneity largely depend on the heat input to the cladding process. Experimentally it was determined that the most acceptable from the point of view of quality of the formed structure (the finest grain size with greater homogeneity, the smallest value of anisotropy of properties) is the structure with the mode of the heat input Q, equal to 425 J/mm. It also should be mentioned that structural macro-defects in the form of pores were detected at the heat input Q less than 190 J/mm.

Fig. 2.
figure 2

Structure of 09G2S steel fabricated at an applied energy of (a) 190 (b) 425, and (c) 475 J/mm.

Full size image

Average grain sizes d formed at different energy modes of the structures of the investigated material were measured. The results obtained in [14] show a linear relationship between the heat input Q and the resulting average grain size d, which allows us to predict the structural state and properties of the formed material in the course of WAAM. Similar dependences were obtained by researchers earlier when analyzing grain sizes after welding of 09G2S steel [15].

Due to the fact that the best mechanical properties are possessed by metals and alloys with minimum grain size, including impact toughness [16, 17], the study of the thermal cycle of the WAAM process was carried out at energy Q = 425 J/mm.

In the general case, during 3D printing, the subsequent layer melts part of the previous layer, and not the molten metal located below undergoes recrystallization (normalization), thus, the structure of the printed workpiece made of 09G2S steel in the first layer should be represented by columnar grains, as in the weld, this is followed by a wide zone passing through all layers and represented by a structure characteristic of steel after normalization, and an upper layer with an overheating zone from the welding arc, also characteristic of a conventional weld.

Next, the influence of the thermal cycle (Tc) on the microstructure and microhardness of the material under study was considered. Below are the modes of thermal cycles of cladding of 50°C (Fig. 3), 350°C (Fig. 4), and 850–900°C (Fig. 5).

Fig. 3.
figure 3

Results of microstructural investigations and estimation of HV microhardness along the height of the clad wall at a thermal cladding cycle of about 50°C.

Full size image
Fig. 4.
figure 4

Results of microstructural investigations and estimation of HV microhardness along the height of the clad wall at a thermal cladding cycle of about 350°C.

Full size image
Fig. 5.
figure 5

Results of microstructural investigations and estimation of HV microhardness along the height of the clad wall at a thermal cladding cycle of about 850–900°C.

Full size image

Figure 3 shows that the sample formed by printing on cooled metal has three characteristic zones. The first layer contains fine grains of ferrite, oriented in the direction of heat removal; the upper layer is represented by an overheating zone with signs of dendrites. Predominantly, including the central part, the sample consists of ferrite and pearlite grains normalized by the cyclic thermal effects of the printing process, the average grain size of ferrite is 15 μm. The structure has a large number of defects in the form of pores and slag inclusions of about 6 μm in size. Evaluation of microhardness over the entire height of the clad wall shows a stable hardness between 177 and 186 HV.

The sample fabricated at a layer temperature of 350°C (see Fig. 4) is identical in structure to the sample produced with significant cooling (see Fig. 3). In this case, the ferrite grain size in the central part is larger and amounts to 20 μm. This is due to the fact that the entire sample received heat from the 3D printing process and was heated to 350°C, which led to an increase in the size of the overheating zone, resulting in a slight increase in grain size. Evaluation of microhardness based on the height of the deposited wall also shows a stable hardness of the middle part of the wall ranging from 165 to 169 HV and a higher hardness of the upper part (last layer) up to 179 HV. Moreover, the microhardness of the entire wall is lower than the microhardness of the wall obtained at Tc = 50°C.

The structure of the sample presented in Fig. 5, which formed in the course of a thermal cycle during which the temperature of the formed part of the workpiece was above 850–900°C (above Ac3), shows that the entire sample except the first layer has signs of dendrites and overheating. The structure mainly consists of coarse ferrite grains with a size of 40–60 μm, oriented in the direction of heat removal. The first layer also has coarse grains, but does not show signs of overheating; this is due to a different heat dissipation compared to subsequent layers, because the formation of the first layer was carried out on a substrate with a temperature of 20°C. Evaluation of microhardness based on the height of the deposited wall shows a stable hardness of the middle part of the wall ranging from 142 to 150 HV and an increased hardness of the upper part (last layer) up to 159 HV. Thus, to relieve internal stresses after cladding, heat treatment should be carried out. The data obtained also correlate with the results of studies of the heat affected zone during welding of 09G2S steel [18].

A cycle of about 350°C was chosen as the optimum thermal cycle for cladding, because samples printed when the layer had completely cooled had a large number of defects in the form of pores and slag inclusions, and samples produced at temperatures above 850–900°C had a coarse grained structure and signs of overheating.

Low temperature impact tests were carried out on samples [19] obtained using various printing modes and differing in the average grain size. The results are presented in [14]. It has been established that lowering the test temperature to –40°C, near which a ductile brittle transition (T50) is observed, has a significant effect on the impact toughness of the samples, reducing the values by up to two times. Thus, the operation of structures made of the metal under study at temperatures below –40°C significantly increases the likelihood of brittle fracture. The impact toughness of the printed samples was shown to be linearly related to the average grain size, a decrease in which significantly increases the resistance against loads at both positive and negative temperatures.

The obtained dependences of impact toughness on grain size confirm their relationship and are similar to the results of previous studies [15, 16].

CONCLUSIONS

(1) As a recommended printing mode, it is necessary to choose a mode that ensures a minimum grain size, which increases the impact toughness. The absence of macrodefects in the structure of a metal is a necessary condition.

(2) The experimental conditions at a cladding energy Q = 425 J/mm and thermal cycle of about 350°C were shown to be optimal from the point of view of forming a fine-grained structure with the minimum number of defects.

(3) The results obtained allow us to apply WAAM parts on real technical objects using the chosen printing conditions.