*Article* **Microstructure and Mechanical Properties of NiTi-Based Eutectic Shape Memory Alloy Produced via Selective Laser Melting In-Situ Alloying by Nb**

**Igor Polozov \* and Anatoly Popovich**

Institute of Mechanical Engineering, Materials, and Transport, Peter the Great St. Petersburg Polytechnic University, Polytechnicheskaya 29, 195251 St. Petersburg, Russia; director@immet.spbstu.ru **\*** Correspondence: polozov\_ia@spbstu.ru

**Abstract:** This paper presents the results of selective laser melting (SLM) process of a nitinol-based NiTiNb shape memory alloy. The eutectic alloy Ni45Ti45Nb10 with a shape memory effect was obtained by SLM in-situ alloying using a powder mixture of NiTi and Nb powder particles. Samples with a high relative density (>99%) were obtained using optimized process parameters. Microstructure, phase composition, tensile properties, as well as martensitic phase transformations temperatures of the produced alloy were investigated in as-fabricated and heat-treated conditions. The NiTiNb alloy fabricated using the SLM in-situ alloying featured the microstructure consisting of the NiTi matrix, fine NiTi+β-Nb eutectics, as well as residual unmelted Nb particles. The mechanical tests showed that the obtained alloy has a yield strength up to 436 MPa and the tensile strength up to 706 MPa. At the same time, in-situ alloying with Nb allowed increasing the hysteresis of martensitic transformation as compared to the alloy without Nb addition from 22 to 50 ◦C with an increase in Af temperature from −5 to 22 ◦C.

**Keywords:** additive manufacturing; powder bed fusion; nitinol

### **1. Introduction**

NiTi nitinol-based alloys are among the most important shape memory and superplastic materials and are of considerable interest for potential applications in aerospace and biomedical fields [1,2]. Nitinol has found its practical applications in the aircraft industry for thermo-power actuators, thermomechanical connectors, in medicine for stents, constrictive fixators, and other medical devices [3,4]. Nitinol-based alloys, unlike Fe- and Cu-based shape memory alloys, are characterized by high strength and ductility and have high corrosion resistance and biocompatibility. However, brittle secondary phases such as the metastable compound Ni4Ti3, which can decompose into Ni3Ti2 and Ni3Ti with increasing temperature, are commonly observed in NiTi-based alloys [5,6]. These phases increase the hardness of the material, but reduce the ductility of NiTi alloys, limiting their application. In this regard, it is relevant to introduce secondary phases with high hardness into the matrix of NiTi-based alloys while maintaining or slightly reducing the level of ductility.

It is known that Nb plays an important role in nitinol-based alloys because the addition of Nb increases the hysteresis of martensitic transformation [7]. Besides, the addition of Nb improves the biocompatibility of NiTi-based alloys [8]. Extension of the temperature interval of martensitic transformation can be used in the manufacture of connecting elements or seals [9]. Ni47Ti44Nb9 alloy is a classic nitinol-based shape memory alloy characterized by an in situ composite microstructure [7]. It is known that the microstructure of the alloy with shape memory effect Ni47Ti44Nb9 consists of NiTi matrix and β-Nb phase [10]. The presence of β-Nb is known to have a significant effect on the shape memory effect and mechanical properties of the alloy [11]. Nb exhibits different properties from the NiTi

**Citation:** Polozov, I.; Popovich, A. Microstructure and Mechanical Properties of NiTi-Based Eutectic Shape Memory Alloy Produced via Selective Laser Melting In-Situ Alloying by Nb. *Materials* **2021**, *14*, 2696. https://doi.org/10.3390/ ma14102696

Academic Editor: Federica Bondioli

Received: 21 April 2021 Accepted: 19 May 2021 Published: 20 May 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

matrix with a shape memory effect [12]. The presence of Nb as a second phase (β-Nb) or as a solution is the primary reason for transformation hysteresis expansion in NiTiNb shape memory alloys [10].

Conventionally, Ni47Ti44Nb9 alloy parts are made by casting or sintering powder materials, which have limitations in terms of part geometry, as well as control of their microstructure and properties [13–15]. In this regard, it is promising to use additive manufacturing (AM) techniques, in particular, the selective laser melting (SLM) method for the manufacture of parts from alloys with the shape memory effect [16].

The SLM process uses metallic powders as a feedstock material. Various research works have shown that pre-alloyed powders allow obtaining a material with a more homogeneous microstructure and stable mechanical properties [17–19]. However, the production of pre-alloyed powders is usually time and labor consuming, especially in the case of custom alloys. The use of an elemental powder blend in the SLM process is an alternative option that can be used for in-situ synthesis or in-situ alloying during the SLM process to obtain a material with a required composition [20–22]. At the same time, rapid solidification and cooling rates typical for the SLM process lead to insufficient diffusion of elements with high melting points, which causes a heterogeneous microstructure. In this regard, additional heat treatment can be applied to improve the chemical homogeneity of the material [23,24]. AM methods, such as SLM or Direct Energy Deposition (DED), involve repetitive heating and high solidification rates leading to distinctive microstructural features of nitinol alloys [25]. For example, rapid solidification promotes the formation of a supersaturated solid solution matrix [25]. At the same time, various process parameters can result in different microstructures and phase composition of nitinol alloys [26], as well as different Ni content resulting in significant changes in transformation behavior [16].

In this work, the feasibility of the SLM process to produce a nitinol-based shape memory alloy via in-situ alloying by Nb was investigated. Using a powder blend of NiTialloy and pure Nb, the influence of the SLM process parameters on the sample's density was studied. Microstructure, phase composition, mechanical properties, and martensitic phase transformation temperatures were studied for the NiTiNb alloy in the as-fabricated and heat-treated conditions.

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

The following materials were used as feedstock materials: gas atomized NiTi alloy powder with a nickel content of 51.4% (at.) and niobium powder with a purity of 99.8%. The NiTi powder had the following particle size distribution: d10 = 27.2 μm, d50 = 50.0 μm, and d90 = 84.9 μm. The niobium powder in the initial state had a spherical particle shape and was pretreated in a thermal plasma jet using Tekna Tek-15 (Sherbrooke, QC, Canada) plasma spheroidization unit to obtain spherical particles. The details of the plasma spheroidization process can be found in [27]. The final niobium powder had the following particle size distribution: d10 = 15.5 μm, d50 = 35.5 μm, and d90 = 72.8 μm. Particle size distribution of the powders was measured by laser diffraction technique with Analysette 22 NanoTec (Fritsch, Idar-Oberstein, Germany).

NiTi and Nb powders were mixed in 85% NiTi: 15% Nb weight ratio for 12 h using a tumbler mixer to obtain a powder blend with Ni47Ti44Nb9 composition. Figure 1a shows a scanning electron microscope (SEM) image of the NiTi and Nb powder blend. The NiTi powder particles have a spherical shape, while some Nb particles have irregular shape due to incomplete spheroidization. Figure 1b,c shows the distribution of Ti, Ni, and Nb elements in the powder blend demonstrating that Nb particles were fairly distributed among the NiTi powder.

The NiTiNb alloy samples were manufactured from the prepared powder blend using AconityMIDI (Aconity3D GmbH, Herzogenrath, Germany) SLM system with varying process parameters. Samples with the size of 10 × <sup>10</sup> × 10 mm<sup>3</sup> were produced to study the microstructure, phase composition, and density of the material. Two groups of samples were fabricated, the difference between which is the use of different powder layer thickness and the laser beam spot size. Laser power, scanning speed, and hatch distance were also varied. The SLM process parameters used in the study are shown in Table 1. Sets A, B, and C use standard laser spot size (~70 μm) and sets G, K, and J use increased laser spot size with an unfocused beam (~300 μm). The volume energy density (VED) calculated according to a standard equation [28,29] was used as a parameter to investigate the effects of SLM process parameters on the samples' density.

**Figure 1.** (**a**) SEM-image of the NiTi-Nb powder blend and chemical distribution of the elements in the blend: (**b**) blue—Ti and red—Nb and (**c**) green—Ni and red—Nb.


**Table 1.** SLM process parameters used to fabricate the samples.

To study the effect of heat treatment temperature and holding time on the microstructure and properties of the samples, annealing was carried out under the following conditions:


The samples were heated at a rate of 10 ◦C/min and furnace cooled. The annealing temperature of 900 ◦C is above the recrystallization temperature of NiTiNb alloys [30], while 500 ◦C is below the recrystallization temperature and its main purpose is residual stress relieving.

The microstructure was studied using a TESCAN Mira 3 LMU scanning electron microscope (SEM) (TESCAN, Brno, Czech Republic) and a Leica DMI 5000 optical microscope (Leica, Wetzlar, Germany). To study the microstructure, polished microsections of the samples were etched using the following etchant: 83% H2O, 14% HNO3, and 3% HF. The samples were cut along the building direction. The chemical composition of the material was investigated using energy-dispersive analysis (EDS) (TESCAN, Brno, Czech Republic). The phase composition of the material was determined using a Bruker D8 Advance diffractometer on CuKα radiation (λ = 1.5418 Å). The density of the material was measured by the Archimedes principle. Tensile mechanical properties were investigated using cylindrical specimens using a Zwick/Roell Z050 testing machine (Ulm, Germany). Temperatures of martensitic transformations of the fabricated alloy were determined by differential scanning calorimetry (DSC). To compare phase transformation temperatures, NiTi samples were fabricated without the addition of Nb particles using the E2 parameter set.

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

Figure 2 shows the effect of VED on the density of NiTiNb samples produced using the NiTi-Nb powder blend with 300 μm (Figure 2a) and 70 μm (Figure 2b) laser spot size. The highest relative density values of around 99.2 ± 0.05% were obtained using E1 and E2 parameter sets, which correspond to 300 μm laser spot size and 26–27 J/mm<sup>3</sup> VED. Increasing VED led to lower density values, which might be attributed to melt pool overheating and instability and formation of gas and keyhole pores [31,32]. At the same time, varying VED at 70 μm laser spot size and 50 μm layer thickness did not lead to a significant change in density.

**Figure 2.** Effect of VED on the density of NiTiNb samples produced by SLM using (**a**) 300 μm and (**b**) 70 μm laser spot size.

As can be seen in Figure 3, there are several types of internal defects in samples produced under different process parameters. In the case of the samples D3 and F1, coarse pores with the size of 300–400 μm and irregular shape can be found along with fine spherical pores. These coarse pores elongated in the direction parallel to the laser path might be the result of lack-of-fusion due to the interaction of various factors: capillary forces, material evaporation, insufficient melting, etc. [33,34]. The smallest number of defects can be seen in the case of the sample E2, which corresponds to 27.5 J/mm<sup>3</sup> VED and 300 μm laser spot and has the highest density value.

**Figure 3.** Optical images of the polished microsections for the samples produced using parameter sets (**a**) D1, (**b**) D2, (**c**) D3, (**d**) F1, (**e**) F2, (**f**) E1, and (**g**) E2.

Figure 4 shows SEM-images of microstructures of the samples manufactured using E1 and A1 parameter sets. In the case of the E1 sample produced with an unfocused laser beam and 100 μm layer thickness, there are coarse solidified melt pools with a width comparable to the diameter of the laser spot as can be seen in Figure 4a. Melt pool boundaries can be clearly distinguished due to the presence of eutectic bands.

**Figure 4.** SEM-images showing the microstructure of the NiTiNb samples produced using parameter sets (**a**,**b**) E1 and (**c**,**d**) A1.

TiNi-Nb alloy system is a eutectic alloy and according to the phase diagram, the eutectic point corresponds to 26 at. % Ni content [35]. The eutectic temperature is 1150.7 ◦C. Hence, NiTi+β-Nb eutectic should be present in the microstructure after solidification and cooling of the Ti-Ni-Nb alloy [36]. The presence of NiTi and β-Nb phase was confirmed by the XRD results (Figure 5). The eutectic is mainly located at the melt pool boundaries as can be seen in Figure 4. Nb has a significantly higher density and melting point compared to NiTi. As a consequence, during the SLM of the powder blend, melting of NiTi will occur first, while the Nb may remain unmelted. Mixing of the alloy components in the melt can be carried out by Marangoni convection [37], which would result in a distribution of Nb in the volume of a melt pool. During the laser melting of the powder layer, the underlying solidified material is partially remelted. This can promote the diffusion at the melt pool boundaries and the dissolution of the elements in these micro volumes. Thus, the eutectic NiTI+β-Nb microstructure can be mainly found at the melt pool boundaries. As can be seen in Figure 4b,d, NiTi+β-eutectic areas are visible near partially melted on unmelted Nb particles where diffusion takes place during the SLM process and promotes the formation of fine eutectic microstructure.

When 50 μm layer thickness and a standard laser spot size were applied during the SLM process (sample A1), the microstructure of the obtained alloy also features melt pool boundaries as can be seen in Figure 4c, but their size is much smaller compared to E1 sample. The eutectic bands width is also smaller in this case, which indicates less degree of Nb diffusion into the NiTi matrix in case of smaller laser spot size. When an unfocused laser beam along with increased laser power and layer thickness is used, the melt pool volume is bigger and a higher volume of powder blend is melted with the laser. At the same time, a bigger melt pool volume leads to a lower cooling rate during the SLM process [38], which promotes the diffusion of elements and formation of the eutectic phase. Thus, a higher volume of eutectic phase is obtained when an increased laser spot size is used.

**Figure 5.** The XRD pattern of the NiTiNb sample produced using the E1 parameter set.

Figure 6 shows the chemical distribution of the elements in the NiTiNb sample produced using the E1 parameter set. It can be seen that there are areas of pure Nb corresponding to unmelted Nb particles. In general, Ti, Ni, and Nb are homogeneously distributed in the volume, and the eutectic regions do not feature a significant change in concentration of the elements compared to the rest of the sample. According to the EDS results, the obtain alloy has the following composition (in at. %): 45.5 ± 0.2 Ti, 44.9 ± 0.2 Ni, and 9.6 ± 0.4 Nb. The obtained composition is characterized by a decreased Ni content due to its evaporation during the SLM process, which is consistent with the results on the SLM of NiTi alloys reported in the literature [16,39].

**Figure 6.** (**a**) SEM-image of the microstructure of E1 sample and EDS-maps showing the distribution of (**b**) Ti, (**c**) Ni, and (**d**) Nb.

To obtain a more homogeneous structure, the samples manufactured using the E1 parameter set were heat-treated under different conditions. SEM-images of the microstructures after heat treatment are shown in Figure 7. As a result of heat treatment, Nb diffusion into the material matrix occurs, which is observed in the form of partial dissolution of Nb particles and an increase in the eutectic volume fraction, but the annealing temperature and time are not sufficient for a complete dissolution of Nb, because it has a low diffusion coefficient. Thus, after annealing at 900 ◦C for 30 min, the alloy structure has not undergone significant changes. As the annealing time was increased to 2 h at 900 ◦C, the former melt pool boundaries became less pronounced, and the fraction of regions with eutectic microstructure increased due to an accelerated niobium diffusion in the alloy matrix.

**Figure 7.** SEM-images of the NiTiNb samples after different heat treatments: (**a**,**d**) 500 ◦C for 2 h, (**b**,**e**) 900 ◦C for 30 min, and (**c**,**f**) 900 ◦C for 2 h.

For the tensile tests, the samples were heat-treated by annealing at 900 ◦C for 2 h since it provides a more homogeneous microstructure of the alloy. The results of the tensile tests are shown in Table 2. While the tensile strength of the in-situ alloyed samples is comparable to the casted Ni47Ti44Nb9 alloy, the elongation is significantly lower. Heterogeneous microstructure and impurities pick-up are believed to be the main factors affecting the elongation of the SLM-ed alloy.

**Table 2.** The results of tensile tests at room temperature for NiTiNb samples in the as-fabricated and heat-treated state.


Figure 8 shows the DSC curves for the fabricated samples. In the as-fabricated condition, the reverse martensitic transformation effect is poorly visible. However, the presence

of a reverse martensitic transformation peak allows for the conclusion that a forward martensitic transformation also takes place during cooling. Annealing at 900 ◦C resulted in Af temperature shift into higher temperature range and made both forward and reverse transformation effects more pronounced. Table 3 summarizes the DSC results of the obtained samples showing martensitic phase transformation temperatures in the as-fabricated and heat-treated states. One of the main functional properties of shape memory alloys are temperatures of direct and reverse martensitic phase transformations. It is reasonable to consider that the solution of Nb in the NiTi matrix can change the kinetics of martensitic transformations [10].

**Figure 8.** DSC-curves of the in-situ alloyed NiTiNb samples (**a**) in the as-fabricated state and (**b**) after annealing at 900 ◦C for 2 h.


**Table 3.** Martensitic phase transformation temperatures of the fabricated alloys.

Annealing at 900 ◦C leads to increased and narrowed temperature intervals of the reverse martensitic transformation compared to the as-fabricated state. The Af temperature shifts to the region of positive temperatures. At the same time, "in-situ" alloying increased the hysteresis of martensitic transformation from 22 to 50 ◦C as compared to NiTi alloy without Nb addition with an increase in the Af temperature from −5 to 22 ◦C. The in-situ alloyed NiTiNb samples showed higher direct martensitic transformation temperatures compared to the casted Ni47Ti44Nb9 alloy, which might be attributed to the difference in Ni content. Due to partial evaporation, the SLM-ed samples have a Ni content of around 45.4 at. % leading to the increase of transformation temperatures.

### **4. Conclusions**

The SLM process of NiTi-based eutectic alloy with shape memory effect obtained by in-situ alloying with Nb has been studied. Based on the results obtained, the following main conclusions can be made:


**Author Contributions:** Conceptualization, I.P.; investigation, I.P.; methodology, I.P.; project administration, A.P.; supervision, A.P.; writing—original draft, I.P.; writing—review and editing, A.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was carried out under the State Contract dated 4 June 2020, No. Н.4щ.241.09.20.1081 (ISC 17706413348200001110).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


### *Article* **Structure and Properties of Ti/Ti64 Graded Material Manufactured by Laser Powder Bed Fusion**

**Evgenii Borisov, Igor Polozov \*, Kirill Starikov, Anatoly Popovich and Vadim Sufiiarov**

Institute of Machinery, Materials, and Transport, Peter the Great St. Petersburg Polytechnic University (SPbPU), Polytechnicheskaya 29, 195251 St. Petersburg, Russia; evgenii.borisov@icloud.com (E.B.); kirill.starikov@gmail.com (K.S.); director@immet.spbstu.ru (A.P.); vadim.spbstu@yandex.ru (V.S.) **\*** Correspondence: polozov\_ia@spbstu.ru

**Abstract:** Multimaterial additive manufacturing is an attractive way of producing parts with improved functional properties by combining materials with different properties within a single part. Pure Ti provides a high ductility and an improved corrosion resistance, while the Ti64 alloy has a higher strength. The combination of these alloys within a single part using additive manufacturing can be used to produce advanced multimaterial components. This work explores the multimaterial Laser Powder Bed Fusion (L-PBF) of Ti/Ti64 graded material. The microstructure and mechanical properties of Ti/Ti64-graded samples fabricated by L-PBF with different geometries of the graded zones, as well as different effects of heat treatment and hot isostatic pressing on the microstructure of the bimetallic Ti/Ti64 samples, were investigated. The transition zone microstructure has a distinct character and does not undergo significant changes during heat treatment and hot isostatic pressing. The tensile tests of Ti/Ti64 samples showed that when the Ti64 zones were located along the sample, the ratio of cross-sections has a greater influence on the mechanical properties than their shape and location. The presented results of the investigation of the graded Ti/Ti64 samples allow tailoring properties for the possible applications of multimaterial parts.

**Keywords:** additive manufacturing; selective laser melting; titanium alloys; multimaterial 3D printing; graded materials

### **1. Introduction**

With the advent of Additive Manufacturing (AM) technologies, it has become possible for designers to improve the technological and functional capabilities of parts by evolutionary design optimization [1,2]. AM technologies have simplified the manufacturing of complex, single-piece products, while opening up the possibility of shaping a specific, given structure [3–5]. One method of part optimization is the use of multiple materials in the fabrication of a single part [6–8]. For example, in a part that is only partially exposed to high temperatures, it is possible to use heat-resistant materials only in the temperature-loaded part. In this case, for the formation of the remaining volume of the part it is reasonable to use less heat-resistant and, at the same time, cheaper materials. In addition, the combination of strong and ductile materials is widely used, for example, in tools for machining and gears, etc. [9,10]. In implants, a very important parameter is the mechanical strength and elasticity of the material. On the one hand, it is necessary to ensure a sufficient strength to avoid fracture. On the other hand, too much elasticity of the material can lead to bone damage due to permanent differences in the strain under load [11].

Recently, an increasing number of studies have appeared in the field of forming parts from several materials during the Laser Powder Bed Fusion (L-PBF) process. The main difficulty of this process is related to the fact that the existing equipment is not designed to use more than one powder material simultaneously. Therefore, much research on the development and modification of the equipment is being carried out [12–16]. For the

**Citation:** Borisov, E.; Polozov, I.; Starikov, K.; Popovich, A.; Sufiiarov, V. Structure and Properties of Ti/Ti64 Graded Material Manufactured by Laser Powder Bed Fusion. *Materials* **2021**, *14*, 6140. https://doi.org/ 10.3390/ma14206140

Academic Editor: Jan Haubrich

Received: 15 September 2021 Accepted: 14 October 2021 Published: 16 October 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

developers, the main difficulty to overcome is the necessity to apply a thin layer of powder material of the heterogeneous chemical composition. At the same time, this heterogeneity must correspond to the computer model of the part on each layer.

Another important direction of research is the study of microstructures and the properties of the multi-material products themselves, which are obtained using the L-PBF [17–20]. In these works, the microstructure and continuity of the transition zone, its phase composition, and mechanical characteristics were investigated.

Currently, there are several research papers devoted to the L-PBF of parts with a graded composition by changing the feedstock powder to build separate parts of a specimen. For example, alternatively using powders of different compositions was applied to fabricate CuSn/18Ni300 [18], NiTi/Ti6Al4V [19], AlSi10Mg/Cr18Ni10Ti stainless steel [21], 316L/CuSn10 [22], or 316L/Cu [23] graded specimens by L-PBF. The authors of [20] investigated the possibility of using the L-PBF process to fabricate graded samples using In718 and Ti6Al4V powders utilizing intermediate layers of mixed powders with a different ratio. The first commercial multimaterial recoating system for the L-PBF machines was recently introduced by Aerosint [24,25] suggesting the importance of multimaterial AM development. It uses mechanical forces to hold the powder on the drum and can release it at the desired location, generating a 2D single material image in a line-by-line manner. Currently, its possibilities have been demonstrated by 3D printing a copper alloy/steel bi-metallic parts [26].

For medical applications, Ti and Ti64 alloys have their advantages and disadvantages. The Ti64 alloy has great strength, but pure Ti has a great resistance to corrosion [27] and does not contain toxic impurities (Vanadium). Therefore, the formation of a graded part, where the advantages of both alloys are used, is relevant.

This work aimed to investigate the microstructure and mechanical properties of Ti/Ti64 graded samples fabricated by L-PBF with different geometries of the graded zones, as well as effects of heat treatment and hot isostatic pressing on the microstructure of the bimetallic Ti/Ti64 samples.

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

Commercially available CP-Ti (grade 2) and Ti-6Al-4V (Ti64, grade 5) alloy powders (Normin LLC, Borovichi, Russia) obtained by plasma atomization process were used as the feedstock material to fabricate the samples. The particles of both powders were spherical shaped (Figure 1) and had a mean size of d50 = 34 μm and d50 = 47 μm for Ti and Ti64 alloys, respectively.

**Figure 1.** Scanning electron microscope (SEM)-images of (**a**) Ti and (**b**) Ti64 powders.

The samples were fabricated using the SLM Solutions 280HL machine (Lübeck, Germany) in an argon atmosphere (99.99% purity) on a Ti64 built substrate. For microstructural characterization and microhardness evaluation, the samples of 20 mm height and

<sup>15</sup> × 15 mm2 section were built. Initially, one of the feedstock materials was used during the L-PBF process to fabricate the first half of the sample (10 mm). After that, the powder in the machine was changed to the second material and the second half of the sample was manufactured. The samples for mechanical tests were fabricated using a similar technique by changing the powder in the machine after building part of the sample. The same L-PBF process parameters were used for Ti and Ti64 alloys that were chosen based on the previous studies [28,29]. The following process parameters were used: scanning speed—805 mm/s, laser power—275 W, hatch distance—120 μm, layer thickness—50 μm. The laser beam size was approximately 80 μm.

Heat treatment of the samples was carried out using a vacuum furnace (Carbolite Gero GmbH & Co. KG, Neuhausen, Germany) at 10-3–10-4 mbar at 950 ◦C for 2 h, followed by furnace cooling. The regime was chosen based on AMS-H-81200A specification for the Ti64 alloy. The same temperature was used for hot isostatic pressing (HIP) of the graded samples, while the pressure was 100 MPa.

The microstructure was studied using a Leica DMI 5000 (Leica Microsystems, Wetzlar, Germany) optical microscope. To study the chemical composition, a Mira 3 (TESCAN, Brno, Czech Republic) scanning electron microscope with an energy dispersive X-ray (EDX) spectroscopy module was used.

Ti/Ti64 samples were scanned on a v|tome|x m300 X-ray computer tomography (CT). The system was equipped with an X-ray source with a maximum voltage of 300 kV. The obtained data were processed and visualized using an extended software package AVIZO for three-dimensional analysis and voxels visualization. Segmentation was performed using global and local gray thresholds.

The hardness of the samples was measured using Zwick/Roell ZHU 250 tester (Zwick GmbH, Ulm, Germany) with a Vickers intender along the material transition area.

Tensile tests were carried out at room temperature using Zwick/Roell Z050 (Zwick GmbH, Ulm, Germany) testing machine. Figure 2 schematically shows the samples used for tensile tests of the graded materials. The gauge length was 45 mm and the width was 20 mm, while the thickness of the specimens was 3 mm. Three samples per point were used for the tensile tests.

**Figure 2.** Schematic representation of tensile specimens' configuration.

Specimens consisting entirely of Ti of Ti64 materials were labeled as I (Ti) and I (Ti64), respectively. The remaining specimens consisting of two materials are shown in Figure 2. Type II specimens were split in half and consisted of 50% Ti and 50% Ti64 materials. Type III and IV had an insert from Ti64 alloy located at the center of the specimen with different orientations of the insert with 50%/50% volume fraction of the alloys. The IV type specimen had two inserts from Ti64 alloy as shown in Figure 2.

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

A sample consisting of Ti and Ti64 materials fabricated by L-PBF is shown in Figure 3a. There are no visible differences between the zones of the sample externally. They have the same color and surface roughness. A small line along the alloy interface, caused by the thermal expansion of the lower zone during fabrication, is noticeable. An image of the transition zone section of the sample in the initial state obtained by computed tomography (Figure 3c) shows internal defects in the form of pores, the average size of which is about 50 μm. It is also possible to see the transition from one material to another using the computed tomography, as the Ti64 material has a lighter shade compared to pure Ti due to the difference in density. The residual pores in the material after the HIP were not detected by CT, but the transition zone from one material to the other can also be seen (Figure 3d).

**Figure 3.** (**a**) A photograph of the Ti/Ti64 sample, (**b**) microstructure of the transition zone, (**c**) as-built, and (**d**) HIPed specimens' volume obtained by CT-reconstruction.

The microstructure of the transition zone between the two materials is shown in Figure 3b. No visible defects in the form of a lack of fusion or cracks were found. After etching, a distinct transition zone can be seen between Ti and Ti64 zones. It can be seen that the transition zone has a thickness of about 50–100 μm, which corresponds to the 1–2 layer thickness used during the L-PBF process.

The microstructure of the Ti zone consists of fine martensitic α' needles, while the Ti64 zone exhibits a needle-like martensitic α' phase within the columnar primary β grains. The high solidification rates typical for the L-PBF process resulted in a metastable microstructure in the case of both alloys. Titanium undergoes an α → β phase transformation above 890 ◦C, and this allotropic phase transformation affects the microstructure and texture of the material. The Ti64 alloy undergoes an β ↔ α + β phase transformation at about 1000 ◦C [30]. However, the L-PBF process leads to metastable martensitic microstructure due to the high cooling rates up to 105 K/s [31]. The L-PBF process, accompanied by rapid solidification, leads to the formation of a martensitic microstructure with elongated grains of the primary β-phase filled with the finely dispersed lamellar α-phase. The partial remelting of the previous layers provides the epitaxial growth of such grains.

The results of a study of the chemical composition of the transition zone (Figure 4) showed that the Al and V content increases smoothly from the zone of Ti to the zone of Ti64. According to the measurement results, the width of the transition zone can be estimated at approximately 200 μm.

**Figure 4.** EDX results showing the change of V and Al composition distribution along the transition zone from Ti (**left**) to Ti64 (**right**) on the sample.

Depending on the heat treatment temperature and cooling rate, the titanium microstructure may have different morphology: equiaxed α-Ti grains inside the primary β-grains, a Widmanstett structure, and a lamellar or needle morphology of the α-phase [32].

The heat treatment of the Ti/Ti64 sample at 950 ◦C resulted in the transition of the martensitic α'-phase to α + β phase in the case of Ti64 zone and α-Ti phase grains in the case of Ti zone (Figure 5).

**Figure 5.** Microstructures of Ti/Ti64 sample after heat treatment: (**a**) Ti64 zone, (**b**) Ti zone, (**c**) the transition zone.

After heat treatment, the Ti zone consists of equiaxed α-Ti grains, which indicate the recrystallization processes [33]. The recrystallization process led to the disappearance of the preferential orientation of the grains. They had an equiaxial shape and size from 80 to 150 μm.

After HIP, the microstructure of the samples underwent changes similar to those after the heat treatment, but there were differences in the morphology. The microstructure of the Ti64 alloy zone (Figure 6a) also consists of α + β phases with lamellar morphology, formed as a result of martensitic α'-phase decomposition to α + β. The formation of the lamellar α-phase with a larger lamellar size and β-phase grains occurs in the Ti64 alloy, in comparison to heat-treated conditions. The increase in the α-phase lamellar size occurs both within and along the grain boundaries, which may lead to an increase in ductility as deformation is mainly found along the grain boundaries.

**Figure 6.** Microstructures of Ti/Ti64 sample after HIP: (**a**) Ti64 zone, (**b**) Ti zone, and (**c**) the transition zone.

The Ti zone (Figure 6b) after HIP has a microstructure of equiaxed α-Ti grains with larger sizes compared to heat-treated ones, which may be caused by the differences in the cooling rate at a different post-processing [29].

Figure 7 shows hardness distribution for the Ti/Ti64 samples along the material transition area of as-built, heat-treated, and HIPed samples.

The hardness of the Ti64 zone is higher compared to the Ti zone for all tested conditions. The as-built condition exhibited the highest hardness values for both the Ti and Ti64 zone due to metastable microstructures formed during the L-PBF process. After heat treatment and hot isostatic pressing, the hardness values decreased for both the Ti and Ti64 zones due to the stress relieving and martensite phase decomposition. Due to different cooling conditions in the vacuum furnace and with HIP, there were differences between the values of hardness. After HIP, the microstructure was slightly coarser in terms of grain and lamellar size compared to the heat-treated samples, which resulted in lower hardness values along all the zones for the HIPed sample.

**Figure 7.** Hardness distribution for Ti/Ti64 samples along the material change area for as-built, heat treated, and HIPed conditions.

Tensile tests of pure Ti, Ti64, and graded Ti/Ti64 material have been made for samples fabricated by L-PBF and subsequent hot isostatic pressing. The results are summarized in Table 1.

**Table 1.** The results of tensile tests of Ti, Ti64, and graded Ti/Ti64 samples produced by L-PBF with the subsequent HIP.


The values of yield and tensile strengths of all the graded samples were between the values for pure Ti and Ti64. The samples of the V type had the lowest strength values; this type had 4 Ti/Ti64 material change interfaces which transversely directed the forces applied during the tensile test. The type II samples turned out to be more strengthened; this type had one interface of Ti/Ti64 material changing, directed along the axis of the tensile during the test. Even the higher strength values were demonstrated by type III samples. This type was characterized by the 2 Ti/Ti64 material change interfaces along the axis of the tensile and the total area of the interfaces were twice as large as those of type II. The highest strength characteristics between the investigated graded materials were found in the type IV samples. The geometry of this sample type had 2 Ti/Ti64 material change interfaces along the axis of the tensile with the largest contact area between the different materials.

The graded samples did not exhibit high elongation values. The changes in the elongation values for different types of samples have a similar tendency to the strength values. The fracture of the type V samples occurred at the material interface. This was potentially due to a partial oxidation of the metal surface, or the cooling of the specimens during a material change. Other sample types had interfaces along the axis of the tensile, as well as a low elongation. Therefore, another possible reason for the low elongation may be the Ti64 zones having a higher yield strength which could limit the elongation of the Ti zones and lead to the formation and failure of stress concentrators with relatively low values of elongation.

It should be noted that the homogeneous specimens, as well as the type V specimens, were fabricated with a build direction along the tensile load, and the other specimen types were fabricated with a build direction perpendicular to the tensile load. As shown in the previous research [34], the strength of the horizontally fabricated samples was higher compared to the vertically built samples, which could contribute to the lower properties of the type V specimens.

The presented results of the investigation of the graded Ti/Ti64 samples allowed tailoring properties for possible applications of multimaterial parts.

### **4. Conclusions**

The investigation of samples with the graded chemical composition, Ti/Ti64, was presented in this work. The study of the transition zone structure showed that these samples had a distinct character and did not undergo significant changes during heat treatment and hot isostatic pressing.

The study of tensile mechanical properties showed that, when the zones are located along the sample, the ratio of cross-sections has a greater influence on the mechanical properties than their shape and location. When the zones are arranged transversely to the specimen, a failure occurs at the interface and the relative elongation is extremely low. Future investigations in multimaterial 3D printing must pay attention to the possibility of creating change interfaces with the smooth changing of chemical compositions and an increasing transient zone.

**Author Contributions:** Conceptualization, E.B., I.P. and V.S.; Data curation, K.S.; Investigation, E.B., I.P. and K.S.; Project administration, E.B. and V.S.; Resources, A.P.; Visualization, E.B.; Writing original draft, E.B. and V.S.; Writing—review and editing, I.P., E.B. and V.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** The research is partially funded by the Ministry of Science and Higher Education of the Russian Federation as part of World-class Research Center program: Advanced Digital Technologies (contract No. 075-15-2020-934 dated 17 November 2020).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


### *Article* **Structure and Properties of Barium Titanate Lead-Free Piezoceramic Manufactured by Binder Jetting Process**

**Vadim Sufiiarov \*, Artem Kantyukov, Anatoliy Popovich and Anton Sotov**

Institute of Mechanical Engineering, Materials, and Transport, Peter the Great St. Petersburg Polytechnic University, 195251 Saint Petersburg, Russia; kantyukov.artem@mail.ru (A.K.); director@immet.spbstu.ru (A.P.); SotovAnton@yandex.ru (A.S.)

**\*** Correspondence: vadim.spbstu@yandex.ru

**Abstract:** This article presents the results of manufacturing samples from barium titanate (BaTiO3) lead-free piezoceramics by using the binder jetting additive manufacturing process. An investigation of the manufacturing process steps for two initial powders with different particle size distributions was carried. The influence of the sintering and the particle size distribution of the starting materials on grain size and functional properties was evaluated. Samples from fine unimodal powder compared to coarse multimodal one have 3–4% higher relative density values, as well as a piezoelectric coefficient of 1.55 times higher values (d33 = 183 pC/N and 118 pC/N correspondingly). The influence of binder saturation on sintering modes was demonstrated. Binder jetting with 100% saturation for both powders enables printing samples without delamination and cracking. Sintering at 1400 ◦C with a dwell time of 6 h forms the highest density samples. The microstructure of sintered samples was characterized with scanning electron microscopy. The possibility of manufacturing parts from functional ceramics using additive manufacturing was demonstrated.

**Keywords:** additive manufacturing; binder jetting; lead-free piezoceramic; barium titanate; sintering; piezoelectric properties

### **1. Introduction**

Functional ceramics are a class of materials that exhibit special properties in addition to those already inherent in ceramics, such as chemical and thermal stability. Functional ceramics typically exhibit one or more unique properties: biological, electrical, magnetic, or chemical. [1]. Due to this, they are used in engine production, aviation, and space industries [2]. The most promising types of functional ceramics include piezoceramics [3]. Piezoelectric ceramics are used to make sensor devices, energy harvesters, and actuators [4–6]. Piezoelectric materials are of particular interest as pressure and temperature sensors in high-frequency environments [7,8]. Piezoelectric ceramics have generated particular interest in the power industry because they can withstand the harsh environmental conditions present in energy conversion systems [9]. However, despite their advantages as sensitive devices, piezoceramics also have the same internal disadvantages that are observed in most ceramic materials: they are difficult to process [10], and their fragility causes low fracture resistance [11]. Therefore, the manufacture of non-standard complex geometries from ceramic materials can be practically impossible using conventional manufacturing methods. The proposed method for circumventing this problem is the manufacture of complex ceramic parts using additive manufacturing (AM) [12]. AM has such advantages as the absence of expensive tools, easy scalability of the process, the ability to implement parts of complex shapes, a high degree of material utilization and minimum production time [13]. One of the most relevant materials for ceramic additive manufacturing is a piezoelectric material since it generates an electric charge when deformed or, conversely, deforms when an electric potential is applied. The use of AM for the manufacture of piezoelectric materials will expand the scope of their application, expanding the

**Citation:** Sufiiarov, V.; Kantyukov, A.; Popovich, A.; Sotov, A. Structure and Properties of Barium Titanate Lead-Free Piezoceramic Manufactured by Binder Jetting Process. *Materials* **2021**, *14*, 4419. https://doi.org/10.3390/ma14164419

Academic Editor: Haim Abramovich

Received: 18 June 2021 Accepted: 4 August 2021 Published: 6 August 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

possibilities of forming multilayer, as well as complex geometries of structures. Therefore, the possibility of integrating piezoelectric materials during the manufacturing process itself would lead to the creation of multifunctional structures within a single processing process. With great freedom in the achievable geometries of piezoelectric elements, the prospect opens for a significant improvement in the performance of many devices based on piezoelectric and ferroelectric properties [14,15].

Some previous investigations of additive manufacturing lead-based piezoceramics have been made using direct writing/FDM [16,17], stereolithography-based processes [18,19], and ink-jetting [20]. Due to the toxicity of lead compounds, the development of new piezomaterials and technologies is moving towards lead-free piezomaterials. Barium titanate (BaTiO3) is one of the most widely used lead-free piezoceramic materials, which became widespread due to its high dielectric and piezoelectric properties [21]. The most promising methods of 3D printing BaTiO3 are direct writing (DW) [22–25], vat photopolymerization (VP) [26–30], and binder jetting (BJ) [31–34]. Samples printed using DW have the best piezoelectric coefficient values (d33 = 200 pC/N [22]) and a density (6.01 g/cm<sup>3</sup> [23]) close to the theoretical density limit of BaTiO3 (6.02 g/cm3). However, the quality of the surface printed layer is rather rough, which may be a limitation for using this technology. It is also worth noting that difficulties arise when using printing nozzles with a diameter less than 500 microns—the nozzle can clog with ceramic powder particles, and this reduces the accuracy of printing parts [23]. Samples printed using VP also have high piezoelectric coefficient values (d33 = 165 pC/N [27] and high material density (5.64 g/cm3 [26]). However, there are next limitations when using this 3D printing technology [29]: (i) using 3D printer with a layer spreading system for viscous slurry with a high solid loading of ceramics in the photopolymer resin; (ii) the high refractive index of UV light for BaTiO3 that limits the curing depth; (iii) a long time of debinding process that directly affects the final result of the subsequent sintering. Samples printed using BJ have low piezoelectric coefficient values (d33 = 74.1 pC/N [31], d33 = 112 pC/N [32] and have a low density (3.93 g/cm3 [31], 2.21 g/cm3 [32]). However, this technology ensures the high quality of printed parts and excludes difficulties of the debinding process.

The BJ additive process is a method where a nozzle print-head jets a liquid binder on a powder layer in places that correspond to the cross-section of the computer model of a part. The result of BJ printing is a green model with low mechanical properties and high porosity. The green model needs further curing, debinding, and sintering. As a result, the characteristics of parts made of the polymer [35], metal [36–38] and ceramics [39] printed on a 3D printer largely depend on manufacturing and postprocessing parameters. Consequently, the behavior of functional ceramics made with additive technologies must be further studied to expand the capabilities of this new technique.

In this paper, BaTiO3 lead-free piezoceramic was used to study the additive manufacturing of piezoelements by using the BJ process. The influence of the manufacturing process on the properties of the material was characterized and discussed, and the dielectric and piezoelectric properties of the manufactured samples were measured.

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

### *2.1. Materials*

Two types of BaTiO3 powder were used for printing by BJ: (i) micron powder with multimodal particle size distribution (PSD) (C-BaTiO3, ZAO NPF Luminofor, Stavropol, Russia) D10—0.1 μm, D50—3.4 μm, D90—25.4 μm, and (ii) submicron powder with unimodal PSD (F-BaTiO3, Acros Organics, Geel, Belgium) D10—0.6 μm, D50—1.1 μm, D90—2.1 μm. Figures 1 and 2 shows images of as-received powders. The C-BaTiO3 powder has particle sizes of about 1 μm, forms agglomerates up to 25 μm (Figure 1). The F-BaTiO3, powder has particle sizes about 1 μm (Figure 2). Figure 3 shows the particle size distributions which demonstrates that the C-BaTiO3 powder consists of agglomerates and reveals three peaks and contains small particles. The unimodal powder is much more homogeneous, while in the case of the multimodal powder one can observe the agglomeration of small particles

**Figure 1.** SEM images of C-BaTiO3 powder at different magnifications: general view (**a**), agglomerate (**b**), powder particles (**c**).

**Figure 2.** SEM images of F-BaTiO3 powder at different magnifications: general view (**a**), powder particles (**b**).

**Figure 3.** Particle size distribution of F-BaTiO3 (**a**) and C-BaTiO3 (**b**).

into larger clusters, which correspond to the third peak in the PSD with a medium size at about 20 microns.

### *2.2. Fabrication*

For the experiment green models of cubic shape with dimensions of <sup>10</sup> × <sup>10</sup> × 10 mm<sup>3</sup> were printed to study subsequent debinding and sintering processes. Also, the two types of cylindrical green models (diameter 15 mm and 10 mm, height 10 mm and 1 mm respectively) were printed to investigate the electromechanical properties.

The piezoceramic samples were manufactured on the ExOne Innovent system (The ExOne Company, North Huntingdon, PA, USA). This system relates to the BJ additive manufacturing process. The original ExOne BS004 solvent binder and CL001 cleaner were used for the printing of the functional ceramic components.

The BJ process can be divided into several stages, a schematic image of which is shown in Figure 4:


**Figure 4.** Process flow for BJ additive manufacturing.

Printed parts are considered "green" and are not suitable for end-use. Thus, these green models require further post-processing, such as sintering or infiltration, to achieve the desired mechanical and functional properties.

After the 3D printing, the platform (together with the green models in the powder surround) is placed in the thermal furnace (Yamato DX412C, Yamato Scientific, Santa Clara, CA, USA) at 180 ◦C for 3 h for curing.

After curing, the green models have sufficient strength to remove excess and loose powder. For green models of a simple shape, removal was done with a brush; for complex shapes removal was done using compressed air.

### *2.3. Thermal Post-Treatment*

Before thermal post-treatment, the green models were placed in alumina crucibles with lids. The debinding process was performed in a muffle furnace (KJ-1700X, Zhengzhou Kejia Furnace Co., Ltd., Zhengzhou, China) at a temperature of 650 ◦C with a dwell time of 60 min. After debinding, the samples were sintered in a muffle furnace at 1300, 1350, and 1400 ◦C with a dwell times of 2, 4, and 6 h under an air atmosphere. The thermal post-treatment profiles are illustrated in Figure 5.

**Figure 5.** The thermal post-treatment profile.

### *2.4. Characterization*

The particle size distribution of the powders was determined by laser diffraction Analysette 22 NanoTec plus (Fritsch, Idar-Oberstein, Germany) with a total measurement range of 0.01–2000 μm.

TGA analysis of BS004 solvent binder was performed using a thermogravimetric analyzer (Q5000, TA Instruments, New Castle, DE, USA). The heating was carried out in an airflow of 30 mL/min in the temperature range 30–700 ◦C at a rate of 10 ◦C/min. The binder was placed in a platinum crucible, after which it was heated from room temperature up to 700 ◦C in air.

The structure of the samples after sintering was studied using a Leica DMI5000 optical microscope (Leica, Wetzlar, Germany) and a Tescan Mira3 LMU scanning electron microscope (SEM) operating at magnifications from 4× to 106× with an accelerating voltage from 200 V to 30 kV. The chemical composition was measured using an energy-dispersion accessory into the SEM.

Optical and SEM-images of sintered samples from C-BaTiO3 and F-BaTiO3 were examined using the ImageJ Software. v.1.52a (Bethesda, MD, USA) The grain sizes were analyzed for various temperature and time sintering conditions.

The density of the sintered samples was measured by the Archimedes method; the calculation of relative density was made in accordance with the theoretical density of BaTiO3 (6.02 g/cm3).

The phase composition was analyzed using a Bruker D8 Advance X-ray (Bruker corp., Billerica, MA, USA) diffractometer (XRD) using CuKa radiation (l = 1.5418 Å) without monochromator.

All samples for the electrical performance test were coated with silver electrodes (paste PP-17, Delta, Zelenograd, Russia) at 700 ◦C for 30 min. The samples were poled in air, at Tc + 20 ◦C (Tc-Curie temperature 120–130 ◦C for BaTiO3). Then, an electric field of 0.6 kV/mm for 30 min was applied to samples, followed by cooling to room temperature. Dielectric constant *ε*´, the loss tangent tgδ, electromechanical coupling coefficient kp, and piezoelectric coefficient d33 were measured and calculated. Dielectric properties were measured on cylindrical samples with a diameter of 10 mm and a height of 1 mm. The capacity of the sample and the loss tangent were measured with an E7-28 immittance analyzer at 1 kHz frequency at 0.5 V effective voltage. The piezoelectric coefficient d33 was determined on polarized cylindrical samples using the APC YE2730A setup by a quasistatic method. The values of the electromechanical coupling coefficient were calculated by the following equation:

$$\mathbf{k\_{P}} = \sqrt{\frac{\mathbf{s\_{P}}}{\mathbf{a\_{P}} + \mathbf{b\_{P}} \cdot \boldsymbol{\delta\_{P}}}},\tag{1}$$

where, ap, bp are the coefficients determined of Planar Poisson's Ratio, δ<sup>p</sup> is the relative resonance gap. The Planar Poisson's ratio value was determined by the frequencies ratio of the third and first (main) overtones of the planar vibration mode on piezoelectric elements in the form of a disk.

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

### *3.1. Investigation of Debinding Process*

The TGA curve showed that when heated to 180.5 ◦C a sharp mass decrease by 86.82% was observed (Figure 6). This is due to the evaporation of two components: ethylene glycol monobutyl ether (EGBE), isopropanol (IPA), and the polymerization of ethylene glycol to polyethylene glycol. The boiling temperatures of evaporating components are much lower at 171 ◦C and 80.4 ◦C, respectively.

**Figure 6.** TGA analysis of the binder.

A further mass decrease occurred at a temperature range of 380–450 ◦C. As a result, the remaining mass of the binder was 0.82% of the initial one. Increasing the temperature leads to a linear decreasing of mass; the binder was almost completely thermally decomposed at 664 ◦C and the residue was 0.21% of the initial mass. Thus, mass loss of the binder is observed in two stages: the first stage—mass decreases on 86.85%, this stage ends at a temperature of 180.5 ◦C. The second stage is the temperature range from 180 to 664 ◦C. Here, from 180 to 447 ◦C, no significant mass loss occurs. From 447 to 664 ◦C, the mass loss is up to 0.21% of the original mass.

The first stage is associated with the transition of ethylene glycol to polyethylene glycol during curing, the second stage is debinding by the burn out of the residue components of the binder [40].

### *3.2. Binder Jetting Process*

For the BJ process, the recoating speed (28 mm/s and 65 mm/s for C-BaTiO3 and F-BaTiO3 respectively) and the frequency of the oscillator (5000 rpm and 4400 rpm for C-BaTiO3 and F-BaTiO3 respectively) were previously optimized to apply a sufficient amount of material to form a smooth thin powder layer. Considering this, the layer thickness for C-BaTiO3 and F-BaTiO3 powder was 100 μm and 35 μm, respectively. The drying time and temperature were also optimized to achieve a uniform layer without cracking and without smearing. The main BJ parameters are shown in Table 1.

**Table 1.** The main BJ parameters for BaTiO3 lead-free piezoceramic powder.


Further, the saturation parameter was investigated. Binder saturation is a computed value used to quantify how much binder is dispensed into each unit volume of powder material. Improper saturation of the binder can cause an inhomogeneous layer of powder as well as inaccurate dimensions of printed parts. The theoretical binder saturation (%) was estimated using the following equation:

$$\mathbf{S} = \frac{1000 \times \mathbf{V}}{(1 - (\frac{\text{PR}}{100})) \times \mathbf{X} \times \mathbf{Y} \times \mathbf{Z}},\tag{2}$$

where V is the volume of binder per drop (pL), PR is the packing rate (%), X and Y are the spacing between binder droplets (μm), and Z is the layer thickness (μm). To obtain the green part with sufficient mechanical strength and surface quality, optimizing the saturation level is critical.

The saturation for C-BaTiO3 powder varied from 40 to 140% with a step of 20%. For F-BaTiO3 powder, the saturation varied from 50 to 200% with a step of 50%. When printing the C-BaTiO3 and F-BaTiO3 samples, no defects were observed on the surface of the powder layer. The powder layer was applied uniformly, the particles did not stick to the roller. After curing of C-BaTiO3 samples printed at 40% saturation, the green model delaminated. At 60% and 80% saturation, the deviation from the computer model size amounted to 0.37 mm along the X and Y-axes and more than 0.1 mm along the Z-axis. For 100% saturation, the C-BaTiO3 samples had clear boundaries and the deviation from the computer model size was about 0.2 mm along the X and Y-axes, and less than 0.05 mm along the Z-axis. At 120% and 140% saturation, the geometry of the green models changed significantly and appeared to be barrel-shaped.

After curing of F-BaTiO3 samples printed at 50% saturation, the green model delaminated since there was not enough binder to bond the layers together. For 100% saturation, the F-BaTiO3 samples had clear boundaries and the deviation from the computer model size was about 0.2 mm along the X and Y-axes, and less than 0.02 mm along the Z-axis. At 150% and 200% saturation, the deviation from the computer model size amounted to

0.36 mm and 0.38 mm along the X and Y-axes and more than 0.21 mm and 0.25 along the Z-axis, respectively.

### *3.3. Investigation of Sintering Process, Shrinkage, Microstructure, Porosity*

For investigation of sintering process BaTiO3 samples, the following samples were selected: C-BaTiO3 samples printed at 60, 80, and 100% saturation; F-BaTiO3 samples printed at 100, 150, and 200% saturation.

To understand the influence of saturation level test-sintering was carried out at 1400 ◦C with a dwell time of 4 h. Samples of C-BaTiO3 printed with 60 and 80% saturation after test-sintering delaminated due to the weak contact between the layers. Also, F-BaTiO3 samples cracked at 150 and 200% saturation, which is due to the high content of the binder. Seemingly, due to the high content of the binder in samples, during debinding and subsequent sintering, the formation of a large amount of gas occurred, leading to the appearance of cracks. However, the C-BaTiO3 and F-BaTiO3 samples printed at 100% saturation after test-sintering were free from defects. Figure 7 shows an image of F-BaTiO3 samples obtained at different saturations after test-sintering. As a result, samples with 100% saturation for both types of powder were selected for further investigation of the sintering process.

**Figure 7.** Images of F-BaTiO3 samples printed by BJ with different saturation after sintering. Scale: each sample has a diameter of 15 mm and a height of 10 mm. Heating rate of 10 ◦C/min to 1400 ◦C with a dwell time of 4 h.

Subsequently, these samples were subjected to sintering in the temperature range of 1300–1400 ◦C for 2–6 h. Sintering experiments at temperature 1500 ◦C led to the melting of BaTiO3 and the destruction of the samples. Initially, the study of the sintering process was carried out for C-BaTiO3 samples at various temperatures of 1300, 1350, and 1400 ◦C. The best value for the density of the material was achieved at 1400 ◦C. Considering that the particle size of the C-BaTiO3 powder is close to the particle size of the F-BaTiO3 powder (but different agglomerates sizes), a further investigation of sintering for the F-BaTiO3 samples was carried out at a temperature of 1400 ◦C with dwell times of 2, 4, and 6 h.

Figure 8 shows graphs of the dependence of sintered samples density on dwell time. The density of the samples increases with an increasing dwell time. The density of F-BaTiO3 samples is higher compared to C-BaTiO3 samples. The density of C-BaTiO3 samples is lower, but the printing speed is higher due to layer thickness difference. During the printing of cylindrical samples from F-BaTiO3 with 15 mm diameter and 10 mm height, the printing time was 7 h, which is 4 h longer than the printing time for similar size samples from C-BaTiO3 powder.

**Figure 8.** Dependence of the density of samples on temperature and dwell time of sintering.

Increasing the temperature and dwell time of sintering leads to grain enlargement. The graphs in Figure 9 show that the grain size of samples from the unimodal powder is more sensitive to changes in temperature and dwell time compared to samples made from multimodal powder. This feature allows adjusting the functional properties of the material in a wider range.

**Figure 9.** Grain size of BaTiO3 samples dependence on temperature and dwell time of sintering.

As a result of the sintering of samples from C-BaTiO3, the shrinkage along the XY direction was 20–25%. The Z-axis shrinkage varied from 24.1% to 24.4%. The measured linear shrinkage of samples from F-BaTiO3 along the XY direction was 24–27%. The Z-axis shrinkage was 25–26%.

The microstructure of the sintered BaTiO3 samples is shown in Figure 10. The structure is a rounded grain formed as a result of sintering the powder material. Some sintered samples have round-shaped pores, these defects may be associated with binder removal since at this stage there is active gas formation, and perhaps a consequence of the nonoptimal sintering mode as well.

**Figure 10.** SEM images of sintered samples microstructures at 1400 ◦C, a dwell time of 6 h from C-BaTiO3 (**a**) and F-BaTiO3 (**b**) powders.

According to EDS measurements, the chemical composition of samples was 59.2% of Ba, 18.8% of Ti, and 22% of O (weight %) which corresponds with BaTiO3 formulation. Figure 11 shows the diffraction patterns of the C-BaTiO3 samples. X-ray diffraction analysis showed that all samples are composed of the tetragonal crystal lattice P4mm of BaTiO3, as evidenced by bifurcated peaks (compare to cubic lattice Pm-3m).

**Figure 11.** XRD of a sample sintered at different temperatures.

To demonstrate the applicability of the developed modes of the BJ process and thermal post-treatment for manufacturing parts with complex geometries, test samples with lattice structures were made from F-BaTiO3 powder (Figure 12).

**Figure 12.** Image of samples with lattice structures printing by BJ from F-BaTiO3 powder before (**left**) and after sintering (**right**).

### *3.4. Investigation of Functional Properties*

The investigation of the functional piezoelectric properties was carried out for C-BaTiO3 samples (a temperature of 1400 ◦C and a dwell time of 6 h) and F-BaTiO3 (a temperature of 1400 ◦C and a dwell time of 4 h). These samples were selected considering the highest density and grain size up to 50 microns. This grain size is due to the fact that, for BaTiO3-based piezoceramic, the high functional properties arise with a grain size of 10 to 50 μm [41]. Table 2 shows the test results of the functional properties of sintered samples manufactured from multimodal and unimodal BaTiO3 powders. Samples printed from C-BaTiO3 powder are inferior in dielectric constant, electromechanical coupling coefficient, and piezoelectric coefficient to samples printed from F-BaTiO3. This can be explained by the non-optimal mode of debinding and sintering, the presence of large pores, and as a result, a decrease of the active phase volume of the sample.


**Table 2.** Piezoelectric properties at 1 kHz of BaTiO3 samples printed by BJ process.

Appreciating the main parameter piezoelectric coefficient d33, it can be noted that using the BJ process allows achieving 72.4% of the piezoelectric coefficient compared to the value obtained by traditional manufacturing technology with multimodal PSD powder and 79.6% of the d33 values obtained with unimodal PSD powder. Pressing and sintering were used as the traditional technology, and a solution of polyvinyl alcohol was used as a binder. Sintering was carried out at a temperature of 1350 ◦C, heating rate 100 ◦C/h, a dwell time of 3 h.

According to the results of studies published in [32], the functional characteristics of AM piezoceramics depend on the direction of measurement. The functional properties along the Z-axis are about 20% smaller in comparison with the XY direction. In the current study, the properties were measured only along the Z-axis, but the achieved values of piezoelectric coefficient d33 = 183 pC/N and dielectric constant *ε*´ = 811 exceed the values obtained by the authors [32] parallel (d33 = 113 pC/N, *ε*´ = 581.6) and perpendicular to the printing orientation (d33 = 152.7 pC/N, *ε*´ = 698). These differences seem to be related to the raw material and the corresponding difference in technological parameters of BJ and subsequent thermal post-treatment.

The presented results demonstrate that the use of a unimodal PSD powder of lead-free piezoceramics barium titanate allows achieving higher piezoelectric properties, and the use of binder jetting technology allows the creation of objects with complex geometry, which has potential in the manufacture of ultrasonic products used in medicine, aviation, marine industry, sensors for monitoring welded joints, pressure sensors in pipelines, etc.

Future research areas that allow for improving piezoelectric properties include the use of new lead-free piezoelectric materials with increased characteristics (such as KNN, BZT-BCT, etc.), as well as the creation of functional gradient systems and the use of multimaterial 3D printing.

### **4. Conclusions**

The paper presents the results of the additive manufacturing of piezoelectric elements using the binder jetting process. Two powders with different particle size distributions were used as raw materials. Binder jetting with 100% saturation for C-BaTiO3 and for F-BaTiO3 allows printing samples without delamination and cracking. Sintering at 1400 ◦C with a dwell time of 6 h forms the highest density samples. It was determined that samples from the unimodal powder are more sensitive to increasing grain size during sintering. The measured dielectric and piezoelectric properties of the samples also demonstrated that samples from unimodal powder F-BaTiO3 have higher values. The results of the functional piezoelectric properties obtained by binder jetting with C-BaTiO3 are d33 = 118 pC/N, *ε*´ = 750, and with F-BaTiO3: d33 = 183 pC/N, *ε*´ = 811.

The future possibilities of improving functional characteristics of samples manufactured with BJ are increasing speed, optimizing sintering modes, and using new lead-free piezoelectric materials with improved functional characteristics.

**Author Contributions:** Conceptualization, V.S.; investigation, A.K. and V.S.; methodology, A.S.; project administration, V.S. and A.P.; funding acquisition, A.P.; writing—original draft, A.K. and A.S.; writing—review & editing, A.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** The research was carried out as part of the work under the State Contract No. H.4щ.241.09. 20.1081 dated 04.06.2020 (ISC 17706413348200001110).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


### *Article* **Structure, Mechanical and Magnetic Properties of Selective Laser Melted Fe-Si-B Alloy**

**Vadim Sufiiarov 1,\*, Danil Erutin 1, Artem Kantyukov 1, Evgenii Borisov 1, Anatoly Popovich <sup>1</sup> and Denis Nazarov <sup>2</sup>**


**Abstract:** Original 1CP powder was studied and it was founded that powder material partially consists of the amorphous phase, in which crystallization begins at 450 ◦C and ends at 575 ◦C. Selective laser melting parameters were investigated through the track study, and more suitable ones were found: laser power *P* = 90, 120 W; scanning speed *V* = 1200 mm/s. Crack-free columnar elements were obtained. The sample obtained with *P* = 90 W, contains a small amount of amorphous phase. X-ray diffraction of samples shows the presence of α-Fe(Si) and Fe2B. SEM-image analysis shows the presence of ordered Fe3Si in both samples. Annealed samples show 40% less microhardness; an annealed sample containing amorphous phase shows higher soft-magnetic properties: 2.5% higher saturation magnetization, 35% higher residual magnetization and 30% higher rectangularity coefficient.

**Keywords:** selective laser melting; soft-magnetic alloy; FeSiB; magnetic properties; additive manufacturing

### **1. Introduction**

Selective laser melting (SLM) technology is an additive manufacturing process by layerby-layer melting of a 20–60 μm thick powder layer of material using a laser [1,2]. One of the most attractive applications of this technology is the production of bulk amorphous alloys.

Amorphous materials are a type of solid materials in which there is no long-range order in the arrangement of atoms [3]. This state of the material is achieved at high cooling rates from the liquid state due to the fixation of atoms in the positions in which they were in the melt. An amorphous metallic material (metallic glass) does not have a crystalline lattice; therefore, its atomic structure lacks the crystalline defects that cause anisotropy of its properties. Metallic glasses based on ferromagnetic alloys exhibit better soft-magnetic properties than crystalline ferromagnetics: lower coercive force values, higher values of saturation magnetization, higher magnetic permeability and electrical resistivity. Due to the higher level of soft magnetic properties of metallic ferromagnetic glasses, their use as a magnetic core material of an electromagnetic device allows for increasing its efficiency by significantly reducing the magnetic field energy losses for remagnetization and eddy currents [4].

High magnetic properties are achieved with amorphous phases, and materials consisting of nanocrystalline inclusions evenly distributed in the amorphous matrix are also known. Currently, amorphous and nanocrystalline materials produced by additive manufacturing techniques are being investigated by many research groups worldwide [1,4–20]. The problem of obtaining samples with amorphous phase from iron-based soft-magnetic alloys using SLM has been solved with varying degrees of success by selecting process parameters [1,4,11,14,18,19] and by developing and applying specially designed materials and laser beam scanning strategies [4,10,11,19]. Obtaining a material with a minimum

**Citation:** Sufiiarov, V.; Erutin, D.; Kantyukov, A.; Borisov, E.; Popovich, A.; Nazarov, D. Structure, Mechanical and Magnetic Properties of Selective Laser Melted Fe-Si-B Alloy. *Materials* **2022**, *15*, 4121. https://doi.org/10.3390/ma15124121

Academic Editor: Jae Wung Bae

Received: 28 March 2022 Accepted: 7 June 2022 Published: 9 June 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

degree of crystallinity involves the suppression of crystallization both when cooling the molten metal and when absorbing the heat of the melt by the already solidified part of the product. This task requires researchers to thoroughly understand the impact mechanisms on the resulting material, which makes investigation of the influence of process parameters on the characteristics of the resulting material a priority step towards its solution.

The aim of this work was to investigate the effect of the selective laser melting process parameters (laser power *P*, hatch distance *h*, offset *m*) on macrostructure and microstructure, phase composition, magnetic and mechanical properties of 1CP magnetic alloy samples and to investigate the effect of thermal treatment of samples on their magnetic and mechanical properties.

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

The flowability of the powder was determined using ISO 4490 "Determination of flow rate by means of a calibrated funnel (Hall flowmeter)". Apparent density measurements were made by pouring the powder into a funnel from which it flowed into a 25 cm3 cup. After filling the cup, the funnel was moved away and excess powder was smoothed out with a trowel. Apparent density was determined by weighing the powder in the cup in grams and dividing by 25 cm3. The skeletal density of the powder was determined in accordance with GOST 22662-77.

The particle size distribution of the powder was determined by laser diffraction on the Analysette 22 NanoTec plus (Fritsch, Germany) with a total measuring range of 0.01–2000 μm. The microstructure of the powder and the obtained samples were studied using a Tescan Mira3 LMU scanning electron microscope (SEM). The etching of the samples was carried out in a 10% nitric acid solution in isopropyl alcohol. The fine structure of the powder was investigated using a Carl Zeiss Libra 200FE transmission electron microscope (TEM) with an energy Ω filter and an operating accelerating voltage of 200 kV. An HAADF detector (STEM mode) and a CCD camera (TEM mode) were used to obtain images. Electron diffraction patterns were measured using an aperture diameter of 1000 nm and a camera length of 450 mm. For the measurement of chemical composition, electron energy loss spectroscopy (EELS) in TEM was used and theoxygen content was measured by infrared absorption and thermal conductivity analysis on LECO TC-500 (LECO Corporation, St. Joseph, MI, USA).

Temperatures of phase transitions were studied using differential scanning calorimeter (DSC) Q2000 (TA Instruments, New Castle, DE, USA) equipped with an automatic sampler, RCS90 cooling system and T-zero baseline alignment technology. Samples were heated in an argon flow to a temperature of 1000 ◦C at a heating rate of 20 ◦C/min followed by second heating of the cooled samples to the same temperature. The phase composition was analyzed with a Bruker D8 Advance X-ray diffractometer (XRD) using Cu Kα (l 1/4 1.5418 Å) irradiation.

The magnetic hysteresis loops for the samples were measured by Lake Shore 7410 vibration sample magnetometer (VSM) (Lake Shore Cryotronics, Westerville, OH, USA) at room temperature (22 ◦C) and under applied different magnetic fields from −18,000 to +18,000 Oe. Magnetic measurements were carried out on 2 sets for each type of sample.

The hardness of the samples was determined using a Buehler Micromet 5103 microhardness tester using the Vickers method at 3 N. To determine the mean value, 5 tests were performed.

Samples were manufactured using an SLM280HL (SLM Solutions GmbH, Lübeck, Germany) selective laser melting system equipped with YLR-Laser (wavelength of 1070 nm and focus size about 80 μm) under a nitrogen atmosphere.

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

### *3.1. Powder Material*

The first stage of the research was to investigate the morphology, phase composition, physical, technological and magnetic properties of the initial 1CP powder, consisting of

iron and alloying elements: boron, carbon and silicon [21]. The chemical composition of the initial powder is presented in Table 1. The technology of the 1CP powder manufacturing was gas atomization [22].

**Table 1.** The chemical composition of 1CP powder.


The SEM image in Figure 1 shows that the powder material is spherical and rounded particles.

**Figure 1.** SEM image of 1CP powder.

The results of the particle size distribution of the powder are shown in Table 2. The initial powder particle size is Gaussian distributed with a mean value of 41.8 μm. This is a typical range for use in selective laser melting [18,23].

**Table 2.** Particle size distribution of 1CP powder.


The results of the investigation of the physical and technological properties of the powder are shown in Table 3. The ability to flow freely through the Hall funnel indicates the possibility of good powder spreading during the formation of thin powder layers in the selective laser melting process. The apparent density is 56.8% of the skeletal density, which indicates an acceptable packing density formed by this powder during the formation of the powder layer.

**Table 3.** Physical and technological properties of the 1CP powder.


The X-ray diffraction pattern of the powder sample is shown in Figure 2. The following phases are present in the sample: solid solution α-Fe(Si) and iron boride Fe2B.

**Figure 2.** X-ray diffraction pattern of 1CP powder.

Figure 3 shows the DSC results of the powder material presented by two curves: the red curve for the primary heating of the original material and the blue one for the secondary heating of the material (cooled down after primary heating). The primary heating curve shows peaks indicating a phase transformation during heating. This process occurs for 1CP alloy powder in the temperature range from 450 to 575 ◦C. The absence of the secondary heating peaks shows the accordance of the peaks to the crystallization process. Based on the DSC data it could be concluded that heating above 450 ◦C would lead to the beginning of crystallization processes of the amorphous phase in case of the presence of the amorphous phase in the samples during heat treatment of samples obtained from this powder. According to this data, it was decided to use annealing heat treatment of samples at 440 ◦C. This annealing temperature corresponds to the recommended temperature for amorphous ribbons from 1CP [21] and the heat treatment mode used in further study: heating at a rate of 10 ◦C/min to 440 ◦C, holding for 30 min, cooling outside the furnace. The halo, which is not clearly visible in the diffraction pattern (Figure 2), is partially visible in the region of 2Θ = 42–47. The absence of an obvious halo can be explained by the small volume content of the amorphous phase in the powder material.

**Figure 3.** DSC heating curves for the 1CP powder ("exo" means that presented peaks are corresponded to exothermic process).

Figure 4 shows the results of investigation microstructure of 1CP powder by transmission electron microscopy.

**Figure 4.** High-resolution TEM images of 1CP metal powder particles (**a**,**b**), electron diffraction pattern of the amorphous phase referring to the light region (**c**), electron diffraction pattern of the crystalline phase referring to the dark region (**d**).

Two phases are present in the studied powder, one of which has a crystalline structure as evidenced by electron diffraction (Figure 4d) and the other has an amorphous structure as evidenced by electron diffraction in Figure 4c. The amorphous phase is present both as separate areas (upper part of Figure 4a) and as areas distributed around the crystalline phase (Figure 4b and lower part of Figure 4a).

The results of the study on the magnetic properties of the powder are shown in Table 4. The hysteresis loop of the powder is shown in Figure 5. 1CP can be considered as a soft magnet with a relatively high coercive force and a low residual magnetization, but a huge saturation magnetization.

**Table 4.** Magnetic properties of the 1CP powder.


**Figure 5.** Magnetic properties measurement results of the 1CP powder: (**a**)—general view of hysteresis loop; (**b**)—enlarged area for coercivity estimation.

### *3.2. Single Track Study*

In order to determine the range of applicability parameters for the selective laser melting process, a series of single tracks were melted on a 1CP substrate using different values of laser power *P* and scanning speed *V*, which were selected after the preliminary tests have been made with various values of laser power and scanning speed and provided continuous tracks.

The modes used for single-track series are presented in Table 5. The linear energy density is calculated as the ratio of a laser beam power to a scanning speed.


**Table 5.** Single track modes.

The SEM images of the tracks presented in Figure 6 show that there are transverse cracks repeated at distances greater than or close to 200–300 μm. A similar pattern is observed for all the tracks. Therefore, it was decided not to use values of one pass laser length exceeding 200 μm in the next experiments.

Figure 7 shows SEM images of the structure of the melted tracks in a cross-sectional view. The geometric characteristics of the resulting tracks are shown in Table 6. Track 1 has acceptable geometrical characteristics, but there is a pore in its cross-section and a crack at the border with the substrate. The linear energy density of the mode of this track is 75 J/m, as well as of track 4, which has good geometrical characteristics and has no visible defects, but the scanning speed used in the growth of the first track was too low for the used power, which led to the formation of defects. Tracks 2 and 3 were formed at lower linear energy densities (60 J/m and 50 J/m), which were insufficient to make a track with acceptable deposit height.

**Figure 6.** SEM images of top-views single tracks (numbers **1**–**5** denotes the building mode of Table 5).

**Figure 7.** SEM images of cross-sectional views single tracks (numbers **1**–**5** denotes the building mode of Table 5).


Tracks 4 and 5 have good geometrical characteristics (sufficient height of the deposited metal and penetration depth) and no visible defects, so modes 4 and 5 are used in the next experiments.

Thus, it was decided to use a melt track length not exceeding 200 μm, with values of *P* = 90 W, *P* = 120 W and *V* = 1200 mm/s.

### *3.3. Selective Laser Melting of Samples Investigation*

As part of the study, eight rectangular samples were manufactured. Samples have been successively made in a nitrogen atmosphere. The plane orthogonal to the height of the cube was divided into cells, each containing two cross-sections of columnar elements, the distance between the centers of which corresponds to the hatch distance parameter h, with the distance between cells corresponding to the offset parameter m. The length of one pass of the laser beam also corresponds to the parameter *h*. The building scheme is presented in Figure 8.

**Figure 8.** Sample building scheme (arrows in the cells indicates the scanning direction).

The image of manufactured samples is shown in Figure 9. The build modes are presented in Table 7, the scanning speed *V* and the thickness of the powder layer *t* were fixed *V* = 1200 mm/s, *t* = 50 μm.

**Figure 9.** Image of samples manufactured by the SLM process from 1CP alloy powder (numbers **1**–**8** denotes the building mode of Table 7).


**Table 7.** Modes of selecting laser melting used for manufacturing samples.

The samples manufactured at *P* = 90 W (Figure 10, 1–4) are less dense than those made at *P* = 120 W (Figure 10, 5–8). Increasing laser power allows the formation of larger structural elements due to the melting of a larger volume of initial powder, which leads to the formation of a denser structure. The increasing value of the offset (Figure 10, 1–4; 5–8) is accompanied by a decrease in the density of samples due to a violation of its structural unity caused by the separation of the columnar elements from each other. The hatch distance parameter h determines the presence of a merger of a pair of columnar elements into a single element: the samples obtained at *h* = 100 μm (Figure 10, 1–3; 5–7) are characterized by united elements, in contrast to the samples obtained at *h* = 200 μm (Figure 10, 4; 8).

**Figure 10.** SEM images of samples (top view) produced by selective laser melting with parameters according to Table 6 (numbers **1**–**8** denotes the building mode of Table 7).

Cross-sectional specimens were prepared for selected columnar elements of samples 4 and 8 (for these samples only the separation of single elements was possible) for examination with a scanning electron microscope. SEM images of the microstructure of the elements are shown in Figures 11 and 12.

The structure of element 8 is characterized by the shape of the layer expressed by the presence of an arc section on the boundary line of each layer. This phenomenon is associated with increased laser power *P*, the value of which for sample 8 was 120 W. In this case, the change in the shape of the layer is associated with deeper penetration of laser irradiation for a separate section of the layer and uneven distribution of thermal energy over the contact spot of the laser with the metal.

The phase composition was investigated by X-ray diffraction analysis. The X-ray diffraction patterns of the samples are shown in Figure 13.

**Figure 11.** Microstructure of columnar element of sample 4, studied in backscattered electrons mode of SEM.

**Figure 12.** Microstructure of columnar element of sample 8, studied in backscattered electrons mode of SEM.

**Figure 13.** X-ray diffraction patterns of samples 4 (**a**) and 8 (**b**).

Based on X-ray diffraction analysis it was found that the following phases are present in the sample: α-Fe(Si) solid solution and Fe2B iron boride. The third phase present in the microstructure images of the samples can be identified as an ordered Fe3Si solid solution. The morphology of the etched cavities is similar to the crystal morphology of this phase [24]. The α-Fe(Si) solid solution has a similar crystallographic structure to the ordered Fe3Si solid solution, due to which the X-ray diffraction analysis may not allow the detection of the reflexes of this phase if the α-Fe(Si) structure prevails [24]. Therefore, researchers [19,23] during the X-ray diffraction analysis of samples of Fe-Si-B alloy obtained by selective laser melting noted the Fe3Si phase together with α-Fe(Si) on the peaks corresponding to α-Fe(Si).

The obtained DSC curves (Figure 14) indicate the almost complete absence of crystallization processes during the heating of the samples. However, the curve of primary heating of sample 4 is characterized by the presence of small peaks, and their absence during secondary heating (which cannot be said for the curve of sample 8), indicating the presence of a small amount of amorphous phase in sample 4.

**Figure 14.** DSC curves of samples 4 (**a**) and 8 (**b**) ("exo" means that presented peaks are corresponded to exothermic process).

Onset crystallization temperatures and enthalpy of the process are presented in Table 8. TEM electron diffraction data presented in Figure 15 proves the presence of the amorphous phase of sample 4.



**Figure 15.** High resolution TEM images of sample 4 (**a**), electron diffraction pattern (**c**) of referring to the region (**b**).

Mechanical and magnetic properties were investigated for the initial and annealed samples. The purpose of annealing was to decrease the level of internal stresses in samples and investigate the effect of it for properties as defined above. The hardness data of the samples (Table 9) indicate that the hardness of sample 4 is slightly higher than that of sample 8. The difference between the mean hardness values is within the standard deviation of the samples (σ<sup>4</sup> = 155, σ<sup>8</sup> = 92). Hence, the laser power has no effect on the samples of this material in the investigated power range. The authors [1] investigated samples of a similar composition alloy and obtained hardness values close to those presented in this study. The annealed samples show an approximately 40% reduction values of hardness.


**Table 9.** Hardness of the samples obtained (HV0.3).

The study of the magnetic properties was carried out for samples 4 and 8. The magnetization curves of the samples are shown in Figure 16. The magnetization curves of the samples after heat treatment are shown in Figure 17. The main parameters of magnetic measurements are summarized in Table 8. The coercivity of the measured samples does not differ significantly from each other. At the same time, there is a difference in the shape of the hysteresis loop (sample 8 achieves a saturation at slightly lower values of field) and the values of residual magnetization. A comparison of the obtained results with the data for amorphous ribbons obtained by melt spinning technology for 1CP alloy [21] shows that the coercivity is much higher and the coefficient of rectangularity is lower for samples made by SLM.

**Figure 16.** Magnetic properties measurements results of samples 4 and 8: (**a**)—general view of hysteresis loop; (**b**)—enlarged area for coercivity estimation.

The change of coercivity of annealed samples is within the margin of error. Sample 4 showed higher values of saturation magnetization (2.5% higher), residual magnetization (35% higher) and rectangularity coefficient (30% higher) after annealing. At the same time, sample 8 after heat treatment shows almost the same values of magnetic parameters as before heat treatment. The changing of magnetic properties for sample 4 is possibly related to a relaxation of internal stresses and the presence of a small amount of amorphous phase, magnetic properties changing of which after annealing is stronger than the crystalline phase. Sample 8 has a lower value of internal stresses due to the higher laser power used for its manufacturing and demonstrates no changing magnetic properties after anneal-

ing. Therefore, the heat treatment mode recommended for amorphous ribbons of this material [21] should be reevaluated for selective laser melting samples.

**Figure 17.** Magnetic properties measurements results of samples 4 and 8 after annealing heat treatment: (**a**)—general view of hysteresis loop; (**b**)—enlarged area for coercivity estimation.

Further research requires the use of scanning strategies with different patterns and multiplicity, substrate heating and cooling experiments, and better optimization of physical and geometric process parameters and reaching more amorphous phase content.

### **4. Conclusions**

The effect of selective laser melting parameters on microstructure and magnetic properties of 1CP alloy was investigated in this work.


**Author Contributions:** Conceptualization, V.S. and E.B.; methodology, E.B.; validation, A.K.; formal analysis, A.P.; investigation, E.B., D.N. and D.E.; resources, V.S. and A.P.; data curation, A.K. and D.N.; writing—original draft preparation, D.E. and E.B.; writing—review and editing, V.S. and D.N.; visualization, D.E. and A.K.; supervision, V.S. and A.P.; project administration, V.S.; funding acquisition, V.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** The research was supported by a grant from the Russian Science Foundation № 21-73-10008, https://rscf.ru/project/21-73-10008.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The VSM measurements were conducted using the equipment of the resource centers of the Research Park of the St. Petersburg State University "Innovative Technologies of Composite Nanomaterials".

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Materials* Editorial Office E-mail: materials@mdpi.com www.mdpi.com/journal/materials

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel: +41 61 683 77 34

www.mdpi.com

ISBN 978-3-0365-4928-6