**About the Editor**

#### **Manoj Gupta**

Prof. Dr. Manoj Gupta received his Ph.D. degree from the University of California, Irvine, USA, in 1992, and is currently an Associate Professor at the Department of Mechanical Engineering at the National University of Singapore. In August 2017, he was highlighted among the Top 1% Scientist of the World Position by The Universal Scientific Education and Research Network, as well as among the top 2.5% by ResearchGate and the top 1% in Stanford's list of researchers. To his credit are (i) the disintegrated melt deposition technique and (ii) hybrid microwave sintering technique, an energy efficient solid-state processing method used to synthesize alloys/microcomposites/nanocomposites. He has published over 600 peer-reviewed journal papers, co-authored 8 books, and owns 2 US patents.

## *Editorial* **Editorial for the Special Issue "Reviews and Advances in Materials Processing"**

**Manoj Gupta**

Department of Mechanical Engineering, National University of Singapore, 9 Engineering Drive 1, Singapore 117576, Singapore; mpegm@nus.edu.sg

In the area of Materials Science and Engineering, the tetrahedron comprising of processing, microstructure, properties and performance as four vertex corners is always key to develop new materials and to convert them to a useful shape for end application with the best properties possible. As can be envisaged, processing is an integral part of this tetrahedron and the proper development and usage of processing for a given end goal is a must to take the product efficiently to the consumer. The processing of the materials comprises of primary processing techniques and secondary processing techniques. Primary processing of materials can be broadly but not limited to classified into four sub-categories irrespective of type of materials. These include: (a) liquid-phase processing, (b) solidphase processing, (c) two phase processing, and (d) vapor phase processing. Each of these subcategories have different types depending on a number of governing factors. Similarly, the secondary processing of materials is also very important for giving a desired shape to the materials and these include, for example, (a) forging, (b) rolling, (c) extrusion, (d) drawing, (e) joining and (f) machining. With the advancement in computing tools and the availability of advanced software, modelling and simulation has also almost become an integral part of the processing process so as to realize the best output from them through optimizing the processing/operating parameters.

In view of the intrinsic importance of 'Materials Processing', especially for the researchers in the area of materials science and manufacturing, this issue was conceived with a specific aim to bring to the notice of readers the latest advances in various processing types currently used for a wide spectrum of materials. A total of 14 papers were accepted and published in this issue. Among them, 3 are review papers and 11 are research papers targeting different research goals.

Among the three review papers, two of the articles address electro-discharge machining of non-conductive ceramics and a combination of materials written by Volosova et al. [1,2]. The critical issues and the way forward in each case is intelligently described by authors. In the third review paper written by TC Yap [3], the effect of cryogenic cooling on the turning of superalloys, ferrous materials and viscoelastic polymers is described along with the challenges faced by the industry in using cryogenic machining.

Eleven research papers written under this thematic issue address a spectrum of topics related to machining, additive manufacturing, oil processing, nanomaterials processing, metallic processing and polymers processing. The article written by Grigoriev et al. [4] described tool electrode behaviour and wear under discharge pulses during electrical discharge machining. Important insight is provided on electrical erosion wear fundamentals. Another paper written by Melnik et al. [5] presents the research on the parameters of vibroacoustic emission for the development of the monitoring and adaptive control system for electrical discharge machining.

Two of the research papers focussed on the secondary processing (joining) of metalbased materials. Tamadon et al. [6] investigated the influence of WC-based pin tool profile and processing parameters during friction stir welding on the microstructure and properties of AA1100 weld. In the second paper, Tamadon et al. [7] used Bobbin Friction Stir Welding

**Citation:** Gupta, M. Editorial for the Special Issue "Reviews and Advances in Materials Processing". *Technologies* **2022**, *10*, 77. https://doi.org/10.3390/ technologies10040077

Received: 14 June 2022 Accepted: 23 June 2022 Published: 24 June 2022

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

**Copyright:** © 2022 by the author. 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/).

and investigated the micro-flow patterns within aluminium weld structure. They used Atomic Force Microscopy as the main tool for characterization.

In the area of processing of non-metallic materials and liquid to be specific, Mazumder et al. [8] presented desulfurization technologies in the context of fuel and fuel cells. They focussed on the efficacy of different types of hydrotalcite absorbents.

Two papers in this issue address the advances in polymer processing. T. Aizawa [9] highlighted the CO<sup>2</sup> assisted polymer compression method to prepare porous polymer materials. These authors provided interesting insight in process development. The paper written by Terekhina et al. [10] presented their work on polymer processing using fused filament fabrication method. Different processing variations were attempted and strength was characterized to understand the effects of these variations. Hartmann et al. [11] highlighted the importance of the design of the additive manufacturing process in the context of industrial adaptability. In their article, they proposed a data-driven geometrical compensation approach.

The paper written by Vasiliev et al. [12] reported a set of computational algorithms to optimize the design and the optical or magneto-optical spectral analysis of 1D magnetic photonic crystals together with a windows software implementation. They also reported the methods and algorithms to obtain absorption coefficient spectral dispersion datasets for new materials.

Related to nanomaterials, two articles are included in the present issue. The article by Liem et al. [13] highlighted the use of eco-friendly microwave synthesis of silver nanoparticles using Mulberry leaves extract and silver nitrate solution. In the second article, Chen et al. [14] presented their work on facile fabrication of macroscopic selfstanding Ni or Co-doped MnO<sup>2</sup> architectures targeted for enhancing catalytic activities for propane oxidation.

In a nutshell, the papers included in this issue address various processing challenges faced by researchers and industry for liquid, metallic, polymers, ceramics and nanomaterials processing (both primary and secondary). The methods described and the solution proposed are of current relevance and of significant importance to move forward for better efficiency and productivity. It is certain that readers will gain much useful knowledge from these articles.

To conclude, I would like to thank all the contributors to this issue for publishing their immensely useful work with '*Technologies*'. I also thank reviewers for critically going through these papers and making useful comments. Their selfless contribution is much appreciated. I would also like to thank the administrative staff for processing these articles at their earliest and providing much needed support to authors, reviewers and editors.

**Funding:** This research received no external funding.

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

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The author declares no conflict of interest.

#### **References**


## *Review* **Electrical Discharge Machining Non-Conductive Ceramics: Combination of Materials**

**Marina A. Volosova 1 , Anna A. Okunkova 1, \* , Sergey V. Fedorov 1 , Khaled Hamdy 1,2 and Mariya A. Mikhailova 1**

	- Minia 61519, Egypt

Received: 24 April 2020; Accepted: 26 May 2020; Published: 28 May 2020

**Abstract:** One of the promising processing methods for non-conductive structural and functional ceramics based on ZrO2, Al2O3, and Si3N<sup>4</sup> systems is electrical discharge machining with the assistance of an auxiliary electrode that can be presented in the form of conductive films with a thickness up to 4–10 µm or nanoparticles - granules, tubes, platelets, multidimensional particles added in the working zone as a free poured powder the proper concentration of which can be provided by ultrasound emission or by dielectric flows or as conductive additives in the structure of nanocomposites. However, the described experimental approaches did not reach the production market and industry. It is related mostly to the chaotic development of the knowledge and non-systematized data in the field when researchers often cannot ground their choice of the material for auxiliary electrodes, assisting powders, or nano additives or they cannot explain the nature of processes that were observed in the working tank during experiments when their results are not correlated to the measured specific electrical conductivity of the electrodes, particles, ceramic workpieces or nanocomposites but depends on something else. The proposed review includes data on the main electrophysical and chemical properties of the components in the presence of heat when the temperature in the interelectrode gap reaches 10,000 ◦C, and the systematization of data on ceramic pressing methods, including spark plasma sintering, the chemical reactions that occur in the interelectrode gap during sublimation of primary (brass and copper) and auxiliary electrodes made of transition metals Ti, Cr, Co, and carbon, auxiliary electrodes made of metals with low melting point Zn, Ag, Au, Al, assisting powder of oxide ceramics TiO2, CeO2, SnO2, ITO, conductive additives Cu, W, TiC, WC, and components of Al2O<sup>3</sup> and Zr2O workpieces in interaction with the dielectric fluid - water and oil/kerosene medium.

**Keywords:** structural ceramic; oxide ceramic; EDM; ZrO2; Al2O3; electrode; thin films; electrical conductivity; white layer; electro physics; chemical reactions; sublimation

#### **1. Introduction**

One of the widespread machining methods is electrical discharge machining (EDM) that allows producing parts regardless of physical and mechanical properties. The only usability condition for EDM is the electrical conductivity of the workpiece material: the higher specific electrical conductivity provides the higher material removal rate [1–3].

In addition, it is one of most precise machining methods as the diameter of the electrode-tool on EDM equipment can vary from 0.6 up to 0.05 mm. The accuracy of electrode positioning on the industrial equipment under certain conditions can reach 1–2 µm (80–100 nm in some cases) when the removal.

achievable roughness *Ra* is about 0.4 µm after roughing. It also allows removing significant volumes of material by a single pass of a wire tool. That can be a considerable advantage in the conditions of the tool industry in comparison with other machining methods based on the mechanical nature of removal. The process of electrical erosion of the material consists of following stages: the initiation of an electric pulse in the interelectrode gap, the breakdown of the dielectric medium by a series of the discharge of pulses,

movements of other axes, expanding the available geometry to six axes (Figure 1).

*Technologies* **2020**, *8*, x FOR PEER REVIEW 2 of 26

industrial equipment under certain conditions can reach 1–2 µm (80–100 nm in some cases) when the achievable roughness *Ra* is about 0.4 µm after roughing. It also allows removing significant volumes of material by a single pass of a wire tool. That can be a considerable advantage in the conditions of the tool industry in comparison with other machining methods based on the mechanical nature of

The electrical discharge machining allows for obtaining complex spatial geometry with a slope of the generatrix up to 30° relative to the Z-axis and making holes with an arcuate or spiral-shaped generatrix on die-sink equipment. The most modern machines are equipped with computer numerical control with additional rotary mechanisms for the workpiece fixture that allows moving

The electrical discharge machining allows for obtaining complex spatial geometry with a slope of the generatrix up to 30◦ relative to the Z-axis and making holes with an arcuate or spiral-shaped generatrix on die-sink equipment. The most modern machines are equipped with computer numerical control with additional rotary mechanisms for the workpiece fixture that allows moving both rotationally and indexally according to the developed program in full coordination with movements of other axes, expanding the available geometry to six axes (Figure 1). the formation of a discharge channel with a temperature of more than 10,000 °C of a cloud of low-temperature plasma, the sublimation of material from the surface of the electrodes, the interruption of the pulses and washing off the cooled-down erosion products out of the interelectrode gap by dielectric medium flows, the restoration of dielectric tightness of the medium.

**Figure 1.** Additional rotary axes of wire electrical discharge machining: (**1**) is a tool, (**2**) is a part, *X0Y0Z<sup>0</sup>* is an absolute coordinate system of the machine, *XYZ* is a relative coordinate system of the lower wire **Figure 1.** Additional rotary axes of wire electrical discharge machining: (**1**) is a tool, (**2**) is a part, *X0Y0Z<sup>0</sup>* is an absolute coordinate system of the machine, *XYZ* is a relative coordinate system of the lower wire attachment point, *UVZ'* is an auxiliary coordinate system of the upper wire attachment point, *X"Y"Z"* is a relative coordinate system of the part, *ABC* is an additional rotary axes.

attachment point, *UVZ'* is an auxiliary coordinate system of the upper wire attachment point, *X"Y"Z"*

is a relative coordinate system of the part, *ABC* is an additional rotary axes. The process of electrical erosion of the material consists of following stages:


For example, during processing steels in water, a subsurface layer of electrodes that often saturated with carbon and oxides is formed [8–10]. This layer is known as a recast or white layer. Due to the high temperatures that are fulminantly reached at the time of the formation of the discharge channel, the electrode material bypasses melting to the liquid phase. It sublimates directly from the solid phase to the condition of low-temperature gas plasma [4–7]. The ions of electrodes and fluid components interact with each other during subsequent pulses interruption and form sediment in the form of gases and droplets or dust that are washed out from the interelectrode gap by dielectric medium, which is, in most cases, water, oil, kerosene. The material of the electrodes and medium in the form of solid substances are partially adsorbed by the eroded surface of the electrodes that usually have a structure of nano-frames of the more resistant to high temperatures components.

For example, during processing steels in water, a subsurface layer of electrodes that often saturated with carbon and oxides is formed [8–10]. This layer is known as a recast or white layer. During machining stainless steel 12Kh18N10T (AISI 321), the nickel of the workpiece reacts with zinc of the brass electrode with the formation of the intermetallide accompanies processing by sparks and a black cloud of erosion dust [11,12]. At the same time, a natural oxide film is formed during processing fusible aluminum, copper, chrome, titanium, and their alloys [13–18]. When attempts to machine aluminum contained materials in oil or kerosene, the aluminum interacts with carbon from the organic medium in the presence of high temperatures when the oxide film is sublimated, with the formation of extremely dangerous aluminum carbide Al4C<sup>3</sup> that may damage the filtration system of the equipment [19–22]. The experiments and chemical analyses of the samples showed that the components of the wire tool penetrate to a depth of 4–10 µm (up to 35 µm). A thermal influence zone is observed at the same depth and characterized by the formation of an austenitic layer in steels [4,10,11,18]. The erosion products depose as well on the walls of the working tank and in filters. After electrical erosion, a crater-like surface topology on the surface of the workpiece is observed [23–26]. of the brass electrode with the formation of the intermetallide accompanies processing by sparks and a black cloud of erosion dust [11,12]. At the same time, a natural oxide film is formed during processing fusible aluminum, copper, chrome, titanium, and their alloys [13–18]. When attempts to machine aluminum contained materials in oil or kerosene, the aluminum interacts with carbon from the organic medium in the presence of high temperatures when the oxide film is sublimated, with the formation of extremely dangerous aluminum carbide Al4C3 that may damage the filtration system of the equipment [19–22]. The experiments and chemical analyses of the samples showed that the components of the wire tool penetrate to a depth of 4–10 µm (up to 35 µm). A thermal influence zone is observed at the same depth and characterized by the formation of an austenitic layer in steels [4,10,11,18]. The erosion products depose as well on the walls of the working tank and in filters. After electrical erosion, a crater-like surface topology on the surface of the workpiece is observed [23–26]. The dielectric properties of ceramics and nanoceramics, in combination with excellent physical

*Technologies* **2020**, *8*, x FOR PEER REVIEW 3 of 26

During machining stainless steel 12Kh18N10T (AISI 321), the nickel of the workpiece reacts with zinc

The dielectric properties of ceramics and nanoceramics, in combination with excellent physical and mechanical properties, limit their machinability - material removal volumes per tool pass, the complexity of the resulting geometry, and, accordingly, the scope of their potential application [27–30]. It restrains the growth of the economy and prevents the transition to the sixth technological paradigm associated with the concept of "nano", the development of innovative materials and technologies, and its introduction into the industry [31–36]. and mechanical properties, limit their machinability - material removal volumes per tool pass, the complexity of the resulting geometry, and, accordingly, the scope of their potential application [27– 30]. It restrains the growth of the economy and prevents the transition to the sixth technological paradigm associated with the concept of "nano", the development of innovative materials and technologies, and its introduction into the industry [31–36]. Today, there are three fundamental approaches to solving this problem (Figure 2). One of them

Today, there are three fundamental approaches to solving this problem (Figure 2). One of them is related to the introduction of a conducting secondary phase to the ceramic matrix of a composite or nanocomposite. It involves the use of powder metallurgy methods, including innovative hybrid sintering using assisting currents in graphite dies. The powder metallurgy methods of production conductive ceramics are known as industrial methods for creating composites from the mid-80s and continuously improves [37]. The idea of creating an electrically conductive ceramic nanocomposite has been developing since the late 80s [38], and many scientists have achieved remarkable results [21,22]. Still, more research is needed before the industrial introduction of this approach, since ceramics, when modified by conducting and non-conducting nanoparticles and platelets, do not always retain their original physical and mechanical properties, whiteness or even transparency, and after modification, even with confirmed improved properties, the material cannot always be electroerosively processed, even with proven conductivity much higher than the percolation threshold. is related to the introduction of a conducting secondary phase to the ceramic matrix of a composite or nanocomposite. It involves the use of powder metallurgy methods, including innovative hybrid sintering using assisting currents in graphite dies. The powder metallurgy methods of production conductive ceramics are known as industrial methods for creating composites from the mid−80 s and continuously improves [37]. The idea of creating an electrically conductive ceramic nanocomposite has been developing since the late 80 s [38], and many scientists have achieved remarkable results [21,22]. Still, more research is needed before the industrial introduction of this approach, since ceramics, when modified by conducting and non-conducting nanoparticles and platelets, do not always retain their original physical and mechanical properties, whiteness or even transparency, and after modification, even with confirmed improved properties, the material cannot always be electroerosively processed, even with proven conductivity much higher than the percolation threshold.

**Figure 2.** Electroerosive methods for processing non-conductive ceramics. **Figure 2.** Electroerosive methods for processing non-conductive ceramics.

The second method for solving the problem of electrical discharge machining ceramics with high values of dielectric impermeability is the approach associated with the introduction of an auxiliary electrode in the interelectrode gap. The auxiliary electrode can be introduced in the form of a thin film produced of a conductive material (aluminum, copper, silver, etc.) on the surface of the dielectric workpiece, to which electric current discharges will be initiated. It is assumed that the erosion of the surface of the ceramic workpiece will be carried out simultaneously with the electrically conductive The second method for solving the problem of electrical discharge machining ceramics with high values of dielectric impermeability is the approach associated with the introduction of an auxiliary electrode in the interelectrode gap. The auxiliary electrode can be introduced in the form of a thin film produced of a conductive material (aluminum, copper, silver, etc.) on the surface of the dielectric workpiece, to which electric current discharges will be initiated. It is assumed that the erosion of the surface of the ceramic workpiece will be carried out simultaneously with the electrically conductive film. For the first time, the idea of introducing the auxiliary electrode was presented in the schematic

diagram of electrical discharge machining dielectrics by soviet scientists [39]. Later similar experiments were carried out by many other scientists, and serious results were obtained - the ceramics were processed electroerosively to a depth of 1.5–2.0 mm [40]. In the described case, after the first operational iterations, when the film-electrode sublimates entirely, the question of reinitiation of current pulses in the interelectrode gap remains open.

The third method can be called an approach in which current discharges are directed to electrically conductive particles uniformly weighted in the volume of a liquid dielectric (water, oil, kerosene) in the interelectrode gap. In this case, it is necessary to monitor the constant concentration of conductive particles in the discharge gap. The first experiments to improve the dielectric constant of the working medium have been carried out since the creation of the method of electrical discharge machining materials by Professors Boris and Natalya Lazarenko [41–45]. The most outstanding works on changing the components of the working fluid during conductive materials processing appeared from the beginning of the 2000s [46]. It was related to the possibility of obtaining a finely dispersed substance and the ability to control the morphology of the used powders. The most relevant studies were obtained using carbon nanotubes and other forms of graphene in the discharge gap [4,47].

The direction of research related to the development of innovative methods of processing super-hard, heat- and wear-resistant materials without losing their physical and mechanical properties, with the possibility of obtaining complex spatial geometry and with minimal losses on material, on operating and auxiliary time, with minimum possible tool consumption, while maintaining high volumes of material removal by a pass of tool, is an extremely relevant task of modern science, as evidenced by multiple publications.

The novelty of the work is in the systematization of ceramic pressing methods, electrophysical and chemical properties of electrodes' and dielectrics' components in the presence of heat, data on possible interactions between components, suitable combinations of the materials for processing structural oxide ceramics as alumina and zirconia that can be applied for functional ceramics and nanocomposites.

This analytical study is designed to answer the question of why some materials should not be combined, why certain phenomena are observed during processing, why the workability of the workpiece is often not dependent on the electrical conductivity and temperature stability of the materials of the primary and auxiliary electrodes, assisting powder and conductive additives.

The tasks of the study include the analytical research on the conductivity of materials in the presence of heat and chemical composition of the subsurface layer of some materials, the investigation of physical properties of structural ceramics in the presence of heat and pressing methods, the sintering cycles under pressure and features of spark plasma sintering, main chemical interaction of alumina and zirconia components with materials of electrodes, assisting powder, additives, and working medium.

The analytical research was conducted, taking into account the basic principles of electrophysics and physical chemistry, the laws of the structures of substances, their structure and transformations, and available theoretical and practical information.

#### **2. Conductivity of Materials and Saturation of Components in Recast Layer**

All conductive materials are subject to electrical erosion to one degree or another. The group of materials to be electroerosively machined include aluminum, chromium, cobalt, copper, beryllium, molybdenum, nickel, niobium, titanium, tungsten, and their alloys including bronze and brass, high-temperature and magnetic cast alloys, cast iron and steels, graphite, hard alloys (carbide ceramics), titanium nitride and boride, conductive ceramic composites and nanocomposites, and some other materials (Figure 3). However, not only the specific electrical conductivity determines the possibility of using electrical discharge machining methods to process all practically conductive materials. The resistance to electrical erosion of metals is determined as well by a combination of thermophysical properties of materials such as a melting or sublimation point, heat capacity, density, structural and microstructural homogeneity, and thermal conductivity [34,48].

ς ( ). **Figure 3.** Specific electrical conductivity (ς) of some materials. (\* The electrical conductivity depends on impurities and suspensions, for distilled water is about 10−<sup>5</sup> S·m−<sup>1</sup> ).

Some dielectric materials change their electrical conductivity when heated and become conductors or semiconductors when metals have another tendency to the opposite phenomenon. It can have a positive effect on the electrical discharge machining a part whose chemical composition contains non-conductive at room temperature components. The conductivity is also affected by the presence of impurities in the material composition. In this case, conductive components increase the specific conductivity of the material, and non-conductive ones decrease. As can be seen, for materials such as carbon and silicon dioxide, the electrical conductivity often depends as well on the structure of the crystal lattice. Carbon the form of graphite calls a conductor when the carbon in the form of a diamond is an insulator, despite its anisotropic conductivity associated as well with the structure of the crystal lattice. Another example, the silicon dioxide in the crystalline form (silica glass) is a better dielectric than in the amorphous form of domestic glass. The same trend is observed for titanium dioxide.

One of the promising areas of research is the introduction into the interelectrode gap of a powder or nanosized powder. It can be some conductive metal or titanium dioxide, cerium dioxide, other ceramics in the powder phase [49,50]. In this case, the ceramics in the presence of heat would acquire conductive properties and provoke more dense discharges of pulses in processing main conductive material—usually metals and metal alloys such as extremely popular nickel alloys, construction and tool steels. The experiments showed that this approach gave a higher performance and improved the quality of machined surfaces related to the even topology.

Typically, the electroerosively machined surfaces have a matte appearance, similar to the surface fired by multiple shots. It contains disordered distributed tiny craters as a result of the material removal by separate discharges. The surface structure with the presence of craters and tubercles can be a favorable indicator for the retention of lubricants in the mechanisms and the formation of working cavities of injection molds. This type of surface is called "shagreen leather" [46]. It is easily distinguishable on any molded plastic products, mainly on computer and television case products, but also in the products for the transportation and sale of foods and drinks, toys for outdoor games.

There remains a layer of adsorbed substance, consisting mainly of components of the electrode-part and sometimes of components sublimated from the electrode-tool and decay products of the dielectric medium on the electroerosively machined surfaces. Mostly carbides are formed during processing in the organic dielectric as oil or kerosene and oxides during processing in deionized water [13,14,19,26]. The carbides are formed as well from the composition of steel during processing steels, as carbon is one of the most refractory materials. It counts as well for other alloys and complex materials when refractory components form the recast or white layer [11,14,46,49–58]. In the case of processing ceramic composites and nanocomposites, the surface layer is saturated with low-melting components in a nano-frame of refractory components [4,29,52]. It was observed that this recast or white layer is more hard and brittle than the main material of the workpiece. The subsurface layer can be easily distinguishable on cross-sections of steels, nanocomposites based on oxide ceramics, other materials and alloys using scanning electron microscopy (Figure 4). In the case of machining steels, carbon saturated layer has a thickness of 2.5–50 µm. It does not exceed 4–10 µm for pure 99% chromium. The recast layer is considered favorable for shaping the cavities of injection molds. In cases where high fatigue strength of the product is required, the white layer can be removed mechanically or electrochemically [51]. part and sometimes of components sublimated from the electrode-tool and decay products of the dielectric medium on the electroerosively machined surfaces. Mostly carbides are formed during processing in the organic dielectric as oil or kerosene and oxides during processing in deionized water [13,14,19,26]. The carbides are formed as well from the composition of steel during processing steels, as carbon is one of the most refractory materials. It counts as well for other alloys and complex materials when refractory components form the recast or white layer [11,14,46,49–58]. In the case of processing ceramic composites and nanocomposites, the surface layer is saturated with low-melting components in a nano-frame of refractory components [4,29,52]. It was observed that this recast or white layer is more hard and brittle than the main material of the workpiece. The subsurface layer can be easily distinguishable on cross-sections of steels, nanocomposites based on oxide ceramics, other materials and alloys using scanning electron microscopy (Figure 4). In the case of machining steels, carbon saturated layer has a thickness of 2.5–50 µm. It does not exceed 4–10 µm for pure 99% chromium. The recast layer is considered favorable for shaping the cavities of injection molds. In cases where high fatigue strength of the product is required, the white layer can be removed mechanically or electrochemically [51].

*Technologies* **2020**, *8*, x FOR PEER REVIEW 6 of 26

distinguishable on any molded plastic products, mainly on computer and television case products,

There remains a layer of adsorbed substance, consisting mainly of components of the electrode-

**Figure 4.** White layer obtained by electric discharge machining 99.9% pure chromium by a brass-wire electrode in a water medium: (**1**) is an epoxy resin that contains a sample for SEM analysis, (**2**) is a white layer of thermal exposure, a saturated component wire electrode—copper and zinc, (**3**) is a cut of the machined part made of chromium. **Figure 4.** White layer obtained by electric discharge machining 99.9% pure chromium by a brass-wire electrode in a water medium: (**1**) is an epoxy resin that contains a sample for SEM analysis, (**2**) is a white layer of thermal exposure, a saturated component wire electrode—copper and zinc, (**3**) is a cut of the machined part made of chromium.

Obviously, the white layer and the erosion products are formed from the components of ceramics, dielectric medium, and auxiliary electrode during processing ceramics, the components of the primary phase and conductive additives of ceramic composites and nanocomposites also interact with other components in processing. Obviously, the white layer and the erosion products are formed from the components of ceramics, dielectric medium, and auxiliary electrode during processing ceramics, the components of the primary phase and conductive additives of ceramic composites and nanocomposites also interact with other components in processing.

In some cases, the formed erosion products may have properties incompatible with the concept of safety-toxic or explosive gas or reactive sediment. Therefore, it is necessary to do carefully analyze the composition of all used main and auxiliary materials in the interelectrode gap, especially before designing the processing conditions for new materials. In some cases, the formed erosion products may have properties incompatible with the concept of safety-toxic or explosive gas or reactive sediment. Therefore, it is necessary to do carefully analyze the composition of all used main and auxiliary materials in the interelectrode gap, especially before designing the processing conditions for new materials.

The ceramics are non-metallic and non-organic materials based on crystal compounds of nonmetals and metals synthesized and consolidated by various methods to impart specific properties and geometry. At the same time, ceramics is a group of phase-sensitive materials—the properties of ceramics with a similar chemical compound may significantly differ depending on precursors,

**3. Properties of Structural Ceramics in the Presence of Heat and Pressing Methods** 

#### **3. Properties of Structural Ceramics in the Presence of Heat and Pressing Methods**

The ceramics are non-metallic and non-organic materials based on crystal compounds of non-metals and metals synthesized and consolidated by various methods to impart specific properties and geometry. At the same time, ceramics is a group of phase-sensitive materials—the properties of ceramics with a similar chemical compound may significantly differ depending on precursors, methods and technology parameters, structure, and phase composition of sintered materials, intermediate processing methods, surface quality, properties estimation methods, exploitation conditions.

The hard-to-melt compounds, which form the basis of structural ceramics, are characterized by low operational ability; in other words, they are difficult in forming blanks, sintering to high-density values, and processing to the required dimensions.

All used ceramics can be divided into structural and functional types. In this context, the structural ceramics provide the integrity and load-bearing capacity of the designs of various products, when physical and mechanical properties are of primary importance, when functional ceramics are aimed at specialized products and devices.

Although WC-Co system cermets still dominate the market of technical ceramics, some oxide ceramics can become a proper alternative in many production areas in the coming years [59–61]. Ceramic cutting tools based on oxide ceramics have gained popularity due to their ability to work in extreme conditions, when high wear resistance, hardness, and low chemical activity are required under operating conditions [62,63]. However, the scope of oxide ceramics as a structural material is still limited due to the low fracture toughness [64–66]. Cracks easily spread and often lead to uncontrolled fracture of specimens [67,68].

There are various methods for improving the mechanical properties of oxide ceramic parts. For example, as was mentioned, the ability to reinforce Al2O3, ZrO<sup>2</sup> by adding secondary phases made of carbide or nitride ceramics, metals, intermetallic compounds, or other materials to the ceramic matrix of composites or nanocomposites. The secondary phase in the form of powders and tubes, nano or micro platelets and fibers is helpful to increase strength and fracture toughness of ceramic matrix composites [69,70]. These methods have proven to be effective as experimental and innovative for increasing the properties of oxide ceramics as reinforced zirconia with Ti[C,N] or CeO<sup>2</sup> particles or reinforced alumina with multidimensional ceramic particles of SiC, WC, TiC, TiN [27,28,59,60,71–73].

Silicon nitride is also one of the most promising materials for critical applications because of its high strength, thermal conductivity, low coefficient of thermal expansion, high resistance to thermal shock, which are superior to oxide ceramics. Silicon nitride and Silicon alumina nitride (SiAlON) can be called the most suitable ceramics for machining superalloys. At the same time, silicon nitride Si3N<sup>4</sup> has low hardness and wear resistance [74,75].

There is a few ceramics compaction technique, sometimes combined with sintering that influences the physical and mechanical properties of blanks [76,77]. Cold static pressing is a unique and straightforward technique for pressing the products of different shapes from any compounds that allow realizing it in the conditions of mass production [78]. However, it does not allow distributing the properties of the product in the volume evenly.

The hot-pressing method solves the problems of the cold pressing method partly since it provides better strength characteristics of the products, the minimum tolerances on the size of the workpieces, reduces of sintering time due to the combination of the pressing and sintering processes. The disadvantage of this method is the rapid wear of the molds and the low productivity of the process [79,80]. Applying a hot-pressing method improves the strength and macrostructural characteristics of the workpieces. However, ferroelectric and ferromagnetic properties deteriorate. That can be a disadvantage for electroceramics but an advantage for structural ceramics. Significant grain growth in zirconia ceramics made by hot-pressing can be observed. In addition, ceramic ferroelectrics obtained by hot-pressing, in comparison with conventional annealing, deteriorate performance due to the presence of residual stresses and violation of stoichiometry of the molar masses of the components.

The hot-pressing method allows obtaining products with a more uniform distribution of density throughout the volume and effective in the production of some types of nanoceramics with superplasticity at elevated temperatures, which can significantly reduce the cost of products due to the elimination of machining. However, the use of the hot-pressing method is limited by high demands on the material of the mold, which must be inert concerning the pressed powders at elevated temperatures, heat-resistant, not having the property of superplasticity, economical.

Hot isostatic pressing is one of the varieties of the method that implemented using various techniques based on the use of elastic media—liquids, gases, polyurethane, etc. by a disposable mold, which avoids the friction forces of the powder granules on the stationary parts of the mold, which is indispensable for the conditions of mass production [81,82].

Other pressing methods are usually aimed at eliminating the drawbacks of previous methods, for example, dynamic pressing methods as isothermal stamping, magnetic pulse (with amplitude up to 5 GPa, when the pressure of conventional pressing methods does not exceed 1 GPa), explosive, hydrodynamic, impact or shock (up to 35 GPa), shear compaction, electrical consolidation, various ultrasound vibration moldings (in the range of 16–22 kHz), pressureless packing [83–89]. They are aimed at improving the density of the workpieces and more uniform density distribution, relieving stresses, preventing grain growth during subsequent sintering. Their use is always limited by the maximum achievable dimensions of the workpiece, the need to use methods of preliminary destruction of agglomerates, and the removal of sorbed gases.

In addition, when choosing ceramics, designers are often guided not only by their physical and mechanical properties but also by their appearance, which is extremely important for some applications and depends not only on the composition of ceramics but also on the methods of their pressing and sintering [77]. For example, in the production of samples by hot-pressing, preforms are predominantly black-gray, while white is predominant in ceramics obtained using cold pressing methods. That correlates to the fundamentals of physical chemistry when other chemical transformations occur in the presence of heat [4,14,29].

The relationship of the basic physical and mechanical properties of some materials to develop requirements for the material is shown in Figure 5. It should be noted that in the presented figure, the range of parameter values for Al2O<sup>3</sup> and ZrO<sup>2</sup> refers to samples obtained by isostatic pressing from powder, and Elbor refers to any trademarks of cubic boron nitride (c-BN).

It should be noted that oxide ceramics as alumina Al2O<sup>3</sup> and zirconia ZrO<sup>2</sup> among the mentioned above properties including low electrical conductivity at room temperatures show other highly relevant for some application properties as chemical inertness even at elevated temperatures or in the presence of an oxidizing agent, hot hardness (or redness, high hardness, and wear resistance to temperatures of red heat), low coefficient of thermal conductivity, and low coefficient of thermal expansion in comparison with more widespread construction materials (Table 1).

As can be seen, despite the quite high thermal conductivity and hardness and low thermal expansion coefficient, alumina Al2O<sup>3</sup> has lower flexural strength and crack growth resistance in comparison with zirconia ZrO2.

It should be noted that electrical conductivity of ceramics that contain vitreous (glassy) phase (amorphous) increases significantly with an increase in temperatures since the concentration and mobility of charge carriers are exponentially dependent on temperature (*t*):

ς = ς0*e* β*t* , (1)

χ = χ<sup>0</sup> *e* −β*t* , (2)

where ς<sup>0</sup> and χ<sup>0</sup> are values of specific electrical conductivity and volume resistivity at 0 ◦C, β is the temperature coefficient.

*Technologies* **2020**, *8*, x FOR PEER REVIEW 8 of 26

indispensable for the conditions of mass production [81,82].

destruction of agglomerates, and the removal of sorbed gases.

transformations occur in the presence of heat [4,14,29].

Hot isostatic pressing is one of the varieties of the method that implemented using various techniques based on the use of elastic media—liquids, gases, polyurethane, etc. by a disposable mold, which avoids the friction forces of the powder granules on the stationary parts of the mold, which is

Other pressing methods are usually aimed at eliminating the drawbacks of previous methods, for example, dynamic pressing methods as isothermal stamping, magnetic pulse (with amplitude up to 5 GPa, when the pressure of conventional pressing methods does not exceed 1 GPa), explosive, hydrodynamic, impact or shock (up to 35 GPa), shear compaction, electrical consolidation, various ultrasound vibration moldings (in the range of 16-22 kHz), pressureless packing [83–89]. They are aimed at improving the density of the workpieces and more uniform density distribution, relieving stresses, preventing grain growth during subsequent sintering. Their use is always limited by the maximum achievable dimensions of the workpiece, the need to use methods of preliminary

In addition, when choosing ceramics, designers are often guided not only by their physical and mechanical properties but also by their appearance, which is extremely important for some applications and depends not only on the composition of ceramics but also on the methods of their pressing and sintering [77]. For example, in the production of samples by hot-pressing, preforms are predominantly black-gray, while white is predominant in ceramics obtained using cold pressing methods. That correlates to the fundamentals of physical chemistry when other chemical

The relationship of the basic physical and mechanical properties of some materials to develop requirements for the material is shown in Figure 5. It should be noted that in the presented figure,

**Figure 5. Figure 5.** The relationship of the basic physical and mechanical parameters for some materials. The relationship of the basic physical and mechanical parameters for some materials.




**Table 1.** *Cont.*

\* Given for reference [98,99].

The electrical conductivity of crystalline ceramics (for titanium(IV) dioxide TiO2, it matches the form of anatase-metastable mineral), changes slower and retain quite low values at very high temperatures.

It should be noted that the porous ceramics significantly increases specific electrical conductivity in the presence of moisture, even in the smallest quantities [95–97].

The study devoted to the evaluation of ceramics workability by electrical discharge machining with an auxiliary electrode [100] describes the properties of non-conductive ceramics as ZrO2, Si3N4, and SiC subjected to electrical discharge machining. The Vickers hardness (*HV*0.3), surface roughness (*Sq*), and flexural strength of samples were compared before and after machining since the use of the methods based on electrical erosion of the materials traditionally causes a decrease in these parameters due to the nature of the material destruction and the zone of thermal influence (or heat-affected zone, HAZ). It was found that the effect shows changes three times higher than the parameters before processing for ZrO<sup>2</sup> samples, and twice for in Si3N4. It can correlate with specific electrical conductivity in the presence of high temperatures (ς 1000◦ ), thermal expansion coefficient (α), and chemical transactions of components that occurred in the presence of heat.

#### **4. Features of Structural Ceramics Sintering**

The ceramics in the form of billets obtained by powder pressing and sintering are usually processed by grinding using abrasive materials or solid tools made of technical diamond or c-BN to give them a more functional shape depending on their applications [6,101–105]. Processing ceramics of complex shapes using traditional processing methods is a complicated and time-consuming process due to their exceptional physical and mechanical properties.

The technological process of powder metallurgy for ceramic sintering consists of four main stages: production of powders, their mixing and homogenizing, pressing or molding, and subsequent sintering [106–112]. In all similar conditions, the method of powder compaction of billets often determines the properties of the final product and plays a decisive role in the formation of the internal structure as uniformity, absence of pores, and cracks [113–115]. Forming billets is often carried out by cold pressing under high working pressure up to 1000 MPa in metal dies of a press that is oriented vertically. A mixture of powders is free-poured into the cavity of the dies between the upper and lower punches or several punches in the case of production with several transitions [116]. The stroke of the lower punch controls the dosage volume. The formed workpiece is pushed out of the cavity of the lower punch. For molding, specialized press equipment with mechanical, hydraulic, or pneumatic drives is used. The resulting billet has the dimensions and shape of the final product and the strength is sufficient for unloading and transporting the pre-form into the furnace for subsequent sintering. The workpiece is sintered to increase its strength and ductility. For the production of high-precision parts, sintering is carried out in furnaces with a reducing (hydrogen, dissociated ammonia) or protective (nitrogen, argon, endothermic gas) atmosphere. Forming is also often carried out using the molding method in combination with sintering by electric current pulses [117].

Traditional sintering methods based on an isostatic process form a ceramic preform in the presence of high temperatures (up to 1100 ◦C). The sintering process can take 1–5 h to obtain a denser structure of the workpiece. At the same time, it is believed that high temperature and prolonged sintering time lead to undesirable grain growth, which may adversely affect the operational properties of the material [38,118,119]. Some chemical transactions can occur with impurities in the ceramic powder during sintering. It leads to the formation of a precipitate that has properties exceeded main material properties and gas evolution. Thus, excessive porosity can be obtained at high temperatures of sintering, which obviously also reduces the physical and mechanical properties of the ceramics.

Spark plasma sintering (SPS) is often used as an alternative method of consolidating powder billets to reduce sintering temperatures. The use of this technology in combination with the use of additional currents facilitates, the rapid consolidation of the powder, makes it possible to produce sufficiently dense ceramic composites and nanocomposites, and prevents grain growth [120–122]. The technology also offers significant advantages, such as a faster and shorter sintering time compared to conventional methods of forming powder billets, which is primarily associated with the complex effect of mechanical pressure on the billet and electrical impulses during processing [123–125].

Spark plasma sintering is an advanced technology with high potential for material processing. Its principles are similar to hot-pressing and differ in the source of heat. SPS uses pulsed direct current (DC), which is passed through the dies and acts in such a way as a heat source that corresponds to the Joule effect. It is used to minimize grain growth and obtain a dense nanostructured material.

A machine for spark plasma sintering consists of a press sintering machine with a vertical uniaxial pressurization mechanism, specially designed electrodes with adjustable water pumping, a cooler, a water-cooled vacuum chamber, a vacuum/air control mechanism, an argon-gas atmosphere control mechanism, a DC pulse generator, a water cooling control unit, a control unit for the position of the upper punch along the *Z*-axis, a temperature measuring unit (Figure 6).

During sintering, the compacted powder is compressed between the die and punches to which the electrodes are connected. After, the spark discharge is supplied, with the help of which high temperatures are obtained. It leads to thermal and electrolytic diffusion between the powder particles that depend on their turn on the electrical properties of the powder material in the presence of heat. Sintering occurs in the process at temperatures between 1000 and 2500 ◦C. The heat distribution and the temperature difference between the core and the edges of the sintered billet are determined by the thermal conductivity of the billet material, which, for example, is ten times lower for zirconia than for alumina (Table 1). Cycle time and sintering power consumption depend on the thermal conductivity of the material [126]. The sintering cycle is selected in such a way as to ensure minimal grain growth of the sintered billets, which should provide higher physical and mechanical properties.

*Technologies* **2020**, *8*, x FOR PEER REVIEW 11 of 26

of the upper punch along the *Z*-axis, a temperature measuring unit (Figure 6).

A machine for spark plasma sintering consists of a press sintering machine with a vertical uniaxial pressurization mechanism, specially designed electrodes with adjustable water pumping, a cooler, a water-cooled vacuum chamber, a vacuum/air control mechanism, an argon-gas atmosphere

**Figure 6.** The configuration of the spark plasma sintering system developed in MSTU STANKIN: (**1**) is an upper electrode, (**2**) is a vacuum chamber, (**3**) is an upper punch, (**4**) is pressed powder, (**5**) is a die block, (**6**) is a heater, (**7**) is a lower punch, (**8**) is a lower electrode. **Figure 6.** The configuration of the spark plasma sintering system developed in MSTU STANKIN: (**1**) is an upper electrode, (**2**) is a vacuum chamber, (**3**) is an upper punch, (**4**) is pressed powder, (**5**) is a die block, (**6**) is a heater, (**7**) is a lower punch, (**8**) is a lower electrode.

During sintering, the compacted powder is compressed between the die and punches to which the electrodes are connected. After, the spark discharge is supplied, with the help of which high temperatures are obtained. It leads to thermal and electrolytic diffusion between the powder particles that depend on their turn on the electrical properties of the powder material in the presence of heat. Conventional electric hot press processes using direct current or power are controlled by the basic process parameters that provide Joule sintering. These parameters or factors are related to the characteristics of the power supply, or high-frequency induction of the elements in combination with the smooth loading compacted powder materials through hydraulic or mechanical pressure.

Sintering occurs in the process at temperatures between 1000 and 2500 °C. The heat distribution and the temperature difference between the core and the edges of the sintered billet are determined by the thermal conductivity of the billet material, which, for example, is ten times lower for zirconia than for alumina (Table 1). Cycle time and sintering power consumption depend on the thermal conductivity of the material [126]. The sintering cycle is selected in such a way as to ensure minimal grain growth of the sintered billets, which should provide higher physical and mechanical properties. The phenomena generated by spark plasma and pressure exclude the influence of gas adsorption, the formation of oxide films, and interactions with impurities and suspended particles that remain on the surface of the powder when a high-temperature field occurs. The action of the electromagnetic field is enhanced, and a high diffusion rate is provided by the rapid migration of ions. The local high-temperature state caused by impulse voltage is accompanied by vapors, solidification, and recrystallization.

Conventional electric hot press processes using direct current or power are controlled by the basic process parameters that provide Joule sintering. These parameters or factors are related to the characteristics of the power supply, or high-frequency induction of the elements in combination with the smooth loading compacted powder materials through hydraulic or mechanical pressure. The phenomena generated by spark plasma and pressure exclude the influence of gas SPS at high pressures is another version of the well-established method of spark sintering for shaping certain materials, such as superhard polycrystalline diamonds, cubic boron nitride, ceramic composites, nanocomposites, including those having refractory properties. Ceramics can also be synthesized into metastable phases or intermetallic alloys using the entire variety of SPS methods [127,128].

adsorption, the formation of oxide films, and interactions with impurities and suspended particles that remain on the surface of the powder when a high-temperature field occurs. The action of the electromagnetic field is enhanced, and a high diffusion rate is provided by the rapid migration of ions. The local high-temperature state caused by impulse voltage is accompanied by vapors, solidification, and recrystallization. Since the sintering process is accompanied by uniaxial compression, the applied force is always limited by the high-temperature tensile strength of a graphite press tool, which is around 100–150 MPa. High pressure prevents the growth of grains in billets. It should be noted that there is a dependence between the physical properties of the sintered workpiece and the particle size of the ceramic powder, which is also determined by the different dispersion of the composition of the powder [129].

SPS at high pressures is another version of the well-established method of spark sintering for shaping certain materials, such as superhard polycrystalline diamonds, cubic boron nitride, ceramic composites, nanocomposites, including those having refractory properties. Ceramics can also be powder [129].

[127,128].

SPS consists of the following stages (Figure 7). In essence, vacuum conditions are created inside the working chamber, the workpiece is warmed up and is subjected to pressure, while at the point of contact of the powder particles there arises a spark discharge with the formed zones of local overheating, with continued heating, the surface of the powder grains reaches melting points, isthmuses form between particles, the mold is cooled down together with the workpiece. SPS consists of the following stages (Figure 7). In essence, vacuum conditions are created inside the working chamber, the workpiece is warmed up and is subjected to pressure, while at the point of contact of the powder particles there arises a spark discharge with the formed zones of local overheating, with continued heating, the surface of the powder grains reaches melting points, isthmuses form between particles, the mold is cooled down together with the workpiece.

ceramic powder, which is also determined by the different dispersion of the composition of the

*Technologies* **2020**, *8*, x FOR PEER REVIEW 12 of 26

synthesized into metastable phases or intermetallic alloys using the entire variety of SPS methods

Since the sintering process is accompanied by uniaxial compression, the applied force is always limited by the high-temperature tensile strength of a graphite press tool, which is around 100–150 MPa. High pressure prevents the growth of grains in billets. It should be noted that there is a

**Figure 7.** The main stages of spark plasma sintering. **Figure 7.** The main stages of spark plasma sintering.

The mold is heated using a sufficiently low voltage of ~10 V and a current of up to 10 kA. The maximum achievable temperature in graphite molds is 2400 °C. The cooling rate reaches 150 °C/min but can be increased to 400 °C/min using additional gas cooling. The mold is heated using a sufficiently low voltage of ~10 V and a current of up to 10 kA. The maximum achievable temperature in graphite molds is 2400 ◦C. The cooling rate reaches 150 ◦C/min but can be increased to 400 ◦C/min using additional gas cooling.

The sintering temperature is usually controlled either inside the wall of the graphite dies using a thermocouple, or superficially using a pyrometer during sintering. It should be noted that there are always differences in temperatures between the measured value and the actual sintering temperature in the mold. It has been experimentally established that the measured temperature during sintering metals and ceramic composites can vary between 50–250 °C in comparison with the actually achievable temperature in the center of the mold. The temperature measured on the surface is always lower than the actual value in the center of the preform. The sintering temperature is usually controlled either inside the wall of the graphite dies using a thermocouple, or superficially using a pyrometer during sintering. It should be noted that there are always differences in temperatures between the measured value and the actual sintering temperature in the mold. It has been experimentally established that the measured temperature during sintering metals and ceramic composites can vary between 50–250 ◦C in comparison with the actually achievable temperature in the center of the mold. The temperature measured on the surface is always lower than the actual value in the center of the preform.

The heterogeneity of heating can lead to increased porosity of the samples, which reduces not only physical and mechanical properties but also thermal and electrical conductivity of the final product [4,5,38,126]. The nature of electrical phenomena during sintering depends on the electrical properties of the raw material. The heterogeneity of heating can lead to increased porosity of the samples, which reduces not only physical and mechanical properties but also thermal and electrical conductivity of the final product [4,5,38,126]. The nature of electrical phenomena during sintering depends on the electrical properties of the raw material.

#### **5. Chemical Interaction of Ceramics Components with Electrodes, Powders, and Medium**

Many materials change their electrical conductivity when heated and become semiconductors and conductors from insulators. A reverse transition is also possible for metals. Such properties during the design of the technology can have a positive effect on the electroerosive machinability of a part whose structure contains dielectric components that changes its conductivity in the presence of heat when the material surface forming the interelectrode gap reaches sublimation temperatures in the conditions of low-temperature plasma when forming the discharge channel to the auxiliary electrode.

The introduction of a powder or nanosized powder of titanium dioxide TiO2, cerium dioxide CeO2, and other ceramic components into the interelectrode gap is one of the most spread approaches in electrical discharge machining of ceramics [4,24,25]. When the discharge channel temperatures reach a certain level to provoke sublimation of the surface to be machined, some of the ceramics acquire conductive properties as it was mentioned above. That causes denser discharges of pulses in the interelectrode gap during processing the main conductive material, leading to higher performance and improved quality of the processed surfaces

Thus, the erosion products and the surface layer of the processed surfaces are formed directly from the components of ceramics, dielectric medium, auxiliary electrode, assisting powder, secondary phase of materials in the case of nanocomposites in the presence of heat, some of the possible chemical interactions for alumina Al2O<sup>3</sup> and zirconia ZrO<sup>2</sup> are provided in Table 2. It is necessary to carefully analyze the composition of all used primary and auxiliary materials before designing the processing conditions [130–137].


**Table 2.** The analytical chemical composition of erosion products and the subsurface layer of the processing surfaces during electrical erosion of some ceramics using a brass or copper electrode tool.


**Table 2.** *Cont.*


**Table 2.** *Cont.*

\* At 20 ◦C.

As can be seen in Table 2, titanium Ti and chromium Cr are most suited for electrical discharge machining of zirconia ZrO<sup>2</sup> in oil or kerosene from all considered transition metals as they provoke the formation of conductive particles in the interelectrode gap.

EDM of zirconia ZrO<sup>2</sup> in oil or kerosene and with assisting carbon particles or nanotubes can form conductive zirconium carbide ZrC by interacting with oil or kerosene decomposition components and elements of the auxiliary electrode or assisting powder. Then the erosion products can contribute to a denser erosion in the interelectrode gap.

Zirconia ZrO<sup>2</sup> tends to change its electrical properties in the presence of heat up to 2000 ◦C (Figure 3), but it is inert to the water medium. The zirconium starts actively to absorb hydrogen H<sup>2</sup> at temperatures around 250–300 ◦C forming solid solution and hydrides ZrH<sup>x</sup> (x = 0.05–2) representing fragile sulfur black powder. The hydride powder is stable at room temperature but ignites at 430 ◦C on the air when the hydride crystals are stable up to 600–750 ◦C and then they decompose with the release of hydrogen and dissociate to the pure metal at 1200–1300 ◦C in a vacuum [153,154]. Thus, the released hydrides may not cause any difficulties during electrical discharge machining.

Processing alumina Al2O<sup>3</sup> in water should not provoke the appearance of new chemically active substances in the interelectrode gap, while processing in oil or kerosene can cause of damage of the filtration system of the equipment due to the formation of explosive and chemically active aluminum carbide Al4C3.

It seems that the use of copper Cu, silver Ag, gold Au, or aluminum Al in the form of the film is the most suitable for EDM of alumina Al2O<sup>3</sup> in water. In normal conditions, alumina of the workpiece, an oxide film of the auxiliary electrode made of aluminum, and formed alumina particles in the process of electrical erosion are inert to water. Ag and Au films and their oxide films do not form compounds that are resistant to high temperatures: the formed gold(III) oxide Au2O<sup>3</sup> is a thermally unstable conductor.

It should be noted that, usually, EDM in oil or kerosene provides more uniform morphology than in water due to more even heat removal from the treatment zone and the damping effect of a more viscous medium to compensate for forced oscillations of the wire electrode during processing [4,11,26,49].

The use of cerium(IV) dioxide CeO<sup>2</sup> (Table 3) as an assisting powder with brass wire or profiled electrode is not highly recommended. It has a similar but brighter nature as the uses of the brass tool during EDM of nickel-containing steel or Ni-coated workpieces. The cerium Ce and nickel Ni react with zinc Zn of brass very actively with the formation of intermetallides. The reaction of nickel Ni and zinc Zn at temperatures above 1000 ◦C has an explosive character that can be observed as a series of sparks in the interelectrode gap during processing [26,44,53,155,156]. The reaction of cerium Ce and zinc Zn has a more violent character that is typical for cerium. The reaction at high temperatures during local heating and consequent fusion of two powders provide a bright flash and a powerful explosion [157,158].

One of the useful properties of titanium carbide TiC assisting powder can be its interaction with nitrogen N at a temperature above 2500 ◦C. It forms conductive titanium nitride TiN that can be an advantage in EDM of non-conductive silicon nitride Si3N<sup>4</sup> [27,28].

**Table 3.** The analytical chemical composition of erosion products and the subsurface layer of the processing surface during electrical erosion of some ceramics using a brass or copper electrode tool [129,159–163].



**Table 3.** *Cont.*

#### **6. Conclusions**

It was shown that the erosion products and subsurface layer of the machined surfaces are formed directly from the components of ceramics, dielectric medium, auxiliary electrode, assisting powder during electrodes sublimation in the presence of heat. The chemical composition of the subsurface layer of the machined surfaces and the machinability of new materials not always depends on the electrophysical properties of the material but also on the combination of the materials of the primary and auxiliary electrode, conductive additives, assisting powder and workpiece.

All possible component interactions should be taken into account when developing the technology for electrical discharge machining non-conductive ceramics since the formation of certain chemicals in the form of insoluble or pyrotechnically dangerous sludge and gas can have dramatic consequences for the quality of the machined surfaces, the service life of the equipment and its units, and even for personnel.

The ceramic pressing methods, electrophysical and chemical properties of components in the presence of heat, data on possible interactions, and suitable combinations of the materials for processing structural oxide ceramics were presented for the most spread aluminum oxide and zirconium dioxide but can be applied for functional ceramics and nanocomposites.

According to the conducted analytical study, electrical discharge machining in oil or kerosene medium gives better qualities of the machined surface. A uniform surface morphology is formed due to more uniform heat removal and the damping effect of a more viscous medium to compensate for forced oscillations of the wire tool.

Titanium and chromium are most suited from the considered transition metals, taking into account the basic principles of physical chemistry, the laws of the structures of substances, their structure and transformations, available theoretical and practical data for electric discharge machining in oil or kerosene. They allow obtaining electrically conductive substances as erosion products, which can contribute to more dense erosion in the interelectrode gap.

During processing in a water medium, the use of silver, gold, aluminum as an auxiliary electrode are more suitable for copper or brass electrode, since aluminum is inert to water, silver and gold do not form compounds that are resistant at high temperatures. Moreover, the formed Au2O<sup>3</sup> is the conductor but not heat-resistant.

Machining alumina in water should not provoke the appearance of chemically active substances in the interelectrode gap, while processing in oil or kerosene can have a negative influence on the filtration system of the equipment due to the formation of explosive and chemically active Al4C<sup>3</sup> in the treatment zone. It is also evident that sintering alumina cannot be implemented in carbon dies, as the consequences of this combination can be non-electrical nature of the workpiece destruction and rapid wear of die paddings.

Probably, applying the brass tool during machining nickel alloys should be revised to the direction of using more chemically neutral to nickel materials.

Cerium dioxide cannot be used as assisting powder in combination with a brass tool as well.

During the development of the technology for ceramic machining, the preferred materials for the auxiliary electrodes, assisting powder and conductive particles should be the materials that provide conductive erosion dust in the interelectrode gap.

The developed approaches in electrical discharge machining structural ceramics can have an impact on the industry if the obtained data will be taken into account.

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

**Funding:** This research was financially supported by the Russian Foundation for Basic Research, grant number 19-08-00899.

**Acknowledgments:** The research was done at the Department of High-Efficiency Processing Technologies of MSTU Stankin.

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

#### **References**


© 2020 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 (http://creativecommons.org/licenses/by/4.0/).

## *Review* **On Electrical Discharge Machining of Non-Conductive Ceramics: A Review**

**Marina Volosova , Anna Okunkova \* , Pavel Peretyagin , Yury A. Melnik and Natalya Kapustina**

Department of High-Efficiency Machining Technologies, Moscow State University of Technology STANKIN, Vadkovskiy per. 3A, 127055 Moscow, Russia

**\*** Correspondence: a.okunkova@stankin.ru; Tel.: +7-909-913-12-07

Received: 4 July 2019; Accepted: 5 August 2019; Published: 8 August 2019

**Abstract:** The inability of ceramic and nanoceramic processing without expensive diamond tools and with a high-material-removal rate hampers the scope of its potential applications and does not allow humanity to make a full shift to the sixth technological paradigm associated with Kuhn scientific revolutions and Kondratieff's waves and restrains the growth of the economy. The authors completed a review on the research state of ceramic and nanoceramic processing by electrical discharge machining, which is possibly solved by two principal approaches associated with the usage of standard commercially available machine tools. The first approach is related to the introduction of expensive secondary phase; the second approach proposes initiate processing by adding auxiliary electrodes in the form of coating, suspension, aerosol, or 3D-printed layer based on the components of silver, copper, or graphite in combination with an improved dielectric oil environment by introducing graphite or carbon nanoparticles, which is hugely relevant today.

**Keywords:** electrical discharge machining; nanoceramics; coatings; auxiliary electrode; electrical conductivity; oxides; nitrides; carbon particles; oil medium

#### **1. Introduction**

By the beginning of the 21st century, ceramics and hard alloys were entirely developed and widely spread in the field of engineering [1–3]. It is difficult to imagine modern production in almost all machinery industries without cutting tools or other products made of ceramics and hard alloys. The scope of their applications is extensive but has one distinction—products of ceramic and hard alloys are used in extreme operating conditions under heat and ultimate loads. In this condition, no other material can provide the necessary performance, heat and wear resistance [4,5].

Due to the excellent exploitation properties of ceramics [6,7] and according to the summary analysis of experts, the volume of the world ceramic market grows at an average annual rate of 9.2%, and by 2020, the market volume will reach \$103 billion [8–11]. The trends for nanoceramics can be even more significant in further perspective. According to the results of recent studies [12,13], the samples of nanoceramics obtained by improved hybrid spark plasma sintering can exceed in main operational parameters (such as hardness, crack and wear resistance) of technical ceramics sintered by traditional isostatic pressing with subsequent vacuum sintering by 30%. The specialists of the subject area associate this phenomenon with the Hall–Petch relation, which describes the growth of the yield strength of a polycrystalline material with a decrease in the grain size [14,15].

However, the introduction of nanoceramics into real life is limited because of their high costs and low workability. The manufacture of each product requires high costs and is very labor-intensive. It requires fewer allowances in processing and the purchase of precision cutting tools based on artificial diamond, as cutting tools made of traditional alloys are worn extremely fast. In its turn, the geometry of cutting tools limits the resulting geometry of the product. Other machining methods such as vibroacoustic machining, laser and electron beam scribing, selective laser sintering with a binder, find their specific place in the processing of ceramics. However, they do not allow the receiving of all geometrically possible shapes of the product without losses on the operational parameters of ceramics [16–18].

One of the most popular methods for the production of parts, regardless of their hardness, is electrical discharge machining (EDM). However, it requires the material to be conductive [19–21]. The process of electrical erosion of the material consists of initiating electric pulses in the interelectrode gap, in the breakdown of the dielectric medium in the gap, in establishing a discharge channel, where the temperature reaches more than 10,000 ◦C (low-temperature plasma). Due to such temperatures, the material is ablated, the drops of cooled material as eroded debris are washed away by the flow of the dielectric medium, and the electrodes again approach to resume the cycle. Thus, a unique crater-like topology on the surface of the workpiece is created [22–25].

The inability of ceramics and nanoceramics to conduct current hampers the scope of its potential applications and restrains the growth of the progress and the switch to the next technological paradigm that most of the scientists associate with the term "nano" [26–32].

Thus, it can be concluded that the study and development of a method for electrical discharge machining of non-conductive ceramics and nanoceramics are incredibly relevant today.

The authors propose to apply the dialectical approach of cognition as a scientific approach to solve the formulated scientific problem. The dialectical approach includes the systematization and updating of existing knowledge and new data about the object of research.

#### **2. Scientific Problem**

The scientific problem of the physical impossibility of superhard ceramic materials machining (as the exception is the diamond grinding method) is considered because of their nearly zero conductivity since the electrical conductivity of ceramics, as for all solids, depends on the concentration of charge carriers, their magnitude, and mobility. The material conductivity is composed of each its constituent phase electrical conductivity: The measurement of the electrical conductivity of especially pure oxide ceramics becomes a complex scientific and technical problem. In many cases, the problem is the impossibility of obtaining the necessary complex spatial geometric shape of surfaces of functional products from high-strength ceramics by traditional methods of processing. There is a strong need of the development of a new inexpensive way of ceramic nanocomposites machining.

There are two main approaches to solving this problem to achieve this goal:


The second approach is the less expensive and may have a high level of technological adaptation to the conditions of real machine-building production, and is capable of producing positive results on a large scale. First of all, these conclusions are related to favorable technological conditions for the sintering of single-phase nanoceramics from the relatively low cost of ZrO<sup>2</sup> and Al2O<sup>3</sup> nanopowders in comparison with the cost of the nanoscale modified electrically conductive phases (for example, graphene, graphene oxide). The other reason is the relatively low cost of materials for the production of a profile tool for EDM.

The design of constructive and technological features of the product allows the withdrawal of the application of superhard materials such as nanoceramics based on ZrO<sup>2</sup> and Al2O<sup>3</sup> on a new production level [45–48]. It contributes to their spread in the industry for creating a new class of wear-resistant parts, which are capable of operating under extreme operating loads without losing their performance. That is extremely important in solving tasks of the transition to advanced intellectual production, new materials, and methods.

#### **3. Current State of Research**

#### *3.1. The Main Scientific Competitors*

The research subject is connected to the research of the method of applying EDM to the processing of nanoceramics made by advanced sintering technologies. The analysis of the leading world scientific competitors in this domain shows that the world community is separated as it was mentioned above. One group of scientists is focused on the modification of nanoceramics with conductive inclusions to create systems such as ZrO2-Ta, ZrO2-TiC, ZrO2-TiCN, Si3N4-TiC, Si3N4-TiCN, Al2O3-TiC, Al2O3-TiCN [49–59], as well as inclusions based on graphene and oxide graphene and graphene nanotubes [23,60–64]. Another group of scientists has focused their research on the processing capabilities of existed ceramics and non-conductive ceramic composites. The processing ability (ability to EDM) of such ceramics can be modified by introducing an auxiliary electrode based on copper, brass, steel, carbon, etc. in the form of particles, nanoparticles or thin-film in the process of electrical discharge machining. In this case, an auxiliary electrode will play the role of an initiator and intensifier of technology.

One of the most successful examples of such work is the work of the scientists at the Universities of Chemnitz and Freiburg and their colleagues at the Toyota Concern [65–68]. The research team at Chemnitz Technical University (Germany) under the guidance of Prof. Andreas Schubert have worked on the processing of insulating materials [23,69]. In particular, they propose replacing expensive diamond cutting with micro electrical discharge machining (µEDM) for specific technical applications. The researchers obtained a result on the µEDM of technical oxide and nitride ceramics (ZrO2, Si3N4-TiN, and others). In this case, the copper electrode moved progressively along the surface, following the developed control program of CNC-machine. The electrode tool motion looks like kinematic motion of milling cutter during 3-axis milling of grooves and pockets.

The search for alternative methods of processing non-conductive materials was also conducted at the University of Freiburg (Germany) [70]. The authors described two technologies for forming an auxiliary electrode in the form of a coating: Applying a suspension of gas soot and polymethyl methacrylate (PMMA, organic glass or acrylic) by hand and applying the suspension by direct printing carbon conductive ink (C-W) on a commercially available Ekra M2 printer. The coating was carried out on cylindrical samples of non-conductive technical ceramics ZrO2. As a result, the fundamental possibility of initiating EDM was proved; eroded debris were involved in processing. However, the authors proposed to use both methods exclusively for super-precision low-cost final finishing since the specific material removal rate could be characterized as modest in comparison with traditional EDM of metals. In their work, the authors criticized the use of plasma-vacuum coating as an initiator of EDM because of its high cost and inaccessibility. The authors of the article, in turn, referred to the successful experience of their Japanese colleagues, in particular, the results of Prof. Mohri.

In Japan, with the cooperation of such scientific schools as the Technological Institute of Toyota, The Technological University in Nagaoka and the Technological College in Tsukuba, in the 1990s, a team of scientists was formed under the guidance of the eminent Prof. Naotake Mohri, with the participation of Prof. Katsushi Furutani. In 1996, this group [71] proposed a new method of EDM of dielectric ceramics, by applying an auxiliary electrode in the form of a metal grid on the surface being machined. The EDM was carried out with a brass or copper electrode using kerosene as a working fluid. Prof. Mohri continues to publish the results of research as part of the same research team [72], in

which work is underway to develop a method for applying a conductive coating for ceramics based on crystallized carbon or carbides and working oil, and explores the phenomenon and mechanisms of initiation and acceleration of electrical erosion of a non-conductive substance.

Their colleagues from the Council for Scientific and Technical Research of Turkey, under the guidance of Prof. Can Ço ˘gun, obtained other impressive results: The process of coarse electrical discharge and electrochemical machining of non-conductive materials KTU-1, TiO2, glass, ZrO<sup>2</sup> (with the addition of Y2O3), and Al2O<sup>3</sup> by adding carbon particles no larger than 30 microns in a dielectric liquid (kerosene) [73–75]. The treatment was carried out using an electrode in the form of a hollow copper tube (a diameter of 3.5 mm). Thus, it was possible to drill the rough holes in workpieces of various shapes, their geometry was obviously faulted, perhaps due to the lack of reliable fixation of spherical blanks on the machine's worktable; the experiment with Al2O<sup>3</sup> turned out to be especially unsuccessful.

The research on the processing of non-conductive materials was also conducted at several research centers in Malaysia. The work published by the research team under the supervision of Dr. Muhammed Abdul Maleque at the International Islamic University of Malaysia (Kuala Lumpur) [76] demonstrated the results of the processing of technical ZrO2, wholly covered with a casing of platinum copper. The obtained result proved the ability and effectiveness of EDM processing in kerosene; copper residues from the surface of the ceramic workpiece were successfully removed. Another Malaysian group of scientists under the supervision of Dr. Mohd Amri Lajis at the Technical University of Malaysia (Malacca) in cooperation with the University of Malaysia described the result [77,78] where an acrylic blank was machined by placing an electrode inside a steel sleeve of 15 mm in diameter. Thus, the authors obtained a conical hole in the non-conducting material following the geometry of the electrode.

The research in the field of EDM processing of dielectric materials is carried out at the Perm National Research Polytechnic University under the supervision of Dr. Timur Rizovich Ablyaz. The authors theoretically and practically investigated the principal ability of EDM processing of non-conductive materials as Al2O3+ZrO<sup>15</sup> ceramics by initiating a discharge channel by placing the erosion products in the interelectrode gap [79–81]. The authors provided an analysis of the EDM methods for dielectric materials and considered modern approaches to optimize technologies proposed by previously mentioned colleagues. Their ideas deserve attention and detailed experimental approbation.

#### *3.2. Production of Nanoceramics by Advanced SPS*

Today the most sought-after researchers are focused on creating new nanomaterials by advanced spark plasma sintering (SPS) [37,39,82–84]. The development and creation of new progressive nanocomposite materials based on ceramic-ceramic and metal-ceramic compositions strengthened with metal nanoparticles and carbides, carbon nanotubes and nanofibers, graphene and graphene oxide are popular due to improved physical and mechanical properties and chemical inertness. The mathematical models developed for SPS assist in the controlling of properties of new materials [85–87].

A technical concept for creating a new class of nanocomposites by spark plasma sintering [88–92] includes the principles of interfacing the components of an innovative spark plasma sintering system, components of a high-current source node equipped with a pressure and temperature controller, and calibration.

Structural nanoceramics can have gradient properties according to particular customer requirements. The developed concept of creating nanocomposite gradient materials consists of creating a new functional mold and using low sintering temperature to minimize residual stresses [93]. The nanostructured powders for this kind of material are prepared according to the known scheme by colloidal processing [37,39].

As is known, one of the critical parameters for sintering materials is the distribution of temperature fields over the entire volume of the sintered product [94,95]. For this reason, the achievement of a uniform distribution of temperature fields in volume is a critical point in the design of tooling and die. Thus, a new model of semi-molds for sintering was designed [37]. The device contains a die of heat-resistant conductive material, insulating a sleeve made of heat-resistant heat-conducting and non-conductive material, a lower punch made of heat-resistant conductive material, an upper punch consisting of two heat-resistant conductive concentric parts, separated by an insulating material. At the same time, punches are processed, and the composition of powder materials is separated through gaskets. This new design allows changing the flow of electrons flowing through the die and, thereby, ensuring its heating. When the percolation of the sintered material is reached, the electrons begin to flow through both the die and the powder material to be processed. Thus, uniform heating of the product throughout its volume is achieved.

Simulated physical processes in order to optimize the technology can help to study the process of sintering and develop solutions for the problems related to the creation of new materials. As part of the study of the problems of technological development for the synthesis of new nanostructured materials by the method of spark plasma sintering, a theoretical description of the sintering process using methods of mathematical and theoretical physics, based on the theory of density functions and nucleation theory, the Keldysh method for electron transfer in a medium consisting of a conglomerate of nanocatrices (tungsten, aluminum and titanium alloys), with additions in the form of crystalline and amorphous mesostructures, causes genuine interest. This method is currently developed in detailed mathematical models of heat and mass transfer in areas of axisymmetric complex shapes of rotation bodies (in connection with the problems of modeling nozzles and nozzles of rockets, aircraft engines). However, with the parameters of the spark plasma sintering process—pulse duration in the range from 1 to 250 ms, average sintering pulse power up to 150 kW, and heating of the mold with powder material up to 2000 ◦C—new approaches are needed to solve the mass transfer equations, vibrational equations, and temperature distributions, with singular parameters of the substance, thermal conductivity and heat capacity, and time modulated by short pulses of sintering current of extra-large power. In this regard, the problem can be solved by simulating a substance in a cylindrical sintering area by directly pulsing the current through a mold with powder material and additional induction heating and obtaining the microphysical parameters of powder materials using the method of asymptotic approximation and singular generalized functions.

Since in the process of pulsed sintering in the mold, nanoparticles of metals (W, Ti, Al) and crystalline nanostructures (nanotubes, graphene, graphene oxides, nano-diamonds) with a particle size of 50 to 500 nm pass through a characteristic time of several tens of phase transformations, with diffusion and threshold processes for the transfer of molecules and electrons, the result of sintering and the physical properties of the substance obtained are not predictable without precisely setting and modeling the initial data and the dynamic behavior of the physicochemical properties of particles during the sintering process.

Preliminary experimental models were developed that use a Langmuir probe, an optical probe–waveguide to remove the spectral characteristics of the plasma during sintering, for measuring the phases of the voltage and current of the sintering pulse to study these properties. The main theory explaining the physicochemical properties of the material obtained after the action of a series of pulses is the theory of the growth of germ clusters and nucleation (Frenkel–Landau).

In the course of the work for the investigation of the process of pulsed sintering, new effects that were not previously considered in international literature were found: Accelerated nucleation, electronic nanojet, nanofilament, and photon nanojet in the field of thermal photons. At present, an important physicochemical process of reducing the threshold power of electronic breakdown for sintering conglomerates of dielectric nanoparticles by introducing additional impurities from metallic conductive nanoparticles was found and studied. It should be noted that for these processes there are important optical analogs. In the visible and infrared range, there is a new effect, not previously considered as an application to the process of a pulse, as well as a laser sintering effect—the photon nanojet (photonic nanojet). This effect is used as an analog of the electronic jet, the spatial effect of high electron density on the surface of nanoparticles, due to the excitation of the Frohlich modes to simulate

the sintering processes. In addition to the theoretical study of nonlinear effects, in the process of pulsed sintering, there is the possibility of the practical application of a photon jet for a wide range of technical applications as a working medium for microlasers, microlenses, radiation detectors, and metamaterials. Interest in the considered processes takes place in the microwave field, in the calculation and design of metamaterials for the microwave coatings of airplanes and rockets, as well as in the calculations of thermal protective coatings.

The properties of sintered Al2O3-TiC-Ni nanocomposite in comparison with the properties of a sample of aluminum oxide (Al2O3) obtained by hot pressing—the traditional sintering method exceed them due to smaller alumina grains [39]. The average particle size of the aluminum oxide powder in the nanocomposite is approximately 0.3 ± 0.1 µm, while in a monolithic material this size is 3 ± 1 µm. The structure of the Al2O3-TiC-Ni nanocomposite contains Ni and TiC nanoparticles. Although the process of sintering occurs at a temperature of 1375 ◦C, the nanoparticles do not significantly increase in size. The hardening of the nanocomposite was expected due to the presence of Ni nanoparticles. The measured relative density of the samples was >98%. The difference in wear resistance between Al2O<sup>3</sup> and Al2O3-TiC-Ni nanocomposite under dry slip conditions depends on the hardness of the material, as well as the size of the aluminum oxide grains in the nanocomposite. Both parameters are most important for improving the tribological properties in working conditions.

#### *3.3. Classification of Technical Ceramics and Its Workability by EDM*

Technical ceramics are classified based on functionality such as biocompatibility, electrical properties, optical properties, mechanical strength, and resistance to high temperatures according to the following types: Medical ceramics; electroceramics; optical ceramics; structural ceramics and ceramics for extreme conditions; ceramics for energy transfer, storage and conversion; and other types of ceramics [96,97].

It should be noted that high-tech industries are experiencing an acute shortage in the production of ceramic materials and high-tech products made from them, which have high performance properties.

On the one hand, this is due to the use of outdated technologies in production and the lack of knowledge in obtaining advanced high-strength ceramic nanomaterials and composites based on them, and on the other hand, using outdated technologies for processing ceramics. The most advanced production is directly dependent on suppliers of equipment and limited in the choice of technological processing modes. This undoubtedly endangers the adequate response to modern challenges.

Electrical discharge machining is a highly efficient method of processing electrically conductive materials regardless of their mechanical properties (Figure 1). The schematic presentation of the electrical erosion steps under discharge pulses is shown in Figure 2. The regular discharge gap forms on the surface of the workpiece during the formation of the cavity to be processed: Two electrodes approach up to the moment of first working discharges, which transforms into the discharge channel (Figure 1a). Then the tool electrode advances in the cavity formed by the first discharges (Figure 1b) and creates a regular cavity in the workpiece (Figure 1c). On this stage, the pulses acting between the two electrodes balance each other (Figure 1d), and electrical discharge machining becomes stable.

μ μ

Δ **Figure 1.** Diagram of the electrical discharge machining (EDM) process by example of cavity formation on the surface of conductive nanoceramics: (**a**) Dielectric breakdown; (**b**) stabilization of discharge pulses; (**c**) formation of regular cavity; (**d**) ongoing processing; *S* is electrode feed speed; ∆ is a discharge gap. Δ

**Figure 2.** A schematic cycle of the material erosion under discharge pulses.

Modern scientists divide the main processing factors that influence the results of electrical discharge machining into four main groups [24,25,98,99] (Figure 3):


**Figure 3.** The main factors of EDM affecting the functionality of the final product.

μ All these factors influence the stability of EDM. The quality of the processed surface can be controlled by measuring final geometrical accuracy, roughness, microhardness, change of chemical content in subsurface layer (up to the depth of ~4–10 µm), homogeneity of the processed surface, and an absence of the visible defects, which can influence the final product appearance and functionality. In this case, the product functionality can be presented as a complex, which consists of its operational ability as part of a more sophisticated unit such as a mechanism or machine and its appearance following customer requirements.

Ω EDM is possible only if the minimum electrical conductivity of materials is about 10-2 Ω·cm-1 . The electrical conductivity of the material is a critical factor in the success of the process of electrical erosion. All metals fulfill the condition in terms of minimum electrical conductivity, which is not the case with ceramics and its composites. Figure 5 shows the main ceramic materials, depending on

**Figure 4.** Schematic diagram of the EDM of non-conductive materials using a specially developed electrically conductive coating: (**a**) Initiating discharge impulses between the coating and the tool electrode; (**b**) electrical erosion of the coating, formation of the erosion products between the conductive film and the electrode tool; (**c**) electrical erosion of the non-conductive workpiece with absorbed erosion products by the workpiece; (**d**) reinitiating electrical erosion of the workpiece that keeps constant presences of the erosion product in sub layer of the workpiece.

**Figure 5.** Main ceramic materials in dependence of its electrical conductivity [100,101].

#### **4. Discussion**

An analysis of the current state of research on this issue and the main guidelines of research in world science showed that the world community is focused mainly on the modification of nanoceramics with current-conducting inclusions to create systems of the type "oxide or nitride ceramics—conductive additives". As well, the most popular but extremely expensive inclusions today for the production of a material system with advanced exploitation properties are nanotubes of graphene and graphene oxide [28,56].

Despite the efforts of the world scientific community to solve the global problem of processing non-conductive ceramics and nanoceramics due to the inclusion of the secondary phase, the issue remains unresolved [34,37,47,48,50,51,88,102]. The main issue that hampers the advances is related to the material of the secondary phase, which often is not widely available for use in the conditions of real production [33,34,36,37,40,44,89,103]. Besides, the material of the secondary phase can significantly change the appearance of ceramics, such as transparency and color, which directly affects and restricts the areas of its potential application [7,8,39,104].

In parallel with the mentioned approach, the world's leading scientists work on the feasibility of processing technical ceramics by including an initiator and intensifier of the process in the form of an auxiliary electrode based on copper, brass, steel, carbon inclusions, etc. [15,23,35,43]. The additional electrode can be presented in the form of coating deposited by different techniques (including thermal spray and coating technologies or by aerosol, 3D-printing, or drying suspension disposed at a workpiece) or in the form of carbon particles introduced into the processing zone. Suspension can be based on the components of silver, copper, or graphite. A dielectric oil environment can be improved by adding graphite particles and nanoparticles (carbon, graphene nanotubes) in combination with a conductive coating or thin film (up to ~4 µm) or aerosol nanoparticles, which is exceptionally relevant today [15,38,42–44,47].

The developed method of the precision shaping of functional products will subsequently contribute to the introduction and distribution of structural nanoceramics in the machine-building industry. It can significantly vary available shapes and properties of nanoceramics regardless of their conductive ability.

The problem of EDM of ceramics needs a fundamentally new approach to find a way of processing by using an economically available solution for direct individual production. Today, the usage of a specially developed conductive coating or thin film with a thickness of ~10-20 µm based on the suspension can be one of the technically available solutions.

The presence of this coating will initiate discharges at non-conductive material. The temperature in the discharge gap can reach 10,000 ◦C, which is suitable for processing any matter. The presence of carbon particles in the oil medium and the eroded debris, which contain the carbon formed from the organic medium, will contribute to processing.

For example, in the case of processing naturally widely available oxide ceramics such as ZrO<sup>2</sup> and Al2O3, their melting points are of 2715 ◦C and 2054 ◦C, respectively [33,49,50,68]. The introduction of an additional process initiator can result in the physical ablation of the ceramics with the formation of the conductive form ZrC and unstable compound Al4C3, which can theoretically assist processing.

This method can be called superior to the proposals of its competitors in terms of its use for processing structural nanoceramics. The development and application of the technique can contribute to widespread structural nanoceramics in such industries as aircraft and aerospace, where resistance to the heat and wear is relevant. The processing of nanoceramics with the usage of standard mechanical equipment such as EDM without adding an expensive secondary phase can significantly reduce the cost in manufacturing to surpass any potential manufacture competitors by several times.

#### **5. Conclusions**

#### *5.1. Research Work*

Based on the results of the research and obtained experimental and analytical results described above, a conductive coating with a thickness of ~10-20 µm for use as an initiator and intensifier of electric erosion corresponds to general world trends. Its application to technical ceramics and nanoceramics in addition to the refusal to modify nanoceramics itself using an expensive conductive secondary phase, which hampers the switch to the next technological order, will allow an advanced and accessible development, in all respects, of a way of processing structural and non-conductive nanomaterials without the loss of their original operational properties such as high resistance to the heat and abrasive wear. In perspective, it can allow the mechanical engineering industry to shift to a new paradigm of production related to the application of nanomaterials, which is unique in its performance properties in terms of its thermal stability and wear-resistance.

#### *5.2. Future Work*

Further experiments with the processing of innovative nanoceramics can give a fundamentally new development for the presented idea. It will reveal the principal ability of processing non-conductive oxide nanoceramics and allow the identification of specific conditions and technological modes, and will draw conclusions about the fundamental laws of the main EDM parameters' influence on the workability of nanoceramic workpieces. That is necessary to form a unique geometry of the typical product sample, for which the solution of the global problem of the operational ability for non-conductive materials is critical. It is important to promote the switch to the next technological paradigm with the elaboration of a new accessible method to process innovative ceramics with equipment, which is widely available at modern mechanical production enterprises.

**Author Contributions:** Conceptualization, M.V.; Methodology, A.O.; Software, N.K.; Validation, P.P.; Formal Analysis, Y.A.M.; Investigation, A.O.; Resources, Y.A.M.; Data Curation, N.K.; Writing-Original Draft Preparation, P.P.; Writing-Review & Editing, A.O.; Visualization, P.P. and N.K.; Supervision, M.V.; Project Administration, M.V.; Funding Acquisition, Y.A.M.

**Funding:** This research was funded by the Ministry of Education and Science of Russian Federation, grant number No. 9.7453.2017/6.7.

**Acknowledgments:** The research was done at the Department of High-Efficiency Machining Technologies of MSTU Stankin.

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

#### **References**


© 2019 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 (http://creativecommons.org/licenses/by/4.0/).

## *Review* **Roles of Cryogenic Cooling in Turning of Superalloys, Ferrous Metals, and Viscoelastic Polymers**

## **Tze Chuen Yap**

School of Engineering and Physical Sciences, Heriot-Watt University Malaysia, Putrajaya 62200, Malaysia; t.yap@hw.ac.uk; Tel.: +60-3-8894-3780

Received: 24 June 2019; Accepted: 31 August 2019; Published: 5 September 2019

**Abstract:** Cryogenic machining is a relatively new technique in machining. This concept was applied on various machining processes such as turning, milling, drilling etc. Cryogenic turning technique is generally applied on three major groups of workpiece materials—superalloys, ferrous metals, and viscoelastic polymers/elastomers. The roles of cryogen in machining different materials are unique and are summarised in this review article. Finally, the challenges in using cryogenic machining in industries are also highlighted.

**Keywords:** cryogenic machining; review; liquid nitrogen; liquid carbon dioxide

#### **1. Introduction**

Conventional cutting fluids are used to reduce the temperature of tool and work in machining. However, conventional cutting fluids are unable to work effectively in high-speed machining of some superalloys such as titanium alloys, inconel alloys and tantalum alloys [1]. The other major problems caused by the conventional cutting fluids are health and environmental problems [2–5]. The conventional cutting fluids can be categorised into semi-synthetic, synthetics, straight, and soluble oils [6]. Such additives are added onto cutting fluids to provide lubricating effects. Some of these additives (aromatic amines, chlorinated paraffines, etc.) might cause a carcinogenic potential [3]. Beside this, Suliman et al. [5] identified several species of bacteria and fungi growth found in cutting fluids used in workshops. The contamination of the cutting fluids by these micro-organisms can cause several types of health problems, such as septic infections, primary allergic bronchopulmonary aspergillosis, dermatomycosis, and neonatal meningitis [4,5].

The conventional cutting fluid can be recycled several times until its quality degrades after the useable period. The storage and proper disposal of these cutting fluids require special processes and extra costs [2,7,8]. Moreover, incorrect disposal of these cutting fluids causes environmental pollution. As such, new machining techniques such as green machining, cryogenic machining, and dry machining have been proposed to substitute conventional machining with cutting fluids.

#### **2. Cryogenics Machining**

#### *2.1. Cryogenic Machining in General*

Cryogenic machining is a relatively new technique in reducing machining temperature by replacing the conventional cutting fluid with extremely cold or sub-zero (below −150 ◦C) cutting fluid at the machining interface. Nitrogen (N2), carbon dioxide (CO2), oxygen (O2), helium (He), etc., in compressed forms, are the potential liquid cryogens to be used. These gases are exist in the atmosphere and can be converted to a liquid form. These liquid cryogens evaporate in room temperature and convert to gas again [9]. Nevertheless, only liquid carbon dioxide (LCO2) and liquid nitrogen (LN2) are commonly used in cryogenic machining studies [10–12]. The boiling points of CO<sup>2</sup> and N<sup>2</sup> are

−78.5 ◦C and −196 ◦C. Clearly, the boiling point of nitrogen is lower than carbon dioxide. Besides, carbon dioxide gas is denser than atmospheric air, and it might accumulate at the ground of the plant and can cause breathing problems to machine operators [11]. Thus, liquid carbon dioxide is not recommended to be used as a cutting fluid in machining, and liquid nitrogen is preferred for use as a cryogenic cutting fluid. In 1953, Bartle used liquid carbon dioxide as the coolant in machining [11]. The earliest investigation on cryogenic machining with liquid nitrogen (LN2) as cryogen was carried out by Uehara and Kumagai [12]. Cryogenic machining with liquid nitrogen LN<sup>2</sup> reduced cutting forces and improved tool life and surface finish of the machined parts. A similar result was reported by Fillippi and Ippolito [10], who conducted cryogenic face milling. However, the usage of cryogenic technology in early days was limited by the high cost. The idea of cryogenic machining was again re-proposed after the 1990's when the cryogenic technology was improved, and higher production rate (high-speed machining) was required [11]. At the same time, the awareness of health and environment issue was also spreading. From 1995–2007, three main groups of researchers have conducted investigations on cryogenic machining. Most of their works focused on cryogenic turning [1,9,11,13–23] and cryogenic grinding [24–27]. The number of reported works in cryogenic milling and cryogenic drilling researches are insignificant, compared to cryogenic turning.

Previous cryogenic turning explorations investigated the tool wear, surface roughness, and cutting forces. Improvements on cutting tool life under the cryogenic conditions have been reported by Hong and co-workers [11,17,18], Wang and co-workers [1,21–23] and Dhar and co-workers [13–16,19,20]. Beside this, better machined surface roughness is obtained through cryogenic machining, as reported by Dhar et al. [14,15] and Wang et al. [1,22,23].

Cryogenic machining technique is commonly applied on superalloys (titanium alloys, inconel alloy, tantalum alloy, etc.), ferrous metals, viscoelastic polymers and elastomers. In general, positive results are reported, regardless of the type of workpiece materials. Nevertheless, roles of cryogen are slightly different in machining different materials.

#### *2.2. Cryogenic Turning of Superalloys*

A superalloy has a combination of high mechanical strength and surface stability at high temperature [28]. Examples of superalloys are titanium alloy, inconel alloy, and tantalum alloy. Due to their characteristics, superalloys are used in various engineering applications. However, high mechanical strength of superalloys also causes poor machinability. Conventional cutting fluids are not efficient to reduce the machining temperature in machining superalloys, and cryogenic fluids were proposed to replace them.

Most of the work materials studied in previous cryogenic machining studies are titanium alloys. One of the earliest reported cryogenic liquid nitrogen machining studies was conducted by Wang and Rajurkar [1]. They designed a liquid nitrogen cooling system for their experiments. Liquid nitrogen was circulated through a sealed cap which was placed on the top of the insert to reduce the cutting insert's temperature. They measured and compared the cement carbide tool wear in turning Ti-6Al-4V with conventional cutting fluids (Mobilmet Omicron) and liquid nitrogen cooling. Liquid nitrogen outperformed conventional cutting fluids in terms of reduction of both tools' flank wear and surface roughness of the machined surface. However, no significant difference was observed in the cutting forces.

On the other hand, Hong and his co-workers did a series of systematic studies on cryogenic turning of titanium alloys in term of tool wear, surface roughness, cutting forces, friction force, and friction coefficient. In their studies, liquid nitrogen was injected on the interface of tool and workpiece, i.e., the liquid nitrogen was directly sprayed onto the tool and workpiece. In 2001, Hong and Ding [17] studied several cooling approaches in cryogenic machining of Ti-6Al-4V. They reported that the best cooling approach was simultaneous cooling rake and flank, followed by cryogenic cooling at the rake face and cryogenic cooling at the flank face. Beside this, they also concluded that by applying liquid nitrogen, the tool temperature can be reduced and the cutting speed of conventional machining of

titanium alloys can be increased (around 60 m/min). In the same year, Hong et al. [18] conducted a similar study but they changed the position of the chip breaker. At the optimum position of the chip breaker and with cryogenic cooling, the tool life was extended by five times compared to conventional machining. Hong et al. [29] furthered the research in measuring the cutting forces. They reported that cryogenic liquid nitrogen hardened the material, and this caused higher cutting force. However, the friction force between the cutting tool and workpiece was reduced due to two main reasons: the material became less sticky at a lower temperature, and liquid nitrogen formed a fluid cushion between the interfaces. As such, they concluded that liquid nitrogen was able to function as a lubricant. Furthermore, Hong [11] also compared the cost of cryogenic machining with conventional machining and showed that cryogenic machining with his patented nozzle was superior in terms of cost, tool life, surface finish, as well as environmental friendliness.

Cryogenic turning of Ti-6Al-4V was also studied by Venugopal et al. [19,20]. They built a special nozzle which was able to inject liquid nitrogen jets on the crater and flank faces of cutting tools. They conducted the cryogenic turning tests in cutting speed above the recommended speed for titanium alloys, (i.e., 60 m/min) under dry, oil, and cryogenic liquid nitrogen environments. From their study, adhesion-dissolution-diffusion wear are the main tool wear mechanisms at the crater in turning Ti-6Al-4V with uncoated carbide; while abrasion and attrition occurred at the flank. Liquid nitrogen cooling was able to control the cutting temperature and then reduced the tool wear (both flank wear and crater wear) at cutting speed of 70 m/min, but this effect decreased at higher cutting speeds (100 and 117 m/min).

Cryogenic turning of Ti-6Al-4V was reported by Dhananchezian and Kumar [30]. They modified the cutting insert used in their experiment by drilling a hole at the rake surface of the insert and applied liquid nitrogen to the hole (Figure 1). The results showed that cryogenic cooling performed better than wet machining in terms of cutting force, surface roughness, and tool wear. Yap et al. [31] studied the influence of low pressure cryogenic liquid nitrogen to the machining forces, friction, tool wear and surface quality under high-speed machining of Ti-5Al-4V-0.6Mo-0.4Fe. The experiments were conducted in dry and cryogenic liquid nitrogen conditions. The 0.034 MPa internal pressure inside a liquid nitrogen Dewar tank caused the liquefied nitrogen to be injected onto the tool–work interface. The experimental results showed that low pressure liquid nitrogen can reduce the friction force and friction coefficient during machining. In additional, surface roughness of the machined titanium alloy was improved and tool wear was reduced. Reduction in machining forces and improvement in surface roughness of machined parts can be attributed to better heat removal by injection of liquid nitrogen. While most of the researchers used only one cryogen in their cryogenic machining works, Kaynak and Gharibi studied cryogenic turning of Ti-5Al-5V-3Cr-0.5Fe with both liquid nitrogen and liquid carbon dioxide and compared with dry machining [32]. At low cutting speed (below 120 m/min), tool wear did not depend on cutting condition (dry/liquid nitrogen/liquid carbon dioxide) and no significant different in wear mechanism was observed on the cutting inserts from three different machining conditions. At higher cutting speeds of above 120 m/min, cryogenic machining with liquid nitrogen outperformed cryogenic machining with liquid carbon dioxide and dry machining. Different wear mechanisms were detected on the cutting inserts from the three different machining conditions.

**Figure 1.** Modified cutting insert for cryogenic cooling method [30].

Inconel is a nickel-based alloy with poor machinability [33]. Low thermal conductivity of inconel alloy (about 11 W/m ◦C) caused higher cutting temperature and high tool wear rate [1] and additional effort was required to reduce the cutting temperature. Wang and Rajurkar reported that machining Inconel 718 with LN<sup>2</sup> cooling was able to delay the tool wear and improve the surface roughness of workpiece surface [1]. Similar results were reported by Pusavec et al. [34]. In addition to tool wear and surface roughness, a thicker compressive zone is noticed beneath the machining surface and smaller grain size is observed after cryogenic machining with liquid nitrogen. Beside this, the sub-surface layer under cryogenic machining is harder and thinner than sub-surface layer under dry or Minimum Quantity Lubrication (MQL) machining. In other words, cryogenic machining is able to improve the surface roughness, and increase the hardness of the machined workpiece. The majority of previous cryogenic machining works supplied cryogen at a constant supply pressure. To study the effect of supply pressure and the corresponding flowrate of the cryogen, Klocke et al. conducted cryogenic machining tests at supply pressures of 7–30 MPa on Inconel 718 and then measured maximum flank wear of the cutting inserts [35]. They reported that the higher the supply pressure and flowrate of cryogen, the lower the maximum flank wear at a cutting speed of 500 m/min. However, they obtained a negative result at cutting speed of 60 m/min, where higher supply pressure and flowrate produced higher maximum flank wear. Earlier researchers used cryogen only in their cryogenic turning experiments, but recently hybrid cooling which combined cryogen with cutting fluids has been proposed. Bagherzadeh and Budak used four different cooling strategies in turning titanium alloy Ti6Al4V and Inconel 718 [36]. They introduced a new method, CMOL, where CO<sup>2</sup> and vegetable oil were mixed in the form of frozen oil particles, before reaching the tool–work interface. The new method, CMOL, is better than cryogenic turning with carbon dioxide cooling only, in terms of tool wear and surface finish. Yildirim compared cryogenic, nanofluids, and hybrid cooling in turning Inconel 625 [37]. Six cooling techniques such as dry, pure MQL, nMQL, LN2, and their hybrid were used in his work. Three types of nanofluids based on Al2O3, hBN, and hBN + Al2O<sup>3</sup> were included in the vegetable cutting fluid in 0.5 vol% and 1.0 vol%. Hybrid cooling (cryogenic liquid nitrogen and 0.5 vol% hBN) outperformed cryogenic cooling only and vegetable oil with nanofluids only in terms of reducing cutting temperature, prolonging tool life and improving surface roughness. Tantalum has low thermal conductivity, low specific heat, high shear strength, high work-hardening capacity, and gummy consistency, which cause it to be difficult to machine. In order to study the roles of cryogenic machining in machining tantalum, Wang and Rajurkar conducted a preliminary study on cryogenic machining of tantalum. They reported that liquid nitrogen was able to reduce the cutting temperature and extend the tool life [1]. In their further research in cryogenic machining of tantalum with liquid nitrogen [22], they found that cryogenic machining reduced tool wear by 70%, improved surface roughness by 200% and reduced cutting forces by 60%.

In general, cryogen is used to reduce the maximum cutting temperature and reduce/delay tool wear in machining superalloys. The majority of the previous works show positive results at selected ranges (cutting speed, or pressure). However, no consensus is obtained on the precise conditions where cryogenic machining is beneficial. Some works reported cryogens work best at lower cutting speed, but some other suggested cryogens only contribute at high cutting speed. The amount of the cryogen used and the location of cooling (workpiece, tool, or both) are important factors in deciding the final results. Overcooling superalloy will produce a negative effect. The roles of liquid nitrogen, whether beneficial or detrimental, depend on workpiece materials, cutting speed, and supply pressure of the cryogen. Furthermore, hybrid cooling (cryogenic machining with lubricant) was evolved from cryogenic machining, and it shows encouraging outcomes in machining superalloys.

#### *2.3. Cryogenic Turning of Ferrous Metals*

Ferrous metals such as steel were also investigated in cryogenic turning research. Zurecki et al. [38] found that the spraying of cryogenic liquid nitrogen on the tool rake during hard turning of AISI 52100 steel decreased the thickness of white layer, maintained the hardness of steel, and improved residual stress distribution of the machined surface. Kumar and Choudhury [39] investigated the effect of cryogenic liquid nitrogen on tool wear and cutting forces during high-speed turning of stainless steel. They obtained positive results in cryogenic machining in terms of reducing tool wear and cutting force. However, they raised their concern in the high consumption of liquid nitrogen because this increases the total cost of machining. Stanford, et al. [40] investigated tool wear in turning carbon steel BS 970-080A15 using uncoated tungsten carbide–cobalt insert in six different cutting environments (dry cutting, flood coolant, compressed air blast, nitrogen gas at ambient temperature, nitrogen gas at −40 ◦C and liquid nitrogen at −196 ◦C). They concluded that liquid nitrogen machining can provide a similar effect as provided by the flood coolant, and thus, liquid nitrogen can replace flood coolants for machining steel. Dhar et al. [14,15] reported that reduction in tool wear produced better cutting surfaces under cryogenic liquid nitrogen. Furthermore, the reduction in cutting temperature was more obvious at a lower feed and lower cutting velocity. Sivaiah and Chakradhar compared cryogenic, MQL, wet, and dry machining conditions in turning 17-4 PH stainless steel and they found that cryogenic turning outperformed in terms of tool wear, surface finish, and chip morphology compared to other three conditions [41]. The improvement was mainly due to retainment of cutting tool edge under cryogenic condition. Sivaiah and Chakradhar furthered their works and identified the optimum cryogenic cutting condition with two optimisation techniques [42]. The two techniques, 'Taguchi incorporated Gray relational analysis' (TGRA) and 'Taguchi coupled Technique for Order Preference by Similarity to Ideal Solution' (T-TOPSIS) were used to optimise the surface roughness, tool wear, and material removal rate of cryogenic turning 14-4 PH stainless steel with tungsten coated carbide. Both techniques were able to optimise the cryogenic turning process but the '*Taguchi incorporated Gray relational analysis*' was a better technique.

In contrast, Yap et al. [43] reported a different result. They studied turning of carbon steel S45C under three conditions; dry, wet, and cryogenic. Their results showed that even through a cryogenic liquid nitrogen jet is able to reduce friction coefficient in turning carbon steel, it failed to produce good surface roughness of the machined surface. Their experimental results suggested that conventional machining with cutting fluid is still the best method to obtain good surface roughness of carbon steel at a machining speed of 226 m/min). While most of the reported works focused on cutting forces, friction coefficient, tool wear, and surface roughness, some researchers investigated the surface integrity and corrosion behaviour cryogenic machined steel. Bruschi et al. investigated the effect of cryogenic machining to the surface integrity and corrosion behaviour of AISI 316L stainless steel [44]. Cryogenic cooling altered the microstructure of machined part, especially the outer surface. Beside this, stainless steel machined under cryogenic condition showed better corrosion resistance behaviour. Similarly, phase transformation from austenite in AISI 347 to martensite was reported by Kirsch et al. [45]. Cryogenic carbon dioxide was used to reduce the temperature of the contact zone during machining AISI 347 and precool the workpiece. Martensite was detected in the surface of the workpiece after cryogenic turning. In additional, microhardness of the surface was also increased after cryogenic cooling. Cryogenic turning can be further explored as a surface hardening improvement integrated in manufacturing methods.

For ferrous metals, cryogenic machining with cryogens such as liquid nitrogen and liquid carbon dioxide can reduce the cutting temperature, alter the microstructure/phase, and harden the surface of the workpiece. Similar to cryogenic turning of superalloy, a correct selection of parameter such as cryogen pressure, machining speed, etc. is important. Cryogen is used to 'reduce' the cutting temperature. However, overcooling of ferrous metals brings negative effects. Overcooling of ferrous metals increases brittleness and reduces toughness in ferrous metal, and then causes higher tool wear and poorer surface finishes. Therefore a proper cooling strategy is prerequisite in cryo-machining of ferrous metals. −

#### *2.4. Cryogenic Machining of Viscoelastic Polymers and Elastomers*

Viscoelastic polymers are hard to be machined precisely at ambient temperature because of their characteristics such as softness, high elasticity, and adhesion [46]. Kakinuma et al. [46] proposed to solve this problem by applying cryogenic machining. They milled polydimethylsiloxane (PDMS) inside a liquid nitrogen chamber (as shown in Figure 2) and obtained positive results. This is because cryogenic liquid nitrogen maintains the glassy state of PDMS during the milling process. Cryogenic milling of PDMS was also investigated by Song et al. [47]. Cryogenic liquid nitrogen improved the surface roughness of machined PDMS with machined temperature lower than −143 ◦C. In addition, the problems of adhesion and shrinkage in machining PDMS were also solved with correct amount of liquid nitrogen flow. A similar technique was applied by Dhokia et al., in machining elastomer [48]. With the technique of cryogenic liquid nitrogen machining, formation of adiabatic shear band was reduced in milling ethylene vinyl acetate (EVA) and Neoprene [49].

**Figure 2.** Jig for cryogenic micromachining [46].

This concept can be applied to cryogenic turning of viscoelastic polymers although less published works in cryogenic turning of viscoelastic polymers can be found in established databases. Putz et al. studied the effect of cryogenic liquid nitrogen on the cutting forces and mechanism of chip formation in turning elastomer [50]. They found that cryogenic machining can improve the machining accuracy of elastomer. This is mainly due to the elastomer's behaviour being changed from viscoelastic to energy-elastic with higher modulus of elasticity, under the influence of liquid nitrogen.

In general, cryogenic machining viscoelastic polymers and elastomers is not as popular as cryogenic machining of superalloys or ferrous steel. The role of cryogen in cryogenic machining of viscoelastic polymers and elastomers is different from the role of cryogen in cryogenic machining of superalloy or ferrous steel. Cryogen is used to change the mechanical properties of the polymer or elastomer during machining, i.e., to increase the modulus of elasticity so that dimensional accuracy can be improved.

#### **3. Challenges of Cryogenic Machining in Industries**

A large number of scientific studies have been conducted for more than 20 years, and most of them showed that cryogenic machining is generally better than conventional machining especially in machining superalloys. However, conventional industries still prefer the conventional machining

technique. The advantages and disadvantage of cryogenic turning are listed in Table 1. As shown in Table 1, one of the major obstructions is the initial set-up cost. Industries are required to invest heavily on the cryogenic cooling system. Beside this, they are unable to reuse the cryogen as how they recirculate cutting fluids in conventional cutting. High consumption of cryogen during cryogenic machining, as reported by Hong [11] and Kumar and Choudhury [39], is another barrier in applying cryogenic machining in industries. To minimise the usage of liquid nitrogen, Hong designed and patented a nozzle to deliver the liquid nitrogen onto the workpiece or/and cutting tool [18,51]. The newly designed nozzle can direct the liquid nitrogen to the rake face, or the flank face, or both faces and inject less liquid nitrogen. A similar concept was also adapted by Ahmed and his research team [52,53]. They used a modified tool to apply cryogenic liquid nitrogen in turning AISI 4340 steel and SUS 304 stainless steel. With their modified tool, a smaller amount of liquid nitrogen was used. However, this special nozzle and modified tool are still not commonly available as commercial products.



The third obstruction in applying cryogenic machining in industries is lack of clear guidelines for optimum combination of machining and cooling parameters for cryogenic machining. As mentioned previously, several factors determine whether cryogenic machining is beneficial or detrimental. The factors are cooling strategy (cooling the chip, cooling the cutting tool and cutting zone), workpiece materials, cutting speed, and supply pressure of the cryogen. Contradictory or different optimum factors/value were reported in literature. Currently, a general guideline to select the correct parameters is still unavailable for machine operators to refer to. The science behind cryogenic machining is a complicated field of knowledge and requires further investigation.

#### **4. Conclusions**

Cryogenic cooling is an alternative technique in machining. Cryogenic machining is applied on superalloys, ferrous metal, and viscoelastic polymers/elastomers. In general, cryogenic cooling during turning improved various performance indicators (surface roughness, tool life, tool wear, cutting forces, etc.) in machining superalloys (titanium alloys, inconel alloys, and tantalum alloys). This is due to cryogen being able to solve the major problem in machining superalloys, i.e., heat accumulated at the cutting zone due to poor thermal conductivities. Beside this, cryogenic machining also showed beneficial improvement in machining ferrous metal, if a correct setting is used. Correct application of cryogen can delay/reduce the tool wear in high-speed machining of ferrous steel and also modify the surface behaviour of the machined parts. On the other hand, cryogenic cooling altered hardness and adhesion properties of viscoelastic polymers, and improved the machinability of viscoelastic polymers. Cryogenic machining has several advantages compared to machining with conventional cooling. Nevertheless, additional set-up cost for a cryogenic cooling system is required. Beside this, the cost of cryogen is another factor that must be considered by industries in order to adopt this technique. A correct selection of machining and cooling techniques/parameters is another challenge for applying this technique in industries. Although cryogenic machining is more than 20 years old, this technique is not widely adopted by machine shops. A general guideline for selection of cryogenic machining parameters is crucial in order to promote cryogenic turning to machine shops.

**Funding:** This research received no external funding.

**Acknowledgments:** The author wishes to thank former mentors who led him to the research in cryogenic machining.

**Conflicts of Interest:** The author declares no conflict of interest.

#### **References**


© 2019 by the author. 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 (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Wire Tool Electrode Behavior and Wear under Discharge Pulses**

**Sergey N. Grigoriev 1 , Marina A. Volosova 1 , Anna A. Okunkova 1, \* , Sergey V. Fedorov 1 , Khaled Hamdy 1,2 , Pavel A. Podrabinnik 1 , Petr M. Pivkin 1 , Mikhail P. Kozochkin <sup>1</sup> and Artur N. Porvatov 1**


Received: 16 August 2020; Accepted: 16 September 2020; Published: 20 September 2020

**Abstract:** This work is devoted to researching the tool electrode behavior and wear under discharge pulses at electrical discharge machining. The experiments were conducted on the workpieces of 12Kh18N10T (AISI 321) chrome-nickel anti-corrosion steel and D16 (AA 2024) duralumin by a 0.25-mm-diameter CuZn35 brass tool in a deionized water medium. The developed diagnostic and monitoring mean based on acoustic emission registered the oscillations accompanying machining at 4–8 kHz. The obtained workpiece and non-profiled tool surfaces were investigated by optical and scanning electron microscopy. Calculated volumetric and mass removal rates showed the difference in the character of wear at roughing and finishing. It was shown that interaction between material components in anti-corrosion steel machining had an explosive character between Zn of brass and Ni of steel at a micron level and formed multiple craters of 30–100 µm. The secondary structure and topology of worn tool surfaces were caused by material sublimation, chemical interaction between material components at high heat (10,000 ◦C), explosive deposition of the secondary structure. Acoustic diagnostics adequately registered the character of interaction. The observed phenomena at the submicron level and microstructure of the obtained surfaces provide grounding on the nature of material interactions and electrical erosion wear fundamentals.

**Keywords:** erosion; tool wear; sublimation; ZnNix; explosive deposition

#### **1. Introduction**

The subject of electrical discharge machining (EDM) and wear of a wire tool electrode is not new, but the physical processes that occur during processing are still not sufficiently studied [1–4]. It is related to the absence of the possibility of visual control over the processes occurring in the discharge gap during EDM especially for large workpieces [5,6]. Plenty of studies are devoted to the processes related to the physical phenomena of erosion wear with various conclusions [7–11]. However, at the industrial level, there is no solution to avoid the negative consequences of the accident wire electrode breakage or dumping of the machined part into the working tank at the end of machining. It is especially actual in the case of splitting two co-dependent parts—die and punch for the injection mold, micro-gears [12,13].

An experienced operator often determines the control over processing and process conditions by the specific noise that occurred in the working area. It grows with an increase in the intensity of processing (roughing or finishing) and varies during the electrical discharge machining of materials with uneven structures—porous material, set of tubes, nanocomposites, or composites. The changes in the specific sound are especially noticeable during wire tool penetration into the workpiece and at the end of machining.

The electrical discharge machining of materials occurs in specific conditions between two electrodes. A bias increase is followed by ionizing the space between two electrodes at the moment of their approach. Dielectric breakdown by spark provokes a discharge channel that creates the conditions of low-temperature plasma with 10,000 ◦C that can be observed in particular conditions (forming intermetallics of Al2Cu [14,15], ZnNi<sup>x</sup> [16], CeNi<sup>2</sup> [17], burning titanium at 1200 ◦C [18,19]) in the form of an instant growing gas bubble surrounded the discharge channel. Then, pulses interrupt to restore the breakdown conditions for the next pulse—the bias in the gap, erosion products' washing away from the working area. The occurred conditions are close to the conditions of lightning formation [20]. The temperature in the interelectrode gap achieves high value in a microsecond [21–23].

All monitoring means can be divided into optical and non-optical—electrical and vibroacoustic. The absence of visual contact with the working zone due to its tiny sizes hampers the application of any optical monitoring means. At the same time, the existed monitoring of the electrical parameters does not provide adequate data on the effectiveness of the discharge pulses since for the modern control systems all the produced pulses in the working zone are counted as working ones when a part of them can be spending on the destruction of erosion products [24,25]. It can be grounded by difficulties that met electrical discharge machining in processing materials with threshold conductivity, uneven structure, or microstructure, and inclined surfaces with a thickness of more than 100 mm.

The vibroacoustic monitoring method does not have this kind of disadvantage, as it counts only effective discharge addressed to the material to be processed on the destruction of the surface that was recorded with the help of the accelerometers placed at the working table of the machine [26,27]. It can be an effective means for not only tool behavior investigation and its influence on the quality of the machined surfaces but also an effective means for adaptive control of electrical discharge machining in real manufacturing conditions.

Simultaneously, research of the character of tool electrical erosion wear, sublimation phenomena [28–30], and the nature of the material destruction of the machined surfaces can give additional, comprehensive, and exhausting data.

This paper is aimed at the research of electrical discharge machining by the developed diagnostic means to obtain new data on the influence of wire tool behavior on the quality of the machined surfaces; wire tool wear at roughing and finishing, the nature of material destruction under discharge pulses, and sublimation phenomena.

The research is conducted for two materials:


The work's scientific novelty is in new data on electrical erosion wear of materials, sublimation phenomena, nature and mechanism of material destruction for two types of structural materials, dependencies between detected acoustic emission and machined surface quality, and classification of the eroded surfaces of the tool.

The tasks of the study are:


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

#### *2.1. Equipment*

A four-axis wire electrical discharge machine, Seibu M500S, was used in the experiments for research of wire tool behavior and wear under pulses. The main characteristics of the machines are presented in Table 1.


**Table 1.** Main characteristics of wire electrical discharge machines used in experiments.

The machines are located in a thermo-constant room to reduce ambient temperature's effect on the results of processing. Workpieces were immersed in a dielectric for 10 min before processing to avoid dimensional fluctuations related to the difference in temperatures between the environment and working fluid. The dielectric height was established at 1 ÷ 2 mm above the workpiece. The upper guide of the machine was placed at a minimum distance above 2 ÷ 5 mm from the dielectric [31–33].

The tool electrode is a brass wire with a diameter *d<sup>w</sup>* of 0.25 mm made of CuZn35 (Cu—65%; Zn—35%) with a processing temperature of 260 ◦C and annealing temperature of 425–750 ◦C.

The choice of the electrode type was made since a brass tool of 0.25 mm in diameter is the most widespread for wire electrical discharge machining and suitable for the broad field of applications when the forced choice of any other electrode is due to a severe technological need and is associated with the need to purchase and reinstall expensive nozzles.

It should be noted that the positive polarization of the workpiece and negative polarization of the tool electrode is traditional for the electrical discharge machining. However, modern machine tools can switch the electrodes' orientation for some particular modes or even during machining uneven and hard-to-machine materials in automatic mode.

A CNC program was prepared manually; path offsets were not taken into account. The EDM-factors were chosen using recommendations mentioned in previously conducted works and developed by other scientific groups [25,34–36] (Table 2). The maximum working voltage *V<sup>o</sup>* varied in the range of 40–70 V with a pitch of 10 V for characterization of the discharge pulses by for oscilloscope research and to provoke the conditions of wire breakage for microscopic research. At least five samples and cuts were produced for each group of parameters.


**Table 2.** Electrical discharge machining (EDM) factors during experimental work.

<sup>1</sup> Provided in equivalent unit of the machine.

#### *2.2. Materials*

The chemical composition of 12Kh18N10T (AISI 321) austenite steel is presented in Table 3; the composition of D16 (AA2024, AlCuMg2) duralumin is in Table 4. The thickness of the samples was 20 ± 0.1 mm for both materials.

**Table 3.** Chemical composition of 12Kh18N10T steel (AISI 321) in wt%.


**Table 4.** Chemical composition of D16 alloy (AA2024, AlCuMg2) in wt%.


The high chromium content of the proposed in the research steel ensures the metal's ability to passivate and causes strong corrosion resistance of steel. The addition of nickel converts steel to austenite class. This property is of exceptional importance, allowing to combine the machinability with an expanded set of performance properties. The addition of strong carbide-forming element titanium eliminates the tendency to intergranular corrosion. In turn, carbon forms a refractory titanium carbide and excludes a decrease in the concentration of chromium by chromium carbides formation. It should be noted that the field of chromium-nickel steel applications dominates in the modern rolled metal market [37–40].

Duralumin D16 is a structural alloy mainly used in the aviation and space industries. D16 is rarely used in its pure form since it has less strength and hardness in the non-quenched state. The alloy is classified as a durable thermo-hardened material [41,42].

A Fischer Sigmascope SMP10 instrument (Helmut Fischer GmbH, Sindelfingen, Germany) measured the specific electrical resistance ρ of the materials used in experiments (Table 5, Figure 1a). The device measures the material electric conductance in Siemens and the percentage of the control

sample's electrical conductance produced from annealed bronze in the range of 1 ÷ 112%. All measured values were converted to <sup>Ω</sup>·*mm*<sup>2</sup> *m* . The melting/sublimation points of the materials *T* provided in Figure 1b [43–46]. ⋅ 

**Table 5.** Specific electrical resistance ρ of some materials at +20 ◦C. ρ

ρ **Figure 1.** Electrophysical properties of the materials used in experiments: (**a**) specific electrical resistance ρ at +20 ◦C; (**b**) sublimation/melting point *T*.

σ The stress-strain curves of chosen materials have the presence of elastic and plastic deformation zones [47,48]. The fracture formation schemes have a ductile phase that increases with the material's plasticity—from 12Kh18N10T (AISI 321) steel and brass to D16 (AA2024) alloy. The reduction area for these materials is ~38 ÷ 42 % for steels [47], ~52 ÷ 53 % for brass alloys [49], and ~75 ÷ 77 % for aluminum alloys [50]. Tensile strength σ<sup>B</sup> (UTS) for these materials are 510–830 MPa for 12Kh18N10T (AISI 321) steel, 450 MPa for CuZn35 brass alloy, and 345–420 MPa for D16 (AA2024) alloy.

Reduction area *SRA* is calculated by the following equation [47]:

$$\mathcal{S}\_{RA} = \frac{\mathcal{S}\_0 - \mathcal{S}\_{\min}}{\mathcal{S}\_0} \cdot 100[\%] \,\text{.}\tag{1}$$

where *S*<sup>0</sup> is an original transverse area, mm<sup>2</sup> , and *Smin* is a minimal area of the final neck, mm<sup>2</sup> .

#### *2.3. Monitoring*

A digital oscilloscope TDS2014B (Tektronix, Berkshire, UK) produced the characterization of the discharge pulses.

The vibroacoustic monitoring was provided by piezoelectric accelerometers mounted on the elastic system of the machine [23,25–27,51,52] (Figure 2). The data received by the accelerometers signals were forwarded to preamplifiers, amplifiers VShV003 (JSC Izmeritel, Taganrog, Russia), and an analog-to-digital converter E440 (L-card, Saint-Petersburg, Russia), and recorded with a personal computer (PC). Data were recorded at 1 min, 30 s, and 5 s before the end of processing. Spectral analysis was performed at frequencies 4–8 kHz. The signal was preliminarily cleaned from low-frequency noise using a high-frequency filter. The signal's effective amplitude was determined. The square of this amplitude is proportional to the signal that arises in the machine's elastic system under disturbing influences of the discharge pulses. The cutoff frequency of filters is 2 kHz.

**Figure 2.** Scheme of monitoring sensor setup at electrical discharge machine: (**1**) is a workpiece; (**2**) is a wire tool; (**3**) is a current sensor; (**4**) is accelerometers; (**5**) is a workpiece fastening system; ADC is the analog-digital converter; PC is a personal computer.

#### *2.4. Physical Relationship of EDM Factors and Vibration Amplitude*

Typically, electrical discharge machining has a very narrow range of working factors for machining every material type. Nevertheless, up to 16 factors can be varied during machining [6]. The papers related to EDM research concentrate on some of them as we have done regarding the research subject. It was decided to vary two of the most important factors—operational voltage, which influences the density of discharge pulses distribution, and wire tension, which influences system stiffness and consequently wire oscillation amplitude. The detailed force diagram is presented in [27].

The amplitude of the wire can be presented by summarized force of working impulses in the system's action Σ*Fimp* and stiffness *kn*:

$$A\_{\rm nl} = \frac{\sum F\_{\rm imp}}{k\_{\rm n}}.\tag{2}$$

At the same time, the stiffness of the system is determined by its mass:

$$k\_n = 4\pi^2 \frac{m\_n}{T^2} \tag{3}$$

 = 4<sup>ଶ</sup> ଶ where *T* is a period of self-oscillations; thus, the signal amplitude has an inverse relationship with the weight.

Regarding operational voltage, it has a physical dependence on the signal's amplitude, since it influences the density of the discharge pulses and, consequently, the summarized force factor. The electrical impulse itself is a short-term burst of electrical breakdown voltage and working current that can be presented as follows:

$$\sum F\_{imp} = I \cdot V\_0.\tag{4}$$

The energy of pulses then will be:

Σ

$$\sum E\_{imp} = I \cdot A\_n. \tag{5}$$

∑ =· The wire tension has dependence that is even more evident—*W<sup>t</sup>* influences the system stiffness:

௱

$$k\_n = \frac{F\_e}{\Delta l'} \tag{6}$$

where *Fe* is a restoring force that is opposite and equal to the applied wire tension *W<sup>t</sup>* and ∆*l* is a change in the wire length. Thus:

$$k\_n = \frac{|\mathcal{W}\_l|}{\Delta l}.\tag{7}$$

Moreover, the height of the workpiece also influences the stiffness of the system and wire amplitude:

$$k\_n = \frac{E \cdot S\_0}{l\_n} \,\text{\,\,\,}\tag{8}$$

where *l<sup>n</sup>* is a wire length, *E* is Young's modulus, and *S<sup>0</sup>* is a wire area.

#### *2.5. Characterization of the Samples, Wear Rate, and Discharge Gap*

An EL104 (Mettler Toledo, Columbus, OH, USA) laboratory balance with a measurement range of 0.0001 ÷ 120 g weighed the obtained samples with an error of 0.0001 g.

The samples' surface roughness was controlled by a high-precision profilometer, Hommel Tester T8000 (Jenoptik GmbH, Villingen-Schwenningen, Germany) with a resolution of 1 ÷ 1000 nm and a measurement error of 2%.

An Olympus BX51M instrument (Ryf AG, Grenchen, Switzerland) provided the optical microscopy; the discharge gap was measured optically.

A VEGA 3 LMH instrument (Tescan Brno s.r.o., Brno, The Czech Republic) with magnification up to 1,000,000× provided the scanning electron microscopy and spectrometry of the sample.

The cross-sections were prepared according to the standard probe techniques by an ATM sample equipment—Opal 410, Jade 700, and Saphir 300 (ATM, Haan, the Netherlands). Epoxy resin with quartz sand provided pouring of the samples as a filler was used.

The worn area of the tool can be calculated by the equation of the circle segment area (*Sw*):

$$S\_{\overline{w}} = \frac{1}{2} r\_{\overline{w}}^2 \left(\frac{\pi \cdot \alpha}{180^\circ} - \sin \alpha\right) \left[mm^2\right],\tag{9}$$

where *r<sup>w</sup>* is a wire radius, mm and α is an angle of the segment, degree. The volumetric (*Rv*) and mass wear rates (*Rm*) are calculated by the following equations [53,54]:

$$R\_v = \frac{\Delta V}{t} \Big[mm^3 \cdot s^{-1}\Big] \tag{10}$$

$$R\_m = \frac{\Delta m}{t} \Big[ \mathbf{g} \cdot \mathbf{s}^{-1} \Big] \tag{11}$$

where ∆*V* is volumetric wear, mm<sup>3</sup> , ∆*m* is a worn mass, *g*, and *t* is the wire length wear time; for a thickness of 20 mm, *t* = 0.343 ± 0.005 s.

The discharge gap is calculated by the following equation:

$$
\Delta\_{DB} = \frac{l\_s - d\_w}{2} [mm]\_\prime \tag{12}
$$

where *l<sup>s</sup>* is the measured width of the slot, mm, and *d<sup>w</sup>* is the wire diameter, mm.

#### **3. Results**

#### *3.1. Electrical Discharge Pulses*

Electrical discharge machining occurs at the value of *Vg*, approximately half of the value of operational voltage *V*0. Idle pulses of a particular amplitude and frequency were detected with the value of the interelectrode gap ∆ (distance between two electrodes) more than the value of the gap of dielectric breakdown ∆*DB* (∆ > ∆*DB*), Figure 3a. The idle pulse repetition frequency is *f* = 10 kHz, and the amplitude depends entirely on the factor of the voltage in the interelectrode gap (*V*0) applied to the electrodes when the electrodes are at a distance ∆.

∆ ∆ ∆ ≥ ∆ ∆ ∆ ∆ ≈ ∆ **Figure 3.** Electrical discharge machining pulses detected by the oscilloscope: (**a**) idle pules at a distance ∆ > ∆DB; (**b**) interrupted introduction of operational pulses at the distance ∆ ≥ ∆DB; (**c**) operational pulses during machining ∆ = ∆DB; (**d**) interrupted operational pulses during unstable wire feed and variation with the factors, ∆ ≈ ∆DB.

∆ ≥ ∆ An infrequent dielectric breakdown of the gap occurs with a decrease in distance ∆ ≥ ∆*DB* by a single series of pointed operational pulses (Figure 3b). Regular operational pulses that are similar to damped harmonic oscillations followed the tool penetration (Figure 3c). The working pulses frequency is quite high and is equal to tenths of MHz and detected at a level of about 0.2 MHz. Furthermore, the idle and operational pulses start to alternate with different frequencies and amplitudes that depend on factors present in the stable mode (Figure 3d).

#### *3.2. Wire Electrode Oscilations*

It was determined that the vibroacoustic signal has a periodic character and decreases gradually during the first 2–3 s after tool penetration. It continues to decline during the next 15–20 s of machining (Figure 4a). At the same time, it increases approximately 20 s before the end of the machining. At 5–7 s before the end of machining, the signal interrupts repeatedly. The character of the signal at the beginning and the end of machining has a parabolic character. It corresponds to the final bridge

destruction under the cutoff sample's weight (Figure 4b) that was detected for the samples of both materials with a weight up to 1.8 g.

**Figure 4.** Recorded signal and its root-mean-square value during electrical discharge machining of 12Kh18N10T (AISI 321) samples: (**a**) at 30 s of the tool penetration into the workpiece; (**b**) at 30 s before the end of the processing; (**c**) octave spectra of the root-mean-square value of the signal amplitude of 24.5 g sample, *V*<sup>0</sup> = 65 c.u., *W<sup>t</sup>* = 35 c.u., where (**1**) at 30 s before the end of the operation, (**2**) at 5 s.

(**c**)

Octave spectra of the root-mean-square value of the signal amplitude (RMS) in Figure 4c showed that RMS differs more than 2.5 times at 60 s and 5 s at a frequency band of 4 ÷ 8 kHz. Simultaneously, the frequency band of 0.125 ÷ 4 kHz does not show significant changes during the recorded periods. The changes in RMS were also observed during the variation of EDM factors and unstable processing.

The changes in a signal's amplitude in the frequency band of 4–8 kHz at 60 s and 5 s before the end of processing has a character that mainly increases with the augmentation of operational factors and weight of the cutting-off sample (Figure 5). The changes at RMS of the signal amplitude at a frequency of 8 kHz were noticeable for 12Kh18N10T (AISI 321) at three various values of the operational voltage Vo. However, the changes in average RMS for the samples of ~2 g from D16 (AA2024) alloy are controversial.

The samples' weight was 24.51 ± 0.0327 g and 4.28 ± 0.0450 g for steel and 10.70 ± 0.0375 g and 1.82 ± 0.1800 g for aluminum for a width of 10 and 2 mm correspondingly based on data of 15 steel samples and five aluminum samples of each width.

**Figure 5.** *Cont.*

**Figure 5.** RMS of the recorded signal's amplitude at 8 kHz for two types of materials at various operational voltage *V*<sup>0</sup> : (**a**) 12Kh18N10T (AISI 321) steel, *V*<sup>0</sup> = 55 V; (**b**) D16 (AA2024) alloy, *V*<sup>0</sup> = 55 V; (**c**) 12Kh18N10T (AISI 321) steel, *V*<sup>0</sup> = 60 V; (**d**) D16 (AA2024) alloy, *V*<sup>0</sup> = 60 V; (**e**) 12Kh18N10T (AISI 321) steel, *V*<sup>0</sup> = 65 V; (**f**) D16 (AA2024) alloy, *V*<sup>0</sup> = 65 V.

The changes in RMS of the signal at 60 s and 5 s before the end of processing with the variation in operational voltage *V*<sup>0</sup> and wire tension *W<sup>t</sup>* are more noticeable for 12Kh18N10T (AISI 321) steel than for D16 (AA2024) alloy, that is more ductile (Figure 6). Adequate data were obtained even for tiny pieces with a weight of 3.7 g and 1.8 g correspondingly. It should be noted that stable processing corresponds to the RMS's minimal value at 60 s. RMS of the signal amplitude is higher at 12Kh18N10T (AISI 321) steel machining (Figure 6a,b), by 12.5%, compared to D16 (AA2024) alloy machining (Figure 6c,d). RMS of steel is in the range of 5 ÷ 14 mV<sup>2</sup> with arithmetic mean of 8.54 mV<sup>2</sup> ; RMS of duralumin is in the range of 4.5 ÷ 9 mV<sup>2</sup> with arithmetic mean of 7.475 mV<sup>2</sup> .

(**a**)

*₀* **Figure 6.** *Cont.*


**Figure 6.** *Cont.*

(**d**)

**Figure 6.** RMS of the received electrical discharge machining oscillation signal at 8 kHz for two types of materials: (**a**) 12Kh18N10T (AISI 321), in dependence on operational voltage *V*<sup>0</sup> ; (**b**) 12Kh18N10T (AISI 321), in dependence on wire tension *W<sup>t</sup>* ; (**c**) D16 (AA2024), in dependence on operational voltage *V*<sup>0</sup> ; (**d**) D16 (AA2024), in dependence on wire tension *W<sup>t</sup>* .

#### *3.3. Morphology of the Samples*

− − Roughness profile *Ra* and recorded signal at 30 s before the end of machining are presented in Figure 7. As can be seen, the density of the signal amplitude is higher at 12Kh18N10T (AISI 321) steel machining (Figure 7a,b), by 20%, compared to D16 (AA2024) alloy machining (Figure 7c,d); approximately 30 µm−<sup>1</sup> and ~24 µm−<sup>1</sup> , correspondingly. The three-dimensional (3D) graphs of the EDM factors' influence on the average roughness *Ra* are presented in Figure 8, where minimal value is associated with the stable machining process and the lowest RMS values of the signal's amplitude. − −

**Figure 7.** *Cont.*

**Figure 7.** Roughness *R<sup>a</sup>* and recorded signal amplitude: (**a**) 12Kh18N10T (AISI 321) steel of 24.5 g, *R<sup>a</sup>* of 2.00 µm, *V*<sup>0</sup> = 55 V; *W<sup>t</sup>* = 35 N; (**b**) 12Kh18N10T (AISI 321) steel of 4.3 g, *R<sup>a</sup>* of 1.89 µm, *V*<sup>0</sup> = 55 V; *W<sup>t</sup>* = 35 N; (**c**) D16 (AA2024) alloy of 10.7 g, *R<sup>a</sup>* of 3.04 µm, *V*<sup>0</sup> = 60 V; *W<sup>t</sup>* = 35 N; (**d**) D16 (AA2024) alloy of 1.8 g, *R<sup>a</sup>* of 2.68 µm, *V*<sup>0</sup> = 60 V; *W<sup>t</sup>* = 40 N.

(**a**)

**Figure 8.** Three-dimensional (3D) graphs of the surface roughness *R<sup>a</sup>* dependences on operational voltage *V*<sup>0</sup> and wire tension *W<sup>t</sup>* : (**a**) 12Kh18N10T (AISI 321) steel; (**b**) D16 (AA2024) alloy.

#### *3.4. Discharge Gap*

Figure 9 shows the measured offset ∆\*DB of the path in the dependence of EDM factors for two types of materials. The offset ∆\**DB* includes the wire radius *r<sup>w</sup>* of 0.125 mm. The optically measured effective discharge gap ∆*DB* is in the range of 45 ÷ 53 µm for 12Kh18N10T (AISI 321) steel and in the range of 71 ÷ 78 µm for D16 (AA2024) alloy. The minimal values—170 µm for anti-corrosion steel and 196 µm for aluminum alloy—are associated with the stable machining process and corresponds to the lowest RMS values of the signal's amplitude. Δ Δ

Δ

(**b**)

Δ **Figure 9.** 3D graphs of the offsets ∆\**DB* dependencies on operational voltage *V*<sup>0</sup> and wire tension *W<sup>t</sup>* : (**a**) 12Kh18N10T (AISI 321) steel; (**b**) D16 (AA2024) alloy.

#### *3.5. Tool Wear*

The tool electrode's rupture point (Figure 10) shows cup neck formation before destruction that corresponds to the ductile properties of the brass with the reduction area:

$$S\_{RA} = \frac{0.049 - 0.003}{0.049} \cdot 100 = 93.8 \text{[\%]},\tag{13}$$

(**a**)

**Figure 10.** Wire tool: (**a**) microphotograph, 5×; (**b**) microphotograph, 10×; (**c**) SEM-image, 2.0k×.

Figure 11 presents the electrode wear after roughing and finishing at the electrical discharge machining of 12Kh18N10T (AISI 321) steel and D16 (AA 2024) duralumin.

Figure 11a shows the conjugation of two interdependent surfaces with two types of wear—lateral at left and front at right. The wear has a different character. The front wear surface has the appearance of the typical eroded surface—sublimated and heat-affected material coated by the secondary structure of electrode components with pores and cracks. Moreover, the surface is covered by craters of solidified secondary material—usually consisting of the metallic material of the first order, solid solutions, and complex compounds of the second order (mostly oxides in the case of machining in deionized water). The craters have an explosive character that is not observed at lateral surfaces. The line of two wear surfaces conjugation is pronounced.

Figure 11b shows lateral wear at finishing. The formed surface has visible edges; the conjoined surface's left side has no presence of wear when the right side is also blank but with clear traces of secondary structure explosive deposition at the blank surface. The lateral wear surface showed two types of material destruction—the classical eroded surface of material sublimation with secondary structure deposition and mechanical wear traces.

Figure 11c shows the conjugation of two surfaces—of lateral wear and blank surface at roughing. The left side of the image—blank surface—has pronounced traces—drops, copious splashes—of explosive character of interaction occurred in the discharge gap at lateral wear. A significant volume of uneven sublimated material coated by the secondary structure with pores and cracks presents the surface with lateral wear at the right side of the image.

The front wear surface at roughing (Figure 11d) has secondary structure pellet formation that coat the sublimated surface. The secondary sublimated surface shows typical nanoframe structure formation—more easy-to-melt material components sublimate from the secondary structure's coating (pellets) and are adsorbed by the refractory matrix.

Figure 11e shows the conjugation of two surfaces—front and lateral wear at roughing. The left side of the image—front wear surface—has a coating of secondary structure.

Figure 11f shows the lateral wear's surface at finishing when the obtained surface has traces of two types of wear—thermal and mechanical, which can be easily identified.

The cross-sections of the electrodes at roughing and finishing are presented in Figure 11g,h. Both the cross-sections showed quite intensive wear with affluent loose of the electrode material. The worn area, volume, and mass of the tool, volumetric, and mass wear rates were calculated using Equations (8)–(10) (Tables 6 and 7).

**Figure 11.** *Cont.*

(**c**) (**d**)

(**e**) (**f**)

**Figure 11.** SEM-images of wire tool electrode sample: (**a**) conjugation of front and lateral wear at roughing of 12Kh18N10T (AISI 321) steel, 916×; (**b**) lateral wear at finishing of 12Kh18N10T (AISI 321) steel, 909×; (**c**) conjugation of lateral wear and blank surface at roughing of 12Kh18N10T (AISI 321) steel, 780×; (**d**) front wear at roughing of D16 (AA2024) duralumin, 909×; (**e**) conjugation of front and lateral wear at roughing of D16 (AA2024) duralumin, 1.61k×; (**f**) lateral wear at finishing of D16 (AA2024) duralumin, 2.04k×; (**g**) cross-section after steel roughing, 696×; (**h**) cross-section after steel finishing, 1.02k×.


**Table 6.** Volumetric wear rate of electrical discharge machining of 12Kh18N10T (AISI 321) steel.

<sup>1</sup> Calculated to the entire cross-sectional area; <sup>2</sup> rewinding rate of 3.5 m/min.


<sup>1</sup> Density of 7.9 × 10<sup>3</sup> kg/m<sup>3</sup> or 0.0079 g/mm<sup>3</sup> at +20 ◦C; <sup>2</sup> rewinding rate of 3.5 m/min.

#### *3.6. Chemical Content*

The chemical content of the tool electrode's cross-section at roughing of 12Kh18N10T (AISI 321) steel is presented in Figure 12.

**Figure 12.** Chemical analyses of the worn tool electrode cross-section after machining: (**a**) chemical elements along the line at front wear; (**b**) chemical elements along the line at lateral wear; (**c**) EDX spectrum at front wear; (**d**) EDX spectrum at lateral wear.

#### **4. Discussion**

#### *4.1. Discharge Pulses and Oscillations Control*

Currently, a large number of EDM factors determine the machining mode, the value of which adaptively adjusts by the CNC system during processing. It is called an adaptive pulse-width modulation based on electrical response (Figure 3). At the same time, the value of interelectrode gap, the stability of processing, and consequently the quality of the obtained surfaces depend on the homogeneity of workpiece structure and microstructure, effectiveness of the erosion debris washout by the working fluid, the workpiece thickness, and electrophysical and electrochemical properties of the materials in the working zone. In this connection, the discharge pulses have a more chaotic, probabilistic nature, depending on many factors.

The vibroacoustic signal reflects the changes in the weight and structure of the workpiece, the main discharging factors that correlated with the force diagram in the working zone [27,55] that influence the amplitude of the signal in the wide range of spectra [23,25,26,51,52].

The recorded signal arises during processing and increases by 1.5 times from the initial level at the end of processing (Figure 4). The signal interruptions can be observed at 5 s before the end of processing, which is associated with the direct contact between the workpiece and tool electrode that occurs during changes in the cutting-off sample position in the working space in relation to the rest of the workpiece. It leads to the consequent clamping of the tool electrode to the workpiece by moving the sample and short circuits.

Optimum electrical discharge machining factors have the least value compared to the closest values (Figure 6a–c, except for the duralumin of 2 g). The EDM factors for the stable electrical discharge machining are *V*<sup>0</sup> = 60 V, *W<sup>t</sup>* = 35 N for steel and *V*<sup>0</sup> = 55 V, *W<sup>t</sup>* = 30 N for duralumin. At the same time, the sensibility of the system grows with its stiffness (Figure 6b,d) and decreases with the weight of the sample (Figure 6c,d).

An increase in RMS of the signal amplitude for stable factors at 8 kHz at 5 s before the end of processing was 40 ÷ 55% for steel and 12.5 ÷ 25% for aluminum alloy compared with data recorded at 60 s (Figure 5b,c and Figure 6).

The developed system showed its controversial response for the samples of 2 g of aluminum and adequate data for the samples' weight more than 4 g for the steel and 10 g of aluminum. The samples' weight varies by more than six times, but the RMS of the recorded signal demonstrates similar trends.

The observed behavior of the signal (Figures 5 and 6) can be correlated with the particular features of elastic and plastic deformation during the ductile failure according to the stress-strain curves and scheme of fracture formation (Figure 2) as aluminum shows better ductility during destruction that actually associated with a stretch of the interatomic bonds [56–60]. It correlates with the recorded signal when more brittle material—steel shows an adequate signal response even for the samples of 4 g when data received for ductile aluminum alloy are less significant but can also be registered for monitoring and control of samples of more than 2 g. That all make a basis for the development of multi-parameter control systems and switch to the next technological paradigm [61–67].

The minimum value of the measured roughness *R<sup>a</sup>* of the samples (Figure 8) correlates with stable machining signals. The same tendency is observed for the measured discharge gap (Figure 9).

#### *4.2. Wire Breakage and Tool Wear*

The optical and scanning electron microscopy (Figure 10) showed that the character of rupture had mechanical nature corresponding to cup neck formation and stress-strain curve of middle ductile material destruction—brass alloy [47–49]. The observed area has the topology of the wire breakage that occurs in most of the cases during electrical discharge machining with unstable factors, in case the surface inclination or uneven structure of the workpiece need to be processed. There is no presence of thermal defects except an ashy shade at the formed cup. Additionally, there is no evidence of the rupture's external origin that can occur during the wire cut.

The excess in bias during wire blockage between the workpiece and cutting-off sample that did not allow adequate debris removal probably caused this rupture, since the current and pulses factors were constant. An increase in bias gave a denser distribution of discharges, while an increase in current gave more expressed discharges [28–30] (Figure 13a).

**Figure 13.** Electrical discharge machining principles: (**a**) dependencies of the discharge pulse character on bias and current, where (**1**) is a tool electrode, (**2**) is a workpiece, (**3**) is a discharge channel, (**4**) is a plasma cloud, (**5**) is dielectric medium, and Fe is discharge force; (**b**) submicron structure of erosion wear.

The calculated enlarged value of the reduction area (*SRA* = 93.8%) is obviously caused not only by mechanical rupture but a mechanical rupture in the softened state [68,69] due to the heat of the discharge gap that was definitely above 600–650 ◦C (dark red color) since the brass's surface around the formed cup neck is covered by the ashy shade of zinc oxide (Figure 10a) [70–72].

The formed craters that have different from the typical erosion morphology are 30–100 µm (Figure 11a). The explosive droplets reached a distance of ~100 µm from the wear edge on the backside surface (Figure 11c). The submicron structure of these droplets (Figure 11e) is different from the typical eroded surfaces as on the lateral surface at roughing and at finishing (Figure 11a,b,e,f).

The observation area of the wire tool presented in the SEM-microphotographs showed:


The samples with typical erosion wear traces (Figure 11c,d) correlate to non-oxide (oxygen unsaturated) structures [73,74]—secondary submicrostructures of the complex compounds (of second order) adsorbed by the eroded surface of the base material—of the first order (Figure 14b) [75–77], which probably contain metastable and insoluble solid solution in the form of adherent and brittle thin film and heat-affected sublayer [78–81].

**Figure 14.** Non-profiled tool electrode wear: (**a**) at roughing; (**b**) at finishing, where (**1**) is a tool electrode, (**2**) is a workpiece, (**3**) is machined surface, (**4**) is front wear, (**5**) is lateral wear, and S<sup>g</sup> is a guidance feed of wire.

The mechanical traces on the surface of the wire at lateral wear correspond to the mechanical destruction of the tool during rewinding (Figure 11b,f) that occurred after electrical erosion (secondary wear). Probably, wire tool pinch rollers or diamond nozzle destroyed the morphology of the lateral wear mechanically.

The thermal traces were very pronounced at roughening (front wear), which are different from the lateral wear morphology that was especially apparent at the conjunction of the front and lateral wear surfaces (Figure 11a,e), and the lateral and blank surfaces (Figure 11c) have a different origin, related to the chemical composition of the tool and workpiece.

2 <sup>ଶ</sup> → 2, <sup>ଶ</sup> → ଶ. This is due to the picture of the erosion process observed visually and based on the interaction of the components of the electrodes—CuZn35 brass alloy and 12Kh18N10T (AISI 321) steel. The nickel and zinc reaction at a temperature of 1000 ◦C has an explosive character and results in the formation of intermetallic ZnNi<sup>x</sup> (x = 0, 5, 10, 15, 20 wt%) [82–86]. It can be easily observed by the formation of non-periodic orange flashes in the discharge gap with the release of abundant black sediment during the processing of anti-corrosion austenite (nickel-containing) steels [25,34,87–89]. Visually, the density of the flashes is less than the density of discharges and occurs with a lower frequency. However, the flashes cannot be detected with a higher workpiece, especially with a height of more than 100 mm that often occurred at machining in tool and mold production, when the visual access to the working zone is absent. Thus, the signal was adequately registered by the developed vibroacoustic diagnostic mean—RMS of the amplitude signal was higher by 12.5% and more intense by 20% for 12Kh18N10T (AISI 321) steel than for D16 (AA 2024) alloy for the thickness of the sample of 20 mm.

As can be seen, the deposition of the secondary submicron structure of the sublimated electrodes' components and working medium in the case of anti-corrosion steel processing occurred explosively (with craters of 30 ÷ 100 µm). The presence of the explosive character of reaction between metals accompanies the electrical erosion wear can be seen in Figure 11c on the blank surface when a clear edge limits the area of deep EDM wear on the lateral surface during roughing. However, the explosive nature of the secondary phase deposition overcomes the wear edge and is visible from the electrode's blank side.

The front wear at roughing has a more pronounced topology that correlates with the non-oxidized erosion wear an explosive reaction between wire and workpiece components (Figure 14), where the deposed film of secondary structure coat eroded base metal surface. The lateral wear at roughing has a less pronounced topology that corresponds to the typical wear that occurred under discharge pulses. It correlated to the degree of the involvement of the sides of the electrode in the formation of the slot when the front surface has the presence of secondary wear of the formed films: the front side is more involved in the formation of the slot, and the side surfaces are involved in the erosion process only partly by secondary "polishing" formed surfaces [24,90,91]. The lateral wear at finishing has a similar character. However, wire tool pinch rollers destroyed the morphology of the lateral wear mechanically.

The electrode's cross-section shows the intensity of the two types of wear (Figure 11g,h). In the considered sample, the front wear does not predominate the lateral one at roughing, and distributes quite even at the periphery of the tool (Figure 11g). The conjugation of the worn surfaces was pronounced for all of the samples.

Analysis of chemical elements along the line and EDX spectrum of the wire tool at roughing and finishing (Figure 12) showed mostly chemical elements except for chemical elements of the brass wire in balance—61.8 ÷ 64.3% of Cu and 34.8 ÷ 35.5% of Zn. However, less than 3.4% of oxygen is proof of semiconductive and amphoteric zinc oxide formation, which usually occurs during brass heating (Figure 10a) [92,93], when copper (II) oxide decomposes in the presence of hydrogen [94,95]:

$$\text{2ZnO} + \text{O}\_2 \rightarrow \text{2ZnO}\_2\tag{14}$$

$$\text{CuO} + \text{H}\_2 \rightarrow \text{Cu} + \text{H}\_2\text{O}.\tag{15}$$

Both of the oxides do not interact with water. Zinc oxide gets yellow with heating and sublimates at 1800 ◦C. It should be noted that that oxygen was present quantitatively more in the samples after finishing and at later wear of roughing, while it was not possible to quantify it along the line in some cases at front wear after roughing. A small amount of carbon that was not quantitatively evaluated (less than 0.2%) is associated with normal atmospheric contamination.

#### **5. Conclusions**

#### *5.1. Monitoring System and Tool Behavior*

A comprehensive study of the tool electrode's wear process during electrical discharge machining was accomplished by the developed monitoring system based on oscillation detecting. That gives detailed data on the character of electrode tool wear and stability of workpiece machining in the high-frequency acoustic band of 8 kHz.

The optimum electrical discharge machining factors are detected by monitoring the vibroacoustic signal—RMS value of the amplitude at 8 kHz for steels and more ductile duralumin with a weight of more than 2 g. The stable electrical discharge machining are *V*<sup>0</sup> = 60 V, *W<sup>t</sup>* = 35 N for steel and *V*<sup>0</sup> = 55 V, *W<sup>t</sup>* = 30 N for duralumin. An increase in RMS of the signal amplitude at 5 s before the end of processing was 40 ÷ 55% for steel and 12.5 ÷ 25% for aluminum alloy compared with data recorded at 60 s. The proposed approach can be used to develop a multiparameter controlling system of EDM-equipment to carry out the modern CNC-systems at a principally new level.

#### *5.2. Wire Tool Topology and Wear Rate*

Classification of the obtained surface topology of the tool electrode determines two types of wear under discharge pulses related to the thermal nature: material sublimation and chemical interaction between components of the working zone when mechanical destruction of the finishing electrode sample has a different origin.

Volumetric wear rate *R<sup>v</sup>* was 1.22 ± 0.04 mm<sup>3</sup> ·s <sup>−</sup><sup>1</sup> at roughing and 0.52 ± 0.002 mm<sup>3</sup> ·s <sup>−</sup><sup>1</sup> at finishing; mass wear rate *Rm*—9.6 × 10−<sup>3</sup> ± 0.01 g·s <sup>−</sup><sup>1</sup> and 4.0 × 10−<sup>3</sup> ± 0.008 g·s −1 , respectively. 41 ÷ 62% of the tool subjected wear under discharge impulses at roughing during electrical discharge machining of anti-corrosion steel when summarized lateral wear exceed front wear by 29.17%. 12 ÷ 24% of the tool sublimates under lateral wear at finishing.

The study showed that the processing of the materials with inadequate process parameters or the not proper combination of tool and workpiece materials causes more intensive wear of the tool correlated with the chemical interaction of the electrodes and dielectric medium components. This leads to the micro explosive character of processing with formation intermetallic ZnNix (x = 0, 5, 10, 15, 20 wt%), with Zn of the brass and nickel of austenite steel that was also registered the mean of vibroacoustic diagnostic. The crater diameter was of 30 ÷ 100 µm; RMS of the amplitude signal was higher by 12.5% and more intense by 20% for 12Kh18N10T (AISI 321) steel than for D16 (AA 2024) alloy.

#### *5.3. Further Procpects amd Paractical Significance of the Work*

As was shown, the amplitude is up to 55% higher for steel and up to 25% higher for duralumin at convenient machining than 5 s before the end of processing that always stays critical for precision cutting, especially in the conditions of tool production—profiled cutters, hot channels, and injection molds. The obtained data were for the thickness of 20 mm when it stays one of the most often used thickness for EDM workpieces in tool production. The developed system proved its reliability for the samples up to 2 g when the standard sample weight for discharge gap and machining mode verifying is 15.6 g for steels and 5.4 g for aluminum for a sample of 10 × 10 mm in the plan with a thickness of 20 mm.

The tool wear under electrical discharge pulses has a complex character related to the thermal type of wear with a heat-affected sublayer, and the upper layer consisted of a secondary structure formed from the components of electrodes with the traces of chemical reactions at a heat of 10,000 ◦C. Thus, electrical discharge machining wear forms in the following stages:


The explosive character of interaction between Zn and Ni should be considered while designing experiments and electrical discharge machining of chrome-nickel anti-corrosion steels. For high-precision and nano-works, machining of nickel-containing steels should be provided by a tool with no Zn in its content—copper, steel, or tungsten wire have a few disadvantages due to the softness of copper, the relatively low electrical conductivity of steels, and heat-resistance of tungsten. However, it is a promising direction for further research.

The obtained knowledge has a fundamental character and can be used as a recommendation for the industrial applications on the choice of the electrode tool material and searching the optimum EDM-factors; in this context, not only structural requirements are addressed for the working and auxiliary surfaces of the final product, but also functionality in the exploitation conditions.

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

**Funding:** This project has received funding from the Ministry of Education and Science of the Russian Federation within the framework of the state task for scientific research, under Grant Agreement No. 0707-2020-0025.

**Acknowledgments:** The research was done at the Department of High-Efficiency Processing Technologies of MSTU Stankin.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **Nomenclature**


#### **References**


© 2020 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 (http://creativecommons.org/licenses/by/4.0/).

## *Article* **On Adaptive Control for Electrical Discharge Machining Using Vibroacoustic Emission**

#### **Yury A. Melnik , Mikhail P. Kozochkin, Artur N. Porvatov and Anna A. Okunkova \***

Department of High-efficiency Machining Technologies, Moscow State University of Technology STANKIN, Vadkovskiy per. 3A, 127055 Moscow, Russia; yu.melnik@stankin.ru (Y.A.M.); m.kozochkin@stankin.ru (M.P.K.); vto@stankin.ru (A.N.P.)

**\*** Correspondence: a.okunkova@stankin.ru; Tel.: +7-909-913-12-07

Received: 16 October 2018; Accepted: 23 October 2018; Published: 24 October 2018

**Abstract:** The article is related to the research of the parameters of vibroacoustic emission for development of the monitoring and adaptive control system for electrical discharge machining. The classical control system based on a response of electrical parameters does not give an adequate data in the cases of a new class of materials processing as conductive ceramics reinforced by conductive nano additives and carbon nanotubes and whiskers. The idle pulses, which are working on the destruction of the erosion products in the gap, count as working pulses. The application of the monitoring and control tools based on vibroacoustic emission gives adequate data about conditions in the working zone. The developed system is available to count only impulses involved in working on the destruction of the workpiece. The experiments were conducted on the samples of materials with a low melting point as austenitic steel and aluminum alloy, and hard alloys. The records of vibroacoustic signals were analyzed for detection of the monitoring and adaptive control criteria.

**Keywords:** electrical discharge machining; vibroacoustic emission; adaptive control; monitoring; discharge gap; erosion products

#### **1. Introduction**

Today it is impossible to imagine the production of high-accuracy parts with complex volumetric geometry and internal cooling system without electrical discharge machining (EDM) [1–3]. The application of the technology was significantly extended by adding from 1 to 3 independent rotating axes [4–6].

Hence, practice shows that up to now EDM has uncontrollable problems related to decreasing of productivity and quality of machined surfaces during conventional processing of easy-to-work materials [7–9]. Mainly it is critical in the case of machining of narrow slots array. E.g., 20 ÷ 24 slots with the sizes in the plan 1.5 × 8 mm, and height of 15 ÷ 25 mm for production of the mold die with push-type ejectors for the part as a body of motor vehicle backlight [10]. In this case, a wire tool can be stuck between a workpiece and a cut-out industrial waste (or part), that leads to the undesirable issues related to process instability, short circuit, wire break, sometimes even to diamond nozzle damage [11–14].

Nowadays, the use of ceramic products in the industry is continuously growing due to its excellent exploitation properties [15,16]. Plenty of scientists aim their study on the development of a new class of materials as conductive ceramic nanocomposites [17–19]. This kind of material should provide mechanical properties up to the level of continent ceramics or even exceed them, but also be suitable for electrical discharge machining. By other words, it should be conductive [20]. The conductivity should be assured by conductive nano-sized additives as TiC, ZrC, Nb(2)C [21–23], when SiCw or carbon in the form of nanotubes or whiskers are responsible for supporting and improving mechanical properties [24,25]. It makes the questions related to ensuring of the uninterrupted electrical discharge machining using in-situ monitoring critical and significant to adapt the electrical parameters of machining and to prevent all undesirable effects during processing.

The tools of diagnostics based on vibroacoustic emission are known for application to the conventional mechanical machining for providing better reliability of the machine tools [26,27]. The vibroacoustic diagnostic tools aim to detect the moment of beating between a tool and a workpiece [28–30]. It allows controlling the operating element movements, preventing accidents. It was proved that based on the vibroacoustic signal it is possible to develop a new class of multiparametric diagnostic and adaptive control system [31]. The significant advantage of vibroacoustic diagnostics is the simplicity of the accelerometers' installation on the elastic system of the machine.

The scientific novelty of the article is in development and application diagnostic system based on vibroacoustic emission to the electrical discharge machining methods, which are known by an impossibility for the visual monitoring tools. The tasks of the current study are:


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

The experiments were carried out at an industrial 4-axis EDM machine with CNC-controller Seibu M500S (Seibu, Fukuoka, Japan). A 0.25 mm-diameter brass wire of CuZn35 was used as a tool. Deionized water was used as a dielectric medium. Electrical discharge machining was produced with the immersion of the workpiece.

The accelerometers (Figure 1) of the developed vibroacoustic monitoring systems were installed at the upper guide (5) of the machine and on the system of the fastening of the workpiece (3). It was found that during electrical discharge machining the upper accelerometer (7) gave a more informative and symmetric signal than the accelerometer from the lower guide (8). It can be explain by presence of the disturbance influence from the working dielectric pump and driving gears of the machine. Further, only the signal from the upper guide was taken into account.

**Figure 1.** The positions of accelerometers in the working area of EDM machine, where (1) is a workpiece, (2) is a worktable, (3) is a fastening system, (4) is a wire electrode, (5) is an upper guide, (6) is a lower guide, (7) is an upper accelerometer, (8) is a lower accelerometer.

A 200 × 20 × 16 mm-blank of austenitic stainless steel AISI 321 [32,33] and a 200 × 16 × 16 mm-blank of aluminum alloy AISI 2024 [34–36] samples (Table 1) were used at the first stage of the experiments for developing of the system of diagnostics and monitoring based on vibroacoustic emission for the needs of electrical discharge machining. The blanks were fastened on the worktable of EDM machine tool.


**Table 1.** The chemical composition of AISI 321 steel and AISI 2024 alloy, %.

The processing was done with the immersion of the blanks in a dielectric for 10 min before machining; it was done according to the EDM standards for high precision processing to exclude the thermal shrinkage of the materials [37]. The level of the dielectric was 1–2 mm above the blanks. The nozzle of the upper guide was installed as close as possible to the level of dielectric [38].

2- and 10 mm-width samples were cut off from the blanks. During the processing, offset of the pass on the value of the discharge gap and the radius of the wire was ignored. The EDM parameters of processing were calculated following the standard machine tool recommendations for the proposed materials. The parameter of wire tension and voltage were varied during processing in the range of ±5 machine measurement units from the recommended values of the mentioned parameters.

The developed system of diagnostic and monitoring was approbated on EDM machine CUT1000 (GF AgieCharmilles SA, Losone, Switzerland). For this purpose, 100 × 12 × 12 mm-blanks of hard alloys M05 [39,40] and P10 [41,42] by ISO (Table 2) were processed according to the prepared manually CNC-program. The standard EDM parameters for SKD-61 alloy [43] were chosen for machining. The frequencies of pulses changed during processing to create conditions for instability and provocation of a series of short circuits and wire breaks.


**Table 2.** The chemical composition of M05 and P10 hard alloys, %.

The developed in-situ monitoring system based on vibroacoustic emission for electrical discharge machining is presented in Figure 2. The signals received from the accelerometers depend on the nature and intensity of the vibrations during processing, which depends on the parameters of electrical discharge machining. The signals were directed to an ADC and recorded. It was decided to take attention on the signal at the moment wire tool penetration into the blank, and at the end of processing: on 1st minute, 30th and 5th seconds before the end of CNC-program. Spectral analysis was performed in the range of frequencies from 1 to 32 kHz.

**Figure 2.** Principal scheme of the system of vibroacoustic monitoring of wire electrical discharge machining: (**a**) is a guide, (**b**) is accelerometers, (**c**) is a wire-tool, (**d**) is a workpiece, (**e**) is a working table, (**f**) is a fastening system, (**g**) is preamplifiers, (**h**) is amplifiers model VShV003 (OOO Izmeritel, Taganrog, Russia), (**i**) is an ADC E440 (L-card, Saint-Petersburg, Russia), (**j**) is a signal-recording device.

The pulsed solid-state laser of diode pumping U15 (RMI, Moscow, Russia) was used for laser ablation of AISI 321 samples (160 × 100 × 5 mm) for better understanding the possibilities of in-situ monitoring and adaptive control during pulsed machining processes.

The characterization of electrical discharge impulses was conducted with the use of the digital oscillograph Tektronix TDS2014B (Tektronix Inc., Beaverton, OR, USA).

The characterization of the machined surfaces and the specific wear of the tool was provided by an SEM VEGA 3 LMH (Tescan, Brno, The Czech Republic) and by an optical microscope Olympus BX51M (RYF AG, Grenchen, Switzerland). For the study, the surfaces of workpieces and a wire, and the changes of chemical composition in the sublayer were controlled before and after processing [10,33].

#### **3. Results**

#### *3.1. Research A Fundamental Possibility of Vibroacoustic Diagnostics of Wire Electrical Discharge Machining*

#### 3.1.1. Mathematical Approach and Evaluation of Wire Amplitude Under Discharge Impulses

The scientists of MSTU Stankin under the supervision of Prof M.P. Kozochkin developed the in-situ monitoring system based on vibroacoustic emission [44,45] and adopted it for the needs of electrical discharge machining (Figures 1 and 2). The preliminary study showed that the oscillations of the wire in the slot during processing might be a reason of instability and turn into the self-oscillation process, which results in the defects of the machined surfaces [46–48].

A diagram of applied forces on the wire during electrical discharge machining presented in Figure 3.

If not taken into attention to the mass of the wire sections *h*<sup>1</sup> and *h*<sup>2</sup> and mass losses on electroerosion wear [49,50], an amplitude of oscillation may be presented as a complex amplitude of the harmonic signal [51–53]:

$$
\overline{A}\_{\mathfrak{n}} = \overline{A}\_0 \cdot e^{\overline{\beta}\tau},
\tag{1}
$$

where *A*<sup>0</sup> is an amplitude in the direction of the wire feed, which is always limited by a discharge gap <sup>∆</sup> and much less than <sup>∆</sup>, *<sup>A</sup><sup>n</sup>* ≤ *<sup>A</sup>*<sup>0</sup> for stable EDM processing; *<sup>β</sup>* is the damping coefficient presented as a complex number in the form (*a + bi*); *τ* is a period of oscillation *T*.

**Figure 3.** A diagram of applied forces during electrical discharge machining: (1) is a wire tool, (2) is a workpiece, (3) is an upper nozzle, (4) is a lower nozzle, (5) is rollers.

Δ Σ Σ Σ *A<sup>n</sup>* is an amplitude of the wire in the inverse direction to the wire feed; ∆ is a discharge gap; *H<sup>n</sup>* is a thickness of workpiece; *Hn'* is a distance between nozzles; *F<sup>g</sup>* is total forces of wire guiding; Σ*Fimp* is total forces of discharge impulses; Σ*F<sup>h</sup>* is total forces of wire holding; Σ*F<sup>w</sup>* is total forces of wire tension; *Ffric* is a friction forces; *S<sup>g</sup>* is a wire feed speed; *S<sup>w</sup>* is a wire rewinding speed; *W<sup>s</sup>* is a torsional moment of rewinding rollers.

$$
\overline{\beta} = \overline{q} - \overline{\mu},
\tag{2}
$$

̅= തെµത ̅ where *q* is an index of excited oscillations; *µ* is a coefficient of dielectric medium resistance. In this case, *q* is

$$
\overline{\eta} = \sqrt{\frac{\sqrt{\overline{\eta}\_{\!\!\!} - \!\!\!F}}{2 \cdot m\_n}},
\tag{3}
$$

ത = ට ଶ∙ *̅* ̅ where *h* is a ratio between *H<sup>n</sup>* and *H*0, *H*<sup>0</sup> is a value of the workpiece thickness for stable electrical discharge machining 80 ÷ 100 mm for the diameter of wire *d<sup>w</sup>* = 0.25 mm and dielectric based on deionized water; *m<sup>n</sup>* is a wire mass; *k* is a ratio between *K<sup>n</sup>* and *K*0, *K<sup>n</sup>* is an index of stiffness and can be presented as:

$$K\_n = \Sigma F\_{imp} / A\_n. \tag{4}$$

Σ *β*̅ Thus, *β* depends on the ratio of workpiece thickness and the ratio of the system stiffness, the EDM parameters and the viscosity of the working medium; it can vary in the limits |1|.

*q* is a criterion of self-oscillation:

̅

̅

$$
\overline{q} = \frac{\lambda}{T} \tag{5}
$$

ത = <sup>ఒ</sup> ் *λ* where *λ* is the logarithmic decrement of the damping ratio of the self-oscillatory and resonant process, describing the decrease in the amplitude of the oscillation process and equal to the natural logarithm of the ratio of two successive amplitudes of the oscillating quantity *A<sup>n</sup>* to the same side:

*λ*

$$
\lambda = \ln(A\_n / A\_{n+1}) \,\tag{6}
$$

The period of self-oscillations *T* depends on the wire mass and the system stiffness, which depends as well on the wire tension:

$$T = 2\pi \sqrt{\frac{m\_n}{K\_n}}\tag{7}$$

If *q* < *µ*, then self- and resonant oscillations do not arise. If *q* > *µ*, then the self-oscillations are formed more intense and higher as q excess over *µ*. *µ* is higher for the oil medium and lower for the water-based dielectric. Thus, electrical discharge machining in the oil medium gives a priori higher accuracy of the machined surface.

Simplified calculations based on the measurement of the frequency of forced impulses in the working area gives a possibility to evaluate the wire amplitude during the processing with varied wire tension (Table 3).


**Table 3.** Results of simplified mathematics for evaluation of wire amplitude under discharge impulses.

<sup>1</sup> Given for reference; <sup>2</sup> Further ignored due to its small value; <sup>3</sup> Taken as an approximate value for evaluation of the amplitude for the first pass of electrical discharge machining of the same height workpiece; <sup>4</sup> Measured in absolute values during experiments.

#### 3.1.2. Experiments on the Fundamental Possibility of Vibroacoustic Diagnostics

During the first stage of the experiments, it was established that miserable changes in the weight of the part influence on the vibroacoustic signal. As an example, a 24.5 g-weight of 10 mm-length AISI 321 sample shows a significant increase of the vibroacoustic signal amplitude at 5 s before the end of processing. For a 4.28 g-weight of 10 mm-length AISI 2024 sample, it was recorded at 2 s before the end of processing. It was noticed in the full frequency range of the signal. As well, it was characterized by instability in the low frequencies of the signal. Thus, it was chosen to use the frequency range higher than 4 kHz for vibroacoustic monitoring and analyzing.

Figure 4 presents a high-frequency spectrum of vibroacoustic signals recorded at 60 and 5 s before the part separation. The octave spectra are shown in the inset of the graph. The effective amplitude in the octave-frequency band is more stable than high-frequency spectra of the vibroacoustic signals due to averaging. The effective amplitude with approaching of the end of processing (the moment of separation of the part) increases approximately in 2 times for the octave-frequency band 4 kHz and in 1.5 times for the octave-frequency band 8 kHz. So as it can be seen, the occurred moment can be evaluated timely and distantly. The changes can be presented by visual demonstration, which may be understandable for any EDM operator.

**Figure 4.** High-frequency spectra of a vibroacoustic signal for electrical discharge machining of AISI 321: (1) is for electrical discharge machining at 60 s before the end of the operation; (2) is for electrical discharge machining at 5 s before the end; the inset shows the octave spectra.

Figure 5 presents two spectra of vibroacoustic signals, which were obtained during the electrical discharge machining of the AISI 2024 sample. Spectrum (1) shows the moment of wire penetration, and spectrum (2) shows the moment when the discharge gap is already formed. The changes in root-mean-square value (RMS) of the vibroacoustic signal at the duration of the moment of wire penetration into the workpiece are shown on the inset of the graph.

**Figure 5.** High-frequency spectra of vibroacoustic signals for AISI 2024: (1) is at the moment of wire penetration; (2) at the moment of the discharge gap formation; RMS-t diagram inset presents the spectra of the signal at the moment of wire penetration.

The experimental research of the influence of the wire tool tension *F<sup>w</sup>* on vibroacoustic signal showed that decrease of the hardness of elements of the technological system [1,56,57] influences negatively on the growth of the vibration. The growth of vibration was detected with the decline tension force. The detected vibrations were also associated with a decrease of flushing ability of the erosion products from the working zone, which influent negatively on the quality of the machined surfaces.

#### *3.2. Adaption of the In-Situ Monitoring Method for the Needs of Electrical Discharge Machining*

The optimization of any machining technology means to ensure its maximum productivity, efficiency, and quality of the machined surfaces up to the required level. It demands the in-time regulation of EDM parameters as the discharge gap ∆, the concentration of erosion product in the gap, the temperature of the working fluid, and its flushing rate. The gap ∆ is the primary parameter, which is responsible for determining the quality of the final product [58–60] A small increase in ∆ may change the conditions in the working area and interrupt the discharge. A decrease in ∆ impairs the yield of erosion products, reduces the productivity, increase accumulation of slag, and provoke short circuits. Electrical discharge machining cannot be effective without automatic regulation of the gap [61,62]. For regulation of the optimal value of the gap, the rate of particle formation *M<sup>p</sup>* in the discharge gap should be equal to the rate of particles leaving the gap *Mex*. The rate *M<sup>p</sup>* is a function of the concentration of the particles *γ*: Δ Δ Δ Δ

$$M\_p = f(\gamma). \tag{8}$$

When *M<sup>p</sup>* and *Mex* are unequal, the change in concentration of particles is *γ*

*γ*

$$
\Delta \gamma = \Delta M \cdot \Delta t / Q\_{\prime} \tag{9}
$$

where *Q* is the volume of the discharge gap, and Δ*γ* Δ Δ

$$
\Delta M = M\_p - M\_{\text{ex}}.\tag{10}
$$

For stable electrical discharge machining ∆*γ* = 0. Δ*γ*

However, it is complicated to ensure constant *γ* due to the many factors acting in the working area. Any fluctuations should be timely eliminated by control signals, which change the parameters of electrical discharge machining. The analysis indicates that the maximum rate *M<sup>p</sup>* decrease as *Mex* decreases. It occurs due to the deterioration of the erosion products' evacuation as wire penetrates in the workpiece, and the number of working pulses reduces [58]. *γ*

Figure 6 presents the dependences of the machining rate *Mp*, the number of pulses n on the discharge gap ∆, where *n<sup>w</sup>* is for working pulses, *nid* is for idling pulses, *nsc* is for short-circuit pulses. Analysis indicates that the gap corresponding to the maximum rate *Mmax* is higher than the gap corresponding to the maximum rate of working pulses as an excess of erosion products at the maximum rate of working pulses and short-circuiting pulses destabilize processing [61]. Δ

Δ **Figure 6.** Dependence of the rate *M<sup>p</sup>* and the number of pulses *n* on the discharge gap ∆.

Thus, the in-situ monitoring systems for electrical discharge machining aim to maintain the efficiency of pulse utilization, which is the ratio of the number of working pulses (*nw*) to the total number *n* of pulses:

$$
\psi = \frac{\mathfrak{n}\_{\overline{w}}}{n} = 0.7 \div 0.9.\tag{11}
$$

The rate *ψ* can be informative for in-situ monitoring and adaptive control, but its use is complicated by the inertia of the required measurement instruments. *ψ*

Hence, there are difficulties in assessing the efficiency of electrical discharge machining concerning *ψ* [33,44]. The energy of the individual pulses is not entirely consumed in the processing of material sublimation when the dielectric medium is contaminated with erosion product. In this case, a part of the energy is consumed in the destruction of the erosion products. Since the number of working pulses is assessed from the total number of the discharge impulses, it results in an imprecise assessment of the electrical discharge machining efficiency. The precise estimate can be obtained by relating the efficiency to the ratio of the useful energy consumption for sublimation of the material and the total incoming energy of the working impulses in the discharge gap. Normally, the energy of the discharge impulses is proportional to the effective discharge current (*Ie*) of independent generators in operation. The use of a current sensor can estimate this factor. *ψ*

#### *3.3. Analysis of the Obtained Data, Search for the Criteria for the Development of the Adaptive Control System Based on Vibroacoustic Emission*

The results of vibroacoustic monitoring of another non-contact precise machining were considered for better-understanding impulse character of electrical discharge machining and possibility to improve its productivity. Laser ablation is close by its nature to the processes of material removal during electrical discharge machining as in both of the cases there are no mechanical contacts, the material removal occurs under thermal influences, initiated by pulses of concentrated energy fluxes. The material removal during electrical discharge machining is related to the complex processes of chaotic material sublimation under discharge impulses [63] as after laser ablation the surface presents the organized wells [64]. The nature of the initial processes is different.

As it was previously determined [61], the vibroacoustic signals coming in the time of processing demonstrate steady increase as the laser power increases and the volume of removed metal increases (Figure 7).

**Figure 7.** Dependencies of the effective amplitude of the vibroacoustic signal (**a**) and productivity (**b**) from laser power *N*: *A<sup>v</sup>* is a vibroacoustic signal amplitude in wide frequency range; *A<sup>v</sup>* <sup>16</sup> is an amplitudes of the signal of 16 kHz octave.

The presented in Figure 7b dependence has a monotonous character. It does not demonstrate any significant material removal at low values of laser power, that is related to lack of thermal energy for

initiation of material ablation. It can be approximated with linear dependence or exponential function with specified accuracy for the industrial needs.

Figure 7a shows the correlation between vibroacoustic signal amplitude *Ava* and productivity of the pulses *M<sup>p</sup>* after data processing. It can be approximated by an empirical function:

$$A\_{\rm cu} = \zeta \cdot M\_p{}^{\Lambda}{}\_{\prime} \tag{12}$$

where <sup>Λ</sup> = 0.72 ÷ 0.76, <sup>ζ</sup> <sup>=</sup> *const*.

It is highly possible that the vibroacoustic signal of electrical discharge machining has the same dependences on processing parameters and instead of (8) is easier to use:

$$A\_{\rm vn} = f(\Delta, \gamma). \tag{13}$$

There is a proved possibility of realization of the in-situ monitoring and adaptive control measures based on vibroacoustic emission for the needs of electrical discharge machining. The parameter of vibroacoustic signals can be observed without any difficulties in the distinction of current pulses productivity.

The dynamic system of electrical discharge machining is linear with dynamic characteristics depending on the state of the dielectric medium. The amplitude spectrum of *U*(*f*) signal received by the accelerometer is defined according to the dynamic system linear properties [44,61]

$$
\Delta L(f) = B\_1(f) \cdot B\_2(f) \cdot \Theta(f),
\tag{14}
$$

where *B*1(*f*) is the frequency response function (FRF) of the operational environment with changes of accumulation of the erosion products, the distance between electrodes, the temperature in the discharge gap (the temperature in the discharge channel is about 7000 K, the temperature of dielectric medium is about 293 K); *B*2(*f*) is FRF of a dynamic influence of the discharge impulses on the elastic system; Θ(*f*) is an amplitude spectrum, *f* is the frequency.

The energy of vibroacoustic signals changes in the dependence on the concentrated energy flux intensity on the machining surface. A part of energy flow is spent on repeated sublimation of erosion products in the discharge gap. As a result, it crates chaotic and overlapped wells on machined, it chanes as well the conditions in the discharge gap. The formed overlaps cause localization of the following discharge. Thus, the impulses created by phase transfer of the processed metal are longer in time.

Therefore, the concentration of erosion products' *γ* increase as short impulses decrease and long impulses increase. The transformations result in changes of dynamic characteristic *B*1(*f*) including changes in FRF of *B*(*f*)*:*

$$B(f) = B\_1(f) \cdot B\_2(f) \tag{15}$$

The current values *B*(*f*) during electrical discharge machining can be evaluated by controlling discharging current and vibroacoustic signal. The experiment showed that *B*2(*f*) might be presented as constant in the short period of observation. Thus, the changes in *B*(*f*) are related to the changes in dielectric medium conditions *B*1(*f*).

#### *3.4. Demonstration of the Opportunity for The In-situ Adaptive Control of Edm by Specific Examples*

The developed in-situ monitoring technique was approbated on the high-precision wire electrical discharge machine CUT1000 (GF AgieCharmilles SA, Losone, Switzerland) during processing of the samples made of hard alloys, which can be actual for the modern machine building industry: M05 and P10, ISO. The vibroacoustic signals and discharge current were recorded during experiments. FRF of *B*(*f*) in the observation channel was fixed at the moment of wire approach to a workpiece and the moment wire break. Figure 8 presents an FRF of *B*(*f*) in the channel of observation. The vibroacoustic signals were recorded in [mV], FRF values were non-dimensional.

**Figure 8.** The record of FRF during electrical discharge machining of M05 hard alloy: (**a**) is in the low-frequency range; (**b**) is in the high-frequency range; (1) is at the moment of wire approach; (2) is before the wire break.

During the period of the initial stage of wire approach (1), the dielectric is clean of erosion products, the wire approaches to the flat surface of the workpiece, the erosion products easily flush from the discharge gap. The highest value of amplitude fixed at the frequency range 11 ÷ 13 kHz.

The wire break occurs at 12 s of electrical discharge machining (2). There were observed a decrease in high-frequency components and an increase in low-frequency components at the frequency range 2.4 ÷ 2.6 kHz. The flushing of erosion products was straitened before the wire break. The short impulses decreased, and long impulses increased for forming of the groove.

It was noticed that FRF changes in the full frequency range are not convenient due to possible disturbance from the work of the equipment. One or two frequency ranges where FRF changes can be easily monitored should be chosen for the development of the adaptive control method.

Figure 9 presents an example of RMS vibroacoustic signal change in various octave bands: from the wire approach to breaking.

**Figure 9.** RMS of vibroacoustic signals during 12 s of electrical discharge machining of M05 hard alloy: (**a**) RMS of vibroacoustic signal; (**b**) proportional ratio of RMS1/RMS2, where (1) is an octave band of 2 kHz; (2) is an octave band of 32 kHz; (3) is the result of interconnection.

The power of discharging current has not any significant changes during changes of RMS of the vibroacoustic signal; hence, the components of the vibroacoustic signal demonstrates the tendency of changes in the conditions of the discharge gap. A gradual increase in RMS of low-frequency vibroacoustic signal (1) and a significant decrease in the RMS of the high-frequency signal (2) was recorded during machining. Graph 3 shows the change in the ratio of RMS of a low-frequency signal to RMS of a high-frequency signal. The ratio gives a more informative picture of a decrease in the conditions of electrical discharge machining. The stable phase of processing was observed until 7 s and then an increase in 10 times of RMS of a vibroacoustic signal from 3 ÷ 4 s was observed; it achieves briefly 15-fold value.

As the control system of the machine reacted only to indirect electrical parameters, in particular, to a coefficient of impulse use, the critical increase of the concentration of the erosion products was left without proper attention. It resulted in overheating of electrodes and consequent wire break.

The exceptional use of data in the electric format of impulses control system results in consideration of the impulses directed to the sublimation of the erosion products in the discharge gap as effective. At the same time, the vibroacoustic signals from such impulses cannot be created and detected at all, or they have a narrow range of frequency spectra.

Figure 10 shows the examples of vibroacoustic signal impulses of various shapes at various spark gaps. It can be seen the discharges current impulses in the form of vibroacoustic signals of less than 0.1 ms in length (a). The length of separate vibroacoustic signals' fragments at the high concentration of erosion products increases up to 5–6 times (b). The long fragments (impulses) provide an increase of low-frequency elements of the spectra of a vibroacoustic signal.

**Figure 10.** Vibroacoustic signals at 10 ms period during electrical discharge machining of M05 hard alloy: (**a**) is at the moment of wire approach; (**b**) is before wire break.

#### **4. Discussion**

In the begging, it was decided to focus attention at the moment of separation of the part from the workpiece, which gives changes in the specific noise accompanied electrical discharge machining. By this vibroacoustic emission, it is possible to recognize a specific noise related to each stage of plastic deformation and brittle fracture of the part separation.

It was shown that an amplitude of the vibroacoustic signal increases gradually with the approach of the moment of the separation (Figure 4). Then the series of peaks of the diagram related to the possible wire stuck in the discharge gap and consequent short-circuiting. It proves the final bridge is weakening due to rapid mass loss under electroerosion pulses. Then the part starts rotated motion relatively lower contact point between a part and a workpiece when the upper part of the bridge is separated. The wire tool stays clasped between a separated part and the rest of the workpiece. The series of the short circuits occur, and the vertical traces of the wire can be observed on the samples of the material with the electrical resistance more than 7.9 <sup>×</sup> <sup>10</sup>−<sup>8</sup> <sup>Ω</sup>·m.

<sup>−</sup> Ω During the experiments, it was noticed that the intensity of vibroacoustic signals increases as the symmetry of the wire tool position in the gap is disturbed. It is related to the prevalence of the discharge current pulses acting on one of the sides of the wire tool. This kind of phenomena was also detected at the moment of wire tool penetration into a workpiece when there is no stable discharge gap between the electrodes. At the moment when the discharge gap is established, and a wire tool penetrated the workpiece and formed a discharge groove, the acting pulses balance each other and the amplitude of the vibroacoustic signal decreases.

It can be detected from the obtained vibroacoustic diagram (Figure 5) of the spectra that at the 5 ÷ 7 s the RMS value of the spectra increases four times in frequency range of 8 kHz. It is noticeable even for alloys of transition metals with the relatively low-melting-point (more than 231.9 ◦C and less than 950 ◦C) and low electrical resistance (2.8 <sup>×</sup> <sup>10</sup>−<sup>8</sup> <sup>Ω</sup>·m). The described changes of the vibroacoustic signal are thoroughly enough for developing in-situ monitoring and adaptive control system for taking measures timely to prevent short-circuiting and consequent defects of the machined surfaces.

The absolute excess of the vibroacoustic amplitude during conventional machining after penetration of the tool in the workpiece can be explained by an excess of the erosion products (solidified particles of the eroded material) in the discharge gap. The excess of the erosion products usually related to the insufficient flushing in the working zone due to the relatively high height of the workpiece (more than 70 ÷ 100 mm for the wire tool 0.25 mm) or relatively low pressure of dielectric in the nozzles (code "1" instead of "2" for the first tool pass, rough cutting). It may also be related to the relatively small discharge gap (less than 0.005 mm) for the materials with the high electrical resistance (more than 1.12 <sup>×</sup> <sup>10</sup>−<sup>6</sup> <sup>Ω</sup>·m) due to the difficulties in controlling it by the servo drives of the machine tool.

Thus, it may be possible that:


Therefore, it was shown that the non-contact electrical discharge machining generates the vibroacoustic signal in the full frequency range that is quite similar by its character to the character of vibroacoustic emission during the convenient mechanical machining. It was proved that the signal might be registered by accelerometers placed on the elastic system of the machine tool wirelessly and at a distance from a working zone to exclude the influence on accuracy positioning of drivers and the parameters of machining. It was determined that the vibroacoustic signal is a result of wire tool disturbance in the discharge gap by discharge current pulses, cavitation processes and intensity of flushing in the working medium, wire tool contacts with the workpiece or erosion products.

The positions of the accelerometers were chosen based on the idea that the accelerometers should be placed as close as possible to the working zone. However, the installation of the lower accelerometer on the lower guide was inconvenient as it is moving under a workpiece and the wires of the accelerometers can hamper the processing. Therefore, it was more suitable to place it closer on the fastening system of the workpiece for monitoring of workpiece vibration. The upper accelerometer was placed on the upper guide for monitoring of wire vibration, and it was most informative and convenient for development of the adaptive control system.

It was found during the study, the upper accelerometer records the vibrations which are associated with the total forces of the working discharge impulses, which are directed in the destruction of the workpiece (Σ*Fimp.w*), but not in the destruction of the erosion products in the discharge gap (Σ*Fimp.idle*). In this case, the total forces of initiated impulses, which can be detected by monitoring only electrical parameters, is:

$$
\sum F\_{imp} = \sum F\_{imp,w} + \sum F\_{imp,idle}.\tag{16}
$$

It should be noted that not all working impulses carry out the same and useful work. Some of them spent a part of the energy or even all its energy on the destruction of the erosion products [65–67]. The experiments showed that the performance of the EDM (volume in mm<sup>3</sup> ) is related to the power of the vibroacoustic signal by a linear dependence (or close to linear) (Figure 7). Besides, it follows that the control of the share of working pulses at high frequencies is associated with great difficulties in the field of circuit engineering and with large errors as it is known from published sources [55,58,59]. It is the advantage of the vibroacoustic signal over the method of controlling the pulse utilization rate (*ψ*).

In other words, during discharge gap breakdown all initiated pulses are considered to be working pulses by measuring electrical parameters, which is not correct because a part of the pulses is spent on the destruction of the erosion products in the discharge gap. It is suitable for the cases when there is a low part of the erosion products in the discharge gap. In the case of vibroacoustic monitoring, only the pulses directed on the destruction of another electrode are taken into account as working impulses.

Moreover, the measurement of the vibroacoustic signal in [mV] is suitable as the amplitude of the vibroacoustic signal is proportional to the signal in [mV] at the output of the measuring channel. In principle, the measurements in [mV] can be converted to the measurement of vibration acceleration [m/s<sup>2</sup> ], but it is unnecessary work called "calibration." For the adaptive control system development, it is enough to have measurement units, which are proportional to amplitude.

The widespread use of vibroacoustic spectra for diagnostics and monitoring is complicated as it is a closed system with a nonlinear dependence of the vibroacoustic signal on the impulse load [68,69]. The dynamic system resembles better a linear model, where the dynamic relation between the load source and the workpiece is permanent.

In other words, the model of the dynamic system is significantly simplified, and its use in the in-situ monitoring and regulation of machining by high-energy fluxes is simplified [70,71].

#### **5. Conclusions**

The application of the developed in-situ monitoring technique using vibroacoustic spectra of the signal in the working zone can provide a current solution for the specific issues related to the insufficient productivity of electrical discharge machining and low quality of the machined surfaces due to technological issues of processing (e.g., wire breakage).

The developed method is based on the effect of the vibroacoustic emission during processing. Physical phenomena of vibroacoustic emission explained the fundamental possibility of this method. It was shown the interconnections between the electrical parameters of machining and vibroacoustic spectra, and the possibility to timely analyze the received vibroacoustic spectra data for development of the system of the adaptive control.

For further development, it is necessary to conduct more intensive research with the high electrical resistant materials to show the principle possibilities to adapt the electrical parameters of the processing for the needs of processing of new classes materials as nanocomposites based on ceramics and made with the use of carbon tubes and whiskers. It was shown that the developed method is suitable for operative positioning of wire tool-electrode or evaluation of wire tool-electrode bending during processing. That can be extremely important for the production of the parts with a complex linear configuration for the needs of the aircraft industry.

The study of processing of workpieces with high energy impulses shows that their effectiveness and parameters of vibroacoustic signals depend on the power of impulses supplied with monotonous analogic dependencies. It allows monitoring of the current effectiveness of flushing in the working zone and optimizes the value of the discharge gap.

The unstable electrical discharge processing is also associated with multiple contacts of electrodes and short-circuiting during processing that cannot be identified timely by modern CNC-system, but this problem can be solved with the adequate data received by monitoring system based on vibroacoustic emission. Analysis of vibroacoustic signals can provide a modern solution for resolving the manufacturing issues, which is not covered by existing methods of electrical parameters control and helps in forming a multi-parameter and multifunctional diagnostic system.

## **6. Patents**


**Author Contributions:** Conceptualization: A.A.O. and M.P.K.; methodology: M.P.K. and A.N.P.; software, A.N.P.; validation: M.P.K., A.N.P., and Y.A.M.; formal analysis: M.P.K.; investigation: A.N.P.; resources: Y.A.M.; data curation: A.A.O.; writing—original draft preparation: M.P.K. and A.N.P.; writing—review and editing: M.P.K. and A.A.O.; visualization: A.A.O. and A.N.P.; supervision: Y.A.M.; project administration: Y.A.M. and M.P.K.; funding acquisition: A.A.O. and Y.A.M.

**Funding:** This research was funded by the Ministry of Education and Science of Russian Federation, grant number No. 9.7453.2017/6.7.

**Acknowledgments:** The research was done at the Laboratories of the Department of High-efficiency Machining Technologies of MSTU Stankin.

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

#### **References**


© 2018 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 (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Influence of WC-Based Pin Tool Profile on Microstructure and Mechanical Properties of AA1100 FSW Welds**

#### **Abbas Tamadon 1, \* , Arvand Baghestani <sup>2</sup> and Mohammad Ebrahim Bajgholi 3,**†


Received: 4 April 2020; Accepted: 5 June 2020; Published: 15 June 2020

**Abstract:** The effect of various tungsten carbide (WC) pin tools and operating parameters on the material structure and properties of an AA1100 friction stir welding (FSW) weld were evaluated. Three different pin shapes were employed (conical, square and threaded). For each tool shape, welds were generated for a set of tool (revolutions per minute, RPM) (710, 1120 and 1400) and advancing speeds (150, 250 and 400 mm/min). Weld samples were tested for mechanical strength by tensile testing. Morphology was examined using optical microscopy, and weld composition with scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy (EDS) and X-ray diffraction (XRD). No weld contamination from the tools was observed. However, a number of structural defects, inherent to the FSW process, were observed (including tunnel voids, kissing bonds and swirling lines). These defects, associated with the stirring action, could not be eliminated. The results show how the operating parameters may be optimized to produce stronger welds.

**Keywords:** friction stir welding; WC; AA1100 aluminium plate; weld contamination; tunnel void; kissing bond

#### **1. Introduction**

Friction stir welding (FSW) is a solid-phase process [1] carried out at lower temperatures than conventional fusion welding [2]. This joining technique was introduced by The Welding Institute (TWI) [3], originally for aluminium alloys. The welding action is conducted by a non-consumable rotating tool, consisting of a penetrating pin at the centre supported by the tool shoulder [4]. The advancement of the tool through the weld-line results in severe plastic deformation [3,5,6], inducing frictional heating in the base material [7,8]. The stirring action mixes the softened material from both sides of the weld-line (advancing side (AS) and retreating side (RS)). The mixing is the result of both forging and extrusion forces underneath and around the tool [9–13]. As the rotating tool leaves the stirring position, material is deposited behind the tool, forming a continuous weld-line [13,14].

FSW can be used for a variety of ferrous and non-ferrous metallic materials [15–17], where solidification-related problems common to fusion welding can be eliminated from the weld structure. Mechanical properties may also be improved. In general, the severe deformation inherent to FSW causes grain fragmentation in the stir zone (SZ) and produces a fine-grained microstructure as the major region of the FSW weld [9,18]. In immediate proximity to the stir zone is the transition zone

where the microstructure differs from the base material, due to dynamic recrystallization (DRX) caused by thermal flux and mechanical strain. These transitional regions are called the thermo-mechanically affected zone (TMAZ) and heat affected zone (HAZ) [19–21]. During cooling, microstructural evolution eventually leads to grain refinement and precipitation, which increases the mechanical strength of the weld [9].

The thermomechanical nature of the process induces DRX, which causes some grain-scale alterations within the weld texture. This microstructural evolution has considerable effects on the final properties of the FSW weld [20].

Depending on the welding parameters (welding speeds and the tool geometry), the frictional heat input and the subsequent flow-based stirring action can establish a relatively complex strain distribution which is stored in the fragmented grain structure [22,23]. By the beginning of the cooling stage, DRX leads to a variety of microscopic transformations throughout the polycrystalline weld structure [24,25].

These thermomechanical alterations need to be accurately studied at a grain-structure level to characterize the origin of microscopic transformations (e.g., grain refinement, morphological alterations and precipitation). However, the FSW process can introduce structural weld defects, such as micro-voids or micro-cracks due to improper welding parameters [26,27]. It is expected that the strength of the weld should reach the same as the parent metal [28] given the homogenous DRX microstructure in the SZ. However the mechanical strength of the weld can be easily affected by the welding speeds—advancing speed (*V*) and rotating speed (ω). Additionally, the geometry of the pin tool has an impact on the weld quality [13,29], it influencing the heating of the substrate and transportation of the plasticised mass flow during the stirring action. The pin tool is subject to large forces and at elevated temperatures may undergo abrasion, possibly contaminating the weld [30,31].

In metal forming processes, the surface quality can be changed during the plastic deformation. In this regards, by using some proper surface finishing procedures (such as burnishing), the high spots of the surface can be flattened by a smooth hard tool. The ball-burnishing [32] is a secondary process that is applied after the FSW to smooth the surface of the welding nugget by a radial feed and exerting a sufficient compressive force. Consequently, the blanked edge of the weld surface is smoothed out by a certain magnitude of compressible residual stress distribution. As a result, it improves the tensile strength or bending strength of the weld in the springback phenomenon.

The inherent formability of aluminium guarantees the applicability of FSW to Al alloys. Deformable non-ferrous materials, such as aluminium, are generally processed using tool steels for the pin-tools. Tool steels provide enough friction and thermal strength for the FSW process, however weld contamination is common. Contaminants displaced from the FSW tool can be trapped into the weld region causing impurity or void formation [33], both of which deteriorate the expected strength properties of the weld. These issues can be worse for FSW of reactive materials such as aluminium. Therefore, enhanced wear resistance and stability at elevated temperatures is required to promote the tool lifetime. Tungsten carbide (WC), which has high hardness (1650 HV), is a promising candidate. However, the performance of WC in FSW conditions [34–36], and its feasibility for FSW, is unknown.

The research objective of this study is to investigate the effect of a WC pin tool on the microstructure and mechanical properties of FSW AA1100 aluminium alloy under different welding conditions. Additionally, the impact of various FSW operating parameters, including tool advancing speed, rotational speed and tool geometry, were assessed. AA1100 aluminium is a commercial aluminium alloy metallurgically stabilized with a standard element composition to attain specific mechanical properties. Weld contamination may be easily detected due to the chemical purity. The alloy was chosen as the first material to test as it is easily deformable, and hence is a good candidate for FSW. If unsatisfactory results are obtained for this alloy, the method will not be suitable for other grades. Three tool geometries were used to generate weld samples: conical, threaded and square, manufactured by powder metallurgy [24,37]. Although WC-based tools are usually used in processing of high strength and high-temperature alloys (e.g., steels or super-alloys), this research was designed to

investigate the possibility of the tool material loss and contamination of the weld region arising from the tool material.

Different tool shapes can provide different mechanical stirring and engagement between the tool and substrate, allowing evaluation of the mechanical instability of the WC tool under the high-friction conditions of the stirring action. By applying different speeds (spindle rotational speed and advancing feed rate), the mechanical performance of the tool can be assessed better.

The thickness of the workpiece and the spindle load were kept constant. The samples were tested for mechanical strength, microscopic structure and weld contamination. Metallographic measurements and mechanical strength can be used as a reliable method to evaluate the optimum microscopic features.

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

FSW welding trials were conducted on AA1100 alloy plates with a thickness of 5 mm. The composition of the AA1100 aluminium plate (utilised as the parent metal in this work) is listed in Table 1, measured using an atomic absorption analysis quantometer (AA-6300, Shimadzu Corporation, Kyoto, Japan).


**Table 1.** Composition of the AA1100 Al alloy.

Three single-piece tools with different pin geometry designs but similar configuration and dimensions were utilised for the weld trials. With a constant shoulder diameter of 20 mm, pins were manufactured to three different geometries; conical, square and threaded (Figure 1). The pin tools were made of WC, by powder metallurgy sintering [38]. The composition of the powder mixture utilized was WC—0.3 wt % C—6 wt % Co. The free-carbon powder was used to prevent porosity, and to avoid formation of oxide phases (e.g., WO and WO3) during the sintering process. Cobalt (Co) was employed as the binder phase to improve the wettability on micro-size WC particles, therefore promoting interfacial bonding strength and wear resistance of the segment. Sintering was done at 1800 ◦C for 90 min under a pressure of 70 MPa. The shoulders of the FSW tools were fabricated from H13 hot work tool steel (hardness 560 HV), a standardised grade suitable for hot working. For each tool, the pin and shoulder were joined by the transient liquid phase diffusion bonding (TLPDB) technique.

The weldment plates were cut into two plates with dimensions of 235 mm in length and 45 mm in width and set side-by-side for a butt weld. For this soft aluminium grade, initial trials showed that lower speeds could not provide a stable welding condition, while higher speeds led to poor weld quality with structural defects (e.g., flash, voids or discontinuities) occurring throughout the weld. Many welds were conducted, but the speed parameters (ω and V) were limited to the ones with better results in preliminary tests. Five feed rate speeds in the range of 100–400 mm/min and eight rotational speeds in the range of 600–1500 RPM (revolutions per minute) were applied to produce the FSW weldments. While some complete welds were produced, there was also a high proportion of only partially successful welds, hence indicating a quality problem. The issue was invariably a partially successful joint with a localised defect (discontinuity line or tunnel void defect). We aimed to better

understand the causes of the quality issues for different pin tool profiles. Therefore, after some trial and error, the feed rates and rotational speeds were optimized to those given in Table 2.

ω

**Figure 1.** Configuration of the FSW tools utilized for the AA1100 weld trials ((**a**) conical, (**b**) square and (**c**) threaded pin tools) with geometry and dimension.

**Table 2.** The operational welding parameters of the friction stir welding (FSW) process for the AA1100 plates.


The aluminium welds were conducted using a 3-axis Computer Numerical Control (CNC) machine with a control unit and motor capacity of 10 kW. The direction of tool rotation was clockwise, and the plates were fixed rigidly by strap clamps at the corners. Clockwise rotation of the tool enables feeding of material into the centre of the weld during rotation (Figure 2). The welding was done without any initial pre-heating, and in an open air atmosphere at a temperature of 18 ◦C. Table 2 presents more details of the process parameters employed for the AA1100 aluminium weld trials. A schematic of the FSW set-up and process is shown in Figure 2. After the test, tensile test samples and metallographic specimens were cut from the middle of the plate, perpendicular to the weld-seam.

To evaluate weld quality, samples were primarily examined by the visual inspection. Weld surfaces were photographed using a high-resolution bridge digital camera (FinePix S9500 Model, Fujifilm brand, Tokyo, Japan), under the light of a fluorescent lamp. For metallographic analysis, mounted samples of the weld cross-sections were polished by standard preparation methods down to 0.5 µm to provide a smooth mirror finish. For microstructural observations, the cross-section samples were etched to identify grain boundaries and morphology. The specimens were etched in an ultrasonic bath using the Keller's reagents (95 mL H2O, 2.5 mL HNO3, 1.5 mL HCl and 1.0 mL HF), at 70 ◦C for 20 s. After etching, the specimens were washed with deionised water, rinsed with ethanol and dried with hot air. The macro- and microscopic observations of the etched cross-sections were done with a typical optical microscope, at magnifications of 50-1000×.

μ

ω

**Figure 2.** Schematic of the FSW process, the FSW tool in interaction with the AA1100 plates as the substrate.

For further microstructural characterisation, after etching samples were examined in a scanning electron microscope (SEM, MIRA model, TESCAN instrument, Brno, Czech Republic) with an energy-dispersive spectroscopy (EDS) detector. The accelerating voltage of the SEM was 20 kV. For phase characterization via X-ray diffraction (XRD), a Philips diffractometer instrument was used with Cu-Kα radiation at angles of 20–110◦ , and an accelerating voltage of 40 kV. α

The tensile test samples, with a dimension of 120 mm gauge length and 30 mm gauge width (Figure 3), were fabricated according to the ASTM E8/E8M standard method. For FSW joints, the sampling position was selected perpendicular to the welding direction. In this configuration, the weld-seam is located in the centre of the dog-bone tensile specimen. Therefore, during the tensile testing, the strength of both the base metal and cross-weld are measured simultaneously [39,40]. Tensile tests were performed with an Instron-5967 tensile tester at a strain rate of 10−<sup>5</sup> s −1 , at room temperature. For each sample, we repeated the tensile test five times and then reported the average value between the three most consistent samples as the selected tensile curve. − −

**Figure 3.** Schematic dimensions of the tensile test sample, made of FSW plates of AA1100.

#### **3. Results**

#### *3.1. General Appearances of the FSW Weld*

*ω* The general appearance of the FSW weld, along with typical structural defects formed during the process, is shown in Figures 4–6. In these figures, the weld appearance varied depending on welding speeds (ω and *V*) and tool features. It is believed that the heat input plays a key role in determining weld appearance, and insufficient heat input can produce defects in the stir zone. Figure 4 shows the crown view at the top surface of the FSW joints welded by the conical-shape tool, at a constant rotating speed (1120 rpm) and two different feed rates (150 and 315 mm/min). The first set of speeds (1120 rpm and 150 mm/min) in Figure 4a shows a uniform pitched surface without any visible defects at the surface. By increasing the feed rate at the same rotating speed (1120 rpm and 315 mm/min), the weld surface became more smooth, while the ripple features disappear and a cavity/groove defect emerges on the advancing side (AS) of the weld surface. Additionally, for both set of speeds, flash defects are visible on the retreating side (RS) edge of the weld-line.

**Figure 4.** Surface appearance of the weld for conical FSW tool in constant feeding rate; (**a**) at the welding speeds of (1120 rpm and 150 mm/min) and (**b**) at the welding speeds of (1120 rpm and 315 mm/min).

*ω* For a better understanding of the surface appearance of the weld, different sets of speeds were studied with the same conical-shape tool, where the feed rate was kept constant and the RPM varied. Figure 5 shows the surface view of the FSW joint for two different speed sets with the same RPM. For the welding speeds set of (250 rpm and 710 mm/min), Figure 5a shows a relatively smooth surface, compared to Figure 5b at a higher feeding rate (250 rpm and 1400 mm/min). Similar to Figure 4, again flash defects were obvious in Figure 5a,b, but the size of the flash defects and the apparent surface roughness increased. In Figure 5, the keyhole position where the tool exits the workpiece was also visible. Clearly, by increasing the feed rate (Figure 5b) the size of the keyhole and the flash defect increased. Our interpretation is that inconsistency in welding speeds (ω and *V*) can intensify the slipping during the stirring action. Therefore, excessive circumferential inertia affects the flow distribution during the stirring, leading to interruption of the flow deposition.

ω The role of the tool geometry on the quality of the weld was studied in Figure 6, where the general appearance of the weld-seam is shown for the conical, square and threaded designed tool. Welding speeds (ω and V) remained constant. As is clear in Figure 6, for the same processing speeds (1120 rpm and 315 mm/min), the weld-line provided by the conical tool shows the best surface quality: a uniform pitched ripple pattern for the weld crest, a filled keyhole and minimal flash defects.

**Figure 5.** Surface appearance of the weld for conical FSW tool at constant RPM; (**a**) at welding speeds of (250 rpm and 710 mm/min) and (**b**) at welding speeds of (250 rpm and 1400 mm/min).

**Figure 6.** Surface appearance of the weld for different FSW tool in constant speed sets of (250 mm/min and 1400 rpm); (**a**) conical pin profile, (**b**) square pin profile and (**c**) threaded pin profile.

The performance of the conical tool can be attributed to the inherent lateral motion of the mass around the pin, which pumps material inwards and simultaneously in a spiralling manner to the top surface. This can avoid mass deficit or ejection as the tool moves forward along the welding locus. Therefore, material loss and discontinuity defects are minimized.

It should be noted that in the FSW process, the material flow plays a key role in the control of defects. In general, insufficient heat input can lead to failure of flow regimes in which the stirred mass is not strong enough to fill the discontinuity, and therefore, a defect occurs.

On the other hand, redundant heat input also can result in the emergence of defects. By increasing welding speeds (ω and V), excessive frictional heat is generated between the tool and the workpiece. The increased heat input can yield more plasticized mass under the shoulder. At higher speeds, this plasticized material flow can split out in the form of the excessive flash defects at the edge of the weld-line. Furthermore, the ejection of the mass results in mass deficit at the weld-seam, where groove defects appear underneath the shoulder, or other subsurface discontinuity defects form in weld.

The relationships between processing parameters and the apparent quality of aluminium welds are illustrated in Figure 5. The photographs, both taken from the top surface of the weld samples at the spindle side position, show a distinctive ejected tail as the flash defect, protruding from the retreating side (RS) of the weld-line. The exit point of the tool shows a keyhole feature where the tool leaves the workpiece. In both samples, the tool leaves the weld-seam by a disruption in the body of the workpiece.

The curved features at the weld surface reveal layered flow patterns that trace the plastic flow between the AS and RS. However, due to disruption at the sides of the weld, the circulation of material and therefore the primary stirring flow-lines are discontinuous.

As the front of the stirred mass around the pin loses stiffness because of the shortage of forwarding material, flow circulation becomes irregular. Thus due to the upward motion of the tool, the plasticized material bulges out and a hole is left at the position of ejection.

In a similar way, where the leading edge of the shoulder approaches the free surface of the workpiece, the stirred material at the retreating side of the tool creates a curve in the surface pattern of the weld-line. During rotation of the tool, the inconsistency in speed ratio (ω/V) can cause excessive stretching at the weld edge, more so than at the middle of the weld-line, whereby the compressed mass can be shredded outwards instead of being deposited at the trailing edge of the tool. This can form a flash sprayed tail as a defect at the RS position of the weld-line.

In Figure 6, constant welding speeds (250 mm/min and 1400 rpm) for different FSW tool geometry (conical, square and threaded shaped pins) show a similarity in the size of the flash defect and the keyhole. The weld samples indicate a similar amount of material loss as the flash defect, and similar shape of the keyhole at the surface. This suggests that the tool geometry plays the main role in the stirring conditions in proximity to the rotating tool.

The surface quality of the weld-seam shows a more uniform pitch-pattern for the conical pin tool (see Figure 6a) compared to the square pin and threaded pin (Figure 6b,c, respectively). This indicates that in similar welding speeds, the presence of the flats (in square pin tool) and the threads (in threaded pin tool) aggregates the plastic flow deposition within the stirring zone, in which needs a more accurate flow control during the FSW position. It should be noted that because of applying a rotating pin, the keyhole profiles could not show the exact flow geometry around the pin. Therefore, the rotating pin leaves a circular pattern at the keyhole by ejection from the weld-line. More specifically regarding the square pin, because of the sharp orthogonal edges of the flats, during the tool ejection, it might cause some material loss from the stirred Al-mass, stuck around the pin (see Figure 1b). Nevertheless, the WC-based square pin tool shows an acceptable mechanical instability during the stirring action without occurring any contamination or tool failure.

The surface features of the weld-line for different welding parameters were illustrated in Figures 4–6. However, the flow aspects of the FSW joints concerning various welding conditions need to be elucidated in more detail through microstructural study of the weld cross-sections. Microstructural observations

can also reveal more flow features within the weld structure, where internal defects arise from flow failure within the stir zone.

#### *3.2. Macrostructure of the Weld Region*

The macrostructure of the cross-section of an FSW joint, processed using three different tools, is shown in Figure 7. The macrostructure readily revealed the SZ basin-shaped pattern in the cross-section of the weld, and onion-ring flow patterns were also visible. In Figure 7c, compared to Figure 7a,b, the weld border was more visible in the middle of SZ rather than at the sides of the weld. Even so, in Figure 7b there is a blurring of the weld border. This is attributed to the tendency for the tool to aggressively remove the substrate material from the base metal (BM) and subsequently backfill the region with heavily worked material, hence making the SZ borders more defined in comparison to the sides.

**Figure 7.** The macrostructure of the cross-section of the FSW weld for three different tools: (**a**) conical, (**b**) square and (**c**) threaded; all in speeds of 250 mm/min and 1400 rpm.

The flow eddy features known as onion rings at the bottom of the SZ were attributed to the mass transport mechanism inside the weld. The tool scouring action during stirring caused the plastically deformed material from the substrate to be aggressively conducted in flow path lines between the advancing side and retreating side. Consequently, because the combined linear and rotation motions of the tool were greater at the bottom surface, this is also the hottest part of the weld region. However, the material was also scoured from the leading edge towards the retreating side, and was deposited at the trailing edge of the tool. The rotation of the tool (clockwise from above) stirred the flowing mass in the same rotational direction, corresponding to a horizontal flow from right to left in the wake of the tool. Consequently, during mass deposition, the stirred flow-lines were packed into the downstream region, and the boundaries between these layered masses were believed to correspond to the onion rings in the cross-section samples shown in Figure 7a–c. The spacing between the boundary lines was also observed to be smaller closer to the centre of the pin, attributed to the flow mechanism during stirring. Additional complexity was introduced by the threads on the tool (Figure 7c). This transports the material downward, hence explaining why the onion rings were preferentially located at the bottom of the SZ.

Inconsistency in flow-lines transport can cause the emergence of internal void defects. As the macroscopic views of the FSW samples were not able to show the grain structure in detail, micro-etching and microscopic observation at higher magnifications is required. The delineation of fine grain size in proximity to probable defects can also explain the origin of defects based on the microstructural evolution of the weld region.

#### *3.3. Microscopic Measurements of the Weld*

The micro-etching of the weld cross-section in Figure 8a shows that the weld region comprised of the SZ and the transition region, which were distinct from the base metal (BM) region. The transition region, including the HAZ and TMAZ, was situated between BM and the SZ. The dynamically recrystallized nature of the FSW weld structure caused each layer to have a different grain size and morphology in comparison with the adjacent layers. μ

**Figure 8.** Microstructure of the grain distribution in the cross-section of the AA1100 FSW weld (processed at 1120 rpm and 250 mm/min); (**a**) general view of the cross-section, (**b**) stir zone (SZ) and (**c**) the base metal at higher magnification.

While the BM shows a directional grain morphology with the average grain size of 50 µm (Figure 8b), this changed to an equiaxed, ultrafine structure with an average grain size of 10–15 µm in the SZ (Figure 8c). This can be interpreted as the direct outcome of grain fragmentation by mechanical stirring, and subsequent post-welding dynamic recrystallization arising from frictional heat generation.

Entering into the SZ, the morphological flow features through the grain structure revealed some inhomogeneous transitional micro-patterns in the form of weld defects. Figures 9–11 demonstrated some of the typical weld defects in the weld region that emerge at the bottom of the SZ adjacent to the base metal (BM).

**Figure 9.** Micro-void features visible in stirring zone of the FSW weld, known as a tunnel void, observed in: (**a**) conical tool and (**b**) square tool welds.

**Figure 10.** Formation of flow-based defects around tunnel voids formed by different FSW tools at constant speed sets of (250 mm/min and 1400 rpm); (**a**) conical pin profile, (**b**) square pin profile and (**c**) threaded pin profile.

**Figure 11.** Formation of the flow-based defects at the structure of the FSW weld; (**a**) kissing bond by conical pin profile and (**b**) swirling zone by threaded pin profile, in speed sets of (250 mm/min and 1400 rpm).

Figure 9 shows the flow patterns around tunnel voids at the cross-section of the FSW weld for the conical and the square tool geometries.

The drawn flow-lines around the tunnel voids show a layered structure representing different gradients in the deposited flow at the BM regions.

Comparison between the samples shows that while the number of voids in the weld sample from the conical-shaped tool (Figure 9a) was greater than for the square-shaped tool (Figure 9b), the size of the voids were greater when processed using the square tool. This can be related to the coarse performance of the square-shaped tool and the sharpness of the square corners compared to the conical geometry.

Further study of the micro-features around tunnel voids is shown in Figure 10. While all three different tool geometries produce some tunnel voids within the stirring zone, the conical (Figure 10a) and threaded (Figure 10c) pin profiles exhibited more complex flow patterns around the tunnel voids compared to the square-shaped pin profile (Figure 10b). The swirling zone in Figure 10a for the conical-shaped pin profile and the kissing bond defect in Figure 10c for the threaded-pin were the direct outcome of flow failure around the formed tunnel void. These flow-based defects could be attributed to the complexities of the stirring flow mechanisms driven by the pin geometry.

Figure 10 also revealed that the swirling zone was situated at the top surface, and the kissing bond had propagated from the bottom surface. Microscopic observation of the weld cross-section at a higher magnification revealed these flow-based features as the kissing bond by conical pin profile (Figure 11a), and the swirling zone by threaded pin profile (Figure 11b). In both cases, the microstructural patterns show a change in the grain direction from the weld surface towards the body of the weld in the ultrafine structure in the SZ.

The fine grains of the SZ reveal the changes in the flow direction, with narrow flow-lines delineating the kissing bond (Figure 11a) and the swirling zone (Figure 11b) as a direct outcome of the tool action within the stirring zone of the weld. These features were indicative of the mechanical stirring action and the subsequent flow inconsistencies and dynamic recrystallization experienced during the FSW welding process.

These dark flow-lines and the existing tunnel voids are representative of flow failure as the microstructure stabilised during cooling to room temperature. The flow inhomogeneity can result in the deterioration in mechanical properties of the weld. Therefore, it is necessary to measure the mechanical properties of the weld to see how the internal defects could affect the strength of the weld.

To investigate the possibility of contamination of the welding region by the tool, the composition of the stir zone texture should be measured. Micromorphology of the grain structure was observed via SEM, and composition was analysed using EDS, as shown in Figure 12. Tungsten atoms have a high atomic number (74) and atomic weight (183.84 g/mol), which can be easily characterized by EDS.

**Figure 12.** Schematic plot of the SEM results; (**a**) micromorphology of the stirring zone and (**b**) EDS composition for the selected region of the stirring zone processed by the square-shape pin tool.

The microstructure of the stirring zone (processed by the square-shape tool) in Figure 12a confirms that the DRX equiaxed grains are uniformly distributed through the weld. Moreover, it can be seen that there were no irregular morphologies indicating agglomerate phases, or any specific inclusions formed due to possible dispersed contamination within the microstructure. In Figure 12b, the EDS spectrum of the selected region from Figure 12a (Spectrum 1) could be seen. This confirms that the composition did not contain any W or C impurities. The presence of Fe (0.2 wt %) was consistent with the initial composition of the parent metal AA1100 (containing 0.182 wt % of Fe in quantometer analysis), so was unlikely to represent contamination from the H13 steel shoulder of the tool.

To provide a better correlation between the FSW processing conditions and the phases formed within the weld region, further phase characterization was conducted by X-ray diffraction (XRD) analysis (Figure 13), where the parent metal AA1100 was compared with the SZ material.

α

α

α

**Figure 13.** XRD analysis of AA1100 aluminium plate, before and after FSW processing by the square-shape pin tool; (**a**) parent metal and (**b**) stirring zone.

Figure 13 illustrates the XRD patterns for the AA1100 alloy parent metal (Figure 13a) and the SZ (Figure 13b) processed by the FSW tool (square-shaped tool). As shown in Figure 13a, the α-Al phase (matrix) and Al2Cu precipitate were identified from the peaks of the diffraction pattern for the parent metal.

According to Figure 13b, the XRD analysis of the SZ region after the FSW processing shows the same phases (α-Al matrix and Al2Cu precipitate), compared to the parent. However, a significant increase in the intensity of the diffraction peaks was evident for both phases of the matrix (α-Al) and precipitate (Al2Cu), upon stirring. This can be attributed to phase stabilization after the stirring process and also an increase in precipitate distribution density due to DRX.

Consistent with the EDS analysis, the XRD analysis also confirmed that the SZ of the weld region was free of any W and C impurities. This shows that the FSW processing provides proper phase homogeneity in the SZ, without any contamination from the tool. This is based on examination of the weld sample processed using the square-shaped pin geometry, which has the highest degree of stress concentration at the sharp corners of the tool.

#### *3.4. Mechnical Strength of the Weld*

Figure 14 shows the tensile strength graphs (stress–strain) for the three sets of weld samples processed by varying different conditions; alteration of RPM (Figure 14a), alteration of feed rate (Figure 14b), and changing of the tool geometry (Figure 14c). As observed in Figure 14a,b, by increasing RPM or decreasing feed rate, the ultimate strength of the weld slightly decreased compared to the base metal. However, at the same time the elongation rate of the sample increased. Regarding the dependence of tensile strength on the geometry of the pin profile, Figure 14c confirms that threaded > square > conical.

Lower rotation speed during FSW results in less frictional heat generation, and consequently poor plastic flow of material. Therefore the tensile strength can be considered a function of the process speeds.

It was observed in Figure 14a that the ultimate tensile stress (UTS) for the base metal (340 MPa) was higher than the weld samples. It can be attributed to the formation of the HAZ region with a negative impact on the strength of the weldment, because of the coarsened grains associated with this region [41].

**Figure 14.** Tensile test results for the AA1100 aluminium welds; (**a**) RPM = 710, 1120 and 1400 at feed rate of 250 mm/min, (**b**) feed rate = 150, 250 and 315 mm/min at 1120 RPM and (**c**) comparison between conical, square and threaded pin tools, processed in constant welding speeds (1120 RPM and 250 mm/min).

However, the percentage elongation of the weld sample was more than the base metal, representative of the ductile behaviour of the weld. Our interpretation is that the processing parameters (rotational speed and feed rate) increase the homogeneity of the weld region during the stirring process, in which lead to the ductility of the stirred material, compared to the parent metal [42]. More specifically, the comparison between the tensile test–strain curve for the samples processed by different tools in constant speeds (1120 RPM and 250 mm/min), shows a higher strength and the percentage elongation achieved by the threaded pin tool, see Figure 14c. This also implied the flow uniformity induced by the tool, leading to an increase of the ductility of the weld [39,43].

It should be noted that all the tensile curves show a kink [44] at the stress level of 25 MPa, approximately. This behaviour is explained as the toe region at the beginning of the tensile test [45,46], where the mechanical loading is less than 2% nonlinear strain for the stress–strain curve [47]. There is not a clear explanation for this behaviour, however, it mainly occurs for the soft materials [48,49]. When the stiffness is reduced, or in the presence of micro-defects, a micro-failure accumulates within the material, which was shortly released as the loading continues. Regarding the AA1100 FSW samples, existing of some flow-based discontinuity or micro-porosities like tunnel voids cause a self-interaction inside the material texture during the first steps of the tensile test loading, therefore the toe region occurs. However, the main aim was to compare the strength of the weld in different processing parameters. Additionally, because the tensile testing method was conducting in the same way, and the toe region was observed at stress level approximately 25 MPa for all samples, it did not affect the results of the tensile test.

#### **4. Discussion**

This work makes the following original contributions to improve the understanding of the FSW welding processed by the WC-based tool for the aluminium alloy, regarding the microstructural evolution and mechanical properties of the weld.

According to the microscopic observations made, at higher speeds the frictional heat generated is higher, which leads to an increase in the plastic flow of material. This can intensify the emergence of flow-based defects (e.g., tunnel voids, kissing bonds and swirling zone) and subsequently results in poor stirring conditions [33,50,51]. This is consistent with observations in the literature for the Al alloys processed by the FSW, which show the control of the plastic flow deformation as the main factor for improvement of the weld quality [42,52,53]. Alternatively, the interruption of the flow integrity can easily deteriorate the quality of the weld structure.

Similar to the speed condition, the pin geometry also can affect the tensile strength. The pin profile in different geometrical configurations (conical, square and threaded pin) can affect the frictional heat. The tensile strength of the threaded tool weld was higher than that of both other types (conical and square). This is because of greater material softening obtained at higher temperatures, as the threaded pin enhances the plastic flow of material due to higher engagement between the pin and material, and extra material contact. According to the literature, the pin profile for the Al FSW welds show a distinct shearing effect, mainly an ultrafine grain refinement at the mid-SZ, altered to the transition region borders (TMAZ and HAZ) towards the base metal. It is also supported by the previous works that the shearing field induced by the rotating tool into the proximity mass flows is the main factor in the microstructure alteration of the weld texture [39,54–56].

As a key finding of the phase characterization, from the XRD and SEM/EDS analysis, it was confirmed that there was no evidence of weld contamination from the WC pin tool. However, optical microscopy revealed some defects, (e.g., kissing bonds, swirling patterns and tunnel voids), which were identified as being inherent to the FSW process, rather than being due to tool contamination. Furthermore, the absence of carbon-containing compounds in XRD and EDS analysis eliminates the possibility of formation of carbon monoxide (CO) gas due to oxidation of the WC at the elevated temperatures occurring during the FSW [24]. Therefore, the stability of the WC pin tool was confirmed, as no oxidative wear degradation occurred during the severe frictional deformation.

This has been affirmed in the literature that the formation of oxide inclusions or other external contaminations can strongly disrupt the flow homogeneity within the plasticized stirred mass [57–59]. Subsequently, the recrystallization and grain refinement also can be affected. The microscopic observations in this study indicate the absence of the formation of the oxide layers of contamination or any other W-based compounds within the weld structure, which confirms the main idea of this research in the successful performance of the WC-based pin tools for FSW processing of the Al alloy as an active material, without occurrence of any metallurgical issue.

#### **5. Conclusions**

The effect of a WC pin in the microstructural quality of the AA1100 FSW weld was evaluated using different sets of welding speeds (RPM and feeding rate) and different pin geometries (conical, square and threaded shapes). Metallurgical analysis confirmed that the weld structure was free of possible W and/or C contamination from the pin. Nevertheless, microstructural observations revealed the formation of flow-based defects (tunnel voids, kissing bonds and swirling patterns) within the weld structure. Since the quality of the weld was directly affected by these structural defects, the tensile strength of the welded sheets was reduced, as the plasticized material of the weld could not sufficiently achieve the strength of the base metal. Moreover, the welded samples experienced a clear decrease in elongation rate by the increase in the feeding rate. The metallographic observations confirmed that the emerging internal defects could not be completely avoided. Therefore, the specific application should adjust the welding parameters according to the required strength.

**Author Contributions:** Conceptualization, A.T. and A.B.; Formal analysis, A.T. and A.B.; Investigation, A.T. and A.B.; Methodology, A.T. and A.B.; Project administration, A.T. and A.B.; Resources, A.T. and A.B.; Supervision, A.T.; Validation, A.T. and A.B.; Visualization, A.T. and A.B.; Writing—original draft, A.T. and A.B.; Writing—review and editing, A.T. and M.E.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors would like to thank Alice Young (University of Canterbury, NZ) for her assistance in proofreading the manuscript.

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

#### **References**


© 2020 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 (http://creativecommons.org/licenses/by/4.0/).

## *Communication* **AFM Characterization of Stir-Induced Micro-Flow Features within the AA6082-T6 BFSW Welds**

## **Abbas Tamadon \* , Dirk J. Pons \* and Don Clucas**

University of Canterbury, Department of Mechanical Engineering, Christchurch 8140, New Zealand; don.clucas@canterbury.ac.nz

**\*** Correspondence: abbas.tamadon@pg.canterbury.ac.nz (A.T.); dirk.pons@canterbury.ac.nz (D.J.P.); Tel.: +64-021-028-12680 (A.T.)

Received: 4 October 2019; Accepted: 6 November 2019; Published: 7 November 2019

**Abstract:** Bobbin Friction Stir Welding (BFSW) is a thermomechanical process containing severe plastic deformation by mechanical stirring and Dynamic Recrystallization (DRX) during recooling. Here we report the three-dimensional characteristics of the micro-flow patterns within the aluminium weld structure. The Surface topography observations by Atomic Force Microscopy (AFM) show the stirred-induced microstructural evolution where the rearrangement of dislocations at the sub-grain scale, and the subsequent High- and Low-Angle Grain Boundaries (HAGBs, LAGBs) exhibit specific alterations in grain size and morphology of the weld texture. The dislocations interaction in different regions of the weld structure also was observed in correlation to the thermomechanical behaviour of the BFSW process. These micro-flow observations within the weld breadth give a new insight into the thermomechanical characteristics of the FSW process during the stirring action where the plastic flow has a key role in the formation of the weld region distinct from the base metal.

**Keywords:** thermomechanical processing; bobbin friction stir welding; atomic force microscopy; AA6082-T6 aluminium alloy; dynamic recrystallization; precipitation

#### **1. Introduction**

Bobbin Friction Stir Welding (BFSW) is a modified variant of Friction Stir Welding (FSW) [1] where the conventional tool is replaced by a bobbin-shaped double-sided configuration [2,3]. The rotating double-shoulder bobbin tool penetrates from the edge through the interface of the side-by-side plates, and mixes the materials into a butt-shaped joint [4–6]. The heat input [4] generated by the friction between the rotating tool and the workpiece plasticises the material from both sides of the interface, Advancing Side (AS) and Retreating Side (RS), and stirs them together to form a bonded structure [6]. The stirring action causes severe plastic deformation [7] at temperatures well below the usual melting point [4]. Hence, the process is suitable for the joining of low temperature deformable alloys [8]. Aluminium is an ideal material for successful processing under BFSW [2,9]. AA6082-T6 aluminium is an industrial marine grade alloy with good machinability which has recently become attractive for FSW processing [10–14].

To achieve a defect-free weld in FSW processes, the material flow regimes have a higher priority than the metallurgical details [15]. Therefore, it is necessary for the continued improvement of the BFSW process to determine the plastic flow patterns in the weld region, which has received minimal attention in the literature compared with metallurgical aspects.

The severe plastic deformation during friction stir welding is the main cause of alteration in grain size and morphology [16–18]. The sub-grain scale analysis of microstructure can elucidate the relationship between the microstructure and the thermomechanical nature of the FSW process, especially shearing and heat generation [19]. The deformation-induced texture varies across the weld, as the shear is the function of the distance of the stirred flow mass from the position of the rotating tool [16,20]. Hence, a better understanding of the grain structure has the potential to contribute to knowledge of the evolution of the thermomechanical mechanism.

This paper presents an innovative study of the three-dimensional topology of the material flow features of BFSW weld texture. By utilising the Atomic Force Microscopy (AFM), the surface topography in the microscopic scale reveals the flow-based characteristics of the weld arising from the stirring action as a severe plastic deformation. This has the potential to give a better understanding of the effect of the microscopic flow regimes on the thermomechanical properties of the BFSW weld texture.

In this work, a high-magnification microscopic measurement was used to observe how the micro-scale plastic deformation affects the microstructure. In this regard, microstructural changes with a focus on dislocation and flow-induced imperfections were analysed in diverse regions of the AA6082-T6 BFSW weld structure.

AFM analysis was used to identify the microscopic details of the weld texture. This provides a greater resolution—to the atomic level—compared to other microstructure measurements. More specifically, the dislocation network and the precipitation evolution within the weld texture can be observed with AFM.

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

The BFSW welding test was conducted with the AA6082-T6 aluminium alloy (Standard; EN AW-6082, ISO: Al Si1MgMn) rolled plates (Aalco Metals Ltd, UK) as the workpiece. The analysed chemical composition of the AA6082-T6 plate with the elemental details is listed in Table 1.


**Table 1.** Chemical composition of the AA6082-T6 aluminium alloy, in elemental detail (wt.%).

The weld samples were in a butt joint configuration comprising two similar plates (250 mm × 75 mm × 6 mm). The geometrically full-featured bobbin tool (included threads, flats and scrolls) was made of H13 tool steel with a hardness of 560 HV [5,6,21]. The schematic of the BFSW process is shown in Figure 1. The BFSW experiments were performed on a 3-axis CNC machining centre (2000 Richmond VMC Model, 600 Group brand, Sydney, Australia) with a Fanuc control unit and 14-horsepower spindle motor capacity. The simultaneous operation speeds consisted of spindle rotational speed (ω = 650 rpm), and weld travel rate (V = 400 mm/min). The direction of tool rotation was clockwise viewed from above.

The welded plates were cross-sectioned perpendicular to the weld-seam and were micro-polished to a mirror level, according to the standard metallographic preparation [22]. The samples then were etched by using of two modified reagents; (Reagent A: 2 mL HF + 2 mL HBF<sup>4</sup> + 10 mL HNO<sup>3</sup> + 20 mL CH3COOH + 33 mL H2O + 33 mL ethanol), and (Reagent B: 0.5 g (NH4)2MoO<sup>4</sup> + 3.0 g NH4Cl + 1 mL HF + 18 mL HNO<sup>3</sup> + 80 mL H2O) per [6,21]. Previous research has shown these etchants successfully delineate grain-boundary microstructure (Reagent A) and micro-flow patterns (Reagent B) [6,21]. Both reagents were used for similar etching time and temperature (90 s, 70 ◦C). For a better cleaning of the

samples, the specimens were cleaned in the acetone bath (60 s, 18 ◦C), under an ultrasonic field of 40 KHz frequency. There is no known corrosive effect of acetone on this material.

**Figure 1.** Schematic of the Bobbin Friction Stir Welding process; Bobbin-Tool in interaction with the workpiece as the substrate.

To observe the macro-flow patterns within the BFSW weld structure, the etched samples were studied with a stereoscopic microscope (Olympus Metallurgical Microscope, Tokyo, Japan). Microscopic observations of the morphological features were conducted with an atomic force microscope (Veeco Digital Instruments Dimension 3100, from Bruker). The in-situ observations were done in contact mode, equipped with TAP300-G (PDMS imprint and replica) tips (BudgetSensors, USA), in dry conditions at ambient conditions (in air, at room temperature). Visualisation of the recorded mapping data and the three-dimensional topography analysis were processed by Gwyddion software (VERSION 2.45).

#### **3. Results**

The macro-etched cross-section of the AA6082-T6 BFSW sample (etched by Reagent A) and micrographs of the relevant regions of the weld are shown in Figure 2. The microscopic images distinguish different morphological microstructure within the weld region from the Base Metal (BM), through to the transition region; Heat-Affected Zone (HAZ) and Thermomechanically Affected Zone (TMAZ), onwards to the Stirring Zone (SZ); flow arm patterns and middle of SZ. Compared to the base metal, the grain size in other regions of the weld shows an extensive grain refinement (reduction of grain size), and grain morphology.

The thermomechanical nature of the FSW process and the subsequent mechanical and thermal behaviours of the weld texture are believed to be the major activators for this microstructural evolution [21,23].

The AFM images of the weld region for the AA6082-T6 BFSW sample (etched by Reagent B) are demonstrated in Figure 3. After FSW treatment, there are microscopic changes that have taken place in different regions of the weld. The topographic features denote that the BM region (Zone 1) is smoother than the SZ (Zone 5), as well as the transition region at the AS/RS borders of the weld (Zones 3 and 4) where bulging lamellar microbands are evident.

**Figure 2.** Macroscopic and microscopic features of the BFSW weld for the etched cross-section of the AA6082-T6 plate (reagent A); (**a**) Macro-etched cross-section of the AA6082-T6 sample, representative of the hourglass shaped weld structure, (**b**) the selected region from AS region at Figure 2a, in higher magnification, distinguishing five different regions for the weld breadth, (**c**) SEM images of 5 different regions of the weld texture, demonstrated in Figure 2b. (AS; Advancing Side, RS; Retreating Side, BM; Base Metal, SZ; Stirring Zone).

The topographic features are interpreted as micro-flow patterns caused by plastic deformation during stirring. The peak-and-valley-like micropattern is apparent throughout the weld section. The figure also shows roughness and texture measures derived from the AFM process. Also given in the figure are descriptions of the grain morphology, and the evolution between these states.

The figure is left for inspection, but there are several broad comments to be made. Inspection of Zones 3 and 4 (SZ) (Figure 3) shows the dominance of wrinkled and linear features, unlike the other regions. Shear is only active in Zones 3 and 4: it does not exist in Zone 1 due to the T6 heat treatment, nor in Zone 2 since this is only affected by heat flux. The shear is less apparent in mid-SZ (Zone 5), which is consistent with a stress relieving process. The Dynamic Recrystallization (DRX) mechanism uses stored strain and heat to nucleate and recrystallize the microstructure into ultrafine grains, thereby erasing the shear bands that might otherwise be expected at this location [23].

We attribute the wrinkled features in Zones 3 and 4 to activation of dislocations due to shear. Zone 3 shows activation at main grain boundaries (High-Angle Grain Boundaries, HAGBs). There is possible evidence for accumulation of dislocations at grain boundaries. In contrast in Zone 4 the dislocations are arrays within the grains (Low-Angle Grain Boundaries, LAGBs) forming sub-grain boundaries. This rearrangement of dislocations is a typical thermomechanical characteristic of DRX.

The 3D topography maps and roughness profiles are not all the same scale, so care must be taken in interpretation of microscopic features. Overall it is apparent that there is a large change in microstructure from the bases metal to the other zones. The base metal is characterised, as expected, by large grains, absence of sub-grain boundaries, an evidence of precipitation inside the grains (because of the artificial ageing T6 cycle [21]). The AFM method is not ideally suited to the larger scale of 50µm where there may be voids and other surface defects that may introduce noise, hence caution is required in the interpretation of some the features at the larger scale. In the transition regions (Zones 2–4) there is evidence of similar roughness and size of features as quantified by the surface profile, but the 3D

images show that the morphology is very different across these zones. In the stirred region (Zone 5) the roughness increases again. We believe this represents the density of grain boundaries, which is confirmed by the 3D image which shows more homogeneity and fineness in morphology.

**Figure 3.** 3D topographic AFM images of different regions of the BFSW weld texture measured by the relevant height profiles of the surface roughness and the in detail explanation of the morphological evolution of the microstructure.

The question might be asked whether the observations might be explained by corrosion (from the etching process) rather than shear. We suggest that the effects are not due to corrosion, for the following reasons. First, the same etchants have previously been shown not to result in corrosion artefacts [6,21]. Second, there is no evidence of pitting or a local depletion of the matrix phase around the precipitate particles, as might be expected from an aggressive etching. Neither do the local line scan profiles show any evidence of pitting or protrusions.

We propose the following interpretation of the process. The mechanical stirring induces a shearing distortion to the lattice of the plasticised mass. To mitigate this during DRX, the microshrinkage positions evolve to be at the location of the accumulated shear bands. The etching procedure leads to more clarity of these shearing microbands in 3D surface tomography scanned by the AFM (Zones 3 and 4 in Figure 3, apparent at nanoscale). It is evident that the etchant has a very different effect on the material in Zone 3, compared to the other zones. The grain boundaries are pronounced, creating a three-dimensional interconnected set of ridges. The inside of the grain is locally dissolved to a greater extent. We proposed that this arises from the selective etching features of the reagent solution, and we attribute this to the shear bands being more sensitive to this etchant. Specific sample preparation was used to avoid over-etching and production of corrosion artefacts.

The grain boundaries and morphologies were revealed because of different reactions of each region of the weld texture to the applied reagent. In this regards, the surface roughness measurements indicate the morphology alteration in different regions after etching [6,24]. All samples were treated with the same polishing and etching procedure.

The surface morphology in the flow arm region (Zone 4) is in agreement with the theory of the stirring action in the FSW process. We suggest that the periodicity of the flow arms (visible in Figure 2b) is caused by the rotary motion of the pin and its flats create. This causes the deposition, in the wake of the tool at the advancing side, of the parallel arm shapes [24–27]. The microstructure of Zone 4 (in Figure 3) was selected from one of these arms, and shows that the sub-grain boundaries (dark lines in Figure 3) are arranged parallel to the flow arm.

The actual roughness in the mid-SZ (Zone 5) was reduced, as the grain size has decreased compared to other regions of the weld. The micrograph shows larger variations, which is because of the higher magnification (smaller scale). The graph of surface roughness confirms that the distance of peaks and valleys as a measure of the surface roughness, is minimized for Zone 5.

#### *3.1. Observation of Precipitate in Mid-SZ*

The most plausible area for location of the precipitate particles is mid-SZ, which experiences a full DRX compared to other regions of the weld. However, from an imaging perspective there is a risk that the etching might eject the fine precipitates from their position within the microstructure. To minimise this risk, an unetched micropolished sample was used for the precipitation analysis. Furthermore, a variety of channelling modes were used for the AFM analysis.

In addition to the usual height channel mapping, frequency channel was used to provide a better resolution for the free edges of the precipitate in contrast with the matrix. Also to indicate the localized torsional stress field at the interface the precipitate-matrix phase, dissipation channel was used. These different channeling modes provide a more comprehensive observation.

The micrographs (Figure 4a,d) and corresponding line profiles (Figure 4e) show an ultrafine particle of size less than 100 nm, a platelet morphology, and an embedment in the matrix. The height channel (Figure 4a) identifies a platelet shape particle without surface etching. The frequency channel (Figure 4b) provides better sharpness of contrast, and shows a localized cleaved pattern at the boundary of the precipitate-matrix. The dissipation channel (Figure 4c) also highlights the dissipated energy from the tapping tip of the AFM probe onto the sample surface. The red contrast, constructed at the interface of the precipitate-matrix phase, shows higher density of dissipated surface energy, attributed to the mechanical torsion at the border of the particle [28]. Frequency channel and dissipation channel both show that the border of the particle and the matrix possesses a distortion which is due to the embedment of the precipitate into the matrix [29,30]. Therefore, it is not an external particle or sediment at the surface.

In general, precipitation requires diffusion of alloying elements, and is function of time and temperature. DRX occurs within seconds during FSW, not enough time for diffusion to occur to produce precipitates and therefore the precipitating phase particles are expected to be ultrafine size [21,29], which indeed is demonstrated here (<100nm). The mechanism of precipitation here in FSW process is attributed to the severe shearing in the SZ and the heat. In the T6 artificially ageing process the cycle time is longer, the temperature is higher, and the process is closely controlled, resulting in control over the precipitates [21,29–31].

**Figure 4.** Selected surface area demonstrating the platelet shape precipitate, scanned by AFM using different channel modes; (**a**) height channel exposure, (**b**) frequency channel, (**c**) dissipation channel, (**d**) phase contrast micrograph, (**e**) line profiles corresponding to the surface roughness of the particle.

The observed platelet morphology has a maximum surface-to-volume, compared to other precipitate morphologies such as spherical or needle shaped. The platelet precipitation can occur in a relatively fast cooling rate at lower reaction temperatures [32,33]. Hence the observation of such a precipitate is consistent with the FSW condition. The density of precipitates was relatively low in the observed samples, less than might be expected from say fusion welding. This and the shape observed, suggest that the low temperature conditions in FSW result in a reduced precipitation.

## *3.2. Dislocations*

Dislocations are the out-of-position of atoms in the crystal structure of grains. The dislocation patterns within different region of the weld are shown in Figure 5. At this high magnification the misarrangement of the crystal layers becomes apparent. The edge dislocations are evident as the crystalline defect in the structure of the aluminium. In general, in polycrystalline structures under shearing (with an intrinsic misorientational angle >0.99) it is expected to observe dislocations as a structural defect through the lattice [19,23].

**Figure 5.** Stripped patterns of the crystalline lattice of AA6082-T6 BFSW weld; (**a**–**e**) The landmarks indicates some of edge dislocations between the lattice planes (crystal layers), in different regions of the weld; (**a**) BM, (**b**) HAZ, (**c**) TMAZ, (**d**) Hourglass-border of SZ. (**e**) AFM map with very high resolution crystal structure with corresponding edge Dislocations (mid-SZ).

During plastic deformation and the subsequent DRX, dislocations are formed in specific preferential orientations within the crystal lattice. Furthermore, during the recooling, interaction and annihilation of dislocation results in rearrangement of dislocation arrays with different density in different regions. This can be related to the amount of the stored strain releasing during the recovery process, also the absorbed heat which is different for each region based on the distance from the frictional stirring action. The dislocations originate from the applied stress during stirring causing shear between the crystal layers. After stirring the DRX process involves movement of the dislocations. They may aggregate to form Low Angle Grain Boundaries (LAGBs) within the grains, or transfer to the grain boundaries and contribute to formation of new High Angle Grain Boundaries (HAGBs) and motion of grain boundaries [21,23].

#### **4. Discussion**

This paper describes joining of an aluminium alloy by bobbin FSW, and investigation of weld microstructural features using the Scanning Electron Microscopy (SEM) and AFM.

A key finding is the identification of three-dimensional micro-flow features with specific changes in grain size and morphology attributed to the stirring action. This has not previously been shown for friction stir welding, and the literature is silent on this aspect.

The AFM technique provides a visualization of features within grains, and by inference provides a record of the flow patterns occurring in a solid-state mechanical stirring by the bobbin-tool FSW [21,23,24]. The surface topographic features for different regions of the weld can also be measured quantitatively to compare the surface roughness corresponding to the shearing regime. A tentative flow-induced thermomechanical mechanism has been suggested for the SZ and the transition region, where the stirring-induced shearing stress affects the weld structure through the thermomechanical behaviour of the BFSW process.

A metallurgical transformation during the DRX process is identified by its effects on changing of grain size to ultrafine. This is evident as grain refinement by increasing density of grain boundaries (see Zone 5, Figure 3). Precipitation is also expected from DRX, but was not readily observed in etched surfaces. Dislocation interaction in the sub-grain scale was visible in the transition regions (Zones 3 and 4, Figure 3), as were HAGBs (Zone 3, Figure 3) and LAGBs (Zone 4, Figure 3). Hence the transformations preceding and caused by DRX have been observed.

#### *4.1. Welding Parameters*

The formation of the weld texture is because of the mechanical stirring action at the proximity of the rotating tool [16]. The welding process causes fragmentation, severe plasticizing, shearing deformation and frictional heating. Therefore, the welding process parameters (tool geometry and welding speeds) can have a major effect on the final microscopic characteristics of the weld texture [34,35]. The complexity of the tool geometry increases the frictional heating generated at the position of the tool-material, inducing more plastic flow through the softened mass [36].

Similarly, the welding speeds (ω, *V*) also can induce more fragmentation and subsequent plasticising, resulting in more strain and hence DRX during the stirring action [37].

All these can intensify the shearing flow during the mass transportation, and potentially elevate the generated heat useable for the DRX mechanism. Therefore, this might be worthwhile to investigate the role of optimised welding parameters in the microscopic evolution of the weld texture characteristics.

#### *4.2. Limitations of this Work and Implications for Future Research*

Our 3D visualization analysis of the microscopic features was limited to the ultrafine microstructural details at the scale of the grain structure of the weld. However, there are some macro-size defects such as tunnel void or cracks which may also have shearing-flow effects. In this regard, because of the limitation of the AFM analysis to ultrafine magnification, the macroscopic defects are better analysed by other microscopic measurements, such as optical microscopy or electron microscopy. Furthermore, fractography analysis could evaluate the crack propagation and failure mechanisms. The formation of these macro-size tunnel void and the micro-cracks adversely affects the strength of the final weld, therefore is unacceptable to industry users.

Another possible future research opportunity could be to use AFM to quantify the grain characteristics for the different weld regions. It may be possible to characterise the surface features, and quantify sub-grain boundaries, and mathematical link these metrics to the weld process. Complementary methods such as electron microscopy (e.g., Electron Backscatter Diffraction (EBSD) and Transmission Electron Microscopy (TEM)) might be considered.

#### **5. Conclusions**

This research determined a physical measurement for describing micro-flow features within the BFSW weld breadth using a three-dimensional surface topography by AFM. It was revealed that the mechanical stirring was associated with complex flow regimes through the stirring zone, also induced shear features at the microscale. These add stored strain to the texture which appears to lead to physical alteration in recrystallization of the weld texture during the post-welding cooling. Therefore, different regions of the BSFW (SZ, TMAZ, HAZ) are identified in different microscopic patterns corresponding to thermomechanical behaviour of the weld.

A key outcome is the use of AFM to better understand the grain structure of the AA6082-T6 material under solid-state friction-stir welding. This is an important industrial material but its thermomechanical behaviour has been poor in this type of welding. The results of this paper elucidate the grain boundaries and precipitates, and thereby show the results of the thermomechanical processes. AFM has been shown to be a useful tool to better understand the gain boundary engineering, dislocation behaviour, and precipitation of this material.

**Author Contributions:** Conceptualization, A.T. and D.J.P.; Methodology, A.T.; Validation, A.T.; Formal analysis, A.T. and D.J.P.; Writing—original draft preparation, A.T.; Writing—review and editing, A.T. and D.J.P.; Supervision, D.J.P. and D.C.

**Funding:** This research received no external funding.

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

#### **References**


© 2019 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 (http://creativecommons.org/licenses/by/4.0/).

#### *Article*
