Multi-dimensional Additive Manufacturing [1st ed.] 9789811579097, 9789811579103

Citation preview

Soshu Kirihara Kazuhiro Nakata   Editors

Multi-dimensional Additive Manufacturing

Multi-dimensional Additive Manufacturing

Soshu Kirihara Kazuhiro Nakata •

Editors

Multi-dimensional Additive Manufacturing

123

Editors Soshu Kirihara JWRI Osaka University Osaka, Japan

Kazuhiro Nakata JWRI Osaka University Osaka, Japan

ISBN 978-981-15-7909-7 ISBN 978-981-15-7910-3 https://doi.org/10.1007/978-981-15-7910-3

(eBook)

© The Japan Welding Engineering Society 2021 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd. The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721, Singapore

Preface

Technologies that confer different properties or functions to the surface of a material have long existed. These technologies have been known by terms such as “surface treatments” and “surface technologies,” which include plating and conversion coating, or “surface heat treatments,” which include carburizing and nitriding. In approximately 1990, a new term, “surface modification,” arrived in Japan from the West. This accompanied the commercialization of new ceramic thin film coating technologies (CVD, PVD) and was intended either to unite these technologies or to differentiate them from the conventional technologies. To accommodate this new technological trend, the Surface Modification Research and Technology Committee was established by the Japan Welding Engineering Society (JWES) in 1989. The committee added thin film coating technologies to conventional surface technologies as well as thick film surface technologies, such as the welding techniques of thermal spraying and hardfacing/ cladding. The committee continues to be active to this day. Recently, the new terms of “3D printing technology” and “additive manufacturing (AM)” arrived in Japan from the West. 3D printing technology was already commercialized in Japan as a method of molding resin and was known by terms such as “stereolithography.” However, the technologies attracting attention now are targeted at metals, i.e., “3D metal printing technologies.” AM was officially defined as a manufacturing technology that encompasses “3D metal printing technologies” by the American Society for Testing and Materials (ASTM) in 2009. From the perspective of welding technologies, thick film surface modification technologies, such as thermal spraying and hardfacing, fall directly under the category of AM because they create 3D shapes by layering thick coatings. In Japan, these welding technologies were applied to more complex 3D shapes in the past, for example, in metal arts and crafts. Unfortunately, the strength of 3D laminated parts manufactured using the previous technologies was not adequately assured, and defects were not adequately prevented. Additionally, ill-defined product demand and applications meant that these technologies were given little attention in Japan. v

vi

Preface

However, with high-quality 3D metal printers supplied from Europe and the USA and items, such as jet engine parts, introduced as example applications, the technologies are suddenly attracting attention. Along with fundamental research at universities and public research institutes, the development of process equipment by machine tool manufacturers and application products by businesses is intensifying. This book intends to deepen the understanding of the relevant technological fields by taking a comprehensive look at conventional 2D surface modification technologies and new 3D metal printing technologies. This has been studied by leading experts in related academic and technological fields. Here, on behalf of the Editorial Board, I would like to thank the authors for their excellent contributions. I would also like to thank Prof. Soshu Kirihara of the Joining and Welding Research Institute at Osaka University for his valuable assistance in planning and editing this book. Finally, this book was published to celebrate the 30th anniversary of the founding of the Surface Modification Research and Technology Committee, JWES. As well as expressing my gratitude to the first committee chairman, Dr. Fukuhisa Matsuda (Professor Emeritus, Osaka University), and to the second committee chairman, Dr. Shoji Miyake (Professor Emeritus, Osaka University), and I would like to take this opportunity to thank all members and relevant parties for their efforts and cooperation in the operation and activities of the committee. Osaka, Japan

Soshu Kirihara Kazuhiro Nakata

Contents

Part I

Lithography

1

Selective Laser Melting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Takayoshi Nakano

3

2

Laser Processing for Metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fujio Tsumori

27

3

Selective Electron Beam Melting . . . . . . . . . . . . . . . . . . . . . . . . . . . Yuichiro Koizumi

35

4

Current Research and Development . . . . . . . . . . . . . . . . . . . . . . . . Naoyuki Nomura

47

5

Stereolithography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Soshu Kirihara

53

Part II

Nano/Micro Lamination

6

Chemical Vapor Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hirokazu Katsui and Takashi Goto

75

7

Aqueous Solution Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yoshitake Masuda

97

Part III

Coating and Deposition

8

Aerosol Deposition Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Kentaro Shinoda and Jun Akedo

9

Cold Sprayed Metal Coatings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Kazuhiko Sakaki

vii

viii

Contents

10 Cold Spray Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Kazuhiro Ogawa 11 Precursor Spray . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 Yasutaka Ando

Part I

Lithography

Chapter 1

Selective Laser Melting Takayoshi Nakano

1.1 Laser Characteristics Light is a type of electromagnetic energy that exhibits the properties of a wave and is formed from the interplay between electric and magnetic fields [1]. Therefore, light can be expressed by spatiotemporal changes, and it is represented by a waveform and its temporal movement. The electric and magnetic fields in each wave are perpendicular to each other, and they move in a linear fashion at the speed of light. The speed of light c is expressed as c = f λ, where λ and f represent the wavelength and frequency, respectively. Light is represented as the range of electromagnetic wavelengths (λ) spanning from ultraviolet rays in a vacuum (λ: ~100 nm) to near-infrared rays (λ: ~800 nm), whose range contains visible light. Lasers are an artificial form of light produced by amplifying light through induced emissions. When laser light is emitted, waves in the same phase sharing a specific wavelength are amplified. Hence, the laser behaves as parallel rays of light (sharing directionality within a diffraction-limited range) that diverge negligibly and share properties such as uniform wavelength (monochromaticity) and matching wave phase (coherence). In particular, the directionality of lasers confers high power and high energy densities owing to the concentrated light. Therefore, lasers are utilized in additive manufacturing (AM) as high-energy heat sources or highly precise reaction sources.

T. Nakano (B) Osaka University, Osaka, Japan e-mail: [email protected] © The Japan Welding Engineering Society 2021 S. Kirihara and K. Nakata (eds.), Multi-dimensional Additive Manufacturing, https://doi.org/10.1007/978-981-15-7910-3_1

3

4

T. Nakano

1.2 Types of Lasers and Laser Absorption Coefficients Lasers can be classified based on the laser oscillating medium (material), laser medium form, laser performance (pulse width, wavelength, etc.), and other criteria [2]. If the medium is used as the classification standard, lasers can be classified into CO2 lasers (λ: 10.6 μm), YAG lasers (λ: 1.06 μm), excimer lasers (λ (ArF gas): 0.193 μm; λ (KrF gas): 0.248 μm; λ (XeCl gas): 0.308 μm), and diode lasers (λ: 0.2–4 μm), among others. If the laser medium form is used as the standard, lasers can largely be classified as either a disk laser, which employs a thin disk as the laser medium, or a fiber laser, which employs optical fibers as the laser medium. Yb fiber lasers exhibit wavelengths approximately equal to that of YAG lasers (λ: 1.07 μm). Fiber lasers seal light within their fibers and oscillate the laser. Therefore, they are renowned for their remarkable reliability. The diameter of the fiber core determines the shape and quality of the beam. A fiber laser with a core diameter of order 10 μm would emit high-quality single-mode light. As the core diameter widens, the laser becomes multi-modal, and as the concentration of light decreases, the laser output increases. Both cases express the spatial characteristics of the laser. A single-mode laser comprises a single horizontal mode, whereas a multi-mode laser employs horizontal multi-modal oscillation, where multiple horizontal single-modal lasers coexist and resonate simultaneously in the laser resonator without mutual interference. Laser light oscillating in a single mode is effective in situations that necessitate micromachining and high luminance, as its output adheres to a Gaussian distribution. Conversely, light fibers oscillating in multi-mode are suitable for high-powered laser emissions, as their output distribution is not Gaussian ; therefore, they are appropriate for heating wide areas. Finally, based on pulse width, lasers can be classified into femtosecond, picosecond, and nanosecond lasers. These lasers are characterized by the presence or absence of heat-affected zones. The laser emission wavelength is determined by the laser medium, which is classified as solid (including semiconductor), liquid, or gas. The laser wavelength must be selected according to the type of manufacturing to be performed, for example, if metal is being used, as shown in Fig. 1.1 [3], the laser absorption coefficient is strongly dependent on the metallic element. As the wavelength decreases, the laser absorption by metal materials increases with high absorption efficiency. Therefore, selection of the laser wavelength is important when lasers are used as a heat source for dissolving metals. Metal additive manufacturing (metal AM) has been conducted with CO2 lasers used for laser material processing; however, the absorption efficiency of these lasers is high when working with iron and steel, and it is low when working with non-ferrous metals such as aluminum alloys. The introduction of high-output Yb fiber lasers has resulted in rapid improvements in AM technology for metallic materials. Metal AM uses metal powder as a starting material. However, laser absorption by the metal powder is dependent on the distribution (Gaussian, bimodal, etc.) of the powder particle sizes, and improves significantly with multiple scattering between the powder particles, when compared to that in bulk materials. Figure 1.2 illustrates

1 Selective Laser Melting

5

semiconductor laser

YAG laser, fiber laser

CO2 laser

0.30 0.25

Absorption ration

Ag

Cu

0.20

Steel Au

0.15

Fe

Al 0.10

Mo

0.05 0

0.1

0.2

0.3

0.5

1

2

4

6

8 10

20

Laser wavelength (μm) Fig. 1.1 Absorption ratio of laser in metals depending on laser wavelength. Reproduced from Schubert et al. [3] with modifications

Fig. 1.2 Absorption ratio of laser in bimodal metal powders depending on powder dispersion. The circle including a big powder (a) is absorbed less than that of small powders (b). Laser multi-scattering is larger in metal powders than in bulk metals

(a)

(b)

different dispersions of bimodally distributed spherical powders. When the vicinity of coarse particles (a) and fine particles (b) are compared, the energy absorption efficiency and the resultant energy transfer efficiency from the laser to the powder were found to be dramatically higher in (b). This property is highly effective for SLM that uses lasers as the heat source [4].

6

T. Nakano

1.3 Types and Characteristics of 3D Layered Manufacturing Processes Since the conception of the fundamental theory of the 3D layered manufacturing process in the 1980s by Hideo Kodama, Yoji Marutani, Herbert, and Hull et al. [5], several researchers and developers have expanded its potential applications to include plastics, ceramics, metals, and their combinations, resulting in further developments and discoveries. The fundamental method converts a 3D model created by computer-aided design (CAD) into 2D slice data. Thereafter, it uses the data to create a 3D structure with complex shapes by shaping and assembling it layer by layer. This process initially received a variety of names, including “rapid prototyping” (RP). In Japan, Takeo Nakagawa named it “layered manufacturing” in 1992 [6], and has been popularly called “3D printing.” At the ASTM F42 meeting in 2009, the term “additive manufacturing” (AM) was defined, which is classified into seven methods as shown in Table 1.1 [7]. Among them, material jetting, powder bed fusion, directed energy deposition, and vat photopolymerization can use lasers as a reaction or heat source. Vat photopolymerization hardens the photocurable resin monomers in the desired position through the photopolymerization response of the material. He–Cd solid-state lasers capable of high, steady output have recently been used in this method. Powder bed fusion and directed energy deposition primarily use high-output Yb fiber lasers, and they can fabricate a metal object directly from a powder. The low laser absorption coefficient of metallic materials was overcome by using, fine powders with a relatively small diameter of 10–45 μm (see Fig. 1.3), with making the layer thickness thin (20–50 μm), resulting in the reduced surface roughness of the product. The process facilitates the fabrication of 3D products that require insignificantsurface post-processing. Direct metal fabrication enables the production Table 1.1 ASTM standard (2009) for defining seven categories [7] Classification

Summary of method

Binder jetting

Liquid bonding agent selectively Plaster, plastics deposited to join powder

Materials

Material jetting

Droplets of build materials selectively deposited

Powder bed fusion

Thermal energy selectively fused Metals (copper, titanium, nickel regions of powder bed alloy, cobalt chromium alloy) Resins (nylon, amide), ceramic

Photocurable resins, waxes

Directed energy deposition Focused thermal energy melts materials as deposited

Metals

Sheet lamination

Sheet material bonded together

Paper, resins, metal foils

Vat photopolymerization

Liquid photopolymer selectively Photocurable resin monomers cured by light activation

Material extrusion

Material selectively dispended through nozzle or orifice

Thermoplastic resins

1 Selective Laser Melting

7

Fig. 1.3 Starting raw metal (Ti-6Al-4V) powders utilized for powder bed fusion using laser beam as a heat source

100μm

of tailor-made industrial products through small-scale customization or through mass customization via smart factories using the technology of the Internet of Things (IoT). Therefore, it is likely to have an extraordinary impact on the industrial world. Figure 1.4 [8] shows the steps in the powder bed fusion process of metal AM. A laser, an electron beam, or an arc discharge can be used as the heat source in powder bed fusion. Since 2008, the process has been used to fabricate metal parts with an almost 100% density. Table 1.2 compares the performance of EOS M290, which uses a laser as its heat source, and Arcam Q10, which uses an electron beam as its heat

Laser beam

Platform

Metal powder

CAD data

Rake

Scanning the beam Lowering the platform and melting the and supplying new first layer powder layer

Scanning the beam Repeating the process and melting the up to produce a second layer designed structure Fig. 1.4 Schematic diagram illustrating the powder bed fusion method used in fabricating a metallic product with a lattice structure. Reproduced from Nakano et al. [8] with modifications

8 Table 1.2 Comparison of product characteristics between laser-beam additive manufacturing system of EOS M290 and electron beam additive manufacturing system of Arcam Q10

T. Nakano EOS M 290

Parameter

Arcam Q10

Laser

Heat source

Electron beam

250 × 250 × 325 mm Fabrication size 200 × 200 × 180 mm Yb-fiber laser

Beam

Electron beam (single crystal line (CeB6 ))

0–400 W

Power

50–3000 W

Max. 7 m/s (min. 0.05 m/s)

Scanning speed Max. 8000 m/s

5–20 cm3 /h

Fabrication rate 125 μm

10–45 μm

Powder size

45–105 μm

20–60 μm

Layer thickness >50 μm

0–200 ºC

Pre-heating

0–1100 ºC

source. As the laser carries no electric charge, it is necessary to physically adjust the galvanometer mirror to regulate its path. Meanwhile, the electron beam comprises negatively charged particles that can be regulated via an electromagnetic induction coil. Consequently, the laser scanning speed is three orders of magnitude smaller than that of the electron beam. However, as the time required for the formation of the powder bed is the rate-determining step during the actual layering process, there is no major difference in the fabricating rate. Nevertheless, if an electron beam is used as the heat source, its high beam-scanning speed can be used to preheat before fabrication. Because a laser beam cannot be used for pre-heating, residual stress (and strain) can accumulate inside the product. Figure 1.5 illustrates an overview of the EOS M290 SLM device, which can fabricate objects with a marginal amount of powder. The EOS M290 SLM is equipped with a 400W Yb fiber laser, and it can operate with almost all metals, including those with high melting points. Figure 1.6 [9] provides a simple overview of the equipment, and Fig. 1.7 [10] shows the optical system. Objects can be fabricated in an atmosphere of N2 or Ar inert gas, with almost no oxygen. A recoater wasused to deposit each layer of fine powder, and the 2D thin layers were selectively melted by laser irradiation according to the 2D slice data. The formation of a powder bed can also be achieved via rollers. In this case, it is applicable to the formation of a powder bed with high apparent density even for pulverized powder with an irregular shape and non-uniform size. The laser diameter on the object surface can be regulated using a lens or galvanometer mirror with a defocus mode.

1 Selective Laser Melting

9

Fig. 1.5 Appearance of laser-beam additive manufacturing system based on powder bed fusion (EOS M290)

When the powder is irradiated by the laser, fumes are generated by evaporation from the powder surface. Simultaneously, the solidified fumes become fine powders, obstructing the irradiation from the laser and posing a safety issue with regard to explosion.

1.4 Melt Pool Formation and Metal Laser Lamination Methods Melt pools are produced as a result of laser irradiation on metal powders. The regulation of their 3D morphology and solidification behavior is closely related to optimal control of the shape and microstructure of the products. There are various control parameters involved: those associated with the powder material and those associated with the SLM fabrication condition, such as the laser power (P), beam diameter (d), scan speed (v), scan pitch (w), and layer thickness (h).

10

T. Nakano

Galvano mirror & lens

Laser oscillator

Atmosphere: Ar or N2 Recoater

Build

PC z (BD) y x

Metal powder

Building platform

Powder supply platform

BD: Building direction

Fig. 1.6 Schematic illustration of powder bed fusion method with a laser heat source based on EOS M290. Reproduced from Nomura et al. [9] with modifications

⑤

④

③ ②

Scanner Collimator Laser

① Beam Expander

Laser beam transmission

⑥

⑧ ⑦

⑨

①: Laser inlet ②: Beam expander (focus position) ③: Beam expander (defocus position) ④: Beam expander exit lens ⑤: Scanner mirror ⑥: FT lens ⑦: Irradiated plane ⑧: Focus position ⑨: Defocus position

Fig. 1.7 Optical system of powder bed fusion method based on EOS M290. Reproduced from Ref. [10] with modifications

1 Selective Laser Melting

11

Fig. 1.8 Powder bed fusion method with a laser heat source based on EOS M290 during selective laser melting fabrication

Generally, the input energy per unit volume (E) is used as an index for the optimization of the fabrication conditions. E is expressed as follows:   E = P/vwh J/mm 3 Furthermore, by creating a process map using E as a function of v, one can obtain information on the optimal fabrication conditions and microstructure control. However, as energy density E does not consider factors such as time-dependent energy loss, owing to the heat transfer and change in free energy during solidification, it can only be used for rough predictions of microstructure formation. The establishment of numerical simulations of thermal distribution reflecting the effects of Marangoni convection generated in the melt pool, for the prediction of microstructure, is critical for the development of computer-aided manufacturing (CAM) in AM technology. Figure 1.8 illustrates the process of laser irradiation during SLM. The process can be used to fabricate products with a variety of internal structures by laser selective irradiation on the powder bed. It is possible to fabricate products with a variety of internal structures simultaneously. Reflection and scattering of the laser and generation of spattering and fumes occur during laser irradiation. Figure 1.9 shows the fabrication process of Inconel 718 products. The process can fabricate intricate products with a resolution equivalent to the laser beam diameter, which is renowned for its capability to fabricate internal as well as external structures. As manufacturing is conducted in an upward direction, the process specializes in the continuous fabrication of products with forms elongated in the build direction. The surface roughness is smaller on the sides or upward surface of the product than on its downward surface. It can fabricate near-net shapes or net shapes, which minimizes post-processing and grinding. Another effective metal manufacturing process is directed energy deposition. In the directed energy deposition process, metal powder is continuously supplied to the

12

T. Nakano

Fig. 1.9 Products during processes of powder bed fusion method with a laser heat source based on EOS M290

Laser beam

Focusing lens

Z positioning

Powder delivery nozzle

Build Base plate X/Y positioning stage

area with laser irradiation (Fig. 1.10) [5], which is referred to as buildup type fabrication. Although the product surface is coarse and the shape accuracy is low, the process has advantages over powder bed fusion because it is capable of producing larger products with an increased manufacturing rate. Various types of powder materials can be supplied continuously, thereby products can be fabricated with a composition gradient or composite material. The powder bed fusion method using a high-power laser will cause the powder to scatter via air flow around the laser-irradiated area. In contrast, in the directed energy deposition method, the powder is directly deposited onto the melt pool formed by laser irradiation. Therefore, it is feasible to use a highpower laser and increase the manufacturing rate linearly with the laser power [11]. Fig. 1.10 Schematic illustration of directed energy deposition method with a laser heat source

1 Selective Laser Melting

13

Because the powder supply is applied only to the area presently being manufactured, it is not necessary to remove residual powder, unlike in powder bed fusion. However, the shape accuracy is low (1–2 mm) in the directed energy deposition compared to ≤0.2 mm of the powder bed fusion. This method is also unsuitable for fabricating the overhanging part. However, in future development, these disadvantages will be overcome.

1.5 Support Design in 3D Data and Measures to Remove Residual Stress In many cases in AM technology, it is necessary to create “supports” in addition to product itself [10]. Although the supports are not required in the actual final product and must be removed, they are necessary during fabrication process because they can keep the product stable by preventing its deformation. For metal fabrication, the temperature distribution can be regulated by heat conduction via the supports. In SLM, where pre-heating for the formation of weak connections between powder particles by neck formation is unavailable, preparation of optimal support is crucial. Supports should be designed prior to fabrication by considering the convenience of removing the supports themselves and the surrounding powder after fabrication, strength as load-bearing parts for the product, prevention of separation caused by thermal strain, necessity of heat conduction, and improvement in the shape accuracy of the product. (Fig. 1.11). Support is necessary empirically in the parts with angles of inclination less than approximately 40°, tunnel shapes with diameters larger than 8 mm, or overhangs (Fig. 1.12). Figure 1.13 shows a Ti-6Al-4V denture fabricated using SLM. Numerous supports were created for the overhanging portion which require high resolution fabrication. Fig. 1.11 Formation of support design in 3D CAD model

14

T. Nakano > 8 mm

Product Support

Support

Support

< 40º

Base plate Fig. 1.12 Conditions for creating support design in 3D CAD model. Reproduced from Ref. [10] with modifications

Fig. 1.13 Support design for metal denture (Ti-6Al-4V) fabricated by laser-beam additive manufacturing system based on powder bed fusion

Furthermore, in SLM where pre-heating is not performed, the accumulation of residual stress inside the product frequently becomes an issue. Residual stress can cause product deformation and rupture after fabrication owing to heterogeneities in the product. When pre-heating is performed, the difference between the pre-heated temperature and the temperature during fabrication can be minimized. Therefore, the rate of thermal expansion/contraction should be small. In contrast, in selective laser melting without pre-heating, there is a significant difference in temperature between the parts being fabricated and those already solidified, generating large residual stress. Controlling 2D and 3D laser scanning patterns (scan strategies) are effective to reduce the residual stress.. A constant laser scan between layers tends to form an anisotropic microstructure in certain directions, which is likely to generate residual stress. The combination of different scan directions (Fig. 1.14) or shift of the start point of the laser scan between layers helps in reducing the residual stress [12].

1 Selective Laser Melting

15

Fig. 1.14 Scan strategy of possible scan pattern for reducing residual stress. Reproduced from Kruth et al. [12] with modifications 15º

y x

1.6 Control of Shape and Microstructure, and Formation of Single Crystalline with SLM SLM can simultaneously regulate the shape and material parameters of the product. Designing isotropic and/or anisotropic functions either through tailoring the shape or through controlling the microstructure of the product is realized. In powder bed fusion, the direction of laser irradiation and layer is fixed (vertical, etc.). Therefore, by appropriately selecting the fabrication conditions, one can obtain a product with anisotropic microstructure and the resultant anisotropic functionality in a specific direction, which is similar to functional structures in nature, such as bones [13, 14]. The shape parameters include the surface morphology and internal structure (cell shape, arrangement of solid parts, etc.). The material parameters address the selection of materials (and different physical properties), grain structure, crystallographic texture, etc. Ultimately, it is possible to regulate both the microstructure and shape simultaneously, for example, single crystalline turbine blades used in airplanes and power generation systems [15].

1.6.1 Shape Parameter and Its Hierarchy The foremost characteristic of AM technology is its capability to create desired complex shapes. From the perspective of imitating structures in nature, it represents the anisotropy and characteristics of the dimensions of each size scale as shape parameters and combines them organically to express functionalities. Even with the latest processing methods, it is feasible to confer shape parameters to a certain degree, and it is impossible to confer characteristics, such as hierarchical and complex inner structures in a single operation. Figure 1.15 [16] shows nominal stress–nominal strain curves for a product with an anisotropic shape, with respect to the x-, y-, and z-axes using a product created from a 27-unit cube structure (3 × 3 × 3) as a model case. Here, 14 of the 27 cube units were solidified, and the remaining were filled with Ti-6Al-4V ELI alloy

16 Fig. 1.15 Anisotropic stress–strain curves depending on the loading axis in powder/solid product. Reproduced from Nakano et al. [16] with modifications

T. Nakano

powder

solid

Empty cube

800

Z

Nominal stress (MPa)

600 600

Y

500

400 300

Z (BD)

X

200

X

100 0

Y BD: Building direction

0

0.05 0.10 Nominal strain

0.15

powder with ~50% packing density. Even in a cube with an internal structure of 27 units, the number of feasible arrangements of the units along fixed x-, y-, and z- axes exceeds 100 million. Therefore, it is necessary to calculate the mechanical characteristics that will be demonstrated prior to fabrication using a mixture rule. Figure 1.15 shows that the Young’s modulus and yield stress along the z-, y-, and x-axes decrease, and the number of support structures (columns) are four, two, and zero, respectively. This demonstrates that the number of columnar parts in the internal structure significantly influences the mechanical properties. Unmelted powder should generally be removed; however, it is purposely left enclosed in the product because of the hierarchical structure design. We refer to this structure as the P/S (powder/solid) complex [17]. Here, the powder (P) part as fabricated has insignificant effect on the mechanical properties of the product. However, as shown in Fig. 1.16 [18], through appropriate heat treatments, necks are formed between powder particles and between the powder and the product wall, resulting in a tenfold increase in the energy absorptive capacity. Hence, from powder particles of the order of tens of micrometers, to product walls, to internal structures of the order of several millimeters and external structures of the order of several centimeters, hierarchical structures are designed that exhibit macroscopic mechanical characteristics [18]. Among AM technologies, the layered manufacturing process of metal materials can change the melting and solidification behavior in individual melt pools, as

1 Selective Laser Melting

17

(b)

50 m

Nominal stress, σ (MPa)

(a)

800

Plateau

400

Energy absorption

200

As-AM without heat treatment

× 0

5 mm

Heat-treated for neck formation

600

0

20

40

60

Nominal strain, ε (%)

80

Fig. 1.16 a Schematic illustration of an AM fabricated product with remaining unmelted powder and photograph of powder part after heat treatment. b Nominal stress–nominal strain curves of products with and without heat treatment that makes powders necked. The product with the necked powder inside becomes highly energy absorptive, due to the stress transmission between powders through the necks. Reproduced from Ikeo et al. [8] with modifications

well as the direction of heat conduction or transmission, and thermal history through fabrication parameters. Therefore, the process can regulate material (microstructural) parameters as well as shape parameters [15, 20–25]. In metals, this process can regulate factors directly to influence the product properties, such as the size, morphology, distribution of crystallites, character of the grain boundaries, crystallographic orientation (texture), and porosity including its gradation. In particular, the process enables regulation of the crystallographic orientation via heat flux directionality. Figure 1.17 shows inverse pole figure (IPF) maps of the beta-Ti alloy products fabricated by SLM. The fabrication was carried out by two types of scanning strategies: one was bidirectional (zigzag) scanning in the x-axis (scan strategy X) and the other was bidirectional scanning with a rotation of 90° between layers (scan strategy XY) that repeats the X-scan and Y-scan [21]. Whereas low-angle grain boundaries with marginal differences in orientation are visible, single-crystal-like microstructures are formed in both cases. The crystallographic orientation differs with only a change in the scan strategy: with scan strategy X, orients in the scanning direction (x) and orients in y- and z-directions, whereas with scan strategy XY, orients in x-, y-, and z-directions [21]. The selection of crystal orientation takes place under the competitive relationship between the stability of the crystal orientation and epitaxial growth in the melt pool. Figure 1.18 shows the morphology of the uppermost surface of the product fabricated via scan strategy X. Figure 1.19 shows the formation of the small-angle grain boundary and the migration of the solid–liquid interface in the melt pool cross-section. Figure 1.19 shows that the -oriented cellular microstructure forms toward the migration direction of the solid–liquid interface, resulting in a tendency of orientation in the building direction. As the fabrication proceeds and crystallographic orientation become stable, the epitaxial

18

T. Nakano

z (BD)

Z-projection

z (BD) y

111

x X-projection

Scan Strategy XY

Y-projection

Y-projection

Z-projection

001

101

z (BD)

Scan Strategy X

X-projection

z (BD)

y x BD: Building direction

100μm

Fig. 1.17 Variations in crystallographic orientation depending on scan strategy, X and XY. Reproduced from Ishimoto et al. [21] with modifications

100µm

Direction of maximum thermal gradient y X (Scanning direction)

Fig. 1.18 Morphology of melt pool observed from upper side after fabrication of product based on powder bed fusion method with a laser heat source

growth dominates the texture formation. The formation of different crystal orientations with respect to the scan strategy is well explained through the mechanism of epitaxial growth shown in Fig. 1.20 [21]. The formation of a single-crystalline-like texture is closely related to the stability of the smooth solid–liquid interface, and it is necessary to prevent nucleation and growth of randomly oriented grains within the liquid phase caused by compositional supercooling [26].

1 Selective Laser Melting Fig. 1.19 Cross-section of melt pool based on powder bed fusion method with a laser heat source

19

z(BD) y

x(SD) 111

001 101 BD: Building direction, SD: Scanning direction

Scan Strategy XY

Scan Strategy X

[101]

[100]

〈110〉

〈100〉 〈100〉 〈100〉

〈100〉 〈110〉

[101]

[100]

500 μm

[001]

x scan

[001] y scan z

y

Direction of maximum thermal gradient Crystallographic

x

z y

x

Fig. 1.20 Mechanism for formation of products with different orientations by a X-scan and b XY scan strategy. Reproduced from Ishimoto et al. [21] with modifications

In the general solidification shown in Fig. 1.21, the solid–liquid interface of the equilibrium partition coefficient k 0 < 1 is considered to be in a steady state, growing from left to right at a constant growth velocity V. Here, C L represents the solute concentration within the liquid phase and DL represents the diffusion coefficient of the solute. The boundary conditions were Z = 0: C L = C L * , Z = ∞ C L = C 0 . GL depicts the thermal gradient at the interface and mL is the slope of the liquidus line. To preserve a stable solid–liquid interface, the thermal gradient G at the solid–liquid interface must satisfy the following equation: G > GL This yields the following formula: G k0 − 1 > m L · C0 · V k0 · D L

T. Nakano

Concentration

Concentration

20

C*

C*S C C0 ~ C*S

S

T1

V

Temperature Temperature

TL Supercooling G

GL

Solid Liquid z

0

Fig. 1.21 Schematic illustration showing compositional supercooling. Reproduced from Ref. [26] with modifications

When this equation holds in the solid–liquid interface in the melt pool, the singlecrystalline-like texture evolution is achieved in metal AM as well. Furthermore, as shown in Fig. 1.22, with a nickel-based superalloy (Inconel 718), a single crystalline can be grown with a specific crystal orientation using a single

z y x

(a)

Product

111

(b) Seed (single crystal)

50 μm

001

101

Fig. 1.22 Seeding along building direction in Inconel 718 alloy fabricated by powder bed fusion method with a laser heat source. Reproduced from Ref. [27] with modifications

1 Selective Laser Melting

21

Fig. 1.23 Products fabricated by powder bed fusion method with a laser heat source for transit metal dislicides

crystalline seed [27]. For heat-resistant materials, the single crystalline form is more effective considering the enhancement in creep properties at high temperatures. For Inconel 718, crystal growth in the direction is beneficial. In SLM, it is feasible to develop single crystalline products and control the orientation direction by optimally selecting scan strategies and seeding conditions. Conversely, Fig. 1.23 shows SLM products of the transition metal disilicide, a potential material with an ultra-high melting point [28]. The material is remarkably heat-resistant, exhibiting anomalous strengthening at 1600 °C [29] and excellent strength. However, its ductility and workability are poor at room temperature, and thus near-net shape production is required. As shown in Fig. 1.23, selective laser melting can be used to fabricate products that include porous media with MoSi2 . Creating [001] fiber textures with good creep properties in the laser scanning direction is a feasible task [24]. New advancements in SLM with regard to heat-resistant materials in the aerospace industry are expected.

1.7 Metal Products with SLM and Quality Management SLM has resulted in added product value in a variety of fields, such as custommade welfare and medical devices [8, 30]; aviation, space, and energy [31]; custom consumer electronics; education; and the service industry. Although expectations are high for high-precision, high-quality SLM products, numerous issues remain.

22

T. Nakano

Fig. 1.24 Turbine blade fabricated by powder bed fusion method with a laser heat source for Inconel 718

Moreover, there is a substantial demand for novel conceptualizations for adding value and opening up markets in ways that are unique to AM. Figure 1.24 shows a product fabricated by SLM. This process can be employed to fabricate turbine blades exhibiting high potential as aviation engine parts. SLM enables the fabrication of internal flow paths for cooling air in one process, which was previously not possible. The production of parts demonstrating excellent creep properties at high temperatures by skillfully combining laminar flow and turbulence to maximize cooling effectiveness cannot be achieved without AM technology. Figure 1.25 shows an example of the application of custom bone plates to bone fractures in animals. Just as human skeletal structures are not identical, the forms of companion animals (pets) also vary significantly in shape and size. To date, human implants have been designed uniformly on the basis of an average human skeleton. AM technology is expected to revolutionize the design concept of implants into customization. The customization would not be limited to the shape; it would be expanded to the microstructure and function considering the patients’ morbidity, medication history, and bone condition. Figure 1.26 shows a custom set of lighting equipment (shade) that combines three layers. The development of unique custom products suited to individual tastes, which provides high-value-added products, is a potential trend in the future development of AM technologies market. Combining AM with IoT and other technologies would be critical for customized products of materials from plastics to ceramics and even metals.

1 Selective Laser Melting

23

Fig. 1.25 Bone plates for rabbit fabricated by powder bed fusion method with a laser heat source for Ti-6Al-4V alloy

(a)

(b)

Fig. 1.26 Custom shade fabricated by powder bed fusion method with a laser heat source for Ti-6Al-4V alloy

The essential technology for the commercialization of AM products is quality management. The products fabricated using AM technology would be featured as tailor-made or small-batch -varied products, which is in contrast with mass production. Therefore, it is necessary to provide a method of quality assessment that differs from those used for mass production. With regard to quality assessment methods that perform destructive inspections, methods providing 3D information are required. These include focused ion beam milling-field emission-scanning electron microscope-electron back scatter diffraction patterns (FIB-FE-SEM-EBSD, Fig. 1.27) and energy-dispersive X-ray spectrometry (EDS). Meanwhile, for noninvasive or minimally invasive quality assessment, μ-computed tomography (μ-CT) and ultrasonic waves are methods that have been attracting attention.

24

T. Nakano

Fig. 1.27 Quality control for tailor-made 3D products with FE-SEM-EBSD method

Japan specializes in craftsmanship and fabrication of high-quality products, and therefore it has to assume a leadership role in AM for continuous progression in the field. Acknowledgements This work was funded by the Grants-in-Aid for Scientific Research (JP18H05254) from the Japan Society for the Promotion of Science (JSPS) and by the Council for Science, Technology and Innovation (CSTI), Cross-Ministerial Strategic Innovation Promotion Program (SIP), Innovative Design/Manufacturing Technologies (Establishment and Validation of the base for 3D Design & Additive Manufacturing Standing on the Concepts of “Anisotropy” and “Customization”) from the New Energy and Industrial Technology Development Organization (NEDO).

1 Selective Laser Melting

25

References 1. M. Ohtsu, T. Tadokoro, Introduction to Optics: Know Your Properties of Light, Optical Technology Series 1 (Asakura Shoten, 2008), pp. 1–214. 2. Laser Platform Association, An Introduction to Laser Manufacturing: From the Basics to Equipment Installation (Sanpo Publications, 2010), pp. 1–159. 3. E. Schubert, I. Zerner, S. Sepold, New possibilities for joining by using high power dipole lasers. Proc. ICALEO 85G, 111–120 (1998) 4. C.D. Boley, S.A. Khairallah, A.N. Rubenchik, Calculation of laser absorption by metal powders in additive manufacturing. Appl. Optics 54(9), 2477–2482 (2015) 5. Y. Marutani, S. Hayano, S. Imanaka, Documents on Layered Manufacturing Technology (Optnics Co., Ltd, 2002), pp. 1–181 6. Y. Marutani, S. Hayano, 3D Printers: Toward Sustainable Development of AM Technology (Optnics Co., Ltd, 2014), pp. 1–208 7. ASTM Standard F2792–12a, Standard Terminology for Additive Manufacturing Technologies (2009). 8. T. Nakano, T. Ishimoto, Powder-based additive manufacturing for development of tailor-made implants for orthopedic applications. KONA Powder Particle J. 32, 75–84 (2015) 9. N. Nomura, Creating metal objects for medical use via selective laser melting. J. Japanese Soc. Biomater. 31(4), 220–227 (2013). 10. EOS GmbH, EOS M290 manuals. 11. TRAFAM, An Introduction to Metal Layered Manufacturing Technologies (Wythup Co. Ltd., 2016), pp. 1–167. 12. J.-P. Kruth, J. Deckers, E. Yasa, R. Wauthle, Assessing and comparing influencing factors of residual stresses in selective laser melting using a novel analysis method. Proc. Inst. Mech. Eng. (2012). doi: https://doi.org/10.1177/0954405412437085. 13. T. Nakano, K. Kaibara, Y. Tabata, N. Nagata, S. Enomoto, E. Marukawa, Y. Umakoshi, Bone 31, 479–487 (2002) 14. T. Ishimoto, T. Nakano, Y. Umakoshi, M. Yamamoto, Y. Tabata, J. Bone Miner. Res. 28, 1170–1179 (2013) 15. T. Nakano, Manufacturing using 3D printers: emphasizing the importance of material and structural parameters. Funct. Mater. 34(9), 5–11 (2014) 16. T. Nakano, H. Fukuda, H. Takahashi, Development of a new powder/solid composite for biomimic implant materials by electron-beam additive manufacturing. Mater. Sci. Forum 879, 1361–1364 (2016) 17. T. Nakano, K. Kuramoto, T. Ishimoto, N. Ikeo, E. Fukuda, Y. Noyama, Energy-absorbing structures and their methods of manufacture, Patent No. 4802277, International Application No. PCT/JP2010/067146, US/EP/ China/Singapore (2011). 18. N. Ikeo, T. Ishimoto, T. Nakano, Novel powder/solid composites possessing low Young’s modulus and tunable energy absorption capacity, fabricated by electron beam melting, for biomedical applications. J. Alloy. Compd. 639, 336–340 (2015) 19. B. Baufeld, O. Van der Biest, R. Gault, Additive manufacturing of Ti–6Al–4V components by shaped metal deposition: Microstructure and mechanical properties. Mater. Des. 31, S106–S111 (2010) 20. P. Nie, O.A. Ojo, Z. Li, Numerical modeling of microstructure evolution during laser additive manufacturing of a nickel-based superalloy. Acta Mater. 77, 85–95 (2014) 21. T. Ishimoto, K. Hagihara, K. Hisamoto, S.-H. Sun, T. Nakano, Crystallographic texture control of beta-type Ti-15Mo-5Zr-3Al alloy by selective laser melting for the development of novel implants with a biocompatible low Young’s modulus. Scripta Mater. 132, 34–38 (2017) 22. S-H. Sun, K. Hagihara, T. Nakano, Effect of scanning strategy on texture formation in Ni-25 at.%Mo alloys fabricated by selective laser melting. Mater. Design 140, 307–316 (2018). 23. S.-H. Sun, T. Ishimoto, K. Hagihara, Y. Tsutsumi, T. Hanawa, T. Nakano, Excellent mechanical and corrosion properties of austenitic stainless steel with a unique crystallographic lamellar microstructure via selective laser melting. Scripta Mater. 159, 89–93 (2018)

26

T. Nakano

24. T. Nagase, T. Hori, M. Todai, S-H. Sun, T. Nakano, Additive manufacturing of dense components in beta-titanium alloys with crystallographic texture from a mixture of pure metallic element powders. Mater. Design 173, 1–10 (2019). 25. T. Nakano, Additive manufacturing technology for developing metallic biomaterials, in Metals for Biomedical Devices, ed by M. Niinomi, 2nd edn (Woodhead Publishing (Elsevier), 2019), pp. 121–125. ISBN: 978–0–08–102666–33.8 26. I. Oonaka, T. Araki, Melt processing. Corona, 40 (1987). 27. T. Nakano, T. Nakamoto, A. Serizawa, T. Inoue, T. Sugawara, N. Shirakawa, M. Yamaguchi, Monocrystalline manufacturing methods, Patent App. 2014–06729 (2014). 28. K. Hagihara, T. Nakano, M. Suzuki, T. Ishimoto, S. Yalatu, S.-H. Sun, Successful additive manufacturing of MoSi2 including crystallographic texture and shape control. J. Alloy. Compd. 696, 67–72 (2017) 29. T. Nakano, M. Kishimoto, D. Furuta, Y. Umakoshi, Effect of substitutional elements on plastic behaviour of NbSi2 based silicide single crystals with C40 structure. Acta Mater. 48, 3465–3475 (2000) 30. K. Hagihara, T. Nakano, H. Maki, Y. Umakoshi, M. Niinomi, Isotropic plasticity of β-type Ti29Nb-13Ta-4.6Zr alloy single crystals for the development of single crystalline β-Ti implants, Scientific Reports, 6 (2016) srep29779. 31. Dongdong Gu, Laser Additive Manufacturing of High-Performance Materials (Springer, 2015), pp. 1–310.

Chapter 2

Laser Processing for Metals Fujio Tsumori

2.1 Processes AM processes can be classified into two main variations: one utilizes polymer materials, whereas the other uses metal powders. UV curable resin and heat-softening or melting materials are employed as AM polymer materials. The former are employed in the stereolithographic process, which is known as the first applied 3D printing process in the world. The latter process is referred to as fused deposition modeling (FDM). In this process, the polymer material is fed as a filament through the heated nozzle to melt and deposit on the stage. The FDM is one of the most popular processes, because the printing machines are so cheap that they can be employed for personal uses, and moreover because available materials such as ABS and nylon are sufficiently strong and tough. Some other processes use polymer powder materials, and inkjet printers are also used for low melting temperature wax or UV curable resin. Nevertheless, metal powders are the most popular materials employed to build metal 3D objects. The powder bed fusion (PBF) process is widely employed, in which a thin layer of metal powder is supplied, and a laser beam or electron beam is scanned on the layer to solidify the designed pattern [1]. If the heated area is sintered, the process is called “selective laser sintering (SLS),” whereas if the area was molten, it is called “selective laser melting (SLM).” The laser power is usually above 200 W, and the laser type is continuous wave oscillation. The electron beam can be used in the same process, and the resulting commercial products are in widespread use [2]. The electron-beam process is performed in a vacuum chamber such that oxidation can be avoided. Another process is referred to as “direct energy deposition.” In this process, the powder material is directly fed in a local heated area by laser or arc plasma. This means that it is enough to prepare the necessary amount of powder, while PBF F. Tsumori (B) Kyushu University, Fukuoka, Japan e-mail: [email protected] © The Japan Welding Engineering Society 2021 S. Kirihara and K. Nakata (eds.), Multi-dimensional Additive Manufacturing, https://doi.org/10.1007/978-981-15-7910-3_2

27

28

F. Tsumori

processes require a much larger amount of powder than that used for products to fill the powder bed. A conventional welding machine can be used for this process, and a metal filament is sometimes used like in the welding process. As mentioned above, powder materials, or partially wire materials, are used for the metal AM process. In these processes, melting and solidification in small areas are repeated by local heating. In the next section, the advantages and shortcomings of this technique are described.

2.2 Advantages of Metal AM Processes The most important advantage of the metal AM processes is that they can output strong metallic parts at the product level directly as “real things” simply by preparing a 3D design drawing on a computer. The conventional machining processes were very difficult or in some cases impossible, as, for example, the products had internal structures. Sometimes, products would be designed to be assembled after fabricating separate parts. With AM processes, the product can be obtained as one part at a time. There are some examples, such as jet engine nozzles and turbine blades, which require an internal cooling flow path in complicated shapes. Since these parts should be heatresistant, the materials applied are usually difficult to machine, such as nickel-based super alloy and titanium alloy. Even if internal structures are not required, it is advantageous that products are manufactured as one part without division, considering the assembling cost. These are the greatest merits when utilizing the AM fabrication processes. Next, the materials are considered. The cooling rate is very fast, since the melting and solidifying process for the metal powder material is performed locally at small heating points, such that fine crystal grains are obtained. Moreover, different material phases could be obtained by rapid heat treatment. This kind of microstructure might be stronger than usual one in some cases, as it could be designed differently from the conventional material developing process. Several studies addressing research and evaluation of AM materials have been reported, along with a rapid increase in AM products.

2.3 Issues and Approaches There are several disadvantages in the metal AM processes. The first is the long processing time and the second is low surface roughness of the products. Moreover, the relatively high cost of the powder materials and process apparatus is a further disadvantage. How these issues are addressed is described in the following.

2 Laser Processing for Metals

29

2.3.1 Processing Time AM processes have been generally used to make prototypes, which is referred to as the “rapid-prototyping process.” However, in recent years, examples show that many AM machines have been introduced to large-scale factories, since sufficiently strong metal products can be obtained for practical applications. In these cases, a large initial investment was made to install many product machines to perform parallel processing. Because the equipment price is high, such investment is relatively large, even if the price of necessary hardware parts such as the laser source has been decreasing. Large companies have also employed strategies to acquire top AM manufacturers, incorporate patents and necessary elemental techniques, and to reduce the cost required in terms of software development. As described above, there is strategy to separate pre- and post-processing from the main laser forming process. Hence, the preparation step and post-treatment step are performed in one place, several laser processing units, as laser processing is a time-consuming step, are prepared and can be operated in parallel. Further, there is another approach where two laser units are installed in the AM system, by which it is possible to halve the scanning time.

2.3.2 Rough Surface and Internal Failures AM depicts a layer-by-layer process, which results in a rough surface of the product, as the layer pitch steps are in line with the contour patterns. Another type of roughness is caused by the metal powder material. As the materials are particulate, the particle structures could remain on the surface of the product. Generally, sieved powder of 50 µm or less is used for the process. As mentioned above, the layer pitch is directly related to the roughness, thus sieving is expected to reduce the pitch. However, the pitch cannot be smaller than the particle diameter. Thus, finer powder material would be preferred. However, another difficulty arises in employing fine powder, since finer powder causes lower fluidity. Consequently, finer powder would be difficult to feed thinly and homogeneously. Finer powder has more surface area, which results in interaction between particles, and finally the fluidity and powder density become lower. The roughness arising from particle size is found on the sidewall surface of the product. Figure 2.1 shows a typical image. The upper surface was formed as a welding bead, while the side surface was covered with powder particles or larger ball structures, which were generated during the melting step. Ball structures are formed by the surface tension of molten metal. These are the same structures formed with sputters of the welding process, and they often cause problems. Figure 2.2 shows an example of the balling process by the laser. In this case, inappropriate conditions were selected to show irregular beads and some rough balls. These balls are usually much larger than material particles.

30

F. Tsumori

Fig. 2.1 Typical rough surface on sidewall of AM-processed part

Fig. 2.2 Example of “balling” in powder bed by laser melting

The coarse balls not only result in a rougher surface, but also lead to more serious problems. Large balls sometimes collide with the blade during the powder feeding phase if the balls are larger than the layer pitch. If the blade is deformed or damaged, the fed layer is disturbed. Figure 2.3 shows a fractured surface of a tensile specimen with long pores, which occurred due to the disturbed fed layer. The failure occurs with a certain probability. Generally, hundreds to thousands of layers are stacked to build one product. In this case, the probability of no defect becomes very low. Even if the failure rate is 0.1%, the probability of producing 1000 layers without defects is about one-third. To overcome this failure, it is necessary to reduce the defect rate to the limit or to employ post-processing such as hot isostatic pressing to eliminate internal pores. However, both of these techniques are costly.

2 Laser Processing for Metals

31

Fig. 2.3 Example of typical failure in SLM part. Horizontally long pores are observed

2.3.3 Cost of Powder Material The fabrication equipment of metal AM processes is relatively expensive. In addition to this initial cost, there is also running costs. The largest running cost is the cost of metal powder. Manufacturers of AM equipment provide metal powder material designed for the machine. The supplier provides the optimized process parameters, so that even unexperienced users can instantly use the equipment by introducing the machine and the powder materials as a set. The powder materials supplied for AM are spherical powders, which are produced by a method like the gas-atomization process. The spherical powder shows high fluidity and highly dense filling characteristic. Powders prepared by crash milling or the electrolytic process exhibit irregular particle shapes, which cause less fluidity. The tap density is likewise lower. Powder materials are generally more expensive than bulk metal materials, and it takes some care to store them. Since powder material has high surface area, metal powder is easily oxidized. The powders should be stored in a dry chamber or sometimes in inert gas or vacuum. The risk of oxidization is related to not only the quality of the products, but also the safety of the factory. The powder could be explosive, and hence it should be treated with extreme caution. These are also classified as dangerous goods class 2 in the Japanese fire service law.

32

F. Tsumori

2.3.4 Consideration of Heat History Each part of the AM product has a different heat history by repetition of local heating and cooling. Sometimes, the microstructure of each part could be different. AM products are formed on a metallic substrate, which means the heating rate at the part near the substrate is higher than that at the upper part, since heat flows easily to the substrate. The difference between lower and upper parts should be considered, even in the fabrication of a simple block shape. Furthermore, the temperature distribution in a more elaborate shape can be more complex. Recently, many studies related to the finite element analysis of heat flow have been reported. In addition, the direction of solidification should also be considered. The melting pool by single laser irradiation is solidified from the bottom to the surface area, so that the solidified structure grows in the upper direction. Therefore, vertically columnar grain structures are often observed. Figure 2.4 shows an example of a

50 m

Building Direction

Fig. 2.4 Example of anisotropic microstructure in AM-processed Inconel 718 [3]

2 Laser Processing for Metals

33

three-dimensional microstructure. The material used was AM-processed Nickelbased superalloy, Inconel 718 [3]. In this specimen, there are vertically long grains of several hundred micrometers. Finer black fiber precipitation structures at the scale of several micros are also observed. Both structures were grown along the vertical building direction. These anisotropic structures resulted in oriented mechanical properties. We should take into account this anisotropy when designing the product. The anisotropic structure can be removed by the post-heating treatment in some cases. A product design that employs anisotropy would also be worth considering.

2.4 Conclusion In this chapter, additive manufacturing (AM) process for metal materials is introduced. Some advantages are explained. Also some issues are shown, and it is described how the issues are addressed.

References 1. L.E. Murr et al., Metal fabrication by additive manufacturing using laser and electron beam melting technologies. J. Mat. Sci. Technol. 28–1, 1–14 (2012) 2. X. Gong et al., Review on powder-based electron beam additive manufacturing technology. Manufacturing Rev. 1, 1–12 (2014) 3. N. Yoshigai et al., Anisotropic mechanical properties of Ni-base superalloy compacts by direct laser forming technology. J. Jpn. Soc. Powder Powder Metallurgy 63–7, 427–433 (2016)

Chapter 3

Selective Electron Beam Melting Yuichiro Koizumi

3.1 Structure of Selective Electron Beam Melting (SEBM) The structure of an selective electron beam melting system resembles that of a Selective Laser Melting (SLM) additive manufacturing system, which is explained in the previous section. Figure 3.1 shows schematic illustrations comparing the SEBM and SLM additive manufacturing systems. First of all, the main component of the apparatus for EBM consists of an electronic optics system and a vacuum chamber. This setup resembles that of an electron microscope (Fig. 3.1a). When compared with a SLM additive manufacturing system, the laser light source is replaced by an electron gun and the galvanometer mirror for laser scanning by a deflecting coil for electron beam scanning. The chamber atmosphere in SLM is mostly inert gas, such as Ar and N2 , whereas the atmosphere must be vacuum to allow electron beam irradiation in SEBM. The vacuum chamber contains a hopper that stores and supplies the metal powder, a comb-shaped stainless plate called a rake that uniformly and thinly spreads, or rakes, powder supplied from the lower part of the hopper, a base plate that is the base of the manufacturing object, and the stage that supports the base plate and adjacent powders and sinks after every layer deposition. Figure 3.2 shows images of model EBM A2X by Arcam (Fig. 3.2a) and the inner view of the chamber (Fig. 3.2b). Besides the main unit, the manufacturing system is composed of a control system (left side of Fig. 3.2a) and a powder recovery system (PRS) that recovers powder that does not melt and solidify.

Y. Koizumi (B) Osaka University, Osaka, Japan e-mail: [email protected] © The Japan Welding Engineering Society 2021 S. Kirihara and K. Nakata (eds.), Multi-dimensional Additive Manufacturing, https://doi.org/10.1007/978-981-15-7910-3_3

35

36

Y. Koizumi

Fig. 3.1 Schematic illustrations of two major powder-bed-fusion-type additive manufacturing systems: a Selective Electron Beam Melting (SEBM), b Selective Laser Melting (SLM)

Fig. 3.2 Apparatuses for EBM. a Main unit of EBM additive manufacturing system (Arcam A2X), b inside of main chamber, c powder recovery system. (Photograph by courtesy of Chibra laboratory, Institute for Materials Research, Tohoku University)

3.2 SEBM Process The manufacturing process is as follows. First, a layer of powder with a thickness of several tens micrometer is on the stage, which is a plate equipped with an elevating mechanism. A metal plate, called the base plate, is placed on top (Fig. 3.3a). The

3 Selective Electron Beam Melting

37

Fig. 3.3 Additive manufacturing process by SEBM: a setting of base plate, b preheating of base plate, c formation of powder layer, d preheating of starting powder, e selective melting, f lowering stage

base plate is usually made of stainless steel. After positioning and leveling, the chamber pressure is adjusted to approximately 10–3 Pa. After the targeted vacuum level is attained, the temperature is increased to the predetermined temperature by scanning and irradiation of a high-energy electron beam on the base plate at high speed. The temperature is monitored by a thermocouple installed on the bottom of the base plate. After the temperature rises to the maximum, the raw material is spread over the base plate (Fig. 3.3b). Here, the rake thinly and uniformly spreads the powder supplied from the hopper. This powder layer is simultaneously heated by a fast-scanning electron beam (Fig. 3.3d). This process is called preheating, and it prevents the charging of powder particles by inducing weak bonding among them. The produced powder layer is irradiated for melting by an electron beam. The beam scans along a 2D slice pattern for each layer, and the powder is selectively fused by melting and solidification (Fig. 3.3e). When scanning of the melting beam along the 2D pattern of a layer is completed, the stage is lowered by the thickness of this layer for lamination (Fig. 3.3f). Another layer of the powder is formed on top, and the 2D pattern is updated to the layer immediately above. The procedure shown in Fig. 3.3c–f is repeated. In this way, an object with the desired 3D shape is obtained. Although the SEBM process resembles SLM, there are several important differences including preheating (Table 3.1). Each difference is discussed in the following.

38 Table 3.1 Comparison between SEBM and SLM

Y. Koizumi Type

SEBM

SLM

System

ARCAM EBM A2 EOS, EOSINT M 280

Heat source

W filament electron gun

Yb-fiber laser

Building space size

200 × 200 × 350 mm3

250 × 250 × 325 mm3

Maximum Power

3500 W

400 W

Beam diameter

0.2–1.0 mm

0.1–0.5 mm

Maximum scanning speed

8000 m/s

7 m/s

Building rate

15–22 mm3 /s

2–8 mm3 /s

(10−1

Atmosphere

Vacuum He-gas)

Preheating temperature

0.5–0.8 T m

Pa

Ar or N2 gas N/A (~90 °C)

3.2.1 Atmosphere The largest difference between SEBM and laser melting is the melting atmosphere. The electron beam, which is the heat source in SEBM, can irradiate only in vacuum atmosphere, while SLM is performed in inert gas atmosphere (Ar, N2 , etc.) as laser irradiation can occur in gas medium. To be precise, SEBM is conducted in a He gas atmosphere of about 10–1 Pa. This apparently suppresses the charging of powder particles and the smoke phenomenon [10], which is explained later. The vacuum atmosphere requirement is sometimes considered as a disadvantage of SEBM. However, manufacturing in vacuum is advantageous in preventing contamination of manufactured objects, including oxidation due to oxygen and moisture [11]. In fact, laser melting apparatus working in high vacuum [12, 13] atmosphere has been developed to suppress oxidation and nitridation during the manufacturing process. Moreover, the chamber is evacuated before inert gas is injected to lower the oxygen concentration of the inert gas atmosphere for the manufacturing process, and therefore the necessity of vacuum is not a disadvantage of SEBM.

3.2.2 Preheating In EBM, the temperature of the base plate and raw powder is typically raised to 50–80% of the melting point on the absolute temperature scale. There are many reasons, however the primary is to prevent charging of powder. The charging of powder particles due to irradiated electrons can result in the electrostatic repulsion between powder particles and most of the powder scatters resembling smoke [10]

3 Selective Electron Beam Melting

39

(“smoke phenomenon” [14, 15]). As a consequence, melting becomes impossible. To avoid this, the base plate is preheated to the designated temperature before forming the first powder layer, and layers of powder before irradiation of the melting beam undergo the same treatment. This preheating is done by scanning a defocused electron beam with several millimeters diameter with a speed higher than 10 m/s. After the predetermined temperature is attained, a powder layer of prescribed thickness of 50–100 µm is piled up. This prevents accumulation of electrons by securing electric conductivity between powder particles and the top surface of the manufactured object and new powder layer or powder particles, and thereby the smoke phenomenon is inhibited. This preheating process is referred to as light sintering, as it has the effect of binding powder particles. Indeed, an object just after manufacturing is buried in lightly sintered matter (Fig. 3.4a). However, the binding among powders in a lightly sintered body is not as strong as imagined from the word “sintering,” but is weak binding that does not result in necking among powder particles. When the binding proceeds until necking is observed, it becomes difficult to crush unmelted powder particles after completion of the shaping process and to extract the manufactured object or reuse crushed powder particles. Preheating is required only to eliminate electric insulation between powder particles, and strong binding is not necessarily required. Increasing the temperature improves the electrical conductivity of the oxide film, which has semiconductor-like properties, that results in a contribution to charge suppression. Effects of preheating other than the suppression of charge include a reduction of thermal stress, fixing of the base plate, and use of a thin and small support. Thermal stress occurs during a manufacturing process with local irradiation by a high-energy electron beam because of the temperature gradients induced with a local increase in temperature. Preheating relieves the temperature gradient and reduces thermal stress. In SEBM, the baseplate is not fixed by screws, but simply placed on the powder bed. Therefore, the formation of lightly sintered bodies by preheating also acts to fix the baseplate. Furthermore, relatively small and light objects can be manufactured without support, because lightly sintered bodies obtained by preheating also play

Fig. 3.4 a Block of lightly sintered powder containing SEBM-built part, b Examples of SEBMbuilt parts: artificial knee joint (left) and acetabular cup (right). (Photograph by courtesy of Chibra laboratory, Institute for Materials Research, Tohoku University)

40

Y. Koizumi

the role of support. Although a support is necessary to manufacture relatively large objects, the thickness and amount of support can be considerably reduced compared with those necessary in laser melting without preheating, because the thermal stress is smaller, and lightly sintered powders act as support (Fig. 3.4b).

3.2.3 Raw Powder Gas-atomized powders of ~40–150 µm diameter are currently used in SEBM. Powder particles with high sphericity and mobility are suitable when preparing a smooth powder bed with uniform thickness. Water-atomized powder particles are generally considered unsuitable for melting manufacturing, as they are not spherical. The ideal particle distribution is uniform, monodispersed, and consists of particles with the same size. The oxide film at the surface should be as thin as possible to suppress the smoke phenomenon in SEBM by securing electric conductivity. This is one reason why gas atomization is preferable to water atomization. Recent findings suggest that gases captured in the powder during the gas atomizing process remain as bubbles in particles of the manufactured product and give rise to fatigue characteristics. Examples where particles manufactured by a method other than gas atomization examples are used are reported, where powder particles were manufactured by plasma atomization [16], centrifugal atomization [17], or the rotating electrode process [18]. The diameter of powder particles used in SEBM, which is 40–150 µm, is larger than the 10–50 µm diameter that is used in laser melting. This difference comes from the characteristics of the processes, and each has their own advantage and disadvantages. In SEBM, powder particles are lightly sintered to secure electric conductivity by scanning a low-density beam that is not strong enough to melt the powder prior to irradiation for melting, as discussed later. Therefore, the manufactured object just after manufacturing is buried in lightly sintered powder (Fig. 3.4a). The lightly sintered powder particles around the manufactured object are crushed in the PRS (Fig. 3.2c) by blasting raw powder. Air is used as the carrier gas of powder particles in this process; therefore, there is a risk of dust explosion when powder particles are too fine. Therefore, the use of powder with relatively large particles, i.e., a diameter >40 µm, is recommended. On the contrary is the surface roughness of the manufactured object, which is a disadvantage of SEBM. Melted powder particles stick locally on the surface of the manufactured object (Fig. 3.5). It is therefore impossible to make the surface smoother than the size of surface powder particles, even if the beam scanning and stage driving mechanism precisions are high. Therefore, the surface obtained from SEBM is rougher than that obtained by laser melting. However, as the inner body is almost perfectly dense, the large size of powder particles does not become problematic when the surface is polished after manufacturing, as shown in Fig. 3.6. In contrast, an advantage of using large particles is the higher manufacturing speed .

3 Selective Electron Beam Melting

41

Fig. 3.5 a SEM image of interface between build part (left) and powder bed (right). (Photograph by courtesy of Chiba laboratory, Institute for Materials Research, Tohoku University)

Fig. 3.6 SEBM-built artificial knee joints after polishing. (Photograph by courtesy of Chiba laboratory, Institute for Materials Research, Tohoku University)

The thickness of one lamination layer in manufacturing is about one to two particles, and thus higher manufacturing speed is attained with larger particles that produce thicker lamination layers.

3.2.4 Selective Melting Process The electric beam for SEBM is obtained similarly to that in an electron microscope: heated electrons are formed by passing electricity over a filament, such as W, LaB6 , and CeB6, and then accelerated. The acceleration voltage in the current apparatus is 60 kV and the maximum power is 3.5 kW. This maximum power is about one order

42

Y. Koizumi

of magnitude larger than the 400 W of the SLM device available in 2010, and it is also few times larger than the 1 kW of the laser melting device that was commercially available in 2016. This is because it is easier to generate electron beams with higher power than laser beams. The energy efficiency is higher in electron beams compared to laser beams. More than 90% of the supplied electric energy can be converted to electron beam energy, while only 10–20% can be converted to laser beam energy. Hence, most supplied electric energy may be converted into electron beam energy. Moreover, increasing the output is comparatively easy to attain and comes at low cost. There is a large difference in the beam energy absorbance in metals. Electron beams can be scanned by a magnetic field produced by deflection coils, and a high-speed scanning of 8000 m/s is possible in currently available devices. This means that the beam irradiation position can be instantaneously shifted, and molten pools can be formed at multiple locations at the same time by instantaneously irradiating a highenergy beam at different points. This is effective in increasing manufacturing speed. In contrast, a laser bean is scanned by a galvanomirror, and the maximum scanning speed is limited to around 10 m/s because of inertia arising from the mirror mass. One may presume that a high-speed scanning of 8000 m/s is not necessary, because the speed of a molten pool is 1 m/s at most. However, forming multiple molten pools simultaneously and moving them independently at the same time is a process available with a device comprising multiple electron guns. Electron beams enable manufacturing processes that are impossible with laser additive melting, because large output is easily obtained, the energy absorption rate in metals is high, and high-speed scanning is possible (Fig. 3.7).

Fig. 3.7 Schematic diagram of electron beam scanning strategy for simultaneous motion of multiple molten pools utilizing the instantaneous motion of electron beam spots

3 Selective Electron Beam Melting

43

3.2.5 Stage Lowering Step The stage lowering step in SEBM is typically 70 µm, and it is changed in the range from 50 to 90 µm in accordance to the powder size distribution. This step size is larger than that of laser additive melting, and the reason is a difference in the powder size. As mentioned before, the powder size used in SEBM is ~40–150 µm, and the average size is 60–70 µm. In contrast, the powder size used in laser melting is 10– 40 µm. It has been proposed that taking the stage lowering step to be nearly equal to the average powder sizes is suitable to obtain satisfactory manufactured objects [19]. However, powder particles larger than the step size may be excluded from raking if the stage lowering step is chosen as the average powder size of 70 µm, when the powder particle size distributes between 40 and 150 µm. Indeed, when there are few laminated layers, such particles with large size are removed. However, the thickness of the solid metal layer accumulated by melting and solidification of the powder layer is roughly half of the thickness of the powder layer (Fig. 3.8). Therefore, as the lamination is repeated, the thickness of the powder layer becomes twice as large as the stage lowering step, and particles with the size at the upper limit of the distribution remain in the powder layer [19].

Fig. 3.8 Examples of parts made with SEBM. (Photograph by courtesy of Chiba laboratory, Institute for Materials Research, Tohoku University)

44

Y. Koizumi

3.3 Conclusions The principle of SEBM apparatus is explained above. Although the basic manufacturing principle is similar to that of SLM, there are various differences that originate from the difference in beam properties. Two categories of differences can be identified: one is intrinsic, due to the differences in physical properties of electron and laser beams, such as different beam energy absorption rate in metals and the need of charge prevention. The other is due to current technical limitations, which include beam power and size of powder used. The former are universal and will not change in the future, whereas the latter may change with future development of techniques. For example, the differences in surface roughness that arise from the difference in the size of used powder particles could be resolved by development of a procedure where the surface is polished during manufacturing. There are discussions about which melting method is superior. However, it is difficult to determine which is better, as they both have advantages and shortcomings. Hence, it is necessary to choose the appropriate method after extensive understanding of the current characteristics. The following section discusses examples of applications. Acknowledgements The author received great cooperation from the people of Chiba laboratory including Professor Akihiko Chiba, Institute for Materials Research, Tohoku University. In addition, some of the contents introduced in this paper were implemented under the support of the Ministry of Education, Culture, Sports, Science and Technology (MEXT) Regional Innovation Cluster Program.

References 1. I. Gibson, D.W. Rosen, B. Stucker, Additive Manufacturing Technologies (Springer, New York, 2010) 2. H. Kyogoku et al., Technology Research Association for Future Additive Manufacturing (TRAFAM) (Introduction to metal Additive Manufacturing, Technology Research Association for Future Additive Manufacturing (TRAFAM), 2016) 3. A. Chiba, Additive manufacturing using electron beam melting (EBM) technique and EBM metallurgy. J. Smart Process. 3, 152–157 (2014) 4. A. Chiba, Fabrication of metal structural parts by electron beam melting. Die Mold Technol. 29, 24–27 (2014) 5. A. Chiba, Exploring the possibility of manufacturing pioneered by 3D printing. Mech. Eng. 63, 64–69 (2015) 6. A. Chiba, Microstructure of alloys fabricated by additive manufacturing using electron beam melting. J. Soc. Instrument Control Eng. 54, 399–404 (2015) 7. A. Chiba, Recent trends of additive manufacturing using electron beam melting. 82, 624–628 (2016) 8. T. Takashima, Y. Koizumi, Y. Li, K. Yamanaka, T. Saito, S. Sun, A. Chiba, Effect of building position on phase distribution in Co-Cr-Mo alloy additive manufactured by electron-beam melting. Mater. Trans. 57, 2041–2047 (2016) 9. L.D. Harwell, M.L. Griffith, D.L. Greene, G.A. Pressly, Energetic additive manufacturing process with feed wire, US Patents US6143378 A (2000). 10. M. Kahnert, S. Lutzmann, M.F. Zaeh, Layer formations in electron beam sintering, in Solid Freeform Fabrication Symposium Proceedings, vol 18 (2007), pp. 88–99

3 Selective Electron Beam Melting

45

11. H.K. Park, Y.K. Ahn, B.S. Lee, K.H. Jung, C.W. Lee, H.G. Kim, Refining effect of electron beam melting on additive manufacturing of pure titanium products. Mater. Lett. 187, 98–100 (2017) 12. M. Hagiwara, Y. Horiba, M. Sasa, S. Nakano, T. Shimizu, K. Matsuzaki, N. Sato, Development of selective laser metal melting equipment with high vacuum atmosphere. J. Jpn. Soc. Powder Powder Metallurgy 61, 223–226 (2014) 13. S. Nakano, M. Hagiwara, T. Shimizu, Y. Horiba, N. Sato, K. Matsuzaki, M. Sassa, Novel selective laser melting solution for metal additive manufacturing using vacuum and a quasi continuous wave laser, in Proceedings of International Conference on Leading Edge Manufacturing in 21st century: LEM21, vol. 7 (2013), pp. 419–422 14. T. Horn, Material development for electron beam melting, https://camal.ncsu.edu/wp-content/ uploads/2013/10/Tim-Horn-2013CAMAL.pdf, Accessed 24 March 2017 15. A. Chiba, Development into new manufacturing industries with electron beam additive technique, in Proceedings of Japan Society of Mechanical Engineering (2015). 16. M. Boulos, Plasma power can make better powders. Met. Powder Rep. 59, 16–21 (2004) 17. W.J. Sames, F.A. List, S. Pannala, R.R. Dehoff, S.S. Babu, The metallurgy and processing science of metal additive manufacturing. Int. Mater. Rev. 61, 315–360 (2016) 18. W.J. Sames, F. Medina, W.H. Peter, S.S. Babu, R.R. Dehoff, Effect of process control and powder quality on Inconel 718 produced using electron beam melting, in Proceedings of the 8th International Symposium Superalloy 718 and Derivatives (2014), pp. 409–423. 19. K. Shinzawa, A study on electron beam melting of TiAl-based alloy, Tohoku University Ph.D thesis (2017)

Chapter 4

Current Research and Development Naoyuki Nomura

4.1 Investigation Method Literature on electron beam melting was surveyed with respect to the number of papers, country of the corresponding author for each publication, topic, and materials using a scientific database (Web of Science, Tomson Reuter) from 2007 to 2016. When a paper studied multiple topics (e.g., microstructure and mechanical properties) or materials (e.g., Ti alloy and Ni alloy), these studies were counted separately.

4.2 Results and Discussion Figure 4.1 shows the increase in the number of research papers on PBF-EB. There were less than 10 studies per year before 2010, a number that was increasing slightly until 2014. However, this trend changed to a remarkable increase after 2014. This was related to the release of Arcam A2X in 2007, which realized the fabrication of large metallic parts and became popular with the growing interest in additive manufacturing. The technical terms of this process, such as rapid prototyping, freeform fabrication, layered manufacturing, and direct metal fabrication, appeared with increasing frequency. However, in ASTM standard F2792-10, the term additive manufacturing was defined as the “process of joining materials to make objects from 3D CAD model data, usually layer upon layer, as opposed to subtractive manufacturing methodologies, such as traditional machining.” Murr’s group used this term in their paper in 2010 [1]. The term selective electron beam melting also appeared in the literature. This may be due to the fact that simply using the term electron beam melting may cause confusion regarding the process for purifying metals or crystal growth, and the N. Nomura (B) Tohoku University, Sendai, Japan e-mail: [email protected] © The Japan Welding Engineering Society 2021 S. Kirihara and K. Nakata (eds.), Multi-dimensional Additive Manufacturing, https://doi.org/10.1007/978-981-15-7910-3_4

47

48

N. Nomura 70

64

Number of research paper

60

54

50 40

35

30 21

20

17 10

10 2

0

14

10

5

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Fig. 4.1 Increase in the number of research papers on PBF-EB

usage of the term enables comparison with the term selective laser melting, which explains powder bed fusion using a fiber laser. Figure 4.2 shows the ratio of the numbers of research papers on PBF-EB by major countries from 2007 to 2016. The USA accounted for one-third of the total number, followed by China and Germany. These three countries were found to be leading countries on PBF-EB. Notably, China has rapidly increased its number over the last Fig. 4.2 Ratio of the number of research papers on PBF-EB by major countries from 2007 to 2016

Singapore 3% France 3% Italia 3%

Others 8% USA 31%

Australia 4% Japan 4% UK 5% Sweden 6% Germany 15%

P. R. China 18%

4 Current Research and Development Fig. 4.3 Materials investigated in the research papers on PBF-EB

49 Co alloys 2% Cu alloys 2% Others 2%

Fe alloys 3%

TiAl 6% Ni alloys 12% Ti and Ti alloys 73%

several years, and the number is now close to that of USA. Sweden, recognized as the cradle of PBF-EB, has produced a few papers recently and remains stable. In addition to China, Japan, Australia, and Singapore carry out PBF-EB research in Asia. Figure 4.3 presents the materials investigated in PBF-EB research papers. Studies on Ti and Ti-6Al-4V alloys accounted for about three-fourths of the total number. Contamination with nitrogen and oxygen, which affects mechanical performance, can be avoided by performing this process in a vacuum. This process is also favorable to Ni-based superalloys containing reactive elements for carbide dispersion. Dense builds without contamination are required for jet engine parts consisting of Ni-based superalloys for aerospace applications. PBF-EB research subjects include Co-based alloys for biomedical applications [2], Fe-based alloys for oxide-dispersion strengthened (ODS) composites [3], and Cu-based alloys for electrically conductive materials [4]. The development of novel materials has attempted the use of PBF-EB with WC-based composites [5]. Figure 4.4 shows the scientific subjects of PBF-EB research papers. Porous structures with complex shapes unique to additive manufacturing accounted for 19% of the total subjects in the area. Their features were evaluated from the viewpoint of tensile or compressive mechanical properties in addition to fatigue [6]. Many studies describe the development of unique microstructures in some alloys, such as columnar grains and their changes. Sun et al. [7] reported that the preferential orientation parallel to [001] was achieved along the building direction, and the mechanical properties were governed by the anisotropy of the build. Custom-made medical devices that can facilitate medical treatments suitable to each patient constitute major applications of this research. Therefore, biological research has been carried out from the viewpoints of cellular and/or biological

50

N. Nomura Building Powder parameters 3% 3%

Others 5%

Surface morphology 3% Mechanical properƟes 23%

Defects 4% Post processing 4% SimulaƟon 6%

Cellular and biological response 9%

Microstructure 21%

Porous structure 19%

Fig. 4.4 Scientific subjects in the research papers on PBF-EB

responses. Regis et al. [8] fabricated a porous titanium trabecular cup consisting of a hexagonal structure by PBF-EB and introduced the mechanical properties and cellular compatibility. They showed that the porous builds had a low Young’s modulus (1.2 GPa) and discussed their applicability to large strains. Porous structures, mechanical properties, and microstructures account for approximately 70% of the total subjects researched. Mechanical property predictions for porous bodies have also been performed and compared with the experimental results [9]. Functionally graded materials (FGMs) were studied using structural modeling and microstructural changes along the building direction [10, 11]. The transition of PBF-EB research will be discussed in the following. From 2007 to 2010, the fabrication of a dense or porous Ti-6Al-4V alloy has been introduced, and its mechanical properties were reported. Interestingly, the evaluation of the cellular or biological response of the builds was reported during the same period. This suggests that the investigation for medical applications was carried out simultaneously with the mechanical and biological aspects. In 2011 or later, research on Ni-based superalloys, TiAl, and Co-based alloys appeared. The design of the structure and reuse of the powder for PBF-EB have been investigated. From 2013 onward, research on post-processing, such as hot isostatic pressing (HIP) and heat treatment, has been reported to improve the properties of the builds [12]. Recently, melting behavior during PBF-EB was analyzed [13] and defects in the builds were evaluated with computed tomography using synchrotron radiation beams [14].

4 Current Research and Development

51

4.3 Conclusion The research areas related to PBF-EB have been expanding, and the phenomena during building remain the focus of analysis. This analysis is fed back to building parameters and contributes to the fabrication of builds with fewer defects. Technologies related to the in situ observation of melting behavior and its feedback system have been developed, as well as a method for detecting thermal stress and its reduction system. The design of lightweight parts using a topological optimization method is effective for the aerospace and automobile industries. Moreover, a specialized powder for additive manufacturing should be developed with the requirement of reusability and low-cost processing. The development of these technologies will expand the application range of PBF-EB from the aerospace and medical fields to general industries in the future.

References 1. L.E. Murr, S.M. Gaytan, A. Ceylan, E. Martinez, J.L. Martinez, D.H. Hernandez et al., Acta Mater. 58, 1887–1894 (2010) 2. L.E. Murr, K.N. Amato, S.J. Li, Y.X. Tian, X.Y. Chen, S.M. Gaytan et al., J. Mech. Behav. Biomed. Mater. 4, 1396–1411 (2011) 3. R. Guo, L. Zeng, H. Ding, T. Zhang, X. Wang, Q. Fang, Mater. Design 89, 1171–1180 (2016) 4. S.J. Raab, R. Guschlbauer, M.A. Lodes, C. Körner, Adv. Eng. Mater. 18, 1661–1666 (2016) 5. H. Peng, C. Liu, H. Guo, Y. Yuan, S. Gong, H. Xu, Mater. Sci. Eng. A 666, 320–323 (2016) 6. M. Jamshidinia, L. Wang, W. Tong, R. Ajlouni, R. Kvacevic, J Mater. Process. Tech. 226, 255–263 (2015) 7. S.-H. Sun, Y. Koizumi, S. Kurosu, Y.P. Li, H. Matsuoto, A. Chiba, Acta Mater. 64, 154–168 (2014) 8. M. Regis, E. Martin, L. Fedrizzi, M. Pressacco, MRS Bull. 40, 137–144 (2015) 9. G. Campoli, M.S. Borleffs, S.A. Yavari, R. Wauthle, H. Weinans, A.A. Zadpoor, Mater. Design 49, 957–965 (2013) 10. W. van Grunseven, E. Hernandez-Nava, G.C. Reilly, R. Goodall, Metals 4, 401–409 (2014) 11. X. Tan, Y. Kok, Y.J. Tan, M. Descoins, D. Mangelinck, S.B. Tor et al., Acta Mater. 97, 1–16 (2015) 12. B. Ruttert, M. Ramsperger, L.M. Roncery, I. Lopez-Galilea, C. Korner, W. Theisen, Mater. Design 110, 720–727 (2016) 13. J. Gockel, N. Kingbeli, S. Bontha, Metall. Mater. Trans. B 47, 1400–1408 (2016) 14. N.V.Y. Scarlet, P. Tyson, D. Fraser, S. Mayo, A. Maksimenko, J Syncrotron. Radiat. 23, 1006– 1014 (2016)

Chapter 5

Stereolithography Soshu Kirihara

5.1 Three-Dimensional Printing 5.1.1 Laser Scanning Stereolithography Geometric patterns in 3D models are modeled by a computer-aided design application. These graphic models are converted automatically into the stereolithography format and sliced into a series of 2D cross-sectional planes of 50-µm uniform layer thickness. The numerical data are then transferred into the stereolithography equipment, creating raster patterns for laser scanning. Figure 5.1 portrays the schematic illustrations of the fabrication process [1, 2]. Photosensitive acrylic resin includes ceramics particles which are 200 nm in diameter at 40% volume fraction. The paste is spread on a flat metal stage by using mechanical knife edge. The thickness is controlled automatically at the same value of 50 µm in the model slicing pitch. An ultraviolet laser of 355 nm wavelength is scanned on the ceramics slurry. Crosssectional planes were formed through the laser drawing at 2000 mm/s scanning speed and 5 µm edge part accuracy. The laser beam is adjusted to a 100 µm spot size and 300 mW irradiation power. After formation of the solid pattern, the elevator stage moved downward 50 µm in the layer thickness, and subsequently the next cross-section is stacked. Three-dimensional structures were fabricated by stacking all two-dimensional layers. The part accuracies of green bodies can be measured and observed using a digital optical microscope. The formed models are dewaxed at 600 °C for 2 h with a heating rate of 1.0 °C/min in air atmosphere, and full ceramic components are obtained after sintering.

S. Kirihara (B) Osaka University, Osaka, Japan e-mail: [email protected] © The Japan Welding Engineering Society 2021 S. Kirihara and K. Nakata (eds.), Multi-dimensional Additive Manufacturing, https://doi.org/10.1007/978-981-15-7910-3_5

53

54

S. Kirihara

Fig. 5.1 Schematic illustration of laser scanning stereolithography

5.1.2 Micro-patterning Stereolithography Three-dimensional micro-patterns are designed using the computer graphic software. The designed models are converted into the stereolithography file format and sliced into the series of 2D cross-sectional data of 10-µm layer thickness. These data are transferred into the micro-stereolithography equipment to automatically create bit map images for micro-pattering. Figure 5.2 shows a schematic illustration of the micro-stereolithography system [3, 4]. Photosensitive acrylic resins, including ceramic nanoparticles of 200-nm average diameter at 40% in volume content are supplied on a glass substrate from a dispenser nozzle using air pressure. This paste is spread uniformly using a mechanically controlled knife edge. The thickness of each layer was set at 10 µm. Two-dimensional solid patterns are obtained on the slurry surface by light-induced photopolymerization. High-resolution images were achieved using a digital micro-mirror device. In this optical device, 1024 × 768 micro-aluminum mirrors of 14 µm in edge length were assembled. Each mirror can be tilted independently by electrostatic actuation. The ultraviolet beam of 405nm wavelength was introduced into the digital micro-mirror device, and the crosssectional image is reduced at 1/5 through an objective lens set and concentrated into the exposing area 1.3 × 1.7 mm2 in size. Through the layer stacking under the computer control, the acrylic resin component with the ceramic particles dispersion is obtained. The composite precursor is dewaxed at 600 °C for 2 h while heating and holding time in the air atmosphere. The full ceramic micro-components are obtained through the sintering heat treatment.

5 Stereolithography

55

Fig. 5.2 Schematic illustration of micro-patterning stereolithography

5.1.3 Ultraviolet Laser Lithography In our evolved stereolithography method, the obtained composite precursors were dewaxed and sintered to obtain dense components without structural deformation. Linear shrinkage corresponding to the particle dispersion amounts was controlled along the optimized heat treatment patterns. Recently, we discovered that an ultraviolet laser beam with a short wavelength and high power could propagate between dispersed fine particles and remove resin menstruum via heat decomposition [5, 6]. Through the optimization of the laser beam spot size, scanning speed, and irradiation power, we sintered the remaining ceramic powder and increased the relative density. The 3D models were automatically converted into a stereolithographic file format with a modeling application and sliced into a series of cross-sectional layers of the 2D patterns. The slicing pitch was defined at 10 µm. The process equipment of stereolithographic additive manufacturing, as shown in Fig. 5.1, is employed in this ultraviolet lithography. The ceramic particles of 300-nm average diameter were dispersed into an acrylic resin at a volume fraction of 50%, without photo- or heatcuring. The obtained paste was spread on a flat glass stage using a mechanical knife edge. The thickness was controlled automatically according to the model slicing pitch. The paste surface was scanned at a speed ranging from 100 mm/s with a laser beam working at 355 nm to create cross-sectional planes through resin dewaxing and powder sintering, involving propagations and absorptions of the ultraviolet ray. The spot size was adjusted to 10 µm, and the irradiation power was varied from 800 mW. After the elevator stage moved down with respect to the layer thickness, the next cross-section was formed and joined to the solid object. The 3D components were

56

S. Kirihara

Fig. 5.3 Cross-sectional schematic of laminated layers through ultraviolet laser lithography

fabricated by stacking all 2D layers. Figure 5.3 shows a schematic of the layer lamination in the ultraviolet laser lithography. Through the ultraviolet laser drawing on the resin paste containing ceramic particles at a volume fraction of 50%, the laminated layer with 20 µm spreading thickness became thinner, reaching a half pitch of 10 µm after dewaxing and sintering. The linear shrinkage of 50% was observed solely for the Z-axis direction, according to the volume fraction of the particle dispersion. The resin dewaxing and the powder sintering were realized via laser beam ablation toward the spread paste without shrinkage for the X- and Y-axes directions. The reaction force of resin vaporization was considered to compress the remaining particles in the direction perpendicular to the laminated layer.

5.2 Artificial Bone 5.2.1 Geometric Formation Tissue scaffolds are required to repair bone defects resulting from illnesses. To encourage osteoconductivity and tissue regeneration, prosthetics mimic bone porosity and optimized flow behaviors are very important. Various techniques to fabricate the artificial bone structures have been investigated, for example, polycaprolactone scaffolds with variable pore size and porosity from 63 to 79% using selective laser sintering [7], modified hydroxyapatite scaffolds by the printing process [8], and periodic micro-arrays using direct ink writing [9]. The scaffold structure requires suitable porosity and pore size to foster tissue regeneration in the human body [10]. Natural bone with graded porosity from 50 to 90% in the volume fraction is mimicked in the fabrication of the artificial bone models. Conventional artificial bones have almost 75% porosity [11, 12]. In this section, creation of the novel artificial bones composed of hydroxyapatite ceramics with effective biocompability and high mechanical strength is demonstrated [13], along with the graded porous structures composed of four coordinate lattices using the laser scanning stereolithography. Evaluating osteogenesis requires long-term clinical experiments. As an alternative

5 Stereolithography

57

to clinical experiments, biofluid flow behaviors were investigated in different types of scaffold structures with the same porosity.

5.2.2 Coordination Number Dendritic lattice structures in the biological scaffolds with 4, 6, 8, and 12 coordination numbers are designed, as shown in Fig. 5.4. The porosity of scaffold models can be controlled in the range from 50 to 90% by adjusting aspect ratios of the rod length to the diameter, as shown in Fig. 5.5. The porosity of all skeletal structures was 75%, which is similar to the porosity of a human bone. These scaffolds have perfect interconnected pores. Fluid circulation in various dendrite scaffolds was visualized with the fluid dynamic solver. Flow velocity in the spatial grids in the scaffold models was calculated through computer fluid dynamics (CFD). The following values of these parameters were used in this simulation [14, 15]. The fluid phase was represented as an incompressible Newtonian fluid with a viscosity of 1.45 × 10−3 Pa s. The inlet velocity applied to the scaffolds was constant at 0.235 mm/s, and the pressure was zero at the outlet. No-slip surface conditions were assumed. Figure 5.6 shows the streamline behavior in the dendrite scaffolds. Figure 5.6a shows a scaffold structure with coordination number 4. The profile indicates that inordinate flow at a low velocity was obtained. This also indicates simulated biofluid

Fig. 5.4 Designed models of dendrite scaffolds. The lattices (a–c) have nodal points with 4, 6, 8, and 12 coordination numbers, respectively

58

S. Kirihara

Fig. 5.5 Numerical variation of porosity percentages toward lattice aspect ratios

Fig. 5.6 Fluid flow behaviors in dendritic scaffolds visualized by CFD. Streamlines (a–d) are formed in lattices with 4, 6, 8, and 12 coordination numbers, respectively

flow to the whole structure, which is expected to provide active tissue regeneration. The fluid velocity in the coordination number of 6 scaffolds is the highest at above 1.0 mm/s, as shown in Fig. 5.6b. There are no blockades from the inlet to the outlet, and the flow becomes linearly stable. The high fluid velocity area above 1.0 mm/s in the scaffold is subjected to shear stress, which can assume the difficulty cell attachment on the scaffolds surface [16]. In the case of a coordination number of 8 scaffold, random fluid flow and high velocity are exhibited in some parts of the

5 Stereolithography

59

structure, as shown in Fig. 5.6c. In the coordination number 12 structure, moderate velocities and flow behaviors are exhibited, as shown in Fig. 5.6d. There are no active flows in some areas because of many rods in the architecture. Smooth fluid flows and propagations are visualized in the four coordination number scaffold without obstacle lattices compared with the larger number of rod coordination.

5.2.3 Biological Scaffold The modulated scaffold models with the graded porous structure are shown in Fig. 5.7 [17–20]. The porosity is distributed gradually through modulations of the aspect ratio in the four coordination number lattices, as shown in Fig. 5.7a. The acryl scaffold with hydroxyapatite particles is shown in Fig. 5.7b. A stereoscopic image was obtained by digital optical microscopic (DOM) observations. Micrometerorder ceramic lattices are successfully fabricated by laser scanning stereolithography. Photosensitive acrylic resin including the hydroxyapatite particles of 10-µm diameter at 45 vol% was used. The porosity of the scaffold form was approximately 75%. The part accuracies of the lattices were measured under 50-µm size difference. The formed precursor was dewaxed at 600 °C for 2 h at a heating rate of 1.0 °C/min and sintered at 1250 °C for 2 h at a rate of 5 °C/min in air atmosphere [21, 22]. The relative density of the sintered hydroxyapatite lattice was measured at 98% by the Archimedean method. The sintered scaffold model of the hydroxyapatite ceramic with the graded lattices is shown in Fig. 5.7c. The linear shrinkage ratios for horizontal and vertical axes were 23% and 25%, respectively. Smaller lattice structures Fig. 5.7 Stereoscopic DOM image of four coordinate lattices with graded porous structures fabricated by laser scanning stereolithography. a Graphic model, b acrylic lattices with hydroxylapatite, and c sintered scaffold

60

S. Kirihara

Fig. 5.8 Microstructure of sintered hydroxylapatite lattice observed by SEM

could be effectively obtained through the controlled body shrinkages on the optimized sintering process. The microstructure of the sintered scaffold is observed by the scanning electron microscope (SEM), as shown in Fig. 5.8. Cracks and pores were not observed. The grain size had a ~4-µm diameter and the relative density reached 98%. The formed scaffold is considered to be effectively biocompatible and has high mechanical strength.

5.3 Solid Oxide Fuel Cell 5.3.1 Energy Generation Solid oxide fuel cells (SOFCs) are promising candidates of next-generation energy conversion systems due to their higher power density and energy efficiency. Yttriastabilized zirconia (YSZ) with added nickel (Ni) has many desirable properties for the SOFC anode, such as high electronic/ionic conductivities and chemical/mechanical stabilities at high operation temperatures. At the anode site, fuel gas diffusions and electrochemical reactions on the electrode surfaces which are composed of YSZ/Ni/gas triple phase boundary (TPB) proceed simultaneously. In addition, activation and diffusion over potentials should cause a decrease in SOFCs’ operating voltages and energy efficiencies. Therefore, anode microstructure design is essential to achieve SOFCs’ higher performances and miniaturizations. Porous YSZ-Ni anodes have been fabricated to realize large surface areas and high activation of electrode reactions [23–25]. Relationships between dispersion ratios of YSZ and Ni particles and SOFCs output characteristics such as energy densities and over potentials were investigated [26–28]. In addition, numerical analyses dealing with material transportations in porous media were attempted to evaluate electrode structures [29, 30]. Moreover, it was confirmed that random vacancy structures in

5 Stereolithography

61

traditional porous anode materials obstruct fuel and produce H2 O gas diffusion and electrode reactions. In this section, the SOFC anode structure was optimized by computer simulation using the finite element method (FEM). Solid electrodes with wide surface areas and smooth fluid permeability were fabricated successfully by using microstereolithography. The dendritic structures constructed from spatially propagating micrometer-order ceramic rods are thought to be a desirable electrode structure, because they have large surface areas and cyclical vacancy structures. Aspect ratios of rod length to diameter were optimized to exhibit the maximum surface area. Subsequently, streamlines, velocities, and stress distributions in the dendritic structures were visualized to evaluate fluid permeability and mechanical strength. The dendritic structure with large surface area and smooth gaseous diffusion property is expected to activate electrode reactions and lower activation and diffusion over potentials.

5.3.2 Porous Structures The electrode structure was optimized through computer simulations. Dendritic lattices of 100 µm in the lattice constant constructed from micrometer-order rods with coordination numbers of 4, 6, 8, and 12 were designed using computer graphic software. Aspect ratios of rod lengths to diameter were varied from 0.75 to 3.00 to investigate the relationships between aspect ratios and surface areas. Relationships between surface areas and aspect ratios determined by rod diameters and lengths are shown in Fig. 5.9. Each dendritic lattice showed maximum surface areas when the aspect ratios were 0.90, 1.17, 2.18, and 2.34, respectively. The dendritic lattice with coordination number of 12 and aspect ratio of 2.18 was verified to exhibit the largest

Fig. 5.9 Calculated variations of dendrite surface areas toward lattice aspect ratios and coordination numbers

62

S. Kirihara

surface area, and the electrode texture is expected to increase TPB points and lower activation potential in the electrode. Gaseous fluid permeability and stress distributions in dendritic structures were simulated by FEM calculations. The analysis model is shown in Fig. 5.10. The number of the arranged dendritic lattice unit cells was 5 × 5 × 1. The static pressures at the inflow and outflow were 1.01 atm and 1.00 atm, respectively. In the mechanical analysis, bottoms of the dendritic structures were fixed. Streamlines and velocities in dendritic structures are shown in Fig. 5.11. Smooth streamlines according to cyclical

Fig. 5.10 Designed model of dendritic lattice for computer simulations

Fig. 5.11 Gaseous fluid behaviors in the dendritic structure visualized by FVM. Streamlines (a– d) are formed in lattices with 4, 6, 8, and 12 coordination numbers, respectively

5 Stereolithography

63

Fig. 5.12 Stress distribution in 12-coordinate lattice calculated by FEM

vacancies were exhibited in the dendritic structure with coordination numbers of 6, 8, and 12. The prompt fuel gas flows can realize reduction of diffusion over potential. However, there were rectilinear streamlines and large flow velocity differences in the dendritic structure with 6 and 8 coordination numbers. These could be the cause of inhomogeneous fuel gas diffusion, for example, back water or vortex. In turn, smooth and homogeneous fluid diffusion was indicated in the dendritic structure with coordination number of 12. In the SOFC anode, H2 O produced in electrode reactions must be discharged from the electrode promptly to avoid lowering concentrations and partial pressures of fuel gases. A stress distribution applied by the fluid flow in the dendritic structure with coordination number of 12 is shown in Fig. 5.12. A stress concentration was not observed, and the dendritic structure can exhibit high mechanical strength by isotropic lattice structures.

5.3.3 Dendritic Electrode The optimized dendritic structure is formed using micro-stereolithography. Slurry paste of photosensitive acrylic resin with nanometer-sized YSZ and Ni particle dispersions was supplied on a substrate from the dispenser nozzle employing air pressure and spreads uniformly at 5-µm layer thickness by a mechanical knife edge. The fabricated dendritic structure composed of YSZ and nickel oxide NiO is shown in Fig. 5.13 [31–34]. The stereoscopic image was obtained by DOM observations. Micrometer-order ceramic lattices with coordination number of 12 were successfully formed. These composite precursors were dewaxed at 600 °C for 2 h and sintered at 1400 °C for 2 h in air atmosphere. The lattice constant of the sintered sample was 98.5 µm. Figure 5.14 shows the microstructure of YSZ-NiO electrode surface observed by a SEM. The YSZ and NiO particles were successfully connected

64

S. Kirihara

Fig. 5.13 Stereoscopic DOM image of 12-coordinate lattices of YSZ and NiO cermet dendrites fabricated by micro-patterning stereolithography

Fig. 5.14 Microstructure of YSZ-NiO cermet lattice observed by SEM

and homogeneously distributed. Micro-voids or pores were not observed. In X-ray diffraction (XRD) patterns of the sintered sample, diffraction peaks of YSZ and NiO were clearly shown. The ideal TPB could be created in the sintered dendritic lattices. The fabricated dendritic electrode with large surface area and smooth fluid permeability is considered to effectively activate electrode reactions and lower activation and diffusion over potentials.

5.4 Photonic Crystal 5.4.1 Bandgap Formation Photonic crystals with periodic arrangement structures represent a dielectric medium that can prohibit electromagnetic wave propagation caused by Bragg reflection and exhibit forbidden gaps in the transmission spectra [35, 36]. Bandgap profiles are controlled arbitrarily via theoretical modulations of cavity introductions into the

5 Stereolithography

65

dielectric arrangements. Electromagnetic waves with typical wavelengths comparable to the cavity sizes resonate in artificial crystal defects, and electromagnetic energy localizations and amplifications can form permission modes of transmission peaks in photonic bandgaps. Three-dimensional photonic crystals with diamond structures are regarded as ideal dielectric patterns that exhibit complete bandgaps, attenuating incident electromagnetic waves from all directions [37]. Diamond photonic crystals can be applied to resonators, wavelength filters, directional antennas, etc. [38, 39]. The dielectric diamond micro-lattices with four coordination numbers can be fabricated via stereolithographic additive manufacturing and powder sintering [40, 41]. Micro-defects in the form of point- or plane-shaped cavities were introduced to control the terahertz (THz) waves with micrometer-order wavelengths [42, 43]. THz waves are expected to detect micro-cracks in material surfaces and structural defects in electrical circuits by fine wave interference. They are used to analyze cancer cells in human skin and toxic bacteria in natural foods through high-frequency excitations [44–48].

5.4.2 Periodic Arrangement A diamond structure unit cell shown in Fig. 5.15 was designed using computer graphics software. To obtain an ideal bandgap, the aspect ratio and volume fraction of the dielectric lattices were adjusted to 1.5% and 33%, respectively. The electromagnetic band diagram of the designed diamond structure was calculated along the symmetry lines in a Brillouin zone via the plane wave expansion (PWE) method, as shown in Fig. 5.16 [49]. A dielectric constant of 100 was selected as the material Fig. 5.15 Designed computer graphic model of a unit cell in photonic crystal

66

S. Kirihara

Fig. 5.16 Electromagnetic band diagram calculated by PWE

parameter of the titania lattices. One hundred and twenty four plane waves propagated into the imaginarily limitless periodic arrangement of dielectric lattices. By modulating the lattice aspect ratio from 1.0 to 2.0, the widest perfect bandgap of 1.5 was obtained. Figure 5.17 shows the frequency variations of the perfect bandgaps according to the lattice constants. The upper and lower solid lines indicate higher and lower edges of the bandgap frequencies, respectively. The gray band from 0.25 Fig. 5.17 Frequency range variations of perfect photonic bandgaps

5 Stereolithography

67

Fig. 5.18 Designed graphic models of diamond photonic crystals

to 0.45 THz indicates the intended frequency range for the photonic bandgap, and the dotted line at 240 µm indicates the preferable lattice constant of the diamond structure. The frequency ranges in the vicinity of 0.35 THz are used in novel imaging and inspection systems for package inspection, quality control, and nondestructive testing [50]. The graphic data for the photonic crystal composed of perfect periodic lattices was finalized by connecting 20 × 20 × 4 unit cells, as shown in Fig. 5.18, to create the crystal model of 4.8 × 4.8 × 0.96 mm3 in dimensions.

5.4.3 Dielectric Lattice The titania micro-lattices with diamond structures were fabricated through resin dewaxing and powder sintering with ultraviolet laser lithography, as shown in Fig. 5.19. The stereoscopic image was obtained by DOM observations. The scanning speed and irradiation power of the laser beam were adjusted to decrease the remaining carbon and increase the part accuracy of the ceramic components. The laser conditions are listed in Table 5.1. The ceramic components in Figs. (a), (b), and (c) were fabricated under the laser conditions A, B, and C given in Table 5.1, respectively. In the case of Fig. 5.19a, the spread resin paste containing nanoparticles was scanned at a moving speed of 50 mm/s with a laser beam of 600 mW irradiation power. The titania particles were coagulated on the lattice surface through thermal conductions in the laser sintering. Residual carbon that accumulated during deficient dewaxing of acrylic resin solvent was observed as black areas in the lattice cross-section. Subsequently, the laser scanning speed was increased to 100 mm/s to reduce thermal conductions for the resin paste, and the coagulation of particles was prevented, as shown in Fig. 5.19 (b). The part accuracy was measured to be ±5 µm

68

S. Kirihara

Fig. 5.19 Titania photonic crystals with diamond structures fabricated by ultraviolet laser lithography. Ceramic lattices (a–c) were fabricated according to the laser conditions A, B, and C listed in Table 5.1, respectively

Table 5.1 Process conditions in ultraviolet laser lithography

Laser conditions

A

B

C

Spot size (µm)

10

10

10

Scan speed (mm/s

50

100

100

Irradiation power (mW)

600

600

700

using DOM. The residual carbon indicated by the black area increased simultaneously in the lattice. Moreover, as laser irradiation power was augmented to 700 mW, the black area disappeared from the lattice cross-section, as shown in Fig. 5.19c. The residual carbon peaks were not observed in the XRD pattern of the formed titania lattice. The crystal structure of the titania was analyzed and revealed as a dual phase of anatase and rutile. The titania particles were considered to be heated in the vicinity of the temperature for the anatase-to-rutile phase transformation. A relative density of the ceramic components of 97% was revealed by the Archimedes’ measurement.

5.4.4 Electromagnetic Property The titania photonic crystal formed by ultraviolet laser lithography was he ated at 1350 °C for 2 h in air atmosphere to transform the crystal structure to the rutile phase. Figure 5.20 shows a top-down view of the titania lattices after the heat treatment [51, 52]. Linear shrinkage due to the treatment was

9 Cold Sprayed Metal Coatings 133

(Low temp. HVOF) Warm spray

(i)

Combustion gases + Nitrogen

Combustion gases + Nitrogen, (air)

Helium?

※

Helium

Gas type

Process gas

Under 4

(Under 1)

2

Under 1

Stagnation pressure (MPa)

1430–2350

(600–2000)

550–900

Under 900

(Under 400)

Stagnation temperature (°C)

– developed by researchers – NIMS/Kagoshima Univ. – no commercial device

Developed by Ottawa Univ. at 2005



Commercial device < >

Making warm spray

Superior coating of Ti

Improvement of HVOF Superior coating of Ti, WC–Co

Al, Cu, SS

Little gas consumptions, controlled by lower than sound speed

Main characteristic

II Impact Innovations GmBH (Rattenkirchen, Germany), NIMS National Institute for Materials Science (Tsukuba, Japan) PG Plasma Giken Co. Ltd. (Saitama, Japan), SM Sulzer Meteco (Truebbach, Switzerland) SST Supersonic Spray Technologies Division in Centerline Ltd. (Windsor, Canada), VRC VRC Metal Systems (South Dakota, USA) ※: Nitrogen, helium, air or those mixed gases

High pressure warm spray

Shock-wave Induced Spray Process (SISP)

(h)

(j)

Low press./high temp.

(g)

Low temp. HVOF

Low press. sonic

Type

(f)

No.

Table 9.1 (continued)

134 K. Sakaki

9 Cold Sprayed Metal Coatings

135

type warm spray (j) that can form a dense titanium alloy coating by increasing the combustion pressure for a higher particle velocity has also been developed [11]. In high-pressure type CS equipment (a), the particle size applied is approximately 5–30 µm, which is often finer than the particle size distribution of conventional thermal spray powders, although commercially available thermal spray powders often cannot be used. Recently developed commercial CS devices can use relatively large particle sizes by extending the nozzle inlet part and increasing the working gas temperature. Along with a high-pressure type, portable low-pressure CS equipment (e) has also been developed in Russia, which heats up to approximately 600 °C using a gun with a built-in heater and spraying air with a pressure of approximately 0.6 MPa or less. In Russia, this equipment is mainly used for part repair and maintenance applications. Initially, portable low-pressure CS equipment has a low pressure, and the particle velocity is limited. In addition, metal particles are formed while mixing hard particles such as alumina with metal powder to activate the surface of the substrate or coating. Presently, zinc, aluminum, and copper can be coated using only metal powder.

9.3 Advantages and Disadvantages of Cold Spray Process The cold spray (CS) process offers several advantages over other coating systems and deposition processes. Because it is a low-temperature process, the cold spray operates at below the melting point of the metals, requires no combustible fuels or gases, and results in extremely low porosity deposits. The greatest advantage of the coating produced by this cold spray process is that there is almost no oxidation or thermal deterioration. That is, because the particles impact and are deposited onto the substrate or coating within a low residence time of several milliseconds in a low-temperature inert gas such as nitrogen or helium, it is possible to form a coating from a material powder that is easily thermally degraded. Therefore, because there is no sublimation or oxidation in the flame, as in other thermal spraying methods, if the particles are subjected to spray conditions exceeding the critical velocity, copper achieves a high deposition efficiency of 95% or higher. The advantages of a CS are summarized as follows: (1) It removes the thermal stress of the coating and suppresses the oxidation and thermal deterioration, (2) achieves a dense coating, (3) results in a coating with thermal and electrical conductivity, (4) allows a thick compressible residual stress coating to be applied, (5) has a high deposition rate (depending on the powder material and particle velocity), (6) results in less fumes, (7) suppresses the heat input of the substrate, (8) achieves a high hardness through work hardening, (9) requires a minimum masking of the substrate, and (10) is applicable using simple equipment. In contrast, the disadvantages of a CS are as follows: (a) There remains an insufficient understanding of the basic coating formation mechanism and an elucidation of the coating characteristics with a small database. In addition, (b) a large amount of consumption gas is required, (c) adhesion and deposition of the powder occurs in

136

K. Sakaki

the nozzle (nozzle clogging issue), (d) the particle diameter used is relatively finer (5–40 µm), and depending on the material, it is necessary to take measures against a dust explosion, and (e) the coating properties differ owing to the difference in bonding state between particles from the impact velocity. Moreover, (f) the adhesion strength differs depending on the spray conditions, (g) the adhesion between the coating and substrate decreases as the coating becomes thicker, (h) the formation of a coating in the micro-pores is difficult to achieve, and (i) a stable supply powder may be unavailable. Finally, (j) the powder design guidelines for a CS have yet to be established, with few commercially available powders exclusively applicable for CS use, and (k) when gas is applied at a pressure above 1 MPa, safety must be taken into consideration. Although a dust explosion of fine particles (d) is not a problem during the operation of a CS when using inert gas such as nitrogen, it is necessary to be careful when collecting dust after applying a CS. In addition, despite a nozzle clogging issue (c) being a significant problem, the adhesion of particles is suppressed and solved by water cooling of the nozzle and by using a nozzle made of polybenzimidazole (PBI) plastic, which has the highest mechanical properties of any plastic at above 200 °C. As for disadvantage (h), if one attempts to form a coating in a fine hole and repair it, a supersonic working gas may accumulate in the hole, and the particles may not decelerate and adhere. It is thus necessary to expand the hole and repair it. In recent years, research and development on the CS process has become active, and the technical defects described above are being overcome.

9.4 Materials and Application Study Cases Table 9.2 shows the types of material particles examined using the CS method. At the beginning of its development, a CS consisted mainly of pure metals such as copper, iron, nickel, and aluminum. At present, as our understanding of the mechanism regarding the coating formation has increased and progress has been made in increasing the temperature of the working gas, along with metals and alloys such as stainless steel, titanium, molybdenum, Ni–Cr, MCrAlY, and tantalum, polymers and Cr3 C2 –NiCr and WC–Co cermets can also be deposited. The deposition of amorphous and quasi-crystalline metals, which change their crystal structure when over-heated, has also been attempted. In addition, a number of composite coatings are being considered. That is, because a metal coating has good heat conduction characteristics, an aluminum composite coating containing diamond and AlN can be produced to improve the heat conduction characteristics, along with copper and tungsten for adjusting the linear expansion coefficient using peripheral materials, and copper-lead-tin composite coating for sliding members. Furthermore, studies have also been reported to control the adhesion rate and hardness of the coating by changing the mixing ratio of a mixed powder of high-carbon and mild steel [12]. With regard to the formation of a cermet coating using a CS, decarburization can be suppressed because of the low temperature, and the WC primary particle diameter

9 Cold Sprayed Metal Coatings

137

Table 9.2 Example of cold spraying materials [22] Pure metal

Cu, Al, Ti, Ag, Ni, Zn, Sn, (Mo), Fe, Ta, Nb, (Si), Cr, Mg

Low-alloy steels

Ancorsteel 1000, Intermetallic Compound: FeAl

Nickel-chrome alloys

50Ni–50Cr, 60Ni–40Cr, 80Ni–20Cr

Nickel-base superalloys

Alloy 625, Alloy 718, Hastelloy C, In738LC

Stainless steel

SUS304/304L, SUS316/316L, SUS420, SUS440

Zinc alloys

Zn–20Al

Aluminum alloys

A1100, A6061, A7075, Al–Sn, Al–17Si

Copper alloys

C95800 (Ni–Al Bronze), 60Cu–40Zn

(MCrAlY)

NiCrAlY, CoNiCrAlY

Amorphous metal

Fe–Cr–Mo–W–C–Mn–Si–Zr–B, 57Ni–18Ti–20Zr–3Si–2Sn, Al–Ni–Ce 54Cu–6Ni–22Zr–18Ti7, Al–Ni–Ce, 41.2Zr–13.8Ti–12.5Cu–10Ni–22.5Be10

Quasicrystal metal Composite material

Ceramic

Al–Cr–Fe–Ti–Co Metal-metal

Al–Cu, Ti–Al, W–Al, Al–Sn, Ni–Al Cu–W, Cu–W–Zn, Cu–Pb, Cu–Pb–Sn, Cu–Cr, Cu–Cr–Al High carbon steel/mild steel

Metal-ceramic

Al-Diamond, Bronze-Diamond bronze-alumina, Al–Al2 O3, Al–Al2 O3 –Zn–Ni, Cu–Al2 O3 Ni–Al2 O3 , Inconel-Al2 O3 , Ag–14SnO2 , Al–SiC, Cu–SiC Ni-B4 C

Intermetallic compound-ceramic

FeAl–Al2 O3 , AlFeAl–Al2 O3 , FeAl–WC

Functional coating

Fe–NdFeB, Nd2 Fe14 B–Al, CNT–Al, MWCNT–Cu, Sn–CNT

Cermets

WC–Co, WC-15% [Fe–18%Cr–8%Ni], Cr3 C2 –NiCr Nano-TiO2 , Nano-TiO*2 , Al2 O*3 , AlN* , TiN* , TiAl2 O5 * Low pressure CS

can be refined to approximately 0.2 µm. In addition, it has been reported that the adhesion rate is increased by using an iron-based alloy as the metal binder layer [13]. With regard to ceramics, primary particles form nano-sized special TiO2 powders under atmospheric pressure and are being studied in photocatalyst applications [14]. Because there is a strong tendency to secretly develop application examples, few application examples of a CS have been published at the commercial level. As the world’s first mass produced application, a copper coating on the back of an aluminum

138

K. Sakaki

heat sink with an optimized thermal conductivity through a soldered copper plate for application in computer CPUs was adopted in Germany in 2003. The following is a summary of the main application study cases (including those that have not reached practical use) that have been published. In addition to these, the number of applications has increased [15]. ➀ ➁ ➂ ➃ ➄ ➅ ➆ ➇ ➈ ➉

 11  12  13  14  15  16  17

zinc coating on automotive steel sheets aluminum alloy coating for corrosion protection of magnesium sputtering target and its repair production of metal near net-shaped members (aerospace and other industries) copper coating for electromagnetic wave shielding corrosion-resistant zinc coating on welds high-temperature resistant coating (MCrAlY) examination of repair of gas turbine blade application of brazing material (radiator fin joint etc.) application to medical equipment parts (pure titanium etc.) diamond blade (composite coating with metal) electrodes made of anode materials for secondary batteries (Si) copper applied to heat dissipation board for power module copper electrode pattern on aluminum nitride substrate repair of aluminum vapor deposited (IVD-Al) coating for aircraft parts repair of aircraft parts iron coating on the surface of aluminum pots for rice cookers.

In many cases, high electrical and thermal conductivity is required owing to the use of copper and aluminum coating, and further heat treatment makes these properties comparable to those of bulk materials. Looking at the status of patent applications, 13 and  14 , have been relatively large the number of power unit-related items, such as  during the past five years. In the United States, MIL-STD-3021 has been used in corrosion resistance, repair, wear resistance, and EMI shielding for military applications. In particular, it is expected to be used to repair aluminum, titanium, and magnesium alloy components 16 helicopters [16]. of aircraft including  As an example application, studies on a large-area thick copper coating applied as a good thermal conductive layer (approximately 5 × 3 m2 , with a coating thickness of 3 mm) have been conducted for a cold sprayed triangular support used in the International Thermonuclear Experimental Reactor (ITER) Vacuum Vessel [17]. However, the design review has eliminated the need for cold spray coatings. If adopted, it is expected that the awareness of the CS technique will improve, which is fortunate. Furthermore, the thick coating formation ability of a CS was also considered, and a formation method combining a near net-shaped processing of aircraft parts with the CS method and end mill cutting was also considered. Therefore, although a CS can be used as a metal spray forming technology as a type of “additive manufacturing,” which has recently attracted research attention, and is the title of this book, there remain certain issues such as the dimensional accuracy and member strength. At

9 Cold Sprayed Metal Coatings

139

Fig. 9.4 CS applications: a electric conductive parts and b rotatable sputtering target for glass coating [21]

present, the CS process is limited to near net-shaped machining, although it may be possible to improve the nozzle diameter and post-processing to a certain extent. Figure 9.4 shows an example of a molded body using the CS method, which has been commercialized in Japan. In Fig. 9.4a, pure copper without thermal deterioration can be easily cut into a flange shape by forming a spray with a thickness of 50 mm or more using a CS and cut into a flange shape. In addition, as shown in Fig. 9.4b, a thick coating 10 mm or more of ZnAl on a long 3 m pipe can be formed, and the high formation ability using a CS is shown. An investigation of the development situation on the CS of the world was conducted in Japan in 2004, and feasibility studies were jointly conducted by industry and academia in 2005 and 2006; moreover, application to gas turbine members, aircraft members, boiler tube members, and wear resistance parts were considered with financial support from the Mechanical Social Systems Foundation [18]. In contrast, a low-pressure type CS device can be deposited using a hand-type gun with compressed air. The impact velocity of the particles and the materials used are more limited than in a high-pressure type CS. However, in Russia as a developing country, more than a thousand units are already on the market, mainly for use in overlay repair applied in the automotive industry, for example, in aluminum wheel rims, automobile sheet metal, and the shape recovery of an aluminum engine block.

9.5 Future Studies on Cold Spray As with other thermal spraying methods, the CS method likewise has many parameters, and it is not easy to control the quality of the coating as an output, such as the interaction of these factors and the degree of variation of each factor. However,

140

K. Sakaki

future research is expected to optimize the main factors, which are becoming clearer. In addition, as mentioned above, the CS method also has certain disadvantages, and an improvement of the bonding between particles has been carried out by increasing the particle impact velocity and applying a post-treatment such as heat treatment and laser heating. Furthermore, as mentioned above, basic research and development into various applications of the CS method are being conducted in domestic and foreign research institutes and companies. However, despite the many case studies, there are still few reports of such cases actually being put into practical use. Research and development are being carried out in search of some breakthrough applications. Moreover, as mentioned above, the application of a CS device to refractory metals is promoted by optimizing the temperature of the working gas (from cold to warm) and the nozzle shape. However, as with other processes, the use of lower-cost equipment (for example, low-pressure, low-temperature type CS equipment) and the reduction of costs with regard to the gas and material powders applied are posing problems, which also hinder CS use. Furthermore, the powder designs of raw materials suitable for a CS have also been studied, some of which are commercially available, although not sufficiently.

9.6 Summary In conjunction with solutions to environmental problems, machine parts are placed under increasingly severe environments to improve machine performance, and a surface modification technology for the point of contact of parts with the outside world is becoming increasingly important. Under such circumstances, the thermal spraying method is increasingly being applied and further developed owing to its advantages. Unlike a conventional coating technology, the solid state deposition feature of a CS method that does not melt the material has the potential to partially overcome the challenges of a conventional thermal spraying process. Further, despite thermal spraying being a conventional technology, it has the potential to be applied in new areas rather than as simply a substitution to current methods. Although a market is developing for the application of a CS, it cannot be stated that this is an innovative thermal spraying method, although it does have the potential to be. Therefore, for further development and practical use, it is necessary to systematically advance interdisciplinary research, such as material science and compressible fluid dynamics (CFD), including an enhancement of the database and an elucidation of the particle deposition mechanism.

9 Cold Sprayed Metal Coatings

141

References 1. A.P. Alkimov, A.N. Papyrin, V.F. Kosarev, N.I. Nesterovich, M.M. Shushpanov, U.S. Pat. 5302414 (1994) 2. V.K. Champagne (ed.), The Cold Spray Materials Deposition Process (Woodhead Publishing Ltd, Cambridge, 2007) 3. K. Sakaki (ed.), Special issue on cold spray. J. Jpn. Therm. Spray Soc. 47(3)–48(3) (in Japanese), on http://www.jtss.or.jp/journal/special_issue_CS_announcement.pdf 4. M. Fukumoto (ed.), Future Openings for the New Particle Deposition Coating Technologies— Cold Spray, Warm Spray and Aerosol Deposition (CMC Publishing Co., LTD., Tokyo, 2013) (in Japanese) 5. J. Villafuerte (ed.), Modern Cold Spray—Materials, Process, and Applications (Springer, Switzerland, 2015) 6. V.R.G. Maev, V. Leshchynsky (eds.), Cold Gas Dynamic Spray (CRC Press, Florida, 2016) 7. C. M. Kay, J. Karthikeyan (eds.), High Pressure Cold Spray, Principles and Applications (ASM International, 2016) 8. H. Assadi, F. Gärtner, T. Thorstenhoff, H. Kreye, Bonding mechanism in cold gas spraying. Acta Mater. 51, 4379–4394 (2006) 9. T. Kubo, K. Sakaki, Y. Tokui, M. Song, S. Kuroda, T. Hosono, Effect of substrate temperature on adhesion strength of metal coating on ceramic substrates in cold spray, in Proceedings of 2013 JTSS Fall (2013), pp. 33–34 (in Japanese) 10. J. Kawakita, S. Kuroda, T. Fukushima, H. Katanoda, K. Matsuo, H. Fukanuma, Dense titanium coatings by modified HVOF spraying. Surf. Coat. Technol. 201(3–4), 1250–1255 (2006) 11. R. M. Molak, H. Araki, M. Watanabe, H. Katanoda, N. Ohno, S. Kuroda, Warm spray forming of Ti-6Al-4V. J. Therm. Spray Technol. 23(1–2), 197–212 (2014) 12. K. Ogawa, S. Amano, N. Yokoyama, K. Ootaki, Improvement of deposition efficiency and control of hardness for cold-sprayed coatings using high carbon steel/mild steel mixture powder. J. Solid Mech. Mater. Eng. 5, 793–802 (2011) 13. K. Sato, H. Furukawa, J. Kitamura, WC-Fe alloy cermet material for new frontier of thermal spray, in Proceedings 4th Tsukuba International Coatings Symposium (2010), pp. 41–42 14. M. Yamada, H. Isago, H. Nakano, M. Fukumoto, Cold spraying of TiO2 photocatalyst coating with nitrogen process gas. J. Therm. Spray Technol. 19(6), 1218–1223 (2010) 15. C.M. Kay, J. Karthikeyan (eds.), High Pressure Cold Spray, Principles and Applications (ASM International, 2016), pp.185–302 16. C.M. Kay, J. Karthikeyan (eds.), High Pressure Cold Spray, Principles and Applications (ASM International, 2016), pp. 227–251 17. N. Espallargas (ed.), Future Development of Thermal Spray Coatings: Types, Designs, Manufacture and Applications (Woodhead Publishing Ltd., Cambridge, 2015), pp. 160–161 18. Feasibility study report on innovative material creation by cold spray—summary—, System technology development research 18-F-3, the Mechanical Social Systems Foundation (2007) (in Japanese) on https://sokeizai.or.jp/japanese/rimcof/images/shisukyou-18–1.pdf 19. M. Fukumoto (ed.), Future Openings for the New Particle Deposition Coating Technologies— Cold Spray, Warm Spray and Aerosol Deposition (CMC Publishing Co., LTD., Tokyo, 2013), p. 39 (in Japanese) 20. S. Kuroda, J. Kawakita, M. Watanabe, H. Katanoda, Warm spraying—a novel coating process based on high-velocity impact of solid particles, in Science and Technology of Advanced Materials, vol. 9 (2008), p. 033002 and adding particle impact photos (Source: Germany Helmut Schmidt University) 21. Photo courtesy Plasma Giken Co., Ltd., and information on http://www.plasma.co.jp/products/ index.html 22. R. C. Tucker Jr (ed.), ASM Handbook Volume 5A Thermal Spray Technology (ASM International, 2013), p. 57

Chapter 10

Cold Spray Technique Kazuhiro Ogawa

10.1 Metal Particle Deposition Mechanism The CS technique has demonstrated significantly successful deposition of metal particles with substantial plastic deformation, as metal particles do not melt, but instead collide with the substrate at a high speed. The CS technique is reported to be useful in depositing various types of materials, including soft materials such as aluminum [1] and copper [2], high-quality materials, such as stainless steel and nickel-based superalloys [3], and cermet particles such as WC-Co (Figs. 10.1 and 10.2) [2, 3]. However, opinions regarding the particle adhesion mechanism of the CS technique are divided, and various mechanisms have been proposed by numerous researchers. For example, Xiong et al. [4] reported that when aluminum particles are adhered to the aluminum substrate, rapid cooling occurs by the Joule–Thomson effect when the particles collide with the substrate, resulting in dynamic amorphization and recrystallization, and the creation of an amorphous layer with specific thickness (of nm size), forming a union. Ning et al. [5] reported that when aluminum-tin alloy is adhered to SUS304, the tin melts upon particle collision, and adherence is achieved by the adhesive action of the melted particles. Additionally, Grujicic et al. [6] reported that during particle collision, a great shear stress acts near the interface; the particles transform to envelop the substrate, producing a mechanical interlocking effect to create adhesion. Various adhesion mechanisms have been proposed to explain the process: union through amorphization [4] and recrystallization near the interface, union through the melting of an ultrathin portion of the particle surface, and mechanical union through shear deformation. However, the detailed mechanism remains unclear. Further, it is difficult to expand the application of the CS technique because of lack of clarity.

K. Ogawa (B) Tohoku University, Sendai, Japan e-mail: [email protected] © The Japan Welding Engineering Society 2021 S. Kirihara and K. Nakata (eds.), Multi-dimensional Additive Manufacturing, https://doi.org/10.1007/978-981-15-7910-3_10

143

144

K. Ogawa

Fig. 10.1 Cold-sprayed copper deposited on aluminum substrate

Deposition

Interface

Substrate Fig. 10.2 Cross-sectional image of cold sprayed Ni base superalloy IN738LC

Therefore, the adhesion mechanism should be promptly identified, and a highly reliable CS coating process needs to be implemented. Here, we explain the implemented aluminum deposition mechanism. Obtained with a scanning electron microscope, Fig. 10.3 depicts an example of a cross-sectional view of pure aluminum powder deposited on a pure aluminum substrate using low-pressure CS. For this example, the spray conditions were as follows: compressed air as a working gas; gas pressure of 0.6 MPa, a nozzle vent gas temperature of 320 °C, a nozzle-substrate distance of 10 mm, and a traverse speed of 0.5 m/s. Figure 10.3 presents evident facts: the aluminum coating is fine, the pores are small, and the substrate-coating interface is of high quality that cannot be

10 Cold Spray Technique

145

Al coating

Al coating

Al substrate

Interface

Al substrate Fig. 10.3 Typical SEM image of Al coating on Al substrate by using the low-pressure cold spray technique

distinguished. However, Fig. 10.4 shows that the low-pressure CS technique delivers a low deposition efficiency (%) in single digits. Figure 10.4 shows the pattern of deposition efficiency that changes in response to coating thickness. Until the coating thickness reaches 200 µm, the deposition efficiency increases with coating thickness; after 200 µm, it is confirmed to stabilize at approximately 7%. The deposition efficiency is calculated by measuring the weight of the material before and after spraying, as well as by the following equation, which takes weight M of the particles left inside the equipment, as reference: D=

Weight of Substrate after Spraying − Weight of Substrate before Spraying × 100(%) Weight of Particles before Spraying − Weight of Particles Remaining After Spraying − M

Fig. 10.4 Deposition efficiency of the low-pressure cold spray technique

Deposition efficiency (%）

Weight of Substrate after Spraying Weight of Substrate before Spraying.

Thickness (µm） ）

(10.1)

146

K. Ogawa

Weight of Particles before Spraying Weight of Particles Remaining After Spraying. This estimation method can be used to calculate various losses during measurement (particles remaining inside the nozzle or piping, etc.). However, the results obtained are not necessarily accurate. Nevertheless, even if these potential losses are taken into account, the deposition efficiency of pure aluminum obtained using the low-pressure CS technique is remarkably low. In contrast, the deposition efficiency is reported to be over 90% in case of aluminum deposition by the high-pressure CS process. Also, the deposition efficiency increases with the increase in thickness of the coating, until 200 µm. The deposition mechanisms differ with adhesion to the substrate-particle interface, which is considered to be a difficult process, as it involves two steps of coating, namely (1) adhesion to the substrate-particle interface, and subsequently, (2) adhesion of further particles atop the particles adhered to the substrate. Therefore, in adhering the particles to the substrate, the potential presence of a “lag time” (time interval during which the particles adhere after collision) has been proposed. As this lag time plays an important role in the adhesion of particles to the substrate, studies have been conducted to address it. Figure 10.5 shows an overview of a lag time evaluation experiment. As it is extremely difficult to directly estimate the time taken by particles to begin to adhere followed by initiating the spray during the study, the particle output was decreased, the substrate was traversed (i.e., the gun moved) at a speed of 0.5 m/s, and the particles were dispersed and collided. Multiple specimens were manufactured with a varying number of traverses; the number of traverses taken by the particles to begin to adhere was reviewed, and from this information, the presence of a lag time was confirmed. In the evaluation, as shown in Fig. 10.5, observations were made

Nozzle

Spray area

Substrate Substrate surface

Fig. 10.5 Schematic illustration of evaluation test for deposition lag time

Observed area

10 Cold Spray Technique

147

1 traverse

2 traverses

6 traverses

8 traverses Deposited particle

Deposited particle Deposited particle

Fig. 10.6 Typical SEM images of deposited particles [1]

using SEM of the number of adhering particles within 5 mm2 area of the spray area, and the number of adhering particles for the number of traverses used in each case was evaluated. As the evaluation has difficulties in distinguishing sprayed particles and contaminants, particles below diameter of 5 µm were excluded from calculations. Figure 10.6 shows SEM images of the substrate surface for a varying number of traverses. Figure 10.7 shows the changes that take place in the number of adhering particles versus the number of traverses. It is observed that no particles adhere during one and two traverses; however, there is a noticeable increase in the number of adhering particles from 6 to 8 traverses. This change confirmed the existence of a lag time until the particles begin to adhere in the deposition of pure aluminum particles on a pure aluminum substrate using lowpressure CS. The presence of a lag time means that conditions such as the hardness, roughness, and degree of activation on the substrate surface change through the repeated collision of particles with the substrate, facilitating the particles to adhere. As stated above, particles do not adhere with a single, but through repeated collisions into the substrate. As the number of particle collisions increases, the roughness of the substrate surface is confirmed to undergo substantial change. This process is also considered to change the structure of the substrate surface, work-harden the substrate surface by particle collision, and activate the substrate surface through particle collision. Hitherto, analysis has clarified that the factor that has a significant effect on particle adhesion of those mentioned above is the activation of particles

Fig. 10.7 Relationship between the number of particles and the number of traverses

K. Ogawa Number of deposited particles (1/5mm2)

148

Rapid increase

Traverse numbers (times)

Fig. 10.8 Number of deposited particles with respect to interval time

Number of deposited particles (1/5mm2)

and the substrate surface. According to conjecture, in the air, a naturally oxidized coating of a particular thickness (nm) is created on the aluminum surface; then this coating is destroyed, which exposes a new surface, initiates the activation, and leads to adhering of particles. Here, changing the interval (waiting period between the initial traverse to the next traverse) for the 0.5 m/s traverse of the nozzle during spraying also influences the degree of activation on the substrate surface. If substrate surface activation creates a new surface through the repeated particle collisions and the destruction of old and naturally oxidized surface, then longer the interval, the more time available to the new surface to re-oxidize, and less the activation phase on the substrate surface. Figure 10.8 depicts changes in the number of deposited particles along with changes in the number of traverses during substrate surface spraying at different intervals of time.

Interval Interval Interval Interval Interval

time time time time time

Traverse numbers (times)

Fig. 10.9 Number of adhering particles for 12 traverses

149 Number of deposited particles (1/5mm2)

10 Cold Spray Technique

Interval time (sec.)

For 1800 and 3600 s, no particle adhesion can be confirmed for traverses 1–10, and adhesion can be measured only when initiating traverse 12. The longer the time interval, the fewer the particles adhere, namely, at 1800 and 3600 s, almost no particles are confirmed to adhere. Furthermore, the number of particles observed to adhere after 12 traverses for different intervals are shown in Fig. 10.9. These results confirm that the number of adhering particles substantially decreases from an interval of 0.6 s to an interval of 60 s. Subsequently, from an interval of 1800 s onward, the number of adhering particles decreases slowly but steadily. From the above, it can be confirmed that the naturally oxidized surface on the aluminum substrate is removed, and a new surface is exposed with the initial traverse. Increasing the duration of interval allows the surface oxidation to reform and redevelop, and the thicker the oxidized coating, the harder it is for particles to adhere. Hence, it is clear that the thickness of the oxidized coating is extremely important for particle adhesion. Moreover, the aforementioned time lag is considered to be the time taken for the removal of naturally oxidized coating through particle collision and for the new surface to be exposed. With high-pressure CS, the particle velocity is much higher than that obtained with low-pressure CS. Therefore, it is believed that natural oxidation could be easily removed exposing the new surface, resulting in a significantly higher adhesion efficiency. It is also considered that analysis of how low-pressure CS could more efficiently remove the naturally oxidized coating from the substrate surface that could also improve the adhesion efficiency. Summarizing the above, at current stage the low-pressure CS may have low adhesion efficiency, but it is capable of depositing fine coatings with few pores. Furthermore, the equipment required is compact, and it can use compressed air as its working gas, which is economical; therefore, the process is expected to make a significant contribution to on-site repairs. Future challenges include studying the low-pressure CS process that can efficiently activate the substrate surface, possibly by combining it with other processes, as increasing the adhesion efficiency is vital.

150

K. Ogawa

10.2 Examples of Ceramic Particle Deposition In recent years, the CS technique could be used to deposit certain ceramic materials. Figure 10.10 shows SEM cross-sections of a TiO2 coating applied on a glass substrate using the CS technique. From the images, it is clear that a fine coating of 100 µm or more is formed. Even though TiO2 particles with a diameter on the order of tens of micrometers are used, deposition cannot be achieved. In this case, as shown in Fig. 10.11, the primary particle diameter is 20–30 nm. These particles clump together, and particles with a secondary diameter of 5–10 µm are used. For the CS technique, using active nanoparticles is considered to enable efficient deposition without using added heat. However, if a fine powder is supplied for the CS process, then active nanoparticles might clump together in the nozzle or elsewhere, causing the nozzle to clog. Therefore, using nanoparticles is difficult, and it is more desirable to allow primary nano-order particles to agglomerate into secondary particles of specific width of several micrometers and rely on using the secondary particles instead. For metal Fig. 10.10 SEM image of cross-section of TiO2 coating on glass substrate by using the cold spray technique

Fig. 10.11 SEM image of TiO2 particles

10 Cold Spray Technique

151

particles, it is difficult to create nano-sized particles, and even though metal particles of specific width of several micrometers exist and can be formed, which can be considered to make a coating, the use of nanoparticles is not considered necessary. In contrast, it is believed that the demand for high speed and highly efficient deposition techniques that do not melt ceramic or use of polymer materials (to be discussed later) will increase in the future. When this happens, it is believed that nanoparticles will be extremely useful. As discussed above, coatings applied by using the CS technique do not undergo high-temperature oxidation or phase transformation during the application. Therefore, they do not undergo as much structural change as those applied with traditional TS process. As the CS technique forms highly reliable coatings, it is believed that this field will undergo increasing development and expansion in the future. Potential applications include corrosion-resistant or abrasion-resistant coatings (where TS was traditionally used), cladding and repair (as the technique is capable of forming thick coatings), and even application in solar power and manufacturing rechargeable batteries. Moreover, as the CS technique is now capable of depositing particular types of ceramic particles, it is also considered capable of creating gradated coatings formed from both metal and ceramic particles, where the ratio of the two substances changes. Further, we present examples of deposition using ceramic materials. The ceramic material used in the experiment was aluminum titanium oxide (manufactured by Ohcera Co., Ltd., Japan) with an average particle diameter size of 2 µm. Aluminum titanium oxide particles with diameter sizes of several nanometers agglomerate into larger particles with an average diameter size of 2 µm. A technique known as “power jet deposition” (PJD) [7] is used for the deposition. The process of the PJD technique is similar to the CS technique, however no heating equipment is used, and the technique can be set to exert even less pressure than the low-pressure CS technique. Experiments were conducted at different compressed air pressures during spraying with the goal of using PJD to optimize ceramic particle adhesion. The number of sprays was also changed to create thicker coatings. For the substrate, Inconel 600, a nickel-based superalloy, and a sintered aluminum titanium oxide were used. Deposition conditions were as follows: the pressure applied in the range of 0.01–0.6 MPa, the traverse speed was 1 mm/s, the nozzle-substrate distance was 1 mm, and the atmospheric pressure was normal. The relationship between the coating thickness obtained from a single spraying and the jet pressure was evaluated for the different substrates. A surface roughness profilometer (a Surfcom 200A, manufactured by Tokyo Seimitsu Co., Ltd., Japan) was used to measure the coating thickness. The injection lines and the lines perpendicular were scanned using the probe of instrument; a low pass filter was used to remove the coarse components, and the undulation component was measured. The maximum height obtained by these measurements was used as the coating thickness. The results are shown in Fig. 10.12. For both Inconel 600 and TiAl2 O5 substrates, it was evident that a pressure of 0.05 MPa produced the highest number of adhered particles. For metal particles, the adhesion did not take place at low pressures applied and particles behaved in

152

K. Ogawa 10

10

Max. deposition at 0.05MPa

Max. deposition at 0.05MPa

5

5

0

0

-5

-5

-10

0

0.1

0.2

0.3

0.4

Pressure (MPa)

a) Inconel600

0.5

0.6

0.7

-10

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Pressure (MPa)

b) TiAl2O5

Fig. 10.12 Relationship between working gas pressure and deposition thickness

a different manner. For the TiAl2 O5 substrate, deposition thickness decreased (into negative numbers) because of the thinning of the substrate by erosion. Figure 10.13 depicts the results of evaluating particle speed using a high speed camera. The results show that at a pressure of 0.05 MPa for particle adhesion, the particle speed was approximately 90 m/s. For metal particles, adhesion was not assumed to occur unless particles were traveling at 300–1500 m/s, whereas the optimal velocity for ceramic particles was far lower. Furthermore, from a pressure of 0.01–0.05 MPa, the amount of particle adhesion increased with an increase in pressure. However, the amount of adhesion tended to experience a gradual decrease. In particular, following 0.4 MPa, particles tended to fail to adhere, and the substrate tended to be lost through erosion. It is believed that these results demonstrate the very limited conditions wherein ceramic particles will adhere, and it is very important to identify such precise adhesion conditions. Figure 10.14 shows the coating thickness obtained from repeated spraying on an Inconel 600 substrate. With the Inconel 600 substrate, a coating of approximate

Fig. 10.13 Relationship between particle velocity and applied pressure

10 Cold Spray Technique

153

Fig. 10.14 Relationship between coating thickness on Inconel 600 substrate and number of traverses

size 70 µm was successfully deposited through 20 repeated sprays. Based on this process, it is clear that ceramic particles with a diameter of 2 µm used in this experiment can overlap and form a layer of thick coating. Hereto, only particles composed of oxide ceramic materials have succeeded in forming the coating. It is believed that the absence of oxygen contributes to adhesion; however, this phenomenon remains unexplained to date. Moreover, particles that grew to a size of 2 µm through agglomeration smashed upon collision with the substrate. This smashing and fracturing is also considered to potentially contribute to adhesion. Figure 10.15 shows an example of a TiAl2 O5 coating on an Inconel 600 substrate. The image elucidates that a fine coating is deposited on the metal substrate. Figure 10.16 shows that, if metal and ceramic particles are combined, a phenomenon occurs where the metal particles begin to adhere once they exceed a certain speed, namely, the “critical velocity.” The particles adhere over a wide range until erosion occurs. In contrast, ceramic particles adhere only under precise speed conditions, as discussed above. Once particle speed accelerates, erosion intensifies, and the substrate tends to be worn away (as shown previously). These results Fig. 10.15 Typical optical microscope image of TiAl2 O5 coating on Inconel 600 substrate

154

K. Ogawa

Fig. 10.16 Schematic of critical particle deposition velocity

show that absolutely precise optimal spray conditions are vital for these and other ceramic particles for deposition using the CS technique. The approach used with metal particles to remove naturally oxidized coating to expose a new surface is usually not applied, which is similar to the ceramic particle adhesion mechanism. Alternatively, adhesion mechanisms that differ from those used for metal particles are under consideration. Chun et al. [8] used the “nano-particle deposition system” technique to deposit aluminum particles on an aluminum substrate. Here, Chun et al. claimed that the aluminum particles collided with the substrate, and the particles were destroyed to nano-size powder-like particles. Subsequently, these particles consolidated to deposit onto the substrate. It is believed that the TiAl2 O5 particles, discussed in this study, were deposited through the same mechanism, although the details regarding this process remain unclear to date. However, while the adhesion mechanism remains unclear, the CS technique can be applied with both metal and ceramic particles, and hence, further developments are expected in the future. Hitherto, we have presented examples of a mechanism for depositing aluminum particles using low-pressure cold spraying and analysis of the optimal spraying conditions for TiAl2 O5 ceramic particles. While the aluminum and TiAl2 O5 represent different types of material, namely metal and ceramic, respectively, it is possible to induce adhesion to a substrate with any of the two materials. However, the adhesion mechanisms for metal and ceramic particles are expected to differ. Metal is considered to adhere by removing a naturally oxidized coating to expose a new, active surface, whereas ceramic particles are considered to adhere following the activation process by fracturing into nano-sized pieces. It is expected that in the future, the deposition mechanisms for metal and ceramic will be clarified in further detail, and that these discoveries will lead to new developments in numerous fields.

10 Cold Spray Technique

155

10.3 Examples of Polymer Particle Deposition The CS technique has also demonstrated potential in the deposition of polymer materials [9–11]. Previous research has succeeded in depositing polycarbonate powder on metal and ceramic substrates by varying spray conditions. However, as the gas temperature increases, some of the material melts, creating a manufacturing process that is relatively closer to thermal spraying than the CS technique. In recent years, some progress was achieved in using the CS technique to deposit ultra-highmolecular-weight polyethylene (UHMWPE) in a partially melted state. Therefore, we introduce this technique in the present study. UHMWPE has certain advantages such as very high resistance to impacts and abrasion, however it is also extraordinarily resistant to flow when it is melted. Therefore, it is difficult to use UHMWPE in injection molding, which is the standard technique of molding and casting polymer materials. This study examined the potential of UHMWPE deposition using the CS technique and succeeded in creating a thick coating of the material with the application of nanoparticles. As shown in Fig. 10.17, the deposited powder comprised UHMWPE with a particle diameter size of 10–60 µm. Metal, polymer, and ceramic materials were examined for the substrates. From each of these groups, pure aluminum, polypropylene, and alumina were selected, respectively. The low-pressure CS technique was used with the cold spraying equipment, the gas temperature was set at 100–250 °C, and the gas pressure varied from 0.2 to 0.8 MPa. To ensure particle exposure time in the high-temperature gas, the nozzle was changed from its standard size of 100–200 mm. Here, even though the gas temperature was 250 °C, the particle temperature was considered to be close to 100 °C. Preliminary testing encountered difficulties in forming a thick coating by applying UHMWPE using low-pressure CS. Only a thin, discontinuous coating could be obtained. Therefore, the UHMWPE particles were combined with activated 40– 90 nm alumina (Al2 O3 ) particles (with the alumina particles combined at 3.8 wt%). Fig. 10.17 SEM image of ultra-high-molecular-weight polyethylene particles

156

K. Ogawa

The composition was calculated to enable the alumina particles to completely cover the UHMWPE surface. Moreover, the alumina particles were combined with the UHMWPE through simple mixing only (without the use of ball mills or other equipment). Figure 10.18 shows a processed sample produced during preliminary testing, where unadulterated UHMWPE particles were deposited on a polypropylene substrate. Only a single layer of uneven and inconsistent thickness could be deposited on the polypropylene substrate, whereas a thick coating could not be formed. This trend continued even if the spray conditions were changed. This shows that it would be difficult to achieve deposition with low-pressure CS equipment using UHMWPE particles only. Further, activated nano-alumina particles were combined with the UHMWPE powder. Examples of the resulting deposition are shown in Fig. 10.19. It is evident that UHMWPE particle deposition was successful for all substrates. For the polypropylene substrate, a coating thickness of approximately 1 mm was obtained; for the aluminum substrate, the coating thickness was approximately 4 mm, and for the alumina substrate, the coating thickness varied from 3 to 4 mm. However, with the alumina substrate, a particular amount of delamination occurred at the edge of the coating because of its thickness.

Fig. 10.18 Example of UHMWPE coating

UHMWPE coang

a) Polypropylene substrate

UHMWPE coang

UHMWPE coang Delaminaon

b) Al substrate

c)

Al2O3 substrate

Fig. 10.19 UHMWPE coatings on polymer, metallic, and ceramic substrates [9]

10 Cold Spray Technique

157

Fig. 10.20 SEM image of UHMWPE coating ( taken from surface)

Figure 10.20 shows an SEM surface observation of a UHMWPE + nanoalumina particle coating on the aluminum substrate. Figure 10.20 points out that the UHMWPE particles have been deposited without melting. Furthermore, Fig. 10.21 shows an SEM cross-sectional view of deposition on a polypropylene substrate. While pores are observed on the coating to a certain extent, a coating thickness of approximately 1-mm UHMWPE particles has adhered, which shows a solid union at the interface. As shown in Fig. 10.22, when alumina mapping is performed on the UHMWPE + alumina nanoparticle coating atop the polypropylene substrate, aluminum (nanoalumina) peaks can be identified around the UHMWPE particles. It is believed that strengthening the interparticle bond using nano-alumina allows a thicker coating to form the infiltration of nano-sized alumina particles between the micro-sized UHMWPE particles, perhaps activating a wedge effect. Therefore, nano-silica (SiO2 ) particles were combined with the UHMWPE particles at 3.8 wt% to confirm whether similar results can be obtained through the use of Fig. 10.21 Cross-sectional SEM of UHMWPE coating on polypropylene [9]

158

K. Ogawa

Fig. 10.22 Al elemental mapping of UHMWPE coating [9]

other nano-ceramic particles. Deposition was attempted under similar spray conditions as used with the nano-alumina particle blend. The results showed that deposition could not be achieved on the aluminum substrate or alumina substrates. The deposition could be achieved only on a polypropylene substrate, at a thickness size of approximately 200 µm. These results suggest that UHMWPE + nano-alumina particle deposition is possible, because of the nano-ceramic particles through an effect apart from the wedge effect. When the characteristics of the nano-alumina and nano-silica used in these experiments were studied, it was discovered that nanoalumina seems to be a substance that carries a positive charge, while nano-silica appears to be a substance that carries a negative charge. At present, whether this difference in charge impacts deposition characteristics is unknown, but the addition of nanoparticles enabled size of micrometer polymer particles to bind, accumulate, and deposit. In the future, the effects of nano-ceramic particles on polymer deposition will be studied in detail, and more effective nanoparticle selection and the development of high-quality polymer coatings using those nanoparticles will be pursued. Because a 3–4-mm-thick coating could be formed, based on these results, it can be expected that polymer-molding technology approximating 3D printing using the CS technique will come into application in future. As mentioned in this study, the CS technique is considered to be advantageous from the perspective of molding speed, because it is quicker than the other processes and can deposit particles at a rate of a particular thickness size of several tens of millimeters per minute.

10.4 Conclusions In recent years, the CS pray technique has exhibited not only a satisfactory track record in the deposition of metal particles, but it has also proven the capability of depositing specific ceramic particles. Nano-ceramic particles have been successfully combined with polymer particles using the CS technique, which resulted in bonding,

10 Cold Spray Technique

159

Fig. 10.23 Particle deposition mechanisms of different types of particles

accumulating, and depositing polymer micrometer sized particles. As the metallic bonds, covalent bonds, ionic bonds, and bonding conditions of these different materials differ, their deposition mechanisms also differ. As shown in Fig. 10.23, it is believed that metallic materials undergo very high plastic deformation, that ceramic particles are fractured and smashed, and that polymer particles are subject to a wedge effect from nano-ceramic particles among other effects (e.g., charge conditions). Here, an example of deposition using ultra-high-molecular-weight polyethylene polymers was introduced. It is expected that further understanding of the deposition mechanisms of different types of materials will lead to the bonding of materials that to date have shown to be resistant to the process.

References 1. H. Ide, Light Mater. Weld. 40, 27–34 (2002) 2. Japan Welding Society, Welding & Joining Handbook, 2nd ed., vol. 4: Welding & Joining Various Materials (Japan Welding Society, 2003), pp. 961–1007. 3. A.P. Alkhimov, A.N. Papyrin, V.F. Kosarev, N.I. Nesterovich, M.M. Shushpanov, U.S. Patent No. 5, 302, 414; April 12, 1994. 4. Y. Xiong, K. Kang, G. Bae, S. Yoon, C. Lee, Appl. Phys. Lett. 92, 194101 (2008) 5. X.J. Ning, J.H. Jang, H.J. Kim, C.J. Li, C. Lee, Surf. Coat. Technol. 202, 1681–1687 (2008) 6. M. Grujicic, J.R. Saylor, D.E. Beaslet, W.S. DeRosset, D. Helfritch, Appl. Surf. Sci. 219, 211–227 (2003) 7. K. Sasaki, Therm. Spray Technol. 21, 29–38 (2002) 8. J. Pattison et al., Int. J. Mach. Tools Manuf. 47, 627–634 (2007) 9. The Mechanical Social Systems Foundation, Research Report on Innovative Parts Manufacturing Using High-Speed Particle Collision, (2005). 10. J. Karthikeyan, International Status and USA Efforts (ASB Industries Inc., 2004), pp. 1–14. 11. K. Sasaki, Therm. Spray Technol. 20, 32–41 (2000)

Chapter 11

Precursor Spray Yasutaka Ando

11.1 History of Precursor Spray Since Karthikeyan et al. presented a thermal spray deposition process using liquid precursors [1], this process has been called solution precursor plasma spray (SPPS). Since SPPS has been called precursor spray (PS) frequently, SPPS is explained as PS in this chapter. The fine particle synthesis method using the non-transferred arc as a heat source, called “Plasma jet heating method (Fig. 11.1),” developed in the late 1960s, is thought to be one of the origin of thermal plasma deposition process [2]. In this process, fine particles were created by irradiating the target material (feedstock) with a plasma jet. Initially, it was common that the same material as that comprising the fine particles was used as the target material. Afterwards, the so-called “reactive fine particle creation process,” which could create composite fine particles by hightemperature plasma reaction between the target material vapor and the reactive gas, has been developed [3]. In both processes, solid phase material was initially used as feedstock. However, in the 1980s, PS, which synthesized fine particles by chemical reaction of reactive gas and liquid phase precursor was developed [4]. The fine particle manufacturing method using these thermal plasmas, including the precursor spraying method, is still practically used as a fine particle manufacturing method for the display phosphor and dielectric of multi-layer ceramic capacitor (MLCC), etc. [4, 5]. Recently, since it was demonstrated that film could be deposited by accumulation of the synthesized fine particles on the substrate located downstream of the plasma jet in PS, PS came to be regarded as a film deposition process [6]. Many functional films, such as diamond [7, 8], SiC [6, 9–11], partially stabilized zirconia (PSZ) [12], TiBC, and TiBN [13] (Table 11.1), were formed.

Y. Ando (B) Ashikaga University, Ashikaga, Japan e-mail: [email protected] © The Japan Welding Engineering Society 2021 S. Kirihara and K. Nakata (eds.), Multi-dimensional Additive Manufacturing, https://doi.org/10.1007/978-981-15-7910-3_11

161

162

Y. Ando

Fig. 11.1 Schematic diagram of non-transfer arc nanoparticle manufacturing equipment

Table 11.1 Typical examples of studies on PS Film

Feedstocka

Plasma

Atmosphere

Diamond [7]

CH4 (g), Ar(g) + H2 (g)

RF

Low pressure

Diamond [8]

CH4 (g), Ar(g) + H2 (g)

DC arc

Low pressure

Diamond [14]

C2 H2 (g), O2 (g)

Combustion flame

1 atm air

SiC [6]

SiCl4 (g), CH4 (g)

DC arc

Low pressure

PSZ [12]

Zr(OC4 H9 )4 (l), Mg(OC2 H5 )2 (l)

RF

Low pressure

TiO2 [15]

PTA solution (l)

DC arc

1 atm Ar

TiBC, TiBN [13]

Metal alcoxide

DC arc

Low pressure

a (g):

Gas phase, (l): Liquid phase

Although the above processes are processes conducted in a low pressure environment, a combustion flame method, which is a diamond synthesis method using acetylene/oxygen combustion flame, was developed by Hirose et al. [14] in the late 1980s. Titanium oxide (TiO2 ) films were deposited on the condition of titanium iso butoxide (TTIB: Ti(OC4 H9 )4 ) as feedstock under an atmospheric pressure Ar ambient gas environment by Futamata et al. [15] in the latter half of the 1990s in Japan. Presently, some oxide films such as yttria stabilized zirconia (YSZ) coatings are practically applied by PS under an atmospheric pressure [16]. Moreover, in United States, a national project for development of a practical PS system was established [17].

11 Precursor Spray

163

11.2 Precursor Spray Equipment Since the PS equipment generates explosive or toxic exhaust gas depending on the feedstock materials, a trap containing a neutralizing agent, etc., is installed in the exhaust system. Except for the exhaust system, the constitution of PS is almost the same as that of the conventional plasma spray equipment. As for the plasma jet (thermal plasma) generation system, although arc discharge has been mainly used for in the case of plasma spray, high-frequency discharge (including RF discharge, microwave discharge) has been used in many cases of PS. The high-frequency discharge has the following advantages. • Contamination-free film can be deposited, since no electrode abrasion and erosion occur during film deposition, because of its electrodeless discharge. • Film with high uniformity can be deposited due to higher uniformity of its plasma profile. • Feedstock materials can be sufficiently activated, because the residence time of feedstock particles in the plasma jet is long, due to low feed stock flight speed. However, high-frequency discharge has the following disadvantages. • Since it is relatively difficult to create an ultra-high temperature field in comparison with arc discharge, highly chemically reactive and non-equilibrium species cannot be generated. • A large-scale facility is required in case of high-power plasma generation, and solutions for electromagnetic interference are required. Therefore, recently, an arc discharge/RF discharge hybrid plasma generating system, which can additionally activate the arc discharge plasma jet by highfrequency discharge plasma, has been mainly used. As for the feedstock feeding system, since the liquid precursor does not contain solid phase materials unlike suspension, the feedstock injection to plasma jet can be conducted without taking into account the sedimentation of feedstock in any cases. Thus, the following feedstock feeding methods can be conducted. ➀ Direct injection of the feedstock (liquid precursor) into plasma jet (including input as mist) ➁ Injection of the vaporized feedstock into the plasma jet. (a) Control of feedstock flow rate in liquid phase state (state before vaporization). (b) Starting material control in gas phase state (state after vaporization). In method ➀, the same feedstock feeding system as SPS can be used. Moreover, the system can consist of low cost equipment, such as a commercial micro syringe (or micro tube pump), micro tube, suction type feeder (carburetor), and an ultrasonic atomizer (nebulizer). However, since the thermal energy of the plasma jet is lost as latent heat of the feedstock during evaporation when the feedstock is injected

164

Y. Ando

Fig. 11.2 Principle diagram of SPPS

into the plasma jet, low quality and non-uniform film is sometimes deposited due to insufficient feedstock activation in the plasma jet. In contrast, in method ➁, since the feedstock is injected into the plasma jet in the gas phase state, the fluctuation of the plasma jet due to the feedstock injection is small, and high quality and uniform film can be deposited. Since injection of the vaporized liquid precursor feedstock into plasma has been used as practical feedstock method in the cases of organic metal CVD (MOCVD: Metal Organic CVD), which is famous as a GaN film deposition method, a light-emitting element of a blue-lightemitting diode, method ➁ is mainly used as reliable method also in the case of PS (Fig. 11.2).

11.3 Reaction Element Process of Reaction in PS In the gas phase synthesis process, as in CVD, the following reaction element process is considered. (a) Fine particles formed through nucleation and grain growth in an atmosphere or plasma are deposited on a substrate to form a film (homogeneous nucleation, Figs. 11.3a, and 11.4 ➀ to ➄). (b) Nucleation occurs due to collision between the feedstock vapor and the substrate, and a film is formed by grain growth or crystal growth on the substrate (heterogeneous nucleation, Figs. 11.3b, and 11.4 ➀ to ➂, ➅). (c) After the feedstock vapor reaches the substrate, nucleation occurs in the process of migration on the substrate, and a film is formed by grain growth or crystal

11 Precursor Spray

165

Fig. 11.3 Illustrations of estimated film deposition mechanism

Feedstock injection Ionization and dissociation of feedstock Precursor creation by chemical reactions Homogeneous nucleation particle growth (nanoparticle creation) Accumulation of nanoparticles on substrate Collision between the feedstock vapor and the substrate Heterogeneous nucleation) Film growth Migration of precursor on substrate Heterogeneous nucleation Film growth

Fig. 11.4 Elementary process of SPPS

growth on the substrate (heterogeneous nucleation, Figs. 11.3b, and 11.4 ➀ to ➂, ➆). Since the low pressure precursor spray was developed as a fine particle manufacturing process, it has been considered for a long time that the film is deposited by the reaction element process (a). However, from the results in many successful cases thereafter, the following phenomena were confirmed. (i) Although a columnar structure film is deposited on the condition of high deposition temperature (substrate temperature), the film consisting of randomly oriented particles is deposited on the condition of low deposition temperature. (ii) The crystal structure and composition of the substrate surface affects the crystal structure and composition of the film.

166

Y. Ando

Therefore, a coating is assumed to have been deposited through all (or any) reaction element process of (a), (b), and (c) at the present time (Note). Note: Columnar structure film can be deposited by the thermophoretic phenomenon due to temperature gradient in the plasma jet and the shadowing effect because of the undulation of the deposited film surface, even on the condition of film deposition with reaction element process (a). Moreover, even in the case of fine particles formed through nucleation → grain growth, the crystal structure of the fine particle easily changes to the same structure as the substrate if the particle size is in the order of several nanometers. Therefore, in the research field of precursor spray, many researchers explain that films are deposited only by the reaction element process of a). However, a columnar structure film can be created since the same phenomenon (tendency) is thought to occur due to the film deposition mechanism with reaction element processes of (b) and (c) in cases of the conventional CVD and plasma CVD. The possibility of the reaction element processes (b) and (c) for the columnar structure film deposition in the case of PS was suggested in this chapter.

11.4 Case Study on Low Pressure PS 11.4.1 SiC Film Deposition Figure 11.5 shows fracture cross-sectional SEM images of SiC films deposited by low pressure PS equipment shown in Fig. 11.6 on the condition shown in Table 11.2 [10]. In the conventional and practical CVD, mono silane (SiH4 ) has been used as feedstock for SiC film deposition. However, since SiH4 is an explosive gas, it is dangerous to use this gas in the case of PS using thermal plasma on the condition of high feedstock rate, which is over ten times higher than those of the conventional CVD

50 μm a) Deposition temperature: 1623 K Fig. 11.5 Cross-sectional SEM image of SiC film

50 μm b) Deposition temperature: 1293 K

11 Precursor Spray

167

M ass flow controller

Scrubber

Plasma torch

Substrate

CH4 SiCl4 Feedstock Plasma jet

Working Gas (Ar + H2)

Exhaust

Heater

DC power source

+ Thermo couple

Vacuum pump

Vacuum chamber

Fig. 11.6 Schematic diagram of low pressure SPPS equipment for SiC film deposition

Table 11.2 SiC film deposition conditions

Working gas

Ar + 20%H2

Feedstock

SiCl4 /CH4

Discharge power

80 V, 300A

Chamber pressure

270 Pa

Deposition time

10 min

Deposition temperature

1133–1623 K

Ar + 20%H2 /SiCl4 /CH4 flow rate

10/1000/1000 sccma

Deposition distance

450 mm

Substrate

Graphite

a sccm:

Standard cubic centimeter per minute

processes. Therefore, silicon tetrachloride (SiCl4 ) was used as feedstock instead of SiH4 in this study. In this study, to avoid air pollution by corrosive exhaust hydrogen chloride gas, a scrubber was equipped in the exhaust system. Consequently, the film deposited consisted of fine columns could be deposited on the condition of 1623 K as the deposition temperature (T = 1623 K). However, although the columnar structure film with almost the same thickness as the film for T = 1623 K could be deposited also in the condition of T = 1293 K, each column of the film was wider than that of the film for T = 1623 K. Furthermore, it was demonstrated that the composition of the deposited film depended on the composition of substrate in this study. Even in the same condition of feedstock gases, though SiC film could be deposited in the case of SiC substrate, carbon-containing SiC film and Si films were deposited in cases of graphite substrate and Si substrate, respectively [11]. Figure 11.7 shows the relationship between degree of supersaturation σ , (σ = (p − pe )/pe , p: Vapor pressure of feedstock in atmosphere, pe : saturation vapor pressure)

168

Y. Ando

Fig. 11.7 Relationship between degree of supersaturation an

and microstructure of film in cases of the conventional CVD and plasma CVD. As shown in this figure, σ increases with lowering the deposition temperature in the condition of constant feedstock feed rate (σ increases with increasing the feedstock feed rate on the condition of constant deposition temperature). As σ increases, the deposited film structure changes from columnar structure film to the film consisting of randomly oriented particles. From these results, it was confirmed that PS had the same relationship between degrees of supersaturation and microstructure of film as those of the conventional CVD and plasma CVD.

11.5 Case Studies of Atmospheric PS 11.5.1 Diamond Deposition by the Combustion Flame Method The combustion flame method is a high rate diamond synthesis method using acetylene/oxygen combustion flame. Since diamonds can be synthesized using low cost commercial acetylene/oxygen welding equipment in the combustion flame method, this process has been studied by many researchers for practical use. Figure 11.8 and Table 11.3 show the schematic diagram of the combustion flame diamond synthesis equipment and diamond synthesis conditions, respectively. Diamonds could be synthesized by continuous irradiation of the substrate with the combustion flame for 5–30 min. As shown in Fig. 11.9, it was demonstrated that the diamond particles with not only cubic, but also octahedral geometry could be synthesized with the combustion flame method. This result suggests that the atmospheric PS has the potential to create films with crystal growth. As for the diamond growth in this process, the diamond particle became large (diamond growth rate became high), whereas number particle density of diamond particles decreased with increasing deposition temperature. Furthermore, it could be confirmed that diamond particle size (growth rate) and number density were also affected by the acetylene/oxygen ratio. The number density of diamond particles increased, whereas the diamond growth rate decreased with increasing acetylene/oxygen ratio. Therefore, since the two-step synthesis method, which consists of

11 Precursor Spray

169

Fig. 11.8 Schematic diagram of combustion flame diamond film deposition equipment

Table 11.3 Conditions of diamond deposition by combustion flame method

O2 flow rate

1.25 SLMa

C2 H2 /O2 flow ratio

0.90–1.30

Deposition distance

10 mm

Deposition time

20 min

Substrate

Mo

Deposition temperature

1073–1773

a SLM:

Standard Litter per Minute

Fig. 11.9 Optical micrograph of diamond particle on each deposition temperature

low temperature diamond synthesis in the first step and high-temperature diamond synthesis in the second step (or high acetylene/oxygen ratio in the first step and low acetylene/oxygen ratio in the second step), was thought to be effective to improve the yield (diamond synthesis efficiency), some researchers attempted to conduct the experiment, and good results were obtained [18–20]. Figure 11.10 and Table 11.4

170

Y. Ando

After the first step After the second step a) Condition #1 (Constant feedstock gas flow ratio, varied deposition temperature)

After the first step

After the second step

b) Condition #2 (Constant deposition temperature, varied feedstock gas flow ratio) Fig. 11.10 Optical micrograph of diamond particles in cases of two-step synthesis method

Table 11.4 Diamond synthesis conditions in the case of two-step synthesis method Condition #

Step#

Combustion gas flow ratio

Substrate temp (°C)

Operation Time (min)

1

1

1.15

1100

10

2

1.15

1200

10

2

1

1.18

1100

5

2

1.15

1100

15

show the deposition conditions in the case of the two-step synthesis method and optical micrographs of the diamonds deposited by the two-step synthesis method.

11 Precursor Spray

171

Fig. 11.11 Schematic diagram of atmospheric SPPS equipment for titanium oxide film deposition

Table 11.5 Titanium oxide film deposition conditions

Condition A

Condition B

Working gas

Ar

Working gas flow rate

20 l/min

Discharge power

75A, 35 V

175A, 25 V

Deposition distance

20–60 mm

100 mm

Deposition time

5 min

Feedstock

C2 H5 OH diluted TTIBa

Volume ratio of C2 H5 OH/TTIB

1/1

1/19

Feedstock feed rate

30 ml/h

100 ml/h

Substrate

430 stainless steel

a TTIB:

Titanium Tetra Iso Butoxide

11.5.2 Titanium Oxide Film Deposition by PS [21, 22] In cases of deposition of carbide film, nitride film, and metal compound film by PS, since oxidation of the film during deposition deteriorates properties of the deposited films drastically, film deposition has been carried out under a low pressure environment for a long time. In contrast, in the case of oxide film deposition by PS, since oxidation of the film during film deposition does not need to be taken into account, the development of practical atmospheric PS has been recently conducted by numerous research institutes.

172

Y. Ando

Fig. 11.12 Cross-sectional optical micrograph of titanium oxide film deposited on the condition B (deposition distance: 100 mm)

Figure 11.11 and Table 11.5 show the schematic diagram of the atmospheric PS equipment for titanium oxide film deposition and the film deposition conditions, respectively. In this case study, the liquid feedstock was injected into plasma jet without evaporation using the micro tube pump. Ethanol diluted titanium tetra iso butoxide (TTIB: Ti(OC4 H9 )4 ) was used as feedstock, and a 430 stainless steel plate was used as the substrate. The titanium oxide film was deposited by irradiating the substrate with an feedstock injected plasma jet for 5 min. Figure 11.12 shows the cross-sectional optical micrograph of the titanium oxide film deposited on the condition B. As shown in this figure, it was proved that atmospheric PS also had a potential for creation of columnar structure film. However, since the strength of the deposited titanium oxide films was only 2B according to the pencil scratch test (JIS K5600-5-4, ISO 15,184), the films were thought to be deposited not by crystal growth on the substrate, but by accumulation of fine particles created during flight in the plasma jet. Figure 11.13 shows Fig. 11.13 XRD patterns of deposited titanium oxide films in the condition B (◯: Anatase, ●: Rutile, ◆: Fe (Substrate), d: Deposition distance)

11 Precursor Spray

173

5 mm

5 mm 0 hour

48 hours a) d=60mm

0 hour

48 hours b) d=20mm

Fig. 11.14 Results of methylene-blue decoloration tests (d: Spray distance)

X-ray diffraction patterns of the titanium oxide films deposited in the condition B. Although anatase rich film could be deposited in case of 20-mm deposition distance (d = 20 mm), the degree of crystallinity of the film lowered with elongating the deposition distance and amorphous film (Note) was deposited in the case of d = 60 mm. Figure 11.14 shows the results of the methylene blue decoloration test. Methylene blue is a blue color reagent, which changes to transparent leuco-methylene blue by a redox reaction due to photocatalytic property of the deposited titanium oxide film. In this test, methylene blue was decolored by irradiating the deposited titanium oxide film located in a dark room with ultraviolet light for 48 h. Furthermore, in this test, the photocatalytic property of the deposited titanium oxide film could be confirmed by the wettability of the methylene blue droplet as well. Consequently, the film for d = 60 mm indicated hydrophobicity (suggesting non-photocatalytic behavior), and decoloration of the methylene blue droplet did not occur after the test. In contrast, the film for d = 20 mm indicated hydrophilicity (suggesting photocatalytic behavior), and decoloration of the methylene blue droplet occurred dramatically after the test. From these results, atmospheric PS proved to have high potential for use on the photocatalytic titanium oxide film. Note: Although no titanium oxide peak was confirmed in the XRD pattern of the titanium oxide film in the case of d = 60 mm, it was confirmed that this film was changed to anatase rich titanium oxide film by post heat treatment on the condition of 673 K for 1 h. Therefore, this film prior to the heat treatment can be regarded as an amorphous film [23].

11.6 Future Work As the issues in atmospheric PS should be solved for practical use, the low adhesion strength between film and substrate and low film strength are pointed out. These problems are thought to be caused because of the termination of dangling bonds of surface atoms of fine particles and the substrate by ambient air interrupting the creation of strong interatomic bonds among fine particles and between fine particles and the substrate. Therefore, to solve these problems, the following methods are proposed.

174

Y. Ando

➀ Addition of reducing gas, such as hydrogen, to the working gas By addition of a reducing gas, the dangling bond of surface atoms of fine particles and substrate are terminated by the reducing gas atoms. Since the bonding strength between the reducing gas atom and the dangling bond is much lower than that between ambient air, the reducing gas atom is easily replaced with another surface atom of fine particles or the substrate. Therefore, a strong interatomic bond between fine particles and between fine particles and the substrate can be created. ➁ Addition of sintering additive to feedstock Since the deposited film is heated by thermal plasma during operation, sintering of the deposited film occurs. Therefore, the film with high strength can be deposited by addition of the sintering additive. ➂ Equipment of reheating plasma torch downstream of the plasma jet By creation of fine particles by PS equipment upstream and the melting of fine particles by the reheating plasma torch located downstream, a high-strength film can be deposited by accumulation of molten fine particles.

11.7 Conclusion It has been over 30 years since the PS was developed in the 1980s. Although films with practical quality and properties could be deposited by PS at the early stage, PS is to date not used in practice due to the difficulties in the optimization of film deposition conditions. However, recent dramatic improvements of the liquid thermal spray technology made it possible to optimize film deposition conditions in precursor thermal spraying. At the present time, PS is practically used as a deposition process of the YSZ film for the TBC topcoat. In the future, further development of PS as a functional film deposition process, including for semiconductor thin films, is expected.

References 1. J. Karthikeyan, C.C. Berndt, J. Tikkanen, J.Y. Wang, A.H. King, H. Herman, Preparation of nanophase materials by thermal spray processing of liquid precursors. Nanostruct. Mater. 9(1), 137–140 (1997) 2. N. Wada, Preparation of fine metal particles by the gas evaporation method with plasma jet flame. Jpn. J. Appl. Phys. 8(5), 551–558 (1969) 3. T. Yoshida, K. Akashi, Preparation of ultrafine particles in a radio-frequency plasma. TetsuTo-Hanage 68(10), 1498–1502 (1982) (in Japanese) 4. A.K. Albessard, N. Matsuda, M. Tamatani, Spherical phosphors for projection CRTs. Toshiba Rev. 56(10), 39–42 (2001) (in Japanese) 5. A. Ema, Manufacturing process of fine particles for MLCC~grinding & classification system and RF plasma. Mater. Stage 7(3), 18–25 (2007) (in Japanese) 6. H. Murakami, T. Yoshida, K. Akashi, High-rate thermal plasma CVD of SiC. Adv. Cream. Mater. 3, 423–426 (1988)

11 Precursor Spray

175

7. S. Matsumoto, M. Hino, T. Kobayashi, Synthesis of diamond films in a RF induction thermal plasma. Appl. Phys. Lett. 51, 737–739 (1987) 8. N. Ohtake, M. Yoshikawa, Effects of oxygen addition on growth of diamond film by arc discharge plasma jet chemical vapor deposition. Jpn. J. Appl. Phys. 32(1), 5A, 2067–2073 (1993) 9. H. Murakami, H. Nagai, T. Irokawa, T. Yoshida, K. Akashi, Super high-rate thermal plasma CVD of ceramics. J. Ceram. Soc. Jpn. 97, 49–55 (1989) 10. Y. Ando, A. Mebata, M. Doi, Y. Kojima, Formation of oxide resistance silicon carbide films by thermal plasma CVD, in Proceedings of the 59th National Meeting of Japan Thermal Spray Society (1994), pp. 9–12 (in Japanese) 11. M. Lackner (ed.), Combustion Synthesis: Novel Routes to Novel Materials (Bentham Science Publishers, Sharjah (UAE), 2010) 12. H. Okayama, A. Nagata, H. Isa, Preparation of zirconia thin films for SOFC by PECVD. Trans. Inst. Electr. Eng. Jpn. 117-A(9), 936–941 (1997) (in Japanese) 13. Y. Fuji, S. Shimada, H. Kiyono, Preparation and wear resistance of TiBC TiBN, SiNX single layer film and TiBC-SiNX and TiBN-SiNX double layer film by thermal plasma CVD. J. Jpn. Soc. Powder Powder Metall. 54(4), 287–293 (2007) (in Japanese) 14. Y. Hirose, S. Amanuma, K. Komaki, The synthesis of high-quality diamond in combustion flames. J. Appl. Phys. 68, 6401–6405 (1990) 15. T. Mukai, M. Futamata, N. Kuramoto, Application of solution plasma spraying method to development of the titanium oxide film, in Proceedings of National Meeting of Japan Welding Society, vol. 67 (2000), pp. 468–469 (in Japanese) 16. “The 100HE Plasma Torch—A Technical Solution for Spraying Solution Precursor and Suspension Feedstocks” Progressive surface Co. web site, www.progressivesurface.com 17. ASM International Web News (Nov. 4, 2011), www.asminternational.org 18. Y. Ando, S. Tobe, T. Saito, J. Sakurai, H. Tahara, T. Yoshikawa, Enlargement of the diamond deposition area in combustion flame method by traversing substrate. Thin Solid Films 457, 217–223 (2004) 19. Y. Ando, T. Saito, S. Tobe, H. Tahara, K. Fujiuchi, T. Yoshikawa, Rapid doping of combustion flame CVD diamonds using supersonic expanding nitrogen plasma jets. Vacuum 73(3–4), 493–498 (2004) 20. M. Takahashi, S. Ito, O. Kamiya, T. Ohyoshi, Synthesis of diamond film on molybdenum substrate surface by combustion flame considering the delamination of the interface. J. Solid Mech. Mater. Eng. 1(2), 223–231 (2007) 21. Y. Ando, S. Tobe, H. Tahara, Photo-catalytic TiO2 film deposition by atmospheric TPCVD. Vacuum 80(11–12), 1278–1283 (2006) 22. Y. Ando, S. Tobe, H. Tahara, Rapid deposition of photo-catalytic oxide film by liquid feedstock injection TPCVD in open air. IEEE Trans. Plasma Sci. 34, 1229–1234 (2006) 23. Y. Ando, Rapid deposition of titanium oxide and zinc oxide films by solution precursor plasma spray. Trans. Inst. Electr. Eng. Jpn. 130-A(10), 935–940 (2010) (in Japanese)