# A review on quality control in additive manufacturing

Hoejin Kim (Department of Mechanical Engineering, University of Texas at El Paso, El Paso, Texas, USA)
Yirong Lin (Department of Mechanical Engineering, University of Texas at El Paso, El Paso, Texas, USA)
Tzu-Liang Bill Tseng (Department of Mechanical Engineering, University of Texas at El Paso, El Paso, Texas, USA)

ISSN: 1355-2546

Publication date: 9 April 2018

## Abstract

### Purpose

The usage of additive manufacturing (AM) technology in industries has reached up to 50 per cent as prototype or end-product. However, for AM products to be directly used as final products, AM product should be produced through advanced quality control process, which has a capability to be able to prove and reach their desire repeatability, reproducibility, reliability and preciseness. Therefore, there is a need to review quality-related research in terms of AM technology and guide AM industry in the future direction of AM development.

### Design/methodology/approach

This paper overviews research progress regarding the QC in AM technology. The focus of the study is on manufacturing quality issues and needs that are to be developed and optimized, and further suggests ideas and directions toward the quality improvement for future AM technology. This paper is organized as follows. Section 2 starts by conducting a comprehensive review of the literature studies on progress of quality control, issues and challenges regarding quality improvement in seven different AM techniques. Next, Section 3 provides classification of the research findings, and lastly, Section 4 discusses the challenges and future trends.

### Findings

This paper presents a review on quality control in seven different techniques in AM technology and provides detailed discussions in each quality process stage. Most of the AM techniques have a trend using in-situ sensors and cameras to acquire process data for real-time monitoring and quality analysis. Procedures such as extrusion-based processes (EBP) have further advanced in data analytics and predictive algorithms-based research regarding mechanical properties and optimal printing parameters. Moreover, compared to others, the material jetting progresses technique has advanced in a system integrated with closed-feedback loop, machine vision and image processing to minimize quality issues during printing process.

### Research limitations/implications

This paper is limited to reviewing of only seven techniques of AM technology, which includes photopolymer vat processes, material jetting processes, binder jetting processes, extrusion-based processes, powder bed fusion processes, directed energy deposition processes and sheet lamination processes. This paper would impact on the improvement of quality control in AM industries such as industrial, automotive, medical, aerospace and military production.

### Originality/value

Additive manufacturing technology, in terms of quality control has yet to be reviewed.

## Keywords

#### Citation

Kim, H., Lin, Y. and Tseng, T. (2018), "A review on quality control in additive manufacturing", Rapid Prototyping Journal, Vol. 24 No. 3, pp. 645-669. https://doi.org/10.1108/RPJ-03-2017-0048

### Publisher

:

Emerald Publishing Limited

## 1. Introduction

Additive manufacturing (AM) technology is a process of making a three-dimensional (3D) object-based on layer-by-layer or drop-by-drop deposition of materials under a computer controlled system. It is composed of various types of printing methods such as binder jetting, material extrusion, material jetting and sheet lamination. AM industries have grown rapidly since 2000, and have shown almost six times the growth during the 2000s as compared with the growth during the 1990s (Wong and Hernandez, 2012). AM technology has a high potential to exploit ideas and applications on various areas. In 2013, Wohlers Associates reported that AM system sales revenue of each sector are industrial products (19 per cent), consumer products (18 per cent), automotive (17 per cent), medical (14 per cent), aerospace (12 per cent), military (5 per cent). Automotive, medical, aerospace and military, which require high precision and reliability, lead 48 per cent in total. The global AM market sales reached up to $2.2bn in 2012 and is forecasted for growth of the market size by$10.8bn in 2020 (Wong and Hernandez, 2012). Therefore, final and functional part productions and relevant researches are growing faster than the general market. Most of AM applications were purposed for prototyping and tooling; however, nowadays, industries invest 10 times more on end-part production than on prototyping (Cotteleer and Joyce, 2014).

Likewise, this new paradigm of AM technology is being dramatically developed and brings high feasible applications into real industry. However, before AM product is directly used for real-functional part fabrication, certified quality control (QC) is of vital importance to address specific requirements, especially quality and reliability. For high-quality products that are used in aerospace and automobile industries, AM is required to certify high precision and mechanical properties to be used and functioned as real part (Wong and Hernandez, 2012). As a variety of materials are being used in the AM technology, such as ceramic, steel, alloy, composites and living tissue, compared with traditional manufacturing, the current AM technology around part quality and performance needs to be constantly improved. Therefore, more the variety of advanced materials being used with AM technology, the more quality processes that will be necessary.

For quality control improvement, there have been studies that focus on quality-approached researches. For example, Kaveh et al. (2015) studied the effect of printing parameters on precision and internal cavity of fabricated parts using fused deposition modeling 3D printer (Kaveh et al., 2015). Another research, Farzadi et al. (2015) investigated the effects of layer printing delay on physical and mechanical properties (Farzadi et al., 2015). If these types of quality-related researches on AM technology keep happening and produce improved quality compared with the traditional manufacturing, AM products will eventually substitute current technologies and expand to areas that require super-advanced technology in the near future.

This paper provides an overview of the research progress regarding the QC in AM technology and the future direction of AM development in the QC aspect, and especially discusses quality issues and the needs to be developed and optimized; further, it suggests ideas and directions toward quality improvement for future AM technology. This paper is organized as follows. Section 2 starts by conducting a comprehensive review of the literature studies on progress of quality control, issues and challenges regarding quality improvement in seven different techniques of AM. Next, Section 3 provides classification of the research findings, and lastly, Section 4 discusses challenges and future trends.

## 2. Overview of quality control in additive manufacturing

### 2.1 Photopolymer vat processes

Photopolymer vat processes (PVP) are based on the selective solidification of a liquid photopolymer using light sources (e.g. ultraviolet; Vaezi et al., 2013). Stereolithography (SL or SLA) is the main technique that is widely used especially in biomedical engineering field such as implant and tissue engineering (Gauvin et al., 2012; Cohen et al., 2009). Melchels et al. (2010) reviewed two different SL techniques: scanning SL and projection SL. The scanning SL builds a structure from bottom with scanning laser, while the projection SL builds from top to bottom with digital light projection (Melchels et al., 2010). The latter technique not only provides better surface finish and dimensional accuracy but also saves materials and fabrication time (Vaezi et al., 2013).

Some researchers (Huang and Jiang, 2003; Adolf et al., 1998) have documented significant heat release owing to thermal reaction between laser source and resin solution during photo-polymerization, which causes residual stress development associated with thermal shrinkage. Likewise, temperature profile has an important role on quality. Corcione et al. (2006) found that temperature increase on vat surface is dependent on energy dose during SL process. Lower scan speed leads to higher temperatures and reaction rates, which involve shrinkage and curl distortion (Corcione et al., 2006). Lee et al. (2007) studied the effects of laser energy with respect to thickness of building part and laser speed. Critical exposure (Ec) and penetration depth (Dp) are two primary parameters of resin determined by windowpane analysis. In addition, molecular weight of polymers (i.e. viscosity) plays an important role to determine final mechanical properties (Lee et al., 2007).

Specific design tools used in the biomedical engineering for patient such as computed-tomography (CT) and magnetic resonance imaging (MRI) enable the SL technique to remove elaborate design process and provide accurate custom-design data. In addition, these design tools allow evaluating the dimensional accuracy by comparing the scanned data with designed model. Cohen et al. (2009) used a commercial 3D workstation (EBW 4.0, Philips medical) which has capabilities of both MRI and CT, and can significantly enhance scanning time from 2 hours to 5 minutes for mandibular reconstruction (Cohen et al., 2009).

Cooke et al. (2003) incorporated the SL technique in manufacturing biodegradable polymeric scaffold for use in ingrowth repairs of human bone defects. In biodegradable resin (mixtures of Diethyl fumarate and poly pylene fumarate), it was observed that approximately 10 per cent of supports did not attach to the build table, which caused problems of dimensional/geometric inaccuracy and rough surface finish built parts. By customizing build tables and hole patterns that control the resin flow rate, the initial layer of cross-linked biodegradable resin mixtures can be securely adhered to the surface of the build table so that it can prevent the part from floating freely in the resin. It was found that the viscosity and density of vat solution influence the lack of adherence as well as design of build table and convex-surface effect (Cooke et al., 2003). Many efforts to fabricate complex tissue scaffold has been attempted to improve part quality with different types of solutions (Gauvin et al., 2012; Dhariwala et al., 2004). These attempts are increasing usage area and better quality control as variety of materials are being used (e.g. epoxy or acrylate or additional additives (PPF/DEF–HA, PDLLA/HA) as based resins with photo-initiators, support material such mix of propylene/polyethylene glycols, glycerin and/or acrylate). Developing materials to be used in the SL technique must take into consideration variety, composition, strength and finishing procedures to increase the fabrication versatility and quality (Gross et al., 2014; Bose et al., 2013).

Kim et al. (2010) decreased contamination when fabricating multi-materials using SL technique. When draining and cleaning the previous material before switching another vat, it was difficult to remove materials entirely owing to high viscosity of the previous resin. By using low viscosity polymer resin and improving a process-planning algorithm, contamination was reduced (Kim et al., 2010). Han et al. (2010) developed an automatic material switching method by cleaning out the current solution in the vat using solvent purging after UV-exposure so that it was able to fabricate high-quality heterogeneous 3D hybrid scaffolds (Han et al., 2010). For high quality fabrication process, Zhou et al. (2011) proposed to remove the previous material sticking on part surface using a soft brush for rough cleaning and then ultrasonic removal for final cleansing before introducing a different material (Zhou et al., 2011).

Orientation in which a design is fabricated can influence the SL performance. Lan et al. (1997) investigated that deposition orientation mostly influences on three significant parameters: surface quality, build time and support structures. On the basis of the parameter consideration, decision criteria and algorithms were established and developed to identify the desirable fabrication orientations (Lan et al., 1997). Pham et al. (1999) developed feature-based system using an object-oriented programming language and solid modeling CAD environment to obtain the best tradeoff by considering factors based on build time, cost and accuracy while having optimal part orientation (Pham et al., 1999).

The SL technique has also been used in micro/nano scale fabrication (Maruo, 2015). Ha and Yang (2014) studied that surface friction factor taken into consideration when fabricating closed structures does not affect open complex structure. Therefore, a micro-scaled open structure was erected securely on the substrate from the mentioned process, as shown in Figure 1. This nanostructure fabrication is capable of having high quality structures with higher efficiency than closed structures. In this study, 10 × 10 × 10 µm open structures were fabricated, CCD camera were used to monitor the nanoscale structure during fabrication (Ha and Yang, 2014).

West et al. (2001) developed a new adaptive slicing algorithm in the design process for improving build performance. This algorithm assists SL operators in developing a process plan by quantifying tradeoffs among surface finish, geometry tolerances and build time. By adjusting process variables within algorithm, it can predict geometric and dimensional properties, and can compare the desired and actual values so that SL operators can approach the specific design requirements (West et al., 2001).

In summary, PVP has advanced in providing accurate data acquisition during design applications that use CT and MRI, which makes this technique particularly useful in biomedical applications. In addition, these procedures allow part evaluation by comparing design and 3D printed model. Finding optimal build orientation to improve the quality and build time was mostly studied by incorporating a new type of solid modeling and slicing algorithm. This indicates that build orientation mostly affects the quality. For printing process, there are still many quality and process problems that occur owing to vat contamination, wasting materials, long process time, material limitation and detachment of build table. Living biological cells can be incorporated using this technique so that human artificial organ or skin can be fabricated. In the near future, more materials will be incorporated which will, in effect, demand more quality related studies on the variety of materials needed.

### 2.2 Material jetting processes

Material jetting processes (MJP) are based on the use of inkjet printing technique to deposit droplets selectively through a nozzle or orifice to build up a 3D structure. This droplet is a liquid that is jetted out of the nozzle and becomes solid after deposit through cooling or UV curing. There are two methods for creating appropriate material droplets: drop on demand (DoD) and continuous inkjet (CIJ). The latter uses fluids of low viscosity and high drop velocity for fast printing speed; in contrast, the former uses smaller droplets and velocity for high dimensional accuracy; therefore, DoD has a higher potential for microfabrication (Vaezi et al., 2013). In the DoD system, two main kinds of pressure pulse methods in the manner of ejecting droplet from ink vat are divided: thermal and piezoelectric actuators. Thermal DoD has a limitation on ink material used, which is volatile, while the piezoelectric DoD has no such limitations, but is required to have droplet generation rates in the range of 3-10 kHz to allow the meniscus to reach capillary equilibrium (Sachs and Vezzetti, 2005).

Vaezi et al. (2013) reviewed the several significant phenomena affecting the quality of the material jetting processes. The shape of deposited droplet influences critically on resolution, precision and accuracy (Vaezi et al., 2013). Beaman et al. (2004) demonstrated through fluid mechanics theory that there is an important relationship between the Reynolds number Re = ρgvd/μ, denoted as a ratio between inertial and viscous forces, and Weber number We = ρ1v2d/σ, denoted as a ratio between kinetic and surface energy, where ρg and ρ1 represent the densities of the process gas and liquid drop, respectively. The variables v, d, µ and σ are droplet velocity, droplet diameter, liquid dynamic viscosity and liquid surface tension, respectively. It was found that Re and We should satisfy 1<Re/We<10 to maintain the overall printing quality of DoD system (Beaman et al., 2004). This author reported that droplets should not exceed more than 10 µm in diameter owing to problems with air resistance. Ko et al. (2010) reported that fundamentally considerable and interconnected conditions for high desired quality control on MJP are the following:

• ink properties (viscosity and surface tension);

• jetting parameters (signal width, voltage magnitude and jetting frequency); and

• environment (pressure, environment and substrate temperature and humidity).

Research on metal nanoparticle inkjet printing investigated thoroughly that proper substrate temperature and sufficient drying of nanoparticle inks in each layer guarantee high quality (Ko et al., 2010).

Jiang et al. (2010) studied on the accurate micro scale-droplet creations. As shown in Figure 2, droplets merge with each other to form bigger droplets owing to different droplet velocities. These merged droplets degrade the droplet’s diameter accuracy, which significantly affects surface roughness and total geometry. It was observed that uniform spherical droplets can be developed at optimal frequency of 2.842 kHz and at low oxygen concentration environment (Jiang et al., 2010). Kullmann et al. (2012) contributed on building 3D microstructures of gold nanofluids with high aspect ratio and low-cost basis using DoD technique. It was discussed that not only are accurate and optimal printing temperatures and voltages essential factors in nanofluids’ control for micro scaled structures, but also that thermal optimization such as substrate temperature and laser annealing are to be taken into consideration (Kullmann et al., 2012).

Inkjet printing technique is used widely in biological tissue engineering and electric circuit design due mainly to its multi-material capability (Nakamura and Henmi, 2008; Mei et al., 2005; Xu et al., 2006; Cui and Boland, 2009; Czyżewski et al., 2009). However, Rengier et al. (2010) rated MJP technique on medical applications as inferior in accuracy and strength in comparison with other techniques (SLA, SLS and FDM; Rengier et al., 2010). Ebert et al. (2009) found the process-related defects as a result of single clogged nozzles owing to either dried up or blocked by agglomerates while printing dental prostheses made of zirconia-based ceramics (Ebert et al., 2009). To resolve this issue, sintering was adopted as post processing to enhance the strength and accuracy, and preheating substrate was implemented to avoid internal stress and bending of layers during drying. In addition, aqueous solution of 10 wt. per cent ethanol was found to be sufficient for nozzle cleaning to prevent clogging.

Wüst et al. (2011) mentioned that protein and biological cells are more likely to be denaturant depending on temperature; therefore, temperature of ink bubbles in thermal-based printing should not exceed more than 41°C (Wüst et al., 2011). Owing to this reason, Wilson and Boland (2003) modified commercial inkjet printers by implementing temperature controller to avoid denaturation of proteins and cells (Wilson and Boland, 2003). Henmi et al. (2007) developed a gelation technique to align printed cell on the substrate without drying and distributed randomly so that living cells and biomaterials can be easily patterned to create sophisticated 3D printed tissue (Henmi et al., 2007). Cui et al. (2010) noted the frequency of piezoelectric inkjet printers to be within a frequency range that causes cell damage and lysis after sonication at 15-25 kHz (Cui and Boland, 2009). For a thermal inkjet printer, its heating temperature at the ejected nozzle increases up to 300°C for a couple of microseconds during the printing. Therefore, there are significant concerns in printing processes that cause cell damage and death (Cui et al., 2010). In 2012, authors comprehensively evaluated cell viability and damage on thermal inkjet cells and found 4-10°C above ambient for only 2 µs avoids cell damage up to an average cell viability of 90 per cent. Therefore, thermal inkjet printing technology is becoming more biocompatible to the living system compared to piezoelectric printing (Cui et al., 2012).

Derby (2010) and Yeong et al. (2006) discussed the importance of fluid properties for quality and investigated the main factors of quality in the MJP technique to be:

• generation of droplet;

• position and interaction of droplet on a substrate; and

• solidification mechanism.

In terms of interaction between droplets and substrates, part quality is significantly determined by properties such as gravitational forces, fluid density, surface energy, drop velocity, droplet size, surface tension, capillarity forces, contact angle, drop spacing and print head traverse speed. In addition, during droplet transition from liquid to solid, which may occur, volume change is important and can be controlled by these considerations such as solvent distribution, fluid flow rate and evaporation rate. Proposed nano/micro-scale patterning and structuring of substrate can overcome this resolution issue limited by drop size and can achieve higher resolution more than drop size (Derby, 2010; Yeong et al., 2006).

Tourloukis et al. (2015) identified factors that affect quality issues and printing performance and studied the potential of computational intelligence algorithms particularly in neural networks to assess and forecast quality of 3D inkjet-printed electronic products. Results validated that one-step ahead predicting model using nonlinear autoregressive neural network with external input brought up better overall prediction behavior compared with multi-step ahead one. However, more investigation is needed with larger data set to be trained with respect to a wider range of values and trends for enhancing predictive performance (Tourloukis et al., 2015).

Recently, Sitthi-Amorn et al. (2015) developed a material jetting systems allowing for self-calibration of print heads, 3D scanning and a closed-feedback loop to enable print corrections. Real-time printing control technique of machine vision built in enables to correct overall platform design and printing parameter in real-time through closed-feedback loop. The image processing algorithm computes depth difference between droplets and adjusts amount of jet droplet. This technique opens up the real-time quality control during printing process. In addition, 3D scanning (i.e. optical coherence tomography, OCT and scanner) provides high resolution (less than 1 µm) to scan the depth map and overlay the model to be printed over the 3D scanned auxiliary component on the build platform. Furthermore, users can interactively translate and rotate the model over the image to facilitate the alignment (Sitthi-Amorn et al., 2015). (Figure 3)

In summary, MJP have significantly advanced in micro-scale fabrication. Most of the studies implemented for quality related issues aim to control the droplet mechanism, which significantly affects resolution and dimensional accuracy. The main issue during printing process is nozzle clogging occurring owing to dry and agglomerates, which are detrimental to medical application. Using thermal inkjet printing technique, sophisticated tissue structures such as biomaterials and living cells can be easily fabricated without cell damage caused by piezoelectric inkjet frequency and high temperature. Positive early stage research was made on data-driven approach using neural networks to predict the quality. A great achievement of MJP is a real-time closed-feedback loop system incorporated with machine vision. This can correct the design and printing parameters in real-time and eventually open up real-time processing during printing process.

### 2.3 Binder jetting processes

Binder jetting processes (BJP) have similar principals to material inkjet process, but based on nozzles depositing droplets of a water-based binder material over surface of a powder bed (Wong and Hernandez, 2012; Stucker, 2012). The powder particles are held together by bonding with deposited binder material and eventually shape 3D structures. This process is followed by lowering the powder bed through a piston and a fresh layer of powder spread over the previous layer and deposited by binder again over the new layer. The possible materials used as powder includes ceramics, metals, shape-memory alloys and polymers (Vaezi et al., 2013). Gibson et al. (2014) found that BJP technique provides better quality control ability for ceramic and metal fabrications to be produced compared with MJP technique. However, a BJP printed part tends to have lower accuracy and surface finish than the part made by MJP. In addition, infiltration processes are typically needed to fabricate dense parts or to ensure good mechanical properties (Gibson et al., 2014).

Deposition ways of the binder are the same as the MJP technique. The CIJ system of BJP technique is designed to print in either uni or bidirectional mode. Bidirectional printing mode is faster than unidirectional one, but causes dynamic shift errors. A part requiring higher quality surface finish should be printed in unidirectional mode. Critical point on quality in this system is a catcher, which helps to achieve a selective printing stratagem in the presence of a continuous stream as shown in Figure 4(A). Figure 4(B) display how the catcher is operating. Deflected plates are guiding droplets to catchers when unwanted droplet and drained for reuse. However, owing to some disadvantages such as binder accumulation, pump clogging and low cross-section of individual catchers, Sachs and Vezzetti (2005) proposed a single catcher system that is shared by all the nozzles, proportional deflection reducing the nozzle gap and 45 degree print head rotation improving dynamic shift and enabling the droplets to land on exact powder bed at certain time to improve surface finish and dimensional accuracy, as shown in Figure 4(C) (Sachs and Vezzetti, 2005). By comparison, the DoD method provides better quality (e.g. resolution and minimum feature width) but is slower in regard to printing process (Cima et al., 2001).

The BJP is capable of achieving accurate geometries in nano-biomaterial fields (Liu et al., 2013). Factors that determine the final part dimensions are accumulative accuracy of deposited layer thickness, accuracy of droplet placement, reproducibility of droplet spread and of dimensional change during curing process. In addition, the dimensional accuracy is strongly influenced by interaction performance between powders and binders such as powder’s surface treatment, size, shape, packing density, distribution, binder’s viscosity, surface tension, droplet size, velocity and temperature.

Chumnanklang et al. (2007) studied the effects of binder concentration in pre-coated particle on part strength. It was found that part strength can be increased by increasing the binder concentration and pre-coated particle size. The hydroxyapatite particle used was coated with maltodextrin and water. Coated powders have easier-to-flow characteristic than raw powders so that it could lead to high flow-ability and uniform distribution. These higher binder concentration and pre-coated particle enables better densification of intra-particles during sintering process (Chumnanklang et al., 2007). For better powder distribution, Snelling et al. (2013) used a sieve test to determine particle size distribution, which is of critical importance to mechanical properties and dimensional accuracy (Snelling et al., 2013).

Lu and Reynolds (2008) analyzed printing variables mainly influenced on integrity, dimensional accuracy and minimal feature sizes that can be printed for less than 500 µm 3D mesh structures (TiNiHf shape memory alloy). This author classified three groups in terms of the variables: 3D printer-, powder- and binder-related variables. Two major 3D printing variables (i.e. printing layer thickness and binder saturation) are evaluated. It was found that breaking strength increases with the binder saturation level up to 170 per cent at the same printing layer thickness, and 35 µm printing layer yields 3D mesh structure with the highest integrity and the lowest dimensional deviation (Lu and Reynolds, 2008).

In bone tissue engineering area, Farzadi et al. (2014) investigated the effects of layer thickness and printing orientation (parallel to X, Y and Z) on dimensional accuracy and mechanical properties of printed porous scaffolds. It was found that concurrence between model’s orientation (i.e. longitudinal direction) and printing head movement results in better dimensional accuracy and mechanical properties, and layer thickness as well as printing direction, which has a significant effect on the compressive strength of scaffolds (Farzadi et al., 2014).

Some of researchers have used mathematical analyses to improve the quality for BJP technique. Asadi-Eydivand et al. (2016) found that the delay time of applying the new layer and printing orientation significantly influences the dimensional accuracy, compressive strength and porosity. A full factorial design of experiment (DOE), signal to noise (S/N) ratio and analysis of variance (ANOVA) were used to investigate optimal process parameters (Asadi-Eydivand et al., 2016). Another research used statistical analysis (ANOVA) to determine significant factors affecting product quality. Parameters of binder saturation, layer thickness, building orientation were optimized through statistical analysis (Uma Maheshwaraa et al., 2008). In addition, BJP technique has been studied for quality in metal casting area (Meisel et al., 2012; Snelling et al., 2013), foundry engineering (Budzik, 2007) and pharmaceutics (Yu et al., 2009).

In summary, BJP have highly advanced in ceramic and metal fabrications compared with other techniques because BJP can fabricate high melting point materials at low temperature using binder and post-treatment. The impressive achievement in BJP was a study on DoD to develop a single catcher system on each nozzle to control individual droplet and pre-coating particle to enhance part strength by solving drawback of BJP such as low density and porosity. In addition, mathematical analyses were conducted to find optimal printing parameters. The main issues during printing process were the interaction performance between binder and powder, accumulative accuracy of deposited layer thickness, droplet placement, the delay time of applying a new layer and dimensional change during post-treatment. These issues significantly influence dimensional accuracy and challenges to investigate for better quality.

### 2.4 Extrusion-based processes

Extrusion-based processes (EBP) deposit material in form of a continuous flow layer-by-layer to build a 3D structure. These processes have diverse technique concepts but are classified into two main groups: melting-based extrusion and non-melting-based extrusion (Vaezi et al., 2013). 3D-bioplotting is the most representative of non-melting-based extrusion techniques and fused deposition modeling in melting based extrusion techniques. In here, fused deposition modeling (FDM) will be the main point of discussion.

#### 2.4.1 Researches on general quality control

In the FDM technique, there are two kinds of deposition methods mainly used: contour and raster. In contour method, extrusion path is following outside contour and offset in until it fills the entire domain. This technique’s, main disadvantage is slow printing operation owing to multiple paths creation. In a raster method, the material is deposited in zig–zag fashion. Raster method is a faster process but less accurate and gives rise to voids in deposition process in comparison with the contour one. This can lead to deterioration in the in-plane properties of the structure. Kulkarni and Dutta (1999) proposed spiral geometric path, which can eliminate multiple paths and inaccuracy to provide faster process and reduce voids. In result, in-plane stiffness spiral method was improved up to 18,000 lb/in2 compared with 10,300 lb/in2 of contour paths (Kulkarni and Dutta, 1999).

Sun et al. (2008) evaluated the effects of processing conditions in the bonding quality of FDM technique. It was found that the neck growth between adjacent filaments of the same layer is expected to be larger in the bottom layers than in the top layers because bottom layer remains longer period than upper layers, indicating voids size in the lower region is clearly smaller than in the upper region. It is the fact that convective heat from adjacent extruded molten layers during process allows neck growth between filaments, which occurs above glass transition temperature. This neck growth can be increased by proposed lateral geometric path to have slow cooling rate. Therefore, bonding strength and voids were improved and minimized, respectively (Sun et al., 2008). (Figure 5)

Boschetto and Bottini (2016) used a design for manufacturing (DFM) method to overcome inconsistent geometrical deviations that cause poor surface quality and depend on layer thickness, deposition angle, etc. Proposed DFM permits a redesign of the components during preprocessing to predict the obtainable dimensional deviations. Modifications carried out in design stage can compensate for the deviations generated in the next manufacturing stage, thus, providing improved part accuracy, as shown in Figure 6 (Boschetto and Bottini, 2016).

Reducing in build time and increasing part surface quality contradict each other and there have been a number of attempts to solve these issues. Pandey et al. (2003a, 2003b) proposed a method of real-time adaptive slicing procedures. This method predicts surface quality expressed by standard Ra value using direct slicing and tessellated model (STL) so that in the design stage, build time and surface quality can be maximized based on the predicted Ra value (Pandey et al., 2003b). Huang et al. (2009) proposed a robust algorithm to generate the support slice data that enable part’s self-support ability and reduce redundant support volume at maximum extent. This slice data based algorithm has the same efficiency as the STL based algorithm, however, it is more stable and robust to generate support slicing process. The advantage of this algorithm provides necessary support for hanging not only structure surface but also vertexes and edges, which can guarantee dimensional accuracy (Huang et al., 2009).

In the area of casting, Kaveh et al. (2015) used high impact polystyrene (HIPS) materials to fabricate wax patterns for casting technology that requires accuracy without nay internal cavity. They designed proper benchmarks regarding printing parameters and statistical equations that can calculate calibration factors (precision and internal cavity) to prevent errors between designed and actual dimensions; eventually, these can be used to determine optimal printing parameters (Kaveh et al., 2015).

#### 2.4.2 Researches on surface quality and post-processing

FDM technique presents limitations in surface quality. Staircase effect influences surface quality of 3D printed part using thick filament approximately 0.24 mm, which is the most used thickness and 0.127 mm at the most. This is lower resolution than other techniques such as SL (0.05 mm), SLS (0.02 mm), BJP (0.05 mm) and MJP (0.016 mm). Other problems are related to thermal distortion and stress resulting from shrinkage (Boschetto and Bottini, 2015).

Armillotta (2006) assessed textured surface quality printed by FDM technique. Critical issues on surface quality are dependent on the rate of material deposition, shrinkage and residual stresses during solidification. It was found that sufficient resolution of surface without stair-stepping, as shown in Figure 7(A), can be obtained under the conditions provided which feature in-plane size ≥ 1 mm on surfaces parallel to the build direction can be fabricated without excessive stair-stepping. Moreover, sinking effects made by contour offsetting, as shown in Figure 7(B), can be avoided when adjacent features are spaced out at least 1 mm away from one another (Armillotta, 2006). Campbell et al. (2002) implemented a visualization algorithm (AutoLisp language within AutoCAD 14) that represents varying surface roughness of 3D designed model as color shading within a CAD image. Using this algorithm, the designer is given clear visual image of the overall surface roughness of a model and any potential problem areas (Campbell et al., 2002).

Galantucci et al. (2009) proved significant improvements on surface finish of FDM 3D printed acrylonitrile butadiene styrene (ABS) part using the chemical post-processing method. Dimethyl ketone (acetone) was used with 10 per cent of water as chemical because ABS has weak interaction with polar solvents such as dimethyl ketone, ester and chloride solvents. Therefore, chemical reaction is controlled by the soaking and exposure time to ABS parts (Galantucci et al., 2009). In biomedical micro-devices, McCullough and Yadavalli (2013) also used the same chemical as dimethyl ketone-based sealing method to modify and seal the surfaces of ABS to render the surface of 3D printed micro-structured features to be impervious to water and increase hydrophilicity and biocompatibility, as shown in Figure 8 (McCullough and Yadavalli, 2013). With this chemical method, Galantucci et al. (2010) studied how this chemical reaction affects mechanical properties as well as surface finish. It was observed that mechanical properties such as ductility, flexural and compressive strength were improved as treated chemicals but tensile strength was slightly reduced because of a different action of the solution on different surface pattern or different reaction of treated filaments (Galantucci et al., 2010). In addition, Percoco et al. (2012) studied compressive strength and surface roughness of FDM 3D printed ABS part using chemicals, 90 per cent dimethyl ketone solution and 10 per cent water. Results showed an increase in compressive strength and reducing roughness up to 90 per cent when the immersion time is at 300 s, as compared with non-treated part (Percoco et al., 2012).

Addanki Sambasiva Rao et al. (2012) used design of experiment, ANOVA, to find significant factors affecting surface finish by analyzing various parameters in chemical treatment process, such as different levels of solution concentration, exposure time, temperature and initial roughness. Two different chemicals, dimethyl ketone and methyl ethyl ketone, were used. In the case of dimethyl ketone, solution concentration, temperature and initial roughness were significant factors that mostly affect surface roughness. For Methyl ethyl ketone, solution concentration, temperature and exposure time were the most important factors to be surface quality (Addanki Sambasiva Rao et al., 2012). Moreover, vapors treatment was introduced by Sratasys. Vapors of tetrahydrofuran created by applied heat interact with surface of 3D part, which settles in the side of the closed vessel with a non-air tight lid (Stratasys, 2016). However, this chemical treatment can dramatically improve surface finish but has been only applied to ABS polymer and drawback limits using to chemical sensitive materials. In addition, it provides poor repeatability, small variations in geometry of the part and unpredictable dimensional accuracy.

Another method used was hot cutter machining, as shown in Figure 9(A). Pandey et al. (2003a, 2003b) addressed a problem of surplus stock material, necessary tooling phase and path planning despite improved surface at low time (Pandey et al., 2003b). A method mostly used for surface finish is a barrel finishing reproducing better surface finish using abrasive or lubricative polishing agents with 3D printed parts put in rotating closed chamber creating friction against each other, as shown in Figures 9(B) and 9(C). This method provides high repeatability and surface finish. Boschetto and Bottini (2015) studied the quality control for predicting surface roughness after barrel finishing by using theoretical models developed through statistical analysis validated in comparison with experimental results (Boschetto and Bottini, 2015).

Many researchers have found that build orientation affects surface quality on FDM parts. Vijay et al. (2011) investigated that surface roughness is proportional to the layer thickness for 20° and 45° build orientation; however, roughness is an inverse proportion to layer thickness when 70° is used for building orientation (Vijay et al., 2011). Sreeram (1995) developed an auto-compute method to decide optimal orientation direction based on variable slicing thickness for a polyhedral part (Sreeram PN, 1995). Frank and Fadel (1995) proposed a user-friendly decision-making tool that recommends the optimal orientation direction before printing (Frank and Fadel, 1995).

Chakraborty et al. (2008) developed a new method for curved layer 3D structure. This curved layer FDM (CLFDM) method has an advantage in fabricating thin and curved part with improved surface finish owing to reduction of stair–step effect as well as the reduction of layers. The method also increased mechanical strength, reduced material waste and build time, as shown in Figure 10(A). In regular FDM, table and head movement do not deviate from vertical (z-axis) when x- and y-axes are being controlled, as shown in Figure 10(B). However, the proposed method uses x- and y-axes with contouring control and z-axis with linear interpolation control. Therefore, the table and head movement can have a 3D linear interpolator for the deposition of curved layers. In addition, they worked on determining adjacent filament paths for proper reproduction of part shape (Chakraborty et al., 2008).

Ahn et al. (2009) proposed a theoretical model to express surface roughness with respect to surface angle θ. As the surface angle is closer proportionally to fabrication direction, surface roughness improves quality. The validity of this proposed model was further proved by comparing empirical data with computed values (Ahn et al., 2009).

#### 2.4.3 Data analytic and algorithm researches

Many researchers used statistical analyses to optimize both processes and product design for quality control. Anitha et al. (2001) performed experiments to assess the effects of the parameters on quality characteristics using the Taguchi method. Analysis of the Taguchi results using methods of signal to noise (S/N) ratio, ANOVA, correlation and regression analyses indicated that without pooling only one layer thickness is effective to 49.37 at 95 per cent level of significance. But on pooling, the layer thickness is mostly effective to 51.57 at 99 per cent level of significance. This result was further strengthened by afore mentioned analyses which indicate a strong relationships with surface roughness. Moreover, based on the S/N analysis, the layer thickness was mostly effective at 0.3556 mm, road width at 0.537 mm and deposition speed at 200 mm/s (Anitha et al., 2001). Similarly, Lee et al. (2005) used the Taguchi method and orthogonal array, main effect, S/N and ANOVA analyses to investigate the optimal printing parameters needed to achieve maximum flexibility of ABS material. Through this statistical analysis, it was concluded that layer thickness, raster angle and air gap has significant impact on the elastic performance of ABS printed structure (Lee et al., 2005). Pandey et al. (2003a, 2003b) studied the improvement of surface finish by staircase machining using the FDM technique. They proposed hot cutter machining method to improve the surface of a staircase structure on 3D printed parts and used ANOVA and fractional factorial design of experiment to analyze the effects of three important machining parameters (i.e. cutting speed and direction) and for developing statistical model, respectively (Pandey et al., 2003a). Sood et al. (2009) adopted the Grey–Taguchi method to obtain optimum level of printing parameters by finding the significant factors and their interactions and to minimize the percentage change in length, width and thickness into a single objective known as the Grey relation grade. This Grey–Taguchi method provided good results on overall improvement in part dimension. In addition, artificial neural networks (ANN) was used to predict overall dimensional accuracy and provide errors between predicted data and observed value varying between 0 and 3.5 per cent. Smaller error percentages mean better suitability of present model (Sood et al., 2009).

Sood et al. (2010) conducted central composite design for design of experiment and ANOVA for validity to investigate the important impact of printing parameters related to mechanical properties such as tensile, flexural and impact strength. It was concluded that the increase in layer thickness, which means decreased number of layers required, leads to minimizing distortion effects within the structure so that total mechanical strength will be increasing (Sood et al., 2010). Sood et al. (2012) studied that compressive strength of 3D structure exhibits high dependence on five important printing parameters (i.e. layer thickness, part orientation, raster angle/width and air gap). This research not only provides an insight into the complex dependency of compressive stress on process parameters but also created a statistically validated predictive equation to find optimum parameters for maximum compressive stress using a quantum-behaved particle swarm optimization (QPSO). However, to solve involved large number of conflicting factors and complex phenomena for part building to predict output characteristics accurately. ANN with back propagation algorithm was adapted to predict the final structure’s compressive stress as shown in Figure 11. In result, this optimized statistically predictive equations that gave a maximum compressive stress of 17.4751 MPa at optimum values of layer thickness, orientation, raster angle, raster width and air gap as 0.254 mm, 0.036°, 59.44°, 0.422 mm and 0.00026 mm, respectively (Sood et al., 2012). Li et al. (2010) also used particle swarm optimization algorithm to obtain appropriate fabrication orientation by optimizing support area, fabrication time and surface roughness so that it was proved to enhance processing efficiency and reduce the costs significantly (Li et al., 2010). Rao and Rai (2016) used the teaching–learning-based optimization algorithm (TLBO) and non-dominated sorting TLBO (NSTLBO) to solve optimization problems on FDM. When it comes to objective function value with a higher convergence rate, it was proved that TLBO shows better performance as compared to QPSO and generic algorithm (GA) in terms of solving single-objective optimization aspect and that NSTLBO demonstrated better performance as compared with non-dominated sorting genetic algorithm (NSGA-II) in terms of solving multi-objective optimization aspect (Rao and Rai, 2016).

Ahn et al. (2002) used design of experiments to study the relationship among the printing parameters of ABS material such as raster orientation, air gap, bead width, color and model temperatures. It was found that the air gap and raster orientation significantly influence the tensile strength of printed part a lot more than bead width, model temperature and color (Ahn et al., 2002). Thrimurthulu et al. (2004) used real coded genetic algorithm to obtain optimum parameters of part deposition orientation, which significantly affects the surface roughness and build time. Also, this optimized parameter considers minimization of support structures estimated by genetic algorithm. This genetic algorithm helped predicting surface quality and build time (Thrimurthulu et al., 2004).

Patel et al. (2012) emphasized that the Taguchi method is a versatile tool for process and product design optimizations and to solving current quality control issues on 3D printing technology (Patel et al., 2012), based on the references (Sood et al., 2009, 2010; Shaji and Radhakrishnan, 2003; Çaydaş and Hasçalik, 2008; Gopal and Chakradhar, 2012).

#### 2.4.4 Researches on composites

Mostafa et al. (2009) investigated the behavior of melt flow of ABS–iron based composite materials deposited by FDM technique. 2D and 3D numerical analysis (i.e. CFD, FEA and ANSYS FLOTRAN/CFX software program) were used to analyze melt flow parameters: mainly temperature, velocity and pressure drop at extrusion. These analyses help prevent clogging and allow smooth flow out of the extrusion and enhance quality of composite based 3D structure by predicting the flow behavior (Mostafa et al., 2009). Bellini and Güçeri (2003) developed a novel FDM technique mounted a mini-extruder on deposition of FDM machine for agile fabrication. It streamlined filament preparation for continuous process of fabrication ECG9/PZT ceramic composites. This high-precision-based extruder was designed for a wide range of materials as well as ceramic based composites. The author studied flow behaviors influencing quality in terms of granules size, nozzle temperature and nozzle size during extrusion and deposition. Two thermocouples were used at both the entrance and exit of the extruder qualifier to determine the influence of temperature in upper and lower parts of the qualifier (Bellini and Güçeri, 2003).

In summary, EBP has been studied in various research approaches as they are the most popular processes commercialized in public. The general quality issues currently identified in FDM technique are the surface quality arose from staircase effect, limited layer thickness, thermal distortion shrinkage causing low resolution and dimensional inaccuracy. Many kinds of post-processing and build orientation methods have been investigated to alleviate the surface and mechanical property issues. This EBP has advanced especially in data analytics and algorithms, which signify that other fabrication techniques can possibly adopt to predict their optimal design and printing parameters. Research in these fields can be the next step to understand the effects of various parameters on the quality of fabricated parts.

### 2.5 Powder bed fusion processes

Most of the processes in powder bed fusion (PBF) system use a high source of thermal energy to melt powder-based materials (i.e. polymer, metal and ceramic) into desired structures and shapes (Stucker, 2012). In the PBF processes, three different techniques have been developed: selective laser sintering (SLS) technique, which uses partial melting to fuse powders together, selective laser melting (SLM), which uses full melting on powders transferring into homogenous parts, as well as fusing and electron beam melting (EBM), which uses an electron beam to design to fuse only metal powders in vacuum environment. For medical applications, SLS technique is widely used for human skeleton or organ reconstruction owing to its high accuracy and repeatability when compared to FDM, LOM and inkjet printing techniques (Silva et al., 2008; Rengier et al., 2010).

Edwards et al. (2013) studied mechanical property improvement of Ti-6Al-4V alloy part printed by EBM. By elevating certain bed temperature, residual stresses were minimized; therefore, post heat treatment could be eliminated. In addition, fatigue performance is highly dependent on rough surface and porosity, which means that the technique highly requires post-process (e.g. peening). However, this approach is not agile and smart enough to enhance part quality. Ideally, the printing parameters need to be quickly and dynamically tuned up by optimized input variables. Accordingly, research attempts have been put into practice to monitor process to control process input variables (Edwards et al., 2013). Melvin et al. (1994) used a video microscopy system to observe sintering and flow behavior in real-time, which can evaluate different sintering characteristics of the materials (Melvin et al., 1994). Benda (1994) developed an infrared light thermal sensor to control laser power for better uniform sintering performance (Benda, 1994). Zeng et al. (2012) reviewed literatures on the thermal modeling method in SLS and SLM techniques. It was summarized that uniform temperature distribution of fields during printing processes leads to better quality; there is a need to provide information from monitoring temperature of the melt pool to be able to control process parameters for part quality. As a temperature monitoring system, pyrometers and thermocouples were used for monitoring its temperature (Zeng et al., 2012).

Pyrometer is defined as the noncontact measurement of body temperature based on emitted thermal radiation. There are two types of pyrometer used depending on the application: photodiodes and digital cameras [e.g. charged-coupled device (CCD), complementary metal oxide semiconductor (CMOS)]. Both detects radiation and light, respectively, and convert them into electric signal. Bi et al. (2007) developed a novel laser cladding head with photodiodes and digital camera (CCD) to enable real time monitoring to monitor the status of machine parts, thus, improve part quality (Bi et al., 2007). Kleszczynski et al. (2012) used a high resolution CCD camera for surface error detection through image processing, as shown in Figure 12 (Kleszczynski et al., 2012). Lott et al. (2011) analyzed images of melt pool size to adjust laser output power by using a CMOS camera with an additional illumination source required for high scanning velocities, resolution, as well as photodiodes (Lott et al., 2011). Kolossov et al. (2004) developed a thermal model that the temperature evolution and sintering formation can be simulated by a 3D FEA to predict thermal properties (i.e. thermal conductivity and specific heat; Kolossov et al., 2004).

Thermocouples, in contrast with pyrometers, are defined as contact measurement of temperature which can reduce the freedom of process. Therefore, pyrometer has been used dominantly more than thermocouples for temperature monitoring in PBF processes. Shishkovsky et al. (2008) used thermocouples to monitor the build of six different intermetallic powders and measure the temperature on the powder bed (Shishkovsky et al., 2008). In addition, Van Belle et al. (2013) used thermocouple attached to the bottom of the base plate with a strain gauge to record residual stresses (Van Belle et al., 2013), and Taylor and Childs (2001) also used thermocouples positioned under the bed’s surface to monitor energy absorption and powder effective conductivity for better understanding of heat transfer in metal powder during laser processing of the powder bed (Taylor and Childs, 2001). Low and Ake (2004) invented a thermocouple control system to improve uniform distribution of temperature on powder bed during part build. The thermocouple was attached inside of the powder bed along with IR sensor and communicates with temperature transmitter through circuitry in real-time (Low and Ake, 2004).

Besides, research was focused on the IR thermography-based monitoring and control system incorporated in the EBM system. Price et al. (2012) demonstrated the feasibility of using a near IR thermal camera for temperature measurement in hatch melting, preheating and contour melting events during the EBS process (Price et al., 2012). Yadroitsev et al. (2014) used a galvanometric scanner system for temperature distribution monitoring of melt pool of Ti6Al4V alloy in the SLM system (Yadroitsev et al., 2014). Dinwiddie et al. (2013) developed continuous data capturing method using the IR camera to demonstrate feasible work to detect porosities inside materials and understand thermal phenomena such as it happens when beam and powder interact with each other (Dinwiddie et al., 2013). Kruth et al. (2007) developed temperature feedback control system by positioning photodiodes and CMOS digital camera on laser beam to stabilize melt pool temperature distribution in SLM system (Kruth et al., 2007). In 2013, Mireles developed automatic feedback control system using IR camera for defect detection, and notification tool to stop printing process when porosity level reaches to certain level. In addition, achieved data can help process parameters to be optimized and stabilize temperature automatically through image processing and auto-decision-making (Mireles et al., 2013). In 2015, Mireles developed in-situ defect monitoring methods using IR thermography inside an EBM machine. This captured IR images provided a good indication of measured defects geometry presence, which can be used to re-melt defect (i.e. porosity) for the purpose of in-situ correction, as shown in Figure 13 (Mireles et al., 2015). This method saved total manufacturing time and decreased altering microstructure compared with HIPing (hot isostatic pressing) process, and eventually increased the mechanical properties throughout the fabricated part.

In summary, the PBF processes, as compared with other processes, have mostly advanced in temperature monitoring system using sensors. Currently, the main quality issues identified was the heterogeneous distribution of powder bed temperature and laser output power. Non-contact method, which is more adaptable owing to the freedom of process, was used to solve these quality issues. In addition, thermal modeling techniques such as 3D FEA were implemented to predict thermal properties. Post processes (e.g. heat treatment, peening, etc.) are necessary after the 3D printing process to alleviate residual stress, rough surface and porosity. Research efforts were made to develop feedback control system for temperature distribution, defect detection and in-situ correction.

### 2.6 Directed energy deposition processes

Directed energy deposition (DED) processes use a laser beam to fuse materials (e.g. metal powder and wire feedstock) delivered from a material deposition head and substrate where materials are deposited (Tapia and Elwany, 2014; Vaezi et al., 2013). PBF process feeds a powder layer using a blade or roller onto the powder bed and selectively melts sections to build 3D structures, while in DED metal powders are fed from the deposition head and melt onto a substrate by energy source coming out from middle of the head. There are different types of techniques such as laser engineering net shape (LENS), laser cladding (LC) and direct metal deposition (DMD). Major considerable processing parameters in DED are melt pool temperature, material delivery rate and nozzle tip-substrate distance.

The process monitoring and control focused on aforementioned major printing parameters have been of increasing interests to improve quality and reproducibility. Bi et al. (2013) developed closed-loop controller that sends monitored temperature feedback and adjust laser power density in real-time to improve dimensional accuracy by using the IR pyrometer that measures melt pool temperature. This closed-loop controller can eliminate the heat accumulation and achieve oxidation-free clad surface with feeding argon gas shielding to avoid disturbance from IR temperature signals, as shown in Figure 14 (Bi et al., 2013). Fathi et al. (2008) used similar closed-loop laser cladding process that can reduce stair–step effects as well as production time (Fathi et al., 2008).

In addition, Bi et al. (2006) investigated printing parameters relationships by looking at microstructure, hardness and dimensional accuracy. It was proved that variations of melt pool temperature and size influencing cooling rate and solidification conditions result in heterogeneous dimensional accuracy and microstructure (Bi et al., 2006). In a similar manner, Hu and Kovacevic (2003) implemented a system to monitor and control the powder delivery rate through a real-time sensing and control using IR high-speed camera positioned with laser beam. Captured images of melt pool by this camera were also used for analysis to adjust laser power for uniform temperature distribution. The reflected data from the sensor helped improve the 3D printed part quality (Hu and Kovacevic, 2003). Similarly, Hofmeister et al. (1999) used a CCD camera mounted with laser beams to capture melt pool images and analyze the melt pool cooling rate. Some researchers studied process monitoring with DOE method (Hofmeister et al., 1999). Ermurat et al. (2013) and Lee (2008) used a high-speed camera and CCD camera to monitor nozzle and powder delivery rate. With its extracted data, DOE is used to analyze the behavior of the nozzle dimension and powder size along with changes of standoff distance and gas flow (Ermurat et al., 2013; Lee, 2008).

Some researchers adopted analytical models to analyze the thermal behavior of the process. Ye et al. (2006) and Wang et al. (2009) reported that DED fabrication requires an in-depth understanding of the entire thermal behavior of the process. In these papers, the finite element method (FEM) was used to obtain the thermal distribution and numerical simulation, known as Thermaviz. Two-wavelength imaging pyrometer was used to provide temperature data such as temperature mapping, cooling and heating rate, modeling and simulation, etc. The numerical simulation and FEM results provide the user further predicting models and parameters (Ye et al., 2006; Wang et al., 2009). Alimardani et al. (2007) also used a 3D numerical method for predicting transient geometrical and thermal characteristics of multilayer laser solid freeform fabrication. Thermal and geometrical properties can be numerically obtained and predicted even with variation of parameters such as the powder feed rate, elapsed time, melt pool temperature, substrate roughness, etc. so that the quality can be more precisely optimized (Alimardani et al., 2007). (Figure 15).

For the studies on material delivery rate, Hu et al. (2011) developed a monitoring and control system using IR diodes and low-power laser. This system is designed for sensing the powder flow rate out of nozzle to ensure consistent powder feed rate (Hu et al., 2011). In addition, Smurov et al. (2012) used high-speed cameras for calculating the powder speed and flux by positioning the cameras next to the nozzle head (Smurov et al., 2012).

Some of the researchers conducted studies on the monitoring layer height through the use of CCD cameras and neural network algorithm. Iravani-Tabrizipour and Toyserkani (2007) used three CCD cameras-based optical detectors positioned at 120 degree of each other around the nozzle head used in LC technique. A novel algorithm composed of an image-based tracking protocol and a recurrent neural network were developed to measure the clad height in real-time. These features tracking and trained network showed a promising results in detecting the clad height in various trajectory angles in real-time, which measured the clad height with 10 Hz and ±0.15 mm average precision independent from clad path with about 12 per cent maximum error (Iravani-Tabrizipour and Toyserkani, 2007). In a different approach, Toyserkani and Khajepour (2006) used a CCD camera-based optical detector in close-loop control to be able to effectively measure not only clad height but also angle of liquid/solid interface in real-time. Using the recorded different images of each layer, it estimates its roughness and product quality (Toyserkani and Khajepour, 2006).

In summary, DED processes have dealt with advanced researches such as process monitoring, real-time closed loop feedback system, thermal behavior and data analytic algorithm. The main quality issues identified were related to heterogeneous melt pool temperature, laser power density and material delivery rate. These heterogeneities can be solved by a closed-loop controller with data acquisition equipment such as CCD and IR high speed cameras, which can sense printing process data in real time and adjust them to optimal values. Researches such as FEM simulation and data analytic algorithm were also executed to predict the quality.

### 2.7 Sheet lamination processes

Sheet lamination processes (SLP) are one of the additive manufacturing techniques to build a 3D structure by sequentially laminating, bonding and cutting 2D cross-section sheets of such polymer, ceramic coated paper or composite paper by either a laser or a knife, and layers bonded by glue or adhesive coating. This technique is also called laminated object manufacturing (LOM). The bonding process is accomplished by heat or pressure applied from a heated cylinder rolling along the sheet. While there is another technique called ultrasonic consolidation (UC), which has the capability of ultrasonic welding of metal sheets, it will be not covered in this paper as this technique is a hybrid fabrication method combining the additive manufacturing with subtractive process (Vaezi et al., 2013; Kechagias, 2007a).

Although LOM process provides faster processes with sufficient quality characteristics on large physical prototypes than other techniques (Kruth, 1991), this technique has many difficulties and drawbacks in manufacturing process, which are poor surface finish before post-processing, difficult disengagement and only good tensile strength properties in laminates direction compared with other techniques (Chryssolouris et al., 2003; Upcraft and Fletcher, 2003).

The LOM process is separated into two methods when it comes to the way of laminating a sheet of material: cut-then-bond and bond-then-cut. The cut-then-bond method, which cuts first and then re-position on workpiece to bond, is effective for de-cubing, however, causes a loss in positioning precision and accumulated error. The bond-then-cut, which bonds a sheet on workpiece and then cuts it by laser, is not able to perform complicated parts (e.g. hollow and vase) owing to difficulty in de-cubing inner waste material so that it consumes high laser power for de-cubing and crosshatching. In addition, it is time-consuming and labor-dependent.

Liao et al. (2003) proposed an online de-cubing process in the bond-then-cut method and with a developed bridge-supporting algorithm. Shielding paper laminated on a sheet was torn for laser cutting a sheet material and then used to remove waste materials online by using adhesion force between shielding paper and waste material, as shown in Figure 16 (Liao et al., 2003). For easy de-cubing process, an algorithm for building bridges from 2D drawing was developed to support and position islands part while de-cubing process (Chiu et al., 2003). This proposed process enabled the fabricating process to perform various geometric shapes with significantly reduced processing time up to 20-50 per cent. Chiu and Liao (2003) proposed a novel laser path planning strategy to improve de-cubing process problem. Burn-out rule proposed in this study was applied in laser path planning to reduce damaging part and bonding strength between the part and waste, eventually saving 68 per cent of de-cubing time. Overlap zone between fabricated part and waste is determined for the burn-out rule with calculated optimal laser energy levels, velocity and burn-out distance to predict better de-cubing process. It is suggested that optimal laser paths can be obtained by ordinary laser power with the burn-out rule and low cutting velocity (Chiu and Liao, 2003).

The most important quality characteristic of LOM part is the vertical surface roughness (Ra). Kechagias (2007a, 2007b) worked on experimental investigation of the surface roughness on LOM part. Improving vertical surface quality leads to reducing post-processing time, easier disengagement, easier de-cubing, process optimization and less finishing (Kechagias, 2007a). Using statistical models, Chryssolouris et al. (1999) found that vertical surface roughness (Ra) is mainly influenced by process parameters such as heater temperature, layer thickness and laser speed when measured on ZX plane, whereas pressure and heater speed have more effects on surface roughness when on ZY plane according to ANOVA analysis (Chryssolouris et al., 1999).

Another critical issue is a delamination between bonded layers owing to weak bonding or process malfunctioning (Vaezi et al., 2013; Kechagias and Iakovakis, 2009). Some researchers proposed mathematical models to enhance the bonding of laminates. Pak (1994) developed a mathematical equation containing a relation between temperature at arbitrary depth of the part (Z-axis) and critical parameters in terms of bonding of lamination such as temperature (Th), speed (ν), and pressure (δ) of heater roller. Reeves and Cobb (1995) presented an analytical model to predict optimal heater temperature needed for good laminate bonding. Sonmez and Hahn (1998) analyzed heat transfer and stress using FEA simulation and found the following:

• surface temperature of heater should not be higher than glass transition temperature of the used paper;

• surface stresses of laminate increase with the decrease in layer thickness or roller diameter; and

• higher rolling speed is more likely to cause delamination or malfunctioning on sheet bonding (Sonmez and Hahn, 1998).

Flach et al. (1998) developed a thermal analytical model with various process parameters for fabrication of functional ceramic and composite parts and founded the following:

• chamber air temperature affects temperature of part being built;

• roller speed significantly affects surface layer in direct contact; and

• delay time of the process affects the temperature of the part being built (Flach et al., 1998).

Some researchers studied on predicting surface roughness. Reeves and Cobb (1995) developed a general model for predicting roughness on sloped surface. The surface angle (θ) and the layer thickness (Lt) are critical parameters in predicting sloped surface roughness. Paul (2001) reported that the orientation angle and sheet thickness are statistically significant effects on surface roughness and proposed a mechanistic model for sloped surfaces considering the two critical parameters (Reeves and Cobb, 1995). Chryssolouris et al. (1999) developed a semi-empirical model in consideration with parameters such as changes of layer thickness, heater speed, heater temperature and platform retraction, based on matrix experiments with different surface angles and cross of laminates direction (Chryssolouris et al., 1999). Ahn et al. (2012) proposed theoretical expressions to quantify average surface roughness according to surface angle variation. In case of upward or downward facing surface (0° < θ < 90°, 90° < θ < 180°), theoretical model was expressed by  R(a)θ=1lA, where R(a) denoted the surface roughness, l the base length of unit profile, which is distance between edge point of sheets and A the total surface area; therefore, R(a) is affected by the base length. On the other hand, in case of vertical surface, roughness layer thickness (t) affected R(a) expressed by R(a)90=1tA. These expressions were verified by comparing with empirical data (Ahn et al., 2012).

Kechagias (2007a, 2007b) used a design of experiment to investigate dimensional accuracy of the LOM part. It was found that parameters such as heater, speed, temperature and nominal layer thickness mainly affect dimensional accuracy in X and Y directions (Kechagias, 2007b). In 2008, he proposed a simple model to predict and optimize LOM process performance using a feed-forward back propagation neural network (FFBP–NN). It was concluded that this model is a quick prediction of each performance measurements and can be used to select optimal process parameters values (Kechagias and Iakovakis, 2009). Vaezi et al. (2013) reported that quality of the LOM also requires well-prepared material sheets to reduce variations in layer thickness and materials content. Depth of cuts varied with variations of layer thicknesses and material content so that these variations lead to more damage to a bonded layer below. In addition, the LOM part can be post-processed to reaction bonding process during a pyrolysis cycle to enhance surface quality and laminates bonding (Vaezi et al., 2013). Chryssolouris et al. (2003) investigated the influence of different process parameters of LOM on the tensile strength of LOM part. It was found that the most significant process parameter affecting the tensile strength of part was layer thickness. Therefore, it can be concluded that layer thickness can provide a prediction on tensile strength of part (Chryssolouris et al., 2003).

In summary, there have been issues on de-cubing process, vertical surface roughness and delamination. Development of online de-cubing process and bridge-supporting algorithm has solved several issues such as saving energy, post-process time and complicated structure arose by de-cubing. Research in mathematical models and neural network were implemented to enhance bonding of laminates and predict optimal process parameters. Vertical surface roughness and delamination are the future challenges to be solved to improve quality. Seven techniques in AM technology was summarized and compared in quality point of view, as shown in Table I.

## 3. Classification of research findings

Researches of QC in AM are being rapidly investigated recently owing to high feasibility, reliability and growth of mechanical and material aspects applied to various areas. The first consideration on AM technology when it comes to QC is that several steps are classified to achieve design requirement, as described by Hollister et al. (2015) who studied a splint used in medical area. This paper designed five steps of the design control process, as shown in Figure 17. The first step is a design input considering all parameters on designing and modeling a part to meet mechanical and specific requirements such as laser power, speed and particle size. Design output as the second step comprises tests performed to measure the design input based on the standard operating procedures by determining the compressive, bending and opening stiffness of the splint by using the method of destructive evaluation. The first two steps are to test sample parts to generate the design variables for optimized final part being considered in design implementation as third step. Design verification as fourth step is to ensure the final manufactured part meets design input via geometric and mechanical evaluation (e.g. calipers, Micro-CT, etc.). The last design validation is to test pre-clinical model and human clinical implantations of final part.

This outline suggested as a good framework for providing steps of QC in medical-related research. However, there are drawbacks when figuring out the design variables. It is time-consuming and wastes materials in making samples for mechanical test and generating optimal design variables. If the design variables are adjustable during the design and printing process, the entire process of QC can be more agile and reliable. Figure 18 shows proposed classification of four main processes to take account of QC during AM cycle and describes detail works recently studied and also included in Section 2.

On the basis of the four steps, modeling, printing, post-treatment, and part evaluation during the entire 3D printing process, the following four categories are described briefly and issues are discussed.

### 3.1 Design process

This is so-called pre-QC considering from raw material’s quality to model design. Material selection is important because it may be contaminated or may have air porosities, which affects directly to the part property. Good design ability by setting up optimized values of design parameters (e.g. layer thickness, infill pattern, printing speed and delay, mesh, etc.) is directly correlated to part’s good quality. Recently, Hollister et al. (2015) studied on designing splint structure meeting quantitative clinical requirement by finding optimized design variables and printing parameters generated from mechanical tests. This approach finally brings the optimized splint part individually customized for patients who needs perfect clinical recovery (Hollister et al., 2015). However, it requires a long time-consuming procedure to fabricate and iterations of mechanical testing depending on patients to generate optimal design variables. Recently, one research (Sitthi-Amorn et al., 2015) investigated on developing real-time printing control technique to correct design and printing factors in real-time through closed-feedback loop. This technique analyzes each printed layer in real-time through the machine vision, computing depth difference between the layers, and sends feedback to central computer to control amount of jet droplet injection. This technique opens up the real-time design process during printing process, but is limited to MJP. Some researchers have studied optimizing design parameters by developing analytical modeling and data-mining based on quality prediction algorithms as mentioned in Section 2. However, this research is limited to BJP, FDM and SLP techniques. There are needs to launch this research in other printing techniques.

### 3.2 Printing process

This is the so-called in process-QC considering from printing setup to process finish of part. Agarwala et al. (1996) reported that internal and surface defects of fabricated wax pattern could be eliminated by the optimization of FDM parameters (Agarwala et al., 1996). Likewise, optimized set-up of printing parameters (e.g. extruder temperature, calibration, pattern, feed and flow rate, etc.) is critical for the quality. In recently, researchers have studied the optimization of printing parameters through real-time monitoring and printing control as mentioned in Section 2. To date, more research has been conducted in terms of the printing process compared to other processes. Farzadi et al. (2015) studied the effects of the layer printing delay on 3D printed part’s physical and mechanical properties for better selections of printing parameters and the best printing conditions (Farzadi et al., 2015). In addition, Kaveh et al. (2015) designed proper benchmarks (e.g. extruded temperature, etc.) and statistical equations calculating calibration factors (precision and internal cavity) used to determine or to optimize the value of printing parameters (Kaveh et al., 2015). Similarly, Hollister et al. (2015) created standard test coupons with consistent geometry and known mechanical properties included in every build providing a standard with which to compare builds at different performing times because printing parameter should be changed according to custom devices for different patients (Hollister et al., 2015). These studies aim to optimize printing process for quality along with different printing parameters in different 3D printing techniques such as PBF, EBP, BJP and MJP techniques. As mentioned above, a research by Sitthi-Amorn et al. (2015) and Mireles et al. (2015) studied real-time optimization of printing parameters by closed-feedback loop control in MJP (Sitthi-Amorn et al., 2015) and in-situ defects correction in EBM technique (Mireles et al., 2015). These techniques might not be limited to other 3D printing techniques when it comes to its application. However, PVP and BJP techniques could not be applied owing to the inaccessibility of the machine vision system and IR thermography. Therefore, there is a need for alternatives for real-time monitoring, closed-feedback loop printing control, and in-situ defect correction processes to be applied in these techniques.

### 3.3 Post process

This is so-called post-QC considering finished part’s quality after printing process and before part evaluation. Some of the techniques such as PBF heavily rely on this process for improving better surface and mechanical property of printed part through sintering and peening processes. Edwards et al. (2013) attempted to remove post process on EBM printed alloy parts by using the elevating bad temperature to minimize residual stresses (Edwards et al., 2013). In the EBP and SLP techniques, post processing was used in chemical, thermal and hot cutter processes for surface quality improvement. However, this research is limited to certain materials such as ABS polymer. Therefore, there is a need to launch research on other materials. In the MJP technique, sintering was adopted as post process for 3D printed Zr-based ceramic dental prostheses.

### 3.4 Evaluation process

This process is to evaluate part’s reliability. This includes all activities measuring external and internal characteristics of printed parts whether it meets specific design and mechanical requirements. The external characteristics are measured by such a calipers, machine vision, roughness tester, etc., while the internal characteristics are measured by destructive or non-destructive evaluations. Some studies have been attempted on optimizing evaluation process. Muller and De Jean (2015) developed a novel stereo microscope adapter, called SweptVue, for cost-effective microfabrication quality control, as shown in Figure 19. Using this optical technique, surface elevation and accuracy were examined over multiple regions on plateaus and hemispherical surfaces, and building errors such as decreased heights were discovered when approaching the edges of plateaus, inaccurate height steps and poor tolerances on channel width (Muller and De Jean, 2015). In addition to external evaluation, Hollister et al. (2015) used a Micro-CT method, which non-destructively builds an image of internal structures allowing quantitative and qualitative measures (Hollister et al., 2015). These two approaches used traditional evaluation methods (e.g. calipers and SweptVue) and high tech evaluation methods (e.g. Micro-CT) for the evaluation of geometry accuracy and internal structure characteristics. However, the latter evaluation method requires high cost and big volumetric facility.

## 4. Challenges and future trends

As a consequence of what has been discussed above, it is evident that vast researches and variety approaches have been studied to improve the quality of AM processes. However, it has been seen that there are still many issues that occur from research reviews. To achieve agile and streamlined AM processes, from design to part evaluation and enhanced quality AM product, novel AM processes for QC could have the potential to address these issues:

• Prediction of optimal printing parameter: Iterations works by external and internal evaluation to find optimal printing parameters; these evaluations are tedious and not agile. If it can predict printing parameter based on data analytic analyses (e.g. data mining) at design process, final part can be expected with better dimensional accuracy and surface quality.

• Prediction of mechanical property: As mentioned above, iterative mechanical testing to meet the requirements eventually will waste materials. If a designing tool (software) before printing process has simulation capability or mechanical prediction module, better design process can be performed by users based on the predicted data.

• Robust real-time monitoring and process control in build process: In MJP, PBF and DED, real-time monitoring and close-feedback loop control and defect correction during build process has been recently developed. These techniques are advanced technology that integrates machine vision, data collection, image processing and closed-feedback loop (Sitthi-Amorn et al., 2015). In PVP, real-time monitoring systems have been used to monitor temperature, layer height, etc. for data collection but image processing (computation) and closed-feedback loop control. If these can be achieved in other processes, they can provide better quality and agile AM process.

• Feedback interaction between design or printing process and part evaluation: Closed-feedback loop recently developed for correcting design and printing factors in real-time is a good example of intercommunication between design and printing processes. There is another research that compares CT scanned images with initial design model to calculate dimensional discrepancy. If there is a module or software that enables feedback of part evaluation (e.g. inner defect, dimensional inaccuracy and surface crack) to be considered when redesigning to remove these discrepancies, high quality of product can be expected in shorter time. For example, on an unresolvable part defect in spite of several printing efforts, design software can adjust design or printing parameter (e.g. printing speed, extrusion rate, etc.) using the feedback data generated from part evaluation; therefore, defects can be prevented before printing process.

• Agile part evaluation: Part evaluation is also time-consuming like the printing process and waste lots of materials owing to iterative destructive evaluation. Non-destructive evaluation (NDE; offline inspection) is needed to enable validation of process performance. However, it is expensive and requires skills to operate the machine. If machine vision or NDE technique can be imbedded into the 3D printer for external and internal quality evaluation as online inspection, part evaluation process can be agile and finished in 3D printers without separated evaluation process. For example, 3D scanners that are commercially available are imbedded into 3D printers can be used for surface quality evaluation through image processing right after printing process.

• High speed fabrication and scale: Future technology will require more rapid and bigger scale customized production skills. However, in the current technology, the faster printing process is, the worse its quality will be. Fabrication speed is significantly correlated to part quality in all techniques and dependent on part price. Increasing this speed with an increase in quality will be the challenge in the future of AM.

• Cyber quality control: AM technology can be incorporated with remote connection for quality monitoring and control. If all the processed data attained from the sensor and printing process are displayable and controllable under remote control, massive and customized AM productions would be feasible in less time and cheaper cost of production in the near future.

## 5. Conclusion

AM technology has reached a crossroad right before being used as end-part production. Many techniques have been developed up to now to respond to the demand of high-quality 3D printed parts but more studies are still required to improve their systems and quality. This paper presented a review on quality control in seven different techniques in the AM technology, and provided detail discussions in each quality process stages. Most of the AM techniques follow the trend of using in-situ sensors and cameras to acquire processing data for real-time monitoring and quality analysis. This effort opens up the in-situ correction by the closed-feedback loop and image processing. However, owing to the closed geometry of the 3D printing system, some techniques such as BJP are not eligible to adopt this technique. More sophisticated process monitoring and agile feedback control are required to enable end-product to be better quality and faster process. Processes such as EBP are more advanced regarding data analytics and predictive algorithms-based research regarding mechanical properties and optimal printing parameters. This data analytics research is critical for optimization and prediction of design and printing process. This research effort would lead to artificial control system that enables to produce perfect quality of AM product. MJP technique has advanced with systems that are integrated with closed-feedback loop, machine vision and image processing to minimize quality issues during printing process. It is expected that future challenges would use these predictive algorithms based research and integrated system in the near future for proved QC-based agile and accurate printing process.

## Figures

#### Figure 1

Schematic diagram of nano-stereolithography apparatus (left) and SEM images of fabricated open structure (right)

#### Figure 2

Water droplet stream using orifice diameter of 150 µm at a gas pressure of 8 kPa

#### Figure 3

(A) Optical coherence tomography 3D scanner: the optical components are arranged in a Michelson configuration. The light source is a collimated, red LED. Polarizers are placed in the LED and camera paths to control the light density. The points in the sample that are at the same distance as the reference mirror cause constructive interference patterns. The platform is moved along the Z axis while the interference patterns are detected by the camera. The scanner features a circular scanning area with a diameter of 15 mm; (B) the machine vision feedback loop proceeds in the stages and image processing pipeline for 3D scanner: We compute the 3 × 3 standard deviation. Next, we choose the depth in which the standard deviation is the highest. We also keep the maximum standard deviation as the confidence map. We perform the depth correction on the depth map from previous stage. Finally, we stitch the depth maps from all the location weighted by the confidence. We assume that the area where the confidence is below certain threshold is a hole, and fill those holes at the end

#### Figure 4

(A) Components of the continuous printing system; (B) flight path of a droplet during printing; (C) binary deflection and proportional one

#### Figure 5

Microphotograph of the cross-sectional area of a 38 × 38 × 30 layer part

#### Figure 6

(A) Modeled (left) and fabricated (right) air fan blade; (B) 3D maps of the thickness deviation measurements for original (left) and modified (right) blade

#### Figure 7

(A) Graphical simulation of the stair-stepping effect on textured surfaces and (B) resolution limits (right): (a) spacing of layer contours; (b), (c) feature sinking

#### Figure 8

(A) Permeability of micro-channels fabricated using FDM before and after treatment; (B and C) FDM ABS feature fidelity before and after treatment; (B) channel and (C) ridge shape fidelity; (D), percentage change in ridge radius of curvature (E) and channel wall incline angle with varying acetone solution concentration, treatment length constant at 4h

#### Figure 9

(A) Different directions of machining using hot cutter; (B) scheme of barrel finishing; (C) 3D map of surface before machining and after 960 min barrel finishing working time

#### Figure 10

(A) Model development of different complex geometries using CLFDM – (a) a hemispherical surface with a straight trim on one side (b) trimmed surface of a part in the interior, discontinuous filament deposition method (c) trimmed surface of a part in the interior, continuous filament deposition method (d) surface made of two patches, continuous filament method; (B) (a) sectional view of a part produced by FDM. The overlapped section of two adjacent layers in FDM is shown by shaded lines. (b) Adjacent filaments of CLFDM

#### Figure 11

The ANN architecture

#### Figure 12

Setup of the CCD camera system in front of machine window (left) and setup of CMOS camera system with illumination source and photodiode for high scanning velocities and resolution (right)

#### Figure 13

Correction of un-melted powder through layer re-melt shown by the build file with the re-melt area shown by the red arrow (top), showing the defect present in the IR image by the black arrow (middle) and showing the defect not shown on the left column owing to the re-melt (bottom)

#### Figure 14

Front view of the deposited sample, the laser power and IR-temperature signals after (left) and before (right) stabilizing IR–temperature signals

#### Figure 15

Finite element mesh and geometry to simulate the LENS process for ten layer wall (left) and molten pool size distribution for each layer at P = 600 W and V = 2.5 mm/s (right)

#### Figure 16

(A) The mechanism of the proposed online de-cubing process; (B) the original sliced 2D contour drawing; (C) the altered 2D drawing developed by bridge-supporting algorithm

#### Figure 17

Schematic outline of the splint design control process by reference

#### Figure 18

The proposed classification of the four main processes during AM cycle

#### Figure 19

(A) A pair of hard-baked photoresist pads used for a microfluidic device, seen in monochrome when imaged through a traditional microscope (left). The SweptVue adapter measures the relative depth of the features in the image and colorizes them in software using a jet color map, with feature heights increasing blue to red (right); (B) depth maps can be exported to MATLAB software for further image processing and 3D plotting; (C) the SweptVue stereo microscope adapter, shown mounted to an Olympus SZX7 stereo microscope

## Table I

Summary and comparison of seven AM techniques in quality point of view

Photopolymer vat processes (PVP) Resin viscosity, density and temperature
Build orientation
Laser scan speed
Exposure energy
Penetration depth
Excellent dimensional accuracy
Good surface finish
Biocompatible process
Microscale fabrication
(∼10 µm)
Only UV curable polymer resin
Material contamination and wastes
Mechanical property limited by resin-based polymer
Thermal shrinkage
Projection SL technique
Accurate custom-design and evaluation processes through MRI and CT
Data analytic and algorithm
Biomaterial fabrication
Nano/micro scale fabrication
Development of variety vat solution
Expensive scanning equipment such as MRI and CT
Resolution
Material jetting processes (MJP) Droplet shape, velocity, diameter, viscosity and surface tension
Jetting frequency, signal width and voltage magnitude
Substrate temperature and evaporation rate
Humidity
Excellent resolution
Fast process
Excellent dimensional accuracy
Multiple material printing capability
Microscale fabrication
(∼100 µm)
Nozzle clogging
Cell damage in certain temperature or frequency
Nano particles agglomerates causes clogging
Low viscosity
DoD technique
Micro-scale fabrication
Biomaterial fabrication
Real-time close-feedback loop for printing correction through image processing on droplet and depth map
Quality prediction
Exposure energy
Nozzle clogging
More fabrication freedom on biomaterials
Binder jetting processes (BJP) Powder surface treatment, size, shape, packing density and distribution
Binder viscosity, surface tension, droplet size, velocity and temperature
Printing layer thickness, orientation, binder saturation and delay time
Excellent quality control ability for ceramic and metal fabrication compared to MJP technique
DoD
Poorer accuracy and surface finish compared with MJP technique
Infiltration process needed for post-treatment
DoD technique
Ceramic and metal fabrication
Interaction performance between binder and powder
Accumulative accuracy of deposited layer thickness
Droplet placement
Delay time of applying new layer
Dimensional change during post-process
Extrusion-based processes (EBP) Layer thickness, build orientation, raster width and angle and extrusion temperature
Post-processing (exposure time and chemical temperature)
Fast process
Good material properties
Fair resolution
Staircase effect
Thermal distortion
Thermal shrinkage
Low dimensional accuracy
Delamination
FDM technique
Data analytic and algorithm
Post-processing
Optimal build orientation
Resolution
Dimensional accuracy
Shape change
Clogging issues on composites materials
Powder bed fusion processes (PBF) Powder bed temperature
Laser/beam output powder
Powder size
Atmosphere
Wide range of materials
Excellent dimensional accuracy
Excellent repeatability
Good material property
Fair resolution limited by particle size
Residual stress
Porosity
SLS technique
Temperature monitoring system
Thermal modeling and image processing
Temperature feedback control system
Real-time defect detection and in-situ correction
Resolution
Laser output feedback control system in real time
Atmosphere control
Directed energy deposition processes (DED) Melt pool temperature
Material delivery rate
Distance between nozzle tip and substrate
Laser power density
Wide range of materials
Good material property
Fair resolution
Thermal stress
Closed-loop controller system for real-time feedback of temperature, laser output, clad height and delivery rate Resolution
Surface post-process
Atmosphere control
Sheet lamination processes (SLP) Heater temperature
Layer thickness
Laser speed/power
Rolling speed/pressure
Chamber air temperature
Delay time
Orientation
Fast process
Sufficient quality on large prototype
Good tensile strength in laminated direction
Poor resolution
De-cubing or crosshatching process
Delamination
Shrinkage
Poor tensile strength in Z direction
Waste
LOM technique
Burn-out rule
Data analytic and algorithm
Online de-cubing process
Resolution
Consistent sheet thickness
Repositioning precision and accumulation error (cut-then-bond process)
Vertical surface roughness
Material damage by laser cut
Tensile strength in Z direction
Delamination

## References

Addanki Sambasiva Rao, M.A.D., Venkatesh, J.V.L. and Ojha, D. (2012), “Investigation post processing techniques to reduce the surface roughness of fused deposition modeled parts”, International Journal of Mechanical Engineering & Technology (IJMET), Vol. 3, pp. 531-544.

Adolf, D.B., Martin, J.E., Chambers, R.S., Burchett, S.N. and Guess, T.R. (1998), “Stresses during thermoset cure”, Journal of Materials Research, Vol. 13, pp. 530-550.

Agarwala, M.K., Jamalabad, V.R., Langrana, N.A., Safari, A., Whalen, P.J. and Danforth, S.C. (1996), “Structural quality of parts processed by fused deposition”, Rapid Prototyping Journal, Vol. 2, pp. 4-19.

Ahn, D., Kweon, J.-H., Choi, J. and Lee, S. (2012), “Quantification of surface roughness of parts processed by laminated object manufacturing”, Journal of Materials Processing Technology, Vol. 212, pp. 339-346.

Ahn, D., Kweon, J.-H., Kwon, S., Song, J. and Lee, S. (2009), “Representation of surface roughness in fused deposition modeling”, Journal of Materials Processing Technology, Vol. 209, pp. 5593-5600.

Ahn, S.-H., Montero, M., Odell, D., Roundy, S. and Wright, P.K. (2002), “Anisotropic material properties of fused deposition modeling ABS”, Rapid Prototyping Journal, Vol. 8, pp. 248-257.

Alimardani, M., Toyserkani, E. and Huissoon, J.P. (2007), “A 3D dynamic numerical approach for temperature and thermal stress distributions in multilayer laser solid freeform fabrication process”, Optics and Lasers in Engineering, Vol. 45, pp. 1115-1130.

Anitha, R., Arunachalam, S. and Radhakrishnan, P. (2001), “Critical parameters influencing the quality of prototypes in fused deposition modelling”, Journal of Materials Processing Technology, Vol. 118, pp. 385-388.

Armillotta, A. (2006), “Assessment of surface quality on textured FDM prototypes”, Rapid Prototyping Journal, Vol. 12, pp. 35-41.

Asadi-Eydivand, M., Solati-Hashjin, M., Farzad, A. and Osman, N.A.A. (2016), “Effect of technical parameters on porous structure and strength of 3D printed calcium sulfate prototypes”, Robotics and Computer-Integrated Manufacturing, Vol. 37, pp. 57-67.

Beaman, J., Atwood, C., Bergman, T., Bourell, D., Hollister, S. and Rosen, D. (2004), “Assessment of European research and development in additive”, Subtractive Manufacturing, Final Report from WTEC Panel, available at: http://wtec.org/additive/report/welcome.htm

Bellini, A. and Güçeri, S. (2003), “Mechanical characterization of parts fabricated using fused deposition modeling”, Rapid Prototyping Journal, Vol. 9, pp. 252-264.

Benda, J. (1994), “Temperature controlled selective laser sintering”, Proceedings of the Solid Freeform Fabrication Symposium, DTIC Document, pp. 277-284.

Bi, G., Sun, C. and Gasser, A. (2013), “Study on influential factors for process monitoring and control in laser aided additive manufacturing”, Journal of Materials Processing Technology, Vol. 213, pp. 463-468.

Bi, G., Gasser, A., Wissenbach, K., Drenker, A. and Poprawe, R. (2006), “Characterization of the process control for the direct laser metallic powder deposition”, Surface and Coatings Technology, Vol. 201, pp. 2676-2683.

Bi, G., Schürmann, B., Gasser, A., Wissenbach, K. and Poprawe, R. (2007), “Development and qualification of a novel laser-cladding head with integrated sensors”, International Journal of Machine Tools and Manufacture, Vol. 47, pp. 555-561.

Boschetto, A. and Bottini, L. (2015), “Roughness prediction in coupled operations of fused deposition modeling and barrel finishing”, Journal of Materials Processing Technology, Vol. 219, pp. 181-192.

Boschetto, A. and Bottini, L. (2016), “Design for manufacturing of surfaces to improve accuracy in fused deposition modeling”, Robotics and Computer-Integrated Manufacturing, Vol. 37, pp. 103-114.

Bose, S., Vahabzadeh, S. and Bandyopadhyay, A. (2013), “Bone tissue engineering using 3D printing”, Materials Today, Vol. 16, pp. 496-504.

Budzik, G. (2007), “Possibilities of utilizing 3DP technology for foundry mould making”, Archives of Foundry Engineering, Vol. 7, pp. 65-68.

Campbell, R.I., Martorelli, M. and Lee, H.S. (2002), “Surface roughness visualisation for rapid prototyping models”, Computer-Aided Design, Vol. 34, pp. 717-725.

Çaydaş, U. and Hasçalik, A. (2008), “A study on surface roughness in abrasive waterjet machining process using artificial neural networks and regression analysis method”, Journal of Materials Processing Technology, Vol. 202, pp. 574-582.

Chakraborty, D., Reddy, B.A. and Choudhury, A.R. (2008), “Extruder path generation for curved layer fused deposition modeling”, Computer-Aided Design, Vol. 40, pp. 235-243.

Chiu, Y. and Liao, Y. (2003), “Laser path planning of burn-out rule for LOM process”, Rapid Prototyping Journal, Vol. 9, pp. 201-211.

Chiu, Y., Liao, Y. and Hou, C. (2003), “Automatic fabrication for bridged laminated object manufacturing (LOM) process”, Journal of Materials Processing Technology, Vol. 140, pp. 179-184.

Chryssolouris, G., Kechagias, J., Moustakas, P. and Koutras, E. (2003), “An experimental investigation of the tensile strength of parts produced by laminated object manufacturing (LOM) process”, CIRP Journal of Manufacturing Systems, Vol. 32, pp. 319-322.

Chryssolouris, G., Kechagias, J.D., Kotselis, J.L., Mourtzis, D.A. and Zannis, S.G. (1999), “Surface roughness modelling of the Helisys laminated object manufacturing (LOM) process”, 8th European Conference on Rapid Prototyping and Manufacturing, Nottingham, pp. 141-152.

Chumnanklang, R., Panyathanmaporn, T., Sitthiseripratip, K. and Suwanprateeb, J. (2007), “3D printing of hydroxyapatite: effect of binder concentration in pre-coated particle on part strength”, Materials Science and Engineering: C, Vol. 27, pp. 914-921.

Cima, M., Oliveira, M., Wang, H., Sachs, E. and Holman, R. (2001), “Slurry-based 3DP and fine ceramic components”, Proceedings of Solid Freeform Fabrication Symposium, Austin, TX, pp. 216-223.

Cohen, A., Laviv, A., Berman, P., Nashef, R. and Abu-Tair, J. (2009), “Mandibular reconstruction using stereolithographic 3-dimensional printing modeling technology”, Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology, and Endodontology, Vol. 108, pp. 661-666.

Cooke, M.N., Fisher, J.P., Dean, D., Rimnac, C. and Mikos, A.G. (2003), “Use of stereolithography to manufacture critical‐sized 3D biodegradable scaffolds for bone ingrowth”, Journal of Biomedical materials Research Part B: Applied Biomaterials, Vol. 64, pp. 65-69.

Corcione, C.E., Greco, A. and Maffezzoli, A. (2006), “Temperature evolution during stereolithography building with a commercial epoxy resin”, Polymer Engineering & Science, Vol. 46, pp. 493-502.

Cotteleer, M. and Joyce, J. (2014), “3D opportunity: additive manufacturing paths to performance, innovation, and growth”, Deloitte Review, Vol. 14.

Cui, X. and Boland, T. (2009), “Human microvasculature fabrication using thermal inkjet printing technology”, Biomaterials, Vol. 30, pp. 6221-6227.

Cui, X., Boland, T., D’lima, D. and Lotz, M.K. (2012), “Thermal inkjet printing in tissue engineering and regenerative medicine”, Recent Patents on Drug Delivery & Formulation, Vol. 6, pp. 149-155.

Cui, X., Dean, D., Ruggeri, Z.M. and Boland, T. (2010), “Cell damage evaluation of thermal inkjet printed Chinese hamster ovary cells”, Biotechnology and Bioengineering, Vol. 106, pp. 963-969.

Czyżewski, J., Burzyński, P., Gaweł, K. and Meisner, J. (2009), “Rapid prototyping of electrically conductive components using 3D printing technology”, Journal of Materials Processing Technology, Vol. 209, pp. 5281-5285.

Derby, B. (2010), “Inkjet printing of functional and structural materials: fluid property requirements, feature stability, and resolution”, Annual Review of Materials Research, Vol. 40, pp. 395-414.

Dhariwala, B., Hunt, E. and Boland, T. (2004), “Rapid prototyping of tissue-engineering constructs, using photopolymerizable hydrogels and stereolithography”, Tissue Engineering, Vol. 10, pp. 1316-1322.

Dinwiddie, R.B., Dehoff, R.R., Lloyd, P.D., Lowe, L.E. and Ulrich, J.B. (2013), “Thermographic in-situ process monitoring of the electron-beam melting technology used in additive manufacturing”, SPIE Defense, Security, and Sensing, International Society for Optics and Photonics, pp. 87050K-87050K-9.

Ebert, J., Özkol, E., Zeichner, A., Uibel, K., Weiss, Ö., Koops, U., Telle, R. and Fischer, H. (2009), “Direct inkjet printing of dental prostheses made of zirconia”, Journal of Dental Research, Vol. 88, pp. 673-676.

Edwards, P., O’conner, A. and Ramulu, M. (2013), “Electron beam additive manufacturing of titanium components: properties and performance”, Journal of Manufacturing Science and Engineering, Vol. 135, p. 061016.

Ermurat, M., Ali Arslan, M., Erzincanli, F. and Uzman, I. (2013), “Process parameters investigation of a laser-generated single clad for minimum size using design of experiments”, Rapid Prototyping Journal, Vol. 19, pp. 452-462.

Farzadi, A., Solati-Hashjin, M., Asadi-Eydivand, M. and Osman, N.A.A. (2014), “Effect of layer thickness and printing orientation on mechanical properties and dimensional accuracy of 3D printed porous samples for bone tissue engineering”, PloS One, Vol. 9 No. 9, p. e108252.

Farzadi, A., Waran, V., Solati-Hashjin, M., Rahman, Z.A.A., Asadi, M. and Osman, N.A.A. (2015), “Effect of layer printing delay on mechanical properties and dimensional accuracy of 3D printed porous prototypes in bone tissue engineering”, Ceramics International, Vol. 41 No. 7, pp. 8320-8330.

Fathi, A., Khajepour, A., Durali, M. and Toyserkani, E. (2008), “Geometry control of the deposited layer in a nonplanar laser cladding process using a variable structure controller”, Journal of Manufacturing Science and Engineering, Vol. 130, p. 031003.

Flach, L., Jacobs, M.A., Klosterman, D.A. and Chartoff, R. (1998), “Simulation of laminated object manufacturing (LOM) with variation of process parameters”, Proceedings of Solid Freeform Fabrication Symposium, pp. 407-416.

Frank, D. and Fadel, G. (1995), “Expert system-based selection of the preferred direction of build for rapid prototyping processes”, Journal of Intelligent Manufacturing, Vol. 6, pp. 339-345.

Galantucci, L., Lavecchia, F. and Percoco, G. (2009), “Experimental study aiming to enhance the surface finish of fused deposition modeled parts”, CIRP Annals-Manufacturing Technology, Vol. 58, pp. 189-192.

Galantucci, L., Lavecchia, F. and Percoco, G. (2010), “Quantitative analysis of a chemical treatment to reduce roughness of parts fabricated using fused deposition modeling”, CIRP Annals-Manufacturing Technology, Vol. 59, pp. 247-250.

Gauvin, R., Chen, Y.-C., Lee, J.W., Soman, P., Zorlutuna, P., Nichol, J.W., Bae, H., Chen, S. and Khademhosseini, A. (2012), “Microfabrication of complex porous tissue engineering scaffolds using 3D projection stereolithography”, Biomaterials, Vol. 33, pp. 3824-3834.

Gibson, I., Rosen, D. and Stucker, B. (2014), Additive Manufacturing Technologies: 3D Printing, Rapid Prototyping, and Direct Digital Manufacturing, Springer, Berlin.

Gopal, A.V. and Chakradhar, D. (2012), “Parametric optimization in electrochemical machining of EN-31 steel based on grey relation approach”, Applied Mechanics and Materials, Vols 110/116, pp. 1649-1656.

Gross, B.C., Erkal, J.L., Lockwood, S.Y., Chen, C. and Spence, D.M. (2014), “Evaluation of 3D printing and its potential impact on biotechnology and the chemical sciences”, Analytical Chemistry, Vol. 86, pp. 3240-3253.

Ha, C.W. and Yang, D.-Y. (2014), “Fabrication of micro open structure using 3D laser scanning method in nano-stereolithography”, International Conference on Manipulation, Manufacturing and Measurement on the Nanoscale (3M-NANO), IEEE, pp. 299-303.

Han, L.-H., Suri, S., Schmidt, C.E. and Chen, S. (2010), “Fabrication of three-dimensional scaffolds for heterogeneous tissue engineering”, Biomedical Microdevices, Vol. 12, pp. 721-725.

Henmi, C., Nakamura, M., Nishiyama, Y., Yamaguchi, K., Mochizuki, S., Takiura, K. and Nakagawa, H. (2007), “Development of an effective three dimensional fabrication technique using inkjet technology for tissue model samples”, AATEX, Vol. 14, pp. 689-692.

Hofmeister, W., Knorovsky, G.A. and Maccallum, D.O. (1999), Video Monitoring and Control of the LENS Process, Sandia National Labs, Albuquerque, NM, and Livermore, CA.

Hollister, S.J., Flanagan, C.L., Zopf, D.A., Morrison, R.J., Nasser, H., Patel, J.J., Ebramzadeh, E., Sangiorgio, S.N., Wheeler, M.B. and Green, G.E. (2015), “Design control for clinical translation of 3D printed modular scaffolds”, Annals of Biomedical Engineering, Vol. 43, pp. 774-786.

Hu, D. and Kovacevic, R. (2003), “Sensing, modeling and control for laser-based additive manufacturing”, International Journal of Machine Tools and Manufacture, Vol. 43, pp. 51-60.

Hu, X.D., Kong, F.Z. and Yao, J.H. (2011), “Development of monitoring and control system for laser remanufacturing”, Applied Mechanics and Materials, Vols. 44/47, pp. 81-85.

Huang, X., Ye, C., Mo, J. and Liu, H. (2009), “Slice data based support generation algorithm for fused deposition modeling”, Tsinghua Science & Technology, Vol. 14, pp. 223-228.

Huang, Y.-M. and Jiang, C.-P. (2003), “Numerical analysis of a mask type stereolithography, pprocess using a dynamic finite-element method”, The International Journal of Advanced Manufacturing Technology, Vol. 21, pp. 649-655.

Iravani-Tabrizipour, M. and Toyserkani, E. (2007), “An image-based feature tracking algorithm for real-time measurement of clad height”, Machine Vision and Applications, Vol. 18, pp. 343-354.

Jiang, X.-S., Qi, L.-H., Luo, J., Huang, H. and Zhou, J.-M. (2010), “Research on accurate droplet generation for micro-droplet deposition manufacture”, The International Journal of Advanced Manufacturing Technology, Vol. 49, pp. 535-541.

Kaveh, M., Badrossamay, M., Foroozmehr, E. and Etefagh, A.H. (2015), “Optimization of the printing parameters affecting dimensional accuracy and internal cavity for HIPS material used in fused deposition modeling processes”, Journal of Materials Processing Technology, Vol. 226, pp. 280-286.

Kechagias, J. and Iakovakis, V. (2009), “A neural network solution for LOM process performance”, The International Journal of Advanced Manufacturing Technology, Vol. 43, pp. 1214-1222.

Kechagias, J. (2007a), “An experimental investigation of the surface roughness of parts produced by LOM process”, Rapid Prototyping Journal, Vol. 13, pp. 17-22.

Kechagias, J. (2007b), “Investigation of LOM process quality using design of experiments approach”, Rapid Prototyping Journal, Vol. 13, pp. 316-323.

Kim, H., Choi, J.-W. and Wicker, R. (2010), “Scheduling and process planning for multiple material stereolithography”, Rapid Prototyping Journal, Vol. 16, pp. 232-240.

Kleszczynski, S., Zur Jacobsmühlen, J., Sehrt, J. and Witt, G. (2012), “Error detection in laser beam melting systems by high resolution imaging”, Proceedings of the Solid Freeform Fabrication Symposium.

Ko, S.H., Chung, J., Hotz, N., Nam, K.H. and Grigoropoulos, C.P. (2010), “Metal nanoparticle direct inkjet printing for low-temperature 3D micro metal structure fabrication”, Journal of Micromechanics and Microengineering, Vol. 20, p. 125010.

Kolossov, S., Boillat, E., Glardon, R., Fischer, P. and Locher, M. (2004), “3D FE simulation for temperature evolution in the selective laser sintering process”, International Journal of Machine Tools and Manufacture, Vol. 44, pp. 117-123.

Kruth, J.-P. (1991), “Material incress manufacturing by rapid prototyping techniques”, CIRP Annals-Manufacturing Technology, Vol. 40, pp. 603-614.

Kruth, J.-P., Duflou, J., Mercelis, P., van Vaerenbergh, J., Craeghs, T. and De Keuster, J. (2007), “On-line monitoring and process control in selective laser melting and laser cutting”, Proceedings of the 5th Lane Conference, Laser Assisted Net Shape Engineering, pp. 23-37.

Kulkarni, P. and Dutta, D. (1999), “Deposition strategies and resulting part stiffnesses in fused deposition modeling”, Journal of Manufacturing Science and Engineering, Vol. 121, pp. 93-103.

Kullmann, C., Schirmer, N.C., Lee, M.-T., Ko, S.H., Hotz, N., Grigoropoulos, C.P. and Poulikakos, D. (2012), “3D micro-structures by piezoelectric inkjet printing of gold nanofluids”, Journal of Micromechanics and Microengineering, Vol. 22, p. 055022.

Lan, P.-T., Chou, S.-Y., Chen, L.-L. and Gemmill, D. (1997), “Determining fabrication orientations for rapid prototyping with stereolithography apparatus”, Computer-Aided Design, Vol. 29, pp. 53-62.

Lee, B., Abdullah, J. and Khan, Z. (2005), “Optimization of rapid prototyping parameters for production of flexible ABS object”, Journal of Materials Processing Technology, Vol. 169, pp. 54-61.

Lee, H.-K. (2008), “Effects of the cladding parameters on the deposition efficiency in pulsed Nd: YAG laser cladding”, Journal of Materials Processing Technology, Vol. 202, pp. 321-327.

Lee, K.-W., Wang, S., Fox, B.C., Ritman, E.L., Yaszemski, M.J. and Lu, L. (2007), “Poly (propylene fumarate) bone tissue engineering scaffold fabrication using stereolithography: effects of resin formulations and laser parameters”, Biomacromolecules, Vol. 8 No. 4, pp. 1077-1084.

Li, A., Zhang, Z., Wang, D. and Yang, J. (2010), “Optimization method to fabrication orientation of parts in fused deposition modeling rapid prototyping”, 2010 International Conference on Mechanic Automation and Control Engineering (MACE), IEEE, pp. 416-419.

Liao, Y., Chiu, L. and Chiu, Y. (2003), “A new approach of online waste removal process for laminated object manufacturing (LOM)”, Journal of Materials Processing Technology, Vol. 140, pp. 136-140.

Liu, W., Li, Y., Liu, J., Niu, X., Wang, Y. and Li, D. (2013), “Application and performance of 3D printing in nanobiomaterials”, Journal of Nanomaterials, Vol. 2013 No. 2013, p. 13.

Lott, P., Schleifenbaum, H., Meiners, W., Wissenbach, K., Hinke, C. and Bültmann, J. (2011), “Design of an optical system for the in situ process monitoring of selective laser melting (SLM)”, Physics Procedia, Vol. 12, pp. 683-690.

Low, S.C. and Ake, B.E. (2004), “Thermocouple control system for selective laser sintering part bed temperature control”, Google Patents.

Lu, K. and Reynolds, W.T. (2008), “3DP process for fine mesh structure printing”, Powder Technology, Vol. 187, pp. 11-18.

McCullough, E.J. and Yadavalli, V.K. (2013), “Surface modification of fused deposition modeling ABS to enable rapid prototyping of biomedical microdevices”, Journal of Materials Processing Technology, Vol. 213, pp. 947-954.

Maruo, S. (2015), “Development of functional devices using micro/nano stereolithography and moulding techniques”, International Polymer Science and Technology, Vol. 42, p. T7.

Mei, J., Lovell, M.R. and Mickle, M.H. (2005), “Formulation and processing of novel conductive solution inks in continuous inkjet printing of 3-D electric circuits”, IEEE Transactions on Electronics Packaging Manufacturing, Vol. 28, pp. 265-273.

Meisel, N.A., Williams, C.B. and Druschitz, A. (2012), “Lightweight metal cellular structures via indirect 3D printing and casting”, Proceedings of the International Solid Freeform Fabrication Symposium, pp. 162-176.

Melchels, F.P., Feijen, J. and Grijpma, D.W. (2010), “A review on stereolithography and its applications in biomedical engineering”, Biomaterials, Vol. 31, pp. 6121-6130.

Melvin, L.S., III, Das, S. and Beaman, S. Jr (1994), “Video microscopy of selective laser sintering”, Proceedings of the Solid Freeform Fabrication Symposium, DTIC Document, pp. 34-41.

Mireles, J., Ridwan, S., Morton, P.A., Hinojos, A. and Wicker, R.B. (2015), “Analysis and correction of defects within parts fabricated using powder bed fusion technology”, Surface Topography: Metrology and Properties, Vol. 3, p. 034002.

Mireles, J., Terrazas, C., Medina, F., Wicker, R. and Paso, E. (2013), “Automatic feedback control in electron beam melting using infrared thermography”, Proceedings of the Solid Freeform Fabrication Symposium.

Mostafa, N., Syed, H.M., Igor, S. and Andrew, G. (2009), “A study of melt flow analysis of an ABS-iron composite in fused deposition modelling process”, Tsinghua Science & Technology, Vol. 14, pp. 29-37.

Muller, M.S. & De Jean, P.D. (2015), “3D microscopy for microfabrication quality control”, SPIE OPTO, International Society for Optics and Photonics, pp. 937607-937607-11.

Nakamura, M. and Henmi, C. (2008), “3D Micro-fabrication by inkjet 3D biofabrication for 3D tissue engineering”, International Symposium on Micro-NanoMechatronics and Human Science, MHS 2008, IEEE, pp. 451-456.

Pandey, P., Reddy, N.V. and Dhande, S. (2003a), “Real time adaptive slicing for fused deposition modelling”, International Journal of Machine Tools and Manufacture, Vol. 43, pp. 61-71.

Pandey, P.M., Reddy, N.V. and Dhande, S.G. (2003b), “Improvement of surface finish by staircase machining in fused deposition modeling”, Journal of Materials Processing Technology, Vol. 132, pp. 323-331.

Patel, J., Patel, C. and Patel, U. (2012), “A review on various approach for process parameter optimization of fused deposition modeling (FDM) process and Taguchi approach for optimization”, International Journal of Engineering Research and Applications, Vol. 2, pp. 361-365.

Percoco, G., Lavecchia, F. and Galantucci, L.M. (2012), “Compressive properties of FDM rapid prototypes treated with a low cost chemical finishing”, Research Journal of Applied Sciences, Engineering and Technology, Vol. 4, pp. 3838-3842.

Pham, D., Dimov, S. and Gault, R. (1999), “Part orientation in stereolithography”, The International Journal of Advanced Manufacturing Technology, Vol. 15, pp. 674-682.

Price, S., Cooper, K. and Chou, K. (2012), “Evaluations of temperature measurements by near-infrared thermography in powder-based electron-beam additive manufacturing”, Proceedings of the Solid Freeform Fabrication Symposium, University of Texas, Austin, TX, pp. 761-773.

Rao, R.V. and Rai, D.P. (2016), “Optimization of fused deposition modeling process using teaching-learning-based optimization algorithm”, Engineering Science and Technology, an International Journal, Vol. 19, pp. 587-603.

Reeves, P. and Cobb, R. (1995), “Surface deviation modeling of LMT processes–a comparative analysis”, Proceedings of the Fifth European Conference on Rapid Prototyping and Manufacturing, Helsinki, June, pp. 59-77.

Rengier, F., Mehndiratta, A., Von Tengg-Kobligk, H., Zechmann, C.M., Unterhinninghofen, R., Kauczor, H.-U. and Giesel, F.L. (2010), “3D printing based on imaging data: review of medical applications”, International Journal of Computer Assisted Radiology and Surgery, Vol. 5, pp. 335-341.

Sachs, E. and Vezzetti, E. (2005), “Numerical simulation of deposition process for a new 3DP printhead design”, Journal of Materials Processing Technology, Vol. 161, pp. 509-515.

Shaji, S. and Radhakrishnan, V. (2003), “Analysis of process parameters in surface grinding with graphite as lubricant based on the Taguchi method”, Journal of Materials Processing Technology, Vol. 141, pp. 51-59.

Shishkovsky, I., Scherbakov, V., Morozov, Y., Kuznetsov, M. and Parkin, I. (2008), “Surface laser sintering of exothermic powder compositions”, Journal of Thermal Analysis and Calorimetry, Vol. 91, pp. 427-436.

Silva, D.N., De Oliveira, M.G., Meurer, E., Meurer, M.I., Da Silva, J.V.L. and Santa-Bárbara, A. (2008), “Dimensional error in selective laser sintering and 3D-printing of models for craniomaxillary anatomy reconstruction”, Journal of Cranio-Maxillofacial Surgery, Vol. 36, pp. 443-449.

Sitthi-Amorn, P., Ramos, J.E., Wangy, Y., Kwan, J., Lan, J., Wang, W. and Matusik, W. (2015), “MultiFab: a machine vision assisted platform for multi-material 3D printing”, ACM Transactions on Graphics ( Graphics), Vol. 34 No. 4, p. 129.

Smurov, I., Doubenskaia, M. and Zaitsev, A. (2012), “Complex analysis of laser cladding based on comprehensive optical diagnostics and numerical simulation”, Physics Procedia, Vol. 39, pp. 743-752.

Snelling, D., Blount, H., Forman, C., Ramsburg, K., Wentzel, A., Williams, C. and Druschitz, A. (2013), “The effects of 3D printed molds on metal castings”, Proceedings of the Solid Freeform Fabrication Symposium, pp. 827-845.

Sonmez, F.O. and Hahn, H.T. (1998), “Thermomechanical analysis of the laminated object manufacturing (LOM) process”, Rapid Prototyping Journal, Vol. 4, pp. 26-36.

Sood, A.K., Ohdar, R. and Mahapatra, S. (2009), “Improving dimensional accuracy of fused deposition modelling processed part using grey Taguchi method”, Materials & Design, Vol. 30, pp. 4243-4252.

Sood, A.K., Ohdar, R. and Mahapatra, S. (2010), “Parametric appraisal of mechanical property of fused deposition modelling processed parts”, Materials & Design, Vol. 31, pp. 287-295.

Sood, A.K., Ohdar, R.K. and Mahapatra, S.S. (2012), “Experimental investigation and empirical modelling of FDM process for compressive strength improvement”, Journal of Advanced Research, Vol. 3, pp. 81-90.

Sreeram, P.D.D. (1995), “Determination of optimal orientation based on variable slicing thickness in layered manufacturing”, ASME Winter Annual Meeting, ASME, San Francisco.

Stratasys. (2016), “Finishing processes”, (accessed 16 July 2016).

Stucker, B. (2012), “Additive manufacturing technologies: technology introduction and business implications”, Frontiers of Engineering: Reports on Leading-Edge Engineering From the 2011 Symposium, National Academies Press, Washington, DC, pp. 19-21.

Sun, Q., Rizvi, G., Bellehumeur, C. and Gu, P. (2008), “Effect of processing conditions on the bonding quality of FDM polymer filaments”, Rapid Prototyping Journal, Vol. 14, pp. 72-80.

Tapia, G. and Elwany, A. (2014), “A review on process monitoring and control in metal-based additive manufacturing”, Journal of Manufacturing Science and Engineering, Vol. 136, p. 060801.

Taylor, C. and Childs, T. (2001), “Thermal experiments in direct metal laser sintering”, Proceedings of Euro RP.

Thrimurthulu, K., Pandey, P.M. and Reddy, N.V. (2004), “Optimum part deposition orientation in fused deposition modeling”, International Journal of Machine Tools and Manufacture, Vol. 44, pp. 585-594.

Tourloukis, G., Stoyanov, S., Tilford, T. and Bailey, C. (2015), “Data driven approach to quality assessment of 3D printed electronic products”, 38th International Spring Seminar on Electronics Technology (ISSE), IEEE, pp. 300-305.

Toyserkani, E. and Khajepour, A. (2006), “A mechatronics approach to laser powder deposition process”, Mechatronics, Vol. 16, pp. 631-641.

Uma Maheshwaraa, N., Arumaikkannu, G. and Gowri, S. (2008), “Three-dimensional reconstruction and rapid prototyping of femur bone using multiple digital X-rays”, Journal of Medical Engineering & Technology, Vol. 32, pp. 30-39.

Upcraft, S. and Fletcher, R. (2003), “The rapid prototyping technologies”, Assembly Automation, Vol. 23, pp. 318-330.

Vaezi, M., Chianrabutra, S., Mellor, B. and Yang, S. (2013), “Multiple material additive manufacturing–part 1: a review: this review paper covers a decade of research on multiple material additive manufacturing technologies which can produce complex geometry parts with different materials”, Virtual and Physical Prototyping, Vol. 8 No. 1, pp. 19-50.

Van Belle, L., Vansteenkiste, G. and Boyer, J.C. (2013), “Investigation of residual stresses induced during the selective laser melting process”, Key Engineering Materials, Vol. 554/557, pp. 1828-1834.

Vijay, P., Danaiah, P. and Rajesh, K. (2011), “Critical parameters effecting the rapid prototyping surface finish”, Journal of Mechanical Engineering and Automation, Vol. 1, pp. 17-20.

Wang, L., Felicelli, S.D. and Craig, J.E. (2009), “Experimental and numerical study of the LENS rapid fabrication process”, Journal of Manufacturing Science and Engineering, Vol. 131, p. 041019.

West, A.P., Sambu, S.P. and Rosen, D.W. (2001), “A process planning method for improving build performance in stereolithography”, Computer-Aided Design, Vol. 33, pp. 65-79.

Wilson, W.C. and Boland, T. (2003), “Cell and organ printing 1: protein and cell printers”, The Anatomical Record Part A: Discoveries in Molecular, Cellular, and Evolutionary Biology, Vol. 272, pp. 491-496.

Wong, K.V. and Hernandez, A. (2012), “A review of additive manufacturing”, ISRN Mechanical Engineering, Vol. 2012 No. 2012.

Wüst, S., Müller, R. and Hofmann, S. (2011), “Controlled positioning of cells in biomaterials–approaches towards 3D tissue printing”, Journal of Functional Biomaterials, Vol. 2, pp. 119-154.

Xu, T., Gregory, C.A., Molnar, P., Cui, X., Jalota, S., Bhaduri, S.B. and Boland, T. (2006), “Viability and electrophysiology of neural cell structures generated by the inkjet printing method”, biomaterials, Vol. 27, pp. 3580-3588.

Yadroitsev, I., Krakhmalev, P. and Yadroitsava, I. (2014), “Selective laser melting of Ti6Al4V alloy for biomedical applications: temperature monitoring and microstructural evolution”, Journal of Alloys and Compounds, Vol. 583, pp. 404-409.

Ye, R., Smugeresky, J.E., Zheng, B., Zhou, Y. and Lavernia, E.J. (2006), “Numerical modeling of the thermal behavior during the LENS® process”, Materials Science and Engineering: A, Vol. 428, pp. 47-53.

Yeong, W.-Y., Chua, C.-K., Leong, K.-F., Chandrasekaran, M. and Lee, M.-W. (2006), “Indirect fabrication of collagen scaffold based on inkjet printing technique”, Rapid Prototyping Journal, Vol. 12, pp. 229-237.

Yu, D.-G., Branford-White, C., Ma, Z.-H., Zhu, L.-M., Li, X.-Y. and Yang, X.-L. (2009), “Novel drug delivery devices for providing linear release profiles fabricated by 3DP”, International Journal of Pharmaceutics, Vol. 370, pp. 160-166.

Zeng, K., Pal, D. and Stucker, B. (2012), “A review of thermal analysis methods in laser sintering and selective laser melting”, Proceedings of Solid Freeform Fabrication Symposium, Austin, TX.

Zhou, C., Chen, Y., Yang, Z. and Khoshnevis, B. (2011), “Development of multi-material mask-image-projection-based stereolithography for the fabrication of digital materials”, Annual Solid Freeform Fabrication Symposium, Austin, TX.

## Corresponding author

Hoejin Kim can be contacted at: hkim4@miners.utep.edu