Binder Removal Behavior of Cu-Al-Ni Green Components Formed by Powder Injection Molding

Abstract

Powder injection molding (PIM) of Cu-Al-Ni components, which exhibit shape memory alloy (SMA) characteristics, is severely challenged by crack formation during the removal of binders in the debinding stage. This study investigated the debinding of Cu-Al-Ni components. For obtaining the feedstock, a Cu-Al-Ni powder mixture was mixed with paraffin wax (PW), high-density polyethylene (HDPE), and stearic acid (SA) binders. This study elucidated that PW and SA removal in green Cu-Al-Ni components increased markedly with temperature during solvent debinding, revealing the material’s pronounced thermal sensitivity. While PW and SA removal peaked at 70 °C, cracking in several components necessitated reducing the solvent debinding temperature to 60 °C, achieving 69.5% binder extraction. Thermal debinding removed the insoluble HDPE and residual soluble binders, as evidenced by thermogravimetric analysis (TGA), while low heating rates (0.1°C/min and 3°C/min) effectively prevented defects. This study presents a debinding route that enables Cu–Al–Ni PIM components to retain their macroscopic integrity during subsequent sintering, providing a basis for further optimization toward SMA applications.

Keywords:
Powder injection molding; Cu-Al-Ni components; Feedstock; Debinding; Thermogravimetric analysis

1. Introduction

Shape memory alloys (SMAs) are a unique type of smart material distinguished by their extraordinary ability to restore their original shape after being deformed, a feature that is driven by two key mechanisms: the shape memory effect (SME) and super-elasticity (SE)¹. These two phenomena are caused by a reversible solid-state phase transition between two different stages: austenite, which is stable at elevated temperatures, and martensite, which develops at lower temperatures²,³. The outstanding characteristics of SMAs have made them useful in various fields, including actuators, robotics, aerospace, automotive, biomedical, and construction³,⁴. Recently, Cu-Al-Ni SMAs have drawn significant attention due to their exceptional thermal stability (up to 200°C), high recoverable super-elastic strain of around 18%, and excellent damping ability⁵,⁶. Cu-Al-Ni SMAs produced via casting, powder metallurgy, and spray deposition suffer from challenges such as coarse grain structures, brittleness, intergranular cracking, and internal porosity⁷,⁸. Considering these constraints, powder injection molding (PIM) offers a potential shaping route for producing small and geometrically complex Cu-Al-Ni components. PIM is a technique for low-cost production of intricate and small structures in net shape, proceeding through four sequential steps such as mixing, injection molding, debinding, and sintering.

PIM is highly dependent on the compatibility between the powder and binder system. Powder characteristics such as particle size, morphology, and packing ability influence powder loading, feedstock viscosity, mold filling, and green-part uniformity. At the same time, the binder formulation must provide sufficient flowability during molding while preserving the shape of the component during debinding. If the powder–binder system is not properly balanced, defects such as binder separation, cracking, slumping, and dimensional instability may occur, reducing the quality of the final product⁹. Accordingly, many studies have examined the effects of feedstock composition, binder formulation, and processing conditions on PIM efficiency and final component properties. For instance, Cicek et al.¹⁰ evaluated Mg, SiO₂, SiC, and 316L powders with different binder systems, including PW–EVA–SA, PW–PP–SA, PW–LDPE–SA, and PEG–PMMA–SA, in a newly designed high-pressure PIM system. They showed that powder geometry and binder formulation strongly affected feedstock preparation, molding behavior, viscosity, porosity, and surface quality, with the 316L/PEG–PMMA–SA combination giving the most suitable molding response. Another research¹¹ investigated the microstructural evolution of gas-assisted MIM-processed Mg–0.5Ca alloy using 40 μm D90 Mg powder and 500 nm D90 Ca powder with a PEG–PMMA–SA binder system. By varying the sintering temperature and holding time, they showed that the alloy reached a metallic structure only after sintering at 600 °C for 5 h, where grain-boundary formation and Mg₂Ca phase development were observed, with an average hardness of 49.9 HV. In a related study¹², 316L stainless steel components were produced using a newly designed high-gas-pressure modified metal injection system with a PEG–PMMA–SA binder formulation and approximately 60 vol.% powder loading. After debinding and sintering at 1100–1300 °C, the produced parts exhibited a density of 6.74 g/cm³, hardness of about 285 HV, corrosion rate of 17.08 × 10⁻³ mm/year, and cell viability of approximately 103%, indicating that process modification and binder-system design can influence final material performance and biocompatibility.

Debinding is a critical stage in PIM because the binder composition must be compatible with the powder characteristics and the selected binder-removal route. This is particularly challenging for multi-component binder systems, where controlled removal is required to preserve component shape and prevent defects before sintering¹³. Accordingly, two-stage debinding, involving solvent extraction of the soluble binder fraction followed by thermal removal of the backbone polymer, is commonly employed in PIM. Solvent debinding removes soluble binder constituents by immersing the green part in a suitable solvent, such as hexane, heptane, ethanol, or acetone, under controlled temperature conditions, thereby forming an open-pore network for subsequent thermal debinding¹⁴. However, solvent penetration depends on component geometry, surface-to-volume ratio, and the intrinsic porosity of the molded part¹⁵. Although higher temperatures can accelerate binder–solvent interaction and diffusion, excessive extraction rates may cause swelling, cracking, or slumping¹⁶. For example, Samanta et al.¹⁷ reported that solvent debinding of 316L stainless steel components containing PW and SA at elevated temperatures ≥50 °C caused surface defects and dimensional distortion. Thermal debinding is then used to remove the remaining insoluble binder through controlled heating, typically under an inert atmosphere or vacuum. Since low-molecular-weight binders evaporate first while high-molecular-weight polymers degrade and release volatile products, careful heating control is required to avoid cracking, bloating, and blistering¹⁵. Similarly, Tseng et al.¹⁸ reported cracking during the early thermal debinding stage of yttria-stabilized zirconia samples containing a vinyl acetate-based polymer binder, further highlighting the importance of maintaining structural integrity during binder removal.

For Cu–Al–Ni SMAs, the control of debinding is particularly important because defects generated at this stage may not only affect the physical integrity of the component, but also influence the subsequent development of its functional shape memory response. In this study, a Cu–Al–Ni system was prepared from Cu, Al, and Ni powders and processed using a wax-based binder system composed of paraffin wax (PW), high-density polyethylene (HDPE), and stearic acid (SA). PW promotes feedstock flow, HDPE provides backbone strength for shape retention, and SA improves powder–binder interaction and feedstock homogeneity⁹,¹⁹. Inadequate debinding of this system may result in cracking, warping, swelling, residual binder-derived contamination, or nonuniform porosity, which can impair sintering behavior through poor densification, dimensional instability, and defect retention. For Cu–Al–Ni SMAs, these issues are especially critical because the final functional response depends on phase stability, microstructural uniformity, and the formation of martensite after quenching²⁰ Therefore, defects or chemical/microstructural heterogeneities introduced before sintering may later affect martensitic transformation behavior, transformation temperatures, and shape memory performance. Although debinding of multi-component binder systems has been widely studied in PIM, rare studies have specifically addressed solvent and thermal debinding behavior in Cu-based SMA systems, particularly Cu–Al–Ni green components containing a PW–HDPE–SA binder system. Therefore, this work investigates the binder-removal behavior of Cu–Al–Ni PIM green components during solvent and thermal debinding, with emphasis on binder removal, defect formation, and retention of component integrity.

2. Materials and Methods

Gas-atomized Cu, Al, and Ni powders were used as the raw materials for the current study. These powders were purchased from Vistec Technology Services, Malaysia. The particle size distributions of the individual Cu, Al, and Ni powders and the mixed Cu–Al–Ni powder were measured using a HORIBA LA-960 laser scattering particle size distribution analyzer. The measurements were conducted on a volume basis in water as the dispersion medium, with circulation and agitation speeds set to 2 and ultrasound turned off. For the individual powders, material-specific optical parameters were used, whereas the mixed powder was evaluated using the Fraunhofer Kernel due to the presence of multiple metallic constituents. The reported D10, D50, and D90 values therefore represent volume-based cumulative particle size values. The pycnometer densities of the powders were measured using an AccuPyc II 1340 Pycnometer. These powder characteristics are summarized in Table 1.

Table 1
Properties of preliminary powders.

Powder	D10 (µm)	D50 (µm)	D90 (µm)	Pycnometer density (g/cm3)
Cu	21.9	52.1	92.3	8.69
Al	5.5	11.9	21.5	2.57
Ni	5.4	8.4	12.7	8.66
Mixture	14.1	24.9	47.3	6.75

As shown in Table 1, the Cu powder exhibited a relatively coarse particle size distribution, with D50 and D90 values of 52.1 and 92.3 µm, respectively. Although such a coarse Cu fraction is higher than that commonly preferred for conventional MIM/PIM feedstocks, the mixed Cu–Al–Ni powder exhibited lower D50 and D90 values of 24.9 and 47.3 µm, respectively. This reduction can be partly explained by the substantial volumetric contribution of the finer Al powder. Based on the nominal 84Cu–12Al–4Ni wt.% composition and the measured powder densities, Al contributes approximately one-third of the powder volume despite its lower weight fraction. Therefore, the fine Al fraction shifts the measured volume-based distribution of the powder blend toward smaller particle sizes. In addition, the mixed powder distribution was measured as an independent blended sample using the Fraunhofer Kernel, rather than being calculated from the individual powder distributions. Thus, the mixture particle size distribution should be interpreted as the measured response of the blended powder system, which may be influenced by sampling and dispersion during laser diffraction analysis. From a processing perspective, the relatively coarse Cu fraction may reduce the overall surface area and assist feedstock flow, but it may also influence interparticle contact, diffusion distance, pore evolution, and densification during subsequent sintering. The morphology of the powder mixture was inspected by employing the field emission scanning electron microscope (FESEM, Zeiss Merlin Compact) and is displayed in Figure 1. As shown, the gas-atomized Cu–Al–Ni powder mixture exhibited predominantly spherical to near-spherical morphology rather than highly irregular particle geometry. This morphology is beneficial for PIM processing because rounded particles can move more easily relative to each other during feedstock flow, reducing interparticle resistance and supporting moldability²¹.

Figure 1
Morphology of the Cu-Al-Ni powder mixture.

Prior to feedstock preparation, the elemental Cu, Al, and Ni powders were physically blended according to the nominal composition of 84Cu–12Al–4Ni (wt.%). This low-energy stirring step was intended to distribute the constituent powders before binder mixing, rather than to induce mechanical alloying or complete chemical homogenization. The weighed powders were placed in a plastic container together with 6 mm stainless steel balls at a ball-to-powder ratio of 2:1 and stirred for 30 min using a plastic impeller at 200 rpm. In this experiment, a binder system consisted of 65 wt.% of PW, 30 wt.% of HDPE, and 5 wt.% of SA, which worked as the main constituent, backbone polymer, and surfactant, respectively. While PW and SA were supplied by Systerm Co., Malaysia, HDPE was purchased from Titanex, Malaysia. The densities of PW, HDPE, and SA were 0.89 g/cm³, 0.94 g/cm³, and 0.98 g/cm³, respectively. The melting point and the decomposition temperature of the binders were determined using a differential scanning calorimetry (DSC) and a thermogravimetric (TGA) analysis, carried out employing TA Instruments Q2000 and the PerkinElmer Simultaneous Thermal Analyzer (STA) 6000. The temperature during the mixing and debinding processes was established using the DSC and TGA data as the baseline values.

Initially, the critical powder volume percentage (CPVP) is required to be measured first in order to determine the optimal powder loading. This study used the oil absorption approach in compliance with ASTM standard 281-31 to perform a CPVP analysis to ascertain the critical powder loading of Cu-Al-Ni powder mixture²². The CPVP was measured based on Equation 1, as can be seen below:

$CPVP\left(\%\right)=\frac{V_P}{V_P+V_0}\times 100$ (1)

where $V_P$ is the volume of powder and $V_0$ is the volume of oleic acid.

A powder loading of 66 vol.% was used for feedstock preparation, as discussed in Section 3.2. The Cu–Al–Ni powder mixture at this loading was mixed with the PW–HDPE–SA binder system using a Brabender W50 EHT mixer for 45 min at 150 °C and 30 rpm to produce the Cu–Al–Ni feedstock. Based on the selected powder loading and the measured densities of the powder mixture and binder constituents, the total binder content in the feedstock was calculated to be approximately 6.49 wt.%. The corresponding mass fractions of PW, HDPE, and SA in the total feedstock were approximately 4.22, 1.95, and 0.32 wt.%, respectively. Therefore, the theoretical soluble binder fraction, corresponding to PW and SA, was approximately 4.54 wt.%, while the insoluble HDPE backbone fraction was approximately 1.95 wt.% of the total feedstock. Afterwards, the feedstock was subjected to injection molding by utilizing a DSM Xplore IM12 semi-automatic injection molding machine to produce defect-free green Cu-Al-Ni components in the form of a tensile bar shape (66 mm in length and 4 mm in thickness). Table 2 outlines the parameters utilized in the fabrication of green components.

Table 2
Injection parameters to produce green components.

Melt temperature (°C)	Mold temperature (°C)	Injection pressure (bar)	Injection time (s)
200	150	12	6

There were two stages associated with the debinding process: solvent and thermal. Soluble PW and SA were eliminated during the solvent debinding stage. Solvent debinding was performed by immersing the green Cu-Al-Ni components in n-heptane at different temperatures between 50°C and 70°C for 8 h by employing a BINDER FDL 115 safety drying oven. By tracking mass loss for five samples under identical conditions, the elimination rate of the binder through solvent debinding was determined. The rate of binder eliminated was obtained based on the following equation²³:

$W_d\left(\%\right)=\frac{W_i-W_s}{W_i}\times 100$ (2)

Where $W_d$ is the mass loss percentage after solvent debinding, $W_i$ is the initial mass of the green specimen, and $W_s$ is the mass of the specimen after solvent debinding. The apparent soluble binder removal was then calculated by normalizing the measured mass loss to the theoretical PW + SA content in the feedstock, according to Equation (3):

$\text{Soluble binder removal}\left(\text{\%}\right)=\frac{W_d}{w_{PW}+w_{SA}}\times 100$ (3)

Where $W_d$is obtained from Equation (2), and $w_{PW}+w_{SA}$ is the theoretical mass percentage of PW and SA in the total feedstock. Therefore, the reported solvent-debinding percentages represent the apparent removal of the soluble binder fraction rather than the mass loss of the entire feedstock. TGA was performed on the solvent-debound Cu–Al–Ni samples to further evaluate the remaining binder decomposition behavior after solvent debinding.

In this study, the insoluble HDPE binder was removed by thermal debinding under an argon environment in a split furnace (RS800/200/200). The samples were heated from 30 °C to 200 °C at a rate of 0.1°C/min in the first stage and then to 510°C at a rate of 0.3°C/min, followed by a holding time of 4 h. The thermally debound samples were then cooled to room temperature at a rate of 3°C/min. TGA was conducted on thermally debound Cu-Al-Ni samples to evaluate the remaining binder decomposition behavior after thermal debinding. To further support the interpretation of binder removal, the carbon content of the samples was measured using a combustion-based carbon analysis method in accordance with ASTM E1019-11 using a LECO CS744 analyzer. The analysis was performed on the initial raw Cu–Al–Ni powder mixture, used as the reference, and on the thermally debound part to evaluate the residual carbon content after thermal debinding. The measured carbon values were used as complementary evidence to the mass-loss and TGA results, particularly for assessing the removal of carbon-containing binder residues. For thermal debinding, the remaining binder was considered to consist of the residual soluble binder fraction left after solvent debinding together with the insoluble HDPE backbone polymer. Since the powder system contained elemental Al, possible oxidation-related mass gain during heating was also considered when interpreting the mass-loss data. Although thermal debinding was conducted under argon to minimize oxidation, the measured mass changes were treated as apparent/net values because minor oxidation-related mass gain, if present, could partially offset binder-removal mass loss. Subsequent to thermal debinding, the specimens were sintered at 900°C for 2 hr in an argon-controlled tube furnace (HTF-15/200–60) to yield the consolidated components.

3. Results and Discussion

3.1. Materials characterization

Based on DSC curves, as can be seen in Figure 2a, the melting points of the PW, HDPE, and SA binders were signified by the peak temperatures, which were 66.3°C, 133.3°C, and 77.6°C, respectively. The presence of two different molecules in PW could account for the two distinct peaks on its DSC curve. As HDPE exhibited a higher melting point in comparison to PW and SA, the mixing procedure in this investigation was performed at a temperature (150°C) greater than HDPE's melting point. As shown in Figure 2b, the TGA was employed to determine the decomposition temperatures of PW, SA, and HDPE. The decomposition of PW, HDPE, and SA commenced at 200°C, 407.9°C, and 182.5°C and came to an end at 327.5°C, 509.4°C, and 340°C, respectively.

Figure 2
(a) DSC and (b) TGA curves of PW, HDPE, and SA.

3.2. Preparation of feedstock

The optimal powder loading of the Cu-Al-Ni powder mixture was determined based on critical powder loading²⁴. Basically, the physical and mechanical characteristics of the sintered component are improved by increasing powder loading, but excessive loading makes it difficult to mix the powder and binder. Likewise, due to elevated interparticle friction, such an intense loading escalates the viscosity of the feedstock, which prevents it from optimally flowing into the mold cavity during the injection molding process²⁵. In this study, CPVP analysis was used to experimentally establish the powder loading behavior of a Cu-Al-Ni powder mixture. The analysis, performed by mixing the Cu-Al-Ni powder mixture with oleic acid using a twin-screw blade mixer, produced a CPVP curve from which a critical powder loading of 70 vol.% was identified (Figure 3a). Based on this value and following the approach reported by German and Bose⁹, an optimal powder loading of 66 vol.% was selected for Cu-Al-Ni system. The powder mixture at this optimal loading was then mixed with the PW-HDPE-SA binder system to prepare a feedstock. The mixing process, in the current investigation, led to an effective dispersion of the Cu-Al-Ni powder mixture in the binder system. The homogeneity of the powder-binder mixture was tested using the twin-screw-blade mixer. While blending the materials, this mixer recorded the mixing torque. Throughout the mixing procedure, the torque was tracked to observe how the consistency of the mixture evolved. The mixing curve of Cu-Al-Ni feedstock is illustrated in Figure 4. As shown in Figure 4, the homogeneity of the feedstock was confirmed by the stable value following a preliminary rise in the mixing torque throughout time. Since homogeneity was attained in 45 minutes, this was the optimal time for mixing to produce the feedstock. Usually, defects during sintering are less likely to occur with a homogenous feedstock. The FESEM micrograph of the Cu-Al-Ni feedstock, as demonstrated in Figure 5, ensures that the particles were completely covered with the binder system.

Figure 3
Critical powder loading obtained for Cu-Al-Ni powder mixture.

Figure 4
Evolution of torque for Cu-Al-Ni feedstock as a function of mixing time.

Figure 5
FESEM micrograph of Cu-Al-Ni feedstock.

3.3. Rheological analysis

Evaluating the flowability of feedstock is crucial to producing successful injection molded parts. In the current investigation, rheological characteristics were evaluated by analyzing the viscosity profile of the Cu-Al-Ni feedstock in relation to temperature and shear rate. Figure 6 depicts the viscosity changes at different shear rates at 160°C, 180°C, and 200°C. It is evident that at all temperatures, viscosity decreased with increasing shear rate, that is, pseudoplastic behavior. Therefore, the particle arrangement within the melt, intact at reduced shear rates, was disrupted, leading to hydrodynamic interaction becoming the key aspect regulating the viscosity changes. This propensity eventually caused the powder particles and polymer chains to realign with the flow direction, which led to interparticle motion and pseudoplastic behavior²⁶. Moreover, pseudoplastic behavior implied that increased shear allowed particles to spread more uniformly, effectively fitting smaller particles between larger ones. Furthermore, such behavior could potentially be correlated with the breakup of particle clusters as a consequence of the elevated shear rate, resulting in better particle packing and more homogeneous feedstock²⁷. As can be seen in Figure 6, the viscosity of Cu-Al-Ni feedstock decreased with the increase of temperature, which could be attributed to the diminished molecular attraction of the binders at higher temperatures. Based on Figure 6, the range of viscosities at various temperatures was consistent with the PIM investigators' recommended limit of less than 1000 Pa.s²⁸. A low viscosity feedstock is essential not only to eliminate short-shot defects but also to ensure that the mold cavity is filled efficiently during injection molding.

Figure 6
Viscosity variation as a function of shear rate at various temperatures.

The relationship between viscosity $\left(\eta \right)$ and shear rate $\left({\rm Y}\right)$ can be described by Equation (4) as the Cu-Al-Ni feedstock demonstrated pseudoplastic behaviour:

$\eta =K{{\rm Y}}^{n-1}$ (4)

where $K$ represents the constant and $n$ denotes the flow behavior index. The value of $n$ typically reflects shear sensitivity, with pseudoplastic behavior occurring when $n<1$. Table 3 displays the calculated $n$ values for the feedstock, as derived from Figure 6, indicating a decline in $n$ values with the rise in temperature. Since a rise in temperature can elevate particle mobility, it leads to a reduction in $n$. A reduced value of $n$ signifies a feedstock with heightened shear sensitivity, exhibiting pronounced pseudoplastic behavior. This trait is particularly advantageous in injection molding, as it facilitates a swift attenuation of the feedstock’s viscosity in response to escalating shear rates during the molding operation²⁹. In this study, the lowest $n$ value was observed at 200°C, which was regarded as optimal for the feedstock and eventually adopted as the melting temperature during the injection molding.

Table 3
Flow behavior index at different temperatures for Cu-Al-Ni feedstock.

Temperature (°C)	Flow behavior index $\left(n\right)$
160	0.42
180	0.40
200	0.35

The temperature-induced variations in viscosity represent a pivotal factor to consider when scrutinizing the rheological properties of a feedstock. The Arrhenius equation, typically represented in the subsequent manner, is extensively utilized to elucidate the correlation between feedstock viscosity and temperature:

$\eta \left(T\right)={\eta }_o$exp $\left(E/RT\right)$(5)

where ${\eta }_o$, $E$, $R$, and $T$ denote the reference viscosity, flow activation energy, gas constant, and absolute temperature, respectively. In this study, E was obtained from the slope of the ln(η) versus 1/T plot, as shown in Figure 7. The flow activation energy of the Cu–Al–Ni feedstock was calculated to be 24.94 kJ/mol, indicating relatively low-to-moderate temperature sensitivity of viscosity. This value is lower than those reported for several metallic PIM feedstock systems, although direct comparison depends on powder type, powder loading, particle morphology, binder formulation, and testing conditions²⁷,²⁹,³⁰. The relatively moderate E value can be attributed to the combined effect of the PW/HDPE/SA binder system and the characteristics of the Cu–Al–Ni powder mixture. PW, as the major low-molecular-weight binder constituent, promotes flowability, while SA improves powder–binder interaction and reduces interparticle friction. HDPE provides backbone strength but remains distributed within the wax-based binder system, preventing excessive temperature sensitivity during flow. In addition, the predominantly spherical to near-spherical morphology of the gas-atomized powders and the relatively coarse Cu fraction may reduce interparticle resistance, contributing to stable feedstock flow. Therefore, the obtained activation energy suggests that the feedstock viscosity was not highly sensitive to small temperature fluctuations during injection molding, which is beneficial for reducing flow instability and molding defects.

Figure 7
$\text{ln}(\eta$) versus $1/T$ graph of the Cu-Al-Ni feedstock.

3.4. Morphology of green part

Figure 8a depicts the green Cu-Al-Ni component that was free of flaws such as short shot, cracking, and jetting. The production of a high-quality injected part is thus demonstrated by feedstock with pseudoplastic behavior, low viscosity, and low activation energy. Basically, the green component is critical in the fabrication process, as it represents the preliminary stage of the material, acting as a precursor to the final product. Figure 8b exhibits the FESEM micrograph of the green Cu-Al-Ni component. According to Figure 8b, the powder particles were homogeneously disseminated throughout the entire component, with each particle adequately enveloped by the binder system. Figure 9 displays the Energy-Dispersive X-ray Spectroscopy (EDS) map of the green component. This validated the relatively homogeneous distribution of Cu, Al, and Ni across the component. Although the FESEM/EDS observations of the green component indicate a relatively uniform distribution of Cu, Al, and Ni at the examined scale, this should be interpreted as powder-scale physical distribution rather than complete alloy formation. Since elemental Al powder is known to develop a stable native Al₂O₃ surface layer, this factor should be considered as a possible barrier to interparticle contact and diffusion during subsequent sintering³¹. Therefore, alloy formation and further homogenization in the present elemental powder system are expected to occur mainly during sintering, while the initial low-energy mixing step serves primarily to distribute the elemental powders before feedstock preparation. Moreover, the uniform distribution of carbon elements implied a homogeneously applied binder system throughout the green component.

Figure 8
(a) Green Cu-Al-Ni component and (b) FESEM image of the component.

Figure 9
EDS map of the green Cu-Al-Ni component.

3.5. Debinding

In this research work, a multi-component binder system consisting of PW, HDPE, and SA was employed; hence, these binder components were eliminated gradually. According to German³², it is important for one of the binder constituents to be present in amounts equivalent to 30% or greater in the binder composition so that it can be preferentially removed during the debinding process in order to prevent the formation of isolated pore pockets. Debinding of the tensile bar shaped green Cu-Al-Ni components was performed in two stages: solvent and thermal debinding. As PW and SA binders comprised 70% and exhibited solubility in n-heptane, they were effectively eliminated from the green components by immersing green components in n-heptane at a temperature range of 50°C-70°C for 8 h. Upon immersing the specimens in the solvent, the PW and SA binders began to integrate the solvent, forming a swollen gel that progressively dissolved as the solvent concentration reached a sufficiently high level³³,³⁴. The partial removal of the PW and SA binders resulted in the formation of pore spaces within the specimens, which expanded towards the interior of the specimens as the debinding time and temperature were increased. The amount of PW and SA extracted with time for the Cu-Al-Ni SMA specimens at various temperatures is shown in Figure 10. Figure 10 indicates that the mass loss of the soluble binders intensified as both time and temperature were elevated. When the temperature in this experiment was raised from 50°C to 70°C, the rate of elimination of PW and SA significantly increased from 56.6% to 72.8%. The solvent debinding process was predominantly regulated by the diffusion mechanism, in which temperature emerged as a critical determinant, modulating the rate of diffusion. As the temperature escalated, a simultaneous increase in binder removal occurred, driven by the accelerated diffusion kinetics³⁵,³⁶. It was noticed that several of the specimens developed cracks when solvent debound at 70°C, as displayed in Figure 11, which could be the consequence of thermal expansion and contraction. When the backbone polymer, HDPE, is immersed in solvent for extended periods, it undergoes swelling, resulting in expansion. Upon completion of the solvent debinding process at elevated temperatures, when the specimens are retrieved from the solvent, the solvent is expelled and evaporates from the specimen’s surface, leading to a rapid decrease in temperature. This abrupt temperature drop induces contraction. The sequence of expansion followed by a sudden contraction generates internal stresses within the specimen, which eventually lead to the formation of cracks¹⁷. Since cracks appeared at 70°C and 69.5% of the PW and SA binders were removed at 60°C, the chosen debinding temperature for this study was 60°C.

Figure 10
PW and SA removal percentage versus debinding time for specimens debound at three different temperatures.

Figure 11
Photograph of crack formation during solvent debinding at 70°C.

Figure 12 shows the characterization of the Cu–Al–Ni component after solvent debinding at 60 °C for 8 h. As shown in Figure 12(a, b), pore features were observed between the powder particles, indicating the development of open porosity after the partial extraction of soluble PW and SA binders. These pore features may provide diffusion pathways for the dissolved binder solution during solvent debinding and for the subsequent removal of the remaining HDPE backbone polymer during thermal debinding. The predominantly spherical to near-spherical morphology of the gas-atomized powders may have reduced severe mechanical interlocking and supported a more regular interparticle arrangement during molding. However, the broad particle size distribution and high powder loading could still produce local variations in pore size and diffusion path length; therefore, the solvent debinding temperature and time were controlled to avoid rapid binder extraction and defect formation. In particular, the relatively coarse Cu particles may have contributed to larger interparticle spaces and facilitated local solvent access, whereas the coexistence of finer Al and Ni particles could reduce local pore size and increase diffusion-path tortuosity in some regions. No visible cracks or delamination were observed on the surface image or in the examined cross-sectional region, suggesting that solvent debinding at 60 °C preserved the structural integrity of the component at the observed scale. The EDS layered map and elemental maps in Figure 12(di) showed a relatively uniform distribution of Cu, Al, and Ni without obvious large-scale elemental segregation. This indicates that the elemental powder distribution achieved during mixing and injection molding was reasonably retained after solvent debinding, although complete chemical homogenization is not expected at this stage.

Figure 12
Characterization of the solvent debound part: (a, b) surface and cross-sectional FESEM images, respectively; (c) cross-sectional BSE image; (d) EDS layered map; and elemental maps of (e) Cu, (f) Al, (g) Ni, (h) C, and (i) O.

TGA was carried out at temperatures between 30°C and 600°C on the green Cu-Al-Ni components before and after solvent debinding. Figure 13 displays these results. Two distinct phases were depicted by the TGA curve of the green part before solvent debinding, as shown in Figure 13a. The initial phase of binder decomposition took place between 180°C and 340°C. Considering the elimination of 69.5% of PW and SA binders from the specimen, and by comparing Figure 13a with 13b, we inferred that the specimen had the remaining 30.5% of PW and SA binders within the decomposition range of 180°C-340°C. The second decomposition phase, which encompassed the temperature ranging from 340°C to 510°C in both TGA curves of the green parts prior to and following solvent debinding, was related to the thermal debinding and expulsion of the HDPE binder from the green body. Since the solvent-debinding percentage was calculated by normalizing the measured mass loss to the theoretical PW + SA fraction, the 69.5% removal obtained at 60 °C represents the apparent removal of the soluble binder fraction rather than the mass loss of the entire feedstock. Accordingly, the TGA comparison in Figure 13 was used to support the interpretation of residual soluble binder and HDPE decomposition behavior after solvent debinding.

Figure 13
TGA of green part: (a) before and (b) after solvent debinding at 60°C.

Thermal debinding was carried out based on TGA evaluation, as illustrated in Figure 13. In this study, the thermal debinding process manifested in two discrete stages to facilitate the elimination of residual binder constituents. In the beginning, as the temperature increased, the remaining PW and SA binders degraded and evaporated, as evaporation functioned as the prevailing mechanism for their elimination. In the succeeding stage, HDPE degraded, shifting to the surface where it disintegrated into gaseous products. Such gaseous emissions were eventually drawn out of the split furnace via argon gas. The implementation of low heating rates, initially 0.1°C/min and eventually increasing to 0.3°C/min, was beneficial throughout the thermal debinding process to prevent the development of cracks and other flaws in components. An increased heating rate can precipitate the buildup of internal pressure and the development of thermal gradients within the injected components. These internal disparities, consequently, have the potential to engender defects such as blistering and cracking within the injected structures³⁷. Figure 14 exhibits the FESEM image of the thermal debound Cu-Al-Ni component, revealing that nearly all of the binders had been effectively eradicated. Moreover, no discernible cracks or other defects were observed, which implied the effectiveness of the implemented thermal debinding procedure. As illustrated in Figure 15, the TGA curves confirmed the comprehensive eradication of the binder system from the component subsequent to the thermal debinding process.

Figure 14
FESEM image of Cu-Al-Ni sample after thermal debinding at 510°C for 4 h.

Figure 15
TGA of solvent debound part: (a) before and (b) after thermal debinding at 510°C.

Table 4 shows the carbon content of the initial raw Cu–Al–Ni powder and the thermally debound part measured by combustion analysis. The initial raw powder contained 0.0215 wt.% carbon, while the thermally debound part contained 0.0302 wt.% carbon. The carbon content after thermal debinding remained very low and close to the baseline value of the starting powder, with only a small absolute difference of 0.0087 wt.%. This indicates that only a limited amount of carbonaceous residue remained after thermal debinding. Therefore, the combustion carbon analysis supports the TGA results by showing that the carbon-containing binder constituents were largely removed during the two-stage debinding process. However, since the thermally debound value was slightly higher than the raw powder baseline, the result should be interpreted as evidence of low residual carbon rather than complete elimination of binder-derived carbon. Moreover, because the powder system contains elemental Al, possible oxidation-related mass gain during heating may partially offset binder-removal mass loss. Therefore, the mass-loss values obtained during debinding should be interpreted as apparent/net mass changes rather than absolute binder-removal values.

Table 4
Results of combustion test for carbon analysis.

Sample	Carbon content (wt.%)
Initial raw Cu-Al-Ni powder	0.0215
Thermally debound part	0.0302

3.6. Sintering

Figure 16 shows the macroscopic appearance of the Cu–Al–Ni component after sintering. The component retained its overall geometry and did not show visible macroscopic cracking, fragmentation, or severe distortion, indicating that the preceding debinding steps provided sufficient structural integrity for subsequent sintering. The sintered component reached a relative density of 91.3%, confirming partial consolidation of the powder structure. However, this density is still lower than that typically expected for fully optimized PIM components, and residual porosity may remain. Therefore, the present result should be interpreted as evidence of macroscopic shape retention and partial densification rather than complete defect-free consolidation at the microstructural level. Further optimization of sintering parameters, together with detailed microstructural, phase-transformation, and functional shape memory evaluations, will be required in future work.

Figure 16
Photograph of sintered Cu-Al-Ni component.

4. Conclusion

The objective of this experiment was to carry out a systematic binder removal process to obtain defect-free debound Cu-Al-Ni components. Contingent upon the aforementioned experimental results and detailed discussions, the following definitive conclusions were drawn:

• Cu-Al-Ni powder mixture was precisely blended with PW, SA, and HDPE binders to generate the feedstock, with the attainment of homogeneity confirmed when the torque values stabilized after an initial increase, indicating the achievement of a well-mixed and optimal material system, which was subsequently used to prepare defect-free green Cu-Al-Ni components.

• A robust solvent debinding framework was established, in which diffusion-controlled extraction of soluble PW and SA binders proceeded uniformly throughout the green components without causing structural or dimensional failures, thereby preserving their integrity and providing a stable foundation for the subsequent thermal removal of the backbone HDPE binder.

• The two-stage thermal debinding strategy, guided by TGA analysis and controlled low heating rates, supported substantial removal of the remaining binder constituents without visible cracking in the examined regions. The debound components retained sufficient integrity for subsequent sintering, although further sintering optimization and microstructural characterization are required to improve densification and confirm defect-free consolidation.

5. Acknowledgments

The authors would like to thank Universiti Kebangsaan Malaysia for the financial support under the grant DIP-2024-020.

6. Data Availability

The data supporting the findings of the current study are available from the corresponding author upon reasonable request.

7. References

• ¹ Zare M, Ketabchi M. Effect of chromium element on transformation, mechanical and corrosion behavior of thermomechanically induced Cu-Al-Ni shape-memory alloys. J Therm Anal Calorim. 2016;127:2113-23.
• ² Shukla U, Garg K. Journey of smart material from composite to shape memory alloy (SMA), characterization and their applications-A review. Smart Mater Med. 2023;4:227-42.
• ³ Rashidi K, Sulong AB, Muhamad N, Fayyaz A, Foudzi FM, Basir A. Martensitic transformation characteristics, mechanical properties and damping behavior of Cu-Al-Ni shape memory alloys: A review of their modifications and improvements. J Mater Res Technol. 2024;29:2732-49.
• ⁴ Jani JM, Leary M, Subic A, Gibson MA. A review of shape memory alloy research, applications and opportunities. Mater Des. 2014;56:1078-113.
• ⁵ Saud SN, Hamzah E, Abubakar T, Hosseinian R. A Review on influence of alloying elements on the microstructure and mechanical properties of Cu-Al-Ni shape memory alloys. J Teknol. 2013;64:51-6.
• ⁶ Suresh N, Ramamurty U. Aging response and its effect on the functional properties of Cu-Al-Ni shape memory alloys. J Alloys Compd. 2008;449(1-2):113-8.
• ⁷ Agrawal A, Dube RK. Methods of fabricating Cu-Al-Ni shape memory alloys. J Alloys Compd. 2018;750:235-47.
• ⁸ Tang SM, Chung CY, Liu WG. Preparation of Cu-Al-Ni-based shape memory alloys by mechanical alloying and powder metallurgy method. J Mater Process Technol. 1997;63(1-3):307-12.
• ⁹ German RM, Bose A. Injection molding of metals and ceramics. Princeton (NJ): Metal Powder Industries Federation; 1997.
• ¹⁰ Cicek B, Sun Y, Turen Y, Ahlatci H. Applicability of different powder and polymer recipes in a new design powder injection molding system. J Polym Eng. 2021;41(4):299-309.
• ¹¹ Cicek B, Sun Y, Turen Y, Ahlatci H. Investigation of microstructural evolution of gas-assisted metal injection molded and sintered Mg-0.5Ca alloy. Sci Sinter. 2022;54(1):25-37.
• ¹² Cicek B, Sun Y, Turen Y, Ahlatci H. Material production and the effect of the system on biocompatibility in the modified metal injection method. Arch Metall Mater. 2023;68(1):195-202.
• ¹³ Hausnerova B. Binder systems for powder injection molding: A review. Adv Mater Process. 2017;2(12):761-8.
• ¹⁴ Basir A, Sulong AB, Muhamad N, Juri AZ, Jamadon NH, Foudzi FM, et al. Influence of SS316L nanoparticles on the sintered properties of two-component micro-powder injection moulded bimodal SS316L/zirconia bi-materials. Materials (Basel). 2024;17(22):5536.
• ¹⁵ Basir A, Sulong AB, Jamadon NH, Muhamad N, Emeka UB. Process Parameters used in macro/micro powder injection molding: an overview. Met Mater Int. 2021;27:2023-45.
• ¹⁶ Enneti RK, Park SJ, German RM, Atre SV. Review: thermal debinding process in particulate materials processing. Mater Manuf Process. 2012;27(2):103-18.
• ¹⁷ Samanta SK, Chattopadhyay H, Godkhindi MM. Thermo-physical characterization of binder and feedstock for single and multiphase flow of PIM 316L feedstock. J Mater Process Technol. 2011;211(12):2114-22.
• ¹⁸ Tseng WJ, Hsu CK. Cracking defect and porosity evolution during thermal debinding in ceramic injection moldings. Ceram Int. 1999;25(5):461-6.
• ¹⁹ Côté R, Azzouni M, Demers V. Impact of binder constituents on the moldability of titanium-based feedstocks used in low-pressure powder injection molding. Powder Technol. 2021;381:255-68.
• ²⁰ López-Ferreño I, Gómez-Cortés JF, Breczewski T, Ruiz-Larrea I, Nó ML, San Juan JM. High-temperature shape memory alloys based on the Cu-Al-Ni system: design and thermomechanical characterization. J Mater Res Technol. 2020;9(5):9972-84.
• ²¹ Mannschatz A, Müller N, Moritz T, Lenk R. Influence of powder morphology on properties of ceramic injection moulding feedstocks. J Eur Ceram Soc. 2011;31:2551-8.
• ²² Reddy JJ, Vijayakumar M, Tallapragada RMR, Ramakrishnan P. Loading of solids in a liquid medium: determination of CBVC by torque rheometry. J Eur Ceram Soc. 1996;16(5):567-74.
• ²³ Zaky MT, Soliman FS, Farag AS. Influence of paraffin wax characteristics on the formulation of wax-based binders and their debinding from green molded parts using two comparative techniques. J Mater Process Technol. 1996;16(18-19):567-74.
• ²⁴ Kong X, Barriere T, Gelin JC. Determination of critical and optimal powder loadings for 316L fine stainless steel feedstocks for micro-powder injection molding. J Mater Process Technol. 2012;212(11):2173-82.
• ²⁵ Sotomayor ME, Várez A, Levenfeld B. Influence of powder particle size distribution on rheological properties of 316L powder injection moulding feedstocks. Powder Technol. 2010;200(1-2):30-6.
• ²⁶ Hausnerova B. Rheological characterization of powder injection molding compounds. Polimery. 2010;55(1):3-11.
• ²⁷ Thavanayagam G, Pickering KL, Swan JE, Cao P. Analysis of rheological behaviour of titanium feedstocks formulated with a water-soluble binder system for powder injection moulding. Powder Technol. 2015;269:227-32.
• ²⁸ Jabir SM, Noorsyakirah A, Afian OM, Nurazilah MZ, Aswad MA, Afiq NHM, et al. Analysis of the rheological behavior of copper metal injection molding (MIM) feedstock. Procedia Chem. 2016;19:148-52.
• ²⁹ Hidalgo J, Jiménez-Morales A, Barriere T, Gelin JC, Torralba JM. Capillary rheology studies of INVAR 36 feedstocks for powder injection moulding. Powder Technol. 2015;273:1-7.
• ³⁰ Askari A, Momeni V. Rheological investigation and injection optimization of Fe–2Ni–2Cu feedstock for metal injection molding process. Mater Chem Phys. 2021;271:124926.
• ³¹ Pieczonka T. Disruption of an alumina layer during sintering of aluminium in nitrogen. Arch Metall Mater. 2017;62(2B):987-92.
• ³² German RM. Powder injection molding. Princeton (NJ): Metal Powder Industries Federation; 1990.
• ³³ Hwang KS, Hsieh YM. Comparative study of pore structure evolution during solvent and thermal debinding of powder injection molded parts. Metall Mater Trans, A Phys Metall Mater Sci. 1996;27:245-53.
• ³⁴ Sharmin K, Schoegl I. Optimization of binder removal for ceramic microfabrication via polymer co-extrusion. Ceram Int. 2014;40(3):3939-46.
• ³⁵ Omar MA, Ibrahim R, Sidik MI, Mustapha M, Mohamad M. Rapid debinding of 316L stainless steel injection moulded component. J Mater Process Technol. 2003;140(1-3):397-400.
• ³⁶ Chen G, Cao P, Wen G, Edmonds N. Debinding behaviour of a water soluble PEG/PMMA binder for Ti metal injection moulding. Mater Chem Phys. 2013;139(2-3):557-65.
• ³⁷ Thomas-Vielma P, Cervera A, Levenfeld B, Varez A. Production of alumina parts by powder injection molding with a binder system based on high density polyethylene. J Eur Ceram Soc. 2008;28(4):763-71.

Edited by

• Associate Editor:
  Aloisio Klein.

• Editor-in-Chief:
  Luiz Antonio Pessan.

Publication Dates

• Publication in this collection
  24 Aug 2026
• Date of issue
  2026

History

• Received
  27 Jan 2026
• Reviewed
  25 May 2026
• Accepted
  08 July 2026