1 Introduction

Additive manufacturing (AM), formerly known as 3D printing and rapid prototyping, provides cost-effective and rapid production at a small and medium scales and enables fabrication of complex structures such as lattices [1, 2]. AM is classified into seven different categories comprising: Material Extrusion (MEX), Directed Energy Deposition (DED), Vat Photopolymerization (VPP), Material Jetting (MJT), Sheet Lamination (SHL), Binder Jetting (BJT) and Powder Bed Fusion (PBF) [3]. PBF has high compatibility and flexibility to process different materials such as polymers, composites, metals, and ceramics. A well-known limitation of the PBF process for large-scale production is the low speed of production when compared to conventional manufacturing methods such as machining, casting and injection moulding. To improve the PBF process and make it competitive with conventional manufacturing, High-Speed Sintering (HSS) was developed and introduced by companies such as Xaar and Voxeljet [3, 4]. Hewlett-Packard (HP) further developed the HSS process by adding thermal agents to improve process efficiency, with the technology known as Multi-Jet Fusion (MJF).

MJF was proven to be more cost-effective than Selective Laser Sintering (SLS) for mass production [5]. Some aspects of MJF have been investigated to characterise the process window comprising build orientation, thermal features and printed part quality. Soskic et al. [6] investigated the cost of polymeric production of MJF and compared the results with powder bed fusion in the semi-solid state. The authors reported that although MJF uses fusing and detailing agents which are costly, due to the substantially low material refreshment rate, the overall material costs of MJF are much lower than SLS. Another key factor for the lower cost of MJF is the use of a dedicated processing station for cooling while SLS cooling occurs within the production machine, extending time between consecutive builds. Emerging MJF technology brings new capabilities in this AM process with increased part quality in comparison with other PBF processes in a rapid and cost-effective way. Different aspects of MJF have been investigated and reported in the literature. X-ray micro-computed tomography (MCT) enables non-destructive investigation of the surface topology and inner structure of materials. It characterises deep enclosed porosity, complex part geometry and subtle changes to topology post-experiment, where porosity can be considered the most important indicator of part quality [7]. Pandelidi et al. [8] reported that the morphology, distribution and size of the pores determine the mechanical properties of MJF components. The elongated pores in MJF can be related to poor layer adhesion which is driven by a lack of energy or fusing agent between particles. Pores with irregular shape are formed due to insufficient packing density or layer compaction, and circular pores are formed due to evaporation and bubble bursting on the specimen surface. Lee et al. [9] reported that the build angle of 45° showed consistency in mechanical properties of MJF samples. This is related to compensation of the negative effect of poor intra-layer powder packing in the horizontal surface and the direction of the inter-layer interface against powder loading direction. MJF was also successfully used to produce aramid fibre reinforced PA12 as a mechanism to improve mechanical performance and flame retardancy for aerospace [10] and biocompatible materials in medical applications [11].

Laser Flash Analysis (LFA or Hyperflash) is a reliable and robust technique to measure thermophysical properties for a range of polymers, including thermal diffusivity, thermal conductivity and heat capacity [12, 13]. LFA operates by exposing the sample to a high-energy pulse of light and measuring the resulting temperature rise, allowing light absorption to be inferred. While LFA has been successfully used to characterise metals [14], polymers [15] and composites [16], no research paper yet reports analysis of Multi-Jet Fusion (MJF) printed polypropylene (PP) while linking observed mechanical properties with absorption. LFA offers direct measurement of diffusivity with reduced uncertainty potentially associated with other methods. For example, Patti et al. [17] reported large variation in thermal conductivity data for PP/PP composites manufactured via non-additive process due to limitations in the measurement method. Accurate determination of thermophysical properties of MJF printed parts is challenging [18] due to extremely low diffusivity of polypropylene polymers. Tavman et al. [19] reported extremely low diffusivity (0.96 × 10 − 7 m2/S) for PP polymers which indicated that current LFA methods applied to other materials is not fully applicable for PP. Li et al. [20] reported similar thermal diffusivity value (1.68 × 10 − 7 m2/S) for flax fiber reinforced HDPE composite in addition to thermal conductivity measurement. The study showed that diffusivity values decrease with increasing temperature. Weidenfeller et al. [21] investigated crystallinity, thermal diffusivity and electrical conductivity of carbon black/PA46 composites. The study reported four probe techniques to measure diffusivity at various temperatures and found that diffusivity decreases with temperature but increases with carbon black content.

The absorptivity of the polymer powder in MJF is manipulated using fusing agent; it is reported that fusing agents increase the absorptivity of materials when irradiated by IR light. In SLS, a CO2 laser with a wavelength of 10.6 µm is used which leads to a low absorption ratio for both metals and polymeric materials. It is reported that [22] for Polyethylene Terephthalate (PET) and PP, the absorption of the laser in SLS is approximately 20%. This value for MJF is around 80% [23, 24] which is mainly attributed to high absorptivity of the powder and fusing agent. To reduce porosity and improve mechanical properties of MJF parts, Hot Isostatic Pressing (HIP) was suggested by Abbott et al. [25]. Results showed that geometry, total volume and location of the components within the build chamber can influence porosity in MJF process. The author reported that applying pressure improved density and mechanical properties of both SLS and MJF parts. Ultimate Tensile Strength (UTS) and Young’s Modulus of MJF parts improved by 4 and 11%, respectively, while elongation at break of MJF components was reduced by 35%. On the other hand, heat treatment exhibited a reduction in mechanical properties, attributed to chain scission and a decrease in in molecular weight observed by Duddleston et al. [26] and Touris et al. [27]. However, no study was conducted around predicting part performance through absorption or reflection indicators or development of methods to quantify the same.

Based on the reviewed literature, no comprehensive results on variation of mechanical properties with absorptivity and reflectivity of light in the MJF process were available. Thus, in this work, the absorptivity of MJF automotive grade PP structure will be conducted through laser flash and porosity correction method. The absorptivity of the material drives the process temperature, crystallinity and mechanical properties. This directly drives the quality of the printed structures. The objective is to identify trends when samples are manufactured with different build location and orientation. The influence of build height and orientation on tensile properties, crystallinity, point temperature and initial porosity of the produced parts are investigated and discussed in detail. Porosity correction after micro-CT analysis will be discussed to improve contrast during the scanning process. Finally, a correlation between corrected porosity and LFA-measured parameters such as half-life and heat loss will be conducted to understand the absorption phenomena within MJF samples and quantify the heat absorption capacity of fusion agent.

2 Materials and process

2.1 Multi-Jet Fusion (MJF)

BASF’s Ultrasint® PP natural 01 powder was used for this study due to excellent plasticity, high elongation, low moisture absorption and high durability. According to the manufacturer’s datasheet, it has a particle size of 60–70 μm, melting point of 140 °C and powder bulk density of 0.33 g/cm3. Proprietary fusing and detailing agents were used to manufacture MJF tensile test coupons in a HP Jet Fusion 3D 5200 printer with a build volume of 380 × 284 × 380 and resolution of 1200 dpi. Furthermore, the specimens were allowed to cool in the powder bed over 16 h, thereby allowing even heat dissipation while the test specimens remained supported by the surrounding powder.

2.2 Compression moulding

Five circular disks of 20 mm diameter and 4 mm thickness were manufactured from BASF’s Ultrasint® PP natural 01 powder material using a compression moulding process. Specimens were produced at 160 °C with 10 kN pressure and had a residence time of 6 min. The pressure was then alleviated and cooled for another 10 min. All samples were dried at 80 °C for 12 h before testing. Manufactured disks were then analysed using a Hyperflash LFA 467.

3 Experimental methods

3.1 Tensile properties

Tensile tests on MJF PP coupons (Fig. 1), printed on a HP Jet Fusion 3D 5200 printer, were performed following ASTM D638. Tests were performed on an Instron 5569 using a 30 kN load cell and a consistent crosshead speed of 20 mm/min and a gauge length of 10 mm for all samples. Minimum 5 samples from each batch were tested with their means and standard deviations reported.

Fig. 1
figure 1

MJF tensile coupons produced based on ASTM D638

Full size image

Specimens were printed at different build height and build orientations as shown in Fig. 2. Throughout this paper, specimens printed in vertical orientation are referred to as Z; those printed in flat orientation are referred to as XY4, XY8, and XY12 where the number indicates the build height in the chamber.

Fig. 2
figure 2

MJF specimens printed in different build locations and build orientations

Full size image

3.2 Crystallinity: differential scanning calorimetry (DSC)

The thermal properties of MJF PP samples were studied by DSC (Discovery DSC 250, TA Instruments). Samples were heated under a 50 ml/min nitrogen flow, from 25 to 400 °C before being cooled back to 25 °C. Both heating and cooling were performed at a rate of 10 °C/min. All characteristic temperatures and associated enthalpies were calculated from the endothermic and exothermic peaks, using the TA Instruments Universal Analysis software. The second heating run was used to calculate the degree of crystallinity (χc) based on the melting enthalpy of a totally crystalline PP material which has been reported to be 207 J/g [28].

3.3 Hyperflash LFA

3.3.1 LFA for coupons

Thermal diffusivity was measured using an LFA 467 series HyperFlash™, Netzsch test apparatus employing the manufacturers HyperFlash software using laser flash methodology. ASTM E1461/ ASTM E2585 — 09(2015) — Standard Test Method for Thermal Diffusivity by the Flash Method was followed. The sample holders used in the apparatus were suitable for square samples with a 10 mm side length. The printed samples were slightly larger than the sample holder, the front, back and sides of the printed sample were coated with graphite 3 times to ensure an optimum emission/absorption capacity prior to analysis. A comparative technique was employed to calculate the specific heat capacity of PP by comparison with a calibration standard reference sample (Pyroceram) with a known specific heat. Once diffusivity was measured and heat capacity was calculated, thermal conductivity was evaluated based on Eq. 1:

$$k=\propto \rho {C}_{p}$$
(1)

where \(\propto =\) thermal diffusivity, \(\rho =\) density and \({C}_{p}=\) specific heat capacity. To that end, the density of PP samples were determined by making use of the Archimedes’ Principle as previously reported by Biran et al. [15].

3.3.2 LFA for powders

PP powders were tested using a pressure mould system which comes with sample holder plate, lower support plate, upper support plate, wrench and fixing screw as shown in Fig. 3. Powder samples are placed inside the sample holder between plates, and a wrench is used to tighten the screw as per desired thickness.

Fig. 3
figure 3

Pressure mould assembly for powder diffusivity measurement in LFA [29]

Full size image

Due to extremely low diffusivity of PP powder [19], it was challenging to characterise using standard LFA measurement parameters. Therefore, a total of 73 single-shot trials (not shown) were conducted to optimise LFA parameters considering a wide range of input pulse width and duration. Interestingly, both PP powder and MJF PP samples needed extremely high pulse width and pulse duration in order to diffuse the light properly through the thickness of samples. Required pulse width was a little lower for powder once a repeatable S shape signal versus time curve was achieved, the next step was to run at different temperatures while comparing the diffusivity values with reference database.

3.4 Micro-CT

Micro-CT scans were performed on a Bruker Skyscan 1275 operating at 20 kV, 100 µA with no filter in place. The object was placed on a brass sample stage and positioned to optimise image pixel size, with height adjusted to image the thickest part of the sample. Samples were scanned with the settings: 0.4° rotation step, 360° scanning, 4 frames averaging, total of 901 shadow projections (raw images). Camera settings were 1944 × 1536 resolution and 67 ms exposure. After scanning, NRecon program was used to calculate cross sections from the raw images with the software determined post processing corrections including smoothing, misalignment compensation, ring artifacts reduction and beam-hardening correction (constant at 30% across all samples). To analyse the volume inside the object, a rectangular shaped region-of-interest (ROI) was set as shown in Fig. 4 (left). The ROI was created over a 2D slice and applied to all slices, making adjustments where necessary. Each slice was checked in CTvox software to make sure there were no inconsistencies in the sample or reconstruction process. Finally, a 3D model based on reconstructed images was created in CTvox software (Fig. 5, right). Finally, CTan software was used to binarize images in order to segregate areas of different density. Cross sections are represented as a binary image of material (1) and air (0). The total porosity and pore volume were calculated from the binarized images.

Fig. 4
figure 4

(Left) ROI and thresholding range correction in MJF printed coupons; (right) 3D model based on reconstructed images created in CTvox software

Full size image
Fig. 5
figure 5

Thermophysical property mapping with respect to temperature for MJF printed coupons. From left top (XY4), left bottom (XY12), right top (XY8) and right bottom (Z) compared to reference PEKK disk

Full size image

3.5 Porosity correction

Initial porosity correction was conducted based on horizontal intensity profile (HIP) values. The target was to achieve a 40–50% HIP range for vertically positioned samples; however, poor contrast (~ 70–80%) was an issue with vertically positioned MJF samples. Therefore, further porosity correction was conducted. In addition to sample orientation within micro-CT chamber, porosity correction method also included resetting the thresholding range based on reselection of ROI.

3.6 Porosity validation

Porosity validation was completed following the measured half-time values collected at each input temperature from the exported data file for each MJF printed samples. Data points were taken manually for each point based on the average of three-shot measurement by the LFA machine.

4 Results

4.1 Effect of build height and build orientation

The effect of build height and build orientation based on tensile, crystallinity, thermophysical properties and porosity of MJF printed coupons was investigated (Table 1).

Table 1 LFA process optimisation PP powder, PP sheet and printed MJF coupons
Full size table

4.1.1 Tensile properties

Table 2 shows tensile properties of MJF printed coupons at 10 mm gauge length. While there is no significant difference in tensile modulus between top (XY4) and bottom (XY12) location of build chamber, but there was a 50% reduction in percentage strain at the break between bottom flat (XY12) and vertical (Z) samples. This is in line with the energy at break value where Z samples showed a 35% reduction in values compared to XY12, which confirmed that Z samples were more brittle than horizontally oriented coupons. Tensile results are in conformity with the Ultrasint datasheet provided by BASF as well as the study by Jafka et al. [30] reported very similar tensile strength and relatively lower Young’s modulus in MJF PP samples when compared to the current study irrespective of build orientation and build location [9, 31]. The study also reported similar isotropic trend in MJF printed PA11 (35) and PA12 (33) polymers. Both studies reported significant decrease in % elongation for vertically oriented (Z) specimens. It is worth mentioning that MJF nylon samples have relatively lower elastic modulus than the MJF PP.

Table 2 Tensile properties of printed MJF coupons at 10 mm
Full size table

In the current study, high standard deviation was observed for Z coupons, which could be because of carbon black in the process that creates a continuous transmission of energy toward the previous layers. However, since the printing surface in the Z direction is smaller than X and Y, higher capillary forces tend to mix the area consisting of fusing agents and detailing agents which increases the presence of defects and variations in mechanical properties [32].

4.1.2 Crystallinity

Crystallinity analysis was based on thermograms measured by differential scanning calorimetry (DSC). Small sections of approximately 10 mm2 were taken from an unloaded portion of the tensile bar before tensile testing. Samples were flattened in the pan to ensure maximum contact between cut polymers and DSC pan. For crystallinity analysis, enthalpy values were calculated from the melting curve. Table 3 below shows powders and coupons printed in Z direction have lower crystallinity compared to horizontal (XY) coupons, which is the possible reason for low percentage strain at break in vertical (Z) coupons.

Table 3 Percent crystallinity, melting temperature and enthalpy of MJF coupons measured by DSC
Full size table

4.1.3 Point thermophysical properties

Point analysis was conducted at 25 °C for all 6 samples including powders and printed coupons. Table 4 shows powder samples possessed had very low diffusivity compared to compression-moulded sheets and MJF coupons. The presence of carbon black in MJF coupons provided a significant increase in diffusivity, conductivity and heat capacity. This observation could be due to inherent heat absorption capacity of carbon black materials leading [33]. Previous studies [19, 20] also reported extremely low diffusivity (0.96 × 10−7 m2/s) for PP powders as well as for PP/HDPE composites. Significantly higher thermal diffusivity and thermal conductivity were observed for specimens printed in the “Z” direction samples compared to XY samples possibly due to high levels of porosity.

Table 4 Point thermophysical Properties of MJF square disks compared with reference Pyroceram samples
Full size table

4.1.4 LFA temperature profile

The LFA temperature profile was completed for all MJF coupons between 25 and 130 °C. Stratasys ANTERO PEKK was considered as a benchmark due to excellent agreement of thermophysical properties with the glass transition temperature reported in our previous study. Kafi et al. [34] analysed a range of additively manufactured ANTERO PEKK where thermophysical properties were found to decrease around glass transition temperature. Given that printed PEKK also has low diffusivity like polypropylene, it was used as a reference sample for heat capacity analysis after diffusivity measurement in LFA. Figure 6 reported thermal diffusivity, conductivity and heat capacity values with temperature. It seems that diffusivity decreases with increase in temperature for all MJF coupons like the reference PEKK. Diffusivity values are also in line with the carbon black/PA46 work reported by Weidenfeller et al. [21] where diffusivity decreases with temperature but increased with carbon black content. Similarly, Li et al. [20] reported diffusivity values decrease with an increase in temperature for PP/Flax bio-composite materials.

Fig. 6
figure 6

Variation of half-time of PP powders, PP sheet and MJF printed samples measured in Hyperflash LFA 467 with temperature

Full size image

4.1.5 Initial porosity

Table 5 shows initial porosity values determined for MJF printed coupons in different build height and build orientations. Porosity values are higher for coupons printed in flat orientation compared with vertically oriented Z. Results presented in Table 5 also show lower porosity values for Z compared to XY-orientated samples. Considering the high diffusivity observed for XY12 and Z coupons (Table 4), porosity was expected to be low for both Z and XY12. However, a large distribution of porosity (~ 2–10%) was observed in both Z and XY samples. Lee et al. [8] reported that the highest volume of porosity was observed at the “Bottom” surface and the smallest pores were found in the “Middle” of the specimen. The authors reported no relation between the porosity and building height and the packing difficulties of intra-layer leads to more porosity in XY direction. However, the paper did not study diffusivity; therefore, its conclusion was based solely on porosity. Liu [35] compounded 124, 226 and 271 µm glass fibre into PA12 powder and found longer fibre led to high porosity and anisotropy within printed MJF. The study, however, did not consider the effect of build orientation or build height like the current polypropylene investigation.

Table 5 Initial porosity values of different MJF coupons
Full size table

Several studies (24,42–44) reported similar porosity level (less than 1%) for MJF printed samples albeit for PA12. Values reported in the current study align with those investigations when porosity values were corrected. Interestingly, [36] reported low porosity values for Z-oriented samples which could be due to different mechanisms of particle fusion in PA12 polymer and laser-assisted fusion instead of carbon blacks used to fuse PP in MJF process.

5 Validation and verification

5.1 Porosity correction method

Initial porosity based on vertical positioning of samples in micro-CT chamber showed large variations in porosity ranging between ~ 5 and 10%. Therefore, repositioning of the sample was required from vertical to horizontal orientation which reduced the porosity values and most importantly the uncertainty associated with them. Table 6 shows higher contrast or low HIP values are achievable for both vertical and horizontally positioned samples. Interestingly, porosity values were only reported for PA11 on vertically positioned samples within the micro-CT chamber by Kong et al. [8]. Given PA11 was not mixed with carbon black, it was not a problem to achieve desired contrast for MJF PA11 coupons in the same micro-CT chamber. The study reported porosities of 0.94% and 0.69% for recycled parts using a mixture of 70–30 recycled to fresh PA11 powder, respectively. Porosity was lower when compared to PA12 SLS parts, i.e., 4–6% [37]. The authors reported a maximum pore size of 700 µm when the recycled powder was used whereas the maximum pore size of 440 µm was reported for virgin powder. Porosity values (after correction) reported in Table 6 are consistent with the study published by Kong et al. [8]. In the current study, significant reduction in porosity are clearly observed and values were distinguishable between MJF PP samples printed in XY and Z orientations when HIP reduced from 70 to 40% (Table 7).

Table 6 Improving HIP/contrast of MJF coupons through sample repositioning in micro-CT chamber
Full size table
Table 7 Effect of horizontal correction method on %porosity and pore volumes, mm3
Full size table

5.2 Porosity validation

5.2.1 LFA half-time

Figure 7 shows the variation of half-time with temperature for PP powders, PP sheet and MJF printed samples measured using Hyperflash LFA 467. It appears the vertically printed samples had a lower half-time value compared to horizontal (XY). Vertically oriented Z samples took less time for Xenon light (thermal energy) to pass through the material, meaning a low temperature gradient or a high backside temperature; therefore, high diffusivity values were observed. This could be due to higher porosity values as reported in Table 8. Samples located on top of the build chamber (XY4) also showed lower half-time values compared to the bottom (XY12), which is also in line with the high porosity data (Table 4). The time required to reach 50% of the maximum temperature during LFA study was highest for XY12 samples. It was most likely due to low porosity in XY12 as the light was mostly absorbed by the polymer at this certain position in the build chamber. Due to high absorption of light, it was expected that XY12 will release more heat compared to Z-oriented samples. Results are in line with the reference sample (Pyroceram) which has the lowest half-time and porosity. The heat loss study in the following section will help to establish the link between absorption and porosity.

Fig. 7
figure 7

Variation of heat loss in MJF printed samples

Full size image
Table 8 Important results summary for discussion
Full size table

5.2.2 LFA heat loss

The occurrence of two-dimensional heat transport through the test piece during LFA xenon light flash exported as heat loss value for each MJF samples. According to previous studies [38, 39], less dense samples usually have low diffusivity or low backside temperature and hence superior insulative properties. As porosity values decreased with build height, it can be assumed that half-life values will be lowest for the bottom-most sample, XY 12. However, in our case, half-life value is increasing with the build height which is most likely due to heat loss phenomenon occurring through convection or radiation.

6 Discussion

Thermophysical properties of MJF samples were determined by measuring half-time value through LFA467 Hyperflash. Before considering this data, it is important to check that the comparison is valid as there are no previous literature which reported half-time values. It is worth mentioning that most studies considered thermophysical properties without validating measured or calculated values, which were not influenced by factors such as heat loss, sample thickness etc. The highest half-time was observed for XY12 samples printed at the bottom of the build chamber compared to samples printed on top of the bed (XY4) or Z samples. This observation is consistent with the observation that higher light absorption in a 3D-printed polymer can indicate lower porosity. Light absorption is related to the material’s ability to absorb and not transmit or reflect light. In the context of 3D-printed polymers, higher light absorption is typically associated with denser and less porous structures.

The observations for light absorption are not consistent with the diffusivity of both XY12 and Z specimens reported in Table 8. It is mainly because half-time is inversely proportional to the diffusivity according to Parker’s equation. Initially, it appears that variation in thickness of the sample (3.5–4.1 mm) printed in different build locations and build orientation is the responsible factor. Pyroceram ceramic samples possessed the highest diffusivity value even though its half-time was significantly lower than all MJF samples. This is also true for thickness variations within MJF printed samples as Z samples had lower thickness than XY-printed samples but higher diffusivity. Therefore, some variation in thickness within MJF samples was not an obstacle to compare different MJF printed samples.

Subsequently, the underlying factors which can alter half-time as well as thermal diffusivity of printed MJF coupons were investigated. As the level of porosity within polymeric structures can alter thermal diffusivity, it is important to consider the implications according to fundamental theory. According to Eq. 1, diffusivity is inversely proportional to density whereas, according to the study by Alqantani et al., conductivity should increase with relative density [40]. Several other researchers [41,42,43,44] also conducted thermal characterization of additively manufactured lattice structures and reported a direct correlation between thermal conductivity and relative density. Table 7 shows highest diffusivity values for XY12 and Z samples, which is most likely due to lower porosity values calculated for XY12; however, porosity was found to be directly proportional to thermal diffusivity before correction. Therefore, the porosity analysis method was revised. Further correction of porosity method significantly improved contrast which helped to detect actual pores when build height was varied. However, Z samples showed the highest porosity. This can be linked to lower heat loss phenomenon which indicated Z absorbed less heat (as indicated by low heat loss) during MJF printing process; therefore, high thermal gradient or diffusivity was expected.

Successively, half-time/heat loss phenomenon was linked to absorptivity of MJF printed samples. Meng et al. [45] reviewed absorption behaviour of graphene-based microwave absorbing composite materials. The study showed absorption is proportional to heat loss phenomena. Taking this statement into consideration, MJF printed PP showed higher absorption with build height, and absorption of light decreased for vertically oriented Z samples which was mainly due to high porosity as reported in Table 8. An increase in heat loss (or half-time) with build height can be attributed to reduced porosity and this reasonable correlation was only achievable for XY samples after porosity was corrected. Therefore, diffusion time (or half-time at 50% of the maximum temperature) increased for XY12 compared to XY4 and XY8. When the sample absorbs less heat or release more heat as evidenced from the heat loss results summary (Table 7) for Z samples, it can be predicted that Z samples experienced more reflection than absorption of light. This phenomenon indicated low diffusivity; however, it is not the case in reality, possibly due to other heat incidents such as radiation or convection occurring in addition to typical conduction within samples. On the other hand, diffusivity values are the highest for reference Pyroceram samples due to extremely low half-time and porosity, and so lowest heat loss, contributed to highest diffusivity.

7 Conclusions

This study proposed a new method for the first time to predict the absorptivity of MJF printed structures through laser flash analysis and corrected porosity. The effects of build height and build orientation were investigated based on tensile mechanical testing, crystallinity, thermophysical properties and porosity of test coupons. While there were no significant differences in tensile modulus between samples printed at different build heights within the MJF print chamber, vertical (Z) samples showed 50% reduction in strain at break compared to XY12 (bottom flat). It was mainly due to marginally lower crystallinity and high porosity of Z samples. As thermal diffusivity (or conductivity) should decrease with increase in porosity, porosity without correction did not support the thermophysical property trend described in the literature. Therefore, further revision of porosity method and subsequent correction was conducted to achieve reasonable porosity-build heigh and, porosity-build orientation correlation. The correction of porosity method significantly improved contrast which helped to identify the tiny pores when build height was varied. However, Z samples displayed the highest porosity even though it had higher diffusivity which contradicted with literature and the hypothesis developed for XY samples. This can be linked to lower heat loss phenomenon which indicated specimens printed in the Z direction absorbed less heat during MJF printing process; therefore, a higher thermal gradient or diffusivity was expected than XY flat-oriented samples. Higher absorption predicted from high heat loss indicated that better interlayer bonding can be achieved in polypropylene by using fusion agent in the MJF process. Higher heat loss in Z or bottom XY (12) indicated that samples printed in the bottom build location and flat orientation are more reliable than the ones printed on top and vertically. This study will provide opportunities for MJF simulator and manufacturer if the IR light absorptivity can be predicted from the printed samples to improve quality of the samples.

8 Future Direction

The future work is directed toward optimising the usage of fusing and detailing agents and analysing the effect of variation of the agents on the crystallinity and mechanical properties. This investigation opens up the possibility of better understanding the effect of the agents on the temperature of the process and overall quality of the produced components.