Mechanical Performance of Pultruded and Compression-Molded CFRTP Laminates: A Comparative Study

Abstract

In this work, the mechanical performance of unidirectional thermoplastic laminates produced via a proprietary non-reactive thermoplastic pultrusion system known as the continuous forming machine (CFM) was compared to the mechanical performance of similar laminates produced via compression-molding in a heated platen press. Using commercially available pre-impregnated unidirectional thermoplastic tapes as the material feedstock for both production methods, a comparison of mechanical performance was executed for six separate material systems ranging from commodity-grade (e.g., polypropylene) to aerospace-grade (e.g., low-melt polyaryletherketone) polymer systems. Mechanical performance was evaluated and compared through tensile testing, compression testing, and short beam strength testing in a universal testing machine. The continuous fiber-reinforced thermoplastic (CFRTP) laminates were composed solely of unidirectional plies and were tested in the longitudinal material orientation. Through testing, it was found that the laminates produced on the proprietary thermoplastic pultrusion platform exhibited mechanical performance characteristics equivalent with those of the laminates produced using heated compression-molding. Furthermore, comparisons to values found in the literature were performed, demonstrating the viability of the CFM’s novel manufacturing process to pultrude thermoplastic parts for axially loaded applications.

1. Introduction

Composite materials are an increasingly popular choice for structural components in the aerospace, automotive, and construction industries [1,2,3,4,5]. This popularity is primarily due to the advantages of composites when compared to traditional materials such as steel due to their high strength-to-weight ratios [6,7] and corrosion resistance [8]. In recent years, thermoplastic polymer composites have attracted additional interest by manufacturers and end users compared to traditionally used thermosetting polymers, primarily due to their increased fracture toughness [2] and ability to be reformed, leading to increased sustainability and recyclability [1,3]. Additional advantages of thermoplastic composites to thermosetting composites include their low emissions of volatile organic compounds during processing, ability to be welded for joining or repair mechanisms, and significantly longer shelf life [6,9,10]. As such, continuous fiber-reinforced thermoplastic (CFRTP) laminates are an area of particular interest in current research efforts.

Although CFRTP laminates are of high interest to industry and in the literature, current manufacturing methods have not yet enabled rapid, high-volume throughput of thermoplastic parts. Primary methods for the fabrication of CFRTP laminates consist of compression molding, autoclave molding, filament winding, and pultrusion [8]. These methods often involve the layup of pre-impregnated thermoplastic feedstocks by hand or via automated tape layup (ATL), followed by consolidation via compression or autoclave molding processes. When combined with necessary post-processes such as machining or drilling, the long cycle times of CFRTP production methods are often viewed negatively when compared to the well-matured fabrication methods typically used for thermoset composites. Recent developments in manufacturing techniques have also enabled the evolution and use of automated fiber placement (AFP) equipment; however, the literature shows that achieving high consolidation quality in situ remains an active area of concern for industry, thus far leading to minimal adoption for manufacturing applications [11,12].

Pultrusion of composite parts has been a common method of production for thermoset materials, due to its ability to efficiently produce constant profile parts in a continuous manner with minimal engineer or operator intervention [13,14]. As such, it is a preferred method of composite production in industry for thermoset composite profiles, since it can produce constant-section profiles of nearly infinite lengths in a cost-effective manner. Although proven to be a reliable manufacturing method for continuous profile composite parts, thermoset pultrusion is limited in operational speeds to the rate of the curing reaction for the polymer being utilized [15,16]. Due to the operational speed limitations of thermoset pultrusion, combined with the recyclability and post-production reformability of thermoplastic materials, there has been significant interest in the literature around exploring the viability of thermoplastic pultrusion platforms. However, it has been demonstrated that there are additional challenges associated with thermoplastic pultrusion that are not typically observed in thermoset pultrusion, primarily due to the high processing temperatures and melt viscosities of thermoplastic polymer resins [3,6,17]. This can lead to difficulties in fully impregnating the reinforcement fibers, requiring both high temperatures and high pressures in the thermoplastic pultrusion process.

Methods of thermoplastic pultrusion have been broadly categorized as reactive or non-reactive by Luisier et al. [18]. Reactive thermoplastic pultrusion uses multi-liquid resin systems which polymerize once mixed, behaving like thermoset pultrusion [3]. Non-reactive thermoplastic pultrusion ranges from separate polymer and fiber feedstocks being mixed in the die, to partially mixed feedstocks (typically powder or fiber polymer forms), to pre-consolidated tape (PCT) materials. A complete review of thermoplastic pultrusion methods was written by Minchenikov et al. [19]. An early example of research into PCT-based pultrusion was conducted at the University of Newcastle upon Tyne, who worked to identify the main process parameters using glass fiber-reinforced nylon 12 (PA12-GF) and carbon fiber-reinforced polyether ether ketone (PEEK-CF) and found that the potential processing speeds to produce well-consolidated composite parts using thermoplastic pultrusion platforms could be significantly higher than those typically used in thermoset pultrusion [20].

Owing to the increasing availability of unidirectional thermoplastic tape feedstocks, and their ability to bypass fiber saturation issues, non-reactive thermoplastic pultrusion processes have been undergoing recent developments. Researchers based out of the Polytechnic of Porto in Portugal used their own powder-coating process to manufacture towpregs as well as PCTs which have been utilized for filament winding, pultrusion, and compression molding, finding that PCTs resulted in better mechanical properties [21]. Their earlier PCTs were made by processing their towpregs through a pultrusion die; however, their later work used direct melt pultrusion to create PCTs as well [22] which were then used in the design of experiments for the optimization of pultruded parts. Recent work by this group compared the mechanical properties of PCTs manufactured into CFRTP profiles with laminates compression-molded from the CFRTP profiles [23]. In all of these studies, glass fiber-reinforced polypropylene was the material system utilized, typically with tensile, flexural, and interlaminar shear testing chosen for mechanical evaluation.

Researchers located at the Skolkovo Institute of Science and Technology are also developing pultrusion using PCTs [24]. They used PCTs made in-house from commingled materials, as well as commercial tape—both using glass fiber-reinforced polypropylene. They made 6 mm diameter bars and 75 mm × 3.5 mm plates and used a combination of tensile, flexural, and interlaminar strength tests for mechanical evaluation. Their testing showed the importance of PCT quality to the final part’s mechanical properties.

The University of Maine’s Advanced Structures and Composites Center (ASCC) has similarly developed a proprietary platform for the non-reactive pultrusion of thermoplastic feedstocks. Known as the continuous forming machine (CFM), the pultrusion system can produce fiber-reinforced structural components at speeds of up to 4.0 m/min [25]. Initially developed to produce thermoplastic reinforcement bar (rebar) for concrete structures, the CFM has since been used to efficiently produce a variety of pultruded profiles such as rods, flat plates, c-channels, and periodic multi-curve structures [25]. Due to the reformability of thermoplastic polymer feedstocks, the CFM pultrusion line can be outfitted with additional co-processes such as rotary stamp forming, roll forming, pull winding, and the induction of curvature to enable in situ modifications of the pultruded parts.

The continuous forming machine, as depicted in Figure 1, features five primary machine zones for the production of pultruded profiles. As seen in Figure 1, rolls of the unidirectional thermoplastic feedstock to be used are loaded onto the creel system within the material input zone and then threaded through the CFM until reaching the drive system zone. The drive system consists of a tracked double-belt puller which clamps down on the pultruded profile shape to continuously move the feedstock material through the heating, forming, and cooling zones. The initial heating zone raises the feedstock material to appropriate processing temperatures as it travels through the CFM system into the forming die. Some co-processes, such as roll forming of periodic multi-curve surfaces, can also occur in the configurable forming zone, while other co-processes, such as filament winding of pultruded rods, are primarily performed after exiting the drive system. Cooling zones, implemented in the forming dies and prior to entering the puller, ensure the part is appropriately cooled before exiting the CFM platform.

Due to its non-reactive nature, the continuous forming machine is capable of processing a variety of unidirectional thermoplastic tape feedstocks. Thermoplastic polymer systems used on the CFM have ranged from commodity-grade (e.g., polypropylene) to high-performance-grade (e.g., low-melt polyaryletherketone) plastics. Additionally, both amorphous and semi-crystalline polymer systems have been utilized, demonstrating the feasibility of the CFM for processing a wide range of polymers from across the entire spectrum of thermoplastic polymers as depicted in Figure 2.

The available mechanical data on CFRTPs made with the continuous forming machine is very limited [25]. Similarly, mechanical performance data on pultruded CFRTPs produced with PCTs is primarily limited to glass fiber-reinforced polypropylene, and some PA12-GF and PEEK-CF data [22,23,24]. Therefore, the purpose of the work conducted here is to increase the available data on pultruded CFRTPs using a wide range of materials, and to assess the continuous forming machine’s capability to pultrude high-quality structural components. To assess the quality of CFRTPs manufactured on the CFM, compression-molding was used as a mature manufacturing method to compare against. This is a similar methodology to the research conducted in Portugal comparing pultruded profiles to compression-molded laminates [23]. The mechanical properties of the resulting CFRTPs were evaluated for testing the longitudinal tensile modulus, longitudinal compressive strength, and short beam strength (anecdotally referred to as interlaminar shear in the literature). Finally, additional comparisons were made to values published in the literature, where available for certain material systems.

2. Materials and Methods

Laminates were fabricated from commercially available, fiber-reinforced thermoplastic tape feedstocks. The feedstocks utilized consisted of unidirectional (UD) pre-impregnated thermoplastic tapes with a common width of 5.08 cm. A total of six material systems were obtained from three separate vendors of UD tapes: A+ Composites (Weselberg, Germany), Avient Corporation (Avon Lake, OH, USA), and Victrex plc (Lancashire, United Kingdom).

Table 1. Material systems’ information.

Material System (Polymer/Fiber Abbreviation)	Material Supplier (Vendor’s Product Code)	Technical Data Sheet
Polyamide 12, Glass Fiber-Reinforced (PA12-GF)	A+ Composites (GF-PA12-53-103-186)	[26]
Polypropylene, Glass Fiber-Reinforced (PP-GF)	Avient Corporation (PolystrandTM IE 7034BU)	[27]
Polyethylene Terephthalate Glycol, Glass Fiber-Reinforced (PETG-GF)	Avient Corporation (PolystrandTM IE 5843.1)	[28]
Polycarbonate, Glass Fiber-Reinforced (PC-GF)	A+ Composites (GF-PC-51-103-193)	[29]
High-Density Polyethylene, Glass Fiber-Reinforced (HDPE-GF)	A+ Composites (GF-HDPE-46-51-296)	[30]
Low-Melt Polyaryletherketone, Carbon Fiber-Reinforced (LMPAEK-CF)	Victrex plc (Victrex AETM 250 UDT)	[31]

presents the six material systems along with their corresponding vendors and product codes. Each of the material systems were used to fabricate continuous fiber-reinforced thermoplastic laminates using two separate methods, compression-molding in a heated platen press and thermoplastic pultrusion using the continuous forming machine.

The compression-molded plates were fabricated in the Advanced Structures and Composites Center’s Alfond Advanced Manufacturing Lab for Structural Thermoplastics (TPL). The TPL features a Dieffenbacher (Eppingen, Germany) FiberForge Relay 2000 automated tape layup machine for the precise creation of thermoplastic preforms, as shown in Figure 3a. All preforms were fabricated using a unidirectional layup of 13 plies. Each ply was staggered laterally throughout the through-thickness direction of the laminate to mitigate any potential voids that could be associated with the stacking of gaps between adjacent courses. The 38 by 38 cm laminates were then consolidated within a Carver Inc. (Wabash, IN, USA) Monarch Series heated compression-molding press as shown in Figure 3b. The heated platen press allowed for controlled heating and cooling cycles for consolidation of the CFRTP laminates. The selection of processing parameters for the consolidation of each material was determined through previous research efforts, with a preference for consolidation recipes that produced the highest tensile elastic modulus as seen through mechanical testing. The consolidation parameters used for each of the six materials are presented in Figure 4. Test specimens were then cut from these plates using a waterjet cutter at the ASCC.

The pultruded laminates were produced on the continuous forming machine (CFM) housed at the University of Maine’s Advanced Structures and Composites Center. For this work, the CFM was outfitted with a custom 150 cm long forming die to continuously pultrude plates with a rectangular cross-sectional profile consisting of a 5.08 cm width by a 0.318 cm height. Processing settings for the pultruded materials were initially determined based on the technical specification sheets provided by the unidirectional tape manufacturers and were then adjusted as needed to produce quality test specimens. The CFM was allowed to reach steady-state operational status before marking plates for test specimens. Test specimens were then cut out of the pultruded plates using the ASCC’s waterjet cutter.

Before specimens could be made on the CFM, the number of tapes required to adequately fill the forming die for each material system needed to be determined. This is primarily due to the differences in thickness for each of the unidirectional tapes used, therefore, a preliminary fill study was conducted for each of the material systems. To achieve this, the 0.318 cm height of the forming die was divided by the nominal tape thickness, as reported by the manufacturer, to provide an initial starting point for the number of tape spools needed for each material. The number of tows fed into the pultrusion machine was then adjusted as necessary to produce test specimens of acceptable quality standards, as determined by operator assessment of poor surface finish and acoustic resonance. The number of tows used to produce pultruded test specimens for each material system are presented in

Table 2. Number of tows used for pultrusion by material system.

Material System	Number of Tows Used
PA12-GF	15 tows at 5.04 cm widths
PP-GF	14 tows at 5.04 cm widths
PETG-GF	13 tows at 5.04 cm widths plus 1 tow at 2.54 cm width
PC-GF	15 tows at 5.04 cm widths plus 4 tows at 0.635 cm widths
HDPE-GF	9 tows at 5.04 cm widths plus 1 tow at 2.54 cm width
LMPAEK-CF	21 tows at 5.04 cm widths

and the creel rack system is depicted in Figure 5.

Two guide plates were used to assist with organization of the unidirectional tape tows as they traveled through the CFM. The first guide plate was used to vertically separate the tows coming off the creel rack so that they maintained a slight separation as they moved through the 183 cm long oven. The second guide plate, positioned between the oven and the forming die, as seen in Figure 6, was used to steer the heated tapes into position within the forming zone. The second guide plate was heated via two cartridge heaters and its surface temperature was monitored using a K-type thermocouple. The UD tapes were first fed through the entire pultrusion system, including the guide plates, until they reached the double-belt puller. The CFM system was then turned on, and all zones in the heating and forming areas were allowed to ramp up to temperature while material was slowly being pulled through. Once the CFM reached steady-state operational status, the pultruded plate was marked to designate the starting line of test specimen production. A minimum of 10 m of pultruded test specimens were then produced and were sliced into approximately 2 m long sections at the cut-off zone.

This process was then repeated for each of the pultruded material systems. Setpoint temperatures and line speeds were adjusted for each material to produce test specimens of adequate visual quality. Quality assessments consisted of preliminary evaluations of geometric consistency, surface finish, and non-destructive estimates of consolidation quality (e.g., acoustic assessment). Depending on the material system, pultrusion line speeds varied from 0.30 to 0.61 m/min and oven zones varied from 72 to 350 °C. A full list of the processing parameters used for pultrusion of each material system on the CFM is presented in

Table 3. Average measured CFM processing parameters for the pultruded material systems.

Material System	PA12-GF	PP-GF	PETG-GF	PC-GF	HDPE-GF	LMPAEK-CF
Line Speed (m/min)	0.457	0.305	0.610	0.305	0.305	0.305
Heater Zone 1 (°C)	119.2	80.2	72.2	80.1	119.8	290.9
Heater Zone 2 (°C)	200.9	99.4	100.3	140.2	136.9	289.7
Heater Zone 3 (°C)	229.8	164.2	158.3	179.6	136.1	349.9
Forming Zone 1 (Guide Plate) (°C)	185.0	167.0	127.6	159.9	134.4	279.7
Forming Zone 2 (Preformer) (°C)	152.1	140.1	122.9	160.0	110.1	274.2
Forming Zone 3 (°C)	120.3	120.0	81.6	132.0	74.1	266.0
Forming Zone 4 (°C)	58.1	94.2	70.2	101.4	50.6	256.1
Forming Zone 5 (°C)	22.2	50.9	45.9	51.2	29.6	230.1

. Figure 7 shows the resulting pultruded flat plate as seen when it exited the forming die.

CFRTP laminates fabricated using both manufacturing techniques were then subjected to burn-off testing or acid digestion, as applicable, to determine the weight fractions of each laminate. These results are summarized in

Table 4. Average measured fiber mass fraction (%).

Material System	CFM	Compression Molded	Technical Data Sheet
PA12-GF	71.7	71.3	74
PP-GF	73.0	70.1	70
PETG-GF	58.0	59.1	58
PC-GF	68.3	68.8	69
HDPE-GF	70.0	68.9	69
LMPAEK-CF	66.9	68.7	66

. As can be seen from Table 4, the differences in fiber mass fractions between manufacturing methods and technical data sheet values were predominantly minor. The material systems with amorphous polymer matrices—PETG-GF and PC-GF—both had higher fiber mass fractions when compression-molded. However, this magnitude of difference is within 1.1% of the values measured using the CFM, a minor effect. The measured fiber mass fractions for the PA12-GF specimens are noticeably lower than the value reported by the preliminary technical data sheet. Since the difference between manufacturing methods is small, within 0.5%, it is presumed that the preliminary data sheet may have overestimated its value.

3. Tensile Testing

3.1. Tensile Testing Methodology

Tensile testing was performed on both the compression-molded and CFM-pultruded laminates for each material system. The performed tensile tests were largely based on ASTM D3039-17 [32], although the specimen thickness exceeded the recommendations of ASTM D3039 due to the fixed geometry of the CFM forming die. All specimens were conditioned prior to testing for a minimum of 40 h at standard laboratory atmosphere and temperature pursuant to ASTM D618-21 [33].

Tests were performed in a servohydraulic universal testing machine, a model 8872 testing system produced by Instron (Norwood, MA, USA), outfitted with a 25 kN load cell. All tension tests were run with a standard crosshead displacement rate of 2 mm/min. Specimens were loaded into the hydraulic wedge grips and were aligned vertically within the test frame. Due to the slightly serrated surface of the grips, no tabbing was used between the specimen and grip interface. Each test sample was outfitted with a strain measurement device, an Epsilon Technology Corp. (Jackson, WY, USA) model 3542 axial extensometer. Tests were run until specimen failure or a noticeable drop in load was observed. An image showing the test setup is presented in Figure 8.

A minimum of seven tensile test samples were performed per material system and CFRTP plate fabrication method. It was observed during tensile testing that some of the specimens would slip within the frictional wedge grips once the loading force reached an appreciable value. This phenomenon was observed to be more pronounced with some materials (e.g., LMPAEK-CF) than with others (e.g., PA12-GF). Due to the varied thickness between the compression-molded laminates and the CFM-pultruded CFRTP laminates and invalid failure modes, values for ultimate strength were not obtained (the related literature has experienced similar challenges [22,23]. Instead, the longitudinal tensile modulus for each laminate type was determined and compared. Erroneous data was removed from the analysis by cutting large jumps in stress–strain data which corresponded to grip slipping, improper specimen failure, or excessive hydraulic fluctuation. To account for initial system compliance and specimen seating, the force–strain data was programmatically zeroed using an artificial preload of 0.075 kN to ensure the reported elastic properties are representative of the material behavior and not the test setup. MATLAB scripts, using MATLAB version R2022a, were developed and used to process the tensile test data and calculate the longitudinal tensile modulus as determined from the 1000 to 3000 με region using the least squares method in accordance with ASTM D3039-17.

3.2. Tensile Testing Results

Figure 9 shows the resulting stress–strain data from the preload to the maximum load for each specimen used and reveals not all materials reached the 1000–3000 με range. For these specimens, only available strain data from the acceptable range was used. If a specimen had no stress–strain data within the acceptable range, it was removed from the primary analysis; this was only necessary for two of the LMPAEK-CF specimens manufactured using compression molding.

The resulting average longitudinal tensile modulus for each material system and CFRTP fabrication method are shown in Figure 10. Figure 10 additionally provides error bars for each value calculated using the standard deviation of the tested samples.

Table 5. Longitudinal tensile modulus testing results.

	CFM-Pultruded Specimens			Compression-Molded Specimens
Material System	Average (GPa)	Standard Deviation (GPa)	CoV (%)	Average (GPa)	Standard Deviation (GPa)	CoV (%)	TDS Value (GPa)
PA12-GF	42	0.8	1.9	44	1.3	2.9	45
PP-GF	42	2.3	5.5	42	4.5	10.7	34.5
PETG-GF	34	5.2	15.4	37	1.2	3.2	30.3
PC-GF	47	5.5	11.6	47	3.6	7.7	44
HDPE-GF	38	4.3	11.2	37	5.2	14.2	35
LMPAEK-CF	122	2.9	2.4	135	15.1	11.2	>150

presents tabulated results for each specimen group alongside the manufacturer’s reported technical data sheet value.

3.3. Discussion of Tensile Testing Results

For the majority of materials tested, the average longitudinal tensile modulus values were nearly identical, with the differences falling within one standard deviation of each other. It was observed that specimens pultruded via the CFM method had lower standard deviations for semi-crystalline materials (HDPE, LMPAEK, PA12, and PP) than seen for the compression-molded specimens, while the opposite was observed for amorphous materials (PC and PETG). It is hypothesized that this result may be due to the presence of fiber wash in the semi-crystalline materials when being processed via compression-molding.

Notably, the LMPAEK-CF specimens pultruded with the CFM exhibited a slightly lower average modulus but also had a much narrower standard deviation than those produced via compression-molding. As a result, the minimum predicted modulus (using a 95% confidence interval) of the CFM specimens would be comparable to that of the compression-molded specimens, making the difference in average modulus a minor consideration in design or application. Due to the LMPAEK specimens having a very limited number of data points within the standard strain range, the lower strain limit was changed to 500με to expand the usable strain range to include all data points from all test specimens and to see what influence there was on the results. This resulted in the average longitudinal tensile modulus of the CFM specimens remaining unchanged, and their standard deviation minorly decreased by 0.2 GPa. The compression-molded specimens, however, saw a significant increase in average modulus to 147 GPa, along with an increase in standard deviation to 27.9 GPa. This value for the updated longitudinal tensile modulus, using the expanded strain range, was seen to approach the manufacturer’s reported value of greater than 150 GPa. Interestingly however, even though using the expanded strain range resulted in an increased average tensile modulus for the compression-molded LMPAEK specimens, due to the subsequent increased standard deviation, the comparative analysis still shows the resultant average modulus for the CFM-pultruded specimens to fall within one standard deviation of the value observed for the compression-molded specimens. Therefore, it was seen that regardless of whether the standard or expanded strain range was used for the analysis, the CFM-pultruded LMPAEK specimens exhibited more consistent mechanical properties and approached but did not fully achieve the higher average modulus observed for the compression-molded specimens. This suggests that additional improvements could be gained through future tuning of CFM pultrusion parameters for the LMPAEK-CF material system.

In summary, the tensile testing results demonstrate that the average values obtained for tensile moduli were found to be similar when comparing the performance of laminates fabricated with either of the two manufacturing methods across the range of CFRTP material systems used. Although data on pultruded thermoplastic laminates is scarce in the literature, additional similarities can be observed through comparison with available tensile data on similarly manufactured CFRTP laminates using glass fiber-reinforced polypropylene material systems. Novo et al. pultruded a variety of PP-GF laminates with a 69.3% fiber mass content and observed through mechanical testing a maximum tensile modulus of 33.9 ± 1.5 GPa [22]. Esfandiari et al. pultruded and then subsequently compression molded their tensile specimens to reach a maximum tensile modulus of 39.4 ± 2.1 GPa with a 71.0% fiber mass fraction [23]. Lastly, Verdernikov et al. achieved a maximum tensile modulus of 26.2 ± 2.2 GPa with a fiber mass fraction of 60.3% using an in-house-made PP-GF feedstock [24]. As seen in

Table 6. Average longitudinal tensile modulus for CFM-pultruded and compression-molded PP-GF laminates compared to values found in the literature.

Source of Data	Longitudinal Tensile Modulus (GPa)	Fiber Mass Fraction (%)
CFM-Pultruded PP-GF	42	73.0
Compression-Molded PP-GF	42	70.1
Novo et al. [22]	33.9	69.3
Esfandiari et al. [23]	39.4	71.0
Vedernikov et al. [24]	26.2	60.3

, when comparing the results in the literature to those obtained in this study through mechanical testing, it was observed that the CFM-pultruded PP-GF laminates exhibited similar tensile moduli to materials with comparable fiber mass fractions. In fact, the CFM-pultruded laminates slightly outperformed the highest value found in the literature—that of Esfandiari et al., who twice consolidated their laminates. These results suggest that the CFM-pultruded laminates, at least for the PP-GF material feedstock, achieved a moderate-to-high level of consolidation quality using the processing parameters presented in Table 3.

4. Compression Testing

4.1. Compression Testing Methodology

Compression testing was performed on both the compression-molded and CFM-pultruded laminates for each material system. The compression testing was largely based on ASTM D6641-23 [34], with specimen geometries based on the available specimen thicknesses for each manufacturing method. All specimens were conditioned prior to testing for a minimum of 40 h at standard laboratory atmosphere and temperature pursuant to ASTM D618-21 [33].

Tests were performed on two different servohydraulic universal testing machines, a model 8801 testing system using a 100 kN load cell, and a model 8872 testing system using a 25 kN load cell, both produced by Instron (Norwood, MA, USA). All compression tests were performed using a standard crosshead displacement rate of 1.2 mm/min. Test specimens were first loaded into a combined loading compression test fixture prior to testing. Tests were then run until specimen failure or a noticeable drop in load was observed.

A minimum of seven compression tests were performed per material system and CFRTP plate fabrication method. For these tests, the maximum compressive stress values obtained were recorded and used for comparison of the different manufacturing methods. To account for initial system compliance and specimen seating, the force–strain data was programmatically zeroed using an artificial preload of 0.04 kN to ensure the reported elastic properties are representative of the material behavior and not the test setup. MATLAB scripts were developed and used to process the compressive test data and calculate the average maximum longitudinal compressive strengths.

4.2. Compression Testing Results

Figure 11 shows the resulting stress–displacement data for each of the specimens tested until they reached their maximum measured stress. Generally, the test specimens linearly increased in stress with displacement until reaching a peak stress state, then subsequently experiencing a drop in load. Some of the specimens, however, reached peak stress and then plateaued.

The resulting average longitudinal compressive strength values for each material system and CFRTP fabrication method are shown in Figure 12. Figure 12 additionally provides error bars for each value calculated using the standard deviation of the tested samples. These results have been tabulated in

Table 7. Longitudinal compressive strength testing results.

	CFM Pultruded Specimens			Compression-Molded Specimens
Material System	Average (GPa)	Standard Deviation (GPa)	CoV (%)	Average (GPa)	Standard Deviation (GPa)	CoV (%)	TDS Value (GPa)
PA12-GF	358	47.1	13.2	280	26.6	9.5	n/a
PP-GF	214	14.0	6.5	228	23.3	10.2	357
PETG-GF	291	31.4	10.8	315	29.5	9.4	621
PC-GF	386	79.5	20.6	351	46.6	13.3	n/a
HDPE-GF	171	16.0	9.3	172	10.8	6.3	n/a
LMPAEK-CF	643	44.4	6.9	668	125.6	18.8	>1200

, as well as the CoV for each specimen group and the manufacture reported technical data sheet value if available.

4.3. Discussion of Compression Testing Results

For most of the materials tested, the average compressive strength values were similar regardless of manufacturing method, with the differences falling within a standard deviation of each other. These results revealed no clear relationship between polymer structure, manufacturing method, and average maximum stress or standard deviation taken as a whole. An additional analysis of observed failure modes was undertaken to allow for further insights into the performance of each manufacturing process.

Notably, the LMPAEK-CF specimens manufactured with compression-molding were observed to have a significantly larger standard deviation than those produced with the CFM. This could suggest a possible presence of fiber wash having occurred during the compression-molding manufacturing process. Inspection of typical tested samples for each of the manufacturing methods shown in Figure 13 reveals the compression-molded specimens typically tended to fail in transverse shear while specimens made via the CFM typically experienced long splitting failures. The long splitting failures of the CFM specimens may indicate a potential weakness in the transverse direction, which could imply suboptimal matrix processing. Alternatively, these failures could have resulted following an initial failure combined with a brittle matrix. Regardless, the average observed strength between both manufacturing methods was similar, suggesting that CFM pultrusion could be a potentially viable platform for aerospace grade materials such as LMPAEK-CF PCTs.

A significant difference was observed when comparing the average compressive strength of the PA12-GF specimens. Those manufactured via the CFM exhibited an average compressive strength value that exceeded the compressive strength of those produced via compression-molding, with the difference falling outside the combined standard deviations. This is dissimilar to the results observed in tensile testing, where the compression-molded PA12-GF specimens had a slightly higher longitudinal elastic tensile modulus than the pultruded specimens. Since the compressive strength of a composite laminate depends more heavily on its matrix properties than its tensile modulus, which is largely a fiber-dominated property, these results suggest the compression-molded specimens may have had inferior consolidation quality. This assumption could be further supported by assessing the failure modes of the tested specimens. The compression-molded specimens demonstrated a mix of transverse shear failure (like the LMPAEK-CF compression-molded specimens) and out-of-plane failure. The specimens produced using the CFM, however, demonstrated only one out-of-plane failure with the remaining specimens failing in transverse shear. Figure 14 shows one of the PA12-GF specimens that exhibited out-of-plane failure. Given that the out-of-plane failure is predominantly a buckling failure mechanism, which is resisted by the matrix, this could also indicate that the compression-molded specimens had an inferior consolidation to the CFM-pultruded specimens.

It was observed that the measured compressive strengths were consistently lower than the manufacturer-reported technical data sheet (TDS) values. This is a common occurrence in composite material testing and is likely a result of fabrication and testing variables such as specimen quality or failure modes. However, because the results were consistent across the sample sets, the data remains useful for a comparison between the two manufacturing methods. Overall, the compressive testing results indicate that the continuous forming machine could be capable of pultruding composite laminates across a range of CFRTP material systems that exhibit compressive properties similar to those produced through conventional compression-molding.

5. Short Beam Strength Testing

5.1. Short Beam Strength Testing Methodology

Short beam strength testing was performed on both the compression-molded and CFM-pultruded laminates for the material systems containing LMPAEK, PA12, PC, and PP polymers. Due to time constraints during testing, the PETG and HDPE polymers were not subjected to short beam strength testing. Unlike the other two test methods, short beam strength testing is predominately polymer-dependent and is typically used by industry for quality control and general assessment of consolidation quality. This makes short beam strength an appropriate comparison tool for ascertaining the relative level of consolidation for the CFM pultrusion and compression-molding manufacturing methods. The testing was largely based on ASTM D2344-22 [35], although the specimen thickness was limited to that of the existing material stock, approximately 3 mm for CFM manufactured specimens, 1.7 mm for the compression-molded LM-PAEK specimens, and 2.4 mm for the remainder of the compression-molded specimens. All specimens were conditioned prior to testing for a minimum of 40 h at standard laboratory atmosphere and temperature pursuant to ASTM D618-21 [33].

Tests were performed in a servohydraulic universal testing machine, a model 8874 testing system produced by Instron (Norwood, MA, USA), outfitted with a 25 kN load cell. All short beam strength tests were run with a standard crosshead displacement rate of 1 mm/min. Specimens were loaded into a test fixture with a 6.0 mm diameter loading nose and 3 mm diameter supports. The specimen span was adjusted as necessary to maintain a specimen span-to-depth ratio of 4:1.

A minimum of five test samples were performed per material system and CFRTP plate fabrication method. Due to the limited thickness of the specimens and the fixture geometry, the specimens continued to gain load after an initial peak due to the specimen being crushed against the supports at large displacements (>1mm). As a result, test data that experienced a 7.5% load drop after a local maximum or a 1 mm displacement was removed from the analysis since it corresponded to invalid failure modes resulting from the test setup geometry. Plots for the full stress–displacement data are presented in Appendix A, see Figure A1. To account for initial system compliance and specimen seating, the force–strain data was programmatically zeroed using an artificial preload of 0.075 kN to ensure the reported stress–displacement curves were representative of the material behavior and not the test setup. MATLAB scripts were developed and used to process the short beam strength test data and calculate the short beam strength of each specimen as determined from the valid data criteria.

5.2. Short Beam Strength Testing Results

Figure 15 shows the resulting stress–displacement data from the preload to the end criteria for each specimen used. Where specimens did not experience a 7.5% drop in load, they dramatically softened and transitioned to asymptotic behavior. This data demonstrates that the compression-molded specimens typically had a stiffer response than those pultruded using the CFM, but overall, they both exhibited similar peak strengths.

The resulting average short beam strength for each material system and CFRTP fabrication method are shown in Figure 16. Figure 16 additionally provides error bars for each value calculated using the standard deviation of the tested samples.

Table 8. Test results from short beam strength testing.

	CFM-Pultruded Specimens			Compression-Molded Specimens
Material System	Average (MPa)	Standard Deviation (MPa)	CoV (%)	Average (MPa)	Standard Deviation (MPa)	CoV (%)	TDS Value (MPa)
PA12-GF	48.6	2.09	4.3	34.6	6.64	19.2	n/a
PP-GF	28.1	1.25	4.4	22.5	5.56	24.7	26
PC-GF	46.6	3.41	7.3	47.8	1.40	2.9	n/a
LMPAEK-CF	85.5	4.40	5.2	83.1	6.49	7.8	n/a

presents the tabulated results for each specimen group alongside the manufacturer’s reported technical data sheet value if available.

5.3. Discussion of Short Beam Strength Testing Results

While the span was adjusted to maintain a consistent span-to-thickness ratio for each specimen type, the specimen thicknesses were likely a confounding variable. There is substantial discussion in the literature that indicates there is a size effect on the apparent interlaminar shear strength of composite materials [36]. For most of the specimen types, the differences are likely minor as their thicknesses are relatively within the same magnitude of each other. However, since the LMPAEK-CF specimens produced on the CFM were nearly twice as thick as the compression-molded LMPAEK-CF specimens, their results should be interpreted cautiously.

From the load–displacement curves, it can be seen that the compression-molded specimens exhibited a stiffer response. It was also observed that the LMPAEK-CF and PC-GF compression-molded specimens experienced test termination from load drop rather than displacement cut-off and were noticeably stiffer than their counterparts manufactured on the CFM. These specimens also appeared to be more brittle. For the LMPAEK-CF specimens, this result was consistent with the compression testing, where the failure mode and lower average strength values indicated potential matrix weakness—therefore indicating the possibility for improvement in consolidation quality when pultruded using the CFM process. This observation was not similar for the PC-GF specimens however, as compression testing for this material system showed higher strength values for pultruded specimens. During both test methods, the pultruded PC-GF specimens had a higher variance—suggesting improvements in the robustness of the pultrusion process for this material could be a subject of future work.

Significantly lower short beam strength values were observed when comparing the compression-molded PA12-GF specimens with their pultruded counterparts. This trend in mechanical performance was also observed when looking at the compressive strength results between manufacturing methods for the PA12-GF specimens. Therefore, the short beam strength results support the hypothesis that the compression-molded PA12-GF specimens had an inferior consolidation quality when compared to their pultruded counterparts.

Lastly, the polypropylene shear strength results indicated the CFM processed the material better than compression molding, exceeding the reported technical data sheet value. Similarly manufactured PP-GF laminates in the literature are also available for comparison. Vedernikov et al. achieved an apparent maximum average interlaminar shear strength of 23.1 ± 1.6 MPa [24], while Novo et al. achieved a maximum average interlaminar shear strength of 27.8 ± 0.6 MPa [22]. These results are within a similar range to the CFM-manufactured (28.1 MPa) and compression-molded (22.5 MPa) specimens, suggesting the values achieved via CFM pultrusion are likely indicative of a well-consolidated PP-GF laminate.

6. Conclusions

The continuous forming machine was used to pultrude fiber-reinforced composite laminates across a range of commodity-grade to high-performance thermoplastic material systems through the use of pre-consolidated tape feedstocks. The pultruded laminates made with these materials were then subjected to mechanical testing to assess their longitudinal tensile modulus, compressive strength, and apparent short beam strength, providing novel data on CFRTP components produced via a non-reactive thermoplastic pultrusion process. To compare the performance of the CFM-pultruded components, compression-molded laminates were fabricated using the same feedstock materials and were also subjected to mechanical testing. Comparative testing showed that, on average, CFM-pultruded laminates exhibited mechanical properties similar to those achieved through compression-molding. These results indicate that non-reactive thermoplastic pultrusion, a relatively novel manufacturing process, can produce components of similar quality to those fabricated via a mature manufacturing method like compression-molding. Additional comparisons were made to data in the literature, where available, on pultruded PP-GF pre-consolidated tapes. The CFM-pultruded laminates were shown to exhibit similar mechanical properties with pultruded PP-GF laminates from other works, further validating the potential of thermoplastic pultrusion using PCTs.

Although this work demonstrated the viability of CFM pultrusion for thermoplastic PCTs, some discrepancies and limitations were observed. These include a potentially suboptimal consolidation quality for the pultruded LMPAEK-CF and compression-molded PA12-GF materials, which could be improved upon in future work. Furthermore, a lack of flexural testing within this study limited additional comparisons to other works in the literature and it is suggested that future work should include additional flexural tests. Other limitations include a lack of tabbing for the testing of tension specimens, which may serve to mitigate slippage of the specimens in future test efforts, along with a limited number of pultruded material systems in the literature to compare against. Additional discrepancies were observed in the comparison of short beam strength values, predominately due to the varied thicknesses of the laminates produced via the two manufacturing methods.

Despite these limitations, this study shows that non-reactive pultrusion via the CFM platform can be used to produce CFRTP components of comparable mechanical performance to that of compression-molded laminates. These results were demonstrated across six different material systems ranging from commodity-grade to aerospace-grade materials, indicating that efficient thermoplastic pultrusion of PCTs could be a viable option for manufacturers regardless of material selection. Finally, through comparison with TDS values as reported by PCT material suppliers, this study shows that CFM-pultruded components could be considered for use by designers in axially loaded application settings.