Comparative Analysis of Impregnation Methods for Polyimide-Based Prepregs: Insights from Industrial Perspective

Abstract

This study presents a comparative analysis of two industrially relevant technologies for manufacturing of prepreg composite materials based on polyimide (PI) resin: hot-melt and solvent-based technology. More specifically, the study focuses on evaluating the relationship between key processing parameters and the final properties of the composite material manufactured with unidirectional (UD) C-fibers and woven fabrics used as reinforcement for both technologies. The impregnation process was carried out using a custom-designed coating equipment developed by Mikrosam D.O.O. Manufactured prepregs were characterized in terms of their resin content, volatile content, weight, width, and quality of the applied resin film. The hot-melt method that involves applying the resin in a semi-molten state with minimal solvent content provided a stable resin content (34–35%) and low volatiles (~1.2–1.5%) in the final product. The solvent-based method, using a resin/solvent ratio of 50:50, enabled deeper resin penetration into the fibers, particularly in woven fabrics (resin content: 34–37%) and lower residual volatiles (~0.3–0.5%). These results showed that the hot-melt technology consistently produced prepregs with very stable resin content, which is critical for structural applications requiring increased mechanical performance. In contrast, the solvent-based method demonstrated better adaptability to different reinforcement forms, improved impregnation depth, and excellent film uniformity, particularly suitable for woven fabrics. Representative SEM micrographs confirmed uniform resin distribution, full fiber wetting, and absence of voids, validating the impregnation quality obtained by both techniques. These findings highlight the technological relevance of selecting the appropriate impregnation route for each reinforcement architecture, offering direct guidance for industrial-scale composite manufacturing, where the hot-melt method is preferred for UD prepregs requiring precise resin control, while solvent-based impregnation ensures deeper and uniform resin distribution in woven fabric structures.

1. Introduction

Carbon fiber-reinforced polymers (CFRPs) are composite materials that a key component in the modern industry due to their exceptional strength-to-weight ratio, superior mechanical properties, corrosion, and chemical resistance, as well as potential for multipurpose applications in different industrial sectors such as automotive, aerospace, marine, energy, and many others. In most of the cases, the polymer matrix is an epoxy resin where carbon fibers are embedded, but other matrices are also possible, such as thermoset ones, like vinyl or phenol; thermoplastic matrices such as polysulfone (PS), polyether ether ketone (PEEK) and polyimide (PI); or others [1,2,3,4]. The main role of the resin is to provide the durability of the composite material by binding and protecting the fibers that actually transfer the load to the structure providing exceptional mechanical strength. Therefore, the composite materials are lighter and stronger compared to traditional materials such as metals or ceramics [5,6,7,8,9].

In the last few decades, CFRPs have played an inevitable role, mainly in the aerospace industry, where CFRPs have replaced the aluminum alloys, for example, in the Boeing 787 Dreamliner (B787)—50% were replaced by CFRPs and contributed to critical plane weight reduction and up to 22% fuel savings—while the Airbus A350 also used CFRPs (~53 wt%), and noted 50% lower maintenance of structures by substantially extending the service interval, higher fuel efficiency, lower CO₂ emissions and maintenance costs, and better design flexibility through parts integration. More specifically, in the last decade, CFRPs have been in high demand for low-cost but large-volume parts in the energy sector such as turbine blades, composite tanks, etc., mainly due to increased production capacity and cost reduction of C-fibers and CFRPs. Overall, it is assumed that demand for CFRPs for different applications has reached 285 kilo-tons/annually (2025) [6,8].

In this context, C-fiber reinforcements in composite materials are supplied in a number of different forms, like uni-, bi-, and multidirectional fabrics; woven fabrics; non-wovens; and random fiber mats. Unidirectional (UD) reinforcements are primarily used in structural applications where maximum strength and stiffness in a specific direction are required, such as aerospace primary load-bearing components. On the other hand, woven fabrics are applied in cases where multidirectional load distribution, dimensional stability, and resistance to delamination are critical, such as in panels, fairings, or energy-sector composite parts [3,8,10,11,12].

Furthermore, C-fiber prepregs are an intermediate (semifinished) preimpregnated product in the preparation of CFRPs, where C-fibers are already impregnated in the polymer matrix. They are a prerequisite for high quality, load-optimized lightweight fiber composites. The global prepreg market size is valued at USD 10.7 Bn (2024) and is predicted to reach USD 37.5 Bn by 2034 at a 13.5% CAGR during the forecast period for 2025–2034 [13]. There has been rapid growth in carbon fiber prepregs, which are currently dominating the current prepreg market, representing over 84% of the total market value. This composite material offers a higher consistency, simplified handling, and greater control over the final product’s properties; it allows the reproducible production of high-quality components. Therefore, industries like aerospace, energy, and transportation can easily manufacture their final products by the use of lightweight preimpregnated materials that can have a relatively high fiber/resin ratio (up to 85%) if needed, a better control of resin distribution, reduced voids, consistent mechanical properties, shorter manufacturing times, are less labor intensive, have better uniformity and quality control, and lead to reduced waste [5,11,12,13,14,15].

In general, prepregs are produced using two basic processes: hot-melt impregnation and solvent-based (or solution-coating). While numerous studies have addressed thermosetting polyimide (PI) resin chemistry and composite performance, only a limited number of studies have compared the practical aspects of PI prepreg production using different impregnation techniques. This study provides a systematic experimental comparison between hot-melt and solvent-based impregnation routes for thermoset PI systems, highlighting their respective process characteristics and limitations.

Parameters, including winding, control, impregnating, and drying, have to be considered and adapted individually depending on the type of reinforcement, the resin matrix, and the required properties of the final product [12,16]. Hot-melt impregnation is used in cases when polymeric matrixes are not soluble in suitable solvents, while solvent-based impregnation is available for all the polymers that are soluble in suitable solvents. In the case of prepregs based on polyimide thermoset resin, both impregnation techniques are possible.

Polyimides (PIs) are widely used in industries due their exceptional mechanical properties and chemical and thermal resilience; they can be used under strict, demanding environmental conditions where long term operational stability is required [14,17,18]. Recent research has explored various thermosetting polyimide systems for high-temperature composite applications, focusing on resin synthesis, processability, and thermal stability [19,20]. However, only a few works have investigated the impregnation behavior and processing challenges of PI prepregs on an industrial scale [21,22,23,24,25]. Recent investigations have explored processability improvements of thermosetting PI prepregs through controlled partial imidization and solvent-assisted impregnation. In this context, the present study provides a systematic comparison between hot-melt and solvent-based impregnation methods using the same thermosetting PI resin formulation, aiming to identify the processing–structure relationships relevant for high-performance prepreg manufacturing. In the hot-melt process, the resin matrix is applied in a semi-melted state without a significant amount of solvent, whereas the solvent-based method involves using a solution of PI to facilitate better penetration of the resin into the fiber/reinforcement structure, followed by evaporation of the solvent [12,26,27]. Despite the recent progress in PI resin development, there is still limited understanding of how different impregnation methods affect the quality and stability of PI-based prepregs at the industrial scale. Most published studies remain focused on laboratory-scale resin preparation or composite characterization, without linking processing conditions to prepreg formation behavior. Therefore, this study aims to bridge the existing gap by systematically and comparatively analyzing the influence of hot-melt and solvent-based impregnation parameters on the quality indicators of thermosetting PI prepregs—such as resin content, volatiles, and film uniformity—using two types of reinforcement materials (unidirectional carbon fibers and woven carbon fabrics). In addition, evaluating the final properties of the resulting prepreg materials obtained by the two methods are discussed; they provide industrially relevant insights into process optimization.

2. Materials and Methods

2.1. Materials

For the purpose of this study, the following materials were used:

o

The polyimide (PI) resin in powder form was supplied by TÜBİTAK (TÜBİTAK Marmara Research Center, Gebze, Kocaeli, Turkey). This resin was specifically developed for high-performance prepreg applications and represents a partially imidized PI pre-polymer. According to the supplier, the general specification range for this PI pre-polymer family is Mn = 1500–20,000 g/mol, while the specific batch used in this study has Mn in the range of 1500–3000 g/mol. [Due to confidentiality restrictions, detailed compositional information cannot be disclosed].

o

Solvent DMAC (dimethylacetamide, purity >99%, Sigma Aldrich), used for solvent-based impregnation.(was obtained from Sigma-Aldrich, St. Louis, MO, USA)

o

Reinforcements:

o

Carbon fibers 12 K, PAN-based, supplied by DowAksa (DowAksa, Gumussuyu, Istanbul, Turkey).

o

Woven carbon fabric, grammage of 200 gsm, was supplied by DowAksa (DowAksa, Gumussuyu, Istanbul, Turkey).

o

Release films and liners:

o

Silicone-coated paper (yellow, 90 gsm, double-sided coated), supplied by Excelitas Noblelight (Heraeus), Kleinostheim, Germany.

o

Silicone-coated paper (white, 90–92 gsm, single-sided coated), supplied by Tireks (Tireks, Prilep, North Macedonia)

o

Kapton® film (50 µm), supplied by Airtech Europe, Niederkorn, Luxembourg.

o

Polypropylene (PP) film (50 µm thick), supplied by Tubitak (TÜBİTAK Marmara Research Center, Gebze, Kocaeli, Turkey).

2.2. Experimental Setup

The impregnation experiments were performed on custom-designed coating units, developed and manufactured by Mikrosam D.O.O. (Prilep, North Macedonia) (Figure 1). The units enable both hot-melt coating and solvent-based impregnation, and they are equipped with the following:

o

Unwinder and rewinder units with adjustable tension control;

o

Resin bath (for solvent process);

o

Three reverse roll coating system for hot-melt resin application;

o

Oven with temperature zones (80–150 °C);

o

Calender rollers (gap adjustable to 0.01 mm precision);

o

Lamination section with heated nip rolls.

The experiments conducted for this study are presented in

Table 1. Experimental setup for conducted tests and individual experiments for each of the prepreg processing methods within this study.

Test	Method	Reinforcement	Experimental Run/Details	Notes
I	hot-melt	UD fibers	Exp. 1–2: film production; Exp. 3: UD prepreg	First application of hot-melt process
II	hot-melt	UD fibers	Exp.1–2: UD prepreg	Confirmation of process reproducibility
IV	hot-melt	UD fibers	Exp.1: UD prepreg	Repeatability test
III	solvent-based	UD fibers	Exp.1: UD prepreg (PI:DMAC 50:50)	Process with standard paper substrate
V	solvent-based	UD fibers	Exp.1: UD prepreg (paper) Exp.1: UD prepreg (Kapton film)	Comparison of substrates (paper vs. Kapton-polymer)
VI	solvent-based	UD fibers	Exp.1: UD prepreg (PI:DMAC 45:65)	Comparison of changing resin/solvent ratio
VII	solvent-based	woven fabric	Exp.1: fabric prepreg with sufficient resin bath. Exp.2: fabric prepreg with low resin solution bath	Comparison of resin bath level on prepreg quality

.

2.3. Methods for Impregnation

Two different technological approaches were applied:

• The hot-melt method (presented in Figure 2) was conducted with composition of PI resin/DMAC = 70:30 (weight ratio). This ratio provided an optimal viscosity (~3000 mPa·s) for a film formation using the three-roll reverse coating unit, enabling uniform coating without excessive solvent retention. The viscosity of resin–hot-melt mixtures was measured using a Brookfield DV3T rheometer on preliminary samples to determine optimal processing ranges at 55 °C. In this case, DMAC was used only to decrease the resin viscosity in order for it to be more easily applied to C-fibers, as has previously been reported by other researchers as well [26]. The film formation and impregnation were carried out sequentially within the same production line, without intermediate storage or transfer. This configuration represents an in situ hot-melt process, where the resin film is formed and directly applied to the fibers in a single continuous operation.

  o

  Resin pre-melted and applied as thin film.

  o

  Impregnation into carbon fibers through heated calendering. The hot-melt film coating was carried out at a roller temperature of 55 °C, which ensured optimal resin flow and adhesion to the paper substrate. The subsequent drying and imidization stages were performed in two heating zones: in oven 1 at 100 °C and in oven 2 at 150 °C. In oven 2, an active ventilation system was engaged to facilitate the removal of residual solvent vapors and prevent their condensation on the film surface. Prior to entering oven 2, the upper paper liner was removed to allow unobstructed evaporation of the solvent. These controlled temperature and ventilation conditions provided uniform resin distribution and ensured that the volatile content remained below 1.5%.

  o

  Minimal solvent evaporation.

• Solvent-based method (shown in Figure 3): PI resin and DMAC (weight ratio of 50:50):

  o

  Carbon fibers/fabrics passed through resin bath;

  o

  Solvent evaporation in multi-zone oven (100–120 °C);

  o

  Final consolidation through heated calenders.

As shown in Figure 2 and Figure 3, the customized impregnation/coating units developed by Mikrosam (Prilep, North Macedonia) include controlled unwinding, coating, drying, and rewinding sections. They illustrate the modular layout of the system, allowing the flexible adjustment of parameters such as tension, temperature, and coating speed. Representative processing parameters were carefully monitored during each of the experimental runs, as presented in Table 1.

2.3.1. Hot-Melt Process

Table 2. Input parameters for tests I, II, and IV and individual experimental setups (hot-melt method, UD prepregs, PI: DMAC = 70%:30%).

Input Parameters	I Tests			II Tests		IV Tests
	Exp. 1	Exp. 2	Exp. 3	Exp. 1	Exp. 2	Exp. 1	Exp. 2
Notes	Film producing	Film producing	UD prepreg	UD prepreg	UD prepreg	UD prepreg	UD prepreg
Number of fibers	/	/	29	30	30	30	30
Unwinder tension paper [N]	45 N 1.5 N/cm	60 N 2 N/cm	36 N 1.2 N/cm	45 N 1.5 N/cm	45 N 1.5 N/cm	36 N 1.2 N/cm	45 N 1.5 N/cm
Rewinder tension—prepreg/paper [N]	60 N 2 N/cm	60 N 2 N/cm	135 N 4.5 N/cm	150 N 5 N/cm	150 N 5 N/cm	150 N 5 N/cm	150 N 5 N/cm
Tension fibers (N)	/	/	87 N (3 N per fiber)	90 N (3 N per fiber)	90 N (3 N per fiber)	90 N (3 N per fiber)	90 N (3 N per fiber)
S1/S2 gap [mm]	0.15 (0.1)	0.15 (0.1)	0.15 (0.1)	0.185 (0.15)	0.185 (0.15)	0.1	0.1
S1/S2 speed ratio	1:0.1	1.5:0.0	1.5:0.15	1.5:0.1	1.5:0.1	1.5:0.1	1.5:0.1
S1/M speed ratio	1/1	1/1	1/1	1/1	1/1	1/1	1/1
Pressure M [bar]	4 bar	4 bar	4 bar	4 bar	4 bar	4 bar	4 bar
Coating Unit and Callender 1 [°C]	55	55	55	55	55	55	55
Film width [mm]	115 mm	115 mm	115 mm	120 mm	120 mm	120 mm	120 mm
Oven 1/Calender Unit 2 [°C]	100	100	100	100	100	100	100
Oven 2/Calender Unit 3 [°C]	150	150	150	150	150	150	150
Ventilation flow rate Oven 2 [m3/h]	600	650	650	650	650	800	900
Gap adjustment at Calender Unit 1 [mm]	/	/	0.3	0.3	0.3	0.3	0.3
Gap adjustment at Calender Unit 2 [mm]	/	/	0.25	0.25	0.35	0.35	0.35
Gap adjustment at Calender Unit 3 [mm]	/	/	0.3	0.3	0.35	0.35	0.35
Gap adjustment at Master-Nip roller/Pulling Unit 4	/	/	0.35	0.35	0.35	0.25	0.25
Gap adjustment at Lamination Unit [mm]	/	/	0.4	0.35	0.35	0.4	0.4
Line speed	1 m/min	1–1.5 m/min	1.5 m/min	1.5 m/min	1.5 m/min	1.5/2	3
Paper type/width	92 g/m2/30 cm	92 g/m2/30 cm	92 g/m2/30 cm	92 g/m2/30 cm	92 g/m2/30 cm	92 g/m2/30 cm	92 g/m2/30 cm

Note: The first two runs within test I (exp.1 and 2) are performed for film manufacture, as shown in Figure 4.

presents detailed information about the hot-melt process in terms of adjusted parameters that are employed during the manufacturing of prepreg material. During the process setup, particular attention was given to tension control of both the paper carrier and the carbon fiber tows. The strain of the paper substrate and fibers were aligned based on their respective elastic moduli to prevent wrinkling or fiber buckling during coating and transport through the processing line. This ensured a uniform resin film transfer and stable prepreg formation. The selection of process parameters (line speed, roller gap, roller speed ratio) was based on preliminary Design-of-Experiment (DOE) studies previously conducted for hot-melt film formation (Samardjioska et al., 2023) [28]. The chosen values maintained adequate resin pick-up and void contents below 1.5%. These parameters correspond to typical industrial prepregging ranges for polyimide systems. Additionally, the oven temperatures and ventilation rates were optimized to achieve efficient solvent removal and stable film characteristics across the production line.

Figure 4 shows the hot-melt resin film production on a paper substrate prior to the impregnation process. The film exhibits a uniform and consistent width (~115 mm) maintained along the full length of the substrate. Although the thickness is quantitatively measured by standard methods, visually it confirms a homogeneous resin layer across the width, indicating a stable film formation within the designed parameters (Table 2).

2.3.2. Solvent-Based Process

Table 3. Input parameters for tests III, V, and VI with individual experimental setups (Solvent-based method, UD prepregs; PI: DMAC = 50%:50%).

Input Parameters	III Tests	V Tests		VI Tests
	Exp. 1	Exp. 1	Exp. 2	Exp. 1
Number of fibers	30	30	30	30
Unwinder tension paper [N]	45 N 1.5 N/cm	36 N 1.2 N/cm	45 N 1.5 N/cm	30 N 1 N/cm
Rewinder tension—prepreg/paper [N]	130 N 4.5 N/cm	120 N 4 N/cm	135 N 4.5 N/cm	120 N 4 N/cm
Tension fibers (N)	90 N (3 N per fiber)	90 N (3 N per fiber)	90 N (3 N per fiber)	90 N (3 N per fiber)
Resin bath temperature [°C]	40	40	40	40
Oven 1/Calender Unit 2 [°C]	100	100	100	100
Oven 2/Calender Unit 3 [°C]	150	120	120	150
Ventilation flow rate Oven 1,2 [m3/h]	700	800	900	800
Calender Unit 1 gap adjustment [mm]	0.3	0.4	0.4	0.3
Calender Unit 2 gap adjustment [mm]	0.35	0.35	0.35	0.35
Calender Unit 3 gap adjustment [mm]	0.35	0.32	0.32	0.35
Master-Nip roller/Pulling Unit 4 gap adjustment [mm]	0.35	0.25	0.25	0.35
Lamination Unit gap adjustment [mm]	0.4	0.45	0.45	0.4
Line speed [m/min]	1/2	2	3	2
Resin system (PI:DMAC)	50:50	50:50	50:50	45:55
Kapton Film/width	/	/	92 g/m2/30 cm	/
Paper type/width	92 g/m2/30 cm	92 g/m2/30 cm	/	92 g/m2/30 cm

presents detailed information about solvent-based impregnation process using UD C-fibers as reinforcement and

Table 4. Input parameters for Test VII and suitable experimental setups (solvent-based method, woven fabric as reinforcement).

Input Parameters	VII Tests
	Exp. 1	Exp. 2 (Low Resin Level in Bath)
Carbon fabric width [mm]	30	30
Unwinder tension fabric [N]	90 N 3 N/cm	90 N 3 N/cm
Unwinder tension film [N]	31 N 0.9 N/cm	35 N 1 N/cm
Rewinder tension—prepreg/paper [N]	135 N 4.5 N/cm	150 N 5 N/cm
Resin bath temperature [°C]	40	40
Oven 1/Calender Unit 2 [°C]	100	100
Oven 2/Calender Unit 3 [°C]	120	150
Ventilation flow rate Oven 2 [m3/h]	900	850
Calender Unit 1 gap adjustment [mm]	/	/
Calender Unit 2 gap adjustment [mm]	0.4	0.4
Calender Unit 3 gap adjustment [mm]	0.45	0.45
Master-Nip roller/Pulling Unit 4 gap adjustment [mm]	0.25	0.25
Lamination Unit gap adjustment [mm]	0.45	0.45
Line speed [m/min]	3	2.5
Carbon fabric/width	200 g/m2/30 cm	200 g/m2/30 cm
Kapton Film/width	50 g/m2/35 cm	50 g/m2/35 cm
Polyethylene Film/width	50 g/m2/35 cm	50 g/m2/35 cm
Paper type/width	92 g/m2/35 cm	92 g/m2/35 cm

presents the same process when woven C-fabric is used as reinforcement for the prepreg composite material. The selection of process parameters (line speed, fiber, and paper tension) was based on chosen values ensuring balanced strain between the carrier paper and carbon fibers to prevent wrinkling or fiber misalignment, while maintaining adequate resin pick-up and void content below 1.5%. The parameters correspond to typical industrial prepregging ranges for polyimide systems. Preliminary experiments and viscosity measurements of PI: DMAC solutions using a Brookfield DV3T rheometer confirmed that the selected ratios are suitable for the intended processes. The 50:50 ratio gives a viscosity of ~350 mPas at 40 °C, appropriate for solvent-based impregnation. The 55:45 ratio results in ~1 500 mPas, which is too high for solvent-based methods and too low for hot-melt applications, making it less suitable. The 45:55 ratio was also evaluated to study the effect of increased solvent content on drying efficiency and resin distribution under controlled oven conditions. These ratios were chosen to represent standard solvent contents typically used for PI prepregs, ensuring better fiber wetting, complete resin penetration through the reinforcement structure, and an optimal balance between impregnation quality, solvent removal, and industrial processability. The oven zones were maintained at 100 °C and 150 °C with active ventilation, ensuring complete solvent evaporation while maintaining process safety within the defined LELs (Lower Explosion Limits).

Figure 5 illustrates the main stages of the solvent impregnation process: immersion of the fiber tows into the resin bath, subsequent solvent evaporation in the oven, consolidation by calendering, and final lamination. The sequence demonstrates how solvent-based impregnation ensures deep wetting of fibers, followed by drying and compaction to achieve uniform resin distribution through the sample.

Figure 6 illustrates the solvent-based impregnation process for woven carbon fabrics. The left-side image in Figure 6a (Exp.1) shows the fabric prepreg positioned between two protective films after passing through the impregnation line; it indicates a good resin penetration under properly adjusted parameters and sufficient resin level in the bath. Figure 6b (Exp. 2) shows the process under reduced bath conditions, where insufficient resin availability affects the uptake and compromises coating uniformity.

2.4. Prepreg Characterization

To evaluate the quality of the produced prepregs, the following parameters were measured:

o

Final prepreg width (mm), measured by digital caliper (Mitutoyo, Kawasaki, Japan) along prepreg length.

o

Resin Content (%) of the prepreg specimens was determined in accordance with ASTM D3171, Method I, Procedure G [29]. This method involves heating the specimen in a furnace under controlled conditions until the resin matrix is completely removed and the reinforcing fibers remain. The mass of the specimen before and after the burn-off was recorded, allowing for the calculation of resin and fiber percentages. For this purpose, the following equipment was used:

o

Muffle furnace capable of reaching 600 °C with accurate temperature control;

o

Analytical balance 204, model CPA2245-0CE (Sartorius AD, Göttingen, Germany);

o

High-temperature-resistant ceramic crucibles;

o

Tweezers and metal tongs for handling heated crucibles

The procedure is adapted from the standard in order to match the material system under study; specimens were prepared and placed into ceramic crucibles for burning. The furnace was set to 600 °C and specimens were heated for approximately 2 h until all resin was removed (as per Procedure G). The specimens were cooled to ambient temperature in a desiccator and mass difference (between final and initial specimen mass) was recorded. Resin content was calculated from the mass difference according to the ASTM D3171.

o

Volatile Content (%) of the prepreg was determined following ASTM D3530 [30]. Specimens were weighed before and after exposure to elevated temperature in an air-circulating oven to remove solvents, retained moisture, and low molecular weight constituents, and the mass loss was expressed as volatile content percentage.

The (adapted) procedure was conducted as follows: prepreg specimens were cut, weighed, and placed in the oven at 150 °C for 2 h. After cooling to room temperature, final mass was recorded. The percentage of mass loss (volatile content) was calculated. The temperature/time conditions were selected based on the material system and within the limits specified by ASTM D3530.

o

Prepreg weight (gsm)—100 mm × 100 mm specimens were cut and weighed on an analytical balance (Analytical balance 204, model CPA2245-0CE, Sartorius AD, Göttingen, Germany).

o

Film quality was evaluated visually and by optical microscopy (Carl Zeiss Jena, Jena, Germany) [this was conducted only for the internal in-house purpose—to inspect the film quality and continue with its processing for prepreg].

o

Representative samples from both hot-melt and solvent impregnation trials were examined by Scanning Electron Microscopy (SEM, JEOL Ltd., Tokyo, Japan) to observe the resin distribution and fiber–matrix interface. Prior to SEM imaging, the samples were mounted on aluminum stubs using conductive carbon tape and sputter-coated with gold (10 nm) to prevent charging. The images were observed under a magnification of 50× (the scale bars 10–100 µm) and at an accelerating voltage of 10 kV. The obtained micrographs illustrate the typical impregnation morphology of PI-based prepregs, and they are used to represent characteristic features such as fiber wetting, resin continuity, and absence of voids, which are consistent with the structural integrity expected for the materials studied.

Each measurement was performed three times under identical process conditions and using calibrated equipment. Selected woven prepregs were analyzed by SEM to assess the local impregnation quality and fiber–matrix adhesion.

3. Results and Discussion

In general, results from this study demonstrated a clear difference between the two impregnation methods. The hot-melt technology consistently enabled a stable and relatively high resin content (34–35 wt%), low volatile levels (<1.5%), and good film uniformity across the unidirectional prepregs. On the other hand, the solvent-based technology provided deeper fiber impregnation, slightly lower volatile contents (~0.3–0.5%), and greater adaptability when processing both UD fibers and woven fabrics into prepregs. The following sections present a detailed discussion of the results, organized according to the reinforcement type and the impregnation method used in this study. Although the experiments were performed on short-run trials due to industrial constraints, the repeated tests (I–VII) confirmed high reproducibility and stable prepreg quality, indicating reliable process control.

3.1. Impregnation with Hot-Melt Technique (UD Prepreg)

The hot-melt impregnation method was applied to unidirectional (UD) carbon fibers using a PI: DMAC ratio of 70:30. The process was performed in situ, where the resin film was formed and immediately applied onto the fibers within the same continuous production line, ensuring stable process control and uniform impregnation quality. The output data are presented in

Table 5. Output parameters for Test I, II, and Test IV (Hot-melt process, UD prepregs, PI: DMAC = 70:30).

	Test I Exp. 3				Test II Exp. 1				Test IV Exp. 1
Parameter	Sample 1	Sample 2	Sample 3	Average	Sample 1	Sample 2	Sample 3	Average	Sample 1	Sample 2	Sample 3	Average
Width (mm)	114.9	115.2	115.1	115.1	119.7	120.2	120.3	120.1	119.8	120.3	120.1	120.1
Volatile [%]	1.2	1.3	1.3	1.27	1.3	1.5	1	1.27	2.1	1.35	1	1.47
Resin [%]	34.7	35.1	35.0	34.98	34.6	35.3	35.0	34.97	34.6	35.3	35.1	35.0
UD prepreg [gsm]	309	311	310.5	310.2	306.0	309.0	307.5	307.5	306.0	309.0	308.0	307.8

and are summarized in the text that follows.

In all cases, the prepregs demonstrated uniform film distribution, stable resin content between 34 and 35%, and very low volatile contents (<1.5%). These parameters indicate that the hot-melt process enables high accuracy in controlling resin content and overall prepreg quality, particularly for UD reinforcements. Furthermore, the hot-melt process proved effective in producing dense and uniform resin films, which are particularly suitable for UD prepregs for Automated Fiber Placement (AFP)/Automated Tape Laying (ATL) applications, where precise resin control and film uniformity are critical [31,32]. Ishida et al. [18] also presented a hot-melt impregnation method for CF in PA6 that was initially dissolved in suitable solvents to lower the resin viscosity. The fabricated composites exhibit a better impregnation quality than those fabricated without the use of solution impregnation. Namely, the impregnation quality and flexural properties improved significantly in the composites impregnated with the solution, while the flexural strength was decreased.

Figure 7 illustrates the formation of unidirectional (UD) prepregs using the hot-melt process. Figure 7a (Test I, Exp.3) shows the initial formation stage, where the resin film is applied onto the fiber tows with paper substrate support. Figure 7b (Test II, Exp.1) demonstrates stable impregnation conditions with uniform film coverage. Figure 7c (Test II, Exp.1) shows the resulting prepreg wound onto a roll, confirming consistent thickness and surface quality. Figure 7d (Test IV, Exp.1) highlights the repeatability of the process, showing prepregs with uniform resin distribution and well-defined edges, free of defects such as wrinkles or resin-rich areas.

More specifically, the main findings for UD prepregs prepared using the hot-melt process are summarized as follows:

(i)

The UD prepregs had with a width of ~115 mm and a stable resin content of ~35%. The obtained prepregs had low volatile content (~1.27%) and a uniform prepreg weight (~310 gsm). As shown in Figure 7a, the hot-melt process enabled the formation of a continuous resin film with stable coating without visibly dry spots, although slight resin overflow near the edges was initially observed and corrected by adjusting side guides. The resulting UD prepreg, presented in Figure 7b, confirmed that the applied film provided uniform impregnation of the fiber tows and stable overall quality.

(ii)

The hot-melt method can be reproduced with high process stability. The prepregs had a width of close to 120 mm, stable resin content (~35%), low volatiles (~1.27%), and uniform prepreg weight (~308 gsm), Figure 7c (Test II).

(iii)

An excellent repeatability of the hot-melt impregnation process was proven. Figure 7d illustrates the UD prepreg (Test IV), showing well-defined edges, consistent width (~120 mm), and uniform resin coverage across the tape, which verifies the repeatability and quality of the hot-melt process.

A representative SEM micrograph (Figure 8) of the UD prepreg produced by the hot-melt method shows well-distributed resin between the fiber tows and good fiber–matrix adhesion, confirming proper impregnation, uniform resin coverage, full fiber wetting along the unidirectional alignment, and absence of dry spots, as well as confirming proper impregnation without detectable voids.

3.2. Impregnation with Solvent-Based Technique (UD Prepreg)

The solvent impregnation method was applied to unidirectional (UD) carbon fibers using PI: DMAC ratios of 50:50 and 45:55. The output data are presented in

Table 6. Output parameters (solvent method, UD prepregs, PI:DMAC = 50:50 and 45:55).

	Test III			Test V						Test VI
	Exp 1			Exp 1			Exp 2			Exp 1
Parameter	Sample 1	Sample 1	Average	Sample 1	Sample 2	Average	Sample 1	Sample 2	Average	Sample 1	Sample 1	Average
Width (mm)	119.3	119.5	119.4	119.7	119.6	119.7	119.9	120.1	120.0	119.5	120.2	119.9
Volatile [%]	0.5	0.3	0.4	0.4	0.3	0.35	0.5	0.4	0.45	0.5	0.4	0.45
Resin [%]	34.5	34.6	34.5	34.7	35.0	34.9	35.2	34.6	34.9	31.8	32.2	32.0
UD prepreg [gsm]	305.5	306	305.8	306.5	307.5	307	308.5	306	307.3	293.5	295.0	294.3

.

In all cases, solvent impregnation of UD prepregs provided slightly lower but deeper penetrating resin distribution within the fiber bundles compared to the hot-melt method. The residual volatiles were minimal (<0.5%), and the resulting prepregs exhibited good handling quality. This indicates that solvent impregnation is more flexible when different reinforcement substrates or process variations are required, although it does not always achieve the same precision in resin content as the hot-melt process.

As shown in Figure 9, the solvent impregnation process with PI:DMAC = 50:50 produced UD prepregs with uniform resin coverage across the full width (~120 mm). The surface appears smooth and defect-free, without visible dry spots or resin-rich streaks, which confirms effective fiber wetting and stable process parameters.

Figure 10 shows the UD prepreg produced on Kapton film (Test V, Exp.2). Compared to the method with a paper substrate (shown in Figure 9), the Kapton-supported prepreg exhibits a smoother, glossier surface and is defect-free. This demonstrates the positive influence of substrate choice on surface uniformity and overall prepreg quality.

Similarly, Kostadinoska et al. [16] concluded that the final prepreg properties of prepreg based on epoxy-glass fibers are dependent on the impregnation process and the processing parameters fed to the system. Moreover, Sun et al. presented a custom-designed solution impregnation for producing filaments with high fiber content (65%), low porosity, and outstanding mechanical properties intended for 3D printing needs [12]. Similarly to the current study, controlled processing parameters, like temperature, speed, and concentration, were the most important for customizing the final properties of the composite material.

Findings for the manufacture of UD prepregs prepared by solvent-based impregnation are summarized as follows:

o

UD prepregs with a width of ~120 mm were produced with a resin content of ~34.5%, and volatile content ~0.4%, indicating almost complete solvent removal. The prepregs demonstrated uniform weight (~306 gsm) and good fiber impregnation. As illustrated in Figure 9 (test III, exp. 1), the solvent-based process allowed the resin to penetrate deeply and evenly, resulting in a smooth surface along the prepreg width. No major defects or dry areas were visible, confirming good wetting of UD fibers with the 50:50 PI: DMAC solution.

o

The process stability of the solvent method was confirmed. Namely, in test V, Exp.1 (standard paper substrate), the prepregs had similar resin content (~35%) and uniform weights (~307 gsm), while in test V, Exp. 2, where Kapton film was used instead of paper, the prepregs showed improved film quality and reduced surface defects, proving the influence of substrate choice on prepreg uniformity. Figure 10 presents prepreg produced on Kapton film, with a smoother and more uniform resin surface than the one produced on paper. This highlights the substrate’s influence on film quality and defect formation. Visual inspection during and after processing confirmed fewer surface defects and better macroscopic uniformity for Kapton-supported prepregs. Although no microscopic analysis was performed, consistent weight and surface appearance clearly demonstrate the positive effect of the substrate on the prepreg quality.

o

Moreover, the study was performed with a different PI: DMAC ratio of 45:55 (test VI, exp. 1), which resulted in a lower resin content (~32%) and correspondingly lower areal weight (~294 gsm). This highlights the sensitivity of the solvent-based process to the resin/solvent ratio, but also confirms its capability for producing thinner, lighter prepregs when required.

To complement the macroscopic analysis, SEM examination was performed to evaluate the internal impregnation quality. As observed in the SEM image (Figure 11), the solvent-impregnated prepreg exhibits full fiber wet-out and homogeneous resin distribution across the tow bundle, without fiber damage or micro-voids, confirming that the impregnation mechanisms discussed above correspond to the observed microstructure.

Microscopic analysis (SEM, Figure 11) of the cross-section of the solvent-impregnated UD prepreg confirmed deeper resin penetration between the individual carbon fiber tows compared to the hot-melt-processed prepreg (Figure 8). The resin phase appears to be continuous and well-distributed, indicating efficient wetting and low void content. This microstructural evidence supports visual observations and verifies that the solvent-based technique provides enhanced fiber impregnation uniformity.

3.3. Impregnation with Solvent Technique (Woven Fabric)

The solvent impregnation method was further applied to woven fabric reinforcements using a PI: DMAC ratio of 50:50.

Table 7. Output parameters (solvent method, woven fabric, PI: DMAC = 50:50).

	Test VII
	Exp. 1				Exp. 2
Parameter	Sample 1	Sample 2	Sample 3	Average	Sample 1	Sample 2	Sample 3	Average
Width (mm)	300.0	300.0	300.0	300.0	300.0	300.0	300.0	300.0
Volatile [%]	0.6	0.7	0.6	0.63	0.8	0.9	0.7	0.8
Resin [%]	37.3	38	37	37.43	27.9	25	27	26.63
UD prepreg [gsm]	318.8	322.6	317.5	319.63	274.5	265	270	269.83

presents the original output data.

In all cases, the solvent method demonstrated higher suitability for woven fabrics than the hot-melt technique. The lower viscosity of the resin/solvent mixture allowed deeper penetration into the inter-yarn spaces, reducing voids and improving overall impregnation uniformity. Therefore, while the hot-melt method is advantageous for UD prepregs where resin precision is essential, the solvent process provides superior adaptability for woven fabrics and complex fiber architectures [26,33].

Figure 12 shows successful impregnation of woven fabric, with uniform resin distribution across the full width of the prepreg. The absence of voids and dry areas confirms the suitability of solvent impregnation for fabric reinforcements.

As shown in Figure 13, when the resin bath level was insufficient, areas with lower resin uptake and dry spots became visible across the fabric. This highlights the critical importance of maintaining optimal bath conditions during solvent impregnation.

These results could be summarized as follows:

o

A fabric prepreg with a width of ~300 mm, resin content of ~37.5%, volatiles ~0.6%, and weight ~320 gsm was produced. The prepreg quality was excellent, with uniform resin distribution across the full width. As shown in Figure 12, the solvent-based process ensures efficient wetting of woven structures.

o

When the resin level in the bath was low, it resulted in reduced resin content (~27%) and lower prepreg weight (~270 gsm). These results clearly illustrate the strong dependence of fabric impregnation quality on maintaining the proper resin bath level. As illustrated in Figure 13, insufficient solvent in the bath led to regions with lower resin uptake and visible dry spots across the fabric, emphasizing the critical need for proper bath conditions during solvent-based impregnation.

The solvent impregnation of woven fabrics was further examined by SEM analysis to evaluate the resin distribution and fiber–matrix adhesion quality. The obtained micrographs (Figure 14, Figure 15 and Figure 16) reveal resin wetting across the carbon fiber bundles, confirming that the impregnation mechanisms discussed above correspond to the observed microstructure. The presented SEM micrographs (Figure 14a–c) illustrate three characteristic regions observed during the impregnation process: (a) dry fabric surface, (b) partially impregnated region, and (c) fully impregnated region with uniform resin distribution and complete fiber wetting.

These microstructural observations demonstrate how variations in resin availability in the impregnation bath affect the quality of fiber wetting and resin distribution. Regions with insufficient resin uptake—resulting from a low resin level in the bath—show partial wetting and locally dry areas, whereas fully impregnated regions exhibit continuous and uniform resin coverage. These findings are consistent with the mechanisms discussed in the previous sections and further confirm the critical importance of maintaining an adequate resin bath level during solvent-based impregnation.

In contrast, Figure 15 shows woven prepregs impregnated under optimal process parameters, where the resin is fully distributed across the yarn cross-section, ensuring complete wetting and intimate fiber–matrix bonding. These images confirm that maintaining proper resin content and drying conditions is critical for achieving void-free and well-impregnated woven prepregs. Namely, SEM micrographs (Figure 15) of woven fabric prepregs obtained by solvent impregnation show uniform fiber wetting and continuous resin phase without voids, confirming complete impregnation quality.

3.4. Comparative Analysis of Used Methods for Prepreg Manufacturing

The comparative evaluation of the two impregnation methods, hot-melt and solvent-based, reveals clear differences in their suitability for different reinforcement architectures and process stability. It should be noted that the different PI:DMAC ratios reflect process-specific requirements for resin viscosity control rather than unequal experimental conditions.

For unidirectional (UD) prepregs, the hot-melt method demonstrated superior precision in resin control, achieving stable resin content in the range of 34–35% with relatively low volatiles (<1.5%). The prepregs exhibited uniform areal weights (~307–310 gsm) and consistent widths of ~115–120 mm across repeated tests. These results confirm the reproducibility of the hot-melt technique and its suitability for applications where tight control of resin content is critical, such as Automated Fiber Placement (AFP) and Automated Tape Laying (ATL) [31,34].

In contrast, the solvent-based method produced UD prepregs with slightly lower resin contents (~32–35%), and minimal but detectable volatile residues (~0.3–0.5%). The process showed higher sensitivity to the resin/solvent ratio (e.g., Test VI with PI: DMAC = 45:55); when the resin content decreased, the areal weight showed larger variability, compared with the process where a higher resin/solvent ratio was used (50:50). However, the solvent-based method provided deeper impregnation of the fiber bundles, reducing the risk of void formation and enhancing the quality of the resin–fiber interface. The use of Kapton film as substrate further improved the surface uniformity, indicating flexibility of the method in adapting to different process configurations.

For woven fabric prepregs, the solvent method proved to be very effective. Namely, it achieved a stable resin content (~37.5%) with uniform distribution across the ~300 mm wide prepreg and minimal volatiles (~0.6%) with this trial. When the resin bath level was low (around 20% in the bath), the resin content dropped (~27%), which directly affected the areal weight and final prepreg quality. This clearly highlights the critical influence of maintaining optimal bath conditions during solvent impregnation.

Overall, the hot-melt technique ensures high accuracy and repeatability for UD reinforcements, whereas the solvent-based process provides superior adaptability and penetration for woven structures. This observation is supported by process trials, where attempts to impregnate woven fabrics using the hot-melt setup resulted in partial wetting and surface resin accumulation, due to the inherently higher viscosity of thermosetting PI and the compact nature of the fabric architecture. It is worth noting that the PI resin family used in this study (PI-1 to PI-4) was gradually optimized by the manufacturer to tailor the viscosity and imidization degree for different processing routes. Such formulation tuning is critical when selecting the appropriate impregnation technique for a given reinforcement type. Therefore, with the currently employed PI formulation used in these experiments, the hot-melt configuration is considered to be less suitable for woven reinforcements, while the solvent-based process ensures complete impregnation and uniform resin distribution. This observation is consistent with the industrial practice of applying solvent-based routes for dense woven reinforcements requiring full fiber wetting. Future work may focus on developing modified PI systems with reduced viscosity to enable more effective hot-melt impregnation of woven architectures. The process parameters were repeatedly verified under the same operating conditions, showing consistent resin content and areal weight values across trials, which confirms the high reproducibility of both impregnation methods. The choice of impregnation method should therefore be aligned with the type of reinforcement and the target application of the final composite.

The cross-section SEM micrograph of the final woven prepreg (Figure 16) demonstrates a continuous resin phase through the full thickness, with uniform wetting and no visible voids, confirming the overall quality and consistency of the produced prepregs. The image confirms good adhesion between the fibers and the matrix, indicating complete wetting and stable process conditions.

It should be noted that the present study is limited to the prepreg manufacturing stage. The variations in resin content, volatile content, areal weight, and width did not exceed ±2% of the mean value, indicating a high level of process reproducibility and measurement precision consistent with industrial standards. The mechanical performance of the final composite parts will be addressed in future research, once the optimized process parameters and PI formulations are implemented in full-scale laminate production.

4. Conclusions, Future Perspectives, and Challenges

This study comparatively evaluated the hot-melt and solvent-based impregnation methods for manufacturing polyimide/carbon fiber prepregs, using both unidirectional fibers and woven fabrics as reinforcements. The study has analyzed the processing parameters and resulting material quality. The main findings are summarized as follows:

(1)

The hot-melt impregnation method ensures stable and reproducible resin content, minimal volatiles, and excellent film uniformity. These characteristics make it particularly suitable for manufacturing unidirectional (UD) prepregs used in high-performance structural applications, including Automated Fiber Placement (AFP) and Automated Tape Laying (ATL) [35,36].

(2)

The solvent impregnation method allows for deeper and uniform resin penetration, particularly in woven reinforcements, ensuring better impregnation of complex fabric architectures compared to the hot-melt based method. It offers higher adaptability to process variations, although it is more sensitive to resin/solvent ratio as well as bath conditions.

(3)

The selection of impregnation method should therefore be guided by reinforcement type and by the end-use application:

o

Hot-melt for UD prepregs where precision in resin control is critical. The integrated (in situ) hot-melt setup demonstrated high process stability and reproducibility, confirming its potential for scalable industrial prepreg production.

o

Solvent impregnation for woven fabrics, where deep resin wetting and uniformity are essential.

Future research will focus on the optimization of process parameters and formulation tuning for thermosetting PI systems. Adjusting resin viscosity for different reinforcement architectures (UD and woven) and exploring hybrid impregnation routes combining hot-melt and solvent stages could further enhance impregnation quality and process robustness. These directions are built directly upon the comparative results discussed in this study and reflect the industrial need for adaptable, high-temperature-resistant prepreg technologies:

(i)

Combining the advantages of both dry and wet methods, e.g., wetting the reinforcement material (woven fabric) using solvent-based technology and continuing with hot-melt processing to tailor the final properties of the prepreg for high-tech targeted applications [32,37].

(ii)

Use of hybrid novel systems in terms of thermoset–thermoplast polymer matrices and/or novel (modified) processes where improved thermal and mechanical properties of final composites are achieved, but also material/energy savings are assumed [37,38,39,40,41,42,43,44,45,46]. Moreover, future research could explore the use of hybrid materials by implementing nanomaterials in composite materials to customize the mechanical, thermal, barrier, or physico-chemical properties of the final composite material’s properties. These hybrid matrices broaden the potential applications of high-performance composites. For example, in a hybrid system where carbon nanotubes (CNTs) were added, enhanced thermo-mechanical properties were obtained; more specifically, the interlaminar fracture toughness, interlaminar shear strength, and flexural strength, at a loading concentration of only 0.5% CNT addition to CF/PI composite, were improved [43,47]. The nano-inclusions can reinforce/bridge the micro gaps between CF and CF/PI, enhancing stress transfer and inhibiting crack propagation.

(iii)

Incorporation of self-healing agents in the form of microcapsules or reversible polymer networks within prepregs that will enable the autonomous repair of microcracks; this will further advance composite materials. In addition, this approach extends service life and reduces the maintenance requirements of composite parts, which is particularly important in aerospace and energy sector applications [48,49].

(iv)

Developing in-line monitoring systems for process control. In situ prepregging concepts are already demonstrated by some researchers, but they need further development because significant challenges remain [50,51,52].

(v)

Tailored Fiber Placement (TFP) allows the precise deposition of UD fibers or prepreg tapes along optimized load paths—it is a promising technology that improves structural performance, reduces material waste, and enables efficient use of prepregs in complex components [53,54].

(vi)

The use of natural fibers as reinforcements is well-aligned with global sustainability goals; namely, natural fibers such as flax, hemp, or jute are being investigated as alternatives or in hybrid combinations with carbon fibers within prepreg composites [55]. In this context, bio-based polymer matrices are also being investigated for a wide range of structural composites, or at least to replace a part of the fossil-based polymer matrices in the prepreg composition. They offer partially bio-based, cost-effective, and recyclable solutions, while reducing environmental impact [3,55]. Recycling opportunities for novel types of composite materials need to be further explored, considering zero-waste technologies and improved socio-economic aspects [56,57,58,59,60,61,62,63].

Overall, this study, with future perspectives, highlights the importance of research into composite materials, particularly prepregs, which require deeper understanding and further design and improvement. This study also emphasizes the potential for future applications that are mainly dependent on the nature of used materials, processing methods, and processing parameters, all of which affect the final quality and properties of the prepregs, and consequently the parts that they are ultimately made into.