Development and Experimental Validation of a Filament-Assisted Chemical Vapor Deposition (FACVD) Reactor Using a Plastic Chamber

Abstract

This study explored the feasibility of using a plastic vacuum chamber for the Filament-Assisted Chemical Vapor Deposition (FACVD) of polymer thin films. Traditional chemical vapor deposition (CVD) methods often require high vacuum and elevated temperatures, which limit their use for heat-sensitive and flexible substrates. FACVD enables polymer deposition under mild vacuum and temperature conditions, providing an opportunity to utilize plastic vacuum chambers as cost-effective and easily machinable alternatives to metallic chambers. In this study, a custom-designed acrylic chamber was fabricated and integrated into an FACVD system. Glycidyl methacrylate (GMA) and tert-butyl peroxide (TBPO) were considered as the monomer and initiator, respectively, for creating thin films under a low-temperature and moderate-vacuum deposition process. Polymeric film (pGMA) contains reactive epoxy groups that allow versatile post-polymerization modifications and are widely applied in coatings and biomedical fields. Preliminary experiments demonstrated the successful growth of pGMA thin films, with Fourier-transform infrared spectroscopy (FT-IR) and X-ray photoelectron spectroscopy (XPS) confirming the characteristic polymer features, including the disappearance of the C=C stretching band as direct evidence of polymerization. Ellipsometry determines a uniformity of film thickness of approximately 85% for the 4-inch wafers’ area, with deposition rates in the range of 18–26 nm/h. These results highlight the potential of polymer-based chambers as cost-effective and versatile alternatives to advanced vapor-phase polymerization processes.

1. Introduction

Chemical and physical vapor deposition (CVD and PVD, respectively) techniques are widely used in thin-film fabrication in the semiconductor, biomedical, and energy sectors. These methods typically operate under high vacuum and temperatures, often exceeding 300 °C, which limits their compatibility with thermally sensitive substrates such as polymers and plastics [1,2]. Furthermore, conventional PVD suffers from line-of-sight deposition limitations, making it challenging to achieve conformal coatings on substrates with complex geometries [3].

FACVD has emerged as a promising alternative to overcome these challenges. FACVD is a gas-phase polymerization process that operates under mild vacuum (~1 Torr) and low substrate temperatures (~30 °C), allowing direct deposition onto a wide range of heat-sensitive and flexible materials [4]. In FACVD, monomers and initiators are vaporized and introduced into a vacuum chamber where the initiator is thermally decomposed and triggers radical polymerization directly on the substrate surface [5]. This technique enables the fabrication of high-purity conformal polymer coatings with excellent chemical control and substrate versatility [6].

Owing to these advantageous operating conditions, FACVD presents a unique opportunity to utilize polymeric vacuum chambers as viable alternatives to traditional metallic vacuum chambers. Although stainless-steel and aluminum chambers offer structural integrity and thermal resistance, they are costly, opaque, and mechanically restrictive in terms of geometry. By contrast, plastic materials, such as acrylic polymethyl methacrylate (PMMA), are cost-effective, lightweight, and highly machinable alternatives for moderate vacuum deposition processes [7]. Additionally, the low operating pressure and temperature of FACVD reduce concerns related to outgassing and thermal deformation, which typically preclude plastic chamber use [8]. However, compared with metallic chambers, plastic chambers have lower mechanical and thermal stability, making them less suitable for applications that require ultra-high vacuum or high-temperature conditions.

Most previous studies on iCVD and FACVD have concentrated on optimizing metallic chamber designs or precursor delivery systems, such as rotating disk reactors [9], advanced gas inlets for flow distribution [10], and roll-to-roll CVD systems for scalable film fabrication [11]. Similar advancements have also been reported in related vapor-phase techniques, including ALD [12] and PECVD [13], which emphasize reactor design and process optimization to improve film conformality and quality. Extensive mechanistic and application-oriented studies of iCVD have been conducted by Gleason and co-workers, demonstrating the versatility of this method for producing functional polymer coatings [14,15,16]. While these approaches have contributed significantly to improving film quality and process scalability, the feasibility of employing polymer-based vacuum chambers has remained largely unexplored, primarily due to concerns regarding vacuum stability and thermal resistance.

In this context, the present study addresses this research gap by introducing an acrylic-based FACVD chamber and experimentally validating its applicability for polymer thin-film deposition. This approach not only demonstrates that plastic chambers can withstand practical deposition conditions but also highlights their potential advantages in terms of cost, machinability, and design flexibility, thereby representing a novel contribution beyond conventional metallic reactor development.

Accordingly, in this study, a vacuum chamber made of acrylic, a commercially available transparent polymer, was designed and fabricated, and its compatibility with fluorescence-activated cell viability detection was evaluated. Thin films were deposited using tert-butyl peroxide (TBPO) as the initiator and glycidyl methacrylate (GMA) as the monomer. GMA possesses both a methacrylate double bond and a reactive epoxy group, enabling rapid deposition even at low temperatures. The preserved epoxy group is advantageous for post-treatment modification and improved adhesion [17,18]. The resulting films were analyzed using Fourier-transform infrared spectroscopy (FT-IR) and X-ray photoelectron spectroscopy (XPS) to assess their chemical structure, film thickness, and surface properties. This study aimed to demonstrate the viability of low-cost transparent plastic chambers for polymer thin-film deposition via FACVD.

2. FACVD and Experimental Setup

2.1. FACVD Process

FACVD is a vapor-phase polymerization process used to deposit thin polymer films under low-temperature and low-pressure conditions. FACVD proceeds through the generation of radicals via the thermal decomposition of an initiator, followed by the condensation of monomers onto the substrate surface and the induction of surface radical polymerization to grow the thin film.

The step-by-step operating principle of FACVD, as shown in Figure 1, is as follows:

(a)

Radical Generation: Vapor-phase initiators (e.g., TBPO) are thermally decomposed on a filament heated to approximately 220 °C, generating reactive radicals. These radicals remain stable in the low-pressure vapor phase and diffuse toward the substrate surface.

(b)

Monomer Adsorption: The substrate is maintained at approximately 25–30 °C, allowing monomer vapors to physically adsorb onto the surface and increase the local monomer concentration. The adsorbed monomers enhance the surface polymerization efficiency and establish the surface reaction-dominant conditions.

(c)

Surface Polymerization: Vapor-phase radicals react with the vinyl groups of the monomers condensed on the substrate surface, initiating free-radical polymerization. Radicals continuously add to the monomers on the surface, promoting chain growth and forming a thin-film network structure as bonds are formed.

(d)

Thin-Film Growth: As polymerization proceeds, a uniform polymer thin film grows on the surface. The film thickness and composition can be controlled by adjusting the monomer/initiator feed ratio, radical generation rate, substrate temperature, and reaction time, thereby enabling high-purity film deposition without solvent residues [5,19,20].

As illustrated in Figure 1, the initiator (TBPO) decomposes at the heated filament to generate radicals, which subsequently interact with adsorbed GMA monomers on the substrate. These surface reactions promote chain growth and thin-film formation through radical-induced polymerization. In addition, the chemical structures and properties of the initiator and monomer are listed in

Table 1. Details of the initiator and the monomer.

Structure	Name (Abbreviation)	Molecular Weight (g/mol)	Chemical Formula
	Tert-Butyl peroxide (TBPO)	146.23	C8H18O2
	Glycidyl methacrylate (GMA)	142.15	C7H10O3

.

In this manner, high-purity polymeric thin films can be deposited through a continuous process of radical generation, monomer condensation, surface polymerization, and thin-film growth. FACVD can be applied to thermally sensitive substrates and allows precise control of the surface chemical properties while depositing uniform polymeric thin films, making it useful for applications in biotechnology, sensors, and flexible electronic devices [21].

2.2. FACVD System and Chamber Configuration

The FACVD vacuum chamber was designed carefully, considering material selection, mechanical integrity, chemical compatibility, and process monitoring requirements. The chamber body was fabricated from polymethyl methacrylate (PMMA), which offers high optical transparency, excellent machinability, and sufficient mechanical strength to maintain moderate vacuum conditions (hundreds of mTorr) for FACVD operation. A wall thickness of 20 mm was chosen, following industry standards for vacuum desiccators and considering the pressure differential that the chamber must withstand during operation. The FACVD system is illustrated in Figure 2 and Figure 3. All vacuum seals and process connections were implemented using 3/8-in stainless-steel tubing and double-ferrule compression fittings to ensure reliable and leak-tight interfaces between the metal pipe and plastic chamber. Flanges were used for the connections between the vacuum pump and outlet line and between the canister and inlet line, ensuring chemical incompatibility and mechanical stress at the plastic-metal interface. The chamber wall thickness was set to 20 mm. The O-rings used in the chamber were made of Viton rubber.

Chemical precursors were delivered to the chamber through two separate feed lines, each equipped with an aluminum (237 W/m·K) canister for containing either the monomer or initiator. The supply lines were fitted with needle valves and on-off valves for precise flow control, and a temperature regulator was installed to control the heating belts. As the particular monomer is vaporized at a temperature of approximately 70–80 °C, silicone rubber heaters were employed to heat the canisters. The monomer line was wrapped with a heating belt and maintained at approximately the same temperature as the canister to prevent condensation. The chamber pressure was monitored in real time using vacuum gauges, enabling independent control of the monomer and initiator flow rates and allowing accurate adjustment of the process pressure (P_M). Figure 4 shows the temperatures measured at the outer and inner walls of the canister heated using a silicone rubber heater. As can be seen, the time required for the canister temperature to reach the set temperature of 80 °C is 2700 s.

A resistively heated filament array was installed at the top of the chamber, as shown in Figure 3b. This filament, composed of nickel-chromium (NiCr) wire, is responsible for thermally decomposing the initiator and generating reactive radicals that initiate polymerization. NiCr allows stable resistive heating even at temperatures above 400 °C and can effectively generate radicals through the thermal decomposition of peroxide-based initiators. Additionally, NiCr exhibits a high electrical resistance and excellent oxidation resistance, maintaining its mechanical stability and durability through repeated thermal cycling [22,23]. The gas inlet (Figure 3b) allows vapor-phase reactants to enter the chamber directly above the substrate. By applying current via a power supply to the filament, it was confirmed that the desired temperature (220 °C) was reached rapidly within 300 s, as shown in Figure 5.

The substrate stage, fabricated from copper (401 W/m·K) and positioned at the base of the chamber (Figure 3c), is thermally isolated from the filament. Directly beneath the stage holder, a thermoelectric (Peltier) module is installed as shown in Figure 2, maintaining the stage temperature at 25–30 °C via proportional–integral–derivative (PID) control throughout the deposition process. This temperature regulation is essential to promote monomer condensation and ensure uniform polymeric film growth. To establish and maintain the vacuum environment, a rotary vane pump (W2V40, WSA, Seoul, South Korea) with a nominal pumping speed of 400 L/min was connected to the chamber outlet, as shown in Figure 2, The system can reach the target process pressure of 250 mTorr within approximately 150 min. The vacuum level was monitored continuously using a pressure gauge installed in the chamber. The chamber pressure time series is shown in Figure 6.

The overall system configuration, combining robust sealing, accurate temperature control, demonstrates the feasibility of low-cost transparent polymer vacuum chambers for advanced vapor-phase deposition processes.

2.3. Deposition Procedure

The P_M/P_SAT ratio is an important FACVD parameter [24]. These parameters are given as follows [25]:

$$P_M=P_{Ch}{\displaystyle \frac{F_M}{F_M+F_I}}$$

(1)

$$\mathit{ln}\left(P_{SAT}\right)={\displaystyle \frac{A}{T_S}}+B$$

(2)

where P_M is the actual partial pressure of the monomer, P_SAT is the saturation vapor pressure of the monomer at a given temperature, P_Ch is the chamber pressure when both the monomer and initiator are introduced, F_M is the monomer flow rate, F_I is the initiator flow rate, T_S is the substrate temperature, and A and B are empirical constants specific to each material.

The deposition rate and thin-film properties were determined using the P_M/P_SAT values. The P_SAT values were based on the saturation vapor pressure of the monomer used (i.e., GMA). In this study, P_SAT = 0.31 at a substrate temperature of 25 °C was used.

To vaporize the precursors, the canister and lines for TBPO were maintained at room temperature (26 °C), while those for GMA were maintained at 80 °C. The filament for radical generation of the initiator was set to 200–220 °C.

The deposition was performed three times. According to Asatekin et al. [26], the recommended P_M/P_SAT range is 0.2–0.8, where uniform thin-film deposition can be achieved. However, Tao and Anthamatten [27] described P_M/P_SAT as a “monomer saturation level.” They reported that when P_M/P_SAT exceeded unity, supersaturation induced monomer condensation, resulting in non-uniform film growth. Based on this interpretation, the present study classified the three experimental specimens as follows: Specimens 1–3, representing high, mid, and low saturation levels, respectively. The deposition conditions are listed in

Table 2. Deposition conditions for each specimen.

	Specimen 1	Specimen 2	Specimen 3
FI (sccm)	1.513	1.797	1.671
FM (sccm)	0.931	0.716	0.430
PCh (mTorr)	950	800	800
Sub (°C)	25	25	25
PM/PSAT	1.167	0.735	0.528

; all specimens were deposited for 4 h. The monomer and initiator flow rates were calculated using the throughput equation defined in ISO 21360, which converts the measured pressure difference between the base pressure and the pressure obtained when introducing only the monomer or initiator with the corresponding flow rate [28]. The deposition conditions listed in Table 2 were selected to investigate the effects of P_M/P_SAT on thin-film deposition. In this study, conditions within the recommended P_M/P_SAT range were chosen, along with one condition exceeding that range, to analyze the effects of P_M/P_SAT variation on the deposition rate, film thickness, and uniformity.

2.4. Analysis Methods

Thin-film characterization was performed by Fourier-transform infrared spectroscopy (FTIR, Vertex70v, Bruker, Bremen, Germany) in ATR mode with 32 scans, a resolution of 2 cm⁻¹, and a spectral range of 600–4000 cm⁻¹. X-ray photoelectron spectroscopy (XPS, NEXSA, Thermo Fisher Scientific, Waltham, MA, USA) was conducted using Al Kα radiation (1486.6 eV). Survey spectra were acquired with a spot size of 400 μm, a pass energy of 200 eV, and a step size of 1.0 eV, while high-resolution spectra were collected with a pass energy of 40 eV and a step size of 0.1 eV. The base pressure was maintained below 1 × 10⁻⁹ Torr, and charge compensation was applied during measurement. Film thickness was measured by fixed-angle spectroscopic ellipsometry (FS-4, Film Sense, Lincoln, NE, USA) at 70° with four wavelengths (465, 525, 590, and 635 nm), and the data were fitted using a Cauchy model.

3. Results

The polymer thin-film deposition results using the FACVD process with an acrylic chamber are shown in Figure 7. These are optical microscopic images of Si wafers after thin-film deposition via the FACVD process. As shown in Figure 7a (high saturation P_M/P_SAT = 1.208), monomer condensation results in droplet formation (dropwise condensation). By contrast, as shown in Figure 7b,c, more uniform films are obtained at mid and low saturation (filmwise condensation) [27]. Figure 7b,c show successful deposition; however, they exhibit differences in the deposition rate according to the P_M/P_SAT value. Ellipsometry measurements indicate that the thicknesses of the films in Figure 7b,c are 104 ± 15.8 and 72 ± 6.1 nm, respectively. Therefore, the film growth rates were determined to be 26 and 18 nm/h, respectively. These results confirm that thin-film deposition occurred and suggest that even with the same deposition time, P_M/P_SAT variations result in differences in film thickness. The upper-left corners of Figure 7b,c show the appearance of the deposited specimen.

Figure 8 shows the FT-IR spectra of the deposited pGMA thin films. A total of five major absorption bands were observed at 2932, 1728, 1275, 1120, and 738 cm⁻¹. In addition, the C=C stretching band of the monomer at ~1635 cm⁻¹ disappeared in the deposited films, confirming that polymerization occurred. The peak at 2932 cm⁻¹ corresponds to C–H stretching vibrations, while the peak at 1728 cm⁻¹ is attributed to the carbonyl (C=O) group. The absorption band at 1275 cm⁻¹ represents the asymmetric stretching of C–O–C, and the peak at 1120 cm⁻¹ is due to C–O stretching vibrations. The band at 738 cm⁻¹ is associated with the out-of-plane bending of C–H bonds. The orange shaded regions in the spectrum highlight the characteristic peaks of pGMA, indicating the presence of specific functional groups in the deposited film. These characteristic peaks confirm the successful deposition of pGMA, as they agree with previously reported FT-IR spectra of pGMA films [29,30,31].

XPS was conducted to investigate the surface chemical composition and bonding states of the deposited pGMA thin films. The survey spectrum in Figure 9a shows characteristic peaks corresponding to C 1s and O 1s, which are the primary constituents of pGMA. Additionally, Si 2p and Si 2s peaks were also observed in Specimen 1, indicating the potential exposure of the substrate owing to insufficient film thickness or non-uniform deposition [32]. Specimens 2 and 3 showed only C 1s and O 1s peaks without a silicon peak, indicating that the films were thicker and more uniform than those of Specimen 1. The C 1s spectra of Specimens 2 and 3 also show three peaks, similar to those of Specimen 1. The C 1s spectra of Specimens 1–3 are shown in Figure 9b–d, respectively. Each spectrum was deconvoluted into three major components: a C–C/C–H peak (284.8 eV, corresponding to the polymer backbone), a C–O–C peak (286.6 eV, attributed to the oxirane ring), and an O–C=O peak (288.8 eV). These peaks are consistent with the chemical structure of pGMA, confirming its successful deposition on all three specimens. Importantly, even when the P_M/P_SAT value was sufficiently high to cause condensation or within the normal range that enabled film formation, the FTIR and XPS analyses consistently confirmed the characteristic chemical bonding structure of pGMA. This indicates that, despite variations in P_M/P_SAT, polymer thin films with consistent chemical compositions were successfully deposited under all experimental conditions.

These results suggest that pGMA was successfully deposited on all three specimens, although the deposition was not completely uniform. These results are consistent with those reported in previous studies [33,34,35]. Finally, these spectroscopic, surface, and visual characterization results are consistent with the expected chemical structure and properties of the FACVD-grown polymer films using GMA precursors.

4. Discussion

In this study, an FACVD system was designed and used to perform polymer thin-film deposition. The results confirmed that polymer thin-film deposition was possible even when a plastic chamber was used. However, the time required to reach the target vacuum level was found to be excessively long, which is believed to be due to the long pipeline length and small diameter between the chamber and the vacuum pump. To improve the operational efficiency in future work, we plan to modify the pipeline configuration to reduce the evacuation time to the desired vacuum level, thereby enhancing the process throughput.

Additionally, the findings of this study confirmed that the P_M/P_SAT ratio plays a crucial role in determining the deposition rate and thin-film growth behavior. When the P_M/P_SAT value was low, the resulting film exhibited a relatively thinner thickness for the same deposition time. Conversely, higher P_M/P_SAT values led to increased deposition rate, resulting in thicker films. The thickness uniformity was determined within 85% across the entire 4-inch wafer area. Moreover, deposition was observed even when the P_M/P_SAT values exceeded the generally recommended range (0.2–0.8). Although supersaturation led to condensation and reduced uniformity, the specimens exhibited the same characteristic chemical bonding structures as those formed within the typical film-forming range. This indicates that the influence of P_M/P_SAT does not determine whether deposition occurs but rather governs the uniformity of the resulting thin films. To address this issue and improve film uniformity, future system modifications will focus on repositioning the initiator inlet toward the center of the chamber or using a showerhead-type diffuser to achieve a more homogeneous radical dispersion across the substrate surface.

Overall, these findings highlight the importance of optimizing both the vacuum system design and thermal management within the chamber to achieve uniform polymer thin-film deposition using FACVD in a low-cost, transparent acrylic chamber system.

5. Conclusions

In this study, the feasibility of applying a plastic acrylic chamber to the FACVD process was successfully demonstrated. Preliminary results indicate that the custom-designed plastic chamber can support uniform film growth under standard FACVD operating conditions, as evidenced by spectroscopic analysis of the deposited films. Therefore, the key findings can be summarized below:

• A plastic acrylic chamber was successfully applied to a FACVD system as a vacuum reactor, demonstrating feasibility for vapor-phase polymerization under moderate vacuum conditions.
• The devised reactor showed a successful demonstration of growing pGMA thin films by controlling the growth rate within 85% of thickness uniformity for a 4-inch wafer’s area.
• Polymerization of pGMA thin films was validated by using FTIR and XPS analysis.

Building on these initial findings, future work will focus on optimizing the chamber design, expanding the range of processable materials, and conducting repeated experiments to establish reliable and reproducible deposition protocols. In addition, while this study focused on demonstrating the feasibility of applying a plastic chamber to the FACVD process, a systematic evaluation of film durability will be conducted as future work. This line of research is expected to contribute to the broader adoption of cost-effective transparent polymer vacuum chambers for advanced vapor-phase deposition processes.