Coupling APS/SPS Techniques for Cu-TiO₂ Antibacterial Coating Deposition: Application to Water Treatment

Abstract

Since the COVID-19 pandemic, efforts in the field of surface decontamination have been redoubled. Finding innovative self-cleaning devices has become a challenge, and several solutions have been proposed in the market in recent years. In this work, an optimized powder/suspension plasma spray process at atmospheric pressure, using a Triplex Pro 210^TM torch, is implemented to produce Cu-TiO₂ surface coatings on stainless steel. The purpose is to investigate the potential improvement of antibacterial efficacy by the reactive surface species generated from TiO₂ photoactivity under irradiation. A water-based suspension, prepared with AnalaR NORMAPUR^TM TiO₂, is used as a precursor to incorporate the photocatalyst into an antibacterial copper matrix. Surface antibacterial tests according to ASTM 2180 standards were performed, and experiments were performed in treated contaminated water. Sub-stoichiometric blue TiO₂ coatings showed complete bacterial elimination after 90 min of visible light irradiation, and Cu-TiO₂ surface coatings were even able to disinfect the surfaces under white light, making the application interesting for bacterial destruction under natural illumination. These materials are also intended for application in water treatment, including both pathogens and chemical micropollutants, which is a pressing issue facing many countries today.

1. Introduction

A widely accepted theory in microbiology posits that microorganisms often flourish in close proximity to surfaces, which facilitates their colonization and growth on a range of materials [1]. The interaction between pathogens and surfaces is of great consequence in the regulation of nutrient cycling and the breakdown of diverse chemical substances in the environment. This process contributes to the rapid dissemination of microbes, the fouling of equipment, and the contamination of medical instruments. It is important to note that microorganisms tend to accumulate in biofilms [2]. The nature of the interaction between the microorganisms and the surface material in question is dependent upon a number of factors, including wettability, molecular topography, adhesion, and responses to external factors such as pH, temperature, and chemicals. As observed in the review by Cloutier et al. [3], the intrinsic and mechanical properties of materials employed in the healthcare sector have undergone notable advancements in recent years. Nevertheless, redox reactions occur in the surface’s initial layers, necessitating considerable effort to enhance the surface condition. The objective is to establish a healthy environment that can meet the increasing demands of modern society. In the context of antimicrobial materials, copper (Cu) is the oldest metal with which human civilization has been acquainted. Its earliest documented use dates back to approximately the 5th millennium BC. The earliest known medical applications of copper are documented in the Smith Papyrus, one of the most ancient surviving texts [4]. Furthermore, recent reviews have investigated the antibacterial potential of two-dimensional materials, biomaterials, and noble metals, including silver (Ag), gold (Au), and platinum (Pt) [5,6,7]. These studies have identified metallic elements as promising candidates for antimicrobial and fungicidal properties. However, their widespread use is restricted by cost and toxicity. Despite their significant presence in the body, these elements can be expensive and toxic in large amounts, both in vivo and in vitro. Copper is indeed an essential trace element in living organisms, playing a role in over 30 known proteins, including tyrosinase (involved in melanin production), lysyl oxidase (involved in collagen cross-linking), and cytochrome c oxidase (which regulates the respiratory chain) [8]. With regard to silver, it has been demonstrated that cumulative doses of approximately 70 to 150 mg·kg⁻¹ of body weight can elicit adverse effects, including dermatological reactions, as well as indications of renal, hepatic, or neurological impairment [9]. Hence, there is an interest in developing these materials as “surface coatings” rather than solids. Beltsios et al. recently reported the preparation of efficient Cu-filled TiO₂ layers by dip-coating technique to prevent infections in cardiovascular implants [10]. Rutkowska-Gorczyca et al. investigated powder precursors for applications in antibacterial coatings by dry processes [11]. They demonstrated the effectiveness of modifying dendritic Cu powder with average amounts of amorphous TiO₂ nanoparticles. They observed a decrease in the inhibition diameter zones of Escherichia coli (E. Coli, gram (−)) and Staphylococcus aureus (S. Aureus, gram (+)) after 48 h of exposure to the modified powders. In a later publication, the researchers compared the mechanical properties of cold Spray Cu and Cu-(10% wt. TiO₂) coatings [12]. The conclusion was that the addition of a ceramic phase to Cu does not reduce the tribological behavior of the metal, which is very interesting for applications of biologically active coatings in everyday use. An interesting study was presented by Wahyuni et al. [13]. They described the preparation of Cu-doped TiO₂ coating by photo-electrodeposition technique. Increasing the Cu concentration shifted the absorbance from 387.5 nm for undoped TiO₂ to 427.5 nm for TiO₂-Cu, which allowed 80% photodegradation of amoxicillin (antibacterial mediation) after 24 h of visible light irradiation. However, although many references could be cited in antibacterial Cu-doped TiO₂, very few studies have reported antibacterial Cu surface functionalization by TiO₂ as coatings. According to Eessaa et al., Cu matrix with different TiO₂ wt.% could be obtained by ball milling and used as chemical vapor deposition (CVD) layers to improve the performance of silicon solar cells [14]. For further applications, scientists have implemented processes capable of producing high-quality coatings by thermal and cold plasmas [15,16]. Indeed, the development of such surfaces is in full expansion, and the choice of the dry coating process, in particular plasma spraying, is based on the ability to control the structure, microstructure, texture (porosity for a specific area), composition, and therefore the different properties of materials. The choice is also based on the feasibility of this type of multifunctional coatings on various metallic substrates such as steel and heat-sensitive substrates such as glass, wood, or plastic, but also on large substrates compatible with industrial requirements. In addition, these techniques can be used to add compounds to the starting material to improve its antibacterial efficacy or even give it new interesting properties.

In this work, a novel process for the preparation of Cu-TiO₂ surface coatings is developed by coupling two types of atmospheric-pressure plasma spraying techniques using powder and suspension. The experiments were performed using a rotating substrate geometry. The photocatalytic and antibacterial efficiencies are also reported in this work. The aim is to develop coatings that are efficient not only against biological elements but also against chemical compounds. The technology development took place on stainless steel substrates but could be transferred to other types of surfaces for medical, public transportation, and daily use applications.

2. Materials and Methods

2.1. Plasma Spray (PS) Coatings

In this study, spheroidal copper particles (−90 ± 38 µm nominal range, Oerlikon Metco 55^TM, Wohlen, Switzerland) were used as the starting material. As for the suspension, the TiO₂ starting suspension (AnalaR NORMAPUR^® nanometric particles ≤ 25 nm from VWR^TM part of Avantor^TM, Radnor, PA, USA) was prepared in distilled water (20%wt. TiO₂) according to the procedure described by Toma et al. [17]. Stainless steel 304L (SS, 25 mm diameter) disks were used as substrates. Prior to deposition, sandblasting (white Corindon F80, 0.5 MPa, 100 mm) was performed to ensure good adhesion of the coating to the substrate (initial Ra ~5–6 µm). Substrates were cooled with compressed air during deposition to avoid excessive heating. The surfaces were then coated using an Oerlikon Metco^TM Triplex Pro 210 (Switzerland) plasma torch with a 9 mm nozzle diameter. This allows a gain in arc stability compared to the F4 torch due to the insulating rings (neutrodes) between the cathode and anode while maintaining a high powder feed rate. The precursors were radially injected 7 mm downstream from the nozzle outlet using a 1.8 mm diameter injector for powder and a 220 µm diameter injector for suspension. Samples were prepared in a rotating cylindrical geometry (Figure 1).

The cylinder rotation speed was set to provide a relative torch/substrate speed of 1 m.s⁻¹ of torch/substrate relative speed. Such a cylindrical geometry is interesting for obtaining thin plasma sprayed coatings since the main bacterial contact and reactions occur on the surface. A detailed explanation of this processing effect’s impact on microstructure, thickness, and stress will be discussed later.

Table 1. Plasma spray conditions for Cu, TiO₂, and Cu-TiO₂ coatings on stainless steel.

Parameter	Conditions
Arc current	250 A (Cu), 400 A (TiO2), 450 A (Cu-TiO2)
Net power	20.1 kW (Cu), 19.8 kW (TiO2), 21.6 kW (Cu-TiO2)
Stand-Off-Distance	140 mm (Cu), 110 mm (TiO2 and Cu-TiO2)
Feedstock feedrate	30 g·min−1 (Cu), 16 mL·min−1 TiO2
Ar/H2 flowrate	85/5 (Cu) LPM, 40/3 LPM (TiO2 and Cu-TiO2)

summarizes the spray conditions.

A low current (250 A) was used for Cu coating to avoid excessive evaporation at higher currents. In parallel, it is necessary to increase the enthalpy of the plasma jet to evaporate the solvent in the TiO₂ suspension. For this, a higher current (400 A) was used for TiO₂. The current of 450 A used for Cu-TiO₂ deposition compromises the required enthalpy for the optimal evaporation of Cu and TiO₂. Moreover, it can be seen from Table 1 that the net power does not vary much by changing the deposition conditions (gas mixture and current).

2.2. Physical/Chemical Characterizations of the Coatings

The surface of the coatings was observed using a Hitachi^TM (TM3000, Tokyo, Japan) tabletop scanning electron microscope (SEM). The cross-sections were observed using FEI Quanta 450 FEG (Hillsboro, OR, USA) and JEOL IT300 LV (Peabody, MA, USA) to reveal the coatings microstructure. Chemical composition analysis by Energy Dispersive Spectroscopy (EDS) was performed on the same instrument. X-ray photoelectron spectroscopy (XPS) was performed using an Axis Ultra DLD-Kratos^TM (Manchester, UK) instrument with a monochromatic Al Kα source (1486.6 eV, 180 W, with 0.5 eV step size). The crystal structure was identified by X-ray diffraction (XRD) using a Bruker^TM D8 Advance Eco (Ettlingen, Germany) with a step size of 0.02° 2ϴ and a step time of 0.5 s. The contact angle (water drop, 3.1 µL at 25 °C, slow speed of 0.397 µL.s⁻¹) was measured on a GBX instruments^TM (Miami, FL, USA) goniometer, using Visiodrop^© software (V15.Ink) for data analysis. Finally, the topography of the coatings was observed using an Olympus^TM DSX1000 (Tokyo, Japan) digital microscope.

2.3. Antibacterial Test Protocol

2.3.1. Step 1: Bacterial Culture and Test Suspension Preparation

One major aim of this work is to provide a proof of concept for the use of the developed coatings as antimicrobial materials. To this end, Escherichia coli (E. Coli) was chosen due to its use as a reference in many studies. In addition, the use of E. coli as an indicator of fecal contamination is justified by its prevalence in wastewater, as referenced by Cabral et al. [18]. The bacterial strain used to evaluate the antibacterial activity of the coatings in this study is gram (–) E. coli (CIP 52.172) because of its ease of growth procedure in the laboratory and its occurrence in humans.

Each sample, extracted from a colony of E. coli strains, is incubated for 24 h at 37 °C with magnetic stirring in a Tryptic Soy Broth (TSB, ref. BK046HA, Biokar diagnostics, France). A bacterial culture (BC) is then obtained. The BC absorbance is measured with a spectrophotometer at 580 nm in order to adjust the drop concentration at 10⁶ bacteria per mL on the sample surface. Intermediate dilutions are also performed to obtain a bacterial suspension (BS). More precisely, this BS is obtained by inoculating a Trypticase Soy Agar (TSA) broth (3 g.L⁻¹) in the supercooled state (50 °C) and an intermediate dilution of the BC. Since different surface conditions are tested in this work, thick agar broth is more suitable than a classical liquid medium in contact with coatings.

2.3.2. Step 2: Bacterial Inoculation of the Sample Surface

With regard to the choice of analytical methods, the standard ASTM 2180 protocol has been selected [19]. This choice was made to limit differences between studies, especially surface states that make comparison difficult [20]. A volume of 87 µL of BS containing approximately 10⁶ bacteria per mL was applied to the coatings using a 10 mm diameter Pyrex^TM cloning cylinder (Figure 2a). The cloning cylinder is essential to ensure the same contact surface on the sample for better reproducibility. The photos of the tests on the different coatings according to the method derived from ASTM2180 are shown in Figure 2b–e.

The contact time between the bacteria and the surface was set to 1 h, at room temperature (estimated to be 25 °C), under sterile conditions. After contact, the solution was collected for analysis using conventional culture techniques.

2.3.3. Step 3: Bacteria Collection and Counting by Culture Technique

After the contact time, both the cloning cylinder and the sample were transferred to a 50 mL tube containing 2.5 mL of saline solution (0.9% NaCl) so that the inoculum deposition zone was completely submerged. This tube was then placed in an ultrasonic bath at 35 kHz for 5 min to disperse and transfer the bacteria from the coating to the liquid. Microbiological analyses were then performed on this stock solution and compared with the initial BS corresponding to the untreated bacteria (control). To quantify the antibacterial effect, enumeration using classical culture techniques was used. This technique is based on bacterial growth: each invisible bacterium gives rise to a colony visible to the naked eye. The procedure is shown in Figure 3.

As shown, 1 mL of the control BS or the contaminated BS was added to 9 mL of sterile aqueous solution (0.9% NaCl), and then 100 µL was ‘‘raked’’ onto a TSA Petri dish. This first dilution was labeled (D1). A duplicate of each dilution was made. Then, 1 mL of (D1) was added to 9 mL of saline solution to make (D2), which was spread on another TSA Petri dish. The process was repeated up to (D10) in this study. This technique facilitates bacterial enumeration and effect quantification.

2.4. Flow Cytometry Analysis (FCM)

Flow cytometry (FCM) analysis was used to evaluate the bactericidal effects of the different coatings. A double staining method using the propidium iodide (PI) and the 5(6)-carboxyfluorescein diacetate (5–6 CFDA) was chosen to assess membrane integrity and intracellular esterase activity, respectively. To explain, CFDA was used as an enzymatic activity marker that allows the detection of active bacteria. Meanwhile, PI, a marker of membrane integrity, was used as an indicator of bacterial mortality. This marker had a sealing agent that intercalated into the minor groove of the bacterial DNA. Thus, it only penetrated bacteria with altered membranes. Figure 4 is shown below to facilitate the understanding of the procedure.

This protocol was adapted from Kolek et al. [21]. Each sample (2 mL) was stained with 22 µM of CFDA (Invitrogen^®, Waltham, MA USA) and incubated for 30 min at 37 °C. Then, 7.5 µM of PI (Invitrogen^®, MA USA) was added, and the samples were immediately analyzed. Analysis of stained samples was performed on a Cytoflex LX^® (Beckman Coulter^®, Brea, CA, USA) flow cytometer device equipped with a blue laser (488 nm, 50 mW). This wavelength corresponds to the excitation of the fluorochromes used (IP and CFDA). This is a relatively interesting technique because of its speed (about 1000 events per second), precision, and sensitivity. Forward scatter (FSC) and side scatter (SSC) signals were used to trigger and define the total cell population. Green (FL1, B525-FITC; 525/40 BP) and red (FL4, Y585-PE; 585/42 BP) fluorescence were recorded to divide into four quadrants as shown in cytograms. Data were analyzed using the Kaluza^® (Beckman Coulter^®, version 5.01) software.

2.5. Photocatalytic Antibacterial Test in Wastewater

In order to investigate the potential applicability of these coatings in water treatment, and more precisely for reclaimed water applications, antibacterial properties were tested in aqueous conditions. Samples of secondary effluent generated with activated sludge treatment at the municipal wastewater treatment plant of Limoges (285,000 population equivalents) were collected. Treated wastewater was then sterilized (20 min and 121 °C), and E. coli (CIP 52.172) was added at the final concentration of 10⁶ CFU/mL E. coli. This step allows for the replication of realistic conditions regarding the physical and chemical characteristics of the wastewater and the control of the contamination level in the sample.

Experiments were carried out under light irradiation conditions using treated wastewater. The coatings were placed in contact with 15 mL of sterilized wastewater contaminated with 10⁶ CFU/mL E. coli. The vials were then placed in the PhotoRedOx EvoluChem (Advion InterChim Scientific, Ithaca, NY, USA). A ventilation system is installed in the apparatus to remove the heat generated by the lamp irradiation, thus maintaining the container at room temperature. For each experiment, a half disk (245.312 mm²) was used and illuminated with a 380 nm lamp. Bacterial reduction percentages were calculated using the plate count method explained in Section 2.3.3.

3. Results

This section presents and describes the results of the study. The interpretations and explanation of the results and the comparison with existing literature will be evaluated in the discussion section (Section 4).

3.1. Copper Powder and Water-Based Suspension Characterizations

The SEM micrograph of the Metco 55^TM Cu starting powder is presented in Figure 5a. In addition, the TiO₂ powder was observed by SEM and characterized by XRD to identify the crystalline phases. The micrograph and the XRD pattern are shown in Figure 5b,c, respectively.

The SEM image of the Cu starting material (Figure 5a) shows a homogeneous particle size distribution with a spheroidal shape. Moreover, this monomodal distribution was confirmed by granulometry. Figure 5b shows the highly porous texture of the TiO₂ powder to be incorporated into the Cu metal matrix. The XRD pattern in Figure 5c was indexed according to JCPDS No. 00-021-1272 previously presented by Mishra et al. [22].

In parallel, the rheological behavior of the suspension (20%wt. TiO₂ in distilled water) was investigated. The viscosity vs. shear rate plot is shown in Figure 6.

The average viscosity of 1.34 mPa.s obtained up to a shear rate of 2000 s⁻¹ is comparable to that previously reported for the same anatase loading in water at 20 °C [23].

3.2. Observation of Coatings and Surface Composition

The surface of the coatings was observed by SEM to compare the different morphologies. The micrographs are shown in Figure 7.

Highly porous Cu coatings are obtained under these plasma conditions (Figure 7a). Some unmelted particles are incorporated into the matrix. A denser coating is observed for TiO₂ (Figure 7b). It can be seen that the surface of the coating shows similarities to the initial powder texture. In the case of the co-sprayed Cu-TiO₂ coating (Figure 7c), the Cu texture observed in Figure 7a is replicated with a higher melt fraction and partial pore filling by TiO₂. In all three cases, full substrate coverage is achieved. An average thickness of 35–50 µm is obtained for the Cu-based coatings (Cu and Cu-TiO₂) and of 10–12 µm for TiO₂, as it will be discussed later, in Section 3.3. Such material thickness is considered in the “thin” range for plasma spraying but remains sufficient to impart antibacterial properties to 304L stainless steel and protect it from external corrosion.

Elemental dispersion and atomic and weight percent estimates were evaluated by EDS. The maps and results are shown in Figure 8.

The Cu coating is slightly oxidized on the surface, although the O estimation is quite difficult due to atmospheric adsorption. EDS maps show a relatively uniform dispersion of Ti on the Cu-TiO₂ surface with about 10% by weight of the total coating (Figure 8c). The O/Ti atomic ratio is 1.72 and 1.85 in TiO₂ and Cu-TiO₂, respectively (Figure 8d). This ratio indicates a substoichiometric TiO₂, which explains its blue color.

XPS Cu2p and O1s are shown in Figure 9 for further information on the different bonding in the Cu coating.

Figure 9a confirms the presence of CuO on the surface of the coating. The O1s data in Figure 9b show both M-O and M-OH bonding, where M refers to the metal (Cu).

Similarly, the Cu2p, Ti2p, and O1s XPS peaks detected on both TiO₂ and Cu-TiO₂ deposited on stainless steel are shown in Figure 10.

These results show that Ti³⁺ states are present either in TiO₂ or Cu-TiO₂. In addition, the O1s peaks are deconvoluted into lattice O^2- and an oxygen vacancy contribution, as reported by Mintcheva et al. [24]. Figure 10c shows the increase of the Cu²⁺ contribution compared to the Cu coating (Figure 9a). This suggests further Cu oxidation during powder/suspension co-spray or mixed-phase formation. This point will be discussed in Section 4.

3.3. Cross-Section and Microstructure Observation of the Coatings

The coatings were cut and polished to estimate their thickness and observe their microstructure. The micrographs are presented in Figure 11.

The thicknesses are in the range of 35, 10, and 50 µm for, respectively, Cu, TiO₂, and Cu-TiO₂ coatings. Figure 11a shows a homogeneous microstructure mainly related to the uniform particle size, which allows a steady thermal treatment even with radial injection. The thinnest coating is TiO₂ (Figure 11b) compared to Cu-based layers since the starting material is vehiculated through a liquid towards the surface. The oxide surface presents agglomerates from the nanometric particles, and a porous network is formed. As expected, Cu coating microstructure and texture are quite dense, and the addition of TiO₂ confers some porosity (Figure 11c). The adherence is improved for Cu-TiO₂ even though it could be improved for Cu and TiO₂ where some voids with the substrate are detected. The addition of ethanol to the suspension will probably enhance the anchoring, and some improvement could be made during the surface pre-treatment by sandblasting.

3.4. Surface Topography and Wettability

As this work highlights the effect of surface state effect on bacterial adhesion and removal, water drop wettability images are shown in Figure 12.

Figure 12a shows the good wettability of stainless steel, even after surface roughening by sandblasting. The contact angle calculated on SS is 25.70° ± 0.25°, averaged on the left and right sides. However, after Cu deposition by plasma spraying, the contact angle value increases up to 106.00° ± 0.84°, indicating a lower surface wettability (Figure 12b). In the case of TiO₂, the contact angle increases slightly to 39.09° ± 0.55° compared to the bare substrate. The coating has a different topography, and its texture is probably less porous than the sandblasted SS substrate, but the surface remains wettable, especially with the OH groups detected in Figure 10b. As for the Cu-TiO₂, a contact angle of 109.36° ± 0.47° was calculated. This value is slightly higher than that of Cu, mainly due to the presence of TiO₂ filling the porosity, as observed in Figure 7c.

To complete the investigation of the surface condition, the samples, before and after deposition, were observed by digital microscopy. The images and roughness color maps are shown in Figure 13.

Figure 13b shows the complete coverage of the surface with Cu. The surface is quite rough, mimicking the topography of the bare substrate. The root mean square (RMS) roughness increases from 1.30 µm on sandblasted steel to 9.15 µm after Cu plasma spraying. Color contrasts are observed on the copper surface as a sign of oxidation, especially some dark spots that could be attributed to tenorite (CuO). In addition, some reddish inclusions detected on the metallic surface could be attributed to cuprite (Cu₂O). However, its concentration is probably very low compared to the other phases and is below the detection limit in XRD, as explained in Section 3.4. The TiO₂ coating surface appears relatively smooth compared to Cu (Figure 13c), with an RMS of 3.25 µm. Although the coverage is relatively uniform, the substrate can be seen in some places. The Cu-TiO₂ (Figure 13d) is quite rough, with an RMS of 13.52 µm and shows blue reflections from the TiO₂ on the surface. This coating has the highest RMS value probably because, as explained earlier, TiO₂ agglomerates block the pores of the Cu surface creating a flake-like effect and thus a rougher surface.

3.5. Identification of Crystalline Phases

The starting AnalaR^TM TiO₂ powder was identified as anatase (Figure 5c). However, the temperature inside the plasma jet during spraying could reach 15,000 K, which induced thermal treatment of the particles, especially at the nanoscale. Since anatase is known to exhibit the best photocatalytic activity due to its indirect band gap compared to the other TiO₂ polymorphs i.e., rutile and brookite, the preservation of anatase at high temperature becomes a challenge. For this purpose, the XRD patterns of Cu, blue TiO₂, and Cu-TiO₂ are shown in Figure 14.

The peaks observed at 43.35°, 50.68°, and 74.19° are attributed to Cu (111), Cu (200), and Cu (220) from the cubic lattice of copper (JCPS No. 040836), as reported by Phul et al. [25]. A small, large peak is detected at 36.68° (Figure 14a). It is attributed to monoclinic CuO (-111), as mentioned by Pawar et al., referring to JCPDS No. 45-0937 [26]. As for the TiO₂ coatings (Figure 14b), anatase (A) and rutile (R) are indexed for a mixed phase according to JCPDS 75-1537, as reported in the work of Morad et al. [27]. It is observed that conversion to rutile occurs at high process temperatures, but a good amount of anatase is still retained. Peaks from the stainless-steel substrate [28] are detected due to the low thickness of the coating. In the Cu-TiO₂ coating (Figure 14c), metallic and oxide Cu peaks are observed, as in the pure Cu coating. Small anatase peaks at 25.23° and 47.92° and rutile at 37.76° are detected as evidence of TiO₂ incorporation into the matrix.

3.6. Antibacterial Efficacy Results

Antibacterial efficacy is evaluated by ‘log10 reduction’ or bacterial reduction. Bacterial reduction is the difference between the log₁₀ of the initial bacterial concentration and the log₁₀ of the final bacterial concentration, as determined by enumeration using the culture incubation technique (24 h, 37 °C). Bacterial reduction was calculated based on 10⁶ colony-forming units (CFU) per ml as the initial concentration. The results after 1 h contact with the samples, according to the ASTM 2180, are shown in Figure 15.

Figure 15 shows that sandblasting the substrate creates porosity that increases the contact surface with E. coli, especially with the high wettability observed previously. In parallel, the Cu coating showed complete elimination of E. coli after 1 h of exposure (6.4 log₁₀), exceeding the sterilization threshold. Without illumination, the blue TiO₂ coating showed a slight increase compared to the substrate, barely exceeding the disinfection limit. However, the activity is still very low compared to pure Cu. An interesting result is observed for Cu-TiO₂. The combination of both compounds reduces the antibacterial efficacy by about 2 log, mainly due to the presence of TiO₂, which presents lower activity. Nevertheless, the reduction is very close to sterilization, and flow cytometry, explained in Figure 16, will give more details about the possible killing/inhibition mechanism.

3.7. Flow Cytometry Results (FCM)

As mentioned above, colony counting after bacterial culture is commonly used to determine antibacterial efficacy. However, this technique does not provide a complete understanding of the elimination mechanism. Therefore, flow cytometry results are presented in this section to explain what might actually happen to bacteria in contact with solid surfaces. Figure 16 shows the comparison of cytograms between the untreated cells and those in contact with the different coatings for 1 h.

CFDA (carboxyfluorescein diacetate) vs. PI (propidium iodide) dot plots show two colors (Figure 16): green for the cells stained by PI (red fluorescence, a fluorochrome that only reaches the bacterial DNA when the membrane is damaged, thus a dead cell marker) and red for CFDA labeling (green fluorescence, a fluorochrome that indicates enzymatic activity, thus a live cell marker). Figure 16a,b show that the percentage of dead cells increases from 0.96% to 1.41% on the Cu surface. In parallel, 19.09% of the cells remain active, and 78.32% of the bacteria show a ‘dormant’ or a ‘inhibited’ behavior. This result seems surprising because in the bacterial culture technique, no bacteria were counted after 1 h exposure to Cu, and the result is repeatable. The TiO₂ coating (Figure 16c) shows a lower inhibition percentage, lower than the untreated bacteria. However, the highest number of dead cells is detected for the oxide coating. The coupling of Cu and TiO₂ (Figure 16d) gives the highest bacterial inhibition, with the lowest percentage of active cells after 1 h of contact. These observations are further discussed in Section 4.

3.8. Bacterial Tests in Reclaimed Water

The minimum performance of urban wastewater treatment plants is regulated by discharge standards in accordance with the French decree of 21 July 2015. Once collected, the wastewater undergoes various treatments in treatment plants to eliminate suspended solids, phosphorus, nitrogen, or dissolved carbon. However, even after this conventional treatment, pathogenic enteric microorganisms, including viruses, bacteria, protozoa, and helminths, persist in the treated wastewater [29]. Detecting all of them is a tedious and very costly task, hence the need to use indicator bacteria [30]. Coliform bacteria, especially E. coli, which are already regulated for reuse of treated wastewater (REUT), are the most widely used indicators to assess the health compliance of water.

For this purpose, the coatings are tested according to the procedure described in Section 2.5. Figure 17 summarizes the main bacterial reduction percentages obtained after 60 and 90 min with the Cu-TiO₂ coatings.

Figure 17a,b show that the activity of the 304L stainless steel substrate is the same before and after illumination, at either 60 or 90 min. The porosity created by sandblasting increases the contact surface and bacterial adhesion, resulting in some activity even without the coating. This activity is significantly increased after copper deposition for both times, which was expected based on the antibacterial test results observed in Figure 15. However, the activity is slightly lower in the aqueous medium compared to the direct contact with the surface, which is logical because the solid/liquid volume ratio is not the same. The interesting observation is in Figure 17b, where an increase in activity is observed after irradiation of the Cu-TiO₂ coating at 380 nm. The relationship between the reactive species and the performance is discussed in detail in Section 4.

4. Discussions

4.1. Discussion of the Suspension Properties

Figure 5c shows that the TiO₂ starting material is a pure anatase polymorph. This information is important because the powder will be subjected to high thermal treatment in the plasma jet, and the rutile phase is expected to be present after deposition. As for the suspension behavior observed in Figure 6, the volume ratio TiO₂/water ø calculated in this study is 0.056 for the prepared suspension, according to the following formula [31]:

ø = ([W/ρ] _TiO2)/([W/ρ] _TiO2 + [W/ρ] _water)

(1)

W is the weight in kg and ρ the density in kg·m⁻³.

The suspension viscosity values at low shear rates (<1000 s⁻¹) could be compared with those obtained by Das et al. at 20 °C and 122.3 s⁻¹ shear rate with water-based anatase suspension for slightly higher volume fractions [32]. In the work of Das et al., the anatase in water suspension using CTAB (Cetyl Trimethyl Ammonium Bromide) or AA (Acetic Acid glacial) surfactant showed a ‘shear thickening behavior’ or an increase in viscosity by increasing the shear rate [32]. However, as shown in Figure 6, the suspension in this work shows a shear rate thinning behavior. Chakraborty et al. reported a shear-thinning flow nature of TiO₂ nanofluids [33]. The authors also mentioned the interesting debate between works where TiO₂ nanofluids show Newtonian behavior (no viscosity variation with shear rate) [34,35] and others that reported viscosity variation with shear rates for TiO₂ nanofluids [36,37]. In fact, the type of dispersant plays a crucial role in the fluidity of the suspension. In this study, the dispersant used was Beycostat C213 (CECA France), which is classified as ‘electrosteric’. In contrast to the cationic CTAB used by Das et al. [30], Beycostat C213 tends to decrease the viscosity by increasing the shear rate, mainly due to the dominant monomeric state of the surfactant at low concentrations, as explained by Chakraborty et al. [33]. The authors observed a similar behavior when using Tween 20, a non-ionic surfactant. Suspension viscosity is also expected to decrease with temperature [32], which is beneficial for flowability and particle acceleration toward the substrate.

As mentioned by Friis et al. [38], particles of the same size will encounter the same thermal treatment in the plasma plume. For that, the microstructure of the coating will be homogeneous since the particles spreading and cooling are quite similar. In this case, the Cu particles are considered on a large scale and will easily enter the plasma core. In parallel, agglomerates of around 1 µm are formed with the anatase TiO₂ powder. As highlighted by Aubignat et al. [39], aqueous droplets present a higher momentum compared to alcoholic droplets, which allows them to enter the plasma according to a trajectory close to the torch axis. Moreover, coatings with aqueous suspension are denser than with the alcoholic but with some porosity due to the particle’s agglomerates. The initial nanometric shape of the TiO₂ particles as well as the relatively low surface temperature with water, will allow the anatase phase conservation. The fact that the Cu and the TiO₂ aqueous droplets enter the plasma following the torch axis is very important to ensure the Cu-TiO₂ co-deposition.

4.2. Interpretation of the Morphology of Coatings

The plasma-sprayed Cu-based coatings were observed by SEM. The surfaces and the cross-sections are shown in Figure 7 and Figure 11, respectively. The calculated plasma jet enthalpy is 7.93 MJ.kg⁻¹ for the Cu coating, a value slightly higher than that obtained by Ranjan et al. [40] (6.75 MJ.kg⁻¹) under relatively similar deposition conditions but using a 9 MB plasma torch and commercially available grit blasted Cu substrates instead of 304L SS substrates. A remarkable difference is observed in the thickness obtained. As mentioned above, approximately 20 to 25 µm are deposited on the sandblasted 304L SS substrate. This highlights the important effect of the substrate type and its initial topography. In addition, as described above, the substrates were rotated during the spraying step (Figure 1). Yang et al. investigated the effect of the substrate holder (fixed or rotating) during the deposition of hydroxyapatite (HA) coatings on titanium, by plasma spraying [41]. The deposition conditions were the same for fixed and rotating samples cooled by compressed air, as in this study [41]. The authors reported larger splats and a rougher surface on the fixed samples. In addition, they explained that increasing the surface speed (i.e., rotation) allows the deposition of thinner layers during each cycle [41]. This is what happens with the plasma sprayed Cu coatings on 304L SS in this study. The application of rotation induces a surface velocity of the sample in front of the plasma jet. Yang et al. also reported a lower residual stress for rotating coatings compared to fixed coatings [41]. These results seem logical because the cooling process is not similar to the fixed processing. The samples are not constantly in front of the torch, and the heat and stress removal are easier and faster with rotation. This approach is interesting because the application is to create a gradient of properties between the core, generally made of low-cost materials such as SS or hard polymers, and the surface, which has properties such as antibacterial efficacy. It could then be observed that the relatively thin Cu coating covers the entire SS surface. Increasing the Cu thickness would imply higher metal leaching, which could be toxic at certain levels, as mentioned in the introduction of the paper. In addition, the samples produced by the rotational geometry generally have a lower surface temperature than the fixed ones due to heat dissipation, as mentioned above [41]. Therefore, under these conditions, Cu plasma spraying experiments could be carried out on thermosensitive substrates such as glass, acrylonitrile butadiene styrene (ABS) plastic, wood, and many other daily-use materials. Regarding the TiO₂ surface in Figure 7b, the splattering characteristic of plasma-sprayed coatings is not observed. In fact, plastic deformation, softening, and melting are more likely to be observed in metallic particles. Ceramic particles tend to fracture before or in contact with the substrate. Therefore, it is important to inspect the coating after deposition to verify its composition and crystalline structure. Finally, in terms of surface morphology, the Cu-TiO₂ coating (Figure 7c) is quite similar to pure Cu. The calculated enthalpy is 26.84 MJ.kg⁻¹, which explains the improvement in splattering splat compared to Cu in Figure 7a. However, some nonmelted particles remain, but this is not critical in this study since roughness is very important for bacterial adhesion and killing.

4.3. Intrinsic Properties of the Coatings

As explained earlier, the EDS elemental percent estimation (Figure 8 c, d) shows the presence of sub-stoichiometric TiO₂. Many references indicate the performance enhancement of blue TiO₂ in organic pollutant degradation [42], visible light harvesting [43], or CO₂ conversion [44]. For antibacterial applications, many researchers reported good efficiencies for doped white or blue TiO₂ [45,46,47], but none mentioned pure blue TiO₂ for such applications. As reported by Yu et al., blue TiO₂ contains Ti³⁺ states and high oxygen vacancy density, which allows absorption in the visible region and preserves absorption in the UV due to its band gap [43]. To explain the formation of blue TiO₂ in this study, H₂ present in the gas mixture is a strong reducing agent at the high temperatures generated in the plasma jet. This creates oxygen vacancies and perturbations in the anatase phase. This phenomenon has been known for many years [48] and has been clearly reported by Kumar et al. [49].

XPS was performed to further explore the intrinsic bonding in the coatings. Cu2p and O1s peaks detected on the Cu/SS coatings are shown in Figure 9. Jiang et al. reported high-resolution XPS spectra of CuO@NiZn nanoarrays [48]. As explained by the authors, the Cu2p1/2 and Cu2p3/2 signals at 940.6, 944.1, and 962.7 eV are satellite peaks. In fact, this phenomenon is often observed in paramagnetic states and is called the “shake-up process” [49]. In addition, two other envelopes are observed in Figure 9a. The first, with a lower binding energy, is attributed to the Cu2p3/2 state, which deconvolves into metallic Cu (Cu⁰) and Cu²⁺ at 932.2 and 933.7 eV, respectively. The higher binding energy envelope is attributed to Cu2p1/2, which deconvolves into Cu⁰ and Cu²⁺ at 952.1 and 953.5 eV, respectively [50,51]. The O1s peak in Figure 9b confirms the presence of copper oxide on the surface. The deconvolution shows the copper oxide and hydroxide at 530.1 and 531.2 eV, respectively, as reported by Jiang et al. [52]. No water adsorption is observed in this case, which generally occurs at around 532 eV [52]. From this analysis, it could be concluded that the coating is mainly composed of metallic copper with low amounts of oxide. This is particularly interesting since particles in the APS process tend to oxidize rapidly as they pass through the plasma toward the substrate. Many references state that copper oxides (CuO and Cu₂O) remain quite efficient for antibacterial applications [53,54,55]. However, metallic copper still shows better antimicrobial efficacy [56]. The hypothesis is that the mechanism is direct with metallic Cu, as opposed to copper oxides, where interactions between the ions and the organic matter of the cell medium, leading to reactive complex formation, are necessary to ensure bacterial elimination [57].

For TiO₂, the deconvolution of the Ti2p peaks (Figure 10a) showed the presence of Ti³⁺ at 463.52 eV [58]; the peaks at 458.46 and 464.30 eV are attributed to Ti⁴⁺ in stoichiometric TiO₂ [58]. Oxygen vacancies are also formed, as shown in Figure 10b [24,58]. The same tendency is observed for TiO₂ incorporated on Cu (Figure 10d,e), with a slight shift for Ti⁴⁺ and O^2- mainly due to the bonding with the metal matrix. The hydroxide groups disappear in Cu-TiO₂ because TiO₂ is more hygroscopic and has an affinity with the ambient humidity. However, changes are observed in Cu-TiO₂ (Figure 10c) compared to the Cu coating. Cu²⁺ is more pronounced than Cu⁰. Co-spraying tends to further oxidize the metallic Cu. Here, the shift to a slightly higher binding energy in the Cu²⁺ peaks could be attributed to a change in the chemical environment and binding distance [59]. A change from Cu-O to Cu-O-Ti could then be hypothesized. Oxygen vacancies are accentuated, and the Ti³⁺ peak is shifted to a lower binding energy.

4.4. Discussion of the Surface State of the Coatings

In addition to the intrinsic composition, the surface state is very important. The contact angle exceeds 90° on Cu, as shown in Figure 12b. Peng et al. demonstrated an increasing hydrophobicity for electrodeposited Cu coatings on steel [60]. The freshly deposited Cu layers were highly hydrophilic at the beginning, then the water contact angle increased up to 151° after 15 days of storage [60]. To explain this phenomenon, the Cu XRD pattern on 304L SS, recorded after a few days of storage, is shown in Figure 14a. Cu oxidation is either a consequence of the storage, which favors the formation of oxides on the surface of the coating or the oxidation of the Cu particles during the flight in the plasma and their incorporation inside the coating. In any case, according to previous studies [60], it is not the copper oxide that has the major influence on the surface wettability, but rather the carbon. In fact, the authors mentioned an increase in carbon content with storage time with the appearance of C-H, C=C, and C=O bonds detected in Fourier Transform InfraRed (FTIR) as evidence of hydrocarbon carbon adsorption with air aging [60]. In this study, the contact angle was measured several weeks after deposition. EDS showed 4 at.% of carbon in the first few days after plasma spraying. Thus, with storage time in air, the adsorbed hydrocarbons in the porosity of the coating may have increased the hydrophobicity. Consequently, this could affect the bacterial killing over time. For this purpose, the variation of the contact angle with aging time should be studied, as well as the carbon percentage tracking. The surface hydroxide groups contribute to the hydrophilicity of blue TiO₂ (Figure 12c) as well as the high porosity due to the nanometric scale of the starting powder. Nevertheless, the wettability of the Cu-TiO₂ was very close to that of pure Cu because the TiO₂ is not fully exposed on the surface and is probably infiltrated into the metallic matrix.

4.5. Crystalline Structure of the Coatings Compared to the Starting Material

Figure 14b shows the preservation of the anatase phase even though the high temperature leads to rutile formation. The anatase volume fraction could be calculated using the following formula [61] and reported in the work of Toma et al. [17]:

C_A = 8I_A/(8I_A + 13I_R)

(2)

where C_A is the anatase volume percent, I_A and I_R are the X-ray intensities of the anatase (1 0 1) and rutile (1 1 0) main peaks, respectively. According to formula (2), the calculated C_A is 0.554, which means that about 55% of the anatase volume is conserved, knowing that the starting material is pure anatase. Gao et al. presented an interesting study on the effect of heat treatment on the crystal structure of nano-TiO₂ [62]. They reported that the plot of 2lnD_t vs. 1/T (where D_t is the grain size at a time t and T is the temperature), derived from the Eastman particle growth theory, does not follow a linear trend. This is attributed to the anatase-rutile transition with increasing temperature. Their apparent activation energy calculations yielded 18.15 and 42.57 KJ·mol⁻¹ for anatase and rutile, respectively. For this, smaller grains are present in anatase at lower temperatures, and the progressive heating helps the defect organization, facilitating the coalescence and transition from anatase to rutile [62]. In this study, many parameters helped to maintain a high fraction of photocatalytic anatase in the coating. The residence time of the particles in the plasma jet was relatively short for the stand-off distance, the use of a small amount of hydrogen in the plasma-forming gas helped to compromise the enthalpy, and substrate cooling during deposition was quite important. Figure 14c shows small amounts of anatase and rutile coupled to Cu and CuO, but the metal peaks probably mask the TiO₂ due to the difference in amount between the two compounds.

4.6. Interpreting the Antibacterial Performance of Coatings

The final part of this discussion section deals with antibacterial efficacy on surfaces and for water treatment applications. As a starting point, the development of a biofilm on a surface is a series of steps, including (a) reversible attachment, (b) irreversible attachment, (c) 3D architecture growth, (d) maturation, and (e) spreading [63,64,65]. Initial attachment is the key to bacterial development and elimination by antibacterial agents. In addition, the solid interface plays an important role in this first step. It is known as the substratum compo sum, evoking the direct relationship between the surface state as well as its physical/chemical properties and bacterial attachment [64]. By analyzing the results in Figure 15, the porosity created on the 304L stainless steel after sandblasting favors the increase in surface area by sandblasting, which allows it to eliminate some bacteria after 1 h of contact. As reported by Nan et al., 304L SS also presents the advantage of anticorrosive surface oxide formation but does not show antibacterial efficacy [66]. However, this bacterial reduction on treated 304L SS does not necessarily mean ‘killing’. It could be a bacterial ‘inhibition’, and FCM analysis will provide more information on this hypothesis. In addition, the biofilm with live bacteria could cause Microbiologically Influenced Corrosion (MIC) due to the metabolic activity of the remaining microorganisms, thereby inducing a poor aging of 304L SS [66]. In parallel, the plasma sprayed Cu layer (Figure 15) showed complete elimination of E. coli after 1 h exposure (6.4 log), reaching the sterilization threshold. In addition, the copper oxide on the metal matrix acts as a passivation against the 304L SS MIC process [66]. The efficacy barely exceeded that observed for blue TiO₂ on sandblasted steel. In fact, Cu is known to have the maximum antibacterial efficacy for applications such as those presented in this work. Therefore, the decrease in activity on the TiO₂ surface was somewhat expected. In addition, the main efficacy of this material is based on its ‘photosensitivity’. UV irradiation is required to generate reactive oxygen species (ROS) to eliminate microorganisms [8]. The oxygen vacancies and Ti³⁺ in blue TiO₂ could also allow its activation under visible light. However, no irradiation test was performed in this study. The aim was to use the bare TiO₂ as a comparative ‘reference’ for Cu-TiO₂. The photocatalytic test on bacteria according to ASTM 2180 should be optimized to ensure that E. coli is not destroyed by thermal effect. The combination of both Cu and TiO₂, with Cu as the matrix, reduces the activity by 2 log compared to the bare metal. This result is a consequence of several causes, to mention the decrease of Cu amount by TiO₂ incorporation, the possible formation of Cu-Ti-O mixed phase as shown by XPS, and the further Cu oxidation, which is a less effective form of copper.

The FCM (Figure 16) provides more detail on the antibacterial results obtained by the enumeration technique (Figure 15). In the bacterial culture technique, no bacteria were counted after 1 h exposure to Cu, and the result is repeatable. However, FCM shows only 1.41% completely dead cells. Zuily et al. proposed a mechanism of action of copper in contact with bacteria [67]. The authors explained the steps required for complete bacterial destruction. Under aerobic conditions, which is the case in this study, ROS generated by Cu Fenton reaction [68] react with macromolecules and induce drastic cellular damage. With prolonged exposure to Cu, ROS and CuS (disulfides) continue to react and cause cellular stress. This phenomenon leads to cytoplasmic reactions as well as protein aggregation, and the cell eventually dies [67]. In this work, all samples were run in flow cytometry immediately after the 1 h exposure to Cu. What probably happens at this stage is that the cells are inactivated because they resist copper through their molecular chaperones, as explained by Zuily et al. [67]. In contrast, for TiO₂ (Figure 16b), E. coli are more ‘killed’ rather than ‘inhibited’ after 1 h of contact. This tendency is also observed for Cu-TiO₂ (Figure 16c), which could be considered as an intermediate in this case. Compared to Cu, Cu-TiO₂ shows more ‘killed’ cells, which means that TiO₂ switches the mechanism from ‘inhibition’ to direct ‘killing’ for the same contact time. This could be explained by a main hypothesis related to the nanometric size of TiO_2, which helps it to infiltrate into the bacteria.

To conclude this study, a test was performed on water contaminated with E. coli. Without irradiation, Figure 17a confirms what was obtained by direct surface inoculation according to ASTM2180. Cu still shows the best antibacterial activity compared to the other materials. Nevertheless, the test assembly in an aqueous medium was easier than on the surface. For this, the Cu-TiO₂ sample was illuminated at 380 nm to check if TiO₂ could have a positive effect on its activity. The system was maintained at room temperature, and a test was performed on a bare aqueous solution containing the same E. coli concentration but without a coating. In this test, hardly any E. coli elimination was observed. Two conclusions can be drawn from Figure 17b. Increasing the contact time increases bacterial elimination even on the bare substrate. In addition, illumination increases bacterial elimination percentage after 60 and 90 min. This highlights the positive effect of TiO₂ addition. This point is important for both biological contaminants (like E. coli) and micropollutants (chemical compounds). Yadav et al. explained the mineralization process of organic compounds induced by ROS generated on TiO₂ after illumination [69]. They also proposed a mechanism for bacteria where OH* and H⁺ act as strong oxidants that disturb the normal functioning of microorganisms [69].

5. Conclusions

In conclusion, this paper proposes the preparation of antibacterial Cu-based coatings on 304L stainless steel substrate by plasma spraying process using Triplex Pro 210^TM torch. It also highlights the importance of surface conditions and composition, either on the first step of biofilm formation, which is ‘bacterial reversible then irreversible attachment’ or on the antibacterial efficacy lifetime and coating aging. The important points to note from this study are as follows:

(1)

A rotating substrate holder allows the deposition of relatively thin Cu, TiO₂, and Cu-TiO₂ coatings at each cycle because the samples are not constantly in front of the torch. This facilitates heat and stress removal;

(2)

XPS results showed the presence of both metallic Cu and CuO on the surface. The antimicrobial mechanism is reported to be direct with Cu and complex-dependent with Cu ions released from CuO. As for TiO₂, oxygen vacancies and Ti³⁺ ions were detected in the coating. This is due to the reducing atmosphere (H₂), which also gives TiO₂ a blue color. Further Cu oxidation was observed in Cu-TiO₂, but TiO₂ was detected in the XRD patterns;

(3)

Sandblasting is important to increase the capillarity and the surface contact for more hydrophilicity. For this reason, sandblasting of the 304L SS substrate is very important in addition to being critical for the Cu coating adhesion. Cu coatings were hydrophobic due to hydrocarbon deposition, while TiO₂ showed hydrophilic behavior due to the surface hydroxide groups on the surface. This wettability is lost when combined with Cu;

(4)

The bacterial reduction reached 6.4 log on Cu after 1 h, exceeding the sterilization threshold, while on sandblasted 304L SS, it barely exceeded decontamination (2 log). This result is remarkable compared to those of the literature and a kinetic follow-up should be performed to estimate the exact interval to eliminate all the E. coli. As expected, TiO₂ was less efficient, and a decrease in efficiency was observed on Cu-TiO₂ without illumination;

(5)

The flow cytograms showed significant bacterial inhibition on the coatings compared to the bare substrates. However, the percentage of dead bacteria was not very high on the samples analyzed directly by FCM after contact. Nanometric TiO₂ switched the mechanism from inhibition to killing due to its high surface area. Counting by culture technique does not allow differentiation between the two behaviors;

(6)

The tests in water contaminated with E. coli showed the same tendency as on the surface, according to ASTM2180. However, the photocatalytic effect of TiO₂ on bacterial elimination was demonstrated after 380 nm light irradiation.

In the future, the contact angle should be measured at several time intervals after deposition to highlight the effect on surface wettability. In addition, an interesting approach currently underway in the laboratory is to fully expose TiO₂ on the surface. This approach may be beneficial to avoid further Cu oxidation and intermetallic compound formation. Tests according to ASTM2180 should be performed under illumination, under optimized conditions. In addition, samples should be analyzed by FCM at several time intervals after contact. More time may be required for the proteins to aggregate and damage the bacterial DNA. For this reason, the bacteria are inhibited rather than killed in an initial window. Further studies will include a complete microbiological analysis including various bacterial indicators such as Gram-positive (Enterococcus faecalis) and spore-forming bacteria (Clostridium perfringens).