Optimization of Cold Gas Dynamic Spray Coatings Using Agglomerated Al–Zn–TiO₂ Powders on Steel

Abstract

Cold gas dynamic spraying (CGDS) enables the production of protective coatings without melting or oxidation. In this study, Al–Zn–TiO₂ composite powders were prepared by wet agglomeration with binders and by dry mechanical mixing, and deposited onto mild steel substrates. COMSOL simulations of gas dynamics and particle acceleration identified optimal parameters (0.6 MPa, 600 °C, 15 mm, 90°), which were then validated experimentally. Coatings formed under these conditions exhibited dense microstructures, minimal porosity (~0.5%), and continuous, defect-free interfaces with the substrate. SEM and XRD confirmed solid-state bonding without new phase formation. Corrosion tests in 3.5% NaCl revealed a tenfold reduction in corrosion current density compared to bare steel, resulting from synergistic sacrificial (Zn), barrier (Al), and reinforcing/passivating (TiO₂) effects. Tribological tests demonstrated reduced friction (CoF ≈ 0.4–0.5) and wear volume. Compared with reported Al- or Zn-based cold- and thermal-sprayed coatings, the optimized Al–Zn–TiO₂ system shows superior performance, highlighting its potential for industrial anti-corrosion and wear-resistant applications.

1. Introduction

Cold gas dynamic spray (CGDS) is a solid-state coating method in which powder particles are accelerated by a high-pressure gas jet (often through a de Laval nozzle) to supersonic speeds and bonded onto a substrate by plastic deformation [1,2,3,4,5]. Because the particles remain below their melting point during deposition, oxidation and other thermal degradation are minimized [1,2,3,4,5], preserving the original powder microstructure. This leads to dense, low-porosity coatings with fine microstructure and primarily compressive residual stresses [1,2,3,4,5]. In contrast to thermal spray, CGDS can coat temperature-sensitive substrates without metallurgical damage. Recent reviews have documented the materials science of CGDS, noting that a wide range of feedstocks—from pure metals to alloys to metal–matrix composites—can be deposited [6,7]. The solid-state bonding mechanism (via severe plastic deformation at impact) enables high-purity coatings and uniform microstructures [8,9]. Critical parameters such as gas pressure, gas temperature, nozzle geometry, and stand-off distance govern the particle velocity and impact conditions [8,9,10], which in turn determine whether particles exceed the critical velocity needed for bonding [8,9,10,11,12,13]. Because CGDS accommodates diverse materials, composite coatings have attracted much interest. Metal–matrix composite (MMC) coatings—combining a ductile matrix with hard reinforcement particles—can achieve a balance of corrosion protection and mechanical strength that pure metals alone cannot provide [14]. For example, ductile matrices such as Al, Cu or Ni have been reinforced with ceramic or nanomaterial additives (Al₂O₃, SiC, WC, TiC, carbon nanotubes, etc.) to improve hardness and wear resistance [15,16,17,18]. Cold-sprayed Cu–CNT [15], WC–Ni [16], Ni–WC [17], Ni–Ti [18], and Cr₃C₂-based [17] composite coatings have all shown superior tribological performance compared to the base metal [19]. To achieve uniform composite powders, advanced feedstock preparation methods such as spray-drying or powder-satelliting are commonly applied [20]. For example, the binder-assisted satelliting of TiC onto Al has been shown to produce a more uniform composite feedstock and more stable spray flow than a simple mechanical blend [20]. Following this principle, a similar agglomeration strategy was adopted here to combine Al, Zn, and TiO₂ into a single feedstock for cold-spraying.

A key application of CGDS is corrosion protection of steel. Zinc and aluminum are widely used as protective coatings: Zn provides sacrificial anodic protection, while Al forms a dense passive barrier [21,22,23]. Cold-sprayed Zn–Al alloy coatings have gained considerable attention because they combine the galvanic protection of Zn with the corrosion resistance of Al [23]. Studies on thermal spraying have shown that Zn–Al systems (e.g., Zn85–Al15) can benefit from post-treatments such as sealing, which further reduce corrosion rates. These findings indicate that Al–Zn composite coatings (sometimes referred to as pseudo-alloys) can significantly enhance steel protection. However, pure Al or Zn coatings tend to be relatively soft or porous, which may limit their wear resistance. To address this issue, hard oxide reinforcements are often incorporated. Ceramic oxides such as alumina and titania [24] are well-known for improving hardness and wear performance. Cold-sprayed TiO₂ coatings have also been successfully produced using nitrogen as the process gas [24], and the addition of TiO₂ particles into a metal matrix is expected to enhance coating hardness without compromising the metal’s corrosion-protective function. Cold-sprayed MMC coatings containing ceramic reinforcements generally exhibit improved mechanical and tribological performance [25]. Recent developments in hybrid metal–ceramic coatings further support the effectiveness of composite strategies. For instance, a recent study [26] demonstrated that integrating ceramic reinforcements into aluminum-based matrices significantly enhances wear and corrosion resistance through improved phase dispersion and interfacial stability. This is consistent with our results, where TiO₂ reinforcement within the Al–Zn matrix contributed to reduced wear volume and improved coating integrity. Therefore, an Al–Zn matrix with dispersed TiO₂ particles could provide both enhanced corrosion protection and increased wear resistance.

Optimizing a multi-component cold-spray process is complex, which makes computational modeling a valuable tool. Various analytical and numerical models have been developed to predict particle dynamics and deposition behavior in cold spray [27,28]. Gas–particle flow in mixed-powder streams has been analyzed to clarify how key parameters such as gas pressure, temperature, nozzle design, and particle size affect the deposition window. Computational fluid dynamics (CFD) and finite element methods have been widely applied to simulate particle trajectories, velocities, and impact stresses under varying spray conditions [27,28,29]. Such models enable the prediction of critical velocities and coating formation thresholds, providing guidance for process parameter selection. Recent works emphasize how advanced modeling approaches are increasingly integrated into cold-spray and additive manufacturing process design [29,30]. In addition, the mechanical performance of cold-sprayed coatings, including fatigue behavior, has been investigated to highlight the importance of producing high-quality deposits [31].

While the fundamentals of CGDS highlight its advantages over thermal spray, a growing research focus has shifted toward composite coatings, where ductile matrices are reinforced with ceramic or nanostructured phases. Such systems offer multifunctional properties, combining corrosion protection with enhanced mechanical and tribological behavior. At the same time, numerical modeling has become an essential tool to bridge processing conditions and coating performance, enabling parameter optimization without exhaustive experimentation. Therefore, this study integrates both experimental and modeling approaches to evaluate Al–Zn–TiO₂ composite coatings.

In this work, we combine COMSOL Multiphysics^® simulations with experiments to optimize CGDS deposition of agglomerated Al–Zn–TiO₂ powders on steel. The aim is to identify spray conditions that maximize particle velocity and bonding for our composite feedstock. We simulate the gas flow and particle impact for various pressures, temperatures, stand-off distances, and spray angles, and predict the resulting coating deposition quality. Guided by the model, we determined that the optimum parameters are a carrier gas pressure of 0.6 MPa, a gas preheat temperature of 600 °C (measured at the heater exit, yielding ~500 K at the substrate after expansion), and a stand-off distance of ~15 mm with normal impingement. These conditions were then applied in experiments to deposit Al–Zn–TiO₂ coatings, and the results were evaluated in terms of microstructure, phase composition, corrosion protection, and tribological performance.

The novelty of this work lies in the combined experimental–numerical optimization of Al–Zn–TiO₂ composite coatings produced by cold gas dynamic spraying. Unlike previous studies, which focused either on binary Al–Zn systems or on oxide reinforcements in other metallic matrices, this study establishes optimized parameters for a three-component Al–Zn–TiO₂ system. The scientific contribution consists of (i) demonstrating the superior coating density and uniformity achieved using agglomerated powders compared to mechanically mixed powders, and (ii) confirming the synergistic action of Zn (sacrificial protection), Al (barrier passivation), and TiO₂ (hard reinforcement improving wear resistance) in one multifunctional coating.

2. Materials and Methods

2.1. Feedstock Powder and Substrate Preparation

Powder Composition: The coating material was a nanocomposite powder consisting of aluminum (Al) as the ductile metallic matrix, zinc (Zn) as a secondary metal for sacrificial protection, and titanium dioxide (TiO₂) as a ceramic reinforcement for hardness and stability (with potential photocatalytic effects). Commercially available Al powder with irregular morphology (particle size ~20–45 µm) and Zn powder with flake morphology (~10–30 µm) were used. The TiO₂ was nano- to sub-micron anatase, with particle agglomerates ranging from 30 to 80 nm. Two methods were explored to prepare a homogeneous Al–Zn–TiO₂ powder blend:

-

Chemical Agglomeration Method: In this approach, Al and Zn powders (pre-mixed at approximately 80:15 wt%) were combined with TiO₂ nanoparticles (~5 wt%) in a liquid medium with surfactants to promote TiO₂ attachment onto Al particle surfaces. Specifically, polyvinyl alcohol (PVA) was used as a temporary binder to adhere TiO₂ to Al, while carboxymethyl cellulose (CMC, 2 wt% in water) was added as a rheological stabilizer. The suspension containing Al, Zn, and TiO₂ was ultrasonically agitated and stirred, then dried and lightly milled to yield agglomerated composite particles. This method produced Al–Zn particles uniformly decorated with fine TiO₂. Coatings deposited using the chemically agglomerated powder exhibited a dense, continuous microstructure with minimal porosity (<0.5%) and strong bonding to the substrate. The superior coating quality is attributed to the homogeneous distribution of TiO₂ and the enhanced inter-particle cohesion provided by the binder during powder preparation.

PVA was selected as a binder due to its strong adhesion to metallic surfaces and its ability to anchor fine TiO₂ particles onto Al cores, while CMC was introduced as a stabilizer to improve suspension homogeneity and prevent nanoparticle re-agglomeration during drying. Alternative binders such as polyvinylpyrrolidone (PVP) or polyethylene glycol (PEG) have also been reported in the literature; however, PVA/CMC was chosen for its low cost, water solubility, and proven effectiveness in dispersing oxide nanoparticles in metallic matrices.

-

Mechanical Mixing Method: In this method, the Al, Zn, and TiO₂ powders were combined by dry blending without any binders. A vibratory ball mixer was used to mix the powders (same ~80:15:5 weight ratio) for 10 min at 57 Hz (360 V), which facilitated distribution of Zn and TiO₂ among the Al particles. This simple mechanical blending produced a composite powder with adequate flowability but less controlled particle dispersion. Many fine TiO₂ and Zn particles remained unattached to the Al particles or formed larger agglomerates due to the lack of chemical binding in the feedstock. Consequently, coatings produced from the mechanically mixed powder were comparatively less dense (with apparent porosity on the order of ~1–2%), with heterogeneous microstructures and weaker adhesion to the substrate. These inferior results are attributed to the non-uniform dispersion of the reinforcement phases and the absence of strong inter-particle bonding prior to spraying. Given these shortcomings, the chemically agglomerated powder was favored for subsequent cold-spray experiments in this study.

The powder size distribution was assumed based on supplier specifications and SEM observations, consistent with the particle size range used in CFD simulations. The feedstock composition for all powders was ~80 wt% Al, ~15 wt% Zn, and ~5 wt% TiO₂, as targeted. Figure 1 shows SEM images of the as-received Al–Zn and TiO₂ powders. The Al–Zn particles exhibited irregular and flake-like morphologies with size distributions between 20 and 50 µm, while TiO₂ nanoparticles formed fine agglomerates in the 30–80 nm range. These micrographs provide a reference for the baseline particle morphology and size distribution prior to blending.

Substrate Preparation: The substrates were coupons of mild steel (structural steel grade S235JR, equivalent to ASTM A36, ArcelorMittal, Luxembourg City, Luxembourg) [32]. Two types of sample geometries were prepared to examine coating behavior on different surface curvatures: (1) flat coupons of 50 mm × 50 mm area and 10 mm thickness, and (2) curved semi-spherical coupons with ~30 mm radius of curvature (dome-shaped, 5 mm thick). A total of 22 substrate samples were prepared, allowing various spray parameter conditions to be tested. Prior to coating, the steel substrates underwent surface preparation to ensure adequate coating adhesion. The steel surfaces were first ground with abrasive paper and degreased with acetone to remove any rust or contaminants. Then, a grit-blasting (abrasive sandblasting) treatment was applied. This abrasive blasting increased the surface roughness and created a textured profile to promote mechanical interlocking of the cold-sprayed particles. Typical surface roughness after blasting was in the range of Ra ≈ 5–6 µm (arithmetic average), with peak-to-valley height Rz on the order of 30–50 µm. The roughened substrates were cleaned of any loose grit and again degreased before coating. This surface preparation follows common practice for thermal spray processes to improve coating adhesion.

2.2. Cold Gas Dynamic Spraying Process

Cold-spraying was carried out using a high-pressure CGDS system (Dymet 404, JSC Obninsk Center for Powder Spraying, Obninsk, Russia) equipped with a de Laval convergent–divergent nozzle. The general appearance of the Dymet 404 cold-spray apparatus is shown in Figure 2. The nozzle had a total length of 100 mm, consisting of a 30 mm converging section, a 10 mm throat, and a 60 mm diverging section. The throat diameter was 2 mm, and the outlet diameter was 6 mm. This configuration was designed to accelerate the gas to supersonic speeds, enabling efficient particle entrainment and deposition. The stand-off distance (nozzle exit to substrate) was varied between 5 mm and 25 mm as a process parameter.

Spray Parameters: Based on preliminary modeling and the literature data, the primary variables selected for optimization were the gas temperature, gas pressure, stand-off distance, and powder feed rate. Compressed air was used as the propellant gas for all runs. The gas temperature was set in the range 400–600 °C (measured at the heater exit, corresponding to ~673–873 K), and the gas pressure (stagnation pressure at the nozzle inlet) was ≈0.8 MPa in the supply line. Due to pressure losses in the system, the effective pressure at the nozzle inlet was 0.6 MPa. Thus, in our modeling and analysis we consider 4.0 × 10⁵ to 6.0 × 10⁵ Pa as the range of nozzle inlet pressures. The powder feed rate was controlled by a twin-screw feeder and was varied between 0.4 g/s and 0.6 g/s. Each coating was built up by multiple passes of the nozzle across the substrate, with typically 10–15 passes applied per sample to achieve sufficient thickness.

Table 1. Cold gas dynamic spray process parameter ranges investigated in this study.

Parameter	Range (Min–Max)
Gas pressure	4.0 × 105–6.0 × 105 Pa
Gas temperature	400–600 °C (673–873 K)
Stand-off distance	10–25 mm
Powder feed rate	0.4–0.6 g/s
Nozzle traverse speed	10 mm/s (fixed)

summarizes the nominal ranges of the main process parameters investigated.

For consistency, all coatings were deposited with the nozzle oriented perpendicular to the substrate (spray angle 90°). The substrate was mounted on a CNC X–Y traverse to ensure uniform scanning of the spray over the surface. The traverse speed was held constant at 10 mm/s for all experiments. Coating thickness per pass was on the order of a few micrometers; multiple passes were used to reach a final thickness of ~50–80 µm, depending on the specific test requirements for each sample. Both mechanically mixed and chemically agglomerated Al–Zn–TiO₂ powders were sprayed using the same equipment and under similar process conditions, as detailed in

Table 2. Deposition parameters used for cold spray-coating of Al–Zn–TiO₂ powders.

Sample	Pressure (MPa)	Temperature (°C)	Standoff Distance (mm)	Feedrate (g/s)
S235JR1	0.4	600	10	0.4
S235JR2	0.5	600	15	0.5
S235JR3	0.6	650	20	0.6
S235JR4	0.6	700	25	0.6

.

The specific deposition conditions applied in the experiments are summarized in Table 2. These include the selected gas pressure and temperature, powder feed rate, nozzle stand-off distance, and number of coating passes used for each spray series. These parameters were optimized based on both preliminary modeling and trial experiments.

Each deposition condition was repeated at least twice, resulting in duplicate samples for each parameter set to ensure reproducibility. In total, 22 coated samples were produced across the four main parameter combinations. Measured coating properties represent averages across repetitions, with deviations remaining within acceptable limits, confirming high repeatability of the CGDS process

2.3. Numerical Modeling with COMSOL Multiphysics

To gain insight into the gas flow dynamics and particle impact conditions in the cold-spray process, a comprehensive numerical model was developed using COMSOL Multiphysics (version 6.2, COMSOL AB). The model coupled several physics interfaces to simulate the multi-phase cold-spray environment: compressible turbulent gas flow, heat transfer, particle motion, and solid mechanics for particle impact. The geometric domain of the simulation included the de Laval nozzle (interior flow) and a region of the ambient atmosphere extending to the substrate surface. A three-dimensional (3D) model was constructed to capture any asymmetries, but given the axial symmetry of the nozzle flow, the results were effectively axisymmetric up to the point of particle impact.

Gas Flow and Thermal Field: The gas (air) flow through the nozzle was modeled using the Turbulent Flow (k–ε) interface in COMSOL, solving the Reynolds-averaged Navier–Stokes (RANS) equations for compressible flow. The inlet boundary was defined with the stagnation pressure and temperature (e.g., 0.6 MPa and 600 °C for a typical run), while the outlet (far-field) was set to ambient pressure (0.1 MPa). The nozzle walls were assumed to be adiabatic with no slip. The flow rapidly expanded in the diverging section of the nozzle, accelerating to supersonic speeds. The simulation captured the formation of a jet plume exiting the nozzle and impinging on the substrate. A Heat Transfer in Fluids interface was included to compute the temperature distribution in the gas. Air properties (specific heat, thermal conductivity) were defined as functions of temperature. The gas was preheated, but due to expansion cooling, the temperature dropped significantly in the jet core. The model showed, for example, gas velocity distributions on the order of ~1200 m/s at the nozzle exit (for 0.6 MPa, 600 °C input) and a temperature drop to ~327 °C in the expanding jet (from 600 °C stagnation). These flow simulations provided the gas velocity and temperature conditions that the powder particles experience in flight.

2.4. Particle Acceleration and Impact (Model Configuration)

To investigate the cold-spray particle dynamics, a coupled computational fluid dynamics (CFD) and discrete particle-tracing model was developed using COMSOL Multiphysics. The nozzle geometry was defined based on a de Laval shape with a throat diameter of 2 mm and an outlet diameter of 6 mm. The gas phase (compressed air) was modeled as compressible turbulent flow using the standard k–ε turbulence model. Heat transfer in fluids was included to account for expansion cooling. The particulate phase was simulated using the Particle Tracing for Fluid Flow module, where particles were treated as spherical, non-interacting Lagrangian entities. The primary particle diameter was set to ~20 µm, with an effective density of 2700 kg/m³, representing Al–Zn–TiO₂ composite particles. Drag force, gravity, and slip correction were included. Particles were injected at the nozzle inlet with initial velocity equal to the local gas flow. The stand-off distance was varied between 10 and 25 mm in the domain to examine its effect on particle velocity and spread.

Modeling Assumptions: In implementing the particle acceleration and impact model, several simplifying assumptions were adopted. First, all powder particles were idealized as perfect spheres (even though the actual Al–Zn feedstock particles had irregular morphologies and carried TiO₂ nanoparticle agglomerates, as seen in Figure 1a,b). This spherical shape assumption simplifies the calculation of drag and momentum transfer. Second, a single representative particle size (20 µm diameter) was used, effectively treating the powder as monodisperse; the actual particle size distribution (approximately 20–50 µm for Al–Zn particles, according to supplier specifications and SEM observations) was not explicitly modeled. Third, particle–particle interactions were neglected: each particle was tracked as an independent entity with no inter-particle collisions or agglomeration in flight. This non-interacting assumption (along with the low overall particle concentration) implies one-way coupling, meaning the particles do not noticeably influence the gas flow field. It was further assumed that particles remain solid throughout their transit (no melting, fragmentation, or phase change) and maintain constant material properties. The initial particle velocity at injection was set equal to the local gas velocity (i.e., particles immediately achieve the carrier gas speed with no slip upon injection). These modeling assumptions, which are common in cold-spray simulations, focus the analysis on the primary physics of particle acceleration and impact while keeping the computational model tractable.

2.5. Corrosion Testing

The corrosion resistance of the cold-sprayed Al–Zn–TiO₂ coatings was evaluated using two complementary techniques: salt-spray (fog) testing and potentiodynamic polarization in a chloride environment.

Salt-spray tests were conducted in accordance with ISO 9227 standards using a 5% NaCl fog chamber [33]. The coated and uncoated steel specimens were exposed to the salt fog environment for 168 h at 35 °C, with the goal of assessing visible red rust formation and coating integrity under accelerated corrosive conditions.

Electrochemical corrosion testing was performed using a Corrtest CS350M potentiostat/galvanostat system (Wuhan Cost Instrument Co., Ltd., Wuhan, China) in a standard three-electrode configuration. The coated specimen served as the working electrode, a saturated Ag/AgCl electrode as the reference, and a platinum wire as the counter-electrode. The electrolyte was an aerated 3.5 wt% NaCl aqueous solution. Potentiodynamic polarization was conducted from −250 mV to +250 mV vs. OCP at a scan rate of 1 mV/s. The corrosion potential (E_corr), corrosion current density (I_corr), and estimated corrosion rate were derived from the Tafel extrapolation of the polarization curves.

2.6. Tribological Testing

The wear resistance and friction behavior of the coatings were evaluated using a dry sliding test setup. A universal tribometer (Anton Paar TRB3, Graz, Austria) was used to perform ball-on-disk wear testing at room temperature under ambient conditions.

The counter body was a 6 mm diameter hardened steel ball (100Cr6, HRC 62), sliding against the coated steel substrate. The applied normal load was 5 N, and the sliding speed was maintained at 10 cm/s. The total sliding distance was 100 m, and the stroke length was 5 mm in the reciprocating linear motion. The coefficient of friction (CoF) was recorded in real-time, and the wear track was analyzed using optical profilometry and SEM to determine the wear volume and mechanism.

3. Results and Discussion

3.1. Simulation Results and Process Optimization

The CFD simulation predicted that at 0.6 MPa and 873 K, the nozzle produced a supersonic jet with gas exit velocities up to ~1200 m/s (Figure 3a). Under these conditions, the Al–Zn–TiO₂ particles accelerated to final impact speeds of approximately 600–700 m/s at a stand-off distance of 15 mm (Figure 3c). Shorter distances (e.g., 10 mm) caused under-expanded flow and excessive impact forces, whereas longer distances (>25 mm) allowed the particles to decelerate, lowering the impact velocity below the critical threshold for bonding.

The impact model showed that, upon collision, the particles experienced severe plastic deformation and flattened into disk-shaped splats. The von Mises stress within the particle peaked around 300 MPa, confirming localized plastic flow (Figure 3b). To put the 300 MPa impact stress into context, this level is an order of magnitude above the yield strength of pure aluminum (on the order of only tens of MPa) and is comparable to or even exceeds the yield of mild steel (roughly 250 MPa). Such a high von Mises stress confirms that the particle is driven well beyond its elastic limit, undergoing intense plastic flow, while the substrate surface also yields locally. These extreme stress and strain-rate conditions favor the formation of an adiabatic shear instability (a localized shear band with thermal softening) at the impact interface rather than any bulk melting of the material [9]. In other words, the bonding remains a solid-state process dominated by severe plastic deformation and interfacial heating, with no evidence of localized melting or uncontrolled shear failure. The observed deformation and interfacial condition are thus consistent with metallurgical solid-state bonding via plastic deformation, rather than indicating any melting or adiabatic shear instability that compromises the integrity of the deposit. The substrate region directly beneath the particle exhibited a stress field approaching its yield limit, resulting in a shallow indentation and minor plastic strain, but without significant substrate damage. The high strain rate and localized heating supported the formation of an adiabatic shear layer, which promotes metallurgical bonding.

Simulation results demonstrated that adhesion efficiency increases markedly with rising gas pressure and temperature, reaching a stable regime beyond ~0.6 MPa and 600 °C. Based on these findings, the cold-spray process was conducted at 600 °C, 0.6 MPa, and a stand-off distance of approximately 15 mm with a moderate powder feed rate (~0.5 g/s), as these parameters ensured sufficient particle velocity and deformation for metallurgical bonding. The resulting coatings exhibited dense lamellar structures with no signs of delamination, in line with previous studies emphasizing the critical role of impact conditions in promoting adhesion [34,35].

3.2. Comparison of Modeling Predictions with Experimental Results

A key aspect of this work is the close interplay between modeling and experimentation. The numerical simulations performed in COMSOL provided detailed predictions of gas dynamics, particle acceleration, and impact behavior, which were in strong agreement with the experimental observations. This integration allowed for a more accurate interpretation of the coating morphology, porosity trends, and deposition conditions. The consistency between simulated particle behavior and the observed microstructural features validates the predictive capability of the model and supports its use in optimizing cold-spray parameters.

Gas Flow and Particle Velocity: The CFD model predicted that at the optimal condition (pressure ~0.6 MPa, temperature ~600 °C, 15 mm stand-off), particle velocities around 500–600 m/s would be achieved (for 20 µm Al particles) (Figure 3a). Experimentally, we did not have a direct measure of particle velocity in flight; however, indirect evidence supports these predictions. The fact that the coating could be deposited at all, and with high adhesion, implies particles exceeded the critical velocity for bonding (often on the order of 500 m/s for Al on steel). In trials at lower pressure (0.4 MPa) or lower temperature (400 °C), coatings were either not adhering or had very poor adhesion, indicating sub-critical velocities. The model’s trend that increasing pressure and temperature increases particle speed was qualitatively confirmed by the improved deposition observed at higher pressure and temperature. Additionally, the particle velocity distribution from the model helps explain the coating thickness uniformity: the plume of particles was concentrated within a diameter of ~10–15 mm on the substrate, consistent with the ~12 mm diameter actual coating spot observed when spraying a single pass at 15 mm stand-off. At very short stand-off (5 mm), the model showed the gas flow was not fully expanded and had strong shock pressures, which could cause particles to erode the substrate or shatter. Indeed, experimentally, at 5 mm stand-off we observed some coating but also signs of particle splashing and substrate abrasion, supporting the model’s indication that 5 mm is too close for stable deposition. At 25 mm, the model showed particle speed dropping significantly (due to drag in air)—the experiment concordantly showed almost no coating build-up at 25 mm, as particles had decelerated below critical velocity. Thus, the stand-off window of ~10–20 mm for effective deposition predicted by the simulation matched the empirical outcome (with 15 mm being optimal).

Interfacial Bonding Mechanisms in Cold-Sprayed Coatings: Numerical simulation of particle–substrate interactions confirmed that at impact velocities above 600 m/s, particles undergo substantial plastic deformation with interfacial jetting—conditions known to be necessary for achieving metallurgical bonding in cold-spray processes. The high strain rates and localized temperature rise at the interface support oxide disruption and intimate contact between the coating material and substrate, resulting in strong adhesion.

Structural analysis of the coatings, including cross-sectional SEM imaging, demonstrated continuous splat boundaries, dense lamellar morphology, and the complete absence of delamination or interfacial defects. These features provide clear evidence of effective bonding between the splats and with the substrate. Coatings deposited under optimized parameters (0.6 MPa, 600 °C) exhibited uniform microstructure and cohesive integrity, while those produced at lower pressures showed increased porosity and disrupted splat interfaces—fully consistent with the modeled reduction in particle velocity and bonding efficiency.

The predicted stress distribution also indicated that the substrate remained structurally unaffected during impact, with deformation localized to the coating region. Collectively, the modeling results and microstructural observations confirm that the deposition conditions used in this study enabled sufficient particle acceleration, deformation, and bonding to ensure the formation of a mechanically robust and metallurgically continuous coating.

3.3. Simulation of Gas Flow and Particle Dynamics

Numerical simulations using COMSOL Multiphysics provided a detailed understanding of the spray process. The gas jet reached supersonic speeds, producing a high-velocity core where particles accelerated to 500–600 m/s under optimal conditions (0.6 MPa, 600 °C, 15 mm stand-off, 90° spray angle) (Figure 3a). The impact simulations indicated that these velocities generated sufficient plastic deformation and interfacial stress for effective bonding. Curved surface modeling showed that particle impact velocity decreased slightly at the edges due to deflection, explaining the minor thickness variations. The numerical predictions matched experimental trends for deposition efficiency, roughness, and coating quality, validating the model as a reliable tool for process optimization.

Optimal Parameter Identification: The optimal deposition parameters were identified through a combined analysis of gas dynamics modeling and experimental observations. Coatings deposited at 0.6 MPa gas pressure, 600 °C gas temperature, and a 15 mm stand-off distance exhibited the most favorable microstructural features—a dense splat packing, uniform lamellar interfaces, and absence of delamination.

These optimal conditions were directly confirmed by cross-sectional SEM analysis (Figure 4). The image shows a dense, continuous coating with a thickness of approximately 20–32 µm, tightly bonded to the steel substrate. No delamination or large interfacial voids are visible, and the splats exhibit uniform stacking across the layer. The compact microstructure, together with the uniform lamellar interfaces, validates that the combination of 0.6 MPa gas pressure, 600 °C gas temperature, and 15 mm stand-off distance yields the most adherent and defect-free coatings.

When slightly deviating from these conditions—e.g., increasing the stand-off distance to 20 mm under the same pressure and temperature—coatings showed reduced thickness and decreased interfacial continuity. Likewise, raising the gas temperature to 700 °C did not further improve coating integrity and approached the thermal limits of the equipment. These outcomes confirm that a high-pressure and high-temperature gas flow, in combination with an optimized stand-off distance, is essential for accelerating particles to velocities that exceed the critical threshold for bonding.

The powder feed rate had a comparatively smaller influence on coating quality. Both 0.4 g/s and 0.6 g/s feed rates yielded similar bonding performance, though the higher feed rate produced slightly thicker coatings. A value of 0.5 g/s was adopted as a balanced setting to ensure consistent particle flux without excessive overspray or clustering.

In summary, the excellent agreement between numerically predicted conditions and observed coating performance demonstrates the effectiveness of using physics-based modeling to guide parameter selection in cold gas dynamic spraying. The results confirm that maintaining sufficient gas flow energy—via elevated pressure and temperature—combined with an appropriate stand-off distance and moderate powder feed rate, ensures optimal particle velocity and deformation for achieving high-quality coatings.

3.3.1. Effect of Process Parameters on Velocity

To validate the selection of optimal spraying conditions, a series of CFD simulations were conducted at a constant gas temperature of 800 K and a fixed nozzle-to-substrate distance of 0.01 m, while varying the working gas pressure from 4 × 10⁵ Pa to 6 × 10⁵ Pa. The velocity fields of the gas–particle flow were visualized for each condition (Figure 5a–c). The simulation results clearly show that increasing the pressure leads to a significant increase in particle velocity at the nozzle exit.

At 4 × 10⁵ Pa, the maximum velocity of particles was approximately 620 m/s. At 5 × 10⁵ Pa, it increased to around 710 m/s. At 6 × 10⁵ Pa, particle velocities exceeded 820 m/s. This demonstrates that higher chamber pressures result in greater acceleration of particles due to increased pressure gradients along the nozzle. Since particle velocity directly affects deposition behavior (e.g., the extent of particle deformation upon impact and bonding), the condition with P = 6 × 10⁵ Pa and T = 800 K is considered optimal for achieving high-quality coatings.

3.3.2. Heat Transfer Analysis

Figure 6a,b show the temporal evolution of the gas temperature field within the nozzle from t = 0 s to t = 0.05 s, modeled with a time step of 0.001 s. At the initial moment (t = 0 s), the temperature remains relatively uniform throughout the nozzle length at ~800 K, indicating steady initial thermal conditions (Figure 6a). As the simulation progresses, a distinct thermal gradient develops by t = 0.05 s (Figure 6b), with higher temperatures concentrated near the nozzle walls and inlet, and cooler regions downstream. This behavior suggests adiabatic expansion and energy dissipation of the working gas along the nozzle. The localized heating zones are indicative of boundary layer development, which influences both gas velocity and particle thermal energy transfer.

3.3.3. Particle Trajectory Analysis

Figure 7 illustrates the evolution of particle trajectories over a simulation interval of 0 to 0.05 s, with data captured at a high temporal resolution (step size 0.001 s). The simulation reflects the behavior of particles under the influence of a supersonic gas jet during cold-spray deposition.

At Time = 0 s, particles are densely clustered near the nozzle inlet, demonstrating their injection into the flow with initial uniform spatial distribution (Figure 7a). By Time = 0.05 s, the particles have progressed significantly along the axis of the nozzle, forming a focused and high-velocity stream. The convergence toward the centerline indicates effective particle acceleration and aerodynamic focusing due to the nozzle geometry and gas flow dynamics.

This trajectory evolution is critical, as it directly correlates with particle impact velocity and deposition efficiency. The highest density of particles reaching the substrate is observed at the stagnation region near the nozzle exit, supporting efficient coating formation under the simulated pressure and temperature conditions.

The CFD and thermal simulations confirmed that increasing gas pressure from 4 × 10⁵ Pa to 6 × 10⁵ Pa (at gas temperature 800 K) led to a substantial increase in particle exit velocity, exceeding 650 m/s at the highest pressure. This indicates enhanced deposition potential under elevated driving forces. The simulations also identified ~15 mm as a favorable stand-off distance, where particles maintain high velocity and focus at the substrate surface—promoting efficient deposition and uniform coating build-up. Thermal modeling showed that gas temperature at the substrate remained around 500 K after expansion, which is safely below the melting point of the materials, thereby avoiding thermal degradation.

Comparison of Particle Velocity—Model vs. Experiment: In-flight particle velocities predicted by the model were compared against experimental indicators to assess the model’s validity. The CFD simulation indicated that under optimal spraying conditions (gas pressure ~0.6 MPa, gas temperature ~600 °C, 15 mm stand-off), particles would reach impact speeds on the order of 500–600 m/s. This predicted velocity range is above the typical critical velocity (~500 m/s) required for aluminum particles to adhere to a steel substrate. Consistent with the prediction, coatings were successfully deposited at 0.6 MPa and 600 °C with strong bonding, implying that the particles indeed achieved velocities high enough to exceed the bonding threshold. By contrast, at a lower stagnation pressure (0.4 MPa) or gas temperature (400 °C), the model predicted significantly reduced particle speeds, and experiments under those conditions resulted in either very poor adhesion or no coating deposition—indicating the particles did not reach critical velocity in those cases. Although a direct measurement of particle flight velocity was not available in our setup, these qualitative outcomes (the presence or absence of effective coating) serve as indirect validation of the simulation results. Furthermore, the model correctly captured the influence of stand-off distance on particle velocity and deposition: it predicted an optimal stand-off window of roughly 10–20 mm (with ~15 mm being ideal) and showed that extreme stand-off distances would be detrimental (particles experiencing excessive drag at long distances or incomplete acceleration at very short distances). This behavior was borne out experimentally—e.g., at 5 mm stand-off, the high-speed gas jet caused particle splashing and substrate abrasion instead of efficient deposition, while at 25 mm stand-off the particles decelerated and produced negligible coating build-up. Such close agreement between the simulated particle velocities and the experimental outcomes strongly supports the accuracy of the COMSOL model. In particular, the model’s ability to predict the conditions for successful versus unsuccessful deposition (through velocity thresholds and trends) reinforces its credibility as a reliable predictive tool for cold-spray process optimization and parameter selection.

Additionally, the modeled behavior of the multiphase powder feedstock—assuming uniform Al, Zn, and TiO₂ dispersion—was supported by experimental SEM, EDS, and XRD observations showing phase retention and homogeneous microstructures. While direct adhesion measurements were not performed, the consistency between model predictions and observed coating morphology suggests that the selected parameters are promising for achieving mechanically and chemically stable coatings. Overall, the synergy between CFD analysis, thermodynamic modeling, and microstructural characterization provides a reliable approach for guiding cold-spray process design and can be extended to other composite systems for application-specific optimizations.

3.4. Microstructure and Composition

Scanning electron microscopy (SEM) confirmed that the chemically agglomerated Al–Zn–TiO₂ feedstock powder (Figure 8a) had a predominantly spherical morphology, with numerous fine Zn and TiO₂ “satellite” particles uniformly attached to the surfaces of larger Al core particles. This engineered structure promotes stable powder flow during spraying and facilitates a more uniform compositional distribution in the final coating.

In contrast, the mechanically mixed composite powder (Figure 8b)—produced without any binder—showed irregular particle shapes, visible TiO₂ agglomerates, and many loose fine particles (Zn and TiO₂) not attached to the Al surface. These characteristics indicate less controlled dispersion and potentially reduced coating uniformity for the mechanically mixed feedstock.

Figure 8 provides a direct comparison between the chemically agglomerated and mechanically mixed feedstock powders. It clearly demonstrates that the agglomerated powder exhibits superior microstructural organization, with homogeneously distributed TiO₂ nanoparticles and minimal clustering. In contrast, the mechanically mixed powder reveals heterogeneous distribution patterns, larger TiO₂ agglomerates, and sporadic detachment of ceramic particles from the Al matrix. This morphological distinction underscores the advantage of chemical agglomeration in achieving uniform phase dispersion at the microscale.

Figure 9 compares cross-sectional SEM images of cold-sprayed coatings obtained from (a) chemically agglomerated and (b) mechanically mixed powders. The coating in Figure 9a (from the chemically agglomerated feedstock) exhibits a dense, continuous structure with tightly packed splats and excellent bonding to the steel substrate. The interface is smooth and well-bonded, with no signs of delamination, indicating efficient particle deformation and interfacial consolidation during impact.

In contrast, Figure 9b shows the coating produced using mechanically mixed powders, where defects such as interfacial gaps, weakly bonded regions, and non-uniform splat stacking are evident. These features reflect inadequate adhesion and inefficient particle anchoring during impact, likely due to the irregular feedstock morphology and poor dispersion of phases.

These microstructural differences confirm the advantage of chemical agglomeration in achieving structurally coherent coatings with strong substrate adhesion—essential for long-term durability in corrosive and mechanical environments.

To further assess elemental distribution within the coatings, Figure 10 presents SEM micrographs with corresponding EDS elemental maps for Al, Zn, and Ti. The results confirm the uniform dispersion of Zn and Ti throughout the Al matrix in the coating derived from agglomerated powder, supporting the superior compositional homogeneity of this sample. In contrast, the mechanically mixed variant shows localized accumulations of TiO₂ and Zn, indicating a less uniform distribution. The uniform splat interfaces and absence of inter-splat voids in the agglomerated sample are indicative of more effective particle deformation and bonding.

Figure 11a shows the surface morphology of the chemically agglomerated Al–Zn–TiO₂ feedstock powder, where fine TiO₂ nanoparticles are uniformly distributed as satellite structures around the larger Al particles. This configuration promotes stable powder flow and improved compositional uniformity during deposition.

Figure 11b presents a representative SEM cross-section of the resulting coating, revealing a well-adhered, crack-free structure with dense splat stacking and a consistent thickness ranging from ~20 µm to 32 µm. No signs of interfacial delamination are observed, confirming the strong bonding between the coating and the steel substrate. These microstructural features are characteristic of successful cold-spray deposition using chemically engineered powders.

On the whole, the SEM and EDS analyses (Figure 8, Figure 9, Figure 10 and Figure 11) demonstrate that using the chemically agglomerated Al–Zn–TiO₂ feedstock, in combination with optimized CGDS parameters, yields a high-performance composite coating characterized by structural continuity, phase homogeneity, and multifunctional protection. The refined dispersion of TiO₂ and Zn phases within the Al matrix—clearly visible in the micrographs and elemental maps—is a key advantage over coatings prepared from the mechanically mixed powder.

To highlight the influence of powder preparation methods,

Table 3. Comparative microstructural characteristics of coatings from different powder preparation methods.

Powder Preparation Method	Estimated Porosity (%)	Surface Morphology	Interfacial Integrity	Elemental Distribution (EDS)
Chemically agglomerated	<0.5	Uniform, dense	Continuous, no delamination	Homogeneous
Mechanically mixed	~1.2–1.4	Rough, porous	Localized cracks/delamination	Heterogeneous

presents a qualitative comparison of key microstructural characteristics for coatings fabricated using chemically agglomerated versus mechanically mixed Al–Zn–TiO₂ powders. These observations are based on visual analysis of SEM cross-sections. Coatings from the chemically agglomerated powder exhibit visibly lower porosity, a more uniform microstructure, and improved interfacial integrity. EDS mapping confirms a more homogeneous elemental distribution. In contrast, coatings from the mechanically mixed powder show higher porosity, localized interfacial heterogeneity (including occasional micro-cracks or delamination), and uneven TiO₂ dispersion.

To complement these qualitative observations,

Table 4. Comparison of porosity characteristics between chemically agglomerated and mechanically mixed coatings.

Coating Type	Average Pore Size (µm)	Porosity (%)
Chemically Agglomerated	2.8	0.95
Mechanically Mixed	5.6	2.15

provides quantitative porosity measurements obtained by image analysis (ImageJ, version 1.54m, National Institutes of Health, Bethesda, MD, USA). This dual approach offers both qualitative and numerical comparisons of microstructure, reinforcing the benefits of using agglomerated powders as CGDS feedstock.

Porosity Study

To complement the microstructural characterization, porosity analysis was performed to assess the internal integrity of the cold-sprayed coatings derived from different feedstock preparation methods. Figure 9 presents cross-sectional SEM images of coatings obtained from chemically agglomerated (Figure 9a) and mechanically mixed (Figure 9b) Al–Zn–TiO₂ powders.

The coating produced from chemically agglomerated feedstock exhibits a dense and uniform microstructure with closely packed splats and no signs of delamination at the interface. In contrast, the coating deposited using mechanically mixed powder reveals scattered pore clusters and visible interfacial separation, indicating weaker adhesion and less effective particle consolidation.

Quantitative image analysis confirmed these observations, as summarized in Table 4. The chemically agglomerated coating demonstrated a significantly lower total porosity (0.95%) and smaller average pore size (2.8 µm), whereas the mechanically mixed coating showed a higher porosity of 2.15% and a larger average pore size of 5.6 µm. These metrics directly correlate with the observed differences in splat-packing density and structural coherence.

The superior performance of the chemically agglomerated powder can be attributed to its improved particle morphology and flow behavior, which promote more consistent particle impact, plastic deformation, and interlocking during cold-spray deposition. In contrast, mechanically mixed powders, due to irregular shapes and insufficient cohesion, lead to variable impact conditions and non-uniform bonding, resulting in porosity and interfacial defects.

These findings highlight the importance of feedstock engineering in cold-spray applications and demonstrate that chemical agglomeration is a more effective approach for producing dense, well-bonded coatings suitable for demanding corrosion- and wear-resistant applications.

3.5. Coating Performance: Corrosion Behavior

The potentiodynamic polarization curves in 3.5% NaCl solution (Figure 12) clearly show that the coated samples exhibited a significantly lower corrosion current density (I_corr) than the bare steel. Specifically, the corrosion current decreased by approximately an order of magnitude, indicating a substantial reduction in the corrosion rate due to the presence of the coating. Additionally, the corrosion potential (E_corr) of the coated samples shifted slightly towards more negative values compared to the bare steel. This behavior is typical for coatings containing sacrificial Zn: the Zn phase corrodes preferentially, maintaining a cathodic protection effect on any exposed steel surface.

The shape of the polarization curves suggests a two-stage protective mechanism. At lower overpotentials, the Zn-rich regions act as sacrificial anodes, dissolving first and thereby protecting the steel beneath. At higher potentials, the passive aluminum oxide (Al₂O₃) layer formed on the aluminum matrix limits further anodic reaction, stabilizing the coating’s overall performance. Furthermore, the presence of finely dispersed TiO₂ particles contributes to the barrier properties of the coating by blocking ionic diffusion pathways and supporting the formation of stable surface oxides [35].

After extended immersion, a thin layer of white corrosion products—primarily Zn hydroxides or carbonates and minor Al oxides—was observed on the coated samples. These corrosion products can help to seal micro-defects and pores in the coating, enhancing its protective capability over time. In contrast, the bare steel developed uniform red-brown rust (Fe₂O₃ and Fe₃O₄) with visible pitting and continuous degradation.

Figure 12 shows the potentiodynamic polarization (Tafel) curves of the cold-sprayed Al–Zn–TiO₂ coatings under different spraying conditions. The untreated steel (“Initial”) demonstrates the highest corrosion current density (I_corr = 5.8 μA/cm²) and the most negative corrosion potential (E_corr = −820 mV), indicating low corrosion resistance. In contrast, coatings deposited under optimized parameters (S235JR3 and S235JR4) exhibit significantly reduced I_corr values (1.8 and 1.2 μA/cm², respectively) and more noble E_corr values (−715 and −690 mV), pointing to enhanced corrosion protection.

This trend correlates well with SEM/EDS observations: better phase homogeneity and denser surface structure likely reduce the number of galvanic micro-couples and active sites for electrochemical reactions.

To quantify the improvement, estimated corrosion rates were calculated based on the Tafel slope and Faraday’s law. As shown in

Table 5. Electrochemical parameters derived from Tafel plots.

Sample	Ecorr (mV)	Icorr (μA/cm2)	Est. Corrosion Rate (mm/Year)
Initial	−820	5.8	0.070
S235JR1	−780	4.2	0.051
S235JR2	−750	3.1	0.038
S235JR3	−715	1.8	0.021
S235JR4	−690	1.2	0.014

, the rate decreases from 0.070 mm/year in bare steel to just 0.014 mm/year in the best-performing coated sample (S235JR4).

Notably, TiO₂ plays a dual role in enhancing corrosion resistance. Firstly, it serves as a passivating component by forming a chemically stable oxide film that inhibits oxygen and chloride penetration, thereby reducing electrochemical reactivity at the substrate interface [36]. Secondly, as an inert barrier filler, TiO₂ increases coating density, decreases porosity, and creates a tortuous diffusion pathway, effectively obstructing the transport of corrosive species such as water, O₂, and Cl⁻ [37,38]. In neutral or mildly alkaline environments, TiO₂ remains chemically inert and supports electrochemical stability. Collectively, these effects—passivation, pore blockage, and electrochemical buffering—synergistically enhance the long-term corrosion resistance of the composite coating.

These results demonstrate that tuning the cold-spray process can significantly enhance corrosion resistance by altering microstructural features that affect electrochemical activity.

Taken together, the combined results demonstrate that the cold-sprayed Al–Zn–TiO₂ coatings deliver robust protection by leveraging the complementary benefits of a dense Al matrix, sacrificial Zn, and stable TiO₂ reinforcement. The polarization behavior confirms that these coatings are highly effective in mitigating corrosion under saline conditions.

3.6. Phase Composition

The X-ray diffraction (XRD) analysis (Figure 13) provides clear evidence of the phase composition and stability of the cold-sprayed Al–Zn–TiO₂ coatings under various process conditions. For the bare steel substrate (labeled Initial in Figure 13), strong diffraction peaks corresponding to the body-centered cubic α-Fe phase are dominant, as expected for mild carbon steel.

After deposition, the coated samples (labeled S235JR1 to S235JR4 in Figure 13) display distinct peaks that match the face-centered cubic (FCC) structure of aluminum at approximately 2θ ≈ 38° (Al (111)), 45° (Al (200)), and 65° (Al (220)). Peaks indicative of hexagonal close-packed (HCP) zinc are also clearly visible, appearing at around 2θ ≈ 36° and 56°, consistent with Zn (002) and Zn (100) reflections. Additionally, the coatings show weak but detectable peaks near 2θ ≈ 25° and 48°, which correspond to the (101) and (200) planes of anatase TiO₂. These peaks confirm that the ceramic reinforcement phase remains intact and well dispersed within the metallic matrix.

Importantly, no new intermetallic compounds (e.g., Al–Fe or Zn–Fe phases) or reaction products were detected within the detection limit of the XRD, supporting that the cold gas dynamic spray process operates entirely in the solid state without local melting or alloying at the interface. This observation aligns with the designed goal of preserving feedstock purity and minimizing unwanted reactions.

A comparison across samples S235JR1 to S235JR4 demonstrates how process parameter variations affect the coating structure. Specifically, coatings deposited under more optimized conditions (S235JR3 and S235JR4) exhibit higher peak intensities for Al and Zn, along with a further reduction in the relative intensity of α-Fe peaks from the substrate. This trend indicates increased coating thickness and more complete substrate coverage under higher gas pressures and temperatures, which enhance particle impact velocity and deposition efficiency.

Furthermore, after exposure to a saline environment (e.g., salt spray or immersion), the XRD patterns (Figure 13) remained consistent with the as-sprayed state, with only slight peak broadening and minor intensity changes attributable to the development of thin surface oxides. Critically, no new crystalline phases were observed, confirming that the composite coating maintains its chemical integrity and phase stability even under corrosive conditions.

Overall, the XRD data presented in Figure 13 validate that the Al–Zn–TiO₂ cold-sprayed coatings successfully retain their designed multiphase structure, ensuring that the sacrificial protection (from Zn), passive barrier effect (from Al), and reinforcement (from TiO₂) are all maintained throughout processing and subsequent environmental exposure.

3.7. Surface Roughness and Feed Rate

Surface profilometry measurements demonstrated that the surface roughness of the cold-sprayed Al–Zn–TiO₂ coatings was influenced by the powder feed rate, as summarized in

Table 6. Surface roughness parameters (Ra, Rz, Rq) for Al–Zn–TiO₂ coatings deposited at varying powder feed rates.

Feed Rate (g/s)	Ra (µm)	Rz (µm)	Rq (µm)
0.4	~2.47	~13.23	~2.84
0.5	2.16	11.41	2.70
0.6	1.84	10.28	2.32

. Specifically, coatings deposited at the lowest feed rate of 0.4 g/s exhibited an average arithmetic mean roughness (Ra) of approximately 2.47 µm, with a corresponding peak-to-valley height (Rz) of about 13.23 µm and a root mean square roughness (Rq) near 2.84 µm.

As the powder feed rate increased to 0.5 g/s and 0.6 g/s, the mean Ra decreased slightly to 2.16 µm and 1.84 µm, respectively, with corresponding reductions in Rz and Rq values. This slight decline in roughness with increasing feed rate can be attributed to the enhanced filling of surface valleys and voids due to a higher particle flux, which promotes more uniform layer build-up and peening action.

In general, the results indicate that the cold-spray process inherently produces coatings with relatively low roughness compared to other thermal spray methods, due to the absence of molten droplets and splat formation. Moreover, the roughness values remain within a narrow range, confirming that the layer-by-layer deposition maintains good surface uniformity across different feed rates.

Profilometry scans also revealed that thickness variation across the spray track was minimal in the central zone, with minor edge tapering, which is typical for cold spray due to the Gaussian velocity distribution of the particle jet. This consistent thickness profile supports the good repeatability of the optimized process parameters.

These values demonstrate that selecting a moderate feed rate (around 0.5 g/s) balances deposition efficiency and surface finish, producing coatings with both sufficient thickness and controlled roughness suitable for corrosion and wear applications.

3.8. Tribological Properties

As illustrated in Figure 14, the average coefficient of friction (COF) for the coated samples decreased notably to approximately 0.4–0.5, while the uncoated steel surface exhibited a higher COF in the range of 0.7–0.8. This reduction indicates smoother sliding and less frictional resistance during contact, which is attributed to the combined effects of the Zn phase acting as a mild solid lubricant and the TiO₂ particles providing localized hardness reinforcement.

Wear volume loss measurements showed that the coated surfaces experienced up to three times less material removal compared to the uncoated counterpart under identical test conditions. Optical microscopy of the wear tracks further confirmed there were narrower and shallower grooves on the coated samples, signifying less severe wear damage and improved resistance to micro-plowing and scratching. This aligns with the role of the hard TiO₂ particles, which impede groove propagation, and the ductile Al–Zn matrix, which absorbs part of the contact stresses.

Figure 14 summarizes the effect of powder feed rate on both surface roughness and tribological behavior. At higher feed rates, the coatings exhibited slightly increased surface roughness (as described previously), yet the wear resistance remained consistently high. This demonstrates that the composite coating maintains its protective function even with minor topographical changes.

When benchmarked against values reported in the literature, these results suggest that the cold-sprayed Al–Zn–TiO₂ coating offers superior tribological performance compared to typical pure Al or Zn coatings produced by cold spray. Conventional thermal-sprayed Zn–Al coatings often contain higher porosity levels (1–5%) and oxidation, which can compromise wear resistance and lead to early coating breakdown. In contrast, the cold-sprayed composite here achieves extremely low porosity (~0.5%) and a denser microstructure, which collectively enhance both wear and corrosion protection.

Additionally, the inclusion of TiO₂ may offer secondary benefits such as localized barrier effects within the coating’s microstructure and potentially improved passivation. Although photocatalytic effects were not directly evaluated in this study, TiO₂ is known for generating protective oxide films under UV exposure, which could be advantageous for outdoor or marine applications.

To elucidate the wear resistance mechanisms introduced by TiO₂, an SEM analysis of the wear track morphology was performed. Figure 15 shows the worn surface of the optimized Al–Zn–TiO₂ coating after dry-sliding wear-testing. The image reveals a relatively smooth wear path with minimal abrasive grooves and no significant cracking or delamination, indicating good resistance to both adhesive and abrasive wear.

Closer examination highlights several microstructural features that support the role of TiO₂ in enhancing tribological performance:

Load-bearing reinforcement: TiO₂ particles, being hard and chemically stable, act as load-bearing constituents within the softer aluminum matrix. Their uniform dispersion (confirmed via EDS mapping, Figure 8) improves resistance to plastic deformation under sliding contact by increasing the effective hardness of the composite.

Crack inhibition: The fine distribution of nano-TiO₂ helps to arrest or deflect microcrack propagation along splat boundaries, thereby increasing the resistance to fatigue-related wear damage. This is evidenced by the absence of microcracks or pull-out pits along the wear path in the SEM images.

Third-body effect: In some regions, embedded TiO₂ particles may act as micro-scale “rolling elements,” reducing friction by preventing direct metal-to-metal contact and thus contributing to the observed lower coefficient of friction (CoF ≈ 0.36).

The marked reduction in CoF for the Al–Zn–TiO₂ composite can be explained by the tribological role of TiO₂ as a solid lubricant and surface conditioner. Titanium dioxide is known to act as a solid lubricant that decreases friction and wear by forming a protective third-body layer at the sliding interface [39]. During dry sliding, TiO₂ particles can be entrained into the contact and serve as a lubricious third body—essentially a micro-scale layer of rolling/sliding particles that self-replenishes—preventing direct metal-to-metal contact and lowering shear stresses [39]. Indeed, tribological studies on oxide-containing composites have observed TiO₂-rich mechanically mixed tribofilms on worn surfaces that function as solid lubricants, correlating with reduced friction coefficients [40]. In addition, the hard TiO₂ phase contributes a polishing effect: the TiO₂-reinforced coating tends to smooth out the wear track, as evidenced by the relatively shallow grooves and lack of severe scratching in the SEM micrographs (Figure 14). This micro-polishing results in a smoother contact surface and a milder wear regime, which also diminishes friction. Literature confirms that incorporating hard oxide particles, such as TiO₂, into coatings can increase surface hardness and simultaneously reduce surface roughness—thereby lowering the friction coefficient [41]. Some TiO₂ particles may also become embedded in the softer Al–Zn matrix or the counterface, where they act as load-bearing asperities that support the contact and facilitate formation of a stable, low-shear, mechanically mixed layer. These combined mechanisms explain the observed drop in CoF (to ~0.36–0.40) for the TiO₂-containing coating. Notably, the SEM images in Figure 15 reveal embedded TiO₂ particles within the wear track and minor delamination or spallation zones at the edges, features that support the interpretation of a third-body lubricating action and surface film formation. For greater clarity, more explicit labeling of these features in the figure (e.g., arrows indicating the embedded oxides and delamination sites) would enhance interpretation of how TiO₂ contributes to friction reduction.

Overall, the results confirm that careful powder feed rate control, combined with the composite’s tailored phase composition, produces a cold-sprayed coating that effectively reduces friction, minimizes wear, and maintains robust surface integrity under dry sliding conditions.

3.9. Performance Benchmarking with Literature

To contextualize the present results,

Table 7. Representative properties of Al–Zn–TiO₂ cold-sprayed coating benchmarked against aluminum-based coatings from the literature. Values are typical ranges reported for as-sprayed or mildly post-processed states.

Coating	Process	Porosity (%)	CoF (Dry)	Wear Rate (mm3·N−1·m−1)	Corrosion Metric (e.g., Icorr in 3.5% NaCl)	References
Al–Zn–TiO2 (this work)	Cold spray	~0.5	0.4–0.5	≤(1/3) × uncoated	Icorr by ~1 order vs. steel	This work
Pure Al	Cold spray	0.3–5	0.7–0.9	8–9 × 10−3 (typ.)	Icorr~10−3 A·cm−2 (typ.)	[42,43,44]
Pure Al	Thermal spray	10–20	—	poor (soft)	needs sealing for durability	[45,46,47]
Zn–Al (e.g., 85/15)	Thermal spray	5–10	0.5–0.7	high (soft)	excellent marine protection	[48,49]
Al + Al2O3 (15–30%)	Cold spray	1–3	0.4–0.6	lower than pure Al	Icorr~10−5 A·cm−2 (typ.)	[50,51]

contrasts the Al–Zn–TiO₂ cold-sprayed coating with representative aluminum-based coatings reported in the literature, including pure Al (cold/thermal spray), Zn–Al alloys (thermal spray), and Al–matrix composites reinforced with Al₂O₃. The comparison covers porosity, microhardness, dry friction coefficient (CoF), wear rate, and corrosion metrics.

Overall, our Al–Zn–TiO₂ coating combines very low porosity (~0.5%), moderate CoF (≈0.4–0.5), and reduced wear volume (up to ~3× lower than uncoated steel under identical conditions) while maintaining excellent corrosion resistance due to the dual action of a dense Al barrier and sacrificial Zn. Compared with typical pure-Al and Zn–Al coatings, the composite shows a better balance of tribological and anti-corrosion performance, and it performs comparably to or better than Al-matrix composites reinforced with Al₂O₃.

3.10. Deposition Mechanism, Process Sensitivity, and Performance Benchmarking

3.10.1. Role of Powder Preparation

The marked difference in coating quality between chemically agglomerated and mechanically mixed feedstocks underscores the critical role of powder preparation in cold-spray technology. Chemically agglomerated powders demonstrated improved flowability and uniform particle morphology, which facilitated consistent particle impact and bonding. In contrast, mechanically mixed powders showed heterogeneous morphology and poor adhesion, as confirmed by SEM images (Figure 9), resulting in irregular splat formation and increased porosity.

Although chemical agglomeration can provide compositional uniformity, our results indicate that it requires careful control to avoid binder-related interfacial weaknesses. In this study, mechanical blending—despite being simpler—resulted in coatings with inferior lamellar structure and poor adhesion, emphasizing the need for robust feedstock engineering in composite systems.

The selected particle size ranges (~20–45 µm for Al and Zn, and nano–micro for TiO₂) proved effective for cold gas dynamic spraying. Larger TiO₂ particles (~10 µm) are typically difficult to deposit alone but can be effectively incorporated when attached to ductile Al particles. This strategy aligns with the literature noting that the successful deposition of ceramics via cold spray is primarily feasible in nanostructured or agglomerated forms.

3.10.2. Influence of Composition on Critical Velocity

Incorporating Zn into the Al matrix may reduce the overall critical velocity required for particle bonding. Due to its lower melting point and higher deformability, Zn may locally soften under impact and contribute to inter-splat cohesion. Although the composition was not systematically varied, the high adhesion observed across samples suggests that the Al–Zn–TiO₂ mixture operated above the critical velocity threshold. Since Zn-rich areas act as local energy-absorbers or “glue zones,” their presence likely aided coating buildup and stability. These results are consistent with published critical velocity values for Zn and Sn alloys (<300 m/s), supporting the hypothesis of a lowered threshold.

3.10.3. Process Parameter Sensitivity

The coating quality was found to be sensitive to variations in gas pressure and temperature. Optimal coatings—exhibiting the most uniform microstructure, reduced porosity, and strong adhesion—were achieved at 0.6 MPa and 600 °C. Deviation from these values resulted in increased surface roughness, interfacial heterogeneity, and less dense splat packing. However, even under sub-optimal conditions, no delamination was observed, indicating reasonable robustness in the parameter window. These findings validate that cold-spray deposition requires tight control over thermomechanical conditions to ensure coating integrity.

3.10.4. Comparison with Conventional Al and Zn Coatings

Compared to conventional cold-sprayed pure Al or Zn coatings, the Al–Zn–TiO₂ composite showed superior performance in multiple aspects. Literature reports often cite high porosity and poor mechanical integrity in Zn coatings, and limited wear resistance in Al coatings due to their ductility and softness [10,16]. In contrast, our composite coatings exhibited lower porosity (~0.95%), reduced coefficient of friction (0.4–0.5), and higher structural homogeneity. The combined effect of Al’s barrier function, Zn’s sacrificial anodic protection, and TiO₂’s hard reinforcement yielded coatings that outperformed single-metal systems in both corrosion and wear resistance.

Moreover, surface roughness analysis showed Ra values of 1.8–2.4 µm for the composite, compared to >4 µm typically reported for pure Al coatings. This indicates more uniform deposition and better splat deformation in the composite system.

3.10.5. Summary and Implications

This study demonstrates that an Al–Zn–TiO₂ nanocomposite coating can be effectively produced by cold gas dynamic spraying using optimized feedstock and process parameters. The integrated use of experimental validation and prior simulation (not detailed here) enabled efficient parameter selection, avoiding exhaustive trial-and-error methods.

The resulting coatings showed stable phase composition, excellent corrosion protection, uniform microstructure, and favorable tribological behavior. These findings highlight the feasibility of using cold spray to fabricate multifunctional composite coatings for advanced surface engineering applications in marine, automotive, and aerospace sectors.

4. Conclusions

Agglomerated Al–Zn–TiO₂ powders were successfully deposited on steel substrates using cold gas dynamic spraying (CGDS), resulting in dense, well-adhered composite coatings with enhanced functional properties. A combined experimental and modeling strategy was employed to optimize feedstock formulation and deposition parameters. The key findings are summarized below:

• Powder Preparation and Feedstock Engineering:
  Chemically agglomerated Al–Zn–TiO₂ powders provided superior coating quality compared to mechanically blended variants. The optimized agglomeration process improved powder flowability, ensured uniform TiO₂ distribution, and minimized porosity and interfacial defects. Mechanically mixed feedstocks, in contrast, showed poor inter-particle cohesion and reduced adhesion to the substrate.
• Microstructure and Phase Composition:
  The cold-sprayed coatings exhibited a fully dense, lamellar structure with tightly packed, plastically deformed splats. XRD analysis confirmed the presence of Al, Zn, and anatase TiO₂ phases without the formation of undesirable intermetallic compounds, supporting the occurrence of solid-state metallurgical bonding during deposition.
• Corrosion Resistance:
  Coatings derived from agglomerated powders showed excellent corrosion resistance, reducing corrosion current density by an order of magnitude and providing stable passivation in 3.5% NaCl solution. The combined barrier effect of aluminum and sacrificial protection from zinc contributed to a positive shift in corrosion potential and long-term durability.
• Tribological Performance:
  The TiO₂ reinforcement significantly improved wear resistance by lowering the friction coefficient (to ~0.4–0.5) and reducing wear volume by approximately 50% compared to uncoated steel. The nanocomposite structure effectively prevented crack propagation and surface damage, indicating its suitability for applications under sliding contact.
• Process Optimization and Simulation Validation:
  Empirical optimization and numerical modeling enabled the identification of optimal CGDS parameters: 0.6 MPa gas pressure, 600 °C gas temperature, 15 mm stand-off distance, and 90° spray angle. These conditions yielded particle velocities above the critical threshold for deposition, ensuring uniform coating build-up even at substrate edges. Experimental results validated simulation outcomes, confirming the reliability of this integrated approach.
• Benchmarking against the representative literature data (Table 6) shows that the Al–Zn–TiO₂ cold-sprayed coating achieves superior overall performance relative to pure Al, Zn–Al, and Al–Al₂O₃ systems due to its dense microstructure and the synergistic action of TiO₂ reinforcement and Zn-assisted corrosion protection.

The developed Al–Zn–TiO₂ coatings possess significant industrial relevance. Their combination of sacrificial anodic protection (Zn), passive barrier protection (Al), and hard reinforcement (TiO₂) makes them highly suitable for marine, oil and gas, and energy sectors. Potential applications include the corrosion- and wear-resistant protection of ship structures, offshore platforms, pipelines, pump housings, and power equipment, where simultaneous resistance to saline corrosion and mechanical wear is critical.

In summary, this study demonstrates that a strategically engineered Al–Zn–TiO₂ nanocomposite coating can be effectively produced using an optimized feedstock and process parameter design. The coatings achieved in this work exhibit a stable phase composition, excellent corrosion and wear protection, a uniform microstructure, and strong substrate bonding. These results highlight the feasibility of using cold spray to fabricate multifunctional composite coatings for advanced surface engineering applications in marine, automotive, and aerospace environments.