A Review of Portable Low-Pressure and Medium-Pressure Cold Spray Repairs: Process Window, Deposition Behaviour, and Deposit Integrity

Abstract

Low-pressure cold spray (LPCS) and medium-pressure cold spray (MPCS) provide field-deployable alternatives to stationary high-pressure cold spray (HPCS) for lightweight metallic component repair, although lower particle velocities and narrower process windows restrict the range of compatible repair materials. This review evaluates portable LPCS and MPCS repair as a distinct class of cold spray application using HPCS as the reference baseline. The article synthesises evidence on process windows, critical velocity, particle velocity ratio, deposition efficiency, deposit build-up capability, substrate preparation, porosity, oxide disruption, ceramic-assisted densification, interfacial bonding mechanisms, adhesion strength, hardness, residual stress, tensile behaviour, and fatigue performance. Attention is given to repair geometry, including blend-out depth, taper angle, aspect ratio, spray angle, and stand-off distance, because these factors govern particle impact conditions and interfacial integrity across practical repair geometries. Available evidence indicates that LPCS can achieve dimensional restoration, surface protection, and, in some cases, fatigue performance comparable with that of the substrate. However, structural qualification remains constrained by limited portable system fatigue data, few direct LPCS–MPCS comparisons under equivalent conditions, and inconsistent reporting of powder characteristics, surface preparation, porosity, and repair geometry. MPCS may extend the range of compatible repair materials while maintaining field deployability, but its peer-reviewed evidence base remains limited. Future work should prioritise controlled LPCS–MPCS comparisons, geometry-sensitive fatigue testing, standardised deposit quality metrics, and qualification procedures linking process window control with repair integrity.

1. Introduction

1.1. Background

Metallic components in aerospace, naval, energy, and other critical engineering sectors undergo complex service degradation involving interacting mechanisms such as fatigue, corrosion, and wear. These degradation processes can result in crack development, surface degradation, and material loss, limiting component reliability and service life. Rising acquisition costs, extended service intervals, and an ageing equipment population have increased the need to restore such components in service instead of replacing them. Deposit integrity is used here to denote the adhesion, cohesion, dense build-up, and residual stress state a deposited layer requires to restore load-carrying capacity under static and fatigue loading, a more demanding criterion than that commonly applied in the thermal spray literature, where deposit performance often denotes corrosion protection or dimensional restoration without structural load transfer [1,2].

Fusion-based repair routes such as weld overlay, laser cladding, and arc-based directed energy deposition can restore damaged material; however, their thermal input may alter the original material properties by introducing a heat-affected zone (HAZ), oxidation, phase transformations, precipitate dissolution, and tensile residual stresses, potentially reducing performance relative to the original microstructure and heat treatment condition. In the precipitation-hardened 7xxx-series aluminium alloys used in aerospace primary structure, hot cracking susceptibility and HAZ strength reductions of 30–50% are significant structural concerns [3,4]. Bonded composite patches are well-established for thin panel repairs but exhibit time-limited adhesive durability, moisture sensitivity, and geometric incompatibility with curvature surfaces [2]. Mechanical fastening via bolted doublers or rivet repairs introduces stress concentrations at fasteners, potentially seeding the fatigue damage the repair was intended to address [5].

Cold spray, invented at the Institute of Theoretical and Applied Mechanics in Novosibirsk in the 1980s and now deployed commercially in aircraft maintenance and naval ship repair [6], has emerged as an alternative. Depositing metal particles in the solid state without bulk melting preserves substrate temper, avoids HAZ formation, and generates residual stresses that are compressive or of manageable magnitude [7]. It has been used to restore corroded aluminium and magnesium airframe features to dimensional tolerance [8], and defence maintenance organisations have reported operational capability for aircraft outer-panel and rotorcraft component repairs [9,10]. Despite this, research has developed along two disconnected tracks: laboratory characterisation under idealised conditions using mostly high-pressure equipment and a smaller set of operationally focused field case studies, leaving engineers without a coherent framework for portable system selection.

Cold spray spans a range of pressure–temperature–velocity regimes with markedly different deposition behaviour. High-pressure cold spray (HPCS), operating up to 7.5 MPa with nitrogen or helium, offers the widest material compatibility and highest deposition efficiency but demands infrastructure that renders the equipment effectively stationary [11,12]. Low-pressure cold spray (LPCS), operating below 1 MPa with compressed air or nitrogen and downstream powder injection (Figure 1), is generally portable, weighing less than 30 kg [13,14], but its constrained velocity envelope limits deposition to ductile metals and composite feedstocks [15]. Medium-pressure cold spray (MPCS), an emerging intermediate class operating up to 3 MPa, likewise uses downstream injection and remains field deployable [16]. The two portable regimes differ by more than nominal pressure. The higher stagnation pressure and gas temperature of MPCS raise particle velocity (up to 800 m/s against 300–600 m/s of LPCS) and hence the velocity ratio (Section 3), which governs interparticle bonding and the accessible material set. LPCS is therefore limited to ductile, low-density metals and metal–ceramic composites, whereas MPCS reaches higher-strength Ti- and Ni-based alloys at the cost of heavier equipment and a nitrogen supply. Data for portable MPCS remain sparse.

Figure 1. Schematic illustration of (a) high-pressure cold spray (HPCS) with upstream powder injection and (b) low-pressure cold spray (LPCS) system configurations. Medium-pressure cold spray (MPCS) shares the downstream configuration shown in (b), operating at higher stagnation pressure and gas temperature.

Deployment of portable cold spray requires four elements: (i) a process window framework accounting for the field constraints (variable spray angle, ambient temperature, substrate condition) that compress the operating margin above critical velocity; (ii) a microstructural understanding of porosity, oxide distribution, and interface bond continuity; (iii) a geometry-sensitive mechanical performance dataset generated on representative substrate flaws under portable deposition protocols, rather than the simplified coupons and HPCS conditions that dominate the literature; and (iv) a qualification framework linking process requirements to deposit acceptance criteria. Work towards the fourth requirement has begun. Stamoulis et al. [17] mapped the airworthiness requirements of the European Union Aviation Safety Agency (EASA), specifically Certification Specification CS-25 for large aeroplanes and Part 21, onto a defined set of established test methods and applied the resulting framework to the repair of a corroded AA6013-T6 wire bundle support bracket from a main landing gear bay using a portable medium-pressure Titomic D623 system. To the best of our knowledge, this is the only qualification and testing framework reported to date for a portable medium-pressure system.

This review examines the relationships between process, microstructure, mechanical properties, and repair performance for LPCS and MPCS. Earlier reviews either conflate the two regimes or treat MPCS as a variant of HPCS. Tregenza et al. [18] confirmed in their review of air-based cold spray that no prior review treats portable systems for surface engineering and repair as a unified topic. Three contributions are made here: substrate blend-out geometry is framed as a deposit integrity decision rather than a preparation detail; the sparse MPCS evidence base is documented in aggregated form as a qualification barrier; and HPCS is retained throughout as a reference baseline.

1.2. Review Methodology

The evidence relevant to portable cold spray is distributed across peer-reviewed research, published standards, and manufacturer and industry white papers. Literature was identified through Scopus for the period of 2000 to July 2026, combining process descriptors (“cold spray”, “low-pressure cold spray”, “medium-pressure cold spray”, “gas dynamic spray”, “portable cold spray”) with application terms (repair, deposition efficiency, adhesion, porosity, residual stress, fatigue, blend-out). However, terminology has evolved throughout the development of cold spray technology, and earlier studies do not always explicitly identify the process as low-pressure or medium-pressure cold spray. In many cases, the applicable pressure regime can only be determined from reported equipment specifications and operating pressure. Earlier publications may also describe related processes using terms such as kinetic spraying or supersonic particle deposition. Therefore, reference lists of retrieved articles were screened for studies the keyword search missed. Figure 2 summarises the process and eligibility criteria. Tabulated values throughout are compiled from studies differing in equipment, feedstock, substrate condition, and measurement protocol; they indicate the range reported for portable systems rather than controlled comparisons, and the limitation applying to each table is noted in its caption.

Figure 2. Flowchart of literature review methodology and thematic synthesis.

The distribution of the retrieved literature is shown in Figure 3. Cold spray publications have increased substantially, from fewer than 10 articles annually in 2000 to approaching 400 articles per year by 2024–2025. Publications specifically addressing low-pressure cold spray emerged in the late 2000s; however, the available literature remains limited, with approximately 250 relevant articles identified compared with 4500+ publications retrieved for cold spray as a broader technology. The low-pressure share has remained close to 5% since 2013 and has not risen as the field has expanded, indicating that research on portable systems has developed at a slower rate than cold spray research overall.

Figure 3. Annual publication output for cold spray in general and low-pressure cold spray from 2000 to July 2026 based on the selected bibliographic search terms. No publication was retrieved using the corresponding search term for medium-pressure cold spray; however, one paper has since been published reporting the use of a portable Titomic D623 medium-pressure cold spray system [17].

2. Fundamentals of Cold Spray

2.1. Process Overview and Bonding Mechanism

Cold spray is a solid-state additive deposition process in which metal powder particles, typically 5–50 μm in diameter, are accelerated through a converging–diverging (de Laval) nozzle to supersonic velocities in the range of 300–1200 m/s and impacted onto a substrate without bulk melting of either particle or substrate [7]. Because gas temperature is maintained well below feedstock melting points, cold spray avoids the extensive thermal exposure associated with fusion-based repair, thereby preserving the substrate alloy temper and feedstock microstructure while minimising oxidation and phase transformation, with residual stresses that are generally manageable and often compressive due to particle impact peening. These attributes distinguish it from flame, arc, and plasma spray processes.

When a particle impacts at or above a material-specific critical velocity, it undergoes extreme plastic deformation at strain rates in the order of 10⁶–10⁹ s⁻¹. The deformation produces localised adiabatic heating at the particle–substrate interface, disrupts native oxide films, and establishes intimate metal-to-metal contact (Figure 4). The resultant bond combines mechanical interlocking, arising from the geometrical interpenetration of the deforming particle into surface asperities, with metallurgical continuity at regions where oxide disruption has occurred [19].

Figure 4. Schematic representation of a cold-sprayed particle undergoing severe plastic deformation and forming a solid-state bond with a roughened substrate upon impact.

The dominant bonding mechanism has been the subject of sustained scientific debate. Assadi et al. [20] attributed bonding primarily to adiabatic shear instability (ASI) at the particle–substrate periphery, wherein the competition between strain rate hardening and thermal softening reaches an unstable threshold that expels interfacial oxides and enables metallurgical contact. Hassani-Gangaraj et al. [21] subsequently argued, based on laser-induced single-particle impact experiments, that hydrodynamic jetting driven by solid pressure waves, rather than ASI, is the primary mechanism of oxide expulsion and contact formation.

This interpretation was contested [22], and the counterargument was rebutted [23]. Both groups agree, however, on the existence of a well-defined critical velocity threshold and on the steep rise in deposition efficiency as particle velocity exceeds this threshold. This feature is of direct practical relevance to portable repair system design.

2.2. Classification of Cold Spray Systems

Cold spray systems are classified by their operating pressure, gas temperature, and powder injection architecture, portability, and operational mode (manual, mechanised, or automated). Eden et al. [24] presented one of the earliest classifications by injection method and operating regime (Figure 5), and Table 1 compares the three regimes against repair-relevant indicators. Increasing stagnation pressure and gas temperature raise attainable particle velocity from 300–600 m s⁻¹ in LPCS to about 800 m s⁻¹ in MPCS [16,25] and up to 1200 m s⁻¹ in HPCS (Figure 6). This widens the processable material set from ductile metals and composite blends to high-strength alloys and steels and raises deposition efficiency from 5–30% to 70–95%. Field deployability moves in the opposite direction, from a 30 kg shop-air LPCS unit to HPCS equipment requiring three-phase power, a high-pressure compressor, and an acoustic enclosure: the characteristics that improve deposition capability also reduce portability.

Figure 5. Classification of cold spray systems by powder injection mechanism and mode of operation. Upstream systems are typically HPCS, whereas downstream systems include MPCS and LPCS. Adapted from Eden et al. [24].

Table 1. Process characteristics and repair-relevant attributes of low-pressure (LPCS), medium-pressure (MPCS), and high-pressure cold spray (HPCS) systems.

Parameter	LPCS (Portable)	MPCS (Portable)	HPCS (Stationary, Baseline)
Stagnation pressure	<1 MPa; compressed air or N2 [13,26]	Up to 3 MPa; compressed air or N2 [16,25]	Up to 7.5 MPa; N2 or He [27,28]
Gas temperature	100–600 °C; limited by portable heater capacity [13]	100–800 °C; higher heater demand than LPCS [16]	200–1100 °C; high-power heaters required [28,29]
Particle velocity	300–600 m/s; ductile metals only [19,30]	Up to 800 m/s; some high-strength alloys [16,25]	Up to 1200 m/s; most engineering alloys [28,31]
Powder injection	Downstream (diverging section); gravity/suction feed [32]	Downstream (per equipment manual description [16])	Upstream (before throat); pressurised feeder [33]
Deposition efficiency (aluminium)	5–30% (air); improves significantly with N2 and Al2O3 addition [34,35,36]	Not reported to date in the peer-reviewed literature	Up to 95% or higher under optimised conditions [29,31]
Portability	Highly portable; handheld gun ~2 kg; complete unit ~30 kg [13,26]	Portable but heavier than LPCS 76–150 kg [16,25]	Effectively stationary; depot-level infrastructure required; newer containerised version available [12,37]
Field infrastructure	Standard 240 V; shop-air compressor; no acoustic enclosure required [13,26]	N2 supply or booster; 400 V; hearing protection required [16,25]	3-phase power; high-pressure compressor; acoustic enclosure (>100 dB) [12,37]
Material set	Al, Cu, Zn, Sn; Al–Al2O3 and Cu–Al2O3 composite blends [15,19,38]	Broader ductile alloys; Ti and Ni-base reported [16,25]	Full engineering alloys, including: Ti-6Al-4V, IN718, steels [31]
Representative systems	Dymet 423/431 [13]; Centerline SST-P/XP [26]	Titomic D623 [16]; Centerline SST EP [26]	VRC Gen III/Raptor; Impact Innovations 5/8; Spee3D WarpSPEE3D [39,40]

Note: Values are representative of commercially available systems and compiled from manufacturer specifications and peer-reviewed sources as cited; they are indicative and not prescriptive.

Table 1. Process characteristics and repair-relevant attributes of low-pressure (LPCS), medium-pressure (MPCS), and high-pressure cold spray (HPCS) systems.

Parameter	LPCS (Portable)	MPCS (Portable)	HPCS (Stationary, Baseline)
Stagnation pressure	<1 MPa; compressed air or N2 [13,26]	Up to 3 MPa; compressed air or N2 [16,25]	Up to 7.5 MPa; N2 or He [27,28]
Gas temperature	100–600 °C; limited by portable heater capacity [13]	100–800 °C; higher heater demand than LPCS [16]	200–1100 °C; high-power heaters required [28,29]
Particle velocity	300–600 m/s; ductile metals only [19,30]	Up to 800 m/s; some high-strength alloys [16,25]	Up to 1200 m/s; most engineering alloys [28,31]
Powder injection	Downstream (diverging section); gravity/suction feed [32]	Downstream (per equipment manual description [16])	Upstream (before throat); pressurised feeder [33]
Deposition efficiency (aluminium)	5–30% (air); improves significantly with N2 and Al2O3 addition [34,35,36]	Not reported to date in the peer-reviewed literature	Up to 95% or higher under optimised conditions [29,31]
Portability	Highly portable; handheld gun ~2 kg; complete unit ~30 kg [13,26]	Portable but heavier than LPCS 76–150 kg [16,25]	Effectively stationary; depot-level infrastructure required; newer containerised version available [12,37]
Field infrastructure	Standard 240 V; shop-air compressor; no acoustic enclosure required [13,26]	N2 supply or booster; 400 V; hearing protection required [16,25]	3-phase power; high-pressure compressor; acoustic enclosure (>100 dB) [12,37]
Material set	Al, Cu, Zn, Sn; Al–Al2O3 and Cu–Al2O3 composite blends [15,19,38]	Broader ductile alloys; Ti and Ni-base reported [16,25]	Full engineering alloys, including: Ti-6Al-4V, IN718, steels [31]
Representative systems	Dymet 423/431 [13]; Centerline SST-P/XP [26]	Titomic D623 [16]; Centerline SST EP [26]	VRC Gen III/Raptor; Impact Innovations 5/8; Spee3D WarpSPEE3D [39,40]

Note: Values are representative of commercially available systems and compiled from manufacturer specifications and peer-reviewed sources as cited; they are indicative and not prescriptive.

Figure 6. (a) Classification of thermal and cold spray processes technologies as a function of particle temperature and velocity, adapted from Ahmed et al. [41], copyright 2021 Springer Nature (b) classification of cold spray regimes as a function of process gas temperature and particle velocity, showing LPCS, MPCS, and HPCS.

Figure 6. (a) Classification of thermal and cold spray processes technologies as a function of particle temperature and velocity, adapted from Ahmed et al. [41], copyright 2021 Springer Nature (b) classification of cold spray regimes as a function of process gas temperature and particle velocity, showing LPCS, MPCS, and HPCS.

3. Process Windows and Deposition Boundary Conditions

3.1. Critical Velocity and the Deposition Threshold

The behaviour of portable cold spray systems is fundamentally governed by whether particles can be accelerated to an impact velocity that exceeds the material-specific critical velocity across the full repair surface, including zones of variable spray angle, varying stand-off distance, and changing substrate temperature inherent to field operations. Deposition occurs when the particle impact velocity $v_p$ exceeds the critical velocity $v_{cr}$:

$$v_p> v_{cr}$$

(1)

One of the earliest quantitative formulations was proposed by Assadi et al. [20], who related critical velocity to material properties through a regression fit of finite-element impact simulations:

$$v_{cr}=667-0.014\rho +0.08\left(T_m-T_R\right)+{10}^{-7}{\sigma }_u-0.4\left(T_i-T_R\right)$$

(2)

where $T_m$ is melting temperature, $T_i$ is particle impact temperature, $T_R$ is a reference temperature (293 K), ${\sigma }_u$ is ultimate tensile strength, and $\rho$ is particle density. Although Equation (2) provides an initial engineering framework, it assumes limited variability in particle characteristics. Schmidt et al. [42] subsequently developed a more rigorous formulation based on the onset of localised severe plastic deformation accompanied by adiabatic shear instability (ASI) at the particle–substrate interface:

$$v_{cr}=F_1\sqrt{c_p\left(T_m-T_i\right)+16{\displaystyle \frac{{\sigma }_u}{\rho }}}$$

(3)

where $c_p$ is specific heat capacity and $F_1$ is a material-dependent fitting factor accounting for strain rate sensitivity. Particle size was also recognised to influence critical velocity:

$$v_{cr}=F_2{\left({\displaystyle \frac{d}{d_0}}\right)}^{-n}\sqrt{c_p\left(T_m-T_i\right)+16{\displaystyle \frac{{\sigma }_u}{\rho }}}$$

(4)

where $d$ is particle diameter, $d_0$ is reference particle diameter, and $n$ is a material-dependent exponent. The predictive reliability of Equations (3) and (4) depends on Johnson–Cook constitutive descriptions extrapolated beyond experimentally validated strain rate regimes (10³–10⁴ s⁻¹), introducing uncertainty for materials with limited high strain rate data. Wiehler et al. [43] demonstrated that gas-atomised aluminium and copper particles exhibit effective yield strengths approximately 81% and 43% higher than bulk material values, respectively, meaning that critical velocities estimated from bulk properties systematically underestimate the actual deposition threshold, a source of unconservative error in process window predictions for LPCS where velocity margins are already tight.

Assadi et al. [33] combined gas dynamic modelling with particle motion analysis and introduced a dimensionless velocity ratio as a compact indicator of deposition conditions. Particle acceleration in the nozzle is governed by a drag–force balance:

$$v_p{\displaystyle \frac{d{v}_p}{d_z}}={\displaystyle \frac{3}{4}}{\displaystyle \frac{C_d{\rho }_g\left(v_g-v_p\right)\left|v_g-v_p\right|}{{\rho }_p{d}_p}}$$

(5)

where C_d is the drag coefficient, ${\rho }_g$ and ${\rho }_p$ are gas and particle densities, and d_p is particle diameter. The dimensionless velocity ratio is

$$\eta ={\displaystyle \frac{v_p}{v_{cr}}}$$

(6)

The velocity ratio η defines a bounded window rather than a single threshold. Below η = 1, impact is too weak to disrupt the surface oxide, so particles rebound and deposition is negligible. Near η ~ 1, only the fastest particles exceed the critical velocity, making deposition highly sensitive to small process variations. Deposition efficiency peaks near η ≈ 1.5 [33]. Figure 7 maps these regimes against gas pressure, temperature, and particle size. Portable LPCS often operates close to the lower boundary, so it is the spatial uniformity of η across the repair surface, not its mean value, that determines whether continuous bonding is maintained at corners, curved surfaces, and regions of variable stand-off.

Figure 7. Cold spray process–windows defined by the velocity ratio η, showing gas stagnation temperature T_o against (a) stagnation pressure p_o at a fixed particle diameter d_p = 20 µm and (b) particle diameter d_p at a fixed pressure p_o = 40 bar. The shaded region denotes optimal deposition near $\eta ~1.5$, while $\eta <1$ indicates no deposition and $\eta >2$ indicates erosion. Replotted from [33].

The models above were developed primarily for HPCS with upstream powder injection, where gas dynamic conditions approximate one-dimensional isentropic flow. For LPCS systems with downstream injection, Martínez-Flores et al. [32] derived an analytical model for particle acceleration in the diverging or downstream nozzle section:

$${\displaystyle \frac{d{v}_p}{dt}}={\displaystyle \frac{3}{4}}{\displaystyle \frac{{\rho }_g}{{\rho }_p{d}_p}}\left(v_g^2-C_d{v}_p^2\right)$$

(7)

This admits an analytical solution $v_p\left(t\right)=\gamma \tanh \left(\beta t\right)$ where $\gamma = {\displaystyle \frac{v_g}{\sqrt{C_d}}}$ is the theoretical velocity ceiling and $\beta = {\displaystyle \frac{3}{4}} {\displaystyle \frac{{\rho }_g}{{\rho }_p}}{\displaystyle \frac{v_g\sqrt{C_d}}{d_p}}$ characterises the system’s acceleration capability.

The model also accounts for particle deceleration in the free jet downstream of the nozzle exit, enabling stand-off distance effects to be parameterised. Its simplifying assumptions of one-dimensional isentropic flow, spherical geometry, constant $C_d$, and ideal gas behaviour limit applicability to preliminary process design and regime comparison rather than quantitative field prediction.

3.2. Practical Implications for Deposition Quality

For primary structural repair, the objective extends beyond achieving deposition. The deposit must exhibit sufficient adhesion, cohesion, and dense build-up to restore load-bearing capability under the service conditions of the deposited component [44,45,46,47]. The process window is therefore not simply the threshold for deposition but the envelope within which deposited material contributes to load transfer and resists crack initiation and growth. Figure 8 illustrates how field variability through spray angle deviation, stand-off variation, and oxide or ambient effects progressively reduces the effective velocity ratio and compresses that envelope toward the deposition threshold.

Figure 8. Schematic representation of the conceptual effect of increasing field variability on the effective velocity ratio (η) in cold spray. The introduction of field-related constraints (e.g., spray angle deviation, stand-off variation, and oxide/ambient effects) reduces the effective velocity ratio, compressing the usable deposition window toward the critical deposition threshold (η = 1), while the erosion limit (η = 2) remains the upper process boundary.

LPCS systems designed to minimise compressed air consumption and power demand typically produce particle velocities concentrated near the lower bound of the stable deposition regime [48]. Variations in stand-off distance, non-perpendicular spray angles imposed by access geometry or manual operation, and substrate temperature variability under ambient conditions collectively reduce effective particle impact velocity while potentially elevating the effective critical velocity through substrate work hardening and oxide growth. The net effect is a narrower operating margin in field repair than in laboratory characterisation, with systematically lower spatial uniformity.

Classical process window models do not account for surface oxide shells or the composite feedstocks commonly employed in LPCS to compensate for limited particle velocity [42]. For Al–Al₂O₃ composite feedstocks, the ceramic phase alters both the effective critical velocity and bonding behaviour in ways not captured by single-component formulations.

Ambient environmental variability constitutes a further, uncharacterised source of process window narrowing. Equipment manufacturer guidelines recommend ambient temperatures between 15 and 35 °C for stable deposition [48], but the structural consequences of operating outside of this range, particularly in marine environments with high humidity, salt contamination, and low substrate temperatures, have not been investigated systematically. Powder exposure to ambient humidity is known to degrade deposition efficiency, consistent with oxide thickening that raises the effective critical velocity [34]. This is a gap of direct operational significance for maritime and arctic deployment.

4. Deposition Efficiency and Build-Up Capability

4.1. Deposition Efficiency

Deposition efficiency (DE) is defined as the fraction of feedstock powder that successfully adheres to the substrate during spraying and is calculated from direct mass measurements:

$$\mathrm{DE}\left(\%\right)={\displaystyle \frac{\mathrm{mass}\mathrm{of}\mathrm{deposit}}{\mathrm{mass}\mathrm{of}\mathrm{feedstock}}}\times 100$$

(8)

DE is a direct indicator of process effectiveness and enables comparison across spray parameters, feedstock compositions, substrate conditions, and system configurations [49]. High DE indicates particles reaching sufficient velocity and plastic deformation for effective bonding and correlates broadly with lower porosity and stronger interparticle interfaces, but it does not by itself guarantee deposit quality, which also depends on powder morphology, gas dynamics, oxidation behaviour, substrate material type, and surface condition.

Optimising DE requires managing several interdependent variables: gas pressure and temperature, particle size, powder feed rate, spray angle, stand-off distance, and substrate preparation. Shorinov et al. [34] identified gas temperature and powder feed rate as the dominant variables governing DE in portable LPCS, with suboptimal conditions producing porous, weakly bonded deposits that cannot reliably transfer load. The same group later demonstrated the strength of this dependence by redesigning the integrated heater of a portable DIMET-405 unit from 3.3 to 5.3 kW, sustaining 605 °C at 1 MPa against 373 °C for the standard heater [50]. This increased the deposition efficiency of aluminium on titanium from 13.4% to 60.6% and the coating thickness more than fivefold under identical conditions.

Table 2 summarises representative DE values for LPCS. DE is improved by adding hard ceramic particles such as Al₂O₃ to ductile metallic powders to form metal–ceramic composite feedstocks [51,52]. Winnicki et al. [53] reported a 7.2% increase in DE when 50 wt% Al₂O₃ was added to a Cu matrix on AA1350. The improvement depends strongly on base powder morphology; however, dendritic electrolytic copper (E-Cu) outperformed spherical gas-atomised copper (S-Cu) under equivalent conditions, and ceramic addition modified both trends (Table 2). High ceramic fractions and powder recycling are in any case often impractical in the field owing to process instability, nozzle wear, and inconsistent powder delivery.

Table 2. Representative values of deposition efficiency (DE) reported for LPCS systems for different powder–substrate combinations and process conditions.

Feedstock Powder	Substrate	DE (%)	Method	Process Parameter	Study Objective	Ref.
Al	Ti (VT3-1)	60.6	Mass-based	Air, 1 MPa, 605 °C, 5.3 kW heater	Increase DE by redesigning the integrated gas heater to sustain high gas temperature at elevated pressure	[50]
E-Cu (dendritic) E-Cu + 50 wt% Al2O3	AA1350	28.6/35.8	Mass-based	N2, 0.9 MPa, 600 °C	Morphology and ceramic effects on corrosion/conductivity	[53]
S-Cu (spherical) S-Cu + 50 wt% Al2O3	AA1350	6.7/4.9	Mass-based	N2, 0.9 MPa, 600 °C	Morphology and ceramic effects on corrosion/conductivity	[53]
Cu-159 (Praxair)	AA7075	34.8	Mass: deposit mass/powder consumed	Air, 0.5–0.8 MPa, 600 °C	Effect of pressure and stand-off on velocity and DE	[54]
Hydroxyapatite + Cu + Zn	316L SS	27	Mass-based	Air, 0.7 MPa, 400 °C	HAP deposits with Cu-Zn blends	[49]

Entries in Table 2 differ in propellant gas, stagnation pressure, and gas temperature. Because DE rises steeply once particle velocity exceeds the critical threshold, modest differences in these variables produce large changes in the reported value. Feedstock morphology acts independently, with dendritic and spherical copper sprayed under identical conditions in one study differing by more than fourfold [53]. Shorinov et al. [50] set out to maximise deposition efficiency, whereas the remaining entries report it incidentally under parameters selected for other endpoints. Direct comparison across the table is therefore not warranted.

Published data for HPCS report DE consistently exceeding 70% for copper and aluminium under optimised conditions [31]. Available commercial specifications for MPCS systems indicate a deposition rate of approximately 60 g/min [16].

4.2. Build-Up Capability and Structural Consequences

The structural implications of DE extend well beyond production economics. For LPCS, the combination of lower DE and lower deposition rate means that achieving millimetres of net material build-up for volumetric structural restoration requires a greater number of spray passes and longer cumulative dwell times over the repaired region [55]. Deforce et al. [47] recommended a minimum aluminium deposit thickness of at least 400 µm to ensure long-term barrier protection, establishing a practical lower bound for repair fill targets.

Traverse speed is a dominant kinematic variable governing build-up rate. Li et al. [56] showed for Cu + Al₂O₃ at 0.9 MPa nitrogen and 400 °C that reducing traverse speed from 200 to 50 mm/s increased deposit area per pass roughly fourfold (Figure 9), with lower porosity and near-linear scaling of deposited mass with pass count, as the denser particle flux raises both peening intensity and the probability of contacting each asperity. Seraj et al. [57] showed that for higher-strength materials, excessively slow traverse produces rebound-dominated erosion of previously deposited layers because thickening of the compressed gas layer beneath the spray spot reduces the normal component of impact velocity. Klinkov et al. [58] reported the same non-monotonic dependence for LPCS.

Figure 9. Transverse thickness profiles of LPCS Cu + Al₂O₃ deposits at a constant stand-off distance of 5mm sprayed at gun traverse speeds (a) 50mm/s, (b) 100mm/s, (c) 150mm/s and (d) 200mm/s, each for 4, 8, 12, 16 passes. Peak thickness after 16 passes falls from approximately 1.15mm at 50mm/s to below 0.4mm at 200mm/s, while the profiles broaden and flatten, showing that slower traverse concentrates particle flux and raises build-up rate per pass. Replotted from [56].

Track spacing controls deposit cross-sectional uniformity through the degree of overlap between adjacent Gaussian deposit profiles. Tracks spaced too widely leave under-filled edge regions with systematic inter-track porosity; tracks spaced too narrowly accumulate heat and can cause localised work hardening that degrades bonding quality in subsequent passes.

The structural consequences are direct. Individual pass increments for LPCS are of the order of tens of micrometres, so a 1 mm net fill requires some 10–20 passes over the same area (Figure 9). In handheld LPCS, traverse speed and track spacing are set by the operator, introducing layer-to-layer thickness variation; successive passes over an initially Gaussian profile alter local deposition efficiency and surface topology, raising the probability of interlaminar bonding defects at pass interfaces that can act as fatigue initiation sites [56]. A possible practical mitigation is to alternate raster scan directions between consecutive layers (i.e., a cross-hatch pattern), distributing track-edge porosity across different orientations and reducing the spatial correlation of inter-track defect planes [59].

4.3. Influence of Substrate Preparation on Deposition Efficiency

The cold spray repair literature has historically under-reported the specifics of substrate preparation and has rarely isolated its effects from those of process parameters. Bonding-focused investigations establish that surface roughness principally affects the interface and the first few deposited layers, whereas the subsequent build-up regime is dominated by particle–particle bonding dynamics [7]. Consequently, more aggressive blasting does not necessarily produce a proportional increase in deposition efficiency or bond strength once coating build-up has commenced. This relationship is often obscured because many studies report only the grit type and blasting pressure without quantifying the resulting surface roughness, cleanliness, or oxide condition.

Grit blasting, sanding, and grinding remain the most widely used preparation methods for portable repairs because they increase asperity contact and promote localised shear instability during particle impact (Table 3). Boruah and Zhang [60] reported higher adhesion for Ti6Al4V deposited onto smoother Ti6Al4V substrates, whereas Marrocco et al. [61] observed higher adhesion for commercially pure Ti deposited onto grit-blasted Ti6Al4V substrates. This apparent contradiction may reflect differences in substrate condition, particle velocity, powder ductility, and material combination. Both average surface roughness, $R_a$, and surface morphology influence deposition efficiency and first-layer adhesion. However, suitable post-blasting roughness ranges for LPCS and MPCS repairs remain poorly defined. Shorinov et al. [35] reported average post-blasting roughness values of 120 to 160 µm for different powder combinations deposited onto EZ33A-T5 magnesium alloy.

Surface preparation also affects fatigue performance. Ziemann et al. [62] found that glass bead blasting increased the fatigue life of commercially pure aluminium deposits on AA2024 by up to 50% at lower stress amplitudes. However, excessively aggressive grit blasting reduced fatigue performance because of increased porosity and greater interfacial irregularity. For field applications, portable blasting units, manual abrasion tools, and power brushing provide practical alternatives. Although these methods produce less uniform surface conditions than controlled laboratory blasting, they still offer substantial improvement over untreated surfaces [63,64]. Quantitative relationships between field-feasible preparation methods, deposition efficiency, adhesion, and long-term mechanical durability remain limited, particularly for portable LPCS and MPCS repair systems.

Table 3. Selected LPCS repair studies classified by process gas, feedstock composition, substrate material, and surface preparation method.

Gas	Feedstock Powder	Substrate	Temperature (°C)/Pressure (MPa)	Stand-Off Distance (mm)	Surface Preparation	Ref.
Air	Al + Al2O3	Q235 steel	400/0.7	20	Grit blast	[65]
	Cu + Al; Cu + Al + Al2O3	S235 JR	600/0.9	10	Grit blast	[66]
	Al + Zn; Al + Al2O3; Al + Zn + Al2O3	EZ33A-T5 Mg	300–600/1.0	10	Sandblast	[35]
	AA2024 + Al2O3	AA2024	600/0.9	12	Sandblast	[67]
	Al + Bronze + Al2O3	Cr12MoV steel	550/0.8	10–20	Sandblast	[68]
	Ni-Al2O3	VT3-1 Ti alloy	537/0.9	11	Grit blast	[34]
	Al + Al2O3	AA7075-T73	350/0.6	5–10	Grit blast	[10]
	Cu + Al2O3; Al + Al2O3	AISI 304 SS	200–600/0.9	10	Grind	[69]
	Hydroxyapatite	TiAl6V4	500–600/0.6	-	--	[70]
	Cu; Al + Al2O3	AA7075-T651	500/0.8	10	Grit blast	[71]
	Al + Bronze	C45	500/0.6–0.8	15	--	[72]
	Cu; Al + Al2O3	304 SS	500/0.8	10	Grit blast	[73]
	Zn + Ni + Al2O3	Q235 carbon steel	400/0.6	25	Grit blast	[74]
	Cu; Al + Al2O3	Q235 CS	413/0.6	20	Grit blast	[75]
	Cu + Al2O3	Polyamide 6	200–300/0.9	10	Acetone clean	[76]
	Ni + CNTs	AA1050	600/0.6	12	Sandblast	[77]
	Fe-Cr-B-C + AA6061	AA7075-T6	350–450/0.6	10	Coarse sandpaper	[78]
	Cu; Ti	GFR epoxy panels	150–450/0.6	50	--	[79]
N2	Al + Al2O3	Polymer matrix composite	360/0.6	10–30	Grit blast	[80]
	Al + Al2O3	AA2024-T3	25–500/0.2–1.7	-	Grit blast	[81]
	Al; Al + Al2O3	Steel plate	500/0.7	10	Grit blast	[82]
	Al; Al + Al2O3	Mild steel; AA7075	500/0.6	10	Grit blast	[51]
	Ni + Al2O3	S235JR	500/0.9	10	Grit blast	[83]
	Al; AA2024; AA5083 + Al2O3	AA2024-T351	600/0.8	12	Sandblast	[84]
	TiO2	Al AW-1054 H14/H24	600/1.0	10	Grit blast	[85]
	Ni + Al2O3	T22 steel	400–600/0.6	20–47	Grit blast	[86]
	AlSi10Mg	Epoxy Vitrimer	200–300/0.6	70	--	[87]

Table 3. Selected LPCS repair studies classified by process gas, feedstock composition, substrate material, and surface preparation method.

Gas	Feedstock Powder	Substrate	Temperature (°C)/Pressure (MPa)	Stand-Off Distance (mm)	Surface Preparation	Ref.
Air	Al + Al2O3	Q235 steel	400/0.7	20	Grit blast	[65]
	Cu + Al; Cu + Al + Al2O3	S235 JR	600/0.9	10	Grit blast	[66]
	Al + Zn; Al + Al2O3; Al + Zn + Al2O3	EZ33A-T5 Mg	300–600/1.0	10	Sandblast	[35]
	AA2024 + Al2O3	AA2024	600/0.9	12	Sandblast	[67]
	Al + Bronze + Al2O3	Cr12MoV steel	550/0.8	10–20	Sandblast	[68]
	Ni-Al2O3	VT3-1 Ti alloy	537/0.9	11	Grit blast	[34]
	Al + Al2O3	AA7075-T73	350/0.6	5–10	Grit blast	[10]
	Cu + Al2O3; Al + Al2O3	AISI 304 SS	200–600/0.9	10	Grind	[69]
	Hydroxyapatite	TiAl6V4	500–600/0.6	-	--	[70]
	Cu; Al + Al2O3	AA7075-T651	500/0.8	10	Grit blast	[71]
	Al + Bronze	C45	500/0.6–0.8	15	--	[72]
	Cu; Al + Al2O3	304 SS	500/0.8	10	Grit blast	[73]
	Zn + Ni + Al2O3	Q235 carbon steel	400/0.6	25	Grit blast	[74]
	Cu; Al + Al2O3	Q235 CS	413/0.6	20	Grit blast	[75]
	Cu + Al2O3	Polyamide 6	200–300/0.9	10	Acetone clean	[76]
	Ni + CNTs	AA1050	600/0.6	12	Sandblast	[77]
	Fe-Cr-B-C + AA6061	AA7075-T6	350–450/0.6	10	Coarse sandpaper	[78]
	Cu; Ti	GFR epoxy panels	150–450/0.6	50	--	[79]
N2	Al + Al2O3	Polymer matrix composite	360/0.6	10–30	Grit blast	[80]
	Al + Al2O3	AA2024-T3	25–500/0.2–1.7	-	Grit blast	[81]
	Al; Al + Al2O3	Steel plate	500/0.7	10	Grit blast	[82]
	Al; Al + Al2O3	Mild steel; AA7075	500/0.6	10	Grit blast	[51]
	Ni + Al2O3	S235JR	500/0.9	10	Grit blast	[83]
	Al; AA2024; AA5083 + Al2O3	AA2024-T351	600/0.8	12	Sandblast	[84]
	TiO2	Al AW-1054 H14/H24	600/1.0	10	Grit blast	[85]
	Ni + Al2O3	T22 steel	400–600/0.6	20–47	Grit blast	[86]
	AlSi10Mg	Epoxy Vitrimer	200–300/0.6	70	--	[87]

5. Microstructural Characteristics of LPCS and MPCS Deposits

5.1. Porosity, Influence of Residual Oxides, and Ceramic Fragmentation

The microstructures of aluminium, copper, nickel, and zinc alloy deposits produced through LPCS have been extensively investigated, particularly for feedstocks containing hard particles or ceramics such as alumina (Al₂O₃) powder to improve DE and deposit densification [34,35,49,67,72,85,88,89,90,91,92,93,94]. Figure 10 illustrates the characteristic distribution of pores in a commercially pure aluminium deposit sprayed using LPCS and Al₂O₃ inclusions in MPCS systems. The comparatively low particle impact velocities associated with LPCS increase the likelihood of incomplete interparticle bonding and, consequently, greater pore populations than those generally obtained under higher-energy spraying conditions. This effect becomes particularly pronounced when the native oxide films at particle contact surfaces are not sufficiently disrupted during impact [84].

Lienhard et al. [95] demonstrated that oxide films on aluminium particles inhibit metallurgical bonding unless they are fractured and dispersed by severe high-strain-rate plastic deformation. This has important implications for LPCS because oxide disruption is strongly influenced by local jetting and interfacial deformation rather than thermal softening alone [21]. The lower impact velocities attained in LPCS may therefore be insufficient to generate effective jetting across all particle contact regions. Consequently, residual oxide-bearing interparticle boundaries may remain within the deposit, acting as planar defects and potential sites for pore formation, crack initiation, and interparticle decohesion.

The addition of hard ceramic particles can partly mitigate these limitations through mechanical tamping, localised deformation, and disruption of surface oxide films during impact. Ceramic particles may also fragment under repeated high-velocity collisions, with the resulting fragments becoming embedded within the ductile metallic matrix. Although this fragmentation can improve particle packing and promote local densification, excessive ceramic content or the formation of angular fragments may introduce stress concentrations, particle pull-out sites, and additional micro voids. The final microstructure therefore depends on the balance between ceramic-assisted densification and defect formation associated with ceramic fracture, poor wetting, or incomplete metal–ceramic interfacial bonding.

Figure 10. Representative SEM micrographs of (a) a pure Al deposit produced through LPCS and (b) Al–Al₂O₃ composite deposit produced through MPCS. The images illustrate differences in pore distribution, interparticle bonding, and ceramic–particle incorporation under different spray energy conditions. The Al–Al₂O₃ composite exhibits comparatively lower porosity, which is attributed partly to ceramic-assisted tamping and enhanced deformation of the ductile Al matrix. (a) Reproduced from [96], copyright 2022 MDPI; (b) reproduced from [17], copyright 2025 MDPI.

Figure 10. Representative SEM micrographs of (a) a pure Al deposit produced through LPCS and (b) Al–Al₂O₃ composite deposit produced through MPCS. The images illustrate differences in pore distribution, interparticle bonding, and ceramic–particle incorporation under different spray energy conditions. The Al–Al₂O₃ composite exhibits comparatively lower porosity, which is attributed partly to ceramic-assisted tamping and enhanced deformation of the ductile Al matrix. (a) Reproduced from [96], copyright 2022 MDPI; (b) reproduced from [17], copyright 2025 MDPI.

Through finite element modelling of multi-particle impact, Weiller and Delloro [97] distinguished two sources of porosity: interfacial porosity, from gaps at particle–particle contacts where flattening is incomplete, and stack porosity, from variations in local particle number density in the gas stream. The two respond differently to process parameters, the first to impact velocity and the flattening ratio and the second to the powder feed rate and gas flow uniformity. Both are more prominent in LPCS, where impact velocities are constrained and radial downstream injection introduces inhomogeneity, than in high-pressure systems [31].

Spray angle exerts a measurable influence on microstructural quality. Schmidt et al. [98] and Binder et al. [99] both demonstrated that deviation from perpendicular particle impact can increase the porosity population and lower the mechanical properties of the deposit, with the effect becoming significant beyond approximately 45° from the ideal 90° impact angle. This sensitivity is important in field repairs, where restricted access can impose oblique spray angles that depart from the near-normal incidence under which most laboratory characterisation data are obtained.

The porosity values collated in Table 4 were all obtained through two-dimensional cross-sectional image analysis, which tends to underestimate the true porosity content of a cold spray deposit. Koivuluoto et al. [100] found that LPCS deposits of Cu, Ni, and Zn alloys that appeared dense in 2D SEM cross-sections nevertheless contained interconnected porosity, revealed by electrochemical corrosion testing, as a 2D section only samples pores intersecting the cut plane. Porosity acceptance criteria derived from 2D measurements are therefore non-conservative, a technical barrier to qualifying primary load-bearing structures; the values in Table 4 should be read as lower bounds.

Table 4. Porosity measurements for LPCS-MPCS deposits under various process conditions. All values were obtained through two-dimensional cross-sectional image analysis and should be read as lower bounds on true volumetric porosity.

Feedstock Powder	Porosity (% of Area)	Measurement Method	Study Objective	Gas	Ref.
Al (CP-Al)	<2.00	2D optical image analysis	Corrosion protection of AA2024-T3 Alclad	N2	[81]
Al (CP-Al)	<1.20	2D image analysis/ASTM E2109	Effect of surface preparation on microstructure and adhesion	He and N2	[101]
Al-2024	1.15 ± 0.61	2D optical image analysis (20 locations)	Mechanical properties for MIL-STD-3021 non-primary structural element applications	He	[102]
Al-7075	0.01 ± 0.01	2D optical image analysis (20 locations)	Baseline properties for cold spray deposit applications	He
Cu (E-Cu; Al2O3-Cu)	4.70	2D optical image analysis	Electrolytic vs. gas-atomised Cu performance	N2	[53]
AA2024-Al2O3	0.55	2D optical image analysis	Effect of Al2O3 on the microstructure of AA2024-T3	N2	[67]
Al–Al2O3	0.40	2D optical image analysis	Qualification and testing framework in cold spray repairs	-	[17]

X-ray computed microtomography (XCT) resolves the deposit volume in three dimensions. Figure 11 shows an AA7075 powder repair of an AA7075-T6 wrought plate sprayed at 6 MPa, with a porosity volume fraction of 0.18% concentrated near the specimen’s surface. Porosity is attributed to incomplete interparticle bonding, insufficient particle deformation, or oxidation during powder atomisation [29]. The pores were irregular in shape, consistent with formation between incompletely bonded particles, and the largest measured about 0.6 mm across. High-resolution XCT is slow, costly, and limited by facility access, so electrochemical corrosion testing has been used as a complementary probe of through-thickness porosity.

Figure 11. X-ray computed tomography (XCT) images: (a) schematic of the specimen; only the section within the dashed lines was scanned through XCT; (b) image showing porosity defects; volume fraction 0.18%; the colour index shows pore diameter ranges of 0–0.6 mm; (c) an enlarged view to show pores with diameters > 0.2 mm; (d) a higher-magnification image of irregularly shaped pores with a scale bar (0.1 mm) to indicate pore size. Reproduced from [29], copyright 2025 Elsevier.

Shorinov et al. [35] demonstrated that gas temperature is a non-monotonic process variable: an increase in gas temperature to 600 °C produced a slight increase in porosity in deposits produced by a Dymet 413 LPCS system, as illustrated in Figure 12, attributed to pre-impact thermal softening reducing the kinetic energy available for interfacial plastic deformation and incipient nozzle clogging by partially softened particles disrupting the stable supersonic jet. An optimum gas temperature in the range of 450–550 °C or incorporation of Al₂O₃ particles suppressed inter-particle porosity through enhanced mechanical peening.

Figure 12. Porosity area fraction as a function of gas temperature for LPCS deposits produced from different aluminium-based powder mixtures illustrating the non-monotonic relationship and the optimum temperature window. Replotted from [35], copyright 2023 IOPScience.

Porosity reduction through ceramic addition is well-documented. Jiang et al. [103] reported a reduction in areal porosity fraction from 6.34% to 2.07% with Al₂O₃ addition in Al/Al₂O₃ composite deposits, which is consistent with the broader composite feedstock literature. Piwowarczyk et al. [104] demonstrated that porosity spatial distribution is as important as volume fraction and that pores adjacent to the interface or brittle ceramic stringers can dominate fatigue crack initiation behaviour even when the bulk volume fraction of porosity is low, an observation directly relevant to repair qualification.

5.2. Effect of Ceramic Particle Size on Deposit Microstructure

Although an optimum ceramic content of approximately 20–40 wt% is generally reported for achieving a balance between DE and mechanical performance, ceramic particle size independently governs the mechanisms controlling deposit formation and the resulting microstructure [51,52,84]. Whereas ceramic content determines the amount of reinforcement introduced into the deposit, particle size dictates how the hard phase interacts with the metallic matrix on impact and therefore governs deposit build-up, defect formation, and structural integrity.

Sova et al. [105] showed that ceramic particle size controls competition between surface activation and erosion. Fine Al₂O₃ (~19 μm) and SiC (~25 μm) particles repeatedly strike the growing deposit, removing surface oxides, generating fresh asperities, and peening the surface, raising the DE of the metallic component above that of pure metal powders and lowering the critical gas temperature by some 60–100 K. Coarse Al₂O₃ (~141 μm) instead eroded the deposit across the whole processing window. Because erosive potential scales with particle volume while surface activation depends on surface interaction, the detrimental effect rises rapidly with diameter, defining an upper size limit beyond which ceramic reinforcement is counterproductive.

Ceramic particle size also exerts a direct influence on deposit microstructure because ceramic particles cannot deform plastically during impact. Instead, sufficiently large particles tend to shatter, producing smaller angular fragments that become embedded within the metallic matrix and act as preferential sites for microporosity and microcrack initiation. Chupong et al. [106] observed fractured Al₂O₃ particles distributed throughout LPCS Al–Al₂O₃ deposits, whereas the incorporation of ultrafine TiN (~1 μm) filled interparticle crevices reduced porosity and suppressed crack formation (Figure 13), demonstrating that finer hard particles improve particle packing and interparticle cohesion. Similarly, Feng et al. [72] reported that decreasing both metallic and Al₂O₃ particle sizes increased particle deformation, reduced porosity, and produced coatings with superior mechanical properties, with optimum performance obtained using particle sizes of approximately 23–25 μm.

Figure 13. SEM (backscattered electron micrographs) of deposits showing shattered ceramics dispersed in (a) Al–Al₂O₃ and (b) Al–Al₂O₃/TiN deposits. Reproduced from [106], copyright 2025 MDPI.

None of this is represented in the critical velocity models of Section 3.1, which treat particle diameter as a single characteristic of a homogeneous material and so describe only the metallic phase, neglecting the independently sized, non-deforming reinforcement. Models incorporating ceramic particle size, and consistent reporting of ceramic size distributions, remain research needs; ceramic particle size should meanwhile be treated as a primary process variable for composite feedstocks.

5.3. Interface Morphology and Bonding Integrity

Bond quality in LPCS repairs is determined at two distinct boundaries: the interparticle boundary within the deposit, and the deposit–substrate interface. Both are limited by the extent to which impact-induced plastic strain fractures and disperses the native oxide film at the contact, although the contact conditions differ, as the substrate is prepared, initially unhardened, and generally of different roughness and hardness from the impacting particle. Within the deposit, LPCS particles arriving at 300–600 m/s develop insufficient strain at the contact to disrupt the oxide across the full contact area once the elevated effective yield strength of gas-atomised powder is accounted for (Section 3.1). Bonding is consequently localised to mechanically interlocked asperity contacts rather than distributed as metallurgical continuity across the splat interface [21,95]. Figure 14a shows the resulting boundary structure. Kundu and Stamoulis [107] observed the mechanical consequence: LPCS AA6061 fractured through brittle quasi-cleavage across flat splat facets, whereas HPCS deposits of identical geometry fractured through bonded metal with tear ridges and dimpling, indicating micro void coalescence instead of separation along particle boundaries. The measured elastic modulus was 62 GPa for the LPCS deposit against 64 GPa for HPCS and 69 GPa for the wrought substrate, a deficit consistent with a higher density of unbonded interparticle area. Both deposits were monolithic, so this contrast is attributable to velocity regime alone.

Figure 14. Fracture surfaces of cold-spray-repaired specimens after tensile testing: (a) LPCS-repaired AA6061 showing flat splat facets in the deposit against a fibrous, ductile substrate; (b) HPCS-repaired AA7075 showing tear ridges and dimpling with a discrete crack at the deposit–substrate boundary. (a) Reproduced from [107] and (b) from [29], copyright 2025 Elsevier.

At the deposit–substrate interface, the same oxide disruption limitation applies, and the interface remains a preferred failure plane even where interparticle cohesion is adequate. Bakir et al. [29] reported a discrete crack along the deposit–substrate boundary in HPCS-repaired AA7075 despite ductile fracture through the deposit (Figure 14b). Raising particle velocity therefore improves interparticle cohesion more readily than it eliminates the interface as a failure plane, which is consistent with the fatigue evidence that cracks initiate at the deposit–substrate interface across materials, geometries, and process regimes (Section 6.4). Initiation resistance is thus governed principally by interface conditions, while resistance to subsequent propagation is governed by the interparticle boundary structure within the deposit. Hard-phase addition acts on both boundaries. At the substrate interface, ceramic impact roughens and cleans the surface and generates asperities, increasing contact area and mechanical anchoring, and Spencer et al. [108] showed that Al₂O₃ reinforcement of cold-sprayed aluminium shifts the failure site from the deposit–substrate interface to cohesive failure within the deposit, indicating that interfacial adhesion had been raised above deposit cohesive strength. Within the deposit, retained ceramic interrupts metallic contact, and Dayı et al. [44] observed microcracks initiating preferentially at ceramic–matrix boundaries under tensile load. Figure 15 shows an Al–Al₂O₃ deposit–substrate interface exhibiting both features: embedded alumina particles and particle penetration into the substrate with localised plastic deformation of the matrix are seen, alongside interfacial gaps and retained porosity. Based on the available evidence, drawn from separate studies rather than from a single system in which both boundaries were measured, hard-phase addition improves adhesion at the substrate interface while introducing an additional crack initiation path within the deposit.

Figure 15. Backscattered electron SEM micrographs of an Al–Al₂O₃ deposit–substrate interface showing interfacial gaps, embedded Al₂O₃ particles, particle penetration, connected and isolated pores, and localised plastic deformation. Reproduced from [106], copyright 2025 MDPI.

The implication for repair selection is that the two boundaries respond differently to the measures available to portable systems. Interfacial adhesion can be improved through surface preparation and hard-phase addition, whereas interparticle bond quality remains limited by particle velocity and is degraded at high ceramic content (Section 5.2). LPCS repairs are accordingly better suited to restoration of surface function than recovery of load-bearing capacity in fatigue-critical members.

6. Mechanical Performance of Cold Spray Deposits

Substantiating a durable cold spray deposit requires evidence at two levels, the material and the assembly (Figure 16). At the material level, the deposit must show adequate through-thickness hardness, tensile properties, wear and corrosion resistance, and an elastic modulus compatible with the substrate; at the structural level, the assembly must show adequate adhesion and load transfer, fatigue life retention under service loading, and a residual stress state that does not promote distortion or cracking. The following sections address each in turn for portable LPCS and MPCS, with HPCS retained as a baseline.

Figure 16. Fishbone (Ishikawa) diagram showing the multi-scale determinants of cold spray repair durability. Material-level attributes and structure-level performance converge in the assessment of durable, high-integrity repair deposits. Developed from information in [59,109,110].

6.1. Adhesion Strength

Adhesion strength testing is the most widely used mechanical screening test for structural repairs. ASTM C633 [111] and the more recent ASM Thermal Spray Society acceptance practice are the standard methods for measuring tensile adhesion and cohesion strength through pull-off testing [112]. The values cannot be used directly in structural design calculations, being influenced by residual stress and the geometry of the deposited part [19]. A further standard, ASTM D4541 [113], provides a portable pull-off test suited to the field, measuring the perpendicular tensile force needed to detach a bonded plug. Maxwell [114] compared seven thermally sprayed and electroplated deposits and found that ASTM C633 returns adhesion strengths 15–35% higher than ASTM D4541 for high-adhesion deposits, a difference attributed to its better load alignment and surface preparation protocol.

A limitation common to both methods is that failure of the bonding adhesive itself may become the limiting failure plane, obscuring the nature of the fracture, that is, whether failure is adhesive (at the deposit–substrate interface) or cohesive (within the deposit bulk). Adhesion test results should therefore be treated as screening indicators of bonding continuity, and qualification must complement them with other mechanical tests, including tensile and fatigue evaluation.

Reported adhesion strengths for LPCS aluminium deposits on aluminium substrates span roughly 15–50 MPa under ASTM C633-type testing (Table 5), with the higher values obtained with nitrogen rather than compressed air and with optimised surface preparation. Bi et al. [102] reported adhesion and mechanical property data for LPCS AA2024 and AA7075 deposits in the context of MIL-STD-3021 qualification for non-primary structure repairs, one of the few datasets framed within a regulatory context. Yandouzi et al. [81] reported 26.9 MPa for LPCS commercially pure Al on AA2024-T3.

Excessive addition of hard particles, however, raises the proportion of weak metal–ceramic bonds within the deposit and reduces its cohesive strength, shifting failure from the interface to the deposit interior [115].

Table 5. Reported adhesion strengths for portable low-pressure cold-sprayed deposits highlighting the influence of process gas, feedstock composition, substrate, and surface preparation.

Gas	Feedstock Powder	Substrate	Temperature (°C)/Pressure (MPa)	Stand-Off Distance (mm)	Adhesion Strength (MPa)	Ref.
N2	Pure Al	AA2024-T3	25–500/0.2–1.7	-	26.9	[116]
	Ni	T22 steel	400–600/0.6	20–47	26.1	[86]
	CP-Al	AA2024-T3	600/0.8	12	46	[84]
	AA2024	AA2024-T3	600/0.8	12	50
	AA5083	AA2024-T3	600/0.8	12	51
	Al–Al2O3	AA2024-T3	600/0.9	12	50.7	[67]
Air	TiO2	Al AW-1054 H14/H24	600/0.5	10	-	[85]
	Al–Zn	EZ33A-T5 Mg alloy	300–600/1.0	10	15.2	[35]
	Al–Al2O3	EZ33A-T5 Mg alloy	300–600/1.0	10	31.9
	Al–Zn–Al2O3	EZ33A-T5 Mg alloy	300–600/1.0	10	31.9
	Ni–Al2O3	VT3-1 Ti alloy	537/0.9	11	34.8	[34]
	Hydroxyapatite (HAP)	TiAl6V4	500–600/0.6	-	22.0	[70]
	Cu-5Sn-Al2O3-Ag	Al7075-T651	500/0.8	10	35.0	[71]
	Al-Bronze-Al2O3	C45	500/0.6–0.8	15	32.0	[72]
	Cu–Al2O3-Graphite solid	304 SS	500/0.8	10	7.5	[73]
	Zn–Ni–Al2O3	Q235 carbon steel	400/0.6	25	37.0	[75]
	Al–Al2O3;	AISI 304 SS	200–600/0.9	10	41.6	[69]
	Cu–Al2O3	AISI 304 SS	200–600/0.9		33.9
	Cu–Al2O3	Polyamide 6	200–300/0.9	10	3.6	[76]
	Sn–Al2O3	AA1350 AA5754	200–400/0.5, 0.7	20	33 34	[104]
	Zn–Al2O3	AA1350 AA5754	200–400/0.5, 0.7	20	32 55
	Al–Al2O3	AA1350 AA5754	200–400/0.5, 0.7	20	55 55

Table 5. Reported adhesion strengths for portable low-pressure cold-sprayed deposits highlighting the influence of process gas, feedstock composition, substrate, and surface preparation.

Gas	Feedstock Powder	Substrate	Temperature (°C)/Pressure (MPa)	Stand-Off Distance (mm)	Adhesion Strength (MPa)	Ref.
N2	Pure Al	AA2024-T3	25–500/0.2–1.7	-	26.9	[116]
	Ni	T22 steel	400–600/0.6	20–47	26.1	[86]
	CP-Al	AA2024-T3	600/0.8	12	46	[84]
	AA2024	AA2024-T3	600/0.8	12	50
	AA5083	AA2024-T3	600/0.8	12	51
	Al–Al2O3	AA2024-T3	600/0.9	12	50.7	[67]
Air	TiO2	Al AW-1054 H14/H24	600/0.5	10	-	[85]
	Al–Zn	EZ33A-T5 Mg alloy	300–600/1.0	10	15.2	[35]
	Al–Al2O3	EZ33A-T5 Mg alloy	300–600/1.0	10	31.9
	Al–Zn–Al2O3	EZ33A-T5 Mg alloy	300–600/1.0	10	31.9
	Ni–Al2O3	VT3-1 Ti alloy	537/0.9	11	34.8	[34]
	Hydroxyapatite (HAP)	TiAl6V4	500–600/0.6	-	22.0	[70]
	Cu-5Sn-Al2O3-Ag	Al7075-T651	500/0.8	10	35.0	[71]
	Al-Bronze-Al2O3	C45	500/0.6–0.8	15	32.0	[72]
	Cu–Al2O3-Graphite solid	304 SS	500/0.8	10	7.5	[73]
	Zn–Ni–Al2O3	Q235 carbon steel	400/0.6	25	37.0	[75]
	Al–Al2O3;	AISI 304 SS	200–600/0.9	10	41.6	[69]
	Cu–Al2O3	AISI 304 SS	200–600/0.9		33.9
	Cu–Al2O3	Polyamide 6	200–300/0.9	10	3.6	[76]
	Sn–Al2O3	AA1350 AA5754	200–400/0.5, 0.7	20	33 34	[104]
	Zn–Al2O3	AA1350 AA5754	200–400/0.5, 0.7	20	32 55
	Al–Al2O3	AA1350 AA5754	200–400/0.5, 0.7	20	55 55

Piwowarczyk et al. [104] reported adhesion strengths exceeding 50 MPa for LPCS repairs of aluminium casting defects using Al, Zn, and Sn powders reinforced with Al₂O₃ measured through a pull-off test, as shown in Figure 17; their cohesive and adhesive fracture mode distributions confirm that the interfacial bond strength was sufficient to shift failure into the deposit interior for Al deposits.

Figure 17. Static tensile (pull-off) adhesion test results for cold-sprayed Al, Zn, and Sn deposits on different substrates, where Al1350 and Al5754 denote the substrate alloys: Al1350 is a commercially pure aluminium (≥99.5% Al), and Al5754 is an Al–Mg (3.6 wt% Mg) wrought alloy of higher strength. Reproduced from [104], copyright 2014 Archives of Foundry Engineering.

A practically important finding is that the beneficial effect of hard particle additions on adhesion strength depends on the failure mode transition: once interfacial adhesion strength is sufficiently strong, the failure site shifts into the deposit interior, indicating that cohesive rather than adhesive strength becomes the limiting property for deposit integrity.

6.2. Hardness and Work Hardening Effects

Cold spray deposits typically exhibit increased hardness relative to the feedstock powder owing to severe plastic deformation imposed by particle impact [36]. The mechanism is strain hardening through dislocation multiplication, which raises the flow stress throughout the deformed volume, so hardening is distributed through the deposit’s thickness rather than confined to a near-surface layer. Hardening remains heterogeneous at the particle scale, however, and is most pronounced at splat peripheries where deformation localises. A distinct contribution arises in the substrate, where peening by impacting particles produces a work-hardened near-surface region together with the compressive residual stress discussed in Section 6.3. While work hardening raises static strength, it reduces ductility and fracture toughness relative to wrought material, a trade-off that is unfavourable where resistance to crack propagation is required [107,117]. Homogeneous deposits provide direct evidence of deformation-induced hardening, as no hard phase contributes to strengthening. Eftekhari and Jahed [118] reported 107 ± 3 HV0.1 for cold-sprayed pure copper, approximately twice the hardness of annealed copper, and confirmed the mechanism through electron backscatter diffraction, observing grain refinement and dynamic recrystallisation concentrated near the substrate’s interface where the accumulated strain was greatest. In composite deposits, the deformation and reinforcement contributions are superimposed and cannot be separated by hardness measurement alone.

Within composite systems, matrix chemistry is the stronger determinant of hardness: Dayı et al. [44] reported 239 ± 23 HV0.1 for Ni–Al₂O₃ on AA7075, appreciably above the Al and Zn matrix composites deposited with the same reinforcement (Table 6), indicating that the matrix alloy rather than the reinforcement sets the attainable hardness level.

Shorinov et al. [35] reported a 32.7% increase in hardness upon adding Al₂O₃ to aluminium sprayed at 450 °C, which they attributed jointly to ceramic strengthening and enhanced peening of the matrix. Ductile reinforcement achieves comparable gains without introducing a brittle phase, as Li et al. [91] demonstrated for Al–Ta relative to Al–Al₂O₃, and post-deposition peening offers a further route; Zhang et al. [119] reported approximately 16% higher hardness for shot-peening-assisted cold spray, although the added process complexity limits field adoption.

Table 6. Summary of microhardness values for portable LPCS metallic and composite deposits. Indentation load and measurement location differ between studies, so values are not directly comparable.

Powder/Substrate	Hardness (HV)	Load (kg)	Location	Process Conditions	Ref.
Ni + Al2O3 on AA7075	239 ± 23	HV0.1	Cross-section	Air, 0.8 MPa, 300 °C, 8 mm SoD,	[44]
Pure Al on Al 1100	196.6	HV0.01	Deposit surface	N2, 0.6 MPa, 300 °C 15 mm SoD	[96]
Cu (satellite-free GA) on Cu	107 ± 3	HV0.1	Cross-section	N2, 1.99 MPa, 400 °C	[118]
Cu + 30% Al2O3 on Al	162.3	HV0.2	Cross-section, 7 locations	N2, 0.9 MPa, 400 °C	[56]
AA2024 + Al2O3 on AZ31B Mg	140	HV0.03	Deposit cross-section	Air, 500 °C, 20 g/min, 10 mm SoD	[120]
Cu+Al2O3 (50/50 wt%) on 304L	100	HV0.1	Cross-section (near-surface, middle, interface)	N2, 0.9 MPa, 600 °C, 10 mm SoD	[69]
AlMg6+C60 + 10% AlN on 1008 steel	259 ± 10	HV0.025	Deposit cross-section	Air, 0.8 MPa, 450 °C, 10 mm SoD	[121]
AlMg6+C60 + 50% AlN on 1008 steel	306 ± 12	HV0.025	Deposit cross-section	Air, 0.8 MPa, 450 °C, 10 mm SoD
TiO2 on AA1350	588 ± 124	HV0.01	Local sintered agglomerates	N2, 0.5 MPa, 600 °C, 10 mm SoD	[122]
TiO2 on mild steel S235 GJ	628 ± 79	HV0.01	Local sintered agglomerates	N2, 0.5 MPa, 600 °C, 10 mm SoD

Table 6. Summary of microhardness values for portable LPCS metallic and composite deposits. Indentation load and measurement location differ between studies, so values are not directly comparable.

Powder/Substrate	Hardness (HV)	Load (kg)	Location	Process Conditions	Ref.
Ni + Al2O3 on AA7075	239 ± 23	HV0.1	Cross-section	Air, 0.8 MPa, 300 °C, 8 mm SoD,	[44]
Pure Al on Al 1100	196.6	HV0.01	Deposit surface	N2, 0.6 MPa, 300 °C 15 mm SoD	[96]
Cu (satellite-free GA) on Cu	107 ± 3	HV0.1	Cross-section	N2, 1.99 MPa, 400 °C	[118]
Cu + 30% Al2O3 on Al	162.3	HV0.2	Cross-section, 7 locations	N2, 0.9 MPa, 400 °C	[56]
AA2024 + Al2O3 on AZ31B Mg	140	HV0.03	Deposit cross-section	Air, 500 °C, 20 g/min, 10 mm SoD	[120]
Cu+Al2O3 (50/50 wt%) on 304L	100	HV0.1	Cross-section (near-surface, middle, interface)	N2, 0.9 MPa, 600 °C, 10 mm SoD	[69]
AlMg6+C60 + 10% AlN on 1008 steel	259 ± 10	HV0.025	Deposit cross-section	Air, 0.8 MPa, 450 °C, 10 mm SoD	[121]
AlMg6+C60 + 50% AlN on 1008 steel	306 ± 12	HV0.025	Deposit cross-section	Air, 0.8 MPa, 450 °C, 10 mm SoD
TiO2 on AA1350	588 ± 124	HV0.01	Local sintered agglomerates	N2, 0.5 MPa, 600 °C, 10 mm SoD	[122]
TiO2 on mild steel S235 GJ	628 ± 79	HV0.01	Local sintered agglomerates	N2, 0.5 MPa, 600 °C, 10 mm SoD

Table 6 compiles reported microhardness for portable LPCS deposits. The values span 100 to 239 HV for the metallic and composite systems but cannot be compared directly across studies because indentation loads range from HV0.01 to HV0.2 and measurement locations vary between deposit surface, cross-section, and interface-adjacent positions. The consequence is visible within the table: the 196.6 HV0.01 recorded for pure aluminium exceeds values reported for aluminium composite deposits measured at higher loads, which reflects the surface location and light indentation rather than a harder material, while the highest entry corresponds to local sintered TiO₂ agglomerates rather than bulk deposit response. Hardness is therefore a useful indicator of deformation state and reinforcement effectiveness within a single study, but it cannot support acceptance criteria or cross-system comparison without a standardised measurement protocol.

6.3. Residual Stress

Cold-sprayed aluminium alloys develop compressive residual stress, arising chiefly from the peening effect of successive particle impacts, in contrast to thermal spray and fusion-based processes, which generate tensile stress through solidification quenching. Higher-strength alloys such as titanium can nevertheless develop tensile deposit stress through quenching and thermal mismatch. Three mechanisms set the final stress state: (a) peening stress arising from high-velocity particle impacts that plastically deform the substrate surface; (b) quenching-analogous stresses arising from constrained relaxation of the rapidly deformed splat; and (c) thermal mismatch stresses resulting from coefficient of thermal expansion (CTE) differences between the deposit and the substrate [123,124,125]. The three contributions superimpose to give the final stress profile. In lower-strength metals such as aluminium, the peening contribution dominates the residual stress state, whereas in higher-strength alloys sprayed at elevated pressures the quenching contribution is significant [124,126].

Neutron diffraction measurements by Luzin et al. [125] confirmed that both copper and aluminium cold spray deposits exhibit compressive residual stress throughout the deposit section regardless of spray parameters, with only the stress magnitude depending on material properties and process energy. Ryabchikov et al. [127] found good agreement between near-surface XRD and incremental hole drilling depth profiles. The stress turns tensile in the substrate to balance the compressive deformation imposed by particle impact, and this tensile field can drive fatigue crack initiation where the substrate or interface contains pre-existing defects.

Modelling studies have provided complementary insights. Boruah et al. [128] showed through a parametric study that the number of deposited layers and the substrate height both influence the magnitude of the through-thickness residual stress, whereas individual layer height has a negligible effect. Manap et al. [88], using smoothed particle hydrodynamics to simulate Al particle impacts on Al, Mg, and Ti substrates under LPCS and HPCS, suggested that LPCS produces lower residual stresses on aluminium substrates and minimises deformation, findings that still require systematic experimental validation before generalisation.

Zhang et al. [67] measured a compressive residual stress of about −48.3 MPa in AA2024–Al₂O₃ composite deposits on AA2024-T3 containing 20 wt% Al₂O₃, the composition giving the best balance of hardness, adhesion, and corrosion resistance. Luzin et al. [129] used neutron diffraction to evaluate through-thickness residual stress in Al–Al₂O₃ and Al–Zn–Al₂O₃ composite LPCS deposits (Figure 18). The profiles combine mechanical peening and thermal–elastic contributions, with a compressive peak at the deposit midpoint falling to near-zero or slightly tensile at the free surface.

Figure 18. Residual stress profiles for Al and Al–Zn LPCS deposits calculated for a 4 mm deposit thickness. The zero point on the x-axis is the deposit–substrate interface. Thicker lines show the combined deposition and thermal contributions. Adapted from [129].

Zhang et al. [84] reported compressive residual stresses in the range of −32.1 to −58.0 MPa for different Al alloy deposit compositions, as shown in Figure 19. The variation in stress magnitude was attributed to alloy resistance to plastic deformation rather than differences in the process conditions.

Figure 19. XRD residual stress measurements at the surface of LPCS deposits for commercially pure Al and different Al alloy compositions showing compressive stresses in the range of −32 to −58 MPa. Replotted from [84].

Reed et al. [130] measured residual stress in cold-sprayed AA7075 repairs through three techniques (Figure 20). Pulstec μ360 XRD gave a uniform compressive surface stress of about 30 MPa in both the x- and y-directions. Incremental hole drilling showed this stress to be non-uniform through depth, peaking 0.24 mm below the surface, the subsurface maximum reflecting plastic deformation by particle impact with slight near-surface relaxation, then decaying to an average of 7 MPa at 1.02 mm. The contour method, evaluated at three positions along the plate, agreed closely: compressive at the surface (14 MPa average), rising to a maximum at 0.27 mm, and turning tensile at an average depth of 1.47 mm where the substrate balances the deposit. The distribution is typical of aluminium cold spray and is governed by the peening mechanism described in [124]. The portable Pulstec μ360 XRD used in [130] is in principle suitable for on-site measurement of field repairs, although such measurements are rarely reported.

Figure 20. Residual stress in a cold-sprayed AA7075 repair measured through three different techniques: (a) XRD contour plot on CS top surface showing the x direction stress (the y direction stress is the same), (b) through the depth stress distribution in CS coating up to 1 mm, both the x and y directions, obtained using incremental hole drilling method, and (c) through the depth distribution in CS and substrate assembly, obtained using the contour method. Reproduced from [130], copyright 2025 Elsevier.

6.4. Fatigue and Tensile Strength Behaviour

Fatigue damage accounts for a large proportion of in-service failures, so whether a repair can sustain a comparable cyclic regime determines its structural integrity. It is also the least studied aspect of portable LPCS and MPCS repair. One limitation must be stated at the outset: most fatigue data for cold-spray-repaired components come from HPCS, which operates at a substantially higher velocity ratio than field-deployed LPCS [32,33]. Whether crack initiation behaviour, fatigue endurance limit, and interface failure mechanisms from HPCS studies can be transferred to portable LPCS deposits, where bond continuity is less consistently achieved, remains open. This unresolved question in the literature is flagged wherever HPCS data are cited. Several studies have found fatigue cracks in cold-spray-repaired specimens initiated from the deposit–substrate interface, mainly from small craters on the substrate produced by particle impact, as shown in Figure 21 [29,131,132,133]. This reframes the deposit integrity problem: bulk deposit properties are necessary but not sufficient for the durability of a repaired part.

Figure 21. SEM images of fracture surfaces of repaired AA7075 specimen showing arrows pointing to craters (yellow dash line) or crack initiation and interface for groove repaired (a,b) and dent repaired (c,d). Reproduced from [29], copyright 2025 Elsevier.

Yandouzi et al. [81] reported fatigue performance of repaired specimens comparable with that of the substrate for commercially pure aluminium deposited on Alclad 2024-T3 under strain-controlled testing (strain ratio R = −1, test frequency 20 Hz). The strain life data (Figure 22) scatter widely, with the fitted curve being an upper bound indicating a fatigue limit corresponding to a strain range of approximately 1.82 × 10⁻³. As one of very few portable LPCS fatigue datasets in the open literature, it indicates that geometry-appropriate LPCS can preserve substrate-level fatigue performance consistent with the compressive residual stress field offsetting the stress concentration at the repair boundary.

Figure 22. Strain life diagram for LPCS-repaired commercially pure Al on Alclad 2024-T3 evaluated under strain-controlled cycling (R = −1, 20 Hz) according to ASTM E739-91. Open circles are failed specimens; arrows denote run-outs. The dashed line shows the fitted fatigue curve. Replotted from [81].

Peng et al. [134] reported in their Phase I durability study that post-deposition annealing substantially reduced fatigue life relative to unannealed repairs, attributed to residual stress relaxation removing the beneficial effect of compressive residual stress. Bakir et al. [29] showed that groove-type repairs to AA7075 can match the fatigue strength of undamaged material at the maximum stress of 200–250 MPa (R = 0.1), provided repair geometry is well-controlled and post-deposition annealing avoided. Both results used HPCS, and their transfer to portable LPCS has not been established directly.

White et al. [135] found that for AA2024-T351 blend-out repairs of fastener hole damage, fatigue life exceeded that of undamaged specimens for repairs placed both inside and outside of the lap-shear configuration, confirming the benefit of the compressive residual stress introduced during material restoration by cold spray. Cavaliere and Silvello [133] found through finite element modelling and experiments that cold spray repair of V-notch damage in AA2099 panels markedly reduced the stress intensity factor, raising crack resistance more than sevenfold, and shifted the maximum stress concentration from the original crack tip to the repaired panel surface, demonstrating that repair geometry can govern the crack initiation site.

The fatigue failure mechanism in cold-spray-repaired specimens is summarised in Figure 23. Two independent studies by Peng et al. [136] and Bakir et al. [29] showed the initiation of fatigue cracks from the aluminium substrate and their subsequent propagation into the substrate. The crack propagation phase in both studies can be seen by the distinct “beach marks” in the substrates. In both studies, the cold spray deposits did not crack until late in the specimen’s life.

Figure 23. Fatigue fracture surfaces showing crack initiation from small craters or dents caused by particle impact during cold spray process: (a) specimen dimension, (b) crack initiation at the substrate interface with lower repair patch, (c) crack initiation from the substrate with upper repair patch, and (d) SEM image of lead crack. Reproduced from [136], copyright 2021 MDPI.

Bakir et al. [29] identified that crack initiation was from very small “craters” on the substrate interfacing with the deposit (Figure 21). These craters were created by particle impacts at supersonic speed during deposition.

Tensile data for repaired assemblies under in-plane loading are also sparse. Dayi et al. [44] examined an LPCS repair of a 1 mm deep groove in an AA7075 substrate using five composite powders (Ni–Al₂O₃, Al–Zn–Al₂O₃, Al–Al₂O₃, Ni–Zn–Al₂O₃, Zn–Al₂O₃). Three specimens were tested per repair condition. Mean curves from the paper are replotted in Figure 24. The stress–strain curve of the undamaged AA7075 specimen showed the expected ductile material behaviour with a strain at fracture just over 14%. All five types of repaired specimens showed brittle fracture behaviour with strain at a fracture of 5%, about 65% lower than that of the substrate material. In terms of strength, the repair using the Ni–Zn–Al₂O₃ powder gave the best result, with a yield stress of 477 MPa and an ultimate tensile strength of 511 MPa, about 13% below the parent material AA7075. Across all five conditions, yield stress was reduced by roughly 10% relative to the parent alloy.

Figure 24. Engineering stress–strain curves for undamaged and LPCS-repaired AA7075 specimens produced using Ni–Zn–Al₂O₃, Al–Zn–Al₂O₃, Zn–Al₂O₃, Ni–Al₂O₃, and Al–Al₂O₃ powders. Each curve represents the mean response of three specimens and shows the influence of repair composition on strength recovery and ductility. Replotted from [44].

Cautious optimism regarding LPCS fatigue performance rests on the compressive residual stress state characteristic of cold spray deposits (Section 6.3), which retards crack initiation and slows stage I propagation through a substrate-governed rather than system-governed mechanism. Provided the interface achieves sufficient metallurgical continuity, an LPCS repair should follow the same fatigue framework as an HPCS repair, differing in magnitude rather than the underlying deformation physics. Table 7 consolidates the mechanical property data across the three pressure regimes.

Table 7. Reported mechanical properties for LPCS, MPCS, and HPCS deposits. Entries are drawn from different material systems and test protocols and may not be directly comparable in some cases.

Property	LPCS	MPCS	HPCS (Baseline)
Elastic modulus (GPa) (AA6061)	62 [107]	-	64 [107]; wrought substrate 69 GPa
Tensile fracture mode (AA6061)	Brittle quasi-cleavage across flat splat facets [107]	-	Ductile; dimples and tear ridges throughout [107]
Adhesion strength -ASTM C633 (MPa) (Al on Al Alloys)	15–50 (Table 5)	34–54 [17]	Higher than LPCS and MPCS (about 60 with N2 and 78 with Helium) [31]
Porosity	<2% (2D Slice, Al) [81]; up to 4.7 (Cu) [53]	0.4 (2D, Al–Al2O3) [17]	0.18 by 3D XCT (AA7075) [29]
Microhardness	100–239 HV, ductile matrix-dependent (Table 6)	70–80 HV0.1 (Al–Al2O3) [17]	140–150 HV0.1 (AA7075) [29]
Residual stress	−32 to −58 MPa, compressive [67,84]	Peening contribution expected to exceed LPCS (Section 7.3)	Can be compressive or tensile depending on the relative contributions of peening-induced compressive and quenching—and thermal-misfit-induced stresses [125,130]
Fatigue strength at endurance-level cycles	139 MPa, R = −1; substrate-level retention under strain control [81]	-	Comparable to undamaged material at 200–250 MPa, 20Hz, R = 0.1 [29]
Deposition efficiency (Al)	5–30% (air) (Table 2)	Expected between LPCS and HPCS (Section 7.3)	Up to 95% [29,31]

7. Substrate Geometry and Interface Design

7.1. Blend-Out Design and Aspect Ratio Effects

The term “blend-out” describes the preparation of a repair by removing material from the damaged area to create a geometry suited to deposition. Its length ratios, corner radius or taper angle, and transition smoothness control deposition quality and efficiency and the bonding of the deposit–substrate interface [131,134]. The repair patch aspect ratios, namely width-to-depth and repair-thickness-to-substrate-thickness, define the elliptical blend-out geometry commonly adopted in the literature (Figure 25). No agreed framework yet exists for the values of a/t and b/c within which a portable cold spray repair can be expected to deliver durable, damage-tolerant performance, irrespective of whether the originating damage is corrosion thinning, a dent, or surface scoring.

Figure 25. Geometry of an elliptical blend-out for repairing surface damage by cold spray. The blend-out is defined by an ellipse with major axis 2c and minor axis 2b. Repair depth a and substrate thickness t define the severity ratio a/t and the aspect ratio of the repair patch b/c.

Evidence from wing skin durability studies and fastener hole repair experiments indicates that blend-out depths of 15–30% of plate thickness, combined with smooth taper transitions, can restore fatigue life to near-undamaged levels [135]. Optimal blend-out geometry depends on the applied stress field, the elastic modulus of the repair material relative to the substrate, and the residual stress profile at the interface. Parameters may vary between LPCS and MPCS deposits even for identical material systems.

Blend-out depths reported across cold spray repair studies range from 15% to 50% of substrate thickness, as summarised in Table 8. Figure 26 shows representative repair geometries presented in the literature.

Table 8. Summary of cold spray repair studies that provided repair geometry and blend-out design. Blend-out ratio is repair patch depth to substrate thickness (a/t). Studies differ in blend-out profile, loading configuration, and specimen geometry. The a/t ratio is reported as a common descriptor and does not capture profile curvature.

Study	Key Finding	Blend-Out Geometry Details	a/t	Substrate	Ref.
Fatigue performance of cold spray repaired AA 7075	Geometry shifts crack initiation site; groove moves crack away from repair edge	Groove and dent; through-width; dent 4 mm wide; 1 mm blend-out	0.20	AA7075-T7351	[29]
On the role of the interface on the damage tolerance and durability of cold spray repairs to AA7075-T7351	Blend-out geometry governs crack nucleation at deposit–substrate interface	Corrosion cut; through-width cut 15 × 17 mm2; 0.675 mm blend-out each side	0.11	AA7075-T7351	[131]
Experimental studies into the analysis required for the durability assessment of 7075 and 6061 cold spray repairs to military aircraft	Stress state governs failure sequence; deposits below ~150 MPa do not crack independently	Through-width centre 50 × 42 mm2; 0.2 mm deep notch	0.15	7075-T6 plate	[132]
Fixing a hole with cold spray	Circular cavity yields best fatigue life; geometry selection informed by stress analysis	Three cavity profiles (trapezoidal, circular, conical); 4 mm blend-out for circular	0.33	A357-T6 Al casting	[55]
Fatigue behaviour of fastener holes in high-strength aluminium plates repaired by cold spray deposition	Blend-out depth and surface placement affect fatigue; repair exceeded undamaged life at optimum	Fastener hole sizes (1/8, ½); depths of 0.79, 1.59, and 3.18 mm around fastener hole	0.12–0.5	AA7075-T651; AA2024-T351	[135]
Cold spray for aircraft structural repair of corrosion damage	30% blend-out with smooth taper showed superior fatigue in screening trials	15% and 30% of 6.35 mm thickness	0.15 & 0.30	AA7075-T7451; AA7050-T7451	[137]
Residual stresses induced by Inconel 718 cold gas repair spray of large cavities- on the influence of the sample geometry	Curvature introduces spray angle and stand-off variability; confirmed by neutron diffraction	Cavity on a cylindrical curved surface; 6 mm wide, 4 mm deep cavity at 60° on shaft	0.1	Inconel 718 shaft	[138]

Figure 26. Geometries of repair specimens: (a) cylindrical samples and their coordinate system, Reproduced from [138] copyright 2025 Elsevier; (b) groove or dent repairs under axial tension, Reproduced from [29], copyright 2025 Elsevier.

7.2. Spray Angle Constraints and Their Effect on Interface Quality

Spray angle has the most direct and quantifiable effect on interfacial porosity distribution within a repair geometry. Petráčková et al. [55] showed that trapezoidal (Kt = 3) and conical (Kt = 2) cavities, which impose oblique spray angles on the sidewalls (Figure 27), produced localised porosity of 43% and 12%, respectively, in the highest-stress interfacial regions, precisely where adhesion failure and fatigue crack initiation are most likely, whereas a spherical cavity holding near-normal incidence throughout reduced porosity to a uniform 1%. The prepared cavity profile is therefore not merely a geometric variable but a determinant of interfacial quality, and uniform deposit integrity cannot be assumed across non-planar repairs.

Figure 27. Schematic cross-sections of trapezoidal, conical, and circular cavity geometries prepared for cold spray repair. Sidewall inclination governs the local spray angle achievable during deposition, with the circular profile maintaining near-normal incidence across the cavity’s surface while the trapezoidal and conical profiles impose increasingly oblique impact on the sidewalls. Adapted from [55].

The depth, sidewall inclination, and transition radius of the damage profile directly determine the spray angles achievable during deposition. On open flat repairs, the nozzle can be held at or near 90° (Figure 28); in recessed grooves, holes, and complex curvature features, it must approach obliquely over at least part of the surface, with the deviation growing as the damage becomes deeper or narrower.

Figure 28. Substrate tilt configuration for the spray angle study; substrates inclined at 15° and 35° relative to the base plane at constant stand-off distance. Adapted [139].

Blochet et al. [139] investigated the effect of spray angle on aluminium deposition onto AA2024-T3, both experimentally and through finite element analysis, using the tilt configuration in Figure 28. Pure Al particles (11–52 μm, nitrogen-atomised) were sprayed between 30° and 90° with a KINETICS 3000-M system at 3 MPa and 300 °C. Two findings bear directly on portable repair design: below a material- and roughness-dependent critical angle θc (Figure 29a,b), all particles rebounded, giving zero deposition efficiency; bond strength, measured through a laser shock adhesion test (LASAT), degraded substantially and continuously between 60° and 30°. The relationship between prepared surface profile and achievable nozzle angle is therefore critical for manual on-site spraying, where an operator cannot hold near-normal incidence across the whole repair.

Figure 29. Schematic illustration of powder sprayed onto a substrate tilted at different angles from 90 to 30 relative to a fixed traversing nozzle: (a) top view; (b) cross-section view adapted from [139].

Spraying above 70° was identified as the threshold for obtaining dense deposits with adequate bond strength in their work (3 MPa, 300 °C); applying this threshold directly to LPCS, where particle velocities are substantially lower, requires caution and has not been independently validated. The authors concluded that near-perpendicular incidence is the only reliable condition for high-integrity deposition.

Li et al. [140] showed that oblique impact generates a tangential velocity component at the contact zone that is absent at normal incidence. As the spray angle falls below 90°, this component grows relative to the normal one, driving the deforming particle laterally across the substrate instead of into it, reducing the splat–substrate contact area and mechanical anchoring and leaving a trailing-edge gap that interrupts oxide disruption and metallurgical bonding. Binder et al. [99] confirmed experimentally that deviation from perpendicular incidence increases deposit porosity and degrades mechanical properties, with degradation becoming significant beyond approximately 45° deviation from normal incidence for titanium deposits, a finding consistent with the aluminium results [139].

These findings place a quantitative constraint on acceptable blend-out geometries for portable repair. Sidewalls imposing spray angles below about 60° on any part of the repair area present a deposit quality problem that process parameter optimisation alone cannot correct. The practical responses are to machine a shallower taper before repair, use a multi-pass robotic strategy that rotates the nozzle to keep near-perpendicular incidence at each surface element, or accept inferior sidewall bond integrity and substantiate the structure accordingly.

Yin et al. [141] also noted that substrate temperature rises non-uniformly under oblique spraying due to the asymmetric heat input distribution, which introduces additional variability in particle deformation behaviour across the repair surface area. In combination, these effects mean that substrate sidewall geometry is a primary deposit process design variable, not merely a manufacturing consideration.

7.3. Process Regime Selection as a Deposit Integrity Decision

Process selection should follow from the specific repair geometry. For shallow, broad damage where the depth-to-thickness ratio is small and the substrate retains an adequate residual section, LPCS is sufficient provided bond continuity can be verified and the blend-out minimises stress concentration [131]. Where damage is deep or volumetric and the deposit carries more of the applied load and the geometry is inherently notch-sensitive, the higher impact velocity of MPCS should reduce incomplete interparticle bonding. In the absence of experimental data, the magnitude of this advantage can be bounded using the gas dynamic model of Section 3.1.

In the analytical solution of Martínez-Flores et al.’s [32] Equation (7), particle velocity follows $v_p\left(t\right)=\gamma \tanh \left(\beta t\right)$, approaching a ceiling of $\gamma = {\displaystyle \frac{v_g}{\sqrt{C_d}}}$ at a rate set by the acceleration coefficient $\beta = {\displaystyle \frac{3}{4}} {\displaystyle \frac{{\rho }_g}{{\rho }_p}}{\displaystyle \frac{v_g\sqrt{C_d}}{d_p}}$. The two parameters respond differently to the process inputs. Because β scales with gas density and ${\rho }_g$ is proportional to $p_o$ at constant stagnation temperature, raising stagnation pressure from below 1 MPa to the 1–3 MPa MPCS range increases β approximately in proportion, so a greater fraction of $\gamma$ is attained within the same nozzle length. The ceiling itself is unaffected by pressure, being set by gas velocity alone. Raising stagnation temperature acts on the other parameter, $v_g\infty \surd \left(\kappa R{T}_0\right)$, so increasing $T_0$ from 600 to 800 °C raises $\gamma$ by approximately 11% while simultaneously reducing gas density and hence β by approximately 10%, as β is proportional to p₀/√T₀. Pressure and temperature therefore act on different terms and partially oppose one another, so the net gain depends on the pressure–temperature combination selected and is not set by pressure alone.

Equation (6) can be evaluated for commercially pure aluminium. Equation (2) gives a critical velocity of approximately 620 m/s for a moderately preheated particle derived from bulk material properties. Wiehler et al. [43] measured an effective yield strength for gas-atomised aluminium approximately 81% above the bulk value, which enters the strength term of Equation (3) and places the critical velocity above 620 m/s. The LPCS particle velocity range of 300–600 m/s therefore corresponds to η at or below unity for a homogeneous metallic feedstock. This is the basis of the dependence of LPCS on composite feedstocks, in which ceramic addition supplies surface activation and densification that velocity alone does not achieve (section effect of ceramic particle size on deposit microstructure). At the upper MPCS particle velocity of 800 m/s quoted by equipment manufacturers [16], the velocity ratio η exceeds unity by a small margin. MPCS therefore operates within the stable deposition regime for pure aluminium, though still below the η ~ 1.5 optimum identified in Section 3.1. For Ti-6Al-4V and Ni-base feedstocks, whose critical velocities are higher, the margin becomes smaller.

Precipitation-hardened alloys reduce the margin further. In Equation (3), the strength term dominates: the flow stress of the 2xxx and 7xxx alloys exceeds that of annealed pure aluminium by roughly an order of magnitude, while the thermal term acts in the opposite direction, as the solidus of these alloys lies below the melting point of pure aluminium and reduces $\left(Tm-Ti\right)$. The net effect is a substantially higher critical velocity. Portable systems nevertheless deposit these alloys, as reported for AA2024 and AA7075 [67,84,102], through three compensating measures: ceramic addition, elevated gas temperature, which raises the particle impact temperature Ti and further reduces (Tm − Ti), and nitrogen propellant in place of air. Table 3 shows that pure Al and Al composite feedstocks predominate, which reflects the extent of compensation required.

Table 8 records the repair geometries used in the blend-out studies. Lang et al. [138] deposited into a cavity on an Inconel 718 cylindrical shaft and found that curvature introduces spray angle and stand-off variability, degrading interface continuity relative to flat coupon testing. Component curvature is therefore a variable influencing the effective process window, alongside blend-out slope and depth.

8. Deposit Qualification and Process Validation Framework

8.1. Regulatory Standards

The standard framework for cold spray repair has been developed largely within the US military. MIL-STD-3021, issued by the US Army Research Laboratory in 2008 and amended in 2015 and 2019, sets process control requirements covering equipment qualification, operator certification, parameter documentation, and inspection [142]. Appendix A, Table A.1 of MIL-STD-3021 covers structural and safety-critical applications and requires a qualification plan approved by the Cognizant Engineering Authority and proof testing appropriate to the intended loading. It does not specify mechanical performance acceptance criteria for individual repairs; that responsibility rests with the engineering authority overseeing each qualified process.

SAE Aerospace Material Specification AMS 7057, co-developed with the SAE International Additive Manufacturing Committee and published in 2024, represents the first cold spray process specification developed under commercial consensus standard processes [143]. Covering both additive manufacturing and repair, it signals a shift from purely military qualification to civil aerospace engagement. Neither the Federal Aviation Administration (FAA) nor the European Union Aviation Safety Agency (EASA) has yet published cold-spray-specific guidance, leaving civil primary structure in a regulatory gap resembling that of metal additive manufacturing a decade ago.

USAF Structures Bulletin EZ-SB-19-01 provides engineering guidance for cold spray substantiation within a damage tolerance assessment framework [144], addressing the requirement to demonstrate that the repaired structure meets the same crack growth and residual strength requirements as the original design. This bulletin is operationally the most relevant qualification guidance for structural cold spray repairs to USAF aircraft, and its principles may be applicable to LPCS and MPCS repairs when the repair process can be shown to produce interface integrity consistent with the accepted damage tolerance model.

8.2. Operational Deployment Records

The deployment record, which covers LPCS and HPCS rather than MPCS, gives indirect evidence of readiness at the system level and of the scale of remaining certification work (Table 9). Corpus Christi Army Depot has restored UH-60 Black Hawk magnesium gearbox housings through cold spray since about 2012, with documented savings across the US Department of Defence (DoD) repair portfolio [109]. The US Navy has approved repair processes at Fleet Readiness Center Southwest encompassing multiple aircraft components, with on-aircraft structural repair of V-22 Osprey windowsill structure demonstrated [145]. The United States Air Force (USAF) 7th Equipment Maintenance Squadron has deployed portable cold spray capability for aircraft outer-panel repairs [9]. These deployments confirm that portable cold spray is operationally mature at the system level, but almost all approved applications are classified as non-primary structures under existing regulatory frameworks; deposit integrity and qualification for load-bearing, safety-critical structures remain open research questions.

Table 9. Examples of field-deployable portable and mobile cold spray repair applications in aerospace, defence, and other industries. Entries drawn from manufacturer and operator reports have not been independently peer reviewed.

Application	System	Deployment	Outcome/Significance	Ref.
V-22 Osprey window frame; H-1 skid tube	HPCS-VRC Metal Systems	Mobile robotic	On-aircraft structural repairs; reduced teardown; enhanced mission readiness	Wilburn [145]
hydraulic components repair (Hydraulex)	LPCS-Titomic D523	Portable	Corrosion-damaged actuators and valves restored to dimensional tolerance	R. Temporal [146]
Mg gearbox casing (military aircraft)	LPCS-custom portable	Portable	Cracked Mg housing restored; minimal thermal input; no distortion	Champagne et al. [109]
H-64 Apache rotor mast support	CenterLine SSM-P3300	Portable	Worn snap ring groove rebuilt with Al composite; part passed fatigue testing	Leyman & Champagne [10]
B-1B Lancer bomber bay panel	HPCS-VRC trailer system	Mobile (trailer)	Panel cracks repaired; ~USD 225k saved per panel versus replacement; integrity restored	J. Miller [147]
NAVSEA ship hull repair	Centreline SST LPCS	Portable	Corroded hull maintenance in situ; dry-docking avoided	Fuentes [148]
Steel valve disc (erosion–corrosion)	LPCS-BAE Systems trial	Portable	Bronze alloy deposited to rebuild disc; passed pressure tests	Pal et al. [45]
Wind turbine heat exchanger	LPCS-Titomic D523	Handheld, remote	Leak repaired in 2 h for under USD 500; USD 80k replacement avoided; aluminium layer deposited	Titomic [149]
Offshore gas platform repairs	LPCS-Titomic D523	Handheld, on-site	Failed deposit replaced with Al in situ; no hot work required	Titomic [150]
Motor bearing housing	LPCS-Titomic D523	Handheld, mobile	Worn bearing seat rebuilt in 1.5 h using ~ 500 g Ni powder; cost under USD 100	Mallee MS [151]
Vintage race car wheel hub	LPCS-Titomic D523	Portable, vehicle-mounted	Worn bearing carrier rebuilt on vehicle without heat; original component preserved	Titomic [152]
Slat track and fuel valve components, USAF	LP/MPCS-ES3 portable/robotic	Handheld/robotic	Dimensional restoration and corrosion mitigation on Al and Mg alloy aircraft components	ES3 [153]

9. Comparative Assessment and System Selection

Selection of a portable cold spray regime for a surface engineering or repair application requires an integrated assessment of trade-offs across process performance, logistics, and substrate geometry. Table 10 compares LPCS, MPCS, and HPCS at the level of individual capability attributes. The figures given are representative of commonly reported values for commercial systems and are indicative rather than prescriptive.

Table 10. Comparative assessment of LPCS, MPCS, and HPCS for surface engineering and repair applications. Values are indicative of commercially available systems and are not controlled comparisons.

Repair Attribute	LPCS Portable	MPCS Portable	HPCS (Baseline Reference)
Velocity ratio	Low: mostly for ductile metals [32,33]	Moderate: enables harder alloys compared with LPCS [16,33]	High: achievable for most engineering alloys [31,33]
Interfacial bonding reliability	Greater risk of incomplete bonding at non-ideal spray angles/stand-off [95,97,98]	Reduced risk relative to LPCS; still sensitive to oxide state [31]	Lowest interfacial defect population; highest bond strength [7,31]
Compressive residual stress	Moderate: beneficial for fatigue crack initiation resistance [125]	No available information to date in the peer-reviewed literature	Can be compressive or tensile depending on the relative contributions of peening-induced compressive and quenching—and thermal-misfit-induced [125]
Deposition efficiency (aluminium)	5–30% (air); 30–50% (N2 + Al2O3 composite) [35,53,154]	Not reported to date in the peer-reviewed literature	70–95% for optimised ductile metals [29,31]
Surface damage repairs	Sufficient with well-designed blend-out and verified interface continuity [81]	Preferred; higher velocity margin η reduces geometric sensitivity [16,29]	Suitable; may be over-specified for shallow repairs [28,29]
Field deployability	Excellent: single operator with shop air [13,26]	Good; can weigh up to four times more than LPCS owing to the larger power generator [16,25]	Poor: depot-level infrastructure required [28]
Qualification status	Qualified under MIL-STD-3021 for non-primary structure; process validation data available [102,142]	Emerging; AMS 7057 applicable; limited primary structure data [143]	Widest qualification precedent; most mature repair qualifications [64,142,155]

The velocity ratio quantifies the margin available to absorb spray angle deviation, stand-off drift, and oxide state variation in the field. LPCS operates closest to the critical velocity threshold, so any parameter perturbation disproportionately reduces bonding reliability. The accompanying deposition efficiency gap compounds this into structural risk through the multi-pass build-up consequences described in Section 4.2.

Field deployability reverses this ranking. LPCS is operable by a single technician on shop-compressed air, MPCS units can weigh some four times more because of their generator requirement, and current HPCS designs are not field-deployable at all. For forward-deployed repair, LPCS is therefore the only viable option, and the selection question is reduced to whether the intended repair falls within its qualification boundary.

10. Conclusions and Future Outlook

This review has established that portable low-pressure and medium-pressure cold spray must be treated as a distinct class of deposition system in terms of process physics and not as a portable variant of high-pressure cold spray. The conclusions below are drawn from the open literature, and the priorities that follow from them are set out in Section 10.2.

10.1. Conclusions

• Portable LPCS and MPCS repairs should be assessed as a distinct class in which process window control, deposit integrity, and repair geometry jointly determine structural suitability.
• LPCS offers clear advantages for field-deployable surface restoration, corrosion protection, and dimensional recovery, but its narrow velocity margin restricts the material set and makes deposit quality highly sensitive to gas temperature, pressure, powder morphology, stand-off distance, and spray angle.
• Bonding in LPCS deposits is predominantly localised at mechanically interlocked asperity contacts and is not distributed as metallurgical continuity, which limits strength and crack growth resistance relative to HPCS.
• MPCS offers a wider material envelope while retaining portability, but peer-reviewed data remain too sparse for robust comparison with LPCS or for confident structural qualification.
• Bulk deposit properties are necessary but not sufficient for repair durability, as fatigue cracks are initiated predominantly at the deposit–substrate interface across materials, geometries, and process regimes.
• Hard particle additions can improve deposition efficiency, densification, hardness, and adhesion, but excessive ceramic content or unsuitable particle size introduces brittle interfaces and fatigue-sensitive defects.
• Compressive residual stress from particle peening is among the strongest arguments for cold spray repair, because it may offset stress concentration at the repair boundary, though post-deposition heat treatment relaxes this benefit.
• Repair geometry is a controlling design variable rather than a preparation detail, with blend-out depth, taper ratio, and local spray angle determining whether adequate velocity ratio and interface continuity are maintained across the repair surface.
• Porosity acceptance criteria based on two-dimensional cross-sectional analysis are non-conservative, as 2D measurement underestimates volumetric porosity.
• The principal barrier to qualification for primary structures is the absence of geometry-sensitive fatigue data for portable systems, with the available evidence drawn predominantly from HPCS and simplified coupon geometries.

10.2. Future Outlook

The conclusions above identify where the evidence base is insufficient. The following priorities address those deficiencies:

• Geometry-sensitive fatigue substantiation: Testing on representative repair geometries should span the range of blend-out depths, taper ratios, and spray angles encountered in field repair, under load spectra relevant to the intended application, so that the interaction between geometry and interface continuity can be quantified.
• Controlled LPCS–MPCS comparison: Direct comparison is needed to establish when the additional pressure and temperature capability of MPCS justify its greater infrastructure requirement. Such a programme should vary process regime alone while holding feedstock, substrate, surface preparation, stand-off distance, spray angle, and repair geometry constant.
• MPCS property datasets: The measurements required, in priority order, are deposition efficiency, porosity, adhesion strength, residual stress, tensile behaviour, fatigue performance, and repair repeatability across commercially relevant alloys.
• Three-dimensional defect characterisation: Standardised X-ray computed microtomography protocols at sufficient spatial resolution are needed to underpin porosity acceptance criteria for load-bearing repairs, alongside field-deployable non-destructive methods, including ultrasonic phased-array inspection, for in-service assessment.
• Combined fatigue and environmental durability: The interaction between corrosion pitting, moisture ingress through porous deposits and galvanic effects at dissimilar material interfaces, and fatigue crack growth at the deposit–substrate interface, has not been investigated for portable repair geometries under service-representative conditions.
• Process window under field environmental conditions: The effects of ambient temperature outside of manufacturer-recommended ranges, relative humidity, salt contamination, and substrate temperature variability on bonding quality and deposit integrity remain uncharacterised, despite their direct relevance to maritime, arctic, and tropical deployment.
• Standardised reporting: Powder morphology, ceramic content and particle size, substrate roughness and oxide state, porosity measurement method, residual stress depth profile, and failure location should be reported as standard. The absence of these parameters obstructs synthesis of existing results and the generation of open, comparable datasets across multiple portable systems and repair geometries. This is a prerequisite for both structural qualification and the predictive process models that would depend on them.