The hybrid corrosion mechanism of titanium in mixed HF-HNO 3 solution during laser engraving/chemical milling manufacturing

Abstract

Laser engraving / chemical milling is a hybrid manufacturing technology which is suitable for the fabrication of the key components in aerospace industries. Here, we originally proposed a molecular dynamics model to study the chemical reactions and corrosion mechanism between TA15 titanium alloy and the mixed HF-HNO₃ solution. Then, experiments on corrosions are carried out to optimise the mixing ratio of the mixed HF-HNO₃ solution. Results indicate that the corrosion mechanism in chemical milling is governed by HF-induced activation-driven dissolution and HNO₃-mediated oxidation-induced passivation. The increase of HF ratio in the mixed solution can lead to higher surface roughness and hydrogen absorption. The increase of HNO₃ ratio in the mixed solution can lead to lower corrosion rates. A formulation is obtained to reveal the relationship between corrosion rate and concentration of HF and HNO₃. By consideration of the corrosion rate, surface roughness and hydrogen embrittlement risk, the optimal mixed ratio is 8% HF concentration and 14% HNO₃ concentration with corrosion rate of 3.405 g·m⁻²·s⁻¹. The proposed model and the summarised formulation can provide theoretical basis and practical process parameters for chemical milling and surface treatment of TA15 titanium alloy.

Keywords

acidic corrosion titanium alloys corrosion modelling corrosion rate

Introduction

Chemical milling has emerged as a critical manufacturing technology in aerospace and military sectors for processing thin-walled, complex-shaped or large-area titanium alloy components, as it eliminates tool-induced stresses and deformation, ensuring structural integrity.^1,2 This non-contact method employs controlled chemical dissolution, typically using acidic solutions containing hydrofluoric and nitric acids with specific additives, to selectively remove material without mechanical force. Titanium alloys exhibit a combination of highly desirable properties, including low-density, high-specific strength, excellent thermal strength, strong corrosion resistance, superior low-temperature performance and good processability. These advantages enable their widespread applications in aviation, aerospace, petroleum, chemical, marine and medical industries.^3–7 However, titanium alloys exhibit high deformation resistance and susceptibility to fracture during cold machining, making chemical milling a preferred process for material removal and complex-shaped workpieces. Owing to its unique advantages in weight reduction and complex structure formation, chemical milling has become an indispensable technology in aerospace and military manufacturing.^8,9

The corrosion dissolution rate during chemical milling of titanium alloys is highly dependent on the composition, proportions and formulation of the chemical milling solution. Precise control of these parameters enables stable and uniform material removal rates. However, in actual production, ideal chemical milling conditions are difficult to sustain. Instability in mixed acid ratios can cause fluctuations in corrosion rates or trigger hydrogen embrittlement risks induced by hydrogen absorption. Therefore, deepening the understanding of the corrosion dissolution mechanism in titanium alloy chemical milling is crucial for achieving precise control of the corrosion process.¹⁰

Regarding the corrosion dissolution of titanium alloys, numerous scholars have conducted extensive research. These studies encompass the corrosion behaviour of titanium alloys in various acidic media, including nitric acid,¹¹ sulphuric acid,¹² hydrochloric acid,¹³ hydrofluoric acid¹⁴ and multiple organic acid systems.^15–17 Li et al.¹⁸ revealed the corrosion behaviour of pure titanium in H₂SO₄, HCl and HNO₃ solutions containing varying concentrations of F⁻. Their findings revealed that pure titanium exhibited the most severe corrosion in fluoride-containing HNO₃. Wang et al.¹⁹ studied the corrosion behaviour of pure titanium in a 0.05 M H₂SO₄ solution containing fluoride ions. They observed that titanium spontaneously passivated in solutions containing 0.0005 M and 0 M fluoride ions, the latter of which accelerated corrosion. Wei et al.²⁰ investigated the corrosion and passivation behaviour of Ti-3Mo titanium alloy in 20 wt% hydrochloric acid solutions at different annealing temperatures. Results indicated that the corrosion rate increased with rising annealing temperature. Chen et al.²¹ investigated the difference in repassivation rates of titanium in 3.5 wt% NaCl and 6 M HNO₃. Findings revealed that titanium exhibited poorer corrosion resistance in 6 M HNO₃ compared to 3.5 wt% NaCl. Wu et al.²² investigated the electrochemical behaviour and passivation film properties of selective laser-melted and forged TA15 alloy in H₂SO₄ solutions. The study indicated that the passivation film formed on the SLM TA15 alloy surface effectively protected the alloy, enhancing its corrosion resistance in H₂SO₄ solutions. Liu's team²³ investigated the influence of Zn²⁺ on Ti corrosion behaviour in H₂SO₄ and F⁻ solutions. Mass loss results demonstrated that the presence of Zn²⁺ in corrosive solutions accelerated Ti corrosion.

Experimental studies are crucial for validating corrosion mechanisms, but corrosion processes in real engineering environments often occur over extended time scales, posing significant challenges for direct observation. Therefore, computer-based simulation techniques hold substantial value for investigating corrosion mechanisms. In recent years, researchers have employed molecular dynamics simulations^24–26 to explore metal corrosion mechanisms at the microscopic scale.^27–30 Chen et al.³¹ revealed the degradation behaviour of Fe₂O₃ passivation films in hydrochloric acid solutions at the atomic scale, discovering that residual stresses and partial relaxation of deformation during corrosion contribute to reduced corrosion rates. Balasooriya et al.³² analysed defect-driven corrosion mechanisms on steel surfaces, investigating the formation of protective layers versus reactive oxide layers under varying oxygen coverage. Chen et al.³³ elucidated the atomic mechanism of Cl⁻-induced Cu₂O corrosion using machine learning potential functions, indicating extensive chemical adsorption of Cl⁻ on the Cu₂O(111) plane, which significantly reduces the dissolution energy barrier for lattice copper. Zr-based amorphous coatings were prepared on TA2 substrates via plasma spraying by Jin et al.³⁴ It was found that the increased diffusion coefficient of oxygen atoms in the ZrO₂ passivation film, compared to its TiO₂ counterpart, resulted in enhanced corrosion resistance over the TA2 substrate. Furthermore, numerous researchers have validated the corrosion inhibition effects of several metallic inhibitors using molecular dynamics simulations, including nicotinic acid hydrazide-grafted chitosan,³⁵ phosphate derivative K₂SrP₂O₇,³⁶ phytic acid,³⁷ water and alcoholic extracts of Mikania micrantha,³⁸ 3-acetylquinoline thiosemicarbazone³⁹ and 1-hexadecyl-2,3-dimethylimidazole hexafluorophosphate.⁴⁰

Recent studies have established critical findings regarding the specific roles of acidic components in chemical milling: hydrofluoric acid acts as the primary etching agent, where highly electronegative F⁻ ions easily form soluble complexes to accelerate active dissolution, whereas oxidising agents like nitric acid promote repassivation by facilitating the rapid formation of a dense TiO₂ barrier layer.

Despite these macroscopic empirical findings, a critical research gap persists. Existing studies on titanium alloy dissolution focus almost exclusively on macroscopic experimental data, leaving the atomic-scale mechanisms largely unexplored. This limitation severely restricts the ability to regulate the dynamic competition between F⁻-induced activation and NO₃⁻-induced passivation at the microscopic level. Due to this lack of fundamental mechanistic insight and the inadequate regulation of component ratios, current chemical milling processes often exhibit significant fluctuations in corrosion rates and poor process controllability. Consequently, the high-precision fabrication of titanium alloy components while strictly preventing hydrogen embrittlement remains a huge challenge.

Regarding this issue, molecular dynamics method is proposed to construct a dynamic model of the titanium–mixed acid interface, revealing the dynamic microscopic mechanisms of corrosion reactions in nitric acid–hydrofluoric acid mixed media. Corrosion experiments are carried out to investigate the regulatory mechanisms of different HF-HNO₃ mixed acid ratios on the corrosion reaction of TA15 titanium alloy. This research provides data support and theoretical basis for the design of titanium alloy milling processes.

Experiment and simulation procedure

Experimental design

Annealed TA15 titanium alloy plates with a thickness of 6 mm were selected as the experimental material in this research. These plates were machined into multiple rectangular specimens measuring 60 mm × 15 mm. The chemical compositions of the TA15 titanium alloy employed in this research are presented in Table 1, which conform to the Chinese National Standard GB/T 3620.1-2016.⁴¹

Table 1.

Chemical compositions of TA15 (wt.%).⁴¹

Materials	Al	Si	V	Zr	Mo	Fe	Ti
TA15	5.5–7.1	0.15	0.8–2.5	1.5–2.5	0.5–2.0	0.25	Bal.

A series of weight loss experiments were conducted to systematically investigate the effect of different ratios of nitric acid–hydrofluoric acid mixed solutions on the corrosion rate of TA15 alloy. Prior to testing, the specimen surfaces were progressively polished using 400-grit, 800-grit, 1200-grit and 2000-grit wet sandpaper in a cross-hatch pattern. Subsequently, samples were polished using a 100-nm silica suspension until the surface exhibited a mirror-like finish with no visible marks. Cleaning involved sequential rinsing with distilled water and anhydrous ethanol, followed by degreasing with acetone. After air-drying with cold air, initial mass was measured on an analytical balance with 0.01 g precision. The acids used in the experiment included 50% nitric acid and hydrofluoric acid, which were mixed in a specific ratio to form an acid mixture. Due to the highly corrosive nature of hydrofluoric acid, all experimental operations were conducted in polytetrafluoroethylene containers.

The complete corrosion testing workflow is systematically illustrated in Figure 1. The procedure initiates with sample polishing and weighing, followed by static immersion in mixed acid solutions with varying HF and HNO₃ ratios. Post-corrosion, the specimens are subjected to systematic surface characterisations, including SEM morphology observation, X‑ray diffraction (XRD) phase detection and hydrogen content analysis, to comprehensively evaluate the corrosion behaviour.

Figure 1.

Corrosion testing flow chart.

The experimental strategy was systematically designed to first validate the molecular dynamics simulations and subsequently optimise the mixed acid formulation. Initially, an acid ratio screening was conducted by fixing the total H⁺ concentration, which effectively isolated the competitive effects of varying HF:HNO₃ volume ratios. Guided by the optimal volume ratio identified in this preliminary screening, progressive concentration gradient experiments were then performed. Ultimately, the final optimal formulation was determined by comprehensively evaluating the corrosion rate, surface roughness and the risk of hydrogen embrittlement.

In the experiment, mixed acid solutions of different concentrations were prepared with nitric acid and hydrofluoric acid ratios of 1:1, 1:2, 1:3, 2:1 and 3:1. To control variables, minimise experimental error and reduce the extent of the corrosion reaction, the hydrogen ion concentration of the prepared mixed acid solutions was fixed at 20%. During the experiment, specimens were immersed in the nitric acid–hydrofluoric acid mixture within each beaker for corrosion testing. At a specific time point (three minutes), the samples were carefully removed and subsequently subjected to a thorough rinsing process with distilled water followed by anhydrous ethanol. Afterwards, the samples were gently blotted with absorbent paper and then allowed to air-dry completely. Subsequently, the specimens were weighed on a precision electronic balance, and the weight loss at different acid concentrations and time intervals was recorded. This data was then used to calculate the corrosion rate. To ensure the reliability of the experimental data, six sets of corrosion tests were completed for each mixed acid solution ratio (each set consisting of three parallel experiments with repeated measurements).

Scanning electron microscopy can analyse the microstructural characteristics of corrosion products on the surface of corrosion specimens. Energy dispersive spectroscopy can analyse the elements and composition of micro-area corrosion products. Therefore, a Japanese SU8600 high-resolution field emission scanning electron microscope equipped with an Oxford UltimMax 100 energy dispersive X-ray spectrometer was employed to conduct scanning analysis of the corrosion specimens. This enabled us to determine the microstructural characteristics of the corrosion products, variations in elemental content and the morphology of the corrosion.

To conduct an in-depth analysis of the constituent phases within the specimens, XRD testing was performed. Copper target Kα radiation was used for the tests, which recorded XRD patterns within the 10°–90° range at a scanning speed of 2°/min. This enabled phase composition characterisation of samples with varying aluminium contents.

Surface roughness is one of the key parameters reflecting the surface quality of a specimen, typically used to characterise microscopic geometric shape errors on the specimen's surface. Surface roughness can be quantified by the arithmetic mean height (Sa) of the three-dimensional surface profile. A smaller Sa value indicates a smoother surface, while a larger Sa value indicates a rougher surface. This study employed a Zygo optical profilometer to scan the three-dimensional morphology of the specimen surfaces. Zygo is a non-contact optical interferometer, which provides more accurate imaging of porous corroded surfaces compared to contact profilometers. The final Sa value for each surface was determined by averaging the surface roughness values measured at three randomly selected locations on the sample surface.

To investigate the effect of different mixed acid ratios on material hydrogen content, this experiment employed an oxygen–nitrogen–hydrogen analyzer to measure hydrogen levels in corroded specimens. Each specimen was cut into three 3 mm × 3 mm × 3 mm cubes, which were then immersed in anhydrous ethanol to remove surface contaminants and minimise measurement errors.

Molecular dynamics simulation

The research investigated the corrosion reaction of titanium substrates in a mixed nitric acid–hydrofluoric acid solution through molecular dynamics simulations. To accurately simulate the corrosion process of titanium alloys via MD, the substrate model was constructed by directly obtaining the CIF file of the titanium unit cell's crystallographic structure. However, the crystallographic structure CIF file cannot be directly used for calculations in LAMMPS software and requires format conversion. Therefore, during the construction of the titanium substrate MD model, the titanium crystal file was imported into Materials Studio. Using the Supercell module, the crystallographic structure CIF file of the titanium unit cell was expanded to complete the initial model construction of the titanium substrate. Real-world metal corrosion environments contain highly complex compositions. Drawing on existing relevant studies,^42,43 this work models the corrosion environment of titanium alloys as a weak electrolyte solution containing only water molecules, hydrofluoric acid and nitric acid. This simplification reduces modelling complexity and shortens computational cycles. To ensure the reliability of this MD model, molecular simulation software was used to construct models of water molecules, hydrofluoric acid and nitric acid. A total of 400 H₂O molecules, 120 H⁺ ions and 120 F⁻ and NO₃⁻ ions in varying ratios (1:1, 1:2, 1:3, 2:1, 3:1) of 120 F⁻ ions and NO₃⁻ ions into a box measuring 66.72 Å × 38.04 Å × 18 Å, then imported into LAMMPS. The ReaxFF potential function^44–47 was employed to simulate corrosion reactions, investigating the effects of varying HF/NO₃ ratios on titanium alloy corrosion behaviour. Figure 2 depicts the corrosion model at the titanium substrate–mixed acid interface. To accelerate the MD simulation process and enable sufficient reaction within limited computational time, the ReaxFF-MD simulations employed an NVT ensemble. The simulation temperatures were set to 300 K, 325 K, 350 K and 375 K. Temperature and pressure were controlled using the Nose–Hoover algorithm and Berendsen algorithm, respectively. The simulation ran for 1 ns with a time step of 0.001 ps and completed 1,000,000 steps. To ensure precise alignment between the substrate model and mixed acid solution model during merging and prevent atomic loss during relaxation, the ReaxFF-MD simulation system employs a shrinkable boundary condition along with the Z-axis and periodic boundary conditions along with the X and Y axes. The reaction force field employed in this research is based on parameters developed by Kim et al.,^48,49 which systematically describes reactions between Ti, O, H, N and F atoms. The ReaxFF methodology ensures the reliability of the force field and enables effective simulation of titanium corrosion behaviour during kinetic processes. Atomic charges were calculated using the charge equilibrium method.⁵⁰ Post-processing of MD results was performed using OVITO software.⁵¹

Figure 2.

Titanium matrix-mixed acid corrosion model (a) HF:HNO₃ = 1:1; (b) HF:HNO₃ = 2:1; (c) HF:HNO₃ = 3:1; (d) HF:HNO₃ = 1:2; (e) HF:HNO₃ = 1:3.

The theoretical foundation of ReaxFF lies in combining bond order formalism with a description of polarisable charges, thereby enabling a unified description of both reactive and non-reactive interactions between atoms. Consequently, this method can accurately simulate covalent and electrostatic behaviours in various materials.⁵² The total energy of a system comprises both bonding and non-bonding energy components, expressed as^48,49:

E_{system} = E_{bond} + E_{over} + E_{under} + E_{lp} + E_{val} + E_{t o r s} + E_{vdW} + E_{Coulomb}

(1)

where E_bond originates from the bond order between atoms. E_over and E_under are used to correct deviations in atomic coordination numbers from ideal valence states. E_lp describes the contribution of lone pairs to the total energy. E_val and E_tor represent the potential energy arising from three-body valence angle distortion and four-body dihedral torsion, respectively. E_vdW incorporates short-range Pauli repulsion terms and long-range dispersion terms. E_Coulomb of the system is evaluated based on charges calculated using the electronegativity equilibrium method to assess electrostatic interactions.^52,53

Bond order and bond energy are key factors determining interatomic interactions. In the ReaxFF potential function used in this research, the bond energy term for each pair of atoms i and j is expressed as^54,55

E_{bond} (i, j) = - D_{e, i j} {BO}_{i j} (r_{i j}) \exp [- p_{b e 1, i j} (\frac{r_{i j}}{r_{0, i j}} - 1)]

(2)

where $r_{i j}$ denotes the interatomic distance, $D_{e, i j}$ represents the reference bond depth (bond dissociation energy scale), $r_{0, i j}$ is the reference bond length, $p_{b e 1, i j}$ is the exponential parameter controlling the rate of bond energy decay with distance. ${BO}_{i j} (r_{i j})$ is the bond order – a continuous function of $r_{i j}$ that enables a smooth description of bond formation and breaking.

The fundamental assumption of the ReaxFF model is that the bond order ${BO}_{i j}$ between atomic pairs can be directly calculated from the interatomic distance $r_{i j}$ . When computing bond orders, ReaxFF distinguishes between the contributions of σ bonds, π bonds and double π bonds. The specifics are shown in the formula^47,48:

{BO}_{i j} = {BO}_{i j}^{σ} + {BO}_{i j}^{π} + {BO}_{i j}^{π π}

= \exp [p_{b o 1} \cdot {(\frac{r_{i j}}{r_{o}^{σ}})}^{p_{b o 2}}] + \exp [p_{b o 3} \cdot {(\frac{r_{i j}}{r_{o}^{π π}})}^{p_{b o 4}}] + \exp [p_{b o 5} \cdot {(\frac{r_{i j}}{r_{o}^{π π}})}^{p_{b o 6}}]

(3)

where $p_{b o 1}$ – $p_{b o 6}$ are empirical parameters for bond order calculation after parameter optimisation, while $r_{0}^{σ}, r_{0}^{π}$ and $r_{o}^{π π}$ are the reference distances corresponding to the respective bonds.

In the ReaxFF potential function, each element is characterised by a set of atomic parameters that govern interatomic interactions, polarisation responses as well as bond formation and breaking properties. The atomic parameter sets for the elements Ti, H, O, N and F are listed in Table 2 in the potential function files used.

Table 2.

Atomic parameters for Ti, H, O, N and F elements used in the ReaxFF potential.^47,48

Element	Cov.r (Å)	Valency	Mass	RvdW (Å)	EvdW (kcal/mol)	GammaEEM	Alfa	GammavdW	ChiEEM	EtaEEM
H	0.4900	1	1.0080	1.3550	0.0930	33.2894	5.0878	33.2894	4.528	10.723
O	0.6799	2	15.999	2.3890	0.1000	13.8449	9.2986	13.8449	5.365	11.760
N	0.7023	3	14.007	1.9324	0.1376	7.84310	9.5108	7.8431	5.318	10.049
F	0.7064	1	18.998	1.8601	0.1200	7.5000	10.8764	7.5000	6.960	12.720
Ti	2.0254	4	47.870	2.2105	0.1574	0.6311	12.7041	16.6482	−1.3647	6.8406

In Table 2, each atomic parameter is coupled to a specific energy term. cov.r represents the covalent radius, defining the effective size of an atom during bonding. It is used to calculate the bond order function, and its value directly influences the equilibrium position of bond length and bond energy. Valency denotes the idealised number of valence electrons in an atom, used to control over- and under-coordination penalty energies. R_vdW and E_vdW represent the van der Waals radius and potential well depth, respectively, which determine the non-bonding interaction term. alfa, gammavdW are the polarisability and van der Waals decay parameters, which affect the smoothness of the potential energy with distance.^52,54–56

E_{vdW} = D_{i j} [e^{α_{i j} (1 - r_{i j} / R_{v d W})} - 2 e^{\frac{1}{2} α_{i j} (1 - r_{i j} / R_{v d W})}]

(4)

gammaEEM, chiEEM and etaEEM are parameters of the electro-negative equilibrium model (EEM), representing the charge shielding constant, electro-negativity and chemical hardness, respectively. They are crucial for the electrostatic energy term and the charge optimisation process⁵²:

E_{Coulomb} = C_{E} \sum_{i < j} \frac{q_{i} q_{j}}{{(r_{i j}^{3} + γ_{i j}^{- 3})}^{1 / 3}}

(5)

A more detailed description of each ReaxFF parameter can be found in the studies by Kim et al.⁵⁶ and van Duin et al.⁵⁷

Results and discussion

Chemical reaction on Ti-mixed acid interface

To investigate the molecular mechanism of nitric acid–hydrofluoric acid mixed solution corrosion on titanium alloys, we performed a 1 ns corrosion simulation using the ReaxFF force field on a titanium substrate–mixed acid interface corrosion model. Figure 3 shows the final geometric structure snapshots of simulation cells with different ratios of nitric acid–hydrofluoric acid mixed solutions after running for 1 ns at 350 K. The bonding reactions between the mixed acid solution and the titanium substrate, along with the separation of numerous titanium atoms from the substrate surface, confirm the vigorous corrosion reaction between the acidic solution and the titanium substrate. The figure reveals multiple small blank areas within the titanium substrate mesh after 1 ns of corrosion. These represent localised defects on the substrate surface caused by mixed acid solution corrosion. As the HF proportion in the mixed acid solution increases, the number of small-area corrosion defects on the substrate surface also rises, leading to the formation of large-area corrosion pits.

Figure 3.

Corrosion morphology of titanium substrates exposed to mixed acid solutions at different concentrations (a) HF:HNO₃ = 1:1; (b) HF:HNO₃ = 2:1; (c) HF:HNO₃ = 3:1; (d) HF:HNO₃ = 1:2; (e) HF:HNO₃ = 1:3.

During the corrosion process, highly electronegative F⁻ ions aggressively bond with surface Ti atoms. This strong interaction significantly lowers the dissolution energy barrier, causing individual Ti atoms to detach from the crystal lattice and generating a multitude of surface vacancies (small-area defects). As the HF proportion increases, the frequency of this atomic extraction rises exponentially. Consequently, the rapid accumulation and coalescence of these adjacent atomic vacancies ultimately drive the formation of the large-area corrosion pits observed on the substrate surface.

A statistical analysis of the change in titanium atom loss over time was conducted to visually demonstrate the impact of nitric acid–hydrofluoric acid mixed solutions on the corrosion behaviour of titanium substrates. Figure 4 illustrates the curve showing the variation in titanium atom loss with simulation time. As evident from the graph, the curve exhibits a characteristic progression from slow to fast: atomic loss is initially gradual, with corrosion reactions occurring at an extremely slow pace. As time progresses, atomic loss increases sharply as the titanium substrate becomes directly exposed and undergoes vigorous reaction with the corrosive solution. An increase in the hydrofluoric acid proportion significantly accelerates titanium atomic corrosion loss, further confirming hydrofluoric acid as the primary factor driving titanium corrosion. Additionally, elevated temperatures lead to a marked increase in the number of corroded titanium atoms, indicating that the corrosion reaction is governed by a thermally activated mechanism.

Figure 4.

Curve of Ti atom loss at different temperatures (a) 300 K; (b) 325 K; (c) 350 K; (d) 375 K.

To investigate the interaction between the Ti substrate and the mixed nitric–hydrofluoric acid solution and quantitatively evaluate the migration rate and motion behaviour of particles within the system under this corrosive environment, the research employs mean square displacement (MSD) analysis.^58–61 The diffusion capability of Ti atoms is one of the key factors influencing their corrosion kinetics. Therefore, accurately obtaining the diffusion coefficient of Ti atoms is crucial for revealing the microscopic dynamic mechanisms of the corrosion reaction.

The research calculated the MSD curves of Ti atoms at four different temperatures: 300 K, 325 K, 350 K and 375 K. The MSD calculation method is as follows⁶¹:

MSD = ⟨ | r_{i} (t) - r_{i} (0) |^{2} ⟩

(6)

where $r_{i} (t)$ and $r_{i} (0)$ denote the positions of the i-th atom at times t and 0, respectively, while the operator <> represents the average value across all particles.

By analyzing the temporal variation of the MSD, the diffusion coefficient D can be further calculated.^62,63 Assuming diffusion behaviour follows the Einstein relation,⁶⁴ D can be determined from the slope of the linear region of the MSD versus time curve⁶⁴:

D = \frac{1}{6} \lim_{t \to \infty} \frac{d}{d t} MSD (t)

(7)

The diffusion coefficient of Ti atoms was further calculated through linear fitting of the MSD data, which are obtained from LAMMPS simulations and are displayed in Figure 5 over time under different temperature conditions. The corresponding results are summarised in Table 3. Data analysis indicates that the diffusion coefficient of Ti atoms exhibits a significant upward trend as the proportion of hydrofluoric acid in the mixed acid increases. This result directly confirms that HF strongly promotes the activation and dissolution of the titanium substrate during the corrosion process, accelerating the material corrosion rate. Additionally, the diffusion coefficient of Ti atoms also increases markedly with rising temperature. Elevated temperatures provide higher kinetic energy for atomic diffusion, thereby accelerating the corrosion reaction kinetics.

Figure 5.

Mean square displacement (MSD) curves of Ti atoms at different temperatures (a) 300 K; (b) 325 K; (c) 350 K; (d)375 K.

Table 3.

Diffusion coefficient of titanium atoms in corrosion reactions.

Temperature/K	HF:HNO₃ = 1:3	HF:HNO₃ = 1:2	HF:HNO₃ = 1:1	HF:HNO₃ = 2:1	HF:HNO₃ = 3:1
300	2.84 × 10⁻⁵	3.98 × 10⁻⁵	4.25 × 10⁻⁵	4.75 × 10⁻⁵	5.26 × 10⁻⁵
325	3.49 × 10⁻⁵	4.21 × 10⁻⁵	4.48 × 10⁻⁵	5.63 × 10⁻⁵	5.75 × 10⁻⁵
350	3.67 × 10⁻⁵	4.31 × 10⁻⁵	4.82 × 10⁻⁵	5.91 × 10⁻⁵	6.28 × 10⁻⁵
375	4.32 × 10⁻⁵	5.24 × 10⁻⁵	5.80 × 10⁻⁵	6.40 × 10⁻⁵	7.08 × 10⁻⁵

The molecular mechanism of the corrosion process involves a complex competitive reaction between the activation dissolution of hydrofluoric acid and the oxidation passivation of nitric acid. Hydrofluoric acid preferentially reacts with titanium to form TiF₄ and hydrogen gas¹⁰:

Ti + 4HF \to {TiF}_{4} + {2H}_{2} ↑

(8)

2Ti + 6HF \to 2Ti F_{3} + {3H}_{2} ↑

(9)

Nitric acid promotes the formation of surface oxides:

Ti + 4 {HNO}_{3} \to {TiO}_{2} + 4 {NO}_{2} ↑ + 2 H_{2} O

(10)

In this reaction system, the dynamic equilibrium between the activation dissolution of HF and the passivation protection of HNO₃ is crucial for achieving efficient and controllable corrosion of the titanium substrate. The overall reaction equation is as follows¹⁰:

3 Ti + 4 {HNO}_{3} + 12 HF \to 3 {TiF}_{4} + 8 H_{2} O + 4 NO ↑

(11)

Figure 6 illustrates the mechanism of corrosion of a titanium substrate by a mixed nitric acid–hydrofluoric acid solution. From the overall reaction equation, hydrogen gas does not appear as a final product. However, the process of hydrogen generation is not eliminated; rather, the produced hydrogen is rapidly oxidised by nitric acid, thereby suppressing hydrogen gas formation. Specifically, during the reaction between hydrofluoric acid and titanium that generates hydrogen gas, nitrate ions in the corrosive solution participate throughout the entire reaction process. Most hydrogen atoms and molecules generated by corrosion react with nitrate ions to form NO. Nevertheless, a small fraction of hydrogen escapes complete oxidation by nitrate. This residual hydrogen either chemically adsorbs onto the titanium surface, forming titanium hydrides or diffuses into the metal interior as hydrogen atoms through defects such as grain boundaries and dislocations, leading to internal hydrogen dissolution.⁶⁵

Figure 6.

Schematic diagram of corrosion mechanism in Ti-HF-HNO₃ system.

To investigate the effect of mixed acid solutions with varying ratios on hydrogen permeation behaviour, this study further measured the number of hydrogen atoms within the titanium substrate under different temperature conditions. The relationship between hydrogen content and simulation time is shown in Figure 7. The figure indicates that as the proportion of hydrofluoric acid increases and the simulation time extends, the number of hydrogen atoms in the titanium substrate significantly rises. This demonstrates that HF not only promotes titanium dissolution during corrosion but also intensifies hydrogen permeation behaviour. Conversely, as the proportion of nitric acid increases, the enrichment of hydrogen atoms within the titanium substrate markedly decreases. This is attributed to the strong oxidising properties of HNO₃, which promote the formation of a passivation film on the surface, thereby inhibiting hydrogen desorption and infiltration. Additionally, elevated temperatures significantly accelerate the corrosion reaction kinetics, further increasing the dissolution of hydrogen within the substrate.

Figure 7.

Hydrogen absorption curve of titanium substrate over time at different temperatures (a) 300 K (b) 325 K; (c) 350 K; (d) 375 K.

Corrosion kinetics and surface characterisation analysis of TA15

To systematically investigate the influence of mixed acid ratios (nitric acid and hydrofluoric acid) on the corrosion behaviour of titanium alloys, this study employed the weight loss method to precisely measure the mass changes of specimens after corrosion testing. Based on the mass loss data, corrosion rates and mass loss rates were calculated for different acid mixtures, with detailed results presented in Tables 4 and 5. The corresponding trend curves showing the variation of corrosion rates and mass loss rates with acid ratios are depicted in Figure 8.

Figure 8.

Mass loss rates and corrosion rates of samples in HF-HNO3 mixed acid.

Table 4.

Corrosion rates of samples in mixed acid at 3 min time intervals.

Time/min	HF:HNO₃ = 1:1	HF:HNO₃ = 2:1	HF:HNO₃ = 3:1	HF:HNO₃ = 1:2	HF:HNO₃ = 1:3
3	1.837 / 1.908 / 1.811	2.201 / 2.309 / 2.301	2.664 / 2.571 / 2.650	1.775 / 1.775 / 1.827	1.529 / 1.570 / 1.471
6	2.291 / 2.275 / 2.333	2.882 / 2.963 / 2.846	3.070 / 3.188 / 3.151	2.284 / 2.205 / 2.231	1.779 / 1.696 / 1.812
9	2.656 / 2.689 / 2.629	3.343 / 3.247 / 3.357	3.561 / 3.660 / 3.622	2.560 / 2.655 / 2.671	2.014 / 1.956 / 2.033
12	2.902 / 3.036 / 2.934	3.485 / 3.440 / 3.559	3.751 / 3.830 / 3.799	2.788 / 2.920 / 2.894	2.214 / 2.209 / 2.298
15	3.268 / 3.278 / 3.222	3.528 / 3.687 / 3.627	3.921 / 4.020 / 3.976	3.117 / 3.193 / 3.098	2.573 / 2.454 / 2.500
18	3.438 / 3.558 / 3.488	3.767 / 3.710 / 3.724	3.852 / 3.722 / 3.896	3.321 / 3.385 / 3.420	2.810 / 2.917 / 2.785

Table 5.

Mass loss rates of samples in mixed acid at 3 min time intervals.

Time/min	HF:HNO₃ = 1:1	HF:HNO₃ = 2:1	HF:HNO₃ = 3:1	HF:HNO₃ = 1:2	HF:HNO₃ = 1:3
3	2.727 / 2.815 / 2.668	3.279 / 3.454 / 3.342	3.957 / 3.812 / 3.863	2.665 / 2.682 / 2.750	2.269 / 2.365 / 2.218
6	3.497 / 3.448 / 3.539	4.472 / 4.503 / 4.330	4.730 / 4.911 / 4.795	3.524 / 3.401 / 3.476	2.672 / 2.632 / 2.808
9	4.207 / 4.219 / 4.132	5.329 / 5.207 / 5.397	5.747 / 5.908 / 5.823	4.104 / 4.264 / 4.277	3.142 / 3.101 / 3.225
12	4.783 / 4.992 / 4.805	5.949 / 5.797 / 5.990	6.443 / 6.571 / 6.465	4.692 / 4.925 / 4.783	3.617 / 3.587 / 3.740
15	5.610 / 5.681 / 5.580	6.489 / 6.750 / 6.738	7.173 / 7.354 / 7.288	5.481 / 5.615 / 5.449	4.320 / 4.126 / 4.274
18	6.276 / 6.490 / 6.426	7.240 / 7.128 / 7.208	7.579 / 7.362 / 7.715	6.185 / 6.288 / 6.370	5.051 / 5.104 / 4.869

The quality loss rate is calculated using the following formula⁶⁶:

Δ m = m_{0} - m_{1}

(12)

Quality loss rate (%) = \frac{Δ m}{m_{0}} \times 100 %

(13)

where Δm denotes the difference obtained by subtracting the instantaneous mass from the initial mass, m₀ represents the initial mass of the sample (g) and m₁ denotes the mass of the cleaned corrosion specimen (g).

The corrosion rate can be calculated from the mass loss using the following formula⁶⁶:

V_{corr} = \frac{Δ m}{S t}

(14)

where S represents the surface area of the specimen (m²), V_corr denotes the average corrosion rate (g·m⁻²·s⁻¹) and t indicates the immersion corrosion time (s).

Analysis results indicate that both the corrosion rate and mass loss rate of the specimens generally increase with prolonged corrosion duration. During the initial corrosion stage, the mixed acid solution primarily reacts with the titanium oxide film (TiO₂) on the titanium alloy surface. Once the oxide layer dissolves, the acid solution further erodes the exposed titanium substrate, accelerating the corrosion process. Notably, when HF:HNO₃ = 3:1, the corrosion rate decreased during the final 3 min. This may result from the HF concentration in the reaction system reaching a depletion threshold, reducing the amount of acid solution participating in effective corrosion reactions. The corrosion rate strongly depends on the volume ratio of HNO₃ to HF in the mixed acid. The HF:HNO₃ = 3:1 group exhibited the highest corrosion rate at all time points, indicating that high-concentration HF rapidly disrupts surface structures and promotes rapid dissolution of titanium. Conversely, the HF:HNO₃ = 1:3 group exhibited the lowest and most stable corrosion rate. This is primarily attributed to the strong oxidising and passivating effects of high-concentration HNO₃, which significantly suppressed the progression of the corrosion reaction.

To evaluate changes in the microstructure of the sample surface, electron microscopy observations were conducted. Figure 9 shows SEM images of the TA15 alloy after corrosion in mixed solutions of nitric acid and hydrofluoric acid at different ratios. All images exhibit typical material loss and morphological changes, but significant morphological differences are observed under varying acid ratios. The surface of the sample with a high HF ratio exhibits numerous dark pits, grooves and coarse protrusions. Local magnifications reveal severe selective corrosion in these areas, likely corresponding to rapid dissolution at grain boundaries, phase boundaries or defects. The high HF concentration causes intense localised corrosion with the fastest corrosion rate but results in high surface roughness.

Figure 9.

Microstructural morphology of samples corroded by mixed acid solutions with different volume ratios (a) HF:HNO₃ = 1:1; (b) HF:HNO₃ = 2:1; (c) HF:HNO₃ = 3:1; (d) HF:HNO₃ = 1:2; (e) HF:HNO₃ = 1:3.

These morphological observations can be mechanistically explained by the chemical interactions between the mixed acid and the titanium substrate. As reported in recent corrosion studies,^66,67 F⁻ ions exhibit an exceptionally strong chemical affinity for Ti. They readily penetrate the natural TiO₂ passive film and react to form highly soluble fluorotitanate complexes, such as [TiF6]²⁻. This aggressive, localised active dissolution continuously extracts Ti atoms from the lattice, driving the rapid formation of the observed macroscopic corrosion pits at high HF concentrations. In contrast, the surface of the high nitric acid ratio sample (e.g., at an HF: HNO₃ ratio of 1:3) appeared relatively flat and uniform. Magnification revealed fine, uniformly distributed shallow pits without large pits or cracks. Additionally, the blocky and flaky structures visible in the magnified image may represent either deposited corrosion products (e.g., TiF₄, TiF₃) or exposed titanium substrate morphology. The addition of sufficient HNO₃ significantly alters the dissolution dynamic. Acting purely as a potent oxidising agent,¹⁰ the abundant NO₃⁻ ions provide strong oxidising power that directly accelerates the repassivation process. This promotes the rapid formation of a thick, dense TiO₂ passivation layer on the surface, which significantly inhibits severe corrosion. Consequently, this dynamic balance between F⁻-induced active dissolution and NO₃⁻-promoted repassivation effectively suppresses localised attack, thereby stabilising the corrosion rate and yielding a uniformly milled, smooth surface.

To systematically investigate the corrosion mechanism of TA15 titanium alloy in mixed nitric acid–hydrofluoric acid solutions at different ratios, the experiment employed energy-dispersive X-ray spectroscopy to perform elemental mapping analysis on the surfaces of corroded specimens (Figures 10 and 11). As shown in Figure 11, the elemental distribution on the corroded surfaces closely correlates with the acid volume ratios, which further confirms the ‘activation-passivation’ mechanism. In the HF-dominated solution (Figure 11(c), HF:HNO₃ = 3:1), a dense distribution of F element is observed, indicating intense active dissolution and the formation the formation and retention of fluoride products such as TiF₄/TiF₃. Conversely, in the HNO₃-dominated solution (Figure 11(e), HF: HNO₃ = 1:3), the surface shows a clear enrichment of N elements alongside a weakened F signal. This confirms that a higher HNO₃ concentration promotes the rapid formation of a nitrogen-doped passivation layer, which effectively blocks F⁻ ion penetration and suppresses further corrosion. Furthermore, the transition of elemental dominance from F to N across the intermediate ratios (Figure 11(a), 11(b) and 11(d)) is in excellent agreement with the atomic-scale molecular dynamics simulations. This confirms that macroscopic corrosion is controlled by the dynamic competition between HF-driven atomic dissolution and HNO₃-promoted repassivation.

Figure 10.

Energy dispersive spectroscopy surface scanning energy-dispersive spectroscopy images of samples corroded with mixed acid solutions of different volume ratios (a) HF:HNO₃ = 1:1; (b) HF:HNO₃ = 2:1; (c) HF:HNO₃ = 3:1; (d) HF:HNO₃ = 1:2; (e) HF:HNO₃ = 1:3.

Figure 11.

Energy dispersive spectroscopy surface scan element distribution map of samples corroded with mixed acid solutions of different volume ratios (a) HF:HNO₃ = 1:1; (b) HF:HNO₃ = 2:1; (c) HF:HNO₃ = 3:1; (d) HF:HNO₃ = 1:2; (e) HF:HNO₃ = 1:3.

To investigate the effect of mixed acid solutions at different ratios on the surface roughness of corroded samples, the experiment measured the three-dimensional surface topography of corroded specimens using a Zygo optical profilometer. Among these measurements, Sa (arithmetic mean height) serves as an extended metric for Ra on the surface. As shown in Figure 12, surface roughness Sa increased from 0.277 to 0.313 μm with rising hydrofluoric acid concentration. This primarily resulted from the rapid formation of large bubbles on the sample surface under high-concentration hydrofluoric acid. The escape of these bubbles intensified pit formation, while the ongoing dissolution process led to the coalescence of corrosion pits, ultimately creating larger and deeper defects that significantly increased surface roughness. The presence of nitric acid mitigated the corrosive action of hydrofluoric acid, reduced bubble formation and prevented pitting caused by localised flaking of reaction products. Specifically, as nitric acid concentration increases, the surface progressively becomes smoother, with roughness Sa decreasing to 0.224 µm and 0.186 µm, respectively.

Figure 12.

Three-dimensional surface morphology after corrosion of mixed acid solutions with different volume ratios (a) HF:HNO₃ = 1:1; (b) HF:HNO₃ = 2:1; (c) HF:HNO₃ = 3:1; (d) HF:HNO₃ = 1:2; (e) HF:HNO₃ = 1:3.

Subsequently, we conducted XRD analysis on the phase structure of the TA15 specimen after corrosion in mixed solutions of nitric acid and hydrofluoric acid at different ratios. The results are shown in Figure 13. The XRD diffraction patterns clearly indicate that the corroded titanium alloy surface primarily consists of a Ti matrix, TiO₂, TiF₄ and TiF₃, with the main corrosion products being TiF₃ and TiF₄. The positions of the diffraction peaks remained stable after corrosion, and the diffraction intensity showed a slight increase trend with extended corrosion time. When the HF proportion was higher, the diffraction peaks of TiF₃ and TiF₄ were relatively stronger, while the TiO₂ peak was weaker. When the HNO₃ proportion was higher, the TiO₂ diffraction peak significantly intensified, while the titanium fluoride peaks were relatively weaker.

Figure 13.

XRD characterisation of specimens (a) XRD diffraction patterns of specimens after corrosion by mixed acid solutions with different volume ratios (b) phase diagram of specimens after corrosion in HF:HNO₃ = 1:2 solution.

Due to titanium alloys’ high sensitivity to hydrogen, hydrogen embrittlement is one of their primary failure mechanisms. Consequently, hydrogen content in chemically milled components must be strictly controlled. To systematically investigate how different volume ratios of nitric acid and hydrofluoric acid affect hydrogen permeation behaviour in titanium alloys, this study measured hydrogen content in specimens corroded by mixed acids at varying ratios, with the results shown in Figure 14.

Figure 14.

Hydrogen content in test specimens after corrosion by mixed acid solutions with different volume ratios.

The quantitative evaluation of hydrogen permeation presented in Figure 14 is of profound significance for establishing a reliable processing window during chemical milling. The figure indicates that hydrogen content is significantly regulated by the mixed acid ratio: the lowest hydrogen content was observed at HF:HNO₃ ratios of 1:2 and 1:3, at 0.0056% and 0.0054%, respectively. Conversely, hydrogen content markedly increased with higher HF ratios (2:1 and 3:1), reaching 0.0107% and 0.0121%, respectively.

Titanium alloys are intrinsically highly susceptible to hydrogen embrittlement, with a critical failure threshold typically ranging from 0.01% to 0.0125%.^68,69 The experimental data reveals that the HF:HNO₃ ratio of 3:1 dangerously pushes the hydrogen content to 0.0121%. Exceeding this threshold may lead to the precipitation of hydride phases, which can trigger hydrogen embrittlement and catastrophic mechanical failure. Therefore, optimising the ratio to 1:2 is not merely conducive to achieving a superior surface finish but is fundamentally critical to preserving the structural integrity, mechanical reliability and overall service safety of the aerospace components.

To systematically investigate the corrosion behaviour of titanium alloys in mixed solutions of HNO₃ and HF, and to precisely optimise their concentration ratio, the research builds upon the previously determined optimal volume ratio (HF:HNO₃ = 1:2). Two additional concentration gradient experiments were designed to determine the optimal concentration range for each component. The specific experimental design is as follows:

With HNO₃ concentration fixed at 15%, the system adjusts the HF concentration ratio. The gradient settings for HF concentration are: 4%, 5%, 6%, 7%, 8% and 9%.

With HF concentration fixed at 8%, the system adjusts the HNO₃ concentration relative to HF under these conditions. The gradient settings are: 8%, 10%, 12%, 14%, 16% and 18%.

To eliminate interference from the natural oxide film on the titanium alloy surface during corrosion rate measurements, all specimens underwent uniform acid pickling pre-treatment prior to experimentation. Subsequently, the mass changes of each specimen after corrosion testing were precisely measured using the gravimetric method. Based on the mass loss data, the corrosion rates for both sets of experimental specimens were calculated, with detailed numerical results presented in Table 6. The corresponding trend curves showing the variation of corrosion rates with acid concentration are depicted in Figure 15, visually revealing the nonlinear influence of the HF:HNO₃ concentration ratio on the corrosion rate of the titanium alloy.

Figure 15.

Relationship curves between HF and HNO₃ concentration gradients and the corrosion rate of TA15 titanium alloy.

Table 6.

Corrosion rate of TA15 titanium alloy under HF and HNO₃ concentration gradient experiments.

HF concentration (%)	Corrosion rate (g·m⁻²·s⁻¹)	HNO₃ concentration (%)	Corrosion rate (g·m⁻²·s⁻¹)
4	1.34409	8	2.32975
5	2.03106	10	2.71804
6	2.47909	12	3.07646
7	2.86738	14	3.40502
8	3.22581	16	3.25568
9	3.28554	18	3.16607

The results indicate that at a fixed HNO₃ concentration of 11%, the weight loss rate of the specimens continuously increases with rising HF concentration. This trend demonstrates that HF plays a dominant activating role in the corrosion process, with its elevated concentration significantly promoting the dissolution reaction of the titanium alloy. Conversely, at a fixed HF concentration of 8%, the weight loss rate of the specimens first increases and then decreases as the HNO₃ concentration rises. This nonlinear variation stems from the dual effect of nitric acid: at lower concentrations, HNO₃ primarily promotes corrosion through its strong oxidising properties. However, once its concentration exceeds a critical threshold, excess nitric acid preferentially induces the formation of a dense passivation layer on the metal surface. This layer effectively inhibits further corrosion, thereby reducing the weight loss rate.

By empirically fitting the experimental data obtained in this study (as shown in Figure 15), fourth-order polynomial functions describing the corrosion rate of the TA15 sample as a function of hydrofluoric acid and nitric acid concentrations were derived as follows:

Relationship between corrosion rate and HF concentration:

y = - 0.0093 x^{4} + 0.24 x^{3} - 2.33 x^{2} + 10.40 x - 16.02

(15)

Relationship between corrosion rate and HNO₃ concentration:

y = 0.0007 x^{4} - 0.037 x^{3} + 0.71 x^{2} - 5.56 x + 17.59

(16)

Ultimately, after comprehensively evaluating factors such as corrosion rate, surface roughness and hydrogen embrittlement risk control, the optimal corrosion solution formulation for this experiment was determined to be: 8% hydrofluoric acid concentration and 14% nitric acid concentration. This ratio not only achieves high material removal efficiency but also significantly improves surface uniformity while effectively suppressing hydrogen embrittlement. Furthermore, the results of this study align with previous research,^7,8 further supporting the feasibility of this ratio for chemical milling of titanium alloys.

Ultimately, the established atomistic-to-macroscopic correlation provides a robust theoretical framework for the aerospace industry to rationally design chemical milling fluids. By precisely controlling the mixed-acid formulation based on this mechanism, manufacturers can enable high-precision fabrication while strictly guarding against catastrophic hydrogen embrittlement failures, thereby ensuring the long-term structural integrity and service safety of titanium alloy components.

While the current study provides profound insights into the competitive corrosion mechanisms of TA15 titanium alloy in mixed HF-HNO₃ solutions, it still has certain limitations regarding real-world engineering applications. Specifically, this research primarily focused on corrosion behaviours within static acid solutions and did not dynamically control the temperature fluctuations during the highly exothermic chemical milling process. In actual industrial manufacturing, fluid hydrodynamics and temperature variations can significantly alter the localised activation–passivation balance and the resulting surface integrity. Therefore, future work will focus on investigating flow-induced corrosion mechanisms under dynamic hydrodynamic conditions and developing precise temperature-controlled systems. This future scope will further bridge the gap between fundamental mechanisms and practical process optimisation for complex aerospace components.

Conclusions

In summary, the research systematically investigated the mechanism by which mixed nitric acid–hydrofluoric acid solutions at different concentrations affect the corrosion behaviour of TA15 titanium alloy through a combination of molecular dynamics simulations and experimental verification. It comprehensively examined the regulatory patterns of acid ratio and concentration changes on the material's microstructure, corrosion rate and surface roughness, providing a critical theoretical foundation and data support for optimising the chemical etching process of TA15 titanium alloy. The main research findings are as follows:

Molecular dynamics simulations revealed the competitive corrosion mechanism at the atomic scale, confirming that F⁻ acts as an activator while NO³⁻ promotes passivation. An increase in HF proportion (from HF:HNO₃ = 1:3 to 3:1) significantly elevates the diffusion coefficient of Ti atoms (from 2.84 × 10⁻⁵ to 5.26 × 10⁻⁵ at 300 K) and exponentially enhances the hydrogen absorption capacity into the lattice.

The mixing ratio of HF to HNO₃ strictly dictates the macroscopic corrosion quality. A high HF ratio (HF:HNO₃ = 3:1) yields rapid dissolution but results in deep pitting with high surface roughness (Sa = 0.313 μm) and pushes the hydrogen content to a critical limit (0.0121%), severely elevating hydrogen embrittlement risks. Conversely, a high HNO₃ ratio (HF:HNO₃ = 1:3) produces a superior smooth surface (Sa = 0.180 μm) with low hydrogen absorption (0.0054%) due to the formation of a dense passivation layer.

After a comprehensive evaluation of corrosion kinetics, surface integrity and hydrogen embrittlement constraints, the optimal chemical formulation was determined to be 8% HF and 14% HNO₃. This formulation achieves a stable and highly efficient corrosion rate of 3.405g·m⁻²·s⁻¹.

Footnotes

Acknowledgements

This work was supported by National Key Research and Development Program of China (2024YFB4610201), National Natural Science Foundation of China (12372191, 52332012).

ORCID iD

Zhao Zhang

Authors’ contribution

Zelong Yu contributed writing – original draft, data curation, methodology and formal analysis. Zhenhe Yu contributed to validation. Xu Guo contributed to writing – review & editing. Zhao Zhang contributed to writing – review & editing, supervision, funding acquisition and project administration.

Funding

The authors disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This work was supported by the National Key Research and Development Program of China, (grant number 2024YFB4610201).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

References

Xie

Fan

Wang

, et al. Stepwise cooling spinning to achieve isotropic strength-ductility synergy of titanium alloy tube. J Mater Sci Technol 2026; 244: 156–172.

Zhang

Wang

, et al. Anomaly in the elastic modulus of metastable B Ti-10V-2Fe-3Al alloy with a plus B dual-phase. J Mater Sci Technol 2026; 243: 237–244.

Zafar

Singh

, et al. New insights into the effect of heat-treatment on the machinability and tool wear in machining of additively manufactured metastable (3-Ti alloy Beta-C (ti-3Al-8V-6Cr-4Mo-4Zr). Tribol Int 2026; 213: 111033.

Park

Kim

, et al. Temperature dependent dynamic restoration mechanisms in Ti-6Al-4 V during isothermal deformation across α plus a phase regions. J Mater Sci Technol 2026; 240: 250–278.

Wang

Liu

, et al. Effect of Fe content on tensile properties of TA15 alloy with equiaxed microstructure produced by laser directed energy deposition. Virtual Phys Prototyp 2025; 20: e2499479.

Sefer

Gaddam

Roa

, et al. Chemical milling effect on the low cycle fatigue properties of cast Ti-6Al-2Sn-4Zr-2Mo alloy. Int J Fatigue 2016; 92: 193–202.

Krishna

Suresh

Mahesh

, et al. Influence of chemical milling on the mechanical and microstructural properties of α cased β-Titanium alloy. J Mater Eng Perform 2024; 33: 11023–11036.

Deshmukh

Kadam

Joshi

, et al. Removal of alpha case on titanium alloy surfaces using chemical milling. Mach Sci Technol 2017; 21: 257–278.

Prasad

Sonwani

. Optimize chemical milling of aluminium alloys to achieve minimum surface roughness in aerospace and defense industry. J Indian Chem Soc 2025; 102: 101537.

10.

Sefer

Dobryden

Almqvist

, et al. Chemical milling of cast Ti-6Al-4V and Ti-6Al-2Sn-4Zr-2Mo alloys in hydrofluoric-nitric acid solutions. Corrosion 2017; 73: 394–407.

11.

Priya

Kaliraj

Ningshen

, et al. Corrosion behavior of Ti-Grade2 dissolver material in nitric acid containing fluoride ions. J Nucl Mater 2025; 612: 155833.

12.

Shliakhetka

Pohrelyuk

Chumalo

, et al. Influence of concentration of sulfuric and hydrochloric acids on corrosion resistance of porous titanium. J Mater Sci 2023; 58: 15047–15060.

13.

Wei

Zhang

Chu

, et al. Superior corrosion-resistant cold sprayed titanium coatings with fully dense microstructure after electron beam modification. J Alloys Compd 2025; 1036: 181978.

14.

Liu

Chen

, et al. An experimental study on corrosion resistance of Ti35 alloy and its high-fluence hydrogen bombardment behavior. J Phys Chem Solids 2025; 201: 112646.

15.

Motasim

Aydogan

Agacayak

, et al. The influence of sodium fluoride on the dissolution kinetics of metallic titanium in citric acid solution using the rotating disc method. Hydrometallurgy 2024; 226: 106297.

16.

Prosolov

Komarova

Kazantseva

, et al. Enhanced corrosion resistance and mechanical durability of the composite PLGA/CaP/ti scaffolds for orthopedic implants. Polymers (Basel) 2024; 16: 26.

17.

Chen

, et al. Visualisation and quantification of biofilm-substrate interface microenvironments based on a fungal-bacterial interaction model: an in-depth investigation into microbially mediated corrosion processes. Corros Sci 2025; 246: 112725.

18.

Wang

Fan

, et al. Understanding the effect of fluoride on corrosion behavior of pure titanium in different acids. Corros Sci 2021; 192: 109812.

19.

Wang

Liu

, et al. The effect of fluoride ions on the corrosion behavior of pure titanium in 0.05 M sulfuric acid. Electrochim Acta 2014; 135: 526–535.

20.

Wei

Pan

, et al. Effect of annealing temperatures on microstructural evolution and corrosion behavior of Ti-Mo titanium alloy in hydrochloric acid. Corros Sci 2022; 197: 110079.

21.

Chen

, et al. Study of the passivation and repassivation behavior of pure titanium in 3.5 wt NaCl solution and 6 M HNO₃ solution. Corros Sci 2023; 224: 111538.

22.

Huang

, et al. Comparison of electrochemical characteristics and passive film properties of selective laser melted and wrought TA15 alloys in sulfuric acid solution. Corros Sci 2024; 236: 112254.

23.

Liu

Govindaraj

Sakairi

, et al. Influence of Zn²⁺ on corrosion behavior of Ti in H₂SO⁴ solution with F⁻. Electrochem 2025; 93: 037010.

24.

Zhang

. The diffusion behaviors in mg/al interface under variated temperatures and surface roughness. Phase Transitions 2025; 98: 85–108.

25.

Kong

Zhang

, et al. Coupling strengthening mechanism of precipitate and solute atoms in edge dislocation motion of Al-Mg-Si alloy. J Mater Res Technol 2024; 30: 134–145.

26.

Deng

Zhang

, et al. A novel synergistic efficient inhibitor of Camellia oleifera fruit shell extract with sodium dodecylbenzene sulfonate for steel corrosion in sulfamic acid. J Mater Sci Technol 2025; 239: 243–261.

27.

Zhang

Liang

Chen

, et al. Ab initio molecular dynamics investigation on Zr corrosion in aqueous environments: the effects of temperature, boric acid, and alloying element Nb. J Chem Phys 2025; 163: 034705.

28.

Hou

Zhang

Hong

, et al. The corrosion deterioration of reinforced passivation film: the impact of defects. Appl Surf Sci 2022; 582: 152408.

29.

Hao

, et al. Investigation of the hydrogen adsorption properties on titanium metal under vacuum conditions. J Ind Eng Chem 2025; 145: 491–505.

30.

Zhang

Ren

Guo

, et al. Experiments and DFT calculations on the effects of interstitial hydrogen on Ti corrosion products in high temperature water. Corros Sci 2024; 232: 112014.

31.

Chen

Wang

, et al. ReaxFF MD simulation on the deterioration of Fe₂O₃ passivation film in the hydrochloric acid coupled with mechanical processing. Surf Sci 2025; 757: 122738.

32.

Balasooriya

Wang

, et al. Understanding steel corrosion: surface chemistry and defects explored through DFT modelling – a review. Processes 2025; 13: 1971.

33.

Chen

Wang

, et al. Atomistic insights into chloride-induced depassivation mechanisms of protective Cu₂O films in marine environments. Corros Sci 2025; 256: 113171.

34.

Jin

Wei

Zhang

, et al. Preparation of Zr-based amorphous composite coatings on TA2 surface and their resistance to extreme corrosion and wear. Appl Surf Sci 2025; 688: 162328.

35.

Njoku

Nwanonenyi

Anyanwu

, et al. Nicotinohydrazide modified chitosan as a new benign bio-macromolecule for corrosion protection of steel in acidic environment: characterization, experimental and theoretical approach. J Build Eng 2025; 106: 112602.

36.

Elouhabi

Timoudan

Bensalah

, et al. Evaluation of the corrosion inhibition performance of the phosphate derivative K₂SrP₂O₇ on carbon steel in 1 M hydrochloric acid solution. Inorg Chem Commun 2025; 180: 114961.

37.

Zhou

Gao

Liu

, et al. Spatial structural effects of phytic acid in Ca-Mg-P composite coating formation: enhanced corrosion resistance and mechanical properties on magnesium. Surf Interfac 2025; 72: 107341.

38.

Tang

Deng

, et al. A comparison of the corrosion inhibition on 1060 aluminum in HCl solution between water and alcoholic extracts of Mikania micrantha. Corros Sci 2024; 228: 111799.

39.

Alwi

MAM

Ahmad

Piah

, et al. Synthesis of bioactive Schiff base thiosemicarbazone analogs with quinoline for high-performance copper corrosion inhibition: insights from RSM-assisted EIS, XPS, DFT, and MD studies. J Mol Struct 2025; 1345: 142961.

40.

Guo

Xue

, et al. Imidazole hexafluorophosphate ionic liquid as a greener corrosion inhibitor for carbon steel in hydrochloric acid and sulfuric acid solutions. ACS Omega 2025; 10: 27976–27991.

41.

Standardization Administration of China. GB/T 3620.1-2016 Designation and composition of titanium and titanium alloys. Beijing: Standards Press of China, 2016.

42.

Modwi

MMY

Feng

Liu

, et al. Study on synthesized corrosion inhibitor of chitosan and menthone based on Schiff base for carbon steel in the acidic environment. J Taiwan Inst Chem Eng 2025; 176: 106322.

43.

Bano

Goni

LKMO

Abdelkreem

SSE

, et al. Synthesis of cyclocopolymer of pyrrolidine motifs containing hydrolyzed methionine residue and maleic acid as a corrosion inhibitor for mild steel in acidizing environments: experimental and theoretical studies. Chem Eng Sci 2025; 318: 122242.

44.

Dan

Wang

Huang

, et al. From macroscopic performance to molecular mechanisms: multi-scale insights into enhanced impermeability of EMS/PAAS composites. J Environ Chem Eng 2025; 13: 118144.

45.

Liu

van Duin

ACT

, et al. Atomic insights into corrosion of Fe-Cr alloy in chloride contaminated environment: development of JAX-ReaxFF force field. Constr Build Mater 2025; 477: 141351.

46.

Wang

Chen

Zhang

, et al. Comprehensive study on the pyrolysis characteristics of end-of-life photovoltaic module backsheet: kinetic analysis, ReaxFF MD simulation and DFT calculation. J Anal Appl Pyrolysis 2026; 193: 107353.

47.

Zydek

Mitoraj-Królikowska

Marchewka

, et al. Atom level insight into corrosion behavior of modified aluminum in chloride solution: reactive atomistic simulations vs experiment. Corros Sci 2025; 255: 113157.

48.

Kim

van Duin

ACT

Kubicki

, et al. Molecular dynamics simulations of the interactions between TiO₂ nanoparticles and water with Na⁺ and Cl⁻, methanol, and formic acid using a reactive force field. J Mater Res 2013; 28: 513–520.

49.

Kim

van Duin

ACT

. Simulation of titanium metal/titanium dioxide etching with chlorine and hydrogen chloride gases using the ReaxFF reactive force field. J Phys Chem A 2013; 117: 5655–5663.

50.

dos Santos

Fernandes

Sampaio

MTG

, et al. DFT-guided design and experimental validation of small molecules to fight against mild steel corrosion: chemical synthesis and extensive evaluation of 4 small enones as corrosion inhibitors. Colloids Surf A: Physicochem Eng Asp 2025; 726: 137874.

51.

Zhao

Singh

, et al. Field-Enhanced catalysis: integrating experiment, theory, and machine learning for catalytic innovation. Appl Catal B Environ Energy 2026; 382: 125901.

52.

Senftle

Hong

Islam

, et al. The ReaxFF reactive force-field: development, applications and future directions. Npj Comput Mater 2016; 2: 15011.

53.

Zhang

Mei

Hou

, et al. Development of a ReaxFF reactive force-field modeling for magnesium nanoparticles and water system. Appl Surf Sci 2025; 700: 163207.

54.

Chenoweth

van Duin

ACT

Goddard

. ReaxFF reactive force field for molecular dynamics simulations of hydrocarbon oxidation. J Phys Chem A 2008; 112: 1040–1053.

55.

Zhang

Liu

van Duin

ACT

, et al. Development and validation of a general-purpose ReaxFF reactive force field for earth material modeling. J Chem Phys 2024; 160: 094103.

56.

Kim

Kumar

Persson

, et al. Development of a ReaxFF reactive force field for titanium dioxide/water systems. Langmuir 2013; 29: 7838–7846.

57.

van Duin

ACT

Dasgupta

Lorant

, et al. ReaxFF: a reactive force field for hydrocarbons. J Phys Chem A 2001; 105: 9396–9409.

58.

Chen

, et al. Wettability of Na₂SO₄-NaCl molten salt on CaO-based sorbent surface at high temperature: in-situ experiment and molecular dynamics simulation. Appl Surf Sci 2025; 708: 163743.

59.

Liu

Hao

Wang

, et al. Role of Ca/Si ratio in carbonate-induced tobermorite corrosion: insights from first-principles simulations. Mater Des 2025; 257: 114481.

60.

Zhang

, et al. Study on the self-diffusion coefficients of binary mixtures of supercritical water and H₂, CO, CO₂, CH₄ confined in carbon nanotubes. Water Res 2025; 283: 123856.

61.

Zhang

Tang

. Molecular simulation of superheated gas injection for CH₄ desorption behavior. Energy 2025; 324: 136023.

62.

Xiao

Zou

Tian

, et al. Experimental investigation and molecular dynamics simulation on the effect of ionic strength and pH on kaolinite flocculation using a novel cationic chitosan-based flocculant. Chem Eng Sci 2026; 319: 122262.

63.

Fang

Zhao

, et al. Effects of ice crystals on the proton transport in proton exchange membranes. J Membr Sci 2025; 736: 124622.

64.

Liu

Han

Niu

, et al. Evaluation of biosurfactant wettability on bituminous coal surface: an integrated experimental and molecular dynamics simulation study. Appl Surf Sci 2025; 709: 163780.

65.

Freitas

BJM

Vidilli

Guerra

, et al. Ti-6Al-4 V and Ti-6Al-4V-0.1Ru alloys in oil and gas exploration: a review of performance and hydrogen embrittlement. J Alloys Compd 2025; 1019: 179258.

66.

Wang

Cheng

Chang

, et al. The study on corrosion resistance of Ti-6Al-4 V ELI alloy with varying surface roughness in hydrofluoric acid solution. Metals (Basel) 2024; 14: 64.

67.

Bezuidenhout

Ter Haar

Becker

, et al. The effect of HF-HNO3 chemical polishing on the surface roughness and fatigue life of laser powder bed fusion produced Ti6Al4 V. Mater Today Commun 2020; 25: 101396.

68.

Jiang

Wang

Shi

, et al. Residual stress in thick titanium-alloy butt-welded plates and effect of hydrogen on fatigue crack propagation. Appl Ocean Res 2025; 158: 104543.

69.

Liu

Wang

Gao

, et al. Surface concentration and microscale distribution of hydrogen and the associated embrittlement in a near α titanium alloy. J Alloys Compd 2021; 862: 158669.