1. Introduction
Titanium alloys are classified according to the β (bcc) phase stability into the alloy classes α(hex), (α + β) and β titanium (Ti) alloys [
1]. Due to the high gas solubility of the β phase and the complete reversibility of the metal–gas reaction, temporary alloying with atomic hydrogen (H) is possible as part of a thermal treatment, a so-called thermohydrogen treatment (THT) [
2]. THT usually consists of the process sequence solution treatment (ST), diffusion-controlled hydrogen uptake (hydrogenation), hydrogen degassing (dehydrogenation) and aging. With few exceptions [
3] it is applied to Ti alloys only, aiming for homogeneous microstructure adaptation [
4,
5,
6,
7,
8,
9,
10,
11,
12] and the generation of microstructural gradient [
13,
14]. A schematic visualization of the process, embedded into the Ti-6Al-4V/hydrogen phase diagram is shown in
Figure 1. The steps of solution treatment (step 1) and aging (step 4) are carried out in ambient atmosphere (red scripture in
Figure 1) at temperatures A and D. Hydrogenation is executed in a gas mixture containing hydrogen (step 2, blue scripture in
Figure 1) at temperature B, and dehydrogenation takes place in vacuum (step 3, black label in
Figure 1) at temperature C. The figure shows two different approaches depending on the H content in the near-surface area of the sample that is established after the hydrogenation. The study follows two different concepts of THT (path A, red dotted lines and Bm blue dotted lines), which differ in terms of maximal hydrogen concentration established in the near-surface area.
Hydrogen-induced effects that influence the microstructure (extrinsic H effects) can have a positive impact on the mechanical properties of the material. The extrinsic H effects can be described by two main phenomena: Firstly, hydrogen is a strong β-stabilizing element in Ti alloys and lowers the β-transus temperature T
β to T
β (H) (see
Figure 1). Hence, the temperature of solution treatment can be lowered, reducing grain growth during the process as compared to conventional heat treatment (solution treatment and aging) of Ti alloys, if hydrogenation is applied before the ST. Secondly, when the hydrogenation concentration in a sample exceeds the maximum hydrogen solubility, hydrogen evokes hydride formation, which is associated with local volume expansion. The hydride-induced volume expansion in the α and β phase reaches a value between 17% and 25%, depending on the underlying solid solution (chemical composition) [
5]. After the hydrides are dissolved during dehydrogenation, dislocations and vacancies remain, which act as additional nucleation sites for precipitates. This phenomenon enhances the α precipitation (and thus a more homogeneous and finer precipitate morphology) as well as the recrystallization kinetics (and thus a potentially reduced grain size) during a subsequent heat treatment [
6]. Previous work on THT has led to increased strength under cyclic and static loading by producing a homogeneous and, compared to conventionally heat-treated Ti alloys, finer microstructure [
7,
8].
The aim of the research project presented is the realization of a local microstructure which, depending on the surface distance, provides a microstructure gradient that improves the properties relevant to technical applications like the strength of components of an aircraft at cyclic loading in HCF and LCF regime. Compared to a homogeneous microstructure, a fine, equiaxed microstructure in the near-surface region is to be established using THT (process comparable to [
9]), which significantly prolongs the fatigue crack initiation phase, while a coarse, lamellar microstructure is to be achieved in the interior, which slows down the propagation of long fatigue cracks [
14,
15]. Berg and Wagner [
16] were able to evoke fatigue life increasing microstructure gradients in Ti alloys using thermomechanical treatments. In contrast to conventional processes enabling surface-near microstructural modifications, such as shot peening or other thermomechanical treatments, THT uses hydrogen from the gas phase as a promoter for the intended local microstructural modifications. Hence, THT enables a local microstructure adaptation of complex geometries, like tubes with a variable wall thickness, that cannot be surface hardened via thermomechanical surface treatments. Therefore, THT is applicable to components which possess a high geometrical complexity and promises an extension of the degrees of freedom for property improvement of applications. The study presented compares the fatigue crack propagation resistance and the fracture toughness in dependence of three different solution treatment conditions. The comparison is used to select suitable solution treatment parameters. The solution treatment must evoke a morphology that is suitable for the surface-far area of the final microstructural gradient [
16]. For the further selection of the hydrogenation and dehydrogenation parameters, different graded microstructural states after hydrogenation and dehydrogenation are evaluated by means of simulated hydrogen concentration (
cH) profiles, measured hardness curves, metallographically recorded microstructural gradients and XRD analysis of the resulting phase fractions.
2. Materials and Methods
The widely used (α + β)-Ti alloy Ti 6Al-4V was used as the test material. For the selection of suitable solution treatment parameters, the solution treatment conditions after [
15] were assessed with respect to the threshold stress intensity range for long cracks (Δ
K0) and the fracture toughness (
KIc or
KQ). Δ
K0 was determined with four-point bending fatigue tests at 100 Hz (stress ratio: 0.1) by using a 4-point bending resonance and a crack length measuring device (Russenberger Prüfmaschinen AG, Neuhausen am Rheinfall, Switzerland) and the load-shedding method, in which the stress amplitude is reduced according to an exponentially decreasing function. Δ
K0 is achieved when the crack propagation rate
da/dN is less than 10
−11 m/load cycle. For the determination of fracture toughness, compact tension tests were carried out at a servo-hydraulic testing machine (MTS Systems Corporation, Eden Prairie, MN, USA) using a strain gauge. The specimen geometries were designed according to the guidelines of ASTM E 647 (Δ
K0) and ASTM E 399 (
KIc or
KQ).
To determine the (de)hydrogenation parameters, (de)hydrogenation of cylindrical samples of 5 mm in diameter and 100 mm in length were heat-treated in a horizontal (vacuum) furnace (University of Siegen, Siegen, Germany) under supply of a He/H
2 gas mixture with 10% H
2 (hydrogenation) or in vacuum (
p = 2 · 10
−5 mbar, dehydrogenation). The sample surfaces were prepared according to [
16], including a Pd coating, that enables hydrogen uptake at comparably low temperatures (500 °C and 600 °C). Moreover, the samples were placed on a Zr foil that acts as a getter of gas impurities, particularly oxygen. The analysis of the hydrogen concentration was conducted by means of carrier hot gas extraction (Leco Corporation, St. Jospeh, MI, USA). The hydrogenation experiments were intended to induce
cH gradients that would promise the desired microstructure after dehydrogenation. Therefore, two approaches (path A and B) were pursued, which should differ regarding the maximum hydrogen concentration generated in the near-surface area of the sample (see
Figure 1). The realization of the different hydrogen concentration values is done by applying different hydrogen partial pressures (path A: 20 mbar, deeper penetration depth; path B: 100 mbar: steeper gradient of microstructural changes).
For the design of path A, samples of the three different initial microstructures, generated by different solution treatments, were hydrogenated under a hydrogen partial pressure (
pH2) of 100 mbar at 500 °C and 600 °C at 2, 3, 4, and 6 h (acc. to [
15]). The parameters for the different solution treatment conditions are listed in
Table 1.
To determine the hydrogenation temperature (TH) and duration (tH), the hydrogenation conditions were selected, which led to the maximum possible cH–value in the near-surface area of the sample without causing surface cracks longer than 100 µm. After the hydrogenation treatment, the samples were water quenched in order to stop a further hydrogen diffusion and release. This should enable a more precise determination of the amount of absorbed hydrogen.
The design of path B is based on the results of Berg and Wagner [
16]. They reported that the near-surface region of the graded microstructure leads to maximum improvement in fatigue strength at a penetration depth of about 750 µm. Hence, hydrogenation experiments were carried out by varying the pressure
pH2 (20, 40, 60, 80 and 100 mbar) to produce a
cH profile in which the local
cH reaches hydrogen concentration values sufficient for hydride formation (>15 at.%) only in a region from the surface to a distance from the surface of 750 µm. The evaluation of the influence of the different
pH2 on the resulting microstructure was done with respect to the maximum possible
pH2 that produced no or negligible surface cracks.
To determine the (de)hydrogenation temperature (
TH,
TD) and duration (
tH,
tD) (de), hydrogenated samples of the three solution treatment conditions dealt with (see
Table 1) were examined by means of SEM (FEI Company, Hillsboro, OR, USA) (used for a qualitative evaluation of the stereological parameters). Accordingly, the referring
cH profiles were numerically simulated using Matlab (R2020, The MathWorks, Natick, MA, USA). The numerical simulation of the
cH profiles consider the
TH- and
pH2-dependent incubation time determined in [
15] and deliver a kinetic surface correction factor (SCF) of the hydrogen adsorption and desorption process, which is the ratio of the experimental and the numerically calculated hydrogenation time. The resulting
cH profiles, the hydrogenation and dehydrogenation times as well as the surface correction factors were calculated according to the method described in [
15] by using finite-element method (FEM) and Matlab (R2020). The input parameters for the simulation consist of thermodynamic data (hydrogen solubility) and kinetic data (hydrogen diffusion coefficient) and are reported in [
15] as well and are displayed in
Table 2 and
Table 3.
The hydrogenation time for path B was calculated by FEM using the surface correction factor of the hydrogenation from path A to estimate the necessary hydrogenation time. Furthermore, local Vickers hardness profiles were determined applying hardness testing (Strues GmbH, Willich, Germany) (2 kp for 10 s). Moreover, the stereological parameters were measured on SEM-BSE micrographs via image processing (Lince 2.4.2 ß, TU Darmstadt, Darmstadt, Germany and ImageJ 1.52a, National Institutes of Health, Bethesda (MD), USA), and X-ray diffraction (XRD) (Malvern Instruments, Malvern, United Kingdom) phase analysis was performed at 45 kV and 40 mA.