Study of the effect of hydrogen content on the mechanical properties in press-hardened boron steel

2008:137 CIV

M A S T E R ' S T H E S I S

Study of the effect of hydrogen content on the mechanical properties in

press-hardened boron steel

Farhad Golkar

Luleå University of Technology MSc Programmes in Engineering

Materials Technology

Department of Applied Physics and Mechanical Engineering Division of Engineering Materials

Abstract

In this master thesis the mechanical aspects of hydrogen embrittlement such as hydrogen content, strength, fracture mode and microstructural changes were studied in press-hardened boron steel.

The specimens were charged electrochemically in a 5 % H2SO4 + 30 mg/l As2O3 solution. By using different current densities and charging times various hydrogen content were obtained.

The hydrogen charged specimens were then subjected to a low strain rate tensile test in order to investigate the effect of hydrogen on the mechanical properties. With increasing hydrogen content a reduced ductility as well as strength could be observed.

Fractography was carried out using Scanning Electron Microscopy (SEM) and light microscopy. Hydrogen assistant fractures were found to be both intergranular brittle and transgranular ductile.

Acknowledgment

This publication is the final part of a Master degree in Space Engineering, specialized in Material Science, at Luleå University of technology (LTU). The work has been performed at R & D department at Gestamp HardTech and the Department of Applied Physics and Mechanical Engineering at LTU. First I would like to thank my supervisor Håkan Andersson at Gestamp Hardtech for the valuable assistant. Further I would like to thank:

• Esa Vuorinen at LTU, for being my academical supervisor.

• Jan Krispinsson and Katarina Lindström, at Gestamp Hardtech, for all your help.

• Johnny Grahn for all the SEM pictures.

• My family and friends for supporting me all the time.

Finally I would like to thank all staff at Gestamp HardTech for your friendly and welcoming atmosphere during my time there.

Luleå, March 2008 Farhad Golkar

LIST OF CONTEST

1 Introduction 3 2 Literature Review 4

2.1 Hydrogen 4

2.2 Press-hardened boron steel 4

2.3 Hydrogen embrittlement 5

2.3.1 Degradation of mechanical properties due to hydrogen 5 2.3.2 Different types of hydrogen embrittlement 6

2.3.3 Hydrogen embrittlement mechanisms 8

2.3.4 Crack characterization and propagation 8

2.4 Introduction of hydrogen into steel 10

2.5 Hydrogen diffusion and solubility 12

2.6 Counteractions against hydrogen embrittlement 17

2.7 Analysis and mechanical testing 18

2.7.1 Analysis of hydrogen in steel 18

2.7.2 Mechanical testing 19

3 Experiments 21

3.1 Press hardened boron steel 21

3.2 Hydrogen charging 22

3.3 Mechanical testing 23

3.4 Analysis 23

4 Results 24

4.1 Hydrogen charging and analysis 24

4.2 Mechanical testing 25

4.3 Fractography 26

5 Discussion 27

6 Conclusion 31

7 Further Work 32

8 References 33

9. Appendices 35

Appendix A 35

Appendix B 36

Appendix C 37

Appendix D 42

1. Introduction

The presence of hydrogen in a material has shown to affect the mechanical properties negatively; for example, the ductility and strength decreases when the hydrogen content increases [16]. The solved hydrogen can either influence the mechanical properties instantaneously, for example decreasing load carrying ability [16], or it can produce delayed failure; that is fracture occurs well after application of a load on the specimen [9]. There exist many types of hydrogen embrittlement types but only three of them; hydrogen reaction embrittlement, internal hydrogen embrittlement, and hydrogen environment embrittlement are frequently mentioned in the literature [11, 13, 15].

There are several processes which the hydrogen can be introduced into the material. Welding with moistened electrodes or heating in atmosphere containing hydrogen are some of processes which should be avoided [10, 11]. The solubility of hydrogen in materials has shown to increase with increasing hydrogen pressure, temperature, transformation in microstructure, and the materials composition [11, 13].

Too prevent catastrophic failure due to hydrogen; one should remove the hydrogen from the material before it is too late, for example by heat-treating the material at right temperature [13]. But to be on the safe side, the material should be checked for its susceptibility to hydrogen. This can be done by first charge the specimen with hydrogen and then test it mechanically [13]. Different testing methods, such as tensile test or creep testing, can be applied to determine the materials sensitivity to hydrogen [13].

2. Literature Review

2.1. Hydrogen

Hydrogen with an atomic mass of 1,00794 g/mol is composed of a single proton and a single electron. Hydrogen is the most simple and abundant element in the universe. It is estimated that 90 % of the visible universe is composed of hydrogen [3]. It is not easy to recognise hydrogen due to its properties such as: being colourless, odourless, non-metallic, and tasteless. Combined with another hydrogen atom, it forms H2, which is very flammable [2].

2.2 Press hardened boron steel

Addition of boron to steel, in a range of 0,002-0,004 %, is a common way to improve steels hardenability [5]. Boron retards the formation of ferrite and pearlite and allows, during the rapid cooling, martensite to be formed, see Figure 1 [5, 6].

Figure 1: Effects of boron on the TTT diagram of steel [6]

Some advantages of boron steels, compared to higher carbon steels and low alloy steels, in hardened condition, are [7]:

• Improved cold formability.

• Improved weldability.

• Lower tempering temperatures, results in saving energy and proper case hardening responses.

2.3 Hydrogen embrittlement

2.3.1 Degradation of mechanical properties due to hydrogen

When hydrogen is absorbed into the material, it can cause a decrease in ductility and load carrying ability. Cracking or catastrophic brittle failure at applied stresses well below the yield strength is also the result of hydrogen presence inside the material [4]. Researcher have showed that with increasing hydrogen content, the ductility and strength decreases considerably, see Figure 3 [12, 17, 20].

Figure 3. Relationship between strength/ductility and hydrogen content for a steel alloy [16]

Hydrogen degradation is more obvious with increasing hydrogen content and with increasing strength of the steel [13, 17, 21].

Once hydrogen has entered a material, it can produce delayed failure (i.e., fracture resulting well after application of a load on the specimen). SHIM et al. describes the delayed fracture:

“Here hydrogen diffusion is an important step; hence a certain amount of time is required during a test to develop a sufficient hydrogen concentration at a location to produce hydrogen-related damage. The method of testing for delayed fracture is simply to apply a static load below the ultimate tensile strength of a notch tensile specimen until failure occurs.

The resulting curve of applied stress vs. time to failure, see Figure 4, provides information on the upper and lower critical stress, incubation time and fracture time. The upper critical stress corresponds to the rupture stress in an ordinary notch tensile test. The lower critical stress or static fatigue limit is the stress below which no delayed failure will occur. The incubation time is that required for crack initiation and the fracture time is the total time until failure.

Generally, the greater the hydrogen concentration, notch acuity and strength level that the steel has, the lower the values of critical stresses, incubation time and fracture time will be”

[9]

Figure 4. Schematic illustration over delayed failure for typical high strength steel alloy [13].

2.3.2 Different types of hydrogen embrittlement

There are three different “types” of hydrogen embrittlement (HE) on steels which is frequently mentioned in the literature; hydrogen reaction embrittlement (HRE), internal hydrogen embrittlement (IHE), and hydrogen environment embrittlement (HEE) [11, 13, 15].

Characteristics and effects of different mechanisms are shortly described by Pilhagen [15] in the Table 1 below:

Table 1: The different types of HE [15]

Type of HE Effects Characteristics Examples

IHE H absorption at elevated temperatures and long exposure

times.

Worst at room temperatures

Almost all alloys, except Al, Cu, and

Ag alloys

HRE Hydride formation.

H2O vapour and CH4

formation.

All alloys contain elements that react

with H2

Ti alloys and C steels

HEE Loss of ductility

Surface cracking Accelerated crack growth

Worst at room temperature

Many Ni-base alloys

HRE

When hydrogen reacts with metals (e.g. Ti, Zr and Nb) and alloying elements, brittle phases and/or vapour are formed (e.g. methane and water). In titanium, HRE occurs due to formation of hydride. On the contrary, the hydrogen does not form hydride in steel.

HRE is an irreversible “process”, i.e. the mechanical properties of the material can not be restored through heat treatment [13, 15].

IHE

When the material is subjected to high hydrogen pressure at elevated temperature for long period of time or during e.g. welding, melting, corrosion, or chemical reactions, the hydrogen is distributed through the bulk material. The hydrogen concentration influences the magnitude of the embrittlement and the cracks initiate mainly inside the material. Using an appropriate heat treatment cycle, the embrittlement can almost be completely prevented (IHE is said to be reversible process) [13, 15].

HEE

Hydrogen environment embrittlement is a result from hydrogen being absorbed by solid metals. Plasticity is required for HEE, because the hydrogen is swept into the material dislocations. Increasing hydrogen pressure results in an increase in HEE. HEE is sensitive to impurities such as oxygen (1 ppm is enough to block the ingress of hydrogen).

HEE is usually most severe at room temperature, mainly because both hydrogen transportation mechanisms (dislocations and diffusions) are active [13, 15].

Strain rate have also shown to influence HEE. With increasing strain rate the HEE will decrease. This is due to the increased motion of dislocations and less hydrogen transport into the material.

2.3.3 Hydrogen embrittlement mechanisms

Of the many suggested mechanisms for the HE, Hydrogen-Enhanced Decohesion mechanism (HEDE) and Hydrogen-Enhanced Localized Plasticity mechanism (HELP) are the two most viable [11, 15, 21].

HEDE

This model is one of the oldest models used to show how the atomic hydrogen changes the material properties. When the hydrogen solubility increases in the tension fields, atomic binding energy of the metal’s lattice decreases. This decrease leads to premature brittle- material fracture along grain boundaries (intergranular cleavage) or transgranular cleavage [15, 21].

HELP

The hydrogen accumulates in the stress fields, i.e. vicinity of the tips of cracks. When an external stress is introduced, dislocation movement is taking place. The existing hydrogen will ease the movement through shielding the emerging dislocation motion against each other and grid defects. The dislocation movement will start at lower stress and thereby increasing the local dislocation mobility, resulting in premature failure [15, 21, 22].

2.3.4 Crack characterization and propagation

M. Wang et al. demonstrated that for hydrogen charged samples, the area fraction of brittle intergranular fracture increased with diffusible hydrogen content [12].

Other experiments have shown that a combination between transgranular ductile and intergranular brittle fracture occurs in hydrogen charged samples, see Figure 5 [17, 19, 25, 28]. Therefore, there is no single fracture micro-mechanism that is characteristic for HE. A completion, for establishing the failure mechanism, is fractographic information such as light microscopy and SEM,.

Figure 5. SEM fractographs, of 3,5Ni- 1,5CR-0,5Mo steel fastener, showing transition from intergranular mode (IG)

to dimple mode of failure.[23]

The crack has also showed that it behaves differently, in different environments. For example, in Figure 6, the cracking behaviour of H-11 steel is illustrated [19, 26].

Figure 6. Slow crack growth in humidified argon (A) and (D) but fast crack growth in water (F) and hydrogen (C). In the presence of oxygen plus argon the crack arrests (B) and (E) [21].

Notable is that oxygen seems to cause total crack arrest, while dry hydrogen and moisture have severe effects on the crack rate. It is believed that oxygen has a greater affinity for iron and therefore forms a protective oxide barrier to block molecular hydrogen to be dissociated in iron [19, 26]. But once the oxygen source is removed, hydrogen can reduce the oxide layer and thereby react with a clean iron surface, resulting once more in increasing crack rate [17].

2.4 Introduction of hydrogen into steel

There are various operations such as; melting, casting, shaping, and fabrication, where the material can interact with atomic or molecular hydrogen [10]. Even under the service life, the material can be exposed to hydrogen. Some common processes where hydrogen is introduced are shown below [10, 11]:

• Arc-welding; if the electrodes contain moisture, water will be decomposed and hydrogen will be liberated.

• Cathodic protection; e.g. with zinc. This procedure is carried out in an electrochemical

• Pickling in acids such as for example hydrochloric -or sulphuric acid. These acids liberate a large amount of hydrogen when they come in contact with the material.

• Heating in an atmosphere containing water vapour or hydrogen.

Whenever a sample, under stress, is exposed to a hydrogen gas atmosphere, hydrogen pickup and associated embrittlement could also be introduced into the metal [11, 13].

Hydrogen can also be introduced, into the material, in controlled forms. This introduction is done, mainly for study how the material is affected by hydrogen. Among various available charging methods, two of them are widely used [16]:

Cathodic method to introduce hydrogen

Samples are charged in an aqua’s solution (electrolyte). Applying external current, hydrogen molecules will be formed on the surface of the sample (cathode). Different electrolytes such as water, NaCl and NaOH can be used, but H2SO4 solved in water is the most common electrolyte. Molecular hydrogen, due to its size, has difficulties to enter the samples.

Therefore, hydrogen recombination poison such as As2O3, which prevents the formation of molecular gaseous hydrogen, is added to the electrolyte [16].

Experiments have shown that with increasing current density and charging time, hydrogen concentration inside the sample will increase [14, 12, 17]. This method is mostly used to study embrittlement of a stainless steels

Gaseous method to introduce hydrogen

Samples will be placed in a furnace with pure or hydrogen rich gas atmosphere. The solubility depends on the partial pressure of the hydrogen; charging normally occurs under high pressures (30 – 50 bar). By increasing the temperature, the charging time will be decreased.

This method is usually used to study embrittlement of high strength pressure vessels [16].

2.5 Hydrogen diffusion and solubility

Hydrogen diffusion and solubility of hydrogen are factors that influence the amount of solved hydrogen in a metal. Factors such as the hydrogen pressure, temperature, microstructure, and composition of the material also influence the solved hydrogen content. At ambient temperature and pressure, solved hydrogen is typically up to few ppms [11, 13]. Hydrogen dissolves interstitially in metals by occupying octahedral and/or tetrahedral interstices [9].

Diffusion

Inside the metal, hydrogen migrates from one interstitial position to a neighbouring one, which is empty. The reason why diffusion occurs is that the metal strives to decrease the Gibbs free energy. Diffusion increases when energy, usually in the form of heat, is injected to the material and diffusion will be going on until its chemical potential, throughout the material, is even. Mathematically it can be described by Fick´s first and second laws.

Skogsmo uses a simplified equation, see Eq 1, for calculating the hydrogen penetration depth of hydrogen in metals [13]:

t D

X = 2 * *

where X is the hydrogen penetration depth, D is diffusion coefficient (usually between 10^-6- 10^-7 cm²s^-1 for steel) and t is time in seconds. The diffusion coefficient is calculated with the Eq 2 [11]:

RT Q

e D D

−

=

where T is temperature, Q is activation energy, D0 is diffusion constant and R is gas constant.

Pressure influence

The relation between hydrogen solubility and pressure is given by Sievert’s law (Eq. 3) [11, 13]:

P k

C = *

where C is dissolved hydrogen concentration, P is the hydrogen gas pressure and k is a constant. The relation is not valid, for simple carbon-steels, at pressures above 120 atm, and temperatures above 200 ^oC. This is because hydrogen will attack the steel and form methane which does not diffuse out of the metal and instead will be collected in the voids, resulting in high pressure and initiating cracks in the steel [11, 13].

Microstructural influence

The hydrogen solubility depends also on the structure of the metal; it is higher in FCC iron than BCC. Therefore, hydrogen rejection can occur during the martensite transformation and at the same time give rise to cracking and flaking [11]. The effect of the steel composition, on hydrogen solubility, is less marked than that of microstructure. A comparison between the effects is shown in Table 2 below. [11, 13]

Table 2: Hydrogen solubility in iron-base alloys at 400 °C and ~1,33 Pa [11, 13]

Alloy H Solubility (ml/100g)

Pure iron (bcc) 0,7

Fe-13%Cr (bcc) 0,4

Fe- 13%Cr ( fcc) 4,8

Fe- 18%Cr-10%Ni ( fcc) 5,8

Temperature influence

At room temperature, the hydrogen solubility in pure iron (α-Fe) is very low. The hydrogen solubility increases slowly with increasing temperature (up to 900 °C). At 910 °C the iron transforms to γ-iron (FCC structure) which clearly leads to an increase in solubility. The solubility increases gradually until γ-iron transforms to δ-iron (BCC). After the transformation the solubility decreases considerably. At the transition from solid iron to molten iron (at temperatures around 1535 ^oC), the solubility increases once again, see Table 3 and Figure 7 [13, 11]:

Table 3: Hydrogen solubility in iron at 1 atmospheric pressure [13].

Physical form Temperature (°C) Solubility, cm³ hydrogen in 100 g iron

α-Fe (BCC) 20 ”very low”

α-Fe (BCC) 900 ~3

γ-iron (FCC) 920 ~5

Solid Fe 1535 ~13

Molten Fe 1535 ~27

Figure 7. Equilibrium iron-hydrogen phase diagram [11]. The hydrogen solubility increases at 910 ^oC and decreases at 1400 ^oC.

Liquid phases such as molten iron or steel can solve more hydrogen than solid phases. This means that weld metal can absorb large amounts of hydrogen, which results in embrittlement or cracking when solidified.

Hydrogen trapping

The hydrogen solubility, in metals, is also influenced by the local distortions or impurities in the crystal. Darken and Smith showed that cold-working of steel, will increase the saturation concentration of hydrogen and that the lattice defects, introduced by cold work, would trap hydrogen. As a result, diffusion of hydrogen will be retarded [9].

Inside the material, the hydrogen can be trapped by binding to impurities, structural defects, or micro-structural entities such as carbides in alloys. This binding can be attributed to stress fields, temperature gradients, chemical potential gradients, physical trapping, or electrostatic forces (proton-defect interaction). The hydrogen traps are either reversible (dislocation, stacking fault), or irreversible (grain boundaries, non-metallic inclusions, precipitates, or individual solute atoms) [9, 11, 13].

Reversible traps are those in which the hydrogen is trapped weakly. Weak traps means that Eb < Es, where Eb is the binding energy between hydrogen and the structural defects and Es ~ 29kJ/mol is the heat of solution (difference between absorbed and released energy) of hydrogen in iron. Irreversible traps (strong traps) are defined as those where Eb > Es and moderate traps are defined as those where Eb ~ Es [11]. Some binding energies, for different traps, are given in Table 3 together with schematic illustrations of the traps in Figure 8 [11].

Table 3: Different traps with correspondingly binding energy Eb and trap number density NT [11].

Trap Eb (kJ/mol) NT(m^-3) Interstitial solutes (N, C) ~ 3-15 10²⁵

Vacancy 46 <10²⁷

Elastic stress field 20 10¹⁹-10²⁶ Dislocation core (screw) 20-30 10¹⁹-10²⁶ Dislocation core (mixed) 59 10¹⁹-10²⁶ Grain boundary ~ 59 10¹⁹-10²³ Fe3C interface 84 10²⁴-10²⁵

Figure 8. (LEFT): Schematic illustration of hydrogen traps types, also showing how dislocation motion can transport external hydrogen to traps. (RIGHT): Potential energy diagram for hydrogen in iron .Em is the lattice

migrational energy for hydrogen [11].

The role of hydrogen traps in embrittlement is twofold. The weakly trapped hydrogen can quickly diffuse and accumulate at crack tips. in the crack propagation stage. The weakly trapped hydrogen is said to be the main reason behind the delayed fracture mechanism. If optimum irreversible traps densities are introduced to the material, the amount of diffusible hydrogen will decrease and also the susceptibility to delayed cracking [11, 13].

However, if the hydrogen concentration exceeds a critical concentration at a defect, hydrogen cracking can occur at strong hydrogen traps sites. The shape and coherency of the defect influence the critical concentration [11, 13].

2.6 Counteractions against hydrogen embrittlement

Degassing the hydrogen out of the metal, before the cracking has started, suppresses the hydrogen cracking. There are different points to take in consideration, regarding the hydrogen content in a metal [13]

• Reducing the corrosion rate will reduce the hydrogen evolution.

• Controlling plating conditions will control the hydrogen pick-up.

• Baking will remove absorbed hydrogen.

• Substitution of the material with one which is less susceptible can also decrease the hydrogen absorption.

• For welding, it is suggested to use electrodes with as low moisture contents as possible.

For some susceptible alloys (after chemical or electrochemical treatments, where hydrogen is produced), the absorbed hydrogen is often removed by being subjected to a heat treatment cycle. For welding cases, especially for high-strength steels, often pre- and post heating is applied, which allows the hydrogen to diffuse out.

By heat-treating the material, hydrogen will either diffuse out or “divide even” in the whole material. One should choose as high temperature as possible (in respect to the material) for hydrogen diffusion. The recommended temperature range is between 180-240 ^oC for the duration of 3-24 hours. For example, removing the hydrogen after zinc-coating will occur at temperatures between 190-210 ^oC with duration of at least 4 hours (if the thickness of the specimen is less than 12 mm). For temperatures above 210 ^oC there is a risk of zink oxidation.

[13]

In cases where the steel has a coated-layer (Zn, Cd or Cu), the hydrogen will likely divide even inside the steel instead of diffusing out. Skogsmo states that with a layer-thickness of 0,5 µm, diffusion decreases by 50 %. With a thickness of 2-5 µm, diffusion out of the steel will totally stop, unless there are mechanical produced pores or cracks in the layer [13].

2.7. Analysis and mechanical testing

It is important to examine the sensitivity of a material against hydrogen embrittlement, so that catastrophic situations such as brittle failures, at stresses below the yield stress, do not occur.

Therefore, one should as a thumb rule, check [13]:

• If the base material is sensitive to hydrogen.

• If hydrogen is formed at the surface treatment process.

• If the base material influences the hydrogen take up.

• If the surface treated details have been affected by the hydrogen.

There is no universal model to establish whether details have reached the risk zone for embrittlement or not. To determine the amount of hydrogen in a material, chemical or physical analyses can be used [13].

It is important to distinguish the difference between residual and diffusible hydrogen.

Residual hydrogen is trapped in the metal, whereas, at or near room temperatures the diffusible hydrogen is considered to be mobile [11]. The sum of these two fractions is defined as total hydrogen. The common hydrogen analyzers are designed to determine the residual or total hydrogen in a specimen. However, some experimental techniques have also been developed for localized analysis of hydrogen [11, 13].

One of the methods, to measure the hydrogen content, is hot extraction [11]. In hot extraction, the sample is heated in a crucible in an inert gas stream, releasing all the hydrogen in gas form. The hydrogen content of the inert gas stream is continuously monitored by a thermal conductivity detector. In some instruments, the surface (diffusible) hydrogen is determined separately by dividing the integration in two parts. First the diffusible hydrogen is released at relatively low temperatures. The strongly trapped hydrogen is released at high temperatures [11].

2.7.2 Mechanical testing

There are many test methods available to determine if a material is sensitive to hydrogen embrittlement. These tests can be performed in hydrogen environment or after hydrogen introduction into the material.

Since every test method has its advantage/disadvantage, the method should be chosen carefully, with respect to the purpose with the examination. Metallographic studies of the fracture and chemical analysis are useful complements to mechanical testing [13]. The common mechanical testing methods are shortly presented here:

Creep testing

The sample is exposed to a constant load until failure. The duration of the test varies from 1 day up to several thousand hours. Usually the testing occurs in intervals, where the load is increased with each interval. The highest load before failure will then depict how the sample is sensitive to the hydrogen embrittlement [13].

Tensile test

One of the most common mechanical stress-strain tests is the “tensile test”. The tension test can be used to determine several mechanical properties of materials such as the ultimate tensile stress, the fracture stress, and the elastic modulus. A gradually tensile load is applied uni-axially along the long axis of the specimen until fracture occurs [1]. The advantage of this method is that it will give a quick answer. It is mainly the failure elongation which deflects the embrittlement. Comparing different hydrogen charged samples, requires same strain rate, since the strain rate can influence the result. For example at high strain rate, the hydrogen has no time to diffuse to areas with high tensions [13].

C-ring testing

A ring formed sample, like a C, with a notch is fastened with a bolt, see Figure 9. The size of the load is determined by connecting the bolt to a “strain-receiver”. First the failure-load is decided and then further samples are loaded up to 75 % of failure-load and load time is up to 200 hours [13].

gure 10. A ring is stretched with a pole [13].

Figure 9. The notched ring is fastened with a bolt, forming a C. [13].

Ring-test

This method is used to find out if a process on the production line, for example surface treatment-process, is the hydrogen source. A ring shaped sample will be treated in different processes and afterward a pole is placed inside the ring, see Figure 10, so that tension is created. The tension is about 90% of the ultime strength and the sample should not fail within 168 hours [13].

Fi

3. Experiments

3.1 Press hardened boron steel

Boron 02 steel is the trade name of the boron steel used in this thesis. It’s a fine grained boron steel produced by SSAB (Swedish Steel AB). In table 5 the measured chemical composition is shown for the Boron 02 steel. The CCT-diagram for Boron 02, see Figure 11, is provided from the manufacturer and is based on austenitization at 900 °C for 300s.

Table 5. The measured chemical composition, in weight percent, for Boron 02.

C (%) Si (%) Mn (%) P (%) S (%) Cr (%) B (%)

Measured 0,25 0,26 1,3 0,018 0,005 0,208 0,0036

Figure 11. The manufacturer’s CCT diagram for the Boron 02 steel.

The boron steel was first press hardened, i.e. the material was heated up to 900°C and then quenched to room temperature. With the help of laser cutting, standard test pieces were then shaped as Figure 12. This geometry was chosen because, concentrated stresses within the

gauge length, is desired during tensile testing. A milling machine was then used for milling away approximately 1 mm of the laser cut edges.

Figure 12. Standard test piece for the tensile testing

3.2 Hydrogen charging

The hydrogen was introduced, electrochemically into the samples, with a solution of 5%

H2SO4. To prevent hydrogen recombination, the electrolyte was poisoned with 30 mg/L As2O3.

Constant current densities between 0,1 - 10 mA/cm ² were applied for various charging times.

The electrodes consisted of one platinum sheet, as anode, and the sample, as cathode. The hydrogen charging was performed in a cylindrical pipe connected in series with a resistor and power supply, see Figure13.

The wanted current densities were first calculated (Appendix A) and the charging times were chosen to 2, 3, and 12 hours.

Figure 13. (LEFT): cathodic charging configuration. (RIGHT) The electrochemical cell used for hydrogen charging.

The samples were rinsed in ethanol before and after the charging. After charging, the samples were kept in liquid nitrogen until the tensile test.

3.3 Mechanical testing

The tensile testing was performed at Gestamp Hardtech. The samples were mounted by its ends into the holding grips of the tensile testing machine. The strain rate used was set to 0,625 mms^-1 .The elongation and instantaneous applied load were measured by an extensiometer and a load cell respectively. As the sample started to elongate, the connected computer simultaneously plotted the figure of stress versus strain. When fracture occurred the instrument stopped and the program presented the Yield strength (Rp0,2) and Ultimate strength (Rm).

A hardness measurement (Vickers) was also carried out for some of the samples.

3.4 Analysis

Directly after the tensile test, a piece of the sample was cut from one of the two broken surfaces. It was kept in liquid nitrogen until the measurement of hydrogen content was performed. The hydrogen analysis was carried out at KIMAB, using the vacuum extraction method.

To investigate the fracture mechanisms of the samples, observation in scanning electron microscopy (SEM) and light microscopy was chosen.

4. Results

4.1 Hydrogen charging and analysis

The samples were milled mainly because to be sure that the edges had not been affected by the laser cutting. After milling the samples took form as in Figure 14.

During the hydrogen charging, bubbles could be seen inside the electrolyte, rising to the surface. The reactions that occurred on the cathode and anode surfaces were:

Cathode: 2H⁺ + 2e^- Æ H2 (gas) Anode: SO42 -2e^- Æ SO2 +O2

The hydrogen content, introduced by cathodic charging, varied with current density and time, see Table 6 (also in Appendix B). The highest concentration obtained was 2,82 ppm when charging occurred at current density 0,75 mA/cm ² for 12 hours.

No macroscopic cracks could be observed before, during, or after the charging. Some samples even became rust-coloured after charging, see Figure 14.

Figure 14. (Left): Before hydrogen charging. (Right): After charging.

Table 6. Hydrogen content for the different samples.

Sample

i [mA/cm²]

time [h]

Diff H [ppm]

Residual H [ppm]

Total H [ppm]

R1 (ref) 0 0 0,04 0,32 0,36

R2 (ref) 0 0 0,04 0,28 0,32

A 5 2 1,14 0,86 2

B 5 2 1,9 0,71 2,61

C 10 12 1,45 0,85 2,3

D 10 12 1,66 1,03 2,69

E 1 12 0,88 0,57 1,45

F 1 12 0,84 0,54 1,38

G 1 12 1,14 0,28 1,42

N 1 12 0,78 0,69 1,47

H 1 3 1,73 0,8 2,53

I 1 3 0,61 1,54 2,15

J 1 3 1,6 0,6 2,2

L 5 2 1,73 0,8 2,53

M 5 2 0,61 1,54 2,15

O 0,5 2 1,61 0,64 2,25

P 0,5 2 1,64 0,55 2,19

Q 0,5 2 0,56 0,58 1,14

R 0,1 12 1,07 0,65 1,72

S 0,1 12 1,93 0,59 2,52

T 0,5 12 1,73 0,47 2,20

V 0,5 12 0,61 0,39 1,00

Z 0,75 12 2,07 0,58 2,65

X 0,75 12 2,20 0,62 2,82

4.2 Mechanical testing

Hydrogen charged samples showed clearly a decrease in tensile strength. Figures in Appendix C show that the ductility has decreased clearly in the hydrogen charged samples. Samples A, I, J, O, and Q were the only samples with strain values larger than 0,2 %, see Figures C1-C10 in Appendix C.

In Appendix D the mean degradation of Ultimate strength versus diffusible, residual and total hydrogen contents are presented. The degradation of ductility versus hydrogen content can also be seen in Appendix D.

The hardness measurement did not show any clear difference between hydrogen charged samples and uncharged samples see table B2 in Appendix B.

4.3 Fractography

The macroscopic examination directly after tensile test showed that for the hydrogen charged samples the broken surfaces were rough and uneven. For the uncharged samples an inclination of 45^o could be seen, see Appendix E.

The fractography carried out by SEM, showed that mixture between intergranular brittle and transgranular ductile fractures had taken place in the samples, see Appendix E.

Some of the samples were observed in light microscope and the result showed that for the hydrogen charged samples, secondary cracks had occurred and that sulphides were in the direction of fracture surface, see Appendix E.

5. Discussion

It was decided, together with supervisors, that the samples should be hydrogen charged in an electrochemical cell. The literature review showed that H2SO4 + As2O3 as electrolyte have worked very well for different steel alloys. Various hydrogen contents had been obtained, for different steel alloys, dependent on charging time and current densities [9, 17, 25]. However, no information could be found for Boron 02 susceptiblity to hydrogen. So to be sure that the electrochemical method would work on Boron 02, test samples charged with hydrogen was sent to KIMAB for evaluation. The results from KIMAB showed that Boron 02 was successfully hydrogen charged and that high hydrogen contents were obtained.

After charging the real samples and tensile testing them, the samples were sent to KIMAB for hydrogen measurement. Dependent on current density and charging time, between 1 – 2,8 ppm (total hydrogen) was obtained. The sample X was charged to the highest level (2,82 ppm total hydrogen) and sample V the lowest (1,00 ppm total hydrogen). Noticeable was that two random samples, which were charged with same current density and charging time, could obtain different hydrogen content. Samples L and M for example showed a difference of about 1,1 ppm (diffusible hydrogen) and 0,4 ppm (total hydrogen), when charged in 5 mA/cm² for 2 hours. First thought was that, by increasing the charging time, the variation would be less. This would give hydrogen enough time to diffuse into the material and therefore make the variation less. Samples R, S, T, V, Z, and X which were charged for 12 hours still showed a considerably large variation. For example samples T and V, gave a difference of about 1,2 ppm (total hydrogen) and 1,1 ppm (diffusible hydrogen). This kind of results were not unique; KIMAB could also observe the same effect, especially when using low current density and low charging time [16]. An explanation for this behaviour has not been found.

The desire to cover hydrogen contents between 0 – 1 ppm (total hydrogen) showed to be difficult. It did not help to decrease the current density and time. This can be due to the usage of As2O3, which releases great amount of atomic hydrogen in the electrolyte. No charging was done without As2O3 in the electrolyte.

Finding a relationship between current density/time and hydrogen content in BORON 02 was difficult due to the variation in hydrogen content. KIMAB also showed that for some steel alloys they couldn’t find the correlation either [16].

The hydrogen charged samples reacted as expected; i.e a decrease in ductility and strength with increasing hydrogen content which other researchers also had seen in other hydrogen charged steel alloys [17, 19, 27, 28]. The stress-strain curve for the uncharged samples R1 and R2, see Figure C1, followed a typical ductile pattern (elastic- and plastic zone). When the hydrogen charged samples were tested they followed the same path, independent of the hydrogen content. The difference between hydrogen charged and the uncharged samples were that the charged samples failed well below the yield- and ultimate strength values, see Appendix C for all the figures. Three cases were found which failed near ultimate strength:

samples A, I, and Q. This result was surprising mainly because sample A and I had total hydrogen contents of 2 and 2,15 ppm respectively. No logic answer was found for this.

When plotting strain versus hydrogen content (diffusible, trapped and total) a clear decrease in ductility was observed with increasing hydrogen content, see Appendix D. By curve fitting the dots an exponentially decreasing line were found. Same curve fitting was found for hydrogen charged notched samples [25]. But KIMAB showed a different curve fitting in their experiments; they used 80 % of the “original ductility property” as a safe point. When the curve reached this point it starts to drop dramatically, see Figure 3 [16]. The start of the curve drop was defined as the critical hydrogen content for the material [16]. This kind of curve fitting could be done by hand in our experiments also, but a mathematical expression could not be found.

A clear curve fitting could not be done easily in the stress versus hydrogen content plots, which differs for example from what KIMAB obtained [16]. On the other hand, a mean decrease of the fracture stress with increasing hydrogen content could be observed, see Appendix D.

Samples with similar hydrogen content showed small variation in ductility but exceptions were found. For example, samples I, M, and V had the same diffusible hydrogen content (0,61 ppm) but sample M and V failed earlier than sample I. Similar examples were found when

The hardness measurement couldn’t be compared to other works in this field, mainly because no literature was found where the hardness of the material, specifically, was measured. The measurements were done on several samples and no noticeable variation could be seen in any sample.

To establish if hydrogen have influenced the failure mechanisms, macro- and microscopically examination was necessary. By initially examine the samples macroscopically, I hoped to get a first indication of the failure mode. I knew that a difference between ductile and brittle failure modes exists; a ductile fracture occurs with large amounts of deformation while the main characteristic of a brittle failure is that the energy is absorbed by the creation of fracture surfaces. Further, brittle fractures usually also show more fragments than ductile fractures [23].

Macroscopical examination of the samples, gave various results; the references showed clearly that shearing had occurred at failure (approximately 45 ^o incline), see Figure E1 in Appendix E. The hydrogen charged samples, on the other hand, showed no or very little shearing (which is special for ductile fracture [23]), see Figures E2-E5 in Appendix E. The limited shearing which had occurred on the charged samples was different from the uncharged samples; the incline was smaller and in another direction. Further, the charged samples showed a rough and uneven failure surface. SEM pictures were also taken, from the side, for some of the samples. This was done because if the samples had failed in a ductile way it would be deformation [24]. The Figure E27 in Appendix E shows that for the uncharged sample, a deformation has occurred and a neck has been formed. The charged samples, Figures E25 and E26 in Appendix E, showed little or no deformation at all.

So the conclusion, after macroscopically examination, was that some kind of mixture between brittle and ductile failure had occurred in the specimens.

The next step was to examine the samples microscopically with the help of SEM and light microscopy. Fractography of the uncharged samples showed that the sample had failed in a ductile transgranular mode. The Figures E6 and E7 in Appendix E showed ductile dimples which are characteristic for a ductile steel alloy. The charged samples, on the other hand, showed that a mixture between intergranular brittle and transgranular ductile fractures had occurred, see the samples in Appendix E. This mixture has also been observed by other researchers [17, 28, 29]. According to Colangelo et al., microcracks initiate internally, often

near inclusions or other interfaces, and propagate intergranularly for an undefined distance [24]. Further, the regions between the adjacent microcracks fail in a ductile manner and present evidence of dimples [28, 29]. Final failure occurs, due to overstress, first after that the cross-section has been sufficiently reduced, and the load bearing capability is sufficiently damaged [24].

When the charged samples were examined in light microscopy, they showed what looked like secondary cracks, see Figures E29 and E30 in Appendix E. These kinds of secondary cracks had been observed by KIMAB as well [16].

Light microscopy showed also that sulphides were in the parallel direction of the failure surface, which also the observation in SEM showed, see Figure E10 in Appendix E. These sulphides are thought to operate like strong hydrogen traps resulting in HIC initiation. The strong hydrogen traps are often due to the existing gap between the sulphide and the matrix, which generates high local pressure that can cause the initiation of HIC [16].

6. Conclusion

The following main conclusions are the results obtained in this master thesis work:

• The cathodic charging worked properly for Boron 02 steel.

• Tensile testing showed that the mechanical properties of Boron 02 steel had been affected negatively by hydrogen. A decrease in ductility and strength was observed with increasing hydrogen content.

• A relationship between hydrogen (diffusible, residual and total) content and ductility was found. For ultimate strength versus hydrogen content a mean degradation could be seen.

• Fractography showed that both intergranular (brittle) and transgranular (ductile) fracture mechanisms have occurred in the samples.

• The material hardness has not been affected by hydrogen

7. Further work

Future work which can be of interest, in the investigation of the effect of hydrogen content on the mechanical properties in boron steel, are:

• The electrochemical method (cathodic) for hydrogen charging can be more effective if the optimal parameters are found.

• Investigate the behaviour of the BORON 02 at very low hydrogen contents (0- 0,5 ppm diffusible).

• Find the reason why tensile strength versus hydrogen content results (Figure D4 - D6) appears as they do and find a way to minimize the scattered results.

• Finding the optimal post-heat treatment parameters, for minimizing the risk of hydrogen embrittlement.

• Investigate if the processes, used by Gestamp HardTech, can cause hydrogen embrittlement in BORON 02.

8. References

[1] Materials science and engineering: an introduction (6^th edition), ISBN-0-471-22471-5, William D.Callister Jr., p:586-588, 113-114, [2003]

[2] Nationalencyklopedin, www.ne.se /väte /, [2007-09-19]

[3] Hydrogen, Thomas Jefferson National Accelerator Facility, http://education.jlab.org/itselemental/ele001.html, [2007-11-12]

[4] Hydrogen embrittlement: An overview from a mechanical fastenings aspect , The fastener engineering &

research association

http://www.fera.org.uk/pdf/Fera%20activity%20listing%20-%20hydrogen%20-%20march%2006.pdf [5] Hardtech Gestamp AB, internal documentation about boron steel

[6] Boron in Steel, G. Haywood & D. Naylor, Steeluniversity.org,

http://www.steeluniversity.org/content/html/eng/default.asp?catid=163&pageid=2081271726, [2007-09-19]

[7] Boron Steels, Corus, http://www.corusgroup.com/en/products/semi_finished_steel/ccandi_boron_steels [2007-09-19]

[8] Boron in Steeel, Daniel H. Herring, Industrial Heating; The international Journal of Thermal technology, http://www.industrialheating.com/CDA/Articles/Column/BNP_GUID_9-5-2006_A_10000000000000100455 [2007-09-19]

[9] A study of hydrogen embrittlement in 4340 steel, I: Mechanical aspects, Materials Science and Engineering A123 (1990) 169-180, I.-O Shim & J.G.Byrne

[10] Nationalencyklopedin, www.ne.se /väteförsprödning/, [2007-11-07]

[11] Hydrogen in steels- a literature study, KIMAB, IM-2002-534, B. Hutchinson, A.Bengtson, R.Pettersson, S.Zajac, T.Siwecki, & L. Ryde

[12] Effect of hydrogen and stress concentration on the notch tensile strength of AISI 4135 steel, Materials science and engineering A 398 (2005) 37-46, M.Wang E. Akiyama & K. Tsuzaki,

[13] Väteförsprödning - mekanismer, orsaker och åtgärder, Institutet for Verkstadsteknisk forskning. ISSN 0349- 063, Jan Skogsmo [1997]

[14] An electrochemical science, ISBN- 0-85109-410-4, J.O´M Bockris, N.Bonciocat & F.Gutmann, p:93-98 [1974]

[15] A literature review of the stainless steel 21-6-9 and its potential for sandwich nozzles, Luleå tekniska Universitet, ISSN: 1402-1617, Johan Pilhagen.[2007]

[16] Critical hydrogen level for mechanical property degradation of high strength steels, KIMAB, IM- 2004-538, O.Wang, S. Adolfsson, & T. Siwecki

[17] Deformation and fracture mechanics of engineering materials, 4^th Edition, ISBN-0-471-01214-9, Richard W. Hertzberg, p: 488-492 [1996]

[18] Hydrogen embrittlement, Universtität des Saarlandes, Saarbrücken http://www.uni-saarland.de/fak8/wwm/research/phd_barnoush/hydrogen.pdf [19] Hydrogen delayed cracking of high-strength weldable steels, Janusz Cwiek, http://www.pg.gda.pl/mech/kim/AMS/012005/ams01200501.pdf

[20] A study of internal hydrogen embrittlement of steels, Material science and engineering, A286 (2000) 269- 281, G.P Tiwari A.Bose et al.

[21] Hydrogen embrittlement and stress corrosion cracking, American society for metals, ISBN 0-87170-185-5, A Troiana Festchrift, p: 13-15 [1984]

[22] Hydrogen related material problems, Topics in applied physics, vol.73, ISBN 978-3-540-61639-9, H. Vehoff

[1997]

[23] Hydrogen embrittlement of 3.5Ni–1.5Cr–0.5Mo steel fastener, Engineering Failure Analysis 15 (2008) 431–439, Abhay K. Jha , P. Ramesh Narayanan, et al.

[24] Analysis of Metallurgical failures, ISBN 0-471-16450-X, Colangelo & Heiser, p: 64-126 [1974]

[25] Fracture criterion for hydrogen embrittlement of high strength steel, Materials Science and Technology;

Feb2006; 22; 2; ProQuest Science Journal, M.-Q. Wang, E. Akiyama & K. Tsuzaki [26] Electrochemical Data, ISBN 0-444-99863-2, D. Dobos, p:27, 42 [1975]

Appendix A

Calculation of the current density:

1. The mean surface area, S, of the sample and Pt-net was first calculated.

2. Multiplying the areas with the desired current density, ICD, gave the needed current, I, see Eq A1.

(Eq A1)

S I I =

*

I = the current, [mA]

ICD = the current density, [mA/cm²]

S = the surface area of the electrode (mean cross-sectional area of the solution), [cm²]

3. calculate the resistance of the electrolyte, RE, by the Eq A2 [26]

S R

1 * l

= κ

RE = resistance, [Ω]

κ = the conductivity of the electrolyte, [Ω^-1m^-1] l = the mean distance between the electrodes, [m]

S = the surface area of the electrode (mean cross-sectional area of the solution), [m²]

4. The resistor, R1, was calculated using Ohm’s Law, Eq A3. The main voltage, E, was set to 2 V

R

I

R

= E −

Appendix B

Table B1. Hydrogen charging and tensile test result

Sample

i [mA/cm²]

tid [h]

Diff H [ppm]

Residual H [ppm]

Total H [ppm]

Rm [N/mm²]

Extension*

A₅₀ [%]

R1 (ref) 0 0 0,04 0,32 0,36 1616 5,92

R2 (ref) 0 0 0,04 0,28 0,32 1610 7,46

A 5 2 1,14 0,86 2 1459,3 1,28

B 5 2 1,9 0,71 2,61 671 0,34

C 10 12 1,45 0,85 2,3 793 0,42

D 10 12 1,66 1,03 2,69 793 0,42

E 1 12 0,88 0,57 1,45 924,4 0,59

F 1 12 0,84 0,54 1,38 995,4 0,73

G 1 12 1,14 0,28 1,42 907,46 0,5

N 1 12 0,78 0,69 1,47 980,55 0,49

H 1 3 1,73 0,8 2,53 689,6 0,34

I 1 3 0,61 1,54 2,15 1634,9 2,08

J 1 3 1,6 0,6 2,2 1437 1,06

L 5 2 1,73 0,8 2,53 832,4 0,47

M 5 2 0,61 1,54 2,15 752,5 0,4

O 0,5 2 1,61 0,64 2,25 1295,2 0,79

P 0,5 2 1,64 0,55 2,19 748 0,38

Q 0,5 2 0,56 0,58 1,14 1689,9 4,34

Rx 0,1 12 1,07 0,65 1,72 1482 2,14

S 0,1 12 1,93 0,59 2,52 1487 1,75

T 0,5 12 1,73 0,47 2,20 1147,3 0,76

V 0,5 12 0,61 0,39 1,00 941,3 0,52

Z 0,75 12 2,07 0,58 2,65 601,2 0,31

X 0,75 12 2,20 0,62 2,82 1248,9 0,94

* measured with a extensometer.

Table B2. Hardness mesurments

Sample Edge Middle Edge

R4* 539 542 529

F 525 520 532

R4 572 561 536

P 525 518 522

N 528 515 534

A 551 515 526

J 529 532 526

C 506 519 503

Q 530 569 556

R1 och R2

0 200 400 600 800 1000 1200 1400 1600 1800

0 1 2 3 4 5 6 7 8

Strain [%]

Stress [N/mm2] Ref1

Ref2

A och B [5mA-2h]

0 200 400 600 800 1000 1200 1400 1600 1800

0 1 2 3 4 5 6 7 8

Strain [%]

Stress [N/mm2]

A B

Figure C2. Tensile curves for sample A and B.

Appendix C

Figure C1. Tensile curves for the uncharged reference samples R1 and R2.

C och D [10mA-12h]

0 200 400 600 800 1000 1200 1400 1600 1800

0 1 2 3 4 5 6 7 8

Strain [%]

Stress [N/mm2]

C D

E,F,G och N [1mA-12h]

0 200 400 600 800 1000 1200 1400 1600 1800

0 1 2 3 4 5 6 7 8

Strain [%]

Stress [N/mm2]

E N F G

Figure C4. Tensile curves for sample E, F, G and N.

Figure C3. Tensile curves for sample C and D.

H,I och J [1mA-3h]

0 200 400 600 800 1000 1200 1400 1600 1800

0 1 2 3 4 5 6 7 8

Strain [%]

Stress [N/mm2]

H J I

L och M [5mA-2h]

0 200 400 600 800 1000 1200 1400 1600 1800

0 1 2 3 4 5 6 7 8

Strain [%]

Stress [N/mm2]

L M

Figure C6. Tensile curves for sample L and M.

Figure C5. Tensile curves for sample H, I and J.

O,P och Q [0,5mA-2h]

0 200 400 600 800 1000 1200 1400 1600 1800

0 1 2 3 4 5 6 7 8

Strain [%]

Stress [N/mm2]

O P Q

Figure C7. Tensile curves for sample O, P and M

Figure C8. Tensile curves for sample Rx and S.

Figure C9. Tensile curves for sample T and V.

Figure C10. Tensile curves for sample X and Z.

Appendix D

Figure D1. The effect of diffusible hydrogen on BORON 02 tensile strain. Curve fitting determined by MATLAB

Figure D3. The effect of total hydrogen on BORON 02 tensile strain. Curve fitting determined by MATLAB

Figure D4. The effect of diffusible hydrogen on BORON 02 tensile strength.

0 0.5 1 1.5 2 2.5 3 0

200 400 600 800 1000 1200 1400 1600 1800

Residual H [ppm]

Sess tr[N/ mm2]

Rm vs. Trap

Figure D5. The effect of residual hydrogen on BORON 02 tensile strength.

0 0.5 1 1.5 2 2.5 3

0 200 400 600 800 1000 1200 1400 1600 1800

Total H [ppm]

Sess tr[N/ mm2]

Rm vs. Tot

Appendix E

Figure E1. Macroscopic examination of the references (uncharged samples).

Figure E2. Macroscopic examination of samples B, C, D, and E.

Figure E3. Macroscopic examination of samples F, G, H, I, and J.

Figure E4. Macroscopic examination of samples K, L, M, N, and O.

Figure E5. Macroscopic examination of samples P, Q, R, S, and T

Figure E6. Fractography, of uncharged sample (R1), showing dimples.

Figure E7. Fractography, of uncharged sample (R2).

Figure E8. Fractography, of sample B, showing mixture between transgranular ductile failure and intergranular.

Figure E9. Fractography, of sample C, showing intergranular mode. Micro-cracks can also be seen.

Figure E10. Fractography, of sample C, showing sulphide trail and micro-crack.

Figure E11. Fractography, of sample D, showing mixture between transgranular ductile failure and intergranular

Figure E12. Fractography, of sample D, showing mixture between transgranular ductile failure and intergranular

Figure E13. Fractography of sample H. Micro-cracks can be seen clearly in the picture.

Figure E14. Fractography, of sample L, showing intergranular mode

Figure E15. Fractography of sample L. Intergranular mode together with micro-cracks can be seen.

Figure E16. Fractography of sample N. Intergranular mode together with micro-cracks can be seen

Figure E17. Fractography, of sample N (another position), showing a mixture between intergranular and transgranular ductile mode.