Methods and results about hydrogen diffusion in metal.

Bucur, Liviu ; Bucur, Gabriela ; Miclosi, Viorel 等

1. INTRODUCTION

The hydrogen is easier solving in metals and diffuses by interstitial mechanism. In this mechanism, the atoms of the solve element jump from a place at network in another, with implications above tension local stage. If the interstitial atom does not "wait" to be taken by a network trouble (vacancy, granule limit, dislocation) for jump, the diffusion coefficient can be relative big (Voiculescu, 2005).

Grace to surrounding atoms, the activation energy is more and more big and the rate of diffusion is relatively low. For all that, the atoms can diffuse along granule limits, interfaces and free surfaces (creaks, voids) of material (Miclosi, 2003).

This paper presents the original methods for quantity and quality analyse of hydrogen distribution in metals.

2. EXPERIMENTS ON OL37 PIECE TEST

The experimental stall for electrolytic filled on homogeneous OL37 piece test is presented in figure 1. This composed by: glace recipient; 8 graffito elements, serial connected on c.c. power source with 2,5V voltage and 0,7A current; OL37 piece test connected with graffito elements. To realize electrolytic process on consider OL37 piece test like anode and the cathode is graffito electrodes battery. The piece test is immersing in distillation water solution. This electrolytic filled was realized in approximate 48 hours.

[FIGURE 1 OMITTED]

On the end of filled period, the piece test was cleaned and submersed in a recipient with glycerin. After 5 minutes, the first balls of hydrogen begin to appear on glycerin surface. The observations were made for 72 hours. Based on these, we can say that for the first 48 hours, the diffusion was accelerated, corresponding for an exponential function (Bucur et al.,2008).

Until that, the diffusion was deferred, the hydrogen diffuse much more slowly, constantly, until a minimum level.

3. MICROSCOPIC OBSERVATIONS

After the macroscopic observations and quality determinations, was appropriate the microscopic evaluation of hydrogen distribution--figure 2. We know the zones with big concentration of hydrogen inside the metal, is the most important causes of micro cracks.

To realize this objective, were necessarily to following the next steps:

* effect metallographic luster for one of piece faces;

* electrolytic filled for piece test using the presented experimental stall;

* fast piece test extraction from electrolytic bath, drying and metallographic attack with 2,3% HNO3 to distinguish the microscopically structure of that;

* microscopic observation using a performing microscope with immersive objective: for preparing the face we applied a small cider oil drop, witch have the propriety to capture the hydrogen balls and also to race the rate of microscope visibility.

The observations were made for 72 hours. All the pictures were captured with a video camera attached by the microscope.

[FIGURE 2 OMITTED]

These images were made for different moments of time. In figure 2 is presented the metallic structure for the moment time = 5 minutes. On every image is attaching one measure scale with we can approximate the dimension of granule and the number of hydrogen balls on surface of granule and also to the limit between granules.

4. MATHEMATICAL MODELING WITH F.E.M.

For mathematical modeling with finite element method (F.E.M.) of microscopic diffusion process, we considere the following (Iordache et al.,2003):

* metallic granule was geometrically considered a regulated hexagonal prism with square base a = 1.4 * [10.sup.-6] cm and high h = 2.2 * [10.sup.-6] cm;

* physical properties was: heat specific [c.sub.p] = 0.63 x [10.sup.-5] [cm.sup.2]/s * K; steel density [rho] = 7.9 x [10.sup.5] g/[cm.sup.3]; volumic diffusion coefficient [D.sub.V] = 7.5 x [10.sup.-17] [cm.sup.2]/s; intergranular diffusion coefficient Dg = 11.1 x [10.sup.-12] [cm.sup.2]/s;

* discretization of model was made for 20 divisions for each square, using tetraedres with 10 peaks, obtaining 217113 nodes;

* boundary and initially conditions are: [C.sub.1] = 0.1 x [10.sup.-9] [cm.sup.3]/100g; [C.sub.2] = 0.65 x [10.sup.-10] [cm.sub.3]/100g; [J.sub.1] = 0.76 x [10.sup.-15]ml/100g x [cm.sup.2] x s ; [J.sup.2] =[ -0.00210.sup.-15]ml /100g. [cm.sup.2] x s, where [C.sub.1] is initial concentration on base 1 of prism, [C.sub.2] is initial concentration on lateral face 2 of prism, [J.sub.1] is input flux on lateral face 3 and [J.sub.2] is output flux on lateral face 4--presented in figure 3. This values was used for moment time = 5 minutes.

For solve the elementary equations system was used ANSYS 10 Program, adapted for diffusion hydrogen problems, making the analogy between termic transfer equation and diffusion equation.

The diffusion equation is:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], (1)

where D is diffusion coefficient, C is hydrogen local concentration, Hv is hydrogen sources density and x, y, z are the coordinates of considered point (Miclosi at al., 2003).

This paper considere that process running in omogen and isotrop environment. Figure 3 presents hydrogen concentration distribution of granule for t=5 min, simulated with Ansys 10.

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

In the same way was determined the volumic and intergranule hydrogen concentration for different moments of time, starting 5 minutes until 4320 minutes. Figure 4 presents the rapport between volumic hydrogen concentration and intergranule hydrogen concentration. We observe that a big quantity of hydrogen is presented inside the granule; then, a part of this hydrogen migrate over intergranular space and starting approximately with 2880 minute we observe a race of volumic concentration again.

We are interested about that because the massive accumulation of hydrogen in these zones is the main causes of steel cracks.

5. CONCLUSION

As the results shown, the hydrogen diffusion is more intense on inter granular space then surface granule.

We can specify that the hydrogen concentration is roughly calculated because the size and density of metallic granules is not the same in all piece test volume.

In some neomogen environment we will study the same things, using the same experimental methods, but, in this case, one piece test will be normalized, wich means the uniformity of size granules, and another piece test will be warming-up for obtaining a race of granule volume.

The results of these experiments will be used for determination the microscopic hydrogen distribution on limit between welding seam and thermo-mechanic influence zone.

6. REFERENCES

Bucur, L., Bucur, G. & Miclosi, V. (2008). Methods and Preliminary Results about Microscopic Hydrogen Diffusion in Metals. Petroleum--Gas University of Ploiesti Bulletin, Technical Series, Vol. LX, No.3A, November 2008 124-130, ISSN 1224/8495

Iordache, F.; Baltaretu, F. & Caracaleanu, B. (2003). Modelling and Simulation of termic transfer for dynamic processes, MatrixRom Publishing House, ISBN 973-685-612-7, Bucharest, Romania

Miclosi, V. (2003). Treatments Heat Associate Steels Welding Fusion, vol.1, Sudura Publishing House, ISBN 973-8359-10-4, Timisoara, Romania

Miclosi, V., Scorobetiu, L., Jora, M. & Milos, L. (1984). Welding Processes Fundaments, Didactica si Pedagogica Publishing House, Bucharest, Romania

Voiculescu, I. (2005). Hydrogen in Steels for Welded Structures, Printech Publishing House, ISBN 973-718-181-6, Bucharest, Romania