PSI - Issue 13

A.M. Polyanskiy et al. / Procedia Structural Integrity 13 (2018) 1408–1413 Author name / Structural Integrity Procedia 00 (2018) 000–000

1412

5

3. Discussion

The carrier gas and the carrier gas with hydrogen released from the sample are conditioned before being supplied to the thermal conductivity cell. Since gases inside the analyzers are flowing, the conditioners are, in fact, heat ex changers, and the heat exchange between the gas medium and solid surfaces is slow. So, it is extremely di ffi cult to ensure the conditioning of the gas flow with the accuracy of maintaining the temperature about fraction of degree. Where can the di ff erence in the temperatures of gas flows come from? It is obvious that the sample and the crucible, heated up to a high temperature, will heat the carrier gas, which flows in direct contact with the walls of the crucible (see Fig. 2). The amount of heat absorbed by the gas depends on the level of the crucible walls grayness and the level of the quartz glass tube inner walls grayness, which tend to be contaminated with metal sprays and vapors released when the samples are heated to a high temperature. Calculation can be based on the concept of the heat capacity of the crucible, sample and carrier gas flowing near the sample during the analysis of hydrogen concentration. The evaluation shows that even if the carrier gas takes only a half of the energy spent to heat the sample, the carrier gas will overheat hundreds of degrees. The empty crucible will heat the gas about two to three times less than the crucible with the sample, due to approximately two to three times less heat capacity. In this regard, the “blank experience”, which would allow to set the background values of hydrogen, is impossible. Hydrogen is contained even in the degassed metal either inside or on the surface GOST (1999). Therefore, it is impossible to produce a “zero” metal sample with zero hydrogen conctent. This point of view is confirmed both in our Polyanskii et al. (2011) and in independent studies Hassel et al. (2013). In our work Polyanskii et al. (2011) we described the measurements of the hydrogen content in 54 calibration samples carved from one hydrogen content standard of the aluminum alloy D16. The measurements were carried out in a certified factory laboratory by a certified hydrogen analyzer RH402 (LECO manufacturer). Three measurements demonstrated zero result and one demonstrated negative concentration of dissolved hydrogen. On the one hand, this means only 8 % of unsuitable measurements, which must be discarded. On the other hand, the negative concentration in the measurement protocol requires explanation. Our analysis of the thermal conductivity call operation makes it possible to give such explanations. The “failure” of the extraction curve below the background can be caused by the closer fitting of a particular sample to the crucible. Because of this, the heat transfer process of the gas will go faster and the gas will warm up a little more. It is also possible that crucibles have di ff erent graying coe ffi cients. Because of this, the relative proportions of the heat emitted and transferred to the carrier by the crucible will be di ff erent, and, consequently, the gas will be heated di ff erently by heat transfer from the surface of the crucible. This e ff ect was observed by independent researchers. Experimental measurements of hydrogen concentration in empty crucibles, probably obtained by using the Juwe H-mat 221 analyzer are described in Hassel et al. (2013). The report does not specify exactly what analyzer was used, but among the “atmospheric” analyzers listed there, only the H-mat 221 possessed the sensitivity necessary for measurements. It is important that all the extraction curves in Hassel et al. (2013) become negative when the empty crucible is heated. They are located at a di ff erent value below the initial background. This value depends on the frequency of the crucible using. Thus, the corresponding integral with respect to the background gives a negative value. This result confirms our estimates and assumptions. There is one more consideration explains the numerous problems that arise when “atmospheric” hydrogen analyz ers are operated. The relative order of the measured quantity is 10 − 6 . Moreover, the situation can not be fundamentally changed while extraction of hydrogen into the carrier gas, since the mass of the carrier gas mass and the standard sample mass are comparable, for example, 1g and 5g. Our estimation allow us to explain the wide scatter observed when measuring the hydrogen concentration in certi fied standard samples in various certified laboratoriesKonopelk´ o et al. (2017).

4. Conclusion

In this paper we analyzed the reasons of the occurrence of systematic errors in measuring the hydrogen con centration in metals by applying modern equipment. These errors can be seven to eight times more than the real concentration.