Quantitative Hydrogen Analysis of Zircaloy-4 Using Low-Pressure

Mar 7, 2007 - ... Nagatsuta, Midoriku, Yokohama 228-8503, Japan, and Department of Physics, Faculty of Mathematics and Natural Sciences, Bandung Insti...
0 downloads 13 Views 115KB Size
Anal. Chem. 2007, 79, 2703-2707

Quantitative Hydrogen Analysis of Zircaloy-4 Using Low-Pressure Laser Plasma Technique Koo Hendrik Kurniawan,*,† Marincan Pardede,† Rinda Hedwig,† Zener Sukra Lie,† Tjung Jie Lie,† Davy Putra Kurniawan,† Muliadi Ramli,‡ Ken-ichi Fukumoto,‡ Hideaki Niki,‡ Syahrun Nur Abdulmadjid,§ Nasrullah Idris,§ Tadashi Maruyama,| Kiichiro Kagawa,⊥ and May On Tjia#

Research Center of Maju Makmur Mandiri Foundation, 40 Srengseng Raya, Kembangan, Jakarta Barat 11630, Indonesia, Department of Nuclear Power and Energy Safety Engineering, Graduate School of Engineering, and Department of Physics, Faculty of Education and Regional Studies, Fukui University, Fukui 910-8507, Japan, Department of Physics, Faculty of Mathematics and Natural Sciences, Syiah Kuala University, Darussalam, Banda Aceh, Nanggroe Aceh Darussalam, Indonesia, Integrated Research Institute, Tokyo Institute of Technology, Nagatsuta, Midoriku, Yokohama 228-8503, Japan, and Department of Physics, Faculty of Mathematics and Natural Sciences, Bandung Institute of Technology, 10 Ganesha, Bandung 40132, Indonesia

A method for hydrogen analysis is urgently needed in some important fields of investigation such as the fields of material sciences and particularly research for industrial applications in energy generation. For example, the regular examination of hydrogen concentration trapped within the metal surface of a fuel cell is required to maintain its optimal performance. Another much needed application is the detection of hydrogen concentration in

nuclear power plants. In a light water nuclear power station, uranium fuel is contained in zircaloy tubes. During the operation of the reactor, hot water reacts with the zircaloy surface to form zirconium oxide and hydrogen gas. Being in the gas phase, the hydrogen atoms readily penetrate into and accumulate in the zirconium tube. The presence of this trapped hydrogen in excessive amount is known to cause a certain structural damage that reduces the strength of the material. Therefore, the accumulation of hydrogen in a zircaloy tube must be examined periodically. A standard technique for detecting hydrogen in such a case involves tedious preparation of the samples by melting a portion of the zircaloy tube in a carbon furnace, before being subjected to analysis by a gas detector. Thus, it is highly desirable to develop an in situ, rapid, and practically nondestructive or minimally destructive technique that would replace the existing time-consuming and destructive procedure, which also exposes the workers to the risk of radioactive hazard. The most popular version of direct laser-induced plasma spectroscopy developed so far is the one employing atmospheric pressure surrounding gas which is commonly referred to as laserinduced breakdown spectroscopy (LIBS) and was first proposed by Radziemski et al.1,2 Despite its demonstrated applicability in rapid quantitative analysis in various fields of elemental analysis,3 very few studies have been reported on its application to the analysis of hydrogen in metal samples.4 This is due to the unfavorable effects of line broadening and severe emission intensity diminution when the plasma is produced at atmospheric pressure, although these effects have been known to be virtually negligible for plasma emissions of heavy elements.5 This undesirable effect was explained in terms of the shock wave excitation

* To whom correspondence should be addressed. Tel: 62-21-5867663, 62-215867660. Fax: 62-21-5867670 or 62-21-5809144. E-mail: [email protected], Homepage: http://www.mmm.or.id. † Research Center of Maju Makmur Mandiri Foundation. ‡ Department of Nuclear Power and Energy Safety Engineering, Fukui University. § Syiah Kuala University. | Tokyo Institute of Technology. ⊥ Department of Physics, Fukui University. # Bandung Institute of Technology.

(1) Loree, T. R.; Radziemski, L. J. Plasma Chem. Plasma Proc. 1981, 1, 271. (2) Radziemski, L. J.; Loree, T. R. Plasma Chem. Plasma Proc. 1981, 1, 281. (3) Laserna, J. J., Omenetto, N., Eds. Laser Induced Plasma Spectroscopy and Applications (LIBS 2004 Third International Conference), Spectrochim. Acta 2005, 60B, 877, and references therein. (4) Colonna, G.; Pietanza, L. D.; Capitelli, M. Spectrochim. Acta 2001, 56B, 587. (5) Lie, T. J.; Kurniawan, K. H.; Pardede, M.; Suyanto, H.; Hedwig, R.; Tjia, M. O.; Kagawa, K.; Maruyama, T. Phys. J. Indonesian Phys. Soc. 2003, A5, 0220-1.

It is found in this work that variation of laser power density in low-pressure plasma spectrochemical analysis of hydrogen affects sensitively the hydrogen emission intensity from the unwanted and yet ubiquitous presence of ambient water. A special experimental setup has been devised to allow the simple condition of focusing/defocusing the laser beam on the sample surface. When applied to zircaloy-4 samples prepared with various hydrogen impurity concentrations using low-pressure helium surrounding gas, good-quality hydrogen emission lines of very high signal to background ratios were obtained with high reproducibility under weakly focused or largely defocused laser irradiation. These measurements resulted in a linear calibration line with nonzero intercept representing the residual contribution from the recalcitrant water molecules. It was further shown that this can be evaluated and taken into account by means of the measured intensity ratio between the oxygen and zirconium emission lines. We have demonstrated the applicability of this experimental approach for quantitative determination of hydrogen impurity concentrations in the samples considered.

10.1021/ac061713o CCC: $37.00 Published on Web 03/07/2007

© 2007 American Chemical Society

Analytical Chemistry, Vol. 79, No. 7, April 1, 2007 2703

model,6 as a consequence of the time mismatch between the generation of the shock wave associated mostly with the heavier and relatively slow moving atoms ablated from the host target, and the very fast movement of the ablated hydrogen atoms due to their much smaller mass. In other words, the hydrogen atoms are likely to pass through the surrounding gas prior to shock wave formation and thus largely miss the subsequent shock excitation process. It was further found that the resulted intensity suppression becomes more severe with increasing surrounding gas pressure. This may partly explain the so-far unsuccessful spectrochemical analysis of hydrogen in solids with conventional LIBS, which also suffers from the Stark broadening effect, known to be detrimental for high-resolution applications. In a more recent experiment,6 hydrogen emission from lowpressure plasma was detected, showing favorable spectral quality for spectroscopic analysis when special care and a certain preliminary process were taken to minimize the presence of surface water on the sample surface. This result promises a useful new technique complimentary to the conventional LIBS technique for hydrogen analysis. However, for its practical implementation, in particular for quantitative analysis, a more detailed study is required concerning the most desirable operational condition involving experimental parameters such as the laser energy, focusing condition, choice of ambient gas, and proper detection time for the spectral data acquisition. Above all, this study should be addressed to the long-standing problem of how to suppress the unwanted effect arising from the ubiquitous presence of water molecules in the ambient gas and on the sample surface, without the need of tedious pretreatment for their removal. Further, a number of samples must be prepared with a variety of predetermined hydrogen concentrations in order to demonstrate the potential application of this technique for quantitative analysis of hydrogen in zircaloy. This experiment was carried in response to those needs. EXPERIMENTAL SECTION As mentioned above, a major issue to be dealt with for the practical application of low-pressure plasma spectroscopy to hydrogen analysis is the spectral contamination originating from the water molecules on the sample surface and in the surrounding gas. We have tried to achieve complete removal of the water molecules by all conventional treatments such as baking at elevated temperature (up to 200 °C), cleansing with benzene and toluene, as well as flushing the chamber with high-purity inert gases, and combinations of all those processes, but to no avail. Therefore, the experiment reported here is more concerned with finding an alternative way of minimizing or suppressing the unwanted hydrogen emission background contributed by the water vapor in the ambient gas and water layer deposited or adsorbed on the sample surface. To this purpose, we have conceived of an empirical approach by taking advantage of the fact that impurity hydrogen atoms in the target and those from the water are separated from their hosts via distinctly different processes. While the release of hydrogen and other atoms from the target sample follows the well-known photoinduced ablation process from the solid surface, those originating from the water molecules involve different processes (6) Kurniawan, K. H.; Kagawa, K. Appl. Spectrosc. Rev. 2006, 41, 99.

2704 Analytical Chemistry, Vol. 79, No. 7, April 1, 2007

Figure 1. Sketch of the experimental setup.

depending on their origins. Specifically, the hydrogen atoms originating from water vapor in the gas are likely dissociated from the water molecules directly, whereas those from the surface water may either be directly ablated from the sample surface or be dissociated from the water molecule after its desorption from the sample surface; all of these may involve complicated multiphoton processes. However, without going into their detailed mechanism, these processes are expected to require lesser overall energies than that needed for the ablation of the more strongly bonded impurity hydrogen atoms in the solid sample. This implies the possibility of applying preferential treatments of the two processes by proper adjustment of the irradiating laser power density. The present experiment is designed to verify the effectiveness of this approach in the minimization of spectral contamination from the water molecules. The basic experimental setup used in this study is shown in Figure 1. In this experiment, a special cylindrical chamber was designed with several entry ports. The inner side diameter of the chamber is 115 mm. The laser employed (Nd:YAG, Quanta Ray Lab Series, 1,064 nm, 8 ns, maximum energy of 450 mJ) was operated in the Q-sw mode at 10-Hz repetition rate with the laser output energy varied by means of a set of filters. The laser beam was focused by a movable lens of 200-mm focal length onto the sample surface at 45° through a quartz window. The He-Ne laser diode is a position marker used to align the invisible Nd:YAG laser with the laser diode illuminated spot on the sample surface. The shot-to-shot fluctuation of the laser was estimated to be ∼3%. The main targets employed in this experiment consisted of a set of zircaloy-4 samples doped with various concentrations of hydrogen (0, 161, 362, 461, 1000, and 4300 ppm). All the samples measured 10 mm × 10 mm in cross-sectional area and 1 mm in thickness. After placing the sample in the chamber, the chamber was evacuated using a vacuum pump to a pressure of 0.001 Torr. Subsequently, the chamber was heated to 200 °C and held for 10 min to remove most of the surface water, before the high-purity helium gas (Air Liquid, 5N) was introduced into the chamber until a pressure of 10 Torr was reached. This gas pressure and the above-mentioned chamber temperature were kept constant during the experiment. An additional sample of titanium (Rare Metallic Co, 4N, thickness of 0.4 mm) was also used in a separate initial experiment using air instead of helium as the ambient gas for

Figure 2. Variations of the ratio of the H I 656.2 nm and Ti I 655.4 nm emission intensities with respect to different positions of the focusing lens. Air at 5 Torr was used as the surrounding gas.

investigation of the basic discriminating effect to be employed in the ensuing experiment as explained earlier. Plasma emission was detected by an optical multichannel analyzer (OMA system, Andor I*Star intensified CCD 1024 × 256 pixels) was attached to a spectrograph (McPherson model 2061 with 1000-mm focal length f/8.6 Czerny-Turner configuration). connected to an optical fiber with its entrance end inserted through cylindrical quartz tube well into the chamber. and kept at a position 6 mm above the sample surface and at a distance of 80 mm from the center of the plasma. At this position, the fiber is expected to collect the emitted light entering within 270° of solid angle. The spectral window, covered by the detector has a width of 20 nm at 500 nm. The accumulated data of 100 detected spectra from each irradiated spot were monitored on a screen. In all experiments, the OMA system was set at a gate delay and gate width of 2 and 50 µs, respectively, for detection of emission from the secondary plasma.6 The spectral resolution of the OMA system is 0.009 nm at 500 nm. The sample surface condition was constantly monitored during the repeated laser irradiation and data acquisition process, using a stereomicroscope (Mini Dia Stereo MDS-40, Nissho Optical Co. Ltd.) through a 50-mm quartz window installed parallel to the sample surface. The working distance between the objective lens and the sample surface was kept at 135 mm. RESULTS AND DISCUSSION As mentioned earlier, in order to demonstrate the effect of laser power density variation on the hydrogen emission from the water molecules, an initial experiment was carried out using hydrogenfree titanium instead of zircaloy as the sample. This material has thermal characteristics (melting and boiling points) similar to that of zircaloy and is easily available in a bigger size. In this experiment, air was used as the surrounding gas so that the resulting hydrogen emission will be dominated by the contribution from hydrogen atoms from water vapor in the air and the possible residual water layer left on the sample surface. Figure 2 shows the emission intensities of H I 656.2 nm and Ti I 655.4 nm as functions of focusing lens position obtained at laser energy of 55 mJ. As displayed in the figure, the hydrogen emission intensity droped quickly as the lens was moved away from its tight focusing position denoted by 0 mm in the figure. The intensity curve features sharp cut off at the lens position further than 5 mm away

from the 0 position. Meanwhile, the titanium emission remains clearly observable even under defocused condition with the lens located at a position more than 10 mm from its tight focusing position. Similar graphs were obtained when the laser energy was increased to 100 mJ, although the hydrogen emission intensity was found to drop to zero at a position farther away (>30 mm from the 0-mm position). The titanium emission intensity featuring a nonmonotonous variation with the defocused positions is a manifestation of the competing effects of two different factors, i.e., the laser power density and the ablated area on the sample surface. The former affects the shock wave strength and hence the effectiveness of the associated excitation mechanism, while the latter determines the number of ablated atoms involved in the emission process. Apparently, an increase in the irradiated area due to defocusing of the laser beam dictates the initial rise of intensity up to a certain defocused condition. Beyond that, the seriously decreased laser power density and the resulting weakening excitation mechanism finally leads to the intensity downturn observed beyond that point. On the other hand, hydrogen emission from the water molecule appears to be mainly determined by the power density without being much affected by the relatively insignificant change in the amount of irradiated hydrogen atoms. This observation suggests the dominant contribution of hydrogen emission from water vapor in the ambient air, which is also consistent with the small likelihood of retaining the surface water at the elevated temperature (200 °C) and low pressure kept inside the chamber. The optimum condition for zircaloy analysis was further investigated by using zircaloy samples of 0 ppm and highconcentration (4300 ppm) hydrogen impurities. This optimum condition was determined by the appearance of minimum (near zero) hydrogen emission from the hydrogen-free sample and maximum hydrogen emission from the sample of highest hydrogen concentration. Taking into consideration the optical damage thresholds of the optical components found in this experiment, and the desirability of enhancing the emission intensity of the impurity hydrogen atoms, the laser was operated at 140 mJ with the focusing lens positioned at 45 mm from its tight focusing condition throughout the following study. It should be stressed that, at this higher laser energy, the larger distance of the lens from its tight focusing position was required to suppress the hydrogen emission from water as indicated in our separate experiment using the hydrogen-free zircaloy as described in the following section sample. For a further lens position, however, the emission intensity from the analyte hydrogen was found to deteriote as well. In other words, the chosen lens position was the one that yields the optimal signal to background ratio. Additionally, the air was replaced by helium gas to take advantage of the meta-stable excited state of helium atoms for further enhancement of emission intensity from the impurity hydrogen. This remarkable enhancement effect has been demonstrated in a previous experiment.6 The spectra of zircaloy samples were taken with helium surrounding gas of 10 Torr. In order to cover an adequate spectral range around the hydrogen emission line, a low-resolution OMA (Princeton Instrument IRY-700) was employed for the detection system. The detector used in this system was a gateable intensified photodiode array with a gating width ranging from 40 ns to 80 ms. The spectral window, covered by Analytical Chemistry, Vol. 79, No. 7, April 1, 2007

2705

Figure 3. Emission spectra of zircaloy-4 samples containing 4300 ppm hydrogen obtained with the focusing lens at the (a) tight focus and (b) defocus 45-mm positions. Low-resolution OMA was used with gate delay and gate width of 200 ns and 50 µs, respectively.

the detector, has a width of 80 nm at 500 nm. The detected signals were monitored on a screen. The OMA system was set at a gate delay and gate width of 200 ns and 50 µs, respectively. The spectral resolution of the OMA system is 0.4 nm at 500 nm. The spectra of zircaloy sample with 4300 ppm hydrogen concentration detected with focusing lens located at (a) 0- and (b) 45-mm positions are presented in Figure 3. It is seen that both the analyte hydrogen emission intensity and its signal to background ratio are significantly higher in the defocused condition (b) as compared to those in the tight focused condition (a). This is partly due to the reason explained for the choice of the defocused position. We have in fact verified separately for hydrogen-free zircaloy sample that the emission intensity of H I 656.2 nm, which is mainly contributed by hydrogen from dissociated water, is greatly suppressed under the chosen defocused condition, consistent with the results presented in Figure 2 and the related discussion. Further, the lower intensity of hydrogen emission exhibited in (a) is probably due to the deep crater produced during the 100 shots of laser irradiation. We have indeed observed a consistent reduction of emission intensity with increasing laser shots and in conjunction with deepening of the crater as monitored by the microscope. It was already shown in a previous experiment6 that the crater would impede the formation of the shock wave and delay the excitation process. This will in turn widen the time mismatch between the passage of hydrogen atoms and the formation of the shock wave, leading to unfavorable condition for the shock wave excitation of the hydrogen atoms as explained earlier. Under the defocused condition, the crater created by the laser shots was found to feature a shallow flat valley in the microscope image. As a result, the time mismatch effect 2706 Analytical Chemistry, Vol. 79, No. 7, April 1, 2007

Figure 4. Emission spectra of zircaloy-4 samples with (a) 1000 and (b) 4300 ppm hydrogen concentrations.

Figure 5. Calibration curve obtained from zircaloy-4 samples with error bars indicating ∼1.3% deviations.

was greatly reduced, resulting in higher emission intensity from the impurity hydrogen atoms. The optimum condition found in the experiment described above was adopted for the study of quantitative analysis of hydrogen in the zircaloy-4 samples prepared. Figure 4 shows the emission spectra of zircaloy-4 samples of (a) 1000 and (b) 4300 ppm hydrogen concentrations. From these spectra, the intensity ratio of H I 656.2 nm/Zr I 655.0 nm is found to be proportional to the hydrogen concentrations. Encouraged by this result, the measurement was repeated for the series of samples with various concentrations of impurity hydrogen. We repeat that the data from each spot on the sample surface are the averaged result of 100 successive shots as stated earlier. This measurement was then repeated at five different spots on the surface of the same sample, which were found to yield highly reproducible results, implying the uniformity of impurity hydrogen distribution in the sample. The averages of these five measurement results for samples with different impurity hydrogen concentration were then plotted in

intensity ratio of O I 777.1 nm versus Zr I 769.3 nm emission lines. Since the presence of surface water will also give rise to additional oxygen emission along with the corresponding hydrogen emission line, measurement of O/Zr intensity ratio can be used to evaluate the amount of ambient water and properly taken into account in the analysis. Figure 6 shows the oxygen spectra taken for zircaloy-4 samples containing hydrogen at (a) 0 and (b) 362 ppm concentrations. It is seen that the O I 777.1 nm/Zr I 769.3 nm intensity ratios are exactly the same, implying the presence of the same amount of ambient water. Based on this result, which was verified in all the zircaloy-4 samples used in this experiment, reliable quantitative analysis can be performed on the basis of the calibration curve.

Figure 6. Emission spectra of oxygen measured for zircaloy-4 samples containing hydrogen at (a) 0 and (b) 362 ppm hydrogen concentrations.

Figure 5 as the calibration curve. It is seen that the curve displays an excellent linear relationship between hydrogen emission intensity and hydrogen concentration in the zircaloy-4 samples. Further, one notes from Figure 5 the small nonzero intercept of the linear calibration line, which is indicative of the recalcitrant effect of the water in the ambient gas. Indeed, we have found that the hydrogen emission intensity from the hydrogen-free sample closely matches the intercept value. In order to take into account this unwanted contribution in the application for quantitative analysis, one can take advantage of the known concentration of oxygen (0.1%) in the zircaloy-4 samples and, hence, the known

CONCLUSION We have shown in this work that the unwanted background contribution from ubiquitous ambient water molecules in the emission spectra of plasma can be effectively suppressed by defocused laser irradiation as well as the proper control of the ambient gas pressure and chamber temperature. In its application to hydrogen impurity detection and measurement in a zircaloy-4 sample employing helium surrounding gas, this experimental approach has demonstrated the possibility of attaining optimal signal to background ratio and a linear calibration curve. Further, a reliable means for determining the residual spectral contribution from the ambient water molecules associated with the intercept of calibration line was also proposed, making possible the application of the proposed experimental method for quantitative hydrogen analysis of the samples, which is urgently needed for regular inspection of the zircaloy tube in a light water nuclear power plant. In view of the general nature of this experimental approach, it should hold promise for its extension to other areas of application where the unwanted spectral contribution of ambient water molecules is known to compromise the detection sensitivity and reliability.

Received for review September 12, 2006. Accepted February 6, 2007. AC061713O

Analytical Chemistry, Vol. 79, No. 7, April 1, 2007

2707