J. Phys. Chem. B 2008, 112, 8099–8105
8099
Equilibrium Isotope Effects in the Preparation and Isothermal Decomposition of Ternary Hydrides Pd(HxD1-x)y (0 < x < 1 and y > 0.6) Weifang Luo,* Don Cowgill, Rion Causey, and Ken Stewart Deptartment of Hydrogen and Metallurgical Sciences, Sandia National Laboratories 7011 East AVenue, LiVermore, California 94551 ReceiVed: February 19, 2008; ReVised Manuscript ReceiVed: March 24, 2008
A Sieverts’ apparatus coupled with a residual gas analysis is used to measure the concentration variations of hydrogen isotopes in the gas and solid phases during exchange and isothermal decomposition of mixed hydrides. β-phase palladium hydrides with known ratios of H:D, Pd(HxD1-x)y (0 < x < 1, y > 0.6), are prepared by H2 with PdDy or D2 with PdHy exchange, and their desorption isotherms are reported here at 323 K. A higher equilibrium pressure in isothermal desorption of mixed hydrides is associated with a higher ratio of D/H in the initial mixed hydrides in β-phase. The composition of the gas desorbed from a mixed hydride varies; i.e., the ratio of D/H in gas decreases with the sum of (H + D) in Pd. The values of the separation factor R during desorption at 323 K and during H-D exchange at 248 K are discussed and compared with those in the literature. Desorption isotherms of mixed isotope hydrides are between those of the single isotope hydrides of H-Pd and D-Pd, however, plateaus slope more than those of pure isotope hydrides. The origin of the plateau sloping in the mixed hydrides can be attributed to the compositional variations during desorption, i.e., the equilibrium pressure is greater when D/H ratio in solid is greater. A simple model is proposed in this study that agrees well with experimental results. Introduction
Dg ) PD2 + PHD/2
Research in the H2-Pd system has been active for more than a century.1 This system is the most extensively investigated metal-hydrogen system.2 A few comprehensive review articles2–5 summarize most of the research activities and findings. Hydrogen isotope effects attract research attention because of their importance for both fundamental and technical reasons.4–14 Hydrogen isotope exchange is a technique that can be applied to isotope separation for enrichment in heavier hydrogen isotopes. Thermodynamic and kinetic understanding of isotope effects is essential. Isotope exchange of β-phase hydride PdmH with D2 is a twostep reaction
Hg ) PH2 + PHD/2
D2 + PdmH ) HD + PdmD HD + PdmH ) H2 + PdmD
(1)
The isotope equilibrium constant KHD in a mixed gas phase is defined as
H2 + D2 S 2HD
KHD ) PHD2 ⁄ (PH2PD2)
(2)
The separation factor, R, is defined as
PD2+PHD/2 R )
PH2+PHD/2 Ds/Hs
)
Dg/Hg Ds/Hs
(3)
where * To whom correspondence should be addresed. Phone: (925) 294-3729. Fax: (925)294-3410. E-mail:
[email protected].
PD2, PHD, and PH2 are the partial pressures of the three isotope species in the gas phase, and Ds and Hs are the isotope concentrations in Pd, i.e., Hs ) H/Pd and Ds ) D/Pd. The separation factor R is the key parameter for characterization of a system for isotope exchange. Strictly speaking the (Dg/Hg)/(Ds/Hs) is a variable for a given system at given conditions before reaching equilibrium. (Dg/Hg)/(Ds/Hs) reaches a constant when the system is in equilibrium and the magnitudes of variables of Dg, Hg, Ds, and Hs will not change. Reliable R values and their dependence on the given working conditions are essential for further kinetic studies and for technical applications of isotope exchange. In this paper the term R is used to stand for the (Dg/Hg)/(Ds/Hs) at both equilibrium and nonequilibrium conditions. There are a number of articles in the literature4–10 providing experimental R values and thermodynamic model calculations. Its dependence on temperature and the isotope compositions in the gas and solid phases for H-D-Pd exchange systems are discussed. Wicke et al.4 reported experimental R values and the dependence on temperature. Brodowsky et al.6 summarized and analyzed a considerable amount of data from the literature and experimental values. They developed models to calculate R values in good agreement with experimental data. On the basis of models they further derived the R dependence at given temperatures in terms of the isotope ratios of Hg/Dg, Hs/Ds, and the sum of hydrogen isotopes in the solid, H/Pd + D/Pd ) Hs+Ds. Andreev et al.5 reported R and its dependence on temperature and the isotope compositions in the gas and solid phases. They used the following equations to calculate the separation factor R in R-phase and R-β two-phase regions
10.1021/jp801487n CCC: $40.75 2008 American Chemical Society Published on Web 06/14/2008
8100 J. Phys. Chem. B, Vol. 112, No. 27, 2008
R phase:
R ) Ks(D2)/Ks(H2)
R/β two phase:
R ) √PD2°/PH2°
Luo et al.
(4) (5)
where Ks(D2) and Ks(H2) are the Sieverts constants and PD2° and PH2° are the equilibrium plateau pressures for D-Pd and H-Pd systems, respectively. It this study we examine the thermodynamic properties during the preparation of mixed isotope hydrides of Pd (β-phase) as well as during isothermal desorption of mixed hydrides. The experiment was carried out in a Sieverts’ apparatus coupled with a residual gas analyer (RGA). With this experimental setup it is possible to monitor the three isotopic species in the gas phase, H2, HD, and D2, during isotopic exchange from which the concentrations of H and D in the solid phase can be calculated. There is limited information for pressure-composition isotherms of mixed hydrides in the literature. Absorption and desorption isotherms reported by Sieverts et al.14 for mixed hydrides are for PH2/PD2 ) 1:1 and 1:3. In this study more detailed isotherms for isothermal desorption of mixed hydrides will be reported and the origin of sloping plateaus will be discussed. Experimental Section Samples consisting of 2.4 g of palladium powder of particle size of 20-36 µm (99.95% from Alfa Aesar) and pure H2 and D2 (99.9995% from Matheson) were employed for the isotopic exchange characterization. Three hydrogen sorption cycles at 298 K were carried out before a mixed isotope hydride preparation. Prior to the preparation of pure PdHy or PdDy for exchange the sample was evacuated at 378 K for 15 min to ensure a complete hydrogen removal, followed by cooling to
Figure 1. Schematic Sieverts’ apparatus for exchange test.
Figure 2. Calibration of RGA pressure readings.
room temperature. No other treatment was employed before collecting isotopic exchange data. A Sieverts’ apparatus, sketched in Figure 1, was used to prepare mixed isotope hydrides. The gas pressures in the system were monitored by MKS gauges (MKS Instruments Inc.). The pressure variation in the sample holder was in a range within 0-20000 Pa. The composition of hydrogen isotopes in the gas phase of the sample holder was monitored by a RGA 200 (by Stanford Research Systems Inc.). The gas flow rate from the sample chamber to the RGA was controlled by a specially designed flow restrictor, as reported previously15 to ensure that the pressure of the gas entering the RGA is below 6.7 × 10-4 Pa, as required by RGA. The gas flow rate was so low that no measurable gas loss occurred during exchange. The correlation between the signal intensity of the RGA and the gas pressure in the sample chamber was calibrated by pure gaseous hydrogen or deuterium separately and their average was used as the calibration value for HD since HD is not commercially available. Figure 2 shows the calibration for RGA readings. It can be seen in this figure that there is a linear correlation between the reading from the pressure gauge and the corresponding RGA readings over a range of 100-14000 Pa. There are, however, some errors in the pressure range of 96%, was obtained between the equilibrium pressure readings from gauges and those from the RGA. Four mixed hydrides, i.e., Pd(HxD1-x)y (x ) 0.25, 0.53, 0.81, and 0.82 and y > 0.6) were prepared by the following steps: (i) React Pd with D2 or H2 to form β-phase of single isotope hydrides with known compositions, i.e., PdDy or PdHy (y > 0.6); (ii) Prepare mixed isotope hydrides by isotopic exchange reaction between the β-phase hydrides and H2 or D2, i.e., H2 with PdDy or D2 with PdHy. This procedure was carried out at room temperature or lower. During exchange the variations of total gas pressure readings from gauge were insignificant; however, the gas compositions, i.e., the partial pressure of H2, HD, and D2 from the RGA, varied
Equilibrium Isotope Effects in Ternary Hydrides
J. Phys. Chem. B, Vol. 112, No. 27, 2008 8101
Figure 5. H, D, and H + D concentrations in Pd during exchange process of D2 with PdH0.678 at a temperature of 248 K. Figure 3. Measured partial pressures of isotopes and total pressure in the gas phase during isotopic exchange for the preparation of sample (Pd(H0.81D0.19)0.68).
Figure 6. The separation factor and exchange equilibrium constant during exchange at 248 K. Figure 4. Hg (PH2 + PHD/2) and Dg (PD2 + PHD/2) during exchange.
significantly. The gas composition values were then used to calculate the concentrations of H and D in the solid. Desorption isotherms for four mixed isotopic hydrides were measured at 323 K. The desorbed gas composition was measured from which the Hs and Ds values were then calculated. Desorption isotherms at 323 K for single isotope hydrides, PdH0.65 and PdD0.63, were measured for comparison. Results and Discussions 1. Preparation of the Mixed Hydride Pd(H0.81D0.19)0.68. A sample of PdH0.65 was prepared at room temperature initially and then cooled to -25 °C. More hydrogen was then added at this temperature until the equilibrium pressure of H2 reached 1470 Pa. This resulted in formation of PdH0.678, which was then used to prepare the mixed hydride by exchange of D2 with PdH0.678 at 248 K. Figures 3–7 depict results in detail, i.e., gas compositions, Hs and Ds in Pd, the isotope separation factor R, and the gas equilibrium constant KHD during preparation of the mixed hydride, Pd(H0.81D0.19)0.68. Figure 3 shows the partial pressures of H2, HD, and D2 in the gas phase measured by RGA during the exchange process. The four red arrows indicate the points where new D2 doses were introduced. It can be seen that the partial pressure changes for all isotopes are very fast when a new dose is added and then the change decreases gradually until the system reaches equilibrium. Within each dose the total gas pressure remain
almost constant. A small pressure drop at the beginning of each dose is believed to result from the thermal interruption upon a new dose of gas entering the sample holder at 248 K from the dosing volume at 296 K. Figure 4 shows the partial pressure values of H and D in gas phase, i.e., Hg and Dg (as “H in gas” and “D in gas”). To facilitate “visualizing” the effect of the exchange, “H2 no-Xchg” (dot-dashed line) and “D2 no-Xchg” (dashed line) are included for comparison. These are the “hypothetic partial pressures” for H2 and D2 in the case of no exchange. The concentrations of H and D in Pd during the exchange process are shown in Figure 5. (H + D)/Pd values are almost constant during the exchange. The observed small increase is believed to result from the total pressure increase as shown in Figure 3. Figure 6 shows the values of PHD2/(PH2PD2) and (Dg/Hg)/(Ds/ Hs) during exchange based on the data shown in Figures 3 and 4. Both PHD2/(PH2PD2) and (Dg/Hg)/(Ds/Hs) vary as a function of time before equilibrium is reached. Table 1 lists equilibrium values of R, KHD, and ratios of Dg/ Hg, Ds/Hs, and total hydrogen content (H + D)/Pd for each dose, as shown in Figure 6. The system reaches equilibrium except for the fourth dose, which is slightly before equilibrium. The average R is 2.82, very close to the literature value, as shown in Figure 7. It can also be seen that R increases slightly over these doses. Brodowsky et al.6 discussed the R variation at a given temperature and concluded that the most influential factor on R
8102 J. Phys. Chem. B, Vol. 112, No. 27, 2008
Luo et al.
Figure 7. Comparison of R values from this work (the purple square) and those from Brodowsky.6
TABLE 1: Summary of Equilibrium Values of the Separation Factor r, Equilibrium Constant KHD, Dg/Hg, Ds/Hs, and (H + D)/Pd at the End of Each Dose during Exchange of D2 with PdH0.678 at 248 K dose no. R KHD Dg/Hg Ds/Hs (H + D)/Pd
0
first
second
third
fourth
2.66
2.71 2.83 0.25 0.095 0.691
2.81 2.80 0.42 0.147 0.691
3.11 2.80 0.63 0.200 0.689
a
0 0 0.678
0.09 0.036 0.685
a
The measured partial pressure of D2 at the end of the first dose was only 40 Pa, which falls in the error range, as mentioned above, and this is believed to have caused a large error in KHD for this dose, and therefore, it is not reported for this dose.
was the sum of H + D in Pd, i.e., Hs + Ds, and the isotopic ratios would affect R in such a way that it decreases with Ds/Hs and increases with Dg/Hg. In the present experiments the Hs + Ds is essentially constant, and the small increase in R seems to be related to the increase of Dg/Hg. Figure 7 shows the R dependence on temperature for R- and β-phases by Brodowsky6 and Sicking,10 with the R value (2.8) from Table 1 as a purple square where a good agreement can be seen. The KHD values are quite constant for the second to fourth doses. The measured partial pressure of D2 at the end of the first dose was only 40 Pa, which falls in the error range, as mentioned above, and this is believed to have caused a large error in KHD for this dose. Different from R values, KHD is calculated from partial pressures of each isotopic species, H2, D2, and HD, and is therefore, more sensitive to the errors in the partial pressure of each species. As a comparison, the value of R at the end of first dose is close to those of the later doses since R depends on Dg, which is the sum of (PD2 + 1/2PHD) and, therefore, less sensitive to PD2. The only KHD values reported in the literature are 2.89 and 3.50 in the range of 200-400 K from the literature.4,11 The value of 2.8 in this study is close but slightly smaller. 2. Preparation of Other Mixed Hydrides. The other three mixed hydrides were prepared by a similar method. The Pd(H0.81D0.19)y and Pd(H0.82D0.18)y were prepared from the exchange of D2 with PdHy while Pd(H0.52D0.48)y and Pd(H0.25D0.75)y were prepared from the exchange of H2 with PdDy. It is a shorter path to prepare the mixed hydride with a higher H/D ratio in solid from PdHy and those with higher D/H
Figure 8. Desorption isotherms at 323 K for Pd(H0.82D0.18)0.65, Pd(H0.81D0.19)0.65, Pd(H0.52D0.48)0.65, Pd(H0.25D0.75)0.65, PdH0.65, and PdD0.6.
ratio from PdDy. The detailed sample preparation and the retrieved data for these three samples are not included here to avoid duplication. 3. Desorption Isotherms for Mixed Hydrides. Desorption isotherms of the four mixed hydrides measured at 323 K are shown in Figure 8. Isotherms for PdH0.65 and PdD0.6 are included for comparison. The results are reproducible as shown by the overlapping of isotherms of mixed hydrides with similar composition, Pd(H0.81D0.19)0.65 and Pd(H0.82D0.28)0.65. It can be seen from this figure that: • Desorption isotherms of mixed hydrides lie between those of single hydride Pd-H or Pd-D, and they slope more than those of Pd-H or Pd-D. • A higher equilibrium desorption pressure and larger slope in plateau region is correlated with a higher ratio of D/H in the initial mixed hydride. 4. Gas Composition Variations during Desorption of Mixed Hydride Pd(H0.52D0.48)0.65. To further understand the desorption mechanism of the mixed isotope hydrides, the desorbed gas composition was monitored by RGA during desorption. The four panels in Figure 9 show sections of the desorption profiles for partial pressures of the three gas species in the following pressure removals from the mixed hydride Pd(H0.52D0.48)0.65. The starting points of new desorption doses are shown in this figure as sharp decreases of the pressure readings. These panels correspond to different regions of an isotherm: panel 1 for the β-phase region, panel 2 for the region between the β-phase and R-β two phase, panel 3 for the R-β two-phase, and panel 4 for the R-phase region. The total pressure (brown line) is the sum of partial pressures of these three isotopic components, and this value agrees with those measured by the gauge (errors < (4%). The partial pressure of D2 (green dashed line) is higher at the beginning of desorption, as shown in panel 1, and it decreases with time and becomes the smallest one as shown in panel 4. The partial pressure of H2 (blue line) is much smaller than those of D2 and HD as shown in panel 1 and then gradually increases and finally becomes the largest at the end of desorption (panel 4). Figure 10 shows the analysis of composition variations in the gas and solid phases during Pd(H0.52D0.48)0.65 desorption. The desorption isotherms for the mixed hydride, and single
Equilibrium Isotope Effects in Ternary Hydrides
J. Phys. Chem. B, Vol. 112, No. 27, 2008 8103
Figure 9. Partial pressures of H2, HD, and D2 in desorbed gas at 323 K during Pd(H0.52D0.48)0.65 desorption.
hydrides, H-Pd and D-Pd, are included in all panels for a convenient correlation between compositional variations and desorption progress. Panel a shows the equilibrium pressure values of Hg and Dg during desorption. In the β-phase region, Dg is much higher than Hg, even though the ratio of Hs/Ds in the hydride is initially close to 1 (52/48). Dg gradually decreases over the two-phase region and becomes smaller than Hg when (H + D)/Pd < ∼0.25, while Hg is almost constant in the plateau region with its value very close to PH2 for the single isotope system, H-Pd. The decreasing Dg is therefore the main cause of the plateau slope in the stepwise desorption of Pd(H0.52D0.48)0.65. The only isotherms reported in the literature are those by Sieverts14 for the absorption at H2/D2 ratio of 1:1 and 1:3 and the subsequent desorption at a temperature of 373 K. After absorption the H/D ratio in the solid was not reported. The slope in these isotherms in the R-β phase regions is apparent; however, no discussion was given as to its origin. A simple model is proposed here to calculate the total equilibrium pressure Ptot, the sum of PH2 + PD2 + PHD, for the plateau regions during mixed hydride desorption. In this model it is assumed that at a given (H + D)/Pd and temperature the Ptot is a function of the fraction of isotope in solid and the equilibrium plateau pressures of H-Pd and D-Pd
When the gaseous phase coexists with two solid phases the phase rule requires that the pressure be invariant and for metal-H systems this region is referred to as the plateau. Frequently these plateaus are found to slope especially for alloys and intermetallic-H systems. Although explanations have been offered for the sloping,15–18 there is no general consensus. In the present case, however, the origin of the slope is unambiguous; it is due to the decrease of the fraction of D in the solid as the sample is desorbed. Panel c shows Dg percentage in the desorbed gas. It can be seen that this decreases monotonically, from 75 to 8%, during desorption. Panel d shows the ratios of Hs/Ds in solid and Dg/ Hg in gas. Hs/Ds gradually increases during desorption, from ∼1 at the beginning to greater than 3 at the end, while Dg/Hg gradually decreases from 2.5 to 0.4. Panel e shows the R variations during desorption in β-, R + β-, and R-phases. The R reaches a constant value of 2.35 in the first half of the R + β-phase region, i.e., (H + D)/Pd >0.22, but decreases with (H + D)/Pd in R- or β-phases. There seems no R data reported for the R-β phase coexisting region for the mixed (H + D)/Pd hydride. Andreev5 suggested using eq 5 to calculate R
Ptot ) fHPH2°(s) + fDPD2°(s) ) fHPH2°(s) + (1 - fH)PD2°(s) fH ) Hs/(Hs + Ds) fD ) Ds/(Hs + Ds) fH + fD ) 1
where the PD2° and PH2° are the equilibrium D2 or H2 pressures in the plateau regions of H2-Pd and D2-Pd single isotope systems, respectively. The value calculated by this equation is 2.16, obtained from desorption plateau pressures of H-Pd and D-Pd. This value is slightly smaller that 2.35 but close enough, since it covers a large range of ratios of H/D in the gas and solid phases during desorption. Wicke et al.4 reported the value R ) 2.25 at 323 K with partial pressure ratio of PH2/PD2 ) 1 in β-phase, which is close to the one obtained in this study. According to Brodowsky,6 at a given temperature, R decreases with the ratio of Dg/Hg. In the plateau region we observed R decreases significantly for (D + H)/Pd 0.6), were prepared by isotope exchange, i.e., H2 with PdDy or D2 with PdHy, in β-phase region at constant temperatures. The separation factor R during exchange was monitored and the results are in good agreement with those in the literature The desorption isotherms of these mixed hydrides were measured at 323 K. The equilibrium desorption pressures of the mixed hydrides vary with the ratio of H/D in the initial hydrides. Higher equilibrium pressures in desorption are associated with high ratios of D/H in the initial mixed hydrides. The D/H ratio in the desorbed gas decreases with the total isotopic content in the solid, (Hs + Ds). The desorption isotherms of mixed isotope hydrides lie between those of single isotope systems H-Pd and D-Pd. Desorption plateaus of mixed hydrides slope more than those for single isotope systems. The plateau slope results from the variation of isotopic fraction in solid, Ds/(Hs + Ds), during desorption, according to the simple model proposed here. Higher desorption plateau pressure corresponds to a higher D fraction in solid, Ds/(Hs + Ds). This model provides a tool useful for estimating isotope fraction in a mixed hydride from its plateau pressure. Acknowledgment. This work was supported by the United States Department of Energy under Contract DE-AC04-
J. Phys. Chem. B, Vol. 112, No. 27, 2008 8105 94AL85000. W. Luo thanks Prof. T. Flanagan for valuable suggestions and advices. References and Notes (1) Graham, T. Phil. Trans. R. Soc. 156 (1866) 415. (2) Lewis, F. A. The palladium Hydrogen System; Academic Press, London, New York, 1967. (3) Flanagan, T. B.; Oates, W. A. Annu. ReV. Mater. Sci. 1991, 21, 269. (4) Wicke, E.; Brodowsky, H. Hydrogen in Metals II. Metal Abstracts; Metal Abstracts Trust. (5) Andreev, B. M.; Magomedbekov, E. P.; Sicking, G. Interaction of Hydrogen Isotopes with Transition Metals and Intermetallic Coumponds; Kuhn, J., Ed.; Springer-Verlag: 1996. (6) Brodowsky, H.; Repenning, D. Z. Phys. Chem. 1979, 114, 141. (7) Wicke, E.; Nernst, G. Ber. Bunsen.-Phys. Chem. 1964, 68, 224. (8) (a) Brodowsky, H. Z. Phys. Chem. 1965, 44, 129. (9) Glueckauf, E.; Kitt, G. Vapor Phase Chromatography; Desty, D., Ed., Butterworth Scientific Publishing: London, Boston, 1957. (10) Sicking, G. Z. Phys. Chem. 1974, 93, 53. (11) Urey, H. C.; Rittenberg, D. J. Chem. Phys. 1933, 1, 137. (12) Folts, G. W.; Melius, C. F. J. Catal. 1987, 108, 409. (13) Leardini, F.; Fernandez, J. F.; Bodega, J.; Sanchez, C. J. Chem. Phys. 2003, 69, 116. (14) Sieverts, A.; Danz, W. Z. Phys. Chem. 1937, 38, 46. (15) Malinowski, M. E.; Stewart, K. D.; Verberkmoes, A. A. Sandia Report, 1991. (16) Park, C. N.; Luo, S.; Flanagan, T. B. J. Alloys Compd. 2004, 384, 203. (17) Luo, S.; Park, C. N.; Flanagan, T. B. J. Alloys Compd. 2004, 384, 208. (18) Naito, S.; Yamamoto, M.; Dol, M.; Kimura, M. J. Chem. Soc., Faraday Trans. 1995, 91 (22), 4143.
JP801487N
