The of and Palladium Alloys

A REVIEW OF RECENT RESEARCHES

By F. A. Lewis, Ph.D. Department of Chemistry, The Queen's University of Belfast

Since the initial investigations of Graham - (I) it has been fairly clearly established that, In this article, to be published in two under almost all conditions of temperature parts, some recent Jindings are included and gas pressures, the amount of in a review of the pressure-concentration hydrogen which can be absorbed* by palla- temperature (P-C-T) relationships of dium greatly exceeds the amount absorbed by the palladium/hydrogen and palladium1 any other element from Group VIII of the s.ystems. A brief account is periodic table and compares (2) with the given of experimental information con- amounts absorbed by the electropositive cerning the structure of, and difusion metals of Groups I to V; for example, at of hydrogen in, the solids which are the temperatures of about 25"C, the atomic ratio products of the absorption of hydrogen (H/Pd) of hydrogen to palladium can readily by palladium. Attention is paid to be made to exceed 0.5. Moreover, in contrast considering errors of interpretation of to what is found for the remainder of Group experimental results that can arise when VIII, absorption of hydrogen by palladium true thermodynamic equilibrium does is an exothermic process (3) albeit with a not exist between the solids and hydrogen fairly low heat of reaction of about 9.5 Kcal/ molecules-both in the presence and mole H, at around 25°C. (Adsorption of absence of aqueous solution hydrogen on to a clean surface is, however, an exothermic process for all Group VIII metals (4.) of the surface and complex slip line structures The original volume of a palladium speci- are observed (5) as a result. men expands by about 10 per cent when It has been claimed (6)that at temperatures H/Pd attains a value of about 0.5 but, in con- of about 25°C hydrogen may subsequently trast to the other transition metals which be desorbed from palladium hydrides without absorb large volumes of hydrogen (i.e. titan- the specimen contracting to its original dimen- ium, zirconium and hafnium in Group IV sions, but more recent work (7, 8, 9, 10)does and vanadium, niobium and tantalum in not confirm these findings. Certainly there is Group V), there is virtually no macro- no dispute that on removal of hydrogen by disruption of palladium specimens which heating at about 300"C, either in air or in undergo this expansion although roughening vacuo, the original dimensions of a palladium specimen are aImost completely restored, *Graham described the absorption of gas as although they may eventually be markedly 'occlusion'. In some subsequent publications it is altered when a large number of cycles of no1 however always clear that 'occlusion' does not refer to adsorption as well as absorption. absorption and violent expulsion of hydrogen

Platinum Metals Rev., 1960, 4, (4), 132-137 132 3 1000 1 Pressure-concentrationisotherms ATM Fig. of the pallodiumlhydrogen system

572 W U W a and p. The ratio of hydrogen to 1 20 ATM I I palladium in both these phases is :I I 0 I generally non-stoichiometric; just I I I- I < I below the critical temperature I I ATM -0 I W 3Ooc I (H/Pd), ~(HPd)~-o.27but over a I zl I 18 rnrnHg the range of contents where u and Y) I ;-2 I /3 phases coexist at 30°C, (I-[/Pd), a I (L I -0.03 and (H/Pd),-o.-j7. It 77 'K _---,' W ld4rnm Hg seems that at even lower tempera- 5 -7 I tures (77°K) the maximum hydro- gen content of the a phase is very low indeed while in p-phase 0.2 04 0.6 0.8 1.0 1.8 2.0 2.2 H/Pd (ATOMIC RATIO) H/Pd-2. A composite diagram is shown in Fig. I. have been undergone. The possibility of P-C-T Data and Electrode dimensional changes has to be carefully con- Potentials sidered when palladium is used as a diffusion The palladium/hydrogen system also lends membrane to separate hydrogen from other itself to quite convenient derivation of P-C-T gases: this problem has been discussed in a data from electrode potential measurements previous issue of this review (I I). (10) by use of the relationships between pressure P, chemical potential p, and electrode Hydrogen Pressure-Hydrogen potential E, i.e. RTlnP = p = 2FE where Content-Temperature (P-C-T) R is the gas constant and F is Faraday's Relationships of the Palladium1 constant. Hydrogen System The advantages to be gained from the Conditions have been attained by several storage of hydrogen by palladium electrodes investigators in which the concentration of are under examination in connection with hydrogen absorbed by palladium is in almost electrochemical cells (I 6) and electrophoresis true dynamic equilibrium with molecular studies (17). hydrogen in the gas phase. Particularly detailed and accurate P-C-T relationships at Crystallographic Investigation temperatures from 0"c-3S0"c and at hydro- The several X-ray studies (2) that have gen pressures up to about 30 atmospheres been reported have generally been carried have been compiled by the research groups out at about 25°C. of Gillespie and Sieverts (12). More recently The weight of evidence indicates that over P-C-T data have also been obtained at higher a-phase concentrations there is a small but pressurcs (13, 14) and at higher and lower continuous increase of the cell constant a, of temperatures (14, IS). the face-centred cubic palladium lattice from Below a critical temperature, T,, of about 3.88A to attain a value of 3.89A when 300°C (critical pressure, P,, -20 arm) there Hpd = 0.03. are regions over which the hydrogen pressure On further increasing the hydrogen content is invariant with change of hydrogen content a new set of X-ray reflections are observed indicating the coexistence of two solid phases, which are characteristic of the p-phase

Platinum Metals Rev., 1960, 4, (4), 133 . The palladium atoms in this phase retain f.c.c. I000 symmetry, but the cell con- stant increases to about 4.0zA. The p-phase reflections (they 800 I exhibit considerable line broad- E ening (18)) increase in in- E b tensity as the total hydrogen 600 w a content increases over the 2 In region where c( and (3 phases In coexist. The expansion of a, 400 to 4.ozA parallels the 10 per cent expansion of the macro- 200 volume observed when the

0: --f 9 transformation is com- pleted (H/Pd N 0.57). Re- I I I I I I sults indicate that still further 0.1 0.2 0.3 0.4 0.5 0.4 absorption of hydrogen then H/Pd (ATOMIC RATIO) causes an additional contin- Fig. 2 Pressure hysteresis between ‘absorption’ and ‘desorption’ uous increase of the (3-phase isotherms. Data obtained with palladium sponge at 103°C cell constant. Neutron dif- (Reference 20) fraction studies on completely transformed P-phase hydride (H,’Pd - 0.65) at 30”C, hysteresis was again observed; have also been reported (19). neither absorption nor desorption pressures altered appreciably on standing at this low Hysteresis of Isotherms temperature, and here it was suggested that Fig. I does not indicate that over regions it was the absorption isotherm which repre- of x and (3 phase coexistence the ‘equilibrium’ sented true equilibrium conditions. pressures measured when the hydrogen con- tent is being successively increased (absorp- Nucleation of P-phase H ydride tion isotherms) often exceed (Fig. z) the There is still disagreement as to whether ‘equilibrium’ pressures when hydrogen is @phase first forms a layer on the surface of successively removed between measurements the solid or whether growth generally develops (desorption isotherms). from nucleation centres in the interior. By using finely particulated palladium and The inward migration of a p-phase layer, by heating to 360°C between measurements, with a concentration gradient extending from Gillespie and his co-workers (12) obtained a hydrogen rich surface to the ct-P phase isotherms in which this ‘hysteresis effect’ was boundary, has been invoked (3, 10)to explain eliminated: the pressures measured in the the slight increase (20) (see Fig. 2) of equili- two-phase region were generally intermediate brium pressures often observed during between the absorption and desorption pres- absorption over the region of GI and sures obtained with more massive specimens Y phase coexistence. (A similar effect can (3). Furthermore, early work in Sieverts’ also be produced by a temperature gradient laboratory (12) with finely divided palladium along the sample but in these circumstances black at III’C,indicated that the absorption sloping isotherms should be observed during and desorption pressures approached a com- both absorption and desorption (21). mon mean on standing. However, during A corollary to this picture is that during further recent work (3) with palladium blacks desorption the ct-p phase boundary retreats

Platinum Metals Rev., 1960, 4, (4), 134 inwards from the surface and over the two- attains a pressure of one atmosphere at phase region the measured pressure is that in -I 10°C while the pressure of hydrogen over equilibrium with a surface consisting solely P-phase hydride only attains one atmosphere of z-phase (3). at -150'C. These features are in line with the fact that deuterium is less strongly The Palladium-Deuterium absorbed by palladium than is hydrogen; the Equilibria heat of absorption(3) of D, (ca. 8.5 Kcal/ Reliable P-C-T data for the Pd/D system mole) being about I Kcal/mole lower than have also been obtained by Sieverts and the value for H, (compared over regions of Gillespie (12). When the data are plotted as cc and P-phase coexistence at 30°C). isotherms these are found to have a similar Factors governing the electrolytic separa- appearance to those of the Pd/H system. The tion of hydrogen from deuterium on a palla- critical temperature, T,, for cc and p-phase dium cathode have been discussed (23). coexistence is about 276°C which is only slightly lower than T, for the Pd/H system Diffusion of Hydrogen in but P, is N 35 atm for Pd/D which is almost Palladium Hydride twice the corresponding critical pressure for When interpreting experimental results Pd/H. Furthermore, in regions where a and related to the diffusion of hydrogen in palla- P phases coexist below T,, the equilibrium dium hydride the possible contribution of the pressure for the Pd/D system markedly following processes must be considered: exceeds the corresponding pressure for the Pd/H system at the same temperature; for (I) Diffusion in the .s-phase example, at 30"C,when x and phases coexist (2) Diffusion of the r-P phase boundary PD, - 80 mm while PH, - 18 mm. Methods (3) Diffusion in phase after completion for the separation of hydrogen from palladium of the a --f P transformation have been designed (22) to take advantage of Methods of analpsing experimental results these differences, which are perhaps best where these alternative possibilities exist have illustrated by plotting the P-C-T data in the been outlined (24). form of isobars. Characteristic 'absorption' The high mobility of hydrogen in palladium isobars in Fig. 3 show that the pressure of hydride has been indicated from nuclear D, in equilibrium with p-phase deuteride magnetic resonance experiments (25), and it is known empirically that processes (I) and (3) proceed extremely rapidly even at room temperature (10). It is, however, only for diffusion of hydrogen in the a-phase that 0.6 detailed experiments have been carried out B and the activation energy for this process has n been reported as about 5 Kcal/mole H,. How- 04 z ever, in spite of recent claims (26) to the con- \ ic trary it does not appear to be finally estab-

0.2 lished that this activation energy refers to a diffusion process in the solid and not to processes of molecular dissociation and asso- ciation at the entrant and emergent surfaces. 60 I00 I20 I40 I60 180 With reference to the phase boundary per- meation, experiments have been carried out Fig. 3 Comparison. of absorption by palladium of in which palladium wires were only partly hydrogen and deuterium under a pressure of one atmosphere as a function of temperature immersed in electrolyte and then charged

Platinum Metals Rev., 1960, 4, (4), 135 with hydrogen by electrolysis. Macro-expan- to-volume ratio such as in the case of finely sion of the section immersed in electrolyte divided palladium blacks or evaporated metal results in the formation of quite a sharp films: for example, when used as diffusion boundary between this part of the wire and membranes, palladium tubes are generally the part not directly charged with hydrogen; impermeable to hydrogen gas at about 25°C. for wires of convenient diameter, say 32 Even when there is incomplete inhibition of SWG, this boundary can be quite readily seen the absorption of molecular hydrogen by under low power magnification (Fig. 4) and more massive specimens the kinetics are generally irreproducible (12). Since the rates of absorption of hydrogen can be varied by annealing and by cold working the palladium, it has been advanced DIRECTION OF (I 2) that both inhibition and irreproducibility BOUNDARY of absorption rates are related to defects M I GRAT I ON present in the solid. However, a likely alter- native solution of these effects is that there is inhibition of surface reaction steps (10,30), the most likely of which is the dissociation of hydrogen molecules. Inhibition of Equilibria in Aqueous Solution Although more massive specimens of palla- dium may be very slow to absorb molecular Fig. 4 Upward difusion, from a gas-electrolyte interface, of $-phase hydride formed by electrol.yysis hydrogen from the gas phase at room tem- perature, they are often found to absorb as a first approximation may be taken as hydrogen readily when they form the cathode separating regions of c( and 9 phase. Rates of of an electrolytic cell (2). For this reason, advance of the boundary into the part of the electrolytic methods have very often been wire not immersed in electrolyte have been employed to introduce hydrogen: the amount calculated (27) to be about 0.2 mm/hour at absorbed can be estimated by using a volta- about 30°C. meter placed above the cathode combined with Phase boundary migration rates of a similar a coulombmeter in series with the cell (31). order of magnitude are also suggested from However, it is only quite recently (9, 10, 32, experiments (28, 29) in which wires were 33, 34) that systematic investigation has been again charged over part of their lengths but made of the equilibrium between palladium the time dependent changes of hydrogen hydride and molecular hydrogen dissolved in content were measured after their removal solution. Results indicate (9, 10, 34) that this from electrolyte. equilibrium is relatively easily poisoned and that in the absence of electrode potential Inhibition of Equilibria in the measurements it is difficult to draw con- Palladium-Hydrogen System clusions as to the ability or inability of the There is much evidence (2) that, without surfaces in many earlier electrolysis studies to very careful preactivation of the palladium equilibrate in this way. sample, P-C-T data, under equilibrium con- It is important to realise that especially if ditions, below temperatures of about 200°C the surface step 5! + 5! + H, is inhibited the can only be conveniently derived in the gas electrolytic discharge of hydrogen ions can be phase if the sample has a very large surface- equivalent to very high pressures of hydrogen

Platinum Metals Rev., 1960, 4, (4), 136 gas (18, 35). Again, it must be noted that 6 G. Chaudron, A. Portevin and L. Moreau, Compt. rend., 1938, 221, 551;see also cupp, once having entered the metal, hydrogen can- ref. 2 not be readily released if the surface recom- 7 P. Wright, Proc. Phys. SOC.,1950, 63A, 727 bination step is poisoned. Although at about 8 F. A. Lewis and A. R. Ubbelohde, J. Chem. SOC.,1954, 1710 20 to 30°C (where the majority of electrolytic 9 J. P. Hoare and S. Schuldiner,J. Phys. Chem., studies have been carried out) the hydrogen 1957,619 399 content of palladium is not strikingly increased 10 T. €3. Flanagan and F. A. Lewis, Trans. Faraday SOC.,1959, 55, 1400, I409 by pressures exceeding one atmosphere, the 11 A. S. Darling, Platinum Metals Rev., 1958, 2, amounts of hydrogen absorbed after pro- 16 longed electrolysis at high current densities, 12 For a complete bibliography see D. P. Smith, ref. 2 and retained after its cessation, have been of 13 P. S. Perminov, A. A. Orlov and A. N. sufficient significance to have encouraged sug- Frumkin, Doklady Akad. Nauk S.S.S.R., gestions (12) that palladium hydrides formed 1952,843 749 14 P. L. Levine and K. E. Weale, Trans. Faraday by electrolysis may in some respects differ S0c.j 1960, 56, 357 from those formed by absorption of initially 15 R. Suhrmann, G. Wedler and R. Schumicki, molecular hydrogen from the gas phase. The Naturwiss., 1959, 46, 600 16 S. Schuldiner, J. Electrochem. SOC., 1959, more recent measurements do not support 106,440 such a difference (10,34). They have stressed 17 R. Neihof and S. Schuldiner, Nature, 1960, that, regardless of the method of introduction 185, 52.6 18 J. N. Andrews and A. R. Ubbelohde, Proc. of hydrogen, comparisons of hydrogen con- Roy. SOC., 1959, 253A, 6 tent can only be made with reference to pres- 19 J. E. Worsham, M. K. Wilkinson and C. G. sure, either directly or by electrode potential Shull, J. Phys. Chem. Solids, 1957, 3, 303 20 See especially B. Lambert and S. F. Gates, measurements, and that when comparisons are Proc. Roy. SOC.,1925, 108A, 456 made there must be no inhibition of any inter- 21 G. G. Libowitz,J. Phys. Chem., 1958,62,296 mediate stage of the absorption equilibria. 22 C. 0. Thomas and H. A. Smith, J. Phys. It may be noted that the hydrogen contents Chew 1959,63,427 23 M. von Stackelberg and W. Jahns, Z. Elek- of certain palladium alloys can be increased trochem., 1958,62, 349 by orders of magnitude by increasing the 24 R. Ash and R. M. Barrer, Phil. Mag., 1959, pressure above one atmosphere, and in such 4, (81, 1197 25 R. E. Norberg, Phys. Rev., 1952, 86, 745 cases hydrogen contents obtained from 26 G. Toda, J. Res. Inst. Catalysis, Hokkaido electrolytic studies can be grossly misleading Univ., 1958, 6, 13 (34) if quoted without the pressure reference 27 D. P. Smith and C. S. Barrett,J. Amer. Chem. SOC.,1940, 62, 2565 provided by electrode potential measure- 28 B. Duhm, Z. Physik, 1935, 94, 434, 95, 801 ments. 29. T. Sugeno and M. Kowaka, Mem. Inst. Sci. Research, Osaka Univ., 1954, 11, 119 30 See Ubbelohde, ref. 2 References 31 See, for example, G. Rosenhall, Ann Physik, I The Collected Chemical and Physical Re- ‘935, 24, (9,297 searches of Thomas Graham; (ed. R. A. 32 M. Breiter, H. Kammermaier and C. A. Smith), Edinburgh, 1876; see also biblio- Knorr, 2. Elektrochem., 1954, 58, 702 graphy given by D. P. Smith, ref. 2 33 R. J. Fallon and G. W. Castellan, J. Phys. 2 See D. P. Smith, Hydrogen in Metals, Chem., 1960, 64, 4, 160 Chicago University Press, 1948; cf. also 34 A. W. Carson, T. B. Flanagan and F. A. A. R. Ubbelohde, Trans. Faraday SOC., Lewis, Trans. Faraday Sac., 1960, 56, 363, 1932, 28, 275, and C. R. Cupp, Progress in 371 1953, Metal Physics (Butterworth: London), 35 F. W. Thompson and A. R. Ubbelohde, 4, 1°5 J. Awl. Chem., 1953,3,27 3 D. M. Nace and J. G. Aston, J. Amer. Chem. S0c-j 1957, 79,3619 4 See, for example, 0. Beeck, Faraday SOC. Disc., 1950, 8, 118 The concluding part of Dr Lewis’s article will be 5 T. J. Tiedema, C. Kooy and W. G. Burgers, published in the Janziary issue of ‘Platinum Proc. K. Ned. Akad. Wet., 1959, 62B, 34 Metals Review’.

Platinum Metals Rev., 1960, 4, (4), 137