Calorimetric Studies of the Chemisorption and Desorptionof Hydrogen on Nickel Films under Ultra HighVacuum ConditionsBY F. J. BROCKER AND G. WEDLERInstitut fur Physikalische Chemie und Elektrochemie der TechnischenHochschule Ilannover, GermanyReceiced 1st February, 1966Heats of chemisorption and desorption of hydrogen on evaporated nickel films have been de-termined at 273°K with an improved Beeck-type calorimeter. The heats of chemisorption aresignificantly lower than those previously reported in the literature. They are independent of thecoverage up to nearly a complete monolayer. When the adsorption became reversible, differentialheats of desorption could also be measured. The change of the electric resistance of the nickel filmswith increasing hydrogen coverage was simultaneously measured.The resistance passes through amaximum. The differential heat of chemisorption is not altered markedly at this coverage. Repeatedadsorption and desorption processes at high coverages indicate a further irreversible sorption ofhydrogen by the nickel films.The chemisorption of hydrogen on evaporated nickel films has been the subject ofnumerous investigations. Volumetric measureinents,l 2 studies of the work function,3the surface potential,49 5 or the contact potentia1,eS 7 investigations of the change ofthe electric resistivity 3 , 8-10 and the heat of adsorption 11-14 yielded much informa-tion on the interaction of hydrogen with nickel surfaces.The results obtained by different investigators and by means of different methods,however, are frequently not in agreement with one another. This may be due to theinfluence of the conditions, under which the nickel films had been prepared.Furthermore it is often difficult to compare the dependence of different adsorptioneffects on surface coverage, since this quantity is not unambiguously defined by theauthors.In order to explain the change of the electric resistance of nickel films due tochernisorption of hydrogen, additional measurements of the heat of adsorption anddesorption were carried out simultaneously on the same film.An improved Beeck-type calorimeter 15 was used, which could be operated under ultra high vacuumconditions. This is essential, since both the change of resistivity 8 as well as the heatof chemisorption *j are strongly affected even by small surface contaminations.EXPERIMENTAL‘The ultra high vacuum apparatus was the same as that described earlier.15 I t wasconstructed of Duran glass without stopcocks or greased joints in the high-vacuum section.The whole apparatus could be baked out at 380-400°C in order to achieve a vacuum betterthan 2 x 10-10 torr.The calorimeter differed from the original Beeck calorimeter in the use of a sphericaladsorption vessel and of a resistance thermoineter that consisted of only 10 p thick tungstenwires, embedded in glass sealed directly to the thin walled bulb.This shape of the calori-meter is more advantageous than a cylindrical tube, because a more uniform distribution of888 CHEMISORPTION OF HYDROGEN ON NICKEL FILMSthe hydrogen is achieved when the gas is admitted to the film.16-18 The sealed-in themo-meter guarantees a very good heat contact.The maximum sensitivity of the calorimetricdevice is a deflection of 6.4 mm on the recorder chart for a temperature change of 1 x 10-5 deg.at room temperature and of 2-6 mm at 77°K.At two spots opposite to one another, platinum foils, 5 ,u thick, were sealed to the innerwall of the calorimeter bulb in order to measure the electric resistance of the films evaporatedfrom the small coil of spectroscopically-pure nickel wire M.R.f~A -. I 1.- *FIG. 1 .-Calorimeter. F, outer jacket ; Thy resistance thermometer ; Rth, connection terminal ofresistance thermometer ; Pt, platinum contact foils ; Rf, connection terminals of the platinum foils ;My nickel wire.During the evaporation of the films the calorimeter was surrounded by a bath of liquidnitrogen.A hydrogen pressure of about 10 torr was maintained in the outer jacket F of thecalorimeter in order to keep the temperature of the films below 150°K. The evaporationof the films was carried out at a pressure less than 2 x 10-10 torr with a rate of about 10 Almin.After the evaporation the films were sintered at 60°C for 1 h, until no further decrease of theresistance of the film occurred due to ordering effects.The adsorption experiments were carried out at a temperature of 273"K, while thecalorimeter was surrounded by a bath of crushed ice, and the outer jacket had been pumpedto high vacuum.The temperature drift was usually about 10-5 deg./min.The hydrogen was purified by diffusion through a palladium tube. It was added in smalldoses. During each adsorption process the change of the resistance of the evaporatednickel film, the change in temperature (resistance thermometer) and the gas pressure bothin the storage bulb as well as near the calorimeter could be measured simultaneously bymeans of four Wheatstone bridges, amplifiers and four recorders. Xt can be estimated, thatthe change in resistivity of the film is only due to the electronic interaction and not to theheat evolved during adsorption.The calibration of Beeck-type calorimeters has been the main difficulty in exact determina-tion of differential heats of adsorption.123199 20 The heat capacity of the calorimeter wasdetermined by means of two methods recently described.15 The " stationary method " issimilar to the '' equilibrium method " used by Wahba and Kemball,l2 the " pulse method "is similar to that used by Bagg and Tompkins.2F. J.B R ~ C K E R AND G . WEDLER 89The stationary method makes use of the equilibrium established between the heat gener-ated by the electric current passing through the evaporated film and the loss of heat by radia-tion. The heat capacity is evaluated bysE2aC =4-18 x IO-~R,A,~/C'where s = deflection of the recorder, when varying the resistance Rth of the thermometerby 10-4Rth, E = potential drop along the film resistance Rf, a = temperature coefficientof the resistance thermometer, A , = deflection of the recorder at temperature equilibrium,Y = Newton's cooling constant, r/C is obtained from the logarithmic plot of the heating orcooling curve.When using the pulse method the electric current is switched on only for a short timeinterval t.With the assumption that the rate of increase of temperature dATldt is pro-portional to the sum of the rates of heat production dQ+/dt and heat loss dQ-/dtdAT --- - l(,Q+ - +- df;) ' (2) dt C dtC can be calculated bywhere A is the deflection of the recorder, when the current is switched off at time t.The values of C calculated by the stationary and the pulse method are slightly different,if only short pulse times are used. For instance, the value obtained with pulse times ofabout 10 sec was 0-573 cal/deg., whereas the value obtained with the stationary method is0-607cal/deg.Thus the pulse method results in a capacity of the calorimeter, which isabout 6 % smaller than the value calculated by means of the stationary method. This isdue to the fact, that eqn. (3) is only valid, if heat equilibrium in the system metal filmlglasslthermometer is established. This seems not to be achieved during short pulse times.Nevertheless, the value obtained by means of the pulse method was used, since the processof adsorption also occurs within such short periods.The heats of chemisorption AU were determined by means of a numerical method 1 5 ~ 2 1taking the kinetics of gas adsorption into account. Frequently the rate of adsorption iscontrolled by the diffusion of the gas resulting in a first-order law.In this case AU isgiven byCNL10-4 k,-riC dAUS k,n dD______ AU = (4)where NL = Avogadro's number, k, = rate constant of the gas adsorption, which can bedetermined from the change of the resistance of the evaporated film, n = number of moleculesadsorbed, dA/dD = slope of the straight line obtained, when A is plotted against{exp (-(r/C)t)-exp (--k,,t)). Eqn. (4) may only be used, if the kinetics of the adsorptionis of first order. Otherwise no such simple numerical evaluation can be carried out. Thisis the case, e.g., when the gas is desorbed by pumping.Following a proposal by Wittig22 the heats of adsorption and desorption were alsocalculated by means of the equation,(dQ+/dt)dt = CdT'+ r(Tk- T0)dt.(5)On integration between the time limits tl and t2 one obtainswhere Tk is the temperature of the calorimeter and TO the temperature of the environment.Eqn. (6) is only valid, if the calorimeter has again attained its original temperature duringthe interval tl to t2 (quasi-isotherm)90 CHEMISORPTION OF HYDROGEN ON NICKEL FILMSQ+ is either the heat generated by the electric current, the adsorption or the desorptionprocess, i.e., the heats of adsorption and desorption may be determined by comparison ofthe areas on the recorder chart. Since this method is independent of the underlying kinetics,it was used to determine the heats of desorption, for which the kinetics are not known.RESULTSFig. 2 shows a typical recorder diagram for the temperature change due to chemi-sorption (a) and desorption (b).The amount of hydrogen chemisorbed was 6-74 x 1015molecules, corresponding to 8.64 x 1013 molecules per cm2 or about 1/20 of a completemonolayer. The monolayer capacity is defined as that coverage, at which eachnickel surface atom is occupied by one hydrogen atom. The maximum increase oftemperature in fig. 2d was 2.9 x 10-4 deg., the heat of chemisorption 15.6 kcal/mole.In the desorption, 8.4 x 101s molecules were pumped off, the temperature was loweredby 2.9 x 10-4 deg., and the heat of desorption was 19 kcal/mole.3Ln 3 '0m0Wtime (min)u% O-00(d - I - .m .c1-3 s-2 -b-3t 1 I I I I I0 I 2 3 4 5 6time (min)FIG. 2.-Variation of the temperature of the calorimeter during the adsorption (a) and desorption (b)of hydrogen.The variation of the resistance of the nickel film and of the heat of chemisorptionwith coverage is shown in fig.3. The doses of hydrogen admitted were about1.1 x 1014 molecules per cm2 of surface.The range, in which the adsorption is reversible is of particular interest. Forequilibrium pressures of higher than 10-5 torr, the heat of adsorption decreases.The residual hydrogen in the gas phase is not adsorbed even after a few hours. Ifthe gas, however, is pumped off, further doses of hydrogen may be adsorbed with ahigh heat of adsorption, as is to be seen from fig. 3. The full lines refer to adsorption,the dotted lines to desorption. The coverage at which the heat of adsorption beginsto decrease is shifted to higher coverages with each new admission.The heats ofdesorption (triangles) lie very well on the expected desorption curves (dotted lines).The electric resistance of the hydrogen-covered film increases when the gas isdesorbed. Further additions of hydrogen result in a new descrease of resistanceF. J . BROCKER AND G. WEDLER 91After pumping for a very long time (about 10 h) the film resistance nearly attains thevalue of the resistance maximum. The adsorption curves of the resistance are shiftedparallel to higher coverages with each new gas admission. At a given hydrogenpressure in the gas phase both the resistance of the film as well as the heat of adsorptionhave a certain value, although the amount of adsorbed hydrogen increased with eachadsorption-desorption cycle.0 0 - 5 1.0 1.5coverage (lo15 molecules/cm2)FIG.3.-Variation of resistance and of heat of chemisorption with coverage.The good reproducibility of the calorimetric measurements is seen in fig. 4, inwhich the results obtained on three different films are compared. The heat ofadsorption is always the same within the experimental error; the only differencebeing that the adsorption capacity of film (a) is slightly higher.DISCUSSIONThe properties of the nickel films prepared as described have been thoroughlystudied in earlier papers. The films exhibit a fairly small internal surface, theirroughness factor being dependent on the thickness. For films used in this paper(about 100 .$) the roughness factor is about two.16s23 No proportionality betweenthe thickness and the adsorption capacity 9 was found, as has been reported byKlemperer and Stone 13 who prepared their films in a different way.The films arefairly homogeneous, consisting of small crystals of nearly equal size.23 The latticeconstant of the f.c.c. crystals is slightly smaller than that of bulk nickel.24 Theelectric conductivity of the films may be explained by the theory of Fuchs andSondheimer,25 showing that they have no island structure26 and that the tunneleffect plays no important role in the conductivity.The change of the film resistance with increasing hydrogen coverage observed inthese experiments is the same as that described in earlier ~apers.3~ 8-10 Thepartial reversibility of this effect at high coverages, however, has so far not beenstudied in detail and will be discussed in connection with the heats of adsorption.The heats of chemisorption, as well as their dependence on coverage, are differentfrom those reported in the literature.For comparison, the results of Beeck, Coleand Wheeler,ll Wahba and Kembal1,lz Klemperer and Stone,l3 Brennan and Hayes,l92 CHEMISORPTION OF HYDROGEN O N NICKEL FILMSRideal and Sweett,27 and of this paper are shown in fig. 5. It has to be consideredthat surface coverage is defined in a different way by the various authors, thereforethe abscissa of fig. 5 has merely a qualitative meaning. The characteristic feature ofthe curves, however, is not effected.The initial differential heats of chemisorption range from 18 to 42 kcal/mole.Since in all cases the differential heat of chemisorption is about 20 kcal/mole, whenthe monolayer is completed, the change in the differential heats with surface coverageis the larger the higher the initial heats of adsorption are.coverage (1015 molecules/cm2)FIG. 4.-Variation of heats of chemisorption of hydrogen with coverage as obtained on three differentnickel films (A, X , e) of about 100 A thickness. The arrows indicate the position of the maximumin the resistance curves.Two factors may be responsible for the different results illustrated in fig. 5, thestructure of the evaporated films and the vacuum conditions.The structure isdependent on the temperature of deposition, the rate of evaporation and the handlingafter the evaporation, i.e., the temperature and time of tempering.The films usedby Beeck, Klemperer and Stone, Wahba and Kemball, and by Brennan and Hayes,were deposited at room temperature. Their structure and their properties will besomewhat different from those of the films used in this work. It is, however, not tobe expected, that the difference is solely due to such a structural effect. This can beconcluded from the fact that the differential heats of adsorption of hydrogen on anickel film deposited at 77°K and annealed at room temperature were the same as thosein fig. 4, referring to nickel films tempered at 60°C. On the other hand it has beenclearly demonstrated by measurements of the photoelectric work function 28 and theelectric resistivity,29 that the properties of nickel films sintered at room temperatureand at 60°C are significantly different.The vacuum conditions are known to have a marked influence on the heat ofadsorption and its dependence on the surface coverage.Recently, this has beenclearly shown by calorimetric studies of hydrogen chemisorption on evaporatedtitanium films.15 For films evaporated and sintered under a pressure smaller than2 x 10-10 torr the heat of chemisorption is 27.5 kcal/mole and is independent of surfacF. J. BROCKER AND G . WEDLER 93coverage up to half a completed monolayer. If the pressure, however, was 1 x 10-9torr, the initial heat of adsorption was 35 kcal/mole and decreased on further hydrogenadsorption.Further experiments have to be carried out in order to study the influence of thestructure of the films and of the vacuum conditions on the heat of adsorption.Fromthe results in fig. 3 and fig. 4, however, it may be deduced, that the films prepared in thedescribed manner, are energetically fairly homogeneous with respect to the adsorptionof hydrogen at room temperature. It would appear that the constancy of the heatof chemisorption is not due to an immobile layer, as suggested by Beeck.11 Thisconcept would not be consistent with the finding that the hydrogen can be pumpedoff to an appreciable extent.‘(1 o l , , , , I0 ol* 5 1.0coverage (monolayer : 8 = I)FIG. 5.-Dependence of heat of chemisorption of hydrogen on coverage for evaporated nickel films.1, Klemperer and Stone 13 - - -; 2, Beeck, Cole and Wheeler 11 -; 3, Wahba and Kernball12 - - - . - ; 4, Rideal and Sweett z7 .. . ; 5, Brennan and Hayes 14 - . . . - . . . - ; 6, this paper - - -.In earlier papers 3 the existence of the resistance maximum was attributed to thepresence of two or more species of adsorbed hydrogen, one of which should increasethe resistance (hydrogen polarized towards H-), the other decrease the resistance(proton +electron). It would be expected that these species would give different heatsof adsorption, i.e., that the heats of adsorption should be altered at a coverage nearthe resistance maximum. The heats of adsorption, however, are practically in-dependent of surface coverage.The slight minimum in fig. 3 may be within theexperimental error. * Although the existence of different species of adsorbed hydrogenhas been experimentally verified,s. 30 the reason for the appearance of the maximumin the resistance curve cannot be given before the nature of these species is known.Both from the behaviour of the resistance as well as of the differential heats ofchernisorption and desorption at coverages beyond the resistance maximum it may b94 CHEMISORPTION OF HYDROGEN ON NICKEL FILMSconcluded that part of the hydrogen is irreversibly sorbed. As mentioned above theresistance as well as the heat of adsorption are functions of the hydrogen pressure.For the heats of adsorption, this has also been observed by Klemperer and Stone.13On the other hand, the coverage is not a function of the gas pressure. Consequentlythat part of the hydrogen, which is irreversibly adsorbed at high coverages does nottake part in the equilibrium between the gas phase and the adsorbed phase.Thesame conclusion has been reached for hydrogen adsorption on palladium films atroom temperature 31 using isotope techniques.The authors thank Prof. Dr. R. Haul for his interest and for numerous discussions,Dr. H. Strothenk, who studied the titanium + hydrogen system, and Dipl. Phys.,P. Wii3mann for their assistance. The support of the Deutsche Forschungsgemein-schaft, the Verband der Chemischen Industrie and the Max-Buchner-Forschungs-stiftung is gratefully acknowledged.1 Gundry and Tompkins, Trans.Faraday SOC., 1956,52, 1609.2 Gundry and Tompkins, Trans. Faraday Soc., 1957,53, 218.3 Suhrmann, Mizushima, Hermann and Wedler, 2. physik. Chem., 1959,20, 332.4 Culver, Pritchard and Tompkins, 2. Elektrochem., 1959, 63, 741.5 Crossland and Pritchard, Surface Sci., 1964, 2, 217.6 Mignolet, Disc. Farday SOC., 1950, 8, 105.7 Mignolet, Rec. trav. chim., 1955, 74, 685.8 Sachtler and Dorgelo, Bull. SOC. chim. Belg., 1958, 67, 465.9 Mizushima, J. Physic. SOC. Japan, 1960,15, 1614.10 Ponec and Knor, Coll. Czech. Chem. Cornm., 1960,25,2913.11 Beeck, Cole and Wheeler, Disc. Faraday SOC., 1950, 8, 314.12 Wahba and Kemball, Trans. Faraday Suc., 1953,49, 1351.13 Klemperer and Stone, Proc. Roy. SOC. A, 1957,243, 375.14 Brennan and Hayes, Trans. Farraday SOC., 1964, 60, 589.15 Wedler and Strothenk, Ber. Bunsenges. physik. Chem., 1966,70, 214.16 Wedler and Fouad, 2. physik. Chem., 1964,40, 12.17 Suhrmann and Wedler, 2. Elektrochem., 1959,63, 748.18 Brennan and Jackson, Proc. Chem. SOC., 1963, 375.19 Brennan, Hayward and Trapnell, Proc. Roy. SOC. A , 1960,256, 81.20 Bagg and Tompkins, Trans. Faraday Soc., 1955, 51, 1071.21 Wedler, 2. physik. Chem., 1960, 24, 73.22 Wittig, private communication.23 Suhrmann, Gerdes and Wedler, 2. Naturforsch., 1963, I&, 1211.24 Suhrmann, Wedler, Reusmann and Wilke, 2. physik. Chern., 1960,26,85.25 Sondheimer, Adv. Physics, 1952, 1, 1.26 Wedler and Fouad, 2. physik. Chem., 1964,40, 1.27 Rideal and Sweett, Proc. Roy. SOC. A , 1960,257,291.28 Suhrmann and Wedler, 2. angew. Physik., 1962, 14, 70.29 Wedler, Brocker, Kock and Wolfing, in Grundprobleme der Physik dunner Schichten (Gottingen,30 Rootsaert, van Reijen and Sachtler, J. Catalysis, 1962, 1, 416.31 Suhrmann, Schumicki and Wedler, 2. physik. Chem., 1964,42,187.1966), in press.* Addedinproof: Further experiments have shown that the minimum in the heat of chemisorp-tion in figs. 3 and 4 is a reproducible effect