J . Chern. Soc., Faraday Trans. I , 1987, 83 ( l l ) , 3271-3282 Hydrazine Reduction of Transition-metal Oxides Donald M. Littrell, Daniel H. Bowers and Bruce J. Tatarchuk* Department of Chemical Engineering, Auburn University, Alabama 36849, U.S.A. The surface interactions of the thermodynamically stable oxides of the first- row transition metals with hydrazine, N,H,, have been assessed by X-ray photoelectron spectroscopy (X.P.S.), and the observed reactivity explained on the basis of the oxides' thermodynamic and acid-base properties. Microgravimetry, transmission electron microscopy (TEM) and X.P.S. studies of copper(1r) oxide (CuO) reduction by hydrazine indicate morphological changes associated with enhanced reactivity. Comparison of hydrazine and hydrogen reduction treatments shows that CuO reduction by hydrazine is more spontaneous and provides a greater increase in surface area than does hydrogen at similar reaction conditions.The interaction of hydrazine, N,H,, with metal oxides has wide and diverse applications for the microelectronic, electrochemical and aerospace industries. Some transition-metal oxides are known to be easily reduced by hydrazine,1-6 with the resulting formation of safe, non-toxic products : an n aM,O, + bN,H, -+ M,,O,, + bN, + 2bH,O where m' m anm' 0 G 7 < - - and am=-+2b. n n n' Surface analysis during metal oxide reduction by hydrazine has received little attention, although the heterogeneous decomposition of hydrazine on various metals has been extensively studied.'-16 In view of the relative lack of data, surface interactions of hydrazine with the thermodynamically stable oxides of the first-row transition metals were assessed by X- ray photoelectron spectroscopy (X.P.S.) to determine those oxides most easily reduced.An explanation of the resultant reactivity has been suggested based on the thermodynamic, physicochemical and catalytic properties of these oxides. In con- junction, morphological changes which occur during CuO reduction have also been examined using X.P.S., microgravimetric techniques and transmission electron microscopy (TEM). Experimental Equipment X-Ray Photoelectron Spectroscopy All X.p. spectra were obtained using a Leybold-Heraeus PAH 10/ 1 1 spectrometer with A1 Kal,, X-rays (1486.6 eV). The angle of incidence of the photons was 60" off-normal, and the emitted electrons were collected normal to the sample surface except where noted.The base pressure of the system during hydrazine dosage was typically Pa. The spectrometer dispersion constant was set by measurement of the energy difference 327 13272 Hydrazine Reduction of Transition-metal Oxides between the Cu 2p3/2 and Cu 3p3,, lines of a sputter-cleaned copper foil. A polycrystalline gold sample integrated into the sample rod was used to calibrate the spectrometer work function; all binding energies were referenced to the Au 4f,,2 peak at 83.8 eV. The full width at half maximum (f.w.h.m.) of the Au 4f,,a peak at typical spectrometer settings was 0.8 eV. Spectra were digitally collected with a Tracor-Northern 1710 multichannel analyser and stored for future retrieval/analysis.Deconvoluted spectra were fitted with Gaussian peaks by a least-squares procedure. X.P.S. samples were treated with hydrazine in situ using a molecular-beam doser. Hydrazine vapour was introduced into the system through a glass capillary array with 10 pm pores positioned 5 cm from the sample position. The expansion effect of this collimated wave of gas allowed an estimated 100-fold enhancement in hydrazine pressure at the sample, compared to the N,H, background pressure based on flux calculations for cosine-type effusion sources, estimated pumping speeds and known background pressures. A stainless-steel jacket surrounding the array permitted the option of an additional differential pumping stage. Micr ograu ime t r ic Analyses A Cahn 2000 recording microbalance, with a capacity of 1.5 g and an ultimate precision of f0.2 pg, was used in the morphological studies of reduced CuO.A rough pump backed by a liquid-nitrogen trap provided an ultimate pressure capability of 0.07 Pa in the Pyrex vacuum chamber surrounding the balance. Pressures were monitored during B.E.T. isotherms with a Wallace and Tierman pressure gauge accurate to f 340 Pa. Buoyancy effects, measured using helium at 77 K, were corrected to a nitrogen equivalent buoyancy by multiplication of the gas density ratio (NJHe). Buoyancy corrections required during reaction with hydrazine were measured at ambient conditions and corrected in a similar manner. For hydrazine reduction studies in the microbalance, the weight of the oxide was monitored during exposure to ca.560 Pa of hydrazine. A helium carrier gas at 100 kPa was bubbled through a fritted Pyrex dispersion tube located in a bulb of liquid anhydrous hydrazine at 298 K. The partial pressure of the hydrazine was reduced to 560 Pa by dilution of this hydrazine-saturated He stream, with an additional helium stream flowing over the balance assembly. This stream was also used to minimize hydrazine contact with the balance mechanism. Transmission Electron Microscopy Micrographs in the morphological study were taken using a Phillips EM-300 transmission electron microscope at 60 kV beam energy. The background pressure in the microscope was < 5 x lop4 Pa. Samples were treated for 9 h in a Pyrex U-tube reactor at 530 Pa of hydrazine in an analogous method to the gravimetric studies.Note that samples were exposed to atmosphere during loading into the microscope. Some contamination was introduced during TEM examination, as evident from a light haze at specimen surfaces. This contamination was believed to be carbonaceous in nature.17 Materials Hy drazine Technical-grade anhydrous hydrazine ( < 1.5 O h H,O) provided by Olin Chemicals was subjected to alternating freeze-thaw cycles under vacuum for removal of residual non- condensible gases. Mass-spectral analysis'' confirmed the composition of the gas phaseD. M. Littrell, D. H. Bowers and B. J. Tatarchuk 3273 to be predominantly molecular hydrazine ; gas-phase photodecomposition or catalytic decomposition from the leak valve’s copper seal had no significant effect, and contamination of the hydrazine by water was minimal.X . P.S. Specimens Oxides of vanadium, iron, cobalt, nickel, copper and zinc were prepared from polycrystalline metallic foils of 0.1 mm thickness (Johnson-Matthey Chemicals, purity > 99.95 Yo). Oxides of chromium and manganese were prepared from flake materials (Electronic Space Products). All metals were degreased with acetone and methanol before being oxidized in air at appropriate temperatures to form the indicated oxides, (i.e. V205, Cr,03, Mn203, Fe203, Co304, NiO, CuO and ZnO). Evidence for the formation of these species was obtained by (i) comparison of the observed X.p. spectra with those obtained from the literature (V205,19-22 Cr203,19920*22-26 Mn 2 0 3’ 2 2 9 2 6 - 2 9 (ii) comparison of the experimentally observed lattice oxygen to metal atomic ratios with those expected for the appropriate oxide stoichiometry (table 1).These comparisons utilized just that portion of the 0 1s signal associated with lattice oxygen and were corrected for photoionization c ~ o s s - s ~ c ~ ~ ~ ~ s , ~ ~ electron escape depths45 and empirically determined spectrometer efficiency Chromium specimens appeared to exhibit multiple oxidation states : predominantly Cr203 with traces of CrO, and CrO,. Fe203,22, 26,28,30-34 Co,04,22,28,35,36 Ni0,22,32,35,37-40 Cu022,35,37,42,43 and Zn022943) and Micrograuimetric Specimens Only CuO samples were examined during microgravimetric studies ; CuO was prepared from copper wool (fine grade 99.9 YO, Carey Electronics).Sample pretreatment before oxidation included cleaning in trichloroethane and acetone to remove organics and subsequent placement in vacuum (0.1 Pa) for 1 h at 298 K to remove adsorbed species. The copper was prereduced in hydrogen at 673 K for 8 h at 100 kPa. Oxygen was removed from the hydrogen stream using, in series, a commercial hydrogen purifier (Deoxo, Fischer Scientific), a copper turnings trap at 523 K and 5X molecular sieves at 77 K. Copper was subsequently oxidized in oxygen (99.99 O h ) at 723-773 K and 100 kPa for 15-60 h depending on the desired scale thickness. These conditions are appropriate for formation of 75-100 % Cu0.47948 The temperature was increased from room temperature to reaction temperature (723 K) at the rate of 100 K h-l.Transmission Elect yon Microscopy Specimens Copper grids (Balzar Union) used in TEM studies are pre-reduced and oxidized in a manner similar to samples prepared for microgravimetric studies. Results and Discussion Thermodynamic Considerations Reactivity of an oxide in hydrazine is based on both kinetic and thermodynamic driving forces. Identification of those oxides most easily reduced by hydrazine was determined by subjecting each oxide to a standard hydrazine treatment in the X.P.S. apparatus: an in situ dose for 30 min at a surface N,H4 pressure of ca. 6.7 Pa using the molecular-beam doser.3274 Hydrazine Reduction of Transition-metal Oxides Table 1. Reducibility of transition-metal oxides in hydrazine" oxygen : metalb reducibility metal experimental theoretical metal oxide factor (R)" V Cr Mn Fe c o Ni c u Zn 2.80 1.32 1.46 1.28 1.08 0.98 0.98 d - 2.50 1 S O 1 S O 1 S O 1.33 1 .oo 1 .oo 1 .oo 0.97 CrO,, CrO, 0.90 0.98 0.99 coio, 0.90 NiO 0.94 CUO 0.05 ZnO 0.95 a Operating conditions : 30 min at 295 K, 6.7 Pa at the sample surface.Lattice oxygen to metal ratio corrected for photoionization cross-sections, electron escape depth and analyser efficiency ; accuracy estimated at 20 %. R = (oxygen : metal) after reduction/oxygen : metal) of the oxidized precursor; includes all observable oxygen species. An elevated 0 : C r ratio results from contamination of the Cr20, with CrO, and CrO,; the 0xygen:metal ratio was not attempted due to the complexity of deconvoluting the 0 1s spectrum. A reducibility factor (R), defined as the oxygen-to-metal ratio of the reduced sample divided by the oxygen-to-metal ratio of the oxidized sample, was used to compare the relative ease of reduction for each oxide (table 1).The reducibility measurement was based on all types of oxygen observed by X.P.S., including contributions from lattice oxygen, hydroxyl species, adsorbed water and carbonyl species. Of the specimens examined, CuO was the most easily reduced (lowest R), losing 95 YO of its initial oxygen content following hydrazine treatment. The nature of oxygen species at the surface is important when considering oxide- N2H4 interactions. The 0 1s photoelectron spectrum of CuO (fig. 1) exhibited multiple oxygen species identified as : (i) lattice 3 5 y 3 7 3 42 at 530 eV, (ii) surface hydroxyl specie^*^-^^ at 531.6 eV and (iii) adsorbed waters0ys2 at 533.5 eV.Water is known to desorb at room temperature, whereas contamination by carbon-containing species may contribute to the peak at 533.5 eV.31 The peak corresponding to surface OH groups was observed in the 0 1s spectra of all transition-metal oxides examined in this study; the position of this peak (531.61 0.2 eV) was independent of the oxide studied, since the contribution to the potential of the inner shell of oxygen from the proton in the OH group over-rides the influence of the cation.49 Although the lattice-oxygen component of CuO was removed by thermal reduction at 325 K, in agreement with the Cu-0 phase diagram,47 the peak assigned to the hydroxyl species remained essentially unchanged at temperatures up to 450 K.Previous have reported the presence of these strongly chemisorbed OH groups on CuO at temperatures approaching 775 K. A decrease in intensity of both lattice and hydroxyl oxygen was noted upon hydrazine exposure (fig. 2). Subsequent deconvolution of the 0 1s spectra showed an increase in the relative proportion of hydroxyl to lattice oxygen, suggesting preferential removal of lattice oxygen and/or generation of surface hydroxyl species resulting from water formation in the N2H4-oxide reaction. Hydrogen abstraction from adsorbates by chemisorbed oxygen has been the subject of previous inve~tigations.~~9 53-55 X.P.S. studiesso* 55 have confirmed the formation of surface hydroxyl species resulting from these reactions. These hydroxyl groups subsequently dehydroxylate through the desorption of 55 Roberts et a1.l'.543275 543 5 2 3 binding energy/eV Fig. 1. Deconvolution of the 0 Is spectrum of CuO. Species identification: (1) 530 eV, lattice oxygen; (2) 531.5 eV, surface hydroxyl species; (3) 533.5 eV, chemisorbed oxygen-containing species. 54 3 523 binding energyfev Fig. 2 . 0 1s spectra of CuO with increasing exposure to hydrazine : (a) oxidized sample, (b) 50, ( c ) 500 ( d ) 3000 and (e) > 50000 laiigniuir N,H, (1 langmuir = Torr s).3276 Hydrazine Reduction of Transition- metal Oxides suggest that strong hydrogen-oxygen bonding provides the kinetic driving force for oxygen-induced chemisorption of hydrogen donors ; the thermodynamic driving force being primarily dependent upon the formation of water. The CuO-N,H, system investigated here may parallel the oxidized copper-NH, reaction reported by Matloob and Roberts." The kinetic driving force for the dissociation of the N-H bonds results from strong hydrogen-oxygen bonding.This interaction results in an increase in adsorption energy of the adsorbate and a corresponding decrease in the activation energy for the dissociation of the N-H bond." The ease of reducibility of CuO may be attributed to its relative thermodynamic and kinetic instability in comparison with the other oxides examined. These oxides exhibited a monotonic increase in their relative ease of reduction us. the heat of formation of the oxide per mole of oxygen (i.e. -AH& per mole of oxygen) as shown in fig. 3.This type of correlation is indicative of a relationship between bulk and surface heats of f o ~ m a t i o n ~ ~ . ~ ~ and has been widely used to qualitatively predict the relative rates of surface reaction^.^*-^' the greater reactivity of Ni oxide over Fe oxide with -NH, groups of ethylenediamine, NH,CH,CH,NH,, was also rationalized on the basis of the oxide's thermodynamic stability (i.e. oxide heat of formation). Ethylenediamine dissociatively adsorbs on metallic iron and nickel to form NH,CH,CH,NH-* at 300 K. Preoxidation of iron inhibits dissociative adsorption at oxygen coverages below half a monolayer, resulting in largely molecular adsorption. In contrast, nickel surfaces with similar oxygen levels still provide dissociative adsorption sites for ethylenediamine with removal of surface oxygen in the sequence shown below: In a similar NH,CH,CH,NH,(g) + 2* 3 NH,CH,CH,NH-* + H-* (2) 2H-* + 0-* -+ H,O(g) + 3*.(3) In this process oxygen vacancies are required to remove a hydrogen atom during dissociative adsorption [reaction (2)] and the subsequent scavenging of surface oxygen [reaction (3)] has an autocatalytic effect from the resulting production of oxygen vacancies. Evidence for this is provided by the predominantly molecular adsorption of ethylenediamine on nickel oxide at saturated coverages, indicating that the -NH, proton is inactive toward the nickel oxide surface and requires oxygen vacancies to promote dissociative adsorption. The molecular processes governing the reaction with oxidized iron are quite different from that on nickel; surface oxygen on iron is less reactive with a weak acid such as ethylenediamine, and dissociative adsorption is not observed on iron surfaces which are partially covered with oxygen. The basis for this phenomenon may result from the relative ease with which an iron surface forms an oxide through the interaction with water at low pressure and room temperature, owing to the much larger heat of formation of iron oxide than that of nickel oxide.The reduction reaction of oxides with relatively small heats of formation (e.g. CuO and Cu,O) by -NH, groups (i.e. hydrazine) may be self-propagating from the resulting formation of metallic sites adjacent to chemisorbed or lattice oxygen. The low degree of reduction exhibited by the majority of the transition-metal oxides examined here may be due to surface kinetic limitations at the temperatures and pressures of this study rather than the respective driving force provided by surface thermodynamics.Acid-Base Interactions Although the high reactivity of CuO might be expected on the basis of thermochemical properties, the CuO/N,H, reaction may also be considered in terms of acid-baseD. M. Littrefl, D . H . Bowers and B. J . Tatarchuk 3277 1.2 1.0 0.8 R 0.6 0.4 0.2 0 Q CUO ~~ 3 375 -AH& per mol oxygen/kJ mol-I Fig. 3. Reduction of first-row transition-metal oxides in hydrazine. Reducibility factor = (oxygen : metal) reduced/ (oxygen : metal) oxidized precursor. Reducing conditions : 6.7 Pa hydrazine at 295 K for 30 min. interactions. Hydrazine decomposes upon adsorption to metal surfaces through the formation and interaction of free radical^.^ One step in this decomposition involves formation of bonds between the nitrogen atoms in hydrazine and the incomplete d orbitals of the Although no simple correlation has been found between electronic structure and reactivity, a metal’s electronic configuration, particularly of the d band, has been proposed as an index of catalytic Dowden et aZ.64 further proposed that metal oxides share a similar dependence based on their metal ion configuration.The observed reactivity of CuO in this study conforms to this theory since CuO possesses a relatively unstable d9 configuration, while Cu,O, although having a complete d shell, is uniquely unstable as assessed by the d-s promotion energy.64 The change in acidity and basicity of oxides follows two simple rules:65 (1) acidity increases with an increase in charge of the transition-metal ion and (2) proceeding from left to right within a specific period, acidity increases.Assuming the first rule takes precedence over the second, these two rules, taken in conjunction, pinpoint CuO as being one of the most basic oxides examined in this study. Chemisorbed oxygen is known to have strong basic characteristics, 53, 66 and hydrazine, although normally considered a Lewis base, is a source of protons. In addition to the Lewis basicity ascribed to it, adsorbed oxygen may act as an ‘adsorption promoter’ by creating Lewis acidity (electron deficiency) at adjacent metal atoms.66 These perturbed atoms are stronger Lewis-acid sites than metallic copper atoms and may interact with lone pairs of electrons on the hydrazine molecule.In this hydrazine-metal interaction, hydrazine would be considered a Lewis base, in accordance with the proposal of Eberstein and Glassrnan6, of bond formation between the nitrogen atoms in hydrazine and the incomplete d-orbitals of the metals. An increased reactivity can be ascribed not only to these acid-base effects but also to3278 Hydrazine Reduction of Transition-metal Oxides changes in the surface coverage of hydrazine. As increased hydrazine surface coverage (0) on the metal oxide is expected owing to (i) promotion of surface acidity by adsorbed oxygen as discussed above, resulting in enhanced N-metal bonding and (ii) an increase in the heat of adsorption caused by strong hydrogen-oxygen bonding following dissociative ad~orption.~~ Morphology of Reduced Copper( n) Oxide X.P.S.Studies In addition to the role of an oxide's physicochemical properties, the reactivity of a metal oxide toward hydrazine may be enhanced by morphological changes in the solid, particularly if such changes provide increased surface areas for reaction. Although X-ray photoelectron spectra of hydrazine-treated samples indicated elemental copper based on the position of the Cu 2p peaks (fig. 4) and the oxygen : metal ratio, visually the specimens did not return to the characteristic colour of this metal. These observations suggest the formation of a metallic layer upon bulk-like CuO. For this reason the morphology of reduced CuO was studied further using angle-resolved X.P.S.A thickly oxidized (i.e. > 1 pm) CuO layer was reduced in situ with hydrazine at conditions appropriate for partial surface reduction (200000 langmuir at 295 K). With the sample plane perpendicular to the analyser, the X-ray photoelectron spectrum of the Cu 2p region showed a mixture of CuO and Cu,O/Cu (fig. 5). However, at an electron take-off angle of lo", with its corresponding five-fold increase in surface sensitivity, a decrease in the CuO content was evident (fig. 5). The principal peaks shifted to the lower binding energies indicative of either metallic copper or Cu,O, while the shake-up satellites characteristic of CuO decreased in intensity indicating a surface enrichment of a more reduced species.The presence of the bulk-like CuO substrate was further confirmed by argon ion bombardment. A fully reduced sample, as assessed by the Cu 2p and 0 1s spectra, was rigorously sputtered to remove the metallic surface layer. Although this technique has inherent chemical reduction tendencies for CUO,~* an increased oxygen content after sputtering (ca. 10 nm) also indicated the existence of sub-surface oxide. Owing to density changes in the crystal lattice upon transformation from CuO to metallic copper, a roughening or pitting of the surface may be expected, resulting in exposure of underlying CuO for further reaction with hydrazine. The newly exposed CuO in these pores may have minimal contribution to angle-resolved X-ray photoelectron spectra, owing to the absence of a line-of-sight electron trajectory from these regions to the analyser.Microgravime tric Studies Confirmation of the proposed change in morphology suggested above was evident from microgravimetric studies that indicated increased surface areas following CuO reduction in hydrazine. Surface areas of copper wool in the sequential stages of prereduction (H2, 673 K, 8 h), oxidation (86 mg 0, per g Cu, ca. 50 nm CuO), and hydrazine reduction (298 K, 10 h) were typically 1.5,0.6, and 2.5 m2 g-l, respectively. The decrease in surface area upon oxidation may be attributed to annealing and/or obscuring of surface defects by the oxide layer.67 The increase in surface area upon reduction in hydrazine appears to result from porosity and roughening created during removal of ca.97% of the oxygen in the CuO layer. Thus the relatively high reactivity of CuO in comparison with the other first-row transition metal oxides may also be due to the increased surface area produced during the oxide-N,H, reaction, since this roughening helps to expose underlying CuO.D . M. Littrell, D. H . Bowers and B. J. Tatarchuk 3279 97 5 950 binding energy/eV 925 Fig. 4. X-Ray photoelectron spectra of the Cu 2p1/2-3,2 doublet obtained from the indicated copper species. (a) Cu, (b) Cu,O, (c) CuO. 975 92 5 binding energy/eV Fig. 5. Angle-resolved Cu 2p1/2-3,2 spectra of partially reduced CuO. (a) 8 = lo", (6) 0 = 90".3280 Hydrazine Reduction of Transition-metal Oxides Although increased surface areas following hydrogen reduction of CuO have been reported previously,68* 69 comparisons of hydrazine and hydrogen reduction treatments suggest that hydrazine reduction is more spontaneous and provides greater increases in surface area than does hydrogen.Oxidized copper wools (ca. 50nm of CuO) with surface areas of 0.6 m2 g-l were reduced in 100 kPa of hydrogen at 423 K for 4.5 h. Temperatures of 423 K were required during hydrogen reduction to provide reduction rates similar to those observed in 600 Pa of hydrazine at room temperature. Hydrogen reduction was halted after removal of ca. 95% of the added oxygen in order to correspond with typical conversion levels provided by hydrazine reduction. Surface areas of hydrogen reduced specimens (1.8 m2 g-') were near those of the initial prereduced copper specimens (i.e.1.5 m' g-I) and ca. 60% less than those produced by hydrazine reduction at 295 K. These differences in surface area may be partially attributed to relaxation effects resulting from the higher temperatures employed during hydrogen reduction, yet thermal annealing of hydrazine reduced specimens proved that the reduced oxides possess surface areas which are thermally stable at temperatures ~ 5 7 3 K. Indeed, specimens with surface areas of ca. 2.5 m2 g-l when heated in vacuo (0.06 Pa) at 423 or 573 K for 4.5 h exhibited no losses in surface area, while annealing at 673 K for 24 h reduced the surface areas to ca. 1.4 m2 g-l. These data indicate that the surface areas created by hydrazine reduction are marginally stable in vacuum, and that increases in surface area cannot be solely attributed to decrease relaxation effects at the lower reduction temperatures required for hydrazine reduction.TEM Studies In support of X.P.S. data and microgravimetric measurements, transmission electron microscopy (TEM) was used to investigate the nature of surface roughness developed during the CuO-N,H, reaction. Micrographs of hydrogen pre-reduced copper grids (plate 1) provide evidence of smooth surface contours associated with these relatively low- surface-area materials. Subsequent oxidation at 623 K and 100 kPa for 12 h (ca. 50 nm of CuO) provided the CuO filaments shown in plate 2, consistent with previous microscopic observations of oxidized copper. 1 7 9 70 These filaments, found predominantly at the curved portions of the grid, are known to be composed of CuO and generally comprise > 1 % of the total oxygen in the scales,17 depending on the oxidation condition^.^^ Subsequent exposure of the oxidized grids to hydrazine (530 Pa, 298 K, 9 h) had two effects, the first being a folding over of the filaments (plate 3) caused by reduction of the filaments to metallic copper.The second was the production of surface roughness corresponding to an increased surface area, as evident from the fissures shown in plates 3 and 4. Thus, increases in surface area may be derived from the formation of reduced oxide whiskers (i.e. metal fibres) and irregular pitting of the surface. Conclusions The reaction between the transition-metal oxides (V20,, Cr203, Mn,O,, Fe203, Co,O, NiO, CuO and ZnO) and hydrazine was qualitatively and quantitatively assessed by X-ray photoelectron spectroscopy.The oxides ' relative ease of reduction increased monotonically with the heat of formation of the oxide per oxygen molecule (i.e. - AGg8 mole oxygen), suggesting a strong correlation between bulk and surface thermodynamics. CuO merited more extensive study based on its relatively high reactivity with hydrazine. It is suggested that its relative reactivity may be explained from consideration of the thermodynamic and acid-base properties of the CuO-N,H, reaction.J. Chem. SOC., Faraday Trans. 1, Vol. 83 part I 1 Plates 1 and 2 Plate 1. Electron micrograph of a prereduced copper grid (H2, 673 K, 9 h) showing a low-surface area sample. Plate 2. Electron micrograph of an oxidized copper grid (02, 723 K, 9 h) showing whisker growth.D. M. Littrell, D. H. Bowers and B. J. Tatarchuk (Facing p . 3280)J . Chem. SOC., Faraday Trans. I , Vol. 83 part 11 Plates 3 and 4 Plate 3. Electron micrograph of a hydrazine-reduced CuO/Cu grid (530 Pa, 298 K, 9 h) showing formation of metallic whiskers. Plate 4. Electron micrograph of a hydrazine-reduced CuO/Cu grid (530 Pa, 298 K, 9 h) showing pits formed on the contours. D. M. Littrell, D. H. Bowers and B. J. TatarchukD. M. Littrell, D. H. Bowers and B. J. Tatarchuk 328 1 Adsorbed oxygen acts as an ‘adsorption promoter ’ by creating Lewis-acid (electron deficient) sites on adjacent metal atoms for increased interaction with adsorbing hydrazine. An increase in hydrazine surface coverage may result from this action as well as from an increased adsorption energy owing to strong hydrogen-oxygen bonding and water formation during the reaction. Enhanced reactivity may also result in part from morphological changes of the oxide which provide increased surface areas for reaction.Microgravimetry and TEM suggest an increase in surface area associated with reduction of CuO by hydrazine at 298 K that is more extensive than reduction by hydrogen at temperatures up to 423 K. The porosity associated with reduction in hydrazine may allow exposure of the underlying oxide, thereby driving to near completion the reduction of thick (ca. 120 nm) oxide layers at room temperature. This work was partially funded by the U.S. National Aeronautics and Space Administration (NASlO-11027) and partially by the Space Power Institute as funded by the SDIO Innovative Science and Technology Office and the Defense Nuclear Agency under DNA contract no.001-85-C-0183. The assistance of Mr Ray Cocco during TEM studies is also acknowledged. References 1 D. M. Littrell and B. J. Tatarchuk, J. Vac. Sci. Technol., Part A , 1986, 4, 1608. 2 D. M. Littrell and B. J. Tatarchuk, Hydrazine Reduction of’ Transition Metal Oxides: In Situ Characterization using X-ray Photoelectron Spectroscopy, presented to the 32nd National Symposium and Topical Conference of the American Vacuum Society, Houston Texas, November 19-22, 1985 (unpublished). 3 L. F. Audrieth and B. A. Ogg, The Chemistry of Hydrazine (Wiley, New York, 1951) and references cited within.4 A. J. Clark and W. F. Pickering, J. inorg. Nucl. Chem., 1967, 29, 836. 5 M. W. Rophael, Surf. Technol., 1982, 16, 235; Chem. Abs., 97: 134199h. 6 W. J. Ward, P. U. Labine and D. A. Redfield, Ammonia Plant S a j , 1979, 21, 57. 7 M. Grunze, Surf Sci., 1979, 81, 603. 8 M. H. Matloob and M. W. Roberts, J. Chem. Res. (3, 1977, 336. 9 V. G. Rienacker and J. Volter, Z . Anorg. Allg. Chem., 1959, 302, 292. 10 M. H. Matloob and M. W. Roberts, J . Chem. Soc., Faraday Trans. 1 , 1977, 73, 1393. 11 J. L. Falconer and H. Wise J. Card., 1976, 43, 220. 12 J. P. Contour and G. Pannetier, J. Catal., 1972, 24, 434. 13 R. C. Cosser and F. C. Tompkins, Trans. Faraday Soc., 1971, 67, 526. 14 R. C . A. Contaminard and F. C. Tompkins, Trans. Faraday SOC., 1971, 67, 545. 15 D.W. Johnson and M. W. Roberts, J . Electron Spectrosc. Relat. Phenom., 1980 19, 185. 16 H. H. Madden and D. W. Goodman, Surf. Sci., 1985, 150, 39. 17 E. A. Gulbransen, T. P. Copan and K. F. Andrew, J. Electrochem. Soc., 1961, 108, 119. 18 J. Block, 2. Phys. Chem. N.F., 1972, 82, 1. 19 C. J. Groeneboom, G. Sawatzky, H. J. De Liefde Meijer and F. Jellinek, J . Orgunomet. Chem., 1974, 20 R. Larsson, B. Folkesson and G. Schon, Chem. Scr., 1973, 3, 88. 21 V. I. Nefedov, Ya. V. Sayln’, A. A. Chertkov and L. N. Padurets, Russ. J. Inorg. Chem. (Engl. Transl.), 22 V. I. Nefedov., D. Gati, B. F. Dzhurinskii, N. P. Sergushin and Ya. V. Salyn’, Russ. J. inorg. Chem. 23 I. Ikemoto, K. Ishii, S. Kinoshita, H. Kuroda, M. A. Alario Franco and J. M. Thomas, J. Solid State 24 G.C. Allen, M. T. Curtis, A. J. Hopper and P. M. Tucker, J. Chem. Soc., Dalton Trans., 1973, 16, 25 G. C. Allen and P. M. Tucker, inorg. Chim. Acta, 1976, 16, 41. 26 J. C. Carver, G. K. Schweitzer, and T. A. Carlson, J. Chem. Phys., 1972, 57, 973. 27 A. Aoki, Jpn J. Appl. Phys., 1976, 15, 305. 28 M. Oku and K. Hirokawa, J. Electron Spectrosc. Relat. Phenom., 1976, 8, 475. 29 M. Oku, K. Hirokawa and S. Ikeda, J. Electron Spectrosc. Relat. Phenom., 1975, 7, 465. 30 H. F. Franzen, and M. X. Umana, J. Solid State Chem., 1976, 18, 363. 76, C4. 1974, 19, 785; Zh. Neorg. Khim., 1974, 19, 1443. (Engl. Transl.), 1975, 20 1279; Zh. Neorg. Khim., 1975, 20, 2307. Chem., 1976, 17, 425. 1675.3282 Hydrazine Reduction of Transition-metal Oxides 31 N. S. McIntyre, and D. G. Zetaruk, Anal.Chem., 1977, 49, 1521. 32 K. Kishi and S. Ikeda, Bull. Chem. SOC. Jpn, 1973, 46, 341. 33 G. C. Allen, M. T. Curtis, A. J. Hooper, and P. M. Tucker, J. Chem. SOC., Dalton Trans., 1974, 34 T. J. Udovic, Ph.D. Dissertation (University of Wisconsin-Madison, 1982). 35 N. S. McIntyre and M. G. Cook, Anal. Chem., 1975, 47, 2208. 36 K. S. Kim, Phys. Rev. B, 1975, 11, 2177. 37 N. S. McIntyre, T. E. Rummery, M. G. Cook and D. Owen, J. Electrochem. SOC., 1976, 123, 1164. 38 K. S. Kim and N. Winograd, Surf. Sci., 1974, 43, 625. 39 K. S. Kim, W. E. Baitinger, J. N. Amy and N. Winograd, J . Electron Spectrosc. Relat. Phenom., 1974, 40 K. S. Kim and R. E. Davis, J. Electron Spectrosc. Relat. Phenom., 197211973, 1, 251. 41 L. J. Matienzo, L. 1. Yin, S. 0. Grim and W.E. Swartz, Jr, Znorg. Chem., 1973, 12, 2762. 42 G. Schon, Surf. Sci., 1973, 35, 96. 43 S. W. Gaarenstroom and N. Winograd, J . Chem. Phys., 1977, 67, 3500. 44 J. H. Scofield, J. Electron Spectrosc. Relat. Phenom., 1976, 8, 129. 45 D. R. Penn, J. Electron Spectrosc. Relat. Phenom., 1976, 9, 29. 46 D. M. Littrell, M. S. Thesis (Auburn University, 1986). 47 A. Ronnquist and H. Fischmeister, J. Znst. Met., 196041, 89, 65. 48 R. F. Tylecote, Metallurgia, 1956, 53, 191. 49 J. Haber, J. Stoch and L. Ungier, J. Electron Spectrosc. Relat. Phenom., 1976, 9, 459. 50 C . T. Au and M. W. Roberts, Chem. Phys. Lett., 1980, 74, 472. 51 T. Robert, M. Bartel and G. Offergeld, Surf. Sci., 1972, 33, 123. 52 C. T. Au, J. Breza and M. W. Roberts, Chem. Phys. Lett., 1979, 66, 340. 53 K. Kishi and S . Ikeda, Appl. Surf. Sci., 1980, 5, 7. 54 L. Moroney, S. Rassias and M. W. Roberts, Surf. Sci., 1981, 105, L249. 55 C. T. Au, S. Singh-Boparai, and M. W. Roberts, J . Chem. SOC., Faraday Trans. I , 1983, 79, 1779. 56 K. Tanaka and K . Tamaru, J. Catal., 1963, 2, 366. 57 R. Ugo, Proc. 5th Int. Cong. Catal. (North-Holland, Amsterdam, 1973), p. 13-19. 58 P. Sabatier, Ber. Dtsch. Chem. Ges., 1911, 44, 2001. 59 W. J. M. Rootsaert and W. H. M. Sachtler, Z. Phys. Chem., 1960, 26, 16. 60 A. A. Balandin, Adv. Catal., 1958, 10, 120. 6 1 A. A. Balandin, Surface Chemical Compounds and Their Adsorption Phenomena (Moscow University 62 I. J. Eberstein and I. Glassman, Prog. Astronaut. Rocketry, 1960, 2, 351. 63 M. McD. Baker and G. I . Jenkins, Ado. Catal., 1955, 7, 1. 64 D. A. Dowden, N. Mackenzie and B. M. W. Trapnell, A h . Catal., 1957, 9, 65. 65 0. V . Krylov, Catalysis By Nonmetals (Academic Press, New York, 1970). 66 M. A. Barteau, M. Bowker and R. J. Madix, Surf. Sci., 1980, 94, 303. 67 T. N. Rhodin, J. Am. Chem. SOC., 1950, 72, 4343. 68 W. D. Bond, J. Phys. Chem., 1962, 66, 1573. 69 T. Takeuchi, 0. Takayasu and S . Tanada, J . Catal., 1978, 54, 197. 70 G. M. Raynaud and R. A. Rapp, Oxid. Met., 1984, 21, 89. 1525. 5, 351. Press, Moscow, 1957). Paper 61 1544 ; Received 28th July, 1986