J . Phys. Chem. 1984,88, 5345-5352
Adsorption and Reaction of
5345
I2 and CH,I with Uranium Monocarbide Surfaces
J. G . Dillard,+ H. Moers, H. Klewe-Nebenius, G . Kirch, G. Pfennig, and H. J. Ache* Kernforschungszentrum Karlsruhe GmbH, Institut fur Radiochemie, 0-7500 Karlsruhe 1 , Federal Republic of Germany (Received: May 1 1 , 1984)
The surface reactions of Iz and CHJ with uranium carbide (UC) at 25 OC have been investigated for exposures up to 200 langmuirs of I2 and 50 langmuirs of CH,I (1 langmuir = 10" torr-s). Adsorption behavior was characterized by Auger electron (electron and X-ray induced) and X-ray photoelectron spectroscopies to obtain quantitative and surface chemical information. The I/U elemental ratio increases with increasing exposure for each adsorbate. Following Iz exposures greater than 10 langmuirs the AES spectra in the U(0PV) (KE = 74 eV) region are identical with that for UI,, while with a Mg Kcu source (1253.6 eV) the U 4f XPS (KE = 874 eV) spectra are principally that for UC. These results indicate the formation of an outer UI, surface layer. The conversion of a UC surface to UI, implies and C carbon. The C/U atomic ratio decreases with increasing I2 exposure, and C 1s photopeaks appear at binding energies higher than the carbide C 1s peak. Above 10 langmuirs the I/U atomic ratio changes more slowly for incremental increases in I2exposure suggesting that the UI, overlayer inhibits the reaction of I2 with UC. No significant binding energy or Auger kinetic energy shifts are evident in the uranium XPS or AES spectra following the adsorption of CH,I on UC. Quantitative analysis of the C/U and I/U ratios and the thermal desorption behavior are consistent with a dissociative adsorption of CH31 on UC.
Introduction Studies of the interaction of iodine-containing molecules with the surfaces of uranium compounds are important for understanding the reactions that occur at the gas-solid interface of actinide elements. Such investigations are also of practical importance in the field of nuclear technology where knowledge of corrosion and degradation processes for reactor fuel elements is vital.' The surface reactions of uranium metal with oxygen-containing compound^^-^ and with 12*and CH319have been investigated to gain information on the stoichiometry of surface products. Detailed surface structural studies revealed5 that a large variety of complex UOz structures are produced when uranium reacts with oxygen. Ellis6 has discovered an anomalous temperature dependence for the reaction of O2 with oxycarbide layers on polycrystalline uranium. Conversion of UOz to a U O surface phase has also been reported! The adsorption of water on uranium metal produces a surface composed of an OH-surface complex with uranium that involves electron transfer from uranium to the hydroxy ~ o m p l e x .The ~ chemical nature of surface oxygen species was elucidated from X-ray photoelectron spectroscopic meas u r e m e n t ~ .It~ has also been noted that the oxygen ls photopeak for C O adsorbed on uranium is indistinguishable from the 0 1s peak at 532 eV for water adsorbed 00 uranium. The reaction of uranium metal with I, yields U13,8and the adsorption of CHJ on uranium is dissociative with the formation of UI, and U C 9 From a practical point of view, studies involving uranium carbide are important because this material is useful in hightemperature reactors.' Thus information on the surface chemical processes that occur in the reactions of iodine-containing fission products would be valuable in considering the chemical form of the fuel. Investigations have been reported where the chemical nature of volatile fission products was probed.'*13 It was found that iodine was released as a uranium iodide when a uraniummolybdenum alloy was heated to 1250 O C in helium.I0 Only elemental iodine was released in an oxidizing atmosphere.1° Similar results were reported for the release of iodine from uranium oxides"J2 where the general behavior was that elemental iodine is produced in oxidizing atmospheres whereas iodide is formed in reducing conditions. More recently13 iodide as cesium iodide rather than elemental iodine has been suggested as the dominant form that would be generated under accident conditions in a light water reactor. Presumably the release of volatile iodine products from a uranium carbide fuel would be similar to those released from metal or oxide fuels. t Permanent address: Blacksburg, VA 2406 1.
Department of Chemistry, Virginia Tech,
0022-3654/84/2088-5345$01.50/0
In this investigation we report the results of X-ray photoelectron and Auger electron spectroscopic studies of the uranium monocarbide (UC) surface following the adsorption reaction with I2 and CH31. Previous XPS studies of metal carbides have been presented by Ramqvist and co-workers.'"16 The goal of the earlier studies was to compare experimentally and theoretically the valence band structure of metal carbides and to provide information on the XPS chemical shifts in the carbides relative to pure metal and graphite carbon. Shifts to lower binding energies for carbon and to higher binding energies for the metal were an indication of electron transfer from metal to carbon. The lower density of electron states at the Fermi level in the metal carbides was correlated with the reduced electrical conductivity of carbides compared to the pure metal. The goals for the current investigation were to obtain XPS and AES spectra for clean uranium monocarbide, to characterize the chemical nature of surface products following the adsorption reaction of Iz and CHJ, and to employ XPS and AES in obtaining information about the thickness of surface layers. Experimental Section The adsorption experiments and the surface characterization measurements were carried out in an ultrahigh-vacuum X-ray photoelectron-Auger electron (XPS/AES) spectrometer (VGESCALAB-5) which has been described p r e v i o ~ s l y .The ~ ~ ~base (1) Naoumidis, A.; Nickel, H. In "Irradiated Fuel Reprocessing, Gmelin Handbook of Inorganic Chemistry", 8th ed.; Springer-Verlag: West Berlin, 1982; Main Vol. Uranium, Suppl. Vol. A4, p 139. (2) Allen, G. C.; Tucker, P. M. J. Chem. SOC.,Dalton Trans. 1973,470. ( 3 ) Fuggle, J. C.; Burr, A. F.; Watson, L. M.; Fabian, D. J.; Lang, W. J. Phys. F. 1974, 4, 335. (4) Ellis, W. P. Surf. Sci. 1976, 61, 37. ( 5 ) Ellis, W. P., Taylor, T. N. Surf.Sci. 1980, 91, 409. (6) Ellis, W. P. Surf. Sci. 1981, 109, L567. (7) Nornes, S.B.; Meisenheimer, R. G. Surf. Sci. 1979, 88, 191. (8) Dillard, J. G.; Moers, H.; Klewe-Nebenius, H.; Kirch, G.; Pfennig, G.; Ache, H. J. J. Phys. Chem. 1984,88, 4104. (9) Dillard, J. G.; Moers, H.; Klewe-Nebenius, H.; Kirch, G.; Pfennig, G.; Ache, H. ,I. accepted , for publication in Surf. Sci. (10) Castleman, A. W.; Tang, I. N. J. Inorg. Nucl. Chem. 1970, 32, 1057. (11) Castleman, A. W.; Tang, I. N.; Munkelwitz, H. R. J. Inorg. Nucl. Chem. 1968, 30, 5 . (12) Nakashima, M.; Tachikawa, E. J. Nucl. Sci. Techno/. 1978,15, 849. (13) Campbell, D. 0.;Malinauskas, A. P.; Stratton, W. R. Nucl. Technol. 1981, 53, 111. (14) Ramqvist, L. J. Appl. Phys. 1971, 42, 2113. (15) Ramqvist, L.; Hamrin, K.; Johansson, G.; Gelius, U.; Nordling, C. J . Phys. Chem. Solids 1970, 31, 2669. (16) Ramqvist, L.; Hamrin, K.; Johansson, G.; Fahlman, A.; Nordling, C. J. Phys. Chem. Solids 1969, 30, 1835.
0 1984 Americap Chemical Society
5346
&-
The Journal of Physical Chemistry, Vol. 88, No. 22, I984
pressure in the system is 3 X 10-l' torr. Photoionization was induced by Mg K a (1253.6 eV) and A1 K a (1486.6 eV) X-radiation. The peak-width-at-half maximum (pwhm) for the Au 4f7,z level of clean gold was 1.2 eV for an analyzer pass energy of 20 eV. XPS atomic ratios were calculated from integrated photopeak intensities (including satellite structure for the uranium 4f levels) which had been corrected for the photoionization cross section17and an empirically determined instrumental sensitivity f a c t ~ r XPS . ~ ~ binding ~ energies are referenced to the Fermi level which we place at the midpoint on the low-energy side of the U 5f photopeak. This assignment procedure is equivalent to that used to establish the Fermi level in uranium metal and uranium corn pound^.^^^^^^^'^ Electron-induced Auger spectra were determined with 5-kV electrons. Auger peak intensities were measured from the first derivative spectra by using peak-to-peak heights, and were used directly for the calculation of AES intensity ratios without correction for different cross sections. The Auger kinetic energies are also referred to the Fermi level by establishing the absolute kinetic energy for the X-ray induced Auger transition and correcting the electron-induced results accordingly. We observed no degradation effects in the spectral parameters or peak intensities resulting from electron or X-ray bombardment of the sample. The precision of the intensity measurements is f 5%. Uranium monocarbide (UC) was prepared by Dr. C. Politis of KM via argon arc melting of uranium and graphite. The sample was characterized by X-ray diffraction and was found to be at least 99% pure UC. The crystal was mounted in a copper holder with stainless steel screws. The U C surface was cleaned to remove traces of oxygen by argon ion bombardment (5 kV) while heating at 600 OC. The sample was then annealed at 600 OC. (Temperatures quoted are nominal values. The actual temperatures are likely to be about 15% lower due to a gradient between the sample and the thermocouple position.) Following the cleaning procedure the concentration of oxygen-containing surface impurities was o0
5
50
C H , I exposure [ L o n g m u i r = 1 d 6 t o r r . s e c ]
Figure 11. I/U intensity ratio (AES) and I/U atomic ratio (XPS) vs.
CH31exposure.
._
380
400
390
5.0
L.0
I
"
I
620
630
binding energy [eV]
Figure 10. XPS spectra: C Is, U 4f, and I 3d following a 50-langmuir CH31exposure at 25 OC.
be likely surface products. This result implies a dissociative process for CH31 adsorption on UC. The XPS and AES energetic results are summarized in Table I1 as single values throughout the adsorption range to indicate the constancy and invariance of the results. In order to suggest the chemical nature of adsorbed iodine, we compare the energetic results at similar I/U ratios for the two adsorbates, since the effects of relaxation and bonding on the binding and kinetic energies would be expected to be similar. The iodine XPS binding energy and the AES kinetic energy results are equal to the values measured for I, adsorption on U C for exposures of 0. l to 0.49 langmuir of I,. From the near equality of the binding energy and kinetic energy results at similar I / U ratios for I, and CH31adsorbed, we conclude that the chemical nature of iodine on U C following CH31exposure is similar to that following I, exposure to UC. This similarity suggests that a UI, surface is formed for CH31adsorption on U C and that CH31 adsorption is dissociative. Several products are possible in the reaction of CH31with UC. The formation of additional carbide requires that hydrogen and iodine from CH31 combine with uranium carbide or are lost to the vacuum as gaseous H2, HI, or CI,. The formation of UI, could lead to the production of gaseous methane, hydrogen, or UH3. Potential reactions are summarized as follows (unbalanced):
Unfortunately, we have no reliable method for determining the carbon surface concentration which can be attributed to adsorption. Thus we can only suggest that uranium iodide and uranium carbide form with some undefined stoichiometry. The I/U ratios exhibit a linear behavior up to an experimental fractional coverage 6 = 0.8 (0 = 1 at saturation) which corresponds
2.0
0
0
"
"
"
"
"
"
'
1.0 2.0 2.6 C H ~ Iexposure [~angmuir=~O~torr.sec]
Figure 12. First-order kinetic plot CH31/UC; -In (1 - 8 ) vs. CH31
exposure. to an XPS atomic ratio of 0.19 and an AES ratio of 0.14 after an exposure to 0.7 langmuir of CHJ. The adsorption results can be fit to a simple first-order model as shown in Figure 12 where the I / U atomic ratios (XPS) were used to calculate the values. An interpretation of the fit to the first-order model is that the adsorption is governed by the number of unoccupied surface sites. A process could be envisioned for CH31adsorption where either the iodide or the methyl group end of CHJ becomes bound to U C initially and then dissociates to yield either a methyl radical or an iodine atom. The subsequent reaction of these odd electron species could occur at another U C surface site. In this process it is reasoned that the initial adsorption reaction is the rate-controlling step. Unfortunately, we have no experimental results to support initial bonding of CH31to the UC surface via C H 3 or I attachment. Summary and Conclusions The adsorption of I, on UC occurs via a dissociative/oxidation process. Evidence from AES and XPS spectra and thermal treatment experiments indicates the formation of U13 and CI, species. We suggest that the CI, species are X I , and -C13. A net decrease in the absolute carbon surface concentration is interpreted to signify the loss of carbon via the formation of a volatile iodocarbon species, perhaps CI,. We are not able to confirm experimentally the loss of CI, using the present experimental apparatus. We imagine that CI, formation could occur via the reaction of 1, with surface X I 2 or -C13 species. The adsorption of CH31on U C leads to the formation of additional uranium carbide and a uranium iodide surface. The formation of carbide and iodide surface products implies that additional surface and/or gaseous species must be formed to account for the stoichiometry of CHJ and the surface products.
5352
J. Phys. Chem. 1984,88, 5352-5359
Among the possible products are surface uranium hydride and gaseous hydrogen, methane, iodomethane, or hydrogen iodide.
Acknowledgment. W e are grateful to Dr. C. Politis, Institut fur Angewandte Kernphysik I, Kernforschungszentrum, Karlsruhe,
for the preparation of the UC sample. J.G.D. expresses gratitude to the Department of Chemistry and to Virginia Tech for granting an educational leave. Registry No. 12, 7553-56-2; CH,I, 74-88-4; UC, 12070-09-6.
Ion Association. Comparlson of Spectroscopic and Conductance Values of Association Constants William R. Gilkerson* and Katherine L. Kendrick Department of Chemistry, University of South Carolina, Columbia, South Carolina 29208 (Received: August 2, 1983; In Final Form: May 14, 1984)
Molar conductances of potassium, rubidium, and cesium picrate and of cesium tetraphenylborate have been determined in 2-butanone at 25 O C over a concentration range of 0.1-1.6 mmol L-I. These A,Cdata together with those for lithium and sodium already determined in this laboratory were fitted by using the Lee-Wheaton (LW), the Pitts (P), the Fuoss 78 (F), and the Justice (J) conductance equations to obtain values of ion association constants, K A , and limiting molar conductances. Spectrophotometric absorbance changes were measured for sodium and cesium picrate and remeasured for lithium and tetra-n-butylammoniumpicrate at 380 nm and 25 OC over a concentrationrange of 0.02-1.4 mmol L-l. Values of the cesium-133 NMR chemical shifts of cesium picrate were determined over a concentration range of 0.1-3 mmol L-' while those for cesium tetraphenylborate were determined over a range from 0.25 to 10 mmol L-l. We calculated values of ion association constants for the various salts using these spectroscopic data. The values of K A determined spectroscopically and the values determined, conductimetricallyare the same within experimental error with one exception; for cesium tetraphenylborate, the NMR value depends on the concentration range of data included in the analysis.
Introduction A variety of spectroscopic techniques has been applied to the study of ionsolvent and ion-ion interactions in electrolyte solutions in addition to the application of more classical physicochemical methods.' Spectroscopic studies of ion-ion interaction have usually focused on structural aspects, but the increasing sensitivity of modern instruments and their availability (FT N M R spectrometers, for example) has made possible the application of these techniques to concentrations as small as a millimolar or less. The association of oppositely charged univalent ions in solution to form ion pairs M + + X- t; M+**.X-t-i M+,X(1) where M+.-X- represents a solvent-separated ion pair and M+,Xrepresents the contact ion pair, becomes important in water when the electrolyte reaches the 0.01-0.1 mol L-' range and at even lower concentrations in less polar solvents. Equilibrium constants for ion association, K A , have usually been determined by conductance methods, but there have been a number of reports of spectroscopic determination^.^-^ In several instances there have been conductance data available so that comparison of the results of the different physical techniques was made; there was agreement between spectroscopic values and conductance values of the association constant in some case^^*^*' and disagreement in other^.^^^^* There has been some discussion in the literature (see Covington (1) See, for example: Faraday Discuss. Chem. SOC.1977, No. 64. (2) Davies, W. G.; Otter, R. J.; Prue, J. E. Discuss. Faraday SOC.1957, 24, 103. (3) Matheson, R. A. J . Phys. Chem. 1965,69, 1537. (4) Dadgar, A,; Khorsandl, D.; Atkinson, G. J . Phys. Chem. 1982, 86, 3829. ( 5 ) Covington, A. K.;Freeman, J. G.; Lilley, T. H. J. Phys. Chem. 1970, 74, 3173. ( 6 ) Janz, G. J.; Miller, M. A. J . Solution Chem. 1975, 4, 285. (7) Yeager, H. L.; Reid, H. J . Phys. Chem. 1976, 80, 850. (8) (a) DeWitte, W. J.; Liu, L.; Mei, E.; Dye, J. L.; Popov, A. I. J . Solution Chem. 1977, 6, 337. (b) Khazaeli, S.;Popov, A. I.; Dye, J. L. J . Phys. Chem. 1982, 86,4238. (9) Hinton, J. F.; Metz, K. R. J. Solution Chem. 1980, 9, 197.
0022-3654/84/2088-5352$01.50/0
et aL5) concerning the significance of any such differences; what has appeared has been in the nature of suggestion or the expression of opinion. The spectroscopic differences between free ions and ion pairs are due to perturbations of energy levels in either the cation, anion, or both, because of interaction between the ions, presumably at short range. Those ions counted as paired (nonconducting) by the conductance technique may include solvent-separated ion pairs, M+--X-, in addition to those contact ion pairs, M+,X- detected spectroscopically. If there are real differences in values of K A obtained by using the different physical techniques, then these differences could possibly be exploited to yield further information about the pair formation process, including details of the role the solvent molecules must play in the final step, formation of the contact pair. If the differences in values of K A that have been determined by the several different techniques are not real but are traceable to some experimental error, then establishment of that fact would mean that one element of controversy could be removed from the use of spectroscopic techniques to probe the state of electrolyte solutions. A program is under way in this laboratory to compare the values of ion association constants as found by the conductance method with those determined by using spectroscopic techniques under conditions such that both kinds of measurements are carried out over the same range of electrolyte concentration and for which the same expressions for ion activity coefficients are used in treating both kinds of data. This is a report of UV-visible spectrophotometric studies of lithium, sodium, and cesium picrates, cesium-133 N M R chemical shift studies of cesium picrate and cesium tetraphenylborate, and conductance studies of potassium, rubidium, and cesium picrate and of cesium tetraphenylborate, all in 2-butanone solvent a t 25 OC. The inclusion of these conductance measurements together with data from a recent study of lithium and sodium picratelo furnishes data on a complete series of alkali-metal picrates. We have analyzed these data using four (10) Jackson, M. D.; Gilkerson, W. R. J . Am. Chem. SOC.1979,101, 328.
0 1984 American Chemical Society
