J. Phys. Chem. 1992,96,2325-2328 Whalley et discuss the thermodynamics of the high-density amorphous ice which forms when ice is compressed to 10 kbar at 77 K. From the enthalpy and volume changes at the highpressure ice-amorph transitions, they show that in order for the high-density amorph to be a metastable continuation of normal water its excess entropy must be about 2 J K-I mol-I. In that respect their conclusion is the same as that reached here, and because it relies on different experimental evidence, the quantitative basis of the argument is independently supported. However, Whalley et a1.26invert the logic of the present paper. They reject the Johari3 and Sceats-Rice19estimates of the excess entropy of amorphous ice on the grounds those estimates do not accord with their premise that the high-density amorph is a metastable continuation of normal water. The high-density amorphous ice transforms into a low-density form, reversibly, by a first-order phase transition when it is dec o m p r e ~ s e d .Whalley ~~ et aLz6estimate that the excess entropy of the low-density amorph is about 1 J K-’ mol-’. That estimate also derives from their premise of continuity between normal water and the high-density amorph. Johari et al.27 report a calorimetric study of a low-density amorph which was formed by compressing cubic ice and then decompressing. They report that it has a glass transition at 130 K and that its crystallisation kinetics are also different from those of water 11. Those observations provide the warning that amorphous water might not be a thermodynamically reproducible metastable state, even at temperatures near 150 K where it freezes.
Discussion Bernal and Fowler’8 and Sceats and have proposed that there exists a metastable, amorphous-solid, random-tetrahedral network of hydrogen-bonded water molecules, which can serve as a model for the underlying structure and properties of liquid water. Studies of vapor-deposited and splat-quenched water have been directed at characterizing that material. Amorphous samples which are prepared by very different routes (vapor deposition and splat quenching) evidently relax into the same metastable state when they are annealed at 130 K. That (26) Whalley, E.; Klug, D. D.; Handa, Y. P. Nature 1989, 342, 782. (27) Johari, G. P.; Hallbrucker,A.; Mayer, E. J . Phys. Chem. 1990, 94, 1212.
2325
metastable material, which I call water 11, has an apparent glass transition at 136 K and it behaves reversibly when the temperature is cycled up and down below 150 K. It is natural to identify water I1 with the above random network models. The paradox deduced above is that if water I1 at 150 K is an amorphous material, then it should have an excess entropy AS( T I ) of 6-9 J K-’ mol-’, in which case it cannot be connected by a continuous reversible path to normal water. Hence, if water I1 is a thermodynamically metastable substance, then it belongs to a distinct phase, disconnected from normal water. It is an irony that the Sceats and Rice model,19920which inherently assumes the continuity of states between glassy and normal water, yields an entropy19 which rules out that continuity. If water I1 is a thermodynamically metastable substance and if it has the excess entropy AS(T,)of 6-9 J K-’ mol-I which is expected3J9for random network material, then thermodynamics requires (as shown by case d in Figure 2) that it becomes more stable than ice not far above 150 K and that, so long as it exists, it remains more stable than both ice and normal water up to 273 K. The oceans and ice fields on earth would then be unstable and, on thermodynamic grounds, should transform to water 11, with catastrophic consequences. That possibility can be rejected on the grounds that if it was possible it would have happened before now. The only way to retain the notion that water I1 is a metastable phase with AS( TI)of 6-9 J K-’ mol-’ and AH( TI) = 1380 J mol-’, while avoiding the above catastrophe, is to propose that water I1 has an upper temperature limit of stability such that it becomes absolutely unstable and ceases to exist. It will freeze before that limit is reached. The stability limit would need to be below the temperature at which AG for water I1 tends to zero. That proposal is consistent with the fact that no metastable amorphous form of water has been observed in the “stability gap” between 170 and 227 K. Supercooled water seems to have a lower temperature limit of stability at Ts= 227 K28*29and it always freezes above 227 K when cooled a t rates which allow it to maintain internal equilibrium. Water I1 always freezes below 170 K when it is warmed. Registry No. Water, 7732-18-5. (28) Speedy, R. J.; Angell, C. A. J . Chem. Phys. 1976, 65, 851. (29) Speedy, R. J. J . Phys. Chem. 1982,86, 982.
Thermochemistry of the Gaseous Osmium Oxides D. L. Hildenbrand* and K. H. Lau SRI International, Menlo Park, California 94025 (Received: October 16, 1991) The gaseous osmium oxide species OsO,, Os03,OsO2,and Os0 were generated by reaction of 02(g)with Os(c) in a heated alumina effusion cell, and thermochemical data were derived from equilibrium measurements made by mass spectrometry. Although OsO, and Os03were observed as low as 1000 K, OsO2and Os0 were detected only at temperatures above 1800 K. From third-law analysis of the gaseous reaction equilibria OsO, = Os03 + 1/202, Os03= 0s02+ 0, and OsO2= Os0 + 0, the bond dissociation energies D(030s-O), D(O2Os-0),D ( 0 0 s - 0 ) , and D ( 0 s - 0 ) were found to be 4 3 5 , 570, 542, and 575 kJ mol-I, respectively, at 298 K. There are no previous results on 0s02and OsO. Results are compared with other information in the literature.
Introduction Recent observation of the removal of metallic osmium films by energetic 0 atoms in low earth orbit’ has stimulated interest in the thermochemical properties of the gaseous osmium oxides. Following earlier mass spectrometric studies by Grimley et al.2 of the reaction equilibrium ( 1 ) Peters, P. N.; Gregory, J. C.; Swann, J. T. Appl. Opr. 1986, 25, 1290. (2) Grimley, R. T.; Burns, R. P.; Inghram, M. G. J . Chem. Phys. 1960,
33, 308.
OsO,(g) = Os03(g) + 1/202(g) (1) Watson et aL3 reinvestigated reaction 1 by the same method and obtained the enthalpy change AH02s8(l)= 176 f 29 kJ mol-’, some 127 kJ mol-I larger than the previous value.2 Both studies2s3 reported evidence for the presence of neutral OsO2but intensities were too low for thermochemical measurements. Watson et aL3 (3) Watson, L. R.; Thiem, T.; Dressler, R. A,; Salter, R. H.; Murad, E. J . Phys. Chem. 1991, 95, 8944.
0022-365419212096-2325%03.00/0 0 1992 American Chemical Society
Hildenbrand and Lau
2326 The Journal of Physical Chemistry, Vol. 96, No. 5, 1992 reported an upper limit of 12.2 f 0.4eV for the ionization potential of Os02.Pedley and Marshall4 estimated the dissociation energy of Os0 as 594 f 84 kJ mol-I by analogy with related oxides, but there have been no direct thermochemical studies of Os0 or OsO2. We have now completed a brief thermochemical study of the Os-0 system, also by mass spectrometry and at somewhat higher temperatures than those attained previo~sly.~,~ Both OsO2and Os0 were clearly identified and characterized thermochemically, and the results are reported here.
Experimental Section Measurements were made with the magnetic-sector mass spectrometer system and heated effusion-beam source described in previous publications.sq6 A powdered sample of Os(c) was contained in an alumina effusion cell with alumina gas inlet tube at the base. The cell contained a 0.15 cm diameter effusion orifice, with a length of 0.35 cm. A thin platinum cover was placed over the alumina cell to provide good blackbody conditions for pyrometer temperature measurement; apertures in the top and side of the Pt cover were aligned with the cell orifice and blackbody cavity. A few experiments were done with a mixture of Os(c) and Ti02(c) in a conventional alumina effusion cell. With the gas inlet cell, the flow of O2was controlled reproducibly with an external molecular leak valve. At most temperatures, gas flow rates were varied to check on attainment of chemical equilibrium. All ion signals were checked for response to displacement of the neutral beam-defining slit between effusion cell and ion source to clarify their origin; except for temperatures above 1500 K, where some background OsO, was observed because of high O2flow rates, the O s 4 species clearly originated in the effusion cell. Ionization efficiency curves were recorded automatically with an x-y recorder and the threshold behavior was interpreted by the vanishing current methoda6 After ion species were identified from their masses and isotopic spectra, and reliable threshold appearance potentials (APs) were determined, neutral partial pressures and equilibrium constants, KO, were evaluated from parent ion intensities determined a few electronvolts (3 eV in this work) above their respective thresholds to avoid overlapping fragmentation contributions. Partial pressures were derived from the ion intensities using a sensitivity constant determined by calibration with a laboratory vapor pressure standard and the application of a cross-section c ~ r r e c t i o n . ~In the O s 4 studies, the ionization efficiency curves and mass spectra indicated that no fragmentation corrections were required to obtain an accurate total ion yield for third-law calculations. This procedure has been found to be quite reliable, and the resulting K O values are estimated to be accurate within a factor of 2.7v8 All other aspects of the experimental p r d u r e and data interpretation are as described p r e v i o u ~ l y . ~ ~ ~ The osmium powder sample of 99.9% purity was obtained from Johnson Matthey, while the 02(g) and Ti02(c) were reagent grade materials obtained from commercial suppliers. Results With O2admitted to the cell containing the Os(c) sample at 1000-1500 K, Os04+and Os03+were observed with threshold APs of 12.3 and 11.3 eV, both f0.3eV. A somewhat weaker 0sO2+signal with an AP of 15.5 eV was also observed in this range and is clearly a fragment ion from Os03. These values agree well with those reported by Watson et al.3 and indicate that Os04+ and Os03+are indeed parent ions under the experimental conditions. Other AP data in the literature for OsO, and Os03 are discussed by Watson et ale3 Above about 1850 K, OsO2+and OsO+ signals with threshold APs of 10.2and 9.5 eV were observed. These APs are in line with (4) Pedley,J. B.;Marshall, E. M. J . Phys. Chem. Ref. Data 1983.12.967. (5) Hildenbrand, D. L. J. Chem. Phys. 1968, 48, 3657; 1970.52, 5751. (6) Hildenbrand, D. L. Inl. J . Mass Spectrom. Ion Phys. 1970, 4, 75; 1971,
7, 255. (7) Lau, K. H.; Hildenbrand, D. L. J . Chem. Phys. 1987, 86, 2949. (8) Hildenbrand, D. L.; Lau, K. H.; Russell, T. D.; Zubler, E. G.; Struck, C . W. J . Electrochem. SOC.1990, 137, 3215.
TABLE I: Equilibrium Constants and Derived Third-Law Enthalpies for Reaction 1
T,K
KO(l),at"/* 3.3 x 10-5 3.8 x 10-5 1.6 x 10-3 2.0 x 10-3 2.0 x 10-3
1013 1013 1288
1288 1288
kJ mol-' 182.3 181.1 189.4 187.1 187.1 av 185.4
TABLE II: Equilibrium Constants and Derived Third-Law Enthalpies for Reactions 2 and 3"
1897 1965 1965 2014 2014 2014 2052 2052 2052 2125
3.0 X 1.9 X 1.4 X 3.9 X 3.8 X 2.6 X 4.8 X 5.4 X lo-* 5.0 X 10" 2.8 X lo-'
2006b 2059b
3.2 X 4.6 X
574.2 564.8 569.8 566.8 567.3 573.6 574.2 572.2 573.5 563.6 av 570.0 568.1 576.9
1.6 X 1.6 X 1.2 X 1.9 X 2.7 X 1.9 X 4.1 X
10" 10" 10" 10" 10" 10" 10"
531.3 531.3 542.1 544.6 538.6 544.6 550.3 av 542.1
From Os(c) + O,(g) system except where noted. bFrom Os(c) Ti02(c) system.
+
anticipated values for the ionization potentials (IPS) of the corresponding neutrals, which would be expected to show increasing Os orbital character as the oxygen ligands are removed successively and to converge on the IP of atomic Os at 8.7 eV. Therefore, the AP data provide unambiguous evidence for the presence of OsO, and Os0 in the effusing vapor. Our AP values for Os02+ and OsO+ are in fair agreement with OsO, and Os0 IPSof 11.2 and 9.7 eV, respectively, estimated by Dillard and Kiser9 from electron impact data on OsO,; they also agree well with measured IP values for R u 0 2 and R u O . ~Not surprisingly, our AP(Os02+) is about 2 eV lower than the upper limit IP(Os02) reported by Watson et al.,3 indicating that OsO2was not present in detectable and amounts in the earlier work. The additional parent ions 02+ O+ were observed in the effusion beam, with APs of 12.0and 13.5 eV, in accord with the known IPS of O2and 0. Parent ion intensities measured at 3 eV above the respective thresholds were used together with the pressure calibration constant and estimated cross-section ratios to evaluate the equilibrium constants K O for reaction 1 and for the additional reactions Os03(g) = Os02(g) + O(g)
(2)
OsO,(g) = OsO(g) + O(g) (3) with the results shown in Tables I and 11; also shown are third-law enthalpies calculated with thermal functions described in the Appendix. K O values were found to be independent of gas flow rate and species partial pressure under these conditions, providing evidence that chemical equilibrium was achieved. For reaction 1, our values of P(1) are in close agreement with S) the selected results obtained by Watson et aL3with V ~ O ~ (as oxidizer; their average third-law enthalpy yields K O ( 1) = 1.9 X (atm)'I2 at 1288 K, the same as our average value from Table I. It is significant that our results are in closer agreement with the data obtained with the V205oxidizer (which they considered to be equilibrium data) than with the 02.3In any event, the comparison shows that reproducible results were obtained. Rather than doing a more complete study of reaction 1, our intent here was only to seek corroboration of the work of Watson et aL3 Our equilibrium data yield the average third-law enthalpy change h f 1 0 2 9 8 ( l ) = 185 f 12 W mol-I, in agreement within the estimated errors with the selected value of Watson et al.? 176 f 29 kJ mol-', (9) Dillard, J. G.; Kiser,
R. W. J . Phys. Chem. 1965, 69, 3893.
The Journal of Physical Chemistry, Vol. 96, No. 5, 1992 2327
Thermochemistry of the Gaseous Osmium Oxides TABLE III: Derived Thermochemical Data gaseous species AIH02QekJ mol-' bond oso, -336.4 i 8 (03os-O) (0,Os-O) -151.0 i 21 eo, 169.4 & 21 (OOS-O) os02 os0 462.3 i 21 (0S-O)
Da2,,a,kJ moP 435 570 542 575
which is an average of second- and third-law results. It is not clear why the earlier second-law results of Grimley et aL2 gave the much lower value M0298(l) = 54 kJ mol-', but it clearly should be discarded in favor of the newer results. It is worth noting that at temperatures above 1500 K, Os04 persisted at noticeable background levels in the ion source, despite fast cryopumping there, and the parent ion did not show the normal sharp slit profile characteristic of effusion cell species, as was seen at lower temperatures. This is no doubt due mostly to the relatively high O2pressures in the cell (approaching lo4 atm) used to enhance the small OsO2+and OsO+ signals. All other signal profiles showed normal behavior. As seen in Table I, the equilibrium measurements on Os04 were made at 1013 and 1288
K. Equilibrium data and derived results for reactions 2 and 3 are shown in Table 11. In order to obtain workable parent ion intensities of Os02+and G O +at AP +3 eV ionizing energy, it was necessary to push O2pressures in the cell to near IO4 atm. At 2052 K, the ratio of parent ion intensities Os03+/Os02+/ OsO+/O+was 281/2.8/1/291.Several data points for reaction 2 obtained with the Os(c) TiO,(c) beam source are in good agreement with those from the Os(c) 02(g) source as seen in Table 11, providing further evidence that the results are reproducible equilibrium values. The selected enthalpies are the average third-law values AW298(2) = 570 f 17 kJ mol-', and AW298(3) = 542 f 17 kJ mol-'. Estimated uncertainties arise from a possible error of a factor of 2 in KO,and 8 J K-' mol-' in the Gibbs energy functions. Because of the low OsO2+and OsO+ signal levels and the limited temperature range, second-law analysis is considered to be unsuitable. From the derived enthalpy changes for reactions 1-3 and the enthalpy of atomization of Os04(g) evaluated from the established enthalpy of formation (see Appendix) and the thermochemical properties of atomic Os and 0, we derive the bond dissociation energies at 298 K, D(0,Os-O) = 435 kJ mol-', D(0,Os-O) = 570 kJ mol-', D(OOs-0) = 542 kJ mol-', and D(Os-0) = 575 kJ mol-', all f20 kJ mol-'. There are no previous literature values on D(020s-0) or D(OOs-0) for comparison. Our value for Os0 leads to the dissociation energy Doo(OsO)= 571 f 20 kJ mol-' (5.92f 0.22eV), which is in accord with the qualitative estimate of Pedley and Marshall4 of 594 f 84 kJ mol-', based solely on the values for related transition metal oxides. As far as is known, there are no experimental thermochemical data for OsO, so that the Doo value reported here is the first direct measurement. The bond dissociation energies and standard enthalpies of formation, AJZ0298, of the Os-0 species derived from this work are summarized in Table 111.
+
+
Discussion It is gratifying that our measured equilibrium constants for reaction 1 at 1013 and 1288 K agree closely with those of Watson et aL3 determined with the Os(c) V205(s)system. Although the invariance of KO( 1) as P(Os04) changed by more than a factor of 10 indicates attainment of chemical equilibrium in our Os(c) + 02(g) measurements, Watson et al.3 inferred from the lack of agreement between second- and third-law enthalpies that equilibrium was not reached in their experiments with 02(g). Osmium and ruthenium are unique in the formation of stable gaseous tetraoxides. The new bond dissociation energy data, listed in Table IV, show that the first three bonds formed as oxygen ligands are added to the central Os atom are of approximately equal strength, varying from only 542 to 575 kJ mol-'. The fourth bond, D ( 0 3 0 s - O ) , is considerably weaker at 435 kJ mol-', indicating that some type of promotion energy must be required in making the transition from the +6 to the +8 oxidation state
+
TABLE IV: Molecular Constpats of O s 4 Species molecule 103~1,g cm2 u R" w;, cm-I 090, 20.7 12 1 974,340 (21,975 ( 3 ) , 3 3 5 j 20.7 20.7 oso, 24.4 6 1 564, 347, 1040 (2), 320 (2) 12.2 12.2 2 3 900, 300 (2), 800 oso, 16.2 7.0 1 10 885 os0 TABLE V: Cibk Energy FUIICHOM of O s 4 Species -(G - H,on)/T,J K-'mo1-I
T,K
os0
oso,
oso,
oso,
lo00 1200 1400 1600 1800 2000 2200
272.2 276.9 281.1 284.9 288.4 291.5 294.5
287.4 294.9 301.8 308.0 313.7 318.9 323.8
321.1 33 1 .O 340.0 348.2 355.6 362.6 369.0
337.3 349.8 361.3 371.7 38 1.3 390.2 398.4
of Os. Ruthenium shows the same type of behavior in that D(03Ru-O) is about 370 kJ mol-', while D(02Ru-0), D(ORu-0), and D(Ru-0) are 448, 490, and 530 kJ mol-', respe~tively.~.'~ The relative weakness of this fourth bond correlates with the chemical instability of the Ru and Os tetraoxides, and with the effectiveness of OsO,(g) as an oxidizing agent. As noted by Dillard and Kiser9 and also in our own laboratory, Ru04 and to some extent OsO, leave a troublesome electrically-conducting film on ion sou~cecomponents, probably from reaction with metal surfaces to form a conducting deposit of Ru or Os lower oxides. It is worth noting that among the platinum group metals, which include Ru, Rh, Pd, Os, Ir, and Pt and involve the filling of the second half of the 4d and 5d shells, the strongest metal-oxygen bonds are formed by Os. The dissociation energies of the gaseous monoxides, which are a good indication, fall in the order 571,523, 410, 402, 385, and 376 kJ mol-' for the monoxides of Os, Ru, Ir, Rh, Pt, and Pd, respectively; values for oxides other than Os0 were taken from Pedley and M a r ~ h a l l . ~ Acknowledgment. This work was supported by the Phillips Laboratory, Hanscom Air Force Base, MA. We are indebted to Lyn Watson for calling OUT attention to the paper on the electronic band spectrum of OsO.
Appendix The thermochemical properties of 0 and O2were taken from the JANAF Thermochemical Tables," while those of atomic Os were taken from Hultgren et al.I2 For calculation of the thermal functions of Os03 and 0s02,the 0s-O internuclear distances were estimated to be 0.175 nm, slightly longer than the electron diffraction value of 0.171 16 nm found in Os04.13J4Other parameters for the Os-O species were selected as described below; the selected molecular constants are summarized in Table IV in terms of the (10) Norman, J. H.; Staley, H. G.; Bell, W. E. In Mass Spectrometry in Inorganic Chemistry; Margrave, J. L., Ed.;Advances in Chemistry 72; American Chemical Society: Washington, DC, 1968; p 101. (11) Chase, M. W., Jr.; Davies. C. A.; Downey, J. R., Jr.; Frurip, D. J.; McDonald, R. A.; Syverud, A. N . J . Phys. Chem. Ref. Data 1985,14 (Suppl. No. 1). (12) Hultgren, R.; Desai, P. D.; Hawkins, D. T.; Gleiser, M.; Kelley, K. K.; Wagman, D. D. Selected Values of the Thermodynamic Properties of the Elements; American Society of Metals: Cleveland, OH, 1973. (13) Seip, H. M.; Stolevik, R. Acta Chem. S c a d . 1966, 20, 385. (14) McDowell, R. S.;Radziemski, L. J.; Flicker, H.; Galbraith, H. W.; Kennedy, R. C.; Nereson, N . G.; Krohn, B. J.; Aldridge, J. P.; King, J. D. J . Chem. Phys. 1978,69, 1513. (1 5 ) Kubaschewski,0.; A l w k , C. B. Metallurgical Thermochemistry, 5th ed.; Pergamon: New York, 1979. (16) Balfour, W. J.; Ram, R. S.J . Mol. Spectrosc. 1984, 105, 360. (17) Krauss, M.; Stevens, W. J. J . Chem. Phys. 1985, 82, 5584.
J. Phys. Chem. 1992, 96, 2328-2334
2328
moments of inertia (I),rotational symmetry number (a), electronic ground-state statistical weight (go), and vibrational frequencies (w). Gibbs energy functions calculated with the selected molecular constants are listed in Table V; the functions are calculated for the ideal gas state at 1 atm pressure. OsO,(g). Spectroscopic constants were taken from the values summarized by McDowell et ale1, The standard enthalpy of formation listed by Kubaschewski and AlcockI5 was adopted. OsO,(g). A planar structure of D3hsymmetry was assumed, along with a 'I:electronic ground state. Following Watson et aL3
the vibration frequencies were assumed to be the same as those of W 0 3 listed in the JANAF tables." OsOz(g). The molecule was assumed to be linear symmetric, with frequencies similar to those given in the JANAF tables" for MOO, and WO,. Because of the four nonbonding 5d valence electrons, the ground electronic state was assumed to be 3Z. OsO(g). The adopted rotational constant and vibration frequency were derived from analysis of the electronic spectrum.I6 The electronic ground state was taken to be 'A, as with RuO and ~ ~ 0 . 1 7
Dynamics of Electron Attachment to AOT/H,O Reversed Micelles G. Bakale, Radiology Department, Case Western Reserve University, Cleveland, Ohio 441 06
G . Beck,+ Bereich Strahlenchemie, Hahn-Meitner Institut, W-1000 Berlin 39, Germany
and J. K. Thomas* Chemistry Department, Notre Dame University, Notre Dame, Indiana 46556 (Received: March 1 , 1991)
Values of the observed rate constant, kob, of excess electrons attaching to aerosol OT (AOT) reversed micelles encapsulating varying quantities of water were measured in isooctane and tetramethylsilane (TMS) at 21 O C using a picosecond-pulseconductivity technique. Dynamic light scattering was used to determine the micellar radii in TMS from which aggregation numbers of the micelles were determined. In the lower-mobility isooctane (we = 5.3 cm2/(V s)), values of kobsapproach the diffusion-controlled attachment rate, kd, at a molar H,O/AOT ratio of 18 and appear to exceed kd at larger values of H,O/AOT. In contrast, values of kobsin TMS (we = 100 cm2/(V s)) are &fold less than kd at the maximum ratio of H,O/AOT = 27 that was studied. Several factors that could contribute to kobsappearing to exceed kd in isooctane are discussed, and the less efficient attachment of electrons to the same reversed micelles in TMS compared to attachment in isooctane is interpreted in terms of the relative attachment time, T,, and the residence time of the electron within the electron-micelle encounter radius, T,. With this model, the T , of electrons attaching to micelles at a diffusion-controlledrate is < I ps and may be identified with the solvation time of electrons in water.
Introduction Micelles, reversed micelles, and microemulsions have been intensively investigated due to their compartmentalization properties which can be exploited in applications that range from artificial photosynthesis and solar energy conversion' to enzyme encapsulation and drug deliverye2 The structural features of these species also present a model system in which unique information concerning the kinetics and dynamics of short-lived reactants such as electrons can be ~ b t a i n e d . It ~ was with the mutual objectives of elucidating the reaction mechanism of excess electrons with reversed micelles in liquids and of using excess electrons to probe the structure of water encapsulated in reverse micelles that the present study was undertaken. In a previous study, we used a picosecond-pulse-conductivity (PPC)technique to measure the rate constant, koh, of electron attachment to aerosol OT (AOT)/H,O reversed micelles in isooctane.4 The AOT/H,O/isooctane system was chosen for that study since it was one of the most thoroughly characterized reversed micellar/microemulsion systems5v6and the electron mobility, pe, in isooctane is of sufficient magnitude to yield a good signal/noise ratio. This latter criterion precluded from study two solvents that also have been well characterized with AOT/H20 reversed micelles, viz., heptane7 and cyclohexane.8 In the earlier work, we found kobs to be strongly dependent on the H,O/AOT ratio and to range from 2 X l O I 3 M-' s-I at R = 0.1 to lOI5 M-' s-l at R = 1.5; R is the ratio by weight of H,O/AOT and can 'Deceased May 24, 1988.
0022-3654/92/2096-2328$03.00/0
be converted to w , the molar ratio of H,O/AOT, by the conversion factor of 24.7 at 293 K. Further, we concluded that the electron-attachment process was diffusion-controlled only if free, nonbonded water were present in the micellar pool. In the present study, we have made additional measurements of kobs in the region of R where the transition from bonded to nonbonded water in the micellar pools occurs. Also, we have used Maitra's detailed characterization of the AOT/H,O/isooctane system to obtain more accurate evaluations of kobe5 In addition, electron attachment to the same reversed micelles was extended to another solvent, tetramethylsilane (TMS), in which pe is 20-fold greater than in isooctane, Le., 100, cf. 5.3 cmZ/(V s) a t 21 0C.9 From the magnitude and the field dependence of pelo
-
( I ) (a) Gratzel, M. Acc. Chem. Res. 1981, 1 4 , 376. (b) Calvin, M. Photochem. Photobiol. 1983, 37, 349. (2) (a) Luisi, P.; Wolf, R. Solution Behavior of Surfactants; Mittal, K . L., Fendler, E. J., Eds.; Plenum Press: New York, 1982; Vol. 2, pp 887-905. (b) Luisi, P. L. Biological and Technical Relevance of Reversed Micelles; Plenum Press: New York, 1983. (3) Fendler, J. H. Annu. Rev. Phys. Chem. 1984, 35, 137. (4) Bakale, G.; Beck, G.; Thomas, J . K. J . Phys. Chem. 1981.85, 1062. ( 5 ) Maitra, A. J . Phys. Chem. 1984, 88, 5122. (6) (a) Eicke, H.-F.; Rehak, J. Hela Chim. Acra 1976, 50, 2883. (b) Zulauf, M.; Eicke, H.-F. J . Phys. Chem. 1979, 83, 480. (7) Robinson, B. H.; Steytler, D. C.; Tack, R. D. J . Chem. SOC.,Faraday Trans. 1 1979, 75, 481. ( 8 ) Day, R. A.; Robinson, B. H.;Clarke, J. H.R.; Doherty, J. V. J. Chem. SOC.,Faraday Trans. I 1979, 75, 132. (9) Allen, A. 0.;Gangwer, T. E.; Holroyd, R. A. J . Phys. Chem. 1975, 79, 25.
0 1992 American Chemical Society
