fl '3 Q ‘. . :' , V 1 ' I r- ». ._ i." SE *2 l in.“ h“; 8 flu an «I? -' T ABSTRACT KINETICS OF ELECTRON-ATTACHMENT REACTIONS IN ANHYDROUS ETHYLENEDIAMINE by Larry H. Feldman The reaction of the solvated electron with water is of particular importance to radiation chemists (1). Preliminary studies by Dewald gt El- (2) using solutions of cesium in ethylenediamine to measure the rate of this reaction by the stopped-flow method have been followed by a more extensive study of this system and extended to the other alkali metals. The stopped—flow apparatus was modified to include a rapid—scanning monochromator and detection system capable of making from three to 150 Spectral scans per second. The reaction system is thermostatted and uses Pyrex mixing cells. This arrangement permits study of rapid changes in the spectrum for suitable moderately fast reactions. The instrument is double-beam and direct reading in absorbance units. Data are stored on an FM tape recorder and played back to an oscilloscope. Tests of the instrument‘s per— formance are given. These studies of the reaction of cesium with water in ethylenediamine verify the earlier rate constant Larry H. Feldman (krW’ZO M-lsec’l), and in many cases show the presence of '1). Similar a slower second—order process (k/W’S M’lsec results are found for the reactions of rubidium and potassium with water. The reaction of sodium with water in ethylenediamine is first-order in metal absorbance, but the order in water varies with the water concentration. Lithium-water reactions yield as many as three second-order rate constants, yet the lithium solutions used showed primarily an infrared band. The results are discussed in light of current models for metal-amine solutions and are correlated with the known absorption spectra (5). Preliminary studies of the rate of reaction of metal- amine solutions with electron acceptors other than water have been made. Reactions with methanol and ethanol show rates similar to those with water. In addition, the presence of an absorbing intermediate was indicated, which will warrant more detailed examination. The rate constants for the reactions of cesium solutions with ethylenediammonium and ammonium ions were found to be 1.7 x 105 and a lower limit of 3 x 106 M‘lsec‘l, respectively. The latter value is in agreement with the results of radiation chemistry (4), while the former has not been previously determined. Possible mechanisms for these electron—attachment reactions are discussed and suggestions for further work are made. 1. 2. Larry H. Feldman REFERENCES M. Anbar, Advances in_Chemistry Series, §QJ "Solvated Electron," edited by R. F. Gould, American Chemical Society, Washington, D. C., 1965, p. 55. R. R. Dewald, J. L. Dye, M. Eigen, and L. de Maeyer, ,1. Chem. Phys., £32. 2388 (1963) . . R. R. Dewald and J. L. Dye, J. Phys. Chem., §§, 121 (1964) . . J. Rabani, p. 242 in Ref. 1. KINETICS OF ELECTRON-ATTACHMENT REACTIONS IN ANHYDROUS ETHYLENEDIAMINE BY x 1 Larry HT Feldman A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1966 To my Parents and Sharon ii ACKNOWLEDGEMENTS The author wishes to express his sincere appreci- ation to Professor James L. Dye for his guidance, encouragement, and keen insight throughout the course of this investigation and preparation of this thesis. The advice of Dr. Robert R. Dewald during the initial phase of this work is gratefully acknowledged. Many thanks are due Messrs. Eklund, Grumblatt, Hansen, Seer, and Wreede for their technical assistance. This research was supported by the United States Atomic Energy Commission and the Rapid Scan Monochromator was purchased in part through a grant from the Petroleum Research Fund of the American Chemical Society. ************ iii TABLE OF CONTENTS I. INTRODUCTION . . . . . . . . . . II. HISTORICAL . . . . . . . . . . . A. Metal-Ammonia Chemistry . . B. Metal-Amine Chemistry . . . C. Radiation Chemistry . . . . D. The Stopped-Flow Method . . III. EXPERIMENTAL . . . . . . . . . . A. General Laboratory Techniques 1. Glassware cleaning . . 2. Vacuum technique . . . B. Metal Purification . . . . C. Solvent Preparation . . . . D. Nitrogen Purification . . . E. Solution Make-up . . . . . F. Fixed Wavelength Stopped-Flow G. The Rapid—Scan System . . . IV. RESULTS . . . . . . . . . . . . A. Introduction . . . . . . . B. Alkali Metals + Water . . . 1. At fixed wavelengths . Cesium . . . . . . . Rubidium and potassium Sodium . . . . . . . Lithium . . . . . . iv Apparatus Page 12 17 25 25 25 25 26 28 52 52 55 59 48 48 48 48 48 51 55 55 TABLE OF CONTENTS - Continued 2. With rapid—scan system . . . . . . . Cesium . . . . . . . . . . . . . . Lithium O O O O O O O O O O O O O Rubidium and Cesium + Methanol . . . . . Cesium + Ethanol . . . . . . . . . . . . Cesium + Ethylenediammonium and Ammonium Ions O O C O O O O O O O C O O I C O O O V. DISCUSSION 0 O O O O O O O O O O O O O O O O O A. B. F. G. Introduction . . . . . . . . . . . . . . Alkali Metals + Water . . . . . . . . . . 1. Cesium, rubidium, and potassium . . 2 O SOdium O O I O O O O O O O O O O O O 5 O Lithium O O O O O O O O O O O O O O Alkali Metals with Methanol and Ethanol . Cesium + Ethylenediammonium and Ammonium Ions O O I O O C O C O O C O O C O O O O Alkali Metals + Pure Ethylenediamine . . Mechanisms . . . . . . . . . . . . . . . Future work 0 O O O O O O O O O O O O O 0 REFERENCES . O C C C O O O O O O O O O O O O O O O Page 58 65 65 67 67 75 75 75 84 85 9O 91 91 94 96 98 TABLE II. III. IV. VI. VII. VIII. IX. LIST OF TABLES Properties of the Hydrated Electron . . . . Rate Constants in Pulse Radiolized Water . Rate Constants for the Reaction of Cesium with Water in Ethylenediamine . . . . . . . Rate Constants for the Reactions of Rubidium and Potassium with Water in Ethylenediamine at Room Temperature . . . . . . . . . . . . Rate Constants for the Reaction of Sodium with Water in Ethylenediamine . . . . . . . Rate Constants for the Reaction of Lithium with Water in Ethylenediamine . . . . . . . Rate Constants for the Reactions of Cesium with Water and Lithium with Watgr using the Rapid-Scan System at 25.0.i 0.2 C. . . . . Rate Constants for the Reaction of Rubidium and Cesium with Methanol in Ethylenediamine Rate Constants for the Reaction of Cesium with Ethanol in Ethylenediamine . . . . . . Rate Constants for the Reactions of Cesium with Ethylenediammonium Chloride and Ammonium Bromide in Ethylenediamine . . . . . . . . vi Page . 20 . 22 . 49 O 54 . 55 . 57 . 59 . 66 . 7O . 72 FIGURE 1. 2. 5. 4. 5. 6. 7. 8. 10. 11. 12. 15. LIST OF FIGURES Lithium break-seal apparatus . . . . . . . . . Freeze purification setup . . . . . . . . . . Ethylenediamine purification train . . . . . . Metal solution make-up assembly . . . . . . . Block diagram of fixed wavelength stopped-flow apparatus . . . . . . . . . . . . . . . . . . Schematic of rapid-scan system . . . . . . . . Comparison of a spectrum of potassium perman— ganate taken with a Cary Model 14 spectro- photometer and the rapid-scan system . . . . . Test of Beer's Law for potassium permanganate at 525 mu 0 O O O O O O I O I O O O O O O O O a- Typical oscilloscope trace for the reaction of cesium with water in ethylenediamine show- ing two rates . . . . . . . . . . . . . . . . b- Plot of log Absorbance gs, time for this trace . . . . . . . . . . . . . . . . . . . . Log K KE’ log (water concentration) for the reaction of sodium with water in ethylenedi- amine . O . O . C . C O O C C C C C C O . O O a- Typical trace for the reaction of cesium with water taken with the rapid—scan system. . b- Typical trace for the reaction of lithium with water taken with the rapid-scan system. . c- Slow oscilloscope sweep for the reaction of lithium with water taken with the rapid-scan system . . . . . . . . . . . . . . . . . . . . Plots of In Absorbance Kg. time for Fig. 12a . vii Page 27 29 5O 54 56 4O 41 44 45 52 52 56 61 61 61 62 LIST OF FIGURES - Continued FIGURE 14. 15. 16. 17. 18. 19. 20. 21. 22. 25. 24. 25. Plots of ln Absorbance yg. time for Fig. 12b . Plot of ln Absorbance yg. time for Fig. 12c . a- Typical trace for the reaction of cesium with methanol in ethylenediamine . . . . . . . b- Typical trace for the reaction of cesium with ethanol in ethylenediamine (no inter- mediate) . . . . . . . . . . . . . . . . . . . c- Typical trace for the reaction of cesium with ethanol in ethylenediamine (intermediate present) . . . . . . . . . . . . . . . . . . . Plots of ln Absorbance ya. time for the traces in Fig. 16 O I C O O O O O O O O O O O O O O 0 Comparison of the spectral shapes for typical traces of Run 7.1 with that of Dewald and Dye. Comparison of the spectral shape of a typical trace of Run 7.2 with that of Dewald and Dye . Comparison of an experimental graph of log (Abs.) yg. time with values calculated con- sidering the reaCtionEI+-OH‘ —+-e + H20 . . . Comparison of an experimental graph of log (Abs.) gs. time with values calculated con- sidering the reaction H + OH‘-—+-e + H20 for a trace showing two rates . . . . . . . . . . Comparison of the spectral shapes for typical traces of Run 7.5 with that of Dewald and Dye. Plots of 1/Absorbance Kg. time for the reb action of lithium with water in ethylenedi- amine . O O O C O O O O C O O C O O O O O O 0 Graph of kapp yg. (Abs.h3for the reaction of lithium with water in ethylenediamine . . . . a- Typical trace for the reaction of cesium with pure ethylenediamine showing initial ab— sorbance buildup . . . . . . . . . . . . . . . b- Typical trace for the reaction of sodium with water in ethylenediamine showing initial fast absorbance decrease and recovery . . . . viii Page 64 65 68 68 68 69 76 81 82 86 88 89 92 92 I . INTRODUCTION Hart (1) has pointed out that solvated electrons, resulting from the interaction of ionizing radiation with water, have been present on this planet for four billion years. So reactive is this species, that its existence has been demonstrated only very recently (2-4). In 1864, Weyl observed that alkali metals dissolve in liquid ammonia (5), and in 1908, Kraus proposed on the basis of electrolytic experiments that the anionic Species was, in fact, the ammoniated electron (6). Metal-ammonia solutions have been studied in great detail in a number of laboratories. At the recent Weyl Colloquium, the physical properties of metal solutions were reviewed and models to explain these properties proposed (7). Since this time considerable work has been done and will be reviewed in some detail in Chapter II. Alkali metals are soluble in several primary aliphatic amines, certain ethers, and other solvents. The existence of solvated electrons in many solvents other than water has been established by radiation chemists (8). An important aspect of radiation chemistry has been the measurement of the rates of reaction of e; with different chemical olv species. Most of the work to date has been with egg as the reducing species. Over 500 different compounds have been investigated for their reactivity with egq (9). These bimolecular kinetic studies are necessarily limited to the range k > 105 M-“lsec”l because of competing reactions with impurities and/or solvent. From the viewpoint of radiation chemists, the reaction e- +H20 ——+ H+OH- aq aq is the most important reaction in aqueous radiation chem- istry. Were it not for the slowness of this reaction, k = 16.: 1 M"lsec‘l (10), only the fastest reactions of the hydrated electron could be studied. The reaction of solvated electrons with water has also been studied by Dewald and co-workers (11,12) who reacted solutions of cesium in ethylenediamine using the stopped-flow method, and followed the decay in electron ab- sorbance as a function of time. A rate constant of 24.7.1 1.5 M‘lsec‘l at room temperature was reported. These studies pointed to possible use of such a system for the study of other electron-attachment reactions. In this research, studies using the cesium-water system were further pursued, but complications developed almost at once. A second, slower reaction was found to occur and the data showed an abnormal amount of scatter from run to run. The resolution of these difficulties has formed a major part of this investigation. Reaction of other alkali metal-water systems in ethylenediamine, as well as a number of other electron—attachment reactions, have been examined. In the course of this work, a number of experimental techniques were develOped for the preparation of anhydrous solvents and metal-amine solutions. The stOppedwflow apparatus has been modified and made more sophisticated by including a rapid-scanning monochromator and associated electronics. II . HISTORICAL A. Metal-Ammonia Chemistry The blue color of dilute metal-ammonia solutions is a familiar phenomenon, and probably is what initially in- trigued most investigators. In concentrated solutions, however, there is a metallic copper-like appearance and, in some cases, phase separation. If proper attention is paid to cleanliness and purity of materials, a stable solu- tion can be obtained. Failure to observe such precautions can lead to false or incomplete results because of the decomposition reaction M+NH3—>'MNH2 +§H2 (1) Using EPR techniques, Kirschke and Jolly (15) have demon- strated that the reverse of this reaction can be made to occur to a slight extent. The physical properties of dilute solutions of alkali metals in liquid ammonia have been studied extensively by a wide variety of methods. Some studies have also been made using solutions of alkaline earth metals. These in— vestigations have been the subject of a number of reviews (7.14-17) and will only be considered briefly here. The following methods have been used: electron paramag- netic resonance, static magnetic susceptibility, nuclear magnetic resonance, conductivity, electromotive force, transference number measurements, x-ray diffraction, photo- electric effects, the thermo-electric effect, optical ab- sorption spectroscopy, vapor pressure studies, compressibility and viscosity, density and volume expansion, photolysis (18). and electron nuclear double resonance (ENDOR) (19). Unfortunately, investigators in this field often pro- pose a model which fits the results of a particular set of measurements, but which is inconsistent with those of another set. Any valid model proposed for these systems must explain: 1. The presence of a single asymmetric optical absorp— tion band that peaks near 7000 cm‘l, has no observable fine structure, and obeys Beer's Law (20-22). 2. The volume expansion of metal-ammonia solutions, including the minimum found with sodium-ammonia and potassium-ammonia solutions (25—25). 5. Electrical properties: the fraction of current carried by the anionic species varies from 0.8 in dilute solutions to more than 0.99 for concen- ‘ trated solutions in the neighborhood of 1 M (17. 26); concentrated solutions exhibit high electrical conductances characteristic of metals (27), while more dilute solutions exhibit electrolytic con- ductance; a minimum in the equivalent conductance occurs at about 0.05 M (28). 4. The results of static field magnetic suscepti— bility measurements (29,50) as well as those of electron paramagnetic resonance (51), which indi- cate the need for diamagnetic species because of a rapid decrease in molar susceptibility with concentration. 5. Knight shifts and relaxation times of alkali-metal and nitrogen nuclei (52) and protons (55) in metal- ammonia solutions. Several models have been proposed, none of which has been fitted to all of the data. These are reviewed below. The earliest model, proposed by Kraus (6), assumed the presence of undissociated sodium atoms, sodium ions, and sol- vated electrons according to the equations Na = Na+ + e (2) e- + x(NH3) = (NH3)Xe— (5) A number of observations led to the realization that the electrons in metal-ammonia solutions are trapped rather than free. 099 (54) postulated that the solvated electron could be considered as trapped in a spherical cavity sur- rounded by ammonia molecules. He assumed that the electron is confined by an infinitely steep spherical potential well. The wave function of the electron according to this model is that for a particle in a spherical box of radius R0, w = 1 sin(vr/RQ) (4) ZWR r o The zero point kinetic energy is ha/Bng while the polari- 2 zation energy of the dielectric is - %_ (1 - l-), where D RD Ds s is the static dielectric constant of ammonia. The total energy is then given by _ h2 .23 .1 E-BTIRE- 2Ro(1-DS) (5) Minimizing E with respect to R0, letting 1/DSC'0, gives 4 E = - fig: = - 0.586V' (6) and 2 R0 = 2284 = 9.9 R (7) This treatment leads to an underestimate of the enthalpy of solution and to an overestimate of the cavity radius. Lipscomb (55) refined these approximate calculations by incorpOration of several additional contributions to the energy of the system: electrostriction, electronic polari- zation of the molecules on the cavity surface, and surface tension. Stairs (56) further improved these calculations by making a better estimate of the wave function using a graphical method of integration. A natural extension of the primitive cavity model assumes that the surrounding solvent molecules do not provide ) an infinite potential well for the electron. Thus the charge distribution of the extra electron can extend into the medium beyond the cavity boundaries (57). Application of polaron theory by Jortner (58) considers the interaction between the electron and the polarization field of the medium. This approach is an improvement upon a similar treat- ment given by Davydov (59) and Deigen (40), who also used polaron theory to theoretically treat metal-ammonia solu- tions. In Jortner's treatment (58) the polarization is con- sidered to be composed of two parts, the total polarization and the electronic polarization.” The difference between these is the "permanent pOlarization" which creates the potential well. The electrostatic potential resulting from this polarization is assumed to be continuous up to R0, the cavity radius, and constant within the cavity. A one para- meter hydrogen-like wave function of the form Wis - exp (-ur) (8) -23 w is used for the ground state and the energy of the ground state is calculated from the variation integral h2v2 81ram + v(r)] gals dv (9) W13 = fv ¢is [- by putting BWIS/ALL= 0. In this model the electronic polar- ization will lower the energy of this state by an additional amount. Agreement between this model and eXperiment is fairly good, considering the number of approximations involved (17,58). An improvement in these calculations has been made by Jortner, Rice, and Wilson (57) using a self- consistent field approach. This treatment, however, does not enhance agreement with experiment. The experimental value of the enthalpy of solution is 1.7.: 0.7 ev, and the value calculated by Jortner (58) is 1.60 eV, while the SCF value (57) is 0.92 eV. A model that has seen considerable popularity is that of Becker, Linguist, and Alder (41) (BLA model), which is sometimes referred to as the cluster model. It is assumed that the metal atom is ionized and that the electron is trapped by the potential produced by the metal ion and the oriented ammonia molecules around the ion. In addition to this "monomer" species, a solvated metal ion, an electron arising from the dissociation of the monomer, and a dimer consisting of two monomers bound principally by exchange forces are postulated. The model involves consideration of the competition of two equilibria M=M+e (10) and M2=M+M (11) Presumably the electrons are solvated, but no explicit mention of their state is made (41). The BLA model has been used successfully to fit conductivity (26,41,42) as well as magnetic susceptibility data (41). 10 Gold, Jolly, and Pitzer (45) have pointed out that the BLA model is not consistent with optical, volumetric, and NMR data. These authors propose instead that the monomer is a simple ion-pair consisting of an ammoniated metal ion and an ammoniated electron. The dimer species, M2, is pictured as a quadrupolar ionic assembly of two solvated electrons and two solvated metal ions in which there is little distor- tion of the solvated species. On the basis of Jortner's calculations (58), they assume that the wave functions for the two egm in M2 will overlap significantly, and further that the singlet will be lower in energy than the triplet by more than kT. In this way, at least qualitatively, optical, volumetric, and Knight shift data are explained (45). It should be pointed out that O'Reilly (44) has recently used the BLA model to successfully fit Knight shift and relaxation time data (52). He was also able to calculate spin densities at the nitrogens of the cavity species which are in good agreement with experiment (44). When calculating equilibrium constants from experi- mental data, either the BLA monomer or an ion pair, M+-e_, can be used for the M species in Equation 10. However, the constants calculated from conductivity data differ from those obtained using magnetic susceptibility data (45). These difficulties can be overcome by introducing another I species,(M)7 and a third equilibrium, (M')'= M + e" (12) 11 Such a scheme has been proposed by Arnold and Patterson (46,47) and by Golden, Guttman, and Tuttle (48). The MI center is diamagnetic and negatively charged. Arnold and Patterson (47) made no assumption about the exact nature of the various centers but confined their discussion to the calculation of the equilibrium constants for the various processes (Equations 10-12). The constant for Equation 10 was determined from conductivity data at low concentration and the other constants were computed from transference number and susceptibility data. Agreement with experiment is good, even at high concentrations, but three adjustable parameters are required. The model of Golden g£_al. (48) introduces the metal anion, M-, for the M. species. Equations 10-12 would then be rewritten as M+ . s" = M+ + s' (10') M+ - M' = M+ + M- (11') and M— + as = M+ + 28- (12') solv’ USing where S is solvent and S- is equivalent to e this scheme and only a single adjustable parameter, they obtain good agreement with vapor pressure, conductivity, Knight shift, and optical absorption data. The last com- parison, however, according to Jolly (49), may be in error. By a thermodynamic argument, Golden g; 3;. (48).show that 12 the M- species is not equivalent to the ion triple, S--M+-S-. This argument is only valid for non-polarizable‘ hard sphere ions and neglects consideration of wave function overlap in the ion triple as a mechanism to lower its energy. The data they used cannot, in fact, distinguish among the various descriptions of the species. The advantage of the treatment of Golden g; 31. over that of Arnold and Patterson is the reduction of the number of adjustable parameters from three to one. B. Metal-Amine Chemistry Alkali metals dissolve in many primary aliphatic amines, several ethers, and in a few other oxygen containing solvents (11). In all cases, the blue color characteristic of metal- ammonia solutions appears. Work with metal-amine solutions has been less extensive than with metal-ammonia solutions. A possible reason was the lack of stability of the former as compared to that of the latter. Analogous to Equation 1, these solutions can decompose with the evolution of hydrogen M + RNH2 —-> M(RNH) + 1‘ H2 (13) Recent results in a number of laboratories have indicated that this reaction is not much faster than the corresponding reaction in liquid ammonia (after adjustment for temperature (differences), and with proper care its effect can be mini- mized. This .could lead to an increase in the number of workers in the field. 15 A summary of earlier studies of optical absorption spectra (50) and conductivities (51) is given by Dewald and Dye. In these papers, and in Reference 11, aside from pre- senting extensive experimental results for metal solutions in ethylenediamine (EDA), the authors pointed out the need for a model for metal-amine solutions which did not simply modify existing models for metal-ammonia solutions (52), but added other species. Dewald and Dye (50) observed one to three very intense absorption bands depending upon the alkali metal used. One of these peaks appeared in the visible region at 650 mu (V band), one in the near-infrared region at 800 to 1100 mu (R band), and the third in the infrared at 1500 mu (IR band). The V band is present in lithium, sodium, potassium, and rubidium solutions and its shape and position are independent of metal, but the intensity relative to the IR absorption depends strongly upon the metal. Solutions of potassium in EDA show the R band at 845 mu, rubidium at 890 mu, and cesium at 1050 mu. All solutions show IR absorption, but in sodium solutions this absorption is negligible compared with that of the V band. The various bands are only slowly interconverted. The conductances of cesium, rubidium, potassium, and lithium in EDA show the same general concentration depend- ence, while sodium solutions exhibit a marked difference in behavior (51). Dye and Dewald (52) postulate that the IR species in these EDA solutions are "ammonia-like" in nature 14 containing species consistent with the model of Gold, Jolly, and Pitzer (45), and that the metal-dependent R band is due to covalent dimers similar to those found in alkali metal vapors. The V band was attributed to an electron trapped by a solvated M: ion in the same way that the electron is presumed trapped in a BLA monomer unit (41). The possibility that the V species was in fact a BLA monomer was ruled out on the basis of the slow conversion of this species (52). These assignments were supported by thermodynamic estimates and the ability of the model to account for the diverse re- sults reported by others (52). The discovery of nuclear hyperfine splitting by the metal nucleus in metal-amine solutions (55,54) has introduced a powerful and sensitive method for examining the paramagnetic ' species in these solutions. Bar-Eli and Tuttle (55) studied the EPR spectra of solutions of potassium and lithium in ethylamine, potassium in methylamine, and potassium and lith- ium in mixtures of these amines. They have also examined rubidium and cesium in ethylamine (56). The hyperfine split- ting obtained in these solutions is strongly dependent upon temperature and solvent composition. A striking example of the latter is seen in the work of Dalton and co-workers (57) on the EPR spectra of potassium in ethylamine-ammonia mix- tures. In samples containing 0.6 to 25.4 molewpercent am- monia the hyperfine splitting changed from 11.52 to 2.64 gauss at 250 C, and corresponding changes were observed at other temperatures. Bar-Eli and Tuttle (55) interpreted 15 these phenomena in terms of a modified BLA model, while Dalton g£_al. (57) quantitatively described the temperature dependence with a model involving equilibrium between atoms and monomers, and introduced an additional monomeric species (one in which an ammonia molecule has replaced ethylamine) to explain the dependence upon ammonia concen- tration. O'Reilly and Tsang (58) have sought to explain the temperature dependence of these solutions in terms of the combined action of solvent-molecule polarizability and an expansion of the monomer with temperature. Ottolenghi and co-workers (59) have assigned the V band to the monomer and the R band to a dimer for metal solutions in ethylamine. This assignment is based upon the observed constancy of the ratio of D550 mu to D850 mu for both potassium and rubidium in ethylamine, where D is absorbance. They also observed that the decay of the EPR absorption closely followed that of the V band for solutions of potas- sium in ethylamine. Similar results for the ratio D200 mu to D have been reported by Ottolenghi and Linschitz 1000 mu (18) in both the original and in photo-regenerated potassium and rubidium solutions in ethylamine. It has been shown conclusively by Dye, Dalton, and Hansen (60) and by Dalton e; 3;. (57) that the species responsible for the V band and that giving rise to the hyper- fine splitting (monomer) are not the same. Their arguments are based upon determination of spin concentrations, opposite 16 changes in V band absorbance and number of Spins as a func- tion of temperature, dependence of spin concentration of the hyperfine species and the absorbance of the V band upon ammonia concentration, and the kinetics of decomposition. The latter was found to yield a D650 to D850 ratio of nearly unity for potassium in ethylamine (57) and for extensive kinetic studies (61) of alkali metals in EDA. Thus the nature of the V species is still largely a matter of con- jecture, but it is pg£_the same as the species observed by EPR. Ottolenghi and co-workers (62) demonstrated a large shift of the maximum of the R band towards the blue with temperature for solutions of potassium, rubidium, and cesium in ethylamine. Such a shift is not consistent with the previous assignment (52) to the lZE't-12; transition of an alkali metal dimer. Recent measurements on the spectra of gaseous Rba and C82 show that the absorption should probably be assigned to the lwh‘<-IZ$ transition shifted to lower energies by the interaction with solvent (65). Alternatively, the intermediate absorption could be due to another type of species. Anbar and Hart (64) have studied the properties of an electron-pulse-generated species in ethylenediamine. They observed only one band with a maximum at 920 mu and a two usec. half—life, and attributed this to the solvated electron. This does not agree with the earlier assignment based on 17 spectra and conductivity (50-52). Recently, this system has been re-examined (102) and the earlier peak was found to be an artifact of the measuring technique. The absorbance of the transient produced by pulse radiolysis rises continu- ously from 700 mp. to the limit of the method (”1100 mu) . This strengthens the assignment of the IR band to the sol- vated electron. C. Radiation Chemistry Historically, the evolution of gas from aqueous solu- tions containing radium salts was one of the earliest radiation-induced chemical reactions to be observed and studied. Both oxidation and reduction of dissolved substances were observed,and as far back as 1914, Debierne (65) suggested that free radicals from the water might be responsible for the chemical action of radiation. In 1959, Barr and Allen (66) showed that the hydrogen "atom" (H.) formed in the radiolysis of neutral aqueous solutions containing hydrogen, oxygen, and hydrogen peroxide by oxidation of hydrogen 0H + H2 —-> H' + H20 (14) reacts with oxygen much faster than with hydrogen peroxide, while for solutions containing only oxygen and hydrogen peroxide, the rates of reaction with hydrogen "atoms" pro- duced by water radiolysis are comparable. They concluded 18 that the two types of hydrogen atoms were different, and that one was atomic hydrogen and the other an acidic or- basic form of the hydrogen atom. The pH dependence in the radiolysis of aqueous systems provided further evidence for this hypothesis (67,68). The species present in acid solu- tion, Ha, can be formed from Hb by + Hb + H30 —-> Ha (15) This reaction has a pK value of about 5 (69). H: was elimi- nated as a kinetically important acid form by Czapski, .Jortner, and Stein (70) who produced H atoms by electrode- less discharge and Showed that the reaction H + H202 9 0H + H20 (16) was unaffected by pH from 1-15. A number of experiments (71) showed H. to be atomic hydrogen, and Czapski and Schwartz (72) utilized the kinetic salt effect to demonstrate that the reducing radical produced in the cobalt-60 radiolysis of water has a unit negative charge and may be considered a solvated electron. Recent work using pulse radiolysis con- firming these results has been reviewed by Matheson (75). Hart and Boag (2,5), using a pulsed beam of electrons to irradiate pure deaerated water, observed a transient absorption band peaking at 7200 3. Since the spectrum was suppressed by known electron scavengers, such as H30+, 02, and N20, and resembled the absorption bands of the solvated 19 electron in liquid ammonia and methylamine, it was identi- fied as that of the hydrated electron. The results of Keene (4) confirmed these observations. The probable existence and properties of the hydrated electron had been predicted earlier by Platzman (74). Since its discovery in 1962, a number of the properties of the hydrated electron have been determined and are summarized in Table I. The initial effect of ionizing radiation on pure liquid water is to strip off electrons from the water molecules in the path of the beam. It is believed that the products of radiolysis (egg, H, OH, H30+, OH-, H202, and H2) are formed within 10'8 sec. What probably happens before this time has elapsed is reviewed by Spinks and Woods (71) and will not be discussed here. A number of investigators include activated water molecules, H20*, in the list of initial radiation products (75). The existence of holes or positive polarons, H20+, has been discussed by Moorthy and Weiss (76) in the irradiation of frozen aqueous solutions at 770 K, but their presence in liquids has not been demonstrated. Reactions of the hydrated electron are numerous, and a recent review by Anbar (9) discusses a number of them in detail. The reaction eaq + H20 —>- H + OHaq (17) is of particular importance to the radiation chemist. A summary of rate constants for the reactions that occur 20 Table 1. Properties of the Hydrated Electron* )‘max Molar exti§ction coefficient at 5780 at 7200 8 (max.) f = oscillator Strength Charge pKa (ea + H20 = H + OHaq) q Hydration energy (calc.) o - + + + E (ea Hgoaq * é‘ H2 H20) q Té (pure water at pH 7.0) Diffusion constant Mean radius of charge distribu- tion (calc.) G - (in neutral water) e aq 7200 R (1.72 e.V.) 10,600 (1 10%) M-lcm-l 15,800 M'lcm‘l 0.65 -1 9.7 2_1.72 e.V. g_2.67 volts 780 usec (Ref. 10) 4.5 x 10'“5 (i 15%) cme/sec 2.5 - 3.0 R 2.6 *- From Matheson (75). he l\') [‘9- (1‘ L) J 21 between many of the primary radiolysis species, taken from Matheson (75), is presented in Table 11. Since Reaction 17 is slow, it does not interfere appreciably with these reactions. The reverse of Equation 17, H + OH ‘—+ eaq + H20 (18) has been studied in detail (77,78) and a rate constant of 2.2 x 107 M"J'sec‘l obtained (78). Combining this with the value of 16 M‘lsec"l (10), obtained for Reaction 17, gives a pKa of 9.6 for the equilibrium eaq + H20 = H + OHaq (19) It has been proposed that e;q reacts with Brdnsted acids following the Brénsted general acid catalysis law (78). Anbar (9) has reviewed evidence contrary to this hypothesis and suggested instead that the electron is incorporated into the acid to form an AH- radical, which may subsequently undergo decomposition. In addition to Reaction 17, e;q may disappear by the bimolecular process eaq + eaq-—+ H2 + 20Haq (20) It has been shown conclusively (77,79) that this reaction produces molecular hydrogen, and that the hydrogen produced does not originate from the recombination of two H atoms or Table II. 22 Rate Constants in Pulse Radiolized Water* Reaction pH Rate Constant (M-lsec‘l) 1. e;q+H20—>H+0H;q 8.4 1511 2. e;q + H30:q -> H + H20 2.1—4. (2.07 1: 0.05) x 101° 5. e;q + 11202 -> OH + 0H;q 7 (1.25 1 0.14) 5: 1010 4° egg + egq ‘* H2 + 2°ng 10.9- (0.9 1 0.15) x 1010 5. qu + H —>’H2 + OH;q 10.5 (2.5.1 0.6) x 1010 5. e;q + on ->- ongq 10.5 (5.0 1 0.7) x 1010 7. e;q + ogq 522+ 205;“; 15 (2.2 .4 0.5) x 1010 s. H + ougq ->- e;q + H20 11.5 (1.5 .4 0.5) x 10" 9. H+H—->H2 2.1 2x1010 10. H + 0H ->H20 5 (0.7 - 5.2) x 1010 11. H+H202+0H+H20 2.1 (9.011) x 107 12. on + OH —> H202 7 (1.25 .4 0.15) x 1010 1:5. 0H + H2 —) H + H20 7 4.5 x 107 14' O;q + H2 "9' H + Ong 15.5 (5 i 4) x 107 15. OH + H202 —> H20 + H02 7 4.5 x 10" 15. 1130;; + 0H;q —-> 2H20 7 1.45 x 1011 * From Matheson (75) 25 from e;q and H in a cage. The exact mechanism of this re- action is not clear. Anbar (9) has speculated that it might proceed through an intermediate hydride ion, H-, since hydride ion formation is thermodynamically feasible and H- reacts with water at a diffusion-controlled rate. Optical absorption spectra of the solvated electron have been reported for a number of liquids other than water. Sauer, Arai, and Dorfman (80) discuss the absorption spectra and yields of the solvated electron in the aliphatic alcohols by pulse radiolysis and the y-radiolysis of ethanol is con- sidered by Myron and Freeman (81). The results of Anbar and Hart (64) in ethylenediamine have been discussed previously (Sect. II-B). A reversible increase in the conductivity of liquid ammonia during y—ray irradiation has been reported by Ahrens, Suryanarayana, and Willard (82) and an optical spectrum attributed to the solvated electron in ammonia was recently reported by Compton and co-workerS (85). Other solvents, along with some theoretical correlations, are reviewed by Dorfman (8). D. The StOpped-Flow Method The stopped-flow method of Chance (84), as modified by Gibson (85), has been used by a number of investigators (86-88) to study moderately fast reactions with half-times from.two milliseconds up to 100 seconds, and requires rela— tively small amounts of reagents. Precision is usually good 24 (ev5%), and the method only requires a rapid measurement of concentration changes. When this technique is applied to systems exhibiting a color change during the course of reaction, several dif- ficulties arise. It is generally not possible to study intermediates, or simultaneous reactions of species whose Spectral bands are close together. For the most part, studies involving changes in absorbance as a function of time using this method have been carried out at one, or at most two, wavelengths at a time. On the other hand, the continuous flow method (87) requires a precisely known flow velocity and rather large amounts of reactant solutions. The advantages to be gained from the ability to scan a large spectral region in a short period of time, have led to the development of several rapid-scanning monochromators (89,90). Incorporation of one of these (89b) into a stopped— flow system is discussed in Sect. III-G. III. EXPERIMENTAL A. General Laboratory Techniques 1. Glassware cleaning. The items were first rinsed briefly with a hydrofluoric acid-detergent cleaning solu— tion (5% hydrofluoric acid, 55% concentrated reagent grade nitric acid, 2% acid soluble detergent, and 60% distilled water, by volume), then ten times or more with distilled water. Next, they were filled with fresh aqua regia and heated to boiling, followed by ten more rinses with dis— tilled water, then by several rinses and/or soakings in double-distilled conductance water. After each use and a preliminary cleaning, they were heated in an annealing oven at 5500 C before being cleaned as described above. 2. Vacuum technique. High vacuum techniques were used wherever pessible with well-trapped mercury vapor diffusion pumps giving pressures between 10"5 and 10"6 torr, as read on a calibrated ionization gauge. Dow Corning high vacuum silicone grease was used on all stopcocks which came into contact with liquid; Apiezon N grease was used elsewhere. Apiezon T grease was used on movable tapers and ball joints, and Apiezon W wax was used on all fixed joints. 25 26 B. Metal Purification Alkali metals of the highest purity commercially avail— able were obtained from the following sources: sodium, J. T. Baker: potassium, Mallinckrodt: rubidium, Fairmount Chemical Company: cesium, a gift from the Dow Chemical Company; and lithium, Lithium Corp. of America. To obtain small lithium samples free from oxide and nitride surfaces, the following procedure was used. A Caemco Model 59 dry-box was evacuated and filled with high purity helium (99.7%) which had been passed through an activated silica gel column at liquid nitrogen temperature. The glass— ware shown in Figure 1, a knife, a forceps, and a piece of lithium that had been washed with benzene were put into the dry-box airlock. After evacuation of the airlock and refill- ing with helium, these items were taken into the main box. The stopcock plug had been greased with Apiezon N grease and was kept apart from the stopcock barrel. A small piece of lithium, cut from the center of the larger piece, was placed in the stopcock against the break-seal as shown. Next a magnet sealed in glass was slipped into the stopcock, and finally the plug was inserted and turned to spread the grease. In most cases, the lithium would remain bright and shiny in the dry-box, but the use of phOSphorous pentoxide as a desiccant seemed to help somewhat. The stopcock was then removed from the dry-box, pumped down to better than 10'5 torr, and sealed off where the glass had been thickened. 27. l g 15/9 6mm right-angle ' /" high-vacuum O Stopcock \ Thickened glass ‘gm’z— Magnet sealed in glass )fi \\\\\ ‘ L Lithium meta l Fragile / break—seal Figure 1. Lithium break-seal apparatus. 28 The other metals were cleaned, cut, and then degassed under high vacuum. Repeated vacuum distillations were used to obtain small metal samples in sealed ampoules which were used to make up metal solutions. Details of the techniques required are given by Dewald (11) and by Vos (91)° C. Solvent Preparation Ethylenediamine (Matheson, Coleman, and Bell 98-100%) was purified in either of two ways. The first method (11,50) involved distillation at atmospheric pressure in a stream of nitrogen, followed by reaction with alkali metals, and re- peated vacuum distillations. In the second method (for which we are indebted to Professor Paul Sears of the University of Kentucky), EDA was purified in a large separatory funnel by slowly freezing it and then melting and discarding the core (see Figure 2). Pre-purified nitrogen (Matheson) was used as a covering gas throughout this procedure. The process was repeated at least four times and resulted in a pure product, specific conductance = 10'7 ohm‘lcm‘l, without the diffi- culties involved in distillation. Following freeze-purification, the EDA was put over sodium pieces in a flask (Figure 5 - flask no. 1) which had been evacuated with a mechanical pump through a trap at 770 K. When the bubbling had quieted down, small pieces of lithium metal were introduced by turning the sidearm, and a deep blue solution formed almost at once. The EDA was 29 4mm 9 stopcock To Variac 4‘ k’r”— S 40/50 18/9 ._ I) 9 5—Liter separatory funnel Heater 10mm 8 Stopcock u ,4- : 14/55 '3 54/45 5 18/9 -§ To trap and rough pump 1-Liter filter flask Figure 2. Freeze purification setup. 50 mmomem m2 4...: .. . O. fl .oz HANQ a'flAI. , 9:59 nmsou T R one menu 09 ”my ‘ + .cflmuu COHDMUAMHHDQ mceEmHomsmamnum mxmmHm m umueqnm M OOZ .m musmflm AOV Edsom> amen OB 51 then cooled using an ice-water mixture and degassed with a rough pump during freezing. Next it was thawed, allowed to stand for about a day, and frozen again. After thawing, the solution was pumped on with high vacuum and, from this point on, only high vacuum was used. The solvent was then vacuum-distilled onto a potassium mirror which had been vacuum-distilled from the sidearm on flask no. 2. It took about a day to distill all of the EDA, and any hydrogen produced was pumped off from time to time. At this point, stopcock A was closed and freeze-purified EDA was again introduced into flask no. 1, and treated as described above. The EDA in flask no. 2, which had remained for about two days with potassium in solution, was next vacuum-distilled into flask no. 5. Then the EDA in flask no. 1 was distilled into flask no. 2, keeping stopcock B closed except when pumping off the hydrogen. After the dis- tillation into flask no. 2 had been completed, stopcocks A and B were closed and the EDA in flask no. 5 was frozen, degassed, and covered with specially purified nitrogen (Sect. III-D). The EDA in flask no. 5 was drawn off as needed through stopcock C, and the EDA in flask no. 2 was vacuum-distilled to flask no. 5 when necessary. Since the metals were used up and the stopcocks gradually became "frozen," it was necessary to clean and reassemble the en- tire purification train for each double batch of solvent prepared. 52 D. Nitrogen Purification Nitrogen (Matheson, pre-purified) was used as a cover- ing gas for all liquids. It was further purified by passing it over copper turnings in a tube furnace, through an Ascarite column, a dry-ice trap, and finally a silica gel column at 770 K in which some condensation occurred. This nitrogen was then stored in two-liter bulbs on the vacuum line. E. Solution Make-up Water solutions in ethylenediamine were made from freshly distilled conductance water that was degassed by re- peated freezing and pumping, and then vacuum-distilled. Next it was either vacuum-distilled directly into a solution make-up flask or was used to prepare fragile glass ampoules, which could be used to make up water solutions of lower con- centration. With proper care, solutions whose concentrations were known to one part per thousand could be made up. Complete details of these procedures are given in Reference 11. Absolute methanol was distlled from a magnesium- magnesium sulfate mixture and then degassed and vacuum dis- tilled. Spectral grade absolute ethanol was distilled after reaction with sodium metal, then degassed and vacuum- distilled. A solution of ethylenediammonium chloride (EDAH+Cl-) was prepared by introducing a known amount of hydrogen chloride gas (Matheson) into a solution make-up 55 flask. An ammonium bromide solution was made using the reagent grade salt. A cesium hydroxide solution was pre- pared using a cesium mirror onto which water treated as above was distilled. The concentration of cesium hydroxide is uncertain because of the separation of a small amount of a liquid-phase when EDA was added. These solutions were made up as soon before a given run as possible, but could be kept for as long as two months, if the flasks were suf- ficiently leak-tight. Metal solutions were made up immediately prior to each run. The apparatus used is shown in Figure 4, which is after Dewald (11). After attaching this apparatus to high vacuum, a sealed capillary containing the required amount of metal was broken open and dropped into the sidearm on the metal make-up vessel, which was then pumped down to better than 10"5 torr. The metal was melted down into B, and A was pulled off under vacuum. The metal was degassed and the flask flamed until the pressure stabilized at better than 10'5 torr. Next, the metal was vacuum-distilled and B was pulled off under vacuum. EDA was introduced, the solution was stirred with the magnet sealed in glass, and any hydrogen produced was pumped off. Nitrogen was then put over the solution, about ten ml. were run into waste, and then about 50 ml. were run into the decomposition vessel for analysis. The metal-ethylenediamine solutions prepared in this manner were stable from eight to 24 hours without 54 .mfiaav UHMBmQ Eouwa SHQEwmmm msamxmfi QOHDSHOm Hmumz .e seamen Hem mos. sowfim omgo ooQ 3 m we? Wanoxmfi H332 mwmfiw cm oedema uocmmz IV. “533356 H.308 no guess a Emmet, 3mm? Al N oh no :00 3500mm 0H. Emma uwo>nmmmh user/Hem so xuoomoum >m31N tmfiz 520m )(s. / 1 N .52 1 .52 v. 5309, am} Oh. (w. m .52 55 visible evidence of decomposition at room temperature. Metal concentration was determined by measuring the amount of hydrogen collected from a solution decomposed with ammonium bromide, since NH4Br + M(in EDA) +15 H2 + MBr + NH3 Lithium solutions could not be prepared as above, so a different procedure was followed. After seal-off of a lithium sample (Sect. III-B), the ampoule was carefully sealed to the metal make-up flask in place of the melt-down sidearm. After evacuation, the flask was filled with EDA and nitrogen covering gas put in as well. The break-seal was shattered only after the flask had been connected to the flow apparatus. Lithium solutions prepared in this fashion were not reproducibly stable, but did last as long as necessary for a run (four hours) in two of the three lithium runs. F. Fixed Wavelength Stopped-Flow Apparatus Only a brief description of this apparatus, shown schematically in Figure 5, will be given here, since con- struction details of a Similar apparatus are given by Dewald (11) and elsewhere (92). The entire system, includ- ing syringes, could be pumped down to better than 10“ torr. Only one concentration of the metal solution was used in a given run, while up to five concentrations of the more 56 .msumummmm SCHMIommmoum numcmam>m3 omxem mo Emummeo xoon .m musmwm MHOEflU mmoomoHHflomo monuoum H030HHOm moonumu mmmcmem one hammsm Hm3om Hummeue wasp coeum>uwmno 6cm Haou msfixflz umaameuassouosm Omcwuhm msemmoum maesmsm monsom unwed _ scumfiounuocoz souezmouonz 57 stable solutions could be prepared in separate vessels and attached to the system. Dilutions with pure solvent were made with a thermostatted buret, and the solutions thoroughly mixed in a thermostatted vessel using a magnet sealed in glass. The syringes were also thermostatted using plastic jackets sealed with O-rings. The thermostatting fluid was circulated by a Lab-Line Hi-Lo Tempmobile. The solutions were delivered gig heavy-walled capillary tubing to their respective syringes. Pyrex ball joints were sealed to the syringes using uranium glass as an intermediate. Three-way high-pressure, vacuum-ground stopcocks, originally made in Germany but also available from H. S. Martin Co., were then turned to the mixing chamber and a lever depressed which pushed the syringes. A four-jet Plexiglas mixing chamber, which had four inlet holes of 0.5 mm bore each arranged almost tangentially about a 1.0 mm central hole, was used. This cell is capable of a mixing efficiency of better than 98% in two milliseconds, as measured by an acid-base reaction. The mixing cell was mounted upright to eliminate bubbles. It was gradually attacked by the solutions, and hadito be changed every few runs.~ To eliminate this difficulty, a four—jet Pyrex mixing chamber of 1.0 mm i.d. was constructed from precision-bore heavy-walled capillary which had been ground and polished by Precision Glass Products Co., Oreland, Pa. This cell was constructed by the University Glass Shop as follows: First 58 a steel template was machined so that four heavy-walled capillary sidearms of 0.5 mm i.d., when sealed to the central tube would have their bores arranged almost tangentially around the central hole and make an angle of about 1050 with the capillary. The sidearms, closed at the end, were sealed to the central tube one at a time, taking care that the template remained hot during the process. A tungsten rod4~’0.5 mm in diameter was twisted almost into the central tube using the radiant heat from the rod to soften the glass ahead of it. The hole was completed from above and the procedure repeated. Then the central hole was closed by melting a fitted glass rod which rested on a tungsten rod that had been inserted through one of the side- arms, and finally this area above the mixing chamber was flattened using a carbon rod. These cells were found to perform as well as the plastic cells. Since Pyrex is not suited for work below about 520 mu, two cells were constructed from fused quartz by W. A. Eberhart, Windsor, Ontario. The basic design is similar to that described above, but the holes into the central tube were precision drilled. The cost of these cells is at least ten times that of the Pyrex cells. When flow was stopped by the filling of a third syringe against a metal plate, a micro-switch triggered the time base of a Tektronix Model 564 Storage Oscilloscope. The vertical deflection on the sc0pe produced by the output of the photomultiplier (RCA 7102 or 6199, depending upon 59 spectral region) was proportional to the light intensity over at least two decades. The trace was then photographed using a Polaroid camera with type 146L transparency film. Sample traces are shown in Chapters IV and V. G. The Rapid-Scan System The principal component of the spectrometer, shown schematically in Figure 6, is a Perkin-Elmer Model 108 Rapid Scan Monochromator. This instrument with appropriate prism is capable of scanning a selected spectral region between 0.2 and 10 u, with from three to 150 scans per second. External controls permit selection both of the interval to be scanned and the center of scan. Different scanning speeds are obtained either by interchanging gears or with a two- speed switch. The spectral scan is produced by a modified double-pass Littrow system with a rotating tilted mirror. The slit is 12 mm high, with bilateral adjustment from zero to two mm, and the monochromator has an effective aperture of f/5. AA photograph of the system is shown in Figure 7. For the present application, a fused quartz prism was used, which can give a usable range with appropriate source and detector of from 0.2 to 2.5 u. A sine sweep generator is incorporated into the monochromator, and should produce an output voltage proportional to the tilt angle of the nutating mirror to drive the horizontal input of an oscillo- scope. The display produced should have a linearity 40 .Ewumwm cmomloflmmu mo UHumesum .m musmflm .oE amaze; 2:3 .556 o» 250255 «9.31m uzozaomoi +1243... N E _ E- ---,, 53003 52.8 c 1 «0:23:85: - we: s... 50.. ..|_|' flue Vm\ 23m 915. E 3 ._.._ (u . _uaoom o u 7 x m _ 2:050 333 758.5 :83 _ 45:3 _ 442G; roz>m 41 .Ewum>m cmomloflmmu mo Smmnmouonm .e musmflm .o D I. A... o o o 0: 42 approaching the wavelength calibration curve of the prism, but, in fact, is non—linear because of a distorted sine out- put. Initial difficulties prevented superposition of the forward and reverse scans so that the signal could be used for triggering purposes only. Even this arrangement was not entirely satisfactory because noise on the synch signal caused random fluctuations in the triggering level. This led to an apparent shift in the spectral peaks under study. The instrument has so far been used only over the range 0.4 to 1.1 u. For this purpose a Bausch & Lamb source employing a tungsten-iodine lamp with a quartz envelope was used. As shown in Figure 6, the light after leaving the monochromator, impinges on a stacked mirror beam-splitter (from a Bausch & Lomb Spectronic 505 Spectrophotometer). The spherical focusing mirrors in this beam—splitter were replaced by two of longer focal length (98 mm) obtained from Karl Lambrecht Co., Chicago, Ill. The parallel beams are focused on the centers of the flow cell and on a reference cell, if desired, and then strike two photomultiplier tubes (RCA 6199 or 7102). The anode currents, which have been balanced as well as possible over the region of interest, are next fed into the log circuit. The log circuit consists of a high impedance input amplifier, a dual logarithmic transconductor, an Operational output amplifier, and a power supply. These components are available from Philbrick Researches, Inc. (Model nos. P25A, 45 PL1-P, P65AU, and PR-50C, respectively). This circuit pro- duces an output voltage that is directly proportional to the logarithm of the ratio of the current from the reference phototube to that of the sample, and hence to the absorbance of the sample. The signal shape and absorbance are inde- pendent of photomultiplier voltage and of slit width, within limits set by low light intensity and spectral bandpass. The typical noise level for a well-resolved spectrum is five mv. peak-to-peak as measured on an oscilloscope. The output voltage, adjusted to one volt per absorbance unit, is equal to absorbance fromvfl'0.01 to at least two absorbance units. Figure 8 shows a comparison of the spectrum of potassium permanganate taken with a Cary Model 14 Spectro- photometer and with our system. Figure 9 shows a test of Beer's Law for potassium permanganate in a one cm cell made with both the Cary 14 and with our apparatus. Also shown here are the results at higher concentration using a 1.0 mm capillary tube and the rapid-scan set-up. The capillary tube must be carefully masked in order to obtain reproducible results and avoid scattered light. During a run, all pertinent data are stored on magnetic tape using an Ampex SP-500 PM Direct Recorder/Reproducer. Four channels are used as shown in Figure 6. The output of the log circuit, the synch signal from the monochromator, a triggering pulse, and an audio signal are each recorded on a separate channel and at least the first three by the same head. 44 .Emummm cmomloflmmn map pom Hmumfiouonmouuommm we Hmooz Sumo m Sue? cmxmu mumcmmcmaumm Edfimmmuom mo Esuuuwmm m mo conflummfiou .m musmflm sumavus; on 8 o. o q A _ a ' Q. nué O, O 0 V 3 a m LQAVH B V C O a N O 3 O n O o . A .15 08-03 580 3 .38 so QDH3 omcflmuno uzmflmc pom coeuflmom O .Umm\mcmom on us Emuwwm cmomloflmmu QDH3 owcflmuno Enuuommm III. .18 mmm um mumcmmcmEumm Edemmmuom How 3mg m.nmmm mo puma .m musmflm 0.52: .0200 .mmd ¢ N 45 q A .uflss mucmnuomnm mmo.o H." o .Emumhw cmomloflmmn on» SDHB 095D mumHHHmmo EE O.H m QH new .MN Mu .uflss mocmnuomnm omo.o H." o .Emumhm cmomloemmu map LDHB HHmU EU 020 m CH HQDMB .MN. 0 .DHCD mucmnHOQO >mo.o H." o .umumEouocmouuommm ea humu QDHB HHmU EU mco m CH HmDMB .MM. 4 BONVGHOSBV 46 A cathode follower circuit must be used when recording the synch signal, if an undistorted signal is to be reproduced. If desired, the data may be monitored on an oscilloscope during recording. The SP-500 in the FM mode has a frequency response of 0-2500 cps with noise level -55 db from one volt rms at 15 ips. At 7% ips, the noise level is ~58 db, but the frequency response is halved. The recorded information is played back into a Tektronix Type 564 Storage Oscilloscope with types 2A65 Differential Amplifier and 5B4 Time Base plug-in units, and photographed. The channel containing the signal from the log circuit is played into the positive side of the differen- tial amplifier, and a blank channel from the same head is played into the negative side to permit common-mode rejection of noise. The time base is triggered externally in the normal mode by the channel containing the synch signal from the monochromator. The 5B4 Time Base is equipped with a magnifier which permits selection of fractions of the dis- play up to 1/50. In this way, every scan need not be examined. Examples of such spectra changing with time are shown in Figures 12a and b. The time between sweeps may be determined from the scanning rate of the monochromator. It should be pointed out that these spectra are sinusoidal rather than linear in wavenumber about the center of scan. 47 If a display of absorbance as a function of time at a given wavelength is desired, it may be obtained in either of two ways. From the data already stored on tape, the pulse may be used to trigger a slow sweep of the oscilloscope, such that the decay of the peak absorbance is displayed. This, as shown in Figure 12c, gives an envelope of absorbance with time at the wavelength corresponding to the highest spectral peak. With the second method it is necessary during a run to select the center of scan, minimize the wavelength interval scanned, and stop the motor. Then the output of the log circuit and triggering pulse are recorded, and the latter will trigger the scope during playback. This method is equivalent to that described for the fixed wavelength system (Sect. III-F). The transparencies were magnified using a photographic enlarger and traced on graph paper. IV . RESULTS A. Introduction The reactions of all the alkali metals with water, cesium and rubidium with methanol, cesium with ethanol, and cesium with hydrogen chloride and ammonium bromide in ethylenediamine have been examined at fixed wavelength. Two runs were made of the reaction of cesium with water and one run of lithium with water in EDA using the rapid-scan sys- tem. These reactions were pseudo-first-order in metal, with solute concentration ten or more times greater than metal concentration. The overall rate constants and their order in solute were obtained by varying the solute concentration. B. Alkali Metals + Water 1. At fixed wavelengths. Cesium. The reaction of cesium with water in EDA was studied over a wide range of water concentrations, using several metal concentrations, and the results are given in Table III. Two rate constants were observed corresponding to the major absorbance changes, while in two runs at low water concentrations, a small initial absorbance change occurred with a faster rate constant varying from 58 to 48 49 Rate Constants for the Reaction of Cesium with Water in Ethylenediamine.* Table III . Run 0.1 (Aug. 10, 1965). At room temp. Cesium conc.««/ 2.5 x 10"4r M (from absorbance). k (M-lseC"1 Water Conc. (M) A(mu) ‘fast intermediate Slow 0.541 950 -- 21.5, 20.5 5.14.6.46,5.65 0.0244 750 86, 200 -- -- 0.0244 950 78 -- -- 0.0244 1100 59.2, 100, 27.0, 24.9 -- '167 Run 0.2 (Aug. 12,-1965). Apparatus broken on this run. Run 0.5 (Aug. 16, 1965). At room temp. Cesium conc. = 1.4 x 10"4 M (by hydrogen analysis). k (M‘lsec'l) Water Conc. (M) 1(mu) intermediate slow 6.11 750 18.5, 24.9 5.2, 5.5 6.11 1100 24.1 6.2 0.541 950 21.2 4.1 0.541 1100 -- 1,75, 5.9, 1.76 0.0244 750 10.8 6.4 0.0244 950 10.5 6.26 0.0244 1100 -- 8.4, 5.9 4.0 with 750 26, 26 11, 11 4.0 CSOH 950 22, 20 6.9, 6.6 4.0 present 1100 14, 14, 51 11 Run 1.1 (Nov. 6, 1965). Water Conc. (M) 5.60 5.60 5.50 0.670 0.670 0.670 At room temp. Cesium conc. 4.1 x 10‘4 M (by hydrogen analysis). k (M‘lsec‘l) 1(mu) fast intermediate slow 1100 -- 28.5 10.7 950 -- 28.0 10.5 750 -- 28.6, 25.5 5.40, 5.45 1150 -- 18.8 6.02, 2.91 950 -- 10.1 2.91 750 -- 6.79 2.42 * All concentrations are after mixing. continued *- Table III - Continued 50 Run 1.1 - Continued k (M'lsec‘l) Water Conc. (M) A(mu) fast intermediate slow 0.0250 1100 56, 105 — 14.5,-15.5, “ 16.5 0.0250 950 82.5, 58.4 16.5, 15.0, -- 15.7 0.0250 750 -- ‘17.1 -- Run 6.5 (Mar. 50, 1965). At 25.0.: 0.20 C. Cesium conc. = 5.8 x 10‘4 M (by hydrogen analysis). k (M‘lsec'l) Water Conc. (M) A(mfl) intermediate slow 1.78? 1100 6.8, 5.24 0.46, 1.15, 1.05, 1.41, 1.20, 1.22 0.928? 1100 4.58, 5.74 1.58, 0.54, 0.59, 0.72,A1.75 0.0728 1100 50.9, 16.7 4.29, 2.15, 2.64, 5.60 *- All concentrations are after mixing. T These solutions stood four months before use . 51 200 M-lsec‘l. The intermediate and slow reactions are firsts order in both metal and water, and show rates of 19.6.: 4.9 and 5.5.: 1.9 M'lsec-l, respectively. Unless Specified otherwise, limits of error are given as average deviations from the mean. It appears, since only reducing species are present in the metal solutions, that these latter reactions are both simultaneous and competitive, and they were so treated. Figure 10a shows a typical oscilloscope trace and Figure 10b shows a plot of log Absorbance Kg. time for this trace. No definite correlations of these reactions with wave- length could be made since the R and IR bands of the metal solution are extremely broad and overlap considerably (50). Since there are two competitive reactions, it would be ex- pected that the initial absorbances would correlate with the appearance of the intermediate rate. At a given wavelength, for many traces that showed two rates, the initial absorb- anCes had about the same values. In other traces at the same wavelength when there was only one rate, initial absorbances were about those corresponding to only the intermediate or the slow rate, depending upon which one was observed. Unfortunately these correlations did not hold in a number of cases, which is probably due to partial decomposition of the cesium solutions in the syringe prior to pushing, and result~ ing shifts in equilibria (52). Rubidium and potassium. Rubidium and potassium solu- tions were studied using a single run for each metal and the ABSORBANCE L06" 52 Time (20 msec/division) Figure 10a. Typical oscilloscope trace for the reaction of cesium with water in ethylenediamine showing two rates. l 5 I I I 0 O-Ol 002 0-05 004 TIME (SECONDS) Figure 10b. Plot of log Absorbance X§~ time for this trace. 55 results are summarized in Table IV. These results are similar to those using cesium. The reaction of rubidium with water gives rateconstants of 157, 12.6 i 2.8, and 5.5.: 1.0 M‘lsec‘l, and potassium + water yields rates of 14.8.: 2.7 and 5.9.: 0.9 M-lsec-l. Sodium. Three runs were made with solutions of sodium in EDA and the results are presented in Table V. The reaction of sodium with water in EDA is also first-order in metal ab- sorbance, and except in a few instances, shows but a single rate. A plot of log K (where K is the pseudo-first-order rate constant) yg. log (water concentration) is shown in Figure 11. The solid line was fitted to the data by the method of least squares using a cubic equation. This line gave the best fit for polynomials of degree one through four. The single point obtained by Dewald (11) is also shown, and is in agreement with these results. Lithium. The reaction of lithium with water was studied in three runs and the results are displayed in Table VI. In Run 2.2, lithium prepared by evaporation of a metal-ammonia solution was used, while in the other runs, lithium in a side- arm was used (Sect. III-B). The absorbance at 700 mu was very low relative to that at 1100 mu (for Run 2.2) which would lead to the assignment of the faster rates to the IR species. Run 6.2, however, consisted of a very dilute solu- tion showing no IR band, and resulted in a Single rate 54 Rate Constants for the Reactions of Rubidium and Potassium with Water in Ethylenediamine at Room Temperature.* Table IV. Run 1.2 (Nov. 9, 1965) Rubidium conc. 1.54 x 10"3 M(by hydro~ gen analysis). k(M‘lsec'l) Water Conc. (M) 1(mg) fast intermediate slow 5.60 750 -- '11.1 5.84 5.60 1100 -- 4.25 1.16 0.580 750 -- 10.9 -- 0.580 1100 -- -- 4.96.5.52 .0.0250 750 182, 89.6 14.2 ~- 0.0250 900 199 17.9 -- 0.0250 >1100 -- 8.91 2.87 Run 5.1 (May 21, 1964). Potassium conc. = 1.15 x 10"4 M (by hydrogen analysis). k(M‘lsec‘l) Water Conc. (M) A(mu) intermediate slow 4.06 700 17.2, 17.2 4.82, 4.82 4.06 850 ' 16.7 4.18 4.06 950 17.7,-14.6, 16.7 4.18, 4.19, 4.44, 12.5 4.52 0.590 700 -- 5.05 0.590 800 15.5, 20.9 5.79, 4.54, 5.20 0.590 1110 11.7, 10.7, 11.5, 1.89, 2.44, 2.44. 15.0, 10.7 6.50 *- All concentrations are after mixing. 55 Table V. Rate Constants for the Reaction of Sodium with Water in Ethylenediamine. Run 2.1 (Mar. 7, 1964). At room temp. Sodium conc. = 5.5 x 10‘4 M (by hydrogen analysis). Water Conc. (M) 1(mg) K(seC'1) 4.07 , 750 6.54, 8.80, 8.15, 8.20, 7.15, 8.40 4.07 850 6.25 2.02 750 1.62, 1.16, 1.08, 1.08 0.589 700 0.556 0.589 750 0.212, 0.224 0.024 700 0.055 Run 5.2 (Oct. 26, 1964). At 25.0.: 0.20 C. Sodium conc. A’1 x 10‘4 M (from absorbance). K(SeC'l) Water Conc . (M) M199) 1* II 5.01 750 10.7, 21.6 ~2.04, 6.52, 2.40, 1.70, 1.54, 5.55 0.295 750 -- 0.0956, 0.0779 0.0572 750 -- 0.0647, 0.0666 Run 6.1 (Mar. 11, 1965). At 25.0.: 0.20 C. Sodium conc. = 1.4 X 10"4 M (by hydrogen analysis). T K(Sec-l) Water Conc. (M) A(mg) I II 1.78 700 15.2 2.56, 2.00, 1.40, 2.40, 1.00 0.850 _ 700 -- 1.00, 0.550, 0.520, 1.15 0.248 700 1.54 0.292, 0.455, 0.052, 0.555 0.0728 700 1.25, 1.22 0.245, 0.117, 0.117 0.0728 750 1.48, 1.56 0.27, 0.20 *- All concentrations are after mixing. fife seCt o v-E o 56 .wcflfimflpmcmawsum CH HmDMB SDHB EnepOm mo sofluummu mnu How Acoflumuucmucoo umuMSV 00H .WN y 004 .aa Guzman 0a one , Aonv awo I 1___:_ i L ____:Lo «.8 I 00.0 .I 1 0- j H _II. Amomuu MO .8553 mom + mle 1. <0m+mZIO l l A: £0.05 Omm + MZIO 1 I 37.83 9.23 0mm + mz IO 1 H 1 I:__Lh _ :_[__ _ :Ler__ _II|.... H.O 0.6 ca 57‘ Table VI. Rate Constants for the Reaction of Lithium with Water in Ethylenediamine.* Run 2.2 (Mar. 12, 1964). At room temp. Lithium conc."V 2 x 10"4 M (from absorbance). k(M‘lsec‘1) Water Conc. (M) A(mu) fast intermediate slow 0.0258 700 -- -- 9.07, 9.07 0.0258 1100 91.6, 69.4, 50.5, 50.5, 7.26.110.9, 109, 102, 50.0, 19.2, 10.2 71.5 50.5 Run 5.5 (Nov. 5, 1954). At 25.0.: 0.20 c. Lithium conc. «I 1 x 10"3 (from absorbance). k(M'lsec‘l) Water Conc. (M) A(mu) fast intermediate Slow 5.01 720 -- 50.4, 14.1 4.25, 5.79, 5.65 5.01 1100 -- 10.5, 16.6 2.81, 5.04, 5.52, 2.89 1.59 1100 -- 25.8, 28.2 8.65, 2.50, 5.07 0.295 720 94.9, 112 -- '1.42 0.295 1100 92.2 -- 1.42, 1.08 0.0572 715 195, 254 58.4, 26.9 5.86, 5.50, 5,16 0.0572 1100 91.1, 64.5, 40.5, 61.1 4.50, 4.22, 215 7.55, 8.71 4.50 Run 6.2 (Mar. 22, 1965). At 25.0.: 0.20 C. Lithium conc. 0’ 10‘5 M (from absorbance). Water Conc. (M) A(mg) l<(M‘lseC“l) 0.0728 700 52.4 0.0728 650 57.8, 25.5, 19.2, 55.7 * All concentrations are after mixing. 58 constant of 29.7.: 5.9 M'lsec‘l obtained by following the decay of the 660 mu absorbance. The third run, Run 5.5, yielded up to three rate constants with values of 140, 27.5 .+_ 5.9, and 4.5 1 1.5 M“1sec-l. 2. With rapid-scan system. .Cesium. Three runs were made using the rapid-scan system--two with cesium + water and one with lithium + water. The results of these runs are given in Table VII. For Run 7.1, a dilute solution of cesium in EDA was prepared. Picture no. 4, obtained from playback of the tape recorder, is shown in Figure 12a and Figure 15 gives plots of ln Absorbance gs, time for this trace made at different wavelengths. Agreement among rate constants for a given trace is reasonable, and the differences in values between traces probably result from an inability to correct the rates for the presence of the second slower rate. Analysis of such traces was made more difficult by the lack of triggering stability, and by the fact that the relative errors increase near the bottom of the trace. Also to be considered is the decreased sensitivity of the spectro- meter for large spectral scans through a one mm capillary tube, since the photomultipliers were operating in a region of poor sensitivity at low light intensity. Again, decomposition and shifts in equilibria may play an important role. The upper limit of the scanning rate, 6.7 msec/scan, made detection of the intermediate rate difficult at high water concentrations (Té~el9 msec). The presence of the intermediate rate resulted Table VII. 59 Rate Constants for the Reactions of Cesium with Water and Lithium wigh Water using the Rapid-Scan System at 25.0.: 0.2 C.* Run 7.1 (July 21, 1965). Pict. No. Water Conc. 0.0569 .0569 .580 .580 .580 .580 .1.75 Cesium conc. = 4.0 x 10‘4 M (by hydrogen analysis). rvA(mu) - k (M‘lsec’l) 1150-14.9: 1080-12.9; 900-15.5? 650-14.8 1150~9.22; 1060-9.90; 1100-25.6711050-21.27 1150-14.5: 1050~11.7: 850-8.45 1100-10.4: 900*18.7 1100-10.97 1050-15.0; 900-11.2 1125-20.2;-1050-24.9; 900-18.6 1100-2.28; 1000-1.867 800-5.17, 1.75 1050-12.0, 2.92 1100-2.287'1050-2.60: 975-2.12 1100-1.55; 950-1.40 1100-4.10, 0.957 950-5.20,.1.06 (M) 900-7.50 950-20.4 Run 7.2 (July 27, 1965). Cesium conc. ”V'2 x 10““3 M (from absorbance). Pict. No. Water Conc. (M) A’A(mu) — k (M~lseC'l) 8 left 0.0569 1150~4.947 1050~24.9, 5.20 8 right 0.0569 1150-26.1, 5.267 1050-9.467 850-4.62 9 0.0569 1150-6.26: 1050~5.98: 850-6.25 10 0.0569 1200~4.167 1050-2.87: 875¢5.15 11 0.580 1100-12.8, 2.27; 900~2.98 12 0.580 1150-5.057 1050-4.56; 850-5.96 15 0.580 1150-4.08; 1075-5.18: 1000*4.68 14 left 0.580 1100-5.427 900-5.11 14 right 0.580 1100~12.6, 2.057 900~5.27 .15 5.18 1050-15.1 *- All concentrations are after mixing. continued 60 Table VII - Continued* — ___. f Run 7.5 (July 50, 1965). 'Lithium conc. -“’4 x 10"3 M (from absorbance). Pict. No. Water Conc. (M) odk(mu) - k (M‘lsec‘l) 1 0.0569 1200-52.1, 8.257 1050-27.4, 7.41 2 0.0569 1200-15.5; 1050-51.5, 8.62 .5 0.0569 1200-40.7, 10.57 1000-52.0, 8.18 4 0.0569 1200-74.6, 18.47 1050-90.9, 15.67 850-59.7 .5 0.0569 1050-58.8, 9.21 7 0.580 1150-85.87 750-75.0 8 0.580 1050-91.1 11 0.580 1050-111, 57.5 15 0.580 1050-154, 62.2 15 5.18 1050-fast rate, 11.8, 5.99 3(- All concentrations are after mixing. 61 Abs. 0.212 sec/scan Figure 12a. Typical trace for the reaction of cesium with water taken with the rapid—scan system. Abs. 0.55 sec/scan I N J I: Figure 12b. Typical trace for the reaction of lithium with water taken with the rapid-scan system. Abs. Time (20 msec/division) Figure 12c. Slow oscilloscope sweep for the reaction of lithium with water taken with the rapid-scan system. 1.0 1.0 + const. ln Abs. 1.0 0.5 62 I 1 I I I I I I I I I l I I T ._ .J _ Run 7.1 No. 4 0 r— at 75 scans/sec —~ — Center of scan = _ _ 12,480 cm"l _ NutatiOn = 10,000 P O .7VV1150 mu ‘ I o : k fl "" "I “ 1 : b. M4050 mu : 1 _ L I "' "I — C. )\~900 mu _: L _ I : — O c-I-I )— .— _ o . )—- --I I I I I I I I I .1 I l 1 I I I Time (1 division = 0.212 sec) Figure 15. Plots of 1n Absorbance gs. time for Fig. 12a. 65 in a large decrease in absorbance between mixing and the first scan after flow had been stopped. The results of Run 7.2, in which a more concentrated cesium solution was em- ployed, were similar to those of 7.1. Analysis of Spectral shape for these runs will be discussed in Chapter V. »Lithium. A run with lithium, prepared in the dry box, was carried out using the rapid-scan system. Figure 12b shows the trace for picture no. 1 of Run 7.5, and plots of ln Absorbance yg, time for this trace are presented in Figure 14. Figure 12c shows the decay of peak absorbance with time for picture no. 15. The oscilloscope sweep rate is slow enough to give a number of complete spectra in a single sweep re— sulting in an envelope whose appearance is similar to the usual decay curves at fixed wavelength. A plot of ln Absorb~ ance yg, time for this trace is presented in Figure 15. The spectrum of this lithium solution shows primarily the IR band, and yet these kinetic studies apparently yield three distinct rate constants. These results agree with those obtained earlier at fixed wavelength. C. Rubidium and Cesium + Methanol The reactions of rubidium and cesium with methanol in ethylenediamine were studied at fixed wavelengths, and the results are presented in Table VIII. While reaction times are comparable to those observed in the water reactions, several marked differences are apparent. In some cases, + const. ln Abs. 1.0 l C) 01 0.5 64 _ J Run 7.5 No. 1 5 7 at 75 scans/sec _ Center of scan = 11,780 cm-l — Nutation = 10,000 cm-l - \ a. I~1200 mu q \. _ \ _-‘ - n I. 1 _ fl 0 b . )x’V1050 mu ' l L J I I I Time (1 division = 0.55 sec) Figure 14. Plots of 1n Absorbance gs, time for Fig. 12b. 65 .UNH .mflm now mEHD .mN.mosmnuomn< CH mo uon .ma mssmem Aomm mmao.0 n sonH>Ho av mafia l I ITF7 is omoe.< A I AIEU 000.0a u sofiumusz HWEU 00>.Hdnucmom mo kusmo omm\mcmom 00a us we .02 m.> cam l 1" l O 'sqv UI 'qsuoo + 66 Table VIII. Rate Constants for the Reaction of Rubidium and Cesium with Methanol in Ethylenediamine.* Run 1.1 (Nov. 6, 1965). At room temp. Cesium conc. = 4.1 x 10’4 M (by hydrogen analysis). k (M'lsec-l) Methanol Conc.(M) h(mu) II III IV 0.767 750 60.0 11.6 5.50 0.767 950 -— 10.4, 10.7 4.26, 6.78, 7.58, 7.20, 4.90, 5.95, 5.90 Run 1.2 (Nov. 9, 1965). At room temp. Rubidium conc. = 1.5 x 10‘3 M (by hydrogen analysis). Methanol Conc.(M) 1(mu) k (M‘lseC’l) 0.767 750 6.05, 7.71 0.767 ' 1100 9.84, 1.56 Run 4.2 (Sept. 21, 1964). At 25.0.: 0.20 C. Cesium conc. = 5.1 x 10"4 M (by hydrogen analysis). k (M‘lsec-l) Methanol Conc.(M) 1(mu) I II III IV 0.105 950 510, 475 70.5 19.4, 12.2, —— 419 29.1, 15.2, 14.5 0.556 ‘ 950 -—- 50.5 15.7, 10.5, 4.95 51.5 12.4 74.6 1.95 950 192, 214, -- -- ~- 150 Run 5.1 (Sept. 25, 1954). At 25.0 _-t 0.20 c. Cesium conc. = 4.1 x 10‘4 M (by hydrogen analysis). k (M‘isec'l) Methanol Conc.(M) 1(mg) I II III IV 0.105 850 581, 168 -- ”12.0, 14.7 -- 0.105 950 450 67.7, 61.2 15.9 -~ 0.105 1100 229, 100 -- 16.2, 24.2, -- 22.5, 19, 14.5 0.555 1100 -- 81.6, 86.1, 15.6, 15.0, 8.50.7.08, 61.2, 65.7 10.6 5.88.4.59 0.551(1 trace) 850 -- 74.6 20.0 7.89 0.556 1100 -- -- 9.55 5.49,7.01, 5.88 * All concentrations are after mixing. 67 using cesium solutions, more than one rate was observed in a given trace, but the constants were not reproducible.) In Table VIII, the second-order rate constants obtained have been arranged into four groups depending upon magnitude. More striking was the fact that the shapes of the traces themselves differed from picture to picture, and in some cases showed what seems to be an absorbing intermediate. Such a trace is displayed in Figure 16a and a plot of In Absorbance ye. time for this trace is given in Figure 17a. D. Cesium + Ethanol Two runs were made with solutions of cesium and ethanol in ethylenediamine. The results of these runs are compiled in Table IX. Most traces were similar in appearance to those obtained with water (see Fig. 16b), but some pictures, Run 4.1 nos. 9-15, showed behavior like that observed with methanol (see Fig. 16c). Plots of ln Absorbance gs. time for the traces in Figures 16b and c are given in Figures 17b and c, respectively. The reaction is clearly first—order in metal absorbance, but no definite order in ethanol can be assigned. Competitive processes and/or reaction inter- mediates could account for this difficulty. E. Cesium + Ethylenediammonium and Ammonium Ions Hydrogen chloride was reacted with cesium in EDA and several traces were obtained. The results of these pushes 68 Time (0.2 sec/division) Figure 16a. Typical trace for the reaction of cesium with methanol in ethylenediamine. Time (0.1 sec/division) Figure 16b. Typical trace for the reaction of cesium with ethanol in ethylenediamine (no intermediate). I T T I I . . . I, . .,. ‘T" I‘I‘I‘I’ 'I'rfl Tr T'T—T’T r-I—r—r—H— 1 "'I '[ I 1‘?— |I Time (0.2 sec/division) Figure 16c. Typical trace for the reaction of cesium with ethanol in ethylenediamine (intermediate present). 1n Absorbance- 69 a. Cesium + methanol in EDA ' A = 1100 mu b. Cesium + ethanol in EDA (no intermediate) 950 mu A Cesium + ethanol in EDA _ ‘ (intermediate present. 950 mu C. Time (1 division 0.1 sec) Figure 17. Plots of ln Absorbance gs. time for the traces in Fig. 16. 70 Rate Constants for the Reaction of Cesium with Ethanol in Ethylenediamine.* Table IX. Run 5-2 (June 9, 1964). At room temp. Cesium conc. = 4 x 10‘4 M (from absorbance). Trace No. 1(mg) Ethanol Conc. (M) K(sec~1) 5 900 0.117 5.10, 1.60 4 900 0.117 5.10, 1.60 6 1100 0.117 2.70, 1.75 7 700 0.117 1.75 8 950 0.117 1.90 9 950 1.04 5.90 10 950 1.04 1.50 15 950 1.04 5.10 14 950 1.04 5.00 16 950 0.504 4.50 17 950 0.504 2.80 18 950 0.504 2.90 19 950 0.504 2.20 21 950 0.504 5.20 22 950 0.504 2.00 25 950 1.61 4.50 24 950 1.61 8.50 27 950 1.61 4.10 At 25.0 :_0.20 C. Cesium conc. = 1.8 x 10'3 M (by hydrogen analysis) and A = 950 mu. Run 4.1 (Aug. 12, 1964). Trace No. Ethanol Conc. (M) K(S§C‘l) 5 1.04 15.7 6 1.04 5.95 8 1.04 5.50 9 0.521 1.94 10 0.521 1.67 11 0.521 1.44 12 0.521 1.57 15 0.521 1.70, 1.50 14 0.521 1.48, 1.14 15 0.521 2.00, 0.527 16 1.61 58.2, 17.6 17 1.61 27.5, 14.5 18 1.61 24.1 19 1.61 20.0 20 1.61 25.6 21 0.557 1.47 22 0.557 1.50 25 0.557 1.50 24 0.557 7.50, 1.57 25 0.557 2.65 *- All concentrations are after mixing. 71 are given in Table X. A rate constant of 1.7.: 0.2 x 105 M'lsec-l was observed. The reaction of ammonium bromide with cesium in EDA was too fast to be followed by the flow method for the two concentrations employed, but a lower limit of 5 x 106 M'lseC“l could be calculated for the rate constant of this reaction. 72 Table X. Rate Constants for the Reactions of Cesium with Ethylenediammonium Chloride and Ammonium Bromide in Ethylenediamine.* 1.. Run 5.1 (Sept. 25, 1964). At 25.0 :.0.20 C. Cesium conc. = 4.1 x 10‘4 M (by hydrogen analysis). + 1 1 EDAH Conc. (M) l(mg) k (M- sec- ) 1.0 X 10':3 1100 (1.7, 2.0, 1.5) X 105 Run 6.5 (Mar. 50, 1965). At 25.0.: 0.20 C. Cesium conc. = 5.8 x 10'4 M (by hydrogen analysis). NHA.+ Conc. (M) 4(mg) Tfi-(msec) k (M‘lseg‘l) 5.5 x 10'4 1100 < 2 > 8 X 10’ 5 x 10‘5 1100 < 2 > 5 x 106 * All concentrations are after mixing. V . DISCUSSION A. Introduction The reaction of dilute solutions of cesium with water in ethylenediamine was studied by Dewald gt _1, (11,12) who obtained a rate constant of 24.7 :.1.5 M-lsec”1. On the basis of Optical absorption spectra (50) and other evidence (51,52), it was assumed that this reaction was that of the solvated electron as well as loosely-bound electrostatic aggregates of electrons and cations with water. This per- mitted correlation with the results of aqueous radiation chemistry. Since solute concentration was in excess, all reactions studied in this investigation appeared to be pseudo-first- order in metal absorbance, so that -d(Abs.)/dt = K (1) then (Abs.) = (Abs.)oexp(-Kt) (2) or 1n(Abs.) = -Kt + 1n(Abs.)O (5) In these equations, K: k(R)n, where k is the overall rate constant and n is the order with respect to the reactant in excess. A plot of log K gs, log (R) will give both k and n. 75 74 For a large number of traces, a plot of ln (Abs.) ye. time yielded two straight line portions with different lepes. Such cases were treated as simultaneous, competitive reactions in the following manner: Letting A J51)- products (4) and B fli), products (5) then A = A0 exp (-Kat) (6) and B = Bo exp (-Kat) (7) so ln (A+B) = ln[AO exp(-Kat) + BO exp(-Kbt)] (8) Next, dln(A+B) = -KaAQeXp(-Kfit) - KaneXp(-Kbt) (9) dt Aoexp(-Kat) + BOexp(—Kbt) At t = 0, dlné:+B) = 353:: 1 £230 5 _Kgpp (10) which simplifies to true = app + .EQ app _ Ka Ka A0 (Ka Kb) (11) where Kipp is the observed higher rate constant, Kb is the lower rate constant, A0 and B0 are the initial absorbances. true of species A and B, respectively, and Ka is the corrected value of Ka. In this derivation it has been assumed that Beer's Law is valid for Species A and B, noting that even if 75 the extinction coefficients are different, the final ex- pression is unchanged. For the most part, the traces obtained with the fixed- wavelength stopped-flow apparatus were of excellent quality. Good first-order kinetic plots were obtained from them. Often the traces obtained in subsequent pushes under the same conditions could be superimposed exactly. In spite of this, over the course of a run there was an abnormal amount of scatter in the rate constants obtained. Analysis in terms of two reactions, as described above, improved the precision significantly, but the deviations were still much larger than the expected instrumental error. The difficulty probably lies with the solutions themselves since they are somewhat unstable and very sensitive to impurities which could lead to partial decomposition and shifts in the various equilibria. B. Alkali Metals + Water 1. Cesium, rubidium, and potassium. From the results at fixed wavelengths, for the reaction of cesium with water in EDA, it was felt that the second-order rate constant of about 20 M‘lsec‘l was compatible with the previous assign- ment to the solvated electron (and loose aggregates). If this were the case, then the slower reaction might be assigned to the reaction of the R band species. Figures 18 and 19 show plots D/D1100 (D = absorbance)_y§. wavelength 76 T I I I I I l 1.1 — /5;,51131 ///5>' C] C] /O C] _ J 1.0 o D Pdll / o D / ... D [,4 _ / 7 D/D 1100 0.8 __ D O// _ / O // 0.7 — / 0 - Pict. No. 4 _ D / / CJ- Pict. No. 12 O / O / ---- IR band from 0.6 — / Ref. 50 ~ / Spectra were normalized at 1100 mu. I I I I ' I I I 700 800 900 . 1000 1100 1200 1500 Wavelength (mu) Figure 18. Comparison of the spectral shapes for typical ’traces of Run 7.1 with that of Dewald and Dye (50). D431100 77 I I I I I T I O O 1A.P j>r’—‘— ,6 / 1.0 - )jfif _ O/ / 0 ./ )— / — 0.9 O / / / O ,/ 008 '— / _. O / / O / o_7__ // _O -Pict. No. 10 _ o / ---- IR band from / Ref. 50 / / 0.6- / Spectra were normalized — / at 1100 mu 4 I I I I I I 700 800 900 1000 1100 1200 1500 Wavelength (mu) Figure 19. Comparison of the spectral shape of a typical trace of Run 7.2 with that of Dewald and Dye (50) . 78 for typical traces from Runs 7.1 and 7.2, respectively, compared to the spectral results of Dewald and Dye (50). The shape of the rapid-scan spectrum in Figure 18, corres— ponding to Run 7.1, picture no. 4, is close to that of Dewald and Dye with some deviation at lowerwavelengths, while the spectral shape for picture no. 12 is quite dif- ferent. Kinetic analysis of picture no. 4 yields only the intermediate rate, and the shape for picture no. 12 (at high water concentration) was obtained after several scans in order to remove any IR band that might be present. These results are in accord with earlier assignments. In Run 7.2, as shown in Figure 19, the shape deviates from that of the IR band, and shows the presence of the R band. This spectrum is similar to that obtained from picture no. 12 in Run 7.1, which is the Spectrum after the intermediate reaction is finished. Since a higher cesium concentration was employed here, the appearance of the R band and the observed kinetic results (namely, the rate of 5 M'lsec‘l) is not surprising. The rapid decrease in absorbance observed at the beginning of some traces (k = 40 to 200 M‘lsec-l) might be due to reaction with impurities. This would require occasional impurity concentrations in excess of 10‘4 molar, however (9;; lithium discussion below). Further study is necessary to determine the origin of this process. Blank determinations of the rate of absorbance change upon mixing 79 with pure solvent indicated auto-decomposition to be much slower than the rate of reaction with the lowest concen- tration of water used. This too, indicates that the re— action with impurities is unimportant, although a catalytic effect could still be present. While only one run each was made with rubidium and potassium with water in EDA, the results were quite similar to those of the cesium-water system. Although runs with the rapid-scan system would be needed before any definite conclusions can be drawn, the reaction of the IR band species with water may be tentatively assigned to the rate, ke’14 M‘lsec'l. Since both rubidium and potassium solutions in EDA possess the V and R bands, as wellIas the IR hand, no assignment can be attempted for the slower rate, kna5.5 M'lsec'l. The presence of a fast rate in a few rubidium- water traces should also be noted. Pulse-radiolysis studies on the disappearance of the hydrated electron through reaction with solvent gives a second-order rate constant of 16.: 1 M-lsecrl (10). This is in good agreement with the value 19.6.: 4.9 M’lsec-l obtained for the reaction of cesium with water in EDA, despite the differences between the two solvent systems. Radiation chemists have shown (77,78) that the disappearance of the hydrated electron is a reversible reaction which can be written ' _ k .— # eaq + H20 k2 H + 0Haq (12) 80 If the same reactions are assumed to take place in the ethylenediamine solvent system, and if H atoms are removed by H+H —k3—>- H2 (15) then, _ 943:1 = k1[e_][H20] - kZIOH‘IIHI <14) and 駧l = k1[e-][H20] - k2I0H‘IIHI - k3[H]2 (15) Assuming that 94%1 = 0, and if [e'] = [e’]O - x and [OH-] = x, Equation 14 may be rewritten as -§§ - k [H 0] (Ie‘I - ) - k [H] x (16) dt _ l 2 o X 2 8.8. Integration of this equation gives k1[H201[§-10 - = 1“ kilH20][e’lo-(ki[H20]+k2[HlXOH‘] ‘k1[H20] + k2[H1)t (17) This equation may be rearranged to yield lnte‘] = 1n IIe-Io - fi:{fi:8§[i’i:[fl]I1-exp(k.IH.01+k2IHIItII (18) 'Figure 20 shows a comparison of calculated values with the experimental graph of log(Abs.) ye. time using a rate constant of 2.2 x 107 M"'J‘sec"l for kg, and k3 = 2 x 1010 M'lsec’l, the values in water. A more favorable case is shown in Figure 21 for an experimental graph that shows a break. Also shown here are the calculated points using a 81 .ONE + I®_Ar.nmo + m coauummu m£u_mcflumpflmcoo UmudeUHmo mmDHm> £UH3 05H» {MN A.mQ¢v mod mo namum Hmucmfiflummxm cm mo consummEoo .ON musmflm Amocouwwv wEHB o.m O.N O.H . _ _ .m.m 2 @10H x >>.a n Hma Hnommalz N.oa x m.m u mxv mmsHm> .UHMU flu T . . .nm.fi| 0mm o.a u u um 2 muoa x m.m u m mfim. snowman: 50H x N.m n axa mwsam> .UHMU nu .IN.HI .l HIEUHIZ .vOH X N .HlUmm “VQMoO H w .2 mNO.o H ”Ommu 2 mafia HmucmEHummxm Do fiHOHI m.OI m.OI eoueqxosqv 501 .mmumu OBu mcHBOLm momma m How Om IH.Im.AI.IdO + m cofluommu mLu mcHHmUHmcoo UmumHsuHmU modam> £ua3 mafia .m> A.mQ .UHmU .mhwumm No.0 u u um Ebroa x m>.N n ”ma fiwummulz boa x N.N n Nxv mmsHm> .Uamo z om.m n Hows“ .anomm em n s tmcfla Housmaflummxm @ ©.OI w.OI eoueq10sqv bot 85 value of k2[H]s.s.= 4.0, a lower value, which decreases the importance of the back reaction in Equation 12. The fit is improved at a few points, but still differs considerably from the observed line. Surprisingly, this treatment gives the lower value of [H]S.s.of 1.2 x 10"7 M (as compared to the initial value of 2.76 x 10'7 M for k2 = 2.2 x 107 M‘lsec'l), but increases kg to 5.5 x 107 M‘lsec'l. Use of this latter value in Figure 20 actually worsens agreement, since k2[H]s.s goes from 4.84 to 5.84 here. An examination of Equation 17 shows that the calculated points can only agree with experiment if the reverse reaction of Equation 12 is unimportant in the water—EDA system. This is probably because hydrogen atoms are removed by another process than that of Equation 13. Such a mechanism will be discussed below in Sect. V-F. In addition, no differences were ob- served in the reaction of cesium with water in the presence of cesium hydroxide. This agrees with Dewald's observations upon adding KOH to the water (11), and further supports the absence of back-reaction. When large pulse currents are used in the radiolysis of water, enough e;q is produced so that it disappears by a bimolecular process, eaq + eaq —-> H2 + 20Haq (19) with a second-order rate constant of 1.1 x 101°M‘lsec'l(77). If this is written as 84 Zesolv + 2H20 ——>-H2 + zossolv (20) then — —L—id gt = k [Remake-12 <21) and the fourth-order rate constant for this equation is 5.57 x 106 M'ssec‘l, in pure water. If this reaction were to take place in the water-EDA system with the same rate constant, then at a water concentration of only 0.08 M, the half-life for this reaction would be equal to that of the reaction of eEDA Wlth water, egDA + H20 ——> i H2 + 0H” (22) No evidence for such a reaction was found. 2. Sodium. The reaction of sodium with water in EDA is first-order in metal absorbance, but the order in water is not constant as shown in Figure 11. Early results (93) indicated an order of 3/2 in water, but additional runs at low water concentrations show that the apparent order is less than one at low concentrations and may be as high as two at higher water concentrations. If the V band species were dimeric as proposed by Dye and Dewald (52), then these. results could be explained by considering two competitive reactions + - + -+H20—->-H+M2 +OH (23) M2 -e and + _ - M2 '6 + 2H20 —_"‘ H2 + 2M+ + 20H (24) 85 Reaction 23 would occur predominantly at low water concen- trations and Reaction 24 at high water concentration. In considering this behavior, it must be recalled that sodium solutions show only a single absorption band with a maxi- mum at 650 mu (50) , and exhibit conductance behavior quite different from that of the other alkali metals (51). The nature of this reaction probably will not be understood until the V band species is itself better characterized. 3. Lithium. From the four runs of lithium with water in EDA several conclusions can be drawn, but a number of questions remain to be answered. The reaction of the V band species with water must be assigned a rate off» 30 M""sec"1 on the basis of the results obtained in Run 6.2. It appears, according to Runs 2.2 and 5.3, that the fastest rate is associated with the IR band, although several traces in the latter run (3;, Table VI) do not support this view entirely. since several traces at 700 mu also show this reaction. Since the V and IR bands do overlap significantly (50), the intermediate rate cannot be positively assigned to either band. The concentration dependence of the species present in these solutions, as well as their relative instability, make them difficult to work with. Run 7.3, carried out with the rapid-scan system, appears to yield three rate constants. The shapes of two lithium solution spectra, Figure 22, show that the solution consisted primarily of the IR band, but the presence of another Species is apparent. This is 86 1-2 I I “I I I I O C) 1.1— II z’fi ’ C] \ ,xfi \ A5 1.0—- l \ I / | / / 0.9— ‘ / ‘| D / / D”31100 I 0/ 0.8— ' D / -— | / ‘ / I , . \ / O-Pict. No. 7 0071'— \ / ———I D \ O / [j— Pict. No. 1 \ / \* / ----- Lithium spectrum 0 6 0 from Ref. 50. I _— D —-I Spectra were normalized at 1100 mu b C? I I I I I I 700 800 900 1000 1100 1200 1300 Figure 22. Wavelength (mu) Comparison of the spectral shapes for typical traces of Run 7.3 with that of Dewald and Dye (50). ‘ -—_. 87 probably another IR species, and recalls the scheme pro- posed by Ottolenghi, Bar—Eli, and Linschitz (62) to explain their flash-photolysis results with alkali metal solutions in ethylamine, which requires an additional species absorb- ing in the IR. Several traces from Runs 5.3 and 7.3 showing the faster rates could be fitted to second-order kinetics over I! about one half-life. Examples of such behavior are shown in Figure 23. However, a second-order process could be ruled out by the following argument: I i If d(Abs.) = 2 dt k(Abs.) (25) then 1 _ 1 (Abs.) ' kt + (Abs.)o (26) and _ 1 1n(Abs.) - -ln[kt + TXBETT; ] (27) Differentiating, d 1n(Abs.) = _ k (28) dt [kt + 1/(Abs.)o] and at t = 0. d 1n(Abs.) = _ = _ aPP dt k(AbS.)o — k (29) A graph of this relation for Run 5.3 is given in Figure 24. This shows that the apparent rate fails to correlate with initial absorbance as required for second—order kinetics by Equation 29. It must be pointed out that by the same 88 .mcHEmaUm:0H>£um CH HmuMB nus? Esflzufla mo :Ofluommu on» How mEfiu .MN wosmnuomnd\d mo manm .mm mnsmflm Aommv wash 0..., o.m 04 o _ a _ o I N .02 m.> cam I flu Iim I fl .02 m6. Sm I o I I ..¢ T In I Im I 1m I D 13 I INH _ _ _ aousqlosqv/I 89 100 I I 7 I I I 90_ O - 80 — - 70 — _ 60 - ‘ 50 - - kapp(M’lsec‘h) 40 r C) _ O 50 ” ‘71.Calculated from 1 the average k for Run 5.3 20 - . . 1L C)Exper1mental p01nt from Run 5.3 10 _ 1 C) O O 0 I I I I I I 0 0.1 0.2 0.3 0.4 0.5 0.6 (AbS.)o Figure 24. Graph of kapp gs. (Abs.k)for the reaction of lithium with water in ethylenediamine. 90 type of argument, the unusual amount of scatter present in the first-order rate constants (for the fast rate) could also be used to rule out this assignment. C. Alkali Metals with Methanol and Ethanol Although the detailed nature of the reactions with methanol and ethanol remains unknown, it is clear from these studies that the overall rate of reaction of the solvated electron with these alcohols is not far different from that with water. Pulse radiolysis studies (94) show minimum half— times of 1.5 and 3 usec for the disappearance of the electron peak in the pure alcohols, but this does not necessarily represent reaction of the electron with the alcohol, since no attempt was made to remove the counter-ion or other radiolysis products. While Freeman (95), on the basis of competition kinetics, is inclined to accept these values as the natural lifetime of the electron in the alcohols, Anbar in alcohols will be solv (9) expects that the lifetime of e much longer than in water. The argument of the latter author is based upon the lower stability of alkoxy ions in the alcohols which would lead to a decreased energy of solvation for the reaction e +ROH—>H+0R— solv solv (50) as R goes from H to alkyl. 91 D. Cesium + Ethylenediammonium and Ammonium Ions While the reaction of the solvated electron with hydro— gen ions is near the diffusion-controlled limit in aqueous and alcoholic systems (94), the reaction of hydrogen chloride in our system, which presumably gives ethylenediammonium ions, is much slower, with k‘“’1.7 x 105.M‘lsec‘1. The am- monium ion, however, reacts with cesium with a rate at least an order of magnitude faster than this, indicating that appreciable proton transfer from ammonium ion to ethylenedi- amine does not occur. From the data of Kelly gt _l. (96) on the reaction of sodium with ethanol in liquid ammonia at -33.40 C, Jolly (49) has been able to estimate an upper-limit of 108 M‘lsec-l for the reaction of the ammoniated electron with the ammonium ion. The rate constant for the reaction of e;q with NHI, cited by Rabani (78) from radiation and photochemical studies,'“/ 1.5 x 106 M“lsec‘1, is in reason— able agreement with our lower limit of 3 x 106 M-lsec'l. E. Alkali Metals + Pure Ethylenediamine When alkali metals in EDA were mixed with solvent alone, there occurred, in some cases,a fast increase in absorbance, followed by a slower decrease, which was then followed by the usual slow decay of absorbance with time due to reaction with the solvent or impurities. An example of this behavior is shown in Figure 25a. A similar phenomenon was also noted in several traces for the reaction of metal with water, as 92 Time (0.2 sec/division) Figure 25a. Typical trace for the reaction of cesium with pure ethylenediamine showing initial absorbance buildup. Time (2.0 sec/division) Figure 25b. Typical trace for the reaction of sodium with water in ethylenediamine showing initial fast absorbance decrease and recovery. 93 shown in Figure 25b. Here, however, there was a fast de~ crease in absorbance, followed by a slower increase, which was then followed by the slower reaction with water. That this behavior was independent of water concentration was demonstrated in Run 5.2 where it was observed at different water concentrations, but with about the same time duration for all concentrations (IV’0.5 sec). Gibson (97) has reported a change in absorbance upon mixing two portions of the same solution in a stopped-flow apparatus. This solution was at a different temperature from the mixing chamber (At = 200 C) and the resulting thermal gradients effectively made the solution act as a lens. A warm solution flowing into a cold tube gave rise to an absorbance decrease (negative lens), while a cold solution in a warm tube yielded an absorbance increase (positive lens). The absorbance change upon mixing observed in our sys—' tem is probably not an instrumental artifact of this type since: 1) it was not reproducible; 2) it occurred in cases where no heat of mixing was involved (metal—EDA with EDA), but did not occur where there was significant heat of mixing (metal-EDA with high water-EDA); 3) in all_cases the mixing cell and solutions were close to 250 C; and 4) the changes were fast compared to bulk diffusion times which would be on the order of several seconds. An instrumental error of a different type could give this kind of behavior if seepage of the reactants into the mixing chamber could occur after 94 the flow stopped. It is puzzling, however, that the ob- served results did not occur with greater frequency if this were the case. If this behavior is not due to instrumental error, it is possible that upon dilution there are shifts in the equilibria among the various species, and that they give rise to the observed behavior. It is likely that im- purities or decomposition products also have an influence on this process. F. Mechanisms Since H20 + EDA k EDAH+ + 0H- (51) .c___ k3 with K < 10‘12 (98), it might be possible for the reaction of alkali metal (represented by e-) with water to go through EDAH+ as an intermediate. The mechanism would consist of Equation 31 and e_ + EDAH+ -——+' H + EDA (52) ' - + Writing --9駗l = kl[e ][EDAH I (55) and + d EgtH = k2[H20][EDA] - k3[EDAH+][0H 1 ~ kl[EDAH+][e-] (54) d EDAH+ then setting dt = 0, and substituting the result into Equation 33 gives _ d|e_| = k1k2[EDA] fit . k1[e'] + k3[OH-] [e-][H20] (35) 95 or ‘ - _ §é%_l = k [e-][H20] (36) Since K 10"2 M, [OH‘J’V 10"6 M by Equation 31. Then (105I(io-2)(10) k< 105(1031) + 1010(10-5T < 1 (57) The value found for k in this calculation represents an upper limit. Since this is at least an order-of-magnitude less than the observed rate constant, and since no dependence of the second-order rate constant upon water concentration was observed, it seems probable that this mechanism does not compete with the direct reaction a" + H20 —kJ-> H + OH' (38) It is likely that this step is followed by rapid ab- straction of an alpha-hydrogen atom from EDA, H + NHgCHgCHgNHg -—> NHgCHgéHNHg + H2 (59) analogous to the way alpha-hydrogens of ethanol react with hydrogen atoms in pulse radiolysis studies (79). The free radical produced may either react with another radical or with the reducing species. Since the overall reaction pro— duces half a mole of hydrogen for each mole of metal reacted, the latter reaction a“ + NH2CHCH2NH2 —a» NH2CH2CH2NH' (40) 96 to ultimately yield a solvent anion is more likely. A steady- state treatment of hydrogen atoms and solvent radicals gives the result -—L—lddf = 2 k3, Ie‘I [H20] (41) Thus with this mechanism, the observed rate constant is twice the rate constant for Reaction 38. G. Future Work 1. (11,12) and the present work The work of Dewald gt have effectively demonstrated the potential of this method for studying electron-attachment reactions. The present work has also pointed up the need for a better understanding of metal-amine solutions, if truly quantitative and meaningm ful results are to be obtained. The state of the theory of these solutions has been discussed previously (Sect. II-B), and a number of general criteria have been mentioned. In direct relation to our kinetic studies, further work is needed on the optical absorption spectra of metal solutions in EDA, particularly of stable dilute solutions. The rapid- scan system could be used in place of a slower spectropho- tometer to achieve this. Certainly the nature of the V band, especially in sodium solutions, needs to be cleared up. Combined EPR and conductivity studies would be of great value here. 97 The use of potassium solutions for kinetic studies should be investigated, since there is a greater separation of IR and R bands in these solutions. Kinetic details of the methanol and ethanol systems, and the apparent inter- mediate formed need to be worked out. Studies to obtain the temperature dependence of these systems would be tedious, but might provide some useful information. Other solvents such as liquid ammonia and 1-3 propanediamine should be investigated. Other electron—transfer reactions could be examined using the flow system. These include reactions such as (IR) + EDA CsR + Na —-> Cs++Na(V) (42) and inorganic reductions (99) such as Zn+2 + e" ——->- Zn+l (43) The rate of reactions such as the latter is too high for direct measurement, but transient spectra could be studied with the rapid-scan system. Metal—amine solutions could be employed to produce organic anions which then could be used to study electron-transfer reactions. Other probable reactants for this system are those which are "unreactive" Math respect to the pulsed radiolysis technique, and include both organic and inorganic compounds (100,101). 4. 5. 6. 11. 12. 13. 14. 15. 16. 17. 18. REFERENCES E. J. Hart, Advances in Chemistry Series, 59, "Solvated Electron," edited by R. F. Gould, American Chemical Society, Washington, D. C., 1965, p. vii. J. W. Boag and E. J. Hart, Nature, 197, 45 (1963). E. J. Hart and J. W. Boag, J, Am. Chem. Soc., §4, 4090 (1962). J. P. Keene, Nature, 197, 47 (1963). W. Weyl, Ann. Physik, 121, 601 (1864). C. A. Kraus, g. A_m_. Chem. Soc., §Q, 1325 (1908). G. Lepoutre and M. J. Sienko, editors, "Metal Ammonia Solutions; Physico-chemical Properties," W. A. Benjamin, Inc., New York, 1964. L. M. Dorfman, p. 36 in Ref. 1. M. Anbar, ibid., p. 55 and references therein. E. J. Hart, S. Gordon, and E. M. Fielden, as quoted in Ref. 9. R. R. Dewald, Ph. D. Thesis, Michigan State University, 1963. R. R. Dewald, J. L. Dye, M. Eigen, and L. de Maeyer, 1. Chem. Phys., 39, 2388 (1963). E. J. Kirschke and W. L. Jolly, Science, 147, 45 (1965). C. Kraus, g. Chem. g9, §_<_)_, 83 (1953) . M. C. R. Symons, Quart. Rev., 13, 99 (1959). W. L. Jolly, Progr. Inorg. Chem., 1, 235 (1959). T. P. Das, Advances ig_Chemical Physics, 4, 303 (1962). M. Ottolenghi and H. Linschitz, p. 149 in Ref. 1. 98 99 19. A. L. Cederquist, Ph. D. Thesis, Washington University, St. Louis, 1963. 20. R. C. Douthit and J. L. Dye, J, Am. Chem. Soc., _2, 4472 (1960). 21. M. Gold and W. L. Jolly, Inorg. Chem., 1, 818 (1962). 22. D. F. Burow and J. J. Lagowski, p. 125 in Ref. 1. 23. C. W. Orgeel, A. M. Filbert, and E. C. Evers, p. 67 in Ref. 7. 24. S. Gunn, ibid., p. 76. 25. W. H. Brendly, Jr., and E. C. Evers, p. 111 in Ref. 1. 26. J. L. Dye, R. F. Sankuer, and G. E. Smith, J, Am, Chem. Soc., 82, 4797 (1960). 27. J. C. Thompson, p. 96 in Ref. 1. 28. C. A. Kraus and E. C. Evers, p. 1. in Ref. 1. 29. E. Huster, Ann. Physik, §§, 477 (1958). 30. S. Freed and N. Sugarman, J. Chem. Phys., 1, 354 (1943). 31. C. A. Hutchison, Jr., and R. C. Pastor, ibid., 21, 1959 (1953). 32. D. E. O'Reilly, ibid., 41, 3729 (1964). 55. T. R. Hughes, ibid., §§, 202 (1963). 34. R. A. 099, Phys. Rev.) £21 668 (1946). 35. W. N. Lipscomb, g, Chem. Phys., 21, 52 (1953). 36. H. A. Stairs, ibid., 21, 1431 (1957). 37. J. Jortner, S. A. Rice, and E. G. Wilson, p. 222 in Ref. 7. 38. J. Jortner, J, Chem. Phys., 39, 839 (1959). 39. A. S. Davydov, g, Expt. Theoret. Phys. U.S.S.R., 18, 913 (1948). 40. M. F. Deigen, Trudy Inst. Fiz., Akad. Nauk Ukr. S.S.R.. .§. 119 (1954). 41.. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 100 E. Becker, R. H. Lindquist, and B. J. Alder, J, Chem. Phys., §§. 971 (1956). . J. L. Dye, G. E. Smith, and R. F. Sankuer, J. Am. Chem. Soc., 82, 4803 (1960). M. Gold, W. L. Jolly, and K. S. Pitzer, J. Am. Chem. Soc., D. E. O'Reilly, J, Chem. Phys., 41, 3736 (1964). J. L. Dye and L. H. Feldman, unpublished results. E. Arnold and A. Patterson, Jr., p. 285 in Ref. 7. E. Arnold and A. Patterson, Jr., J. Chem. Phys., 41, 3089 (1964). S. Golden, C. Guttman, and T. R. Tuttle, Jr., J, Am, Chem. Soc., 81, 135 (1965). W. L. Jolly, p. 27 in Ref. 1. R. R. Dewald and J. L. Dye, J. Phys. Chem., 68, 121 (1964). R. R. Dewald and J. L. Dye, ibid., 128. J. L. Dye and R. R. Dewald, ibid., 135. K. D. Vos and J. L. Dye, ,1. Chem. Phys., 38, 2055 (1963) . K. Bar-Eli and T. R. Tuttle, Jr., Bull. Am, Phys. Soc., .Q. 352 (1963). K. Bar-Eli and T. R. Tuttle, Jr., J. Chem. Phys., 49, 2508 (1964). T. R. Tuttle, Jr., and K. Bar-Eli, Abstracts, 146 Eh National Meeting of the American Chemical Society (January 1964), p. 7D. L. R. Dalton, J. D. Rynbrandt, E. M. Hansen, and J. L. Dye, submitted to J, Chem. Phys. D. E. O'Reilly and T. Tsang, J, Chem. Phys., 42, 3333 (1965). M. Ottolenghi, K. Bar-Eli, H. Linschitz, and T. R. Tuttle, Jr., ibid., 2Q, 3729 (1964). 101 60. J. L. Dye, L. R. Dalton, and E. M. Hansen, Abstracts, 149 £h_National Meeting of the American Chemical Society (April 1965), p. 455. 61. J. L. Dye, personal communication. 62. M. Ottolenghi, K. Bar-Eli, and H. Linschitz, J. Chem. Phys., éé. 206 (1965). 63. A. Ulland and J. L. Dye, unpublished results. 64. M. Anbar and E. J. Hart, J, Phys. Chem., 69, 1244 (1965). 65. A. Debierne, Ann. Physique (Paris), 2, 97 (1914). 66. N. F. Barr and A. 0. Allen, J. Phys. Chem., 63, 928 (1959). 67. J. T. Allan and.G. Scholes, Nature, 187, 218 (1960). 68. E. Hayon and J. Weiss, J. Chem. Soc., 5091 (1960). 69. P. J. Dyne, D. R. Smith, and J. A. Stone, Ann. Rev. pf Phys. Chem., 14, 313 (1963). 70. G. Czapski, J. Jortner, and G. Stein, J. Phys. Chem., §§, 964 (1961). 71. J. W. T. Spinks and R. J. Woods, "An Introduction to Radiation Chemistry," John Wiley & Sons, New York, 1964, Chap. 8. 72. G. Czapski and H. A. Schwartz, J, Phy . Chem., fig, 471 73. M. S. Matheson, P. 45 in Ref. 1. 74. R. L. Platzman, "Basic Mechanisms in Radiobiology," U. S. Nat. Acad. Sci. Pub. No. 305. 1953, p. 34. 75. T. J. Sworski, p. 263 in Ref. 1. 76. P. N. Moorthy and J. J. Weiss, ibid., p. 180. 77. M. S. Matheson and J. Rabani, J, Phys. Chem., 69, 1324 (1965). 78. J. Rabani, p. 242 in Ref. 1. 79. L. M. Dorfman and I. A. Taub, J, Am, Chem. Soc., 85, 2370 (1963). 80. 81. 82. 83. 84. 85.‘ 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 102 M. C. Sauer, Jr., S. Arai, and L. M. Dorfman, J, Chem. Phys.. 43. 708 (1965). J. J. J. Myron and G. R. Freeman, Can. J. Chem., 43, 381 (1965). R. W. Ahrens, B. Suryanarayana, and J. E. Willard, .J. Phys. Chem., §§. 2947 (1964). D. M. J. Compton, J. F. Bryant, R. A. Cesena, and B. L. Gehman, submitted to Rad. Res. B. Chance, J, Franklin Inst., 229, 455, 613, 737 (1940). O. H. Gibson, J, Physiol., 117, 49P (1952). F. J. W. Roughton and B. Chance, Technique 9; Organic Chemistry, VIII, edited by S. L. Friess, E. S. Lewis, and A. Weissberger, Interscience Publishers, New York, 1963, p. 703. E. F. Caldin, "Fast Reactions in Solution," John Wiley & Sons, Inc., New York, 1964, Chap. 3. B. Chance, gt 31., editors, "Rapid Mixing and Sampling Techniques in Biochemistry," Academic Press, Inc., New York, 1964. (a) Electronic Scanning Spectrometer, ITT Industrial Laboratories, Fort Wayne, Indiana. (b) Rapid Scan Monochromator, Perkin-Elmer Corp., Norwalk, Connecticut. (c) Millisecond Wavelength Scanning Spectrometer, Control Instrument Division, Warner & Swasey Co., Flushing, New York. W. Niesel, D. W. LUbbers, D. Schneewolf, J. Richter, and W. Botticher, Rev. Sci. Instr., 35, 578 (1964). K. D. Vos, Ph. D. Thesis, Michigan State University, 1962. G. Dulz and N. Sutin, Inorg. Chem., 2, 917 (1963). L. H. Feldman, R. R. Dewald, and J. L. Dye, p. 163 in Ref. 1. I. A. Taub, D. A. Harter, M. C. Sauer, Jr., and L. M. Dorfman, J, Chem. Phys., 41, 979 (1964). G. R. Freeman, personal communication to J. L. Dye. 103 96. E. J. Kelly, H. V. Secor, C. W. Keenan, and J. F. Eastham, J, Am. Chem. Soc., 82, 3611 (1962). 97. O. H. Gibson, p. 115 in Ref. 88. 98. M. Stafford, M.S. Thesis, Michigan State University, 1963; and J. L. Dye, personal communication. 99. G. E. Adams, J. H. Baxendale, and J. W. Boag, Proc. Chem. Soc., 241 (1963). 100. E. J. Hart, S. Gordon, and J. K. Thomas, J, Phy . Chem., .§§. 1271 (1964). 101. J. K. Thomas, S. Gordon, and E. J. Hart, ibid., 1524 (1964). . 102. E. J. Hart, E. M. Fielden, L. R. Dalton, and J. L. Dye, personal communication. (“mm “‘I’I‘I‘I‘I‘II'IIII‘I AYE UNIVE II I II‘II'IIIIIIII'I‘II'I‘I‘“ 93 03056