摘要:

J. Chem. Soc., Faraday Trans. 1, 1982, 78, 2387-2397 Structure and Dynamics of Graphite Intercalation Compounds Part 2.-Kinetics of Formation of C,KDj and C,KHj BY JEAN-PIERRE BEAUFILS Institut Laue Largevin, Grenoble, France AND TIMOTHY TREWERN, ROBERT K. THOMAS AND JOHN w . WHITE* Physical Chemistry Laboratory, Oxford University, South Parks Road, Oxford OX1 342 Received 7th August, 1981 The kinetics of the intercalation of hydrogen and deuterium into C,K have been studied by neutron powder diffraction. There appear to be two different mechanisms for the reaction of hydrogen or deuterium with C,K depending upon the temperature. Below ca. 370K the activation energy derived from the Arrhenius plots is ca. 50 kJ mol-1 and the reaction order is < 1. In this range the ratio of the rates for hydrogen and deuterium is large.At high temperatures the activation energy was not measurable and the order of the reaction is rather higher, possibly 2. In this temperature range also the isotope effect has decreased. A mechanism involving dissociation and diffusion is invoked to explain these results. In the previous paper1 we have confirmed that the first-stage compound C,K absorbs hydrogen and forms the second-stage intercalate C,KH+, in which bilayers of potassium are alternated with double graphite sheets. This reaction 3 C,K + H, --+ 3 C,KH$ (1) is reversible in the temperature range above 300 K and the thermodynamics have been thoroughly studied by Colin and Herold., The reaction is a remarkable solid-state transformation requiring the large-scale rearrangement of the graphite lattice and perhaps the macroscopic migration of dislocations, rafts of alkali-metal ions and hydrogen, in the same way as supposed for formation of two-component second-state intercalates them~elves.~-~ We suppose that the initiation of such a transformation requires lattice-dynamical phenomena akin to those responsible for displacive phase transitions in solids as well as a driving force (in this case presumably the formation of the rather stable potassium hydride sheet of the second stage C,KH$).Since little is known of the chemical kinetics of reaction we have started by determining the rates of reaction and their orders. Neutron diffraction has been used to follow the rate of reaction of C,K with hydrogen and deuterium as a function of pressure and temperature and is convenient because of the very different structure factors for C,KH+, C,K and C,KD .A subsequent paper will report the hydrogen- deuterium exchange, again followe d by neutron diffraction. EXPERIMENTAL The same sample of C,K made from powdered pyrolytic graphite [sample 9 of ref. (l)] was used for all experiments. Each experiment consisted of first carefully outgassing at 603 K the powder specimen contained in a thin-walled silica tube and then introducing hydrogen or 23872388 GRAPHITE INTERCALATION COMPOUNDS deuterium at a pressure, p , while the system was maintained at a temperature, T. The pressure was measured by a Baratron gauge and the temperature of the specimen was measured by a thermocouple.The apparatus for the experiment is shown in fig. 1. Patterns were accumulated over times ranging from 1 to 10 min using the high-flux powder diffractometers DIB and D2 at the Institut Laue Langevin, Grenoble. Whilst initial experiments were done on D2, the longer wavelength and higher resolution of DIB was required for precise measurements and all kinetics runs used this. Each pattern from DIB was corrected to constant monitor counts (1000 counts = 45 s). In each case the reaction was followed to completion. It was not possible to study the reverse reaction because of limitations in the furnace system. to gas handling system t vanadium heater thermocouple I water cooled diffractometer terminal table FIG. 1 .-Experimental arrangement for kinetics experiments using neutron diffraction at high temperatures.RESULTS PRELIMINARY OBSERVATIONS A N D DATA TREATMENT Fig. 2 (a) and 3 (a show the diffraction patterns from samples of C,KD; and C,KHi by driving off the hydrogen (or deuterium) at 603 K are also shown (the background of the silica sample container has not been subtracted). The (001) peaks of the C8K produced by dehydrogenation are narrower than the (001) peaks of the parent hydrogen compounds. W-ith the higher dispersion and good focussing of the DIB instrument the patterns had the same general features, but backgrounds around the characteristic peaks of C,KHi, C8KDi and of C8K itself were flat after subtraction of the scattering from the empty silica sample container. A typical scan of the region 3 < 2 8 / O < 15 for the shortest measuring time (60 s) used is shown in fig.4(b) along with a computer profile [fig. 4(a)] fit used to determine peak area, widths and heights. A typical run consisted of measuring ca. 35 such patterns, although where necessary longer runs were taken (table 1 ) . For reactions at lower taken at A = 1.26 B with D2. For comparison the patterns from the C8K producedJ-P. BEAUFILS, T. TREWERN, R. K. THOMAS A N D J. W. WHITE 2389 b ~~ (Q 1 (001 1 (003) (004) (b) (004) (008) (040) (22 0) & 200- n 0 , , , , 1 , , , , 1 , , , , , , , 1 I 0 10.0 20.0 30.0 40.0 2e lo FIG. 2.-Diffraction pattern of C,KD outgassing (D2 diffractometer I = 1, and of C,K (b) regenerated from this sample by heat and ). INT = scattered neutron intensity (counts per monitor).temperatures the run times for each spectrum extended up to 10 min, with corresponding improvements in the statistical accuracy of peak profile fittings. For the instrument DIB, the two most prominent features of the diffraction pattern of mixed C,K and C8KHj during a reaction were peaks at scattering angles, 28, of 11.2 and 25.2', corresponding to the first (001) reflexions of the product C8KHi and reactant C,K, respectively. The progress of the reaction could be followed qualitatively by observing both the decrease in intensity of the C8K peak and the increase in intensity of the C,K& peak. To put the observations on a more quantitative basis two parameters were determined for each of the two peaks for each pattern: the peak height, H , and the peak area, A .A background was subtracted, taken to vary linearly with angle. This procedure allowed the extraction of a third parameter, the peak width, W, which is defined for the, assumed, triangular line by: Wx = A,/H,. The results of all experiments are summarised in table 1. TEXTURE OF THE SAMPLE D U R I N G THE REACTION Since all experiments were carried out on the same sample of C,K, regenerating it by heating after a hydrogenation run, the width parameter of the C8K (004) and C,KHi and C,KDi (001) peaks was monitored to detect any irreversible changes in sample texture resulting from the sequence of reactions. For C,K the (004) reflexion width returned to an approximately constant value afterN TABLE l.-SUMMARY OF RESULTS time per width of C,KHt initial limiting rate constant experiment temperature pressure pattern no.of width average standard area k (W2) no. T/K p/Torr /min patterns W(C,K) W(C,KH,) deviation A(C,KH,) /min-' / 1 0-5 s-' HO 1 H02 H03 H04 H05 H07 H08 H09 H10 DO 1 DO2 DO4 DO5 DO6 DO7 DO8 401 401 303 368 337 368 337 303 337 40 1 337 368 368 401 337 337 170.0 392.3 393.7 377.0 375.4 166.7 166.2 167.3 261.1 376.1 376.6 374.4 166.3 166.8 165.0 3 16.4 1 1 5 1 2 2 2 10 2 1 5 2 33 1 5 5 31 8 123 12 24 15 39 58 28 31 32 19 36 46 57 28 - 1.863 1.532 1.621 1.570 1.664 1.868 1.700 1.526 1.760 1.636 1.468 1.939 1.852 - 3.43 3.43 2.89 3.32 3.74 3.60 3.84 3.87 4.23 4.24 4.3 1 4.35 4.62 4.47 4.54 - 0.07 0.06 0.15 0.1 0.03 0.04 0.04 0.11 0.03 0.06 0.06 0.03 0.28 0.09 0.07 - 3300 3900 4600 5100 4600 4900 4700 4000 5000 3 100 3500 3200 3000 1800 2300 2000 0.076 0.53 0.0065 0.17 0.103 0.19 0.088 0.0041 0.09 0.21 0.026 0.14 0.045 0.041 0.0097 0.019 11.5 61 .O N/L' N/L 5.1 17.0 3.6 NILu 15.0 48.0 NILu 7.4 2.9 6.3 N/Lu NILu a N/L: non-linear plot.J-P.BEAUFILS, T. TREWERN, R. K. THOMAS AND J. W. WHITE 2391 0 10.0 20.0 30.0 40.0 201" 7 0 10.0 20.0 30.0 2e1° 40.0 FIG. 3.-Diffraction pattern of outgassing (D2 diffractometer and of C,K (6) regenerated from this sample by heat and ). INT = scattered neutron intensity (counts per monitor). each regeneration, thereby indicating that at least the c-axis texture of the starting material for each run was similar. It was not possible to monitor the ab-axis texture in the same way, because the complex peak shape of the (220) and other reflexions of the layer lattice need good statistics to make measurements worthwhile.By contrast with the C,K peak, the width of C,KHi (001) at saturation with hydrogen, when monitored at the end of each run, appeared to increase slightly from one experiment to the following. We believe that the variation of the width and limiting area at saturation of the C,KHj peaks is connected to the presence of an intermediate or disordered phase before crystallisation of C&Hf as the reaction proceeds (see below). During a reaction the width of the C,K peak increases as the peak area decreases, a feature which we interpret as due to decrease in the C,K crystallite sizes. The apparent size of crystallites (in the c direction), I,, was calculated from the Warren formula6 3" B cos 8 I, = - where A is the wavelength, B is the broadening (in rad) and 28 is the scattering angle (in ").Typical values at the start of the reaction are ca. 100-200 A.2392 v) Y 1 8 GRAPHITE INTERCALATION COMPOUNDS 6 11 16 21 26 2810 6 11 16 21 26 2e1° FIG. 4.-Computer fit profiles (a) and the measured diffraction pattern (b) during a reaction of hydrogen and C,K. The pattern was taken in 6 s on the DIB diffractometer (I.L.L.) 1 = 2.52 A. -0- 4 0.3 0.4 0.5 log w FIG. 5.-Variation of C,K peak area, A , with peak width, W, during the course of a hydrogenation reaction.J-P. BEAUFILS, T. TREWERN, R . K . THOMAS AND J . W . WHITE 2393 If one assumes that the linear dimensions of the C,K crystal along the a, b and c directions change in step as the reaction proceeds and if all grains were uniform spherical particles, the volume of the C,K scattering at any instant (and hence the diffraction peak area) should be proportional to B-3 and hence to the inverse cube of the peak width, W(C,K).Graphs of log A(C,K) for separate experiments have qualitatively the same shape and broadly demonstrate this criterion. An example is shown in fig. 5. That the gradient is not - 3 over the whole range shows that the model is too simple. The two most probable causes of this deviation are the distribution of crystallite sizes within the sample and systematic errors in background subtraction to determine the peak height of broad, weak peaks toward the end of the reaction. 3000 2000 cd 2 4 % 1000 V 0 0 0 O O 0 0 0 O 0 0 0 0 0 0 0 0 0 0 O O O O O 0 0 0 0 0 0 I I I 1 I I I 0 10 20 30 time/ min FIG.6.-An example of the kinetics of disappearance of the C,K peak as the reaction proceeds. Log peak area (C,K) is plotted as a function of time (min). EXTRACTION OF REACTION KINETICS PARAMETERS DISAPPEARANCE OF C,K Graphs of log A(C,K) against time were linear, allowing the reaction rate constants to be determined from their gradients. An example is shown in fig. 6. These rate constants were, in turn, used to construct Arrhenius plots as, shown in fig. 7. Such plots are a convenient way of presenting the data, which are too erratic for there to be any significance in drawing straight lines through the points. However, on joining the points as shown it is seen that all four plots have similar slopes in the lower temperature range.Three of the four plots have much smaller, or even negative, slopes in the upper temperature range. (Actual figures for the activation energies are collected in table 2.) The tentative conclusions are that the activation energies are2394 GRAPHITE INTERCALATION COMPOUNDS TABLE 2.-sOME KINETIC PARAMETERS FOR THE REACTION OF C,K WITH HYDROGEN AND DEUTERIUM temperature T/K pressure parameter isotope p/Torr 303 337 368 40 1 activation H energy H /kJmol-l D D order H D ratio of - rates for H, and D, - - 170 54 - 46 - 68 - 17 - 375 - 51 - 170 56 - 375 - - - - - - - 0.7 - < O 0.54 4.85 - 4.22 - - 1.3 - - - - - - 170 3.96 - 375 - - - - < O < O 31 - - 15 - 2 2 1.85 1.9 - - - - -1 -2 Y .5 -3 -4 -5 -6 I I I I I I I I I 2.5 3.0 103 KIT FIG.7.-Arrhenius plots for hydrogenation and deuteriation of C,K at two different pressures: x , H, p x 375 Torr; A, H, p M 170 Torr; +, D, p z 375 Torr; 0, D, p x 170 Torr.J-P. BEAUFILS, T. TREWERN, R. K . THOMAS AND J. W. WHITE 2395 ca. 50 kJ mol-l at lower temperatures and that they decrease substantially at higher temperatures. However, the errors are obviously large. From graphs of log k against log p a reaction order with respect to hydrogen or deuterium was determined (see fig. 8). Again the results are collected in table 2. PRODUCTION OF C,KH$ Consider the reaction at an arbitrary time t. If all the C,K were immediately converted into crystalline C,KHj then the expressions -A(C8K)t/A(C&3K)0 and A(C8KH$)t/A(c8KH$)~ should equally well represent the progress of the reaction.A graph of A(C,KH& against A(C&)t should therefore be linear. As can be seen from fig. 9 this was not the case. There are at least two possible interpretations; the passage from C& to C,KH; is via an amorphous intermediate; the crystallinity of the C,KHg itself varies and the compound may have a high concentration of defects. The correct interpretation may be a combination of these possibilities. - 0.8 - 1.2 Y 00 - - 1.6 - 2.0 FIG. 8.-Dependence of the rate constant on pressure at 337 K. The slopes are of the order of or less than unity. 0, H; 0, D. It has been found possible to adopt a semi-quantitative approach to the increase of the C,KHi peak. If the growth of the C,KHg peak was determined by a diffusion process then the peak area should vary as ti.In most cases, graphs of A(C,KHi) against ti were found to be linear, the exceptions occurring at the lower temperatures and higher pressures.2396 3000 2 g 2000 T 1000 0 GRAPHITE INTERCALATION COMPOUNDS O \ O ‘\ 0 o\ 0 >, O \ 0‘ 0;’ 1 I I 1 I I , ol\,o, 0 1000 2000 3000 4000 A (C&Hw) FIG. 9.-An example of the test for equivalence of the rates determined from disappearance of C,K and appearance of C8KH1. The sense of the deviation from linearity suggests as intermediate, an amorphous phase or defect C8KH1 lattice. DISCUSSION INITIAL STEPS The variation of A(C,K) with time is only a practical means of calculating an initial rate. We cannot propose a simple model to account for it. There appear to be two different mechanisms for the reaction of hydrogen or deuterium with C,K depending on the temperature.Below ca. 370 K the activation energy derived from the Arrhenius plots is ca. 50 kJ mol-l, the reaction order is < 1 (possibly 0.5), and the ratio of the rates for hydrogen and deuterium is large, In this range the rate-determining step is probably the surface dissociation of hydrogen. At higher temperatures the activation energy was not measurable, the order is higher (possibly 2) and the isotope effect has decreased. The absence of an activation energy suggests that a diffusive process may determine the rate of reaction. If this is so then the high order is not easily explained and the ratio of the rates for the different isotopes of 1.8-2.0 is higher than the predicted value of d2.We can only say that, after dissociation, H creates a species X and that the solid becomes partly amorphous, a process that implies the diffusion of species X. FINAL STEPS During each hydrogenation the CEKHj peak width does not vary. Either the macroscopic disorder (mosiac spread) is insensitive to completion of the reaction or the increase of the peak area corresponds to an increase of the number of CEKHt crystallites and not to a growth of each individual crystallite. Assuming this hypothesis and noting that the rate of this nucleation follows a ti law we can consider that the nucleation of CEKH3 takes place when a sufficient number of X species have migrated. Again from the linear dependence of the peak area on ti and by assuming that the radius r of the particles in which diffusion occurred is known, a value for the diffusion constant D can be obtained.Values of D/r2 range from ca. 3 x to 60 x s-l. Assuming a particle radius of lo-* places D in the region of m2 s-l.J-P. BEAUFILS, T. TREWERN, R. K. THOMAS A N D J. W . WHITE 2397 SUMMARY OF THE REACTION Evidence from this study has allowed an overall picture of the hydrogenation of potassium graphite C,K to be devised. (1) Hydrogen is dissociatively chemisorbed on the C,K surface. (2) H, produces a species X which migrates between the graphite layers, giving an amorphous solid. (3) X collects to produce nuclei of C,KHg. (4) The nuclei eventually grow. Step (1) is rate-determining below ca. 370 K. Above this temperature, step (2) is rate-determining for the destruction of the C,K lattice.In any case step (2) determines the rate of the nucleation. It is plausible that the species X is either atomic hydrogen alone (produced by the dissociative chemisorption at the C,K surface), or atomic hydrogen closely associated with dislocations in the graphite stacking. In the low-temperature range, below ca. 370 K, production of X is the rate-determining step although at higher temperatures this ceases to be the case. Production of X is followed by its diffusion into the interior of the reacting crystalite to allow C,KHi clusters to form. The estimated time-scale of this process suggests that it is not observable by neutron quasielastic scattering. The atomic hydrogen causes the alkali-metal species to cluster forming an amorphous intermediate and accounting for the deviation from linearity of graphs of the types shown in fig. 9. The final stage of the reaction is the migration of the clusters to form imperfectly crystalline C,KHg, consistent with the model of Lagrange et aL4 The defects in the structure take the form of distortions of the graphite layers, although more gross defects, for example deviations from the regular stacking sequence or amorphous regions, are probably present. The trends of increasing width and decreasing area of the limiting C,KH$ peaks possibly arise from an increase in the concentration of dislocations with repeated hydrogenation-dehydrogenation. In view of the catalytic properties of potassium graphite it would be useful to examine the reaction under conditions where the dissociative chemisorption of hydrogen is not rate-determining and to study the effects of using carbons with a wide range of initial dislocation densities. ' T. Trewern, R. K. Thomas, G. Naylor and J. W. White, J. Chem. SOC., Faraday Trans. 1, 1982,78, 2369. M. Colin and A. Herold, Bull. SOC. Chim. Fr., 1971, 1982. A. Herold, N. Daumas, C. R. Acad. Sci., Ser. C, 1966, 286, 373. P. Lagrange, Carbon, 1978, 16, 235. B. E. Warren, Phys. Rev., 1941, 59, 693. * P. Lagrange, A. Metrot and A. Herold, C.R. Acad. Sci., Ser. C, 1974, 278, 701. (PAPER 1 / 1249)

ISSN:0300-9599    

DOI:10.1039/F19827802387

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. Sac., Faraday Trans. I , 1982, 78, 2399-2410 Structure and Dynamics of Graphite Intercalation Compounds Part 3.-Crystal Dynamics and Binding in C,K, C,K& and C,KDg by Inelastic Neutron Scattering BY TIMOTHY TREWERN, ROBERT K. THOMAS AND JOHN W. WHITE* Physical Chemistry Laboratory, Oxford University, South Parks Road, Oxford OX1 3 4 2 Received 7th August, 1981 Neutron inelastic scattering from C,K, C,KHt and C,KDI has been measured with the objective of determining the vibrational spectrum for the hydrogen and hence the site symmetry. A model for the site symmetry consistent with the observed spectra is proposed and this is related to the diffraction measurements made on powder and partially oriented specimens. The phonon dispersion curves for C,K at room temperatures have been obtained and give a measure of the change in stiffness between carbon sheets as a result of intercalation.The vibrational spectra of the alkali-metal intercalates C,K, C,KH; and C,KDr provide complementary structural information to that available from diffraction. In particular the hydrogen vibrations are sensitive to the local symmetry, an element of the structure which is rather inaccessible through diffraction measurements because of the limited long-range order. Neutron inelastic scattering is the method of choice because the materials are opaque and because measurements at non-zero phonon wavevectors are needed. The lack of large single-crystal specieslg precludes measurements of the phonon dispersion curves in the classical way and so coherent scattering methods developed for partially ordered polymers3 and density of phonon states measurements4 have been used.EXPERIMENTAL The samples were prepared by methods previously given,' and table 1 of ref. (1) summarises experiments done on all samples. In all cases spectra were taken with the sample contained in thin-walled silica tubes whose spectra were separately recorded to allow subtraction of background where this caused interference. Time-of-flight spectra from C,K and C,KH, were recorded with the 4H5 and 6H cold-neutron spectrometers (A.E.R.E., Harwell) at 300 and 100 K. The elastic energy resolution of these instruments was ca. 8% AE/E or cu. 200 peV. In the search for diffusion of hydrogen in C,KH, at elevated temperatures the IN10 backscattering spectrometer at I.L.L.Grenoble, with quasielastic resolution of 1 peV, was used. Beryllium filter spectrometers on Pluto (at A.E.R.E.) and INlB (at I.L.L., Grenoble) were used to determine the amplitude-weighted density of states of C,K, C,KH, and C,KD, at energy transfers above 200 cm-l. 23992400 GRAPHITE INTERCALATION COMPOUNDS RESULTS LATTICE DYNAMICS C8K INCOHERENT SCATTERING Fig. 1 shows the background uncorrected time-of-flight spectra of C,K at 300 K taken at scattering angles, 29, of 63, 54, 45 and 36O using the 6H spectrometer at Harwell. The steadily rising background on which the spectrum sits is an experimental artifact. Energy transfers above ca. 100 cm-l are associated with momentum transfers greater than ca. 2 A-l, sufficient to allow the incoherent approximation, at least for excitations along the c dire~tion.~ The peaks correspond to maxima in the density of states and are listed with closely related beryllium-filter data in table 1.The beryllium-filter spectra of C,K at 80 K taken at Harwell(3-axis machine) and at I.L.L. (INlB) are shown in Fig. 2. The energy transfers of all these peaks are the same within experimental error, a feature consistent with their assignment to incoherent scattering. The peak intensities are quite different in the two different spectroscopic methods owing to the non-linear resolution function in time-of-flight and to the larger momentum transfers in beryllium-filter experiments. energy transferlcm-' 5.0 I 30 x 5.0 2.0 O 1 *A0 ' d o I rdbo ' 1;oo I 4 neutron time of flightlps m-' FIG.1.-Time-of-flight spectra for C,K at 300 K at scattering angles, 28, of 63, 54, 45 and 3 6 O , showing phonon peaks as well as density of states maxima.T. TREWERN, R. K. THOMAS A N D J. W. WHITE 240 1 TABLE ANG FREQUENCIES IN THE INCOHERENT SCATTERING SPECTRA OF C,K ATTRIBUTABLE TO MAXIMA IN THE PHONON DENSITY OF STATES. ENERGY TRANSFERS ARE IN OPTICAL WAVENUMBERS (cm-l). beryllium filter (80 K) time of flight (300 K) scattering mean/ Harwell INlB anglelo.. . 27 36 45 54 63 72 81 90 cm-l /cm-l /cm-l uncertainty/cm-l f 20 170 140 140 165 165 140 125 - 149+lO 170f10 - & 40 245 220 230 255 245 245 230 240 239f10 230f 10 - +250 620 595 595 570 550 510 510 530 560+50 - 590f 10 640 780 1100 1265 1350 1400 1500 x x x x .....**. x x x .........*...'. " . .... ;......:-.".*,~. .. .=.. . . . . 1 ....*--' ... . ..... ... ..-......:. ..... -.... .... .. ...... 1 1 1 1 1 1 1 1 1 1 1 1 ! I I I I l 500 1000 1500 cm-1 energy transfer/cm-' FIG. 2.-Beryllium-filter spectra of C,K at 80 K taken with the Harwell (BF, 0 ) and I.L.L. (INlB, x ) instruments which cover complementary energy ranges. (Note the INlB spectra are on a greatly enhanced count scale compared with those of BF.) C,K COHERENT SCATTERING In contrast to the high-energy peaks in fig. 1 there are peaks between 90 and 0 cm-l in the C,K time-of-flight spectra whose energy transfer is a function of scattering angle (momentum transfer). These are the average phonons observable when a particular phonon-dispersion surface is intersected by the instrumental dispersion curve.3 78 FAR 12402 GRAPHITE INTERCALATION COMPOUNDS The condition for coherent inelastic scattering is Q = 2m-I-q where Q is the momentum-transfer vector defined in terms of the incident and outgoing neutron momenta, k , and k’, by Q = k’ = k,. Q is a function of the energy transfer n2 2m AE = - ( ~ k ’ ~ 2 ~ k o ~ 2 ) as well as the scattering angle 28.In the above expression, 7 is a reciprocal lattice vector and q is the momentum transfer associated with the creation (+) or annihilation (-) of the phonon. All phonon modes can, in principle, contribute to the neutron spectrum but in practice the situation is greatly simplified because the coherent inelastic intensity is determined by the square of the dynamical structure factor where (b), is the coherent scattering length of the atom at p in the unit cell, Mp is its mass and exp (- W,) is Debye-Waller factor. For longitudinal acoustic modes the effect of the first exponential is to boost the intensity when Q is near 2nz, whereas the optical and transverse modes are conversely affected.Another simplification occurs also as a result of the limited momentum-transfer range accessible under the experimental conditions of neutron energy gain by incident cold neutrons, a feature which comes especially into play for intercalation compounds for the following reasons. (a) The intercalate is structurally anisotropic with a large c-axis spacing, which means that only the Debye-Scherrer ring corresponding to the first strong powder reflexion is accessible in the range of momentum transfer covered.(Reciprocal lattice points off the c axis are widely spaced compared with the Q resolution of the spectrometer.) (b) The amplitude of vibration Uj in the scalar product (Q.Uj) is greater for vibrations perpendicular to the layer than for those in the layer plane. In theory a step-change in intensity should occur at the phonon frequency, but, in practice, this is not resolved by the spectrometer. The mid-point of the steep rise in intensity has been identified5 with the ‘cut-on’ and so both the phonon frequency u) and the momentum transfer Q may be determined. The values of Q and u) for the excitations seen with two spectrometers at 300 K are given in table 2. HYDROGEN MOTIONS IN C,KH$ A N D C,KD$ QUASIELASTIC SCATTERING Both C,KH$ and C,KD$ were examined at 300 K on the time-of-flight spectrometers at Harwell to search for possible diffusive motions of hydrogen by quasielastic scattering.Even at the highest momentum transfers used (ca. 2.5 A-l) there was no significant difference between the widths of the elastic peak obtained from the intercalates and that from the vanadium standard under identical conditions. All peaks contained some truly elastic Bragg scattering owing to the poor Q resolution of the instrument, but, this notwithstanding, it was concluded that any diffusive processes present gave quasielastic widths less than the range of measurement of the instrument (< 100 peV) at these temperatures. Accordingly a search for the quasielastic scattering from C,KH$ at temperatures up to 453 K was made using the much higher resolution of lNlO at Grenoble (1 peV).T.TREWERN, R. K. THOMAS A N D J. W. WHITE 2403 TABLE 2.-cOHERENT INELASTIC FEATURES FROM A C,K POWDER SPECIMEN ON THE HARWELL TIME-OF-FLIGHT SPECTROMETERS 4H5 6H frequency/cm-l, frequency/cm-l, scattering Q/A-l, scattering Q/A-l, angle/O reduced wavevector angle/O reduced wavevector 33 38 43 50 58 66 74 82 90 64.5 1.45 0.474 1.44 0.457 1.11 0.16 1.16 0.025 5.5 1.22 0.08 1 2.06 0.485 2.14 0.349 2.16 0.315 2.24 0.178 54 14 + 5 52 44 32 26 34 1.33 0.269 64 2.00 0.588 75 2.65 0.521 57 2.63 0.487 18 27 36 45 54 63 72 81 90 62.5 1.48 0.52 1.37 41 20.5 1.27 0.163 4 1.19 0.027 75 2.61 0.445 - - - - - - 29 1.05 0.21 2 1.10 0.127 1.14 0.059 2.00 0.594 2.07 0.475 21 11.5 65 56 - - - No quasielastic broadening was observed even though the hydrogen vapour pressure above the sample was of the order of 100 mmHg* at the highest temperatures.INCOHERENT INELASTIC SCATTERING Fig. 3 shows the time-of-flight spectrum of C,KHi at 300 K taken with the 6H chopper at A.E.R.E. Harwell at two scattering angles (28 = 36 and 54') to the incident-beam direction. The sample-container background has been subtracted in each case. The spectrum at higher scattering angles allows modes near 70 and 40 cm-l energy transfer to be seen more clearly. The beryllium-filter spectrum of C,KH+ at 80 K is shown in fig. 4, which compares the results from the Harwell (BF) and Grenoble (INlB) instruments in the range where * 1 mmHg = 13.5951 x 980.665 x lo-* Pa. 78-22404 GRAPHITE INTERCALATION COMPOUNDS I I I 4 1200 1600 2000 neutron time-of-flightlps m-I FIG.3.-Time-of-flight spectra from C,KH1 at 300 K and two scattering angles, 36 and 5 4 O , to the incident-beam direction. FIG. 4.-Beryllium-filter spectra of C,KHt at 80 K. [Note the INlB ( x ) and BF (0) count scales are very different .]T. TREWERN, R. K. THOMAS AND J. W. WHITE 2405 oooooooo 0 0 0 O O 0 0 0 0 0 0 &O0 O O O O 0 0 "0 O 0 0 0 0 0 0 00 0 0 00 0 200 400 600 80 0 FIG. 5.-Beryllium-filter spectra of C,KHg (0) and GKDg (e) at 80 K compared- TABLE 3.-FEATURES IN THE BERYLLIUM-FILTER (BF) AND TIME-OF-FLIGHT SPECTRA OF C8KHi AND C8KD, ATTRIBUTABLE TO MAXIMA IN THE PHONON DENSITY OF STATES instrument energy transfer/cm-l C8KHi - HarwellBF 8 K - - - 21 5 450 575 760 539d - 6H 298 K 550d - 4H5 328 K INlB 80 K - - - - - 600 760 1000-1750 - 36" 77' 158' - - 36" 76' 155' - - - C8KDg Harwell BF 80 K - 90 - 170-230 (325-550) - - - a Mean of two measurements; mean of four observations; mean of six measurements; mean of nine measurements.they overlap. The spectra are much more intense than the spectra of the C,K starting material (fig. 2). To assign the hydrogen vibrations a comparison was made between the beryllium-filter spectra (measured on the Pluto 3-axis machine at Harwell) of C,KHz and C,KDg. This is shown in fig. 5 . Table 3 lists the peak maxima found for C,KHi using the time-of-flight and beryllium-filter instruments with some data for C,KDg.2406 GRAPHITE INTERCALATION COMPOUNDS DISCUSSION LATTICE DYNAMICS OF C8K Since there is little change in the Graphite a, b lattice parameter upon formation of CaK or CaKHi we concentrate on changes in the c-axis binding as the most obvious consequence of intercalation.Since also the supplementary zone boundaries due to the longer c-axis of the intercalates seem to provoke little discontinuity in the longitudinal acoustic (LA) dispersion curves for C8Rba we may compare qualitatively the excitations in related compounds using an extended zone scheme based upon the repeat distance of neighbouring layers. This comparison is the basis of spectra O0 I / 95.5 cm 1 I I 1 I 1 I I 0 0.1 0.2 0.3 0.4 0.5 reduced wavevector FIG. 6.-Phonon dispersion curve for the LA(OO1) branch in C,K at 300 K: 4H: V, 33; A, 38; 0, 43; A, 50; 0, 58; A, 66; 7, 74; 0, 82; ., 90'; 6H: +, 36; x , 45; El, 54; 0, 63; V, 72; *, 90°.T.TREWERN, R. K. THOMAS AND J. W. WHITE 2407 assignments made below. All of the phonons in table 2 belonged to the same branch though they did not fall in the same Brillouin zone. The zero of frequency occurred at a value of the momentum transfer coincident with the (004) reflexion of C,K as determined by diffraction. This allowed a reduced wavevector to be calculated and a dispersion curve constructed as shown in fig. 6. The error bars span the maximum systematic error and represent the i-3 height points of the phonon cut-on. They have been drawn along the loci of the instrumental dispersion curves for each scattering angle. The dispersion relation in the harmonic approximation for a diatomic lattic with a single force constant, K, can be written6 as M1M,~4-2K(M1+M2) u2+2K2(1 -cosqc) = 0 where cis the lattice parameter and M, and M , are the masses of the oscillating species.Using this, the force constant has been extracted from fig. 6 by plotting w 2 against and making a least-squares fit for the gradient and Brillouin zone boundary intercept. The numerical values of K were 1.310 N m-l, obtained from the gradient, and 1.319 N m-l, obtained from the intercept. Taking the value of K to be 1.31 N m-l allows the optic branch to be calculated in the above approximation (fig. 7). The acoustic-branch maximum frequency appears at ca. 95 cm-l. The optic branch cut-off at the zone centre occurs at a frequency of ca. 175 cm-l. The group velocity & was found to be 4.06 x lo3 m s-l, which compares with the values for pyrolytic graphite of 4.1 x lo3 (Dolling and Brockhouse)s and for Graphon of 3.46 x lo3 m s-l (Gamlen and White4).160 12 0 - I 5 \ h $! 80 5 40 0 0.2 0.4 reduced wavevector FIG. 7.-Reconstructed acoustic and optic branches of the c-axis phonon dispersion curve for C,K at 300 K.2408 GRAPHITE INTERCALATION COMPOUNDS The stiffness constant, C, is related to the groups velocity by the relation c= v,zp where p is the density. Taking the crystallographic density of C,K to be 2.0 x lo3 kg m-l gives a value for C of 3.30 x 1Olo N m-2. As a result of the high incoherent scattering from the hydrogenated intercalate it was impossible to determine the dispersion curve of C&H$ in the same way.TABLE 4.-c-Ax1s LATTICE DYNAMICS PARAMETERS FOR GRAPHITE AND SOME ALKALI-METAL INTERCALATES calculated frequencies experimental from mass frequencies interplane mass of unit scaling/cm-l /cm-l force LA sound intercalate cons tan t velocity substance layer LA, A LO, LA, A LO, r /N m-l /103m s-l - graphi te7 96 86 125 86 125 4. I 5.15 39.1 86 195 95 175 1.31 4.06 C8Rbs 85.5 86 132 108 154 - C8K - - 137 - - CsKH, 78.8 86 The interlayer force constants (C-K for example) and the Brillouin zone centre and boundary frequencies allow the extent to which intercalation modifies the graphitic van der Waals forces to be judged. Table 4 compares the relevant parameters known at present and the expectations from scaling graphite frequencies by the square root of mass differences.For the A point in the zone no mass scaling applies, since the longitudinal acoustic eigenvector involves no K atom displacement. In C,Rb the graphite-rubidium interplanar force constant is greater than that between graphite sheets; the reverse is true in C,K. The near equality of the c-axis longitudinal acoustic sound velocities in C8K and graphite facilitates scaling of the graphite dispersion curvesg for interpreting the incoherent scattering spectra of C,K. Apart from the mass rescaling, recent Raman spectra of C&s, C,Rb, C8K and pristine graphite suggest that the Mlg level in all of these materials should be lowered to 560 cm-l. This is consistent also with the graphite neutron density of states4 which has a strong 'cut-on' at ca. 550 cm-l, rising to a strong Van Hove singularity at 620 cm-l.The neutron density of states also suggest a greater separation of the single and doubly degenerate modes in Mlg, at the A4 point. The frequency separation of the flat regions of these branches between the M and K points may be more like ca. 200 cm-l rather than the 50 cm-l of Nicklow et aL9 There is a persistent strong maximum at 230 cm-' in all spectra of C8K and C,KHg which could be due to the follow through of the Blg mode between r and C or to removal of the degeneracy of the in-plane TAI mode at M. The low frequency of this mode implies a smaller force constant for the shear of graphite and potassium planes.T. TREWERN, R. K. THOMAS AND J. W. WHITE 2409 VIBRATIONS IN C,KHi AND C&Dg LATTICE DYNAMICS Peaks in the amplitude weighted density of states for C,KHg and CgKDg are compared with those for graphite, C,K and NaH in fig.8. For C,KHg and C,KDg below 300 cm-l, there exists a qualitative correspondence with the excitations in graphite, and the symmetric high density-of-states points on the graphite dispersion surface have been adopted for reference. energy transfer/cm FIG. 8.-Relationship between density-of-states maxima in graphite and its stage-one (C,K) and stage-two (C,KHg and C,KDI) intercalation compounds. Symmetry labels are those of the graphite dispersion surface. The three modes at 450, 580 and 760 cm-l (fig. 4) are all strongly associated with the absorption of hydrogen and are shifted down in frequency in C,KDi. These excitations almost certainly cover at least one of the graphitic and C,K M-point modes, but from their great intensity (e.g.in fig. 3) they must be primarily localised modes of hydrogen. The clearest case of a frequency shift with deuteration is of the strongest mode, at 580 cm-’ in C,K& (ca. 400 cm-l in C,KD$) (frequency ratio 1.45). Interstitial hydrogen atoms and ions in alkali-metal and rare-earth halides give rise to characteristic infrared frequencies in the region 400-1000 cm-l,lof l1 and so by analogy the observed modes could be interpreted as the localised mode of a trapped hydride ion.12 At the other extremum of behaviour shown by a crystalline hydride, two infrared frequencies corresponding to the transverse and longitudinal optical frequencies are seen. l3 The sandwich structure proposed from diffraction measurements1 could be expected to have intermediate behaviour.In fact by making a full F-G matrix calculation, the spectrum shown in fig. 4 can be fitted quite well for a Dzd site symmetry. The observed excitations below 800 cm-l correspond to the two in-plane and the out-of-plane vibrations of the hydride ion and the bands above 1100 cm-l are harmonics and combinations with intensities as expected from the simple theory of incoherent inelastic neutron scattering.2410 GRAPHITE INTERCALATION COMPOUNDS CONCLUSIONS The lattice dynamics of graphite and its alkali-metal intercalation compounds are closely related. In the stage-one compound C8K the c-axis longitudinal excitations develop a marked energy gap and the interplanar forces are softened relative to graphite itself.This is also probably true for at least one of the transverse acoustic modes. The density-of-states spectrum from C8K is consistent with little change to the in-plane graphitic modes upon intercalation. For the two-stage compounds the hydrogen vibration spectrum is consistent with the model of a three-layer ‘potassium hydride ’ sandwich, although the alternative hypotheses of hydrogen-molecule ions cannot be ruled out. The lattice vibrations once again bear a remarkable correspondence to those in graphite. Hydrogen diffusion in C,KHt was unobservable between 293 and 453 K. Experiments at higher hydrogen pressures and all temperatures will be needed to detect the quasielastic scattering. T. Trewern, R. K. Thomas, G. Naylor and J. W. White, J. Chem. SOC., Faraday Trans. I , 1982,78, 2369. J. P. Beaufils, T. Trewern, R. K. Thomas and J. W. White, J. Chem. SOC., Faraday Trans. I , 1982, 78, 2387. J. F. Twisleton and J. W. White, in Neutron Inelastic Scattering (I.A.E.A., Vienna, 1972), p. 301. P. H. Gamlen and J. W. White, J. Chem. SOC., Faraday Trans. 2, 1976, 72, 446. D. K. Ross, J. Phys. C, 1973, 6, 3525. G. Dolling and B. W. Brockhouse, Phys. Rev., 1962, 128, 1120. Yu, W. Novikov and M. E. Vol’pin, Russ. Chem. Rev., 1971, 40, 733. W. D. Ellenson, D. Semmingson and J. E. Fischer, Muter. Sci. Eng., 1977, 31, 137. R. Nicklow, N. Wakabyashi and H. G. Smith, Phys. Rev. B, 1972,5,4951. lo G. Schaefer, J. Phys. Chem. Solids, 1960, 12, 233. l1 B. Fritz, J. Phys. Chem. Solids, 1962, 23, 375. l2 R. J. Elliott, W. Hayes, G. D. Jones, H. F. Macdonald and C. T. Sennett, Proc. R. SOC. London, Ser. l3 A. D. B. Woods, B. N. Brockhouse, M. Sakamato and R. N. Sinclair, Int. Symp. Inelastic Scattering A, 1965, 289, 1. of Neutrons in Solih and Liquih, Vienna, 1960, (I.A.E.A., Vienna, 1961), p. 487. (PAPER 1 / 1250)

ISSN:0300-9599    

DOI:10.1039/F19827802399

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. Soc.. Faradoy Trans. 1. 1982. 78, 241 1-2421 Effect of Substituent on the Behaviour of the Excited Singlet and Triplet States in Carbonyl Derivatives of Anthracene of the Type 9-X*CO.A BY SATOSHI HIRAYAMA Faculty of Textile Science, Kyoto Technical University, Matsugasaki, Sakyo-ku, Kyoto, 606 Japan Received 24th August, 198 1 Kinetic and spectroscopic studies relevant to the excited singlet and triplet states have been carried out on eighteen carbonyl compounds of anthracene, principally of the type 9-X CO . A (where A is anthracene), aiming to reveal distinctly how a substituent X controls the excited-state behaviour. The compounds examined are classified into six groups based on the nature of X. Group I compounds, where X is an alkyl group, are non-fluorescent in any solvent at room temperature, but they start to fluoresce strongly near 77 K with fluorescence lifetimes of ca.10 ns. When X is capable of conjugation with the carbonyl group, the derived compounds (group 11) totally lack fluorescence even at 77 K. The carbonyl groups of anthracene-9-carbaldehyde and 2-methylaceanthren-1-one (group 111) differ from the others in that they form hydrogen bonds in protic solvents. Only group V compounds, with an electron-donating substituent such as an amino, methoxy or hydroxy group, are modestly fluorescent at room temperature. Intersystem crossing from S, to Trill+, whose location relative to S, is affected by X, appears to play an important role in the appearance of fluorescence. Group VI compounds show that the introduction of another substituent at the 10-position also affects the fluorescence properties. In contrast to the wide variation seen in the fluorescence properties, the lowest triplet states are found to be similar to each other and only slightly different from those of anthracene.The shape of the phosphorescence and triplet-triplet absorption spectra and the lifetimes and energies of the lowest triplet states are also similar to those of anthracene. This similarity suggests that the lowest triplet states of the examined compounds are not of nn* but of m* character, originating from the anthracene group. The very weak or practically absent fluorescence in those organic compounds which consist of both pi and lone-pair electron systems has often been explained in terms of the so-called El-Sayed rule.' This states that intersystem crossing of the type nn* rw, ~ n * or vice versa is the most important process in depleting the lowest excited singlet state S,, although in some exceptional cases an internal conversion to the ground state would become dominant.2 In particular, when the radiative transition from S, is allowed, as is frequently so for SIIIn* (but not for Slnn*), the lack of fluorescence means that very efficient intersystem crossing occurs on the picosecond time-scale .In anthracene it is well recognized that chemical substitution affects the radiative properties through a substantial change in the intersystem crossing rate from S1,3 while the radiative transition probability itself remains practically constant. When a substituent contains a carbonyl group the effect of substitution may well be amplified, since it adds nn* states, which are expected to play an important role as intermediate states in the radiationless process depleting S,.This paper is a study of such a substituent effect on the photophysical processes of the rneso-substituted carbonyl derivatives of anthracene of the type 9-X * CO *A, where A and X stand for anthracene and the substituent, respectively. By changing X systematically, it will be shown how the substituent controls the radiative properties of the derived compounds. 241 12412 EXCITED-STATE BEHAVIOUR OF SUBSTITUTED ANTHRACENE EXPERIMENTAL The absorption spectra were obtained on a Shimadzu 210A spectrophotometer and the fluorescence and excitation spectra were recorded on a Shimadzu RF 502 spectrofluorophoto- meter provided with a Hamamatsu R 928 photomultiplier.The excitation spectrum of anthracene paralleled its absorption spectrum, showing the high precision of the excitation spectra obtained. The fluorescence lifetimes were measured with a Hitachi MPF 4 spectro- fluorometer equipped with a 10 ns flash-lamp and processing devices, as described in a previous paper. The triplet-triplet (T-T’) absorption spectra at 77 K were taken either spectrophotographically or photometrically by means of conventional flash excitation. A rectangular sample cell (10 x 20 x 50 mm with a 20 mm path length) fitted with a glass pipe for evacuation was made of Pyrex and its upper half was kept in good contact with a metal jacket filled with liquid nitrogen.The whole part of the cell (the glass cell and metal jacket) was immersed in a transparent Dewar vessel which was also filled with liquid nitrogen. Its level, however, was kept slightly below the bottom of the metal jacket so that the lower half of the cell was in touch with liquid nitrogen but bubbling did not interfere the light path of an interrogating beam. When the triplet lifetimes were > 1 ms, a commercially available standard camera flash was found to be very useful to observe a transient absorption or its decay. When the triplet lifetimes were Q 1 ms, a Q-switched ruby laser (30 ns pulse width) was used as an excitation s ~ u r c e . ~ The decay curves were monitored with a Techtronics 475 oscilloscope. The triplet lifetimes were calculated from the decays of T-T’ absorption monitored at the maximum absorption wavelengths. The phosphorescence spectra were recorded on a home-built phosphorometer which consisted of a Hamamatsu R 928 photomultiplier, a Ritsu ML 20 monochromator and a rotating can with two windows.A 500 W high-pressure mercury arc was used as an excitation light source. The spectra obtained were not corrected for the spectral sensitivity of the apparatus. The phosphorescence lifetimes were also measured with this phosphorometer, but signals were fed in a Union RA-450 data processor and thus smoothed curves were obtained for display on an X-Y recorder. In several cases the triplet-state lifetimes obtained from T-T’ absorption decays were compared with those from the phosphorescence decays in order to establish the identity of the origin of the phosphorescence and T-T absorption. All of the anthracene derivatives except 9-anthryl methoxy ketone were synthesized according to the literature and purified either by sublimation in uucuo or by thin-layer chromatography.Relevant references to the synthesis of the compounds are given in table 1. The solvents used were either of a spectroscopic or a guaranteed grade. The sample solutions were not degassed since no significant oxygen effect was observed at 77 K. The concentrations of the solutions were ca. mol dm-3 for T-T’ absorption and phosphorescence spectrum measurements and ca. 2 x mol dm-3 for fluorescence and lifetime measurements. Thus the excitation spectra obtained are practically free from reabsorption.RESULTS A N D DISCUSSION CLASSIFICATION OF THE COMPOUNDS The eighteen carbonyl compounds examined are classified into six groups based on the nature of the substituents, as given in table 1 . The first group of compounds (called group I etc. hereafter) are derived from 9-X - CO - A, where X is an alkyl group, i.e. 9-anthryl methyl ketone (l), 9-anthryl ethyl ketone (2), and 9-anthryl propyl ketone (3). The compounds in which X is capable of conjugation with the carbonyl group are gathered in group 11. They are 9-anthryl vinyl ketone (4), 9-anthryl phenyl ketone (5), 9-anthryl styryl ketone (6) and 9-anthryl 1 -naphthyl ketone (7). Anthracene- 9-carbaldehyde [X = H, (S)] and 2-methylaceanthren-1-one (9) are placed in group 111, since in these molecules the planes expanded by the carbonyl groups are in a broad sense co-planar rather than orthogonal to the anthracene ring.Thus, from aS. HIRAYAMA 2413 TABLE 1 .-SPECTROSCOPIC PROPERTIES AT ROOM TEMPERATURE OF THE CARBONYL DERIVATIVES OF ANTHRACENE EXAMINED absorptiona fluorescenceb? compound 0-0 band 0-0 band v(C=O) (9-X * CO * A) /nm /nm 77 Kd /cm-l ref. I n JII IV V VI 9-Me - CO - A (1) 9-Pr * CO - A (3) 9-Et * CO * A(2) 9-CH2:CH*CO*A (4) 9-Ph. CO * A (5) 9-C6H5 * CH : CH * CO .A (6) 9-Naph - CO * A (7) (Naph = Naphthyl) (9) (see text) 9-H*CO*A (8) 9-CH2Br - CO * A (10) 9-CH3 - CHBr - CO A (1 1) 9-NH,.CO.A (12) 9-COOH-A (13) 9-CH30. CO A (14) 9-Ph. CO- 10-C 1 * A (15) 9-Ph.CO-1O-NO;A (17) 9-Ph. CO- 1 0-CN * A (18) 9-Ph * CO- 1 0-Br * A (16) 382 382 38 1 382 383 382 384 400 418 383 384 380 380 38 1 394 394 388 406 none none none none none none none 500e 470"g 2o none none 390 388 4 76 none none none 450 387 1703 6 387 1708 6 387 1703 7 none 1663 8 none 1669 9 none 1633 10 none 1660 11 463" 1682 12 430 1703 8 430 1729, 1700 7 none 1703 7 387 387 410 13 14 15 none 1670 16 none 1670 16 none 1670 16 415 1673 17 a In cyclohexane; 'none' indicates that the fluorescence quantum yield is < lop4; in cyclohexane except for compounds 8 and 9 for which ethanol was used; in ethanol; due to the hydrogen-bonded form.The italicised values are for the absorption or emission maxima. molecular-structure point of view, these two compounds are in sharp contrast to the other compounds, where the carbonyl groups are approximately perpendicular to the anthracene ring.In the fourth group, two a-bromoalkyl anthryl ketones [(lo) and (ll)] are listed. These compounds can be derived by the bromination of the corresponding compounds in group I. In contrast to the parent compounds, the compounds thus derived undergo several interesting intramolecular photochemical reactions which are thought to reflect the presence of efficient energy-dissipation channels in the carbonyl derivatives of anthracene.sv l8 In group V we list the compounds which have electron-donating substituents such as amino, hydroxy and methoxy groups, i.e. anthracene-9- carboxamide (12), anthracene-9-carboxylic acid (13), and its methyl ester (14). Compounds of the sixth group are derived from (5) by introducing another substituent at the 10-position.ABSORPTION SPECTRA The wavelengths of the 0-0 absorption bands measured in cyclohexane at room temperature are summarized in table 1. When the absorption spectrum is too broad to determine the position of the 0-0 band unambiguously, the maximum absorption wavelengths are given instead. For any 9-substituted compound other than those in24 14 EX C I TED-S T ATE BE H AVI OUR 0 F SUBSTITUTED AN THR A CE NE group 111, the position of the 0-0 band shifts to the red (by ca. 8 nm) only slightly compared with that of anthracene. This minute substituent effect on the absorption spectrum arises from the large steric hindrance caused by peri-hydrogen atoms which put the carbonyl group in a position perpendicular to the anthracene ring in the ground state, making the conjugation of the C=O double bond with the anthracene group insignificant.Consequently, the substituent effect on the absorption spectra is nothing more than that due to a methyl group.19 On the other hand, in group I11 compounds this conjugation is very effective because their molecular structures favour co-planarity, making the 0-0 band (or the maximum) shift to the red larger (by ca. 30 nm). To illustrate this difference, the absorption spectra of 5, 7 and 8 are shown in fig. 1. In addition to this large red shift, group I11 compounds are characterised by a large solvent effect on the electronic spectra, since the carbonyl groups in these compounds can form hydrogen bonds in protic solvents. In fact the fluorescence in protic solvents is dominated by the hydrogen-bonded forms in these compounds, as will be shown later.A detailed spectroscopic study of 9 has been given elsewhere. 2o wavelength/nm FIG. 1.-Absorption spectra of 5 (-.--), 7 (-) and 8 (----) in cyclohexane at room temperature. The absorption spectrum for 8 is exceptionally broad and red-shifted compared with the other compounds. In contrast, compounds other than group I11 do not exhibit a specific solvent effect on the electronic spectra which might arise from the presence of a carbonyl group within the molecule. By introducing another substituent at the 10-position, the red shift mentioned above is expected to increase in an additive way, as is exemplified by the 0-0 absorption wavelengths of the group VI compounds shown in table 1.The effect is greatest for the carbonitrile group.21 These additional shifts appear to play an important role in the radiationless processes seen in these molecules, as will be discussed later. FLUORESCENCE SPECTRA There has long been discussion as to which kinds of substituents increase or decrease fluorescence efficiency.Z2* 23 The carbonyl group is widely recognized as a quencher of fluorescence rather than an enhancing group. At first glance, this seems to be the case for the compounds listed in table 1. All compounds other than those in groups I11 and V are practically non-fluorescent, at least at room temperature, irrespective ofS. HIRAYAMA 2415 300 350 400 450 500 550 wavelength/nm FIG. 2.-Fluorescence and excitation spectra of 3 observed in EPA at 77 K.For comparison, the fluorescence spectrum of 12 in ethanol at room temperature is shown by the broken line. TABLE 2.-FLUORESCENCE LIFETIME IN EPA AT 77 K compound +sa 1 2 3 9 12 13 14 18 ~~ ~ 11.2 11.4 11.6 9.9 13.1 11.2 ( 1 2. 5)b 11.1 a The standard deviations of zf are 0.1-0.3 ns. The fluorescence from 10 was too weak to determine a reliable lifetime. Except 8, all of the other compounds not listed in this table are non-fluorescent even at 77 K. In cyclohexane at 22 OC. See T. C . Werner, T. Matthews and B. Soller, J. Phys. Chern., 1976, 80, 533. solvent polarity. However, when lowering the temperature, we see that such a simple empirical argument does not necessarily hold. For instance, group I compounds become strongly fluorescent near 77 K.24 As a typical example, the fluorescence spectrum of 3 is shown in fig.2 together with that of 12. The fluorescence lifetimes measured at 77 K in EPA are given in table 2 for most of the compounds fluorescent at 77 K. They are of the order of magnitude expected for anthracene derivatives whose electronic transitions to S , are allowed (with radiative rate constant k , z lo8 s-l), indicating that the fluorescence quantum yields at 77 K are of an appreciable magnitude. In other words, radiationless processes, which are so important at room temperature, are virtually frozen at 77 K. With a conventional fluorophotometer, the lowest detection limit of fluorescence would be a quantum yield2416 EXCITED-STATE BEHAVIOUR OF SUBSTITUTED ANTHRACENE of ca.and hence the lack of fluorescence implies that the fluorescence quantum yield is < lod4. Thus an increase in the fluorescence quantum yield up to, say, 0.5 indicates the enhancement of fluorescence by a factor of 5 x lo3 in going from room temperature to 77 K. With this enormous increase, group I compounds would provide a very interesting case among those compounds whose fluorescence is temperature-sensi tive. 25 9 26 Group I1 compounds exhibit a great contrast to group I compounds with respect to the temperature dependence of fluorescence. The conjugating groups X examined vary widely from the vinyl to naphthyl groups but the derived compounds are always totally non-fluorescent even at 77 K.? How well this holds can be seen by examining the effect on fluorescence caused by breaking a conjugation between the carbonyl group and X, e.g.as in 9-anthryl phenethyl ketone, which is obtained by saturating the a-double bond of 6.1° It behaves like the group I compounds, i.e. non-fluorescent in solution at room temperature but highly fluorescent at 77 K. Thus we see that conjugation of the group X with the carbonyl group is essential to make the molecule totally non-fluorescent even at 77 K. 0. a Q) 2 -fl 3 0.4 9 wavelength/nm FIG. 3.-Absorption, fluorescence and excitation spectra of 8 observed in ethanol at 77 K. The absorption spectrum is shown by the dotted line. The red-shifted and structured excitation spectrum (thin solid line) monitored at 490 nm is caused by the hydrogen-bonded form, whose absorption is not prominent in the absorption spectrum.It was carefully confirmed by examining the excitation spectrum at several concentrations of 8 that the disagreement between the absorption and excitation spectra is not due to experimental artifacts such as inner-filter effects. Group I11 compounds are fluorescent in solution at ambient temperature, but their yields are low20 and the fluorescence arises mainly from hydrogen-bonded forms. To show this, the fluorescence spectrum of 8 and its excitation spectrum are presented in fig. 3 together with the absorption spectrum at 77 K. Apparently the absorption and excitation spectra do not coincide and the red-shifted and more structured excitation spectrum is due to the hydrogen-bonded form. t Note that this holds only for 9-substituted anthracenes.In the case of 1- or 2-carbonyl substituted anthracenes, e.g. 1 - and 2-Me. CO * A or 1 - and 2-Ph * CO * A, appreciable fluorescence has been observed.2’. 28 One of the reasons for this difference might be that the above-mentioned compounds are able to form hydrogen bonds in protic solvents, in contrast to those of groups I and 11.S. HIRAYAMA 2417 Contrary to the expectation that group IV compounds behave like those of group I, the consequence of bromination is remarkable. As has been reported previously,s* 21 both compounds undergo facile intramolecular photochemical reactions ; 10 yielding 9-bromoanthracene and aceanthren- 1 -one as the major photoproducts and 11 yield- ing 4 and 9, whose relative yields depend critically on temperature.Since both reactions of 10 require large structural change, it is not surprising to find that these reactions are completely prohibited in rigid matrices at 77 K. On the other hand, photo- dehydrobromination of 11 to yield 4 occurs readily even in frozen matrices at 77 K. Presumably because of this difference (10) starts to fluoresce weakly near 77 K, but i l l ) is practically non-fluorescent even at 77 K. Certain vibrational modes, which in the extreme lead to the dehydrobromination reaction, may well act as efficient promoting modes of radiationless transition from S, of 11 to the ground state. Group V compounds present another interesting example which illustrates how remarkably the nature of the substituent X controls the fluorescence. All compounds in this group are modestly fluorescent in solution at room temperat~re,'~ in contrast to compounds in the other groups.Taking account of the electron-donating nature of the substituent~,~~ it is quite likely that these groups shift the Tnn* state to a higher energy than that of S,, making the process of Slnn* mTnll* unfavourable energetically. T-T' ABSORPTION SPECTRA A N D TRIPLET-STATE LIFETIMES A T 77 K The T-T' maximum absorption wavelengths and the lowest triplet-state lifetimes determined by decay of the T-T' absorption are listed in table 3. The T-T' absorption spectra of 5 and 9 are compared in fig. 4 to illustrate the spectra in the two extremes with respect to the extent of conjugation to which the carbonyl groups contribute. When a carbonyl group is not in conjugation with the anthracene, the T-T' absorption spectra are similar to that of anthracene and have well-resolved peaks at ca.407 and TABLE 3.-TRIPLET-STATE PROPERTIES OBSERVED IN EPA AT 77 K FOR THE CARBONYL DERIVATIVES OF ANTHRACENE triplet-state T-T' absorption phosphorescence compound lifetime/ms Amax/nm 0-0 band/cm-l 1 2 3 4 5 6 7 8 9 10 12 13 15 16 17 29.5 32.1 34.2 31.3 33.1 34.3 30.7 (29.8)" 1.7 5.9 (5.6)u (20.9)" 39.6 34.5 3.8 0.17 16.05 429 430 430 430 430 430 (broad) 427 (broad) 455 45 5 433 43 1 432 430 43Y - 14 720 14 730 14 740 14 720 14 730 14 740 14 690 13 790 13 830 14 350 14 820 14 760 14 160 14 140* 14 390 a From the phosphorescence decays. * The phosphorescence spectrum was measured by removing the rotating can from the phosphorophotometer.2418 EXCITED-STATE BEHAVIOUR OF SUBSTITUTED ANTHRACENE 1.aJ c .fl 0. s1 I, m '0.. "'%..o. Q.,, - 400 410 420 430 440 450 460 470 480 wavelength/nm FIG. 4.-T-T absorption spectra observed for 5 (-) and 9 (--O--) in EPA at 77 K. The two peaks at ca. 407 and 430 nm are characteristic of the compounds examined here except for 8 and 9, although some may show broader absorptions. 430 nmO3O The lifetimes centre around 30 ms, which is again close to that of anthra~ene.~~? 30 On the other hand, when a carbonyl group comes into conjugation with the anthracene group, the T-T' absorption spectra become broad with largely red-shifted absorption maxima, as is shown by the broken line in fig. 4. Furthermore, the lifetimes become much shorter than those for compounds in groups I, 11, V and VI.Thus again the compounds in group I11 are in great contrast to the other carbonyl derivatives with respect to the shape of the T-T' absorption spectrum and triplet-state lifetime. Those compounds containing a heavy atom such as chlorine or bromine naturally show shorter triplet-state lifetimes owing to the heavy-atom-induced T,-S, transition. PHOSPHORESCENCE SPECTRA AT 77 K The 0-0 band of the phosphorescence in anthracene lies in a long-wavelength region owing to the low energy of the lowest triplet state, T,. The substituent effect on the energy of T, has been thought to be minute and, in fact, the phosphorescence spectra observed for the carbonyl derivatives of anthracene are very similar to that of anthracene in their spectral shape and energy (14927 cm-l for anthra~ene,~~ see table 3) except that the phosphorescence in group I11 compounds are very weak and largely red-shifted.The phosphorescence spectra of 7 and 9 in EPA at 77 K are shown in fig. 5. The observation of phosphorescence is of great significance from the following two points. First, we can locate the T, state and identify its electronic nature. Taking into account the low energies of T, and the similarity between the shapes of the phosphorescence spectra and that of anthracene, it can safely be said that the lowest triplet states of the compounds studied here are of nn* rather than nn* character.t The shortening of lifetimes found in the group I11 compounds may be due to a slight mixing of Tnl* with the lowest triplet state T1,,, , since the coplanar configuration of the carbonyl group against the anthracene ring is favourable for this mixing, as has previously been The lowering in T,-state energies of these compounds (ca.1000 cm-l) would not be large enough to cause a large increase in the non-radiative rate constant k,, of Tlnl*-SO to give such shortened triplet-state lifetimesP2 Secondly, t The energy gap between S, and T, (ca. 1 1000 cm-l) is too large for T, to be assigned as originating from an nn* state.S. HIRAYAMA 2419 A wavelength/nm FIG. 5.-Phosphorescence spectra observed for 7 (-) and 9 ( * - - .) in EPA at 77 K. In either case the separation between the first and second band groups is in the range 1400 cm-’, indicating the electronic states of the lowest triplet states are of KK* character.Compared with 7, the phosphorescence in 9 is weaker, and hence an expanded spectrum (ca. 8 times) is given in this figure. it should be emphasized that most of the rneso-substituted anthracenes studied here give triplet states in appreciable yields even at 77 K. Other rneso-substituted anthracenes with no carbonyl group are reported to yield no triplet state at 77 K;33 i.e. #f is practically unity at this temperature, since the pathway of S,-T, (n = 2 or 3) intersystem crossing is completely shut off at low temperatures because of its endothermicity. Typical examples are 9-methylanthracene and 9,lO-dichloro- anthra~ene.~~ In fact neither phosphorescence nor T-T’ absorption was observed in these compounds with the apparatus employed in the present paper.Thus, the rneso-substituted carbonyl compounds which start to fluoresce at low temperatures are unique among the anthracenes in the sense that their df values at 77 K are still less than unity. CONCLUSIONS We have so far seen how chemical substitution determines excited-state behaviour in compounds of the type 9-X - CO .A. Although numerous cases are known in which chemical substitution affects the kinetics, reactivities, spectroscopic properties, etc. in a way predicted by theory, only a limited number of examples34 are known where a substitution effect on the excited-state behaviour in single-family compounds has been so widely examined as in the present study. The fluorescent properties found in the compounds of groups I-VI are so widely varying that, at a first glance, it may appear to be difficult to provide a unified picture of the radiationless processes pertinent to these compounds.When we look at the results more closely, however, we see that the location of the triplet nn* state could well play a critical role in determining the lack of fluorescence in many of these compounds, although its location has not yet been disclosed spectroscopically for2420 EXCITED-STATE BEHAVIOUR OF SUBSTITUTED ANTHRACENE them. Consequently, in order to explain the non-fluorescent properties seen for several carbonyl derivatives, it is reasonable to invoke a very rapid intersystem-crossing process of the type Slnn*nnryTnn, ,which is faster than the radiative one by at least three orders of magnitude.' A previous picosecond-laser study of intersystem crossing in some carbonyl derivatives of anthracene has revealed that triplet-state population occurs on the picosecond time-~cale.~~ 35 Electron-donating substitutents such as NH, shift the nn* level to higher energy,29 making the process Slnn* 'vvYT,,, less efficient.As a result fluorescence can be observed, as is seen in group V compounds. When X is a conjugating group, it does so with the carbonyl group, although the anthracene part itself cannot come into conjugation because of steric hindrance. The triplet nn* level will then be lowered enough to be located below Slnn*. Thus it would be expected that Slnn*-+Tnn* would become fast enough to make the compounds non-fluorescent even at 77 K. However, when Slnn* is already low enough, as is found in 18, the fluorescence process could start to compete with radiationless processes.21 Another important factor in determining the fluorescence efficiency may be related to rotational motion of the carbonyl group upon photoexcitation, as is implied from the photochemical reactions found in compounds of group IV.The restriction of the rotational motion caused by lowering the temperature would serve to shut off the efficient radiationless transition. For group I compounds this appears to play a dominant role in the appearance of fluorescence at very low temperatures. This problem, however, has already been fully discussed elsewhere. Finally, note that most of the compounds presented here deserve further detailed quantitative study, e.g. by picosecond laser photolysis, because of their unique rapid radiationless processes.It would also be interesting to test the present classification of carbonyl derivatives of anthracene for compounds other than those mentioned here. I am grateful to Prof. K. Hamanoue for the use of a Q-switched ruby laser. I also thank Mr Y. Kajiwara for his help in measuring the triplet-state lifetimes using the ruby laser. S. K. Lower and M. A. El-Sayed, Chem. Rev., 1966, 66, 199. H. Giisten, M. Mintas and L. Klasinic, J. Am. Chem. SOC., 1980, 102, 7936. A. Kearvell and F. Wilkinson, Transitions Non-Radat. Mol., 20th Reunion SOC. Chim. Phys., 1969; J. Chim. Phys., special no., 1970, 125. S. Hirayama, J. Am. Chem. SOC., 1981, 103, 2934. K. Hamanoue, S. Hirayama, T. Nakayama and T.Teranishi, J. Phys. Chem., 1980,84, 2074. E. L. May and E. Mosettig, J. Am. Chem. SOC., 1948, 70, 686. T. Matsumoto, M. Sat0 and S. Hirayama, Bull. Chem. SOC. Jpn, 1975, 48, 1659. 'I R. Calas and R. Lalande, C. R. Hebd. Seances Acad. Sci., 1958, 246, 277. !3 P. H. Gore and J. A. Hoskins, J. Chem. SOC., 1964, 5666. lo H. J. Williams, J, Chem. SOC., Perkin Trans. I , 1973, 1852. l2 L. F. Fieser and J. L. Hartwell, J. Am. Chem. SOC., 1938, 60, 2555. l3 E. L. May and E. Mosettig, J. Am. Chem. SOC., 1948, 70, 1077. l4 H. G. Latham Jr, E. L. May and E. Mosettig, J. Chem. SOC., 1948, 1079. l5 T. C. Werner and D. M. Hercules, J. Phys. Chem., 1969, 73, 2005. P. Rona and U. Feldman, J. Chem. SOC., 1958, 1737. J. W. Cook, J. Chem. SOC., 1926, 1282. P. J. Gore, J. A. Hoskins, R. L. W. Lefevre, L. Radom and G. L. D. Ritchie, J. Chem. SOC. B. 1967. 227. T. Matsumoto, M. Sat0 and S. Hirayama, Chem. Phys. Lett., 1974, 27, 237. M. Martynoff, M. Chauvin, M. Grumez and N. Lefevre. BUN. SOC. Chim. Fr.. 1958, 164. T. Matsumoto, M. Sat0 and S. Hirayama, Chem. Phys. Lett., 1973, 18, 563. 'O S. Hirayama, Bull. Chem. SOC. Jpn, 1975, 48, 1127; 2653.S . HIRAYAMA 242 1 22 Th. Forster, Fluoreszenz der Organischer Verbindungen (Vanderhoeck und Ruprecht, Gottingen, 23 D. M. Hercules, Fluorescence and Phosphorescence Analysis (Wiley, Univ. of Tokyo Press, 1966). 24 T. Matsumoto, M. Sat0 and S. Hirayama, Chem. Phys. Lett., 1972, 13, 13. 25 T. Wismonski-Knitlel, G. Fisher and E. Fisher, J. Chern. Soc., Perkin Trans. 2, 1974, 1930. 28 Pill-Soon Song, Quae Chae, M. Fujita and H. Baba, J. Am. Chern. Soc., 1976,9?3, 819. 27 S. Hirayama, Rev. Phys. Chem. Jpn, 1972, 42,49. 2B T. Tamaki, Bull. Chem. Soc. Jpn, 1978, 51, 2817. 28 R. Ditchfield, J. E. Del Bene and J. A. Pople, J. Am. Chem. Soc., 1972, 94, 703. 30 M. V. Alfimov, N. YaBuben, V. L. Glagolev, E. S. Kuyumdzhi, Yu. V. Pomazan and V. N. Shams- 31 M. R. Padhye, S. P. McGlynn and M. Kasha, J. Chem. Phys., 1956, 24, 588. 32 J. B. Birks, Photophysics of Aromatic Molecules (Wiley-Interscience, London, 1970), p. 147. 33 E. C. Lim, L. D. Laposa and J. M. H. Yu, J. Mol. Spectrosc., 1966, 19, 412. 34 S. Murata, C. Iwanaga, T. Toda and H. Kokubun, Ber. Bunsenges. Phys. Chern., 1972,76, 1176. 35 S. Hirayama and T. Kobayashi, Chern. Phys. Lett., 1977, 52, 55. 1951). hev, Opt Spectrosc., 1977, 42, 267. (PAPER 1/1361)

ISSN:0300-9599    

DOI:10.1039/F19827802411

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1982, 78, 2423-2433 Reaction Kinetics in Acetyl Chemistry over a Wide Range of Temperature and Pressure BY CHRISTOPHER ANASTASI* Shell Research Ltd, Thornton Research Centre, P.O. Box 1, Chester CH1 3SH AND PAUL R. MAW Department of Physical Chemistry, University of Cambridge, Lensfield Road, Cambridge CB2 1EW Received 24th August, 198 1 The molecular modulation spectrometer has been used to study the complex chemical kinetics involved in acetyl radical chemistry. This has involved direct monitoring of both acetyl and methyl radicals in the same experiment and over a variety of temperatures (263 < T / K d 343) and total gas concentration (0.3 < [M]/lO1@ molecule ,< 2.7) conditions. These measurements have been complemented by a non-linear least-squares analysis of the experimental data and simple product studies.Rate data on four reactions and the absorption cross-section of the acetyl radical at 223 nm have been determined in this way. Unimolecular rate theory, based on Kassel integrals, has been applied to the pressuredependent formation and decay of the radical to extract limiting values for the rate constants at T = 303 and 343 K. The acetyl radical is an important intermediate in the slow combustion of hydrocarbons and their cool flames.l In particular, cool-flame models must accurately reflect the competition between the decomposition of this radical and its oxidation, which subsequently leads to an important branching agent, acetyl hydroperoxide. There have been numerous studies on the kinetics of the acetyl radica12-10 based primarily on product analysis and usually on the pressure-dependent thermal- decomposition process.However, these studies have yielded conflicting results. The agreement has been made worse in many cases by the use of Lindemann-Hinshelwood (LH) plots to express the data, with long extrapolations to low- and high-pressure regimes. We present a study of several reactions involving the acetyl radical, using an ultraviolet spectrum obtained with the molecular modulation spectrometer (m.m.s.)l0 to monitor this species directly. The methyl radical, usually present in acetyl chemistry, was also monitored by its absorption in the same spectral region and in the same experiment. The m.m.s. information derived in this way has been complemented by simple product-analysis studies and a computer-based parameter estimation routine for extraction of the kinetic and spectroscopic information.Unimolecular rate theory, based on Kassel integrals, has been used to obtain limiting rate constants for the pressure-dependent reactions. The reactions have been studied over wide temperature (263 < T/K < 343) and concentration (0.3 < [M]/10lB molecule cmU3 < 2.7) ranges. EXPERIMENTAL MOLECULAR MODULATION SPECTROMETER The m.m.s. and the experimental method have been described in detail elsewhere." In the present series of experiments, periodically interrupted photolysis of azomethane (ca. lo1' molecule ~ m - ~ ) in the presence of CO [(0.3-2.7) x 1019 molecule ~ m - ~ ] produced methyl and 24232424 REACTION KINETICS I N ACETYL CHEMISTRY acetyl radicals which were monitored by their absorption at 21 6 and 223 nm, respectively.The concentration of each of the species is modulated, and averaging techniques were used to extract the information. The absorptions led to in-phase and in-quadrature signals relative to the photolysis light and were expressed in reduced form as a function of reduced period of pho to1 ysis. PRODUCT ANALYSIS EXPERIMENTS The apparatus12 has been described before; briefly, six 20 W ‘dark-light’ fluorescent tubes photolysed azomethane in the presence of CO in concentrations similar to those described above. Samples were withdrawn at set intervals and analysed using a Pye Unicam GCD chromatograph equipped with a flame-ionisation detector.A Gas-Chrom Q column coated with 20% w/w dinonyl phthalate was used in all the experiments, and this gave good separation for both the precursor (azomethane) and the products (ethane, acetone and biacetyl). No other products were observed either with this column or with a Porapak Q column, and conservation of CH, groups in each experiment confirmed that no photodissociation of these products occurred. Retention times and calibration curves were obtained by sampling prepared mixtures from the reaction cell by the same method. ANALYSIS OF DATA Differential equations were set up, two of which described the rate of formation of CH, and CH,CO radicals according to the reaction scheme below. Equations were also derived by differentiation of the general in-phase and in-quadrature m.m.s. equations” for each of the radicals.A standard computer package based on Gear’s method was used to solve the equations numerically. Estimation of unknown kinetic and spectroscopic information relied on repeatedly solving the differential equations and on a non-linear least-squares parameter estimation routine applied to fit the experimental data for both radical species. MATERIALS Azomethane was prepared using the method of Renaud and Leitch.13 CO (B.O.C.) was passed through three molecular sieve traps maintained at 195 K. Ethane (Air Products) was taken directly from the cylinder, while acetone (B.D.H.) and biacetyl (Aldrich) were thoroughly outgassed before use. RESULTS The formation and decay of methyl and acetyl radicals in the m.m.s.and product studies are governed by a series of reactions initiated by the photolysis of azomethane in the presence of CO: (1) CH,N,CH,+ hv ( A M 350 nm) --+ 2CH,+N, CH, + CH,( + M) C,H,( + M) CH, + CO( + M) + CH,CO( + M) (3) CH, + CH,CO --+ CH,COCH, (4) CH,CO + CH,CO + (CH,CO), ( 5 ) ( 6 ) CH,CO( + M) -+ CH, + CO( + M) where M denotes a diluent gas (CO in the present series of experiments). The rate constants for reactions (3)-(6) and the absorption cross-section for the acetyl radical, 0, are unknowns. The reactions are interdependent, making parameter estimation difficult even with kinetic information from the m.m.s. on both CH, and CH,CO radicals. Product analysis was used to reduce the number of unknowns as well as to confirm the reaction scheme used.C.ANASTASI A N D P . R. MAW 2425 RELATIONSHIP BETWEEN k,, k, AND k, The rates of formation of the products are given by d[C,H,]/dt = k,[CH3I2 = E d[CH,COCH,]/dt = k,[CH,] [CH,CO] = A d[(CH,CO),]/dt = k,[CH,CO]z = B where square brackets denote concentration, t the time and k the rate constant. These expressions may be arranged to give k, = A(k, k,)$/(BE)$ or k4/(k2 k5)i = A/(BE)i. k, is knownll and A/(BE)$ was determined using the gradients of the product profiles. r photolysis timelnumber of lamps X min FIG. 1.-Product analysis of the photolysis of azomethane (10.3 x 10l6 molecule ~ m - ~ ) in CO ( 1 . 1 x 1019 molecule ~ m - ~ ) . Fig. 1 shows a typical experiment, with the products ethane, acetone and biacetyl increasing as the azomethane was consumed.The experiments were carried out at room temperature under a variety of azomethane and CO concentrations and the results are summarised in table 1 ; as expected, the constant was independent of the concentrations of both gases. Ten separate experiments yielded a mean value for k,/(k,k,)i of 1.73 0.06. P A R AMETER-EST1 MATION RESU L T S The estimation routine was used to fit the experimental data from the m.m.s. to extract values for k,, k,, k, and a(acety1) while k, was measured experimentallyL1 as2426 REACTION KINETICS IN ACETYL CHEMISTRY 1 .oo- 0.90 0.80 0.70 0.60 Q s 2 0 . 5 0 - ou I 0.40 0. 30 0.20- 0.10- TABLE I.-k,/(k2k,)4 AS A FUNCTION OF AZOMETHANE AND CO CONCENTRATION - - - - - - r [azornethaneJ/ 10ls [CO]/lOl* molecule ~ r n - ~ molecule ~ r n - ~ k,/(k2 k6)i 6.74 6.91 7.00 7.03 6.97 6.97 9.78 10.00 10.01 10.05 3.37 3.43 11.37 15.80 24.46 24.59 3.43 11.50 15.80 24.46 1.73 1.66 1.68 1.77 1.80 1.72 1.70 1.66 1.80 1.83 x I 0 0 0 FIG.2.-Predictions of the least-squares estimation routine (-) fitting the experimental points for acetyl ( x , in-phase; 0, in-quadrature) where [azomethane] = 9.7 x 1Ol6 molecule and [CO] = 2.5 x lo1@ molecule an+. (See text for definition of A and B.)C. ANASTASI AND P. R. MAW 2427 before; k , and a(methy1) are knownll and k, can always be related to k,. The solid line in fig. 2 is the result of this procedure for an acetyl run; in addition to the signal count and photolysis period, z, the axes contain the groups A and B where and A = $Fl(a/k,) B = (x, NLk, k2)i. Here F is the calibration factor relating count rate to absorption, I is the cell length, x, the azomethane concentration, a the absorption cross-section for the methyl radical and NL the number of photolysis lamps.Table 2 summarises all the kinetic and spectroscopic data derived in this way for 263 < T/K < 343 and 0.3 < [CO]/1019 molecule cm-, < 2.7. TABLE 2.-KINETIC AND SPECTROSCOPIC DATA DERIVED USING THE PARAMETER ESTIMATION ROUTINE AND k, = 1.73 (k,k,)i [CO]/1OLo o(CH,CO)/lO" k,/lO-I8 k,/10-I1 k,/ lo-" (k,/ks)/lO's T/K molecule ~ m - ~ cm* molecule-' cms molecule-' s-I cm8 molecule-' s-' cms molecule-' s-' k,/s-' molecule ~ m - ~ 263 2.1 219 1.9 290 2.0 295 2.1 303 0.3 0.6, 1.1 1.1 1.6 1.6 1.9 2.4, 2.4 2.4, 1.7, 2.2 343 0.3 0.6 1.1 1.6 2.0 2.1 323 1.1, 2.2, mean confidence - limits (%) 1 .o I .o 1 .o 1 .o 1 .o 1 .o 1 .O 1 .o 1 .o 1 .o 1 .o 0.8 I .o 1.1 I .o 1 .o 1 .o 1 .o 1 .o 1 .o 1 .o 1 .o 1 .o 1 .o 18, +30 - 1.3 2.0 3.5 4.4 3.6 4.2 4.I 5.5 5.6 7.4 6.6 1.6 1.3 1.9 I . 1 10.1 10.4 15.0 6.1 8.4 12.0 16.0 15.0 16.6 15, + I 5 4.0 4.0 4.0 4.0 4.0 4.0 4.0 4.0 4.0 4.0 4.0 3.3 4.2 4.2 3.6 3.6 3.8 3.6 3.5 3.5 3.5 3.5 3.5 3.6 - 1.3 1.3 1.3 1.3 1.3 1'.3 1.3 1.3 1.3 1.3 1.3 0.9 1.5 1.5 1.1 1.1 1.2 1.1 1.1 1.1 1.1 1.1 1.1 1.1 - 30, + 50 13 16 19 30 31 21 40 55 49 52 21, -28 5.1 3.2 3.3 3.4 3.3 3.1 The experimental runs for T>303 K at the highest pressures, usually ca. 2 x 1019 molecule ~ m - ~ , were the most reliable with radical absorptions of ca. 3 x lo-, at 1 s photolysis period and 100 s experimental time. As the pressure was reduced the acetyl signal declined and the methyl signal increased; at 3 x 10l8 molecule cm-, the acetyl signal was only just above the limits of detectability.Therefore, for [CO] < 2 x lOl9 molecule ~ m - ~ , k,, which may be assumed to be pressure-independent under our experimental conditions, and a(acety1) were set to the values derived at the highest concentration, allowing only k, and k, to vary. As the temperature was reduced the acetyl signal again declined and the methyl signal increased. For T < 295 K, the room-temperature values for k, and a(acety1) were used, allowing only k, to vary. The acetyl signal at T = 263 K and [CO] = 2.7 x 1019 molecule CM-~ was again at the limit of detectability and set a lower limit on the temperature.2428 REACTION KINETICS I N ACETYL CHEMISTRY It was not possible to obtain an estimate for k6 at room temperature and below because of its small influence on the overall chemistry.The maximum reported value for k6 at room temperature is 4 s-l at the high-pressure limit.' A value of 2 s-l, for our experimental conditions, leads to an increase of only 1 % in k , when 0 and k, are decreased by 5 and 12%, respectively, at the highest concentrations. These changes are well within the mean uncertainties for the rate constants, which are also included in table 2. DISCUSSION THE CONSTANT k,/(k,k,)i The value of 1.73 derived from the product-analysis study at room temperature is in excellent agreement with the value of 2 predicted from the simple cross-combination rule for reactions (2), (4) and (5).Adachi et aL9 also report a value of 2, but with significantly higher values for k, and k, than those derived in the present work (4.0.). A value of 1 has been reported by Gandini and Hackett,8 although these authors suggest that the source of deviation from the simple cross-combination rule may be experimental; an error of only 10% in their azomethane actinometry would increase their value for this ratio from 1 to 2. REACTION OF CH, WITH CO The variation of k , with total gas concentration is shown in table 2. The mean value of 7.6 x 10-l8 cm3 molecule-l s-' at the highest concentration and at 303 K is in very good agreement with the value of (6+ 1) x 10-l8 reported by ParkeslO for similar conditions. Although this preliminary study used the same m.m.s.apparatus, a quite different kinetic and spectroscopic analysis was used. Parkes also carried out experiments at 3 x 10l8 molecule cm-,; however, his analysis did not allow for the lamps being on for only half the time. This manifests itself in a k, value raised by a factor of 2, i.e. from 1.8 x cm3 molecule-l s-l, which is identical to that derived in the present study. The level of agreement may, however, be fortuitous considering the limitations of both sets of data at this concentration; the present study relies to a large extent on methyl signals only and Parkes' value of k, is the result of a long extrapolation. Clearly, at our concentrations the reaction is passing through its transition region between third- and second-order kinetics, and rate constants in the limit of low pressure (k:) and high pressure (k?) cannot be obtained directly from the experimental results.To find k: and k p , the empirical procedure of Troe14 has been used, which relies on matching the experimental pressure dependence by curves that are calculated by evaluating Kassel integrals. This procedure has been outlined in detail by Troe and used for association reactions in other l6 Fig. 3 shows the result of this exercise for the data at T = 303 and 343 K. Table 3 gives the limiting values and compares them with those reported in the open literature. The absolute value for k? at 303 K is considerably higher than that given by Ker-r and Calvert., However, these authors did their experiments with a maximum total gas concentration of < 8 x l0l8 molecule ~ m - ~ , relying on a Lindemann-Hinshelwood (LH) straight-line extrapolation to the high-pressure limit.Such plots are known to be curved over a wide pressure range, so that extrapolation from a low-pressure regime would lead to a considerable underestimate of k?. Watkins and Ward7 have also used an LH extrapolation to obtain k?, but these authors conducted their experiments at room temperature and in up to 7 x 1019 molecule cm-,, of which 90% was SF,, a highly efficient collision partner. They were therefore much closer to the high-pressure limit, to 3.6 xC. ANASTASI AND P. R. MAW 2429 - 1 0 1E I I I I ) 18.5 19.0 19.5 20.0 log,, ([CO] /molecule ~ r n - ~ ) FIG. 3.-Comparison of calculated fall-off curves with experimental k, data: (a) T = 343 K, Sk = 5.3, B k = 19.5; (b) T = 303 K, Sk = 4.7, B, = 19.4.&(CO) = 0.3. [These terms are defined in ref. (14).] 0, Experimental points. TABLE 3.-RATE CONSTANTS FOR CH3 + CO ( 4- M) -+ CH3C0 ( + M) CH,+CO+M -+CH,CO+M k!/ cm6 k,"/ 1O-l' cm3 T/K s-l molecule-' s-l ref. - 298 0.1 4 1.2 7 303 4.1 7 343 - 5.7 1.6 this work 10.7 5.7 this work -2430 REACTION KINETICS I N ACETYL CHEMISTRY which was confirmed by a much higher value of k?. Even in this study, however, the maximum value attained for k3/kF was estimated by these authors to be 0.7, still some way short from the high-pressure limit. The k? values at T = 303 and 343 K suggest an activation energy of 6.6 kcal mol-l. Also, if the data at constant total concentration of 2 x lOlS molecule ~ m - ~ in the temperature range 263 < T/K -c 343 are expressed in an Arrhenius form (fig.4), an activation energy of 6 kcal mol-l and pre-exponential FIG. 4.-Arrhenius plot 2 . 8 3.0 3.2 3.4 3.6 3.8 4 . 0 for k, at a constant pressure of [CO] x 2 x 1019 molecule an-,. (The point at 103 KIT lo3 K/T = 3.8 is at 2.7 x lOI9 molecule and3.) factor, A , of 1.2 x cm3 molecule-l s-l are evaluated. The activation energy is in excellent agreement with thosederived by Watkins and Ward,7 who report 6 kcal mol-l, and Kerr and Cal~ert,~ who suggest ca. 5 kcal mol-l, which was 1 kcal mot1 higher than their experimental value derived at low pressures. The pre-exponential factor, however, is lower than the 2.6 x 10-13 cm3 molecule-' s-l derived by Watkins and Ward.CH3C0 DECOMPOSITION Table 4 compares the results of this study with previous work and where possible under similar conditions of pressure and temperature. The early product-analysis s t ~ d i e s ~ - ~ are omitted, since these were done at low total gas concentrations ( < 5 x lOls molecule cmd3), and relied on LH plots to obtain the high- and low-pressure limits. As indicated in the previous section, such plots lead to an underestimate of both limits. Similarly, the study by Frey and Vinal15 at 326 K suffers from the same limitations, with their k? expression derived from data at < 1.3 x 10l8 molecule ~ m - ~ . However, table 4 shows reasonable agreement with one run conducted by these authors at a higher total concentration of 2 x 1019 molecule ~ m - ~ .There are two values, differing by two orders of magnitude, for k? at this Both are derived using LH plots, although once again Watkins andC. ANASTASI AND P. R. MAW 243 1 TABLE 4.-RATE CONSTANTS FOR CH,CO (+ M) + CH, + CO ( + M) CH,CO + M + CH, + CO k,/s-l ref. W l / T/K molecule cm-, 323 00 0.04 6 00 36.0 7 2.2 x 1019 24.5 this work 326 2.0x 1019 14.0 5 343 00 0.26 6 00 172 7 m 186 this work 338 k: = 1.6 x cm3 molecule-' s-' 4 343 k: = 3.1 x cm3 molecule-' s-l this work Ward7 conducted their experiments at higher total concentrations. Our value of 24.5 s-l at 2.2 x 1019 molecule cm-, favours the k? value of 36.0 s-l derived by these authors. The empirical procedure based on Kassel integrals can also be applied to the acetyl decomposition data at 343 K to evaluate k: and k?.The rate constants are less reliable than the corresponding k, values, so that 30% error limit's are realistic bounds for k! and k? given in table 4. The k, value at 3 x lOls molecule cm-, was not included in this analysis since the k,/k, equilibrium constant is considerably higher than that at higher concentrations, as shown in table 2. The difference is probably due to low acetyl signals at the lowest total gas concentrations. Values of 2.9 x 10l8 and 3.3 x 10l8 molecule cm-, for kt/k! and k?/kF, respectively, suggest the lower equil- ibrium constant. Once again, the k p value derived in this study favours the higher value of Watkins and Ward. The only other reported limiting low-pressure rate constant is that of Kerr and C a l ~ e r t , ~ who did their experiments at ca.338 K. Their value of 1.6 x is a factor of 2 lower than that reported here, the difference highlighting the limitations of the LH approach. THERMOCHEMISTRY The equilibrium constant k,/k, at T = 343 K in table 2 can be used to evaluate the standard free energy change, AGe, for CH,CO (+M) + CH,+CO (+M) using AG* = -RTln(k,/k,). Ignoring the data at 3 x 10l8 molecule cm-,, a mean value for k,/k, of 3.3 x 10l8 molecule cm-, gives A G e = 1.3 kcal mol-l. Table 5 compares this with A G e values derived using the expression AG* = A H e - T A P , where A H 0 and AS* have been r e p ~ r t e d . ~ l l ~ - l ~ The values for A G e range from 1.2 to 4.2 kcal mol-l. Table 5 also shows that 80% of the difference in these two values is due to the difference in AH* for the reaction, and, in particular, to the difference in AHfe (CH,CO) which ranges from -4.17 to -5.817 kcal mol-l.When normalised to the same AHf* (CH,) value,19 this range becomes - 3.8 to - 5.8 kcal mol-l. Other literature values include -4.0 f 2.04 and 5.1 f 2.02* kcal mol-l. Although our measurements at 343 K are limited our results are consistent with the lower values for A H F (CH,CO).2432 REACTION KINETICS IN ACETYL CHEMISTRY ks k3 TABLE 5.-THERMOCHEMISTRY FOR THE REACTIONS CH3C0 ( 4- M) CH, + CO( + M) AHe/ A S e / AGe/ AHfe(CH,CO)/ sp/ kcal cal kcal kcal cal mol-l mol-I K-l mol-la mol-I. mol-l K-' ref. 11.7k0.7 30.6 1.2 -4.1 f 1.2 62.6 7 - -3.81f: 1.2 63.0 7b - - 13.3 29.1 3.3 - 5.4 64.5 19 14.2 29.1 4.2 - 5.8 & 0.4" 64.5 17 13.7 29.1 3.7 - - 1 7d this work - - 1.3" - - a AGe = AHe- T A P ; T = 343 K.Recalculated using AH, and S* values for CH, and Recalculated CO given in ref. (19). AH, (CH,CO) quoted in ref. (17) is taken from ref. (18). using AH, and S* values for CH, given in ref. (19). AGe = -RTln (k6/k3); T = 343 K. THE CH,-CH,CO AND CH,CO-CH,CO REACTIONS There have been two recent measurements for k , and k5,9v10 both yielding higher values than those given in table 2. ParkeslO in his preliminary study has reported values of 7.0 x 10-l' and 3 x cm3 molecule s-l fork, and k,, respectively. The agreement with the present study is fair and is slightly better if the assumed cross-combination constant of 2 used by Parkes is replaced by our measured value of 1.7. Adachi et aL9 report considerably higher values of 1.3 x and 7.5 x 10-l' cm3 molecule-l s-l for the same rate constants.It is possible to estimate k , from the known decomposition rate of biacetylZ0 at 730 K, if the equilibrium constant for k-5 k5 CH,COCOCH, + 2CH,CO is calculated using known thermodynamic quantities. This exercise has been carried out before by Szirovicza and Walsh;6 using AH,e = - 4.1 kcal mol-1 and S e = 63.3 cal mol-l K-' for CH,CO and the values of Szirovicza and Walsh for AHF and S e for biacetyl (- 78.0 kcal mol-1 and 86 cal mol-1 K-l, respectively) leads to k, = 9.4 x cm3 molecule-' s-' at 730 K if k-, is 7.6 x lo-, s-l. Assuming zero activation energy for k,, this is close to the value derived in the present study. It is also instructive to compare the rate constant for the combination of acetyl radicals with that for the combination of methyl radicals. Simple collision theory, assuming diameters d(CH,) = 3.8 A and d(CH,CO) = 4.4 A, shows the acetyl reaction to be characterised by a 20% smaller collision number.Also, in practice, the rate constant for the mutual reaction of alkyl radicals decreases as the size of the species involved increases.21 For example, the rate constant for the combination of ethyl radicals is a factor of 3 lower than the corresponding reaction involving methyl radicals. There is good evidence then to support the rate constants k , and k , observed in this study. These values are smaller than those assumed in several product-analysis studies2* 57 * at room and elevated temperatures.Assuming zero activation energy for both rate constants, the present work suggests that the results from these studies may need to be reinterpreted.C. ANASTASI AND P. R. MAW 2433 THE ACETYL ABSORPTION CROSS-SECTION AT 223 nm Parkedo has reported a value of 1 x cm2 molecule-' for a(225 nm), derived using k , = 3 x lo-'' cm3 molecule-l s-'. If our value for k , is used, a = 0.7 x lo-'' cm2 molecule-' is calculated, which is in good agreement with our value of 1 x cm2 molecule-' s-l. This is, however, a factor of two lower than that reported by Adachi et al.9 at the same wavelength. As in the case of k, and k,, the cause of the difference between the two studies is not obvious at this time. We thank Mrs D. U. Hancock for conducting most of the experimental work, Mr A. Prothero and Dr R. M. Furzeland for their help in the computer analysis of the data and Dr D. R. Blackmore for useful comments. P. M. thanks the S.R.C. for a CASE award. M. P. Halstead, A. Prothero and C. P. Quinn, Proc. R. SOC. London, Ser. A , 1971, 322, 377. J. G. Calvert and J. T. Gruver, J. Am. Chem. SOC., 1958, 80, 1313. H. O'Neal and S. W. Benson, J. Chem. Phys., 1962, 36, 2196. J. A. Kerr and J. C. Calvert, J. Phys. Chem., 1965, 69, 1022. H. M. Frey and I. C. Vinall, Int. J. Chem. Kinet., 1973, V, 523. L. Szirovicza and R. Walsh, J . Chem. SOC., Faraday Trans. I , 1974, 70, 33. K. W. Watkins and W. W. Ward, Int. J. Chem. Kinet., 1974, VI, 855. A. Gandini and P. A. Hackett, J. Am. Chem. Soc., 1977, 99, 6195. H. Adachi, N. Basco and D. G. L. James, Chem. Phys. Lett., 1978, 59, 502. lo D. A. Parkes, Chem. Phys. Lett., 1981, 77, 527. l1 D. A. Parkes, D. M. Paul and C. P. Quinn, J. Chem. SOC., Faraday Trans. I , 1976, 72, 1935. l 2 L. J. Kirsch and D. A. Parkes, J. Chem. SOC., Faraday Trans. I , 1981, 77, 293. l 3 R. Renaud and L. C. Leitch, Can. J. Chem., 1954,32, 545. l4 J. Troe, Ber. Bunsenges. Phys. Chem., 1974, 78, 478. l5 C. Anastasi and I. W. M. Smith, J. Chem. SOC., Faraday Trans. 2, 1976, 72, 1459. l6 C. Anastasi and I. W. M. Smith, J. Chem. SOC., Faraday Trans. 2, 1978, 74, 1056. l7 R. F. Hampson and D. Garvin, Natl Bur. Stand. ( U S . ) , Spec. Publ., 1978, 513. l8 J. A. Devore and H. E. O'Neal, J. Phys. Chem., 1969, 73, 2644. l9 S. W. Benson, Thermochemical Kinetics (Wiley, London, 1976). 2o K. J. Hope and M. F. R. Mulcahy, J. Phys. Chem., 1969,73, 177. D. A. Parkes and C. P. Quinn, J . Chem. SOC., Faraday Trans. I , 1976, 72, 1952. (PAPER 1 / 1362)

ISSN:0300-9599    

DOI:10.1039/F19827802423

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1982, 78, 24352446 Interaction of CO with Ni( 100) Adsorption Isotherms and Thermal Decomposition BY FREEK LABOHM,* CHARLES W. R. ENGELEN, ONNO L. J. GIJZEMAN, JOHN W. GEUS AND GOSSE A. BOOTSMA~ Van't Hoff Laboratory, University of Utrecht, Padualaan 8, 3584 CH Utrecht, The Netherlands Received 28th September, 198 1 The reversible adsorption of CO on Ni(100) in the temperature range 100-150 "C has been investigated with ellipsometry and Auger electron spectroscopy, as has the thermal decomposition at higher temperatures. From thermodynamic data the (coverage-independent) heat of adsorption was found to be 134 & 4 kJ mol-'. An isotherm equation is given, valid for temperatures up to 225 O C , consistent with the hexagonal structure of the CO adlayer at saturation.Decomposition, which occurs measurably above 150 "C, leads to the deposition of carbon and is influenced by an operating ion gauge. Decomposition occurs according to CO(a) + CO(g) -. CO, + C(a). The rate of disappearance of CO has a small pressure dependence that is explained by a mechanism in which the formation of a highly reactive CO(a) molecule from adsorbed CO is rate-determining. For this step the availability of a site not poisoned by carbon is necessary. It is shown that the ion gauge reduces the number of sites poisoned by one carbon atom. The reaction rate for thermal decomposition has an activation energy of 130 kJ mol-l. In this paper investigations on the adsorption of CO on Ni( 100) below 150 OC and the decomposition of CO above this temperature are described.The results of many previous studies of CO adsorption on Ni single crystals cannot readily be interpreted, because no attention was paid to the cleanliness of the surface and to the electron- beam-stimulated decomposition of CO during low-energy electron diffraction (LEED) and Auger electron spectroscopy (AES) experiments. 1-4 Onchi and Farnsworth3 and later Tracy5 were the first to investigate the influence of small amounts of surface carbon on the adsorption of CO. It is our aim to determine the heat of adsorption of CO on a Ni(100) surface that is free of carbon not only before but especially during the adsorption. To meet this condition the measurements from which thermodynamic data are obtained must be made as quickly as possible to avoid the build-up of carbon becoming detrimental.We therefore measured the extent of adsorption ellipsometrically. Since this technique allows a very rapid measurement to be made, the carbon deposition remains limited to an extent that does not markedly effect-the adsorption. Moreover, the light beam of the ellipsometer does not influence CO during or after adsorption. The interaction of CO with Ni(lO0) surfaces at higher temperatures ( T 2 175 "C) has not been much investigated, despite its importance in the production of methane from carbon monoxide.6 That CO decomposes above 175 OC has been stated by several authors5-13 without mentioning further details about the mechanism. According to our work the build-up of carbon from CO and the influence of carbon on the t Permanent address: Department of Applied Physics, Twente University of Technology, P.O.Box 217, 7500 AE Enschede, The Netherlands. 243 5 19-22436 INTERACTION OF co WITH Ni(l00) adsorption of CO are complicated. They depend on pressure, crystal temperature and the operation of an ionisation gauge. This leads to the conclusion that both adsorbed and gaseous CO are involved. EXPERIMENTAL The experiments were carried out on a Ni(100) single crystal, mounted in a Varian U.H.V. system equipped with facilities for argon-ion bombardment and simultaneous Auger electron spectroscopy and ellipsometry. The base pressure was cu. Pa. Disc-shaped crystals (diameter 8 mm, thickness 1 mm) were spark-cut from a 5N nickel rod within 0.5 O of the (100) orientation.The base pressure was cu. Pa. The angle of incidence of the laser beam of the ellipsometer (A = 632.8 nm) was 68 O. Only off-null irradiance measurements were performed. AES was used to measure the carbon coverage before and after CO exposure. Use was made of a Varian cylindrical-mirror analyser with an on-axis electron gun, adjusted to yield an anode current of 400 pA at an electron energy of 2 keV. A modulation frequency of 17 kHz and a modulation amplitude of 10 V peak-to-peak were used. The cleaning procedure consisted of sputtering at room temperature in an argon atmosphere of 6.6 x Pa for 1 h with 450 eV Ar+ ions at 10 pA cm-2 and at an angle of incidence of 45 O. After this the crystal was annealed at 6OOOC for 10min. The gases used, i.e. CO (purity 99.99%), C,H, (99.95%) and argon (99.999%), were supplied by 1'Air Liquide.The pressure in the vacuum chamber was measured with a Varian nude-ionisation gauge not in line of sight with the nickel crystal. RESULTS ADSORPTION OF co ON Ni(100) The change in the ellipsometric parameter A [6A = A - A, A corresponding to a clean Ni(100) surface] was used to monitor the CO coverage during CO expositions and was determined by means of off-null measurements. The change in the other ellipsometric parameter v/ was too small for accurate determination. The increase in 6As (equilibrium value of 6A) at crystal temperatures varying from 100 to 150 O C was followed during a stepwise raising of the pressure from 1.3 x to 0.1 Pa until BAS did not increase further with pressure.This latter saturation value is denoted by 66,. By applying this procedure, measuring ca. 10 points of one isotherm took ca. 20 min at a crystal temperature of 100 O C and 5 min at 150 OC. At the lower temperature the limiting factor was the time required for the establishment of equilibrium, while at the higher temperature it was the time required to adjust the CO pressure. Below 100 O C 6Am was already reached at CO pressures near the base pressure, and therefore no reliable isotherms could be obtained. Evacuation, causing 6A to return to zero, shows that the adsorption was completely reversible in this temperature range. After prolonged evacuation (5-120 min depending on the crystal temperature) at < Pa, when 6A had become zero, Auger spectra reveal a small amount of surface carbon [always < 0.05 monolayer (ML)] increasing with the crystal temperature at which the CO exposure had been made and with the duration of the experiment.Oxygen was present at coverages < 0.01 ML. Above 150 "C the rate of carbon deposition was too large to measure a complete adsorption isotherm. As expected 6As was a function of temperature and pressure only. However, 6A, was scattered between 0.28 and 0.32 O, independent of temperature. The isotherms shown in fig. 1 were rescaled to SA, = 0.30 *. The low-coverage parts of the isotherms could not be measured at all temperatures, because the corresponding equilibrium pressures were too close to the background pressure. At constant pressure 66, increased with decreasing temperature and at a lower temperature 6Am was reached at lower pressures.F.LABOHM et al. 2437 3.6 0.4 e 3.2 I 1 I I :0-7 10-6 10-5 10-4 PcolPa FIG. 1.-Adsorption isotherms of CO on Ni(100); 6As and 0 as a function of CO pressure at the temperatures indicated (the pressure was increased until 66, became pressure-independent) ; 0, calculated according to eqn (4). DECOMPOSITION OF CO ON Ni(100) The decomposition of CO adsorbed on Ni( 100) was monitored ellipsometrically at temperatures between 175 and 225 OC and at pressures varying from to 0.1 Pa. Below 175 OC the decomposition rate was very small. Above 225 OC at pressures up to 0.1 Pa the CO adsorption was difficult to measure because of the small extent of adsorption and the very rapid decomposition (see fig. 3).In order to obtain a reliable carbon coverage the carbon 272 eV peak height (h,) in the differentiated spectrum was first corrected for the slope of the background14 and then divided by the Ni 848 eV peak height (hNi) to eliminate variations in the electron-gun and electron-multiplier currents. The Ni 848 eV peak height was virtually constant for sub-monolayer carbon coverages. After each AES measurement CO was readmitted and the decomposition was observed to proceed as if no evacuation had taken place. After the system was evacuated, no change in A took place (see fig. 2). Carbon deposition from CO lowered the A value of the CO free surface. The time FIG. 2.-Decomposition run, 66 as a function of time, schematically represented. Interrupted by evacuation (*) to determine 8, by AES .6A consists of 2 components, 6Aco and 6Ac.2438 INTERACTION OF co WITH Ni(100) 0.3 1 I I 1 0 .0 a" Lo 0.1 0 10 20 30 t/min FIG. 3.-Decomposition of CO at 6.6 x lop4 Pa and the temperatures indicated; 6Aco as a function of time. measured 6A value is thus the algebraic sum of a positive contribution owing to adsorbed CO (6Aco) and a negative contribution owing to deposited carbon (6Ac), 66, being the change in A with respect to the clean surface (see fig. 2). The value of 6Aco was therefore calculated by adding to the measured dA the value of 66 corresponding to the CO-free carbon-covered surface. The latter initially varied linearly with time. Fig. 3 shows the decomposition of CO (dAco against time) at a CO pressure of 6.6 x Pa and different crystal temperatures.With increasing temperature the initial values of 6Aco decrease, whereas the rate of CO removal [d(6A)/dt] increases. At all temperatures above 175 OC the CO coverage (6Aco) eventually decreased to zero and carbidic carbon12 was deposited on the surface. The amount of oxygen never exceeded 0.01 ML. The effect of CO pressure on the decomposition of adsorbed CO at 175 OC is shown in fig. 4 (for higher temperatures the behaviour is qualitatively the same). The influence 0.2 0 . 8 m 0.1 0 10 2 0 3 0 t/min FIG. 4.-Decomposition of CO at 175 OC; 6Ac0 as a function of time: (-) ion gauge on; (---) ion gauge off. A, 0, according to eqn (13) and the data in table 1 . Values of pco/Pa as follows: (a) 6.6 x lo+, (6) 6.6 x (c) 6.6 x ( d ) 6.6 xF. LABOHM et al.2439 FIG. 5.-Influence of the ion gauge on the relation between the absolute value of dAc and hc/h,, derived from interrupted decomposition runs as in fig. 2 at several temperatures and pressures. A, ion gauge off; 0, ion gauge on. of the ionisation gauge on the decomposition is evident from a comparison of the solid curves (filament on) and the dashed curves (filament off). Surprisingly, at all pressures the CO coverage decreased faster with the ionisation gauge off. The ionisation gauge also influences the relation between dAc and hc/hNi. When the filament is switched off during the decomposition dAc is larger for a given amount of carbon than with the gauge operating. This is illustrated in fig. 5. With the information thus obtained O,, against 0, diagrams were constructed.They are shown in fig. 6. The CO coverage was calculated by assuming: (i) dAm = 0.30 O corresponds to a COcoverage of 0.69 ML59 and (ii) dAco = A Oc0 with A independent of temperature and carbon coverage. The carbon coverage was calculated using hc/hNi = 0.17 for 0.50 ML of carbon obtained by the thermal decomposition of C,H, on Ni(lO0).l4$ l5 The solid curves were determined with the ionisation gauge switched on, the dotted lines with the gauge off. After dAco had reached zero we could not exceed 0.5 ML of carbon even by prolonged exposures at 175 O C at CO pressures of 0.4 0 m 0.2 ( I I I I I ' \ . b\\\ \ \ . \ *C FIG. 6.-f?co as a function of OC at 175 OC: (-) ion gauge on; (----) ion gauge off. Values ofp,,/Pa as follows: (a) 6.6 x low4, (6) 6.6 x2440 INTERACTION OF co WITH Ni(l00) Oe3 * 0 1 0 2 0 3 0 flmin FIG.7.-Effect of switching on the ion gauge. Decomposition of CO at 6.6 x Pa and 175 O C : (1) ion gauge on; (2) ion gauge switched on after 6 min; (3) ion gauge off. 10-3-0.1 Pa or 0.25 ML at pressures < Pa. The existence of two pressure ranges was evident at all temperatures between I75 and 225 OC, whether the ionisation gauge was on or off. At higher temperatures, pressures > Pa were required to establish a saturation coverage of 0.5 ML. The filament only influenced the shape of the curve in the 8,, against 8, diagram (see fig. 6). Further evidence for the existence of ionisation-gauge effects is shown in fig. 7. Initially the filament is switched off. When, after several minutes the gauge is switched on, a sudden decrease in the gradient occurs.The curve then continues parallel to the one obtained with the ionisation gauge on. Exposing a Ni(100) crystal to ethylene at 6 x lop5 Pa and at temperatures above 200 OC results in the deposition of 0.5 ML carbidic carbon with ap4g structure.15 This feature was used to determine whether the ionisation gauge has any influence on the structure of the carbon deposit from CO. When < 0.5 ML of carbon was deposited from CO with the ionisation gauge switched on, this amount could always be made up to 0.5 ML by means of ethylene exposure. If the initial amount had been deposited with the ionisation gauge switched off, C,H, exposure always resulted in a saturation coverage of < 0.5 ML.Not only the mechanism, but also the rate of carbon deposition is influenced by the ionisation gauge. Although over the same lapse of time the change in 6Aco was larger (see fig. 4), less carbon was deposited when the ionisation gauge was off. This effect was independent of temperature and pressure, DISCUSSION ADSORPTION OF CO ON Ni(100) By applying the Clausius-Clapeyron equation to the isotherms shown in fig. 1 we were able to evaluate the isosteric heats of adsorption. The corresponding Clausius- Clapeyron plots are shown in fig. 8. The isosteric heat of adsorption - A H(66,) = - R[d lnp/d( 1 / T)] is shown in fig. 9 as a function of 6As and 0. As can be seen, the heat of adsorption of CO on Ni(lO0) is independent of coverage for 0.2 < 0 < 0.55 and equals 134+4 kJ mol-l.Beyond 0 = 0.55 AH could not be determined owing to the lowF. LABOHM et al. 244 1 2.4 2.5 2.6 2.7 103 K I T FIG. 8.-Clausiu~-Clapeyron plots at constant dAS/O. eco 0.2 0.4 0.6 1501 I 1 I I .--. _._.-. . \ . . . \ 0 0 .I 0.2 0.3 6 q0 FIG. 9.-Isosteric heat of adsorption, -AH(a), for different CO coverages expressed as 8 and dAs. (----) Assumed behaviour for 6As+ 0.3. gradient of the isotherms, but it probably decreases to zero when 0 increases to 0.69. Our initial heat of adsorption (also 134 kJ mol-l) is in good agreement with the values found by Klier et aZ.,16 Tracy5 and Bordoli et ~ l . , ~ respectively 11 1, 128 and 129 kJ mol-l. The latter two groups also found AH to be independent of coverage. The discrepancy, although small, could be a result of a difference in measuring time and thus in the amount of carbon deposited during the adsorption measurements. An initial heat of adsorption of ca.134 kJ mol-l was also found by Mesters et al.17 for CO on a Cu(llO)-Ni surface, supporting their conclusion that CO adsorbs initially on (clusters of) Ni atoms in the copper surface. From statistical thermodynamics the following isotherm equation can be deduced :la where p = l/kT (2)2442 INTERACTION OF co WITH Ni(l00) and (3) Here q(a) and q(g) are the molecular partition functions of adsorbed and gas-phase molecules, whose ratio is proportional to exp [ -p AH(a)]. V is the volume of the system, w the interaction energy between molecules on nearest-neighbour sites and n is the number of nearest-neighbour lattice points.This equation pertains to an ensemble of localised, distinguishable lattice points, each capable of binding at most one adsorbed molecule. From this equation a coverage-independent heat of adsorption can be derived if cc) -+ 0 or w -+ 00 (zero or infinite repulsion between molecules on nearest-neighbour sites). In the case co = 0 eqn (1) becomes the Langmuir isotherm, which does not fit the observed data in fig. 1. If we assume infinite repulsion between molecules on nearest-neighbour sites, eqn (1) reduces to Surprisingly, the best fit with this formula was obtained for n = 3, whereas one would intuitively expect n = 4 for an underlying (100) square lattice. Points thus calculated are included in fig. 1. The success of eqn (4) might of course be purely accidental.Assuming its validity each adsorbed CO must have three nearest-neighbour sites (all of them empty). Thus, at saturation the adsorption layer has the structure given in fig. 10, which bears no resemblance to the underlying Ni( 100) lattice. However, the predicted hexagonal structure of the adlayer at saturation is in agreement with the LEED studies of Tracy.5 For lower CO coverages, structures in registry with the Ni lattice have been reported. They are of course not reproduced by the model, which contains only nearest-neighbaur interactions. One may speculate that if the Ni surface is energetically rather smooth for adsorbed CO, the adsorption sites are fuzzy and a 'square' arrangement of CO is possible on the rectangle shown in fig. 10, with lengths 0.32 and 0.37 nm (averaging 0.35 nm, the Ni lattice constant).This will still leave potential nearest-neighbour sites empty . From the values of the constant A in eqn (4) the heat of adsorption can be evaluated. FIG. 10.-Adsorption lattice in which each site has 3 nearest neighbours. At saturation the adsorbed atoms 0 form a hexagonal lattice; the hatched area represents the area of the Ni(100) unit cell.F. LABOHM et al. 2443 The value obtained, 136 kJ mol-l, equals the value obtained from a purely thermo- dynamical approach. Being now convinced that eqn (4) is at least empirically valid in the temperature range 100-1 50 OC, we may check its validity at higher temperatures. From the equation we have calculated the isotherms at higher temperatures and compared them with the measured initial dA,, obtained from fig. 3 and similar measurements.The initial values of dAs in these experiments are not believed to be affected by carbon deposition. As can be seen in fig. 11 the agreement is very good. We therefore conclude that eqn (4) describes the adsorption of CO on Ni( 100) quantitatively, although its theoretical justification is dubious. 0.3 I 1 I I 1 0.2 0 d m 0.1 0.6 0.4 e 0.2 10-4 10-3 10-2 r ) d P a FIG. 1 1 .-Test of the validity of the isotherm eqn (4) above 175 OC. A, 0, 0, measured values at 175, 200 and 225 OC, respectively ; (---) calculated isotherm. DECOMPOSITION OF CO ON Ni(100) In view of the experimental evidence presented, there are two mechanisms that can be excluded for the decomposition. (1) Dissociation CO(a) -+ C(a) + O(a).This reaction hardly takes place, as can be concluded from the small amounts of oxygen deposited on the surface during decomposition. There is of course a possibility that after dissociation the adsorbed oxygen reacts with gaseous or adsorbed CO: Under similar experimental conditions we have investigated the interaction between CO and the Ni(100) surface covered with 0.25, 0.5 and 3 ML of oxygen. For CO pressures up to 0.1 Pa and temperatures up to 250 OC no reaction was found, so that this mechanism can be' excluded. O(a)+CO -+CO,. (6) (2) Reaction between adsorbed CO molecules CO(a)+CO(a)-+ CO,+C(a). (7) This reaction is rather improbable, because in this case the retarding influence of the ionisation gauge on the disappearance of CO is hard to envisage.This is only possible if a reaction takes place between adsorbed and gaseous CO. Gas molecules that cross2444 INTERACTION OF CO WITH Ni(lO0) the ionisation gauge may be affected and may thus react differently. The decrease of the reaction rate cannot be attributed to a lower sticking coefficient of some exited CO species. This would mean an effective reduction of the CO pressure, and thus a lower CO coverage, which was not observed. Furthermore, the carbon deposited with the ion gauge on was clearly of a different nature than that deposited with the gauge Off. Since carbon has to be deposited and not oxygen, only the following reaction remains possible : However, this reaction cannot be an elementary one, since in that case the decompo- sition rate should be proportional to the pressure, which is obviously not the case.The pressure was varied over three orders of magnitude, but in fact the rate of CO removal actually decreases at the highest pressures (cf. fig. 4). Nevertheless, decomposition has to occur according to this overall reaction in order to rationalise the observed ion-gauge effects. The pressure dependence can be reduced by assuming the following elementary steps, adding up to eqn (8): CO(a)+CO(g)-+ CO,+C(a). (8) k, (Ni) + CO(a) + CO(a)* (9 a) CO(a)* is some thermally activated, highly reactive CO molecule, possibly with its interatomic axis parallel to the surface. (Ni) is an adsorption site without CO and not poisoned by carbon. If the first step, eqn (9a), is rate-determining the rate of carbon deposition is simply (10) where a is the number of CO adsorption sites poisoned by one carbon atom and OF8.is the limiting CO coverage on a clean Ni( 100) surface. The pressure dependence now arises only from the pressure dependence of the CO coverage, which is rather weak in the pressure range studied. On the other hand, the number of sites poisoned by one carbon atom is determined by eqn (9b), which involves gaseous CO and thus depends on the operation of an ion gauge. The preceding argument can be made more quantitative by the following analysis. At any time during the decomposition O,, is determined by the ambient gas pressure and the amount of carbon deposited so far. It is thus reasonable to write 8, = k, O,,(O~~x - a0, - Oco) which reduces to the adsorption isotherm given in eqn (4) if 0, = 0.The isotherm now gives O,, as a function, f ( p ) , of the CO pressure. This function is rather complicated (it involves solving a third-order equation), but for our purposes we only need the property thatf(p) increases monotonically with p and has a maximum value of unity. From eqn (1 I) we deduce immediately, using eqn (lo), Solving this differential equation givesF. LABOHM et al. 2445 The decomposition curves can be fairly well fitted to this expression, as can be seen in fig. 4, where the data of table 1 have been used. The initial decomposition rate can be written as It is proportional to the factor a, the number of sites blocked by one carbon atom. From fig. 4 it follows that the initial value of a is smaller when the ionisation gauge TABLE 1 .-VALUES OF a AND f(p) FOR DECOMPOSITION AT 175 "C AND THE PRESSURES INDICATED 6.6 x 8.4 2.8 0.37 6.6 x 4.2 2.8 0.63 6 .6 ~ 4.2 1.4 0.87 6.6 x 2.1 1.4 0.93 For all pressures k , = 0.18 ML s-' and Om,, = 0.69 ML; ax, ion gauge off; a,, ion gauge on. is switched on, all other parameters being the same. This is in agreement with the observation that the amount of carbon deposited can only be made up to 0.5 ML with ethylene if the carbon had been deposited with the ion gauge operating. Unfortunately we had no way of investigating in just what way the ion gauge affects CO, whether it produces ionized molecules or perhaps a vibrationally or electronically exited species. Bertolini and Tardy13 found that thermal decomposition of CO leads to 0.5 ML of carbon exhibiting the p4g LEED pattern.Like us, they found that a surface saturated with carbon does not adsorb any CO. Since at the highest pressures used f ( p ) approaches unity, the initial rate of CO disappearance given in eqn (14) decreases with increasing pressure, as is observed experimentally. From the calculated values of k, at 175, 200 and 225 O C (0.003, 0.015 and 0.09 ML s-l, respectively) we evaluated an activation energy of 130 kJ mol-1 for the rate-determining step. T0ttrupl9 found for the decomposition of CO on a supported nickel catalyst a similar value of 138 kJ mol-l. We may even attempt to extrapolate our equations to much higher pressures. From the data published by Goodman et a1.12 on the CO decomposition on Ni(100) at a pressure of 24 Torr (ca.3 x lo3 Pa) and 177 O C the initial rate of carbon deposition is found to be 1.8 x ML s-l. From our equations and with the values for the constants appropriate for the low-pressure regime we calculate this rate to be 2.5 x ML s-l, in reasonable agreement with the former value, considering an extrapolation over four orders of magnitude. We may thus conclude that our analysis describes the decomposition of CO on Ni( 100) satisfactorily. R. L. Park and H. E. Farnsworth, J . Chem. Phys., 1965, 43, 2351. D. Lichtman, T. R. Kirst and R. B. McQuistan, Phys. Lett., 1966, 20, 129. M. Onchi and H. E. Farnsworth, Phys. Lett., 1968, 26, 349. R. A. Armstrong, in The Structure and Chemistry of Solid Surfaces, ed. G. A. Somorjai (Wiley, New York, 1969). J. C. Tracy, J. Chem. Phys., 1972, 56, 2736. M. A. Vannice, Catal. Rev. Sci. Eng., 1976, 14, 153. A. M. Horgan and 1. Dalins, J. Vac. Sci. Technol., 1973, 10, 523.2446 INTERACTION OF CO WITH Ni(100) * T. Fleisch, G. L. Ott, W. N. Delgass and N. Winograd, Surf. Sci., 1979, 81, 1. @ R. S. Bordoli, J. C. Vickerman and J. Wolstenholme, SurJ Sci., 1979, 85, 244. lo T. E. Madey, D. W. Goodman and R. D. Kelly, J. Vac. Sci. Technol., 1979, 16,433. l1 D. W. Goodman, R. D. Kelley, T. E. Madey and J. T. Yates Jr, J. Catal., 1980, 63, 226. l 2 D. W. Goodman, R. D. Kelley, T. E. Madey and J. M. White, J. Catal., 1980, 64, 479. l3 J. C. Bertolini and B. Tardy, Surf. Sci., 1981, 102, 131. l5 J. H. Onuferko, D. P. Woodruff and B. W. Holland, Surf. Sci., 1979, 87, 357. *e, K. Klier, A. C. Zettlemoyer and H. Leidheiser, J. Chem. Phys., 1970, 52, 589. E. G. Keim, F. Labohm, 0. L. J. Gijzeman, G. A. Bootsma and J. W. Geus, Sug. Sci., 1981,112,52. C. M. A. M. Mesters, A. F. H. Wielers, 0. L. J. Gijzeman, J. W. Geus and G. A. Bootsma, Surf. Sci., 1982, 115, 237. l8 T. L. Hill, Introduction to Statistical Thermodynamics (Addison-Wesley, Reading, Mass., 1960), chap. 7 and 14. * @ P. B. Terttrup, J. Catal., 1976, 42, 29. (PAPER 1 / 1503)

ISSN:0300-9599    

DOI:10.1039/F19827802435

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. Soc., Faraday Trans. 1, 1982, 78, 2447-2454 Dehydrogenation of Propan-2-01 on Zinc Oxide Powders as a Structure-insensitive Reaction BY GI~RALD DJEGA-MARIADASSOU* AND LEON DAVIGNON Laboratoire de Cinetique Chimique, Universite Pierre et Marie Curie, 1 rue Guy de la Brosse, 75230 Paris Cedex 05, France AND AFONSO R. MARQUES Instituto Militar de Engenharia, Rio de Janeiro, Brazil Received 29th September, 1981 The dehydrogenation of propan-2-01 over zinc oxides has been studied in the 473-548 K temperature range. These oxides show oriented textures. At low temperatures and conversions (a < 6-7%), the reaction was zero order in alcohol pressure. The activity per unit area increased by a factor of four with respect to the ZnO origin. From crystallographic data and on the assumption that every surface zinc ion is an active centre, the turnover frequency has been calculated.The essentially constant value obtained shows that the dehydrogenation of propan-2-01 over zinc oxides must be considered as a structure-insensitive reaction. Furthermore, the turnover frequency is of the same order of magnitude as that observed for the dehydrogenation of butan-2-01 on supported-copper catalysts. Metal oxides generally expose cations, oxygen ions and hydroxy groups in unusual coordination states. 9 These coordinatively unsaturated sites (cus) may be considered as the ‘active centres’ in catalysis by oxide~.~ The surface of many oxides can also be studied in terms of acidic and basic sites4 in addition to reducing or oxidizing centres.Attempts have been made to calculate the absolute number of active sites per unit surface area. These data are necessary for the calculation of turnover frequencies and to make a quantitative comparison between activities of different heterogeneous catalysts. For oxides there are three main ways to obtain the number of active sites. The first involves the reactant itself as a probe m01ecule.~ The number of surface intermediates formed per unit surface area gives directly the density of active sites, L. We may try to infer L from the adsorption characteristics determined at some lower temperature (where the reaction occurs either slowly or not at all); however, it is misleading to extrapolate these results as the temperature is raised, since a given adsorbate may change its mode of chemisorption.6 The second method involves specific chemisorption or specific poisoning of the surface oxide with the eventual displacement of the preadsorbed species by In this case it may be difficult to discriminate between surface sites that adsorb the poison and those that are true active centres.Haller and Saint-Just8 have estimated the site density of chromium. oxides by determining the amount of CO irreversibly adsorbed at 195 K which could desorb at 296 K. Nevertheless the question arose as to whether the adsorption of CO is a correct measure of the number of cus. The amount of but- 1 -ene strongly adsorbed at the reaction temperature was also determined, but 6 * 244 72448 S TR U C T U R E-I NSE N SIT I VE RE A C T I 0 N 0 N ZnO this strongly adsorbed butene does react and desorb.This surface species was also studied by displacement with an olefin of different carbon number. These difficulties have made the third method appear attactive. It involves a detailed analysis of the absolute rate in order to estimate L from the pre-exponential factor.g* lo This kinetic method has been discussed by MaatmamlO and is perhaps the best way to determine L under true experimental conditions. There are few data available on turnover frequencies in catalysis by oxides.ll Rates per unit surface area are generally used for activity measurements, and the concept of reaction sensitivity to the structure of the oxide catalyst is only now being developed.ll9 l2 The decomposition of propan-2-01 over various oxides has been studied by Krylov et aI.l3 For MgO, ZnO, CdO and HgO dehydrogenation predominates and the method of preparation of the oxides has been found to have a definite effect on their catalytic activity .We assume in this work that a powder of oriented texture is an interesting intermediate between a single crystal and a non-oriented powder, and our purpose is to discuss (i) the calculation of turnover frequency and (ii) the structure sensitivity of propan-2-01 dehydrogenation on ZnO. EXPERIMENTAL Oriented-texture powders were prepared by thermal decomposition of zinc carboxylate~,~~-~~ as shown in the electron micrograph (plate 1). The characteristics of some of the zinc oxides are presented table 1. TABLE 1 .-CHARACTERISTICS OF SOME ZnO POWERS orientation of surface average B.E.T.crystallographic particle surface ZnO origin planea diameterb/nm area/m2 g-' ex-ace tate ex-oxalate ex-hydroxyoxalate ex- hydrox ycarbona te ex-formate (00.1) (00.1) (00.1) (1i.o) (21 .O) 20-25 25 20-25 20-25 20-25 28-30 15 45-85 50- 100 5-6 a Determined from selected area electron-diffraction patterns. Determined by transmis- sion electron microscopy and X-ray line-broadening, using the Scherrer equation. The total surface areas were measured by the B.E.T. method with nitrogen at 77 K on a conventional volumetric apparatus. X-ray line-broadening (using the Scherrer equation) and transmission electron microscopy were used to determine the average particle diameter. A JEOL electron microscope, type JEM 120, was also used for determining the orientations of the crystallographic planes of the oxide particles, together with their shapes.The dehydrogenation of propan-2-01 was carried out in a dynamic differential microreactor at atmospheric pressure. Alcohol pressures were generated by two saturators thermostated at +_ 0.1 K, and helium was used as carrier gas. Unchanged propan-2-01 and the reaction products were analysed using gas chromatography. Each oxide was preheated in a vacuum Torr = 133.3 x N m-2) at 573 K for 12 h before reaction with propan-2-01.J. Chem. SOC., Faraday Trans. 1, Vol. 7 8 , part 8 Plate 1 P L A ~ 1 .-Electron photomicrograph of ZnO ex-hydroxyoxalate: G = 60000; SAED: (00.1) reciprocal section. G. DJEGA-MARIADASSOU, L. DAVIGNON AND A. R. MARQUES (Facing p . 2448)G.DJkGA-MARIADASSOU, L. DAVIGNON A N D A. R. MARQUES 2449 RESULTS Zinc oxide is known to have a high selectivity in propan-2-01 dehydrogenation. Dehydration is < 1 % in the 450-500 K temperature angel'-^^ and the equation ZnO (CH3)2CH0H(g) * (CH3)2C0(g)+H2(g) can be considered as the overall reaction occurring in this temperature range where thermal decomposition does not occur. For conversion degrees < 6% we observed a zero-order law with respect to alcohol pressure. This result has already been established by Teichner et aZ.17 under static conditions. Rates per unit area (r') for three zinc oxides are shown in table 2, for several alcohol pressures. As can be seen when po is varied the rate per unit area TABLE 2.-RATE PER UNIT AREA FOR PROPAN-2-OL DEHYDROGENATION AT 513 Ka ro/ 1 0-8 mol s-l m-2 po/Torr F'/ 10-8 mol s-1 m-2 a/ 10-8 mol s-l m-2 ZnO 20 30 40 60 ex-formate (21 .O) 9.5 - 8.6 8.9 9.0 ex-hydroxyoxalate 2 0 .6 - - 18.0 19.3 ex-hydroxycarbonate 4.9 4.4 5.1 - 4.8 (00.1) (1i.o) 0.4 1.3 0.3 a Total system pressure = 760 Torr; carrier gas = He; a (degree of conversion) x 2-3%; ro is the rate per unit area calculated from the production of acetone obtained over a 5 x lop3 or 10 x g sample, with no extrapolation and assimilated to an initial rate; To is the arithmetic mean value of the specific rate; a is the standard deviation; po is the initial pressure of propan-2-01. remains essentially constant, but at a value depending on the origin of the catalyst. From one oxide to another ro varies by a factor of 4, and consequently the oxide with the higher surface does not have the greater activity per unit area.These results are in good agreement with those of Krylov et aZ.13 Following Kokes et al.24 we now consider the surface structures exposed for each stoichiometric oxide and discuss the site density, L. SITE-DENSITY CALCULATIONS FOR ORIENTED Z I N C OXIDE POWDERS The unit cell of ZnO is shown in fig. 1. Its space group is P6,rnc with a = 3.249 A and b = c = 5.206 A. Fig. 2(a)-(c) shows the three crystallographic planes previously considered (table 1) and the geometric arrangements of the zinc and oxygen ions in the ZnO unit cell. The bulk structure of ZnO is such that both the zinc and oxygen ions are tetrahydrally coordinated. As can be seen the (00.1) plane contains the zinc ions coordinated to the three lattice oxygens, the remaining tetrahydral position being unsaturated (or occupied by an adsorbed ion such as a hydroxy group or by an adsorbed molecule).The stoichiometry of the cleavage process requires the (00.1) face to be an oxygen-ion surface or a reconstructed surface such as those proposed by2450 STRUCTURE-INSENSITIVE REACTION ON ZnO FIG. 1.-Unit cell of ZnO. FIG. 2.--Structure of three ZnO faces: (a) (00. l), (6) (1T.O) and (c) (21 .O).G. DJkGA-MARIADASSOU, L. DAVIGNON A N D A. R. MARQUES 2451 Dent and K o k e ~ ~ ~ and Atherton et aZ.25 The situation for the (11.0) plane is quite different and leaves each cation attached to two hydroxy groups if the surface is hydroxylated. Finally, in the case of the (2T.O) plane each cation is coordinated to three lattice atoms, but the three-fold axis is no longer normal to the surface.From crystallographic data, using these plan views and assuming that every cus is an active centre, site densities for each surface plane are obtained and are given in table 3. TABLE 3 .-SITE DENSITIES FOR THREE CRYSTALLOGRAPHIC PLANES oxide plane Llpmol m-2 Llsite cm-2 ex-formate (21 .O) 11.3 6.8 x 1014 ex-oxala te ex-acetate ex-hydroxyoxalate (00.1) 18.2 1.09 x 1015 ex- hydroxycarbonate (1i.o) 9.82 5.9 x 1014 Two experimental proofs support our assumption concerning the behaviour of the oriented-texture powders. The first concerns the kinetic method of determining L through the pre-exponential factor lo For zinc oxide ex-formate, A was found to be 5.8 x lo8 mol s-l m-2.14 From transition-state theory, an order of magnitude for the site density can be estimated if we consider a zero value for the entropy (the activated complex configuration must not be very different from that of the molecule in the adsorbed phase).where k is Boltzmann’s constant and h Planck’s constant. From this expression, in our temperature range, L can be estimated as A = (kT/h)L (1) L = (5.8 x 108)/(l.l x 1013) = 5.3 x mol m-2 or 3.2 x 1015 site cm-2. If we compare this value with those presented in table 3 we obtain the same order of magnitude. This result gives us confidence in using our L values to calculate turnover frequencies. The second observation concerns the value of the adsorption equilibrium constants, K, of propan-2-01 on various oriented oxides.Kinetic measurements showed14 that K varies with the nature of the oxide, but for two oxides presenting the same crystallographic plane (ZnO ex-oxalate and ZnO ex-acetate) the same value was observed. INSENSITIVITY OF PROPAN-2-OL DEHYDROGENATION OVER ORIENTED ZINC OXIDES The calculated densities of zinc or oxygen ions depend on the nature of the crystallographic planes and must be considered as upper estimates of the total numbers of active sites per square metre.28 We have shown that rates per unit area obtained in the case of the zero-order kinetic law vary by a factor of 4. If we now divide the rates per unit area by the corresponding site density, table 4 shows that the values of N , the turnover frequency, are not markedly different. Since ro = NL, we may plot the rate per unit area against the site density for three different oxides.As can be seen fig. 3, the best line through the points does not pass2452 s TR u c T URE-IN SEN s I T I v E RE A c TIO N ON ZnO TABLE 4.-TURNOVER FREQUENCIES, N , FOR THE DEHYDROGENATION OF PROPAN-2-OL OVER ZnO L - r0/1Ol2 site site (4 x 1014) molecule N area area /1014 site N,,,, ZnO cm-2 /s-l /nm2 /wa cm-2 s-l ~ ~~ ex-formate 5.42 0.79 0.15 0.56 2.8 1.9 ex-hydroxycarbonate 2.89 0.49 0.17 0.65 1.9 1.5 ex- hydroxy oxala te 11.62 1.06 0.09 0.35 6.9 1.8 a w is the cross-sectional area of the propan-2-01 molecule. 11 Tm 9 ,,b2 / / / / / / / 8 / ?; I / / / 8 1'0, / / / I / / / I I I 0 2 4 6 8 10 ~ / 1 0 l ~ site cm-2 FIG. 3.--rO plotted against L, for three zinc oxides (1, ZnO ex-formate; 2, ZnO ex-hydroxycarbonate; 3, ZnO ex-hydroxyoxalate). through the origin.* However, if the same value, 4 x 1014, is subtracted from each site density, then the straight line will pass through the origin.This indicates that a number of sites (4 x 1014 site cm-2) are unavailable for propan-2-01 dehydrogenation. The area of each type of site can be deduced from crystallographic data [fig. 2(a)-(c) and/or table 31. Results are reported in table 4. These values must be compared with the cross-sectional area of the propan- 2-01 molecule. The molecular area, w, can be calculated according to the equation 6 M where p is the density of liquid propan-2-01, M is its molar mass and NA is Avogadro's Eqn (2) gives w = 0.26 nm2 molecule-l.A comparison between site areas (table 4) and w clearly shows that a propan-2-01 molecule is much larger than a catalyst * This observation was noted by one of the referees.G. DJEGA-MARIADASSOU, L. DAVIGNON A N D A. R. MARQUES 2453 site; consequently a steric effect must exist. This type of theoretical analysis had already been considered by Frennet et in connection with hydrocarbon chemi- sorption over metal catalysts. The plot of rate per unit area against L may now be interpreted as follows: some of the surface sites are hindered during the chemisorption of alcohol on one site, and the same number of sites (4 x 1014 site cm-2) are hindered whatever the nature of the plane; nevertheless the fraction of free surface sites varies. The number of available sites [L - (4 x lo1*)] is also reported in table 4 for each oxide.In terms of surface areas, to claim that 4 x 1014 site cm-2 are hindered for L site cm-2 is equivalent to considering that (4x 1014)-1 cm2 are covered whereas the site area is L-l cm2; i.e. during the adsorption process 0.25 nm2 are covered per site, which is precisely the cross-sectional area of a propan-2-01 molecule. It is clearly now relevant that one propan-2-01 molecule adsorbed on one surface site occupies 0.25 nm2. This result suggests that the step in which an alcohol molecule is chemisorbed needs a patch of nearest-neighbour sites, and since the same value (4 x must be subtracted from each site density (fig. 3) this means that the alcohol molecule occupies the same area whatever the nature of the site.Using the available site densities [L - (4 x 1014)] we are now able to correct the initial N values. The corresponding values (N,,,,) are reported table 4. Agreement between the three Ncorr values is better than that previously obtained for N . This result shows that the concept of turnover frequency is valid for oxide catalysts, and the conclusion can be drawn that propan-2-01 dehydrogenation over zinc oxides is a reaction insensitive to structure. DISCUSSION A comparison between our results and data already published can be established. The first point concerns the kinetic law describing propan-2-01 dehydrogenation. Our dynamic measurements verify the zero-order law established by Teichner et a1.l’ for static conditions. The zero-order law of this monomolecular decomposition enables us to use the kinetic method to determine Various examples have been published where the site density for chemisorption turned out to be only of the order of magnitude of 1012-1013 site cm-2 or a fraction of the total sites.6 The necessity for a patch of nearest-neighbour sites may be an explanation of these results. The value found by K01boe~~ for the dehydrogenation of propan-2-01 over ZnO ex-ZnCO, was ca.IOl4 site cm-2, which agrees perfectly with our density range. One of the aims of turnover-frequency calculations is a comparison between totally different catalysts, such as metal and oxide catalysts.28 Let us consider the work of Teichner et ~ 1 . ~ ’ on the dehydrogenation of butan-2-01 over alumina-supported copper.The same mechanism applies, the rate-determining step being the monomolecular reaction of the adsorbed alcohol. The value of the initial rate has been measured under conditions such that the rate is zero order with respect to the alcohol. Accordingly, the turnover frequency is expressed in the form N = 3 . 0 1 7 ~ 103exp(-6250/T) (3) lo where 6250 is the activation temperature in K.28 This equation takes into account the total number of surface atoms (1.7 x 1015 site cm+). Three main conclusions must be stated (i) the authors found that every surface copper site is active for the reaction, but that a patch of 7.3 Cu atoms is necessary for the dehydrogenation to take place; (ii) the turnover frequency on 18 different catalysts was found to be perfectly constant,2454 s TR u c T UR E-I NS E NS I TI v E RE A c TIO N ON ZnO so the authors concluded that the reaction is structure-insensitive on copper: this conclusion must be compared with our main result; (iii) eqn (3) enables us to calculate the turnover frequency at 5 13 K.The result, 1.5 x s-l, is close to our values. Thus for the same family of reactions the comparison between two totally different types of catalyst (ZnO and Cu/AI20,) is remarkable. We can now explain the behaviour of the oxide powders which present an oriented texture. It seems that their statistical orientation must be taken into account in studies of the sensitivity of reactions over oxide catalysts. It is also clear that a variation in L by a factor of 4 may be significant and may be correlated to the corresponding rate of evolution per unit area.Further investigations are being carried out to verify this assump tion. We thank Dr J. M. Manoli for the computer program used in the calculations and representations of the ZnO crystallographic planes. H. P. Boehm, Ado. Catal., 1966, 16, 179. H. Knozinger, Adv. Catal., 1976, 25, 184. R. L. Burwell Jr, G. L. Haller, K. C. Taylor and J. F. Read, Adv. Catal., 1969, 20, 1. K. Tanabe, Solid Acidr and Bases: Their Catalytic Properties (Kodansha, Tokyo and Academic Press, New York, 1970). R. J. Kokes, Proc. 5th Int. Congr. Catal., ed. J. W. Hightower (North Holland, Amsterdam, 1973), vol. 1, p. Al. G. Munuera and F. S. Stone, Discuss. Faraday SOC., 1971, 52, 206. ’ R. E. Davy, G. D. Parfitt and J. Peacock, Discuss.Faraday Soc., 1971, 52, 215. * G. L. Haller and J. Saint-Just, Proc. 6th Int. Congr. Catal., ed. G. C . Bond, P. B. Wells and F. C. * D. E. Mears and M. Boudart, AIChE J., 1966, 12, 313. lo R. W. Maatman, Catal. Reu., 1974, 8, 1. l1 J. E. Germain, Franco-American Seminar on Heterogeneous Catalysis (unpublished lectures, Paris- l2 J. C. Volta, W. Desquesnes, B. Morawek and J. M. Tatibouet, Proc. 7th Int. Congr. Catal., ed. T. l3 0. V. Krylov, S. Z. Roginski and E. A. Fokina, Izu. Akad, Nauk SSSR, Old. Khim. Nauk., 1957,421 l4 G. Djega-Mariadassou, Thesis (Paris University, 197 1). l5 G. Djega-Mariadassou, G. Pannetier and R. Giovanoli, J. Microsc. (Paris), 1972, 15, 323. l7 Y. Dechatre and S . J. Teichner, Bull. SOC. Chim. Fr., 1967, 8, 2804; 1968, 5, 1865. Tompkins (The Chemical Society, London), vol. 1, p. 235. Lyon, 1980). Seiyama and K. Tanabe (Kodansha, Tokyo, 1980), preprint C4. (Chem. Abs., 1957, 51, 152330. L. Ponsolle, Thesis (Lille University, 1962). G. F. Garcia and G. Kremenic Olandini, An. Real SOC. ESP. Fis. Quim., 1958, 2, 98. A. A. Balandin and V. Vasserberg, Acta Physiocochim. URSS, 1964, 21, 678. A. Eucken, Naturwissenschaften, 1945, 32, 61. 2o J. C. Balaceanu and J. C. Jungers, Bull. SOC. Chim. Belg, 1952, 60, 476. 22 T. Kwan, K. Okada and S. Matsushita, Bull. Inst. Phys. Chem. Res. Tokyo Univ., 1944, 23, 173. 23 J. E. Germain, J. Bigourd, J. P. Beaufils, B. Gras and L. Ponsolle, Bull. SOC. Chim. Fr., 1961, 1504. 24 A. L. Dent and R. J. Kokes, J. Phys. Chem., 1969, 73, 3772, 3781. 25 K. Atherton, G. Newbold and J. A. Hockey, Discuss. Faraday Soc., 1971, 52, 33. 26 S. Kolboe, J. Catal., 1969, 13, 208. 27 B. Echevin and S. J. Teichner, Bull. SOC. Chim. Fr., 1975, 7-8, 1487, 1495. 28 M. Boudart and G. DjCga-Mariadassou, La Cinttique des Rtactions en Catalyse HttProgt?ne (Masson, 28 A. V. Kiselev and Ya. I. Yashin, La Chromatographie Gaz-Solide (Masson, Paris, 1969), p. 155. 30 A. Frennet, G. Lienard, A. Crucq and L. Degols, J. Catal., 1978, 53, 150. J. E. Germain, B. Gras and J. P. Beaufils, An. Real SOC. ESP. Fis. Quim., 1965, 61, 233. Paris, 1982), in press. (PAPER 1/1510)

ISSN:0300-9599    

DOI:10.1039/F19827802447

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1982, 78, 2455-2466 Conductance of Calcium Sulphate and Magnesium Sulphate in Aqueous Solution at High Pressure BY AN-KONG HSIEH, KOK-P,ENG ANG* AND MENG CHANG Department of Chemistry, National University of Singapore, Kent Ridge, Singapore 051 1 Received 9th October, 1981 The conductance of calcium sulphate and that of magnesium sulphate were measured at different temperatures (1 5-35 "C), pressures (1-1 500 bar?) and concentrations (up to lo-, mol dm-3). The association constant, KA, equivalent conductance at infinite dilution, A", partial molar volume change, At', and the distance of closest approach, a, have been calculated. The maximum value of I\Op/A; for CaSO, and that for MgSO, occur at higher pressures than that for KC1; this observation is interpreted in terms of the dissociation of the ion pair at high pressure.With increasing pressure Walden's rule can no longer be applied, owing to a change in size of the hydrated ion.The phenomenon is closely related to the change in the dielectric constant of the solvent and the distance of closest approach, in agreement with the prediction of Bond's electrostatic model. The ion pair of calcium sulphate is more likely to be a contact ion pair than that of magnesium sulphate. Plots of the specific volume as a function of change in partial molar volume at three temperatures extrapolate to a point which may be taken tp represent the situation where the structure of water around the ion is similar to that of bulk water. Magnesium sulphate in solution has received particular attention because of the high sound absorption of the solution (ca.30 times greater than that of water). In sea water the attenuation of ultrasonic sound waves is due to the pressure-induced dissociation of the ion pair of magnesium su1phate.l Much work has been done recently on the interpretation of the structure of the ion Other 2-2 metal sulphates exhibit similar sound-absorption proper tie^.^ Solutions of calcium sulphate have received comparatively little attention. From conductance measurements Osugi and coworkerslO deduced that there were intervening water molecules between the ions in the ion pair, in agreement with the conclusion of Eigen and Tamm derived from ultrasonic absorption studies" of ion pairs. Millero et a l l 2 made use of the deviation in the values of the apparent molal volume from those calculated from the limiting law and the principle of additivity to estimate the partial molal volume change for the formation of the calcium sulphate ion pair.They concluded that calcium ion formed an ion pair with sulphate ion to a greater extent than magnesium, as the partial molal volume change for calcium sulphate was 25 & 3 cm3 mol-l as against a value of 7.3 cm3 mol-l for magnesium sulphate. Evidence for the formation of the ion pair was also obtained using data from ultrasonic absorption spectra,l3, l4 the temperature-jump method15 and the pressure- step method.ls A conductance method for evaluating the structure of the ion pair is presented here. The evaluation of the structure is not only based on the partial molar volume change but also on the change in equivalent conductance, Walden product and distance of closest approach.A range of thermodynamic quantitites can be obtained from conductance measure- ? 1 bar % 105 Pa. 24552456 CONDUCTANCE OF AQUEOUS CaSO, AND MgSO, ments at various temperatures and pressures, and this can be used to yield information on the ion pair and the structure of the electrolytic solution. However, owing to an insufficient temperature range having been studied, only a few quantities relevant to the structure of the ion pair and the structure of the electrolytic solution are discussed here. EXPERIMENTAL The experimental details are described in a previous paper." The conductivity of the solution, x, is calculated after subtracting the conductivity of water.The experimental equivalent conductance, AeXpt, is then obtained by using the equation 1 0 0 0 ~ - x") 2 c Aexpt = where C is the molar concentration of the electrolyte and xo is the conductivity of water. The equivalent conductance at high pressure is calculated by eqn (2) A p = - xp A1 x1 P r where A, is the equivalent conductance at 1 atm,? x p and x1 are conductivity at pressure p and 1 atm, respectively, and pr is the density of water at pressurep relative to that at 1 atm.l** l9 The values of A', KA and u at different temperatures and pressures were calculated by feeding the related data into Fuoss's computer program (program 158) based on the Fuoss-Hsia equation2O' 21 (3) where y is the degree of dissociation, f is the mean activity coefficient of the solute, KA is the association constant, S is the Onsager limiting slope and E and L are the coefficients of the higher terms in the Fuoss-Onsager equation22 which involve A', solvent properties (S and E ) and the ion-size parameter.The mean activity coefficient,f, was calculated from eqn (35) of ref. (23). The values of the viscosity of water were graphically interpolated from the results of Bett and C a ~ p i . ~ ~ The dielectric constant, D, of water at different pressures and temperatures was obtained from the work of Owen.26 The computer program was modified for 2-2 electrolytes, and Ao, KA and u were computed by a series of successive approximations, starting with the extrapolated value of AExpt and the estimated value of a (4.5 A).$ The iteration proceeds until the successive values of Ao differ by less than 0.01 % and those of KA by not more than 2%.The values of A for CaSO, and MgSO, at different concentrations, temperatures, and pressures are shown in table 1-6. The computed values of log,, KA, Ao, AZxpt and a at different A = Ao - SCi yi +ECy In (6ECy) + LCy - KA Cyf 2A TABLE l.-VALUES OF EQUIVALENT CONDUCTANCE (n-' Cm2 eqUiV.-') OF &SO4 SOLUTION AT DIFFERENT PRESSURES AT 15 OC c / lo-, mol dm-3 p/bar 16.73 12.71 11.15 7.818 5.576 4.887 3.909 1.955 1 85.56 88.66 90.03 93.50 96.40 97.40 98.88 102.84 250 87.81 90.76 92.43 95.43 98.19 99.12 100.59 104.33 500 89.58 92.47 93.76 96.96 99.64 100.57 101.95 105.62 750 90.95 93.79 95.03 98.10 100.72 101.52 102.93 106.51 1000 91.90 94.68 95.87 98.87 101.41 102.29 103.57 107.01 1250 92.48 95.18 96.33 99.27 101.75 102.59 103.84 107.20 1500 92.85 95.48 96.65 99.45 101.87 102.69 103.93 107.16 t 1 atm = 101 325 Pa.1 1 A = m.A-K. HSIEH, K-P. ANG A N D M. CHANG 2457 TABLE VALUES OF EQUIVALENT CONDUCTANCE (n-l cm2 equiv.-l) OF CaSO, SOLUTION AT DIFFERENT PRESSURES AT 25 OC c/ mol dm-3 15.59 12.67 9.746 7.796 4.873 3.898 1.949 1 250 500 750 1000 1250 1500 108.45 11 1.38 110.55 113.38 112.19 114.97 113.32 116.03 114.17 116.75 114.70 117.18 114.92 117.35 114.78 116.65 118.13 119.06 119.68 120.00 120.06 117.38 122.35 119.22 123.98 120.63 125.18 121.48 125.92 122.10 126.40 122.40 126.60 122.39 126.51 124.43 129.42 125.99 130.79 127.14 131.80 127.82 132.25 128.21 132.53 128.33 132.49 128.18 132.22 TABLE VA VALUES OF EQUIVALENT CONDUCTANCE ( n - l cm2 equiv.-l) OF CaSO, SOLUTION AT DIFFERENT PRESSURES AT 35 'C c/ mol dm-3 ~ ~ _ _ _ ~ ~ ~ ~ p/bar 16.64 15.56 12.64 9.725 7.779 4.868 3.889 1.945 1 130.20 131.33 134.82 139.00 142.24 148.32 150.84 157.12 250 132.10 133.21 136.61 140.63 143.80 149.64 152.11 158.15 500 133.62 134.68 138.02 141.89 144.96 150.60 153.03 158.81 750 134.72 135.78 139.01 142.74 145.74 151.22 153.57 159.16 1000 135.52 136.58 139.72 143.34 146.18 151.54 153.83 159.19 1250 135.95 137.04 140.09 143.58 146.32 151.52 153.78 158.99 1500 136.11 137.17 140.13 143.55 146.24 151.27 153.47 158.50 TABLE 4.-vALUES OF EQUIVALENT CONDUCTANCE (n-' Cm2 equiV.-') OF MgSO, SOLUTION AT DIFFERENT PRESSURES AT 15 O C c/ loT4 mol dm-3 p/bar 16.55 12.31 11.03 8.210 5.516 5.131 4.105 2.053 1 82.67 85.53 86.55 89.08 92.02 92.54 93.98 97.65 250 84.64 87.45 88.44 90.94 93.78 94.18 95.62 99.19 500 86.15 88.92 89.87 92.26 95.07 95.48 96.86 100.35 750 87.29 90.00 90.92 93.25 96.02 96.39 97.74 101.12 1000 88.13 90.77 91.66 93.92 96.66 96.99 98.33 101.60 1250 88.65 91.23 92.10 94.30 96.97 97.29 98.57 101.78 1500 88.94 91.45 92.30 94.45 97.05 97.36 98.61 101.73 temperatures and pressures are shown in tables 7 and 8.The value of the association constant at 25 O C and 1 atm for CaSO, is 221.8, which is comparable with the value of 204 given by the solubility methodz6 and the e.m.f. method.27 The value of Ao obtained by Osugi is greater than the present value by 1%. In the former case the value was obtained by extrapolation of the A against C : curve to zero concentration.2458 CONDUCTANCE OF AQUEOUS C&04 AND MgSO, TABLE 5.-vALUES OF EQUIVALENT CONDUCTANCE (n-l Cm2 eqUiV.-') OF MgSO, SOLUTION AT DIFFERENT PRESSURES AT 25 OC c/lO-* mol dm-3 p/bar 16.50 12.28 11.00 8.186 5.500 5.117 4.093 2.047 1 103.43 107.18 108.44 111.66 115.65 116.40 118.23 122.97 250 105.18 108.75 109.95 113.20 117.07 117.74 119.48 124.24 500 106.53 110.10 111.20 114.20 118.17 118.77 120.53 125.15 750 107.51 110.95 112.08 115.15 118.89 119.40 121.12 125.66 1000 108.21 111.65 112.70 115.72 119.33 119.84 121.54 125.90 1250 108.66 111.93 113.05 115.98 119.45 120.00 121.62 125.85 1500 108.85 112.07 113.10 116.02 119.40 119.90 121.49 125.61 TABLE 6.-VALUES OF EQUIVALENT CONDUCTANCE (a- ' Cm2 eqUiV.-') OF MgSO, SOLUTION AT DIFFERENT PRESSURES AT 35 O C c/lO-, mol dm-3 p/bar 16.46 10.98 8.168 5.488 5.106 4.084 2.042 1 126.38 132.60 136.74 141.68 142.58 145.03 151.19 250 127.87 133.98 138.00 142.86 143.69 146.09 152.10 500 128.99 135.00 138.94 143.65 144.47 146.85 152.68 750 129.82 135.68 139.53 144.10 144.89 147.19 152.86 1000 130.40 136.11 139.91 144.33 145.10 147.35 152.84 1250 130.69 136.29 139.97 144.39 145.06 147.23 152.57 1500 130.65 136.12 139.70 143.93 144.66 146.78 151.98 On the other hand, the value of the association constant for MgSO, at 25 O C and 1 atm pressure is 200.4, which is outside the range 169-182.' The association constant varies greatly with the conductance equation used.Moreover, since different values of a were used in the calculation the difference in KA values is to be expected.However, the difference in KA values will not have a great influence on the value of the partial molar volume change, AT, which is used here to interpret the effect of pressure on the ion pair. Values of the equivalent conductance at infinite dilution for CaSO, and MgSO, at 25 OC and at various pressures obtained by various authors are listed in table 9 for comparison. All the thermodynamic quantities were obtained from the derivative of the association constant. The least-squares method was used to determine the analytical form of the equation; the derivative at the respective points was then obtained by numerical differentiation using a computer. RESULTS AND DISCUSSION EFFECT OF PRESSURE ON EQUIVALENT CONDUCTANCE AND WALDEN'S The effect of pressure on the equivalent conductance of aqueous electrolytic solutions can be analysed by comparing the observed pressure depedence of equivalent conductance with that expected from the pressure-induced changes in specific volume RULEA-K.HSIEH, K-P. ANG A N D M. CHANG TABLE VALUES OF log,, KA, Ao AND a FOR CaSO, SOLUTION 2459 A'lf2-l cm2 AZXpt/F-l cm2 equiv.-l equiv.-l 1 250 500 750 1000 1250 1500 2.323 2.271 2.230 2.194 2.162 2.136 2.109 T = 15.00 *C 110.96 112.2 112.15 113.3 113.19 114.2 1 13.89 115.1 114.18 115.3 114.19 115.3 1 13.97 114.9 5.05 5.19 5.32 5.43 5.52 5.58 5.64 T 1 2.346 250 2.304 500 2.266 750 2.23 1 1000 2.202 1250 2.174 1500 2.150 = 25.00 "C 140.04 141.6 5.49 141.06 142.5 5.63 141.73 142.8 5.75 141.91 143.1 5.87 141.91 143.0 5.94 141.62 142.6 6.02 141.08 141.9 6.07 T = 35.00 OC 1 2.373 170.62 172.1 6.01 250 2.336 171.19 172.5 6.12 500 2.301 171.44 172.8 6.23 750 2.270 171.39 172.6 6.32 1000 2.240 171.05 172.2 6.42 1250 2.214 170.80 171.7 6.50 1500 2.190 169.69 170.7 6.55 and viscosity of the solution.If one assumes that the radius of a hydrated ion remains constant at high pressure, by Walden's rule the maximum value of Ao and the minimum value of viscosity should lie at the same pressure. However, the maximum values of A' for the system of CaSO, and MgSO, do not coincide with the minimum value of the viscosity of water at the same pressure, as shown in fig. 1. The pressures corresponding to the maximum value of A' for CaSO, and MgSO, are found to be higher than the pressure for the minimum in the viscosity by ca.300 bar. On the other hand, for KC1 solution this pressure difference is relatively small. The effect of pressure on conductivity can be very complicated. Among other things it changes the strength and number of hydrogen bonds, modifies the water structure, decreases the average number of vacancies and hinders vacancy formation, giving rise to different effects on the conductivity. At a constant pressure its value also depends on the nature of electrolyte, concentration and temperature. The normal conduction of CaSO, and MgSO, is caused by the translational movement of the hydrated species through water. The determining step is believed to be the formation of a 'hole' in the solvent and this process is facilitated by the open structure of water.The initial increase of A;/A; with increasing pressure can be attributed more to the destruction of local water structure near the ion than to the destruction of the bulk structure. As the pressure is further increased, the viscosity2460 CONDUCTANCE OF AQUEOUS CaSO, AND MgSO, TABLE &-VALUES OF log,, KA, A" AND a FOR MgSO, SOLUTION T 1 2.254 250 2.21 1 500 2.175 750 2.143 1000 2.114 1250 2.087 1500 2.063 = 15.00 105.10 106.45 107.43 108.06 108.39 108.39 108.18 "C 106.4 5.48 107.8 5.64 108.4 5.75 109.3 5.85 109.5 5.93 109.5 6.01 109.2 6.07 T 1 2.302 250 2.268 500 2.237 750 2.208 1000 2.181 1250 2.156 1500 2.133 = 25.00 OC 132.98 134.4 5.82 133.99 135.4 5.9 1 134.65 136.0 5.99 134.89 136.4 6.07 134.94 136.2 6.13 134.66 135.7 6.20 134.19 135.2 6.23 T = 35.00 "C 1 2.347 164.25 165.7 6.23 250 2.318 164.80 166.3 6.28 500 2.289 165.10 166.5 6.32 750 2.264 164.89 166.2 6.38 1000 2.237 164.58 165.8 6.44 1250 2.216 163.98 165.2 6.47 1500 2.195 163.08 164.4 6.50 TABLE 9.-cOMPARISON OF A" VALUES FOR CaSO4 AND MgSO, AT VARIOUS PRESSURES AND 25 OC AOl0-l cm2 equiv.-l CaSO, MgSO4 plbar this work ref.(10) this work ref. (1) 1 140.04 140.7 132.98 133.07 500 141.73 142.9 132.65 135.1 1000 141.93 142.9 134.94 135.5 1500 141.08 - 134.19 133.4A-K. HSIEH, K-P. ANG A N D M. CHANG 246 1 1.02 1.01 0 - c Ez c \ 1.00 ( a9 9 4 I 0 500 1000 1500 plbar FIG. 1 .-Variation in AJA1 with pressure at 25 'C of MgSO,, CaSO, and KC1 compared with the variation in the viscosity of water (q) with pressure.TABLE ~~.-WALDEN PRODUCT FOR CaSO, AND MgSO, AT 25 OC AND VARIOUS PRESSURES CaSO, MgSO4 change in change in change in change in change in plbar tt (%) A0 (%) A0 tt PA) A0 (%) A0 tt(%) ~~ 250 - 0.6 0.7 0.1 0.8 0.2 500 - 0.9 1.2 0.3 1.3 0.3 750 - 1.0 1.3 0.4 1.4 0.4 1000 - 0.4 1.3 0.9 1.5 1.1 1250 0.3 1.1 1.4 1.3 1.6 1500 1.5 0.7 2.3 0.9 2.4 of water increases rapidly, causing the transportation of ions to be more difficult. This effect is counter-balanced by the continual destruction of local water structure around the ion and an increase in the number of ions owing to the dissociation of ion pairs, which could be the dominant factor in the region of Ai/Ay maximum. However, for solutions of potassium chloride the effect of pressure on dissociation is small, and the pressure difference corresponding to maximum Ao and minimum viscosity of water is found to be relatively smaller.The above discussion is closely related to the Walden product, which varies with the pressure. For pressures < 750 bar, Walden's rule holds roughly, but with a further increase of pressure, the Walden product increases with increasing pressure and Walden's rule no longer applies. This may be owing to a slight change in the size of2462 CONDUCTANCE OF AQUEOUS CaSO, AND MgSO, the hydrated ion caused by a partial stripping of their primary hydration shells. Hole formation could be made more difficult owing to the crowding of water molecules. The effect of stripping and hole formation could be counter-balanced by further dissociation of the ion pair.This may be the reason for the increase in Aoq with increasing pressure. The values of the Walden product Aoq for CaSO, and MgSO, in aqueous solution at 25 O C are listed in table 10. EFFECT OF PRESSURE ON THE ASSOCIATION CONSTANT, K A , AND THE DISTANCE OF CLOSEST APPROACH, a The effect of pressure favours the formation of ionic species, and hence the value of KA decreases with increasing pressure. Fisher1 found the plot of log,, KA against pressure to be linear for magnesium sulphate, but a curve was found in each case for manganese sulphate2* and lanthanum sulphate20. Similar curves were obtained for calcium sulphate and magnesium sulphate. The plots of log,, KA against 100/0 for both sulphates show a slight curve concave to the log,, KA axis. Similar concavity was also shown by MgSO, in the MeOH + H 2 0 By the Fuoss equation if one assumes that a is independent of pressure, then a plot of log,, KA against 1/D should give a straight line.A non-linear relationship could be caused by the dependence of a on pressure. As shown by a kinetic analysis of the ultrasonic absorption spectra of the bivalent sulphates,l1Y 31 the interaction between the hydrated metal and the sulphate ion involves a number of partially solvated species which participate in a series of equilibria represented by M2+(aq) + SOi-(aq) $ M2+(H20),SOg-(aq) S M2+(H,0)SOi-(aq) f M2+SO;-(aq). 2.3 1 1 5 O C 2.0 2.2 2.4 2.6 1 03/aD FIG. 2.-Plot of log,, KA for CaSO, against 1 / a l l .A-K. HSIEH, K-P. ANG A N D M. CHANG 2463 I 2.1 t 2.0 1 I I I I L 1.8 2.0 2.2 1 03/aD FIG.3.-Plot of log,,K, for MgSO, against l / a D . The effect of pressure shifts the above equilibria towards the left-hand side, resulting in a decrease in KA and an increase in a. If the formation of an ion pair is mainly caused by the electrostatic interaction between the metal ion and the sulphate ion, a linear relationship between log,, KA and l/aD would be expected. These relationships are shown in fig. 2 and 3. P r ~ e ~ ~ stated that for most of the ion pairs to be present in the form of solvent-separated ion pairs, a will hardly be less than 6 A. The values of a obtained for CaSO, and MgSO, at 25 O C and 1 atm are < 6 A; a = 5.49 A for CaSO, and 5.82 A for MgSO,, which is not consistent with the formation of only solvent-separated ion pair species.Because of the opposing effect between the coulomb force and thermal movement, a series of equilibria may well exist with some ion pairs being contact pairs while others are solvent-separated. The sum of the ionic radii for CaSO, and MgSO, are 3.72 and 3.38 A, respectively, indicating a separation between the metal ion and sulphate ion of 1.8 and 2.4 A for CaSO, and MgSO,, respectively. Taking the diameter of a water molecule to be 2.76 A the estimated percentage of the contact ion pairs will be 35% for CaSO, and 13% for MgSO,. The low value of a could also be explained in terms of ‘localized hydrolysis’.33 Water molecules in the inner coordination sphere of the cations are strongly polarized, and hydrogen atoms are suitably placed to form strong hydrogen bonds with an oxygen atom of the sulphate However, the percentage change in a is relatively small, and this indicates a small shift in the equilibria under the pressure range examined here.EFFECT OF PRESSURE O N PARTIAL MOLAR VOLUME The partial molar volume change at constant temperature was evaluated by the equation2464 CONDUCTANCE OF AQUEOUS CaSO, AND MgSO, where A t is the partial molar volume change owing to the formation of ion pairs and can be defined as where I: tion pairs is the sum of partial molar volumes of the ion pairs and C Vions is the sum of partial molar volumes of the free ions. In an electrolytic solution, A T is always positive because of the cancellation of charges and consequential release of water molecules on ion pair formation.More ions will be formed at high pressure owing to the dissociation of ion pairs; therefore the value of A V becomes smaller at high pressure. The values of A T at various pressures and temperatures are listed in tables 1 1 and 12. The consequence of the effect of pressure on the association constant KA is that more solvent-separated ion pairs will be formed at higher pressure; this will result in a smaller partial molar volume change, AV. In comparing the values of AV for calcium sulphate and magnesium sulphate, the smaller value for MgSO, suggests that the latter's ion pair is much more hydrated than that of calcium sulphate. Any break-up of hydration shells and possible cancellation of charges will result in a large positive volume change. The com- paratively higher value of A t for CaSO, than for MgSO, indicates that there is a TABLE 11 .-ISOBARIC PARTIAL MOLAR VOLUME CHANGE (cm3 mol-l) OF CaSO, SOLUTION AT VARIOUS PRESSURES T/OC 1 1 1.90 f 0.72 10.05 f 0.50 9.05 k 0.42 250 10.85 f0.21 9.05 f: 0.18 8.40 f 0.09 500 8.80 & 0.15 8.10 f: 0.13 7.75 kO.11 750 7.55 f 0.13 7.25k0.10 7.08 f 0.15 1000 6.60 f 0.20 6.50 f 0.14 6.35 f 0.18 1250 5.85 f 0.35 5.80 k 0.22 5.70 f 0.24 1500 5.18 & 0.52 5.10f0.36 5.02 f: 0.35 TABLE 12.-ISOBARIC PARTIAL MOLAR VOLUME CHANGE (Cm3 m0l-l) OF MgSO, SOLUTION AT VARIOUS PRESSURES T/OC 1 9.85 f 0.48 7.80 & 0.40 6.92 f 0.42 250 8.55 _+O.18 7.35f0.10 6.60 f 0.1 2 500 7.48f0.12 6.80 f 0.08 6.25 f 0.14 750 6.72 f 0.10 6.30 f 0.08 5.95 & 0.18 1000 6.10 f 0.22 5.80 f 0.12 5.60 & 0.12 1250 5.55 f0.28 5.43 f 0.15 5.30f0.20 1500 5.20 f 0.35 5.05 f 0.30 4.95 f 0.30A-K.HSIEH, K-P. ANG A N D M. CHANG 2465 0.93 0.95 0.97 0.99 V/cm3 g-' FIG. 4.-Plot of partial molar volume change of CaSO, against specific volume of solution. I I I I 1 I I 0.92 0.93 0.94 0.95 0.96 0.97 0.98 0.99 1.00 V/cm3 g-I FIG. 5.-Plot of partial molar volume change of MgSO, against specific volume of solution. significant difference in the degree of disturbance of the surrounding water molecules, and therefore the associated ion pair tends to be more of a contact-pair type for CaSO, than for MgSO,. The values of AV for CaSO, and MgSO, at various pressures and temperatures could be easily converted to those at various specific volumes and temperatures with the aid of thepVTrelation for water.18 The value of A rdecreases with specific volume, (fig.4 and 5), and the extrapolated curves appear to originate at a common point. This behaviour is also seen in other s y s t e r n ~ . ~ ~ . ~ ~ A common point generally implies 80 FAR 12466 CONDUCTANCE OF AQUEOUS CaSO, AND MgSO, the conjoint variation of at least two parameters. At a higher temperature the rate of self-exchange of water molecules around the ions is surprisingly high, and hence the amount of structural water decreases owing to thermal movement. On the other hand, at higher pressure the amount of structural water may increase owing to dissociation of ion pairs, especially in the immediate vicinity of the ions. Therefore the common point at higher specific volume may correspond to a point at which the structure of water around the ion is the same as the structure of water in the bulk.F. H. Fisher, J. Phys. Chem., 1962, 66, 1607. A. R. Davis and B. G. Oliver, J. Phys. Chem., 1973,77, 1315. R. M. Chatterjee, W. A. Adams and A. R. Davis, J. Phys. Chem., 1974, 78, 246. F. J. Millero and W. L. Masterton, J. Phys. Chem., 1974, 78, 1287. G. Atkinson and S. Petrucci, J. Phys. Chem., 1966,70, 3122. A. D. Pethybridge, and S. S. Taba, Faraday Discuss. Chem. SOC., 1977, 64, 274. A. D. Pethybridge and S. S. Taba, J. Chem. SOC., Faraday Trans. 1, 1980,76, 368. G. S. Verma, Rev. Mod. Phys. 1959, 31, 1052. .5 F. J. Millero, G. K. Ward, F. K. Lepple and E. V. Hoff, J. Phys. Chem., 1974, 78, 1636. lo E. Inada, K. Shimizu and J. Osugi, Rev.Phys. Chem. Jpn, 1972, 42, 1. l1 M. Eigen and K. Tamm, 2. Elektrochem., 1962, 66, 107. l2 F. J. Millero, F. Gomber and J. Oster, J. Solution Chem., 1977, 6, 269. l 3 J. Stuehr and E. Yeager, in Physical Acoustic, ed. W. Mason (Academic Press, New York, 1965), l4 G. Atkinson and S. K. Kor, J. Phys. Chem., 1962, 66, 107. vol. IIA. M. Eigen and L. Demaeyer, in Techniques of Organic Chemistry, ed. S . Friess, E. Lewis and A. Weissberger (Interscience, New York, 1963), vol. VIII, part 2. l6 H. Hoffmann, J. Stuehr and E. Yeager, in Chemical Physics of Ionic Solution, ed. B. E. Conway and R. G. Barradas (John Wiley, New York, 1966). l7 A. K. Hsieh, K. P. Ang and M. Chang, J. Chem. SOC., Faraday Trans. I , 1977, 73, 920. G. S. Kell and E. Whalley, Philos. Trans. R. SOC. London, 1965, 258, 565. l8 C. W. Burnham, J. R. Holloway and N. F. Davis, Am. J. Sci., Part A, 1969, 267, 70. 'O R. M. Fuoss and K. Hsia, Proc. Natl Acad. Sci. U.S.A., 1967, 58, 1550. 22 R. M. Fuoss and L. Onsager, J. Phys. Chem., 1957, 61, 668; 1958, 62, 1339. 23 R. M. Fuoss, Proc. Natl Acad. Sci. U.S.A., 1961, 47, 818. 24 K. E. Bett and J. B. Cappi, Nature (London), 1960, 207, 620. 25 B. B. Owen, R. C. Miller, C. E. Milner and H. L. Cogan, J. Phys. Chem., 1961, 65, 2065. 26 R. P. Bell and J. H. George, Trans. Faraday SOC., 1953, 49, 619. 27 R. G. Ainsworth, J. Chem. SOC., Faraday Trans. I , 1973, 69, 1028. 28 F. H. Fisher and D. F. Davis, J. Phys. Chem., 1965, 69, 2595. 29 F. H. Fisher and D. F. Davis, J. Phys. Chem., 1967, 71, 819. 30 G. P. Johari, J. Am. Chem. SOC., 1.965,38, 5552. 31 G. Atkinson and S. K. Kor, J. Phys. Chem., 1965, 60, 128. 32 J. E. Prue, Experientia Suppl., 1964, 11, 157. 33 R. A. Robinson and H. S. Harned, Chem. Rev., 1941, 28, 419. 34 J. W. Larson, J. Phys. Chem., 1970, 74, 3392. 35 G. J. Hills, in Chemical Physics in Ionic Solution, ed. B. E. Conway and R. G. Barradas (John Wiley, 36 S. B. Brummer and G. J. Hills, Trans. Faraday SOC., 1961, 57, 1816. Y. C. Chiu and R. M. Fuoss, J. Phys. Chem., 1968, 72,4123. New York, 1966). (PAPER 1 / 1578)

ISSN:0300-9599    

DOI:10.1039/F19827802455

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1982, 78, 2467-2477 Pulse Radiolysis of Methyl Viologen in Aqueous Solutions BY SONJA SOLAR, WOLFGANG SOLAR AND NIKOLA GETOFF* Institut fur Theoretische Cliemie und Strahlenchemie der Universitat Wien und Ludwig Boltzmann Institut fur Strahlenchemie, A- 1090 Wien, WahringerstraBe 38, Austria AND JERZY HOLCMAN AND KNUD SEHESTED Accelerator Department, RISO National Laboratory, DK-4000 Roskilde, Denmark Received 12th October, 198 1 Pulse radiolysis of air-free aqueous methyl viologen (MV2+) solutions was carried out at various pH. The attack of eiq on MV2+, with k(eiq+MV2+) = 7.5 x 1Olo dm3 mo1-l s-l, leads to the formation of the long-lived radical cation (MV' +), which possesses two absorption maxima at 392.5 nm (E,,,,, = 4200 m2 mol-') and 600 nm ( E ~ , , = 1450 m2 mol-l).The H-atoms react with MV2+ at pH 1 forming two species, which have superimposed absorption bands. By means of a computer simulation they are resolved in the absorptions belonging to: (1) a protonated form of the radical cation (MV'+H+), which is produced with k(H + MV2+) = (3.5 f 0.2) x lo8 dm3 mol-l s-l, has 2 absorption maxima at 390 nm (E,,, = 1700 m2 mol-I) and 595 nm ( E ~ , : , = 760 m2 mol-l) and decays by second-order kinetics with k = 3.5 x lo9 dm3 mol-' s-l; (2) an H-adduct (MVe2+H) on the ring carbon, which is formed with k(H + MV2+) = 2.5 x lo8 dm3 mo1-I s-l, absorbs at 3 10 nm (E,,, = 900 m2 mol-l) and 470 nm (E,,, = 630 m2 mol-l) and decays by conversion into MV'+H+ in a first-order process with k = 6 x lo3 s-l.MV'+H+ e MV' + + H+ For the equilibrium pK = 2.9 used as an electron acceptor in various devices for the utilization of solar energy. 0.1 was determined. The presented data explain, at least partly, the instability of MV2+ when Methyl viologen [(also called paraquat) 1,l '-dimethyl-4,4'-bipyridinium dihalide ; MY2+] has previously been studied because of its herbicidal and toxicological activity. 1-3 Owing to its ability to act as a strong electron acceptor it is also an important component in various photochemical and photoelectrochemical cells for the utilization of solar en erg^.^*^ The formation and reactivity of its reduced form (MV'+, radical cation), produced by a one-electron reduction, is of particular interest for both types MV2+ MV'+ The rate constant of reaction (1) determined by pulse radiolysis is very high: 7 x 1O1O dm3 mol-l s-l [ref. (6)] or 8.4 x lOlo dm3 mol-1 s-l [ref.(7) and (8)]. Two absorption maxima of MV'+ at 600 and 395 nm were reported, but the reported values of the corresponding molar extinction coefficients differ widely, E~~~ from l O l O 0 to ca. 1370 m2 mol-l10 and E,,, from 33006 to 4500 m2 mol-l.lo A relatively strong consumption of MV2+ is established when it is used as an electron acceptor in various devices for hydrogen production by solar energy in aqueous solutions. In such systems, MV'+ appears as an intermediate, and is usually assumed 2467 80-22468 PULSE RADIOLYSIS OF METHYL VIOLOGEN to be regenerated in the presence of a suitable catalyst (e.g. Pt0,ll). This regeneration is in most cases not quantitative.l29 l3 The reason for that is not precisely known.Since in such devices the precursors of H, are the H-atoms, which can partly diffuse away from the catalyst’s surface, these could be involved in the decomposition of MV2+. To elucidate this problem, the intermediates produced by the attack of H-atoms on MV2+ in acidic aqueous solutions were studied by pulse radiolysis. Some experiments were also performed at pH 7-9.8 in air-free aqueous solutions in order to verify the previous spectroscopic and kinetic data concerning MV’+. EXPERIMENTAL IRRADIATION FACILITIES* Two sets of pulse radiolysis equipment were used. (1) The 10 MeV LINAC at RISO (Haimson Research Corp., HRC-712; pulse duration variable from 10 ns to 1 p s , equipped for optical detection of transients14 was combined with Nicolet Explorer I11 digital storage oscilloscope, where the data were stored on disks.Further treatment of the data was performed using an on-line PDP 8 computer. (2) A 3 MeV Van de Graaff accelerator (type K, High Voltage Engineering Co., Burlington, USA) provided 0.4 ,us pu1ses.l5+ l6 The kinetic spectroscopy measurements of the short-lived species were performed in combination with an on-line PDP 11 computer. The applied dose in both cases was 2.5-6 J kg-’ (0.25-0.6 krad) per pulse. PREPARATION OF SOLUTIONS The aqueous solutions were prepared using triply distilled water. Methyl viologen dichloride @.a. quality, B.D.H. Chemicals Ltd, Poole) and all other chemicals [NaCl, t-butyl alcohol, HClO,, NaOH and Ba(OH),; E.Merck, p.a. grade] were used without further purification. Before irradiation the solutions were purged with high-purity argon for ca. 1 h in order to remove oxygen. As appropriate OH scavengers, one series of experiments used C1- ions [k(OH + C1-) = 4.3 x log dm3 mo1-l s-l]17 and a second t-butyl alcohol [k(OH) + t-C,H,OH) = 5.5 x lo8 dm3 mol-l s-l].lS COMPUTER SIMULATIONS Because of the complexity of the kinetics, computer simulations of the experimental data for the attack of H-atoms on MV2+ were performed. For this purpose the CDC-computer of the Vienna University was used. RESULTS AND DISCUSSION REACTIVITY OF MV2+ WITH eiq The rate constant for the attack of eiq on MV2+ [reaction (l)] was determined using airfree solutions of lop5 mol dm-3 MV2+with 5 x lop3 mol dm-3 t-butyl alcohol at pH 8.2.The pH of the solutions was adjusted with a NaOH+Ba(OH), mixture in order to remove bicarbonate. The measurements were performed at 720 nm, following the first-order decay of eiq (dose : 2-4 J kg-l per 0.4 ,us pulse). After matrix corrections a mean value of several determinations was obtained : k(eiq+MV2+) = (8.3 1.0) x 1O1O dm3 mol-1 s-l. In a second series of experiments (air-free solutions of mol dm-3 MV2+ with lo-, mol dm-3 t-butyl alcohol at pH 9.7; dose: 1.5-2 J kg-l per 1 p pulse) the rate * We would like to thank Mrs Hanne Corfitzen and Mr Torben Johansen (RISO) and Dipl. Phys. F. Schworer, Mr K-H. Toepfer, Mr F. Reikowski and Mrs G. Kampmann (MPI fur Strahlenchemie) for their valuable help.SOLAR, SOLAR, GETOFF, HOLCMAN A N D SEHESTED 2469 constant was determined by the transient build-up at 392.5 nm (absorption maximum of MV'+).The mean value of several experiments is: k(eLq + MV2+) = (6.7 & 1 .O) x 1O1O dm3 mol-1 s-l. The mean value for the rate constant, determined by both methods is therefore: k(e, + MV2+) = 7.5 x 1O1O dm3 mo1-I s-l. This value agrees with previous data.6-8 The absorption spectrum of the methyl viologen radical cation (MV'+), produced by the reaction of MV2+ with eCq, was measured in air-free solutions [ ( 2 - 4 ) x mol dm-3 MV2+, ca. pH 91 in the range 300-700 nm. Two maxima were found, at 392.5 nm (E392.5 = 4200 m2 mol-l) and at 600 nm (E,,, = 1450 m2 mol-l), in agreement with the previously published spectrum.lO REACTIVITY OF MV2+ WITH H-ATOMS Two series of experiments were performed, one using C1- ions and a second using t-butyl alcohol as OH-scavengers.THE MV2+/C1- SYSTEM In acidic aqueous solutions the C1- ions (present in the original compound) act as efficient scavengers for OH radicals:17 k2 k: ci- + OH f CIOH- ( 2 ) k, = 4.2 x lo9 dm3 mol-l s-I ki = 6.1 x lo9 s-' s-] k , = 2. I x 1O1O dm3 mol-l s-l k4 Cl-+Cl. + Clg- ( 4 ) k, = 2.1 x 1010 dm3 mol-l s-1 k5 Cl, - + Cl, - + c1, + c1- 2k5 = 7.8 x lo9 dm3 mol-l s-l. The absorption maximum of Clg-, at 340 nm (&340 = 880 m2 mol-l), and of ClOH-, at 350 nm (E,,, = 370 m2 mol-l), and also their formation and decay kinetics are taken into account in the subsequent studies. The total value for the reactivity of H-atoms with MV2+ was determined by following the formation of the transient absorption at 390 nm, which is the absorption maximum of MV'+H+ species (the protonated form of MV'+; see later).The air-free solution at pH 1 used contained 2 x mol dm-3 C1- as OH scavenger (dose rate: 5 J kg-I per 0.4 ,us). The apparent rate constant resulting from several determinations, after correction for the chlorine radical, was found to be: k(H+MV2+)total = (6f0.5) x lo8 dm3 mol-1 s-l. mol dm-, MV2+ and 5 x2470 PULSE RADIOLYSIS OF METHYL VIOLOGEN Similar rate constants were obtained under the same experimental conditions at 470 and 595 nm. They have been used as the starting parameters in the subsequent computer simulations. In order to determine the reactivity of MV2+ with OH-radicals, needed for completion of the reaction mechanism, preliminary pulse-radiolysis studies were carried out with 2 x lo-* mol dm-3 MV2+ saturated with N20 (pH 6.3). In this case eiq is converted into OH (total GOH = 5.5) and the transient spectrum showed absorption maxima at 395,470 and 570 nm.Following the build-up of the transients at 470 and 570 nm, and after appropriate corrections for the species resulting from the reaction of H with MV2+, a rate constant for the reaction of OH with MVz+ was k(OH + MV2+) = 4 x lo8 dm3 mol-1 s-l. obtained : Note that in neutral solution C1- does not act as OH scavenger, because reaction (3) does not take place. A detailed report on this subject will be published later.19 The absorption spectrum of the methyl viologen transients, formed by the attack of H-atoms, was measured in air-free acid solutions (pH 1) using (1 -4) x 1 O4 mol dm-3 MV2+ and 5.8 x mol dm-3 C1- (dose: 3-6 J kg-l per 0.4 ,us pulse) and is presented 5 --..a 0 wavelength/ nm FIG. 1.-Absorption spectra of transients produced by attack of H-atoms on methyl viologen in air-free solution at pH 1 . A: 4 x rnol dm-3 C1-, corrected for C1;- absorption. B: 2 x mol dm-3 MV2+, 0.1 mol dm-3 t-C4H,0H. C: 4 x loF4 mol dm-3 MV2+, 0.5 mol dm-3 t-C,H,OH. Insert: Determination of pK for the equilibrium MV'++H+SMV'+H+ (see text): 0, 600 and e, 392.5 nm. mol dm-3 MV2+, 5.8 x in fig. 1, spectrum A. The spectrum is corrected for Cli- and shows absorption maxima at 3 10,390,470 and 595 nm. The kinetics of the disappearance of the transient at these wavelengths were quite different.At 390 and 595 nm the decay is pure second order (k x 2.6 x lo9 dm3 mol-1 s-l), whereas at 310 nm it is mixed and at 470 nm predominantly first order (k x 4 x lo3 s-l). This requires the assumption that the measured absorption belongs to more than one species. Theoretically, the H-atomsSOLAR, SOLAR, GETOFF, HOLCMAN AND SEHESTED 247 1 can attack the MV2+ molecules at different positions. Hence the formation of the following intermediates is assumed : Radical (A), (MV'+H+), is denoted as the protonated form of the radical cation (MV'+), whereas radical (B) represents one type of the various H-adducts (MV' 2+H) on the ring carbon. The transient (C), resulting as a third possibility from the attack of H-atoms on MV2+ by abstraction of an H-atom from the -CH3 group, seems to be less probable, since reactions of this type usually proceed with a rate constant of the order of lo6 dm3 mol-1 s-1.20-22 COMPUTER SIMULATION OF THE MV2+/Cl- SYSTEM The composite spectrum (fig.1A) and the reaction kinetics resulting from the multisite attack of H-atoms on MV2+ can be resolved by a computer ~imulation.~~ This has previously been applied to the elucidation of the pulse-radiolysis data of aqueous methylene blue, thionine and acridine ~ r a n g e . ~ ~ - ~ , In the present case the observed rate constants for formation and decay of the methyl viologen species (given above) and the measured molar extinction coefficients, calculated from fig. 1 A (which are approximate values), have been taken likewise as starting conditions for the computations.All reactions, together with the corresponding rate constants which can influence the reaction mechanism, are summarized in table 1. Assuming a homogenous distribution of H-atoms in the solution, the corresponding differential equations expressing the kinetic course of the reaction mechanism were set up. By applying a non-linear least-squares fitting procedure, the unknown parameters (k6, k,, klo, kill as well as and &go) have been computed, so that OD/cm values calculated for the reaction scheme (table 1) best reflect the experimental OD/cm data from the pulse-radiolysis experiments for all given wavelengths. In this optimisation procedure the unknown parameters were varied until the sum of the squared deviations between calculated and experimental OD/cm values reached a minimum.28* 2v The squared deviations for every set of data at each desired wavelength were weighted according to their different absorption inten~ities.~~ For solving the stiff differential equation system based on the reaction mechanism (table l), Gear's30 numerical algorithm was applied.This method contains an efficient self-adjusting optimum step-size control, which is necessary in this case because of the large differences in the rates of production and consumption of the various radical species. &h0, &tg5,2472 PULSE RADIOLYSIS OF METHYL VIOLOGEN TABLE REACTIONS AND RATE CONSTANTS IN PULSE RADIOLYSIS OF THE MV2+/C1- SYSTEM no. reaction rate constants /dm3 mol-1 s-la 1 2 2a 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 MV2++e- + MV'+ C1-+ Ofiq + ClOH- t l 0 H - + C1- + OH * ClOH- + H+ + C1' + H,O C1' +Cl- --+ c1;- Cl;,-+Cl,- + Cl,+Cl- MV2++H' MV'+H+ (A) E K * Z + H (B) MV2++OH' + MVef20H MV'2+H (B) -+ MV'+H+ (A) 2MV' +H+ (A) + products H' +H' + H, OH' +OH' + H,02 H' +OH' -+ H,O eiq+OH' -+OH- eaq + eiq -+ H, + 20H- eiq+H' +H,O -+ H,+OH- H++eiq + H' MV * +OH -+ products k, = 7.5 x IO'O k, = 4.3 x lo9 kia = 6.1 x lo9 k , = 2.1 x 1O'O k, = 2.1 x 1Olo k, = ? k , = ? k, = 10' k, = 4 x 10, 2k, = 7.8 x lo9 2kl0 = ? k;, = ? 2k,, = 2.3 x 1O'O 2k13 = 1.2 x 10" k,, = 2 x 1O'O k,, = 3 x 1O'O 2k,, = 1.2 x 1O'O k,, = 2.5 x 1O1O k,, = 2.3 x 1O1O ki9 = 4~ 10, a k' is given in s-l.The rate constants are taken from ref. (17), (31) and (32). The computations were performed for three MV2+ concentrations 2 x and 4 x mol dm-3) in the presence of (1-5) x lop3 mol dm-3 C1- at pH 1.Based on the similarity of the kinetics, the absorption bands at 390 and 595 nm were assigned to species (A) (the protonated form of the radical cation, MV'+H+). The absorption at 310 and 470 nm belongs predominantly to the H-adducts (B) (MVS2+H). The kinetic course of the various species absorbing at 310 nm, which was resolved by the described computation method, is presented in fig. 2. It can be seen that the measured total absorption (MTA) and computed total absorption (CTA) are in agreement, indicating the correctness of the applied reaction mechanism. This is also the case at 390 nm, as shown in fig. 3, where only two species are superimposed, MV'+H+ and Cl, -.The MVS2+H species are not absorbing at this wavelength. The absorption band at 470 nm (fig. 4) is caused by both methyl viologen transients, MV'+H+ (the protonated form of the radical cation), and, predominantly, MVS2+H (H-adduct of MV2+). At 595 nm only one transient, (MV-+H+), is absorbing, hence its kinetics is not presented in detail. The final results of the computer simulations are shown in table 2. Based on the formation kinetics (table 2) it has been calculated that at pH 1 ca. 60% of the H-atoms produce the protonated form of the radical cation (MV'+H+) and ca. 40% form H-adducts. It was established that the observed first-order decay of the MVS2+H species (k = 6 x lo3 s-l) was independent of solute concentration and dose rate up to 10 J kg-l per 0.4 ,us pulse.Above a dose of 25 J kg-' the decay follows second-order kinetics. On this basis one can conclude that at low dose rate, where first-order reactions areSOLAR, SOLAR, GETOFF, HOLCMAN AND SEHESTED Q 0 0 I I I I 2473 15 6 10 m I 2 n \ 0 5 0 I I I I I I I 1 o 10 20 30 LO 50 60 ),IS time/ps FIG. 2.-Plot of optical density as a function of time at 310 nm for computed formation of MV'+H+ (A), MVa2+H (0) and Cl;- (0) as well as measured (MTA, x ) and computed (CTA, 0 ) total absorption. Solution: rnol dm-3 MV2+, 5.1 x mol dmP3 C1-, pH 1, saturated with argon. E m I 2 \ n 0 FIG. 3.-Plot of optical density as a function of time at 390 nm for the formation of MV'+H+ (A) and of CL- (0). as well as for the measured (MTA. x ) and comDuted (CTA. 0 ) total absomtion.Solution: loT4 mol dm-3 MV2+, 5.1 x mol dm-3 C1-, pH 1, saturated with argon.2474 PULSE RADIOLYSIS OF METHYL VIOLOGEN f IX :& e o I 4- I I I I I I ' 0 10 20 30 (0 50 60 ps time/ps FIG. 4.-Plot of optical density as a function of time at 470 nm for the formation of MV'+H+ (A) and of MVS2+H (O), as well as for the measured (MTA, x ) and computed (CTA, 0 ) total absorption. TABLE 2.-KINETIC AND SPECTROSCOPIC DATA FOR TRANSIENTS FORMED BY ATTACK OF H-ATOMS ON METHYL VIOLOGEN IN AIR-FREE AQUEOUS SOLUTIONS AT pH 1, OBTAINED BY THE OPTIMIZATION PROCEDURE rate constant, k /dm3 mol-' s-' /m2 mol-' molar extinction coefficient transient formation decay 310nm 390nm 470nm 595nm MV'+Hf (A) (3.5k0.2) x lo8 (3.5k0.3) x 108 350+_10 1700+30 140k 10 760k 10 MV'ZfH (B) (2.5k0.2) x lo8 (6k0.5) x 900+20 - 630+20 - a First-order reaction (k' ih s-l).more easily observable, the adducts (half-life 2/2 z 100 ps) at pH 1 are converted into protonated radical cations (MV' +H+): MVe2+H -+ MV'+H+. (1 1) (B) (4 However, at higher concentr.ations of MV'2+H, in addition to the conversion process (1 1) two further reactions become important, namely: (a) reaction of MV' 2+H with Cl; - radicals by charge-transfer regeneration of MV2+ : MV' 2+H + Cli- + MV2+ + 2C1- + H+ (20)SOLAR, SOLAR, GETOFF, HOLCMAN AND SEHESTED 2415 and (b) dimerisation or/and disproportionation : 2hV2+*H dimers The second-order decay of the protonated radical cations (MV'+H+) might also be due to dimerisation or/and disproportionation, as shown for MV' 2+H species in reactions (21) and (22).The possibility of reaction of MV'+H+ with Cli- was examined by a set of simulation computations. It was established that such a process, if possible, does not play an essential role under the experimental conditions. DETERMINATION OF pK By measuring the optical density at 390 and 595 nm as a function of pH in the range from 0.5 to 5.6 it was possible to determine pK for the equilibrium MV' + + H+ e MV' +H+ (23) using the Hammett33 relationship: - OD, -OD OD - OD, pK = pH-lOg where OD, and OD,, are the optical densities of the pure acid and base forms. Plotting the logarithmic expression from eqn (24), log A , as function of pH, pK='2.9+0.1 was obtained (insert, fig. 1). THE MV2+/t-BUTYL ALCOHOL SYSTEM In these series of experiments the solutions contained (1-8) x mol dm-3 MV2+ and 5 x 1 0-2 to 0.5 mol dm-3 t-butyl alcohol (as OH scavenger ; pH 1) and were purged with high-purity argon before use.5 x mol dm-3 t-butyl alcohol was sufficient to scavenge all OH radicals, so no Clh- radicals were formed. The formation of methyl viologen transients was rapid in the first 30 ,us after the end of the pulse, and was followed by an additional slow component. The reaction was not completed before 120-150 p s . This indicates that not only the H-atoms but also the t-butyl alcohol radicals are involved in the reduction of MV2+, by analogy with the methyl alcohol radical:6 MV2+ + t-c,H,OH + MV' +H+ + C4H,0. (25) The effect of t-butyl alcohol is illustrated by two examples, using : (a) 2 x 1 0-4 mol dm-3 MV2+ with 0.1 mol dmP3 t-C,H,OH (fig.1 B) and (b) 4 x mol dm-3 MV2+ with 0.5 mol dm-3 t-C,H,OH (fig. 1 C). Both solutions were air-free at pH 1. As can be seen from the fig. 1 the transient absorption is enhanced at 390 and 595 nm and diminished at 310 and 470 nm, when the t-butyl alcohol concentration is increased. Using a concentration > 5 x mol dmP3 MV2+, some of the H-atoms are scavenged by the alcohol. In order to verify the role of reaction (25), a computer simulation was carried out for (1-4) x lov4 mol dm-3 MV2+ in the presence of (4-16) x mol dm-3 t-butyl mol dm-3 t-C4HgOH in the presence of2476 PULSE RADIOLYSIS OF METHYL VIOLOGEN alcohol. The reactions taking place in this system, together with the corresponding rate constants, are shown in table 3.The kinetic and spectroscopic data from table 2 were used as input parameters for the computer simulation. It was established that the computed values for this system are consistent with experiment, assuming a rate constant for reaction (25) of k = (1 _+ 0.4) x lo7 dm3 mol-l s-l. TABLE 3.-REACTIONS AND RATE CONSTANTS INVOLVED IN THE PULSE RADIOLYSIS OF THE MV2+/t-BUTYL ALCOHOL SYSTEM no. reaction rate constants used /dm3 mol-l s-la 1 MV2++e- -+ MV'+ 6 MV"+f' MV'+H+ (A) 7 8 E ~ * P + H (B) 9 MV2++OH' -+ MV'2+OH 10 2MV'+H+ (A) + products 11 12 H'+H' +H2 13 OH' +OH' -+ H,O, 14 H' +OH' -+ H,O MVm2+H (B) -+ MV'+H+ (A) 15 eiq-+OH' -+ OH- 16 17 e;q+H'+H,O-+H,+OH- 18 H++eiq -+ H' 19 MV'2+OH -+ products eaq +eiq -+ H, + 20H- 26 27 25 28 2t-C4H,OH + (t-C4H80H), t-C4H80H + H' -+ H, + t-C4H80H t-C4H80H +OH' + H,O + t-C,H,OH t-C4H80H 4- MV2+ -+ MV'+H+ (A) + C4H80 k, = 7.5 x 1O'O k, = ? k, = ? k, = 4 x 10, k8 = 10, 2k10 = ? k;, = ? 2k12 = 2.3 x 10,' 2k1, = 1.2 x 1O'O k,, = 2 x 10'O k,, = 3 x 10'O 2k,, = 1.2 x 1O'O k,, = 2.5 x 1O1O kig = 4 x 10, k,, = 3 x lo5 k,, = 5.5 x lo8 k,, = ? k18 = 2.3 X 10" 2k2, = 1 .4 ~ lo9 a k' is given in s-l. The rate constants are taken from ref. (17), (31) and (32). Any further enhancement of [t-butyl alcohol] (> 2 x 10-1 mol dm-9, however, leads to an additional increase in the absorption at 390 and 595 nm (fig. 1). This indicates that the yield of the t-butyl alcohol radicals is enhanced accordingly, probably due to the influence of spur reactions and the beginning of the direct alcohol radiolysis.CONCLUSIONS Methyl viologen (MV2+) in aqueous solutions is very reactive towards eCq, H-atoms and OH radicals. Reaction (1) of MV2+ with eiq leads to the formation of just one transient (MV'+) and the measured rate constant as well as its absorption spectrum are in good agreement with previously reported data.6-s The determined molar extinction coefficients of the first (&Goo = 1450 m2 dm-3) and second (E,,,., = 4200 m2 dmP3) absorption bands of the radical cation (MV'+) partially agree with published data.lo Using the pulse-radiolysis technique in combination with a computer simulation method23 it was possible to resolve the composite transient spectrum obtained by the reaction of H-atoms with MV2+ (fig. 1). Two intermediates were observed: the protonated form of the radical cation (MV'+H+) and the H-adducts on the ring carbon (MV ' 2+H).Their kinetic and spectroscopic characteristics are summarized in table 2.SOLAR, SOLAR, GETOFF, HOLCMAN AND SEHESTED 2477 When t-butyl alcohol is used as OH-scavenger, the t-C,H,OH radicals are involved in the reaction mechanism, leading to the MV'+H+ transients. This is demonstrated by the decrease of the transient absorption at 310 and 470 nm in the presence of the alcohol, whereas the absorption at 390 and 595 nm increases strongly. The computed rate constant, k(MV2+ + t-c,H,OH) = (1 0.4) x lo7 dm3 mol-1 s-l, is less than the corresponding value for methyl alcohol radicals, k(MV2+ + cH,OH) = 3 x lo8 dm3 mol-l s - ~ . ~ Based on the present experimental results it can be concluded that the primary products of water radiolysis (eLq, H, OH), which also appear during photoelectro- chemical processes, can initiate the decomposition of MV2+ [see reactions (21) and (22)] in devices for the utilization of solar energy.Two of us (N.G. and S.S.) thank Prof. Dr D. Schulte-Frohlinde, Max Planck Institut fur Strahlenchemie, Miilheim/Ruhr, F.R.G. for permission to use the pulse-radiolysis facility, and the RISO National Laboratory, Roskilde, Denmark for the financial support. We also thank the Bundesministerium fur Wissenschaft und Forschung, Austria, for financial assistance. H. K. Fischer, J. A. Clements and R. R. Wright, Am. Res. Resp. Dis., 1973, 107, 246. 2 J. S. Bus, S. D. Aust and J. E. Gibson, Biochem. Biophys. Res.Commun., 1974, 58, 749. A. P. Autor, Life Sci., 1974, 14, 1309. N. Getoff, K. J. Hartig, G. Kittel, G. A. Peschek and S. Solar, Hydrogen as Energy Carrier. Production, Storage and Transport (in German) (Springer, Vienna, 1977). Book of Abstracts, 3rd Int. Con$ Photochem. Conv. Storage of Solar Energy, ed. J. S . Connolly, Colorado, 1J.S.A. (1980). fi L. K. Patterson, R. D. Small Jr and J. C. Scaino, Radiat. Res., 1977, 72, 218. J. A. Farrington, M. Ebert, E. J. Land and K. Fletcher, Biochim. Biophys. Acta, 1973, 314, 372. D. Meisel, W. A. Mulac and M. S. Matheson, J. Phys. Chem., 1981, 85, 179. @ E. M. Kosower and J. L. Cotter, J. Am. Chem. Soc., 1964, 86, 5524. 10 J. A. Farrington, M. Ebert and E. J. Land, J. Chem. SOC., Faraday Trans. I , 1968, 74, 665. l 2 M.Gohn and N. Getoff, Z. Naturforsch, Tie1 A, 1979, 34, 1135. l 3 0. Johansen, J. E. Lane, A. Launikonis, A. W-H. Mau, W. H. F. Sasse and J. D. Swift, in Book of Abstracts, 3rd. Int. C o n , Photochem. Conu. Storage of Solar Energy, ed. J. S. Connolly, Colorado, U.S.A. (1980). K. Kalyanasundaram, J. Kiwi and M. Gratzel, Helv. Chim. A d a , 1978, 61, 2720. l4 K. Sehested, H. Cofitzen, H. C. Christensen and E. J. Hart, J. Phys. Chem., 1975, 79, 310. l 5 N. Getoff and F. Schworer, Radiat. Res., 1970, 41, 1. Ifi N. Getoff and F. Shworer, J. Radiat. Phys. Chem., 1971, 3, 429. la M. Gohn and N. Getoff, J. Chem. Soc., Faraday Trans. I , 1977, 73, 406. l9 S. Solar, W. Solar, N. Getoff, K. Sehested and J. Holcman, to be published. *O G. C. Stevens, R. M. Clarke and E. J. Hart, J. Phys. Chem., 1972, 76, 3863. 22 N. Getoff, M. Pruchova and F. Schworer, to be published. 23 S. Solar, W. Solar and N. Getoff, Radiat. Phys. Chem., in press. 24 S. Solar, N. Getoff, W. Solar and F. Mark, Radiat. Phys. Chem., 1981, 17, 107. 25 S. Solar, W. Solar, N. Getoff and F. Mark, Can. J. Chem., 1981, 59, 2719. 27 S. Solar, W. Solar and N. Getoff, Z. Naturforsch., Teil A, 1982, 37, 78. 28 U. Hoffman and H. Hofman, Eignfiihrung in die Optimierung (Verlag Chemie, Weinheim, 1971 2@ D. W. Marquardt, Soc. Ind. Appl. Math. J., 1963, 11, 431. 30 J. A. 1. Craige, in A Variable Order Multistep Method for Stif Systems of Ordinary Diferentii ! 31 M. Anbar, M. Bambenek and A. B. Ross, SelectedSpecijic Rates of Reactions of Transients from Water 32 M. Anbar, Farhataziz and A. B. Ross, Selected Specific Rates of Reactions of Transients from Water 33 P. Hammett, J. Am. Chem. SOC., 1935, 57, 2103. G. G . Jayson, B. J. Parsons and A. J. Swallow, J. Chem. SOC., Faraday Trans. I, 1973, 69, 1597. B. Hickel, J. Phys. Chem., 1975, 79, 1059. S. Solar, W. Solar and N. Getoff, Radiat. Phys. Chem., in press. p. 178. Equations (University of Manchester, Numerical Analysis Report No. 1 1, 1975). in Aqueous Solutions, Part I (National Bureau of Standards, Washington D.C., 1973). in Aqueous Solutions, Part II (National Bureau of Standards, Washington D.C., 1975). (PAPER 1 / 1 589)

ISSN:0300-9599    

DOI:10.1039/F19827802467

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1982, 78, 2479-2481 Partial Molar Volumes, Partial Molar Expansibilities and Viscosity of Benzene Solutions of Tri-n-octylammonium Halides BY SPELA PALJK AND CVETO KLOFUTAR* J. Stefan Institute, E. Kardelj University of Ljubljana, 61000 Ljubljana, Yugoslavia Received 13th October, 198 1 The densities of benzene solutions of tri-n-octylammonium chloride, tri-n-octylammonium bromide and tri-n-octylammonium iodide up to 0.25 mol kg-' at 293.15, 298.15, 303.15, 313.15 and 323.15 K were measured. The partial molar volumes and partial molar expansibilities of the solutes were found to be independent of concentration, and the partial molar expansibilities were found to be equal for all the solutes investigated. For these systems viscosity measurements were also made and the viscosity coefficients B and D determined.In addition, the relative viscosity was interpreted on the basis of the theory of rate processes and regular solution theory, and thermodynamic functions of activation for viscous flow and the solubility parameters of the solutes investigated were calculated at 298.15 K. The conductance of a system which contains solute molecules as hydrogen-bonded ion pairs of relatively high dipole moment dissolved in a non-polar solvent suggests that even at highest dilution the dissociation of ion pairs to free ions and the formation of triplet ions can be negligible, and that the non-ideal behaviour of the system can be attributed to interactions among the solute molecules and, to a lesser extent, among the solute-solvent pairs.In the case of an aprotic solvent of low permittivity, the solvent effects may be considered ~nimportant.l-~ The aim of this work was to determine some volumetric and transport properties in such a system where ion pairs, due to the dipole-dipole interactions, undergo intensive aggregation processes, while the solvent effects are ignored. With this in mind the benzene solutions of tri-n-octylammonium halides were investigated at 293.15-323.15 K. The partial molar volumes and partial molar expansibilities of solutes and solvent, respectively, and the viscosity of these solutions were determined. The relative viscosity was interpreted on the basis of the viscosity coefficients B and D, and by rate process t h e ~ r y ~ ' ~ and regular solution The limiting values of thermodynamic functions of activation for viscous flow and the solubility parameters of the solutes were also calculated.In addition, the effects of the size of the halide ion and also the effects of aggregation of the investigated solutes on the values obtained are discussed. EXPERIMENTAL Tri-n-octylammonium chloride (TOAHCI), tri-n-octylammonium bromide (TOAHBr) and tri-n-octylammonium iodide (TOAHI) and their benzene solutions were prepared as in ref. (9) and (10). The densities of the investigated solutions at a definite temperature were determined as in ref. (1 1). 24792480 BENZENE SOLUTIONS OF TRI-n-OCTYLAMMONIUM HALIDES The viscosities of these solutions were determined with a Cannon-Fenske viscometer. The absolute viscosity values were calculated by means of the equationI2 where ql, is the absolute viscosity of the solution (kg rn-' s-,), d l , is the density of the solution (kg dm-9, t is the flow time(s) and C and E are constants characteristic of the viscometer.The viscometer constants C = 1.6061 x (& 5 x 10-lo) m2 s - ~ and E = 0.805 f0.069 m2 were determined by a least-squares fit to eqn (1) of the literature data for the absolute viscosity13 and densityI4 of water at the respective temperature. The temperature of the water bath was maintained to k0.05 K. RESULTS AND DISCUSSION The density determinations were made over the concentration range 0.0 1-0.25 mol kg-' at temperatures from 293.15 to 323.15 K. The density of the investigated solutions dl, is given by dl,2 = dl+ac (2) where dl is the density of the solvent (kg dm-3), c is the concentration of the solution (mol dm-3) and a is a constant characteristic of the solute and the temperature.The values of constant a of eqn (2) for the solutions investigated and the density of benzene at the temperatures studied are given in table 1. The values of dl obtained at 293.15, 298.15 and 303.15 K are close to values given in the literature.'* TABLE VALUES OF COEFFICIENT a OF EQN (2) AND THE DENSITY OF BENZENE, d,, IN THE TEMPERATURE RANGE FROM 293.15 TO 323.15 K T / K 293.15 298.15 303.15 313.15 323.15 ~~ ~ ~~ ~~ a/ kg mol-l solute TOAHCl 0.0035 f 0.0002 0.0053 f 0.0002 0.0057 k O.oOo3 0.0085 0.0002 0.01 10 & 0.0003 TOAHBr 0.041 5 f 0.0005 0.0432 & 0.0003 0.0444 k 0.0007 0.0475 & 0.0006 0.0500 f 0.0009 TOAHI 0.0850 & 0.0007 0.0856 f 0.0005 0.0869 If: 0.0007 0.0896 f 0.0008 0.0922 f 0.0007 solvent benzene 0.8796+_0.0001 0.8740+0.0001 0.8679 +_0.0001 0.8564f 0.0002 0.8448 k 0.0001 dJkg dmP3 The values of the coefficient of expansibility of the solutions investigated, al, 2, were calculated from15 a', = a, + bc where a, = -(6dl/6T)p/d, is the coefficient of expansibility of the solvent and b = - (&/a T)/d, is a constant.The coefficient of expansibility of benzene at 298.15 K, calculated from the density data for benzene given in table 1, is 0.001 33 +O.OOOOl K-l. In ref. (14) the value of a, for benzene in the temperature interval from 273.15 to 353.15 K is 0.001 38 K-l. The derivative aa/aTfor all solutions studied was found to be independent of the anion of the solute and is 0.00026_+0.00002 kg mol-l K-l.The apparent molar volume, q4G, for the solutes investigated is given by (4) 1 - dl 4V - - (M2-a)s. PALJK AND C. KLOFUTAR 248 1 where M , is the molecular weight of the solute (kg mol-l). From eqn (4) it follows that the apparent molar volume is independent of concentration and equal to the partial molar volume of the solute, c. At 298.15 K the following values of the apparent molar volume were obtained (dm3 mol-l) : 0,440 & 0.0002 (TOAHCl), 0.4478 +0.0004 (TOAHBr) and 0.4531 +0.0006 (TOAHI). The value of for TOAHCl is close to the value given in ref. (1 1). The apparent molar expansibility, 4E2, for the solutes investigated is given by15 By combining eqn (3) and (9, the apparent molar expansibility is expressed by the equation 4E2 = b+a, &.(6) From eqn (6) it is evident that the apparent molar expansibility is independent of the concentration and thus equal to the partial molar expansibility of the solute, g2. Furthermore, the values of the apparent molar expansibility for the solutes investigated are, within experimental error, equal and are 2.96 x dm3 mol-1 K-l at 298.15 K. The viscosity determinations were made in the same concentration and temperature ranges as the density measurements. The experimental data were analysed according to16 (7) qr=-= l+Bm+Dm2 (_+ 5 x V l , 2 71 where vr is the relative viscosity, v ~ , ~ is the absolute viscosity of the solution, ql is the absolute viscosity of the solvent and m is the concentration of the solutions (mol kg-l).Eqn (7) is valid for the concentration dependence of the relative viscosity of non-electrolytes. The viscosity coefficient B may be given by B = r n - 0 lim igr-lmDm2 In terms of the volume fraction, 42 = m c dl, eqn (7) can be given in the form The parameters B and D are characteristic for a given solute-solvent pair and include the solute-solvent and solute-solute interactions, respectively. The values of coefficients B and D, given in table 2 with the relevant errors, were obtained as the intercepts and slopes of the lines of plots (vr- l)/m against rn by the method of least squares. The values of the viscosity coefficients B and D for TOAHCl at 298.15 K are, within experimental error, equal to those determined previously.ll As can be seen from table 2, the coefficient B shows a slight dependence on the anion of the solute and the temperature.The temperature dependence of coefficient B is linear and nearly equal for the solutes investigated; the average value of dB/dT = 0.0043 f0.0004 kg mol-1 K-l. The positive value of dB/dTmay be due to changes of conformation structure of the alkyl chains in the alkylammonium ion. On the other hand, the coefficient D shows pronounced dependence on the nature of the anion of the solute, as well as on the temperature. The values of viscosity2482 BENZENE SOLUTIONS OF TRI-n-OCTYLAMMONIUM HALIDES TABLE 2.-vALUES OF COEFFICIENTS B AND D OF EQN (7) FOR THE INVESTIGATED TERTIARY n-OCTYLAMMONIUM HALIDES IN BENZENE SOLUTIONS AND THE ABSOLUTE VISCOSITY OF BENZENE, ql, IN THE TEMPERATURE RANGE FROM 293.15 TO 323.15 K T / K 293.15 298.15 303.15 B/kg mol-l D/kg2 moP2 B/kg mol-l D/kg2 mo1-2 B/kg mol-l D/kg2 mo1-2 solute TOAHCl 0.8940.01 1.56f0.07 0.91 f0.02 1.29f0.07 0.93f0.01 1.04f0.08 TOAHBr 0.95 2 0.05 3.08 f 0.20 0.97 f 0.04 2.73 f 0.12 1 .OO f 0.10 2.40 f 0.15 TOAHI 1.00f0.03 5.56f0.19 1.02f0.03 4.93fO.11 1.05f0.06 4.36f0.25 solvent kg m-l s-l benzene 0.647 f 0.001 0.599 f 0.002 0.558 f 0.002 313.15 323.15 B/kg mol-l D/kg2 mo1-2 B/kg mol-l D/kg2 mo1-2 T/K solute TOAHCl 0.98 0.02 0.52 & 0.10 1 .OO f 0.02 0.24 f 0.09 TOAHBr 1.04f0.09 1.60f0.30 1.09+0.10 0.80f0.30 TOAHI 1.09 4 0.05 3.16 f 0.30 1.13 f 0.06 1.96f 0.35 ql/ 1 0-3 kg m-l s-l solvent benzene 0.489 f 0.002 0.434 f 0.002 coefficient D of the investigated solutes at a definite temperature increase in the same order as their ability to form higher aggregates.The temperature dependence of coefficient D is linear, specific for each solute and negative: dD/dT (kg2 moF2 K-l) = - 0.045 & 0.002 (TOAHCI), - 0.076 0.005 (TOAHBr) and In fig. 1, the concentration dependence of the relative viscosity of benzene solutions of TOAHCl, TOAHBr and TOAHI at 298.15 K is shown. The concentration dependence of the relative viscosity of benzene solutions of TOAHI at the temperatures studied is shown in fig. 2. In fig. 1 and 2 the curves were drawn on the basis of eqn (7), using the coefficients B and D at definite temperatures from table 2. The values of the viscosity increment, u = B/r2dl, of the investigated solutes at 293.15 K are: 2.30 (TOAHCI), 2.42 (TOAHBr) and 2.52 (TOAHI).The viscosity increment u increases with increasing temperature; the average value of du/dT is 0.014 K-l for all the solutes investigated. From the values of the viscosity increment it may be assumed that the solute particles behave as spheres in a continuum.17 The values of the viscosity increment D/rEdy increase with increasing size of the anion and at 293.15 K are 10.43 (TOAHCl), 19.93 (TOAHBr) and 35.35 (TOAHI). Their temperature dependences are: d(D/ pi d:)/dT (K-l) = - 0.343 f0.002 (TOAHCI), - 0.49 & 0.01 (TOAHBr) and - 0.744 _+ 0.006 (TOAHI). Considering the theory of rate processes applied to viscous flow, the relative viscosity of a solution is given by5 - 0.120 & 0.005 (TOAHI). rr = - exp [(AG:, - AG,*)/RT] (2)s.PALJK AND C. KLOFUTAR 2483 1.500 1.400 1.300 77r 1.200 1.1 00 1.000 0.050 0.100 0.150 0.200 rnlmol kg-' FIG. 1 .-Concentration dependence of qr of benzene solutions of TOAHCl (O), TOAHBr (0) and TOAHI (A) at 298.15 K. where K, is the average molar volume and AG:, is the change of the average Gibbs free energy of activation for viscous flow of the solution, is the molar volume of the solvent, AG,* is the change of the Gibbs free energy of activation for viscous flow of the solvent, R is the gas constant and T is the absolute temperature. For dilute solutions eqn (7) can be give- as In v,. = Bm + Dm2. (1 1) Combining eqn (10) and (1 1), the following relation can be written6 where subscript 0 indicates the values at infinite dilution of the solute. on solute concentration, Since for the investigated systems the molar volume of solution is linearly dependent K , 2 = v;+(G- Q X , (13) where X2 is the mole fraction of the solute, the partial molar volume of solvent, c, is identical to its molar volume, = Ml/dl, where M, is the molecular weight of solvent (kg mol-I), and the partial molar volume of solute, <, is independent of =2484 BENZENE SOLUTIONS OF TRI-n-OCTYLAMMONIUM HALIDES 1.600 1.500 1.400 Qr 1.300 1.200 1.100 1.000 I I I I 1 u.u3u U.IUU u.13u U.LUU U.L3U rnlmol kg-' FIG.2.-Concentration dependence of qr of benzene solutions of TOAHI at 293.15 (O), 298.15 (A), 303.15 (O), 313.15 (0) and 323.15 K (m). concentration and equal to its volume at infinite dilution, i.e. for dilute solutions the approximation Xz = mM, can be made.= c, o. Furthermore, So, the first term of eqn (12) may be replaced by and the term (AG?, - AGr)o by W?, 2 - m > o = mM1 (A&, 0 - AP?,o) (15) where ApZo is the change of the partial molar Gibbs free energy of activation for viscous flow of the component i at infinite dilution. From eqn (14) and (15) it is seen that the first two terms on the right-hand side of eqn (12) are linearly dependent on molality. Therefore, these terms can be identified with the Bm term of eqn (12), relating only to the solute-solvent interactions, while the Dm2 term, related mainly to the solute-solute interactions, can be identified with f (m>* Consequently, the relation for the viscosity coefficient B can be given bys.PALJK AND C. KLOFUTAR 2485 Eqn (16) shows that the viscosity coefficient B also includes, besides the partial molar volumes, the contributions of changes of partial molar Gibbs free energy of activation for viscous flow of solvent and solute. TABLE 3.-vALUES OF Ap:, AT 293.15-323.15 K AND AS:, AND AR, AT 298.15 K FOR THE SOLUTES INVESTIGATED IN BENZENE SOLUTIONS ______~ ~~ T / K 293.15 298.15 303.15 313.15 323.15 298.15 298.15 A%, 0 A R , 0 solute Ap:, 0/104 J rnol-l /J rnol-1 K-' / lo4 J rno1-I - 3.7 TOAHCl 4.7 4.8 4.9 5.2 5.4 - TOAHBr 4.9 5.0 5.2 5.4 5.7 -285f5 - 3.5 - 3.3 TOAHI 5.1 5.2 5.3 5.6 5.9 - From the known values of the viscosity coefficient B, the partial molar volumes of solute and solvent at infinite dilution, the density of the pure solvent and the change of partial molar Gibbs free energy of activation for viscous flow of solvent, the change of partial molar Gibbs free energy of activation for viscous flow of the solute, Ap:, o, can be calculated.The values of Ap:, for the investigated solutes in the temperature range studied, together with the partial molar entropy, As:,,, and partial molar enthalpy, AP:, o, of activation for viscous flow at 298.15 K are given in table 3. The value of AG;" = Apl, of 9397 & 16 J mol-I was calculated from the relation 271 = w*/ G ) exp (AG;"/RT) (17) where h is Planck's constant and NA is the Avogadro constant, using the values of the density and absolute viscosity of benzene at the temperatures studied from tables 1 and 2. As can be seen from table 3, the partial molar entropy of activation for viscous flow of the solutes investigated, s:,o, calculated via was found to be negative and independent of the anion of the salt.The values of the partial molar enthalpy of activation for viscous flow of the investigated solutes, A@, o, were calculated from the Gibbs-Helmholtz relation As can be seen from table 3, the Ap:, values for all of the solutes studied are positive and show an increase with increasing size of the anion of the salt and with temperature. The Ap:, values are nearly five times greater than that of the pure solvent, indicating the formation of a less favourable transition state in the presence of solute.6 The AH:, values at 298.15 K are negative and increase with the size of the anion. by Through eqn (1 6) the temperature dependence of the viscosity coefficient B is given (20) dB M - dT = - ( E 2 , 0 - '1,o) ' 1 + ' 1 (E, 0 - <) -1 (ARz, 0 - AH;", 0) R T 22486 BENZENE SOLUTIONS OF TRI-n-OCTYLAMMONIUM HALIDES where El,, is the partial molar expansibility and AR,, is the partial molar enthalpy of activation for viscous flow of solvent at infinite dilution.From eqn (20) it is obvious that dB/dTdepends on three terms, i.e. the differences of partial molar expansibilities, the partial molar volumes and the partial molar enthalpies of activation for viscous flow of solute and solvent at infinite dilution. However, the contributions of the first and second terms of this equation are negligible compared with the third term and lie within the experimental error of dB/dT.Taking the ratio of the change of molar energy of vaporization, AEV, to the change of Gibbs free energy of activation for viscous flow as constant for solution or solvent, i.e. AEV/AG* 2 2.5,5 and the change of molar energy of vaporization of the solution, AE;,,,7 as AEY,2 = X1AEY+X2AE:-AEM (21) where AEM is the change in the energy of mixing of both components and for regular solutions given by (22) AEM = Wl K + x2 vz) (61 - 621, 41 #2 where & is the molar volume of the solute and 6, the solubility parameters of solvent and solute, are8 6, = (AEY/E)h and 6, = (AE,V/K)i, then the relative viscosity of solution is given byl8 On the basis of eqn (23) the solubility parameters for the investigated solutes in benzene solutions at 298.15 K were calculated by the method of least squares.The obtained values are: 6, (Ji dmd) = 0.57 (0.5818) (TOAHCI), 0.60 (TOAHBr) and 0.65 (TOAHI). From the above results we conclude that the nature of the anion affects neither the viscosity coefficient B nor the respective partial molar thermodynamic quantities at infinite dilution (e.g. & , o , p:,,, AS:,,, A@,,) of the solutes investigated. Since the partial molar volumes of solutes are independent of concentration, it may therefore be anticipated that the partial molar volume of i-meric species is i times the partial molar volume of the monomer. From the viscosity increment u it may be assumed that the molecules of the solutes investigated behave as spheres in a continuum. Furthermore, it was shown that the aggregation of similar ion pairs depends mainly on their dipole moments.', 4 r la Thereby, the effects of aggregation are visible in the viscosity coefficient D.Thus, the aggregation of solute molecules, and consequently the values of viscosity coefficient D, increase with increasing dipole moment of the simple ion pairs and decrease with increasing temperature. From the values of the solubility parameters of solutes, which are close to the value of the solubility parameter of the pure ~olvent,~ it follows that the investigated solutes are highly soluble in benzene. We thank Mrs J. Burger for her skilful technical assistance. We also thank the Slovene Research Community for financial support. C. A. Kraus, J. Phys. Chem., 1956, 60, 129. C. Klofutar, 5. Paljk and M. Zumer, J. Chem. SOC., Perkin Trans. 2, 1978, 292. C. Klofutar and s. Paljk, J. Chem. SOC., Faraday Trans. 1, 1979, 75, 825. C. Klofutar and 5. Paljk, J. Chem. SOC., Faraday Trans. I , 1981, 77, 2705.5. PALJK AND C. KLOFUTAR 2487 S. Glastone, K. J. Laidler and H. Eyring, The Theory of Rate Processes (McGraw-Hill, New York, 1941), p. 477. D. Feakins, D. J. Freemantle and K. G. Lawrence, J. Chem. Soc., Faraday Trans. 1, 1974, 70, 795. J. H. Hildebrand, J. M. Prausnitz and R. L. Scott, Regular and Related Solutions (Van Nostrand Reinhold, New York, 1970), pp. 8, 82 and 213. C. Klofutar, s. Paljk and M. Ostanek, J . Inorg. Nucl. Chem., 1976, 38, 1045. * H. M. N. H. Irving and J. S. Smith, J . Inorg. Nucl. Chem., 1968, 30, 1873. lo C. Klofutar and s. Paljk, J. Inorg. Nucl. Chem., 1978, 40, 515. l1 5. Paljk, C. Klofutar and M. Zumer, J. Inorg. Nucl. Chem., 1976, 38, 293. l2 M. R. Cannon, R. E. Manning and J. D. Bell, Anal. Chem., 1963,32, 355. l3 L. Korson, W. Drost-Hansen and F. J. Millero, J. Phys. Chem., 1969, 73, 34. l 4 J. A. Riddick and W. B. Bunger, in Techniques of Chemistry, ed. A. Weisberger (Wiley, New York, l5 H. S. Harned and B. B. Owen, The Physical Chemistry of Electrolytic Solutions (Reinhold, New York, l6 T. T. Herskovits and T. M. Kelly, J. Phys. Chem., 1973, 77, 381. l7 A. Einstein, Ann. Phys., 1906, 19, 289; 1911, 34, 591. l 8 C. Klofutar and s. Paljk, Vestn. Slou. Kern. Drus., in press. 1970), vol. 11, pp. 66 and 107. 1950), p. 262. 5. Paljk and C. Klofutar, J . Chem. Soc., Faraday Trans. 1, 1978, 74, 2159. (PAPER 1 / 1596)

ISSN:0300-9599    

DOI:10.1039/F19827802479

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1982, 78, 2489-2496 Exchange of Metal Atoms in Ni(H,O),Ni(CN),-4H,O and Cs,, S bVCl,( S bI'I Cl& BY SAM N Y A R K U ~ AND ALFRED G. MADDOCK* University Chemical Laboratory, Lensfield Road, Cambridge CB2 1 EW Received 2 1 st October, 198 1 The kinetics of the exchange of metal atoms in Ni(H,O), Ni(CN);4H20 and in Cs,, SbVCI,(Sb"'C1,), have been investigated. The data for the former cannot be interpreted in terms of a combination of first-order processes: the data for the latter compound can. Treatments of this compound, with either ionising radiation or ultraviolet light before heating, alter the proportions, but not the energies of activation, of the various exchange channels. The first report on exchange of normal lattice entities concerned ligand-anion exchange.l Recent work has s h o ~ n ~ - ~ that in solids containing an element in two different environments in the lattice, exchange may be able to take place by a quasihomogeneous mechanism, or mechanisms, within the solid.Kinetic data on such processes might provide some clue as to the nature of these mechanisms. In the case of Tl,TlCl, it was found4 that the exchange does not follow the usual linear relation between logfand the time, t (wherefis the fraction not yet exchanged at time t ) , commonly found for exchange in solution. The progress of the exchange reaction showed the same characteristics as the annealing reactions following nuclear reactions in and it has been reported that other exchange reactions of this kind show similar beha~iour.~ Now the Mackay relation,f = exp (- kt), is based on two premises.(i) That a single channel for exchange, characterized by the values of the energy of activation, E, and the frequency factor, v, is open to all exchanging atoms. (ii) That the exchange is proceeding to a dynamic equilibrium. In solids it would not be surprising to find that different exchange channels, with different E and v, are open to differently situated exchanging atoms. For instance, atoms at dislocations, or perhaps near vacancies, might exchange by a mechanism with a lower value of E than is possible for those in a more perfect part of the crystal lattice. The Tl,TlCl, data support this s~ggestion.~ Certainly, since these are exchange reactions, one would appear justified in attempting to analyse the results in terms of a combination of processes (i.e.channels) with different Ei and vi. We report two investigations of this kind, one dealing with a compound with an element in two different lattice sites, but in the same oxidation state and the other of an analogous type to Tl,TlCl,. Ni(H,O), Ni(CN), 4H20 There are three structurally closely related nickel tetracyanonickelate(II).* The quasi-cubic violet form with a structure resembling that of Prussian blue is unsuitable for specific labelling. The blue-green compound, precipitating on addition of a Drive, Corner Brook, Newfoundland, Canada A2H 6P9. t Present address : Memorial University of Newfoundland, Sir Wilfred Grenfell College, University 24892490 METAL-ATOM EXCHANGE PROCESSES solution containing Ni2+(aq) to a solution of tetracyanonickelate(rr), after washing with alcohol and air drying, or in a vacuum dessicator in a refrigerator, has the above composition.It consists of two-dimensional continua of planar Ni(CN), units each linked, through the nitrogen, octahedrally to Ni(H,O),. The planes are separated by the water molecules attached to the nickel and they stack in an eclipsed configuration. The remaining water is accommodated in the inter-lamellar region. Above 333 K four molecules of water are lost, giving yellow Ni(H,O),Ni(CN),, and the stacking of the planes changes to a staggered arrangement. Long has shown that the hydrated compound can be specifically labelled and redissolved and the cationic and anionic nickel separated without very much e ~ c h a n g e .~ EXPERIMENTAL MATERIALS Potassium tetracyanonickelate(I1) was prepared by the method described in ref. (10). Samples of Ni(H,O),Ni(CN), - 4H,O specifically labelled in the cation or anion were prepared as described above, by using labelled Ni2+(aq) or Ni(CN)i-. Analysis for nickel, carbon and nitrogen confirmed the composition of the products. The long-lived 63Ni was used for labelling. The product slowly loses water in a good vacuum at room temperature (ca. 3H,O are left after one week). At ordinary pressure, four water molecules are lost at ca. 333 K and the remainder above 373 K. The dehydrated products were hygroscopic. No evidence for other decomposition was found in this temperature range. The product was stored in a refrigerator at 263 K.METHODS SEPARATION FOR ACTIVITY MEASUREMENTS Ca. 100 mg of the compound was suspended in 1 cm3 of water and 5 cm3 of 1 % alcoholic solution of dimethylglyoxime added. An equal volume of hot 0.2 mol dm-3 ammonium succinate solution was added and the mixture digested at 353 K for a few minutes. The cationic nickel yields bis(dimethylglyoximato)nickel(n). The red precipitate was separated on a demountable sintered filter, washed and dried at 383 K. The filtrate and combined washings were digested with a few cm3 of concentrated hydrochloric acid. After cooling a further 5 cm3 of the 1 % alcoholic solution of dimethylglyoxime was added together with sufficient ammonium hydroxide to bring about complete precipitation of the previous anionic nickel.The precipitate was separated and dried as before. The dimethylglyoxime precipitates were counted as infinitely thick samples, using a windowless gas-flow proportional counter. The 63Ni is a pure j? emitter, giving very soft (EBmax = 67 keV) electrons. Preliminary experiments showed that the extent of exchange accompanying preparation of the specifically labelled material, although varying from one preparation to another, was small enough to permit precise study of the exchange. Further an exchange reaction took place at, or above, room temperature and could be followed both in samples initially labelled in either anion or cation. Generally, anionically labelled material was used. Different preparations showed quantitatively different behaviour and so sets of measurements were made on material from a single preparation.EXCHANGE TREATMENTS Isothermal and isochronal exchange measurements were made on 100 mg samples sealed in small tubes in air. Isothermal treatments were conducted in a thermostatically controlled air oven (& 1 K) or in an ice bath. Isochronal treatments were made in a thermostatted oil bath (& 1 K) for 24 h. Under these conditions the samples lost < 2% of their water during the heating. Further, there was no change in their powder X-ray patterns, beyond a little line broadening.S. NYARKU A N D A. G. MADDOCK 249 1 PRETREATMENTS One set of samples was irradiated to a dose of 27.8 kGy at a dose rate of ca. 1 . 1 1 kGy h-l at room temperature. Another batch of labelled compound was irradiated as a thin layer in air for 50 h at a distance of 15 cm from a Hanovia lamp.The material was cooled by ice during irradiation. RESULTS Exchange isotherms for untreated material are shown in fig. 1 (a). Measurements are corrected for zero time exchange [f = (a? - at,)/(@? - a:) for anionically labelled material, where subscript c denotes the cationic activity and superscripts 0, t and GO indicate zero, time t and the equilibrium value of the activity]. Fig. l(b) shows the variation between different preparations. /*-+-+-+- +-+-+- 10 20 LO 60 80 100 120 1LO 160 180 200 220 t l h FIG. 1 .-(a) Exchange isotherms for one preparation of selectively labelled Ni(H,O),Ni(CN), .4H,O. (b) Exchange isotherms at 313 K for different preparations, 0, + and A, of Ni(H,O)Ni(CN),-4H,O.Both the U.V. and gamma pre-irradiations accelerated the exchange, as is seen in fig. 2(a) and (b). Isochronal data show interesting stepped curves for both untreated and irradiated samples, the pretreatments apparently increasing the depth of the steps, but not changing the temperature at which they occur (fig. 3). The effect of the U.V. irradiation increased for times up to ca. 90 h and then produced no further change. The gamma irradiation also reaches a maximum effect at a dose of ca. 50 kGy.2492 60 50 - ~x-x-x-x-x-x-x-x- )(/x- - 7' +-+-+-+-+-+- +- - :;kT (0) - 20 + - % 10- - 5 . 1 1 1 1 1 1 ' 1 1 1 5 7 0 - * e - x~ x- x- x-x-x-x- x- x- 60:/,+-+-+-+-+-+-+- 50 x ,+" I 30 y+ - 4 4 /+ - 20 5 - 10 L - I t I I I I I I I I I I 0 20 LO 60 80 100 120 140 160 180 200 2ol 10 80 70 60 W M 5 50- 5 2 10- & 30 0 1 1 I I 1 I 1 I 1 I 1 I I 293 297 301 305 309 313 317 321 325 329 333 337 341 TI K complex; + , gamma-irradiated complex.FIG. 3.-24 h isochronal exchange data: 0, untreated Ni(H2O),Ni(CN),*4H,O; x , u.v.-irradiated - - - - DISCUSSION There seems no doubt of the reality of the two steps in the isochronal curves. But to obtain such steep steps, extending over only 10 K, using a combination of exchange channels, with a fraction xi of the atoms having a channel with Ei and vi accessible to them, it would be necessary to have very large values of E,, and since the exchangeS. NYARKU A N D A. G. MADDOCK 2493 is proceeding at very modest temperatures this will demand impossibly high vi. For instance, the isochronal data for the untreated compound can be fitted quite well with a low El with x, = 0.27; E, = 3.50 x lo5 J with x, = 0.25; E3 = 3.70 x lo5 J with x, = 0.21 and a fraction of 0.27 exchanging only by some inaccessibly high E4. But these values must be combined with v M Such a value of v is physically impossible. Further iffis plotted against In t using the isothermal data, then for pairs of points at constant values o f f , In v = (T, In t , - In tl)/(q- c).6 From the isothermal data treated in this way v M 1013.An iterative treatment of all the data, isothermal and isochronal, gives a best fit for El = 1.06 eV,xl = 0.225;E2 = 1.17 eV,x, = 0.384andE3 = 1.32 eV,x, = 0.391 with v = 3.69 x lo1,. But such a combination does not give the stepped structure shown by the experimental curve.We are driven to conclude that our data cannot be adequately represented by a combination of first-order exchange processes with different E and/or v. The process still seems to be affected by pretreatment of the labelled material by U.V. and gamma radiation. It is possible that analysis of the kinetic data in terms of various E and v fails because the second condition for the Mackay equation is lacking. Perhaps the exchange accompanies movements of the lattice water or even some dehydration of the material. Slight changes in colour are noticed during heating. Such processes may not lead to a dynamically reversible equilibrium and may account for the discontinuities in the isochronals. CSl,SbVC16(Sb111C16)3 A number of mixed-valence halogenoantimonates have been described.ll Specific labelling of some of these is possible.12 The compounds of the type M,Sb,X,, have dimensionally slightly different SbX; and SbXi- anions, and are semiconductors.This has been attributed to hole migration in the SbXi- lattice. They have consider- able thermal stability and are therefore attractive for these studies. EXPERIMENTAL MATERIALS C S , S ~ ~ C ~ , S ~ ~ ~ ~ C ~ , The compound was prepared as described by Setterberg et uZ.13 The almost-black crystals are precipitated from a hot solution. Use of labelled SbV or Sb"' permitted some specificity of labelling. The composition of the product was verified by analysis for both antimony and chlorine. Csl,SbVCl,(Sb111C1,)3 Cold solutions of SbC1, and SbCl, in hydrochloric acid were mixed with excess caesium chloride.A violet precipitate was obtained. Analysis for Sb"', SbV and chlorine showed it was CS,,S~~C~,(S~~~'C~,),. The corresponding ammonium salt has been described by Weinland and Schmid.14 Preparations were made labelled with lz4Sb in the SbI" and also in the SbV. For separation of Sb"I and SbV for activity measurements the samples were dissolved in cold 8.0 mol dm-, hydrochloric acid, also 2.0 mol dm-3 in magnesium chloride, and previously saturated with di-isopropyl ether. The di-isopropyl ether used was freed from peroxides by shaking with acidified ferrous sulphate solution. The solution of the sample was immediately extracted with an equal volume of the ether and aliquots of the two phases, the ethereal phase containing the SbV and the aqueous phase the Sb"', taken for measurement of the activity.ACTIVITY MEASUREMENTS METHODS The lZ4Sb was measured in the two phases using a well-type NaI/T1 scintillation counter. A small correction for differences in efficiency in the two media was made. Periodic checks on the activity balance were made.2494 METAL-ATOM EXCHANGE PROCESSES THERMAL STABILITY OF PRODUCTS Samples of the products and samples that had been subjected to the maximum U.V. and gamma radiation doses used were heated in sealed ampoules for one week at 686 K. There was no loss in weight and the SbV/SblI1 ratio, determined analytically, did not change. Above 706 K the compounds slowly changed colour and some reduction of SbV took place. EXCHANGE STUDIES Heating was carried out on samples sealed in glass tubes and wrapped in aluminium foil.The isochronals were of 3 h duration. The U.V. and gamma irradiations were made as for the nickel compound. RESULTS CS,SbVC16Sb111C16 Measurement of the zero-time exchange showed that the specificity of labelling of this compound was poor. Values of between 18 and 25% exchange were obtained and reproducibility was poor. This is not surprising, since hot solutions were used in the preparation and the rather complex SblI1-SbV solution exchange is not very slow under these l6 Some isothermal measurements of the exchange were X I+-/-+----+-+-+- 373 /+-+-+-+-+- 353 & +-+--+-+- 333 ( a ) +-+-+-+--+-+-303 o l 1 1 1 I I I I I I 293 313 333 353 373 393 413 433 453 t l h FIG. 4.-(a) Exchange isotherms for Cs,SbVC1,SblllC1,.(b) Exchange isotherms for CS,,S~'C~,(S~"~C~,),.S. NYARKU A N D A. G. MADDOCK 2495 made, showing that exchange takes place in the solid at temperatures at or above room temperature. The data are shown in fig. 4(a) (fwas calculated as for the nickel compound, vide supra). Cs,,SbVC1,(SblllC1,), The zero-time exchange for this material was much lower, ca. 5 + 1 %, and reasonably reproducible. The results of isothermal exchange are shown in fig. 4(b). Isochronal measurements were made on the original macrocrystalline material and on some material that had been ground to a fine powder in a mortar. There was no difference in the behaviour of the two samples (fig. 5). Samples gamma irradiated to a dose of 53.3 kGy showed no change in exchange isochronal beyond the experimental error (ca.f 1.7%). However, a dose of 213 kGy gave a marked acceleration of the thermal exchange. There was no direct effect at either dose. An irradiation with U.V. under the same conditions as the nickel compound also led to enhanced thermal exchange (fig. 6). 80 70 6 0 - 50- 4 0 - 30- rJ m 2 0 - .- CI m - - ox 8 0 - 70 & 6 0 - $ 50- - 4 0 - 2 3 0 - 5 .c e, - .- * X 0 XI xoc 0 - 01 1 I I 1 I I I I I I 323 343 363 383 403 423 TI K FIG. 5-3 h isochronal exchange data for: 0, untreated Cs,,~bvC1,(Sb"'C1,),; x , crushed CS,,S~~C~,(S~"~C~,),. I '"Fl I I I , , I , s 32 3 343 363 383 4 0 3 L23 TI K material; +, 2.1 x lo5 Gy gamma-irradiated material. FIG. 6.-3 h isochronal exchange data for: 0, untreated Cs,,SbVC1,(SblllC1,),; x , 48 h u.v.-irradiated2496 METAL-ATOM EXCHANGE PROCESSES DISCUSSION If the isothermal data for Cs,Sb,Cl,, are plotted as f against In t it proves to be impossible to superpose all the different isothermals by fixed displacements on the In t axis.Thus it is not possible to account for these data in terms of channels of different Ei but fixed v . ~ The data for the other compound seem more amenable to analysis. An examination of a similar f against In t plot shows that v z lo5. An iterative treatment of the isochronal data shows no statistically significant improvement in fit with more than three Ei with El = 0.572 eV, x, = 0.178; E, = 0.725 eV, x, = 0.386; E3 = 0.847 eV, x, = 0.436 and v = 2.61 x lo5 the discrepancy of the experimental and calculated data is f 1.4%, which is within experimental error. If the data for gamma- and u.v.-irradiated material are treated similarly the computation converges on Ei and xi for a different value of v.However, if v is fixed at 2.61 x lo5 the Ei values obtained are not significantly different from the values for the untreated material. The xi become x, = 0.178, x, = 0.41 3 and x, = 0.409 for the u.v.-irradiated material and x, = 0.178, x, = 0.440 and x, = 0.382 for the gamma- irradiated material. Although this is not the best mathematical fit the discrepancy between the experimental and calculated data, f 1.7% for the gamma-irradiated and f 1.4% for the u.v.-irradiated materials, remains within experimental error. This result resembles the effect of crushing on Tl,T1C16;3 the pretreatments do not affect the Ei for the different exchange channels, but rather the proportions of atoms for which a given channel is accessible.CONCLUSIONS It would appear from these two studies that there are a number of mechanisms for such exchange processes. Although the antimony compound resembles the thallium in that pretreatment affects xi and not Ei, note that exchange in the thallium compound is affected by crushing and not by ionising irradiation but the reverse is true for the antimony compound. For the nickel compound, however, some quite different types of process must be involved. G. B. Schmidt, I.A.E.A. Symposium on Exchange Reacfions (I.A.E.A., Vienna 1965), p. 236. * E. Lazzarini and A. L. Fantola-Lazzarini, J. Znorg. Nucl. Chem., 1975, 37, 407. E. Lazzarini and A. L. Fantola-Lazzarini, J . Znorg. Nucl. Chem., 1976, 38, 655. S. M. Valverde-Fernandez, G. DuplPtre and A. G. Maddock, J. Znorg. Nucl. Chem., 1978, 40, 999. S. Nyarku, A. M. Passaglia Schuch and A. G. Maddock, Inorg. Nucl. Chem. Lett., 1979, 15, 69. A. G. Maddock and M. M. De Maine, Can. J . Chem., 1956, 34, 275. ' A. G. Maddock, in Physical Chemistry, an Advanced Treatise, ed. H. Eyring, D. H. Anderson and W. Jose (Academic Press, London, 1975), vol. 7, chap. 9. * Y. Mathey and C. Mazikres, Can. J. Chem., 1974, 52, 3637. A. F. Long, J. Am. Chem. SOC., 1957, 73, 537. lo W. C. Fernelius and J. J. Burbage Inorg. Synth. 1946, 2, 227. * l M. B. Robin and P. Day Adv. Inorg. Radiochem., 1967, 10, 382. l2 A. Turco and L. Mazzoni, Ann. Chim. (Rome), 1953,43, 865. l3 C. Setterberg, Oefvers. Akad. Forhandl. Stockholm, 1882, 6, 27. l 4 R. F. Weinland and H. Schmid, Benchte, 1905, 38, 1080. l5 C. H. Cheek, N. A. Bonner, and A. C. Wahl, J. Am. Chem. Soc. 1961,83, 80. l6 W. Goishi and N. A. Bonner, J. Am. Chem. SOC., 1961, 83, 85. (PAPER 1 / 1635)

ISSN:0300-9599    

DOI:10.1039/F19827802489

出版商:RSC    

年代:1982

数据来源: RSC