摘要:

J . Chem. SOC., Faraday Trans. I, 1984,80, 2981-2987 Redox Reactions of Thionine and Leucothionine with Iron Chelate Compounds BY KAZUKO TANAKA Department of Inorganic Chemistry, The Institute of Physical and Chemical Research, Wako, Saitama 351, Japan AND EMIKO ENDO AND YOSHIMI KURIMURA* Department of Chemistry, Ibaraki University, Mito, Ibaraki 3 10, Japan Received 16th January, 1984 The kinetics of the reduction of thionine by several iron(I1) chelate compounds (FeI'Y) have been investigated under a nitrogen atmosphere at pH 7.0 and 25 "C by the stopped-flow method. The reactions obey first-order kinetics with respect to the concentration of both thionine and F:e"Y. A linear relationship has been observed between log kobs and log [ K F e ~ ~ ~ Y / K F e ~ ~ Y ] , where kobs is an apparent rate constant for the reduction of thionine by Fel*Y, and K F p Y and.KFenY are the formation constants of the iron(II1) and iron(I1) chelates, respectively.Activation parameters for the reduction of thionine by Fe"Y have been calculated from the temperature dependence of the rate constants. The oxidation of leucothionine by iron(Ir1) chelates has also been investigated using the same technique. A mechanistic scheme for the thionine-leucothionine redox system is discussed in the light of these results and thermodynamic considerations. The photochemical reactions of thionine and its polymeric derivatives with the Fe2+ ion have received much attention over many years,1-18 partly because of the importance of this system in photogalvanic-cell applications. Detailed studies have revealed that photoactivated thionine in the triplet state is quenched by the Fe2+ ion to produce semithionine, followed by dismutation to the starting material, thionine, and leucothionine.Leucothionine is subsequently reoxidized to thionine by the Fe3+ ion. However, most of these studies are limited to acidic solutions mainly because of the instability of Fe2+ and Fe3+ ions in the higher-pH region. However, iron(I1) and iron(Ir1) species can be stabilized by complexation with chelating agents, which is also advantageous for varying the redox potentials and steric structures of the iron chelate species. Such variations allow us to examine the mechanistic picture in more detail. This paper deals with kinetic studies of the reduction of thionine by various iron(r1) chelates and the oxidation of leucothionine by iron(II1) chelates in aqueous solutions at pH 7.0.EXPERIMENTAL Thionine (Eastman Kodak) was recrystallized from water and dried at 105 O C . 4 Chelating Solutions of iron(r1) and iron(rr1) were prepared by dissolving the required weight of ferrous agents (Dojin Kagaku) used were of analytical-reagent grade.? t Abbreviations of the chelating agents (in the acid form) used in the present studies are: EDTA, ethylenediamine-N,N,N',N'-tetra-acetic acid; CyDTA, 1,2,-cyclohexanediamine-N,N,N',N'-tetra-acetic acid; DTPA, diethylenetriamine-N,N,N',N",N"-penta-acetic acid; TTHA, triethylenetetramine- N,N,N',N",N'",N'"-hexa-acetic acid; GEDTA, glycolethylenediamine-N,N,N',N'-tetra-acetic acid; EDTAOH, N-(2-hydroxyethyl)ethylenediamine-N,N',N'-triacetic acid ; NTA, nitrilotriacetic acid. 298 12982 REDOX REACTIONS OF THIONINE AND LEUCOTHIONINE ammonium sulphate and ferric ammonium sulphate, respectively, in dilute perchloric acid solutions. In the former case the perchloric acid solution was previously carefully deaerated by flushing with pure nitrogen.Iron@) chelates were prepared by mixing a weakly alkaline solution of chelating agent (in 20% excess) with the iron(I1) solution under nitrogen. All the iron(u) chelate solutions used in the kinetic study were freshly prepared. Iron(1Ir) chelate solutions were made by adding the iron(m) solution to a small excess of chelating agent in dilute alkaline solution. Solutions of leucothionine were prepared in a nitrogen atmosphere by reduction of thionine with a large excess of glucose in dilute alkaline solution at 60 "C.The solutions used for the kinetic studies were prepared from stock solutions of the reagents prepared as above, and the hydrogen-ion concentration was adjusted to pH 7.0 using phosphate buffer. The ionic strength was maintained at 0.2 mol dmP3 with NaClO,. The reduction of thionine and the oxidation of leucothionine were monitored by observing the decay and growth, respectively, of thionine absorbance at 600 nm with a stopped-flow spectrophotometer (Yanagimoto model SPS- 1). There is no appreciable absorbance from other species in the system in this region. All solutions submitted to these measurements were carefully deaerated by bubbling with nitrogen.The temperature of the reaction solutions in the mixing chamber was controlled by thermostatted circulating water. RESULTS AND DISCUSSION REDUCTION OF THIONINE BY IRON(II) CHELATES The reactions were followed with iron(I1) chelates (FelIY) in large excess, as required for pseudo-first-order kinetics. In all cases a linear correlation between log ( A , -A,) and the reaction time, t, was observed over at least two half-lives, where A , and A , are the absorbances at time t and after the completion of the reaction, respectively. The rate constants calculated from linear plots for three concentrations of FeIIEDTA at a constant concentration of thionine agreed well, indicating that the reaction obeys first-order kinetics with respect to the concentrations of both thionine(T) and FeITY, as defined by -d[Tl/dt = k,,,[T] [FeIIY].(1) The apparent rate constants, kobs, for several chelates are summarized in table 1, together with the formation constants of the corresponding compounds FeIIY and FeIIIY, KFetiY and KFeltiY, respe~tively.~~ The values of log kobs are plotted as a function of log [ K F e ~ ~ ~ Y / K F e ~ t Y ] in fig. 1. The fairly good linear relationship observed indicates that a linear free-energy relationship holds in these reactions. A similar linear relationship has been reported20 between log k and log [ K F e ~ ~ ~ Y / K F e ~ ~ Y ] , where k is the second-order rate constant for the reaction of FeIIY with dioxygen where Y = NTA, EDTAOH, EDTP, EDTA, HEDTA, DTPA, CyDTA and HDTPA.* The values of log k for the last of these three cases fall below the best straight line on the plot of log k against ~ O ~ [ K ~ ~ I I I ~ / ~ ~ ~ ~ I I ~ ] .This result indicates that, as a result of steric hindrance arising from the bulky ligands, the reaction of Fe"Y with dioxygen may proceed with an inner-sphere electron-transfer mechanism. On the other hand, no evidence of such steric effects is observed in the case of the present results shown in fig. 2, which may suggest that the reaction proceeds via an outer-sphere electron-transfer process. Table 2 shows activation parameters for the reduction of thionine by FeIIY, determined from the temperature dependences of the rate constants. The negative entropy changes are probably ascribable to a decrease in the degree of freedon of the reactant molecules in the formation of the activated complexes, and also suggest that there are no large positive entropy changes associated with the release of solvent * EDTP is ethylenediamine-N,N,N,N-tetrapropionic acid, HEDTA is the monoprotonated form of EDTA and HDTPA that of DTPA.K.TANAKA, E. END0 AND Y. KURIMURA 2983 Table 1. Rate constants for the reduction of thionine by iron@) chelates (25 "C, pH 7.0, I = 0.2) and formation constants of the iron(I1) and iron(m) chelates compounda ~ reduc tan t Fe"EDTA 1.4,' 1.2,d 1.3" 14.3 F e W y DTA l . l d 18.2 FeIIDTPA 1 .9d 16.0 Fe"TTHA l . l d 17.1 FeIIGEDTA 0.20d 11.9 Fe'IEDTAOH 0.083d 12.2 FeI INTA 0.06tid 8.8 25.1 29.3 27.5 26.8 20.5 19.8 15.9 a Formation constants of ferrous and ferric chelates (25 "C, I = 0.1) are cited from ref.(19). 2.0 x mol dm-3 thionine. Errors of the rate constants were within k 6%. 2.0 x mol dm-3 Fe"Y. 4.0 x lop4 mol dmP3 Fe'IY. 6.0 x lop4 mol dm-3 Fe"Y. 5 Fell EDTA Fe'ITTHA \ , Y qFeiiDTPA F e *' C y D TA / I C \ F e " E D T A O H F e " N T A I I I I I I 6 8 10 12 log WFe"'Y IKFe"Y) Fig. 1. Linear relationship between log kobs and log (KFe~~~Y,KFe~~Y). molecules from the reactanct molecules on going from the separate reactants to the transition state. The effect of added Fe1I1Y on the rate of reduction of thionine by FeIIY was investigated. The plot of log (A, - A , ) against time was almost linear for at least 1.5 half-lives, and apparent rate constants were determined from these plots. The values were found to decrease with increasing concentration of added FelIIY, as shown in fig.2, which suggests that the leucothionine and/or the semithionine formed as an intermediate are reoxidized to the starting material by FeIIIY.2984 REDOX REACTIONS OF THIONINE AND LEUCOTHIONINE 15 " 'v, E 10 rn -a 4 LI E 0, 2 5 - Yo m Table 2. Activation parameters for the reduction of thionine by Fe'IY (25 "C, pH 7.0, I = 0.2)a - 0 - reductant AH/kJ mol-l AS/J mo1-l K-l FeIIEDTA Fe"Cy DTA FeIIDTPA FeIITTHA FeIIGEDTA FeI'EDTAOH FeIINTA 41 36 30 41 39 40 32 - 27 - 50 - 63 -31 - 50 - 54 - 84 a The temperature range studied was 12.5-34.0 "C and the accuracies of the temperatures were within &O. 1 "C. Standard errors of all the activation parameters were within +4%. 0 0 0 0 V 0 - 2 4 [ FeIIIEDTA 1 / 1 0-5 mol dm -3 Fig.2. Effect of the concentration of added FeIIIY on the apparent rate constants for the reduction of thionine by FeIIY. OXIDATION OF LEUCOTHIONINE BY FelIIY When a colourless solution of leucothionine was mixed with a solution of FelIIY, the blue colour of thionine rapidly appeared. The rate of generation of thionine, d[T]/dt, was monitored spectrophotometrically with FeIIIY in large excess. The apparent rate constants, khbs, were determined from the slopes of the log (A, - A,) time plots, which were linear over at least two half-lives. The apparent rate constants for the oxidation of leucothionine by several FeIIIY thus obtained are presented in table 3. Good agreement was observed for the rate constants obtained at several concentrations of FeIIIY as shown in table 3, indicating that this reaction obeys first-order kinetics with respect to the concentrations of both leucothionine and FeIIIY, according to d[TJ/dt = khbs [L] [FeIII].(2) For the oxidation of leucothionine by FelIIY, no correlation between kbbs and log [KFe~~~Y/KFe~~Y] was observed.K. TANAKA, E. END0 AND Y. KURIMURA Table 3. Apparent rate constants for the oxidation of leucothionine by iron (111) chelates (25 "C, pH 7.0, I = 0.2)u 298 5 oxidant k;,,J lo4 dm3 mol-' s-l FeII'EDTA 4.8,b 4.6,c 4.9d FeIIIDTPA 7.6b FeII'CyDTA 20b Fe'I'TTHA 4.3b FeIIIGEDTA 2.1b Fe'I'EDTAOH 0.16b FeI'INTA 0.28b o, 2.0 x mol ~ i m - ~ thionine. Errors in the rate constants were within f 5%. 2.0 x mol dmP3 FelIIY. 4.0 x mol dm-3 FeII'Y. 6.0 x mol dm-3 FeII'Y.MECHANISTIC CONSIDERATIONS One-electron reduction of thionine by FeIIY and one-electron oxidation of leucothionine by Fe1I1Y produce semithionine (S). The semithionine radical produced may undergo further reduction by FeIIY, oxidation by FeIIIY or a dismutation reaction to thionine and leucothionine. The following reaction scheme can be proposed for the thionine - leucothionine equilibrium in the presence of FeIIY and FeIIIY: I I 1 I k-d k-d where kred,, kred,, koxl and kox, are the rate constants of each step and k, and k-, are rate constants of following reactions: 'd 2S'T+ L. k-d (3) It is of interest to compare the rate constants for the reactions involved in this scheme with values for the much-studied photoinduced reduction of thionine by ferrous ions.The evaluation of the various rate constants in the above scheme depends on a knowledge of the equilibrium constants. These are not easily accessible from experiment, but may be estimated as follows from electrochemical considerations. Rabinowitch3 proposed that the standard electrode potential for the reduction of thionine to leucothionine, @,L, is the arithmetic means of each step (@/s, E&) and that the difference between qIs and qL is related to the dismutation constant, Kd, of semithionine to leucothionine and thionine as defined by2986 REDOX REACTIONS OF THIONINE AND LEUCOTHIONINE Table 4. Rate constants for the reactions of thionine, semithionine and leucothionine at 25 "C rate constant/mol dm3 s-' reaction Fe"'EDTA/Fe"EDTAU Fe3+/Fe2+ T + Fe" -+ S + Fe"' T* + Fe" + S + Fe"' S + Fell --+ L + Fe"' L + FelI1 -+ S + Fe" 1.3 x 104 L O X 104-1.ox 106d 4.8 x 104 - S + Fe'II -+ T + Fe" 2S+T+L - T+L + 2s - 7 x lo5-7 x - 107 7.9 x 104 2.4 x 109 + 105 - 2.6 x lo2 a pH 7, I = 0.2 mol dm-3.Obtained by Hatchard and Parkers from a kinetic analysis of T* is excited-state the photoreduction of thionine by ferrous sulphate (0.5 mol dmP3 H2S04) thionine. Obtained by assuming Kd = 10-2-10-5. We can derive an expression for Kd as follows: @,s - %L log& = 0.06 where 0.06 V is the value of RT/F used in the Nernst equation (at 25 "C). A value of 0.06 V for E@L at pH 7.0 has been reported.* Using this value approximate values of @, and qL as a function of Kd are estimated from eqn ( 5 ) and (6): J!?@s = 0.06+0.03 log Kd q~ = 0.06-0.03 log &.(7) (8) Hence eqn (10) can be derived for the equilibrium constants of reaction (9):t and combining eqn (10) with eqn (7) we derive log K = 1 + 0.5 log Kd - 1 7,!?$?eIIIy,~e11y. (1 1) The logarithmic equilibrium constant of reaction ( 12) may then be expressed as in eqn kred2 (13): S + FE"Y 'L + Ferl'Y k O X 2 (12) lOgK' = 1-0.5 log&-@elIly/~eI1y. (13) * The value of Kd was determined to be 2.3 x lo9 mol-I dm3 s-I at pH 6" and 2.4 x lo9 mol-' dm3 s-' in 0.05 mol dm-3 H,S0,,8.g and it has been demonstrated that the pH of the solution has little influence on this value." t The value of P(Fe"'Y/Fe"Y) was obtained by the use of l?(Fe1I1Y/Fe1'Y) = P(Fe"'/Fe") - 0.06 log (KpAIIIY /Kw,,~~v) and P(Fe3+/Fe2+) = 0.7 V.21K. TANAKA, E. END0 AND Y.KURIMURA 2987 The equilibrium constants K and K', and the rate constants kox, and kred,, can be evaluated using eqn (1 1) and (1 3), respectively, assuming kred, = kobs and kOXz = k&. This assumption is consistent with the first-order dependence of kobs and k& with respect to the concentration of thionine (leucothionine) and FeIIY (FeIIIY). Since the value of Kd has been estimated to be 0.01-0.05 at pH 2.03 and is expected to be smaller at higher pH values5 (Kd < lop4 at pH 6), Kd was given values in the range 10-2-10-5, and these are shown in table 4 together with the rate constants of photoreduction of thionine in the presence of Fe3+ and Fe2+ ions in 0.05 mol dm-3 H,SO, solutions.* It is seen from a comparison of the rate constants that the oxidation reactions of leucothionine and semithionine by FelI1 EDTA in pH 7.0 are faster than those by the Fe3+ ion in 0.05 mol dmP3 H2S0, solution.This might be of considerable significance in the design of photoelectrochemical systems consisting of FeII, a chelating agent and thionine. We thank Dr R. Tamamushi and Dr M. Hoshino for many helpful discussions and suggestions. We also thank Dr H. P. Bennett0 for help in preparing the manuscript. K. Weber, 2. Phys. Chem. (Leipzig), 1932, B15, 19. J. J. Weiss, Nature (London), 1935, 136, 794. E. Rabinowitch, J. Chem. Phys., 1940, 8, 551; 560. E. Rabinowitch and L. F. Epstein, J . Am. Chem. SOC., 1941, 63, 69. R. Hardwick, J . Am. Chem. SOC., 1958, 80, 5667. J. Schlag, Z. Phys. Chenz., Neue Folge, 1959, 20, 53. S. Anisworth, J. Phys. Chem., 1960, 64, 715. C. G. Hatchard and C. A. Parker, Trans. Faraday SOC., 1961, 57, 1093. H. Fisher, 2. Phys. Chem. (Frankfurt am Main), 1964, 43, 177. lo S. Matsumoto, Bull. Chem. SOC. Jpn, 1964, 37, 491. 11 R. Bonneau, J. Faure and J. Joussot-Dubien, Ber. Bunsenges. Phys. Chem., 1968, 72, 263. l 2 P. D. Wides, N. N. Lichten and M. Z. Hoffman, J . Am. Chem. SOC., 1975,!V, 2288. l 3 M. Isabel, C. Ferreira and A. Harriman, J . Chem. SOC., Faraday Trans. 1, 1977, 73, 1085. l4 P. Gomer, Electrochim. Acta, 1975, 20, 13. l 5 W. D. K. Clark and J. A. Eckert, Solar Energy, 1975, 17, 147. l 6 M. Kaneko, M. Sugai and A. Yamada, Makromol. Chem., 1978, 179, 2431. l 7 K . Shigehara, H. Matsunaga and E. Tsuchida, J. Polym. Sci., Polym. Chem. Ed., 1978, 16, 1853. l8 W. J. Albery, W. R. Bowen, F. S. Fischer, A. W. Foulds, K. J. Hall, A. R. Hillman. R. G. Egdell and l 9 J. Bjerrum, G. Schwarzenbach and L. G. Sillen, Stability Constants ZZ (The Chemical Society, 2o Y. Kurimura, R. Ochiai and N. Matsuura, Bull. Chem. Soc. Jpn, 1968, 41, 2234. A. F. Orchard, J. Electrocznal. Chem., 1980, 107, 37. London, 1958). A. J. de Bethune and N. A. Swendeman Loud, Standard Aqueous Electrode Potentials and Temperature CoeSJicients at 25 "C, ed. C. A. Hampell (1964). (PAPER 4/090)

ISSN:0300-9599    

DOI:10.1039/F19848002981

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1 , 1984,80, 2989-2998 Conversion of Methanol into Light Hydrocarbons on Erionite-Offretite(T) Zeolite BY STANISLAW CECKIEWICZ~ Groupe de Physico-Chimie Minerale et de Catalyse, Universite Catholique de Louvain, Place Croix du Sud, 1, 1348 Louvain-la-Neuve, Belgium Received 23rd January, 1984 The reaction of methanol on T-type zeolites has been studied at 400 "C in a flow reactor [weight hourly space velocity 1 g (methanol) per g (catalyst) per hour]. The results demonstrate the influence of decationization and dealumination of zeolites on catalytic lifetimes (maximum 4 h). To reduce the deactivation of HT zeolites by coke formation, Ni- and Pt-exchanged samples have been used, and the reaction has been studied in the presence of hydrogen. The deactivation of Pt-HT is much slower than that of HT but only paraffinic hydrocarbons are produced.Observations by infrared spectroscopy of the conversion of deuterated methanol in the 20-400 "C region are embodied in a mechanism of propylene formation invoking dimethyl ether, methoxy groups and zeolite-T hydroxy groups. The possibility has recently appeared of replacing naphtha steam cracking for olefin production by the selective synthesis of hydrocarbons from methanol. The use of methanol homologation to ethanol followed by dehydration to ethylene, and the direct catalytic conversion of methanol to olefins on Mobil zeolite catalysts apparently shows some economic potential.'t Among the traditional zeolites, those of the erionite-offretite type could be especially useful for light-olefin production.Well characterized zeolite-T3 composed of offretite and erionite intergrowth belongs to the class of shape-selective, narrow-pore zeolites. The potential of this zeolite for the selective formation of ethylene and propylene from methanol has been described previo~sly,~-~ and these findings are confirmed in this work. However, a serious problem remains, namely the fast deactivation of the catalyst by coke formation. The aim of this work is to study the deactivation of different protonic and cationic (Pt, Ni) forms of zeolite T and to gain some insight into the mechanism of converting methanol into hydrocarbons. The general reaction pathway for methanol conversion may be summarized by the following sequential reaction scheme : olefins { aromatics.methanol f dimethyl ether + water + olefins + paraffins (1) Although numerous studies of this reaction have been undertaken, many questions remain concerning (i) the role of dimethyl ether as a necessary intermediate, (ii) the identification of ethylene and/or propylene as the primary olefins produced and (iii) the type of C , coupling species participating in the initial C-C bond formation. Answers to these questions are controversial at present. t On leave from the Jagellonian University, Institute of Chemistry, Karasia 3, 30-060 Cracow, Poland. 29892990 METHANOL CONVERSION ON ZEOLITES Table 1. Sample preparation cation- crystalline- cation dealumination" exchange phase sample exchangea (% ) (% 1 solution content (% ) calcining in air HT/43 43 18 HCl 85 350 "C, 6 h HT/58 58 23 HCl 73 350 "C, 6 h HT/77 77 0 NH,Cl 94 500 "C, 16 h HT/84 84 7 NH,C1 86 500 OC, 16 h a Calculated from chemical analysis.25 Accepting some simplification, the situation may be summarized in the following way: (i) it is widely accepted that dimethyl ether is a possible intermediate for olefin f ~ r m a t i o n , ~ - ~ ~ (ii) a large number of data supports the opinion that propylene is a primary product13-18 and (iii) the mechanisms for C-C bond formation mainly involve surface alc~xyls,~~ l 9 ~ 2o carbene89 1 9 9 21-23 or oxonium ions.l0* 11, 16, 24 EXPERIMENTAL MATERIALS The samples with various extents of decationization and dealumination were obtained from an initial zeolite T, synthesized in the Institute of Chemistry, Jagellonian University, Cracow, and described by the chemical composition : 0.13 Na,O + 0.87 K,O -A1,03 * 7.4 Further details are given in table 1.The 0.2 wt % platinum-exchanged HT zeolite, Pt-HT, was prepared by competitive ion exchange of the calcined HT/77 sample (0.5 g) in 5 cm3 of an aqueous solution (pH 3) of H, (PtCl,) 6H,O (reagent grade, Merck). The 0.2 wt % nickel-exchanged HT/77 zeolite, Ni-HT, was obtained from 0.5 g of zeolite in 5 cm3 of an aqueous solution (pH 6) of Ni(N03), . 6H,O (reagent grade, Merck). The prepared mixtures were stirred for 24 h at room temperature and maintained in solution for one day without stirring. They were then filtered and dried overnight at 110 "C. The final platinum content was measured colorimetrically by the iodide method, whereas nickel content was analysed by atomic absorption spectroscopy.Methanol (reagent grade, Merck), deuterated methanol, CH30D (reagent grade, Institute of Nuclear Research, Swierk, Poland), and commercial-grade helium (99.995 % ) and hydrogen (99.90 % ) supplied by 1'Air Liquide were used without further purification. PHYSICOCHEMICAL CHARACTERIZATION X-RAY DIFFRACTION X-ray diffraction patterns were obtained using a Kristalloflex 805 Siemens diffractometer with Cu K, radiation (Ni filter). The crystalline-phase content (table 1) was estimated from the average ratio of intensities of four chosen X-ray reflections (20/O = 7.8, 20.6, 23.8 and 3 1.6) measured for the prepared samples and for an initial unexchanged zeolite-T sample. INFRARED SPECTROSCOPY Infrared spectra were recorded in the absorbance mode with a Perkin-Elmer 180 spectrograph in a heated i.r. cell attached to a conventional vacuum apparatus.For spectroscopic measurements, self-supporting wafers (1 3-16 mg ern-,) were prepared and evacuated in situ at 450 "C overnight. After evacuation, all reference spectra were recorded at room temperature. The experimental procedure was as follows.S. CECKIEWICZ 299 1 (a) Studies ofpyridine adsorption. Wafers of HT/43, HT/58, HT/84 were exposed to pyridine vapour at 150 "C for 30 min and outgassed at this temperature for 1 h. The i.r. spectra were measured at room temperature. (6) Studies of CH,OD adsorption and conuersion. Wafers of HT/77 were exposed for 1 h to CH30D vapour at various temperatures (20, 300 and 400 "C), cooled to room temperature, then outgassed at 150 "C for 1 h and cooled again to room temperature, after which the i.r.spectra were recorded. Spectra were calibrated for absorbance using a polystyrene film (0.06 mm). CATALYTIC TESTS The reactions were carried out in a glass reactor tube (i.d. 4 mm) attached to a conventional flow apparatus working at atmospheric pressure. Methanol was supplied to the zeolite powder (0.1 g forming a bed of ca. 20 mm height) in a stream of helium (or helium mixed with hydrogen) saturated with methanol at 0 "C (methanol partial pressure 28.7 Torr, i.e. ca. 3.83 kPa) and preheated to 100 "C before entering the reactor. In all experiments, the average weight hourly space velocity (WHSV) was 1 g (methanol) per g (catalyst) per hour.Experiments were carried out at 400 "C. All products (except coke) were analysed by on-line gas chromatography (Intersmat IGC 120 ML) using a Porapak Q (2 m) column. Analyses were carried out every 30 min in order to monitor deactivation of the catalysts as a function of time on stream. After each catalytic test a stream of pure helium was passed through the reactor at the reaction temperature for ca. 1 h to remove all adsorbed species from the zeolite. The amount of coke on the used catalysts was analysed by burning off the carbonaceous deposits at 1200 "C in oxygen and absorbing the carbon dioxide evolved in formaldimethylamide ; titrations were then carried out with non-aqueous tetra-n-butylammonium hydroxide.26 The pretreatment of the catalysts was carried out according to two different procedures. (i) In a stream of helium (2.6 dm3 h-l): the zeolites were heated at 450 "C for 2 h and at 400 "C for 1 h before each run (samples HT/43, HT/58, HT/77 and HT/84).(ii) In a stream of helium mixed with hydrogen (2.6 dm3 h-l, volume ratio H,: He = 3 or 1 : 7.5): the zeolites were heated at 500 "C overnight and at 500 "C for 1 h before each run (samples HT/77, Pt-HT and Ni-HT). We detected no influence of hydrogen on the activity of HT/77 during the pretreatment. RESULTS AND DISCUSSION DEACTIVATION STUDY Plots of the conversion of methanol into hydrocarbons as a function of time are presented in fig. 1 and 2. Fast deactivation of protonated zeolites was observed in all cases. The rate of deactivation at high temperatures (400 "C) depends mainly on two parameter^,^^ i.e.the degree of cation exchange for protons and the percentage of amorphous material in the samples. Fig. 1 shows that the shortest lifetime (ca. 1.5 h) was observed for HT/58, which contained the highest amount of amorphous material (cf. table l), and for HT/84, which had the highest degree of ion exchange and therefore the largest number of Bronsted-acid sites. Sample HT/43, with the lowest degree of exchange, had the longest lifetime ; nevertheless deactivation was nearly complete after 4 h. To improve the catalytic stability of HT/77 and reduce coking, Ni- and Pt-exchanged zeolites were used, and the reaction in the presence of hydrogen was studied. Fig. 1 indicates that hydrogen has a pronounced stabilization effect, by inhibiting coke deposition, on the Pt-HT sample when an H2 +He mixture of volume ratio 3 was used.Hydrogen activated on metallic platinum particles probably spills over onto the acid sites of the zeolite and reacts with the carbocation intermediates involved in coke formation.28 As a result almost total conversion of methanol was observed, but the2992 METHANOL CONVERSION ON ZEOLITES 0 C w ." 0 0 1 2 3 4 5 time on stream/h Fig. 1. Deactivation of the zeolite catalysts during reaction at 400 "C: x , HT/43; 0, HT/58; 0, HT/77; 0, HT/84 and 0, Pt-HT. products were paraffins. A decrease in the hydrogen concentration by using a H, 1 : 7.5 H, +He mixture caused a decrease in the above-mentioned effect. The deactivation of Ni-HT in the stream containing hydrogen was comparable to that of HT/77.This can be explained by the lower activity of Ni compared with Pt in producing spill-over hydrogen. We did not try to use a higher Ni loading in the zeolite because this bears a higher risk of pore blockage. The left-hand side of fig. 2 illustrates the increasing ease of olefin formation with increasing degree of ion exchange in decationized zeolite-T [histograms (b)-(4]. HT/58 [histogram (a)] is not taken into consideration because of its poor crystallinity, which caused the excessively rapid deactivation discussed previously. In this last case the observed traces of ethylene (the only olefin found) are probably merely a result of coke cracking. For HT/77 and HT/84 the selectivity to ethylene and propylene reached 75 wt% after 1.5 h on stream, but this yield diminished with longer reaction times.Increasing amounts of dimethyl ether and methanol were then observed in the products. On the right-hand side of fig. 2 the influence of Pt and Ni in zeolite HT/77 on changes to the conversion of methanol can be seen. The presence of hydrogen increased the ratio of paraffin to olefin in the products up to the complete absence of olefins over Pt-HT [histogram (b)]. Histograms (c) and (e) corresponding to HT/77 are similar. This indicates that the influence of hydrogen on methanol conversion is negligible when there is no Pt in the sample. CARBONACEOUS DEPOSITS Table 2 reports results concerning the amounts of coke deposited on the zeolite samples after their total deactivation, except for Pt-HT, where the measurement corresponds to a catalyst which worked for 10.5 h with only 50% deactivation.The HT catalysts showed a considerable tendency for coke f~rmation.~ The lowest percentage of coke was observed for zeolites with the lowest degree of ion exchange, i.e. HT/43 and HT/58. For the latter, imperfections in crystallinity accelerated deactivation. The smallest amount of carbon deposited on HT/58 may point to theS. CECKIEWICZ 2993 40 (d1 100 20 50 40 20 h &? c. t n a l k e n e 20 Cl-c.6 c 2 c 3 c4 ( b ) 40 20 c 40 c 20 L straight-chain hydrocarbons Fig. 2. Hydrocarbon distribution in the products of methanol conversion at 400 "C after 1.5 h on stream. Histograms for runs in helium: (a) HT/58, (b) HT/43, (c) HT/77 and ( d ) HT/84.Histograms for runs in helium + hydrogen : (e) HT/77 (H, :He = 3), (f) Ni-HT (H, : He = 3), (g) Pt-HT (H2: He = 1 : 7.5) and (h) Pt-HT (H2: He = 3). Table 2. Determination of coke content methane selectivity (wt%) sample T/"C (after time on stream/h) coke (wt % ) HT/43 400 5.6 (1.5) 21.6 (4) 7.9 HT/77 400 5.4 (1.5) 15.7 (3.5) 12.0 HT/84 450 8.5 (1.5) 66.0 (2.5) 12.0 HT/58 400 32.9 (1.5) 32.9 (1.5) 6.0 HT/84 400 3.3 (1.5) 26.3 (3) 11.6 Pt-HT 400 3.1 (1.5) 7.0 (10.5) 9.02994 METHANOL CONVERSION ON ZEOLITES 3700 3500 3300 wavenumber/cm -' 1600 1500 1400 wavenum berlcm-' Fig. 3. Left-hand side: infrared spectra in the hydroxy-group region for: (a) HT/43, (b) HT/58, (c) HT/77 and (d) HT/84. Right-hand side: infrared spectra for adsorbed pyridine after outgassing at 150 "C for: (b) HT/43, (c) HT/58, (d) HT/77 and (e) HT/84; spectrum (a) constitutes an example of i.r.spectra obtained after high-temperature evacuation at 450 "C. The vertical lines correspond to 3610 cm-' (left-hand side) and 1490 cm-' (right-hand side). growth of macromolecular coke precursors on the surface, leading to a retarding effect on methanol conversion, as has been suggested elsewhere.6 Temperature is an important factor in coke formation. The results obtained at higher temperatures (450 "C) on HT/84 (table 2) support the idea that the mobility of adsorbed species on zeolite increases and that hydrogen-transfer reactions lead to the formation of increasing amounts of methane6 as well as polyaromatic coke strongly attached to the zeolite surface.18 It is remarkable that the extent of coke formation on HT increased with the ability of these catalysts to form olefins [see fig.2, histograms (c) and (41. Pt-HT showed a medium amount of coke as a result of successively accelerating deactivation with time on stream. In the presence of platinum, the amount of coke was reduced by 25% for Pt-HT in comparison with HT/77, and channels of Pt-HT zeolite were still open after > 10 h on stream.S. CECKIEWICZ 2995 3000 2800 2600 2400 2200 2000 w avenurn berlcm -' 1700 1600 1500 1LOO wavenumberlcm-' Fig. 4. Infrared spectra: (a) HT/77 after evacuation at 450 "C; (b)-(d) HT/77 after adsorption of CH,OD at: (6) 20, (c) 300 and (d) 400 "C (followed by outgassing at 150 "C). The vertical line corresponds to 2660 cm-'.INFRARED STUDIES SURFACE ACIDITY Fig. 3 presents the OH stretching region, between 3800 and 3300cm-l, of decationized zeolite-T samples outgassed at 450 "C. Three absorption bands at 3640-3650, ca. 3610 and 3560 cm-l are due to surface OH groups, in agreement with previous 2 9 9 30 The band at 3740 cm-l is assigned to non-specific SOH hydroxyls. The intensity of OH bands characteristic of Bronsted-acid sites increased significantly with the degree of ion exchange in HT. - Fig. 3 also shows the i.r. spectra obtained after chemisorption of pyridine at 150 "C on the samples studied. Several bands appear in the 1650-1400 cm-l region. The band at 1540 cm-l is attributed to the interaction of pyridine with Bronsted-acid sites, whereas the band at 1455 cm-l corresponds to the interaction with Lewis-acid The bands at ca.1620 and 1490 cm-l are common for both of these types of acid sites. Generally the surface acidity increases with the degree of ion exchange, with the exception of HT/58 in which the amorphous material can be an obstacle to pyridine chemisorption. In this case the recorded bands are weaker than they should be, when we take into consideration the concentration of OH groups in HT/58 (see the region 3800-3300 cm-l). CONVERSION OF CH,OD Fig. 4 presents the i.r. spectra of HT/77 wafers exposed to deuterated methanol (CH30D) at various temperatures (20, 300 and 400 "C) and outgassed in each2996 METHANOL CONVERSION ON ZEOLITES Table 3. Main i.r. frequencies (cm-l) and their assignments CH,OD adsorption temperature/'C ca.3010 2960 2930 2850 2750 2660 - 15 10- 1490 1475-1450 - 2955 2915 2850 2750 2660 1520 1 500- 1490 1410 1400-1370 - 1475- 143 5 - 2955 2915 28 50 2660 1605 ca. 1520 - 1475-1 435 1410 1400-1370 CH stretching CH stretching CH stretching CH, stretching CH stretching OD stretching OD stretching C-C stretching C-C stretching OCO stretching CH, deformation =CH, deformation CH, deformation experiment at 150 "C. The i.r. bands characteristic of chemisorbed surface species are listed in table 3. We observe conspicuous differences between the i.r. spectra measured after adsorption of CH,OD on HT/77 at 20 and 300 or 400 "C. The exposure of zeolite to deuterated methanol at 20 "C, followed by outgassing at 150 "C, leads to zeolite methoxylation, manifested by the presence in the spectra of OCO bands at 1510-1490 cm-l accompanied by two shoulders at ca.3010 and 2930 cm-I [fig. 4(b)]. The band patterns in the spectra corresponding to adsorption at 300 and 400 "C are caused chiefly by methyl and methylene groups constituting structural parts of the chemisorbed hydrocarbon species [fig. 4(c) and (41. Additionally, for 400 "C the presence of bands near 1605 and 1520 cm-l is characteristic of vibrations in aromatic rings representative of carbonaceous deposits. After the chemisorption of CH,OD on HT/77, indicated by the OD bands at 2750 and 2660 cm-I at different temperatures, there is a complete absence of the bands at 2220 and 2070cm-l corresponding to C-D vibration in CD,. This is proof that CH,OD does not transform with the formation of C-D bonds even at higher temperature (400 "C) in our case.The behaviour of the zeolite hydroxy groups after the introduction of CH,OD into the i.r. cell was also investigated. As a rule, a weakening of the OH band intensities, increasing with reaction temperature, particularly for the band at ca. 3610 cm-l, was observed. This i.r. spectroscopic study corresponds well with the previous one1* giving a general network of methanol transformations on HT. MECHANISM In the interaction of methanol or deuterated methanol with surfaces, the surface 2 o q 3o play a prominent role: (2) acidity of the catalyst (HZ) and surface methoxylation7* CH,OD + HZ -+ CH,Z + DOH. This reaction is followed at ca. 150 "C or higher temperatures by the formation of dimethyl ether: CH,Z + CH,OD + CH,OCH, + DZ.(3)S. CECKIEWICZ 2997 This is manifested in the i.r. spectra (fig. 4) by the presence of bands characteristic of OD groups and OCO bonds. At higher temperatures (300-400 "C) dimethylether molecules, being in the vicinity of acidic sites (OH or OD groups) and methoxy groups, lose oxygen, with the simultaneous rearrangement to the propylcarbenium ion which is finally transformed to propylene : H Z + H Z CH,OCH, + CH,Z CH,CHCH, + Z- + HOH CH,-CH=CH, + HZ + H,O or in the case of deuterated methanol: (4) D Z + HZ CH,OCH, + CH,Z + CH,CHCH, + Z- + DOH CH,-CH=CH, + HZ + DOH. ( 5 ) The desorbed products from the earlier stages of the reaction consist mainly of dimethylether and propylene [ref. (1 8), table 4, and ref.(20), table 11. In reaction (4), published ear lie^,^ protons belonging to HZ surface acidic groups may be supplied by deuterium, as in reaction (5), without the formation of C-D bonds in the propylene molecule (see fig. 4). The simple simultaneous decomposition of surface complexes composed of the dimethylether molecule, the methoxy group and acid OH sites, suggested previ~usly,~ is rather doubtful. A multi-step mechanism proposed recentlyl0' l1 seems to be more probable, e.g. the propyl carbocation could be formed via a methylethylether intermediate, involving a Stevens-type rearrangement of the trimethyloxonium ion. Nevertheless the general idea,7 expressed by reaction (4), remains the same. I thank Prof. B. Delmon for helpful comments and Dr A. Cichocki for the gift of the zeolites.I acknowledge the financial support of the Services de Programmation de la Politique Scientifique (Belgium) for a stay at the Universitd Catholique de Louvain. J. P. Leonard and L. H. Weiss, Energy Progress, 1981, 1, 41. T. Inui and Y. Takegami, Hydrocarbon Process., 1982, November, 117. R. L. Gorring, J. Catal., 1973, 31, 13. C. D. Chang, W. H. Lang and A. J. Silvestri, US. Patent, 1977, 4,062,905. S . Ccckiewicz, Bull. Acad. Pol. Sci., Ser. Sci. Chim., 1979, 27, 629. E. Langner, Appl. Catal., 1982, 2, 289. S. Ccckiewicz, J. Colloid Interface Sci., 1982, 90, 183. C. D. Chang and A. J. Silvestri, J , Catal., 1977, 47, 249. P. Dejaifve, J. C. Vbdrine, V. Bolis and E. G. Derouane, J. Catal., 1980, 63, 331. lo T. Mole and J. A. Whiteside, J.Catal., 1982, 75, 284. l1 G. Perot, F. X. Cormerais and M. Guisnet, J. Mol. Catal., 1982, 17, 255. l2 M. B. Sayed and R. P. Cooney, Aust. J. Chem., 1982, 35,2483. l3 J. R. Anderson, K. Foger, R. Mole, R. A. Rajadhyaksha and J. V. Sanders, J . Catal., 1979,58, 114. l4 B. J. Ahn, J. Armando, G. Perot and M. Guisnet, C.R. Acad. Sci., Ser. C, 1979, 288, 245. l5 C . D. Chang, W. H. Lang and R. L. Smith, J. Catal., 1979,56, 169. l6 J. P. van den Berg, J. P. Wolthuizen and J. H. C. van Hooff, in Proc. 5th Znt. Conf. Zeolites, ed. L. V. C. Rees (1980), p. 649. F. X. Cormerais, G. Perot, F. Chevalier and M. Guisnet, J . Chem. Res. (S), 1980, 362. S. Ccckiewicz, J. Chem. SOC., Faraday Trans. I , 1981,77, 269. l9 P. Salvador and W. Kladnig, J . Chem. Soc., Faraday Trans. I , 1977, 73, 1153. 2o Y. Ono and T. Mori, J. Chem. Soc., Faraday Trans. I , 1981, 77, 2209. 21 P. B. Venuto and P. S. Landis, Adv. Catal., 1968, 18, 259. 22 E. A. Swabb and B. C. Gates, Znd. Eng. Chem., Fundam., 1972, 11, 540. 23 C. D. Chang and C . T-W. Chu, J . Catal., 1982,74, 203.2998 METHANOL CONVERSION ON ZEOLITES 24 W. W. Kaeding and S. A. Butter, J . Catal., 1980, 61, 155. 25 A. Cichocki, Krist. Techn., 1978, 13, 991. 26 R. F. Jones, P. Gale, P. Hopkins and L. N. Powell, Analyst (London), 1965, 90, 623. 27 S. Ceckiewicz, React. Kinet. Catal. Lett., 1981, 16, 11. 28 N. S. Gnep, M. L. Martin de Armando and M. Guisnet, Spillover of Ahorbed Species, ed. G. M. 29 T. J. Weeks, Jr., C. L. Angel1 and A. P. Bolton, J. Catal., 1975, 38, 461. 30 S. Ceckiewicz, React. Kinet. Catal. Lett., 1980, 13, 297. 31 G. Poncelet and M. L. Dubru, J . Catal., 1978, 52, 321. 32 M. Falk and E. Whalley, J. Chem. Phys., 1961, 34. 1554. 33 I. Kanesaka and K. Kawai, Spectrochim. Acta, Part A , 1982, 38, 549. 34 B. Kontnik-Matecka, M. Gorska, J. Eysymontt and A. Salek, J. Mol. Struct., 1982, 80, 199. Pajonk, S. J. Teichner and J. E. Germain (Elsevier, Amsterdam, 1983), p. 309. (PAPER 4/ 122)

ISSN:0300-9599    

DOI:10.1039/F19848002989

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1984, 80, 2999-3009 Reaction of Hydrogen Atoms with Propane in the Temperature Range 298-534 K BY ROGER M. MARSHALL, HOWARD €'URNELL* AND ADRIAN SHEPPARD Department of Chemistry, University College of Swansea, Singleton Park, Swansea SA2 8PP Received 3 1st January, 1984 The reaction of hydrogen atoms with propane has been investigated at a helium pressure of ca. 5 Torr in the temperature range 298-534 K using a conventional flow-discharge system coupled to a quadrupole mass spectrometer. Two sets of pseudo-first-order experiments were performed: (i) where [C,H,] 9 [HI, in which case decay of [HI was measured, and (ii) where [C,H,] + [HI, in which case formation of CH, was measured. In order to evaluate the rate constant, k,, of the reaction assumptions concerning the stoichiometry of the overall process have to be made.It is shown that previous assumptions are wrong. We have used computer integration of a detailed mechanism for the overall process to evaluate self-consistent values of k, from our experimental data. We obtain the Arrhenius parameters log(k,/cm3 mol-l s-l) = (13.81 *0.07)-(28.8+0.6) kJ molF1/2.3 RT from our results for T 2 370 K. At lower temperatures there is evidence for curvature of the Arrhenius plot and the activation energy drops to 21.8 f 1.8 kJ mol-l for temperatures < 370 K. In recent years we have investigated the reactions of hydrogen atoms with methanel and with ethane2., with particular emphasis on the suggested curvature of the Arrhenius plots for these reactions.It was concluded that the plots were, in fact, straight over the very wide temperature ranges 400-1800 K for H+CH4 and The present paper extends this work to the reaction H + C,H,. The rate constant for this reaction has been measured by several groups of workers, the measurements falling mainly into two classes. The first is an indirect and difficult method involving the addition of propane to a reacting H, + 0, The second is the use of a flow-discharge system with suitable monitoring of the reacting specie^,^? this method being, in principle, direct but open to the difficulty of an assumption about the overall stoichiometry, n, of the reaction, i.e. the total number of H atoms used per initial attack of H on C,H,. Previous workers, believing methane to be the sole product, have assumed the overall reaction to be 330-1 350 K for H C2H6.6H + C,H, -, H, + 3CH4 and hence n = 6. However, it is clearly impossible for the situation to be as simple as this since after the initial metathesis H+C,H, -+ H,+C,H, 29993000 REACTION OF HYDROGEN ATOMS WITH PROPANE the most likely reactions are H + C3H, + H, + C,H, H +C,H, + C3H,* and then the C,H,* can decompose or be collisionally deactivated. The problem is further exacerbated by the production of propene which, at low temperatures, reacts with H at a rate that is orders of magnitude faster than that of propane and thus secondary reactions must be taken into account. In these circumstances it is to be expected that the value of n will be dependent upon the experimental conditions and must be evaluated for each individual experiment.We have used computer integration of a detailed mechanism, set out in table 1, for this purpose. This mechanism is based upon the best current understanding of the corresponding reactions with methane, ethane, ethene and propene. EXPERIMENTAL Details of the flow-discharge apparatus, mass spectrometer and experimental procedure have been described previ~usly.~ THEORY To simplify the kinetic analysis of our results we have attempted to work in pseudo-first-order conditions. There are two ways of achieving this : either with a large excess concentration of C,H, over H or vice versa. Different methods of data treatment are required in the two cases. EXPERIMENTS WITH AN EXCESS OF PROPANE The appropriate condition required to achieve pseudo-first-order conditions is GH,I D [HI where n is the total number of hydrogen atoms used per molecule of propane reacting via reaction (1).When carrying out this type of experiment one measures the concentration of hydrogen atoms at the inlet of the mass spectrometer when a constant flow of propane is introduced to the reactor successively via the seven inlet jets, i.e. seven different reaction times are used. Simple theory then gives In ([Hl~O/[Hlt) = nk1[C3H81 where the subscripts refer to concentrations measured with no propane flowing and to those with propane flowing for a reaction time t. The ratio of concentrations is conveniently replaced by the ratio of mass-spectrometric peak heights h, (appropriately corrected for background etc.), at m / e 1 .We thus have (1) Note that [C3H,] is effectively constant on account of the large excess of C,H, over H. The evaluation of k, is then dependent on an evaluation of n. This is described in detail below. In h, = -nk,[C,H,] t+constant. EXPERIMENTS WITH AN EXCESS OF HYDROGEN ATOMS In principle, to achieve pseudo-first-order conditions we requireR. M. MARSHALL, H. PURNELL AND A. SHEPPARD 300 1 Table 1. Mechanism and rate constants assumed in computer integration reaction rate constanta reference H + C,H, -+ H, + i-C,H, H + i-C3H, -+ C,H,* H + i-C3H, -+ C3H6 + H, M + C3H,* -+ C3H, + M C3H,* -+ CH, + C2H5 C3H,* -+ CH, + C2H4 H + C3H6 -+ i-C3H: H + C3H6 -+ CH, + C2H4 M + i-C,H,* -+ i-C3H, + M H + i-C3H: -+ CH, + C,H5 H + C2H4 -+ C2H5 H + C,H5 -+ C2H: -+ 2CH3 CH, + H -+ CH, CH, + i-C3H7 -+ i-C,H,, 2CH3 -+ C2H6 H -+ 0.5 H, - 3.7 x 10130 2.7 x 10136 1.6 x 1 0 5 b 1.3 x 105b 4.1 x 1012.7 exp (- 5230/RT)C exp (- 1 1 500/RT)C 1012.6 4.1 x 10'2b,d 2.0 x 1014b 3.7 x 1013e 1013.1 exp (- 1000O/RT)C 1.5 x 10l2 4.0 x 10l2 6.3 x 10l2f measured directly 8 7 7 see text 7 1 1 - a Units: s-l or cm3 mol-1 s-l as appropriate.Constraints applied making ratios of rate Units Assumed equal to k, because of the identical nature of the Assumed equal to k2 because of the identical nature of the reactions. f Low value constants in agreement with reported measurements of k,/k,, k5/k,, k,/k, and k,,/k,. of activation energy: J mol-l. reactions. used in error but immaterial because of insensitivity of calculations to value chosen.in which case [HI is effectively constant and [C,H,] decays with time. Such conditions are, in fact, impossible to achieve because of the inevitable occurrence of the heterogeneous process wall H-0.5 H, and the consequent decay of [HI along the reactor tube which, in turn, precludes any simple algebraic treatment of the results. Indeed we are forced to carry out a much more complex experiment. This involves first the determination of the absolute value of [HI as a function of distance (and hence time) along the flow tube. The technique used was our previously described ethene-titration technique,, in which a large excess of ethene is added to the flow reactor via one of the inlet jets. The mass spectrometer was then used to measure the yields of the saturated products (CH,, C,H6 and C,H,) from which the value of [HI in the vicinity of the inlet jet was evaluated.From these results the value of k16 was determined via a plot of In [HI against t since we have In [HI = - k16t + constant for a first-order process. The experiment proper then consisted of adding a known flow rate of propane via an inlet jet for which [HI was known. In principle we could then have measured the propane surviving at the mass-spectrometer inlet. However, it was found that the accuracy of measurement of the small decay of propane was very poor on account3002 REACTION OF HYDROGEN ATOMS WITH PROPANE Table 2. Eff'ects of changes in assumed rate-constant values on calculated values of stoichiometry and of methane yield stoichiometry, n methane yield, rn T = 315 K T = 515 K T = 315 K T = 515 K rate constant x 2 xO.5 x 2 xO.5 x 2 xO.5 x 2 x0.5 1 293 4,5,6 7 8 9,lO 11 12 13 14 15 ratio of rate constants 0.91 1.06 0.92 1.10 1.81 0.52 1.00 1.00 1.02 0.97 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.07 0.92 1.06 0.96 0.99 1.02 1.01 1.00 1.01 1.00 1.02 0.99 1.00 1.00 1.00 1.00 1.00 1.00 1.02 0.98 1.01 0.99 1.00 0.99 1.00 1.00 1.00 0.99 1.00 1.00 1.01 0.98 1.02 0.99 1.00 1.00 0.99 1.01 1.00 1.00 1.00 1.00 1.00 1.00 0.99 1.00 1.00 1.00 T = 315 K T = 515 K T = 315 K x 1.1 xO.9 x 1.1 xO.9 x 1.1 xO.9 1.43 0.61 1.00 1.01 1.00 1.00 0.99 1.01 1.02 0.98 1.00 1.00 1.01 0.99 1.00 1.00 1.02 0.96 0.99 1.00 0.99 1.00 T = 515 K x 1.1 xO.9 kJk2 1.02 0.98 1.00 1.00 1.04 0.97 1.02 0.99 k5lk4 1.01 0.99 1.00 1.00 1.01 0.98 1.01 0.98 k,lk* 1.00 1.00 1.00 1.00 1.01 0.99 1.01 0.98 klOlk9 1.00 1.00 1.00 1.00 1.02 0.98 1.01 0.99 All calculations done for assumed reaction conditions: P = 5 Torr and FHe = 1000 pmol s-l.Other flow rates/pmol s-l assumed were FH = 1.00 and FC3Hs = 5 for calculation of n and FH = 15 and E;c3H8 = 0.5 for the calculation of rn. Typical reaction times for 3 15 and 5 15 K were used, as appropriate. The values shown in the first part of the table are the factors by which the calculated values of n or of m are multiplied when the indicated rate constants are multiplied by 2 or by 0.5. Where a set of rate constants is indicated, the value of each member of the set is multiplied so as to maintain the constraints referred to in footnotes b, d and e of table 1. The effects of relaxing these constraints by factors of 1.1 and 0.9 are shown in the second part of the table.of the low extents of reaction, particularly at the lower temperatures, and it was more convenient and accurate to measure instead m, the yield of the only observable product, methane. The value of k, was then evaluated using the computer integration procedure described, in detail, below. Briefly, using the known values of [HI, and [C,H,],, the initial concentrations at the inlet jet, together with the measured value of k,, and known reaction time, t , we can assume a value for k, to compute the value of m at the mass-spectrometer inlet. Comparison with the measured value permits adjustments in the assumed value of k, to be made until agreement between computed and measured values of rn is obtained.The value of k , giving this agreement was assumed to be the 'true' value. COMPUTER INTEGRATION PROCEDURE As described above the evaluation of k , requires -either a calculation of the stoichiometry, n, or of the yield of methane, m. To do this we postulate a detailed mechanism for the reaction together with values for the rate constants of theR. M. MARSHALL, H. PURNELL AND A. SHEPPARD 3003 elementary reactions involved. Table 1 shows the mechanism and the values of the rate constants assumed. It should be pointed out that reaction (1) is assumed to produce only i-C,H,, which permits us to assume rate-constant values derived in our study of the reaction of hydrogen atoms with propene. Discussion of this assumption and of the mechanism as a whole is deferred to a later section.The integration of the coupled differential equations derived from the assumed mechanism and rate-constant values was achieved using a library routine which uses a variable-order, variable-step-length Gear m e t h ~ d . ~ Since our method of data analysis depends upon the computed values of n and m we have carried out extensive computations to determine the sensitivity of the computed values to changes in the assumed rate-constant values. The results of this analysis are summarised in table 2, which contains two sections each describing changes in the calculated values of n and rn for typical reaction conditions at both ends of the temperature range of the experiments. The first section shows the factor by which the calculated values of n and rn are altered when individual rate constants are multiplied by a factor of 2 or by a factor of 0.5.Constraints were applied to the rate constants for reactions (2)-(6), (9) and (10) so as to maintain rate-constant ratios in agreement with experimental measurements. The second section of table 2 shows the effects of relaxing these constraints by factors of 0.9 and 1.1. All of these computations show a remarkable lack of sensitivity to the assumed rate-constant values. Thus the calculated values of YI are only significantly affected by changes in the value of k, and k , when the constraints are applied, and the effect is always < 10% even for a 100% change in assumed rate constant. Moreover, the effect of a change in k , is irrelevant since its value is measured directly in this work.Calculated values of m are effectively only dependent upon the assumed value of k,. When the constraints are relaxed there is a relatively small effect on the calculated values of n and m, particularly at the higher temperature. From the results of these computations we conclude that our evaluations of n and m provide a reliable basis for the determination of k , from the presently obtained experiment a1 results. RESULTS Experiments were carried out at temperatures in the range 298-534 K at a total helium pressure of ca. 5 Torr corresponding to a flow rate in the range 400-1 100 pmol s-l and flow velocities in the range 4.2-19.3 m s-l. Hydrogen-atom flow rates were in the range 0.8-23.5 pmol s-l, the flow rate of propane being adjusted as necessary to ensure the required reaction conditions, i.e.an excess either of hydrogen atoms or of propane. Complete tables of results are available.1° Table 3 shows typical results for the excess of propane experiments where the condition n[C,H,II[HI $- 1 must be achieved. Values of this ratio were almost always > 6, which is adequate for the condition to be deemed to be obeyed. This is confirmed by our inability to detect any significant curvature on any of the plots of lnh, against t from which values of nk, were obtained. Fig. 1 shows typical examples of such plots. Table 4 shows typical results for experiments in which hydrogen atoms are in excess. Note that the method of data treatment used does not, in fact, require any condition to be satisfied by the flow rates of hydrogen atoms and propane.It was experimentally convenient to operate in conditions of an excess of hydrogen atoms because in those circumstances significant amounts of reaction occurred over the whole length of the reactor and the observable products consisted predominantly of methane, which3004 REACTION OF HYDROGEN ATOMS WITH PROPANE Table 3. Typical results for experiments in an excess of propane [C,H,I P/Torr a/m s-' /nmol cmP3 [HI, /nmol cmP3 kl n / lo9 cm3 mol-l s-l 303 340 392 444 485 530 5.2 4.54 1.22 5.2 4.94 1.26 5.4 4.82 1.08 5.2 6.60 1.22 5.3 7.3 1 0.61 5.2 7.98 0.76 0.36 0.33 0.34 0.37 0.28 0.3 1 3.9 1 .o 3.7 2.5 3.2 8.7 2.8 27 2.7 52 2.5 87 \ 0 2 5 50 75 t/ms Fig. 1. Typical plots of In h, (h, in arbitrary units) against t/ms.Experimental conditions, T/K and [C,H,]/nmol ern-,, respectively: (a) 298 and 1.54, (b) 332 and 1.74, (c) 392 and 1.08 and (d) 485 and 0.61. makes the observed yield an excellent monitor of the amount of reaction occurring. The results given in table 4 show the self-consistency of the values of k , obtained by our computer fitting procedure for several different reaction times in a given set of experimental conditions. This was typical over the whole range of experiments. Fig. 2 shows an Arrhenius plot of all the values of k,, the points arising from the two different methods of measurement being distinguished by the use of different symbols. It is immediately apparent that the results obtained by the two methods are in excellent agreement over the'whole temperature range.For the results at lo3 K/T < 2.7 we calculate loglo(kl/cm3 mol-1 s-l) = (13.81 *0.07)-(28.80f0.58) kJ mol-l/2.3 RT whence the error limits here and subsequently are 95% confidence limits calculated via least-squares methods.R. M. MARSHALL, H. PURNELL AND A. SHEPPARD 3005 Table 4. Typical results for experiments in an excess of hydrogen atoms (a) T/K = 315, P/Torr = 5.4, [C,H,],/nmol CM-~ = 0.122, kls/S-' = 6.7 reaction time/ms 40 52 63 74 86 96 rnlnmol cm-3 0.040 0.052 0.061 0.074 0.088 0.098 (b) T/K = 516, P/Torr = 5.2, [C,H,],/nmol cm-3 = 0.081, kls/s-l = 22.2 [H],/nmol 4.0 4.3 4.6 5.1 5.5 5.7 fitted k,/108 cm3 mol-' s-' 2.2 2.1 2.0 2.0 2.0 2.0 reaction time/ms 31 37 44 51 57 [H],/nmol cmP3 2.1 2.2 2.4 2.7 3.1 rnlnmol cm-3 0.16 0.18 0.20 0.21 0.23 fitted kl/lOs cm3 mol-l s-' 6.7 7.7 7.9 8.1 8.0 0 2 .o 2.5 3 .O Fig.2. Arrhenius plot for k,: 0, measured in an excess of H and 0, measured in an excess of C,H,. The plot is drawn as two intersecting straight lines, as described in the text. lo3 KIT3006 REACTION OF HYDROGEN ATOMS WITH PROPANE Above lo3 K/T = 2.7 the scatter of the points becomes greater, largely on account of the very small decays we are required to measure (cf. fig. 1). Nevertheless, it is quite clear that the Arrhenius plot curves upwards. For this range we calculate loglo(kl/cm3 mol-1 s-l) = (1 2.8 1 0.28) - (21.82 1.76) kJ mol-l/2.3 RT parameters which are significantly different from those obtained in the higher temperature range. Note that it is convenient to draw the plot as two intersecting straight lines.As we have pointed out before,,, results of almost unattainable accuracy are required to actually draw a smooth curve through such experimental points. DISCUSSION MECHANISM The mechanism given in detail in table 1 assumes that the initial attack of H on propane generates only i-C,H, radicals. Plainly this is not correct in that the reaction must have two components, one generating 1-C,H, and the other n-C,H,: H + C,H, -+ H, + i-C,H, H + C,H, -+ H, + n-C,H,. The only direct measurement of the ratio kli/kln seems to be that of Campbell et a2.,l2 who for the range 318-450 K (conveniently overlapping most of the presently used temperature range) obtained log (kli/kln) = - (0.11 f 0.1 1) + (437 42)/T which evaluates to give kli/kln = 22.2 1.6 at 300 K = 5.2k0.4 at 530 K and it is clear therefore that formation of 1-C,H, is the major process throughout the temperature range of the present experiments; indeed, at the low-temperature end it is the predominant process.Thus, given the insensitivity of the computer integrations to the values assumed for the rate constants we conclude that our assumed mechanism and the calculations based on it are not significantly affected by our neglect of the reactions of n-C,H,. All the rest of the steps in the mechanism, with the exception of reactions (7), (8), (1 1) and (16), are given rate constants taken from the literature and assumed to be independent of temperature. This assumption is certainly reasonable since these reactions are (i) radical + radical processes, (ii) collisional deactivation processes or (iii) reactions of an excited state, all of which would be expected to have little, if any, temperature dependence.In any case, the insensitivity of the computer integrations to assumed values ensures that there could be no significant effect even if there were some temperature dependence of these reactions. Of the four remaining reactions the appropriate value of k,, was measured directly in this work and Arrhenius parameters were taken from the literature for k, and k,. The parameters for k,, are estimated using the activation energy evaluated by Marshall and Purne1P3 with the pre-exponential factor chosen so as to give a value kll/cm3 mol-1 s-l = 2.2 x lo1, at 298 K in line with rneasurement~~~ made in helium at pressures of ca.5 Torr. Thus the pressure dependence of this rate constant is, at least partly, taken into account. However, once again any uncertainties in rate- constant values are cancelled out by the insensitivity of the computer integrations to assumed values.R. M. MARSHALL, H. PURNJZLL AND A. SHEPPARD 3007 REACTION STOICHIOMETRY As pointed out in the Introduction, the usual assumption, n = 6, cannot be correct and the stoichiometry must depend on the experimental conditions. This is illustrated in table 3, which shows calculated values of n covering the range of experimental conditions used in this work. It is immediately clear that the assumption n = 6 is never valid since n ranges from ca. 4.0 at 300 K to ca.2.5 at 530 K. Note that there is some dependence also on the reactant concentrations and that we are discussing here the gross effects of changes in temperature for the conditions of an excess of propane. This variation of n can be readily rationalised. At low temperatures, the rate constant of reaction (1) is very much lower than are those of reactions (7), (8) and (1 1) and so any ethene or propene formed react further. Moreover, the concentrations of any radical produced are so low that their only radical+radical reaction is that involving H. Thus the only observable product is methane, the other products being H, and reformed propane. In these circumstances the overall process approximates to 4H+C,H8 -+ H2+0.5 C3H8+ 1.5 CH, i.e. n = 4. Reaction (1) has a much higher activation energy than have reactions (7), (8) and (1 1) and, thus, at a sufficiently high temperature the reactions of H with ethene and with propene will be insignificant in comparison with the reaction with C,H,, which is, of course, present in excess.Thus one would expect the major part of the overall reaction to correspond to reactions (1)-(3), i.e. n = 2.0, with, of course, some extra €4 used up reacting with the decomposition products of C,H,*. Thus a value of n greater than 2.0 would be expected, i.e. as shown in table 3. Note that this is not the lower limit to the value of n since at very high excesses of propane, at a sufficiently high temperature, one would expect the value of n to approach unity, corresponding to a mechanism consisting solely of reaction (1) and the removal of C,H, radicals by their mutual combination and disproportionation reactions, which are negligible in the presently used conditions and hence not included in the mechanism given in table 1.COMPARISON WITH PUBLISHED VALUES OF k , The only data with which the presently obtained values of k, may be compared unambiguously are the high-temperature data,,? all obtained by the difficult method of addition of propane to reacting mixtures of H2+02, this method requiring no assumption of reaction stoichiometry. Such data are widely scattered, covering a range of a factor of ca. 10 at temperatures of ca. 700-900 K. We may, however, com- pare our results with the two most recent studies. Baker et aZ.* determined the single value k,/cm3 mol-1 s-l = 5.4 x lo1' for 753 K, which may be compared with the value 6.5 x loll calculated from the presently obtained Arrhenius parameters.Bearing in mind the long extrapolation, over 200 K, the agreement is most satisfactory. Azatyan15 obtained log (k1/cm3 mol-1 s-l) = 13.73 - 30.1 kJ molk1/2.3 RT for the temperature range 843-933 K. At 900 K, this expression gives k,/cm3 mol-1 s-l = 1.0 x 10l2, which may be compared with the value 1.4 x 10l2 obtained by extrapolation of the presently obtained parameters. Clearly, the absolute values of k, are in excellent agreement bearing in mind the long extrapolation of almost 400 K and, moreover, the Arrhenius parameters agree to within any reasonable estimate of experimental error. Comparison with low-temperature data is usually ambiguous because of the3008 REACTION OF HYDROGEN ATOMS WITH PROPANE uncertainties about the stoichiometry, n, in the reaction systems used by the various authors.The most comprehensive set of results is that of Azatyan et aL5 determined using an e.s.r. method for the temperature range 283-485 K but there is insufficient information in their paper for us to recalculate accurately the appropriate values of n. However, we estimate n = 2.4k0.3 at the high end of their temperature range and n = 3.8 f 1 .O at the low end. Given these uncertainties the agreement with the presently obtained results, particularly at the low-temperature end, is entirely satisfactory, being within 0.1 log units of our line on an Arrhenius plot. The results of Azatyan et al.also imply curvature of the Arrhenius plot similar to that observed here. The results of Kazmi and Le Roys are well documented and permit us to recalculate the values of n. However, this study led to very scattered values of k, (ca. 0.1 log units at a given temperature), which may be symptomatic of problems with the very difficult experimental technique of these authors. We do not regard the fact that the recalculated values of k, fall 0.1-0.3 log units below our line as being at all significant. k,/cm3 molt1 s-l = 1.5 x lo8 using an assumed value n = 6. This value of k , is a factor of 6 lower than the comparable value of this study and is clearly anomalous even if a more appropriate value of n were used. The reason for the anomaly is not clear. The low-temperature results of Yang17 are independent of assumed stoichiometry but are dependent upon assumed values for the rate constants of H addition to olefins.As pointed out above, for both ethene and propene there is uncertainty in these values and this carries over into Yang's values for k,, which are a factor of 2 to 3 lower than the presently obtained values. We do not regard these discrepancies as significant. The study of Lede and Villermaux16 gave a single value of COMPARISON WITH OTHER ALKANES In our earlier work we have measured the rate constants for H attack on methane, log (k,/cm3 mol-1 s-l) = 13.88 -49.9 kJ mol-l/2.3 RT and on ethane2. obtaining, respectively, the results for the range 400-1800 K and log (k,/cm3 mol-1 s-l) = 14.12 - 39.2 kJ mol-l/2.3 RT for the temperature range 330-1350 K, the Arrhenius plots being straight lines within these ranges.The Arrhenius parameters obtained here for H + C3H8 are in line with these values. The pre-exponential factors are of very similar magnitude, as would be expected, and the activation energies diminish in parallel with the diminishing strength of the C-H bonds being attacked, remembering that it is mainly the secondary bond which is attacked in propane. We can carry the comparison with ethane a stage further by using the result of Campbell et all3 for k,i/k,, to split our value of k , for the higher end of our temperature range into its two components. In this way we obtain log (kln/cm3 mol-1 s-l) = 13.75 - 36.2 kJ mol-l/2.3 RT log (kli/cm3 mol-1 s-l) = 13.64- 27.8 kJ mol-l/2.3 RT.As would be expected, the parameters for k,, are very similar to those for k,. Moreover, for the middle of the presently used temperature range, 450 K, they evaluate to give log k values differing by only 0.02, a remarkable agreement well within any reasonable estimate of experimental error.R. M. MARSHALL, H. PURNELL AND A. SHEPPARD strengths of the primary and secondary C-H bonds in propane. 3009 The differential activation energies associated with k,, and kIi reflect the relative CURVATURE OF THE ARRHENIUS PLOT Curvature has been observed in the Arrhenius plots for the metathetical reactions of methyl and of ethyl with various alkanes.' It has also been suggested18 as characteristic of the similar reactions of H with methane and with ethane but, as described above, this suggestion is in error, at least for the temperature ranges 400-1800 and 330-1 350 K, respectively.We believe that our present observation of curvature for H + C,H, is not inconsistent with our previous conclusions since the curvature is only really apparent because we have obtained many results for temperatures < ca. 350 K, i.e. completely outside the range of validity of our conclusions for H + CH, and very nearly outside the range for H + C,H,. Also there are very few measurements for H + C,H, at these low temperatures. In their paper, Azatyan et aL5 suggested that curvature of the Arrhenius plot was due to competition between reactions (In) and (li). This cannot cause the observed curvature since kIi > k,, throughout the range of the present work whereas curvature would be observed when kli = k,, which, on the basis of the parameters presently obtained, does not occur until temperatures of ca. 4000 K. A. Sepehrad, R. M. Marshall and J. H. Purnell, J. Chem. Soc., Faraday Trans. I , 1979, 75, 835 and references therein. D. Jones, P. A. Morgan and J. H. Purnell, J. Chem. SOC., Faraday Trans. I , 1977,73, 131 1. R. R. Baker, R. R. Baldwin and R. W. Walker, Trans. Faraday Soc., 1970, 66, 2812 and references therein. V. V. Azatyan, S. B. Filippov and M. S. Khachatryan, Kinet. Katal., 1971, 12, 5 and references therein. H. A. Kami and D. J. LeRoy, Can. J. Chem., 1953,50, 14. H. Gg. Wagner and R. Zellner, Ber. Bunsenges. Phys. Chem., 1972, 76,440. C. W. Gear, Numerical Initial Value Problems in Ordinary Dzflerential Equations (Prentice Hall, Englewood Cliffs, N.J., 1971). * P. Camilleri, R. M. Marshall and J. H. Purnell, J. Chem. SOC., Faraday Trans. I , 1974, 70, 1434. ' M. J. Lexton, R. M. Marshall and J. H. Purnell, Proc. R. SOC. London, Ser. A, 1971,324,433. lo A. Sheppard, Ph.D. Thesis (University of Wales, 1983). l1 P. Camilleri, R. M. Marshall and J. H. Purnell, J. Chem. SOC., Faraday Trans. 1, 1975, 71, 1491. l2 J. M. Campbell, 0. P. Strausz and H. E. Gunning, Can. J. Chem., 1969, 47, 3759. l3 R. M. Marshall and J. H. Purnell, J. Chem. Soc., Chem. Commun., 1972, 764. l4 J. V. Michael, D. T. Osborne and G. N. Suess, J. Chem. Phys., 1973, 58, 2800. l5 V. V. Azatyan, Arm. Khim. Zh., 1967, 20, 577. J. Lede and J. Villerrnaux, Can. J. Chem., 1978, 56, 392. l7 K. Yang, J. Am. Chem. SOC., 1962,84, 3795. l 8 T. C. Clark and J. E. Dove, Can. J. Chem., 1973, 51, 2147. (PAPER 4/ 174) 98 FAR 1

ISSN:0300-9599    

DOI:10.1039/F19848002999

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1984,80, 3011-3019 Photo-oxidation of Tris(2,2’-bipyridine)ruthenium(11) by para-Substituted Benzene Diazonium Salts in Acetonitrile Two-compartment Photoelectrochemical Cell Applications BY HERMINIA CANO-YELO AND ALAIN DERONZIER* Laboratoires de Chimie, Laboratoire d’Electrochimie Organique et Analytique (LA C.N.R.S. NO32 l), Departement de Recherche Fondamentale, Centre d’Etudes Nucleaires de Grenoble, 85 X, F.38041 Grenoble Cedex, France Received 3 1st January, 1984 The photo-oxidation of tris(2,2’-bipyridine)ruthenium(11) [Ru(bpy):+] by substituted benzene diazonium salts (p-RC,H,Ni) has been studied in acetonitrile. Quenching of the excited state [Ru(bpy)i+*] byp-RC,H,Ni takes place at diffusional rates (k, 2 1.6 x 1Olo dm3 mol-1 s-l) and leads to the effective build-up of the oxidized species Ru(bpy):+ with a high quantum yield (# > 0.12).# depends on the nature of the substituent R. This Ru(bpy):+*/p-RC,H,Nl system has been applied in some two-compartment photoelectrochemical cells. The photogenerated Ru(bpy)g+ oxidizes an electron donor D in the dark compartment. Features of this cell are reported, especially the influence of D on the open-circuit photovoltage and the photocurrent passing through a 1000 0 resistor. The performance of this cell has been improved by the use of a second irradiation compartment where D is photogenerated as the reduced form of the 1,l ’-dimethyl-4,4-bipyridinium dication. The quantitative photogeneration of the strong oxidant1 Ru(bpy):+ and the strong reducing agent2 Ru(bpy),’ by electron-transfer quenching of the excited state of the tris(2,2’-bipyridine)ruthenium(11) complex [Ru(bpy)g+] is a rare reaction.The effective build-up of these two species is generally prevented by the electron-transfer back reaction which returns the system to the ground ~ t a t e . ~ In a previous study4 on the use of Ru(bpy)g+ as a photocatalyst for the Pschorr reaction, we noted that the 4-bromobenzene diazonium salt quenched Ru(bpy)t+, leading to the quantitative formation of Ru(bpy)g+. It has been shown recently5 that this type of diazonium salt can act as an electron-transfer quencher of triplet photosensitizers such as anthracene. As an extension of our previous we report here the quantitative photogeneration of Ru(bpy):+ in acetonitrile using, as the irreversible oxidative quencher, various para-substituted benzene diazonium salts (p-RC,H,N:).We also report the use of this system in some photoelectrochemical cells. EXPERIMENTAL Ru(bpy):+(BF;), was prepared from its chloride salt (Strem) by exchange with AgBF, and recrystallized from an acetonitrile + benzene mixture. The para-substituted benzene diazonium tetrafluoroborate salts were prepared and purified by a standard procedure.6 The electron donors tris(N,N-diethyldithiocarbamato)iron(~~r),~ N,N-bis(2,4,6-trimethoxyphenyl)methyla- mines and tris(p-bromophenyl)amineg were prepared according to the literature methods. The preparation and purification of phenothiazine, N-methylphenothiazine and 3,7-dithiocyano- phenothiazine have been described previously.1° Tetraethylammonium chloride (Eastman Kodak) and tetra-n-butylammonium bromide (Fluka) were used as supplied.Tetra-n-butylam- monium benzilate was synthesized as reported previously.ll Acetonitrile (Touzart et Matignon) 301 1 98-23012 PHOTO-OXIDATION OF Ru(bpy):+ BY p-RC6H4Ni and tetra-n-butylammonium perchlorate (Southwestern Analytical Chemicals) were purified as reported elsewhere.12 For the determination of quantum yields at 436 nm, the desired mercury emission line from the 250 W Hg lamp (Applied Photophysics type ME/D) was isolated by using band-pass filters (Oriel no. 5645). The intensity of light absorbed by the sample was measured by a calibrated calorimeter (Scientech 36-000 1) equipped with an X-t recorder. Quantum yields were deter- mined spectrophotometrically at the wavelength of the maximum absorption of Ru(bpy);+ (454 nm) by comparing the absorption of the sample before and after irradiation.All electronic spectra were recorded on a Beckman Acta IV spectrophotometer. Luminescence quenching measurements were made using a Jobin Y von JY 3C fluorescence spectrophotometer. In the presence of the diazonium salts, the luminescence intensity of Ru(bpy);+* began to decrease immediately after opening the excitation shutter. Consequently, the initial emission intensities were determined by extrapolating the intensity against time plots to the time of the shutter opening. Cyclic voltammograms were obtained using a P.A.R. model 173 potentiostat, a P.A.R. 175 universal programmer and a Sefram TGM 164 XY recorder.All potentials are reported us an aqueous saturated calomel electrode (SCE). H.p.1.c. analyses were performed with a Waters Ass. Instruments high-performance liquid chromatograph equipped with a printer, plotter, integrator and data module. The quantity of hydrocarbons was determined on a C18 radial pack Waters cartridge ( Z module) with the mixture CH,OH+H,O as eluant. Assays were made by the method of external standards. The photolysis apparatus and the cell design for the photoelectrochemical experiments were identical to those described previously.ll To prevent slow thermal decomposition of the diazonium salts during the longest irradiation (6-8 h), the cell was refrigerated at 0 f 2 "C by passing a stream of ethanol through a glass jacket sealed around the cell.The cooling of the solutions resulted in a 10% lowering of the photocurrent Il compared with the room-temperature experiments. A second identical photolysis apparatus was used for the double-irradiation- compartment cell experiments. All photoelectrochemical cell experiments were made with a stirring rate of 820+20 revolution min-l. All solutions were purged with dry argon and kept under an argon atmosphere during the experiments. RESULTS AND DISCUSSION PHOTOLYSIS OF Ru(bpy):+ IN THE PRESENCE OF p-RC6H4NB IN ACETONITRILE All the para-substituted benzene diazonium salts studied here quench the lumines- cence of the excited state Ru(bpy):+* in deaerated acetonitrile with diffusion-controlled rate constants evaluated from Stern-Volmer plots of the luminescence intensity quenching (table 1).The quenching process is a simple bimolecular reaction since the Stern-Volmer plots are linear. Continuous irradiation by visible light (A > 400 nm) of an oxygen-free dry acetonitrile solution containing ca. 1.5 x mol dm-3 Ru(bpy):+ and ca. 1.2 x mol dm+ p-RC,H4Ni gave the Ru(bpy):+ species quantitatively. Fig. 1 shows the spectral changes during photolysis with benzene diazonium (R = H). The last spectrum (e) is identical to the authentic spectrum of Ru(bpy):+ obtained by electrochemical oxidation of Ru(bpy)t+ at 1.5 V in acetonitrile. The quantum yields of the transformation at 436 4 nm are high and depend on the substituent R (see table 1). The highest quantum yields are obtained for the most easily reducible para-substituted benzene diazonium salts (R = Br and H).A permanent build-up of Ru(bpy):+ arises quenching reactions : hv RU(bPY )32+ + because of the following electron-transfer WbPY )32+ * (1) Ru(bpy)$+ +p-RC6H4N,. (2)H. CANO-YELO AND A. DERONZIER 301 3 Table 1. Results for the quenching of Ru(bpy),2+ by p-RC,H,N,f in CH,CN EJV us SCE, 0.1 mol dmP3 quenching rate in quantum 0.1 mol dm-, Bu,NClO, + CH,CN, hydrocarbon substituent, yields at Bu,NClO, + CH,CN, sweep rate 100 mV yields (% Ib R 436 nma kq/dm3 rno1-I s-l a s - ~ a H 0.33 & 0.01 2.1 kO.1 x 1010 CH, 0. I2 & 0.02 1.8k0.1 x 1O1O OCH, 0. I7 & 0.04 1.6k0.1 x 1 O 1 O Hr 0.34 rfr0.03 2.6 & 0.1 x 1O1O -0.16 > 65 - 0.23 > 80 -0.23 > 70 - 0.09 > 70 a At room temperature.At 0 Ifi 2 "C. 500 600 '0° h/nm Fig. 1. Spectral changes during irradiation with a 250 W Hg lamp at A > 400 nm of [Ru(bpy)i+] = 1.55 x mol dmP3 in CH,CN at room temperature: (a) t = 0 s, (b) t = 5 s, (c) t = 16 s, ( d ) t = 26 s and (e) t = 88 s. mol dm-, with [C,H,Nl] = 1.22 x The back-electron-transfer reaction between Ru(bpy)i+ and p-RC,H,NH is largely suppressed by the fast irreversible evolution of the p-RC,H,NH radical produced by the electron-transfer quenching. As will be demonstrated, the main final product of p-RC,H,Ni evolution is the corresponding hydrocarbon derivative RC,H, obtained witha high yield ( 2 65 % ). The formation of RC,H, can be explained by dediazotization of p-RC,H& leading to the corresponding aryl radical p-RC,H,N;+RC,Hi + N, (3)3014 PHOTO-OXIDATION OF RU(bpy)i+ BY p-RC,H,N; which abstracts hydrogen from acetonitrile CH,CN RC,& --+ RC,H,.(4) A similar has been proposed for the electrochemical reduction of benzene diazonium salts in acetonitrile, where formation of the corresponding hydrocarbon was observed with a good yield (> 50%). In our photochemical system the short-lived transient RC,H& cannot be characterized by the nitrone method since the resulting nitroxide derivative, being electr~activel~ at 0.7 V, is easily oxidized by Ru(bpy)i+ = 1.3 V)3 into a non-paramagnetic species. On another hand, although diazonium salts are soluble in water, no photo-oxidation of Ru(bpy);+ takes place in their presence in water. This difference in behaviour can be explained by the different reduction mechanism of the benzene diazonium salt in the two solvents.It has been shown that electrochemical reduction in water leads to the hydrazine C,H,NHNH,, which is easily 0~idized.l~ PHOTOELECTROCHEMICAL CELLS PRINCIPLE OF THE CELL The Ru(bpy):+/p-RC,H,N$ system in acetonitrile with 0.1 mol dmP3 Bu,NClO, as electrolyte can act as a two-compartment photoelectrochemical cell by its coupling with the oxidation of an electron donor D. The photogenerated strong oxidant Ru(bpy):+ in an illuminated compartment oxidizes D in the dark compartment through a system of two platinum electrodes connected by a 1000 l2 resistor. This type of photoelectrochemical cell is known to be an excellent practical device for visible-light conversion, and recently several examples have been reported,16 based on the irreversible oxidative quenching of Ru(bpy)i+* in water.The operation of our photoelectrochemical cells is illustrated in scheme 1. 0' + t I I I I r - P t electrodes-, \ PI* anodic I pho toca thodic I compartment cornpar tment Scheme 1. Continuous irradiation of the photocathodic compartment containing ca. 3 x lo-, mol dm-3 Ru(bpy):+ and ca. mol dm-3 p-RC,H,N: gives a rapid exponential rise of the photocurrent with time. At the same time the typical colour of the stable oxidized form D'+ of the electr< 'I donor appears in the dark compartment if one uses a reversible donor. The liniimg photocurrent Il is reached rapidlyH. CANO-YELO AND A. DERONZIER - . . . 301 5 4 0 0 . 4 *r- 200 1500 2000 l o o o t l s 500 (< 1 min) and it remains quite stable until almost complete consumption either of the diazonium salt [fig.2(a)] or of the donor D [fig. 2(b)]. In the later case another, lower, limiting photocurrent I,. is obtained if D O + is able to be oxidized. Fig. 2(b) shows this phenomenon where D is N-methylphenothiazine. It is well knownlo that the primary oxidized species D O + undergoes a further oxidation x 1.3 V) into the dication D2+ which reacts with residual water giving the sulphoxide. The photo-3016 PHOTO-OXIDATION OF Ru(bpy):+ BY p-RC,H,N; 600 LOO . 9 +- 200 0 0.5 EIVvsSCE ' Fig. 3. Plots of limiting photocurrents Il against experimental open-circuit photovoltage E for different electron donors D at T = 0 "C : (a) tris(p-bromophenyl)amine, (b) 3,7-dithiocyano- phenothiazine, (c) N-methylphenothiazine, ( d ) phenothiazine, (e) N,N-bis(2,4,6-trimethoxy- pheny1)methylamine and (f) tris(N,N-diethyldithiocarbamato)iron(III). Table 2.Data relating to the photoresponse of the single-irradiation compartment photoelectrochemical cell donor, Da I I1 I11 IV V VI Ecalculateci/Vb 0.27e 0.54d 0.64d 0.79d 0.93e 1.06f EexperirnentaJV 0.38 0.60 0.65 0.84 0.93 1.06 r/n 600 550 600 600 600 600 IJPA 200 370 390 500 584 640 a I, tris(pbromopheny1)amine; 11, 3,7-dithiocyanophenothiazine; 111, N-methylpheno- thiazine; IV, phenothiazine; V, N,N-bis(2,4,6-trimethoxyphenyl)methylamine; VI, tris(N,N- diethyldithiocarbamato)iron(IrI). From the difference between the El,2 of Ru(bpy):+/ Ru(bpy);+ [ref. (20)] and D/D'+ couples. Ref.(18). Ref. (10). Ref. (8). f Ref. (19). current Zl* is significantly lower than 1, because of the higher D'+/D2+ potential and the irreversibility of the process. In order to characterize the final products of the reduction of p-RC,H,N; we have analysed by h.p.1.c. the solution in the photocathodic compartment after irradiation until Il = 0, implying total consumption of the diazonium salt. The analysis shows that the main product of the transformation of p-RC,H,Nz is the corresponding hydrocarbon RC,H,. Table 1 gives the yields of RC,H,. Small amounts of the corresponding biphenyl derivatives are also detected (yields < 1 % ). Visible-spectral analysis of the photocathodic solution after irradiation for 1 or 2 h seems to indicate that the Ru(bpy):+ complex is not chemically modified after this time.Although Elofson et a1.l' have shown that pyridine can be phenylated with good yield by the electrolytic reduction of a benzene diazonium salt in acetonitrile, the very reactive RC,H; radical does not seem to react with the bipyridyl ligands of the complex. However, no definitive conclusion can be drawn since for longer irradiation (6-8 h), corresponding to the complete disappearance of the quencher, partial mixing of both compartments by diffusion does not allow us to follow by visible spectroscopy the possible evolution of the Ru(bpy):+ complex.H. CANO-YELO AND A. DERONZIER 3017 INFLUENCE OF THE DONOR As reported previouslyll the open-circuit photovoltage, E, of this type of photo- electrochemical cell, with a reversible electron donor D, depends directly on the choice of D.E is approximately equal to the difference between the El,, of the two electrochemically reversible couples Ru(bpy):+/Ru(bpy):+ and D/D'+. By applying Ohm's law, the limiting photocurrent Il is equal to E / ( R + r), where R is the external circuit resistance and r the internal cell resistance. The internal resistance depends mainly on the cell geometry and the conductivity of the solution and is independent of D. In this cell r was found to be 600 f 50 0. By using a constant external resistance (R = 1000 0) the variation of Il with E (and hence with D) is a straight line, as can be seen in fig. 3. All the data for these photoelectrochemical cells are summarized in table 2. The higher values of photovoltage and photocurrent are obtained with an tris(diethyldithiocarbamato)iron(m) complex as D.We have also verified that, for the same D, E and Zl are independent of the choice of p-RC,H,Nz. By using an irreversible donor D, this cell induces a net chemical reaction in the dark anodic compartment. Oxidation of anions such as C1-, Br- or benzilate [(C,H,),C(OH)CO;] can be achieved by this way. For instance, with the benzilate anion, at the end of the photoelectrochemical oxidation benzophenone was obtained in good yield (> 40%, as determined by h.p.1.c. analysis) with benzilic acid: -2e 2(C,H5),C(OH)CO; d ( C 6 H 5 ) , C O + (C,H,),C(OH)CO,H + CO,. (5) This result is identical to that obtained by direct electrochemical oxidation of benzilate.21 During the operation of the cell, a steady photocurrent of 340 pA was measured.TWO-IRRADIATION-COMPARTMENT PHOTOELECTROCHEMICAL CELL We have previously describedll a photoelectrochemical cell based on the visible irradiation of a charge-transfer complex formed between the paraquat dication (1,l '-dimethyl-4,4'-bipyridium dication, PQ2+) and the benzilate anion in acetonitrile containing 0.1 mol dm-3 in Bu,NClO,. A large photocurrent was obtained. We have shownll that the limiting photocurrent Il was reached more rapidly by addition of a small amount of Ru(bpy)i+ while the value of Il was not modified. We have coupled this (C,H5),C(OH>CO,/Ru(bpy)i+/PQ2+ system with the Ru(bpy)i+/ p-RC,H,N$ system to make a two-irradiation-compartment photoelectrochemical cell. Scheme 2 illustrates the operation of this double photoelectrochemical cell. h v Pt electrodes i-R-C6 H l / 2 ( c6 H 5 ) c ( OH) co; I P I A I I 112 H' photoanodic compart ment Scheme 2.compartment Nf Ru ( bpy):?Q p-c6 5 ) 2 OH ) pho toea t hod i c3018 PHOTO-OXIDATION OF Ru(bpy)g+ BY p-RC6H4N: O a 2 I 500 1000 1500 ti s Fig. 4. Photocurrent of the cell arising from irradiation at A > 400 nm at room temperature. Photoanodic compartment: [PQ2+] = 1.35 x lop3 mol dmp3, [(C,H,),C(OH)CO;] = 5 x mol dm-, and [Ru(bpy):+] = 5.9 x lop4 mol dm-3 in 0.1 mol dm-, Bu4NC104+ CH,CN. Photocathodic compartment: [p-CH,OC,H,Nl] = 1.18 x lop2 mol dm-, and [Ru(bpy):+] = 5.9 x mol dm-, in 0.1 mol dmp3 Bu4NC10, +CH,CN. A, First limiting photocurrent Zll ; B, second limiting photocurrent Z12.By irradiating only the photocathodic compartment, a steady photocurrent Ill is reached rapidly (1 min see fig. 4), comparable to that obtained with the single- irradiation photoelectrochemical cell with benzilate as D (vide supra). The photo- generated Ru(bpy):+ oxidizes the benzilate anion, according to reaction (9, into the non-irradiated compartment through the external circuit. The photovoltage of this cell increases continuously with time because of the irreversibility of the oxidation of D. Then, if the other compartment is irradiated, a new, significantly larger, steady photocurrent 112 is obtained (fig. 4). This double photoelectrochemical cell has a photocurrent (Ilz = 1.02 mA) and an open-circuit photovoltage ( E = 1.76 V) greater than those obtained from any single-irradiation-compartment photoelectrochemical cell reported above.Here, the photogenerated Ru(bpy):+ oxidizes the photogenerated PQ*+ species through the external circuit of the cell. The observed open-circuit photovoltage value is in good agreement with the calculated photovoltage obtained from the difference between the E1,2[Ru(bpy)g+/Ru(bpy)z+] and E1,2[PQ2+/PQ'+] couples (Ecalc. = 1.79 V). Note that the value of the internal cell resistance r remains the same when the second compartment is irradiated. This means that the improvement in the photopotential by using a two-irradiation-compartment device implies an improvement in the electrical power of the cell.H. CANO-YELO AND A. DERONZIER 3019 CONCLUSIONS In summary, it appears that para-substituted benzene diazonium salts are excellent irreversible oxidative quenchers of Ru(bpy)i+* in organic solvents, expecially in acetonitrile.This system has applications in some two-compartment photoelectro- chemical cells. The distinctive feature of our cell is that it was possible to have a two-irradiation-compartment photoelectrochemical cell by using two different photoactive systems. A substantial improvement in cell performance has been obtained using this new device. On another hand, the great ability of the Ru(bpy)i+*/ p-RC,H,Nl system to produce the very reactive phenyl radical RC,H; should, in principle, be interesting in organic synthesis. We thank Prof. G. Cauquis for his interest in this work and J-C. Lotito for technical assistance. (a) H. D. Gafney and A.W. Adamson, J. Am. Chem. SOC., 1972, 94, 8238; (b) G. L. Laurence and V. Balzani, Inorg. Chem., 1974, 13,2976; ( c ) F. Bolleta, A. Juris, M. Maestri and D. Sandrini, Inorg. Chim. Acta, 1980,44, L.175; ( d ) G. Giro, G. Casalbore and P. G. Di Marco, Chem. Phys. Lett., 1980, 71, 476. H. Cano-Yelo and A. Deronzier, Nouv. J . Chim., 1983, 7 , 147. For a recent review see K. Kalyanasundaram, Coord. Chem. Rev., 1982, 46, 159. H. Cano-Yelo and A. Deronzier, J. Chem. Soc., Perkin Trans. 2, 1984, 1093. J. C. Scaiano and N. Kim-Thuan, Can. J . Chem., 1982,644 2286. A. Roe, in Organic Reactions, ed. R. Adams (Wiley, New York, 1949), vol. 5, p. 203. A. W. White, R. Roper, E. Kokot, H. Waterman and R. G. Martin, Aust. J. Chem., 1964, 17, 294. D. Serve, Nouv. J. Chim., 1980, 4, 497.T. N. Baker, W. P. Doherty Jr, W. S. Kelley, W. Newmeyer, J. E. Rogers Jr, R. E. Spalding and R. J. Walter, J . Org. Chem., 1965, 30, 3714. A. Deronzier and F. Esposito, Nouv. J . Chim., 1983, 1, 15. (a) G. Cauquis, A. DeronZier, D. Serve and E. Vieil, J. Electroanal. Chem., 1975, 60, 205; (b) G. Cauquis and A. Deronzier, J . Inorg. Nucl. Chem., 1980, 42, 1447. lo G. Cauquis, A. Deronzier, J-L. Lepage and D. Serve, Bull. SOC. Chim. Fr., 1977, 3-4, 295. l 3 F. F. Gadallah and R. M. Elofson, J. Org. Chem., 1969, 34, 3335. l4 A. J. Bard, J. C. Gilbert and R. D. Goodin, J . Am. Chem. Soc., 1974, 96, 620. l5 0. Orange, C. Elfakir-Hamet and C. Caullet, J. Electrochem. SOC., 1981, 128, 1889 and references therein. (a) B. Durham and T. J. Meyer, J. Am. Chem. SOC., 1978,100,6286; (b) D. P. Rillema, W. J. Dressick and T. J. Meyer, J . Chem. SOC., Chem. Commun., 1980, 247; (c) M. Neumann-Spallart, K. Kalyana- sundaram, C. Gratzel and M. Gratzel, Helv. Chim. Acta, 1980, 63, 11 11 ; (d) M. Neumann-Spallart and K. Kalyanasundaram, Ber. Bunsenges. Phys. Chem., 198 1,85,704; (e) M. Neumann-Spallart and K. Kalyanasundaram, J . Chem. SOC., Chem. Commun., 1981,437; cf) W. J. Dressick, T. J. Meyer and B. Durham, Isr. J . Chem., 1982,22,153 ; (g) W. J. Dressick, T. J. Meyer and B. Durham, Inorg. Chem., 1982, 21, 3451. R. M. Elofson, F. F. Gadallah and K. F. Schulz, J. Org. Chem., 1971, 36, 1526 I B R. F. Nelson and R. N. Adams, J. Am. Chem. SOC., 1968, 90, 3925. l9 G. Cauquis and D. Lachenal, Inorg. Nucl. Chem. Lett., 1973, 9, 1095. 2o N. E. Tokel-Takvoryan, R. E. Hemingway and A. J. Bard, J . Am. Chem. SOC., 21 A. Deronzier and F. Esposito, J . Electroanal. Chem., 1983, 146, 207. 973, 95, 6582. (PAPER 41 175)

ISSN:0300-9599    

DOI:10.1039/F19848003011

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1984, 80, 3021-3036 Magnetic Resonance Studies of Cation Radicals from Chromans Part 2.-Nuclear Magnetic Resonance, Electron Spin Resonance, ENDOR? and TRIPLES Spectroscopy of Some Tricyclic Chromans BY IAN M. SMITH AND LESLIE H. SUTCLIFFE* Donnan Chemistry Laboratories, The University, Liverpool L69 3BX AND STEFAN WIESNER, WOLFGANG LUBITZ*§ AND HARRY KURRECK Institut fur Organische Chemie (WE02), Fachbereich Chemie, Freie Universitat, Takustrasse 3. 1000, Berlin 33, Federal Republic of Germany Received 6th February, 1984 N.m.r. spectroscopy has been used to study the electron-exchange reaction between the substrate-radical pairs of 2,3,4,7,8,9-hexahydro-2,2,5,7,7,lO-hexamethylbenzo[ 1,2-b: 4,5-b’]- dipyran, 2,3,4,5,6,7-hexahydro-2,2,7,7,9,lO-hexamethylbenzo[ 1,2-b : 4,3-b’]dipyran and 2,3,6,7- tetrahydro-2,2,4,6,6,8-hexamethylbenzo[ I ,2-b : 4,s-b’ldifuran and their cation radicals.The rate constants and energetics of activation of the reactions have been determined, along with absolute signs of the averaged proton hyperfine coupling constants. ENDOR and TRIPLE measurements provided accurate values of the individual cpupling constants and their relative signs: the n.m.r. and ENDOR data taken together gave a complete assignment. In Part 1 of this series1 we presented the results of a detailed e.s.r. investigation and of an ENDOR study of the cation radicals formed from the following three tricyclic chromans : 2,3,4,7,8,9-hexahydro-2,2,5,7,7,lO-hexamethylbenzo[ 1,2-b : 4,5-b’]dipyran [henceforth referred to as (l)] and its isomer, namely, 2,3,4,5,6,7-hexahydro- 2,2,7,7,9,10-hexamethylbenzo[ 1,2,-b: 4,3-b’]dipyran [henceforth referred to as (2)].The third chroman is 2,3,6,7-tetrahydro-2,2,4,6,6,8-hexamethylbenzo[ 1,241 : 4,5- b’ldifuran [henceforth referred to as (3)]. We have extended our measurements in several ways: (i) an n.m.r. study has enabled absolute signs of proton hyperfine coupling constants of the cation radicals from (1)-(3) to be determined; it has also provided a measure of the rates and energetics of electron exchange between the radicals and their substrates; (ii) a complete ENDOR 1- Electron-nucleus double resonance. 1 Electron-nucleus-nucleus triple resonance. Q Present address : Physics Department, University of California, San Diego, La Jolla, California 92093, U.S.A.302 13022 MAGNETIC RESONANCE STUDIES OF CHROMANS examination of the radicals has been carried out, supplemented with general TRIPLE experiments. We have shown earlier2 the value of this multiple approach; indeed, the additional data obtained from these measurements have led to important revisions in the assignments of the e.s.r. spectra, as there is no ambiguity in the n.m.r. assignments. EXPERIMENTAL MATERIALS Compound (1) was prepared from 2,Sdimethyl- 1 ,Cbenzoquinone (Eastman Chemical Co.) and 2-methylbut-3-en-2-01 (Aldrich Chemical Co.) using the method described by Frampton et a1.3 Compound (2) was prepared from 2,3-dimethylquinol and 2-methylbut-3-en-2-01 by the method of Al-Kha~at.~ Compound (3) was prepared as follows: 2,5-dimethylhydroquinone was formed in 80% yield by reducing 6 g of Eastman Kodak 2,5-dimethylbenzoquinone in 100 cm3 60% aqueous acetic acid with 7.5 g of zinc dust; to 5.0 g (0.036 mol of 2,5-dimethylhydroquinone and 5.2 g (0.072 mol) Aldrich 2-methyl-2-propen-1-01 in 200 cm3 formic acid were added two drops of concentrated sulphuric acid, and the mixture was stirred at 100 "C for 12 h and then poured on to crushed ice; this was extracted several times with diethyl ether and the extracts washed with water followed by aqueous sodium carbonate solution; the ether was evaporated off and the residue dissolved in a mixture of 5 cm3 concentrated hydrochloric acid and 100 cm3 methanol and refluxed for 15 min in order to hydrolyse the formate ester.After removal of the solvent, the residue was dissolved in a 1 : 9 v/v mixture of diethyl ether + 6&80" petroleum ether and refluxed for 10 min; this process dissolved the desired difuran and on cooling deposited any unreacted hydroquinone.The difuran was eluted from a silica-gel column with a 1 : 4 v/v mixture of diethyl ether + petroleum giving an overall yield of ca. 20%. The cation radical salts of compounds (l)-(3) were prepared by the addition of a 2: 1 molar ratio of a 10% solution of antimony pentachloride (Aldrich Chemical Co.) in dichloromethane to a solution of (l), (2) or (3) in di~hloromethane:~ the salts were precipitated with diethyl ether and were isolated as stable green, yellow-brown and green powders, respectively. Other methods of preparing the cation radicals are given in Part 1 of this series.' N.M. R. MEASUREMENTS All n.m.r. spectra were recorded with a Bruker WM250 spectrometer and its associated Aspect 2000 computer. Proton spectra of the compounds were acquired using 16k data points and Fourier-transformed into 32k data points. Any artificial line broadening caused by ' zero filling' or by small changes in magnetic field inhomogeneity were corrected for by measuring linewidths relative to internal chloroform. The variable-temperature unit of the spectrometer had been pre-calibrated and found to be accurate to & 1 "C over the temperature range used. MEASUREMENT OF LINEWIDTHS AND CONTACT SHIFTS To a known concentration of (l), (2) or (3) in 1 cm3 of chloroform was added progressively known small volumes of the appropriate cation salt in chloroform.Proton spectra were recorded for each addition: Fig. 1, 4 and 8 show the spectra in the presence and absence of cation radicals. After Fourier transformation, each spectral line was plotted on an expanded scale, the linewidth was measured and the contact shift relative to the chloroform reference obtained. An external reference was also used in order to monitor the shift of the internal reference line caused by the radical.6 The external reference comprised a 1 mm 0.d. capillary tube into which chloroform was sealed; the capillary was held centrally in the standard 5 mm 0.d. n.m.r. sample tube with PTFE spacers. The disposition of the spacers relative to the receiver coils was adjusted until optimum resolution was achieved.The paramagnetic shift of the internal solvent line was used both as a measure of the radical concentration and to check that the radical concentration remained constant during a run.I. M. SMITH et d. 3023 DETERMINATION OF HYPERFINE COUPLING CONSTANTS AND ELECTRON-EXCHANGE RATES These were determined from the contact shifts and linewidths using the theory of De Boer and MacLean:' B o { ( l +[fD z$(aH/2)']/(1 +2zpT;,')}-' 4k T 6, = where 6, is the proton contact shift (in Hz), q , l , , is the electron exchange contribution to linewidth (in Hz), fp and f D are the fraction of paramagnetic and diamagnetic species, respectively, z p is the lifetime of paramagnetic species (in s), qe is the electron spin-lattice relaxation time and all other longitudinal relaxation effects (in s), aH is the proton hyperfine coupling constant (in Hz), ge is the free-electron g-factor, pe is the Bohr magneton (in J T-l), B, is the external applied field (in T), k is Boltmann's constant (in J K-l) and T is the temperature (in K).De Boer and MacLean identified two limiting expressions for fast exchange and for slow exchange. cfp ~ $ ( a ~ / 2 ) ~ + 22, Tcl) -g 1. (i) Fast exchange, i.e. Eqn (1) then reduces to and eqn (2) reduces to 6 c =- aHfp ge P e B,/4kT AT,'ex = f p 2P(aH/2)'- This expression is useful when aH > 0.6 mT. (ii) Slow exchange, i.e. f D z;(aH/2)' B (1 + 22, T G ~ ) . The exchange contribution to the linewidth is AT<\, = T ~ I , which means that all lines broaden identically. The rate constant (k,,) for electron exchange between the diamagnetic and paramagnetic species is given by k,, = (rp[D])-l or ke, = (zD[P])-' (3) where [PI and [D] are the concentrations of the paramagnetic and diamagnetic species.For the slow-exchange limit, the line broadening is directly proportional to the rate: In order to apply eqn (1) and (2) quantitatively it is necessary to determine q,, which is estimateda to be 3 x lop6 s, in accord with the few values obtained experimentally. 9 9 l o For this value of q,, the slow-exchange limit only applies for k,, > lo5 dm3 mo1-I s - l . Our data do not conform to this limit because the lines in our proton n.m.r. spectra broaden to different extents. For each temperature there is only one value of zp, but each line in the n.m.r. spectrum corresponds to a proton having different e.s.r.hyperfine coupling constant, aH. Values of aH and z p were found by putting trial values of them into eqn (1) and ( 2 ) until a self-consistent set of data was obtained which matched the linewidths and contact shifts : the accuracy of coupling constants is poor when their magnitude is > ca. 0.5 mT. The direction of the contact shift gives the absolute sign of aH. The rate constant of electron exchange for each temperature was calculated from eqn (3) and the energetics from the Eyring equation: KkT h kex = - exp (- AG*/RT) where the symbols have their usual meanings. E.S.R . MEASUREMENTS E.s.r. spectra were recorded with a Varian E4 X-band spectrometer. The procedures used and sample preparations are described in Part 1 .l It was particularly convenient to prepare radical solutions from the hexachloroantimonate salts of the cation radicals.3024 MAGNETIC RESONANCE STUDIES OF CHROMANS ENDOR-TRIPLE MEASUREMENTS The spectrometer used for ENDOR and TRIPLE spectroscopy consisted of a Bruker ER 220 D e.s.r.spectrometer equipped with a Bruker cavity (ER 200 ENB) and home-built n.m.r. faci1ities.l' ENDOR spectra were accumulated by using a Nicolet averager 1 170 employing 1 k data points; typically 32 or 64 sweeps were taken, with 30 s per scan and a time constant of 40 ms. The temperature (& 1 K) was varied over the range 180-310 K with a Bruker B-VT temperature-control unit and checked by means of a thermocouple. A microwave power of 200 mW was used and the proton n.m.r.frequency of 14.4 MHz was employed at a power of ca. 0.4 mT in the rotating frame: the 10 kHz FM of the n.m.r. field had an amplitude of 25-40 kHz. DETERMINATION OF HYPERFINE COUPLING CONSTANTS FROM ENDOR SPECTRA According to the ENDOR condition1'? l2 where v, is the n.m.r. resonant frequency for a given nucleus, each set of equivalent nuclei in a radial in isotropic solution gives rise to only one pair of lines. In the case of protons where vH > laH/21 is valid (as for the chroman radicals here) the lines are displayed symmetrically about the free-proton frequency vH (ca. 14 MHz for X-band e.s.r. magnetic fields) and are separated by the isotropic hyperfine coupling constant a. The latter can be obtained directly from the spectra with high precision ( f 10 kHz).Although ENDOR reaches only a few percent of the e.s.r. sensitivity, the spectral resolution is considerably higher, depending on the complexity of the radica1.l' The saturated linewidths for the protons in our spectra are typically ca. 100 kHz. It has to be borne in mind that ENDOR line intensities do not reflect directly the relative numbers of nuclei contributing to the lines: the intensities are a consequence of saturation and the individual relaxation behaviour of the nuc1ei.l2 We have performed special TRIPLE13 experiments which provide a better guide to numbers of nuclei contributing to a line. We have simulated successfully the e.s.r. spectra using the accurate data provided by ENDOR. DETERMINATION OF RELATIVE SIGNS OF HYPERFINE COUPLING CONSTANTS FROM GENERAL TRIPLE Pumping one transition in the ENDOR spectrum with a second saturating radiofrequency field gives a general TRIPLE which shows characteristic variations of line intensities.From the changes in high- and low-frequency line amplitudes with respect to the 'pumped line', the relative signs of the hyperfine coupling constants can be deduced. The theory and applications of this technique have been described recently in detail.". l2 RESULTS CATION RADICAL OF (1) N.M.R. MEASUREMENTS Fig. 1 shows the hydrogen n.m.r. spectrum of (1) and includes the unambiguous assignment of the bands. Addition of the cation radical hexachloroantimonate caused the lines to broaden and shift as may be seen in fig. 1 : data for the variation with radical concentration of half-height linewidth and contact shift are listed in table 1.Note that linewidths for the triplet arising from the methylene group at positions 3 and 8 were measured when the radical concentration was sufficiently high to convert the triplet into a single broad line. The triplet from the benzylic protons at positions 4 and 9 broaden too much for reliable linewidths or contact shifts to be measured with precision. For the same reason, contact shifts were also difficult to measure for the aromatic methyl groups at positions 5 and 10; however, it was established thatI. M. SMITH et d. 3025 1 I 1 2.5 3 3.5 (PPm) Fig. 1. Hydrogen n.m.r. spectra at 250.133 MHz of (1) in deuterochloroform at 296 K in the presence of fractions up) of its cation radical: (a) 0.00, (b) 3.22 x (c) 6 .4 4 ~ ( d ) 10.7 x and (e) 16.1 x lo-*. the contact shifts are negative (positive hyperfine coupling constants) for these and the benzylic protons. From the data of the type given in table 1 we calculated the motionally averaged hyperfine coupling constants for all the protons in the radical and obtained their absolute signs: positions 5 and 10 positions 4 and 9 positions 3 and 8 positions 2 and 7 + 0.30 k 0.05 rnT +0.13+0.01 mT - 0.01 7 k 0.002 mT - 0.002 k 0.00 1 mT. We also calculated the rate constant for electron exchange for various temperatures3026 MAGNETIC RESONANCE STUDIES OF CHROMANS Table 1. Dependence of half-height linewidths and contact shifts at 296 K of compound (1) on the concentration of its cation radical contact shift/Hz linewidth increase/Hz radical fraction of CH, at CH, at CH, at CH, at CH, at CH, at concentration radical positions positions positions positions positions positions /10-5moldm-3 (104fp) 3 and 8 2 and 7 5 and 10 4 and 9 3 and 8 2 and 7 4.35 8.70 14.50 21.75 29.00 36.25 43.50 58.00 72.50 87.00 101.50 116.00 3.22 6.44 10.74 16.1 1 21.48 26.85 32.22 42.96 53.70 64.44 75.19 85.93 0.2 f 0.2 0.9 f 0.2 2.0 f 0.4 3.0 f 0.5 3.5 f 0.7 5.2 & 0.7 - Ok0.2 33f2 35&2 1 1 + 1 O.Ok0.2 Ok0.2 64+3 73+3 16+1 0.1k0.2 Of0.2 - - 22*1 0.1 f0.2 Ok0.2 - - 33k2 0.2k0.2 Of0.2 - - 43k2 0.3k0.2 Of0.2 - - 48k3 0.4f0.2 0.2k0.2 - - - 0.5k0.2 0.5k0.2 - - - 1.0k0.2 1.4 f 0.2 0.7k0.2 - - - 0.8f0.2 - - - 1.5 k0.2 l.Ok0.2 - - - 1.9 & 0.2 1.5k0.2 - - - 2.4 f 0.2 Table 2.Rate constants for electron exchange between compound (1) and its cation radical 240 1.55 0.07 260 1.76 f 0.13 280 2.23 f 0.03 296 2.60 k 0.04 (see table 2) and obtained the following energetic parameters: AG,$, = f 36.4 0.1 kJ mo1-I AH&6 = + 3.1 f 0.4 kJ mol-l As,$, = - 1 I I f 10 J mo1-l K-l.E.S.R. AND ENDOR MEASUREMENTS In Part l1 we reported a detailed analysis of the e.s.r. spectrum along with kinetic data for the rate of intramolecular motion. However, we had not determined the relative signs of the hyperfine coupling constants; hence it was necessary to carry out general TRIPLE experiments. The n.m.r. data given above provide motionally averaged hyperfine coupling constants for all the axial and equatorial positions in the radical, apart from the 5 and 10 methyls, which are only slightly affected by out-of-plane bending, unlike the other protons.In comparison with the e.s.r. values, the magnitudes suggest that some of the slow-limit values could be negative. Fig. 2 shows two superimposed TRIPLE spectra of the radical from (1): the dotted and full lines are general TRIPLE experiments with pumping being applied in turn at the low- and high-field line pair corresponding to the largest coupling in the spectrum. Fig. 3 isI. M. SMITH et al. 3027 Fig. 2. Two superimposed general TRIPLE spectra of the hexachloroantimonate salt of the cation radical of (1) in CH,Cl, at 180 K; 100 scans for each spectrum. The full lines and dotted lines are for pumping at 18.65 and 10.32 MHz, respectively. Mi Fig. 3. Two superimposed general TRIPLE spectra of the hexachloroantimonate salt of the cation radical of (1) in CH,Cl, at 180 K; 10 scans for each spectrum.The full lines and dotted lines are for pumping at 18.65 and 10.32 MHz, respectively.3028 MAGNETIC RESONANCE STUDIES OF CHROMANS a narrow sweep of the centre of this spectrum, obtained under the same pumping conditions as used for fig. 2: it shows clearly that the two intense lines belong to two very similar but distinguishable coupling constants which have opposite signs since very little TRIPLE effect occurs. The coupling constants and their relative signs extracted from the spectra depicted in fig. 2 and 3 are as follows: - +0.2967, kO.2369, kO.0350, JO.0343, T0.0134 and T0.00418 mT. Combining the ENDOR results with those from n.m.r.we can make the following assignments of the hyperfine coupling constants (in mT) along with their absolute signs : CH,, positions 5 and 10 CH axial, positions 4 and 9 CH equatorial, positions 4 and 9 0.2967 0.2369 0.0350 CH axial, positions 3 and 8 CH equatorial, positions 3 and 8 CH, axial, positions 2 and 7 - 0.0343 -0.0134 - 0.004 1 8 CH, equatorial, positions 2 and 7 ca. 0. The average couplings for positions 4 and 9, 3 and 8 and 2 and 7 (0.136, - 0.024 and -0.002 mT, respectively) are in reasonable agreement with the n.m.r. data. E.s.r. spectra from the slow to the fast exchange rates of interconversion can be simulated from the ENDOR values and assignments listed above. This more detailed assignment does not affect the kinetic and energetic data obtained for the out-of-plane intra- molecular motion of the radical given previously. CATION RADICAL OF (2) N.M.R.MEASUREMENTS The hydrogen n.m.r. spectrum of compound (2) is shown in fig. 4, where it can be seen to be of a similar type to that of (1); this is to be expected since they both contain the same spin systems: the effect of adding the cognate radical cation is also shown. The data obtained for linewidths and contact shifts by adding the cation radical salt of (2) are given in table 3. From this type of data we derived the motionally averaged hyperfine coupling constants (in mT) and their absolute signs for all the protons in the radical: positions 9 and 10 positions 4 and 5 0.17 k 0.02 0.35 k 0.05 positions 3 and 6 -0.012_+0.002 positions 2 and 7 - 0.003 1 f 0.0002. Again the rate constant for electron exchange was determined at several temperatures (see table 4) giving the following energetic data: AG2*96 = 33.3k0.3 kJ mol-1 AH:g6 = 15.2 0.6 kJ mo1-l As,*,, = - 60 k 2 J mol-' K-l.E.S.R. AND ENDOR MEASUREMENTS In our previous report1 of this radical no ENDOR results were available; this deficiency has now been made good and has led to important revisions of the e.s.r.I. M. SMITH et al. 3029 B 1 . 5 I 2 2.5 (PPm) Fig. 4. Hydrogen n.m.r. spectra at 250.133 MHz of: (2) in deuterochloroform at 300 K in the presence of fractions (f,) of its cation radical: (a) 0.00, (b) 0.91 x (e) 14.0 x lop4 and (f) 34.0 x lop4. (c) 3 . 2 0 ~ ( d ) 6.85 x Table 3. Dependence of half-height linewidths and contact shifts at 300 K of compound (2) on the concentration of its cation radical contact shift / Hz linewidth increase/Hz _______ radical fraction of CH, at CH, at CH, at CH, at CH, at CH, at concentration radical positions positions positions positions positions positions mol dm-3 (lo4&) 3 and 6 2 and 7 4 and 5 9 and 10 3 and 6 2 and 7 0.62 0.95 1.56 2.17 3.12 4.64 6.17 9.49 15.60 23.05 30.85 0.9 1 1.40 2.30 3.20 4.60 6.85 9.10 14.00 23.00 34.00 45.50 - O .l t 0 . 2 O.Ok0.2 2 3 f 1 19+1 0.0 f 0.2 0.3k0.2 O.Ok0.2 35k2 27+1 - 0.0 f 0.2 0.4k0.2 0.1k0.2 5 0 f 3 37k2 - 0.0 f 0.2 0.3Ifr0.2 0.11fr0.2 6 4 f 4 4 6 f 3 - 0.1 f0.2 0.3f0.2 0.2k0.2 - - 9.8k0.3 0.1 f0.2 1.7f0.5 0.2k0.2 - - 14.3k0.6 0.1 k0.2 1.8k0.7 0.2k0.2 - 17.0 k 0.6 0.1 & 0.2 2.0k 1.0 0.4k0.2 - 25+ 1 0.2k0.2 7.2k1.5 0.8f0.2 - - 39+2 0.4f0.2 0.9Tfr0.2 - - - 0.5k0.2 1.2k0.2 - - - - 1.1k0.2 - - -3030 MAGNETIC RESONANCE STUDIES OF CHROMANS Table 4.Rate constants for electron exchange between compound (2) and its cation radical T / K k,,/108 dm3 mo1-I s-l ~~ ~ ~~ 240 1.73 & 0.02 260 2.75 f 0.08 280 6.16 f0.18 300 9.71 f0.28 ;; f Fig. 5. Two superimposed general TRIPLE spectra at 190 K of the cation radical from (2) with CF,COOD in CH,Cl,; 100 scans. The full lines and dotted lines are for pumping at 16.25 and 12.12 'MHz, respectively. assignments. Fig. 5 shows the complete ENDOR spectrum in the form of two superimposed general TRIPLE experiments. The central portion of fig. 5 was examined more closely as shown in fig. 6. The magnitudes and relative signs of the hyperfine coupling constants are: - + 0.4987, & 0.1532, f 0.0365, T 0.01 80 and T 0.0034 mT.In our original analysis1 we extracted a coupling constant of 0.623 mT from the e.s.r. spectrum near room temperature. The absence of this value from the ENDOR data led us to reconsider the e.s.r. analysis. Consequently, fig. 6 in ref. (1) cannot be the conformational slow-limit e.s.r. spectrum. By taking a solution of the hexachloro- antimonate salt of the cation radical of (2) we were able to obtain an e.s.r. spectrum at 180.8 K (fig. 7) that can be regarded as the true slow-limit spectrum. The variation of the e.s.r. spectrum over the temperature range 181-348 K can best be explained from the assignment a,'tlo = 0.1532, = 0.4987, a:,, = 0.1532, a?&, = 0.0365 andI. M.SMITH et d. 303 I MHz Fig. 6. Two superimposed general TRIPLE spectra at 190 K of the cation radical from (2) with CF,COOD in CH,Cl,; 10 scans. The full lines and dotted lines are for pumping at 16.25 and 12.12 MHz, respectively. 0.4 mT - Fig. 7. First-derivative X-band e.s.r. spectrum of the hexachloroantimonate salt of the cation of (2) in CH,Cl, at 180.8 K.3032 MAGNETIC RESONANCE STUDIES OF CHROMANS a&e = 0.0180 mT: the spectra change with the rate at which the respective axial and equatorial hydrogens interchange. The simulations match only when it is assumed that the coupling constants for the 4 and 5 axial and equatorial hydrogens have the same sign and also those for the 3 and 6 axial and equatorial hydrogens have the same sign. At the fast limit the main lines of the spectrum comprise a 1 : 4: 6: 4: 1 quintet having a spacing of 0.3260 mT.The fast- and slow-limit spectra are rather similar, apart from a loss of resolution in the former. By considering together the e.s.r., n.m.r. and ENDOR data the following assign- ments can be made: CH,, positions 9 and 10 CH axial, positions 4 and 5 CH equatorial, positions 4 and 5 0.1532 0.4987 0.1532 CH axial, positions 3 and 6 CH equatorial, positions 3 and 6 CH, axial, positions 2 and 7 - 0.0365 -0.0180 - 0.0034 CH, equatorial, positions 2 and 7 ca. 0. This new assignment caused us to revise the rate of chair-to-chair inversion required to simulate the alternating linewidth spectrum. The rate constant is ca. lo7 s-l at room temperatuE, this value is similar to that found for the radical from (1) at this temperature.Also, the ratio of the 4 axial to the 4 equatorial coupling constants is 3.3, which indicates15 that C-He bond is ca. 3" from a right angle to the plane of the aromatic ring for the cation radicals from both (1) and (2): note that the angle for (1) was given wrongly in Part 1. The result tells us that the ring structures of the radicals have similar geometries. Because of their small magnitudes the separate axial and equatorial coupling constants for positions 3 and 6 become averaged at relatively low temperatures. CATION RADICAL OF (3) N.M.R. MEASUREMENTS Fig. 8 shows the n.m.r. spectrum of (3) along with spectra obtained in the presence of various fractions (f,) of its cation radical: there is no doubt about the spectral assignments in these simple spectra. The linewidth and contact-shift data are listed in table 5 : the motionally averaged hyperfine coupling constants for the cation radical of (3) and their absolute signs are as follows (in mT): positions 3 and 7 0.6 + 0.1 0.1 + 0.01 positions 4 and 8 positions 2 and 6 - 0.007 0.002.The rate constants for electron exchange were measured at several temperatures (table 6 ) giving the following energetic data: AG&* = 31.6k0.3 kJ mol-l AH&,, = 26.5 f 2.0 kJ mol-1 AS&s = - 17+7 J mol-1 K-l. E.S.R. AND ENDOR MEASUREMENTS Fig. 9 shows two superimposed general TRIPLE spectra of the cation radical of (3). Since there is a suggestion of a splitting of the centre line a smaller scan was usedI. M. SMITH et al.3033 I 1 I I I 3 2.5 2 1.5 (PPm) Fig. 8. Hydrogen n.m.r. spectra at 250.133 MHz of compound (3) in deuterochloroform at 300 K in the presence of fractions (f,) of its cation radical: (a) 0.00, (b) 1.13 x lop4, (r\ 2 26 x lop4. tdl 3.39 x lW4 and ( ~ l 4.52 x lW4. 'Table 5. Dependence of half-height linewidths and contact shifts at 300 K of compound (3) on the concentration of its cation radical contact shifts/Hz linewidth increase/Hz ~~ radical fraction of CH, at CH, at CH, at CH, at CH, at CH, at concentration radical positions positions positions positions positions positions /lop5 mol dmp3 (104fp) 4 and 8 3 and 7 2and 6 4and 8 3 and 7 2and 6 ~~~ ~ 0.57 1.13 -0.6f0.3 -1.1 f0.6 0.08f0.20 15+ 1 40+ 1 0.3f0.2 1.15 2.26 - 1.2f0.6 -2.0+ 1.0 0.30f0.20 38f2 69f2 0.2f0.2 1.72 3.39 -1.3k1.0 -1.8f1.0 0.60f0.20 60+3 8 1 f 4 0.7f0.2 2.30 4.52 -5.1k1.0 -1.8kl.O 0.90f0.20 87-63 148-66 0.7f0.2 as depicted in fig. 10.These spectra, the n.m.r. spectra and the e.s.r. spectra reported mrlim-1 vipldpd the fnllnwinu hvnprfinp rniinlino rnnctnntc fin mT\ 2nd thpir nhcnliitp 0.1374 0.61 82 signs : CH,, positions 4 and 8 CH,, positions 3 and 7 CH,, positions 2 and 6 - 0.0041 5.3034 MAGNETIC RESONANCE STUDIES OF CHROMANS Table 6. Rate constants for electron exchange between compound (3) and its cation radical 240 (9.9k 1.2) x 105 256 (2.54k0.28) x lo6 270 (5.65 f 0.69) x lo6 300 (1.86 k 0.07) x lo7 4.4 Y I Fig. 9. Two superimposed general TRIPLE spectra at 233 K of the cation radical prepared from (3) with 1 : 4 v/v C1,CHCOOD + CH, Cl,; 64 scans.The full lines and dotted lines are for pumping at 22.79 and 5.42 MHz, respectively. Although e.s.r. spectroscopy is intrinsically incapable of resolving the small splitting of 0.00415 mT, the broader lines of the e.s.r. spectrum of the radical from (3) can be accounted for by the fact that it has twelve such protons whereas the other two radicals have only six: this conclusion is consistent with the fact that the furan rings are planar on the e.s.r. timescale. DISCUSSION PROTON HYPERFINE COUPLING CONSTANTS The cation radicals of all three compounds have relatively large magnitudes for the benzylic proton coupling constants; as may be seen from fig. 2, 5 and 9, these protons give rise to intense ENDOR signals: this phenomenon may be a useful diagnostic aid to assignment.The radicals from (1) and (2) are geometric isomers and can be regarded as hexasubstituted benzene cation radicals having the unpaired electron in an antisym-I. M. SMITH et a l . A 3035 , I , I , I 0 1 ! I * I j 17.15 MHz , I I 1 , I * I ' I ! I I ' I 0 , , . Fig. 10. Two superimposed general TRIPLE spectra at 233 K of the cation radical from (3) with 1 : 4 v/v C1, CHCOOD : CH, Cl,; 64 scans. The full lines and dotted lines are for pumping at 22.79 and 5.42 MHz, respectively. metric molecular orbital.16 The magnitudes of the proton hyperfine couplings of the aromatic methyl and the benzylic protons show that virtually all the spin density resides at the corresponding a carbon atoms in the benzene ring. Consequently, a check can be made of our assignments from the following simple relationships : Cation from (1) : Cation from (2): 'aH = 6aFH3(9, 10) +2agH2(4, ~ ) ~ X + ~ ~ F H Z ( ~ , 5)eq H + 2aFHz(3, 6)ax + 2aCHz(3, 6)eq = 2* mT.The close agreement between the two totals confirms the validity of our assignments. The radical from (3) is planar and the saturated rings are five-membered; hence a further comparison is not possible. RATES AND ENERGETICS OF ELECTRON TRANSFER The rates of electron exchange between the three substrates and their cation radicals are similar as are the free energies of activation. The much smaller entropy of activation for the benzodifuran (3) system is probably a consequence of the planarity of this compound and its cation radical.3036 MAGNETIC RESONANCE STUDIES OF CHROMANS CONCLUSIONS We have shown that it is necessary to employ three major magnetic-resonance techniques in order to obtain a full knowledge of the hyperfine structure of the chroman cation radicals.Accurate values of the hyperfine coupling constants and their relative signs have been deduced from high-resolution ENDOR and TRIPLE spectroscopy of the solutions. Multiplicities were obtained using these coupling constants by computer simulation of the partially resolved e.s.r. spectra. The latter also proved to be valuable in obtaining rates and energetics of the out-of-plane motions of the cation radicals from the two isomeric benzodipyrans. Although the hyperfine values are motionally averaged and not particularly accurate from n.m.r. linewidth measurements, this method provides an excellent tool for the assignment of hyperfine couplings to specific protons in the molecule.Also the method gives absolute signs of couplings from the contact shifts and allows rates and energetics of electron exchange to be measured. We are convinced that the combination of these magnetic-resonance techniques will be very useful for unravelling the hyperfine structure of complex radicals in future investigations. We thank the University of Liverpool for a maintenance grant for I. M. S. and the S.E.R.C. for a travel grant to L.H.S. We are grateful to Miss D. L. Ball and Mr S. D. Johnson for experimental assistance. We are also indebted to Dr W. R. McIlwaine for considerable help with computing facilities. H. K. thanks the Fonds der Chemischen Industrie and the Deutsche Forschungsgemieinschaft for financial support. S. A. Fairhurst, L. H. Sutcliffe and S. M. Taylor, J. Chem. SOC., Faraday Trans. I , 1982, 78, 2743. S. A. Fairhurst, I. M. Smith, L. H. Sutcliffe and S. M. Taylor, Org. Magn. Reson., 1982, 18, 231. V. L. Frampton, W. A. Skinner, P. Cambour and P. S. Bailey, J. Am. Chem. SOC., 1960, 82,4632. F. A. Bell, A. Ledwith and D. C. Sherrington, J . Chem. Soc. C, 1969, 2719. M. L. Martin, J. J. Delpuech and G. J. Martin, Practical NMR Spectroscopy (Heyden, London, 1980), p. 177. * I. Al-Khayat, Ph.D. Thesis (Liverpool University, 1980). ' E. De Boer and C. MacLean, J. Chem. Phys., 1966, 44, 1334. * N. M. Atherton, Electron Spin Resonance (Halsted Press, London, 1973), p. 266. J. W. H. Schneurs, G. E. Blomgren and G. K. Fraenkel, J. Chem. Phys., 1960, 32, 1861. lo M. P. Eastman, R. G. Kooser, M. R. Das and J. H. Freed, J. Chem. Phys., 1969, 51, 2690. l 1 H. Kurreck, B. Kirste and W. Lubitz, Angew. Chem., Int. Ed., Engl., in press. l2 K. Mobius, M. Plato and W. Lubitz, Phys. Rep., 1982, 87, 171. l3 K. P. Dinse, R. Biehl and K. Mobius, J. Chem. Phys., 1974, 61, 4335. l4 R. Biehl, M. Plato and K. Mobius, J. Chem. Phys., 1975, 63, 3515. l5 J. R. Morton, J. Chem. Phys., 1964, 41, 2956. l6 G. Vincow, in Radical Zons, ed. E. T. Kaiser and L. Kevan (Interscience, New York, 1968). (PAPER 4/202)

ISSN:0300-9599    

DOI:10.1039/F19848003021

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. I , 1984, 80, 3037-3040 Radiotracer Study of Alkane Homologation over Platinum Black BY ZOLTAN PAAL,* MARIA DOBROVOLSZKY AND PAL TEYENYI Institute of Isotopes of the Hungarian Academy of Sciences, Budapest, P.O.Box 77, H-1525 Hungary Received 6th February, 1984 Homologation over platinum black has been studied by a radiotracer method. The results are in agreement with the mechanism involving carbene addition to open-chain surface species. Radioactivity values prove that carbenes originate from deep fragmentation of the parent alkane. The side methyl group of 3-methylpentane participates to a lesser extent in carbene formation. Benzene alkylation (if it occurs at all) is only of minor importance in toluene formation. As distinct from the well known hydrogenolysis (degradation) reaction, homolog- ation, i.e.a building up of hydrocarbons over metallic catalysts, has been studied only recently. Preliminary studies reported propane formation from ethane over W, Rh and Pt films’ and toluene formation from n-hexane over Ni/Al,03.2 Guczi et al.3 claimed that toluene was produced from n-hexane via benzene alkylation, the cracking of benzene being responsible for creating alkylating fragments. No toluene formation was reported over Pt.3 Following Peter and .Clarke,4 O’Donohoe et u I . ~ studied systematically homologation over several metal films. Both these groups of authors and Sarkany et ai.6 found little activity for Pt films or Pt black. Since the reaction was attributed to carbene addition to surface ~lefins,~ the low activity of Pt was associated with poor carbene formation over this metal.The addition of methylene and an alk-1-ene occurs predominantly at the C, p o ~ i t i o n . ~ t ~ The main product of chain growth for pentane is benzene, since homologation is favoured by hydrogen- deficient condition^,^^^ which also favour alkane ar~matization.~ Margitfalvi et a1.* proposed another pathway : the oligomerization of C6 units over Pt/Al,03 followed by degradation to various aromatics (including toluene), and to C,, C,, etc. species. However, the bifunctional character of their catalyst might play a role in this type of build-up processes. The present work reports details of radiotracer experiments carried out with a Pt-black catalyst. Such studies represent unique possibilities for catalytic studiesg The experiments were carried out in a pulse-microcatalytic apparatus.1° Labelled compounds: [14C,]-n-hexane and [3-14C]methylpentane were used ; their radioactivity was monitored by a proportional counter attached downflow to the gas chromatograph.Relative molar radioactivity values were used for evaluation, with corrections for mole-number changes. Homologation of n-hexane and 3-methylpentane was observed at or above ca. 650 K. 1-1.5% toluene, ca. 0.5% C, aromatics and, with 3-methylpentane feed, up to 0.5% saturated C, products were formed.ll A study of the molar radioactivity of c6 and C, products permitted us to estimate how the chain-growth mechanism operates. The probability that a C, unit picked up by a c6 hydrocarbon will be radioactive 30373038 HOMOLOGATION OVER Pt BLACK is 1/6.This is valid with one labelled atom in the parent molecule and assuming complete breakdown of a c6 alkane into C, units. However, if methylene had originated from removal of the methyl group only, i.e. CGH14 + C5H12 + CH,=M then this probability increases to 1/2 for n-hexane. The relative molar radioactivity of toluene should be 7/6 = 1.17 in the first case and 3/2 = 1.50 in the second case. Table 1 shows that with n-hexane the activity of toluene is close to the former expected value. These data show that, in spite of the fact that platinum is typically singly hydrogenolysing metal,', i.e. the number of fragments per hydrogenolysed molecule, [, is 2, the low amount of multiply fragmented molecules form those active CH,=M surface species that are responsible for chain growth.Thus each carbon atom of the n-hexane chain are equivalent in this respect: -* - 5c1+ *c, . (1) With 3-methylpentane, however, where the label is in the methyl side group, lower toluene activities are obtained. This can be attributed to two reasons. If the alkylating CH,=M units contain no radioactivity, the molar radioactivity of the C, products should be 6/7 = 0.83. This is close to the actual values (table 1) and it may mean that the methyl group containing the 14C label does not produce surface CH,=M species if it is attached to a tertiary centre. The breaking of the primary-tertiary (CI-CIII) bond is hindered in hydrogenolysis over Pt.l3t l4 This is in agreement with the higher activation energy of the ' iso-unit ' type of hydrogen01ysis.l~ This should mean that the splitting of the CII-CIII bond is also slower.Here terminal demethylation may supply most C, units: * * This difference is supported by an analysis of the hydrogenolysis products of these two alkanes. The [ fragmentation factor for n-hexane is 3.3 1 f 0.06 (six runs), whereas the same value for 3-methylpentane is 2.03 f 0.07 (five runs). Thus the fragmentation of n-hexane is indeed deeper. 1.5-2% of its C, and C5 products suffer further isomerization, giving isobutane and isopentane. The amount of isobutane from 3-methylpentane is, however, 6 7 % of the total amount of fragments. This cannot be ascribed to isomerization and points to the validity of reaction (2).Secondly, a peculiar C,-addition-abstraction mechanism may occur here,16 i.e. some of the build-up products may also suffer subsequent degradation before desorption. Thus some toluene may have originated from the demethylation of C, aromatics: * * * * h .... h - .... hJ= b.+. ..@I . (3) . ,."' The conversion of n-octane and 4-methylheptane yielded 4 7 % C, aromatics (more with n-octane) and 3 4 % toluene." This confirms directly the possibility of such a demethylation, whose relative importance may be stronger with branched-chain feed.z. PAAL, M. DOBROVOLSZKY AND P. TETENYI 3039 Table 1. Relative molar radioactivity values of various products formed from radioactive alkanesa relative molar radioactivity product n-hexaneb 3-methylpentanec fragments 1.115f0.06 1.090 f 0.02 isomers 0.98 f 0.02 1.01 f0.08 unreacted feed 0.99+0.01 1.01 k0.08 methylcyclopentane - 1.06 & 0.13d benzene 0.95 f 0.02d - C,-saturated - 0.65f0.10 toluene 1.16 f0.03 0.87 f 0.07 a Catalyst: Pt black, 0.4 g; pulse system: 1 mm3 pulses into hydrogen carrier gas at T = 673 K. Relative molar radioactivities were calculated by dividing the percentage of radioactivity in each substance by the molar percentage of feed used for formation of the product in question. In this way an increase (with fragments) or decrease (with build-up products) of mole number was corrected for. Average of 5 runs. Benzene (Bz) and methylcyclopentane (MCP) were not separated properly. It could be estimated, however, that small amounts of MCP were formed with a large amount of Bz from n-hexane; the reverse was true for 3-methylpentane. Average of 4 runs.Table 2. Relative molar radioactivities obtained with a mixture of [14C,]n-hexane and unlabelled benzenea relative molar radioactivity in the unreacted feed product mixtureb after reactionb fragments isomers n-hexanec benzenec toluened - 2.72 0 3.18 2.88 2.70 0.13 0.72 a See footnote (a) to table 1. Average of two runs each. Initial composition: 37% n-hexane + 63 % benzene. Ca. 8-9 % of n-hexane was transformed into benzene ; this represented ca. 4-5 % of the benzene present after reaction. Toluene content of the effluent was ca. 0.5%. Similarly, degradation of surface oligomers would lead to a partial loss of the original label, so that a still lower molar radioactivity ought to be observed.Our results disagree with this hypothesis.8 The possibility of alkylation of aromatic molecules3 (as an alternative to cyclization of chain-lengthening products) was investigated by reacting the mixture of radioactive n-hexane and inactive benzene (table 2). Here the molar radioactivity of toluene formation from n-hexane is at least 5-6 times faster than alkylation of benzene molecules. This might be a consequence of the low surface coverage of benzene under the experimental conditions used. Benzene yields from pure n-hexane were ca.3040 HOMOLOGATION OVER Pt BLACK 12-14% ,11 compared with 8-9% in its mixture with benzene (table 2). This indicates that benzene displaces n-hexane on some of the active sites for dehydrocyclization, thus hindering this reaction by 3040%.Its coverage cannot be negligible, and it is clear that adsorbed intermediates produced from n-hexane could undergo alkylation more easily. Essentially the same results (i.e. alkane-chain lengthening instead of benzene alkylation) was found with reverse tracer studies, i.e. with the mixture of labelled benzene and inactive 2-meth~1pentane.l~ Separate experiments showed that the cracking of benzene under these conditions is negligible. To summarize: our experiments confirm that homologation occurs by the combi- nation of C, units to (presumably dehydrogenated) open-chain surface specie^.^ The source of the former component is the deep fragmentation of the parent alkane, which is unimportant but not negligible on Pt. Our experiments do not fit the theory of surface oligomer degradation as a main pathway to homologation, but the loss of one or two methyl groups is possible with 3-methylpentane feed.We thank Dr J. Volford and Mrs E. Bursics-Szekeres for synthesizing the labelled 3-methylpentane and n-hexane, respectively. We also thank one of the referees for drawing our attention to certain quantitative aspects of alkylaromatic demethylation and competitive adsorption in hexane +benzene mixtures. J. R. Anderson and B. G. Baker, Proc. R. SOC. London, Ser. A, 1963, 271, 402. Z. Pa61 and M. I. Rozengart, Acta Chim. Acad. Sci. Hung., 1966,49, 395; 1967, 54, 284. L. Guczi, J. Kalman and K. Matusek, React. Kinet. Catal. Lett., 1974, 1, 51; L. Guczi, K. Matusek and P. Tbtenyi, React. Kinet. Catal. Lett., 1974, 1, 291. A. Peter and J. K. A. Clarke, J. Chem. Soc., Faraday Trans. 1, 1976, 72, 1201. C. ODonohoe, J. K. A. Clarke and J. J. Rooney, J. Chem. Soc., Chem. Commun., 1979,648; J . Chem. SOC., Faraday Trans. 1, 1980,76, 345. A. Sarkany and P. TCttnyi, J. Chem. SOC., Chem. Commun., 1980, 525; A. Sarkany, S. Palfi and P. TCtenyi, React. Kinet. Catal. Lett., 1980, 15, 1. J. Margitfalvi, M. Hegediis, S. Gobolos, E. Kwaysser, L. Koltai and F. Nagy, Acta Chim. Acad. Sci. Hung., 1982, 111, 573. S. J. Thomson, in Characterisation of Catalysts, ed. J. M. Thomas and R. M. Lambert (Wiley, Chichester, 1980), p. 214. ' Z. Paal, Adv. Catal., 1980, 29, 273. lo I. Manninger, Z. Paal and P. TCtenyi, Acta Chim. Acad. Sci. Hung., 1978, 97, 439. l1 M. Dobrovolszky, Z. Paal and P. TCtenyi, Acta Chim. Hung., in press. l3 G. Leclercq, L. Leclercq and R. Maurel, J. Catal., 1977, 50, 87. l4 Z. Paal and P. Tetenyi, React. Kinet. Catal. Lett., 1979, 12, 131. l5 K. Foger and J. R. Anderson, J. Catal., 1978, 54, 318; 1979,59, 325; 1980, 61, 140. l8 A. Sarkbny, S. Palfi and P. Tetknyi, Acta Chim. Acad. Sci. Hung., 1982, 111, 633. Z. Paal and P. TetCnyi, Nature (London), 1977, 267, 234. (PAPER 4/204)

ISSN:0300-9599    

DOI:10.1039/F19848003037

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1984, 80, 3041-3050 Diffusion in Dilute Aqueous Solutions of Phosphoric Acid Verification of the Limiting Law for Diffusion of Weak Electrolytes BY DEREK G. LEAIST Department of Chemistry, University of Western Ontario, London, Ontario N6A 5B7, Canada Received 9th February, 1984 The diffusion coefficient of aqueous phosphoric acid has been measured by a conductimetric technique at 25 "C over the concentration range from 0.0012 to 0.0826 mol dmW3. Below 0.01 mol dm-3 dissociation of neutral H3P0, molecules to H+ and H2P0, produces a sharp increase in the diffusion coefficient of the phosphoric acid component. When a small correction is included to allow for the association reaction H,PO, + H,PO; H,P20;, the results are in good agreement with predictions based on the limiting law for diffusion of a weak electrolyte. Analysis of the data gives (0.87 f 0.01) x m2 s-l for the limiting diffusion coefficient of the aqueous phosphoric acid molecule.Within experimental error this value is identical to the limiting diffusion coefficient of H,PO;. Diffusion of weak electrolytes takes part in a wide range of chemical and biological processes. Since these components are transported in solution as molecular and ionic species, in proportions that vary with concentration, their diffusional properties are more complicated than those observed for either strong electrolytes or non-electrolytes. For example, the mobility of a weak electrolyte component can be unusually sensitive to c0ncentration.l Conveniently, however, most of the concentration dependence of the mobility of an incompletely dissociated solute can be attributed to changes in the relative proportions of the diffusing species, rather than to changes in the mobilities of the individual species2 This means that the diffusion coefficient can be predicted if the dissociation constants and mobilities of the various species are known.3 Conversely, diffusion measurements can provide useful information about the concentrations and mobilities of species in solution.l? 2* 4-6 Very recently, Harned's conductimetric technique for the determination of diffusion coefficients of salt solutions was successfully extended to a weak-electrolyte system.' Diffusion coefficients determined by this technique for dilute aqueous acetic acid were in good agreement with values predicted by the limiting law for diffusion of a weak electrolyte.However, at 0.00 1 mol dm-3, the lowest concentration accessible to experiment, only ca. 10% of the acetic acid molecules are dis~ociated.~ So that a more sensitive test of the limiting law can be made for diffusion in this important class of solutions, we report here diffusion coefficients for dilute aqueous solutions of phosphoric acid. Since the first dissociation constant of phosphoric acids is two orders of magnitude larger than the dissociation constant of acetic acid, diffusion of the former can be measured and compared with theory up to much higher degrees of dissociation (ca. 90% ) than is practicable for acetic acid. Other reasons for the choice of aqueous phosphoric acid include the general importance of this system as well as the availability of precise therrnodynamics-ll and ionic mobility12-14 data required for a detailed analysis of the diffusion measurements.Although phosphoric acid molecules 99 304 1 FAR 13042 DIFFUSION IN SOLUTIONS OF PHOSPHORIC ACID show a marked tendency to dimerize, corrections for this effect are readily made at the concentrations used here, as are corrections for electrophoresis and viscosity changes of the solutions. This has enabled us to obtain a reliable value for the mobility of the neutral phosphoric acid molecule, which we compare with the mobility of the structurally similar H,PO; ion. EXPERIMENTAL Edwards and Huffman14 have used Gouy interferometry to measure the diffusivity of aqueous phosphoric acid at concentrations from 0.036 to 16 mol dm-3.Below 0.14 mol dmP3, however, accurate analysis of the interferograms was difficult owing to skewed, non-Gaussian boundaries produced by the strong concentration dependence of the diffusion coefficient. In very dilute solutions, where most of the acid is ionized, the refractive index along the diffusion column becomes too small to be measured precisely. PROCEDURE So that precise diffusion data could be determined at low concentrations of theoretical interest, we used the conductimetric method mentioned previously.3* A calibrated syringe was used to introduce a small volume of electrolyte solution into the bottom of a vertical chamber [height (4.564k0.002) x loT2 m] filled with solvent.The electrolyte was allowed to diffuse upward for a period of several days. The rate of diffusion was determined from the rate of change of the electrical conductance of the solution measured at one-sixth of the distance from the top and bottom of the column. Further details of the procedure and equipment have been described previou~ly.~~ l5 Conductances were measured six times per day over a period of four days. The readings were taken from one and a half days after introduction of electrolyte. After this waiting period the concentration gradient in the cell was very small. This helped to ensure that the measured diffusion coefficient corresponded to the differential value at the final cell concentration.16 MATERIALS Reagent-grade phosphoric acid was used without further purification.Solutions were made up from deionized, doubly distilled water. Corrections for solvent conductivity (6 x lop5 Q-l m-l) were negligible. Acid concentrations were determined by titration against standardized base. TREATMENT OF EXPERIMENTAL DATA The first and second thermodynamic ionization constants8, of phosphoric acid are Kl = 7.121 x mol dmP3. In the solutions used here, the concentrations of H+ and H,PO; arising from the first stage of ionization are approximately equal and much larger than the negligibly small concentration of HP0,2- (ca. 6 x mol dm-3) produced by the second ionization. Therefore, computation of the diffusion coefficient from conductances was based on the difference equation and K2 = 6.3 x derived previously for a 1 : 1 weak ele~trolyte,~ where D is the differential diffusion coefficient, a is the height of the diffusion column, KB and KT denote conductances measured at the bottom and top electrode pairs and r = k,/kT is the ratio of electrode cell constants.The second logarithmic term in eqn (1) is a small correction (< 3% of D) that allows for the non-linear change with concentration of the conductivity of the dissociating electrolyte. The symbols a, and A refer to the apparent degree of dissociation and the molar conductivity of the electrolyte, both of which are evaluated at the final cell concentration c,.D. G. LEAIST 3043 We obtained A values by graphical interpolation of the conductances reported by Mason and C~1vern.l~ Degrees of dissociation were calculated from the equilibrium relation Values for the mean ionic activity coefficient y+_ and the activity coefficient of the molecular acid ym were calculated from the semi-empirical Pitzer equations.l7 The parameters of the equations were determined by Pitzer and Silvester,ll who fit the equations to the e.m.f. data of BateP and the osmotic coefficient data of Platfordlo and Elmore et aL9 The molal activity coefficients yi obtained from this source were converted to molar-scale coefficients y z required here by using published densities13 and the relations1ga Y , = mpo Y i / c (3) (4) c/m = [p - (cM/ 1 OOO)] where m is the solute molality, p is the solution density, po is the density of the pure solvent and M is the molecular weight of phosphoric acid. RESULTS Table 1 contains characteristic data from one of the diffusion experiments.The D values shown in the fourth column were calculated from eqn (1) by pairing each conductance reading with the corresponding reading taken on the following day (At “N 24 h). Over each four-day period during which the conductance readings were taken the values of D were sensibly constant. This behaviour confirmed that the diffusion cell was operating correctly. Also, it showed that the measured diffusion coefficients were differential values,16 even though D changes rapidly with c. The results of the diffusion measurements are recorded in the second column of table 2. These values were obtained by averaging the eighteen D values determined for each experiment, as illustrated in table 1.In each case the standard deviation from the mean was ca. 0 . 1 4 2 % of D. When uncertainties in the cell dimensions and electrode cell constants are included, we estimate that the accuracy of the data is - +0.5%. Between 0.036 and 0.082 mol dmP3, our conductimetric results overlap the Gouy data of Edwards and Huffman.l* At the lower end of this range the conductimetric data are ca. 1 % lower than the Gouy data. In that region, however, the Gouy results may be slightly too high20 because the effects of the strong concentration dependence of D on the interference patterns were neglected. Nevertheless, the two sets of data are in close agreement, well within the suggested uncertainties of the two methods. The good agreement between the two independent methods lends confidence to the conductimetric procedure used here.For a binary solution containing solvent and a single solute component, irreversible thermodynamics provides the relation21 D = RTUf between the diffusion coefficient and the solute mobility U, where f is a dimensionless thermodynamic factor defined by f = “(F) RT OC T , p where p is the chemical potential of the solute component. Provided activity data are available for the evaluation o f 5 eqn ( 5 ) can be used to derive the mobility of the solute from diffusion measurements in cases of arbitrarily complex solution chemistry.22 99-23044 DIFFUSION IN SOLUTIONS OF PHOSPHORIC ACID Table 1. Dataa for the computation of the diffusion coefficient of aqueous phosphoric acid at 25 "C 132 207 138 616 144 297 149 548 153 992 161 966 219 068 226 293 230 556 236 793 243 177 247 334 305 749 310 856 317 462 323 120 328 197 234 097 391 185 396 676 401 664 407 444 413 358 417 181 0.599 018 0.639 304 0.673 854 0.704 473 0.730 560 0.775 374 1.051 91 1.080 60 1.096 79 1.119 68 1.142 53 1.156 87 1.325 73 1.338 63 1.352 87 1.364 84 1.376 58 1.387 54 1.485 80 1.493 47 1.500 15 1.507 48 1.514 85 1.519 50 3.199 80 3.158 19 3.123 15 3.091 14 3.065 44 3.020 97 2.762 35 2.735 37 2.719 59 2.696 74 2.674 68 2.660 67 2.495 68 2.484 21 2.468 71 2.455 30 2.445 17 2.431 42 2.328 69 2.320 22 2.312 49 2.303 57 2.294 98 2.289 64 1.161 5 1.160 6 1.161 1 1.161 8 1.162 1 1.162 5 1.163 4 1.164 1 1.162 1 1.162 0 1.161 4 1.161 6 1.160 9 1.160 6 1.163 0 1.163 0 1.163 5 1.163 0 a c, = 0.007 10 mol dm-3, a = 0.045 64 m, r = 0.8530, a,c,A/1000 (1 -ao) k, = 2.985 x f2-l.Average value D = 1.1621 x m2 s-l, standard deviation 0.0010 x m2 s-l. In order to derive mobilities for the system under discussion, thermodynamic factors from the relation which is by differentiation of eqn (2) and the expression p = p0+2RT1n (aocy+) we calculated (7) for the chemical potential of ionized phosphoric acid. Upon substitution of activity coefficients from ref. (1 1) into eqn (7),ftakes the values listed in table 2. The striking concentration dependence off is in sharp contrast to the much weaker concentration dependence observed for solutions of non-electrolytes4 or strong electrolyte^.^^ In very dilute solutions, dissociation of the phosphoric acid molecules effectively doubles the number of solute particles per formula weight of the electrolyte.This has the effect of increasing the free-energy gradient (the driving force for diffusion) per unit concentration gradient. In this way dissociation accounts for most of the concentration dependence of both D andfobserved at low concentrations. Values for the mobility of the aqueous phosphoric acid component (in units ofD. G. LEAIST 3045 Table 2. Diffusion coefficients and supplementary dataa- for aqueous solutions of phosphoric acid at 25 "C C Dobs Dcalc RTU A,+A, Di / 1 0-3 mol dmP3 lo-, m2 s-l a P f rllrl" 0.00 - 1.606' 0.803 0.000 - 1.000 0.000 2.000 1.000 1.20 1.405 1.415 0.799 -0.015 (0.827) 0.879 0.000 1.758 1.000 1.80 1.348 1.366 0.796 -0.018 (0.804) 0.837 0.000 1.693 1.000 2.84 1.298 1.307 0.805 -0.021 (0.849) 0.780 0.001 1.613 1.001 4.26 1.229 1.253 0.798 -0.024 (0.815) 0.722 0.001 1.539 1.001 5.68 7.10 8.56 13.8 22.7 22.7 36.0 36.2 39.7 48.6 62.6 62.9 71.4 71.7 82.6 .216 1.216 0.817 -0.027 (0.872) 0.643 0.002 1,488 1.001 .162 1.188 0.802 -0.029 (0.822) 0.643 0.002 1.449 1.002 .153 1.165 0.815 -0.030 (0.849) 0.613 0.002 1.415 1.002 .lo4 1.113 0.822 -0.035 0.857 0.537 0.004 1.343 1.003 .060 1.067 0.830 -0.040 0.862 0.460 0.007 1.277 1.006 .061 1.067 0.831 -0.040 0.863 0.460 0.007 1.277 1.006 .041d 1.032 0.850 -0.045 0.885 0.396 0.010 1.225 1.009 1.043d 1.032 1.028 1.025 1.020d 1.012 1.003 0.997 1.002 0.997 0.991d 0.990 0.994d 0.989 0.990 0.982 0.851 0.846 0.853 0.854 0.853 0.851 0.854 0.859 - 0.045 - 0.046 - 0.048 - 0.05 1 - 0.05 1 - 0.053 - 0.053 -0.055 0.888 0.875 0.882 0.880 0.879 0.874 0.878 0.883 0.395 0.383 0.357 0.327 0.327 0.313 0.312 0.297 0.0 10 0.01 1 0.013 0.016 0.0 16 0.018 0.0 18 0.020 1.225 1.009 1.215 1.010 1.196 1.013 1.174 1.016 1.174 1.017 1.164 1.019 1.164 1.019 1.153 1.022 a Dy = 9.315 x Di = 0.87, x lop9 and Dy = 0.80 x lo-, m2 SKI.Solvent viscosity1Yd Nernst value calculated q0 = 8.90 x from eqn (21). kg m-' s-' and solvent dielectric constant30 E = 78.3. These data obtained by the Gouy method.I3 m2 s-l) derived from the diffusion measurements are listed in table 2. Although highly mobile hydrogen ions are generated, the overall mobility of the phosphoric acid component actually decreases with the degree of dissociation. This behaviour is to be expected,lgb however, since a single H3PO4 molecule can move through the solution with less frictional resistance than two separate H+ and H,PO; ions.DISCUSSION In addition to H3P0,, H+ and H,PO; species, aqueous solutions of phosphoric acid contain smaller amounts of the dimer H,P,O, and its dissociation product H,P,OT.~~ Before a detailed analysis of the overall mobility of the phosphoric acid component can be given, the concentrations of the various solute species must be established. Consider the following equilibriallq 25 H,PO, H+ + H,PO;, Kl = 7.121 x mol dmP3 H,P,O, + H+ + H,P,O;, H,PO, + H,PO; + H,P,O;, KO = 1.3 dm3 mol-1 K4 = 0.3 mol dm-3. Since the concentrations of the major species can be given as [H3P04] z (1 - a,) c and [H+] z [H,PO;] z a,c, we have [H,P,O;] z (1 -a,) a, c2K, and[H,P,O,] = (1 - a,) a; c3K,/K4.3046 DIFFUSION IN SOLUTIONS OF PHOSPHORIC ACID At 0.0826 mol dmP3, the highest concentration used here, we find [H,P,O;] z 0.0018 mol dm-3, which is small but not negligible, whereas [H,P,O,] is only 0.0002 mol dmP3.Therefore, we will include the species H3P04, H+, H,PO, and H,P,O, in our analysis of the mobility of phosphoric acid, but neglect H,P,O,. In terms of the degree of dissociation of H,POg, a = [H+]/c, and the extent of formation of H,P,O; which, for convenience, we define as P = [H,P,O;]/c, the concentrations of the diffusing species are given by [H,PO,] = (1 - a-P) c, [H+] = ac, [H,PO;]' = (a -P> c and [H,P,O;] = Pc. Values of a and P are readily calculated by iterative solution of the equilibrium relations (10) P (1 -a-P>(a-P>c' KO = Note that we have omitted activity coefficients from eqn (10).This approximation is justified on the grounds that values of y , are close to unity. Also, yH,P,O, in the numerator will nearly cancel yVHZPO, in the denominator. In order to simplify the calculations further, we used values of y , and y , obtained from ref. (1 I), as described previously, even though these coefficients are based on the assumption that the only solute species are H,POg, H+ and H,PO;. Since only small amounts of H,P,O, are formed, these approximations will lead to negligible errors in the values determined for a and /? at the conentrations used here. Calculated values for these parameters are listed in table 2. Having identified the major solute species and estimated their concentrations, we turn now to the problem of determining the effect of the solution chemistry on the mobility. We start with the rigorous relation22T 26 u = [vT($l$T)-' v]-l/c (1 1) between the mobility of the solute component and the transport coefficients &k of the solute species.v and 0 are matrices of stoichiometric coefficients in which vip and Giq refer to the number of moles of constituent ion i per mole of solute component p and per mole of species q , respectively, and superscript T denotes the transpose. If we adopt the following numbering arrangement for the components: 1 = H3P0,; constituent ions: 1 = H+ and 2 = H,PO;; species: 1 = H+, 2 = H,PO;, 3 = and 4 = H,P,O;, we find / 1 \ $=(I O 1) 0 1 1 2 . Except for an electrophoretic correction that will be made later, we will assume that the off-diagonal elements of 1 are zero and invoke the reasonable approximation that the mobilities of the species are constant.We then have'. 27 where 6 i k is the Kronecker delta and ti and DY refer to the concentration and limiting diffusion coefficient of species i.D. G . LEAIST 3047 Upon substitution of eqn (12)-( 14) into eqn (1 1) we obtain and hence* a(a -p) DY Di + 4apDy Di + (a -p)pD; 040 a(A1 + A2) aDy + (a -p) D; +PDl + 2 (16) In obtaining this expression for D we have followed the frequently used procedure whereby the predicted diffusion coefficient is multiplied by the ratio of the viscosity of the pure solvent to the viscosity of the solution. Values for this ratio obtained from data given in ref.(28) are listed in table 2. In order to allow for the effects of interionic attractions on the ionic mobilities, we have also included the Onsager-Fuoss electrophoretic c~rrection~~f(A, + A2)/2 appropriate for a 1 : 1 weak electr01yte.l~~ In the usual notation,'. l g b ? 31 the first- and second-order correction terms are given by with A ---( e2 ~ e x p ( ~ a ) (0.5772+ln(2~a)+ ( - 2 ~ a ) ~ 27~47~ I + K ~ n=1 n(n!) 2 - 4nNe2 tiz; K2 = ____ 1OOOEkT i (19) evaluated at ionic strength ac. In order to obtain the above expressions for the electrophoretic terms, we have neglected the fact that the solutions contain minor amounts of H,P,O;. Since the electrophoretic correction is small (< 2% of D), this approximation will lead to negligible error in the calculated D values.We based our calculations of A, and A2 on the arbitrary although reasonable value a = 5 x m for the ion-size parameter (an accurate value is not required). Other parameters used in the electrophoretic calculations are listed in table 2. The DY values for the ionic species are obtained from limiting molar conductances according to the convenient relation For H+, we have the accurate valuel29 lgC A; = 0.034981 m2 R-l. Since modern estimates13 of the limiting conductance of H2PO; range from 0.003218 to 0.003 335 m2 R-l, the value of A; is more uncertain. Following Mason and Cu1vern,l3 we will use 1; = 0.00330 m2 i2-l. Based on these values, the limiting diffusion coefficient of phosphoric acid is given by the Nernst relation for fully dissociated H+ + H,PO;.* In very dilute solutions where formation of H,P,O; is safely neglected, eqn (16) reduces to the result25 obtained previously for a simple 1 : 1 weak electrolyte.3048 DIFFUSION IN SOLUTIONS OF PHOSPHORIC ACID Since less frictional resistance is offered when H3PO4 and H2P0, merge into one diffusing entity, formation of H5P20, will tend to increase the mobility of the phosphoric acid component.* At the concentrations used here, however, this increase amounts to < 2% of D . Since this is clearly a small effect, an accurate value of di for H5P20, is not required; we will use the value 1; = 0.003, m2 R-l suggested by Elmore et ~ 1 . ~ ~ All the values necessary for the determination of Di for the neutral HaPo4 molecule are assembled in table 2.Upon substitution of each experimental value for D into eqn (1 6) and solving, one obtains the values given in table 2. At concentrations from ca. 0.01 to 0.0826 mol drn-,, Di values calculated in this manner are essentially constant. Averaging the values obtained within this range gives Di = (0.87 kO.01) x m2 s-l. When this value is substituted into eqn (16), the calculated curve for D plotted in fig. 1 is obtained. The agreement between the measured and predicted diffusion coefficients shown in fig. 1 is excellent at concentrations > 0.01 mol drn-,. Below this limit the observed coefficients fall sightly below the calculated curve. We will return to this small discrepancy. It is evident from table 2 that the values computed for Di for concentrations c 0.01 mol dm-, are several percent lower than the value 0.87 x m2 s-l obtained from analysis of data at higher concentrations.However, this discrepancy might reflect an incorrect value used for the molar conductance of the anion H2PO; rather than a significant departure from the diffusional behaviour predicted by theory. If we consider the approximate relation for Di Di z [(2 - a) Dobs - aDo]/2( 1 -a) which is obtained by omitting from eqn (16) minor corrections for activity coefficients, viscosity changes, dimerization and electrophoresis, it follows that an error SD: - in the value used for the limiting diffusion coefficient produces the error a SDO, SD" - -____ 2(1-a) - 3 - in the value calculated for Di. Taking 0.00003 W1 as an estimate of the probable error in A;,13 the corresponding error in 0: will be SDO, = 0.013 x m2 s-l.At phosphoric acid concentrations of 0.001 2-and 0.008 56 mol drn-,, the resulting errors in Di will be -0.047 x m2 s-l, respectively. Since SDi is appreciably magnified at low concentrations by the factor 1 - a appearing in the denominator of eqn (23), we have therefore omitted data obtained below 0.01 mol dm-3 from the determination of Di. Had we based our calculations on the value A; = 0.003 27, which is slightly lower than the recommended value of 0.003 30 m W1, but a plausible value nevertheless, the value Di = (0.87+0.01) x m2 s-l would be obtained over the entire range of concentration investigated here. Also, the slightly lower value for A; would bring the observed and predicted D values into excellent agreement at concentrations < 0.01 mol drn-,.In a previous study, Edwards and Huffmann14 reported the approximate value 0.76 x m2 s-l for the limiting diffusion coefficient of the H3P0, molecule, a value 13% lower than our result for Di. The earlier value was calculated from optical data by extrapolating the apparent molecular diffusivity4v5 and -0.010 x * In addition, formation of H,P,O; will decrease the thermodynamic factorf. However, since thefvalues shown in table 2 were obtained from experimental data, this effect is included implicity.D. G. LEAIST 3049 I I I I I I I 1 I I I 0.16 0.20 (ac)+/mol* d m t Fig. 1. Comparison of observed and predicted diffusion coefficients for aqueous solutions of phosphoric acid at 25 "C: (-) eqn (16); 0, present conductimetric results; 0, optical data of Edwards and Huffman.l* to zero concentration, where 1 / Q is an empirical correction14 for viscosity changes of the solution approximately equal to q"/q.Because the plot of DL against c showed pronounced curvature at low concentrations, an accurate value for Di was not obtained. The present determination of Di is based on data obtained for concentrations c 0.1 mol dm-3. In this region viscosity corrections are small and unambiguous. Explicit corrections are included in the analysis to allow for electrophoresis and the formation of H,P20;. Also, we have taken advantage of accurate expressions for phosphoric acid activities that have recently become available. For these reasons we consider 0.87 x loa9 m2 s-l to be a more reliable estimate of Di.m2 s-l, of the H2PO; anionisidentical withinexperimentalerrortoDi = (0.87 kO.01) x lo-, m2 s-lobtained here for the undissociated H3P04 molecule. This comparison suggests that the charge on the anion does not produce stronger solvation. However, as Stokes4 has pointed out, a reduction in the anion's mobility resulting from stronger ion-solvent interactions could be masked by a structure-breaking effect of the ion on nearby water molecules, which would 'lubricate' movement of the ion through the solvent. Limiting diffusion coefficients of anions of aqueous formic, acetic, propionic and n-butyric acids are 3-10% lower than the diffusion coefficients of the corresponding molecular specie^.^^^ On the other hand, the diffusion coefficient of the dihydrogen citrate ion is 23% larger than the diffusion coefficient of molecular citric acid.4 The ratio Di/Di z 1.00 for phosphoric acid is thus typical of values found for other ani on-molecule pairs.To sum up, the present investigation has shown that Harned's conductimetric technique is well suited to the difficult problem of determining accurate differential diffusion coefficients in cases where D varies rapidly with concentration. Since the measurements reported here for phosphoric acid cover a wide range of values for the degree of dissociation, yet at concentrations low enough so that the necessary approximations made in the analysis of the data are justified, the good agreement Note that the limiting diffusion coefficient, Di = 0.87, x3050 DIFFUSION IN SOLUTIONS OF PHOSPHORIC ACID between the observed and predicted diffusion coefficients provides compelling evidence for the validity of the limiting law for diffusion of weak electrolytes.With the value determined for the diffusion coefficient of the H,PO, molecule, the diffusional properties of more complicated phosphate-containing systems, such as phosphate buffers, can be predicted.31 We thank Tom Richards for technical assistance and we gratefully acknowledge financial support by the Natural Sciences and Engineering Research Council of Canada. C. W. Garland, S. Tong and W. H. Stockmayer, J. Phys. Chem., 1965, 69, 1718. J. M. Creeth and B. E. Peters, J. Phys. Chem., 1960, 64, 1502. E. L. Holt and P. A. Lyons, J. Phys. Chem., 1965, 69, 2341.G. T. A. Muller and R. H. Stokes, Trans. Faraday SOC., 1957, 53, 642. L. A. Dunn and R. H. Stokes, Aust. J. Chem., 1965, 18, 285. V. Vitagliano and P. A. Lyons, J. Am. Chem. SOC., 1956, 78,4538. ' D. G. Leaist and P. A. Lyons, J. Solution Chem., 1984, 13, 77. R. G. Bates, J. Res. Natl Bur. Stand., 1943, 31, 205. K. L. Elmore, C. M. Mason and J. H. Christensen, J. Am. Chem. Soc., 1946,68, 2528. lo R. F. Platford, J. Solution Chem., 1975, 4, 591. l1 K. S. Pitzer and L. F. Silvester, J. Solution Chem., 1976, 5, 269. l 2 B. B. Owen and F. H. Sweeton, J. Am. Chem. Soc., 1941, 63, 281 1 . l 3 C. M. Mason and J. B. Culvern, J. Am. Chem. SOC., 1949, 71, 2387. l4 0. W. Edwards and E. 0. Huffman, J. Phys. Chem., 1959,63, 1830. I5 D. G. Leaist and P. A. Lyons, Aust. J . Chem., 1980, 33, 1869. H. S. Harned and R. L. Nuttall, J. Am. Chem. SOC., 1947, 59, 736. K. S. Pitzer, J. Phys. Chem., 1973, 77, 268. R. G. Bates, J. Res. Natl Bur. Stand., 1951, 47, 127. Is R. A. Robinson and R. H. Stokes, Electrolyte Solutions (Academic Press, New York, 2nd edn, 1959), (a) p. 31, (b) chap. 1 1 , (c) appendix 6.1, (d) appendix 1 . 1 . 2o J. A. Rard and D. G. Miller, J. Solution Chem., 1979, 8, 701. 21 A. Katchalsky and P. F. Curran, Non-equilibrium Thermodynamics in Biophysics (Harvard University Press, Cambridge, Mass., 1965), chap. 9. 22 W. H. Stockmayer, J. Chem. Phys., 1960, 33, 1291. 23 D. G. Leaist, Can. J. Chem., 1983, 61, 1494. 24 H. S. Harned, Discuss. Faraday SOC., 1957, 24, 9. 25 K. L. Elmore, J. D. Hatfield, R. L. Dunn and A. D. Jones, J . Phys. Chem., 1965, 69, 3520. 26 D. G. Leaist, J . Chem. SOC., Faraday Trans. I , 1982, 78, 3069. 27 J. M. Creeth and R. H. Stokes, J. Phys. Chem., 1960, 64, 946. 28 0. W. Edwards and E. 0. Huffman, Ind. Eng. Chem., Chem. Eng. Data Ser., 1958, 3, 17. 2s L. Onsager and R. M. FUOSS, J. Phys. Chem., 1932, 26, 2689. 30 T. A. Renner and P. A. Lyons, J. Phys. Chem., 1974, 78, 1667. 31 D. G. Leaist and P. A. Lyons, J. Phys. Chem., 1981,85, 1756. (PAPER 4/234)

ISSN:0300-9599    

DOI:10.1039/F19848003041

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. I, 1984, 80, 3051-3058 Effect of Ionic Strength on the Size-exclusion Chromatography of Aqueous Polystyrene Latices BY MARK G. STYRING, CHRISTOPHER J. DAVISON, COLIN PRICE AND COLIN BOOTH* Department of Chemistry, University of Manchester, Manchester M 13 9PL Received 16th February, 1984 Purified polystyrene latices with narrow particle-size distributions have been prepared and studied by size-exclusion chromatography (s.e.c.) with columns packed with porous glass beads. Mean elution volumes and peak widths are found to increase as the ionic strength, I, of the eluant is increased. This can be interpreted quantitatively in terms of an effective exclusion radius of a latex particle which includes a contribution from the electrical double layers that is proportional to 1-i.The chromatographic separation of aqueous latices according to particle size has attracted interest since the early work of Small et a1.l and Krebs and Wunderlich.2 The topic has been reviewed re~ently.~ Attention has been paid to several chromato- graphic method^,^ and elution through packed beds of porous beads seems to offer resolution over the widest range of particle sizes. This method is generally termed size-exclusion chromatography (s.e.c.). Aqueous polystyrene latices are ideal systems for investigation of the separation in s.e.c. At room temperature the latex particles are rigid spheres with reproducible surface properties after p~rification.~ Methods are available for the preparation of polystyrene latices with narrow particle-size distributions, some of which 5-7 we have used in our work.The stability of polystyrene latex particles in an s.e.c. system is determined by a balance of repulsive and attractive interactions between particles and packing. The former originate in the overlap of electrical double layers8 and the ' steric' repulsion of adsorbed layers of non-ionic chain molecule^,^ and the latter in van der Waals The ' steric' repulsion of short-chain molecules adsorbed on particles and packing is short range in nature, and incorporation of (say) a non-ionic surfactant in the s.e.c. system increases the size of the core of the particle. The repulsion caused by overlap of electrical double layers is of longer range. Moreover, the range depends on the ionic strength ( I ) of the aqueous medium.Consequently, provided that a critical value of I (analogous to the critical value in DLVO theory8) is not approached, a latex particle will elute without adsorption at a volume determined by its effective exclusion radius, re, i.e. the core radius of the particle in the system plus a contribution from the repulsion of double layers. This latter contribution can be quantified as the sum of the thicknesses of two double layers, one on the particle and one on the packing. The effect of ionic strength on the thickness of an electrical double layer is well known. In consideration of colloidal stabilitys it suffices to use the approximations of Gouy and Chapman, whereby the thickness of the double layer varies as I-;. Consequently it may be predicted (vide infra) that the s.e.c.elution volume of a polystyrene latex will decrease approximately linearly with I-;, when other factors are kept constant, provided that the parameter which determines separation in s.e.c. 305 13052 SIZE-EXCLUSION CHROMATOGRAPHY OF POLYSTYRENE LATICES depends on the effective exclusion radius, re. The core radius of the particle will determine separation only when the double-layer thickness is zero, i.e. when I-+ 00. This condition cannot be approached experimentally, because of adsorption, but may be examined by extrapolation of data obtained at finite values of I to I = co. Surprisingly this prediction has not yet been tested directly. It has been shownlO that the s.e.c. elution volume of a polystyrene microgel in dimethylformamide with added lithium bromide varies linearly with I-;, but this system is far from ideal since the microgels are not rigid spheres.As mentioned earlier,l0 data in the literature1 are consistent with the proposition. Here we present a study of the effect of ionic strength on the s.e.c. elution volume of well characterised polystyrene latices which supports the argument presented above. EXPERIMENTAL PREPARATION AND CHARACTERISTICS OF LATICES Seven polystyrene latices were prepared by emulsion polymerisation procedures designed to produce narrow distributions of particle size. Four samples, designated 3, 4, 6 and 7 (see table 1) were prepared by the method of Woods et aL5 with mixed anionic (ultrawet K) and non-ionic (Triton X-100) surfactants. Sample 5 was prepared by a method suggested by Dunn6 with a single anionic (Aerosol MA-80) surfactant.Samples 1 and 2 were prepared by a method reported by Lichti et al.' with a single anionic (sodium dodecyl sulphate) surfactant but short-stopped by addition of quinone at 1&30% conversion. Each of the latices was purified by the mixed-bed ion-exchange technique of McCann et aL4 Particle diameters were determined with a JEOL lOOCX electron microscope operated at 100 kV and 13 000 x magnification. Specimens for examination were prepared by allowing drops of a dilute suspension of the purified latex (ca. 5 g drn-,) to dry on a thin carbon substrate supported by a copper grid. Average particle sizes were calculated from measurements made on enlarged prints; the overall magnification was ca.40000 x . Values of average diameters are listed in table 1. These are the number-average diameter, D,, and the root-average cube diameters, (0;); and (0;);. The ratio D&/Di, equivalent to the ratio M,/M, for spherical particles, serves to indicate the width of the particle-mass distribution. The number-average molar masses of the particles, calculated assuming a density of 1.06 g ~ m - ~ , vary from ca. 7 x 106 g mol-' (sample 1) to ca. 280 x 106 g mol-l (sample 7). VISCOMETRY Intrinsic viscosities, [v], were determined at 25 & 0.5 "C for the latices in aqueous solutions similar to those employed in the s.e.c. experiments : i.e. aqueous solutions containing inorganic salt (NaC1 or NaNO,) at concentrations < 0.1 mol dm-3 plus Ultrawet K (sodium dodecyl- benzene sulphonate, 0.4 g dm-3).A Ubbelohde viscometer with solvent flow time near 170 s was used and values of [v] were obtained from the common intercepts of Huggins and Kraemer plots for 5 concentrations. The kinetic energy correction was CQ. 2%. No density correction was necessary since the densities of polystyrene and water are similar. SIZE-EXCLUSION CHROMATOGRAPHY Four stainless-steel columns (length ca. 1 m, diameter ca. 6 mm) were packed with glass beads varying in porosity from 7.5 to 200 nm (CPG- 10 series, 12&200 mesh, Electro-Nucleonics Inc.) to 500 nm (Fractosil, 120-230 mesh, Merck Ltd). The overall plate count was 750 m-l. The eluant was degassed distilled water containing NaN, (0.1 g dm-3, antifungal agent), Ultrawet K (0.4 g dm-l, above the c.m.c.), buffering salts (pH ca.6, NaH,PO, and Na,HPO,, maximum concentration ca. 20 mmol drn-,) and NaNO, in sufficient quantity to bring the eluant to the required ionic strength, in the range 8-250 mmol drn-,. Latex solutions (ca. 4 g dmP3 in aqueous solutions of the same composition as the eluant) were filtered (Millipore, 10 mp) before being introduced into the eluant stream by means of a 2 cm3 loop and a Rheodyne injection valve. The eluant flow rate was maintained at 1 cm3 min-' by means of a Du Pont model 870 pump. The detector was a Waters Associates model R403 differential refractometer.M. G. STYRING, C. J. DAVISON, C. PRICE AND C. BOOTH 3053 Table 1. Diametersa of polystyrene latices from electron microscopy 1 28 29 33 1.46 2 30 31 33 1.24 3 58 59 62 1.14 4 67 69 71 1.12 5 74 74 75 1.04 6 92 93 96 1.06 7 94 95 98 1.06 a Error & 5% Elution volumes were measured relative to those of a low-molar-mass marker (either tetrahydrofuran or benzyl alcohol) added to the injected solution.Wolkoff and Larosel’ have discussed the problems of measuring elution volumes in aqueous s.e.c. by the customary method with a syphon, and have pointed out that variation in ionic strength leads to variation in surface tension and thereby to variation in volume per syphon count. Our investigations confirm this effect. Moreover, with an eluant of given ionic strength and a given latex and marker, reproducibility of elution volume measured by a syphon and reproducibility of emergence time were both no better than 1 % .However, the ratio of emergence time of latex to that of marker ( E ) was constant to better than f 0.1 :d. S.e.c. curves were analysed by a procedure similar to that described by Pickett et a1.12 in order to find the weight-average emergence-time ratio, E,. This quantity should correlate more closely with particle size than the emergence-time ratia of the peak, E ~ ~ . RESULTS AND DISCUSSION INTRINSIC VISCOSITY Irrespective of the ionic strength, the intrinsic viscosities were all within the range [q] = 2.8 f0.3 cm3 g-l. Values reported by othersl37 l4 for stabilised latices in aqueous solutions with I in the range 1-100 mmol dmP3 are similar ([q] = 2.8 f0.2 cm3 g-l) and also independent of ionic strength. The measured value of [q] is slightly higher than Einstein’s predicted value of 2.5 cm3 g-l, which may be because of the adsorbed layer of surfactant.The near conformity with Einstein’s prediction confirms that the particles can be well modelled in their hydrodynamic behaviour by hard spheres. S.E.C. CURVES Selected s.e.c. curves are illustrated in fig. 1. All the latices gave broad peaks with half-widths at 0.6 height (a) in the range 6-8 cm3 at elution volumes of 100-120 cm3. Comparable values of a have been reported for similar s.e.c. systems: e.g. ca. 5 cm3 by Husain et aZ.15 At I = 250 mmol dmP3 the latex particles do not emerge from the columns, presumably because of adsorption. EFFECT OF IONIC STRENGTH The s.e.c. elution volume of a particle of given size is given by v,= V,+KC ( 1 ) where V, is the void volume (i.e.total volume minus bead volume), 5 is the pore volume and K is the partition coefficient for the particles between pores and void which depends upon the effective size (e.g. upon the radius of a sphere13). Any additional3054 SIZE-EXCLUSION CHROMATOGRAPHY OF POLYSTYRENE LATICES A n . 1 1 1 I I 1 c 0.5 0.7 E 0.9 0.5 0.7 E 0.9 Fig. 1. S.e.c. curves of refractive-index difference (An) against elution volume relative to that of benzyl alcohol ( E ) for (a) latex 1 and (b) latex 7. In each case the ionic strength ( I ) of the eluant is 8 (lower E ~ ~ ) or 110 mmol dm-3 (higher E ~ ~ ) . exclusion will decrease K . For additional exclusion caused by overlap of electrical double layers, K will be decreased by an amount dependent on the double-layer thickness, which, in turn, is proportional to I-; for surfaces of constant potential, The partition coefficient K can be written16 K = [(a - r,)/aI2 = [ 1 - (r,/a)], where a is the equivalent cylindrical pore radius of the packing and re the effective particle radius.With re set equal to the sum of contributions from the core and the double layer, i.e. re = r,+6 K can be rewritten where Kl = [I -(r,/a)I2 is the partition coefficient when 6 = 0 (i.e. I = co) and k, = 2[1 - ( r , / a ) ] - ( d / a ) . Substitution for K in eqn (I), with 6 = k, I-;, where k , contains the double-layer parameters including constant yo, gives V, = V,-kI-i K = Kl - kl(6/a) where V, = V, + K, 6 is the elution volume when 6 = 0 ( I = co) and k = k, k , a-l.M. G. STYRING, C.J . DAVISON, C. PRICE AND C. BOOTH 3055 0.8 E, 0.7 0.6 I I I I J I 1 1 1 1 1 1 1 0 2 4 6 8 1 0 1 2 I-k/dmQ m o l 3 100 - 4- -0 E $ 60- E I \ 20- 1 I I I 1 I I I I 1 0 ’ 20 40 60 80 rc /nm Fig. 2. (a) Weight-average elution volume relative to that of benzyl alcohol (E,) against inverse root ionic strength (I”) of eluant for (0) latex 2, (a) latex 3 and (a) latex 7. (b) Negative slope (-k) of the plot of E, against 1-h against particle radius from electron microscopy ( 0 , / 2 x rc) for latices 1-7. Plots of 6 against I-; should be linear at high ionic strength, when 2[ 1 - (rc/u)] b (S/u). For particles of different size the slope of & against I-; should decrease as particle size is increased unless rc 3 a. Plots of E , (proportional to V,) against I-; are shown in fig.2(a). As mentioned earlier, a choice* may be made between &pk and E,, the difference between the two quantities being small (e.g. at I = 8.26 mmol dm-3, for latex 1 E , = 0.668 and &pk = 0.664, and for latex 7 E, = 0.581 and cpk = 0.565). The plots are linear with slopes (k) which decrease as the particle size of the latex is increased, in accordance with k = (a-rc)/k, where k, is a constant dependent on the system. A plot of k against rc is shown in fig. 2(b). It is assumed here and later that rc equals the radius obtained by electron * It may be noted that the present results can also be represented by plots of log,, D, against &pk without change of conclusions.3056 SIZE-EXCLUSION CHROMATOGRAPHY OF POLYSTYRENE LATICES 2.0- 1 .9- 1 .8 - 1 . 7 - 1.6- 1 . 5 - 1 .L- 0 9 P - ' 0:6 0'.7 0.8 EW Fig. 3. Calibration curves of logarithm of particle diameter (log,, Do, D, x 2r,) against weight-average elution volume relative to that of benzyl alcohol (E,) for latices 1-7 in eluant of ionic strength ( I ) of (0) 12.4, (a) 27.8 and (0) 110 mmol drnp3. microscopy, the small contribution of the adsorbed surfactant layer being neglected. The radius chosen is the harmonic mean, i.e. rc = iD,,, where Do = (DL D3,):, since this average correlates best with the weight-average elution The value of the effective cylindrical pore radius, a, obtained from the intercept in fig. 2(b) is ca. 90 nm. CORRELATION WITH PARTICLE SIZE The conventional method of presenting s.e.c. results is to plot the logarithm of particle size against the elution volume.A suitable plot for the present results is of log,, Do against E,. Selected plots are shown in fig 3. The results fit well to a set of straight lines, the slopes ( S ) of which become more negative as Z is decreased. This effect of ionic strength on S is as expected. The contribution to the effective exclusion radius (re) resulting from overlap of double layers (6) is not greatly dependent on particle core radius (rc). (This can be demonstrated by calculating the repulsive potential energy between a charged spherical particle and a similarly charged infinite platela and defining the effective exclusion radius by equating this potential energy to the thermal energy:19 e.g. a decrease in 6 of < 20% is predicted on changing the core radius from 45 to 15 nm when the surface charge density is 6 mC m-2 and the ionic strength is 10 mmol dm-3 at 298 K.) Consequently the effective exclusion radius relative to the core radius ( r e / r c ) is greater for a small particle than for a large one.PEAK BROADENING The s.e.c. peaks increase in breadth as the ionic strength of the eluant is increased; e.g. for latex 7 the peak half-width at 0.6 height (a) is 6.8 cm3 for Z = 8 mmol dm-3 and 8.4 cm3 for Z = 110 mmol dm-3. As can be seen in fig. 3 the resolution of the method, measured by the inverse slope of the calibration curve (i.e. S-l), is increased as Z is increased, because of the smaller contribution to exclusion from the double layer. In fact the resolution is an approximately linear function of I-;, as seen inM.G. STYRING, C. J. DAVISON, C. PRICE AND C. BOOTH 0.6 1 -s 0.5 i / 0 /O / 3057 ~ l , l l l l 1 l l l l ' 0 2 4 6 8 1 0 1 2 If/drn+ m o l a 0 2 4 6 8 1 0 1 2 la/drn*moI*' Fig. 4. (a) Negative slope (- S) of the calibration curve (fig. 3) against inverse root ionic strength (I-$) of eluant for latices 1-7. (b) Ratio of peak half-width (a) to resolution ( S l ) against inverse root ionic strength (1-1) of eluant for (a) Iatex 3 and (0) latex 7. fig. 4(a). If the increase in peak width is due entirely to the increase in resolution, then the ratio of peak width (0) to resolution (S-l) should be constant across the range of ionic strength investigated. This is found to be so : results for samples 7 and 3 are shown in fig. 4(b). CONCLUSIONS Provided experiments are carried out below a critical value of the ionic strength ( I ) , polystyrene latex particles will elute without adsorption through packed beds of porous glass beads at a volume characterised by their effective elution radius (re).The latter quantity is determined by the core radius of the latex particle plus a contribution from the electrical double layer. For our system with eluant of fixed I , a calibration plot of logarithm of particle size (determined by electron microscopy) against elution volume is linear. The slopes of such plots become less negative as I is increased, as predicted from the dependence of the double-layer thickness on I-$. This behaviour means that the resolution of the s.e.c. method increases as I is increased, and it is linked with the observed broadening of the s.e.c.peaks. Note that each calibration plot established for eluant of given I relates to a set of latex particles having identical surface potential. Such a calibration plot cannot be3058 SIZE-EXCLUSION CHROMATOGRAPHY OF POLYSTYRENE LATICES used directly to interpret elution volume obtained for latex particles of a different type. A reliable s.e.c method for determining the size of undefined latex particles requires elution volumes and calibration plots for eluant of infinite ionic strength (by an extrapolation procedure), thereby avoiding the effect of electrical interactions. We thank Mr D. J. Roy and Miss R. B. Stubbersfield for help with the experiments, one of the referees for kindly pointing out the relevance of ref.(1 8) and (19), and the S.E.R.C. for financial support. H. Small, J. Colloid Interface Sci., 1974, 48, 147; H. Small, F. L. Saunders and J. Solc, Ado. Colloid Interface Sci., 1976, 6, 237; H. Small, Adv. Chromatogr., 1977, 15, 113. V. K. F. Krebs and W. Wunderlich, Angew. Makromol. Chem., 1971,20, 203. A. Husain, A. E. Hamielec and J. Vlachopoulos, J. Liq. Chromatogr., 1981,4(52), 295; A. Penlidis, A. E. Hamielec and J. F. MacGregor, J. Liq. Chromatogr., 1983, 6(S-2), 179. G. D. McCann, E. B. Bradford, H. J. van den Hul and T. W. Vanderhoff, in Clean Surfaces: Their Preparation and Characterisation for Interfacial Studies, ed. G. Goldfinger (Dekker, New York, 1970). M. E. Woods, J. S. Dodge, I. M. Krieger and P. E. Pierce, Proc. Paint Res. Inst., 1968, 40, 527, 541. A. S. Dunn, personal communication. G. Lichti, R. G. Gilbert and D. H. Napper, J. Polym. Sci., Polym. Chem. Ed., 1983, 21, 269. B. V. Derjaguin and L. Landau, Acta Physicochim. URSS., 1941, 14, 633; E. J. W. Verwey and J. Th. G. Overbeek, The Theory of the Stability of Lyophobic Colloids (Elsevier, Amsterdam, 1948). E. J. Clayfield and E. C. Lumb, J. Colloid Interface Sci., 1966, 22, 269: F. Th. Hesselink, A. Vrij and J. Th. G. Overbeek, J. Phys. Chem., 1971, 75, 2094. lo C. Booth, J-L. Forget, I. Georgii, W. S. Li and C. Price, Eur. Polym. J., 1980, 16, 255. l1 A. K. Wolkoff and R. H. Larose, J. Chromatogr. Sci., 1976, 14, 51. l2 H. E. Pickett, M. R. J. Cantow and J. F. Johnson, J. Polym. Sci., Part C, 1968, 21, 67. l3 J. Stone-Masui and A. Watillon, J. Colloid Interface Sci., 1968, 28, 187. l4 Y. L. Wang, J. Colloid Interface Sci., 1970, 32, 633. l5 A. Husain, A. E. Hamielec and J. Vlachopoulos, J. Liq. Chromatogr., 1981, 4, 425. l6 E. F. Casassa, Sep. Sci., 1971, 6, 305. l7 C. van der Linden, Polymer, 1980, 21, 171. l 8 G. M. Bell, S. Levine and L. N. McCartney, J. Colloid Interface Sci., 1968, 33, 335. S. L. Brenner, J. Phys. Chem., 1976,80, 1473. (PAPER 4/267)

ISSN:0300-9599    

DOI:10.1039/F19848003051

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

f. Chem. SOC., Faraday Trans. I, 1984, 80, 3059-3070 Adsorption and Conductivity Studies in Oxychlorination Catalysis Part 3.-The Ethene-Transition-metal Chloride Interaction BY PETER G. HALL,* PHILIP HEAT ON^ AND DAVID R. ROSSEINSKY Department of Chemistry, University of Exeter, Stocker Road, Exeter EX4 4QD Received 16th February, 1984 Gas adsorption chromatography (g.a.c.) has been used to obtain isotherms and isosteric heats of adsorption of C,H, at temperatures < 150 "C on VCl,, CrCl,, CrCl,, MnCl,, FeCl,, CoCl,, NiCl,, CuCl, and CuCl. Physical adsorption is indicated for CrCl,, MnC1, and CuCl and chemisorption for NiCl, and CuC1,; the others occupy an intermediate position. The results are discussed in terms of d-electron configuration and ion charge/(radius)2 ratio.G.a.c. studies of C,H, on PdCl,, PtCl, and PtC1, indicate that chemical reaction occurs at room temperature. For those adsorbents for which heats of adsorption had been obtained, further reaction was initiated by increasing the temperature above 150 "C. Reaction products, e.g. HCl, vinyl chloride, dichloroethane and polymers, were identified, where possible, by mass spectrometry or chromatography. The capabilities of a microreactor to study the copper chloride-C,H, system have been evaluated. This paper reports the results of C,H, adsorption studies on a wide range of transition-metal chlorides (and MgCl,). The investigation was conducted in three main parts. (a) Low temperatures ( < 150 "C) : gas adsorption chromatography$ (g.a.c.) was used to obtain isotherms and isosteric heats of adsorption (qst) for those cases where there was no chemical change in the adsorbate.(b) High temperatures (> 150 O C ) : for those adsorbents for which qst values had been obtained, the solid was further investigated by increasing the temperature to initiate a reaction and identification of reaction products was by mass spectrometry or chromatography. (c) The capabilities of a microreactor to study the kinetics of the system copper chloride-C,H, were evaluated. EXPERIMENTAL GAS ADSORPTION CHROMATOGRAPHY The same apparatus as described in Part 2 of this series was used.] All adsorbents were anhydrous and of analytical-reagent grade (Fluka AG: CrCl,; B.D.H. : VCl,, CrCl,, CoCl,, NiCl,, PtCl,, PtCl, and PdCl,; ICN Pharmaceuticals : MnCl,, MgCl, and 99.999% FeCl,; CuCl, and CuCl were prepared as in Part 2).During weighing and column packing all materials were handled in a dry box over P,O,. Column resistance was dependent on particle size and for some solids it was found necessary to pre-mix the adsorbent with glass beads (ca. 100 mesh, B.D.H.). Ethene (CP Grade, BOC Special Gases) was the only adsorptive used and a sample flask was prepared as in Part 2. Adsorbents were preheated at 250 "C for 1 h. Elution peaks were analysed as described in Part 2 to obtain isosteric heats of adsorption and isotherms over the temperature range where there was no chemical reaction, typically 50-1 50 "C. .t Present address : Johnson Matthey Research Centre, Blounts Court, Sonning Common, Reading, Berkshire.t Aiso referred to as gas-solid chromatography. 30593060 OXY CHLORINATION CATALYSIS G. A. C. WITH MASS-SPECTROMETRY DETECTION, G. C . M .S. In g.a.c. the column exit gas passes directly into a flame-ionisation detector or katharometer detector. Replacement of these detectors by a mass spectrometer (Micromass 16F, Vacuum Generators Ltd) allowed the determination of whether there was any chemical change in the injected adsorbate, as well as qualitative identification of reaction products. THE MICROREACTOR The microreactor was constructed of 4 mm i.d. Pyrex glass and temperature was regulated by an Ether controller/metal-block furnace or a chromatography oven. The adsorptive was pulsed over the catalyst by syringe injection and the pulse was analysed by chromatography.The analysis of the products of the chlorination of ethene was performed by a 9 ftt long, 4 mm i.d. glass column of 15% MS 550 silicone oil on Chromosorb W.AW-DMCS 60-80 mesh (Phase Sep Ltd) at 80 "C, with a Pye 104 flame-ionisation or katharometer detector. For some of the results nitrogen (from a liquid-nitrogen reservoir) was used as the carrier gas with a water trap (type 5A, 30-60 mesh, molecular sieve) and an oxygen trap (manganese oxide, regenerated by heating at 375 "C in hydrogen). Samples were packed and weighed in air, while the ethene sample flask was prepared by thoroughly flushing with nitrogen before injecting a large volume of C2H4 into the flask. This was not as satisfactory as flushing with pure C,H4 because of the impurities in the nitrogen.In other cases, helium (grade A, BOC Special Gases) with a drying trap of molecular sieve (type 5A, B.D.H.) was used as the carrier gas. Samples were packed and weighed in a dry box over P20,, while an adsorptive sample flask was prepared as in Part 2. It was found that the % yield, defined by moles of EDC moles of (C2H4 + EDC) % yield = x 100% where EDC is dichloroethane, varied slightly with the dose size and therefore a constant dose size of 1 cm3 was used. In the calculation of contact time, half of the volume occupied by the catalyst was assumed to be void space. Thus contact time was defined as volume occupied by catalyst x 0.5 flowrate contact time = RESULTS AND DISCUSSION G.A.C. RESULTS FOR FIRST-ROW TRANSITION-METAL CHLORIDES In order to evaluate the importance of d electrons to the bonding of ethene, adsorbents of varying d-electron configuration were investigated. However, control experiments were first performed using magnesium chloride as an example of a non-transition-metal chloride and CuF,, CuCl, and CuBr, to investigate the effect of the anion.Elution peaks were narrow and symmetric for MgC1,-C,H, and following the method described in Part 2 a plot of log (net retention time) against reciprocal temperature corresponded to an isosteric heat of adsorption at zero coverage (qo) of 26 kJ mol-1 (see fig. 1). The only other adsorbent without delectrons to be successfully studied (Part 2) was alumina, A1,0,, for which qo was 26 kJ mol-l. When the cupric salt was changed from the anhydrous chloride to the bromide, after thermal activation at 250 "C for 1 h (a standard procedure for all adsorbents), it was observed by mass spectrometry that some direct bromination and fluorination occurred with ethene at room temperature.Without thermal activation there was no adsorbate-adsorbent interaction. Thus isosteric heats could not be obtained but it can be concluded that changing the anion effects adsorption considerably,2 in part because 7 The assistance of R. Bailey in operating the mass spectrometer is gratefully acknowledged. $ 1 ft = 0.3048 m.P. G. HALL, P. HEATON AND D. R. ROSSEINSKY 306 1 2.: - E - 1 . 1 - 0 .! 0 2.75 3.0 3.25 lo3 KIT Fig. 1. Plot of retention time (tk) against T-' for C,H,-MgC1,. of the differing electronegativity, structural geometry and ion size.Henceforward, only the anhydrous chlorides were studied in order to eliminate this effect and because oxychlorination is really the process of interest. The results obtained for a series of first-row transition-metal anhydrous chlorides are summarised in table 1. Because of the complicating effect of metal-ion charge, adsorbents were restricted when possible to the +2 valence state. For all the adsorbents, elution peaks were asymmetric, necessitating the construction of isotherms and isosteres as shown for NiCl, and MnCl, in fig. 2 and 3. The parameter n,, the amount of C,H, adsorbed per gram of adsorbent, is readily converted to surface concentration r by dividing by surface area. Assuming an effective cross-sectional area of 2 x m2 for the ethene molecule also allows the fraction of surface coverage 8 to be estimated; surface coverage fractions were all generally < 0.01.Specific surface areas by one-point B.E.T. gravimetric analysis (carried out by Winnington Analytical Laboratory, I.C.I.) were generally low, in the range from ca. 2 to 17 m2 8-l. From table 1 it is evident by comparison with the heat of liquefaction for ethene (14.45 kJ mol-l) that for CrCl,, MnC1, and CuCl there is only physical adsorption of C,H,. Chemisorption seems to be occurring for NiCl, and CuCl,. The adsorption of ethene on VCl,, FeCl,, CrCl, and CoCl, is not readily categorised as physical or chemical, being intermediate in magnitude.3062 OXY CHLORINATION CATALYSIS Table 1. Isosteric heats of adsorption of C,H, on anhydrous chloride adsorbents moles of C,H, isosteric heat adsorbed per gram of adsorption of adsorbent, adsorbent qst/kJ mol-1 r ~ , / l O - ~ mol vc1, CrCI, CrCl, MnCl, FeCI, COCl, NiCl, CUCl, CUCl 29 27 25 42.5 44 26.5 24.5 23.5 13.5 13 12 1 1 1 1 35.5 34 33 25 20.5 19 52 52.5 53 53.5 54 58 50 44 44 43 13 13 13 0.5 0.8 1.2 1 .o 1.5 10 20 30 0.2 0.35 0.2 0.3 0.4 0.2 0.3 0.4 0.15 0.25 0.35 0.3 0.6 0.9 1.2 1.5 12 29 58.5 88 117 1.3' 1.95 2.6 G.C.M.S.RESULTS FOR PALLADIUM(II) AND PLATINUM(II, IV) CHLORIDES G.a.c. indicated that for these adsorbents chemical reaction between the surface and the ethene was occurring at room temperature. G.c.m.s. was used to confirm this and identify the products. These solids are of particular interest because of their wide usage as industrial catalysts with alkenes.For PdCl, the mass spectrum for the product (shown in fig. 4) is dominated, excluding the background air spectrum, by a broad peak between mass numbers 21 and 26. This is due to decomposition of a metastable ion, i.e. an ion produced in the ion source is unstable and decomposes to a resultant ion of lower mass before detection. However, the original ion may be identified by the position of this broad20 16 ’ij 12 E 2 a 1 La ‘ 8 4 0 P. G . HALL, P. HEATON AND D. R. ROSSEINSKY 3063 57.3 65.3 69 2.75 2.87 2.99 CzH4 partial pressure/ 1 O-’ atm lo3 K/T Fig. 2. (a) Isotherms for the adsorption of C,H, on NiC1, at T = 57.3-102 “C (marked on the curves). (b) Isosteres for the adsorption of C,H, on NiC1, at n, = (3-15) x mol (marked on the curves).- 8 . 7 -9.3 -9.9 2.74 2.86 2.98 3.1 lo3 KIT Fig. 3. (a) Isotherms for the adsorption of C,H, on MnCl, at T = 43.8-99 “C (marked on the curves). (b) Isosteres for the adsorption of C,H, on MnCI, at n, = (2-4) x lo-@ mol (marked on the curves).3064 OXYCHLORINATION CATALYSIS mass number C0 Fig. 4. Mass spectrum of the products of reaction between C,H, and PdCl, at room temperature. Table 2. Possible combinations causing a metastable peak position of original resultant metastable ion ion peak [C4H7ClI+ cyclobutane cyclo butene cyclo bu tene [C4H7I+ :cc11+ 24.54 :C3HI+ 24.45 GI+ 24 :C&I+ 25.35 :cm+ 24.89 peak as it is equal to the (mass of resultant ion),/rnass of original ion. A C, cyclic molecule would be expected to be under strain, with any unsaturated character adding to the tendency of the molecule to undergo decomposition/rearrangment. Inspection of fig.4 shows the presence of vinyl chloride (VC) and thus the possibility of chlorine cannot be discounted. In table 2 some possible original/resultant combinations are suggested. They all give good agreement with the experimentally observed maximum at 24.7. Other worker~~-~ have studied the decomposition in solution of (PdCl,C,H,), and reported linear butene formation. Mechanisms involving the coordination structuress have been propo~ed.~-~* ' 7 For PtCl, the mass spectrum for the product desorbing from the surface after reaction with CzH4 at room temperature did not show any VC but again featured a metastable peak at mass numbers 23-25. Use of lower electric-field potentials again failed to avoid decomposition of the parent ion.The PtCl, used was itself analysed after injections of C,H, by placing it on a silicaP. G. HALL, P. HEATON AND D. R. ROSSEINSKY 3065 probe and heating in the ion source at up to ca. 200 "C. The resulting spectrum was that of a polymer hydrocarbon with chain length reaching ca. C30. The spectrum was free of (i) any chlorine or isotope splitting patterns, (ii) metastable peaks and (iii) any Pt signal. Clearly PtCl, would not be of use under these conditions where it has been reported that PtCl, does not catalyse the dimerisation of C,H,3 and decomposes on heating above 180 "C to give chlorinated C, hydrocarbon^.^ However, other workers have recorded dimerisation and further polymerisati~n.~? s t Platinum(1v) chloride was shown by g.c.m.s.to chlorinate ethene at room tempera- ture. VC, EDC, hydrogen chloride and C,H,CI (where n = 1-4 and x = 0-8) ions were observed. As the reaction temperature was increased C,HuCl,, C,H, and CH, were additionally formed at 70 "C. At 140 "C further products such as C,Cl,H, and C,Cl,H, were seen. It is concluded that the mechanism of interaction between C,H, and PtCl, is not like that for PtCl,, but is a substitution/radical mechanism. RESULTS FOR THE IDENTIFICATION OF REACTION PRODUCTS FROM ETHENE AND FIRST-ROW TRANSITION-METAL CHLORIDES? Two methods of identification were used: ( a ) the pulse microreactor operated with a chromatographic separating column and (b) g.c.m.s. The first method was least satisfactory because hydrogen chloride created spurious peaks by reaction with the separating column and caused tailing by attacking the katharometer detector.However, g.c.m.s. has the disadvantage that only reaction temperatures < 200 "C could be studied. The results obtained by methods ( a ) and (b) are given in tables 3 and 4, respectively. The two techniques show that: (i) VCl, is very volatile at high temperature, which makes product identification difficult, and at 250 "C some vinyl chloride (VC) was observed but it is not known if 1,2-dichloroethane (EDC) was formed, (ii) CrCl, forms VC and EDC at 250 "C, but a hydrocarbon polymer (non-chlorine-containing) was also formed, (iii) there is no reaction between NiC1, and C2H4 at < 350 "C and (iv) CuC1, predominantly forms EDC above 150 "C with little VC until above 300 OC, although a mass spectrum which has been completely analysed15 showed the chlorine isotope splitting patterns, clearly identifying the products.MICROREACTOR RESULTS FOR COPPER(I, 11) CHLORIDE-ETHENE The effectiveness of the microreactor was found to be dependent on the precision of the temperature thermostatting and the absence of oxygen. When the catalyst temperature was regulated by an Ether controller and block heater, the temperature oscillated by 3 "C with gradients inside the sample. Direct chlorination of C2H4 by CuCl, was very temperature sensitive between 230 and 270 "C, the yield trebling every 20 "C. Thus for isothermal experiments scatter in the results was observed. At 250 "C, EDC was the only major product from injections of ethene over ca.0.5 g of copper(1, 11) chloride. The yield was dependent on [Cu2+] and never greater than 2,% for a contact time of ca. 0.2 s. The small yield represents an advantage of the pulse mode of operation since experiments can be carried out at an effectively constant solid composition. The yield of EDC from copper(1) chloride (containing small amounts of Cu2+) was very sensitive to oxygen. There were experimental problems in excluding oxygen totally, particularly from the sample flask and the results seemed dependent on the success in excluding it. The effect of pretreating copper(I1) chloride with HCl+O, and 0, at 250 "C to remove any Cu+ was investigated. The initial yield of EDC after pretreatment f' The assistance of N.Warrender is gratefully acknowledged.3066 OXY CHLORINATION CATALYSIS Table 3. Identification of reaction products from anhydrous metal chlorides and ethene by use of a microreactor with chromatographic detection reaction products reaction solid temp./"C vc EDC other NiC1, cuc1, VCl, 250 300 CrC1, 250 300 350 250 300 350 50 100 150 200 250 300 350 400 J unknown J J J unknown J unknown - J J - J J J J J - J J J J - J Table 4. Identification of reaction products from anhydrous metal chlorides and ethene by use of g.c.m.s. reaction products reaction solid temp./"C HCl vc EDC other 40 J no no solid decomposition - - .- - CrCl, 35 J - - polymer 1 00 CUCI, - - 50 - - - - - 110 J 150 J J J 210 J J J - - generally was found to be smaller and increased with increasing number of ethene injections, whereas the overall tendency was for the yield to decrease slightly as [Cu2+] fell, provided no regenerative pathways were available.The selectivity of the chlorination reaction in producing EDC was always high ( > 95 % ) but occasionally other products were observed : (i) vinyl chloride, presumably from the thermal cracking of EDC [this also suggests that hydrogen chloride must be present (not detected by f.i.d.)] and (ii) methylene chloride, possibly fromP. G. HALL, P. HEATON AND D. R. ROSSEINSKY 3067 I I I I I 0.2 0.4 0.6 0.8 tls Fig. 5. First-order kinetics plot for C,H4-CuC1, at 250 "C. 2CuO - CuC1, + C,H, -+ ~ C U , + 2CH2C12 as this is consistent with a pink/red coloration of the solid afterwards. When the thermostatting of the microreactor was controlled by an oven, isothermal experiments were possible to an accuracy of k0.5 "C.Results obtained were reproducible to within 8% and devoid of significant scatter. The first property of the CuCl,-C,H, system to be examined was the change in yield of EDC with contact time. Variation of the flowrate over the catalyst effectively changes the extent of reaction and therefore reaction order can be determined. The first-order plot of In [a/(a-x)] against t at 250 "C (shown in fig. 5), where a is the number of moles of reactant that undergo reaction in time t , showed that the rate constant was 4 x lo-, s-l, indicating that the chlorination of C2H4 to EDC at 250 "C is a fast process. Freshly prepared copper(1) chloride gave no detectable yield of any product with C2H4 at 250 "C.The change in catalyst activity with age is shown in fig. 6. It appears that the activity stabilises after 6 h. A sample of CuCl, heated at 250 "C for 40 h showed a decrease in surface area? from 3.2 to 2.9 m2 g-l. The effect of sintering would therefore appear not to be important. However, heating will also cause some disproportionation of cu2+. Further kinetic work has been carried outf with at least 6 h being left for activity to stabilise before injections of C,H, were made. The variation of yield with temperature was established15 and yield against contact time data were collected at several temperatures. The activation energy for the chlorination of C2H4 to EDC was determined to be 40 4 kJ mol-l. -f Determined by gravimetric N, adsorption by Winnington Analytical Laboratory, I.C.I.1 The assistance of N. Warrender is gratefully acknowledged.3068 OXYCHLORINATION CATALYSIS 24 n E 8 18 3 w rcl .- x 12 I I I I 0 6 12 18 24 tlh Fig. 6. Change in catalyst (CuC1,) activity with age at 250 "C. DISCUSSION A range of reaction products has been identified (1,2-dichloroethane, vinyl chloride, dimers and hydrocarbon polymers), but no catalyst has been seen to rival CuCl, in the selective formation of 1,2-dichloroethane. The microreactor has demonstrated its potential for qualitative and quantitative work. However, results seem to depend on experimental conditions, e.g. method of thermostatic control, degree of oxygen exclusion, sample pre-treatment and dose size.THE ELECTRON CONFIGURATION FOR FIRST-ROW TRANSITION-METAL CHLORIDES The results described above compare the strength of the M-C,H, bond, as reflected by the isosteric heat of adsorption qst, for a range of transition metals M. Weak bond strengths are not readily measured and therefore the heat of adsorption is a valuable guide. To simplify the data of table 1, qst values were extrapolated to the zero coverage value qo, with the uncertainty in qo being a function of the variation of the measured (qst, n,) values in table 1. Fig. 7 shows the variation of qo with the number of delectrons and indicates that the bonding of C,H, is not favoured by d4, d5 and d10 configurations. This may be explained by the molecular-orbital bonding approach as follows: (i) a d5 high-spin octahedral metal ion will have no completely filled orbitals to participate in back-bonding into the empty n*-antibonding orbitals of ethene, nor any completely empty orbitals to accept the n electrons of ethene, (ii) a d4 Jahn- Teller-distorted high-spin ion will have no completely filled orbitals for back-bonding, nor appreciable a-bonding as the dz2 orbital is partially filled, and (iii) a d10 molecule will have complete back-bonding but the fully occupied orbitals will not favour o-bonding to C,H, by acceptance of n electrons.However, this argument does not explain the large change in qo between d9 and dlO.P. G. HALL, P. HEATON AND D. R. ROSSEINSKY 3069 60 20 I -AgH OL I I I I I I I I 1 I d o d' d 2 d 3 d' d 5 d 6 d7 d s d9 d'' no. of d electrons Fig.7. Isosteric heats of adsorption of C,H, on transition-metal chlorides: effect of d-electron configuration. (AkH is the heat of liquefaction of C,H,.) 60 20 0 2 4 6 8 charge/(radius)2 (arb. units) Fig. 8. Isosteric heats of adsorption of C,H, on transition metal chlorides: effect of charge/(radius)2. In fact d'O-ethene complexes are often stable because of the electronic configuration, e.g. [Ag(C,H,)]+ with a-bonding to the unoccupied 5s orbital; the bonding molecular orbitals allow a variety of ethene-transition-metal-ion orientations.16 THE ION-POLARISING EFFECT The g.a.c. results in Part 2 established that the unsaturated nature and therefore the higher polarisability of the adsorbate was important. For a constant adsorbate, i.e. ethene, it is possible to understand the results of fig. 7 further using the ion-polarising effect in terms of the ratio of the ionic charge/(ion radius)2.3070 OXYCHLORINATION CATALYSIS Assuming for representation purposes only a physical interaction, fig.8 shows qo plotted against e/r2 for the first-row transition-metal chlorides (and MgCl,): it is seen that the majority of them lie fairly close to a common line through the origin, presumably corresponding to the polarisation effect on C,H,. Cu+ lies above the line because the van der Waals forces probably do not permit q,, to fall much below the heat of liquefaction (14.45 kJ mol-l). Mn2+ lies below the line for reasons explained earlier. Since the value of qo for CoCl, was 30 _+ 5 kJ mol-l, Co2+ may well lie close to the line. Good agreement is seen for Mg2+, Cr2+, V3+ and Cr3+, thus only for FeCl,, NiCl, and CuCl, are other factors altering q,,, such as electronic configurations allowing for back-bonding.PdCl,, PtCl, and PtCl, gave q,, values > 65 kJ mol-l and therefore have a greater interaction with C,H, than can be accounted for by the polarisation effect. These salts are known to have different crystal structures than those of the other chlorides, which suggests that structure may be a major factor. Financial assistance for this work was provided by an S.E.R.C./Imperial Chemical Industries P.L.C. (Runcorn) CASE award. We also thank Dr R. A. Hann (I.C.I.) and Dr J. Wolstenholme (formerly I.C.I.) for helpful discussions. P. G. Hall, P. Heaton and D. R. Rosseinsky, J. Chem. Soc., Faraday Trans. I , 1984, 80, 2785. E. R. Gilliland, J. E. Seebold, J. R. Fitzhugh and P. S. Morgan, J. Am. Chem. SOC., 1960, 61, 1939; J. S. Anderson, J. Chem. SOC., 1934, 971. J. T. van Gemert and P. R. Wilkinson, J. Phys. Chem., 1964, 68, 645. A. D. Ketley, L. P. Fisher, A. J. Berlin, C. R. Morgan, E. H. Gorman and T. R. Steadman, Znorg. Chem., 1967,6, 657. R. G. Schultz and D. E. Gross, Adv. Chem. Ser., 1968, 70, 97. J. N. Dempsey and N. C. Baenziger, J. Am. Chem. SOC., 1955, 77, 4984. ' C. M. Harris and S. E. Livingstone, Rev. Pure Appl. Chem., 1962, 12, 16. Y. Kusunoki, R. Katsuno, N. Hasegawa, S. Kurematsu, Y. Nagao, K. Ishii and S. Tsutsumi, Bull. Chem. SOC. Jpn, 1966, 39, 2021. A. S. Gow and H. Heinemann, J. Phys. Chem., 1960,64, 1574. R. S. Berger and E. A. Youngman, J. Polym Sci., Part A, 1964, 2, 357. lo J. Smidt, W. Hafner and J. Sedlmeier, British Patent 887362, 1963; Chem. Abs., 1963, 58, 3521e. l2 N. Phung and G. Lefebvre, C.R.Acad. Sci., Ser. C, 1967, 265, 519. l 3 F. N. Jones, J. Org. Chem., 1967, 32, 1667. l4 H. C. Volger and K. Vrieze, J. Organomet. Chem., 1968, 13, 495. l5 N. Warrender, internal report, Exeter University, 1981. l6 J. W. Moore, Acta Chem. Scand., 1966, 20, 11 54. (PAPER 4/269)

ISSN:0300-9599    

DOI:10.1039/F19848003059

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

,I. Chem. SOC., Faraday Trans. I, 1984, 80, 3071-3094 Cobalt-59 Nuclear Magnetic Resonance Studies of the Kinetics and Mechanism of Deuteration for Tris(ethy1enediamine) Cobalt Chloride BY ROBIN K. HARRIS* AND ROBERT J. MORROW School of Chemical Sciences, University of East Anglia, Norwich NR4 7TJ Received 17th February, 1984 Dissolution of [Co(en),]Cl, in D,O gives spectra which are time-dependent because of the progressive replacement of protons by deuterons. The isotopic chemical shift is 5.0 ppm per 'H --+ 2H substitution and all thirteen isotopomers can be detected, even using a low-field spectrometer. Isotopic distributions have been obtained by bandshape-fitting. The exchange rate has been measured under a variety of conditions and is found to depend on the [1H]/[2H] concentration ratio and on the pH.A simple model for the exchange is proposed and tested. An isotopic effect influences the rate constants. Four separate values for rate constants have been extracted from the data. These values, which range from 8 x lo6 to 1 x lo6 dm3 mol-l s-l, indicate that the rate of removal of D by OD- > H by OD-. > H by OH- > D by OH-. Cobalt-59 has a natural abundance of 100% and a receptivity relative to carbon of 1572, making it one of the more readily detectable nuc1ei.l It has a nuclear spin I = 7/2 and a quadrupole moment of 0.4 x m2, which causes the spectral lines of species of low symmetry to be very broad (linewidths in solution > 30 kHz are known), while octahedrally symmetric species may have linewidths < 100 Hz. The chemical-shift range of the nucleus is large, extending over 18 000 ppm, corresonding to 0.43 MHz in a field of 2.35 T.As a corollary, small structural vkriations within i i molecule may give rise to drastic spectral changes, such as a large displacement in signal frequency and, in cases of departure from octahedral symmetry, a marked increase in signal linewidth. A case in point is that of isotopic replacement at a position remote from the cobalt nucleus under study. Of course, deuterium-induced frequency shifts are generally larger and easier to observe than isotope shifts induced by other nuclei because the fractional mass change is greatest in the substitution of lH by 2H (excepting ,H). Very large deuterium-induced isotope shifts of ca. 5 ppm per deuterium atom substitution were reported by Bendall and Doddrel12 to occur for 59C0 n.m.r. of aqueous solutions of [Co(NH,),]Cl, and [Co(en),]Cl,.These shifts were also noticed at about the same time by Burton and Harris., The object of the present paper is to show that this isotope shift can be used to extract kinetic and thermodynamic information on H/D exchange between cobalt(II1) compounds and the solvent. Earlier research into such H/D exchange has used i.r. and lH n.m.r. to measure exchange rates in acidic solutions of pure H,O and pure I>,0.4-10 The studies reported here have used 59C0 n.m.r. to determine the H/D exchange rates in solutions of [Co(en),]Cl, - 3H20 in H 2 0 + D20 mixtures of variable acidity and variable deuterium content. At ambient temperature the exchange takes of the order of hours or days and so can be readily monitored by 5 9 C ~ n.m.r.307 13072 5 9 C ~ N.M.R. STUDIES OF DEUTERATION OF [Co(en),]Cl, Table 1. Compositions of all the solutions used in the kinetics and equilibrium experimentsa 1 2d 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1 9e 20 0.08 0.08 0.09 0.09 0.09 0.09 0.09 0.09 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08 - - 0.05 0.09 0.07 0.07 0.05 0.05 0.10 0.10 - 76.4 53.6 78.5 23.0 11.5 64.8 44.6 78.3 90.2 81.5 77.5 76.0 81.5 75.8 82.0 75.7 75.3 1.4 107.2 1.4 32.1 53.4 29.0 85.5 95.9 43.3 63.6 29.8 17.7 26.9 29.5 30.0 26.9 29.3 24.0 32.4 32.6 107.1 1.1 107.1 1.1 1 2.02 2.03 2.15 2.20 2.25 2.30 2.50 2.55 3.03 3.45 3.69 3.80 4.02 4.51 4.85 5.90 3.34 6.02 3.80 a The concentrations are expressed as mol dmP3.Some solutions were buffered with potassium hydrogen phthalate (PHP). pH,, refers to the pH meter reading of a solution at ambient temperature (ca. 25 "C) after the meter was calibrated with buffers dissolved in pure H20 to give standard buffer solutions of pH 4 and pH 7. The pH,, of each solution was read after the completion of each kinetics experiment. The initial isotopomer concentrations in solution 2 were 0.035, 0.002, 0.01 and 0.025 mol dm-3 for Do, DlO, D,, and D,,, respectively. The initial isotopomer concentrations in solution 19 were ca. 0.02 and ca. 0.06 mol dmP3 for D,, and D12, respectively. Clearly there are thirteen possible isotopomers of [Co(en),]Cl, with differing degrees of H/D substitution at the nitrogens of the ethylenediamine groups. For convenience, the isotopomer containing i deuterium atoms, viz.[Co(en), - diI3+, will be designated Di. Of course, many of the isotopomers will occur in a variety of stereochemical forms, but these will be ignored in the discussion here. EXPERIMENTAL The solvent for each sample was prepared by mixing weighed amounts of H,O and dilute hydrochloric acid with weighed amounts of D20. Some samples were buffered with potassium hydrogen phthalate (PHP) and hydrochloric acid or sodium hydroxide solution. For each kinetic measurement a 5.0 cm3 portion of the freshly prepared solvent sample was pipetted into an 18.0 mm 0.d. n.m.r. tube and the tube was then placed in the spectrometer probe, being left there for 30min in order to warm the sample to the ambient probe temperature. An accurately weighed amount, ca.0.16 g, of [Co(en), - do]C13 - 3H20 was quickly added to and dissolved in each 5.0 cm3 sample. The compositions of all the solutions studied are listed in table 1. The time at which the complex dissolved was taken as zero time and 59C0 spectra were recorded at subsequent intervals. A mercury thermometer set in an 18 mm 0.d. n.m.r. tube containing ethylene glycol was used to measure the sample temperature before and after eachR. K. HARRIS AND R. J. MORROW 3073 n 500 Hz P Fig. 1. Evolution of the cobalt-59 spectra from a 0.08 mol dm-3 solution of [Co(en),-d,]Cl, dissolved in a buffered mixture of 76 mol dm-, H,O and 30 mol dm-, D,O (solution 12, see table 1). The time, in min, after dissolution before the signal was recorded is given beside each spectrum.kinetic run. The sample temperature could be altered by a few degrees by adjusting the flow rate of the sample air supply. Every experiment was performed at a temperature of 33 f 2 "C. The 59C0 spectra were recorded using a Varian XL-100-15 spectrometer and a Nicolet MONA accessory tuned to 23.91 1 MHz, with the observe offset dialled to 32002 Hz. The 19F external (second-sample) lock was used throughout. The proton decoupler was not employed as the heat generated in H,O solutions can set up convection currents in 18 mm 0.d. tubes. The optimum conditions for spectral accumulation were found to be a 0.0125 s acquisition time (256 points) with a spectral width of 10 kHz to high frequency of the carrier.Accumulation of 500 transients 100 FAR 13074 5 9 C ~ N.M.R. STUDIES OF DEUTERATION OF [Co(en),]Cl, 1380 Hz - Fig. 2. Equilibrium W o signals from 0.1 mol dm-3 solutions of [Co(en>,-d,]Cl, in 100% H,O and in 99.8% D,O. using 90" sampling pulses, without any delay other than the acquisition time, gave an adequate signal-to-noise ratio of > 20 to 1 in most cases. The memory was zero-filled to 8 192 points before Fourier transformation. The 90" pulse was 26 ,us, but the r.f. magnetic field B, suffered from considerable inhomogeneity . For the kinetics experiments a time interval between sets of data acquisitions was introduced into the standard FT/16K99410-D program. Another minor alteration to the program enabled up to 32 free induction decays to be recorded for any one kinetic run.Selected free induction decays could then be transformed and plotted using a slightly modified version of the program1' OCTAFID. The listings of these patches are available on request.12 Some reactions were recorded over a period of several days, the longest time being one week. In these cases the solutions were, to facilitate other kinetic measurements, removed from the spectrometer probe and placed in a thermostatted water bath set to the same temperature as the probe, 33 "C. The pH of each solution at or near equilibrium was measured using a WG Pye model 290 pH meter. Some experiments used mixtures of D,,, D,, and Dlo. This mixture was prepared by dissolving 1 g of the fully protonated isotopomer in 5 cm3 of D,O to give a near neutral solution, ca.pD 6. At this pD deuteration proceeds rapidly. After one hour most of the D,O had boiled off. Another 5 cm3 of D,O was added and the process repeated. The residue was collected and dried in a vacuum desiccator. RESULTS AND DISCUSSION OBSERVATIONS OF THE 5 9 C ~ SPECTRA IN THE REGION pH < 7 A typical example of the evolution with time of a single-peak spectrum of Do into a multiplet is shown in fig. 1. The molar ratio of H,O to D,O is 2.5 : 1 in this case and up to 9 peaks (corresponding to the Do-D, isotopomers) appear. Fig. 2 compares the equilibrium spectrum of a Do solution, 0.1 mol dm-, in 99.8 % D,O, with the same concentration in 100% H,O. The 5 9 C ~ peak in D,O is shown to be shifted 1380 Hz or 58 ppm to lower frequency and to have a small shoulder on its high-frequency side, partly originating from the protons of the initial Do isotopomer.Both spectra were obtained using the 19F external lock, one tube being replaced by another in the probe. To obtain an in situ comparison the spectrum of a different solution was taken. In this a mixture of Do and D,, (lesser amounts of D,, and D,, were present) was dissolved in a 1 : 1 molar ratio of H,O + D,O (solution 2). Fig. 3 is a record of the spectrum immediately after preparation and shows a Do-D,, shift of 1450 Hz? or t This in sztu value is more accurate than the one given in fig. 2 because the latter is obtained from spectra measured at different times (and, possibly, slightly different temperatures).R. K. HARRIS AND R. J.MORROW 3075 A 4 9 . 3 Fig. 3. Cobalt-59 spectra of a solution of Do and D,, in a 1 : 1 mixture of H,O and D,O (solution 2., see table 1) as a function of time after preparation (given in h beside each spectrum). The spectra record the formation of the other eleven isotopomers. 100-23076 5 9 C ~ N.M.R. STUDIES OF DEUTERATION OF [Co(en),]Cl, Fig. 4. Four typical examples of LORENTZA simulations of the 59C0 kinetic spectra of [Co(en),JCl,. In these examples the simulations are plotted over the experimental kinetic spectra of (a) solution 5, 18.3 h; (b) solution 11, 3.4 h; (c) solution 10, 4.0 h; (d) solution 18, 6.67 h. The different spectra are not aligned vertically.R. K. HARRIS AND R. J. MORROW 3077 60.7 ppm. The adjacent spectra were recorded at intervals along the approach to equilibrium.At an intermediate stage all isotopomers are present in quantities sufficient for detection, as shown by the 13 peaks visible in some of the spectra. At 23.9 MHz all peaks in any multiplet overlap. Each peak is invariably separated by 120 & 2 Hz from its nearest neighbours, while linewidths are 97 5 Hz in solutions without PHP. The degree of overlap rules out the obtaining of isotopomer amounts directly from spectral peak heights. In order to extract the required information, the computer program LORENTZA was written.', This program can simulate any experimental spectrum of up to 24 Lorentzian lines of differing linewidths and separations. It uses the same data stores and plotting facilities as those used for XL-100-15 experimental spectra and enables the true peak intensities of a large number of spectra to be readily extracted.Fig. 4 illustrates the typical degree of rnatching between the experimental and LoRENTzA-simulated spectra. Table 2 lists the percentage concentration of each isotopomer present at various times after preparation for solutions 1-5. These concentrations have been extracted from the experimental spectra using LORENTZA and are quoted in molar percentages of the total [Co(en),]Cl, concentration. Values for the other solutions have been compiled.12 Successive additions of potassium hydrogen phthalate (PHP) to a neutral solution of [Co(en),-d,]Cl, in H,O shifted the 5 9 C ~ signal to lower frequency and increased its linewidth. Table 3 lists the magnitude of these changes.To ascertain whether buffering with potassium hydrogen phthalate and sodium hydroxide or hydrochloric acid solution had any effect on the H/D exchange rate some solutions prepared for the kinetics experiments were left unbuffered. A comparison between the measured H/D exchange rates in the buffered and unbuffered solutions showed that buffering at pH < 7 had no noticeable effect on the H/D exchange. ANALYSIS OF THE KINETICS Isotopomers with the same number of deuterons, differing only in the positions in which these are arranged, give single peaks, 5 9 C ~ n.m.r. at 23.9 MHz being unable to resolve the resonances in which the deuterons are geminal (-ND,) or further away. If position shifts exist they are at least an order of magnitude lower than the shifts arising from an increase in the number of deuterons.Consequently H/D exchange was analysed without regard to the neighbours, H or D, adjacent to the atom undergoing exchange; that is, no distinction was made between the rate at which a proton exchanges when adjacent to another proton (-NH,) and the rate when adjacent to a deuteron (-NHD). These exchange rates apparently have rate constants whose values lie too close for the techniques described here to distinguish between them, and thus each isotopomer is regarded as having essentially a single rate constant for, say, deuteration. This is the basic simplification used to form a set of rate equations that describe the full exchange process. It is also considered that each substitution step involves the exchange of one deuteron or proton. and so the exchange can be represented by Furthermore, the rate at which each isotopomer exchanges protons for deuterons will be proportional to the number of sites available for substitution by deuterons, and this of course corresponds to the number of exchangeable protons in the molecule.Exchange in the other direction, substitution of deuterons for protons, will occur at a rate dependent on the number of deuterons present. Finally, the rates of substitutionw 0 4 00 Table 2. Isotopomer concentrations for solutions 1-5 (see table 1) as a function of time, obtained using the program LORENTZA (see text) molar percentage of each isotopomer solution time Do D, D, D, D, D5 D, D, D8 /days 1 0.3 0.6 0.7 1.5 3.6 4.7 6.0 6.7 /hours 2 0.0 5.5 10.5 14.2 21.0 31.2 38.0 53.5 61 .O 94.2 87.6 86.6 73.5 47.9 38.9 33.5 29.7 50.0 20.2 11.0 8.6 3.3 1.9 1.3 0.0 0.0 5.7 12.3 13.3 24.2 36.8 38.3 38.4 38.7 0.0 19.4 18.1 15.6 10.1 6.2 4.4 2.3 1.5 0.0 0.0 0.0 2.2 12.4 18.2 21.5 22.8 0.0 8.2 13.9 15.1 14.2 11.0 9.0 5.5 4.0 0.0 0.0 0.0 0.0 2.7 4.1 5.6 7.5 0.0 2.2 6.3 8.6 12.5 13.2 12.3 9.8 8.6 0.0 0.0 0.0 0.0 0.0 0.2 0.7 1.1 0.0 0.0 0.6 2.6 6.8 10.1 11.5 11.4 12.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.7 5.4 7.8 11.5 13.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.9 2.6 4.8 10.9 13.9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.8 3.8 5.7 11.0 13.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.5 2.7 5.8 7.9 9.4 11.4 13.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.9 2.5 3.2 9.6 7.6 13.8 8.0 14.3 11.1 14.4 12.3 12.8 12.4 11.3 11.3 8.7 10.8 5.9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 13.5 19.1 17.0 15.5 11.4 9.0 7.1 4.5 2.6 Dl 2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 33.1 18.0 9.1 9.0 4.9 3.8 3.0 1.7 0.5/hours 3 3.7 7.2 12.0 12.5 14.7 16.2 18.0 24.5 35.5 /hours 4 2.16 4.14 6.12 7.74 12.06 18.24 23.94 29.28 /hours 5 0.2 2.2 4.2 6.0 12.0 20.4 24.2 73.7 63.1 47.4 47.1 43.6 39.4 35.6 25.1 16.3 56.5 28.4 18.3 13.8 5.9 2.7 1.6 1.3 95.2 50.5 26.1 15.6 4.3 2.3 1.3 21.0 30.1 35.4 35.8 36.7 37.3 38.5 35.9 29.4 31.4 34.9 30.8 25.7 14.4 6.1 3.2 2.7 4.8 34.4 33.6 28.3 10.9 4.5 2.8 4.3 5.8 12.8 14.0 15.4 17.5 19.3 24.6 27.1 9.8 21.8 26.3 27.0 21.9 12.4 6.7 5.0 0.0 11.8 22.4 26.5 19.4 8.6 5.6 1 .o 1 .o 3.5 2.6 3.4 4.7 5.9 10.6 15.9 2.4 9.3 15.0 18.2 22.6 18.8 12.2 8.9 0.0 3.2 11.0 16.9 22.4 14.9 9.9 0.0 0.0 0.9 0.5 0.9 1.1 0.7 3.1 7.2 0.0 3.9 6.6 9.4 17.0 20.5 17.6 14.0 0.0 0.0 5.7 8.0 18.8 19.5 16.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.7 2.9 0.0 1.2 2.5 4.2 9.9 16.9 19.5 17.7 0.0 0.0 1.2 3.8 12.0 19.7 19.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1 .o 0.0 0.4 0.5 1.6 5.3 11.0 16.2 17.9 0.0 0.0 0.0 0.9 6.8 15.2 18.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.3 6.5 11.0 14.3 0.0 0.0 0.0 0.0 3.7 9.5 13.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.8 3.5 6.5 8.9 0.0 0.0 0.0 0.0 1.7 4.5 7.9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.6 3.7 5.1 0.0 0.0 0.0 0.0 0.0 1.2 3.9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.7 2.1 0.0 0.0 0.0 0.0 0.0 0.0 1.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.03080 5 9 c ~ N.M.R.STUDIES OF DEUTERATION OF [Co(en),]Cl, Table 3. Effect on the cobalt-59 signal {from a neutral, 0.09 mol dm-, solution of [Co(en), -do] C1, * 3H20 in H20 at 33 "C> caused by additions of potassium hydrogen phthalate (PHP) PHP concentration measured 59C0 linewidth 59C0 signal offset /mol dm-, /Hz /Hz 0 0.09 0.28 0.55 97 109 118 136 0 - 190 - 450 - 680 must also depend on the proton and deuteron concentrations present in the exchange medium. These studies were performed in H20+D,0 and the concentration was chosen such that [Co(en),Cl,] < [HI, [D], so that any direct intermolecular exchanges of H or D that may occur between isotopomer molecules are disregarded in view of the overwhelming preponderance of protons and deuterons from the aqueous solvent.This large inequality means that [HI and [D] effectively remain constant throughout each reaction. It is anticipated that the rate constant for substitution may be isotope-dependent, i.e. that the rate constants for H + D and D -+ H substitution (designated kD and k,, respectively) may differ. The above considerations imply that when the fully-protonated isotopomer Do is dissolved in H20+D20 it will be consumed according to d[DOl/dt = k,[HI [D,1- 12 k , [Dl [Do1 H substitution D substitution while the rate of production of fully deuterated isotopomer D,, follows the equation d[D121/dt = kDID1 LD 1 - l 2 kH [D121- D s u b s t i t u t k H substitution The rate at which any other isotopomer Di is produced or consumed is given by d[D,]/dt = k,(12- i+ 1) [D] [DiJ +kH(i+ 1) [HI [Di+J -kHi[H] [Di]-kD(12-i)[D][Di].(3) This set of linear differential equations was solved numerically by means of hlerson's form of the Runge-Kutta method, and the computation used to obtain the solution was based on the DO~YAF sub-program from the NAG library at the University of East Anglia's Computing Centre. DO~YAF extrapolates from known boundary conditions the course of dependent variables such as the isotopomer concentrations of eqn (1)-(3). The initial concentrations of each isotopomer are the only applicable boundary conditions. Simulations of the kinetics of H/D exchange as it proceeds to equilibrium were carried out and the results were compared with the experimental data.An average deviation between the experimentally determined is0 topomer concentrations and their simulation values was defined and the parameters kH and kD varied until the minimum deviation was found. For each measurement at a particular time the deviation is taken to be120 100 4 0 : N 2 8 0 60 R. K. HARRIS AND R. J. MORROW 0.8 100 95 4 El 2 a N 9 0 8 5 1 .o 1.2 6k,/ 10’ h” 1 . 4 308 1 I I I 2 .o 4 - 0 6 .O 6k,/ 10’ h” Fig. 5. Plots of the average errors (between the experimentally determined isotopomer concentrations and their simulation values) against k , for (a) solution 2 and (b) solution 5. The number beside each trace is the k,/k, ratio used in the simulation.Deviations are then averaged over the total number of measurements. Sets of these average deviations were plotted against the parameter kH, each set having a different value of k,/k,. Sample plots are given in fig. 5 . The absolute minimum of k , / k , is taken to yield the correct values of k , and k H . For those solutions with 25-30 mol % deuterium this minimum occurs near k,/k, = 1.2. At higher deuterium concentrations kn/k, decreases. The minimum for the solution with 80% and 90% deuterium occurs when k,/kH is < 1.0. The results for the rate constant kH are listed in table 4. These values were inserted in eqn (1)-(3) along with the predetermined proton and deuteron molarities to simulate the courses of the reactions. The molar percentages of each isotopomer, plotted against reaction times, are displayed as continuous lines in the examples shown in fig.6(a)-(c). Superimposed on these simulations are the actual3082 100 L n €5? 4 2 W 50 0 a c s1 .A 0 I00 h * E - 0 v 50 a c s1 .A 0 5 9 C ~ N.M.R. STUDIES OF DEUTERATION OF [Co(en),]Cl, 0 * time/days 4 \ Do \ 0 2 4 6 8 10 time/h Fig. 6. Isotopomer distributions as a function of time for (u) solution 1, (b) solution 10 and (c) solution 18. The solid lines represent computer-simulated distributions (see text).100- 75 n @. 2 50 3 v c a c1 ._ 2 5 0- 0 100 200 300 time/min Fig. 6.(c) For legend see opposite. 400 500 w 0 00 w3084 59C0 N.M.R. STUDIES OF DEUTERATION OF [Co(en),]Cl, Table 4. Values of 6k, corresponding to the minimum deviations between the experimentally determined and simulated isotopomer concentrationsa solution 6k,/s-l solution 6k,/s-l 1 2 3 4 5 6 7 8 9 10 3.08 x 2.94 x lo-‘ 2.5s x 10-7 5.30 x 10-7 5.60 x 10-7 5.70 x 10-7 7.40 x 10-7 6.10 x lo-’ 9.00 x lop7 3.00 x lop6 11 12 13 14 15 16 17 18 19 6.67 x 8.61 x lop6 1.42 x 10-5 1.40 x 10-5 1.17 x 10-4 8.00 x 10-4 1.23 x 10-3 4.17 x 10-5 5.60 x 10+ a The program described in the text gives 6 k , as output, rather than k , itself.percentages, denoted by symbols. Each symbol type represents one of the thirteen isotopomers. Most of the experimental data lie close to that simulated, average deviations varying between 0.2 and 2.0 mol % . THE EXCHANGE MECHANISM A great deal of evidence has already been a c c ~ m u l a t e d l ~ - ~ ~ to show that in transition-metal amines such as [Co(en),]Cl,, hydrogen exchange occurs via the amido conjugate base -H+ +I)+ * Co-NH,+ Co=NH* 0 CO-NHD.The exchange mechanism in aqueous media is known to be SN 1, catalysed by OH-. In H,O+D,O mixtures both OH- and OD- must be contributing to the exchange. To quantify these contributions a mechanism of exchange in H,O + D,O is proposed from which equations are derived relating the experimentally determined rates to the mechanism rate constants. The mechanism of a proton replacement by a deuteron may be summarised as k - m Di + OH- DSbH20 k-H2 Di + OD- + D;” + DOH k+D D;”+ DOH + Di+l +OH- 2k+ 1) D;”+ D,O + Di+l + OD- 2k+H k + H k-I)i k-132 (4) (7) where DFb denotes the conjugate base molecule containing i deuterons.Of course, reactions (4t(7) are incomplete in the region of high acidity since they do not takeR. K. HARRIS AND R . J. MORROW 3085 into account reactions involving H,O+ and its deuterated isotopomers. We have ignored this complication, since even at the lowest pH used [H,O+] 4 [H,O]. A distinction is made between a replacement generated by OH- and that by OD-. Subscripts - H 1 and - D 1 refer to H+ and D+ abstraction by OH-, respectively, while - H2 and - D2 refer to H+ and D+ abstraction by OD-, respectively. The subscripts +H and +D are used to indicate abstraction by a conjugate base of a water proton or deuteron and no attempt is made to distinguish between abstraction from H,O or DOH or between abstraction from D,O or DOH. Eqn (4)-(7) may be substituted into eqn (1 )-(3) to relate the mechanism rate constants to the measured constants k , and k I ) .It is assumed that in the course of the reaction the concentration of the complex conjugate base derived from any two adjacent isotopomers remains constant so long as those isotopomers are present in detectable amounts. The concentrations of the conjugate bases remain very small for the pH range studied and the reaction rate of each isotopomer is monitored only when the contribution of the particular isotopomer is > ca. 0.5 mol %. Combining reactions (4)-(7), taking account of the number of protons or deuterons (as appropriate) in each isotopomer, and setting d[Dfb]/dt to zero gives [D;'] = (( 12 - i) (k-,1[OH-] + ~-H,[OD-] [Di]) + (i+ 1) (k-~i[OH-l+ k-1xJ0D-l) [Di+il>/Q (8) where Q = k+ H [HI ~+L)[DI (9) and 2[H,O]+[DOH] and 2[D,O]+[DOH] are abbreviated to [HI and [D], the total hydrogen and deuterium concentrations, respectively.The total rate of change of [Di] through formation from Dfb and D;Jl and removal by proton and deuteron abstraction is give by d[DJ/dt = k+,+[DFb] [HI + k+D[D:L'J [D] - (12 - i) (k-H1[OH-] +k-,,[OD-] [Di])-i(k-,,1[OH-] +k-,,[OD-] [Di]) (10) Substituting the expression for [DY"] and the analogous expression for [Df!l] and employing some algebraic manipulation gives 4Dil/dt = (i+ 1) k+ H(k-l>i[OH-l +k-,,,[OD-I) [HI [Di+illQ term (i) +(12-i+ l)k+,(k-,l[OH-]+k-H,[OD-])[D] [Di-,]/Q term (ii) -(12-i)k+~(k-~l[OH-]+k-i~,[O~-]) [Dl [Dil/Q term (iii) - ik+t1(k-ni[OH-I + k-m[OD-l) [HI [Dil/Q- term (iv).(1 1) A comparison of these four terms with the corresponding terms involving [HI [Di+J, LD] [D,-l], [HI [Di] and [D] [Di] in eqn (1)-(3) shows that k , and k , are composite products of the form k~ = k+H(k-l)i[oH-l + k-dOD-l)/Q (12) k1, = k+,(k-,,[OH-] +k,,[OD-])/Q. (13) k+H and kcr, are considered to be approximately equal and the mechanism rate constants k-Dl, k-,,, kPH1 and k-", may be found from the experimental rate parameters kH and k , at different hydrogen-to-deuterium ratios, provided the hydroxide and deuteroxide concentrations are known. Since the acidity rather than the alkalinity of each [Co(en),]Cl, solution was monitored (using a glass electrode), hydroxide and deuteroxide concentrations had t.0 be obtained indirectly, as discussed below.3086 59c~ N.M.R.STUDIES OF DEUTERATION OF [Co(en),]Cl, ESTIMATION OF THE CONCENTRATIONS OF [OH-] AND [OD-] IN H20 + D20 MIXTURES The relation between the pH meter reading and total acidity in H20 + D20 mixtures has been discussed by Glasoe and Long.ls They found that any glass-electrode pH meter calibrated using H+ in pure H20 gave, in a solution of D+ in pure D20, a meter reading (pH,,) 0.4 pH units less than a solution of H+ of the same concentration in pure H20. That is PD(pure D ~ O ) = ~ H m r - 0.4 (14) and this difference of 0.4 was found to be constant into the high-alkalinity region. For H20 + D20 mixtures they found that the meter reading pH,, was an almost linear function of mol % of deuterium. On the basis of these findings, total acidity ([H+] + [D+]) of the H20 + D20 mixtures is taken to be related to the reported pH meter reading by the equation (1 5 ) where fD is the deuterium mole fraction [D]/([H] + [D]).The distribution of acidity between H+ and D+ has been determined from the data of Korman and La Mer," who obtained values for the various equilibria that occur in H20+D20 mixtures. In particular, use is made of their reported equilibrium constants (at 25 "C) for the equilibria log,,([H + 1 + [D+D = log10 at = - (~Hrnr + 0.4fD) D++H,OeH++DOH and H20 + D20 e 2DOH From eqn (16) = 3.27. [DOHI2 W2OI P2OI [H+l 7. 1[H20] [D+] - = [DOH] ' -- (17) The ratio [H,O]/[DOH] may be determined from eqn (1 7) as this can be rearranged to give [DOH] in terms of initial concentrations of H20 and D20, i.e. the concentrations used in preparing the mixture.In order to isolate separate expressions for [OH-] and [OD-] it is proposed here that the concentration of H+ in a H20+ D20 mixture can be considered as derived from a hypothetical solution of H+ in pure H20 which has undergone dilution in D20 [H+l = [H+l(pure H , O ) ~ H (19) [OH-] = [OH-] (pu re H 01 f~ * (20) Thus [H+l [OH-I = KW(H20)fi (21) and for deuterium [D+l [OD-] = KW(D,O)fZ, (22) where fH is the protium mole fraction, [H]/([H] + [D]), and similarly where Kw(H20) and Kw(D,O) are the respective ion-product constants. The values for the ratio [H+] [OH-]/[D+] [OD-] obtained from eqn (21) and (22) agree well with the corresponding values predicted from the data of Korman and La Mer."R. K. HARRIS AND R. J.MORROW 3087 Table 5. Hydroxide and deuteroxide concentrations of the [Co(en),]C13 solutions solution [OH-]/mol dm-3 [OD-]/mol dm-3 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 1.68 x 1.45 x 1 0-l2 9.39 x 10-13 4.47 x 10-13 2.21 x 10-13 1.95 x 10-l2 1.42 x 4.24 x 5.47 x 10-l2 1.49 x 10-11 3.78 x 10-l' 6.50 x 10-l' 8.78 x lo-" 1.40 x 10-lo 4.63 x 10-lo 9.17 x 10-lo 1.03 x lo-* 1.87 x lo-* 3.04 x 10-13 4.14 x 10-14 5.71 x 10-13 3.13 x 10-13 7.82 x 10-13 9.44 x 10-13 6.14 x 10-13 1.07 x 1.22 x 10-l2 1.26 x 1 0-l2 2.86 x 8.41 x 10-l2 1.50 x 1.69 x 10-l' 3.16 x lo-" 7.85 x 10-l' 2.31 x 10-lo 1.57 x 1.06 x 10-'0 2.61 x 10-9 Inserting a, and a, [eqn (1 5) and (1 8)] into eqn (21) and (22) gives [OH-I = Kw(H2O)f2, (a, + 1)/ar at [OD-] = Kw(D201-G (a, + l)/at- (23) (24) Eqn (23) and (24) were used to estimate the concentration of [OH-] and [OD-] present in each solution, these values being listed in table 5 .The values of a, and a, refer to measurements at 25 "C while the values of K,(H,O) and k,(D,O) are taken at 33 "C. However, the error introduced by this is regarded as small, since a, and a, are considered to have small temperature coefficients. Another source of error must arise from the uncertainty of the ionization constants of ion products (K,) of H,O and D,O, no data being available for their values in 0.1 mol dm-, [Co(en,] Cl,. The extrapolation of the literature data1*, l9 for K,(H,O) and K,(D,O) (at zero ionic strength) to 33 "C gives values of 0.18 x lo-', and 0.28 x respectively, so these are used in this study.In 1 mol dm-, KC1 at 34.3 "C, K,(H,O) and K,(D,O) are quoted as 0.195 x and 0.38 x loy1*, respe~tively.~3 * The systematic error in the value of K,(D,O) and thus [OD-] is therefore likely to be greater than the error in K,(H,O) and [OH-]. Each pH meter reading is considered to have a limit of accuracy of f 0.05 pH unit, which would give rise to errors in hydroxide and deuteroxide concentrations of ca. 10%. ESTIMATION OF RATE CONSTANTS OF THE EXCHANGE The experimental data have been plotted in various ways to test the validity of eqn (1 2) and (1 3). Fig. 7 shows one such plot, that of In 6k, against In [OH-] for solutions with proton-to-deuteron ratios in the range 2.6-3.4, corresponding to values of3088 -J, 1 I , , . , , , , , * 5 9 C ~ N.M.R. STUDIES OF DEUTERATION OF [Co(en),]Cl, Fig. 7.Plot of In 6k, against In [OH-] for the solutions indicated by the numbers (see table 1). [OH-]/[OD-] from 4.0 to 6.0. All the points lie close to a straight line, the least-squares gradient of the line being 0.97. A ‘worst’ straight line drawn through the points suggests a maximum error in the gradient of 0.07. Thus at a constant value of [OH-]/[OD-] the exchange rate, within experimental error, is proportional to the hydroxide-ion concentration. Using this proportionality and extrapolating to unit hydroxide concentration allows rate-constant information to be extracted. This gives the value of In 6kH at ln[OH-] = 0 as 1 1.8 f 0.3. Substituting this result and the values 0.2(+0.04) for [OD-]/[OH-] and 108.0 for [H]+[D] into the equation In kH = In {(kD1 + [oD-]k-D2/[oH-])/([H] + [D])} + In [OH-] (25) gives k - ~ , + 0 .2 k - ~ , = 2.5 X lo6 dm3 m01-l S-’. The In 6kH was also plotted against the ln[OD-] (fig. 8) and treated in a similar manner to the previous plot. All points of roughly constant [OH-]/[OD-] values lie on a straight line of least-squares gradient 0.93 with error f 0 . l . Thus the exchange rate at constant [OH-]/[OD-] is, within experimental error, proportional to the deuteroxide-ion concentration. The value of In 6k, at ln[OD-] = 0 is 13.5f0.3. Substituting this result and the values 5.0 for [OH-]/[OD-] and 108.0 for [H]+[D] into the equation In k , = In (([OH-] ~-D~/[OD-] 4- k,,)/([H] 4- [D])} +In [OD-] (26) gives 5.0 kPDl + k-,, = 1.3 x 10’ dm3 mol-l S-’ or a result close to that obtained from the previous plot. kD1 + 0.2 k-,,, = 2.6 x lo6 dm3 mol-l s-’R.K. HARRIS AND R. J. MORROW 3089 Fig. 8. Plot I I I I - 30 - 2 5 - 20 In [OD-] of In 6k, against In [OD-] for the solutions indicated by the numbers (see table 1). A more exacting test of the validity of eqn (12) is to plot k,/[OH-] against [OD-]/[OH-]. Such a plot should be linear, yielding kWD, and LD, from the intercept and slope as Fig. 9 shows the experimental data plotted in this way. The graph is considered linear, within experimental error, validating eqn (1 2), but the accuracy with which the rate constants can be determined is limited because of the nature of the distribution of the points. The least-squares slope and intercept give (using a value of 108.0 for [HI + [D]) k-D2 and k-I)l as (8.3 1.7) x lo6 dm3 mol-l s-l and 1 .O x lo6 dm3 mol-l s-l, respectively.There is a large degree of error associated with the last value. However, it can definitely be concluded that the rate constant for abstraction of D by OH-, kD1, is significantly smaller than that of the rate constant for abstraction of D by OD-, k.-I>2. The relative error associated with k-,,, is correspondingly larger than that associated with kD2. This is reflected by a plot of k,/[OD-] against [OH-]/[OD-] where the large error associated with each point masks the linearity that is, from the previous plot, considered to be present. However, least-squares fitting of the points gives kPD, as ca. 6.7 x lo6 dm3 mol-1 s-' and kPD, as ca. 1.7 x lo6 dm3 mol-1 s-l.One measurement was made of the rate of protonation of a mixture of D12, D,, and D,, isotopomers in H20 (solution 19). The equilibrium concentration of hydroxide ion in this solution is 170 times less than the concentration of deuteroxide, so that [OD-] can be eliminated from eqn (1 2) giving k H = k-,,[OH-I/(CHI + [DI) (28) where k , = 0.2 x lop3 dm3 mol-l s-l and [OH-] = 1.87 x lo-* mol dm-3, giving a3090 5 9 C ~ N.M.R. STUDIES OF DEUTERATION OF [Co(en),]Cl, Fig. 9. Plot of 6k,/[OH-] against [OD-]/[OH-]. value of kD1 equal to 1.1 x lo4 x 108 = 1.2 x lo6 dm3 mol-1 s-l, close to the value obtained from the least-squares plot of fig. 9. However, no great reliance can be placed on a one-off measurement and this must merely serve as a check on the reasonableness of the quoted value.EQUILIBRIUM SPECTRA Each equilibrium distribution of isotopomers should also reflect the inequality between k , and kD. As these distributions can be predicted in terms of kD/kH and [H]/[D] a similar investigative procedure to the one adopted for the analysis of the kinetic experiments was used. In the equilibrium case simulations were performed with different values of the only variable parameter, k,/kH, and the value of this ratio which gave the minimum deviation between simulation concentrations and those experimentally determined was found and asserted to be the true ratio. At equilibrium each isotopomer must have a concentration weighted according to 12Ci, as each deuteron occupies one of the i sites on a molecule which has 12 possible sites.If the exchange between pairs of isotopomer combinations can be expressed by reactions of the formR. K. HARRIS AND R. J. MORROW 309 1 0-4 8 1 i l l 1 1 1.4 0.8 1.8 0.7 kDlkH k 1.7 Fig. 10. Average errors for (a) solution 4, (b) solution 10 and (c) solution 17 at equilibrium (see text) plotted against k,/k,. and so on, where the superscript ’ indicates a combinatorial form of the isotopomer, i.e. Di = Di/12Ci, then at equilibrium tD’lle = (kDIDl/kHIHI) ID;], (29) and P i l e = (k,[DI/kdHI) P ; I e (30) or generally CDi1e = (kdDl/kHIHl)i [D;le* (31) As [DJ, = l2Ci[DJe and each isotopomer will have an equilibrium concentration given by [Di] = total [Co(en),]Cl, concentration, denoted by [Co], i Varying k D / k , in eqn (32) and holding [H]/[D] at its known value produced sets of simulated results, which may be compared with the experimentally determined concentrations.The deviation or average error between both sets of results is taken as (1 / 13) Z I(mol% of Qexperimental - (mol% of 9simulationl i and this is plotted against a range of k , / k , values. Fig. 10 shows how the average errors of some of the solutions vary with k , / k , and table 6 lists the values of k,/kH which give the minimum deviations, together with the experimentally determined concentrations. For those solutions where the concentration of deuterium is 25-30%, k , / k , is ca. 1.15, while at higher concentrations of deuterium k , / k , is reduced. This is in agreement with the kinetics results. The limiting value of kD/k,, as measured from the equilibrium spectra of [Co(en),]Cl, - 3H,O in 99.8 % D 2 0 (solution 1 8), is ca.0.6. The limiting value at the other extreme, pure H20, measured from the equilibrium spectrum of solution 19, is ca. 1.25.Table 6 . Experimental isotopomer concentrations of most of the solutions at equilibrium at 33 "C" v, molar percentage of each isotopomer 8 Z solution [H]/[D] Do D, D, 2 4 5 9 10 12 13 14 15 16 17 18 19 20 1 .o 0.0 0.27 0.0 0.12 0.0 5.1 9.8 3.0 I .o 2.5 4.7 3.0 1.1 2.6 3.5 3.4 4.2 2.3 0.3 2.3 0.0 0.01 0.0 97.4 86.6 0.01 0.0 0.0 0.0 0.0 22.3 7.9 8.4 8.3 7.8 11.5 3.4 2.9 0.0 13.4 0.0 0.0 0.0 0.0 26.5 18.9 14.8 18.5 15.1 19.9 11.4 12.3 0.0 0.0 0.0 D, D4 D5 D6 D, D8 D9 DIO Dll DlY k,/kFIb % 2.5 7.8 16.0 22.4 22.8 17.0 8.9 2.6 0.0 0.0 1.21 2 sd 0.0 0.0 0.0 4.4 13.0 22.2 26.5 22.5 11.3 0.0 0.75 c( U 20.5 11.2 6.3 2.6 0.8 0.0 0.0 0.0 0.0 0.0 1.16 g 25.6 23.0 14.7 6.4 1.9 0.6 0.0 0.0 0.0 0.0 1.22 20.5 19.9 15.4 9.3 4.9 2.0 0.0 0.0 0.0 0.0 1.25 8 19.6 19.4 16.1 9.6 5.1 2.6 1.3 0.0 0.0 0.0 1.18 w v1 0.0 0.0 0.0 0.0 2.4 7.7 18.2 29.0 28.3 14.4 0.64 v1 M 25.2 22.5 14.9 6.9 2.1 0.5 0.0 0.0 0.0 0.0 1.22 ;;1 22.8 18.1 11.8 6.8 3.9 0.9 0.0 0.0 0.0 0.0 1.22 5 19.5 23.8 20.2 12.9 6.2 2.3 0.0 0.0 0.0 0.0 1.26 5 5 21.5 25.4 20.7 12.5 4.3 0.3 0.0 0.0 0.0 0.0 1.18 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.9 16.2 82.9 ca.0.6 2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ca. 1.25 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.4 17.6 81.0 ca. 0.6 3 - W a Solution 1 had not reached equilibrium after two months at room temperature. These are the k,/k, ratios which produced the best simulations of the measured isotopomer concentrations.WR. K. HARRIS AND R. J. MORROW 3093 Substitution of these limiting values into eqn (1 3) can give estimates of the values of the rate constants kH1 and k - H 2 directly. Given that eqn (12) and (1 3) reduce to and (33) (34) in the region of low percentage deuteroxide and that kD/kH in this region is 1.25, it follows that kH1 = 1.25 kD,. The rate constant k-Dl was, with an error comparable to its value, estimated to be 1.2 x lo6 dm3 mol-1 s-l, and kH1, with a similar error attached, has a value of ca. 1.5 x lo6 dm3 mob1 s-l. Considering the other extreme, the region of low percentage hydroxide, eqn (12) (35) and (1 3) can be written as k H = b,[OD-I/[Dl and kD = b,[OD-I/[DI (36) and taking k,/k, here as 0.6 gives k-H2 % 0.6kD2.The rate constant kl), has been determined as (8.3 1.7) x lo6 dm3 rnol-' s-l so that k-Hz must have a value of ca. 5.0 x lo6 dm3 mol-l s-l. This value of kHz agrees (within experimental error) with that obtained from the one-off measurement of [Co(en),]Cl, in 99.8% D, (solution 18). The rate parameter kD for this solution was found to be 5.6 x lop7 dm3 mol-l s-l and [OD-] was estimated as 1.57 x mol dm-3, giving kH2 = 3.9 x lo6 dm3 mol-l s-l. Therefore, placed in order of descending value : k - D 2 > k - H z > k-H1*> k - D , which may be restated as: rate of removal of D by OD- > H by OD- > H by OH- > D by OH-. The greater abstracting power of OD- over OH- must be partly attributed to OD- being a better nucleophile than OH-.As the present paper was sent to press, we became aware of the short publication of Russell and Bryant,20 which discusses the use of the isotope effect on the 5gCo resonance of aqueous hexa-amminocobalt(Ir1) chloride to determine the isotopic composition of the solvent. A further paper by Peterson et al.,,, which became available to us only when the present paper was accepted, gives further details of the proton/deuterium isotope effect on equilibria for hexa-amminocobalt(In) solutions under a variety of conditions, but the results cannot be directly related to those we give herein. We thank Dr R. D. Cannon for supplying the sample of tris(ethy1enediarnine) cobalt chloride and for very helpful discussions about cobalt chemistry. One of us (R.J.M.) thanks the Northern Ireland Department of Education for a Research Students hip. NMR and the Periodic Table, ed. R. K. Harris and B. E. Mann (Academic Press, London, 1978). M. R. Bendall and D. M. Doddrell, Aust. J . Chem., 1978, 31, 1141. D. J. Burton and R. K. Harris, unpublished work. F. Basolo, J. W. Palmer and R. G. Pearson, J . Am. Chem. Soc., 1960, 82, 1073. J. W. Palmer and F. Basolo, J . Phys. Chem., 1960, 64, 778. ti B. Halpern, A. M. Sargeson and K. R. Turnbull, J . Am. Chem. Soc., 1966, 88, 4630.3094 59C0 N.M.R. STUDIES OF DEUTERATION OF [Co(en),]Cl, ' D. A. Buckingham, L. G. Marzill and A. M. Sargeson, J. Am. Chem. SOC., 1967, 89, 825; 1967, 89, 3428; 1968, 90, 6028; Inorg. Chem., 1968, 7 , 915. G. H. Searle and F. R. Keene, Inorg. Chem., 1972, 11, 1006. D. A. Buckingham, P. J. Cresswell and A. M. Sargeson, Inorg. Chem., 1975, 14, 1485. lo U. Sakaguchi, K. Maeda and H. Yoneda, Bull. Chem. SOC. Jpn, 1976,49, 397. l1 R. K. Harris and R. H. Newman, J. Magn. Reson., 1976, 24, 449. l 2 R. J. Morrow, Ph.D. Thesis (University of East Anglia, 1982). l 3 J. Burgess, Inorg. React. Mech. 1971, 1, 179. l4 J. Burgess, Inorg. React. Mech. 1972, 2, 171. l5 J. Burgess, Inorg. React. Mech. 1974, 3, 219. l6 P. K. Glasoe and F. A. Long, J. Phys. chem., 1960, 64, 188. l7 S. Konnan and V. K. La Mer, J. Am. Chem. SOC., 1936, 58, 1396. l9 Handbook of Chemistry and Physics, ed. R. C . Weast (CRC Press, Cleveland, Ohio, 55th edn, 1974), 2o J. G. Russell and R. G. Bryant, Anal. Chim. Acta, 1983, 151, 227. 21 S. H. Peterson, R. G. Bryant and J. G. Russell, Anal. Chim. Acta, 1983, 154, 21 1 . R. W. Kingerley and V. K. La Mer, J. Am. Chem. SOC., 1941,63, 3256. p. D131. (PAPER 4/273)

ISSN:0300-9599    

DOI:10.1039/F19848003071

出版商:RSC    

年代:1984

数据来源: RSC