摘要:

J . Chem. SOC., Faraday Trans. I, 1986, 82, 3141-3148 Effects of Electrostatic Repulsion on the Aggregation of Azo Dyes in Aqueous Solution Kunihiro Hamada, Seiji Take and Toshiro Iijima* Department of Polymer Science, Tokyo Institute of Technology, Ookayama, Meguro-ku, Tokyo 152, Japan Shigetoshi Amiya Central Research Laboratories, Kuraray Co. Ltd, Sakazu 2045, Kurashiki 710, Japan The aggregation of an azo dye containing one trifluoromethyl and two sulphonate groups (m-FTR) in aqueous solution has been investigated by means of 19F n.m.r. and electronic absorption spectroscopy. The 19F n.m.r. signals shifted to lower magnetic field, indicating that the fluorine atoms are located outside the aromatic ring of an adjacent dye molecule in the aggregate. On the other hand, in the case of an azo dye carrying one trifluoromethyl and one sulphonate group (m-FTS), the fluorine atoms were oriented above the adjacent aromatic ring. This difference gives a different aggregate structure between m-FTR and m-FTS. The absorption spectra of m-FTR aqueous solution show a single monomer/dimer equilibrium.On the basis of the results, the aggregation constant K was determined. The value of K for m-FTR was half that for m-FTS, suggesting that the latter dye aggregates more easily than the former owing to a lower electrostatic repulsion. The aggregation of small molecules in solutions and at interfaces is a current subject of research in biological, colloid, surface, textile and photographic chemistry etc. The behaviour of surfactant and dye in aqueous solutions is a pertinent example of aggregate formation.Dye solutions at high concentration usually exhibit a deviation from the Beer-Lambert law. Based on this phenomenon the aggregation behaviour of various dyes has been widely in~estigated.l-~ Azo sulphonic dyes, e.g., Methyl Orange and its homologues, were often used for the experiments. Cyanine and acridine dyes have been investigated in detail and discussed extensively, with regard to their aggregation not only in aqueous solution but also in polymer solution. Cyanine dyes are currently being investigated as sensitizers of solar-energy conversion systems through photo- chemical processes, with special attention being paid to their J-aggregates. n~~ HO S0,Na We have studied the aggregation behaviour of an azo dye containing one trifluoro- methyl and one sulphonate group (m-FTS, 1) by means of absorption spectra and 19F n.m.r.measurement~.~~ Chemical shifts determined by the n.m.r. measurements made it possible to presume an aggregate model. Blears et ~ 1 . ~ ~ 9 l5 and Veselkov et a1.16 made a lH n.m.r. investigation of the aggregation of acridine dyes. However, in the case of 31413142 Aggregation of Azo Dyes IH n.m.r., the assignment of peaks is complicated and the change in chemical shifts is small. On the other hand, the chemical shifts in 19F n.m.r. exist in the region from -450 to 550 ppm and are sensitively changed by the ring current and carbonyl n-electrons. Furthermore, we can obtain information solely for fluorine atoms. Because of these advantages, 19F n.m.r.investigation has been performed on the interaction between proteins and small molecule^.^^ a disulphonic azo dye containing one trifluoromethyl group (m-FTR, 2) was synthesized and its As an extension of our previous work on a monosulphonic azo S0,Na aggregation behaviour was studied by means of electronic absorption spectra and 19F n.m.r. measurements. In this work the effect of the electrostatic repulsion on dye aggregation is focussed to elucidate the molecular interaction of this system. Experimental Compound (2) was synthesized by coupling diazotized rn-trifluoromethylaniline with an R acid (2-naphthol-3,6-disulphonic acid) in alkaline conditions. The dye obtained was purified by repeated recrystallization from aqueous acetone and finally by passage through a silica gel column.The purity was confirmed by elemental analysis (obsd: C , 39.56; H, 2.62; N, 5.08; S , 12.47; F, 10.46%; calcd: C , 39.24; H, 1.74; N, 5.38; S, 12.32; F, 10.95%). lgF n.m.r. and absorption spectra were determined by using a Jeol GX-500 spectrometer and Hitachi 556 double-wavelength double-beam spectrophotometer, respectively. All the measurements were performed under the same conditions as in the previous work .I3 Results Absorption spectra obtained in an aqueous phosphate buffer solution of pH 7.10 at an ionic strength of 0.108 are given in fig. 1. In this figure isosbestic points are defined, in- dicating the existence of a single aggregation equilibrium in this concentration range. As the shapes of the absorption spectra in pure water and aqueous NaCl solution are the same, the addition of salts seems only to facilitate aggregate formation, leading not to a change of the equilibrium system but to an increase of the aggregation constant.The extinction coefficient at 500 nm [a shoulder of the maximum absorption (483 nm) at the higher wavelength] changed with the dye concentration to a greater extent than at 483 nm. The absorption at 500 nm is estimated as the result of the interaction between n-electrons in the dye aggregate. The chemical shifts of the fluorine atoms of (2) and (1) as a function of the dye concentrations are shown in fig. 2. The fluorine signal of (2) shifted to lower fields as the concentration increased. This shift is opposite to that of (l), suggesting differences in their aggregate structures.At low concentration the chemical shifts are found to saturate, but the saturated values vary from dye to dye owing to differences in their hydration states and substituent groups.3143 U 300 400 500 600 wavelength/nm Fig. 1. Absorption spectra of (2) aqueous phosphate buffer solution (298 K, pH 7.10, ionic strength 0.108). (-) 1.44 x mol dm-3, (------) 1.44 x lo-* mol dm-3, (---) 5.76 x mol dm-3, (---- ) 1.44 x mol dm-3. 15.5 15.4 2 a (.o W 15.3 15.2 -6 -5 -4 -3 -2 -1 log (Co/mol dm-3) Fig. 2. Dependence of chemical shift 6 on the concentration of aqueous dye solution (298 K). 0, (1); e, (2).3 144 40 30 - s z . h d I z rn v E 5 20 . h 3 I s G . 3 5 10 0 0 Aggregation of Azo Dyes 5 10 A€/ 10’ m2 mol-’ Fig. 3. ( A E / C F - ~ ) ~ ’ ~ us.A& at 500 nm. Aqueous phosphate buffer solution of (2) (298 K, pH 7.10, ionic strength 0.108). n: 0, 2; a, 3; A, 4. 40 d 3 I 0 E 30 3 . n I N . 0 g u, d, 20 - / 0 5 10 Ae/ 1 O2 m2 mol-’ Fig. 4. ( A E / C , ) ~ / ~ us. Ae at 500 nm. Solution of (2) (298 K). 0, Water; a, phosphate buffer (pH 7.10, ionic strength 0.108).K . Hamada et al. 3145 0 5 10 A€/ 1 O2 m2 mol-' Fig. 5. (A&/Co)l12 us. A& at 476 nm. Solution of (1) (298 K). 0, Water; a, phosphate buffer (PH 7.10, ionic strength 0.087). Table 1. Aggregation constant K (dm3 mol-') at 298 K n.m.r. absorption spectra water water (0. 108)" (0.020)" (0.05 1)" (0.102)" buffer NaCl NaCl NaCl m-FTS, 1 430+40 510+50 - - - - m-FTR, 2 230 & 20 260 & 20 860 &- 20 340 f 20 440 5 30 470 f 30 a Ionic strength.Discussion From the result of absorption spectra, the following single equilibrium can be assumed: nD D,. (1) Denoting the aggregation constant as K, K = C,/Cr where C, and C , are the concentrations of monomer and n-mer, respectively. These concentrations can be written with the total dye concentration Co: c1 = [(AED, - AE>/AEDnl cO (3) cn = (AE/AED,) cO (4) where Ac = Jc-cEgJ, A E ~ , = JED, -cDI and E is the molar extinction coefficient observed at a certain wavelength. cD and cDn are the molar extinction coefficients of monomer and n-mer, respectively. Substituting eqn (3) and (4) into eqn (2), we obtain ( 5 ) (Ac/C;--l)lln = - (nk/Acg;l)l/nAe + (nkAcDn)lIn.3146 15 N . 4 I -z 10 E E N . ‘u N . d 5 . G 5 a 2 . - h a 9 0 Aggregation of Azo Dyes 0 0.1 0.2 A6 (ppm) Fig.6. (A6/C,)1/2 vs. Ad. Aqueous solution of (2) (298 K). Fig. 3 shows the relations between (A&/C:--l)lln and A& at 500 nm with varying aggregation number, n. The plot of n = 2 showed good linearity with a correlation function of more than 0.99, while those of n >, 3 were curved. This result indicates that the aggregation number of (2) is two. The linearity was kept unchanged with the ionic strength of the aqueous dye solution (fig. 4), i.e. the aggregation number did not change with the ionic strength. On the other hand, the plot for (1) became less linear with an increase of ionic strength (fig. 5). This is attributed to the polyaggregate formation of (1) at a high ionic strength as was described in the previous paper.13 As it is difficult to determine the multiple aggregation equilibrium constants, the case for (1) in the presence of salt was not analysed.The aggregation constant Kcalculated from the slope and intercept of eqn ( 5 ) is shown in table 1. K of (2) increased with increasing ionic strength. The aggregation was enhanced in aqueous phosphate buffer solution in comparison with aqueous NaCl solution. These salt effects are due to an increase of the dye activity through the hydration of the individual ionic species of these salts. In pure water the K value for (2) reduces to half of that for (1) indicating that the aggregation of (2) is hidered much more by the repulsive force of an extra anionic charged group, -SO;. The stronger repulsive force is expected to change the equilibrium mode as well as the aggregation constant. The aggregation constant K was also obtained by 19F n.m.r.information given in fig. 2. Assuming the monomer/dimer equilibrium as mentioned above, we obtain the following equation :le where Ad = 1d-d,1, AbD2 = ldD,-d,l and 6 is the observed chemical shift. 6, and d,, are the chemical shifts of the monomer and dimer species, respectively. The plot of (AC~/C,)~/~ against Ad gives a straight line (fig. 6), the slope and intercept of which gave K and do, as 230 f 10 dm3 mol-1 and 15.5469 ppm, respectively. WhereK. Hamada et al. 3 147 Fig. 7. A model of the dye aggregates in aqueous solution. S, = 15.3200 ppm, was used as an extrapolated value to the extreme dilution in fig. 3. The aggregation constant thus obtained agrees quite well with that calculated from the electronic absorption spectra as shown in table 1.Here it is worthwhile to note that in both cases of the dimer formation model in the present study and the isodesmic model used in the previous the plot of (Ad/Co)1/2 against Ad gives a straight line.18 This means the aggregation mode cannot be presented solely by the 19F n.m.r. data. The aggregation constant obtained with assuming the dimer formation is half of that calculated from the isodesmic model. When the 19F n.m.r. data in fig. 2 are analysed by the isodesmic model, two-fold larger numerical value than in table 1, i.e. 460 f 20 dm3 mol-l will be obtained. This value leads to a false argument about the effect of the chemical structure of the dye on the aggregation.In this context the electronic absorption data and the 19F n.m.r. data are the complementary information to establish the aggregation model. Spin-spin relaxation times T, were calculated from the width of the n.m.r. signals. T2 of (2) is kept at 100 ms even in the concentrations > mol dm-3, whereas for (1) a pronounced decrease of T2 from 100 to 6.3 ms at 3 x mol dm-3 was 0bser~ed.l~ This is an additional proof that (1) forms polyaggregates but (2) does not. Duff et aL8 and Jones et aLg pointed out the strong effects of sulphonate groups on the mechanism of the aggregation. However, they did not argue the aggregate model of the molecular level. The 19F n.m.r. data gave us another prominent information concerning the location of fluorine atoms. The change of chemical shifts, either to higher or lower magnetic field, indicates a relative spatial location between fluorine atoms and aromatic rings. From fig.3, the fluorine atoms of (1) seems to be oriented above the plane of the aromatic rings of adjacent dye molecules, while those of (2) are not located directly above the aromatic rings. Taking into account the above results, together with the repulsion between the negative charges, a model of the aggregate is proposed in fig. 7. It is thus clear that both hydrophilic and hydrophobic moieties of the dye affect the aggregation process as well as the mode of the aggregate. 104 F A R 13 148 Aggregation of Azo Dyes References 1 A. H. Herz, Photogr. Sci. Eng., 1974, 18, 323. 2 A. H.Herz, Adv. Colloid Interface Sci., 1977, 8, 237. 3 E. Wyn-Jones and J. Gormally, Aggregation Processes in Solution (Elsevier, Amsterdam, 1983), p. 271. 4 T. Takagishi, S. Fuji and N. Kuroki, J. Colloid Interface Sci., 1983, 94, 114. 5 A. Datyner, M. J. Delaney and H. Holliger, Aust. J. Chem., 1971, 24, 1845. 6 A. Datyner, A. G. Flowers and M. T. Pailthorpe, J. Colloid Interface Sci., 1980, 74, 71. 7 A. Datyner and M. T. Pailthorpe, J. Colloid Interface Sci., 1980, 76, 557. 8 D. G. Duff, D. J. Kirkwood and D. M. Stevenson, J. Soc. Dyers Colour., 1977, 93, 303. 9 F. Jones and D. R. Kent, Dyes Pigm., 1980, 1, 39. 10 P. Dan, I. Willner, N. S. Dixit, and R. A. Mackay, J. Chem. SOC., Perkin Trans. 2, 1984, 455. 11 H. Hada, R. Hanawa, A. Haraguchi and Y. Yonezawa, J. Phys. Chem., 1985,89, 560. 12 L. M. Natoli, M. A. Ryan and M. T. Spitler, J. Phys. Chem., 1985, 89, 1448. 13 K. Hamada, H. Kubota, A. Ichimura, T. Iijima and S. Amiya, Ber. Bunsenges. Phys. Chem., 1985,89, 14 D. J. Blears and S. S. Danyluk, J. Am. Chem. Soc., 1966, 88, 1084. 15 D. J. Blears and S. S. Danyluk, J. Am. Chem. Soc., 1967, 89, 21. 16 A. N. Veselkov, L. N. Dymant and E. L. Kulikov, Khim. Fiz., 1984, 3, 1108. 17 See e.g., L. J. Berliner and J. Reuben, Biological Magnetic Resonance (Plenum Press, New York, 1978), 18 J-L. Dimicoli and C. Helene, J. Am. Chem. Soc., 1973,95, 1036. 859. vol. 1, p. 139. Paper 512159; Received 9th December, 1985

ISSN:0300-9599    

DOI:10.1039/F19868203141

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1986, 82, 3149-3161 Phosphorus and Proton Nuclear Magnetic Resonance Studies of Transition-metal Complexes of Triphosphate and Pyrophosphate in Aqueous Solution Olive Laurie, John Oakes,* Jeff W. Rockliffe and Edward G. Smith Unilever Research Port Sunlight Laboratory, Quarry Road East, Bebington, Merseyside L63 3J W The structures of transition-metal complexes of triphosphate and pyro- phosphate have been examined in aqueous solution by phosphorus 31P and water proton n.m.r. For pyrophosphate, 1 : 1 bidentate complexes containing four water molecules are formed with Ni2+, Co2+ and Mn2+ in the pH range 7-10, but the Cu2+ complex contains two pyrophosphate molecules. 1 : 1 complexes are also formed between triphosphate and Ni2+, Co2+ and Mn2+ but there exists a dynamic equilibrium between bidentate and tridentate forms, the former being favoured by Mn2+ and the latter by Ni2+ and Co2+.A further equilibrium involving a complex containing two triphosphate molecules becomes important for Cu2+ and this predominates in alkaline media. The lifetimes for triphosphate bound to transition metals, Cu2+, Mn2+, Co2+, Ni2+ and Fe3+, have been determined as a function of temperature, and activation energies for ligand exchange are 4.6,23.4, 32.1, 34.4 and 25.8 kJ mol-l, respectively. The rotational correlation times determined for triphosphate complexes of Mn2+ and Cu2+ are 5.6 x and 4.0 x s determined from 31P n.m.r., but the effective correlation times determined from water-proton n.m.r. are 4.5 x lo-" and 5.0 x s, respectively.The electronic relaxation time for Mn2+ decreases from 1.5 x to 5 x s on complexation with equimolar amounts of tripolyphosphate. Sodium triphosphate (STP) is used widely in the detergents industry as a water softener because of its affinity for sequestering metal ions. Although a wide range of metal-ion complexes have been ~haracterizedl-~ and their stability constants determined, particu- larly for 1 : 1 complexes, there has been very little reported'? 2~ about their structure or lability. In the work described here, n.m.r. spectroscopy has been used to elucidate the structure of triphosphate (TP) complexes of ions of the first transition series and to ascertain the dynamics of ligand exchange involved. In particular, water-proton relaxation measurements have been made, using procedures described previo~sly,~-~ to determine the number of water molecules bound to the complexes.Relaxation times have also been measured for the a and j3 phosphorus atoms in triphosphate complexes to provide information about their stereochemistry and lifetimes. To illustrate this point, where TP acts as bidentate ligand, an equilibrium between two forms, I and 11, with equal probability might be anticipated, viz. 0 0 0 0 0 0 0-P-0-P-0-P-0 e 0-P-0-P-0-P-0 1 I I I I I I I ( I ) Mh+ (11) ++ 0 I 0 I \\ /' 0 I 0 I 0 0 \ / \ / \\ // 3149 104-23150 N.M.R. Studies of Phosphate Complexes Thus, under conditions of fast exchange, the central phosphorus atom (p) is expected to be relaxed twice as efficiently as the terminal phosphorus (a).For a tridentate complex, on the other hand, less discrimination in relaxation rates might be expected. In addition, investigations for the pyrophosphate (PP) complexes of the corresponding metal ions have been made, primarily as an aid to interpretation of the results for the triphosphate system. Experiment a1 Materials Anhydrous sodium triphosphate (Albright and Wilson) was purified by recrystallising twice from aqueous ethanol to yield the hexahydrate. AnalaR tetrasodium pyrophosphate decahydrate was obtained from B.D.H. Solutions containing MnSO,, CuSO,, CoCl,, NiC1, and FeC1, were prepared using water which had been deionized and doubly distilled from permanganate. Complexes were formed by addition of solutions containing an excess of TP or PP to the above salt solutions and the pH adjusted accordingly.For the preparation of the Fe3+ complexes it was essential to start with freshly prepared solutions in order to avoid the irreversible formation of polymeric iron oxides (yellow colour). The TP complex of Fe3+ is colourless, indicating the absence of charge-transfer bands characteristic of bound hydroxy groups. For all complexes studied it was important to have ligand to metal ratios > 2 to prevent precipitation of the complexes or their hydrolysis above pH 8; this occurs particularly for copper and iron complexes. N.M.R. Measurements Phosphorus 31P T,, (spin-lattice) n.m.r. relaxation time measurements were carried out on a Bruker CXP300.HP spectrometer operating at 121.4 MHz and using a variable temperature probe accurate to within f 1-2 "C.The high-resolution Fourier-transform (f.t.) spectrum for TP consists of a doublet and a triplet in the intensity ratio 2: 1 and separated by ca. 14 ppm, corresponding to the a and /3 phosphorus resonances, respectively. The separate T, relaxation times for the two groups were determined from a semi-log plot of peak intensity versus pulse separation, z, in a standard 180°-z-900 inversion recovery experiment, where the f.t. spectrum for each setting of z is obtained from the free induction decay (f.i.d.) signal following the 90" pulse. The values of 3.7 and 3 s for the q of 01 and /3 resonances at 333 K were high enough to indicate the absence of any serious interference from relaxing metal-ion impurities. T, (spin-spin) relaxation times were obtained using the conventional spin-echo 90-( 180), pulse programme of Carr Purcell/Meiboom and Gill, described previously.sT All 31P relaxation time measurements were carried out in solutions prepared in D,O.Water, W, Relaxation-time Measurements These were performed on a Bruker 3223 variable-frequency wide-line pulse n.m.r. spectrometer equipped with variable-temperature probe. magnetization recovery curves were established from the f.i.d. amplitude following the 90" pulse in the 18O"-z-9O0 inversion recovery method, and & was measured as for the experiments. In cases where = T,, for convenience only the spin-spin relaxation time was measured.0. Laurie, J . Oakes, J . W. Rocklifle and E. G. Smith 3151 Theory General The effect of adding transition-metal ions such as Mn2+ to aqueous solutions of triphosphate is to increase the relaxation rates, c1 and cl, of both the phosphorus nuclei and the water protons.The measured relaxation rates are given by [T,, G1-l = [T,, TIP1 + [T,, GI2 ( 1 ) where [T,, &];;l is the diamagnetic contribution to relaxation arising from unbound ligand and water molecules and [q, TJpl is the paramagnetic contribution, given bylo, l1 [TIP' = ~ ( T , M + 7 ~ I - l (2) and ( 3 ) wherefis the fraction of water or ligand bound to the paramagnetic ion, zM the lifetime of water or ligand in the bound state, A WM is the shift between resonances for the free and bound states in the limit of long zM, and T,,M and G,M are the relaxation times of bound water protons or phosphorus nuclei.In the limits of slow and fast exchange12 [T,lP1 = [GIG1 and [T,, &]pl =fikl (slow exchange) [T,, G]pl =f[T,, &, (fast exchange). However, for conditions of intermediate exchange, where then and Proton, lH, Relaxation When the ligand is a water proton there is an additional contribution6? to relaxation arising from secondary solvation effects. However, fast exchange prevails under all conditions, so that the simple expression given in eqn ( 5 ) is applicable at room temperature. The relaxation rates for bound water protons are given by7 where p is the effective magnetic moment of the paramagnetic ion, A is the electron nuclear hyperfine coupling constant, W, is the electronic precessional frequency, r is the ion-magnetic nuclear distance and z,l = z,l + z-,1+ r;l (9) z,l = ZMl++Z,' (10) where z, is the rotation time of the complex and z, is the electronic relaxation time.3152 n 9, 0 l 0 N.M.R.Studies of Phosphate Complexes 0 - x - 2 v M 6 - x -d cd n r - 2 h --. M m 2 - x - x - 00. Laurie, J. Oakes, J . W. Rocklie and E. G. Smith P, n P v P I P 31533154 N.M.R. Studies of Phosphate Complexes Water hydration numbers, n, are determined using the equation7 where [R,] is the molar relaxation rate and no refers to the hydration number of a standard sample. The bold print refers to the corresponding ratios for unknown and standard samples. Phosphorus Relaxation On addition of low concentrations of paramagnetic ions to triphosphate solutions, broadening of the two distinct a and /? phosphorus resonances occurs but no spectrum for bound TP is detected, i.e.slow exchange prevails. Consequently, only a single T, spin-spin relaxation time is recorded, given by eqn (4) enabling direct calculation of ~ k l , the frequency of exchange of TP between free and complexed states, for complexes of known stoichiometr y . On the other hand, as the temperature is raised, two distinct spin-lattice relaxation rates [TJ-l, corresponding to a and /? phosphorus atoms and a maximum in the [TI-' us. temperature behaviour, are detected [see fig. 1 (a)-(e)]. Evidently, fast exchange conditions are applicable for spin-lattice relaxation rates, and eqn (2) and (6) are tenable over the temperature range examined. This allows calculation of [ for phosphorus nuclei is given by Assuming the contribution to relaxation is dominated by the dipolar term, the [ T,, [T,, M3-1 = 1.35 x 10-33p2r-6 (12) Results 31P Relaxation The variation with temperature of the normalized? contribution to relaxation rates (f T,, T,)pl for phosphorus nuclei in solutions of TP containing different metal ions is shown in fig.1. Relaxation rates for Mn2+, Cu2+ and Fe3+ were recorded at pH 7 as a precaution to avoid hydrolysis of complexes at higher pH. Nevertheless, relaxation rates measured at room temperature were found to be independent of pH between 7 and 10. Room temperature spin-lattice relaxation rates were also measured for some pyro- phosphate solutions and were found to be comparable to those for corresponding solutions of triphosphate, e.g.for a solution of Mn2+ and PP, relaxation rates were only 1.1-fold greater than that for the /? phosphorus atom in TP solution and were (1.2-1.3)-fold greater than the a phosphorus when corresponding solutions are compared for Ni2+ and Co2+, respectively. The most striking feature of the results in fig. 1 is the observation that the /? phos- phorus nucleus of TP is relaxed more efficiently than the a phosphorus for solutions containing Mn2+ and Cu2+, whereas the converse is true for Fe3+, Co2+ and to a lesser extent Ni2+. Values of T~ can be obtained directly from fig. 1 using eqn (4) and are shown in table 1. The activation energies for ligand exchange have also been calculated in each case using the expression (13) and are included in table 1. T~ = 7; exp (EIRT) t Values off shown in fig.1 (aHe) correspond to 1 : 1 complexes.0. Laurie, J . Oakes, J. W. Rocklife and E. G. Smith 3155 Table 1. Variation of rM with temperature for metal-ion-TP complexes 1 0 3 ~ / T metal ion 3.36 3.27 3.18 3.1 1 3.0 Eau/kJ mol-1 Mn2+ 1.4 x 1.1 x 8.3 x 6.7 x lop6 5 x 23.4 C U ~ + ~ 1.1 x lop5 9.6 x 9.2 x 8.6 x lop6 8 x 4.6 Fe3+b 1.0 x 7.4 x 5.9 x 4.4 x 3.2 x 25.8 Ni2+ 5.3 x 10-4 3.3 x 10-4 2.6 x 10-4 1.7 x 10-4 1.1 x 10-4 34.4 C02+ 1.6 x 10-5 1 x 10-5 6.5 x 10-6 4.8 x 10-6 3.4 x 10-6 32.1 Activation energy. Assuming 2: 1 TP:metal-ion predominates (see later). Table 2. Calculated values of z, for different metal-ion-TP complexes metal ion ZC Mn2+ 5.6 x loplo C U ~ + ~ 4.0 x Ni2+ 5.2 x 10-l' CO2+ 9 x 10-12 Fe3+u 1.2 x lop1' Assuming 2: 1 TP:metal-ion. The exchange rates, zkl, are slightly lower in magnitude than those for water exchange between bound and free envir~nments,~~ but follow the same general pattern, i.e.Cu > Mn > Co > Ni > Fe. Following the determination of zM, the spin-lattice relaxation rates for bound phosphorus nuclei, [&I$, can be calculated using eqn (2). If it is assumed that this rate for the phosphorus is twice that for the a phosphorus atom in a bidentate metal complex and that they are equal in a tridentate complex, then the proportion of bidentate complex in equilibrium with the tridentate complex can be estimated. On this basis it is calculated that 60% of the Mn2+ and 80% of Cu2+ exist in bidentate form. On the other hand, the results for Co2+, Ni2+ and Fe3+ favour the tridentate form as the dominant species.? Assuming that the central phosphorus atom (p> is situated 3.3 A distant from the metal then z, can be calculated for all complexes from eqn (1 2).These values are listed in table 2. The application of eqn (12) for complexes of Mn2+ and Cu2+ is beyond question, since the dipolar term dominates15 and indeed z, = z,. Accordingly, there is close agreement between values of z, for the two complexes. However, values of z, are considerably higher (approximately 10 fold) than the corresponding z, calculated from water-proton relaxation data (see later). This reflects the different degrees of rotational freedom15 for bound water and bound TP molecules. Although the relevant correlation time for complexes with Co2+, Ni2+ and Fe3+ is z, (table 2), in fact [&];;r]- is dominated by scalar intera~ti0ns.l~ t On formation of complexes with diamagnetic cations, Ca2+, Mg2+ and Li+ the central phosphorus resonance is shifted5 more than the outer phosphorus resonance, implicating the presence of bidentate complex.3156 N.M.R.Studies of Phosphate Complexes ""1 v 20 - v 8 9 10 PH Fig. 2. Dependence of water proton relaxation time (&) on pH for solutions of PP and TP complexes of Co2+, Cu2+, Ni2+ and Fe3+: 0, CoTP; ., CuPP; 0, CuTP; a, CoPP; A, NiTP; A, NiPP; V, FeTP; V, FePP. lH W ater-proton Relaxation Water-proton spin-spin relaxation times as a function of pH for solutions of PP and TP complexes of Co2+, Cu2+, Ni2+ and Fe3+ are shown in fig. 2. The observed relaxation times are higher than those of free (aquated) transition-metal ions (table 4) owing to complexation.However, complexation is not complete for cobalt complexes until pH 8 (fig. 2). Similarly, not all the phosphate groups are deprotonated3v l6 until the pH rises above 8. Nevertheless, like in aminopolyphosphonates,8 these protons are too distant from the metal ion to contribute to relaxation. The spin-spin relaxation times (T,) for the TP complex of Mn2+ are given in fig. 3 for two different ratios of TP to Mn2+. The spin-lattice relaxation time is independent of the TP:Mn2+ molecular ratio and this is also shown in fig. 3. Inspection of this figure shows that complexation of Mn2+, like cobalt, is incomplete until around pH 8. It is also clear that the electronic spin relaxation time, z,, which dominates T,, strongly depends upon the ligand-to-metal ratio.Both these observations have been independently confirmed17 by monitoring changes in the linewidths of the e.s.r. spectrum of Mn2+. relaxation time, measured in aqueous solution, for the PP complex of Mn2+ is similar to that for the TP complex at the same high ligand-to-metal ratio and is omitted for clarity of presentation. However, the Although generally, for a given metal ion, the relaxation time for a PP complex is lower than that of the corresponding TP complex, an intriguing situation arises for Cu2+, The for the PP complex is illustrated in fig. 3.0. Laurie, J. Oakes, J . W. Rocklife and E. G. Smith 3157 3 4 5 6 7 8 9 10 11 PH Fig. 3. Dependence of water proton relaxation times (q, T,) on pH for solutions of PP and TPcomplexes of Mn2+.0, T, (TP:Mn = 2: 1); m, T, (TP:Mn = 100: 1); @, (TP:Mn ratio independent); 0, (PP:Mn = 100: 1). 2c 15 v1 --- 2 i a 5 0 \ \ i 1 I 1 I I b 0 10 20 30 40 50 60 experimental frequency/MHz Fig. 4. Frequency dependence of relaxation rate, [TJ1, for solutions of Cu2+ and Mn2+ complexes with TP: ., MnTP; @, CuTP.3158 N.M.R. Studies of Phosphate Complexes Table 3. Calculated values of primary relaxation rate, [ T J l , and primary molar relaxation rate, Rl,p, for complexes concentration complex /mol dmP3 [ TI- 1 / s- 1 K1;I Rl,p x 10-2 CuTP CUPP NiTP NiPP CoTP COPP FeTP FePP MnTP MnPP 6 x lop3 6 x 2 x 102 2 x 102 2 x 102 2 x 102 6 x lo3 6 x lo3 1 x 10-3 1 x 10-3 3.03 2.33 8.00 1.72 2.32 10.0 16.7 33.3 10.0 9.52 2.28-28.3 1.58-1.13 7.0-6.6 8.96-8.0 1.241.14 1.84-1.74 14.9-1 3.4 3 1.5-30.7 8.37-7.94 8.85 ~ 4.1-3.1 2.6-1.9 3.5-3.3 4.844.0 0.62-0.57 0.92-0.8 7 25-22 53-5 1 84-79 89 a The primary relaxation rate, [ q,p] is obtained by subtracting secondary solvation contributions (table 5 ) and the diamagnetic contribution from free water molecules. Table 4.Calculated molar relaxation rates, Rl,p, for free ionsa concentration complex /mol dm-3 [ T]-l/s-l Rl$ x 10-2 CU2+(H20), 6 x lop3 4.55 4.25 7.6 x lo2 Ni2+(H20)6 2 x 13.33 13.03 6.5 x lo2 Co(H20)6 2 x 10-2 2.96 2.66 1.3 x lo2 Mn2+(H20), I x 10-3 9.09 8.73 8.8 x lo3 Fe3+(H20)," 6 x 66.6 66.3 11.1 x 103 a These were recorded at pH 5 except for Fe3+ at pH 1. where the converse is true, indicating that the PP complex is less hydrated than that of TP.This is considered in more detail later. The frequency dependence of the spin-lattice relaxation rate for solutions of Cu2+ or Mn2+ complexes with TP are shown in fig. 4. After taking into account secondary solvation contributions, values of the rotational correlation time, z, can be obtained by comparison of relaxation rates at several different pairs of frequencies. This produced values for z, of 4.5 x 10-l' and 4.0 x s for Mn TP and Cu TP complexes. Values of z, for the respective aquated ions determined earlier6? are 3 x s. Using eqn (8), it can be shown that z, the electron spin relaxation time, for Mn2+ decreases from 1.5 x to 5 x s upon complexation with TP, for a Mn2+: TP ratio of 1 : 2. This value is still higher than that found for aminopolyphosphate8 complexes of Mn2+ (1.62 x s) for two reasons: (i) it occupies fewer coordination positions and (ii) it lacks nitrogen donors.Molar relaxation rates have been calculated for TP and PP complexes of all the transition-metal ions at pH 9 and the results are given in table 3. Molar relaxation rates for the free aquated ions are given in table 4. Secondary solvation contributions to relaxation rates have been estimated using procedures adopted in earlier assuming coordination by two, three and four phosphate groups (table 5). Using all the and 3.6 x0. Laurie, J . Oakes, J . W. Rocklife and E. G . Smith 3159 Table 5. Calculated molar secondary solvation contributions to relaxation rates from1$ 7 * corresponding EGTA complexes EGTA complex 2P 3P 4P CuEGTA 7.5 x lo1 1.1 x lo2 1.5 x lo2 NiEGTA 3.7 x 10' 5.6 x 10' - CoEGTA 9.0a 1 4a 5 FeEGTA 2.55 x lo2 3.8 x lo2 5.0 x lo2 MnEGTA 8.5 x lo2 1.28 x lop3 - a This contribution has been adjusted, since the values obtained for the EGTA complex are too owing to changes in p2 and 7, on complexation with EGTA but not TP.Table 6. Calculated values of hydration numbersa for ion-TP/PP complexes MnTP MnPP CuTP CUPP NiTP NiPP CoTP COPP FeTP FePP 0.95-0.9 1.01 0.54-0.46 0.34-0.4 1 0.69-0.62 0.71 0.47 0.5 5-0.5 1 0.48-0.44 0.23-0.22 1 1 1 1 1 1 1 1 1 1 1 1.5 1 1.5 I 1.4 1 1.2 1 1 1 1 1 1 1 1 1.3 0.87 1.3 0.87 3.8-3.6 4.04 2.3-1.8 2.0-1.5 3.43.1 4.1-3.7 3.g2.6 4.2 2.1-2.0 4.3-3.2 a Values of p2, r6 and fi(z) are taken from earlier reference^^-^ excepting for Mn2+ and Cu2+ complexes: f,(z) = 32, + 7.r,/( 1 + W,Z z:). data, water hydration numbers have been calculated for all complexes using eqn (1 1) and the results are given in table 6 .These values are accurate to within lO-15%, with the exception of Fe3+ complexes, where there is uncertainty in the values of r.7 Further details of the calculation can be found in table 6 . Discussion Pyrophosphate Complexes The observation of single 31P spin-lattice relaxation (TJ behaviour for solutions of PP containing Mn2+, Co2+ and Ni2+, together with values for water hydration numbers of ca. 4 (table 4), indicate that 1 : 1 bidentate complexes predominate under the conditions t r increases by 0.12 A upon binding' of EDTA, although the increase is expected to be less for the PP complex, partly owing to higher water coordination numbers.3160 N.M.R.Studies of Phosphate Complexes investigated. On the other hand, the low value of n obtained for the PP complex of Cu2+ suggests a structure, uiz. 0 I 0 I I \ \ -. 0 .. I .. 0 - - 0-P-0 0-P-0 I 0 I 0 in which two PP molecules coordinate* to Cu2+. This supports earlier suggestions18 that this complex predominates in the pH range 7-10. Low concentrations of polynuclear species Cu,(PP), indicated from other worklg cannot be ruled out here. Evidently, the Fe3+ complex of PP has a higher hydration number than that for Cu2+ (table 6), suggesting a lower tendency to form 1 : 2 complexes. This may partly be associated with its higher charge and smaller ionic ratio (0.64 vs.0.69 for Cu2+). Triphosphate Complexes The hydration numbers for TP complexes of Mn2+, Co2+ and Ni2+ (table 6) and the relative magnitudes of spin-lattice relaxation for a and a atoms for each complex can be interpreted in terms of the equilibrium? 0 0 0 I I I 0 0 0 I I I 0-P-0-P-0-P-0 0-P-0-P-0-P-0 I 0 I 0 I 0 I 0 I 0 I 0 M n+ M n+ which, for Mn2+ (like Ca2+, Mg2+ and Li+) lies predominantly to the left (bidentate species), whereas the equilibrium for Co2+ and Ni2+ lies more in favour of the tridentate species. (The observation of higher relaxation rates for the a phosphorus atom than the phosphorus atom, together with a hydration number, n < 3 for the Co TP3- complex may suggest the simultaneous binding of two oxygen atoms of an end phosphate group similar to that proposed for lanthanide complexes.2o) The results for the copper complex are best explained in terms of an equilibrium between CuTP and the predominating Cu(TP), species, viz. * The low value of n may also suggest longer Cu-OH, bond lengths than in the aquo-ion.t Examination of Co2+ and Ni2+ complexes of TP by electronic absorption spectroscopy indicate octahedral symmetry. zo0. Laurie, J . Oakes, J . W. Rocklife and E. G . Smith 3161 0 0-P-0 0 i i 0 I I I I I 1 0-P-0 0 0-P-0 0 0-P-0 0 c / c / #- / .-- :cuz+ 1 1 1 1 0-P-0 0 0-P-0 0 This assignment fits in well with (i) the prediction that Cu(TP), predominates3 over CuTP at pH > 9 and (ii) the finding that the hydration number is marginally higher than that obtained for the corresponding PP complex.A polynuclear complex of stoichiometry Cu,(TP), has previously been identifiedlg and is possible in the present investigation. The situation for TP complexes of Fe3+ is less clear but is likely to involve small polynuclear complexes in which TP coordinates as a tridentate and possibly a tetradentate 1igand.t We thank Dr I. D. Campbell, Oxford University, for helpful discussions in the early phases of this work. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 H. Ellison and A. E. Martell, J. Znorg. Nucl. Chem., 1964, 26, 1555. G. Andergg, Helv. Chim. Acta, 1965,48, 1712. P. E. Sturrock, E. D. Loughran and J. I. Walters, Znorg. Chem., 1962, 1, 467. S. M. Lambert and J. I. Walters, J. Am. Chem. SOC., 1957, 79, 5606. M. M. Crutchfield and R. R. Irani, J. Am. Chem. SOC., 1965, 87, 2815. 3. Oakes and E. G. Smith, J. Chem. SOC., Faraday Trans. I , 1983, 79, 299. J. Oakes and E. G. Smith, J. Chem. SOC., Faraday Trans. I , 1983, 79, 543. 3. Oakes and E. G. Smith, J. Chem. Soc., Dalton Trans. 1983, 601. J. Oakes and C. G. van Kralingen, J. Chem. SOC., Dalton Trans., 1984, 1133. T. J. Swift and R. E. Connick, J. Chem. Phys., 1962, 37, 307. Z. Luz and S. Meiboom, J. Chem. Phys., 1964,40, 2686. R. A. Dwek, Nuclear Magnetic Resonance in Biochemistry (Clarendon Press, Oxford, 1973). F. A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry (J. Wiley, New York, 4th edn, 1980). F. F. Brown, I. D. Campbell, R. Henson, C. W. J. Hirst and R. E. Richards, Eur. J. Biochem., 1973, 38, 54. H. Sternlicht, R. G. Schulman and E. W. Anderson, J. Chem. Phys., 1965, 43, 312. A. Johansson and E. Wannima, Talanta, 1963, 10, 769. M. Sharples, personal communication. J. I. Watters and A. Aaron, J. Am. Chem. SOC., 1953, 75, 61 1. M. Bobtelsby and S. Kertes, J. Appl. Chem., 1955, 5, 675. C. G. van Kralingen, personal communication. M. S. Nieuwenhuizen, J. A. Peters, A. Sinnema, A. P. G. Keiboom and H. van Bekkum, J. Am. Chem. SOC., 1985, 107, 12. Paper 5/2162; Received 9th December, 1985

ISSN:0300-9599    

DOI:10.1039/F19868203149

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Furuday Trans. I, 1986’82, 3 163-3 173 Ordered Solution Structure of a Monodispersed Polystyrene Latex as studied by the Reflection Spectrum Method Tsuneo Okubo Department of Polymer Chemistry, Kyoto University, Kyoto 606, Japan Ordered lattice structures, face-centred cubic (f.c.c.) and body-centred cubic (b.c.c.) lattices, are systematically analysed by the reflection spectrum measurements of monodispersed polystyrene latices in deionized solution. The nearest-neighbour interparticle distance (20,,,) is determined from the wavelength at the reflection peaks. 20,,, decreases with increasing latex concentration and is smaller than the calculated average values for the f.c.c. or b.c.c. distribution (20,). The results are consistent with a ‘two-state structure’ which has ordered and disordered regions.The f.c.c. and b.c.c. lattices are stable at high and low concentrations, respectively. The b.c.c. structure often appears at a high solution temperature and in the presence of a small amount of foreign salt. The size of the ordered region is estimated from the width of the reflection pattern and also from microscopic photo- graphs. The number of crystal layers is estimated to be several hundred and depends on latex concentration and solution temperature. The regular lattice structure of monodispersed polystyrene latices has been investigated by a number of researcher~.l-~ Our work with ultramicroscopic observations has clarified some fundamental properties of solution structure. We found that the observed interparticle distance (20,,,) in the ordered state was smaller than the distance calculated from the concentration of the particles by assuming a uniform distribution throughout the solution (2D0).*~ This result strongly indicates that ordered regions where particles are distributed regularly such as in a ‘crystal’lO9 l1 or ‘liquid crystal’12 coexist with disordered regions where particles exhibit vigorous random motions.l39 l4 We have called this situation a ‘ two-state 9 7 l5 The ultramicroscope method, which was first applied to latex solutions by ha chis^,^ is very effective for studying the distribution of particles.However, the latex particles must be large enough to allow direct observation by visible light, i.e. the particles should be >ca. 250 nm in diameter.On the other hand, optical methods are useful for studying solution structure of smaller particles.6T ‘7 lo, l1, l 6 , l7 Luck et al.,lo9 l1 Krieger and ~ o w o r k e r s , ~ ~ - ~ ~ Hoffman20 and van de Ven and c ~ w o r k e r s ~ ~ - ~ ~ extensively studied the reflection spectrum of latex especially at high concentrations (ca. 10-40 vol % ). In this paper we discuss the reflection spectrum of latices (80-300 nm in diameter) in solution at relatively low concentrations (0.3-5 vol % ). Following a discussion of the interparticle distances in the f.c.c. and/or b.c.c. lattices and two-state structures as a function of polymer concentration, salt addition and solution temperature, the size of a liquid-crystal-like ordered region is estimated by Scherrer’s equation and also by the direct microscopic observation.Experiment a1 Monodispersed Latex Particles The D1 series of latex, monodispersed polystyrene latices was purchased from Dow Chemical Co., and the N-150 series from Sekisui Chemical Co. The characteristics of 31633164 Ordered Structure in Polystyrene Latices Table 1. Properties of latices used diameter charge no. charge density latex /nm per particle /pC cmP2 ~ DlC25 85+6 3 010 2.1 DlC27 91 +6 4 550 2.8 DlB22 109 f 3 8 480 3.6 N-150 150f4 15 200 3.3 DlB72 173 & 7 6 060 1 .o DlB34 246 4 22 400 1.9 Fig. 1. Optical set-up for the reflection spectrum measurements. these series are listed in table 1 . The values of the particle diameters and their dispersions were obtained by electron microsocopy by the manufacturers.The charge densities of the particles were determined by conductometric titrations with a Wayne-Kerr auto- balance precision bridge, model B33 I mark 11. The charge number per particle is defined as the number of sulphate residues detected by conductometric titration with NaOH. The charge density was obtained from the charge number multiplied by the electric charge and divided by the surface area of the particle. These latices were carefully purified several times by an ultrafiltration cell (model 202, membrane: Diaflo XM300, Amicon Co., Lexington, Mass., U.S.A.). Then the samples were treated on a mixed bed of cation- and anion-exchange resins [Bio-Rad, AGSOl-X8(D), 20-50 mesh] for at least 7 days. The resulting solution was believed to contain only the macroions and their counterions (protons). All latex samples now displayed brilliant iridescent colours and their concentrations were ca.8 vol% in a stock solution bottle. Water used for the purification and for solution preparation was deionized by using cation- and anion- exchange resins (Amberlite IR-120B and IRA-400) and further purified by the Milli-Q reagent-grade water system. Its specific conductance was 8 x l2-l cm-l at room temperature (ca. 20 "C). Reflection Spectral Measurements The reflection spectra were recorded on a modified fluorescent spectrophotometer (type FS-401, Union Giken, Hirakata, Japan). The optical system is shown in fig. 1. The sensitivity of the photomultiplier was so high that the widths of the slits were typically between 0.05 and 0.2 mm.The optical path length of the cell was usually 1 mm, but several other kinds of cells were used, as will be described later. Ultramicroscopes and Photography Photographing the latex solution was carried out with a reversed-type metallurgical microscope, Axiomat IAC (Carl Zeiss, West Germany), and with a polarized metallurgicalT. Okubo 3165 wavelengt h/nm Fig. 2. Reflection spectra of DlC27 latex solution. 1.46 vol % . (a) quartz cell, (b) triangular cell, (c) square cell. microscope, Diaphot-TMD (Nikon, Japan). The camera used to take close-up pictures was a Nikon FE2 with a zoom close-up lens (Kenko Co., Tokyo). The solution cell used was the same as that described previously.8 Results and Discussion Iridescence of the Latex Solution The latex solution in the optical cell looked transparent, although the same sample in a text tube displayed iridescence and was milky and opaque.Other important features seen were as follows. First, the colours of the reflected light (with Bragg diffraction) and of the transmitted light were different. Secondly, in some cases a striped pattern was observed. This pattern appeared immediately when the latex solution was poured through a constriction pipette (Bie & Berntsen, Denmark, 200 pm3) into the cell from the stock solution. This orientation effect caused by the stress of the solution flow at the time of pouring could be maintained for a long time, as is often the case with liquid crystals, Thirdly, we often observed brilliant and sparkling patches, which are called ‘islands’, ‘crystallites’ or ‘liquid crystals’ in the s o l ~ t i o n .~ ~ ~ 26 These islands correspond to the ordered regions described previously. As already reported by Vanderhoff et aZ.,12 the islands seem to flicker and disappear immediately when the solution is shaken. However, they reappear within several seconds when the solution is left to stand. Our group is now undertaking a kinetic study of this phenomenon, i.e. the determination of the structural relaxation time by using fast reaction techniques such as U.V. stopped-flow and conductance stopped-flow methods.27 Reflection Spectra with Various Optical Cells In order to check the reliability of our spectral system, reflection patterns for a DlC27 latex solution of 1.46 vol% were obtained with several kinds of quartz cells (see fig.2).3166 Ordered Structure in Polystyrene Latices Type (a) is a quartz cell (1 mm path-length), whereas types (b) and (c) are triangular- and square-type cells, respectively. The peak positions for cells of types (a) and (b) were close to each other (546 and 544 nm, respectively), while that for type (c) was shorter (406 nm). Following the treatment of Hiltner and Krieger,18 the interparticle distance in the face-centred cubic (f.c.c.) or body-centred cubic (b.c.c.) lattice (2Dexp, , or 2Dexp, b ) is given by eqn (1) for types (a) and (b) and in eqn (2) for type (c): where Amax denotes the wavelength of the primary reflection peak. The refractive indices of sample solution, air and quartz glass are designated by n,, n, and ng, respectively.In this work n, was taken as 1.333 at 25 "C, which is the refractive index of water,2s because the measurements were carried out in relatively dilute concentrations of latex. n, was 1 .OO. Although ng decreased with the wavelength of light,28 a mean value of 1.455 was taken. The incident angle 0 was 45" in this work. Note that the reflection spectrum gives information on the Bragg distance only, which is the distance between the parallel plates. There is no information on the structure of the two-dimensional plates themselves. Evidence of the existence of f.c.c. or b.c.c. lattices cannot therefore be obtained from the reflection spectrum alone. However, many reports have supported the existence of the f.c.c. (or hexagonal close-packed) and b.c.c. structures hitherto.l-s* 16-26 Thus our method of calculating 2Dex, is believed to be sound.The calculated values of 2DeX, were 296, 295 and 293 nm for cell types (a), (b) and (c), respectively. The agreement is excellent. Cell type (a) was also used to obtain the transmitted-light spectrum, and the wavelength of the extinction peak was 644 nm for the same latex sample, from which 2Oex, was evaluated as 296 nm. The experimental error is believed to be within f 3 7;. Details of the transmitted-light spectra! measurements will be reported in a separate paper.29 Concentration Dependence of the Nearest-neighbour Interparticle Distance (20,,,) Fig. 3 shows the reflection (or diffraction) pattern of the latex solution. The number- averaged value of the diameter of the latex used was 109 & 3 nm and the concentrations were between 0.63 and 6.31 vol% .Generally speaking, the shape of the reflection pattern is a single peak, double peak or single peak with a shoulder. The two peaks corresponding to the wavelength were always close together, with a difference of only 1.03 in the ratio of the wavelengths of the two peaks. This difference supports the idea that the peak appearing at longer wavelength is ascribed to the f.c.c. lattice and the shorter one to the b.c.c. lattice in the ordered latex solutions. Eqn (3) and (4) relate the nearest-neighbour interparticle distance (2D0, , or 2D0, b), the Bragg distance (d, or db) and the lattice constants (a, or a,): (2D0, ,/0.707) = 2/3 d, = a, (2D,, ,/0.866) = 2/3db = ab. (3) (4) Subscripts f and b denote the f.c.c.and b.c.c. lattices, respectively. From eqn (3) and (4) the ratio df/db (or a,/ab) is derived as df/db = 2D0, f/2D0, b = 1.0284. ( 5 ) The peaks denoted A, C and D in fig. 3 belong to the b.c.c. lattice structure, and the shoulder to the f.c.c. structure. Note that the reproducibility of the profile of the diffraction pattern was not very good, although the peak positions never changed. Thus we obtained spectra repeatedly by using different cells. Secondary peaks, although weak, appeared at half the wavelength of the primary peak. This means that the (1, 1 , l ) planeT. Okubo 3167 D I 300 4 I\ I \ A wavelength/nm Fig. 3. Reflection spectra of DlB22 latices at various concentrations (~01%) at 25 "C. (A) 0.63, (B) 0.88, (C) 2.10, (D) 3.16, (E) 4.21 and (F) 6.31.of the f.c.c. lattice or the (1,1,0) plane of the b.c.c. lattice was parallel to the wall of the cell. The number of planes which contribute to the Bragg diffraction is not known. However, this number must be large, probably several hundreds or more, because the latex sample used in this work is relatively transparent. As is clear in fig. 3 , the peak position shifted to the shorter region of the wavelength as the latex concentration increased. Typical experimental results for 2Dexp, and 2Dexp, b evaluated from the peak positions using eqn (1) and (2) are shown in fig. 4 for the DlB22 latex as a function of the polymer concentration. The f.c.c. or b.c.c. lattice was assigned from the relative position of the reflection peak, and 2Dexp, and 2Dexp, b values are given by open circles and crosses, respectively.When the assignment was impossible, 2Dexp is shown by triangles. Note that the f.c.c. structure has been reported to be more stable at higher concentrations; this was first by Williams and Crandall,30 then by Hachisu et al.5931 and later by Yoshiyama et al.32 Uniform f.c.c. or b.c.c. distributions were assumed when the average values of 2D,, and 2D,, b were calculated from the latex concentration. The inequality relations 2Dexp, < 2D,, and 2DeXp, b < 2D,, b hold and give strong support to the existence of a two-state structure in solution which consists of ordered regions of a higher solute density and less dense disordered regions. A further comparison of 2Dexp and 20, for other latices is made in fig.5, in which the relative difference (2D,xp-2D,)/2D, is given. The difference seems to become larger with decreasing concentration and decreasing size of the latex particles. However, its dependence on the surface charge density of the latex is not yet clear. The inequality 2D,,, < 20, is thus believed to hold for latex solutions, although it was obtained only by ultramicroscopic and spectrophotometric methods. For these measurements, a cover glass or quartz glass is always used, and several latex layers from the surface of the cell are observed. Therefore, the specific interaction between the glass surface and the latex particles must be carefully examined before there can be any quantitative discussion of interparticle distance. Deformation of the ordering or a different stacking order has not yet been clarified.Recently, Pieranski et al. reported the existence of a different stacking mode in latex ordering in the thin layer (not in the thick layer) between two glass plates.2s* 33 However, the mean interparticle distance was the same both in the square layers and in the triangular layers. Furthermore, as is seen in fig. 2, the same value of 2Dexp was obtained from various kinds of observation cells,3168 Ordered Structure in Polystyrene Latices 600 500 \ f u 9 c 2 400 - .C.( c1 c 8 # * 300 200 1 I 2 4 6 [DlB221(vol%) Fig. 4. Concentration dependence of interparticle distance of DlB22 latex. 0, 2 0 , , , , , ; x ? 2oexp, b; A, 2oexp, f Or 2Oexp, b; (-1 ~ O O , f ; (----I 2 0 ~ , b. 0, -0.11 X >c.x x X nT.Okubo 3169 h m U .- E -2 W A E c U .C U .C I I I I 3 5 ~ 640 650 660 670 680 690 waveleng t h/n m Fig. 6. Change in reflection spectra with solution temperature (in "C). [DlB22] = (a) 1.05 and (b) 1.58 vol% . i.e. types (a) (optical path 1 or 2 mm), (b) and (c). At present, therefore, we believe that the glass-latex interaction does not influence 2DeXp values. Further comments are necessary concerning the interaction between latex particles in the ordered state. Hachisu et al.34 proposed that the ordering is due to the effective hard-sphere model, where the Debye length is added to the true radius of the latex particle. This model is consistent with the Alder transition with the electrostatic repulsive interaction between particles, and with the Derjaguin-Landau-Verwey-Overbeek theory (DLVO theory) of colloidal stability35 that the interparticle interaction in two-particle systems consists of electrostatic repulsion and van der Waals attraction. The potential energy of attraction and repulsion between latex particles arises from their charge, collision diameters and thermal motion.In the disordered region vigorous thermal motion of the latex particles including the Debye length was observed by Hachisu et al.5 and by o ~ r s e l v e s . ~ ~ ~ l4 We believe that the main cause for the existence of the ordered region is due to an apparent interparticle attraction in the ordered region caused by a slight excess microstructural pressure from the disordered region. In other words, repulsive forces between particles lead to a coexistence of gas-like and solid-like 36 Thus the microscopic two-state structure, which is the coexistence of ordered solid and disordered liquid states, can be explained by the effective hard-sphere model and/or DLVO Moreover, the two-state structure, where solid-like and gas-like partitions coexist in a deionized latex solution, is similar to the theory of significant structure of 38 which also considers gas-like and solid-like molecules.However, there are some differences between the two, such as the fact that the repulsion is a predominant feature in the former. Influence of Solution Temperature Reflection spectra were measured in the range 15-65 "C. Typical patterns are shown in fig. 6(a) and (b) and 7(a) and (b). The spectra were taken ten minutes after the sample3170 Ordered Structure in Polystyrene Latices n E c, .- 2 W x t: Y .I c, .r( I 1 I I (a> 610 620 630 640 wavelength/nm I I I I I ( b ) l 5 k 0 Fig.7. Change in reflection spectra with solution temperature (in "C). [DlB22] = (a) 2.10 and (b) 6.31 volx . solution, thermostatted with circulated water, had reached a given temperature. The spectrum did not change within this time. However, at elevated temperatures the ordering structure gradually broke up with time, owing to the convection flow of the latex solution in the quartz cell. In our experiments an accurate 'melting' transition temperature was not determined. However, the melting temperatures are believed to be usually around 60-80 "C. The diffraction pattern is sensitive to changes in solution temperature and polymer concentration. The following tendencies were clearly seen : (1) the reflection intensity decreased with rising temperature and ( 2 ) the f.c.c.lattice structure, which is stable at low temperatures, changed to a b.c.c. lattice with rising temperature. The 20,,, values of the D 1 B22 and D 1 C27 lattices at various temperatures and latex concentrations are given in tables 2 and 3. The changes in the refractive index of the solution with temperature were taken into account in the estimation of 2D,,,. However, 2D,,, was insensitive to solution temperature, which is contrary to the results of the optical microscopy study.9 The size of the ordered regions or islands, L, was estimated by Scherrer's equation from the half-width of the reflection spectra:39 L z A S 1 (6) where S is the scattering vector, given by 2 sin @/A, and A S denotes S, - S,, with S, and S, referring to the largest and the smallest wavelengths corresponding to the half-width of the reflection peak.Thus the number of layers of latex, N , is given by L/d, where d is the Bragg distance. The values of N are also listed in tables 2 and 3. They vary between 50 and 200 and seem to decrease with rising temperature and with increasing latex concentration; they are also based on several assumptions and much uncertainty. However, we succeeded in taking photographs of the two-state structure by using a metallurgical microscope with a modified illumination method, whereby light hits theT. Okubo 3171 Table 2. Estimation of crystal size by Scherrer's equation for the D I B22 latex solution [latex] (~01% ) T/"C 1.05 1.58 2.10 6.3 1 15 2DeXp/nm 414(f) 369 (f), 362 (b) N 220 ( f ) 210 (f) 25 2Dexp/nm 414(f) 369 (f), 361 (b) N 230 (f) 140 (f) 35 2DeXp/nm 414(f) 369 (f), 361 (b) N 230 (f) 180 (f) 45 2DeX,/nm 414 (f+ b) 369 (9,36l(b) N 160 (f) 55 2Dex,/nm 413 (f+b) 370 (0,360 (b) 140 ( f ) , 130 (b) N 65 2Dex,/nm 412 (b) 360 (b) N - 110 (b) - - 339 (f) 240 (f+b) 190 (0 339 (f) 240 (f + b) 190 (f) 339 (f) 240 (f + b) 150 (f) 340 (f) 240 (f + b) 150 (f) 339 240 (b) 130 61 (b) - 241 (b) 56 (b) - - - - - Table 3.Estimation of crystal size by Scherrer's equation for the DlC27 latex solution [latex] (~01% ) T/OC 0.329 0.658 1.10 2.19 15 25 35 45 55 65 2Dexp/nm 494 382 N > 140a 220 2Dex,/nm 494 382 N >120a 210 2Dexplnm - 382 N 190 2Dexplnm - 382 110 N 2Dexplnm - 382 N 210 2Dexplnm - 383 210 N - - - - 323 140 323 I40 323 1 40 323 140 323 120 324 I30 a Calculated from secondary peaks.cover glass of the sample solution cell diagonally. From the photographs we could estimate the size of the islands and obtain N values very similar to those resulting from the reflection spectrum method. Photographic details are presented later. Influence of a Foreign Salt Fig. 8 shows the effect of salt on the reflection pattern. In the absence of salt there are two peaks corresponding to the f.c.c. and b.c.c. structures. However, when a small amount of KCl, 3.3 x lod6 mol dm-3, was added, the peak for the f.c.c. structure disappeared completely and the remaining peak was assigned to the b.c.c.structure. A further addition of KCl shifted the peak to larger wavelengths, and all the peaks disappeared completely when the addition of KC1 was > 1.33 x mol dm-3.3172 Ordered Structure in Polystyrene Latices n c U .3 2 v x c c1 .3 Y .- I . I . : . . . . . : . . . . . . . . 1 I I 600 620 640 E wavelength/nm ;0 Fig. 8. Reflection spectra of D1 B22 latex solution with KC1 at 25 "C. [D 1 B22] = 1.94 vol % ; (-) 0, (----) 3.33 x lop6, (.. ..) 6.67 x and (----- ) 1.00 x lop5 mol dmp3. Microscopic Observation of a Two-state Structure and Size Estimation of the Ordered Region Vanderhoff et a l l 2 first reported the existence of the ordered region of liquid crystals or crystallites in a monodispersed polystyrene latex in deionized aqueous solution.We used several methods to obtain microscopic photographs of our latex solutions : (1) using a metallurgical microscope with a standard illumination device (type 1 ), (2) a metallurgical microscope with diagonal illumination (type 2), (3) with a polarized metallurgical microscope (type 3) and (4) close-up photography of the cover glass containing latex solution (type 4). In type 1, the particles in solution were illuminated from the bottom of the solution cell through the object glass, as described in detail by Hachisu et aZ.5 In type 2, the cover glass of the solution cell was illuminated diagonally with a tungsten lamp through the object glass (incident angle ca. 45"). Although the size distribution is rather dispersed, it can generally be concluded that the size decreases with increasing latex concentration.A rough estimation of the number of latex layers, N , was made from the photographs. The results for the DlC27 latex are listed in table 4. The values of N derived by this method agree well with those obtained from the reflection spectrum method, although in a qualitative sense. The size of the ordered regions also decreased with increasing latex concentration (0.5-5 vol % ) for DlC25, DlB22 and DlB72 lattices. I thank Professor N. Ise of this Department for his encouragement of this work.T. Okubo 3173 Table 4. Estimation of N from the microscopic observation of ordered regions of the D 1 C27 latex solution [latex] diameter of (~01% ) islands/,um N NU 0.22 180f70 400 - 0.33 140 & 70 350 - 0.66 95 f 50 300 210 1.10 40 f 30 150 140 2.19 20f 15 100 140 a From reflection spectra (taken from table 3 of this paper).References 1 R. H. Ottewill and J. N. Shaw, Discuss. Faraday SOC., 1966,42, 154. 2 J. C. Brown, P. N. Pusey, J. W. Goodwin and R. H. Ottewill, J. Phys., 1975, 8, 664. 3 J. W. Goodwin, R. H. Ottewill and A. Parentich, J. Phys. Chem., 1980, 84, 12; 1580. 4 P. A. Hiltner, Y. S. Papir and I. M. Krieger, J. Phys. Chem., 1971, 75, 12; 1881. 5 A. Kose, M. Ozaki, K. Takano, Y. Kobayashi and S. Hachisu, J. Colloid Interface Sci., 1973,44, 330. 6 N. A. Clark, A. J. Hurd and B. J. Ackerson, Nature (London), 1979, 281, 6; 57. 7 M. Tomita, K. Takano and T. G. M. van de Ven, J. Colloid Interface Sci., 1983, 92, 367. 8 N. Ise, T. Okubo, M.Sugimura, K. Ito and H. J. Nolte, J. Chem. Phys., 1983, 78, 536. 9 N. Ise, K. Ito, T. Okubo, S. Dosho and I. Sogami, J. Am. Chem. SOC., 1985, 107, 8074. 10 W. Luck, M. Klier and H. Wesslau, Ber. Bunsenges. Phys. Chem., 1963, 67, 75; 84. 11 W. Luck, M. Klier and H. Wesslau, Naturwissenschafren, 1963, 50, 485. 12 W. Vanderhoff, H. J. van de Hul, R. J. M. Tausk and J. Th. G. Overbeek, Clean Surfaces: Their Preparation and Characterization for Interfacial Studies, ed. G. Goldfinger (Marcel Dekker, New York, 1970). 13 N. Ise, T. Okubo, K. Ito, S. Dosho and I. Sogami, J . Colloid Interface Sci., 1985, 103, 292. 14 N. Ise, T. Okubo, K. Ito, S. Dosho and I. Sogami, Langmuir, 1985, 1, 176. 15 N. Ise and T. Okubo, Acc. Chem. Res., 1980, 13, 303. 16 I. M. Krieger and F.M. ONeill, J. Am. Chem. SOC., 1968, 90, 3 1 14. 17 I. M. Krieger and F. M. O'Neill, J. Am. Chem. SOC., 1968, 90, 12. 18 P. A. Hiltner and I. M. Krieger, J. Phys. Chem., 1969, 73, 2386.. 19 P. A. Hiltner, Y. S. Papir and I. M. Krieger, J. Phys. Chem., 1971, 75, 1881. 20 R. L. Hoffman, Trans. SOC. Rheol., 1972, 16, 1; 155. 21 M. Tomita, K. Takano and T. G. M. van de Ven, J. Colloid Interface Sci., 1983, 92, 367. 22 M. Tomita and T. G. M. van de Ven, J. Colloid Interface Sci., 1984, 99, 374. 23 M. Tomita and T. G. M. van de Ven, J. Colloid Interface Sci., 1984, 100, 112. 24 M. Tomita and T. G. M. van de Ven, J. Phys. Chem., 1985,89, 1291. 25 R. Williams, R. S. Crandall and P. J. Wojtowicz, Phys. Rev. Lett., 1976. 37, 348. 26 P. Pieranski, Contemp. Phys., 1983, 24, 25. 27 T. Okubo, presented at the 2nd Symposium on Chemical Reactions, Okazaki, Dec. 1985; Polymer 28 Handbook of Chemistry and Physics (CRC Press, Cleveland, 56th edn, 1968). 29 T. Okubo and K. Ito, J. Chem. SOC., Faraday Trans. 1, 1986, 82, 3185. 30 R. Williams and R. S. Crandall, Phys. Lett., 1974, 48A, 225. 3 1 A. Kose and S. Hachisu, J. Colloid Interface Sci., 1974, 46, 460. 32 T. Yoshiyama, I. Sogami and N. Ise, Phys. Rev. Lett., 1984, 53, 2153. 33 B. Pansu, P. Pieranskii and L. Strzelecki, J. Phys. (Paris), 1983, 144, 531. 34 S. Hachisu and Y. Kobayashi, J. Colloid Interface Sci., 1974, 46, 470. 35 E. J. W. Verwey and J. Th. G. Overbeek, Theory of the Stability of Lyophobic Colloids (Elsevier, 36 W. G. Hoover and F. M. Ree, J. Chem. Phys., 1968,49, 3609. 37 H. Eyring, T. Ree and H. Hirai, Proc. Natl Acad. Sci. USA, 1958, 44, 683. 38 H. Eyring and M. K. Jhon, Significant Liquid Structures (Wiley, New York, 1969). 39 R. W. Jones, The Optical Principles of the Diflraction of X-Rays (Cornell University Press, Ithaca, 1947). Preprints (Japan), 1986, 35, 1 1 12. Amsterdam, 1948). Paper 5/2210; Received 16th December, 1985

ISSN:0300-9599    

DOI:10.1039/F19868203163

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1986, 82, 3175-3183 Transmitted Light Spectrum Measurements A New and Convenient Technique for the Study of the Ordered Structure of a Monodispersed Polystyrene Latex in Solution and Film Tsuneo Okubo Department of Polymer Chemistry, Kyoto University, Kyoto 606, Japan Transmitted light spectra of monodispersed polystyrene latices in solution and film have been obtained using a commercially available spectrophoto- meter with an optical quartz cell (path lengths 1 and 2 mm for the solution). The spectra provide much information on the ordered microstructure of the polymer latex. The reliability is checked by comparing the lattice constant obtained from the transmitted light spectrum with that from two other independent techniques, the reflection spectrum method and direct ultra- microscopic observation.This unique and convenient method should aid future studies on the static and dynamic characteristics of ordered latex structures. Monodispersed polystyrene latex is an excellent system to investigate the correlation between ordered structure in solution and the interparticle interaction. Much has been published on its various aspects, such as the direct observation of the latex structure with a metallurgical microscope, its reflection spectrum, together with studies of light scattering, dielectric dispersion, viscoelasticity, viscosity, diffusion constant and ultrasonic absorption. Among these studies the optical technique has proved particularly effective, especially for solution studies of small-diameter latex particles (usually < 200 nm).It is surprising that no report has yet been published of a systematic analysis of the transmitted light spectra of a latex solution. As is well known, measurements of transmitted light spectra (or extinction) have often been used to study the structure of liquid crystals in solution. For example, the pitch of the spiral in an anisotropic solution of hydroxypropylcellulose was determined from extinction spectra.l* In this report the usefulness of analysis by transmitted light spectrum is demonstrated. Experimental Materials Diameters and other characteristics of the monodispersed polystyrene latices used in this work are listed in table 1. Data on the particle diameters and the monodispersity were obtained by electron micrography. Sample G530 1 was obtained commercially from Japan Synthetic Rubber Co.(Tokyo). The D1 -series of latices were purchased from Dow Chemical Co. The N- and C-series, the latter modified by carboxylation, were obtained from Sekisui Chemical Co. Details of the purification of these latices have been given in a separate paper.3 The pure water used was obtained through a Milli-Q reagent-grade water system. Spectrophotometric Measurements The reflection spectrum was recorded on a modified fl~orimeter.~ The U.V. spectro- photometers used for the extinction measurements were models SM-40 1 (Union Giken, 31753176 Ordered Structure in Polystyrene Latices Table 1. Properties of the latices used charge density latex diameter/nm /pC cm-2 D1C25 DlC27 DlB22 D1B72 D1B34 N- 1 50 N-200 C-30 1 N-300 G530 1 DlA92 D l C l l N-60 1 N- 1000 85k6 91 k 6 109f3 150f4 173f7 220+ 5 246 f 4 300k 10 320 & 10 369f 10 497 f 6 600k 12 804 5 1020 f 20 2.1 2.8 3.6 3.3 1 .o 4.2 1.9 34.3" 5.6 7.2 2.8 19.5 2.4 2.6 a Concentration of carboxylic acid groups.Hirakata), UV-190 (Shimadzu, Kyoto), UV-200s (Shimadzu) and EPS-3T (Hitachi, Tokyo). Quartz optical cells (path lengths I and 2 mm) were used for the reflection and transmitted light spectrum measurements. Ultramicroscope 0 bservation A metallurgical microscope, Carl Zeiss Axiomat IAC, was used. Details of the observation were reported by Hachisu et al.49 Principle of Transmitted Light Analysis of a Latex Solution The main features of an optical system are as follows. The incident light hits the surface of the quartz cell (1 mm path length in most cases) at right angles in fig.1. When the cell contains an 'ordered' latex solution the light is turned back by Bragg diffraction [(b) in fig. 13. The selective extinction (c), which will be analysed, comes from the reflected light, i.e. when the light is reflected selectively by the structural layers in the sample solution; the light, of course, cannot go straight through. However, the transmitted light has other important contributions. One is the extinction of light by latex particles and solvent molecules themselves (arrow d) and the other is the forward scattering (arrow e). The contribution of the former is not very important for polystyrene latices except for that at ca. 270 nm due to the phenyl groups.The contribution of the latter is usually very significant, particularly for randomly distributed latices in solution, and is interpreted quantitatively by the so-called Mie theory5 However, the forward scattering can be reduced to a negligible amount by using 2 mm pinholes fixed to the transmission cell. Note that multiple scattering is also important. Fortunately, these forward-scattering and multiple-scattering terms become very low in relation to the selective extinction term, when the particles form a lattice structure. Furthermore, the latex solution in a glass test tube is usually milky and opaque, but in a plane optical cell the solution becomes surprisingly tran~parent.~ This means that the diffraction planes of ordered latices become parallel to the wall of the cell glass.Thus, for the deionized and structure-formed latex solutions and latex film, transmitted light analysis is very useful.T. Okubo 3177 absorption by particles and solvents( d ) incident light (a), I forward scattering(e )* reflection light (b) ,11 selective absorption (c ) c 11 ! Fig, 1. Transmitted light spectrum. wavelengt h/nm Fig. 2. Transmitted light spectra of DlC27 latex solutions (in ~ 0 1 % ) : (a) 3.65, (b) 2.19, (c) 1.46, (a) 1.10, (e) 1.46 with 0.0001 mol dmd3 KCI and (f) I .46, reflection spectrum. Results and Discussion Transmitted Light Spectra of an Ordered Latex Solution Curves (a)-(d) in fig. 2 are the transmitted light spectra in the visible region of DlC27 latex (91 nm in diameter) at the concentrations of 3.65, 2.19, 1.46 and 1.10 ~ 0 1 % .A quartz optical cell of 1 mm path length was used. It is apparent that the extinction peaks are very sharp. Curve (e) in the figure shows the disordered latex solution at the same polymer concentration as that of curve (c) (1.46 vol % ) in the presence of mol dm-3 KCl. The background extinction of the ordered solution is certainly lower than that of the disordered solution. In our system of transmitted light spectral measurements, the scattering angle (28) is 180°, leading to the equation (1) AJnS = 2d sin 8 = 2d where d denotes the Bragg length and n, is the refractive index of latex sample solution (taken as that of water in this work). Al is the wavelength at the extinction peak. For3178 Ordered Structure in Polystyrene Latices wavelength/nm 700 800 1 1.6 c I \ n -0.6 Y c X * 0.4- wavelengthlnm Fig.3. Transmitted light spectra of DlB22 latex solutions (in ~ 0 1 % ) : (a) 1.94, (b) 0.97 and (c) 0.68. both face-centred (f.c.c.) and body-centred cubic (b.c.c.) lattices, the interparticle distance (2Dexp, or 2DeXp, b) is given by (2) 2DeXp, was estimated as 293 nm from the Al (644 nm) for curve ( c ) of fig. 2 and from eqn (2). The wavelength at the reflection maximum, Amax, for the solution is 547 nm, and 2D,,,, becomes 296 nm. The agreement is excellent. The decreasing tendency of A, with increasing latex concentration means a decrease in 2De,,. Fig. 3 shows the transmitted light spectra of DlB22 latex in deionized solution. In curve (a) two adjoining peaks appear, which are assigned to f.c.c.and b.c.c. lattice structures, as discussed in a separate paper.3 Furthermore, curves ( b ) and (c) display two extinction peaks, one sharp primary peak at longer wavelength and a secondary one at shorter wavelength. In these cases the wavelength at the primary extinction peak (A,) is twice that of the secondary peak (A2). The peak at A, is a Bragg diffraction of adjoining layers of the (1,1,1) plane (f.c.c.), and 1, is from that of alternate layers. In fig. 2 and 3 the transmitted light curves are not shown in the shorter wavelength region of 200-400 nm, where the extinction peak at ca. 270 nm is ascribed to phenyl groups of the latices. The reliability of the extinction spectrum measurements was checked by using two other kinds of spectrophotometer, UV- 190 (Shimadzu) and EPS-3T (Hitachi).Both instruments gave very similar patterns and the wavelengths of the extinction peaks coincided within 2 nm of each other. No influence of the slit width was found in the 0.03-0.5 mm range. Fig. 4 shows the time dependence of the transmitted light spectra of DlC27 solution at 2.19 vol % [this is the same as curve (b) of fig. 21. In the course of time, the background 2Dexp, f (= 2Dexp, b) = (2/3/2/2)d = O.6124(Ws).T. Okubo 3179 0.8 1 I I I 500 600 700 wavelength/nm 800 Fig. 4. Time dependence of the transmitted light spectra of DlC27 latex solution. 2.19 vol % . extinction, especially at shorter wavelengths, decreased and the extinction pattern became sharp. Much time is necessary for the (1, 1 , l ) plane of the f.c.c.lattices to become perfectly arranged parallel to the cell wall. This time dependence was often observed when the latex concentration was ca. 2 vol% for the DlC27 latex. Next, reflection and transmitted light spectral measurements were taken independently for a series of latex samples (diameter 85-300 nm). The relationship between the extinction maxima (A,) and reflection maxima (Amax) is seen in fig. 5. Clearly, the experimental points are located on the theoretical line. Since the reflection spectrum method is an established one, it is concluded that the transmitted light method is also sound. 2DeX, was also determined from the direct observation of the latex distribution using a metallurgical microscope. The results are compiled in table 2 for DlB72 (173 nm diameter) and DlB22 (109 nm diameter) latices.Experimental errors are estimated as 8% for the transmitted light spectrum, reflection and microscope methods, respectively. The agreement between them is satisfactory. 5 % , 5% and Measurements on a Monodispersed Polystyrene Latex Film The ordered solution structure of a monodispersed latex is not very stable and can be broken up with a small amount of ionic impurities from the cell glass etc. This section discusses films of monodispersed and deionized polystyrene latices on a cover glass. These films are white and opaque and also display brilliant iridescent colours. These colours are attributed to the diffraction of visible light by the latex particle crystallites. From a photograph taken by an optical ultramicrograph of a polystyrene latex film, the monodispersed spheres are seen to be closely packed in a hexagonal array.The photograph clearly shows the existence of many small ordered aggregates or crystallites and grain boundaries, both of which are quite similar features seen in an ordinary solid crystal. Photography was successful for the latex films when their diameters were >300 nm. For very large latex particles such as DlCll and N-1000, the regularity of 105 FAR 13180 400 Ordered Structure in Polystyrene Latices f' d - 0 - 8 00 c t 600 . x X' @ /O' B 3" 0% Fig. 5. Relation of Al and Amax in the latex solution. 0, D 1 B22 ; A, D 1 C27 ; 0, D 1 C25 ; x , N-200; a, N-150; ., C-301; A, DlB72; V, DlB34; (-) theory for f.c.c. and b.c.c. lattices.Table 2. Comparison of 20,,, values estimated by extinction and reflection spectra, and ultramicroscope methods for aqueous solutions of DlB72 and DlB22 latices [latex] latex (~01%) DlB72 2.02 3.03 4.04 DlB22 0.97 1.46 1.94 extinction reflection microscope 515 522 478 506 480 437 433 445 420 445 46 1 388 389 41 5 345 382 367 ~~ ~~~ - the distribution was not very good. The 20,,, values taken from the photograph appear in table 3 . The transmitted light spectra of the latex film are seen in fig. 6 and 7. The extinction peaks are clear, although they are broad compared with those of the latex solution. This is due to the fact that (i), there exist many grain boundaries, and (ii), the Bragg plane is not quite clear because the latex particles touch each other in the closely packed array. In some cases secondary peaks were observed in addition to the primary peaks. These peaks obviously came from the selective Bragg diffraction between the adjacent ( l ? l ? 1) planes of the hexagonal closely packed lattices.Note that the extinction peak was observed even for films of very large spheres (ca. 1000 nm in diameter) in the near-infrared region. The interparticle distance 20,,, was obtained from eqn ( 3 ) and the value of A1T. Okubo 3181 Table 3. Comparison of 2Oe1, values estimated by extinction and reflection spectra and ultramicroscope methods for polymer latex films 24,plnm diameter latex /nm extinction reflection microscope D1B72 C-30 1 N-300 65301 D 1 A92 D l C l l N-60 1 N- 1000 173 300 320 369 497 600 804 1020 163 294 325 344 464 575 778 973 _- 168 293 298 318 318 337 356 - 472 - 604 800 - 993 - 400 500 600 700 800 wavelength/nm Fig.6. Transmitted light spectra of latex films. (a) DlB72, (b) C-301 and (c) N-300. at the extinction peak. To estimate the refractive index of the film, the volume fractions of the polystyrene sphere and air were taken as 0.74 and 0.26, respectively. The refractive indices of latex particles decrease with increasing wavelength (AJS-l2 The 2D,,, values thus obtained are shown in table 3. A typical example of the reflection spectrum of the latex film is given in fig. 8. The interparticle distance 2DeXp was estimated from eqn (3) of ref. (3) and the reflection peak (Amax). The results are also given in table 3. The three independent measurements resulted in 2D,,, values very close to each other for the same film sample.Thus the transmitted light spectrum method is evidently one of the most convenient and reliable techniques for the analysis of the ordered structure of a latex film. 105-23182 Ordered Structure in Polystyrene Latices Fig. 7. Transmitted light spectra of latex films. (a) G539 1, (b) D 1 A92, ( c ) N-601, ( d ) D 1 C 1 1 and (e) B1000. n m + .r( f -$ W 4-3 .r( c" U c .d 300 400 500 600 700 800 wavelength/nm Fig. 8. Reflection spectra of latex films. (a) DlB72, (b) C-301, ( c ) N-300 and (d) G5301T. Okubo 3183 I thank Professor N. Ise of this Department for his encouragement of this work. Professor J. P. Kratohvil, Dr Y. Onogi, Dr M. Sishido and Dr H. Matsuoka are also acknowledged for their valuable suggestions on the transmitted light spectrum method. References 1 Y. Onogi, J. L. White and J. F. Fellers, J. Polym. Sci., Polym. Phys. Ed., 1980, 18, 663. 2 F. Fried and P. Sixou, J . Polym. Sci., Polym. Chem. Ed., 1984, 22, 239. 3 T. Okubo, J. Chem. SOC., Faraday Trans. I , 1986, 82, 3163. 4 A. Kose, M. Ozaki, K. Takano, Y. Kobayashi and S. Hachisu, J. Colloid Interface Sci., 1973, 44, 330. 5 N. Ise, T. Okubo, M. Sugimura, K. Ito and H. J. Nolte, J. Chem. Phys., 1983, 78, 536. 6 G. Mie, Ann. Phys., 1908, 25(4), 377. 7 R. Gans, Ann. Phys., 1912, 37(4), 881. 8 M. Nakagaki and W. Heller, J . Appl. Phys., 1956, 27, 975. 9 Handbook of Chemistry and Physics (CRC Press, Cleveland, 56th edn, 1968). 10 F. Robillard and A. J. Patitsas, Can. J. Phys., 1973, 51, 2395. 1 1 F. Robillard and A. J. Patitsas, Can. J . Phys., 1974, 52, 1571. 12 F. Robillard, A. J. Patitsas and B. H. Kaye, Powder Technol., 1974, 10, 307. Paper 5/2211; Received 16th December, 1985

ISSN:0300-9599    

DOI:10.1039/F19868203175

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J . Chem. SOC., Furaduy Trans. I , 1986, 82, 3185-3196 Ordered Solution Structure of a Monodispersed Polystyrene Latex as studied by the Transmitted Light Spectrum Method Tsuneo Okubo Department of Polymer Chemistry, Kyoto University, Kyoto 606, Japan The microstructure of monodispersed polystyrene latices in deionized solution has been systematically investigated by the transmitted light spectrum method. Latices used are between 85 and 246 nm in diameter and the charge density of the latex surface is between 1 .O and 36 pC cmP2. Clear peaks showing the selective Bragg diffraction are observed in the extinction spectrum. The ordered regions of the face-centred cubic (f.c.c.) and body- centred cubic (b.c.c.) lattices are assigned from the sharp peaks. The nearest-neighbour interparticle distance, 2Dexp, is estimated from the wave- length at the extinction peak.The 20,,, value decreases with increasing latex concentration and is close to (but smaller than) the calculated average value, 20,. The coexistence of ordered and disordered regions, or a two-state structure, is clarified. At low temperatures the f.c.c. structure is stable compared with the b.c.c. lattice, whereas the b.c.c. structure exists predomi- nantly at elevated solution temperatures. The addition of a small amount of a foreign salt (KCl) shifts the lattice structures in the ordered region from f.c.c. to b.c.c. The 20,, value increases on the addition of KCl and then begins to decrease upon further addition. The ordered structure is completely broken down by the addition of KCl above 2 x loP5 mol dm-3.In a previous paper1 the author reported that the transmitted light spectrum measurements from a conventional spectrophotometer gave very clear-cut information on the micro- structure of a monodispersed polystyrene latex in solution and of latex films. This paper reports the experimental results for the transmitted light spectrum measurements, especially on the face-centred cubic (f,c.c.) and body-centred cubic (b.c.c.) lattices present in the ordered regions in solution, and on the dependence of the latex concentration of the interparticle distance. The changes in the ordered structure with solution temperature and salt addition were also investigated. The comparison of these results obtained by the transmitted light spectrum method with those by the reflection spectrum method reported previously2 was made.Experimental Transmitted Light Spectrum Measurements Transmitted light spectra in the ultraviolet, visible and near-infrared ranges were recorded on spectrophotometers from Union Engineering Co. (SM-401, Hirakata, Japan) and Hitachi (EPS-3T, Tokyo, Japan). Quartz optical cells (path length 1 mm) were used. The reflection spectrum was recorded on a spectrophotometer as described previously.2 The conductance was measured by a Wayne Kerr autobalance precision bridge (B-33 1 mark 11) at a frequency of 1592 Hz. A Jones-Ballinger type cell was 31853186 Ordered Structure in Polystyrene Latices Table 1. Properties of the latices used charge density latex diameter/nm /pC cm-2 DlC25 85&6 2.1 DlC27 91 f 6 2.8 DlB22 109f3 3.6 DlB72 173+7 1 .o DlB34 246 f 4 1.9 -1 .o I I 300 400 500 600 700 800 wavelength/nm Fig.1. Transmitted light difference spectra of DlC27 solutions (in ~01%). (a) 0.55, (b) 0.73, (c) 0.97, (d) 1.28, (e) 1.83, df) 2.74, (g) 3.65 and (h) 4.56. The reference solution is a latex solution with 1.67 mmol dm-3 KC1. Materials Monodispersed polystyrene latices were purchased from Dow Chemical Co. The characteristics of the latices are listed in table 1. Details regarding purification of the latex samples were reported previously.2 Pure water for solution preparation was obtained through a Milli-Q reagent-grade water system. Results and Discussion Latex Concentration Dependence of the Transmitted Light Spectra All latex samples exhibited very strong iridescence. The difference spectra of D lC27 latex (91 nm in diameter) solution are shown in fig.1 . The reference solution contains 1.67 mmol dm-3 of KCl and is milky and completely disordered in structure. The concentration of latex varies from 0.55 to 4.56 vol % . Clear primary and secondary peaks due to selective Bragg diffraction were observed. The significantly negative extinction means that the background extinction of the ordered solution is much lower than thatT. Okubo 3187 wavelength/ n m Fig. 2. Transmitted light spectra of DlB22 solutions (in ~01%). (a) 6.0, (b) 4.21, ( c ) 3.16, ( d ) 2.10, (e) 1.94 and (f) 1.46. The reference cell contains pure water. of the disordered solution. The wavelength of the peak decreased with increasing latex concentration, as reported in a previous paper.l In curves (e) and (f) the shoulders adjacent to the sharp peaks correspond to the Bragg reflection of the (1, 1,l) planes of the b.c.c.lattice, which are parallel to the cell wall. The sharp peak is assigned to the f.c.c. lattice, or the (1, 1,l) plane. Details of the assignment were reported previously.' Fig. 2 shows the transmitted light spectra of DlB22 latex (109 nm in diameter) solution. Two observations were made. (1) The sharp peaks ascribed to the f.c.c. lattice were seen and peaks or shoulders of the b.c.c. structure appeared, particularly in diluted concentrations. Generally speaking, the f.c.c. structure was stable at high latex concentrations ( > 2 ~ 0 1 % for D 1 B22), whereas the b.c.c.lattice was observed predominantly at lower concentrations (< 1.5 vol% in the case of DlB22). Curves ( d ) and (e) each displayed an independent peak. (2) The peaks shifted to lower wavelength with increasing latex concentration. A similar trend was observed for other latex samples, D 1 C25, D 1 B72 and D1 B34, although their graphs are omitted. For D 1 B72 and DlB34 latices the secondary peaks appeared between 400 and 600 nm. The primary peaks would appear between 800 and 1200 nm, but the experiment was not carried out on these samples in this spectral region. By using eqn (2) of ref. (1) the nearest-neighbour interparticle distances in f.c.c. and b.c.c. lattices (20,,,, and 2D,,,, J were estimated accurately from the transmitted light peaks. The refractive indices of the sample solution, air and quartz glass were taken as 1.333, 1 .OO and 1.455, respectively, as described in ref.(2). The concentration dependence of 20,,, are shown in fig. 3-6 for DlC25, DlC27, DlB22 and DlB72 latex samples, respectively, together with their average values calculated on the assumption of a uniform f.c.c. or b.c.c. distribution, i.e. 2D0, (solid curve) and 2D0, b (dashed curve). In these figures, open circles and crosses denote 20,,,, and 20,,,, b, respectively. When the assignment was impossible, triangles are shown instead for the experimental values of 2D,,,, or 2DeXp, b e In many cases, the f.c.c. and b.c.c. lattice structures coexisted.3188 Ordered Structure in Polystyrene Latices 400 300 5 a 2 1 0 E4 200 0 I I I I 2 4 [DlC25](vol%) Fig.3. 100 500 400 E 6 s" --.- N 0 N 300 200 L- 2 4 [ D 1C27 I (~01%) Fig. 4. i Fig. 3. Interparticle distance of DlC25 solution estimated by the transmitted light spectrum Fig. 4. Interparticle distance of DlC27 solution estimated by the transmitted light spectrum method at 25 "c. 0, 20,,,, f ; X , 20,,,, b; (-) 2 0 0 , f ; (----) 200, b. method at 25 "C. 0, 20,,,, f ; (-) 2D0, f ; (----) 2 0 0 , b e However, the f.c.c8 structures appeared more often in our samples. For DlB22 especially, the b.c.c. lattices often appeared at dilute concentrations. This trend coincides with previous results.l? 2 * 4-8 From the figures it is evident that the values of 2Dex, are smaller than those of 20, by 5-20 % , i.e. 2Dexp -= 20,. This inequality has already been reported by our research group.2q 9 7 lo This work also supports the occurrence of a two-state structure, which is the coexistence of ordered and disordered regions.For D 1B34 latex, 2Dex,, was 614 nm at 3.83 vol% from the secondary peak, while the value of 2Dexp, from the reflection spectrum was 618 nm. The 2Dexp, values are certainly smaller than the 2D0, value of 660 nm. In order to account for the two-state structure by the effective hard-sphere model,,, the Debye length D, was estimated in the absence of added salt. D, is given by (4nBn)d. Here, B denotes the Bjerrum length (0.719 nm at 25 "C) given by e2/&kT, where e is the electronic charge, E the dielectric constant of the solvent and n is the concentration of the free-state gegenions. The fraction of free gegenions was assumed to be 0.10.This value is the mean obtained from transference and conductometric measurements of latex s o l ~ t i o n s ~ ~ - ~ ~ and is definitely smaller than those reported for soluble polyelectrolytes (ca. 0.4). Further, we assumed that the lowest ionic concentration of H+ and OH- ions in solution is ca. mol dm-3 because of water dissociation. Fortunately, this contribution was negligible in our case. The effective diameter of the latex, Deff, is given by the sum of 2 0 , and the true diameter of the latex. The results are compiled in table 2. Note that the effective diameter, which includes the Debye length, agrees with theT . Okubo I 3189 600 5 00 5 5f -.. 0 3- 400 a 300 I I 600 z 1 $ 500 a N 2 400 I I I I I 1 1 2 4 6 200 [DlB221(vol%) Fig.5. I I 0 2 4 [D1B72] (~01%) Fig. 6. Fig. 5. Interparticle distance of DlB22 solution estimated by the transmitted light spectrum Fig. 6. Interparticle distance of DlB72 solution estimated by the transmitted light spectrum method at 25 "c. 0, 20,,,, f ; x , 20,,,, b; A, 20,,,, or 2o,,,, b; (6-1 200, f ; (----> method at 25 "c. 0, 20,,,, f ; X , 2D,,,, b ; (--) zoo, f ; (----) zoo, b. 2o0, b. corresponding 2De,, value within experimental and calculation errors. Thus, this equality supports the validity of the effective hard-sphere model. The oscillation of particles around the equilibrium points in the ordered region (the G-value) has been reported to be within 0.1 .15 G denotes A2DeXp/2D,,,, where 2Dex, and A2Dex, are the average centre-to-centre interparticle distance in the ordered array and its standard deviation, respectively.The existence of the ordered region is therefore due to the slightly excess microstructural pressure from the disordered region, where particles move very vigorously.15* l6 In other words, the apparent ' attractive' interaction by which ordering seems to be caused is actually induced by the vigorous motion of the particles in the disordered regions, because the ordered regions (' islands ') are surrounded by disordered areas (' sea'). Moreover, the ordered and disordered regions coexist and fluctuate with time. Hachisu et al.6 described this in the following terms: 'At one moment an ordered portion.. . disintegrated into disorder and in the next moment an ordered aggregate appeared at another place., .'.Solution Temperature Dependence of Transmitted Light Spectra The influence of solution temperature on the transmitted light spectra is seen in fig. 7-9 for six different concentrations of DlB22 latex. At high latex concentrations such as 2.92 vol in fig. 7 ( a ) , the profiles are quite insensitive to the temperature between 15 and 55 "C. The calculated values of 2De,,, were 303 nm for all temperatures examined,3190 1 ‘650 700 750 0 Ordered Structure in Polystyrene Latices 800 Table 2. Estimation of Debye length (Dl) and effective diameter for various latex solutions concentration n Dl Deff (~01%) /moldm-3 /nm /nm 1 2 4 1 2 4 1 2 4 6 2 3 4 5 3.83 DlC25 (85 nm) 1.6 x 10-5 109 3.1 x 10-5 77 6.2 x 10-5 54 1.9 x 10-5 98 3.8 x 10-5 69 7.7 x 10-5 49 2.1 x 10-5 94 4.2 x 10-5 66 8.3 x lop5 47 1.2 x 10-4 39 7.4 x 10-6 160 1.1 x 10-5 130 1.5 x 10-5 110 1.9 x 10-5 100 3.59 x 10-6 226 DlC27 (91 nm) DlB22 (109 nm) DlB72 (173 nm) D 1 B34 (246 nm) 300 240 190 290 230 190 297 24 1 203 187 490 430 400 370 700 240 210 170 330 260 210 420 330 270 240 520 470 440 420 660 wavelength/nm Fig.7. Transmitted light spectra of DlB22 solution: temperature dependence. (a) 2.92 and (h) 2.10 vol”/, .T. Okubo 3191 E 0 E .- c .* c a, w avelength/nm Fig. 8. Transmitted light spectra of DlB22 solution: temperature dependence. (a) 1.94 and (bj 1.46 vol % . I I I I I I 400 450 500 550 600 650 700 750 800 wavelength/nm 0 1 Fig. 9. Transmitted light spectra of DlB22 solution: temperature dependence. (a) 0.97 and (b) 0.68 vol% .3192 Ordered Structure in Polystyrene Latices This means that the crystal structure is quite stable.At 2.10 vol "/o, however, the spectra are now sensitive to solution temperature, and f.c.c. lattices are broken up with increasing temperature, whereas the b.c.c. structures are stable even at 55 "C. At 1.94 vol % , the change of the profiles is similar to that at 2.10 vol % . However, at high temperatures even b.c.c. lattices are gradually broken up. At latex concentrations < 1.46 vol% , only the b.c.c. lattices were present and the structure was broken up by degrees with rising temperature. 20,,,, b also remained constant, irrespective of solution temperature. Note that (1) the f.c.c. lattices are transformed to b.c.c. with temperature, but (2) the lattice constant in the ordered regions is insensitive to solution temperature.Furthermore, (3) the lattice structure is stable even at diluted concentration (0.68 vol% in fig. 9) and at elevated temperature (65 "C). The transmitted light spectra of DlC25 were examined at 1.25 and 3.42 vol% between 15 and 65 "C although the graphs are omitted here. The f.c.c. structures were quite stable at 3.42 vol "/o even at 65 "C and the spectrum did not change much. A slight shift of the peak to shorter wavelength was observed, which corresponds to the temperature dependence of the refractive index of water. The 20,,,, values remained at 184 nm. At 1.25 vol% of DlC25 latex, the peak position did not change much but the extinction at the peak decreased with rising temperature.The values of 2D,,,, were between 253 nm (at 15 "C) and 254 nm (at 65 "C). The spectra of DlC27 solution (1.46 ~ 0 1 % ) were also insensitive to solution temperature. The values of 20,,,, ranged from 296 nm (at 15 "C) to 297 nm (at 65 "C). However, the shoulder corresponding to b.c.c. lattices gradually appeared with increasing temperature. Influence of Added Salt on the Transmitted Light Spectra Fig. 10 shows the spectral change when KCl was added to DlC27 latex solution. Clearly. the wavelength at the peak first increased with the addition of a small amount of KCl (< 7 x 1 OP6 mol dm-3) and then began to decrease [see fig. 10 ( d ) ] . The peak disappeared with the further addition of salt (> 1.3 x lop5 mol dm-3). It is surprising that the absorbance at the peak increased significantly with increasing salt concentration.We could see with the naked eye more colour to the sample solution in the quartz cell and an increased amount of flickering in the crystallite regions in the presence of a small amount of KCl. Similar but more complex profiles containing an f.c.c.-b.c.c. lattice transition were observed for DlB22 latex solution with KCl (see fig. 11-13). At high concentration [2.92 v o l x , fig. 11 (a)] and in the absence of KC1, the f.c.c. structure was stable. When KCl was added, the f.c.c. peak shifted slightly to longer wavelength and the b.c.c. lattice appeared [fig. 11 ( d ) ] . When the salt concentration was 2 x lop5 mol dm-3, almost all the f.c.c. structure disappeared and the b.c.c.structure appeared instead. When more KCl was added, the ordered structure was broken up [fig. 11 (f)]. The DlB22 solution of 1.94 vol % (fig. 12) also showed this systematic change in solution structure from f.c.c. to b.c.c. with addition of KCl, although the curves seem complex at a first glance. The spectra in fig. 13 at 1.46 vol% showed that the lattice constant of the b.c.c. structure increased with increasing KCl concentration (< 6.67 x lop6 mol dm-3), and decomposition of the ordering occurred at KCl concentrations > 1 x lop5 mol dmp3. Also, the flickering crystallites became more conspicuous at this level of KCl. The size of the crystallites was 1 .O mm under optimum conditions, and they quickly disappeared when the solution was shaken, although they reappeared later.The conductance of the latex solution containing KCl was measured using a cell (cell constant = 16.1 cm-I). The specific conductances of DlB22 solution (1.46 vol%) were 7.4 x lo+, 8.5 x 1.03 x and 1.19 x lop5 W1 cm-l at [KCl] = 0, 3.33 x 6.67 x and 1.33 x mol dm-3, respectively. The solution conductance increased monotonically with KC1 concentration, and no discontinuity in conductance was found in the region of KCl concentration where degradation of the ordered structure had occurred.T. Okubo 3193 580 600 62 0 640 wavelengt h/nm Fig, 10. Transmitted light spectra of DlC27 solution : salt concentration dependence. [DlC27] = 1.64 ~ 0 1 % . [KCl]/mol dm-3 = (a) 0, (b) 3.33 x (e) 1.33 x ( f ) 1.67 x and (g) 3.33 x lop5. (c) 6.67 x (d) 1 x I 10 620 64 0 660 680 700 wavele ng t h/nm Fig.11. Transmitted light spectra of D lB22 solution: salt concentration dependence. [DlB22] = 2.92 ~01%. [KCl]/mol dm-3 = (a) 0, (b) 5 x (c) 1 x (d) 1.5 x (e) 2 x and cf) 2.5 x3194 Ordered Structure in Polystyrene Latices wavelengt h/nm Fig. 12. Transmitted light spectra of D 1B22 solution: salt concentration dependence. [DlB22] = 1.94 ~01%. [KCl]/mol dm-3 = (a) 0, (b) 3.33 x (c) 6.67 x (d) I x (e) and cf> 1.67 x 1.33 x G Y 0.6 2 0.4 6" 5 3 0.2 -0.4 .3 Y 0 76 0 780 800 82 0 840 660 wavelength/nm Fig. 13. Transmitted light spectra of D1 B22 solution: salt concentration dependence. [DlB22] = 1.46 ~01%. [KCl]/mol dm+ = (a) 0, (b) 3.33 x (d) 1 x lop5, (e) and (f) 1.67 x (c) 6.67 x 1.33 xT. Okubo 3195 400 380 360 --- 6 N a 340 3 20 300 \ .----* 0 I X ~ O - ~ 2 ~ 1 0 - ~ [ KCl]/mol dmV3 Fig.14. Salt effect on the interparticle distance of DlB22 solution at 25 "C. [DlB22] (vol "/o ) = (a) 1.46, (b) 1.94 and (c) 2.92. 0,20,,,, f ; @, 20,,,, b; (-) 20,, f ; (----) 20,, b. Size of the circles is proportional to the extinction of the peak. The dependences of the wavelength and the extinction at the peak on the salt concentration are shown in fig. 14 for three different concentrations of DlB22 latex. The size of the open and solid circles is proportional to the extinction of the peak in the figure. Although both the interparticle distance and the extinction of the peak first increased with increasing KCl concentration, they then began to decrease. The inequality 2D,,, < 20, held even in the presence of salt, as is clear from the figure, and this indicates that the coexistence of the ordered and disordered regions (the two-state structure) is valid even in the presence of a small amount of foreign salt. The increase in 2D,,, with KCl is due to the fact that the stability of the ordering structure is lowered, so that the two-state structure becomes imperfect because of the thinning of the electrical double layers surrounding the latex particles and the weakening of the apparent attraction between latex particles in the ordered regions.I thank Professor Norio Ise for his encouragement. This work was supported by grants-in-aid from the Kurata Foundation and from the Ministry of Education, Science and Culture, Japan for the special project research ' Design of Multiphase Biomedical Materials ' . References 1 T. Okubo, J. Chem. Soc., Faraday Trans. I , 1986, 82, 3175. 2 T. Okubo, J. Chem. SOC., Faraday Trans. 1, 1986, 82, 3163. 3 G. Jones and M. Ballinger, J. Am. Chem. SOC., 1931,53, 41 13196 Ordered Structure in Polystyrene Latices 4 V. W. Luck, M. Klier and H. Wesslaw, Naturwissenschaften, 1963, 14, 485. 5 P. A. Hiltner and I. M. Krieger, J. Phys. Chem., 1969, 73, 2386. 6 A. Kose, M. Ozaki, K. Takano, Y. Kobayashi and S. Hachisu, J. Colloid Interface Sci., 1973, 44, 330. 7 R. Williams and R. S. Crandall, Phys. Lett., 1974, 48A, 225. 8 T. Yoshiyama, I. Sogami and N. Ise, Phys. Rev. Lett., 1984, 53, 2153. 9 N. Ise, T. Okubo, M. Sugimura, K. Ito and H. J. Nolte, J. Chem. Phys., 1983, 78, 536. 10 N. Ise, K. Ito, T. Okubo, S. Dosho and I. Sogami, J. Am. Chem. Soc., 1985, 107, 8074. 11 S. Hachisu and Y. Kobayashi, J. Colloid Interface Sci., 1974, 46, 470. 12 J. W. Vanderhoff, H. J. van de Hul, R. J. M. Tausk and J. Th. G. Overbeek, Clean Surfaces: Their Preparation and Characterization for Interfacial Studies, ed. G. Goldfinger (Marcel Dekker, New York, 1970). 13 S. Alexander, P. M. Chaikin, P. Grant, G. J. Morales, P. Pimcus and D. Hone, J. Chem. Phys., 1984, 80, 5776. 14 K. Ito, N. Ise and T. Okubo, J. Chem. Phys., 1985, 82, 5732. 15 N. Ise, T. Okubo, K. Ito, S. Dosho and I. Sogami, J. Colloid Interface Sci., 1985, 103, 292. 16 N. Ise, T. Okubo, K. Ito, S. Dosho and I. Sogami, Langmuir, 1985, 1, 176. Paper 512212; Received 16th December, 1985

ISSN:0300-9599    

DOI:10.1039/F19868203185

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SQC., Faraday Trans. 1, 1986, 82, 3197-3204 Adsorption and Decomposition of Ethylene and Acetylene on Flat Ru(001) and Stepped Ru( l,l, 10) Surfaces Chikashi Egawa,* Shuichi Naito and Kenzi Tamaru Department of Chemistry, Faculty of Science, University of Tokyo, Bunkyo-ku, Tokyo 113, Japan The adsorption and decomposition of ethylene and acetylene on flat Ru(001) and stepped Ru(l,l, 10) surfaces have been investigated by u.p.s. and t.d.s. On the flat surface ethylene and acetylene were adsorbed molecu- larly at 120 K and converted into ethylidyne species, ZC-CH,, at 280 K. Heating to 450 K in vucuo caused the ethylidyne species to decompose, with the evolution of hydrogen and probable formation CCH(ads) species; the latter were changed to CH(ads) at 550 K and finally to C(ads) above 650 K.However, the presence of CO retarded the decomposition of the ethylidyne species. In contrast, cleavage of the C-H bond occurred at 120 K on the stepped surface, presumably with the formation of CCH(ads) or CCH,(ads) species. The formation of ethylidyne species around 280 K was not observed, which also indicates the higher reactivity of the stepped structure on the Ru surface towards the decomposition of alkenes. This is further supported by a shift of the desorption peak to lower temperature in H, temperature- programmed desorption spectra. In Fisher-Tropsch synthesis on transition metals it has been concluded that hydrocarbon products are formed through the hydrogenation and propagation of carbonaceous species produced by CO diss~ciation.l-~ In our laboratory4r5 the mechanism of CO hydrogenation has been examined over a silica-supported Ru catalyst.In situ infrared spectra5 revealed that the Ru surface was mostly covered by adsorbed CO, and that adsorbed long-chain aliphatic hydrocarbon species accumulated during the reaction. While exchange between the hydrogen of the adsorbed hydrocarbon species and D, in the gas phase was slow under the reaction conditions when CO was present, an appreciable fraction was easily decomposed in vacuo (i.e. after CO was desorbed) to form active carbon species which were hydrogenated to methane at room temperature. In addition, by 13C isotope labelling experiments4 (i.e. switching from l2CO to 13CO), it was confirmed that initial hydrocarbon products contained only 12C and that some parts of the adsorbed hydrocarbon species that had accumulated during the reaction were incorporated into the products.The addition of ethylene during the CO + H, reaction also provided similar active carbon species as those produced by CO dissociation. In this study, therefore, the adsorption and decomposition of ethylene and acetylene was investigated on flat Ru(001) and stepped Ru( 1, 1 ,lo) surfaces with regard to the adsorbed hydrocarbon species as mentioned above. There have been numerous studies on the interaction of hydrocarbons with transition metals in which various surface techniques have revealed the presence of surface intermediate^.^-^^ On the 3d transition metals, ethylene was molecularly adsorbed at lower temperatures and was dehydrogenated to acetylene at ca.300 K on Ni( 1 1 1)6* and on Co foil.l5 Adsorption on Ni( 100) and Ni( 1 10)12-14 caused highly distorted species, and subsequent heating induced CCH(ads) or CH(ads) species. On the other hand, on the 4d and 5d series, the presence of di-a molecular species was confirmed on Pd(l1 1),24-27 Pt( 1 1 1),33-35 Ru(O01)20T 21 and Rh( 1 1 1) ;47 this was usually converted into ethylidyne 31973198 Decomposition of Ethylene and Acetylene on Ru Surfaces species at ca. 300 K, although the formation of CCH(ads) species has recently been reported on Pd(100).26 On these metals hexagonal surface structure seems to play an important role in the formation of ethylidyne species. On the other hand, it has frequently been reported that a stepped surface structure has a higher reactivity for the scission of C-C and C-H bonds on Pt,38 Ir3, and N i 7 On a Ni stepped surface acetylene was dehydrogenated to C,(ads) species at 150 K.We have also demonstrated that a stepped structure on the Ru surface is effective for the dissociation of C-0,399 40 N-0 and N-H41 bonds. Accordingly, in this study we focused our interest on the effect of the stepped structure on the formation of surface intermediate, especially ethylidyne species, from ethylene and acetylene adsorption. Experimental Experimental details were as described previo~sly.~~~ 42 U.P.S. measurements were carried out in a u.h.v. chamber (V. G. Escalab 5). The photoelectrons were collected in a direction normal to the samples, and accordingly the direction of analyser was inclined by 17.6" from the normal to the (001) terrace plane to (551) direction for the stepped surface.Ethylene and acetylene were dosed by a stainless-steel capillary onto the surface, where the pressure near the sample was ca. two magnitudes higher than in the background. Thermal desorption (t.d.) experiments were performed with another stainless-steel u.h.v. chamber. The sample was resistively heated by d.c. and was cooled to 150 K through the sample holder. The temperature was monitored by a chromel-alumel thermocouple attached to the sample edge. Ru samples were cleaned by 0, treatment, Ar sputtering and subsequent heating in vacuo. The cleanliness of the surface was checked by Auger electron spectroscopy and low-energy electron diffraction.The effect of the sample edge was eliminated by treating it with H,S followed by Ar sputtering. The purity of the gases used was checked by mass spectrometer. Results and Discussion Ethylene and Acetylene on the Flat Ru(OO1) surface Fig. 1 shows He(I1) ultraviolet photoelectron spectra for the adsorption of ethylene on the Ru(OO1) surface at 120 K and subsequent decomposition. Spectrum (a), obtained from the clean surface, has two characteristic peaks at 5.5 and 7.5 eV besides the d-band region; these are assigned to the critical points I?; and I': of the bulk Ru emission.43 Because of some overlapping of these peaks, the analysis of the adsorbed species is rather complicated over this energy region, but the adsorption of ethylene at 120 K (ca. 3 LT) gave new peaks at ca.4, 6.5 and 8 eV with shoulders at 9 and 11.5 eV. These peaks correspond to the lblu, lb,,, 3ag, lb,, and 2ss* (2A,) orbitals of the ethylene molecule, respectively. However, the peak separation of lblg and lb,, (2.5 eV) is smaller than that (3.1 eV) for gaseous ethylene, as observed on Ir(100),31 which indicates that the C-C bond distance is stretched upon adsorption. Also, from the decrease in the work function (- 1 .O eV) caused by adsorption, the ethylene molecule is considered to be bonded parallel to the Ru surface by the donation of a x-electron. This is consistent with a recent HREELS experiment21 in which, from two strong losses at 11 10 cm-l (C-C stretch) and 2910 cm-l (CH, stretch), the ethylene molecule is considered to be di-a-bonded to Ru atoms, with rehybridization of the carbons from sp2 to sp3, reducing its C-C bond order.At 280 K the adsorbed species caused emissions at 6-8 and 11.6 eV which are identical to those for acetylene adsorption on Ru(001) as shown later. These features are in reasonable agreement with ethylidyne species reported on Pd( 1 1 1),27 Pt( 1 1 1)35 and a Co complex.44 The formation of ethylidyne species oriented perpendicular to the Ru t 1 L = 1 Torr s (1 Torr = 101 325/760 Pa).C. Egawa, S. Naito and K. Tamaru 3199 15 10 5 O=E, binding energy/eV 15 10 5 0 =E, binding energy/eV Fig. 1. (A) He(I1) u.p. and (B) difference spectra for the adsorption of ethylene on Ru(001) surfaces: (a) clean, (b) 3 L at 120 K, A 0 = - 1.0 eV, (c) (b) heated to 160 K, ( d ) heated to 210 K, A@ = -0.9 eV, (e) 2.5 L at 280 K, A@ = -0.9 eV, (f) (e) heated to 450 K, A@ = -0.8 eV, (g) heated to 550 K, A@ = -0.5 eV and (h) heated to 650 K.surface was also found by the HREELS method,21 where the losses at 1120, 1340 and 2910 cm-l were assigned to C-C stretching, CH, bending and CH, stretching modes, respectively. The spectrum at 210 K [fig. 1 (d)] had peaks at 7-10 and 1 1.4 eV, which are different from those for the ethylidyne species at 280 K, and is therefore tentatively assigned to a vinyl-like species, H,C=CH(ads), which Demuth3' has observed as an intermediate of the ethylidyne species. The peaks at ca. 9 and 12 eV remained on heating to 450 K [fig. 1 (f)] and further heating to 650 K gave the emission at 9.2 eV, which is induced by C ( a d ~ ) .~ ~ T.p.d. spectra for H, from ethylene decomposition (see fig. 3 later) show that almost one hydrogen atom per ethylene molecule remains on the surface above 450 K. Accordingly, the spectral changes from (f) to (g) in fig. 1 indicate the decomposition of surface species such as CH or C,H. In the HREELS experiment21 it was concluded that a sp2 CH(ads) species is formed at 460 K, and further heating to 570K caused the transformation to another type of CH(ads) species, such as a bridge-bonded one. On the other hand, the formation of CCH(ads) from the decom- position of CCH,(ads) has been reported recently on Pd(100) and (1 1 1),26 which gives similar EELS spectra to the inclined CH(ads). Therefore, the two peaks at ca. 9 and 12 eV observed at 450 K may be interpreted as the 30 and 2ss* orbitals of the CCH(ads) species, which decomposes to give CH(ads) and C(ads) at higher temperatures.He(I1) u.p. spectra for the adsorption of acetylene on the Ru(OO1) surface are shown in fig. 2. The adsorption of acetylene (1.5 L) at 120 K demonstrated two emissions at 8.8 and 11.8 eV. They are assigned to 30, and 20, orbitals of the acetylene molecule, respectively. The peak separation (2 eV) is the same as that of gaseous acetylene. Taking3200 Decomposition of Ethylene and Acetylene on Ru Surfaces 15 10 5 O=E, binding energy/eV (c)-clean ----+/ (b)-clean 4 (a)-clean 4 15 10 5 O=E, binding energy/eV Fig. 2. (A) He(I1) u.p. and (B) difference spectra for the adsorption of acetylene on Ru(OO1) surfaces: (a) 1.5 L at 120 K, A 0 = - 1.4 eV, (b) (a) heated to 160 K, (c) heated to 210 K, A 0 = - 1.3 eV and ( d ) 2.5 L at 280 K.I 250 300 400 500 600 700 TIK Fig. 3. H, t.p.d. spectra from ethylene adsorbed on Ru(001) surfaces (3 L at 230 K): (a) clean [(----) indicates acetylene], (b) preadsorbed CO surface, O,, = 0.25 and (c) under a CO pressure of 2 x Pa (adsorbed ethylene was preheated to 300 K before desorption).C. Egawa, S. Naito and K . Tamaru r 15 10 5 0 binding energy/eV 320 1 15 10 5 0 binding energy/eV Fig. 4. He(I1) u.p. spectra for the adsorption of ethylene on Ru( 1 , 1,lO) surfaces: (a) clean, (b) 0.4 L at 120 K, A@ = -0.2 eV, (c) 1.2 L at 120 K, A@ = -0.7 eV and ( d ) (c) heated to 220 K. the work function (3.9 eV) into consideration, the relaxation energy is found to be 4 eV, which is a little larger than on other metal surfaces.18 This indicates that the acetylene molecule is more closely bonded to the Ru surface.Warming the surface to 160 K, however, caused an increase in the peak separation (2.5 eV), indicating an increase in the C-C bond distance. Further heating to 210 K gave the two comparable peaks at 9 and 11 eV, probably due to 30, and 20, orbitals of the CCH,(ads) species, which is postulated as an intermediate on other metal surfaces.lg The spectral change above 280 K is similar to the case of ethylene adsorption, which gives two peaks at 6-8 and 12 eV. This is consistent with H, t.p.d. spectra (fig. 3), where although the peak area below 500 K from ethylene decomposition is much larger than that from acetylene, the peak shape above 500 K is identical in both cases.As mentioned before, the lower peak is caused by the decomposition of ethylidyne species, and three peaks in the higher- temperature region are considered to reflect the decomposition of hydrocarbon species such as CCH(ads), CH(ads) etc., which have different stabilities on the surface. A further investigation was made of the effect of adsorbed CO on the decomposition of ethylidyne species. As is demonstrated in fig. 3 (b), adsorbed CO reduced the amount of H, desorption and caused a shift in the peak to lower temperatures. This indicates that CO adsorbed on top sites on Ru hinders the formation of ethylidyne species by site blocking. As a result, the amount of ethylidyne species decreased below saturation coverage and only after CO desorption above 450 K could further decomposition of hydrocarbon species proceed easily using empty Ru sites.Ethylene is reported to be decomposed to C(ads) and H(ads) on hydrogen-preadsorbed Ru(O0 1),21 instead of forming ethylidine species; this indicates that preadsorbed hydrogen, probably adsorbed on three-fold hollow prevents the formation of ethylidyne species on the same sites . On the other hand, the decomposition of adsorbed ethylene in the presence of CO (Pco = 1.3 x Pa) shows an H, desorption peak at 390 K as shown in fig. 3 (c). This means that the ethylidyne species is stabilized by coadsorbed CO, which will be more noticeable under the normal CO pressure range. The accumulation of long hydrocarbon chains was observed over a supported Ru catalyst under reaction condition^.^3202 Decomposition of Ethylene and Acetylene on Ru Surfaces 15 10 5 O=Ef binding energy/eV 15 10 5 0 =Ef binding energy/eV Fig. 5.(a) He(I1) u.p. spectra for the adsorption of acetylene on Ru(l,l,lO) surfaces: (1) 0.5 L at 120 K, A@ = -0.9 eV, (2) 1.5 L at 120 K, A 0 = -0.9 eV, (3) (2) heated to 210 K, (4) 2.5 L at 280 K, ( 5 ) (4) heated to 450 K, (6) heated to 520 K and (7) heated to 650 K. (b) He(I1) u.p. difference spectra for the adsorption of acetylene on Ru( 1 , 1,lO) surfaces. Ethylene and Acetylene on the Stepped Ru(l,l,lO) Surface Fig. 4 shows He(I1) u.p. spectra for ethylene adsorption on the stepped Ru(l,1,10) surface. A small exposure (0.4 L) of ethylene at 120 K induced emissions at 6, 7 and 10 eV and a slight decrease in the work function (A@ = -0.2 eV).Further exposure (1.2 L) added new emissions at 6, 7.6, 9 and 11.4 eV and reduced the work function to 4.2 eV; these are typical features for molecularly adsorbed ethylene, as shown in fig. 1. Similarly, a low exposure (0.5 L) of acetylene on the stepped surface at 120 K also caused emissions at 6.7 and 10 eV and additional peaks at 8.8 and 10.8 eV derived from molecularly adsorbed acetylene (see fig. 5). The intensity of the latter peaks was increased by further exposure (1.5 L). Accordingly, it is considered that ethylene or acetylene is initially decomposed on step sites and then adsorbed molecularly on terrace sites; scission of the C-H bond is likely to occur on the stepped surface at temperatures as low as 120 K, forming CCH(ads) or CCH,(ads) species, where the emission at 7 and 10 eV may be ascribed to 30, and 20, orbitals14 and the 6 eV peak to adsorbed hydrogen atoms.48 On a stepped Ni surface,' which has almost the same structure as in the present study, acetylene is dehydrogenated to C,(ads) species and ethylene is partially dehydrogenated to CHCH,(ads) species at 150 K.The difference in reactivity between the two molecules is thought to be caused by a change in the hydrogen concentration on the step sites. On the other hand, the formation of CH,(ads) or CH(ads) species by C-C bond cleavage is believed to occur on an open Fe(100) surface at 123 K;ls this may be excluded from the present study because the peak positions and their separation are not related to their derived ~a1ues.l~ At 210 K adsorbed acetylene and ethylene gave the same spectra, with two peaks at 7.5 and 11 eV.Further heating to 285 K or adsorption at 285 K shifted them to lowerC . Egawa, S. Naito and K. Tamaru 3203 250 300 400 600 Fig. 6. H, t.p.d. spectra from ethylene and acetylene adsorbed on Ru(l,l,lO) surfaces (3 L at 230 K): (-) ethylene; (----) acetylene. TIK binding energies of 7 and 10 eV, resembling the spectrum obtained at low exposure at 120 K. It is clear that the ethylidyne species is not so stable at 285 K as on the flat surface. The result also corresponds well to the lack of sharp desorption peaks in the H, t.p.d. spectra (fig. 6), where only a broad peak at 300 K is obtained. The assignment of surface species at 210 K is not clear at present.As demonstrated in fig. 6, H, desorption started at 240 K, and three or more hydrogen atoms per adsorbed ethylene molecule are already desorbed at 450 K. Therefore, the peak at 8.8 eV in the u.p. spectra at 450 K is assigned to CH(ads) species. Further heating to 650 K produced the broad emission at 6-9 eV which is derived from C(ad~).~Oy 43 The higher reactivity of the stepped surface is also reflected in the peak shift towards lower temperatures in H, t.p.d. spectra, which is consistent with results on other A comparison of the behaviour of hydrocarbons on flat and stepped surfaces may shed some light on the reaction mechanism of CO hydrogenation over silica-supported Ru catalysts. The hydrogenation or 'hydrogen exchange reaction of the ethylidyne species is considerably slower on a saturated surface,46 which is in reasonable agreement with the slow exchange between the hydrogen of the long-chain hydrocarbon species and D, under reaction condition^.^ Also, isotope labelling shows that a limited part of the long-chain hydrocarbon species is incorporated into the products, although active carbon species are produced by the decomposition of those hydrocarbon chains on a bare surface.Taking into consideration the fact that alkyilidyne species require hexagonal surface structure, such long hydrocarbon chains may be formed on the flat surface of Ru and stabilized by the surrounding CO, and they are converted into the active carbon species on stepped sites where the dissociation of CO molecules occurred and active carbon species were formed.The active sites for chain growth, however, are as yet less known. Summary (1) On a flat Ru(OO1) surface, ethylene and acetylene were adsorbed in their distorted molecular forms at 120-160 K, which were then converted into ethylidyne species at 280 K via CHCH,(ads) and CCH,(ads) species, respectively. (2) Although the ethylidyne species was decomposed to CCH(ads) with a sharp H, desorption peak at 350 K in vacuo, it was stabilized by coadsorbed CO. (3) On a stepped Ru(l,l,lO) stepped surface, ethylene and acetylene were initially decomposed on step sites and then adsorbed3204 Decomposition of Ethylene and Acetylene on Ru Surfaces molecularly on terrace sites at 120 K. (4) The absence of ethylidyne species at 280 K and the shift of the desorbed peak to lower temperature in H, t.p.d. spectra over the Ru( l , l , 10) surface show the higher reactivity of the stepped surface.References 1 M. Araki and V. Ponec, J. Cutul., 1976,44,439. 2 P. Biloen, J. N. Helle, F. G. A. van der Berg and W. M. H. Sachtler, J. Cutal., 1983, 81, 450. 3 P. Winslow and A. T. Bell, J. Catul., 1984, 86, 158. 4 Y. Kobori, H. Yamasaki, S. Naito, T. Onishi and K. Tamaru, J. Chem. Soc., Faraday Trans. I, 1982, 5 H. Yamasaki, Y. Kobori, S. Naito, T. Onishi and K. Tamaru, J. Chem. Soc., Furuday Trans. I, 1981, 6 J. E. Demuth and D. E. Eastman, Phys. Rev. Lett., 1974, 32, 1132. 7 S. Lehwald and H. Ibach, Surf. Sci., 1979, 89, 425. 8 J. E. Demuth and D. E. Eastman, Phys. Rev., 1976, 13, 1523. 9 H. Ibach and S.Lehwald, J. Vac. Sci. Technol., 1981, 18, 625. 78, 1473. 77, 2913. 10 J. E. Demuth and H. Ibach, Surf. Sci., 1978, 78, L238. 11 J. E. Demuth, Surf. Sci., 1977, 69, 365. 12 K. Horn, A. M. Bradshaw and K. Jacobi, J. Vuc. Sci. Technol., 1978, 15, 575. 13 J. A. Stroscio, B. R. Bare and W. Ho, Surf. Sci., 1984, 148, 499. 14 J. E. Demuth, Surf. Sci., 1980, 93, 127. 15 P. Tiscione and G. Rovida, Surf. Sci., 1985, 154, L255. 16 G. Broden, G. Gafner and H. P. Bonzel, Appl. Phys., 1977, 13, 333. 17 W. Erley, A. M. Bar0 and H. Ibach, Surf. Sci., 1982, 120, 273. 18 C. Brucker and T. Rhodin, J. Catal., 1977, 47, 214. 19 U. Seip, M-C. Tsai, J. Kuppers and G. Ertl, Surf. Sci., 1984, 147, 65. 20 T. E. Fisher and S. R. Kelemen, Surf. Sci., 1978, 74, 47. 21 M. A. Barteau, J.Q. Broughton and D. Menzel, Appl. Surf. Sci., 1984, 19, 92. 22 M. A. Barteau, P. Feulner, R. Stengel, J. Q. Broughton and D. Menzel, J. Catal., 1985, 94, 51. 23 B. E. Koel, B. E. Bent and G. A. Somorjai, Surf. Sci., 1984, 146, 21 1. 24 J. A. Gates and L. L. Kesmodel, J. Chem. Phys., 1982, 76, 4281. 25 L. L. Kesmodel and J. A. Gates, Surf. Sci., 1981, 111, L747. 26 L. L. Kesmodel, G. D. Waddill and J. A. Gates, Surf. Sci., 1984, 138, 464. 27 D. R. Lioyd and F. P. Netzer, Surf. Sci., 1983, 129, L249. 28 L. L. Kesmodel, J. Chem. Phys., 1982, 79, 4646. 29 E. M. Stuve, R. J. Madix and C. R. Brundle, Surf. Sci., 1985, 152/153, 532. 30 M. A. Chesters, G. S. McDougall, M. E. Pemble and N. Sheppard, Appl. Surf. Sci., 1985,22/23, 369. 31 G. Broden, T. Rhodin and W. Capehart, Surf. Sci., 1976, 61, 143. 32 B. E. Nieuwenhuys, D. I. Hagen, G. Rovida and G. A. Somorjai, Surf. Sci., 1976,59, 155. 33 H. Ibach and S. Lehwald, J. Vac. Sci. Technol., 1978, 15, 407. 34 L. L. Kesmodel, L. H. Dubois and G. A. Somorjai, J. Chem. Phys., 1979,70, 2180. 35 M. R. Albert, L. G. Sneddon, W. Eberhardt, F. Greuter, T. Gustafsson and E. W. Plummer, Surf. 36 T. E. Fisher and S. R. Kelemen, Surf. Sci., 1977, 69, 485. 37 J. E. Demuth, Surf. Sci., 1979, 80, 367. 38 R. K. Herz, W. D. Gillespie, E. E. Petersen and G. A. Somorjai, J. Cutul., 1981, 67, 371. 39 E. Shincho, C. Egawa, S. Naito and K. Tamaru, Surf. Sci., 1985, 149, 1. 40 E. Shincho, C. Egawa, S. Naito and K. Tamaru, Surf. Sci., 1985, 155, 153. 41 C. Egawa, S. Naito and K. Tamaru, Surf. Sci., 1983, 125, 605. 42 H. Shindo, C. Egawa, T. Onishi and K. Tamaru, J. Chem. Soc., Faraday Trans. I, 1980, 76, 280. Sci., 1982, 120, 19. 43 F. J. Himpsel, K. Christmann, P. Heimann, D. E. Eastman and P. J. Feibelman, Surf. Sci., 1982, 115, L159. 44 N. C. V. Costa, D. R. Lloyd, P. Brint, T. R. Spalding and W. K. Pelin, Surf. Sci.. 1981, 107. L379. 45 M. A. Barteau, J. Q. Broughton and D. Menzel, Surf. Sci., 1983, 133, 443. 46 B. E. Koel, B. E. Bent and G. A. Somorjai, Surf. Sci., 1984, 146, 21 1. 47 R. J. Koestner, M. A. van Hove and G. A. Somorjai, Surf. Sci., 1984, 121, 321; L. H. Dubois, G. G. Castner and G. A. Somorjai, J. Chem. Pkvs., 1980, 72, 5234. 48 P. Hofmann and D. Menzel, Surf. Sci., 1985, 152/153, 382. Paper 5/2214; Received 16th December, 1985

ISSN:0300-9599    

DOI:10.1039/F19868203197

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I, 1986, 82, 3205-3214 Solubilisation of Two-component Oil M Surfactant Solutions Brendan J. Carroll Unilever Research, Port Sunlight Laboratory, Quarry Wirral, Merseyside L63 3JW xtures by Micellar Road East, Bebington, Single-component oils are solubilised in different amounts and at different rates by a given micellar surfactant solution, the differences originating in the size and chemical nature of the solubilising oil molecule. This paper examines theoretically the case where solubilisation without fractionation occurs in a binary oillsingle-component (micellar) surfactant solution. The amounts solubilised at intermediate bulk phase compositions are shown to depend on the ratio of the values for the two composition extremes and also on packing in the micelle.The rate of solubilisation is calculated using the expressions for the amounts solubilised. The theoretical predictions have been compared with some recent experimental data. The theory is shown to give a reasonable fit to both equilibrium and kinetic solubilisation data. The topic of solubilisation of mixtures of oils by micellar systems is of interest because of the possible existence of a differential solubilisation process in certain circumstances. This could lead to the development of fractionation techniques which have the merit of not being energy-intensive. Separation processes based on surfactants might be of analytical or preparative importance in the laboratory and might also be of commercial relevance in e.g. crude oil fractionation. Relatively few papers on the topic of mixed solubilisation have a~peared,l-~ and so far no coherent picture has emerged.This may be because of the choice of system and conditions, however. In many of those systems investigated the likelihood is that different ‘ sites’ in the micelle were occupied by the different solubilising components. For example, in micelles of non-ionic surfactants of the ethylene oxide type, the ‘core’ and ‘palisade layer’ tend to be associated with non-polar and polar solubilisates, respectively, and it may be problematical to allocate ‘sites’ to different solubilisates. This paper is concerned with the question as to what happens when the ‘site’ is the same for all the solubilising components. Consider an oil phase of two components, neither of which is significantly soluble in water.When this oil phase is contacted with an aqueous micellar solution, both components become solubilised in the micelles. As the oil phase composition is varied from pure component 1 to pure component 2, the amounts of 1 and 2 solubilised must vary between the amounts found for the respective single-component oil systems. Since the rate of solubilisation in these systems is proportional to the amounts solubilised,1° this quantity must likewise vary between its values for the two single-component oil systems. The functions describing both these quantities must be continuous. The question at once arises as to whether the composition of the solubilised oil is the same as or differ- ent from that of the bulk oil phase.If different, the composition of an oil phase of finite size (such as a droplet) will vary continuously as solubilisation proceeds. Two important consequences of such a variation would be, first, the fractionation of one component from the original mixture and, secondly, the concomitant variation of the rate of solubilisation, which will tend towards its value for the least solubilised of the two components. 32053206 Solubilisation of Oils by Micellar Surfactants This report is concerned with the simplest possible behaviour a two-component oil/aqueous micellar system can show: the oil is solubilised without change in composi- tion. That is, the solubilisation process is non-selective. In such systems the initial rate of solubilisation is steady because the composition of the oil drop remains constant and no significant saturation of the micellar system has yet become apparent (see later).As the bulk oil composition is varied, the requirement then is for the amount of oil solubilised to vary continuously while having a composition identical with that of the bulk. The amount of oil solubilised must thus assume the respective one-component values when the bulk oil composition goes to pure 1 or pure 2. In what follows expressions which relate the amount solubilised to the bulk oil composition and which satisfy these constraints are considered, and attention is focused on forms which admit of a simple physical interpretation. Systems most likely to exhibit this type of behaviour will consist of chemically rather similar oils of non-polar character. Such oils tend to be solubilised in the micelle core, so that the change in chemical environment experienced by an oil molecule on solubilisation is not very great.The positing of non-polar character also effectively eliminates from present consideration one of the two principal solubilisation mechanisms, viz. diffusion of dissolved oil from the aqueous phase into the micelle, leaving the mechanism involving diffusion of micelles to the oil phaselo to predominate. The interpretation of data (especially kinetic data) is much simplified when only one mechanism is in operation. Theoretical Consider an oil phase consisting of two components 1 and 2, mole fractions x, and x2 (= 1 -xl), respectively. If a surfactant micelle (consisting of an aggregate of a constant number of monomers) solubilises b, molecules of 1 and b2 molecules of 2, the condition that the compositions of bulk and solubilised phases are identical gives x, = bl/(b, + b2); x2 = b,/(bl +b2) (1) or x1Ix2 = bl/b2' (2) The total number of molecules solubilised per micelle, b, is equal to b, + b,.molecules solubilised per micelle are by and bi, respectively, then If, when 1 and 2 are respectively solubilised as single-component oils, the numbers of b = by at x, = 1 (3 4 b = bi at x, = 0. (3 b) The functional form of b can be deduced. To satisfy eqn (2), b, and b, must be of the whereJTx,) (= b) is required by eqn (3) to have the limiting values JT1) = by; JTO) = bi. (5) in which m m (7) The physical significance of the parameters m, p and q is discussed at later points in this paper.B.J. Carroll 3207 0 x1 1 Fig. 1. Interpretation of eqn (8) and (9). There is thus a set of equations satisfying the conditions (2) and (3). It will now be shown how the first two members of this set of equations (for which m = 1 and m = 2) have a physical basis, and their use in the interpretation of experimental data is afterwards discussed. When m = 1 , b[ = f ( x , ) ] takes the form b = x , b ~ + x , b ~ . (8) (9) From which it is seen that b is simply a linear interpolation between the two pure oil datum points ( 1 , b:) and (0, b:) (fig. 1). Thus m = 1 corresponds to a linear dependence of the total number of molecules solubilised, b, on the bulk oil phase composition, x,. In order to satisfy eqn (2), the individual oils must solubilise in amounts It is readily shown that an equivalent form of eqn (8) is b = hi +x,(b; - bi) b, = x,(x, b;+x, bi).(lob) [Expression (8) also results when the assumption is made that the amounts solubilised are proportional to the bulk phase mole fraction i.e. b, = by x , etc. However, expressions of this type do not satisfy eqn (2). It is evident that, comparing this type of solubilisation with that expressed in eqn (10a and b), the condition (2) does not alter the total number of molecules solubilised per micelle, but it does affect their relative numbers.] When m = 2 the equation corresponding to eqn (6) is 4 = f l x d l = b:bl X l + P2 xf) + b&, x2 + q 2 x i ) ( 1 1) 6 = x l b ~ + x , b i - x , x , ( p , b ~ + q , b ~ ) (12) which, with the aid of eqn (7), reduces to in which p , and q2 may take values of between 0 and 1.Expressions for b, and b, then follow from eqn (4) and (12). Eqn (12) is related to eqn (8), to which it reduces when p , = q2 = 0. The cross-term in x , x , can be interpreted as a reflection of the constraints imposed on the packing of type 1 and type 2 molecules in the micelle, and will be larger when 1 and 2 differ in structure (although not in chemical character). This cross-term will be related to the configurational entropy of the solubilised system. The maximum value the term can take3208 Solubilisation of Oils by Micellar Surfactants I \4 3 0 - --. r4 9 9 0 9 9 4 0 - k 2 1 1.0 On5 x1 0 Fig. 2. Theoretical dependence of amounts solubilised on oil composition according to eqn (1Oa and b) (m = 1).is x,x,(bY+b;), corresponding to p , = q2 = 1. In general, the term can be written as x, x2 by P, where P may take values between 0 and (1 + bi/by). It is possible to analyse the behaviour of b, and b, in the two cases considered atrove. Considering first the case m = 1, differentiation of eqn (10a) gives db,/dx, = 2xl(by - bi) + bi. It follows that a turning point (a maximum) in b, occurs at x1 = b;/2(b:-bY). This maximum only appears in the range 0 < x, < 1 when bi > 2by, which latter is therefore a condition for the appearance of a maximum in b, in these systems. Differentiation of eqn (10 b) gives from which it is readily shown that, if by < bi, no turning point exists in the range 0 < x, < 1.Further, the gradient of the function b, becomes decreasingly negative as xl+ 1. Fig. 2 illustrates the predicted behaviour of b, and b, in terms of several values of the ratio B = bi/by. b, and b, are both linear when by = bi, but when by < bi (the behaviour when by > bi is completely analogous), the curve for b, is concave to the x, axis while that for b, is convex. The function b, begins to exhibit a maximum value when bi > 2hy, as expected. The function b = 6, + b, (omitted from fig. 2 for the sake of clarity) would join the coordinates of bi/by = B (selected ratio) at x, = 0 and of bi/by = 1 at x, = 1 by a straight line. In the case when m = 2 the behaviour of b, and b, as x, is varied can be similarly analysed. Allowing p , and q2 to take their maximum permitted values of 1, the curve db,/dx, = 2x1(bi - by) + by - 2biB. J .Carroll 4c 0 0: 5 3209 X1 Fig. 3. Theoretical dependence of amounts solubilised on oil composition according to eqn (12) (m = 2). for b, is found to inflect at values of x, equal to 2bi/3(by + bi) and to show a maximum for values of hi > 3by. The function b,(x,) is in all cases monotonically decreasing. Fig. 3 illustrates behaviour for representative values of B( = bi/b;). The function b = b, + b, is again omitted. It is in general a curve, concave to the abscissa, linking the points bi/by = B and bi/hy = 1. Solubilisation Rate The rate of solubilisation is defined as the rate of transport of oil from the bulk phase across unit interfacial area :lo rate = - (1 / A ) (d V/dt).(13) As it is defined in eqn (1 3) in volume terms, the prediction of rate requires knowledge - of b, and b, separately rather than of their sum. If the molecular volumes are and G, respectively, the solubilisation rate can be written as rate = constant x (b, + b, K) (14) where the constant term depends on the micellar system properties.lO9 l1 Experimentally, the solubilisation rate is especially useful for the study of the present type of system, since it is more readily measured than are the amounts solubilised at equilibrium.3210 Solubilisation of Oils by Micellar Surfactants Table 1. Solubilisation of Mixed Oils, from ref. (9) label oils benzene/hexane benzene/ hexane benzene/hexane benzene/hexane benzene/hexane benzene/cyclohexane hexane/cyclohexane benzene/cyclo hexane surfac tan t CPCU DACb SDSC DTACd AAYe CPC CPC AAY concentration /mol dmP3 0.1 0.1 0.1 0.1 0.3 0.1 0.1 0.3 B = bi/by (exptl) 3.75 3.12 4.29 9.38 4.73 2.6 1 1.54 2.67 a Cetyl pyridinium chloride.Dodecylammonium chloride. Sodium dodecyl sulphate. Dode- cyltrimethylammonium chloride. Aerosol AY. 1.0 O n 5 x1 0 Fig. 4. Data for CPC-hexan-yclohexane (1, hexane; 2, cyclohexane) (system A, table 1). Drawn lines represent theory for B = 1.54, P = 0 [ref. (9)]. Comparison with Experiment Data fitted by rn = 1 Equations Recently, Ruckenstein and co-workersg have published equilibrium solubilisation data for mixtures of hydrocarbons (permutations of two from benzene, hexane and cyclo- hexane) in micellar systems of cationic, anionic and non-ionic surfactants.The systems are summarized in table 1. It is apparent that, for a given oil mixture, the values of B = bi/b: do not vary very greatly with the surfactant, with the notable exception of system D, for which there seems to be no ready explanation. The experimental B values reported were used with eqn (1Oa and b) to predict the values of b, and b, for the several systems and agreement with theory ranged from good to fair for all systems save system D, for which the fit was very poor. Some typical data fits are shown in fig. 4-6. The fit is naturally sensitive to the accuracy with which the experimental parameter B is known, but the general features of the data are quite well forecast. In particular, the position and heights of the maxima in fig. 5 and 6 are in good accord with theory, and the absence of any maximum in the data in fig.4 ( B = 1.53)B. J . Carroll 3/ 2 0 - -. P( 9 9 b 0 - 9 9 -2 1 321 1 0 1 Oo5 x1 0 Fig. 5. Data for system H, table 1. 1, Cyclohexane; 2, benzene. Drawn lines represent theory for B = 2.67, P = 0 [ref. (9)]. 1.0 x1 0 0.5 Fig. 6. Data for system B, table 1. 1, Hexane; 2, benzene. Drawn lines represent theory for B = 3.12, P = 0 [ref. (9)]. FAR 1 1063212 Solubilisation of Oils by Micellar Surfactants 0.5 1.0 x1 Fig. 7. Data for system dodecane-hexadecane-C,,E,. Drawn lines represent theory for B = 3.06, (a) P = 0, (b) P = 1.3 and (c) P = P,,, [ref. (1 1)) is also in agreement with the prediction that such maxima only appear when B > 2. Note that the values of B quoted refer in all cases to the experimental ratio of the experimental quantities bi and b:, and a better fit to the data could in some cases be made by taking a different value for B within the range possible because of experimental error.Very probably the m = 1 equations fit the data well without recourse to the cross-term of eqn (1 2) because the solubilisate molecules are small and fairly equal in volume, thus minimising packing constraints in the micelle (corresponding to a high mixing entropy). The extent to which theory and experiment agree is sometimes surprising: the usually accepted picture for the solubilisation of aromatic molecules is that they tend to solubilise at least partially in the palisade layer of the micelle, so that in these systems the postulate of one-site solubilisation may not be valid. This could, however, be why some of the data for benzene lie somewhat above the expected values (see e.g.fig. 6 ) , whereas the fit for the less polar oils is generally good. Data fitted by m = 2 Equations Systems in which the solubilisates are long-chain molecules are associated with greater constraints on packing (i.e. they have a lower mixing entropy) in the micelle than are systems of smaller molecules, so it is to be expected that the cross-term in eqn (1 2) will taken on appreciable values. In fact, it is found in such systems that the neglect of the cross-term results in predicted solubilisation rates which are much higher than those found experimentally. It is therefore essential to include this term. The quantity P then assumes non-zero values, up to a maximum of (1 +B).Systems which have been studied kinetically include n-dodecane-n-hexadecane/ 1 %B. J. Carroll 321 3 1 0 0.5 x1 Fig. 8. Data for system hexadecane-squalane-C,,E,. Drawn lines represent theory for B = 17.6, (a) P = 0 and (b) P = PmaX [ref. (12)]. w/w C,,E, [ref. (1 l)] and n-hexadecane-squalane/ 1 % w/w C,,E, [ref. (1 2)] (squalane is a branched-chain alkane, C3,,H62). The rate of solubilisation of a small drop of the mixed oil was studied using the drop-on-fibre technique,1° in which the oil/surfactant ratio is so low that only initial rates are obtained throughout the life of the oil drop. In all the systems studied it was found that the rate of solubilisation was independent of time, indicating that the drop composition was not changing.(This is a most useful check on the absence of fractionation and so on the suitability of a system for the application of the theory developed above. The kinetic experiment is so arranged that complete solubilisation of the test drop occurs without any appreciable saturation of the micellar system. Constancy of the rate therefore implies constancy of the oil composition .) From the kinetic data the amounts solubilised can be deduced, and these are plotted in fig. 7 and 8. It is found that for the dodecane-hexadecane system a value of the coefficient P of 1.3 gives an excellent fit to the data (fig. 7). (Theoretically, this coefficient may take values from 0 to 4.38.) The very poor fit obtained when P = 0 is also shown.In the case of the hexadecane-squalane system, however, the best fit is obtained when the coefficient is given its maximum allowable value Pmax = 1 + B (fig. 8). This reflects the severity of the constraints imposed on the packing of two long-chain molecules of disparate length and structure into the micelle core. 106-23214 Solubilisation of Oils by Micellar Surfactants Conclusions A theory for the non-selective solubilisation of a binary oil mixture by a micellar system has been developed. Starting from data describing the respective one-component oil systems, equations are developed which describe the amounts solubilised and the rate of solubilisation in these systems. The theory employs a single parameter, P which may only take values in a certain range and which is related to packing constraints in the micelle. The theory is most pertinent to systems consisting of mixtures of non-polar oils, which tend to solubilise in the micelle core. Good agreement with some experimental results for such systems is obtained. References 1 F. Tokiwa, Bull. Chem. SOC. Jpn, 1970, 43, 939. 2 F. Tokiwa and K. Tsujii, Bull. Chem. SOC. Jpn, 1973, 46, 1338. 3 F. Alhaique, D. Giacchetti, M. Marchetti and F. Riccieri, J. Pharm. Pharmacol., 1976, 29, 401. 4 T. Lovgren, B. Heikius, B. Lundberg and L. Sjoblom, J . Pharm. Sci., 1978, 67, 1419. 5 B. Lundberg, T. Lovgren and B. Heikius, J. Pharm. Sci., 1979, 68, 542. 6 V. A. Volkov and A. I. Shulitskaya, Kolloidn. Zh., 1981, 43, 752. 7 R. Nagarajan and E. Ruckenstein, Sep. Sci. Technol., 1981, 16, 1429. 8 D. C. Thomas and S. Christian, J. Colloid Interface Sci., 1981, 82, 430. 9 M. A. Chaiko, R. Nagarajan and E. Ruckenstein, J. Colloid Interface Sci., 1983, 99, 168. 10 B. J. Carroll, J. Colloid Interface Sci., 1981, 79, 126. 11 P. Faulkner, M S c . Thesis (National University of Ireland, 1985). 12 B. G. ORourke, B. J. Carroll and A. J. Ward, to be published. Paper 512229; Received 18th December, 1985

ISSN:0300-9599    

DOI:10.1039/F19868203205

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1986,82, 3215-3231 Decay of High-valent Manganese Porphyrins in Aqueous Solution and Catalysed Formation of Oxygen Anthony Hardman,* Paul A. Christensen and George Porter Davy Faraday Research Laboratory, The Royal Institution, 21 Albemarle Street, London WlX 4BS Kim Morehouse Radiation Laboratory, University of Notre Dame, Notre Dame, Indiana 46556, U.S.A. Pedatsur Neta and Marie-Claude Richoux Center for Chemical Physics, National Bureau of Standards, Gaithersburg, Maryland 20899, U.S.A. Manganese(m) porphyrins (MnII'P) are easily oxidised to the corresponding MnIVP in alkaline aqueous solution. At pH < 5 the oxidation product is a MnI'IP 71-radical cation. These oxidised metalloporphyrins have limited stability in water and they revert to the original Mnl"P upon standing in the dark.The rate and mechanism of this inherent reduction process depends upon pH, with lower pH giving the higher rates. The inherent reduction appears to involve disproportionation and rearrangement of the MnIVP but it does not lead to formation of molecular 0,. Addition of colloidal RuO, - 2H,O, a good 0,-evolving catalyst, has a pronounced effect upon the reduction process. The oxidised metalloporphyrin is bound to the catalyst particles by electrostatic forces, and at pH < 11 the bound material decays more slowly than the free compound. For 8 c pH c 11, decay of the bound metalloporphyrin involves oxidation of water to 0,, but the yield of 0, is much less than the stoichiometric value. Previous have described the redox equilibria established between the various oxidation states of the central manganese ion sited within the porphyrin ring.The stable form is the manganese(n1) porphyrin (MnIIIP), but it is easily reduced to the corresponding MnIIP. This latter species is stable only in alkaline solution (pH > 8) and in the absence of 0,. Oxidation of MnIIIP forms the corresponding MnIVP, which has limited stability and is reduced back to the original MnIIIP upon standing in the dark even at pH 14.5 Several research group^^-^ have shown that the rate of this reduction process increases with decreasing pH. Actual reduction rate constants have been published only for the MnIV/MnlI1 haematoporphyrin coupleQ, where it has been reported that the side-chain provides the electrons necessary for the reduction of the central metal ion.' In this paper we report kinetic parameters associated with the decay of various MnIVP obtained by derivatisation of the meso-tetraphenylporphyrin class.The process under investigation is of interest because of the possible involvement of manganese in green-plant photosynthesis. Much speculation exists about whether or not a manganese complex functions as the natural water oxidation catalyst. If it does then surely it is a high-valent manganese complex that is the active catalyst. Here we give particular attention to the possible oxidation of water to 0, by MnIVP with and without an added redox catalyst. Experiment a1 Manganese(111)tetrakis(4-sulphonatophenyl)porphyrin (sodium salt) (MnIIITSPP), man- ganese(111)tetrakis(4-~arboxyphenyl)porphyrin (sodium salt) (MnIIITCPP), mangan- 321 53216 Decay of Manganese Porphyrins in Solution ese(rr1) tetra(4-pyridy1)porphyrin (MnIIITPyP) and manganese(II1) tetrakis(N-methyl- 4-pyridy1)porphyrin (chloride salt) (MnIIITMPyP) were prepared and purified as described beforelo or purchased from Midcentury Chem.Co. All other materials were of the highest available purity and were used as received. For the pulse- radiolysis studies, water was obtained freshly from a Millipore Milli-Q system. For the stopped-flow and membrane polarographic detector studies, water was deionised and doubly distilled from a Fisons quartz still. It was then redistilled from alkaline permanganate to ensure that no reducing impurities remained.In all cases water was used within a few hours of final purification. The pH was adjusted by addition of the appropriate buffer; HCl/NaCl pH < 4; NaH,PO,/Na,HPO, 4.5 < pH < 9.0; Na,B,O,/KOH, 8.8 < pH < 10.8; KOH pH > 11. Where possible, the concentration of buffer was 10-3moldm-3 and all kinetic measurements were made at a fixed ionic strength of 0.05 mol dm-3. Colloidal RuO, 2H,O was made as before.ll The general conditions used in the pulse-radiolysis experiments have been described in detail.,, l2 Pulses of 8 MeV electrons from an ARCO-LP-7 linear accelerator were usually of 10 ns duration with doses (300-600 rad) that produce (2-4) x mol dm-3 radicals from water. An electronic shutter and the proper interference filters were used to protect the reaction solution from undue photolysis by the analysing light.Dosimetry was performed with N,O-saturated KSCN s01utions.l~ Aqueous solutions of the MnIIIP [(5 x 10-6)--10-4 mol dm-3] containing KBr (lo-, mol dmP3) at the required pH were saturated with N,O before irradiation. The course of reaction was followed by optical absorption and the resultant signals were averaged.12 Stopped-flow experiments were performed with an Applied Photophysics or a Durrum instrument (dead-times 2 ms). Solutions of the MnrTIP (4.5 x lop6 mol dm-3) in water containing buffer and sufficient Na,SO, to give a total ionic strength of 0.05 mol dme3 were mixed with an aqueous solution of Br, (2 x lo-, mol dm-3) at the same pH and ionic strength. Again, the course of reaction was followed by absorption spectroscopy and decay kinetics were analysed by iterative computer fitting to pseudo-first-order processes. At pH > 12 oxidation of a MnIIIP can be achieved by addition of the (electron equivalent) stoichiometric amount of NaOC1.l The product is sufficiently stable to be used as a reactant in further stopped-flow experiments.Thus MnTVP was prepared by stoichiometric oxidation of the corresponding MnIIIP at pH 13 and mixed in the stopped-flow instrument with an aqueous solution containing acid and/or buffer. The rate of decay of MnrVP was followed at 430 nm. In all the stopped-flow experiments the reported rate constants were the average of at least three independent measurements. Concentrations of dissolved 0, were monitored with a purpose-built membrane polarographic detector (m.p.d.)14 and were displayed on a x-t chart recorder.Aqueous solutions of the MnIIIP (2 x lop5 mol dm-3) at pH 13 were oxidised within the reaction chamber of the m.p.d. by addition of the (electron-equivalent) stoichiometric amount of NaOCl; and the solution was purged with 0,-free N,. An aliquot of N,-purged acid and/or buffer was added to lower the pH and the concentration of evolved 0, measured. Measurements were repeated at least 5 times at each pH value. In related experiments, the MnIIIP solution was thoroughly purged with N, within the chamber of the m.p.d. and the required amount of N,-purged Br, or NaOCl solutions added via a septum. Again, the change in concentration of dissolved 0, was measured.', Blank experiments were made in all cases.These measurements consisted of repeating the above 0, determinations in the absence of each of the reactants. Operating the m.p.d. in its most sensitive mode gives a baseline drift of ca. 3 mm h-l. Injection of N,-saturated water does not perturb the baseline, but injection of sufficient 0,-saturated water to give a final 0, concentration of 5 x lo-' mol dm-3 corresponds to a 22 mm deflection on the chart recorder. The risetime for this deflection is ca. 15 s.A . Harriman et al. 3217 Experiments involving colloidal RuO, - 2H,O were made by identical procedures to those outlined above. In all cases the colloidal catalyst was mixed with the aqueous solution of the MnIIIP before the experiment. Results and Discussion Oxidation of Manganese(@ Porphyrins In alkaline aqueous solution (pH > 12), an Mn'IIP is readily oxidised to the corres- ponding MnIVP, which possesses reasonable stability under such conditions.'? Oxida- tion can be achieved with a wide variety of oxidants, including H,O,, K,Fe(CN),, KMnO,, NaOC1, KBrO, and Na,S,O,, and rate constants for the oxidation step have been reported.2 Identical oxidations proceed under pulse-radiolytic conditions using Br; as ~ x i d a n t .~ Electrochemical studies15 have shown that the redox potential for one- electron oxidation of an MnIIIP increases with decreasing pH throughout the range 7 < pH < 14. This effect is manifest in the rate constant for oxidation of a given MnlIIP as a function of pH. Thus, with Br; as oxidant, the bimolecular rate constant for oxidation of the various MnIIIP (kox) undergoes4 an approximate tenfold decrease upon lowering the pH from 13 to 7.Even so, oxidation with Br; under pulse-radiolytic conditions in neutral solution is fast and the formation of MnIVP is quantitative.* Using Br, as oxidant, k,, for oxidation of MnIIITMPyP was measured as a function of pH using the stopped-flow technique : 2MnIIIP + Br, -+ 2MnIVP + 2Br-. (1) At a constant ionic strength of 0.05 mol dm-3, k,, varied considerably with pH, giving values of 1 x lo2, 4 x lo5 and 1.5 x lo7 dm3 mol-1 s-l, respectively, at pH 7, 9 and 11. These large changes are due to hydrolysis of Br, at alkaline pH : 3Br, + 30H- $ BrO; + 5Br- + 3H+. (2) As a consequence of such changes, electrostatic forces between the reactants and the redox potential of the oxidant vary over a wide range so that changes in the observed k,, values cannot simply be ascribed to changes in the structure of the MnIIIP.In the present investigation we are concerned with decay of the MnIVP, and it is important that the oxidation step is complete well before decay begins. In the pulse- radiolysis experirnent~,~ the formation reaction is complete well within 100 ps of the pulse, so that decays occurring on the millisecond timescale can be followed without complication. Chemical oxidations performed at pH > 12 require only stoichiometric amounts of oxidant, and the product lives for at least many minutes.'? Again, there are few problems associated with such studies. At lower pH the rate of oxidation with Br, is slower and excess oxidant must be used.For the stopped-flow studies the concentration of MnIIIP was 4.5 x mol dm-3, and for 8.2 < pH < 12.0 it was necessary to use a Br, concentration of 2.5 x mol dm-3 to ensure that oxidation was complete before the onset of decay. At pH 7 a large excess of Br, was needed to achieve oxidation, and this precluded meaningful measurement of decay kinetics. Thus, the stopped-flow studies were restricted to pH > 8.2. Even in this range the decay kinetics will be distorted because of the excess of Br, present. Absorption spectra recorded for the MnIVP product formed at 7 < pH < 14 showed no obvious differences apart from the known pK transitions' which occur at ca. 7.5 and 10.5. However, pulse-radiolytic oxidation of various MnIIIP at pH < 3 gave4 a product with an absorption spectrum markedly different from that characteristic of MnIVP1 [see fig.7D in ref. (411. The strong absorption found4 around 650-700 and 490 nm suggested that the product was a metalloporphyrin ;n-radical cation. Addition of sufficient HNO,3218 Decay of Manganese Porphyrins in Solution Fig. 1. Differential absorption spectrum of MnlIITMPyP+ as formed by adding sufficient HNO, to a solution of MnIVTMPyP at pH 13 such that the final pH falls to 2.2. PH Fig. 2. Energy-level diagram showing the pH dependence of the various redox couples used in this work. The MnIV/MnlIITMPyP data are taken from ref. (15) and the other redox potentials have been taken from standard textbooks. The half-reactions involved are : (a) Br, + e = 2Brr; (b) C10-+2H++2e = C1-+H20; (c) Ru04+4H++4e = Ru02+2H,0; (d) 02+4H++4e = 2H,O; (e) MnIVTMPyP + e = Mn'IITMPyP; (f) MnlIIP+ + e = MnIIIP.The redox potential for process df) has not been measured in water and the value given is that determined for MnII'TPP in CH,C12 solution as ref. (2). to a solution of MnrVP prepared at pH 13 to give a final pH of 2 gave a transient absorption spectrum similar to that found in the pulse-radiolysis experiments (see fig. 1). Thus at low pH the oxidising equivalent seems to be centred on the porphyrin ligand rather than on the central Mn ion. This is to be expected, since the redox potential15 for the MnlI1/MnIV couple increases by 118 mV per pH unit upon decreasing the pH below 10, whereas the redox potential for oxidation of the porphyrin ring should be independent of pH (fig.2). Consequently, high pH favours formation of MnIVP and low pH favoursA . Harriman et al. 3219 formation of the n-radical cation. The transition point depends on the porphyrin structure and is at a higher pH for TMPyP than for TSPP: Mn(I11)P + oxidant ' \ (3) Oxidation of Manganese(1v) Porphyrins It is known2 from chemical studies performed with MnTCPP that oxidation with NaOCl at pH 14 results in formation of MnVTCPP. This was confirmed by pulse-radiolysis experiments in which MnIVP was formed chemically by oxidation with Br, at pH 11-14 and then oxidised further with Br; formed in the pulse: (4 a ) MnIVP + Br; -+ MnVP + 2Br- or MnIVP + Br; -+ MnIVP+ + 2Brr. (46) The MnIIIP was chemically oxidised in a mixing chamber just before entering the cell.In the radiolysis cell Br;, produced by the electron pulse, oxidised the MnIVP. This oxidation was very rapid and the rate constants for oxidation of the various MnIVP are collected in table 1. The differential absorption spectra for oxidation of the MnIVP are shown in fig. 3-5. The spectra observed in the case of MnTCPP and MnTSPP at pH 14 [fig. 3(a) and 4(a), respectively] exhibit a very sharp absorption band at ca. 430 nm, along with other bands around 550-600 nm. By comparison, with the published2 spectra of MnIVTCPP and MnVTCPP, these spectra are assigned to the corresponding MnVP. However, as the pH is decreased, oxidation of MnIVTCPP [PH 13, fig. 3(6)] and MnIVTSPP [pH 1 1, fig.4(6)] results in formation of products having broad absorptions around 500 and 600 nm. These products are more characteristic of metalloporphyrin n-radical cations than high-valence metalloporphyrins.16-18 MnIVTMPyP seems to be oxidised to the n-radical cation even at pH 14 (fig. 5). The structure of these species will be discussed further in a later section. Decay of MnVP The products of oxidation of MnIVP decay relatively quickly. The first half-lives derived from the pulse-radiolysis experiments (table 1) vary from 60 ms for MnTSPP to 4 ms for MnTMPyP, both at pH 14. The oxidised form of the TMPyP complex is found to be always less stable than that of the TSPP derivative. In the present case, the nature of the oxidation products is also different, i.e.MnVTSPP and MnIVTMPyP+. Further details on the structure and mechanism of decay of these species are given below. Decay of MnwP As discussed above, oxidation4 of MnIIIP at pH < 5 results in formation of the corresponding MnlIIP+ rather than the MnIVP. It is not possible to oxidise the Mn1I1P at pH < 5 by chemical methods, but pulse-radiolytic oxidation4 with C1; proceeds smoothly at pH 3. Also, addition of acid to an aqueous solution of MnIVP prepared at pH 13 so that the final pH falls to < 5 leads to intermediate formation of MnlIIP+. The n-radical cation decays fairly slowly, the kinetics giving a reasonable fit to a3220 Decay of Manganese Porphyrins in Solution Table 1. Rate constants for the oxidation of various MnTVP by Br; radicals and the first half-lives for the oxidation products at that pH Wl-1 k4/ 1 O8 porphyrin pH /mol dm-3 dm3 rnol-' s-l films MnIVTSPP 14 0.10 I 1 60 12.9 0.0 1 6.0 20 11.1 0.01 7.8 - MnIVTCPP 14 0.10 18 4 12.9 0.01 16 4 MnIVTMPyP 14 0.0 1 4.5 4 60 40 20 0 -20 -40 0 5 5 -60 2 3 .3 9 8 +-I I a 4 0 -44 - 8 300 400 500 600 700 800 X/nm Fig.3. Differential absorption spectra observed upon the one-electron oxidation of MnIVTCPP with Br, in N,O-saturated aqueous solution. (a) KBr (0.1 mol dm-3), KOH (1.0 mol dm-3), recorded 0.5-0.7 ms after the electron pulse; (b) KBr (0.01 mol dm-3), KOH (0.1 mol dm-3), recorded 50-200 ps after the electron pulse. first-order process. For a given porphyrin, the rate of decay is independent of the initial concentration of Mn1I1P+, ionic strength and presence of dissolved 0,.Measurements made at pH 1.8 showed that the first-order decay rate constant depended upon the nature of the porphyrin periphery groups. For example, kred values of 300 & 8, 34 f 8, 60 10 and 100 _+ 12, respectively, were obtained with MnTMPyP, MnTPyP, MnTSPP and MnTCPP. In all cases, the rate of decay of MnlIIP+ measured at 430 nm agreed well with the rate of formation of MnIIIP measured at 470 nm.A . Harriman et al. 3221 - 1 5 1 10 5 0 -5 - 1 0 -15 I I I I I 1 300 400 500 600 700 800 h/nm Fig. 4. Differential absorption spectra observed upon the one-electron oxidation of MnIVTSPP by Br; in N,O-saturated aqueous solutions. (a) KBr (0.1 mol dmP3), KOH (1 .O mol dm-3), recorded 80-140 ,us after the electron pulse; (b) KBr (0.01 rnol dm-3), KOH (0.1 mol dm-3), recorded 160-280 ,us after the electron pulse; (c) KBr (0.01 mol drn-9, pH, 11.1, recorded 0.5-0.8 ms after the electron pulse.The mechanism for reduction of the n-radical cations, under such conditions, is far from clear. Relative to n-radical cations derived from diamagnetic metallop~rphyrins,~~~ 2o MnIT1P+ is fairly stable, and the observed dependence upon periphery group is consistent with known trends.lg Experiments made with an m.p.d. showed that the reduction step does not involve oxidation of water to 0,. More probably, MnlIIP+ reacts with trace impurities in the solution, undergoes disproportionation to form a n-dication,21 or adds a hydroxide ion at the meso position.22 At pH 13, quantitative oxidation1 of MnTIIP can be achieved by addition of the stoichiometric amount of NaOCl or NaBrO,. The MnIVP product is sufficiently stable (t1/2 > 2 h) for its reactions to be studied by conventional techniques5 Thus, stopped-flow studies performed with MnTVTMPyP showed that quantitative reduction occurred upon3222 Decay of Manganese Porphyrins in Solution 12 6 W flm e $ 0 W > .- w - 2 -6 -1 2 300 400 50 0 600 700 8 00 A/nm Fig.5. Differential absorption spectra observed upon the one-electron oxidation of MnIVTMPyP by Br; in N,O-saturated aqueous solution containing KBr (0.1 mol dm-3) and KOH (1.0 mol dm-3). Spectra recorded 0, 10-25 ps and 0, 0.2-0.3 ms after the electron pulse. mixing with H2S (kred = 2.8 x lo4 dm3 mol-1 s -l) or benzhydroquinone (kred = 2.0 x lo3 dm3 mol-l s-l): ( 5 ) 2MnIVP + H2X -+ 2MnIIIP + 2H+ + X.Even in the absence of an added reductant, MnlIIP is re-formed upon standing in the dark overnight, although there is some loss (ca. 10 % ) of porphyrin. The kinetics for this inherent reduction were found to be complex. Approximating the decay rate of MnIVTMPyP to a pseudo-first-order process gave kred = (2.5 f 1 .O) x s-l, which would correspond to a bimolecular rate constant for reduction by water of (4.5 f 1.8) x Lowering the pH of the bulk solution by addition of buffer or performing the oxidation at pH c 13 leads to a marked increase in the rate of decay of the MnIVP product. Throughout the range 1 1 < pH < 14, the decay kinetics were complex and the decay rates were slow. Decay profiles recorded at 430 nm (decay of MnIVP) and 470 nm (return of MnIIIP) were similar, showing that the principal decay process involved reduction of MnIVP to MnIIIP, although there was ca.10% overall loss of porphyrin in each case. Fig. 6 gives some typical traces showing the decay of MnIVTMPyP at different pH, as monitored at 430 nm using stopped-flow techniques with bromate as the oxidant. Bleaching of the porphyrin is evidenced by the observation that the final absorbance falls below the initial baseline. This effect is more pronounced at higher pH. The decay profiles give poor fits to second-order kinetics and better fits to first-order processes. A first-order fit to the decay profile obtained at pH 12.5 is given as an insert to fig. 6. Analysis of the decay data by an iterative fitting procedure gave first-order rate constants for reduction of MnIVTMPyP of 0.029, 0.019 and 0.014 s-l at pH 11.3, 11.9 and 12.5, respectively (each rate constant being the average of at least three independent measurements).Reduction rate constants measured for the other MnTVP showed a similar trend with pH but were generally slower. For example, at pH 12.5 kred values of 3 x 1 x and 1 x respectively, were obtained with MnIVTPyP, MnIVTSPP and MnIVTCPP. As mentioned above, the oxidised phenylporphyrins are more stable than the analogous pyridylporphyrins. dm3 mo1-I s-l.A . Harriman et al. 3223 256 * 128, -0 40 80 120 160 200 t l s 256r t l s Fig. 6. Inherent decay of MdVTMPyP at different 0.4L 0 0 40 80 120 160 t l s pH as measured at 430 nm after stopped-flow formation with Br,.(a) pH 12.5, (b) pH 11.9, (c) pH 11.3. Trace (a) shows an atternpied fit to first-order kinetics for the decay profile observed at pH 12.5. Oxidation of MnIIIP in aqueous solution between pH 5 and 11 results in formation of the corresponding MnIVP. There is no spectroscopic evidence to show that the oxidation product is dimeric, although the absorption spectra are quantitatively and qualitatively similar to those obtained for the MnIVP at pH > 11. Pulse-radiolysis experiments have not resolved an absorption change that could be due to dimerisation, but such changes might be small. Consequently, the structure of the oxidation product is not clear, but it is possibly monomeric. Other high-valent metalloporphyrins are to favour formation of p-0x0 dimers at higher pH, so that the observed pK transition found around pH 1&11 might reflect both changes in the state of axially bound water molecules and the changeover from monomer to dimer.Measurements made with MnIVTMPyP showed that decay of the oxidised species occurred by two well resolved, first-order steps throughout the range 7 < pH < 1 1 (fig. 7). As before, monitoring at 430 and 470 nm gave similar profiles, suggesting that MnIIITMPyP is re-formed in two separate steps. The first-order rate constant observed for the fast step in the reduction reaction (k,) is pH dependent, as shown in fig. 8 and table 2. For MnIVTMPyP k , shows a non-linear dependence upon pH (the other porphyrins show similar behaviour). At pH < 5, k , refers to reduction of MnlIIP+ whilst at pH > 1 1 the fast (k,) and slow (k,) steps cannot be resolved.There is a clear increase in k, upon decreasing the pH below 11, suggesting that the p-0x0 dimer undergoes a pK transition at this pH. Earlier work,, relying on absorption spectroscopy, reported a pK for MnIVTMPyP of ca. 10.5. The kinetic measurements infer that the protonated species, with its higher redox potential (fig. 2), disproportionates faster than the basic form. This pK transition might involve dissociation of the p-0x0 dimer into monomeric MnIVP. In this case it is necessary that the monomer is reduced faster than the dimer.3224 4- 3 - 2 - 1 - - -Y M 4 -1 -2 -3. Decay of Manganese Porphyrins in Solution - - 25 6 192 - L d 5 128 .- v) 64 t i s tl S Fig, 7.Typical profile for decay of MnIVTMPyP at 430 nm following stopped-flow formation with Br, at pH 10.6. . -\\ \ \ \ \ \ \ -41 I I I I I 1 1 0 2 4 6 8 1 0 1 2 1 4 PH Fig. 8. Plot of log k , us. pH for decay of MnIVTMPyP as measured at 430 nm following stopped-flow formation with Br,. Each point is the average of at least three independent measurements. Within the limited amount of data, it appears that k, shows no pH dependence throughout the range 7 < pH < 11. However, the magnitude of k, does depend upon the nature of the porphyrin ring, as shown by the data in table 2. Detailed experiments performed with the m.p.d. showed that the inherent decay of various MdVP in water at 2 < pH < 13 did not result in measurable yields of 0,. On the basis of such studies, we calculate that if 0, is formed as a reaction product (andA .Harriman et al. 3225 Table 2. Rate constants for reduction of various MnIVP in aqueous solution at pH > 7 and of Mnl'IP+ at pH < 7" PH TMPyP 1.8 2.2 3.7 5.1 7.1 8.2 8.8 9.4 10.6 11.3 11.9 12.5 13.0 300 280 - 4.2 - 2.2 (0.010) 2.3 0.42 (0.01 1) 0.085 (0.008) 0.029 0.0 19 0.014 TSPP - 60 70 - 1 .o 0.03 0.009 (0.002) 0.007 (0.002) 0.004 0.00 1 - - - - TPyP 34 10 8 6.5 0.024 0.014 (0.003) 0.010 (0.002) 0.01 2 (0.002) 0.007 (0.002) 0.005 0.005 0003 0.00 1 ~~ a At pH > 11 and pH -= 7 the decay profile was fitted to a single exponential process but for 7 < pH < 11 the decay kinetics required a two-exponential fit corresponding to k , and k,. The numbers in parenthesis refer to k,.it is not consumed by reaction with MnlIP) then its yield must be < 10% of the total expected from the stoichiometric oxidation of water: 4MnlvP + 2H,O --+ 4MnIIIP + 4H+ + 0,. (6) In separate experiments it was shown that both MnIIIP and MnIVP are able to induce decomposition of H,02 : 2H20, -+ 2H,O + 0,. (7) but extensive loss of the porphyrin occurs. Thus, formation of H,O, has not been confirmed in our experiments. Mechanisms of Decay Previous work2 has suggested, on the basis of magnetic susceptibility measurements and absorption spectroscopy, that the MnIVP formed at pH 13 exists in the form of a p-0x0 dimer : (8) PMnIV-0-MnIVP. Assuming that the oxidation product is indeed a p-0x0 dimer, the most likely mechanism for the reduction process involves disproportionation within the dimer, explaining the observed first-order kinetics : (9) As mentioned earlier, the preference for formation of MnVP or MnIVP+ depends upon the nature of the porphyrin periphery groups and upon pH.In most cases, the PMnlV-O-MnlVP -+ MnIIIP + MnIVP+ (or MnVP).3226 Decay of Manganese Porphyrins in Solution disproportionation product is expected to be the MnIVP+. This species is fairly short-lived (t1,2 x 10 ms at pH 13) and it decays to the original MnIIIP: MnIVP+ + 2H20 --+ MnlIIP + H202 + 2H+. (10) The poor stability of the MnIVP+ product shows that the rate-determining step in the reduction of the MnIVP, under such conditions, is the primary hydrolysis of the pox0 dimer. Previous worker^^^^ have shown that the inherent decay of manganese(1v) haemato- porphyrin results in formation of peroxide. We did not identify H202 as a reaction product in our systems, but in control experiments it was found that H,O, bleached the porphyrin ring.Thus the observation that the inherent reduction of an MnIVP involves some overall loss of porphyrin is consistent with the formation of H,02. Studies performed with a m.p,d. showed that the reduction process did not result in formation Decay of MnrVP at 7 -= pH < 11 occurred by two first-order steps, again consistent with a mechanism involving disproportionation within the dimer. If the MnIVP does not exist predominantly as a dimer at pH < 11, then we assume that it equilibrates rapidly with a small percentage of dimer which disproportionates. To account for the slow step in the re-formation of MnIIIP it is necessary that the MnIVP+ is reduced to MnIIIP more slowly at pH < 11 than observed in the pulse-radiolysis experiments at pH > 11.This is certainly the case with other metalloporphyrin radical cations,lg? 2o but we could not confirm it for MnIVP+ because the poor stability of MnIVP at pH < 12 prevents its use as a reactant in the pulse-radiolysis set-up. The products of disproportionation [reaction (9)J formulated as MnVP or MnIVP+ may exist in several forms: of 0,. I While the form favoured at very high pH is written as MnVP, recent has proposed an alternative assignment as MnIVP with the extra oxidising equivalent centred on an axially coordinated oxygen. Whatever the real nature of this species, it is known from the pulse-radiolysis studies that it is stable only at very high pH.Lowering the pH (to < 14 for MnTMPyP or c 13 for MnTSPP) favours formation of MnIVP+ (fig. 2). It is possible that lowering the pH further favours formation of Mn11rP2+, but there is no evidence for such a product in our experiments. The formulation of the high-valent MnP as 'OMnIVP provides a direct route for formation of hydrogen peroxide via a peroxo dimer : 28A . Harriman et al. 3227 0' I I 2 M n(IV) P - PM n(Iv) --Or\-Mn(rv) P 1 OH I OH OH / 2 H 2 0 OH 1 I 2M'n(IV)P + H202 I OH A two-electron change, involving intermediate formation of a MnIIP is a possible alternative to the above one-electron scheme. However, the characteristic absorption spectrum1 of MnIIP has not been observed, even as a transient, and the presence of 0, has no effect upon k,.Furthermore, MnIIP undergo acid-catalysed demetallati~n~~ and, although this reaction is fairly slow at pH > 5, we might have expected to detect some demetallated porphyrin. Catalysed Decay of MdVTMPyP A series of stopped-flow experiments was performed to investigate the influence of redox catalysts upon the decay kinetics of MnTVTMPyP. Experiments were made in two ways. First, chemical oxidation of MnIIITMPyP was achieved by addition of the stoichiometric amount of bromate at pH 13. The resultant solution was mixed with acid and/or buffer containing a known concentration of colloidal Pt or CoSO,. Decay of MnIVTMPyP was measured as a function of the concentration of added catalyst at a particular pH.Throughout the range 2 < pH < 13 neither colloidal Pt ([Pt]) < 2 x lo-, mol dm-3) nor CoSO, ([Co2+] < 5 x lo-, mol dm-3) had any observable effect upon the rate of decay of MnIVTMPyP. In the second set of experiments, buffered aqueous solutions of MnIIITMPyP containing various concentrations of colloidal RuO, ' 2H20 were mixed with aqueous solutions of Br, at the same pH. Again, the rate of decay of MnrVTMPyP was measured as a function of the concentration of added catalyst. The decay profiles were analysed in terms of pseudo-first-order reactions by iterative computer fitting. As found with the inherent uncatalysed decay, reaction with the catalyst led to re-formation of the original MnIIITMPyP. At pH 11 to 13, addition of colloidal3228 Decay of Manganese Porphyrins in Solution 0.6 0.5- I v1 1 -Y 0.4 0.3- 0.2 0.1 0.0 128 64 I - I I I 0 I I I - I I I I - \ 96 - ' I I I 0 5 10 15 0.20- 0.15- d 9 0.10 - -Y 0.05 - (aiii) 1 I I 1 8 16 24 32 1 t l s 0 I 1 1 1 0 0.5 1.0 1.5 2.0 2.5 3.0 [ RuO,]/ mol dm-3 (aii) I I I I 8 16 24 32 t/ s Fig.9. For legend see opposite.A . Harriman et al. 3229 RuO;2H20 resulted in a small but significant increase in the rate of decay of MnIVTMPyP. Typical decay profiles obtained in the presence of various amounts of catalyst are shown in fig. 9(a). Throughout this limited pH range there appears to be a linear correlation between the pseudo-first-order decay rate constant (kred) and the concentration of added RuO, - 2H,O, as shown by fig.9 (b). However, careful study of the experimental records shows that colloidal RuO, 2H,O does more than simply increase the rate of decay of MnrVTMPyP. In the presence of catalyst, the absorbance at 430 nm observed immediately after mixing [fig. 9 (a)] is higher than found in the absence of RuO;2H20. Also, high concentrations of catalyst seem to retard the rate of oxidation of MnIIITMPyP. These findings suggest that the positively charged metalloporphyrin is bound to the negatively charged colloid surface by electrostatic forces. Such adsorption would inhibit oxidation with bromate owing to Coulombic repulsion and also prevent dimerisation of MnrVTMPyP. Under such conditions the oxidised metalloporphyrin would be a surface-bound, monomeric MnrV- TMPyP, which is expected to have a higher molar extinction coefficient than the pox0 dimer.30 With lower concentrations of RuO, - 2H,O there is not enough colloid surface for complete binding and MnIVTMPyP will be distributed as surface-bound monomer and non-adsorbed p-0x0 dimer.The observation that the rate of decay of MnlVTMPyP increases in the presence of colloidal RuO, * 2H20 can then be explained in two ways. Either the surface-bound monomer has a higher inherent kred than the p-0x0 dimer, or monomer and/or dimer can abstract an electron from the colloid. Over the pH range 1 1-14, the measured15 redox potential for the MnIV/MnI1ITMPyP couple suggests that water oxidation is unlikely (fig. 2), and studies with the m.p.d. failed to detect 0, as a reaction product. Therefore, reduction of MnrVTMPyP under these conditions does not appear to involve oxidation of water to 0,.Consequently, the most likely explanation for the observed increase in kred upon adding colloidal RuO, * 2H,O is that adsorbed, monomeric MnIVTMPyP shows a higher inherent decay rate than does the corresponding p-0x0 dimer at that pH. At pH < 11, the presence of colloidal RuO;2H20 also has an effect upon the rate of reduction of MnIVTMPyP to MnIIITMPyP, as shown by fig. 9(c). The data given refer to pH 8.2, but similar plots were found at pH 8.8, 9.6 and 10.6. Even allowing for the large experimental uncertainty involved in such measurements, the relationship between kred and concentration of catalyst is not linear. In fact, addition of low concentrations of catalyst appear to slow down the rate of reduction, whilst further additions of RuO, 2H20 lead to an increase in kred until a plateau is reached.In each case, the presence of ca. 2 x lo-* mol dm -3 catalyst seems to give the maximum kred, although the final value is not as high as found in the complete absence of catalyst. (Unfortunately, the concentration of colloid is limited to ca. 3 x lo-* mol dm-3 because of light absorption and precipitation.) As for pH > 11, the rate of oxidation of MnlIITMPyP becomes slower in the presence of catalyst, suggesting adsorption of porphyrin onto the colloid surface. However, the magnitude of the absorbance at 430 nm observed immediately after mixing does not change systematically with increasing concentration of catalyst. This infers that the oxidation product is monomeric MnTVT- MPyP with and without catalyst.At pH < 11, adsorbed MnIVTMPyP seems to be reduced at a slower rate than the Fig. 9. (a) Typical decay traces observed at 430nm following stopped-flow oxidation of MnII'TMPyP at pH 11.3 in the presence of colloidal RuO;2H20. Concentration of colloid; (i) 0, (ii) 1.0, (iii) 1.6 and (iv) 2.3 x mol drnp3. (b) Plot of pseudo-first-order rate constant for decay of MnIVTMPyP as a function of the concentration of added RuO;2H20. (c) Plot of the pseudo-first-order rate constant for decay of MnIVTMPyP at pH 8.2 as a function of the concentration of added RuO, - 2H20.3230 Decay of Manganese Porphyrins in Solution non-adsorbed form, possibly reflecting a change in mechanism. Under the conditions of the stopped-flow experiments, the concentration of MnP was 4.5 x mol dm-3.The plateau region observed at ca. 2 x mol dm-3 RuO, * 2H,O correspondsll to a colloid particle concentration of ca. 5.4 x lo-' mol dm-3. Of the 3700 molecules of RuO, * 2H,O that comprise a single particle, ca. 20% are located on the surface. Thus, the onset of the plateau region corresponds to some 10% occupancy of the surface. Dilution of the surface-bound metalloporphyrin does not affect kred. Increasing the concentration of surface-bound MnP by lowering the concentration of colloid leads to a decrease in kred. Lowering the concentration of colloid even further results in only partial binding of the MnP. At high ratios of MnP to colloid, the porphyrin will be held weakly at sites more remote from the colloid surface since all the preferred sites will be filled.Under such conditions, there is a distinct decrease in kred. This latter finding implies that adsorbed MnIVTMPyP decays, at least partially, by abstracting an electron from the colloid particle. The measured redox potentiaP shows that water oxidation is thermodynamically possible at pH c 11 (fig. 2), and studies performed with the m.p.d. showed that small amounts of 0, were formed in the presence of RuO, * 2H,O. Oxidation of MnIIITMPyP at pH 8-13 was achieved by addition of NaOCl solution (fig. 2). The final concentration of NaOCl was limited so as to give 95% conversion of MnIIIP into MnIVP. This involved stoichiometric addition at pH > 11 but a sixfold excess at pH 8.Monitoring the reaction with the m.p.d. showed that no 0, was evolved during the inherent reduction of MPTMPyP. This finding confirms the studies made with bromate as oxidant. However, addition of colloidal RuO, 2H,O (2 x mol dm-3) prior to oxidation with NaOCl resulted in formation of small amounts of 0, throughout the range 8 < pH < 11. For an initial MnIVTMPyP concentration of 5 x mol dm-3, concentrations of evolved 0, of 0.9, 1.1, 1.2 and 0.6 x mol dm-3, respectively, were obtained at pH 10.6, 9.4, 8.8, and 8.2. Thus the optimum pH seems to be around 9, where the yield of evolved 0, is ca. 10% of the stoichiometric amount. No 0, was detected at pH > 11 or in experiments conducted at pH 8-1 1 if either porphyrin, catalyst or NaOCl was omitted. This work has provided the first example of water oxidation by a MnP, although it appears that an active redox catalyst is necessary for 0, evolution to be observed. Thus, the MnP itself is not able to mediate water oxidation despite its wide range of available oxidation states.In the absence of the added catalyst, water is probably oxidised to H,O,, which attacks the porphyrin. Such findings suggest that the active 0,-evolving catalyst in green-plant photosynthesis is not a MnP, although it could be a manganese containing enzyme. Furthermore, the poor efficiency for 0, liberation together with the narrow pH range over which 0, formation occurs seem to preclude the use of a MnP as a relay in water-splitting systems. As a consequence, it appears that the next step in the design of an effective 0,-evolving photosystem is to find ways to stabilise strongly oxidising metalloporphyrin radical cations.We thank Dr R. E. Huie for the use of his stopped-flow instrument and P. Ouellette for technical assistance with the stopped-flow experiments. This work was supported by the S.E.R.C., G.E. (Schenetady) and the Office of Basic Energy Sciences of the US. Department of Energy through contracts to N.B.S. and N.D.R.L. Certain commercial equipment, instruments, or materials are identified in this paper in order to specify adequately the experimental procedures. Such identification does not imply recommendation or endorsement by the National Bureau of Standards, nor does it imply that the material or equipment are necessarily the best available for the purpose.A .Harriman et al. 3231 References 1 M. Calvin, Science, 1974, 184, 375. 2 N. Carnieri, A. Harriman and G. Porter, J. Chem. Soc., Dalton Trans., 1982, 931. 3 N. Carnieri, A. Harriman, G. Porter and K. Kalyanasundaram, J. Chem. SOC., Dalton Trans., 1982, 4 K. M. Morehouse and P. Neta, J. Phys. Chem., 1984,88, 1575. 5 A. Harriman, G. Porter and I. A. Duncan, Photosynthetic Oxygen Evolution, ed. H. Metzner (Academic 6 A. Harriman and G. Porter, J. Chem. SOC., Faraday Trans. 2, 1979, 75, 1543. 7 P. A. Loach and M. Calvin, Biochemistry, 1963, 2, 361. 8 I. Tabushi and S. Kojo, Tetrahedron Lett., 1974, 1577. 9 P. A. Loach and M. Calvin, Biochim. Biophys. Acta, 1964, 79, 379. 1231. Press, London, 1978), p. 393. 10 A. Harriman and G. Porter, J. Chem. SOC., Faraday Trans. 2, 1979, 75, 1532. 1 1 P. A. Christensen, A. Harriman, G. Porter and P. Neta, J. Chem. Soc., Faraday Trans. 2, 1984, 80, 12 L. K. Patterson and J. Lilie, Int. J. Radiat. Chem., 1974, 6, 129. 13 R. H. Schuler, L. K. Patterson and E. Janata, J. Phys. Chem., 1980, 84, 2088. 14 A. Mills, A. Harriman and G. Porter, Anal. Chem., 1981, 53, 1254. 15 A. Harriman, J. Chem. Soc., Dalton Trans., 1984, 141. 16 A. Harriman, M-C. Richoux and P. Neta, J. Phys. Chem., 1983,87, 4957. 17 R. H. Felton, The Porphyrins, ed. D. Dolphin (Academic Press, New York, 1978), vol. 5, chap. 3. 18 M. Gouterman, The Porphyrins, ed. D. Dolphin (Academic Press, New York, 1978), vol. 3, chap. 1. 19 A. Harriman, P. Neta and M-C. Richoux, Homogenous and Heterogenous Photocatalysis, ed. E. 20 A. Harriman, P. Neta and M-C. Richoux, J. Phys. Chem., submitted for publication. 21 D. Dolphin, R. H. Felton, D. C. Borg and J. Fajer, J. Am. Chem. SOC., 1970, 92, 743. 22 J. A. S. Cavaleiro, B. Evans and K. Smith, Porphyrin Chemistry Adoances, ed. F. R. Longo (Ann Arbor 23 R. S. Srivastava and E. B. Fleisher, J. Am. Chem. SOC., 1970,92, 5518. 24 J. T. Groves and W. J. Druper, J. Am. Chem. SOC., 1979, 101, 7613. 25 D. H. Chin, A. L. Baulch and G. N. La Mar, J. Am. Chem. SOC., 1980, 102, 1446. 26 J. W. Buchler, in The Porphyrins, ed. D. Dolphin (Academic Press, New York, 1978), vol. 1, chap. 10. 27 0. Bortolini and B. Meunier, J. Chem. Soc., Chem. Commun., 1983, 1364. 28 C. A. Reed, Adv. Chem. Ser., 1982,201, 333. 29 P. Hambright, Inorg. Nucl. Chem. Lett., 1977, 13, 403. 30 R. L. Fulton and M. Gouterman, J. Chem. Phys., 1964, 41, 2280. 1451. Pelizzetti and N. Serpone, NATO AS1 Ser. C. (D. Reidel, Amsterdam, 1986), vol. 174, p. 123. Science, Michigan, 1979), p. 335. Paper 61116; Received 15th January, 1986

ISSN:0300-9599    

DOI:10.1039/F19868203215

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. I , 1986, 82, 3233-3244 A Non-equilibrium Configuration Theory of Polyelectrolyte Adsorption William Barford,* Robin C. Ball and Christopher M. M. Nex Cavendish Laboratory, Madingley Road, Cambridge CB3 OHE The adsorption of polyelectrolytes is poorly understood, with experiments suggesting irreversible behaviour beyond the scope of equilibrium theories. We propose a new sequential adsorption theory which has quantitatively different predictions from the equilibrium case. In particular, the predicted threshold value of the surface concentration at which no more polymer can adsorb is much smaller and thus there are more likely to be extended states of adsorbed polymer. 1. Introduction Recent experiments on polyelectrolyte adsorption have cast doubts on the hitherto accepted mechanisms of polyelectrolyte adsorption.Klein and Luckhaml have measured the force profile between two surfaces covered with both neutral and charged adsorbed polymer ; their results show significantly different behaviour for the two cases. The experiments on neutral polymers1 are relatively easy to explain and there is general agreement between theory and experiment. Essentially, in this case the polymer molecules are allowed to adsorb freely onto the surface by diffusion so that there is full coverage after an ‘incubation time’ of a few hours. When the two surfaces are brought together a force profile is measured as a function of distance between the surfaces. If the surfaces are subsequently withdrawn and the force measurement immediately repeated then the range of the force observed is considerably reduced. However, if a few hours elapse before the measurement is repeated then the initial force profile is observed ; thus there is no hysteresis for a sufficiently slow cycle of measurement.The interpretation seems relatively clear : while the polymers are allowed to adsorb onto the surface during the incubation time an equilibrium profile is set up. When the surfaces are brought together the polymer is squeezed flat onto the surfaces and a few hours are needed for the polymer to relax to its equilibrium profile. There has been much theoretical work on calculating the equilibrium profiles of neutral polymers : there are the mean field calculations of Jones and Richmond and the scaling arguments of de Gennes.2$ The experiments on charged polymers4 are much less easy to explain, however.In these experiments the polyions are once again allowed to diffuse onto the surface to achieve full surface coverage in an incubation time of ca. 16 h. When the surfaces are initially brought together a force profile is observed which is extended approximately 600 A from the surface. However, when the measurement is repeated, after withdrawing the surfaces and waiting for periods up to at least a day, no force is observed from the polymer. Since desorption has not taken place the polymer has been squeezed irreversibly, on the time scale of a day, onto the surface; this is shown in fig. 1. If we assume that the final compressed profile is the equilibrium profile then the initial extended profile cannot be in equilibrium; this is the view favoured by Luckham and Kleir~.~ If, on the other hand, we do not assume that the final profile is the equilibrium profile we would deduce that the time scales over which equilibrium is reached are very long, too long for the initial profile to be in equilibrium. 32333234 lo4 - I lo3 5 5 n Lx lo2 lo', Theory of Polyelectrolyte Adsorption 1 I 40 80 120 Dlnm - 10 2c Dlnm Fig.1. Normalised force (FIR) plotted against distance (D) profile between mica surfaces with adsorbed polyelectrolyte as measured experimentally by Klein and Luckham. (a) , Measurement on first approach and @, first withdrawal. (b) All subsequent compression and decompression measurements. Reproduced from ref. (4).Hesselink5 and Papenhuijzen et al.6 have developed equilibrium theories of polyelec- trolyte adsorption. An important difference between the adsorption of charged and neutral polymers is that the former have an electrostatic attraction to the surface (the binding energy per segment might be many times kT) and are therefore much more strongly bound to it than the latter, where the binding energy per segment is of the order kT. We would therefore not necessarily expect the mechanisms of adsorption to be the same in both cases. Klein and Luckham have already suggested that in general adsorption processes are irreversible. In particular, during the squeezing of the polymer no desorption takes place and the surface layer of polymer may remain fixed or at least varies only s1owly.l In an attempt to explain these experiments on polyelectrolytes we propose that the initial profile observed is a non-equilibrium one obtained by the sequential adsorption of polyions.This idea will be explained in more detail in section 4. First, in section 2 the model of polyelectrolytes used in this theory will be outlined, and in section 3 the self-consistent field theory of polymer adsorption will be reviewed and applied to pol yelectrolytes. 2. Polyelectrolytes in Salt Theoretical work on the conformation of strongly charged polyelectrolytes, i.e. those in which electrostatic interactions dominate, in salt-free and salt-containing solutions has clarified some of their behaviour.'j In this paper we consider only polyelectrolytes in salt solution.To model the polyelectrolyte it is necessary to consider the effect of electrostatic interactions between segments of a chain, between segments and ions in solution and between segments of different chains. In the model used below it is assumed that the effect of the electrolyte is simply to screen the Coulombic inter- and intra-chain interactions to a Debye-Huckel form assuming polymer concentrations to be sufficiently small that linearising the Boltzmann factor is valid. It is then possible to model the polyion in terms of a parameter, known as the persistence length, first introduced by Odijkg and Skolnicks and Fixman.lo The idea is that the polymer is locally stiffenedW. Barford, R. C. Ball and C. M . M . Nex 3235 owing to Coulombic forces, but at larger length scales these forces are sufficiently screened out so that the polymer behaves ideally.( R 2 ) = Ll Thus the average end to end distance squared is where 1 = li+le. li is the intrinsic persistence length and 1, is the screened electrostatic persistence length. If the polymer is intrinsically fairly stiff and for low electrolyte concentrations: 1, = ~ (e’A)2n2 E 7.13 A (4)2 at 293 K 4nco kT in water, where A is the distance between charges and R is the Debye-Hiickel screening length. Thus the polyion can be modelled as a Krakty-Porod worm-like chain with Kuhn lengths of thickness R and length 1. If the number of rods is sufficiently large then excluded-volume effects become important and a further parameter is introduced, namely that of the excluded-volume of a Kuhn segment: v = 112.3. Equilibrium Adsorption 3.1. Introduction The self-consistent field calculation for neutral polymer adsorption has been given by Jones and Ri~hmond.~ For small polymer concentrations it is known to be deficient;ll however, it offers the simplest, non-trivial approach to polymer problems. There have been many papers and reviews on this subject, so we will not go into details here. The self-consistent field equation, due to Edwards,12 is for a polymer of length L : i3G 1 V -- = - V,Z G ---p”(r) G i3L 6 212 where p”(r)/L is the polymer probability density at r such that j T / ( r ) d3r = 1. For adsorption onto the yz plane the problem becomes one-dimensional, and by expanding G(r, r’) in terms of the eigenfunction expansion, letting L + 00, and keeping only the lowest eigenvalue : where V(x) is the self-consistent potential at x and is given by V(x) = ( v / 2 P ) Cyyx). Also P(x) is normalised lompdx = 1 so that C is equal to the total length of polymer per unit area.Theory of Polyelectrolyte Adsorption convenient to introduce dimensionless parameters : let x -+ x / l and then the equation becomes in terms of the dimensionless parameters E = vC/Z2 and a2 = - 2E, 1. To model the adsorption we use the de Gennes boundary condition at the surface, which has a self-consistent potential term to take account of polymer density in the surface layer :13 J- Ax) dx 2=0 = - [KO - y&f”(o)] = - K,ff(&). The essential difference between the adsorption of charged and neutral polymers is that for the latter K,, < 1, the so called weak coupling limit, while for the former K , is the order of or greater than unity.A derivation of K , , based on electrostatic attraction, is shown in Appendix 1. For the simplest case of a polymer without excluded volume whose decay length is given by $co it can be seen that the charged polymer will be much more tightly bound to the surface. y is the thickness of the surface layer. In the following it has been put equal to 1 (i.e. to I in real lengths), as this is consistent with the continuum model used. It can, however, be kept in the formalism if a further renormalisation of length scales is performed. 3.2. A Solution without Bulk Excluded Volume It is useful to solve eqn (1) with the boundary condition, eqn (2), for no bulk excluded volume, since only this limit can be solved analytically for the problem of sequential adsorption below.Physically this implies that the polymer density at the surface screens the surface attraction, but there is no self-consistent potential in the bulk. A solution which -+ 0 as x --+ GO isf(x) = (2a)i exp (-ax) with a = K ~ / ( 1 + 2 ~ ) . (3) If a = 0 then no more adsorption can take place and the threshold adsorption concentration has been reached. For this case notice that the theory predicts that saturation never occurs and is thus unrealistic. 3.3. A Solution with Bulk Excluded Volume If the bulk excluded volume is kept in eqn (1) then the solution is: with and (2a2/~); = sinh a(x + x,) KO - &/2(& + 1) 1 +2& a = xo=Lln(~+:). 2a The details are shown in Appendix 2.In this case saturation does take place when eth = [( 1 + 8rco)t - 1]/2 and a -+ 0 more quickly because of the bulk excluded volume. For x Q x, Ax) has a power law behaviour: (2a2/&); exp (-ax,) l+ax ( 7 4 Ax) =W. Barford, R. C . Ball and C. M . M . Nex while for x 9 x, it has an exponential form: 3237 Thus for distances > x, the polymer density, Cf2(x, E ) , falls away very rapidly, so that x, defines an interaction range. For a given E x, and 1 / a are decreasing functions of K , , and for K , > 1 are rather small for E < 1; this means that for small surface concentrations the polymer is confined to a thin layer. At the threshold concentration a power law dominates: (2/&)+ + q as a - 0 x, -P 2/&* Only as the threshold concentration is reached does the profile extend as the exponential tail flattens off.4. Sequential Adsorption 4.1. Introduction We know from the above analysis, and from experiments on adsorbed polyele~trolytes,~~ that for small surface concentrations the polymers are tightly bound to the surface and have a narrow profile. Only at large surface concentrations does the equilibrium analysis predict that the excluded-volume effects will conspire to make the profile more extended. We are thus led to a sequential model of polyelectrolyte adsorption to explain the profile set up as the polyions diffuse onto the surface. Suppose that a polyion diffuses and sticks onto a bare surface; since the attraction is strong it will assume a narrow equilibrium profile. This molecule will screen the attractive potential of the surface (reduce Keff) and will set up a repulsive, but screened, potential owing to its charged segments. We now assume that this profile remains fixed in space.Now suppose that another polyion diffuses onto the surface. As the first polyion is very strongly bound the second cannot displace it in any way, and so it binds less tightly assuming an equilibrium profile due to the now fixed profile of the first polymer. This process now carries on: each new polyion adsorbing less strongly because of the increased screening of the surface attraction and because of the repulsive potential in the bulk due to the fixed profile of the already adsorbed polymer. As the eigenvalue, a:, of the eigenfunction,f,, of the nth adsorbed polymer decreases the polymers will extend further away from the surface until no more adsorption can take place.This can be set up mathematically as follows [cf. eqn (1)]: where BE = BCu/P, d&Z:-y&2 is the potential due to the already adsorbed polymer and 6C is the fraction of each new piece of adsorbed polymer such that nBC is the total amount of adsorbed polymer. Also f,"(x)dx = 1 s with the boundary condition -Keff = -( Kg -6& 5 ~ ' ( 0 ) ) . i=l3238 In the where Theory of Polyelectrolyte Adsorption continuum limit the equations become -Ax, E ) I r ( x , E’) de’ = a2(E)f(x, E ) Kf”(x, E‘) dd = V(x, E ) is the self-consistent potential term and Note that this is the simplest way of considering sequential adsorption, but it is at least fairly tractable.It would be more realistic to fix only the segments stuck to the surface and to keep the loop distribution the same while allowing the shape of the loops to change. 4.2. Without Bulk Excluded Volume This is the only case which we have solved analytically for this model. As in section (3.2) set V(X,E) = 0 in eqn (8). Then the solution is J ~ X ) = (2a)i exp (-ax) applying the boundary condition (9) -a = - K~ + 2 Jlf a(&’) de’. Therefore -da(E)/dc = 2a and lna = -2e. Thus %q(4 = KO exp (- 2&) (10) since %q(O) = KO. When this result is compared with the analogous result in section 3.2, eqn (3), it can be seen that for large e the deviation between %q(~) and aeq(E) is large and that asq decays much more quickly, although once again saturation never occurs. 4.3.Perturbation Calculation In general, of course, it is necessary to include the repelling potential from the loops of the adsorbed layer; this means keeping V(x, E ) in eqn (8). Unfortunately no analytical solution exists for this general problem; however, insight can be gained by performing an expansion in E offle) and or2(&) for small E . A second-order expansion is required. Setting f T E ) = f o + E f i + E 2 f i + . . . a2(e) = @ +&a: + c2ai + . . . substituting into eqn (8) and (9) and equating coefficients of E the following expansion is obtained: gq(&) = K i - E K O ( 4 K , + 1)+$2(12Kg+K0/12+ 1/4)+ .... This should be compared to the analytical result: agq(E) = ICE - E K ~ ( ~ K , + 1) + E ~ ( 1214 + K ~ / 12 + 1 /4) + .. . . atq(&) - a&(&) = $ E ~ ( 1 2 ~ ; + I C ~ / 12 + 1 /4). So, the difference between the eigenvalues at e isW. Barford, R . C. Ball and C. M . M . Nex 3239 Hence, we can deduce that, since this difference is an increasing function of K , , then the physical differences between sequential and equilibrium adsorption will be more pronounced for polymer systems with a large coupling constant, IC,. The result, a&(€), will also be useful for comparing to the numerical calculation at small concentration given below. 4.4. Numerical Calculations To understand the behaviour for general E and to calculate the sequential profile j:p(x, E’) ds’ we turn to numerical calculations. For a general potential V(x,s) the differential equation (8) is solved self-consistently for As) and a2(&).The solution is incremented in E , given the solution for E = 0. The details are shown in Appendix 3. 4.5. Results The numerical calculations have shown the following results. (i) For a general K , aS9(c) tends to zero more quickly than meq(e) as E is increased; however, this becomes more noticeable for larger K , , and in particular for K , > 1. This is shown for K , = 0.1 and 5.0 in fig. 2. Hence saturation occurs earlier in the sequential case, and for large IC, this effect becomes quite marked. Indeed, (&&h becomes almost independent of IC,. This is shown quite clearly in fig. 3. (ii) For a general E the density of sequentially adsorbed polymer is larger at the surface and at distances greater than a certain critical distance than for the equilibrium density profile.The reason that the polymer density at the surface is larger is because there is a small weighting for states with a large eigenvalue. Similarly there is a larger density away from the surface because there are states with smaller eigenvalues than the equilibrium case. It must be admitted, however, that the difference between the profiles for sequential and equilibrium adsorption is rather small when E is far removed from (iii) As (ssq)th is reached the states for small %q(~) extend much further from the (&sq)th- E e Fig. 2. A theoretical plot of a us. E for (a) K, = 0.1 and (b) K, = 5.0. When a reaches zero saturation has occurred. (1) Sequential adsorption and (2) equilibrium adsorption.E is related to the adsorbate concentration by E = vC/P.3240 Theory of Polyelectrolyte Adsorption 3 - 3 Fig. 3. The scaled threshold concentration E~~ us. the surface coupling constant IC, for sequential and equilibrium adsorption. The deviation in the values of threshold concentration becomes significant for K,, > 1. (1) Sequential and (2) equilibrium adsorption theories. X X Fig. 4. A log plot of the theoretical scaled density profile for sequential and equilibrium adsorption as a function of scaled distance from the surface when the adsorbate concentration is at the threshold value for sequential adsorption. (a) IC, = 0.1 and (b) K , = 5.0. ( I ) Sequential and (2) equilibrium. surface. In this case the differences between the two profiles is quite significant, since the equilibrium profile is determined by aeq(&), which in general is quite large at ( E , ~ ) ~ ~ .The profiles are shown for rc0 = 0.1 and 5.0 in fig. 4. 4.6. Discussion From the above results we would suggest that in the experiments carried out by Luckham and Klein the polyelectrolytes have adsorbed sequentially onto the surface until the threshold adsorption concentration has been reached. As this is reached the final polyions to be adsorbed will extend far from the surface, and it is these which cause the initial onset of the force. What happens after the polymer has been squeezed onto the surface is less clear. One possibility is that the polymer readjusts itself to an equilibrium profile, and indeed we know that if the surface concentration remains fixed then the equilibrium profile is much narrower. However, presumably there is more polymer in the solution which can onceW. Barford, R.C. Ball and C. M. M. Nex 3241 again adsorb sequentially given a repulsive potential due to the adsorbed polymer. If this did happen then extended states would again be observed on the next squeeze. Another possibility is that the polymer is squeezed irreversibly onto the surface so that no more polymer can adsorb because either the surface layer is saturated or the dense adsorbate presents a barrier to prevent further adsorption. The results of Luckham and Klein4 suggests that during the cycle of force measurements no more polymer does adsorb. This evidence strongly favours the second possibility.5. Conclusions We have presented a new theory of non-equilibrium adsorption for polyelectrolytes. That an equilibrium theory of adsorption for charged polymers is inadequate is shown by experiments which cannot be explained entirely, if at all, by such a theory. We therefore believe that a non-equilibrium theory is necessary. Although the model used above is rather simple it predicts quite different behaviour, especially for large K , , and in particular a much smaller threshold surface concentration. However, it could be refined, and these refinements might cast light onto the problem of why more polymer does not seem to adsorb after the adsorbed polymer has been squeezed onto the surface. Appendix 1. Derivation of the Coupling Constant, K , Following Edwards,I2 the probability weight for an ideal chain configuration r(s) in a potential V is G[r(s)] Sr(s) = 9[r(s)] exp -- i 2 ( s ) ds - y[r(s)] ds ( :s,” r ) where V = W/RT, W = 40, q5 is the surface potential and o is the linear charge density.W is therefore the potential energy per unit length. For the one-dimensional problem of adsorption onto an infinite surface in the y z plane In terms of eigenfunction expansions I q - V(x) f*(x) = - E,f(x) 2 dx where E, is the smallest eigenvalue, and is negative for bound states. the surface is at x = - y : We will model the attractive potential of range y as a 6 function of strength Vat x = 0; 1 d2f 2 dx2 --+Ef = -yVS(x)J: Integrate about x = 0: 1 2 - [Vf( + S) - vf( - S)] = - y Vf(0). For a sufficiently smoothf3242 then and Theory of Polyelectrolyte Adsorption In this paper we have set y = 1, so The critical potential for adsorption, yC, is This is shown to be a desirable result when entropic considerations are taken into account, since the critical potential energy per chain is L K N , the number of segments on the chain.Let us choose some numbers: A = 20 A, 0 = 0.1 C/A and 4 = 20 mV, then 1 = 30 A and K , in normalised units, i.e. K , -+ k0 = 2. Appendix 2. Reassessment of Equilibrium Adsorption with Bulk Excluded Volume Solve : dzf(x> = Ebfj(x) + ay(x) dx2 subject to the boundary conditions: and normalisation: df -(x)+O a s x - + co dx q = -[Ko-E,f"O)] f d x x=o /om~(x) dx = 1. (9 (ii) (iii) Here we have put subscripts on the excluded-volume parameters so that &b and E, signify excluded volume in the bulk and surface layer, respectively.The solution obeying eqn (i) is = sinh2 a(x + x,) Eqn (iii) gives cotanh(ax,)- 1 = ~ , / 2 a 2cc Eqn (ii) gives a cotanh (ax,) = K , - ~,f2(0).W. Barford, R. C. Ball and C. M . M. Nex 3243 Therefore eqn (iv) and (v) give a = K , -c,[fL(O) +3. Also, = BEb( 1 + 4a/Eb). K , - q E , + 1) 1+2E, Therefore the final result is e 2 a = Appendix 3. The Computer Algorithm We will outline the algorithm used for calculating the eigenfunctions and eigenvalues of the differential equation : - f i x , E ) j:’(x, E’) dE’ = a2(E)f(x, E ) with the boundary conditions f’ f’=f-+O as x+co Iompdx = 1. and the normalisation condition Perform the transformation t = x/(x+ 1) x = t / ( l - t ) with 0 c x < m and 0 c t c 1 ; then (1 -t4))j;(t)-2(l - t 3 ) j - a 2 y - y y2(E’)dE’ = 0 s,‘ In view of the normalisation condition make the further substitution: P = y / ( l - t ) = f/(l - t ) then again (1 - t)4 F-- 4( 1 - t)3 P+ (1 - t)2[2 - T(t)] = a2F 107 FAR 13244 Theory of Polyelectrolyte Adsorption The boundary conditions become F(1) = F(1) = 0 F(O)+[a(O)- V(0)- 13 F(0) = 0 with V(0) = F2(0, E ’ ) dd.Joe For a given V(X, E ) , this differential equation can be solved numerically by calculating the coefficients of the differential operators in terms of Chebyshev polynomials, integrating twice and solving the resulting matrix equation. For details see Nex.15 Given a general F(t,E) we can calculate F(~,E+&) by iteration since as V(t, E ) = F2 dE’ J: it is seen that since l q t , E + d E ) = v(t,E)+S&Fyt,E)+ ...K(t, E + S E ) = P(t, E ) + SEC? V/2& + . . . to first order, where the subscript refers to the number of iterations. Having found a first approximation to a2 and F calculate a new one, i.e. in general calculate the ith iterate using the trapezoidal approximation for the integral j: F2 dc’ so that &(t, E + 6 E ) = V(t, E ) + (SE/2) [F2(t, E ) + Ft-&, E + de)]. This calculation is conveniently carried out as the functional dependence on t is represented in the Chebyshev series, so that operations may be carried out term by term on the coefficients. When self-consistency has been achieved to the desired accuracy e can be incremented and the solution advanced as a function of E . We increment in E given the solution at E = 0: (1 - t ) ~ ( t , 0) = f ( x , 0) = [2a(0)]4 exp [ - a(0) XI, a ( ~ ) = K,,. Accuracy was checked by varying the number of terms in the Chebyshev expansion of the functions, and by repeating the calculation with different increments in E . W. B. is grateful for a CASE Studentship from the S.E.R.C. and Unilever Research Ltd. References 1 J. Klein and P. F. Luckham, Macromolecules, 1984, 17, 1041. 2 I. S. Jones and P. Richmond, J. Chem. SOC., Faraday Trans. 2, 1977, 73, 1062. 3 P. G. de Gennes, Macromolecules, 1981, 14, 1637. 4 P. F. Luckham and J. Klein, J. Chem. Soc., Faraday Trans. I , 1984, 80, 865. 5 F. T. Hesselink, J. Colloid Interface Sci., 1977, 60, 488. 6 J. Papenhuijzen, H. A. Van der Schee and G. J. Fleer, J. Colloid Interface Sci., 1985, 104, 540. 7 P. G. de Gennes, P. Pincus and R. M. Velasco, J. Phys. (Paris), 1976, 37, 1461. 8 T. Odijk, Macromolecules, 1979, 12, 688. 9 T. Odijk, J. Polym. Sci., Polym. Phys. Ed., 1977, 15,477. 10 J. Skolnicks and M. Fixman, Macromolecules, 1977, 10, 944. 11 J. Des Cloizeaux, J. Phys. (Paris), 1970, 31, 715. 12 S. F. Edwards, Proc. Phys. Sac., 1965, 85, 613. 13 P. G. de Gennes, Scaling Concepts in Polymer Physics (Cornell University Press, Ithaca, 1979). 14 M. C. Cafe and I. D. Robb, J . Colloid Interface Sci., 1982, 86, 41 1. 15 C. M. M. Nex, Comput. Phys. Commun, 1980, 20, 1 . Paper 61204; Received 29th January, 1986

ISSN:0300-9599    

DOI:10.1039/F19868203233

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

J . Chem. Soc., Faraday Trans. I , 1986, 82, 3245-3254 Kinetics of Metal Oxide Dissolution Oxidative Dissolution of Chromium(II1) Oxide by Potassium Permanganate Michael G. Segal* and William J. Williams Central Electricity Generating Board, Berkeley Nuclear Laboratories, Berkeley, Gloucestershire GL13 9PB A detailed study of the oxidative dissolution of chromium(Ir1) oxide (Cr,O,) in potassium permanganate is reported. The kinetics of the reaction follow an inverse-cubic rate law under both acid and alkaline conditions; no inhibition by MnO, product is observed. In alkaline solution, in the temperature range 43-140 "C, the rate exhibits dependences on permanga- nate and hydroxide concentrations which are explained in terms of a Langmuirian adsorption mechanism. The rate of dissolution is controlled by electron transfer within a surface complex in which both OH- and MnO; are bound.Thermodynamic parameters imply chemical bond formation between permanganate and ions in the oxide surface. A study in dilute HNO, shows that adsorption of MnO, is also a prerequisite for reaction under acid conditions. The dissolution of metal oxides is of considerable importance in a wide variety of fields. Many oxides are difficult to dissolve in mineral acids or chelating agents, for kinetic rather than thermodynamic reasons. We have shown that oxides containing iron(m), such as Fe20, and NiFe204, can be dissolved very rapidly with reducing agents.l* The kinetics of dissolution of the ' insoluble ' nickel ferrite3 with the reductant tris(pico1in- ato)vanadium(II) have been described in detail.2 Many other one-electron reductants have been used to dissolve iron-based oxides: the rate of dissolution is determined by the rate of electron transfer at the oxidewater interface.This topic has been reviewed re~ently.~ The presence of chromium in mixed metal oxides, such as the spinels of general composition NixFegCr3.x.p04, inhibits the rate of dissolution very markedly. This has been demonstrated both with reducing agents based on vanadium(11)~ and with oxalic and nitrilotriacetic acid.6 Chromium-rich oxide films can be dissolved by oxidising reagents which can remove chromium by an oxidative dissolution process. The most commonly used oxidant is permanganate,'q although others, such as lo persulphatell and peracetic acid,12 have been tested.Up till now, no detailed studies have been made of the kinetics and mechanisms of the oxidative dissolution of metal oxide^.^ In this paper we describe the reaction of chromia (Cr20,) with MnO;, in alkali and acid; the reactions of mixed spinels will be described in a subsequent paper.13 Experimental Materials Solutions of potassium permanganate, sodium hydroxide and nitric acid were prepared from AnalaR reagents, in triply distilled water. Chromium(rI1) oxide was prepared from chromium(II1) nitrate solution by the carbonate method2 using sodium carbonate to precipitate chromium(II1) hydroxide. The precipitate was washed in triply distilled water and calcined at 1000 "C under argon. Sizing of the oxide was achieved by grinding and sieving using Endicott sieves.The particle size chosen for the study was < 53 pm. 3245 107-23246 Kinetics of Metal Oxide Dissolution Apparatus and Procedure The majority of kinetic runs were carried out in glass jacketted reaction vessels; by recirculating water from a thermostatted bath through the jacketted vessel, the desired temperature (40-90 "C) could be obtained. Temperature control was within f 0.5 "C. A measured quantity of Cr,O, (ca. 0.003 g) was transferred to the reaction vessel. The required quantities of water and alkali or acid were added. The resulting solution was kept deaerated by a slow bleed of purified argon fed directly into the solution via thin PTFE tubing. The solution was magnetically stirred. Access to the reaction vessel was through hypodermic needles piercing a serum cap.The reaction at 140 "C was studied using apparatus described previously.6 To start the dissolution run, potassium permanganate was injected from a glass syringe so the final volume of the reaction mixture was 100 cm3. Samples (ca.1.5 cm3) were withdrawn at appropriate times to determine the chromium dissolution. The reaction was quenched in ice, and 1 cm3 of solution was transferred to a 10 cm3 graduated flask ; potassium permanganate and manganese dioxide were destroyed with a few drops of hydrogen peroxide-nitric acid solution. The resulting colourless solution was made up to the mark using 0.1 rnol dm+ HC1 (B.D.H. Convol). The solutions were filtered through a 0.22 pm Millipore Millex GS filter to remove any undissolved oxide and sealed to await analysis.The concentration of dissolved chromium was determined on an atomic absorption spectrometer (Baird A5100). At the end of each run, the remaining solution was cooled to room temperature and the pH measured (Radiometer pHm 62 standard pH meter). All reactions were studied under pseudo-first-order conditions, with at least a ten-fold excess of permanganate and acid or base over the oxide to be dissolved. Results Variation of the Amount of Oxide Dissolution with Time The dissolution of chromium(II1) oxide in permanganate solution is consistent with the reaction occurring at the surface of the particles at a rate proportional to the instantaneous surface area, i.e. drnldt = -kA where m is the mass of undissolved oxide and A is the total surface area of the oxide. Assuming that the particles are uniform spheres and expressing the changes in terms of concentrations leads to the equation where C, and C , are the instantaneous and final concentrations of chromium in solution, ra is the initial radius of the particles and p the density.Thus a plot of [ 1 - (C,/C,)]i vs. time, t , yields a straight line with slope kobs (= k/r,,p). Derivation of this equation has been published previously.2 Examples of the function plot are given in Fig. 1. for the dissolution in permanganate, in 0.01 mol dm-3 NaOH and 0.01 mol dm-3 HNO,. Reaction rate constants, kobs, were calculated from the slopes of plots of [ 1 - (C,/C,)]i against time. Values of C , were based on the weight of oxide present initially.In a number of runs where the dissolution was rapid, the reaction was allowed to go to completion, and C , determined experimentally. Agreement with calculated values was within f 10%.M. G. Segal and W. J. Williams 3247 t/min Fig. 1. Inverse-cubic function plots for dissolution of Cr,O, in potassium permanganate solution at ca. 80 "C. 0 , KMnO, 0.03 mol drn-,, NaOH 0.01 mol dmP3; 0, KMnO, 0.04 mol dm-,, HNO, 0.01 rnol dmP3. [KMn04]/10" mol dm-3 Fig. 2. Variation of observed rate constant kobs as a function of potassium permanganate concentration at 81 "C and various concentrations of NaOH: 0 , O . O l ; .,0.02; 0,0.04; m, 0.07; 0, 0.10 mol dmP3. Lines are calculated according to eq. (8) using constants shown in table 2. Reaction under Alkaline conditions Rate constants obtained under a variety of permanganate and hydroxide concentrations [(2.5-80) x mol dm-3, respectively] at 81 "C are shown in fig.2. At all hydroxide concentrations, the reaction becomes less than first-order in manganesefvn) as the concentration is increased, and observed rate approaches a limit with increasing permanganate. It is also clear that the limiting rate rises with increasing [OH-]. Reciprocal pots, l/kobs vs. l/[MnO,], are linear at constant [OH-]: Fig. 3 shows an example, at [OH-] = 0.01 mol dm-3. mol dmA3 and (1-10) x3248 Kinetics of Metal Oxide Dissolution 201 I I I 0 1 2 3 0""' ( l/[KMnO4])/1O2 dm3 mol-' Fig. 3. Reciprocal plot of observed rate constant us. permanganate concentration at 81 "C; [NaOH] = 0.01 mol dmP3. Table 1.Rate constant at 43 and 140 "C 43 0.0 1 0.01 7.1 x 10-5 1.0 x 10-4 43 0.03 0.01 1.1 x 10-4 1.1 x 10-4 140 0.0083 0.01 2.7 x 2.2 x 10-3 a Rate constants calculated according to eqn (S), using mean values of the rate and equilibrium constants k' = 2.4 x lo-, min-l, K3 = 1.2 x dm6 rnol-, and values of AW and A F for K,, all obtained from 62.5 and 81 "C data. Results obtained at three other temperatures are given in table 1 and fig. 4. At 43 and 62.5 "C, and constant [OH-], the rate again reaches a limit with increasing [MnO,]; this is shown in fig.4 for the results obtained at 62.5 "C, with [NaOH] = 0.01 and 0.10 mol dmP3. The observed rate at 43 "C is 20-50 times lower than 81 "C; at 140 "C, however, the rate is not significantly higher, a rate constant of 2.7 x min-l being obtained at [MnO;] = 0.0083 mol dm-3 and [OH-] = 0.01 mol dm-3, compared with a value of 2.1 x min-l at 81 "C, the same hydroxide concentration and [MnOy] = 0.01 mol dm-3.Reaction under Acid Conditions In dilute nitric acid solution the reaction also obeys an inverse cubic rate law, as shown in fig. 1. Reaction is much faster than in alkaline solution. The variation of rate with permanganate concentration, at a constant acid concentration of 0.005 mol dmd3 and at 80 "C, is shown in fig. 5. The rate again approaches a limit with increasing [MnO;]. A reciprocal plot of l/kobs vs. l/[MnO;] is linear, as shown in fig. 6.M. G. Segal and W. J. Williams 3249 8 I I I I I I a n /o- 0 0 I I I I 1 2 3 4 5 6 [KMn04]/10-2 mol dm-3 Fig. 4. Variation of observed rate constant kobs as a function of permanganate concentration at 62.5 "C.0, [NaOH] = 0.01 mol drn-,, @, [NaOH] = 0.10 mol dmP3. Lines are calculated accord- ing to eqn (8), using constants shown in table 2. [KMn04]/10-2 mol dm-3 Fig. 5. Variation of observed rate constant kobs as a function of permanganate concentration at 80 "C; [HNO,] = 0.005 mol dm-3. The line is calculated according to eqn (9), using constants given in the text. Discussion The Oxidative Dissolution Process under Alkaline Conditions Under alkaline conditions, the reaction between permanganate and chromia is governed by the stoichiometric eqn (2): Cr,O, + 2Mn0; + 20H- = 2Cr0;- + 2Mn0, + H,O. Cr,O, + 2Mn0; + H,O = 2HCrO; + 2Mn0, (2) (3) Under acid conditions, the overall reaction is3250 Kinetics of Metal Oxide Dissolution 015 l!O 1!5 2!0 (1/[KMn04])/102 dm3 mol-' Fig.6. Reciprocal plot of observed rate constant us. permanganate concentration at 80 "C; [HNO,] = 0.005 mol dm-3. Reduction of MnO; in acid solution may produce Mn2+, but this would react with excess managese(vI1) to yield the manganese(1v) oxide. The variation of the amount of oxide dissolved with time establishes clearly that the dissolution process involves reaction at the particle surface, under both alkaline and acid conditions. The rate-determining step involves MnO; in each case, and there is no evidence for inhibition of reaction by the solid product, MnO,, to at least 65% dissolution. The Role of MnO, Pick1, has discussed the role of MnO, in the reaction of permanganate with oxide films on metal surfaces.The isoelectric point (or point of zero charge) of MnO, is low, with values such as 2.55 reported;15 under alkaline conditions the product carries a strong negative charge, but in dilute acid it is only very weakly charged, with a sign dependent on the pH. Cr203, on the other hand, has an isoelectric point at pH 6.516, and so is negatively charged under alkaline conditions but positively charged even in dilute acid. It is not surprising, therefore, that the formation or MnO, as a product does not inhibit the reaction between Cr,O, and MnO, at pH values in the range 12-13, since both the solid reactant and the solid product are strongly negatively charged and so tend to repel. In dilute acid (pH 2.5) there is evidence for MnO, deposition inhibiting the reaction with oxide films on certain alloys,13 but in the reaction with pure chromium oxide particles at pH ca.2 this effect does not appear to be significant. The Concentration Dependences The dependences of observed rate on permanganate and hydroxide concentrations are consistent with a mechanism in which both are involved in a pre-equilibrium at the particle surface. Three absorption equilibria must be considered prior to the electron- transfer step, which results in dissolution of chromium(vI), reactions (4)-(7) : K , >Cr, + MnO, + >Cr-MnO,M. G. Segal and W. J. Williams 325 1 Table 2. Rate and equilibrium constants for reaction of Cr,O, with alkaline permanganate at 8 1 and 62.5 "Ca ~ ~~ T/"C KJdrn-, mol-l K,/dm3 mol-' &/lo3 dm6 mo1-2 kJ10-2 min-' 81 0 32k8 1.5f0.2 2.7 f 0.2 62.5 0 200 0.94 2.1 a Equilibrium and rate constants apply to reactions (4)-(7), as defined in the text.K3 >Cr,+OH-+MnO,g,'Cr-OH-MnO, (6) k > Cr-OH-MnO, + dissolution (7) >Cr, is a free chromium active site, >Cr--OH,=Cr-MnO, and >Cr-OH-MnO, are surface complexes. The formulae written for these complexes are not intended to imply detailed molecular structures, but merely represent attachment of the relevant anions at active sites. Kl, Kz and K3 and are equilibrium constants for this process (K3 for the stepwise addition of OH- and MnO, is unknown order), and k is the rate constant for the reaction of >Cr-OH-MnO,, which ultimately produces dissolved chromium species. Since the rate tends to zero at zero permanganate or hydroxide, we can neglect direct reaction of either >Cr-OH or >Cr-MnO, to give dissolution.This mechanism leads to the following expression for the rate of reaction: k'K,[OH-][MnO,] (8) rate = 1 + K,[OH-]+K,[MnO~]+K,[OH-][MnO~] where k' = k[Cr,] and [Cr,] is the total number of chromium active sites per unit area of powder. The reciprocal of eqn (8) show how a linear plot is obtained from 1 /kobs cs. 1 /[MnO,] at constant [OH-] such as in fig. 3: K2 1 K1 + 1 1 -+ - -- rate k' k'K3[MnOJ K'KJOH-] +k'K3[OH-][MnO;]' (91 All the 35 data points obtained at 81 "C were fitted to eqn (8) using a non-linear least-square regression (STATSYS on the CEGB's mainframe IBM computer). Allowing all four parameters K,, K,, K3 and k' to vary, the regression yielded a small negative value for K,, the binding constant for OH- alone.This was therefore fixed at zero. The values obtained for the other parameters are shown in table 2. The goodness of the fit of eqn (8) to the data at this temperature is shown in fig. 2, where the curves are calculated according to this equation, using the parameters given in table 2. Whilst there is inevitably some scatter, heterogeneous reactions of this kind being intrinsically less reproducible than homogeneous ones, the fit is satisfactory over the whole range of hydroxide and permanganate concentrations. Effect of Temperature Accurate temperature coefficients for the rate and equilibrium constants of the reaction scheme (4)-(7) cannot be calculated from the limited data obtained at other temperatures.However, the data at 62.5 "C are sufficient to yield reasonable estimates for these constants, derived from linear regression of the reciprocal plots, like that shown in fig. 3. The lines plotted in fig. 4 are calculated using these values. From these results it is3252 Kinetics of Metal Oxide Dissolution possible to calculate approximate thermodynamic parameters for reaction (5), the binding of MnO, (i.e. for K2): AH; = - 101 kJ mol-1 AS; = -254 J K-l mol-l. The temperature dependence of the other reactions is small, and the data (particularly at the lower temperature) too imprecise to yield meaningful thermodynamic or activation parameters. The large negative entropy for binding of MnO, at the oxide surface is consistent with the unfavourable electrostatic interaction between the anionic reagent and a strongly negative charge surface.17 The magnitude of A H O , however, is considerably gres :ir than would be expected on electrostatic grounds;17 the equilibrium is clearly driven by the favourable enthalpy, implying chemical bond formation and not simple physical adsorption.The accompanying reaction with OH- to form the reactive complex must have a positive enthalpy of similar magnitude. Although these values of the thermodynamic parameters are no more than estimates, confidence is gained by using them to derive calculated values for the rate constant at the extreme temperatures studied, 43 "C and 140 "C. Assuming that reactions (6) and (7) are essentially independent of temperature, and taking mean values of K3 and k', we have calculated rate constants at the relevant permanganate and hydroxide concentrations at 140 and 43 "C.Table 1 shows the comparision between observed and calculated rate constants at both extremes of temperature. This shows that very satisfactory agreement is obtained over a range of virtually 100 "C. The Mechanism These results imply a mechanism for reaction under alkaline conditions in which both permanganate and hydroxide ions bind reversibly at the particle surface. Neither is sufficient alone to react with chromium(II1) in the oxide and cause it to dissolve: dissolution only occurs at sites where both are bound. Clearly the electron-transfer step between chromium(Ir1) at the oxide surface and bound manganese(vI1) is slow when the surface is exposed to weakly alkaline solution (pH < 12); binding of a hydroxide ion (or deprotonation of bound water) is required to create an environment around the Cr ion more conducive to oxidation. It is important to recognise that a large number of hydroxide ions are bound at the surface at all the concentrations studied here (0.01-4.10 mol dm-3); these results show that whatever the equilibrium state at pH < 12, reaction with pennanganate is slow unless a further hydroxide ion binds also (or a further deprotonation takes place).The value of zero obtained for Kl above implies that the equilibrium constant for this further hydroxide binding, in the absence of bound MnO,, is small compared to that for permanganate. If we assume that reaction occurs predominantly at ' kink ' sites,5 then the chromium(Ir1) ions at the active sites will have several coordination sites available for substitution (probably three).The suggestion that reaction occurs largely at kink sites or other surface defects is supported by the following observation: when a sample of oxide prepared in the same way but calcined at 1400 "C was treated with permanganate and hydroxide, both at 2 x lop2 mol dm-3, at 80 "C, the rate of dissolution was ca. 50 times slower than for the 1000 "C-calcined oxide. The role of hydroxide is almost certainly to deprotonate coordinated water, or to remove a proton from coordinated hydroxide. Either process will lower the Franck-Condon barrier to electron transfer by creating an environment around the chromium ion more similar to that in the product, CrOi-.M .G. Segal and W. J. Williams Reaction under Acid Conditions 3253 The variation of rate with permanganate concentration at constant [H+] is similar to that observed in alkaline solution at constant [OH-]. A simplified form of eqn (8) applies: where k , is the limiting rate at the acid concentration studied, 0.005 mol dm-3. Values of k , and K were calculated from linear regression on the reciprocal plot (fig. 6): k , = 1.54 x 1 0-2 min-l K = 90 dm3 mol-l. Although this system has not been studied in detail, a positive dependence on acid concentration has been observed. Clearly these derived constants are functions of acid concentration, and an equation similar in form to eqn (8) probably applies.Tn the absence of acid, the rate is very much slower. The form of eqn (10) again implies a pre-equilibrium, which we attribute to binding of MnO, at the surface as a requirement before the dissolution reaction takes place. The role of H+ ions in the reaction is not clear: it seems unlikely that protonation of oxide or hydroxide coordinated to Cr"' would increase the rate of oxidation. A more probable explanation is protonation of the MnO; ion; although this ion is only very weakly basic (pK, - 2.25lS), addition of a proton to permanganate bound at the surface may increase the rate of the electron-transfer reaction. This work was performed under contract RPl329-1 with the Electric Power Research Institute, and is published by permission of the Central Electricity Generating Board.Many of the dissolution experiments described here were performed by Messrs S. M. Pimblott and R. Cant, and we also thank Drs T. Swan and R. M. Sellers for their advice and encouragement. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 M. G. Segal and R. M. Sellers, J . Chem. Soc., Chem. Commun., 1980, 991. M. G. Segal and R. M. Sellers, J . Chem. Soc., Faraday Trans.1, 1982, 78, 1 149. D. G. Wickham, Inorg. Synth., 1967, 9, 152. M. G. Segal and R. M. Sellers, Ado. Inorg. Bioinorg. Mech., 1984, 3, 97. M. G. Segal and T. Swan, in Water Chemistry of Nuclear Reactor Systems 3 (British Nuclear Energy Society, London, 1983), p. 187. R. M. Sellers and W. J. Williams, Faraday Discuss. Chem. SOC., 1984, 77, 265. M. E. Pick, in Decontamination of Nuclear Facilities (Canadian Nuclear Association/American Nuclear Society, 1982), session 3, p.5. R. Riess and H. 0. Bertholdt, in Decontamination and Decommissioning of Nuclear Facilities, ed. M. M. Osterhout (Plenum Press, New York, 1980), p. 47. J. Arveson and H. P. Hermansson, in Decontamination of Nuclear Facilities (Canadian Nuclear Associ- ation/American Nuclear Society, 1982), session 3, p. 1. J. Torok, in Decontamination of Nuclear Facilities (Canadian Nuclear Association/American Nuclear society, 1982), session 3, p. 37. P. Spekkens, in Decontamination of Nuclear Facilities (Canadian Nuclear Association/American Nuclear Society, 1982), session 2, p. 21. R. L. Clark, A. B. Johnson Jr and R. A. Shaw, in Decontamination of Nuclear Facilities (Canadian Nuclear Association/American Nuclear Society, 1982), session 3, p. 63. A. B. O'Brien, M. G. Segal and W. J. Williams, J . Chem. Soc., Faraday Trans. I , paper 6/371. M. E. Pick, in Water Chemistry of Nuclear Reactor Systems 3 (British Nuclear Energy Society, London, 1983), p. 61. J. W. Muray, J . Colloid Interface Sci., 1974, 46, 357.3254 Kinetics of Metal Oxide Dissolution 16 D. E. Yates and T. W. Healey, J . Colloid Interface Sci., 1975, 52, 222. 17 R. M. Izatt, D. Eatough, J. J. Christensen and C. H. Bartholomew, J . Chem. SOC. A , 1969, 47. 18 N. Bailey, A. Carrington, K. A. K. Lott and M. C. R. Symons. J. Chem. SOC., 1960, 290, quoted in Stability Constants of Metal-ion Complexes, ed. L. G. Sillen and A. E. Martell (Special Publication No. 17, The Chemical Society, London, 1964). Paper 61217; Received 31st January, 1986

ISSN:0300-9599    

DOI:10.1039/F19868203245

出版商:RSC    

年代:1986

数据来源: RSC