|
1. |
Front cover |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 92,
Issue 1,
1996,
Page 001-002
Preview
|
PDF (1258KB)
|
|
摘要:
THE ROYAL SOCIETY OF CHEMISTRY Journal of the Chemical Society Faraday Transactions Scientific Editor Managing Editor Prof. A. Robert Hillman Dr. Rosemary A. Whitelock Department of Chemistry The Royal Society of Chemistry University of Leicester Thomas Graham House University Road Science Park Leicester LE1 7RH. UK Milton Road Cambridge CB4 4WF. UK Senior Assistant Editor: Mr. D. Bradley Production Editor: Mrs. S. Shah Assistant Production Editors: Dr. J. S. Humphrey, Dr. J. C. Thorn Editorial Secretary: Ms. J. Bladen Faraday Editorial Board Prof. M. N. R. Ashfold (Bristol) (Chairman) Prof. J. A. Beswick (Paris) Prof. A. R. Hillman (Leicester) Dr. D. C. Clary (Cambridge) Prof. J. Holzwarth (Berlin) Dr. L. R. Fisher (Bristol) Dr. D. Langevin (Bordeaux) Prof.B. E. Hayden (Southampton) Prof. S. K. Scott (Leeds) Prof. J. S. Higgins (London) Dr. R. K. Thomas (Oxford) Dr. R. A. Whitelock (RSC, Cambridge) (Secretary) International Advisory Editorial Board R. S. Berry (Chicago) R. H. Ottewill (Bristol) A. M. Bradshaw (Berlin) V. A. Parsegian (Bethesda) A. Carrington (Southampton) R. Parsons (Southampton) G. Cevc (Munich) S. L. Price (London) M. S. Child (Oxford) F. Rondelez (Paris) G. R. Fleming (Chicago) D. K. Russell (Auckland) R. Freeman (Cambridge) P. J. Sarre (Nottingham) H. M. Frey (Reading) J. P. Simons (Oxford) H. H. J. Girault (Lausanne) I.W. M. Smith (Birmingham) H. lnokuchi (Okazaki) S. Stoke (Amsterdam) J. N. lsraelachvili (Santa Barbara) J. Troe (Gottingen) M. L.Klein (Philadelphia) J. Wolfe (Kensington, NSW) A. C. Legon (Exeter) C. Zannoni (Bologna) R. A. Marcus (Pasadena) R. N. Zare (Stanford) B. J. Orr (North Ryde) A. Zecchina (Turin) C. Zhang (Dalian) Journal of the Chemical Society. Faraday Transactions (ISSN 0956-5000) is published twice monthly by The Royal Society of Chemistry, Thomas Graham House, Science Park, Milton Road, Cambridge CB4 4WF, UK. All orders accompanied with payment should be sent directly to The Royal Society of Chemistry, Turpin Distribution Services Ltd., Black- horse Road, Letchworth. Herts. SG6 1HN. UK. NB Turpin Distribution Services Ltd., dis- tributors, is wholly owned by the Royal Society of Chemistry. 1995 Annual subscription rate EEA €822.00, Rest of World €863.00. USA $1 510.00, Canada €863 (excl.GST). Customers should make payments by cheque in sterling payable on a UK clearing bank or in US dollars payable on a US clearing bank. Second class postage is paid at Rahway, NJ. Airfreight and mailing in the USA by Mercury Airfreight International Ltd. Inc., 2323 Randolph Avenue, Avenel, NJ 07001, USA and at additional mailing offices. USA Postmaster: send address changes to Journal of the Chemical Society, Faraday Trans- actions, c/o Mercury Airfreight International Ltd. Inc., 2323 Randolph Avenue, Avenel, NJ 07001. All despatches outside the UK by consolidated Airfreight. PRINTED IN THE UK. Q The Royal Society of Chemistry, 1995. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form, or by any means, electronic, mechanical, photographic, recording, or otherwise, without the prior permission of the publishers.Advertisement sales: tel. +44(0)171-287-3091; fax. +44(0)171-494-1134. INFORMATION FOR AUTHORS The Royal Society of Chemistry welcomes submission of manuscripts intended for pub- lication in two forms, Research papers and Faraday Communications. These should describe original work of high quality in the areas of physical chemistry, biophysical chemistry and chemical physics. Research Papers Full papers contain original scientific work which has not been published previously. However, work which has appeared in print in a short form such as a Faraday Communi- cation is normally acceptable.Four copies including a top copy with figures etc. should be sent to The Editor, Faraday Transactions, at the Editorial Office in Cambridge. Authors may, if they wish, suggest the names (with addresses) of up to three possible referees. Faraday Communications Faraday Communications contain novel scientific work whose interest to the reader- ship and impact on the subject warrants rapid publication. The total length is restricted to four pages of the double-column A4 format. Submission of a Faraday Communication should be made to The Editor, Faraday Transactions, at the Editorial Office in Cam- bridge. Faraday Research Articles Faraday Research Articles are occasional invited articles which are published follow- ing review.They are designed to be topical articles of interest to a wide range of research scientists in the areas of physical chemistry, biophysical chemistry and chemi- cal physics. All authors submitting work for publication are required to sign an exclusive copyright licence. All submissions should be accompa-nied by a completed form (a blank for photocopying is reproduced at the end of the Information for Authors in Issue 1). without which publication cannot proceed. Full details of the form of manuscripts for Articles and Faraday Communications, con- ditions for acceptance etc. are given in issue number one of Faraday Transactions, published in January of each year, or may be obtained from the Editorial Office. Instruc- tions for authors can also be accessed on the World Wide Web at the URL http://chemistry.rsc.org/rsc/fapifa.htm There is no page charge for papers published in Faraday Transactions. Fifty reprints are supplied free of charge. Prof. A. R. Hillman, Scientific Editor. Tel. : Leicester (01 16) 2525226 (24 hours) E-Mail (JANET):ARH7@UK.AC.LElCESTER Fax: (0116) 2525227 Dr. R. A. Whitelock, Managing Editor. Tel.: Cambridge (01223) 420066 E-Mail (INTERNET): FARADAY@RSC.ORG Fax: (01 223) 420247
ISSN:0956-5000
DOI:10.1039/FT99692FX001
出版商:RSC
年代:1996
数据来源: RSC
|
2. |
Back cover |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 92,
Issue 1,
1996,
Page 003-004
Preview
|
PDF (2378KB)
|
|
ISSN:0956-5000
DOI:10.1039/FT99692BX003
出版商:RSC
年代:1996
数据来源: RSC
|
3. |
A method to calculate vibrational frequency shifts in heteroclusters: application to N2+–Hen |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 92,
Issue 1,
1996,
Page 11-15
David F. R. Brown,
Preview
|
PDF (659KB)
|
|
摘要:
A method to calculate vibrational frequency shifts in heteroclusters: Application to N2 +-He, David F. R. Brown, Jonathon K. Gregory and David C. Clary* Department of Chemistry, University of Cambridge, Cambridge, UK CB2 1E W A method is proposed whereby vibrational frequency shifts can be calculated in weakly bound heteroclusters and is applied to the ionic N,+-He, clusters. We use diffusion Monte Carlo to simulate the 3n degrees of freedom of the N,+-He, interactions while the N-N vibrational motion is solved by an adiabatic method. In addition, we report minimum energy structures, vibrationally averaged structures and binding energies for N,+-He, clusters for n = 1-12. In the last few years there has been considerable interest, both theoretical and experimental, in weakly bound ion-atom and ion-molecule clusters.Whilst there have been many studies on neutral clusters1-3 using experimental techniques such as microwave, IR, visible and UV spectroscopy, as well as numerous theoretical there has been surprisingly little work on their ionic analogues, despite their chemical importance. Elucidating the structure and properties of ionic clusters is important in determining ion-neutral reaction pathways as many are predicted to proceed via long-lived complexes. Such reactions play a vital r81e in biological and atmospheric systems. In addition, they are useful models for investigating solvent-solute interactions as they are sufficiently small to be probed by high-resolution spectroscopy which may be per- formed size selectively, so it is possible to examine properties experimentally progressing gradually from an isolated ion to a fully solvated ion.Bgcii: and co-workers have performed such a study theoretically on the neutral Ar,-HF system using a 5D quantum bound state calc~lation,~ successfully showing that the vibrational frequency shift has a strong dependence on both geometry and cluster size, and Lewerenz' has pre- dicted the vibrational frequency shifts for C1,-He, clusters. Here we describe a simple but general method for calculating vibrational frequency shifts in molecular clusters. A major problem with studying AB-Rg,, clusters is the lack of three-body potential surfaces and furthermore that many of these potentials have no dependence on the internuclear separation of the chromosphore species. One type of system which is the subject of current spectroscopic and theoretical examination is N2+-He, .8-'1 These clusters have been shown to have the general characteristics usually associated with weakly bound van der Waals clusters and a 3D potential surface for N,+-He has been fitted by Billing and co-workers', to a set of ab initio points due to Berning and Werner.' Bieske et aL8 have measured vibrational predisso- ciation rates in N,+-He, and deduced some structural infor- mation using high-resolution electron spectroscopy. However, despite the work carried out on N,+-He, over the last few years, little information regarding the structure, zero-point energies or vibrational frequency shifts of N2+ for varying n has yet been found.Diffusion Monte Carlo (DMC) has been used extensively and quite successfully in the last few years to solve the many body Schrodinger equation in a variety of different situations. The diffusion simulation was developed by Anderson14 and has been used extensively in electronic structure problems. It was then applied to molecular vibrations by Suhm and Watts." DMC simulations are particularly well suited to the investigation of weakly bound clusters ; all anharmonicities, couplings and degrees of freedom, which in the case of weakly bound systems can cause problems for basis set calculations, are automatically included in the simulation. In addition, DMC scales vary favourably with system size and is much more straightforward to implement as there is no need for complicated basis sets or large integrals.The large memory requirements of basis set methods are not a problem in DMC. This, coupled with the favourable scaling, allows one to simu- late clusters including all their degrees of freedom far larger than is feasible with more conventional rnethods.l6 In this paper, we report calculations performed using the DMC method on N,+-He, (n = 1-12) clusters and describe a method by which vibrational frequency shifts may be calcu- lated. The paper is set out as follows. In the next section, an outline of the method of calculating vibrational frequency shifts by DMC will be given. The results of the calculations are then presented ; minimum-energy structures, vibrationally averaged structures, binding energies and vibrational shifts are described and discussed.Finally we draw some conclusions and discuss future applications of this method. Theory Diffusion Monte Carlo The DMC algorithm has been described extensively in pre- vious publications' 5*1'-' and therefore only an outline of the method and details specific to this work will be described. The Schrodinger equation for N bodies is written in atomic units and imaginary time, 7 = it where I,$ is the wavefunction, V is the potential energy, x the positional coordinates and rn, the mass. yefis a reference energy introduced to stabilise the ground state. This is analo- gous to a diffusion equation with an additional source/sink term :Wx, t, -DjV2C(X,t) -kC(x, t) at j=l This implies that the Schrodinger equation may be treated as a diffusion process and solved by a random walk.The linear- ity of the Schrodinger equation allows many non-interacting random walkers (replicas) to be used in each simulation. These replicas are defined in terms of Cartesian coordinates, which are changed randomly to simulate the kinetic energy term in eqn. (1) and have a weight which varies to simulate the poten- tial energy. The population is propagated along 7,by using finite time intervals (steps). Consider the formal solution to J. Chem. SOC., Faraday Trans., 1996,92(1), 11-15 11 He N N Fig. 1 Coordinate system used to describe the cluster.All Riand Oi are labelled X in the text. eqn. (1) $(x, T)= C Cn4n ex~C-(En -V,ef)Iz (3) n where $,, and E, represent the eigenfunctions and eigenvalues of the quantum state. As z becomes large, the dominant term in eqn. (3) corresponds to the lowest eigenstate since the higher eigenstates will have decayed to zero. The DMC simu- lation begins with a population of replicas, each having a weight of 1.0, distributed randomly on a potential surface. The variable z is increased by an amount AT at each timestep, after which the population is adjusted; this is done in three stages. The first stage involves a random movement (Ax) of each replica to simulate the kinetic energy in eqn. (1).According to the Einstein equation, Ax is related to the timestep (AT) by Ax2 1D=-=-2A2 2m, (4) where the diffusion coefficient is defined in eqn.(1). Each atom is moved randomly in three dimensions by an amount taken from a Gaussian distribution with a mean of 0 and standard deviation (Az/mk)'/2.The second stage is to update the weight of each replica by continuous weighting,' % = exp[ -( V' -Kef)A~] (5) where V' is the potential energy of replica i. When using con- tinuous weighting, it is found ne~essary'~~~' to remove those replicas whose weight falls below a certain critical value and to duplicate the replica with the highest weight. The total weight is conserved in this process and the duplicate is allowed to evolve independently. This 'repacking' of replicas is the third stage of the simulation, which is done prior to the adjustment of the reference potential, qef.The ground-state energy of the system may then be obtained from the average value of the reference potential (Fef),or the potential energy (V).The wavefunction is equivalent to a histogram of the posi- tions of the replicas with respect to the coordinate in question.Each of these properties is obtained by averaging over a large number of steps in the simulation. All the simulations carried out in this work are unguided; no importance sampling is used, as this would introduce bias, particularly in relation to the number of solvation shells. Since the form of $(X, r) is known, it is possible to calculate expectation values of any non-differential operator p; we employ descendent weighting as described by Kalos2' to do so.Vibrational frequency shift The vibrational frequency shift is given by the difference between E(u = 1)-E(u = 0) for the free N,' ion and E(u = 1) -E(u = 0) for the solvated ion, where u is the vibra- 12 J. Chem. SOC.,Faraday Trans., 1996, Vol. 92 tion quantum number of the N,+ chromophore. This is a value which usually may be obtained quite accurately from experiment and therefore provides a useful probe of the poten- tial surface as well as giving information regarding the solvent-solute interaction. The two main problems which arise from using DMC to calculate vibrational frequency shifts need to be overcome. First, the standard DMC algorithm can rigorously find only ground-state wavefunctions and energies ; we require both the ground (u = 0) and excited vibrational states corresponding to N,+(u = l)He,.The second problem is that the vibrational motion of the N2+ is of a much higher frequency than the motion of the helium atoms. This makes it very difficult to acquire accurate results including all degrees of freedom in the DMC simulation as it is hard to obtain a high degree of accu- racy for low-frequency motion when high-frequency vibra- tions are also included.,, The fact that the formation rate for the N, +-He, clusters is not found experimentally to be affected by the initial vibra- tional state of the N2+ core8 indicates that there is only a very weak coupling between the ion core and He atoms.Similar approaches to the method we are proposing have been per- formed in various basis set calculations on the bound states of weakly bound c~mplexes.~~-~~ The method described below is also closely connected to the method of Quack and Suhm26 that calculates adiabatic potential curves and has been applied to (HF), . Consider the matrix element for the cluster containing n helium atoms : < Y(X,r)I 2I VX,r)) (6) where Y is the total wavefunction, 2the total Hamiltonian, r the N2+ vibration coordinate and X all the other coordinates of the system. Because of the weak coupling described above, Y may be considered approximately as the product of a func- tion dependent solely on r, q"(r),and a function +(X; r) dependent explicitly on all the other system coordinates X and parametrically on r.The Hamiltonian may be factorised similarly : 2= 3" + 3x (7) This gives: and by DMC we obtain 2x &X; r) = E(r)4(X;r) (9) giving <vu(r)I {2NN + ~(r))I qU(r)) (10) Thus the problem is effectively reduced to a one-dimensional oscillator. For the system N,+-He,, the N2+ vibrational coordinate r is frozen at some value and a DMC simulation performed with a potential 5 that has the N,+ potential V"(r) sub-tracted from the full potential V including all the other coor- dinates X. The wavefunction $(X; I) may be obtained directly from this and the energy E(r) may be found by performing more DMC simulations with other values of r and fitting a curve to the energy points acquired.The vibrational states of N,+ may then be found by diagonalising the Hamiltonian 2" + E(r)where This equation is solved numerically using LeRoy's LEVEL pr~gram,~'which employs the Numerov method. Numerical methods To simulate the full potential surface, a pairwise combination (Fig. I) of the 3D N,+-He potential due to Billing and co- workers12 and the He-He interaction potential by Aziz et aL2* were used. No three-body forces were considered as they are small when compared with the other parts of the potential of this system; it would have been possible to add a three- body tripole-dipole term but, in the case of He, this term is negligible. We have carried out energy minimisations for each of the N2+-He clusters (n = 1-16).The minimisations were carried out by selecting at random around loo0 geometries from a DMC simulation for the particular cluster; these DMC points should collectively cover all important regions of the potential-energy surface. We then performed minimisations for all geometries using Powell's method3' with the N-N unit fixed with a bond length of 2.109 a,. The global minima obtained are reported later in this article. A large number of local minima were also found, especially for the larger clusters, which is not surprising since the He-He interactions are rela- tively weak. For each cluster, a total of nine DMC simulations were carried out with r held fixed at a selection of values between 1.3 and 5.0 a,.In order to freeze r in the simulation, it is necessary to use a modified DMC algorithm. The separation of the degrees of freedom is accomplished by treating the N2+ ion core as a rigid body by adding a subroutine after the drift term to move the N atoms along the internuclear axis in such n =1 n =2 8 8 0 n =5 ............ ee n =10*:n =9 .......... -0.-......... ...............---@ ...Q............a .......Q ................ ......... n =13 0 e 8 a way as to keep r fixed but to allow the centre of mass to freely translate and the ion to rotate.31 The elimination of the high-frequency motion allows a much longer timestep to be used than would otherwise be possible.A fifth-order polynomial fit was performed on the resulting points in order to obtain the function E(r) in a form in which it could be used easily to find the vibrational energy levels in LeRoy's LEVEL pr~gram.~'A timestep of A7 = 200 h/Eh was found to give negligible timestep error compared with the statistical uncertainties associated with the DMC algorithm for the clusters (n= 1-8). For clusters with more than eight atoms, we found it necessary to reduce the timestep to 100 h/Eh. All simulations were allowed to equilibrate for 2 x lo6 h/E, and then propagated for a further 3 x lo6 h/Eh to obtain in most cases a statistical uncertainty of less than 1%. A cluster containing two helium atoms took ca. 30 min. of cpu time of a DEC Alpha workstation using the above parameters for each DMC simulation.The time taken rose to ca. 8 h for the n = 12 cluster. Results and Discussion Structure and energy The minimum-energy structures are shown in Fig. 2 and the minimum energies in Fig. 3. The energies show a linear decrease in total minimum energy up to n = 12. With n b 13 a slight departure from linearity is observed, which may be n =3 n =4 0 0 8 0 n =7 n =8 * @ 0.-... .....,;a e 8 ........... .-.e n =11 n =12 e.................@* ........... ........0::: .............. --e-0.......e 0.-....... ........0' ..................e*@&I. 3 .......3 0.0..........e-.'-'Q.-e.............a........ n =14 Fig.2 Minimum-energy structure for n = 1-14. Dotted lines are for clarity only. J. Chern. SOC.,Faraday Trans., 1996, Vol. 92 13 0, I 0 -500 -0 0 0 --1000 0 ? E 0 0-1500 -0F a, 0 --2000 0 O 0 I -3000 I I 0 2 4 6 8 10 12 14 n Fig. 3 Variation of minimum energy with n attributed to the formation of a second solvation shell in which the He atoms are slightly less well bound (see Fig. 3). In general, the clusters show structures in which there is an equal number of helium atoms around each N atom when n is even, and the helium atoms are arranged to give as even a distribu- tion as possible when n is odd. The He atoms do show a ten- dency to group together (see especially n = 8) showing that the He potential is slightly attractive.The n = 10 and n = 12 clusters show pentagonal and hexagonal antiprismatic struc- tures, respectively, and beyond 12 helium atoms, a second sol- vation shell is observed, with the extra He capping the prism. The radial plots for the helium wavefunctions in the vibra- tionally averaged structure are shown in Fig. 4 and the angular plots for the n = 2 and 6 clusters in Fig. 5. They show that the structure is highly delocalised with large amplitude motions, and that the barrier to internal rotation is small, as observed by Bie~ke.~The wavefunctions for the clusters n = 1-7 for the helium atoms are almost identical, indicating that the atoms can comfortably fit round the ion core and are not interacting repulsively with each other.The n = 8 cluster shows a slight broadening of both the radial and angular parts of the wavefunction indicating that although they are not suficient to force the formation of a second solvation shell, He-He repulsive forces are coming into play. The expec- tation values for R do show a slight increase when n > 8; however the uncertainty due to the diffuseness of the He wavefunction means that these figures are not conclusive. By constructing the perturbed wavefunctions for the N, vibra-+ tion, it was possible to find the expectation value for r; this does not change significantly with cluster size. 1 0.8 0.6 w 0.4 02 0 Fig. 4 Wavefunctions for selected clusters as projected onto the Ri coordinate to give the radial part: n = 2 (-), 6 (--), 8 (---) and 12 (* -.) 14 J.Chem. SOC.,Faraday Trans., 1996, Vol. 92 1 08 0.6 w 04 0.2 ~0 0 10 20 30 40 50 60 70 80 90 8/degrees Fig. 5 Wavefunctions for the n = 2 (0)and 6 ( x ) clusters, projected onto the 8,coordinate Table 1 contains values for Do along with the average binding energy per helium atom for each of the clusters. The first six clusters where the energy is almost exclusively due to the N,+-He interaction give an average binding energy of ca. 122.5 cm-' which is in good agreement with the value calcu- lated by Buchenko et aL3, for one helium atom using an approximate basis set method of 123.3 cm-' with the same potential, considering the errors in each calculation and that the fitting parameters may be slightly different. Clusters with more than seven helium atoms show a gradually decreasing average binding energy.Vibrational frequency shifts Table 2 and Fig. 6 show the variation in vibrational frequency shift Av as n varies. It may be seen that the shift varies linearly up to n = 8 and then a departure from linearity is observed for n > 8 with the plot apparently tending towards a horizon- Table 1 Doand binding energy, E,, for N,+He, per helium atom as obtained from the DMC simulations, with r fixed at (r) E, per He atom n /cm -1 -121.9 (2) 121.9 2 -244.3 (2) 122.2 3 -367.1 (2) 122.4 4 -489.9 (3) 122.5 5 -613.0 (10) 122.6 6 -734.7 (10) 122.5 7 -854.1 (12) 122.0 8 -971.4 (10) 121.4 9 -1086.2 (12) 120.7 10 -1189.4 (13) 118.9 11 -1287.1 (20) 117.0 12 -1366.0 (30) 113.8 The statistical uncertainty is given in parentheses.Table 2 Calculated vibrational frequency shifts for N2+in N,+-He, (n = 1-12) vibrational frequency shift n /cm -0.65 (1) 1.32 (1) 2.72 (1) 4.17 (4) 5.59 (7) 6.66 (14) 7.31 (14) Statistical uncertainties are shown in parentheses. a I 0 0 n Fig. 6 Variation of N,+ vibrational frequency shift in N,+-He, with n, calculated using the method outlined in the second section tal asymptote. It is useful to compare these results with other theoretical studies of similar systems and some of the experi- mental work done by Bieske et al.*,’ The 5D quantum bound state calculations performed by BZcik and co-workers on the Ar,-HF system suggest that the vibrational frequency shift Av depends on the number of atoms in the cluster, the geometry of these atoms, and the interaction potential between the atom and the chromophore.Thus the linearity of our results up to n = 8 may be explained by the fact that the He atoms are not interacting significantly (Fig. 5) and are effectively identical as they add to the cluster due to the large amplitude motion. The departure from linearity arises as the heliums are sufficiently crowded to interact, pushing each other away slightly from the N,+ core, hence reducing their ability to perturb it. BZcik and co-workers’ work on the Ar,-HF system’ shows the same general trend as the results presented here, but since that system is much more rigid, the variation is less smooth as the argon atoms take on specific geometries to minimise the Ar-Ar interaction energy.In general, the vibrational shift per atom is significantly greater than that for the N,+-He, system. Lewerenz has performed variational Monte Carlo (VMC) calculations on the more closely related He,-Cl, clus-ter~,~again finding a much more irregular relationship between Av and n, and a much larger vibrational shift. In this case, the difference may be due to the relative magnitudes of the He-He and Heechromophore interactions. Whilst Bieske et al. have not reported the vibrational fre- quency shifts, they have obtained the vibrational predissocia- tion rates which depend on a similar matrix element to Av.They show an increase in vibrational predissociation rate with n, although they have as yet only gone as far as six helium atoms and the errors in the results are sufficiently large as to make the exact form of the curve uncertain. Conclusions A method has been described for calculating the vibrational frequency shifts for weakly bound heteroclusters using diffu- sion Monte Carlo. It has been applied to the N,+-He, clus-ters and structural data has also been acquired. The main feature of the numerical data obtained is that an essentially linear dependence on n up to n = 8 is displayed, indicating little He-He interaction occurs up to this point.The growing significance of the He-He interaction beyond IZ = 8 causes the He wavefunction to become more diffuse about the ion core. Consequently the N,+-He interaction is lessened and a departure from linearity in the results is observed. The method we use is general and may be applied to much more complicated systems than N, +-He,, provided suitable potential surfaces are available. The surface must contain at least a parametric dependence on the intermolecular coordi- nates of the chromophore. It is hoped that the data presented here will be useful in interpreting future experimental work on N, +-He, related systems. We would like to thank Robert LeRoy for supplying the LEVEL 5.2 program and Marius Lewerenz for sending pre- prints of his recent work on heteroclusters. This work was supported by the Engineering and Physical Sciences Research Council.References 1 D. H. Levy, Adv. Chem. Phys., 1981,47,323. 2 D. J. Nesbitt, Chem. Rev., 1988,88, 843. 3 S. Goyal, D. L. Schutt and G. Scoles, Ace. Chem. Res., 1993, 26, 123. 4 S. Y. Liu, Z. Biicik, J. W. Moskowitz and K. E. Schmidt, J. Chem. Phys., 1994,100,7166. 5 S. Y. Liu, Z. BgciC, J. W. Moskowitz and K. E. Schmidt, J. Chem. Phys., 1994, 101, 6359. 6 H. C. Chang, F. M. Tao, W. Klemperer, C. Healey and J. M. Hutson, J. Chem. Phys., 1993,99,9337. 7 M. Lewerenz, J. Chem. Phys., in the press. 8 E. J. Bieske, S. Nizkorodov, A. Friedmann and J. P. Maier, Int. J. Mass Spectrom. Ion Processes, 1994, 135, 19.9 E. J. Bieske, J. Chem. SOC.,Faraday Trans., 1995,91, 1. 10 S. Miller, J. Tennyson, B. Follmeg, P. Rosmus and H-J. Werner, J. Chem. Phys., 1988,89,2178. 11 S. K. Pogrebnya and D. C. Clary, Chem. Phys. Lett., 1994, 219, 366. 12 V. A. Zenevich, W. Lindinger and G. D. Billing, J. Chem. Phys., 1992,97,7257. 13 A. Berning and H-J. Werner, J. Chem. Phys., 1994,100, 1953. 14 J. B. Anderson, J. Chem. Phys., 1975,63, 1499. 15 M. A. Suhm and R. 0.Watts, Phys. Rep., 1991,204,293. 16 R. N. Barnett and K. B. Whaley, Phys. Rev. A, 1993,47,4082. 17 J. B. Anderson, Int. Rev. Phys. Chem., 1995, 14, 85. 18 C. J. Umrigar, M. P. Nightingale and K. J. Runge, J. Chem. Phys., 1993,99,2865. 19 M. Lewerenz and R. 0.Watts, Mol. Phys., 1994, 81, 1075. 20 H. Sun and R. 0.Watts, J. Chem. Phys., 1990,92,603. 21 M. H. Kalos, Phys. Reu. A, 1970,2,250. 22 J. K. Gregory and D. C. Clary, Chem. Phys. Lett., 1994,228, 547. 23 S. C. Althorpe, D. C. Clary and P. R.Bunker, Chem. Phys. Lett., 1991,187,345. 24 S. L. Holmgren, M. Waldman and W. Klemperer, J. Chem. Phys., 1977,67,4414. 25 K. D. Kolenbrander, C. E. Dykstra and J. M. Lisy, J. Chem. Phys., 1988,88,5995. 26 M. Quack and M. A. Suhm, Chem. Phys. Lett., 1991,183,187. 27 R. J. LeRoy, LEVEL 5.2, University of Waterloo Chem. Phys. Research report, CP330R2, 1993. 28 R. A. Aziz, F. R. W. McCourt and C. C. K. Wong, Mol. Phys., 1987,61, 1487. 29 D. J. Margoliash, T. R. Proctor, G. D. Zeiss and W. J. Meath, Mol. Phys., 1978, 35, 747. 30 W. H. Press, S. A. Teukolsky, W. T. Vetterling and B. P. Flan- nery, Numerical Recipes, Cambridge University Press, Cam- bridge, 1992. 31 V. Buch, J. Chem. Phys., 1992,97,726. 32 A. A. Buchenko, A. Y. Baisogolov and N. F. Stepanov, Chem. Phys. Lett., 1994,220,93. Paper 51048876;Received 24th July, 1995 J. Chem. SOC., Faraday Trans., 1996, Vol. 92 15
ISSN:0956-5000
DOI:10.1039/FT9969200011
出版商:RSC
年代:1996
数据来源: RSC
|
4. |
Anthracene singlet quenching by indoles in media of different polarities: mechanism and chemical photoreaction |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 92,
Issue 1,
1996,
Page 17-22
M. V. Encinas,
Preview
|
PDF (704KB)
|
|
摘要:
Anthracene singlet quenching by indoles in media of different polarities : Mechanism and chemical photoreaction M. V. Encinas; C. M. Previtalib and S. Bertolottib a Facultad de Quimica y Biologia, Universidad de Santiago de Chile, Casilla 307-2, Santiago, Chile Departamento de Quimica y Fisica, Universidad Nacional de Rio Cuarto, 5800 Rio Cuarto, Argentina The quenching of excited singlet anthracene by indole derivatives in solvents of different polarity has been studied using static and time-resolved techniques. The quenching process is dependent on the nature of the indole derivative and on the polar properties of the medium. In non-polar solvents photoreaction takes place with unit efficiency in compounds which have a hydrogen atom bound to the nitrogen. The efficiency decreases in polar solvents and is considerably suppressed in aqueous solvent mixtures.The results here obtained indicate that charge transfer is the primary step for the photoreaction in polar solvents. For compounds with the indolic N-H group, direct hydrogen transfer is the dominant mechanism in non-polar media. The importance of radical ion formation in these anthracene-indole systems is investigated as a function of the indole structure and solvent polarity. Singlet quenching of polycyclic aromatic hydrocarbons by ground-state donors in solution has been the subject of active research. Much attention has been paid to the deactivation of the singlet state of pyrene by aromatic and aliphatic amines. It has been well established that the interaction results in the formation of short-lived exciplexes, followed by electron trans- fer from the amine to the arene, giving an ion pair where hydrogen-atom transfer reactions can be produced.'-4 The formation of several kinds of radical ion intermediates has been directly observed using transient absorption and tran- sient photocurrent st~dies.~,~,~ Indoles have not been employed as donors to such an extent, despite their high electron donating ability and the high reactivity of the hydrogen on the N atom. Pyrene fluores- cence is quenched by indoles through a charge-transfer The rate constant of this mechanism depends strongly on the solvent polarity, being close to that found for diffusion-controlled reactions in highly polar media.' Recent-ly, Montejano et a/.' have evaluated the radical anion and the triplet quantum yields produced from the interaction of pyrene with indole. These studies showed that the geminate ion pair decays to the ground state or triplet state with a solvent-independent rate constant.In non-polar solvents, indole derivatives have been shown to form emissive inter- molecular exciplexes with pyrene and cyanopyrene.' '*' ' In order to get more information on the quenching mech- anism by indoles and to evaluate the importance of the chemi- cal photoreaction, we have carried out a study of the interaction between excited singlet anthracene and several indole derivatives. Anthracene is of particular interest owing to its high tendency to undergo photochemical reactions.Pho- toreactive quenching of singlet anthracene has been reported with amine~,"--~~ h ydroxylamines' '*' and secondary piperi- dines.' Experimental Anthracene (Fluka A. G., puriss.) was used without further purification. Indole and its derivatives (Aldrich or Sigma) were used after recrystallization or sublimation. All solvents used were of spectroscopic grade. Fluorescence-quenching experiments were carried out by measurements of the fluorescence intensity and/or fluores- cence lifetime. Steady-state measurements were performed with a Perkin-Elmer LS-5 spectrofluorimeter, and fluores- cence time-resolved experiments were made by using phase and modulation techniques, in a GREG 200 multifrequency fluorimeter.Laser-flash photolyses were carried out with a set-up described previously.' Briefly, a nitrogen laser [Laseroptics, 5 ns full width at half-maximum (FWHM) and 5 mJ per pulse] was employed. The signal was collected by a digitizing scope where it was averaged and then transferred to a computer. All measurements were performed in de-aerated solutions at 298 K. The number of absorbed photons was calibrated by using a standard solution of zinc tetraphenylporphyrin (ZnTPP) in acetonitrile, with a triplet quantum yield and molar absorp- tion coefficient of 0.9 and 7.1 x lo4 1 mol-' cm-' (470 nm), respectively." Bleaching experiments were carried out by irradiation with light from a medium-pressure lamp, using a glass filter to isolate the 366 nm band. Anthracene consumption was evalu- ated from the change in its near-UV absorption in a Shimadzu-160 spectrophotometer.The absorption band shape does not change on the addition of indole or with the irradia- tion time. Carbon tetrachloride was used as reference." Quenching of Singlet Anthracene by Indoles Excited singlet anthracene is strongly quenched by indole and its derivatives. Fluorescence quenching gave linear Stern-Volmer plots, showing that no significant indole-anthracene ground-state complex is formed. From the slope of these plots, and considering the lifetime of singlet anthracene in the absence of indoles, bimolecular quenching rate constants have been obtained in different solvents ranging from n-heptane to ethanol-water mixtures. Representative values are given in Table 1.In acetonitrile, the rate constants decrease with increasing oxidation potential of the indole, as can be seen in Fig. 1. This can be taken as an indication that the quenching of singlet anthracene by indoles most likely proceeds through a charge-transfer mechanism implying electron transfer from the indole to the aromatic molecule. This holds for all indoles studied in highly polar solvents, independent of the nature of the substit- uents in the indole ring. In non-polar solvents such as n-heptane, the rate constants do not follow a simple correlation with the donor capability of the indole, as deduced from the J. Chem. SOC.,Faraday Trans., 1996,92 (l), 17-22 Table 1 Quenching rate constants for the deactivation of singlet 11.0 1 anthracene by indoles in different solvents indole EJV" solvent k,/109 1 mol-' indole 1.24 heptane 8.6 ethyl acetate 1.3 ethanol 3.3 methanol 5.1 acetonitrile 6.0 ethanol-water 5.2 methanol-water(25Ib 6.3 U7Ib 1-methylindole 1.23 heptane ethyl acetate 0.8 0.5 ethanol 2.3 methanol 5.2 acetonitrile 8.2 ethanol-water 6.5 1,2-dimethylindole heptane ace t oni t rile (25Ib 4.3 16 5-methoxyindole 1.18 heptane 11.4 ethanol 7.9 2-methylindole 1.10 heptane ethanol 13.9 8.0 acetonitrile 14.7 3-methylindole 1.07 heptane ethanol 16.3 8.3 acetonit rile 18.1 2,3-dimethylindole 0.93 heptane ace toni t rile 19 21 One-electron reduction potential for indolyl radical cations taken from ref.20 and 21. The number in parentheses indicates the per- centage of water (v/v) in the ethanol-water mixture. lack of correlation shown in Fig. 1. Furthermore, methyl sub- stituents in the 9- and 10-positions on the anthracene mol- ecule strongly decrease the reactivity towards indole, in both polar and non-polar media. The effect of solvent polarity on the rate constants is mark- edly dependent on the substitution at the N atom of the het- erocyclic ring, as shown in Fig. 2. For the case of l-methylindole, where the hydrogen atom on the N atom is replaced by a methyl group, the rate constants decrease stead- ily with the relative permittivity of the solvent.For indole, in the region of high relative permittivity, the trend is similar to that of l-methylindole, but at low polarity a bizarre behaviour can be seen (Fig. 2): For solvents with values of relative per- mittivity less than 10, the rate constants increase abruptly, reaching a value close to the diffusional limit found for n-heptane. 10.5 10.0 2z 0 -0 9.5 9.0 0.9 1 .o 1.1 1.2 1.3 E112N Fig. 1 Quenching rate constants of singlet anthracene as a function of the oxidation potential of indoles in acetonitrile (B) and in n-heptane (0) 18 J. Chem. SOC., Faraday Trans., 1996, Vol. 92 t 2.z 9.5t 0 10 20 30 40 relative permittivity Fig. 2 Quenching rate constants of singlet anthracene as a function of solvent relative permittivity : (B)indole; (0)l-methylindole The results in polar solvents are similar to those reported for the interaction of the singlet state of pyrene with indole compounds.8 In this case, the quenching has been reported to be electron transfer-like in ~haracter,'.~ with k, values strong- ly dependent on the solvent polarity.The rate constant values decrease by two orders of magnitude when the solvent is changed from ethanol-water (2 :3) to ethanol.' In the case of anthracene, this dependence is much weaker (Table 1). This is because the rate constants for anthracene are close in value to the diffusional limits found for most solvents, while for pyrene-indole mixtures they are two orders of magnitude smaller. This difference can be explained by considering that the energetics of the process measured by E""(indo1e) -Ered(hydrocarbon)-E"(excited singlet) is 0.1 eV more negative for anthracene than for pyrene.The diverse behaviour at low values of relative permittivity shown in Fig. 2 for indole and l-methylindole (both of almost identical oxidation potential2') points to a change in the quenching mechanism for the former. The quenching efi- ciency by l-methylindole, where an N-H bond is not present, is ten times lower in non-polar solvents than that obtained for indole. This is in agreement with the correlation of the quenching efficiency with the solvents with higher relative per- mittivity. This particular effect must be related to the presence of the N-H bond.This conclusion can also be reached by examining the trend in the rate constants for 2-methylindole and 1,2-dimethylindole. For the former, the rate constants are very similar on changing the solvent from acetonitrile to n-heptane, while for the latter, with methyl substitution on the N atom, the rate constants decrease steadily with decreasing relative permittivity. Anthracene photobleaching in the presence of indoles Anthracene photobleaching is promoted by indoles unsub- stituted at the nitrogen atom. The consumption of anthracene was evaluated in several solvents by following the decrease in the absorbance at the band of longest wavelength. For all sol- vents, it was observed as a clean photobleaching (up to 90% disappearance). The quantum yield of the chromophore loss increased proportionally to the singlet fraction quenched by the indole compound.In Fig. 3, a double inverse plot for the photobleaching quantum yield as a function of the indole con- centration in n-heptane can be seen. A good linear correlation is obtained and, from the slope and intercept, a Stern-Volmer constant (I&) is obtained which agrees with that from fluorescence-quenching experiments. These results indicate that the photoreaction involves exclusively the excited singlet. If the efficiency of the photobleaching (a) is defined as the number of anthracene molecules consumed per quenching Table 2 Effect of solvent on photobleaching efficiency indole solvent U13.0 1-methylindole heptane <0.04 .z 2.5 -f0 \ 7 20-15 -I ." 0 20 40 60 80 100 [indole]-'/dm3 mol-' Fig.3 Double inverse plot of anthracene photobleaching as a func- tion of indole concentration in n-heptane event, its value can be obtained from eqn. (l), @bleaching @quenching where @quenching is given by -KdQI @quenching -1 + WQI Photobleaching efficiencies obtained using several indoles are given in Table 2. These data show that the photoreaction takes place with an appreciable quantum yield only in com- pounds with an H atom on the nitrogen. This fact indicates that N-H bond cleavage must be the primary photochemical process. Similarly, hydrogen abstraction at the -NH indole ring to give an N-centred radical has been reported for rad- icals such as tert-b~toxyl,~~,~~ peroxylZ4 and highly oxidizing radicals.'9' Furthermore, the value of the photoreaction efficiency is highly sensitive to the nature of the solvent. The photoreac- tion proceeds quantitatively in non-polar solvents, proceeds to a lesser degree in polar solvents, and is considerably sup- pressed in aqueous solvent mixtures. For a given solvent, a values are nearly independent of the nature of substituents in positions other than on the N atom. Laser flash photolysis experiments The transient absorption spectrum produced by the photolysis of anthracene in acetonitrile in the presence of indole is shown 0.06 8 0.04 acetonit rile ethanol ethanol-water (25)" 1,2-dimethylindole heptane indole heptane ace tonitrile ethanol methanol ethanol-water (25)" 5-methox yindole heptane ethanol ethanol-water (17)" 2-methylindole heptane acet onit rile ethanol ethanol-water (17)" 3-methylindole heptane acetonit rile ethanol ethanol-water (25)" 2,3-dimethylindole heptane <0.03 <0.01 <0.02 <0.03 1.o 0.7 1 0.65 0.65 0.41 1.o 1.o 0.58 0.28 0.68 0.67 0.28 1.o 0.71 0.64 0.25 1.o " The number in parentheses indicates the percentage of water (v/v) in the ethanol-water mixture.in Fig. 4. The broad absorption band around 500 nm can be assigned to the neutral indolyl radi~al,'~-'~ while the 425 nm peak corresponds to triplet anthra~ene.'~ The inset in Fig.4 demonstrates that the absorptions at 425 and 500 nm belong to different species. The absorption typical of the radical cation of indole, centred at 580 nm,21,2s*28is almost negligi- ble, indicating the absence of long-lived radical ions. The quantum yield of triplet anthracene formation in the presence of indole was determined from the initial absorption at 425 nm by using a molar absorption coefficient of 64700 1 mol-' cm-l 27 The value of the quantum yield was 0.20, and corre- sponds to the quantum yield of the non-quenched singlet with the indole concentration used in this experiment, indicating that induced spin inversion does not occur. The decay kinetics at 425 nm of the triplet-triplet absorption of anthracene obey a mono-exponential law.The decay rate constant is unchanged in the presence of indole, hence triplet quenching by indole can be disregarded. When similar experiments were carried out using n-heptane as solvent, no signal on a ps time-scale could be detected in the region 450-700 nm. Considering that the quantum yield of the photoreaction is unity, this .o 02 -001 c (d ' 000% a timev) (d \ JL0.02 In.&I\ 0.00 400 450 500 550 600 wavelength/n m Fig. 4 Transient absorption spectrum of anthracene in the presence of 0.1 mol 1-' indole in acetonitrile, taken 5 ps after the laser pulse. Inset: decay profiles at 420 and 500 nm J. Chem. SOC.,Faraday Trans., 1996, VoZ. 92 19 0.09 I I I I R .3An40.06 Q)0 c ae2 -0 (d 0.03 0.00 400 500 wavelength/nm Fig.5 Transient absorption spectrum of anthracene in the presence of 0.2 mol I-' result suggests that the radicals formed are too short-lived to be observed under our experimental conditions. Fig. 5 shows the transient absorption spectrum for anthra- cene in the presence of 1-methylindole in acetonitrile solution. The 550-600 nm absorption is characteristic of the indole radical ati ion.'^,'^,^^ The broad absorption band centred at ca. 700 nm can be ascribed to the radical anion of anthra- cene.' The relative absorbances at these wavelengths agree with the ratio of the molar absorption coefficients for these species. The decay timescales of these signals are consistent with the presence of free ion pairs, and can be fitted to second- order kinetics.The presence of long-lived radical ion interme- diates in this case is as expected, from the lack of H atoms available for proton transfer owing to the substitution by a methyl group. From the absorption at 425 nm, the triplet yield in the presence of 1-methylindole can be evaluated as 0.28. Taking into account the unquenched fraction of anthra- cene singlets, the induced intersystem crossing yield is 0.23. In n-heptane, there was no observed transient absorption in the region 450-700 nm. Discussion The results presented in this work show that indoles quench singlet anthracene with high efficiency. The quenching mech- anism is dependent on the nature of the indole compound and the polar properties of the medium.Thus, 1-methylindole shows strikingly different behaviour to indole. Considering that the Gibbs energy change, AGO, of electron transfer from polar / 6-6+/An*+HN [h--HN ]\ \ non polar solventI [m'---'N/ \] Products (@= 1) Scheme 1 20 J. Chem. SOC.,Faraday Trans., 1996, Vol. 92 An '-In*+ I\ 600 700 800 1-methylindole in acetonitrile, taken 1 ps after the laser pulse both indole compounds to excited anthracene is almost identi- cal, the different behaviour can be explained as being due to the presence of the H atom on the nitrogen in the case of indole. The results obtained with indole not substituted at the N atom are consistent with the mechanism depicted in Scheme 1.In polar media, the geminate ion pair is likely to be pen- etrated by the solvent to give a solvent-separated ion pair. Based on the timescale of our laser photolysis experiments, the neutral or radical ions observed are those that escape 'cage recombination'. The lack of signal found in the region where the anthryl and indolyl radical ions absorb indicates that diffusional rupture of the ion pair to give separated ions is a negligible decay pathway of this intermediate. On the other hand, using a molar absorption coefficient of 2000 1 mol-' cm-',26 a value of 0.45 can be estimated for the indolyl radical yield under conditions of total singlet quenching. This value is lower than the photobleaching quantum yield (0.7 1).This fact would suggest that indolyl and anthryl long-lived free radicals are not the only species that lead to photoreac- tion (i.e. products have to be formed prior to the separation of the neutral radicals). In other words, the diffusional separation of radicals competes effectively with hydrogen abstraction within the solvent cage. The absence of long-lived transients, indolyl radicals or radical ions, when anthracene is irradiated in the presence of indole in n-heptane as solvent, and the photoreaction effi-ciency close to unity, would indicate that the geminate ion pair decays exclusively by fast intramolecular hydrogen trans- solvent -/$ An+ HN \ AnH' + 'N,/ Products / polar / 6-6+/ solvent An* +RN -[An---RN] -\ \ non polar IsolventI /An + RN' An3 + RN \ \ Scheme 2 fer from the N-H bond (Scheme 1).However, the lack of correlation of the quenching rate constants for indole with the solvent relative permittivity in media of low polarity (Fig. 2), and deviations to the electron donating ability of indoles with an N-H bond in the region of high oxidation potential, suggest that in non-polar media, another pathway to the quenching of singlet anthracene is present. Owing to the high reactivity of the H at the N-H bond, direct hydrogen abstraction most likely contributes to the quenching process in media of low polarity. The 'almost diffusional' rate con- stant obtained in n-heptane indicates that in this solvent direct H atom transfer must be the dominant mechanism.When the polarity of the solvent increases, proton transfer preceding electron transfer will be favoured. Lissi et a!.15have suggested that in the deactivation of singlet anthracene by diethyl- hydroxylamine in benzene, a free-radical type process involv- ing direct transfer of hydrogen atom could dominate the quenching process. Analysis of the photoreaction efficiencies shows that photo- bleaching decreases in polar solvents. However, the diminu- tion does not follow a clear relationship with the solvent relative permittivity. Thus, in media of equal relative permit- tivity, such as ethanol-water (25% water) and acetonitrile, the photobleaching quantum yield is much lower in the aqueous mixture.This behaviour holds for all indoles containing an N-H bond. This result would indicate that the electron transfer back to the ground state species is more favourable in the aqueous solvent. Hydrogen-bonding interactions in the ionic pair possibly reduce the rate of the proton transfer which competes with the non-radiative transition to the ground state. A decrease in the reaction yield with the solvent polarity has been also reported for the photoreaction of anthracene with aniline' and diethylhydroxylamine.' This effect has been attributed to a competition between proton transfer and formation of the dissociated ion pair. The lack of long-lived ionic transients in the anthracene-indole system points to a lack in importance of the diffusional separation of the radical ions.Thus, back electron transfer should be the main pathway in high protic media. The correlation of k, with the oxidation potential (Fig. 1) and the observation of radical ions as transient species (Fig. 5) for indoles substituted on the N atom, points in this case to an electron-transfer mechanism for the quenching. In these com- pounds, where the radical cation cannot be deprotonated, the geminate ion pair breaks up through three competing path- ways : diffusional separation of ions, spin-forbidden inversion and back electron transfer (Scheme 2). The first of these leads to the formation of separated radical ions. The yield of anthryl radical anions, estimated using the initial absorption at 720 nm and taking a molar absorption coefficient of 12000 1 mol-' cm-' ,27 was approximately 0.65 in acetonitrile.The above calculated quantum yield for the induced spin-forbidden transition was 0.23. Then, the recombination of radical ions inside the solvent cage leads mainly to triplet for- mation. Formation of the triplet state on fluorescence quenching of polycyclic aromatic hydrocarbons has been observed in several systems, in polar as well as in non-polar media.' 92.29-32 Delouis et reported that enhancement of triplet formation of pyrene to a quantum yield of unity takes place on quenching the pyrene fluorescence with DABCO in cyclohexane. Ha~himoto~~ has explained that the enhanced formation of the triplet state in the pyrene fluorescence quenching by purines in aqueous solution is due to weak electron-transfer interactions in the exciplex intermediate.Spin-forbidden inversion with a quantum yield of 0.09 has been reported for the quenching of pyrene by indole in aceto- nitrile.' The rate contant for recombination to the triplet state can be estimated from the usually assumed value (1 x lo' 1 mol-' s-l) for k,,, in a~etonitrile.~'~~,~~ From the quantum yields of the radical ions and return to the triplet state, a value of kin" of ca. 3 x LO9 1 mol-' s-' for anthracene-1-methylindolecan be estimated. This value is higher than that reported for the pyrene-indole system, but it is in agreement with the more favourable energy gap found in the case of anthracene, although electronic coupling must also play a role.In conclusion, the quenching mechanism of the excited singlet state of anthracene by indoles is markedly dependent on the substitution at the nitrogen atom and on the polar properties of the medium. In non-polar solvents, fast direct proton transfer can be considered as the main deactivation process for indoles unsubstituted at the N-H group. However, in polar solvents the quenching mechanism can be described by a charge-transfer mechanism, for all indole deriv- atives. Triplet formation could be observed only for the quenching by 1-methyl-substituted indoles in acetonitrile. We are grateful to DICYT (Universidad de Santiago de Chile) and CONICET (Argentina) for financial support.The use of the GREG-200 fluorimeter was made possible through grants FONDECYT-1864012 and Fundacion Andes C-12302. References 1 A. Weller, 2.Phys. Chem., Neue Folge, 1982,130, 129. 2 N. Mataga, Pure Appl. Chem., 1984,56, 1255. 3 N. Mataga, T. Osaka, Y. Kanda and H. Shioyama, Tetrahedron, 1986,42,6143. 4 H. Lemmetyinen, R. Ovaskainen, K. Nieminen, K. Vaskonen and I. Sychtchikova, J. Chem. SOC., Perkin Trans. 2, 1992, 11'3. 5 Y. Hirata, Y. Kanda and N. Mataga, J. Phys. Chem., 1983, 87, 1659. 6 A. Hirata, T. Saito and N. Mataga, J. Phys. Chem., 1987, 91, 3119. 7 N. Miyoshi and G. Tomita, Photochem. Photobiol., 1979,29, 527. 8 M. V. Encinas and E. A. Lissi, Photochem. Photobiol., 1985, 42, 491. 9 H. A. Montejano, J. J. Cosa, H.A. Garrera and C. M. Previtali, J. Photochem. Photobiol. A: Chem., 1995,86, 115. 10 J. P. Palmans, M. Van der Auweraer, A. M. Swinnen and F. C. De Schryver, J. Am. Chem. SOC., 1984,106,7721. 11 M. M. H. Khalil, N. Boens and F. C. De Schryver, J. Phys. Chem., 1993,97,3111. 12 N. C. Yang and J. Libman, J. Am. Chem. SOC., 1973,95,5783. 13 S. Vaidyanathan and V. Ramakrishnan, Indian J. Chem., 1975, 13, 257. 14 M. V. Encinas, C. Majmud, J. Garrido and E. A. Lissi, Macro-molecules, 1989,22, 563. 15 E. A. Lissi, M. A. Rubio and M. Fuentealba, J. Photochem., 1987, 37, 205. J. Chern. SOC., Faraday Trans., 1996, VoI. 92 21 16 P. Bortolus, N. Camaioni, L. Flamigni, F. Minto and S. Monti, J. Photochem. Photobiol., A: Chem., 1992,68,239. 27 T.Shida, Electronic Absorption Spectra of Radical Ions, Elsevier, Amsterdam, 1988. 17 V. Avila, J. J. Cosa and C. M. Previtali, An. Asoc. Quim. Argent., 1990,78, 279. 28 J. F. Baugher and L. I. Grossweiner, J. Phys. Chem., 1977, 81, 1349. 18 J. K. Hurley, N. Sinai and H. Linschitz, Photochem. Photobiol., 1983,3$, 9. 29 30 M. Ottolenghi, Acc. Chem. Rex, 1973,6, 153. G. J. Kavarnos and N. J. Turro, Chem. Rev., 1986,86,401. 19 M. V. Encinas, M. A. Rubio and E. A. Lissi, J. Photochem., 1982, 31 H. Masuhara, H. Shioyama, T. Saito, K. Hamada, S. Yasoshima 20 3,97. G. Merenyi, J. Lind and X. Shen, J. Phys. Chem., 1988,92, 134. 32 and N. Mataga, J. Phys. Chem., 1984,88,5868. F. Lewitzka and H. G. Lohmannsroben, Z. Phys. Chem., Neue 21 22 S. V. Jovanovic and S. Steenken, J. Phys. Chem., 1992,96,6674. E. Leyva, M. S. Platz, B. Niu and J. Wirz, J. Phys. Chem., 1987, 91,2293. 33 Folge, 1990, 169, 203. J. F. Delouis, J. A. Delaire and N. Ivanoff, Chem. Phys. Lett., 1979,61, 343. 23 M. V. Encinas, E. A. Lissi, C. Majmud and A, F. Olea, Int. J. Chem. Kinet., 1991,23, 761. 34 35 S. Hashimoto, J. Phys. Chem., 1993,97,3662. S. L. Mattes and S. Farid, Science, 1984,226,917. 24 J. E. Packer, J. S. Mahood, R. L. Willson and B. Wolfenden, Int. J. Radiat. Biol., 1981,39, 135. 36 N. Mataga, T. Asahi, Y. Kanda, T. Okada and T. Kakitani, Chem. Phys., 1988,127,249. 25 S. V. Jovanovic and M. G. Simic, J. Free Radicals Biol. Med., 1985, 1, 125. 26 D. V. Bent and E. Hayon, J. Am. Chem. SOC., 1975,97,2612. Paper 5/04511H; Received 10th July, 1995 22 J. Chem. SOC.,Faraday Trans., 1996, Vol. 92
ISSN:0956-5000
DOI:10.1039/FT9969200017
出版商:RSC
年代:1996
数据来源: RSC
|
5. |
ν2Band region of nitrate as an indicator for contact ion pairing in aqueous lithium and calcium nitrate solutions |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 92,
Issue 1,
1996,
Page 23-28
Gerhard Fleissner,
Preview
|
PDF (937KB)
|
|
摘要:
v2 Band region of nitrate as an indicator for contact ion pairing in aqueous lithium and calcium nitrate solutions Gerhard Fleissner, Andreas Hallbrucker and Erwin Mayer* Institut fiir Allgemeine, Anorganische und Theoretische Chemie, Universitat Innsbruck, A-6020 Innsbruck, Austria The FTIR spectra of 2 and 10 mol dm-, LiNO, solutions, and of 0.5, 3 and 5 mol dm-, Ca(NO,), solutions in D20 are shown in order to demonstrate the usefulness of the v2 band region of nitrate as an indicator for contact-ion pairing in aqueous alkali-metal and alkaline-earth-metal nitrate solutions. In fourth-derivative curves of the original composite bands, the develop- ment of a second band and its increase with concentration is seen at low frequency of the band, arising from the 'free' nitrate.The band at low frequency is assigned to the contact ion pair, and the assignment is supported by a parallel study of the v4 band region of the same solutions. The separation of the two bands in the v2 band region is between 3.3 and 5.2 cm-',and for 10 mol dm-LiNO, solution it is even less than half of the average half-bandwidths of the two component bands. Reliable curve-fitting of the composite bands is possible, despite their strong overlap, by comparison of the fourth derivatives of the original composite bands with those of the sum of the curve-fitted component bands. We suggest that this comparison of fourth derivatives can, for spectra of high signal-to-noise ratio, provide reliable curve fits even when the separation of the peak maxima is less than half of the average of the half-bandwidths of the component bands.The nitrate ion is probably the most popular model system for investigating both contact ion pairing by vibrational spectros- copy and the effects on this pairing of e.g. concentration, tem- perature and solvent (reviewed in ref. 1-3). The 'free', i.e. unperturbed, hydrated nitrate ion, of D,, symmetry, should exhibit the following four-band spectrum: vl (A;, Raman) at ca. 1050 cm-', v2 (A:, IR) at ca. 830 cm-', v3 (E', Raman, IR) at ca. 1380 cm-',and v4 (E', Raman, IR) at ca. 720 cm- '. The well known splitting of the v, band into two components is caused by a lifting of the degeneracy owing to lowering of the symmetry of the hydrated ion.' Contact ion pairing generates a new set of bands, somewhat displaced from those of the hydrated ion.However, although the number of nitrate bands is essentially doubled by the presence of both free and contact ion-paired (or bound) nitrate, their separation and quantitat- ive evaluation is often impossible owing to insufficient separa- tion of the bands. Contact ion pairing is seen most clearly through the development of a second band in the v4 band region at ca. 740 cm-' and, sometimes, the doubling of the non-degenerate v1 band.'*, Interpretation of the v3 band region is difficult and in most cases not unambiguous. For the non-degenerate v2 band region at ca. 830 cm-', several reports by Irish and co-workers have described new bands arising from contact ion pairing from Zn(NO,), in methan01,~ aqueous Gd(NO,), s~lution,~,~aqueous Ce(NO,),,' aqueous Mg(NO,),* and AgNO, in waterg and in acetonitrile." In these reports, curve fitting of the two com- ponent bands from free and bound nitrate was only possible for systems where strong perturbation of the anion by the cation occurred, since only then were the two bands sufficient- ly separated.For aqueous Mg(NO,), solution, where pertur- bation of the nitrate ion by Mg2+ is weak, a shift of the peak maximum to lower frequency was reported on going, with increasing concentration, from free nitrate to the monodentate and the didentate nitrate ion (825, 821 and 819 cm-', respec-tively, see Table 2 in ref. 8), but separation of the component, bands by curve-fitting was apparently not possible with spectra obtained by dispersive instruments. Of the many solvents used for studying ion pairing, water is probably the one with most problems.'Y2 Nevertheless, we chose to use water as a solvent, first because of our continuing studies of the increase in contact ion pairing in the glassy state of hyperquenched aqueous electrolyte solutions,' '9' and sec- ondly, because of its biological significance.We show here for aqueous LiNO, and Ca(NO,), solutions that separation of the component bands arising from free and bound nitrate in the v2 band region, and their reliable curve-fitting, is possible. This requires FTIR spectra of very high signal-to-noise (S/N) ratio and a comparison of fourth-derivative curves as criteria for artefact-free subtraction of background and good quality curve-fits.Ca(NO,), was chosen as solute as part of our con- tinuing study on the temperature dependence of its contact ion pairing."Y1' LiNO, was chosen since the separation of the component bands is much less than for Ca(NO,), , and, there- fore, LiNO, nicely demonstrates the power of the technique. In addition, the extensive literature on the spectral features of both electrolytes in aqueous solution is helpful for assignment and discussion.' 1-22 The approach shown here even allows the contact ion pairing in aqueous solutions of biologically important alkali-metal and alkaline-earth-metal cations to be followed quantitatively, via the non-degenerate v2 band.Note that even this improved curve-fitting method does not enable separation of bands due to free and bound nitrate in the v1 band region. Therefore, for Li+ and Ca2+ (and possibly the other alkali-metal and alkaline-earth-metal cations), the v2 band region of nitrate is the only useful non-degenerate mode for following contact ion pairing in aqueous solution. (Note that James and Frost's22 study of ion-ion-solvent interactions using Fourier transformation and band analysis of the v1 band of nitrate is basically a different approach). We had pre- viously pointed out that curve-fitting of the v2 band region gives, for Mg, Ca and Sr nitrate salts in D20 solution, the least reliable results because of severe band overlap.' ' This now has to be changed because of the additional sensitive comparison of fourth-derivative curves.This study goes beyond the evaluation of the v2 band region in that it shows where the limits of reliable curve-fitting are for strongly overlapping bands in spectra with a high S/N ratio. It is an extension of the comparison of second derivatives sug- gested by Gans2, and used in our earlier studies.' ',12 Experimental Anhydrous LiNO, (Ventron) and Ca(NO,), .4H20 (Merck, pa quality) were dried over P20,, and solutions were made with D20 (99.7%). D,O was preferred to H20 as a solvent J. Chem. SOC.,Faraday Trans., 1996,92 (l),23-28 Table 1 Curve-fitting analysis of the v2 and v4 band region in FTIR spectra of aqueous 10 mol dm-3 LiNO, solution in D20, and the comparison with Raman spectra v,,, FWHH ~~~ ~ relative area and /cm-' /cm-' % Gauss assignment (%) v2 (FTIR," 829.1 10.0 20 64, F best fit) 825.8 8.1 88 36, B v2 (FTIR," 828.6 13.6 44 60 poor fit) 826.8 8.5 100 40 v4 (FTIR") 717.9 14.9 78 21, F 734.2 27.5 95 79, B v4 (IRb) 719 18, F 735 70, B 752 12 v4 (Ramanb) 719 67, F 734 33, B This work, spectra recorded at 320 K.From Riddell et al." (determined from Fig. 7), spectra recorded at 298 K. v,,, is the peak frequency and FWHH is the full width at half height of the curve- fitted component bands. F indicates free nitrate; B, bound, i.e. contact ion paired nitrate. % Gauss are the values obtained when using a sum of Gaussian and Lorentzian peak shapes.because the intense librational band is shifted to lower fre- quency. ZnSe windows and a 15 pm spacer were used. The spectra were recorded at 300 K except for the 5 mol dm-, Ca(NO,), solution, the spectrum of which was recorded at 295 K. The temperature was kept constant to k0.1 K during the measurements. Transmission FTIR spectra were recorded on a Biorad FTS 45 at 2 cm-' resolution (UDRl), by co-adding between 250 and lo00 scans. Peak maxima are given in the figures to an accuracy of 0.1 cm-' and were found to be reproducible to within a few tenths of a wavenumber. The FTIR spectrum of water vapour was subtracted from each spectrum. Data pro- cessing was performed with Spectra-Calc (Galactic).For sub- traction of the librational band of water in the v4 band region, the spectrum of 0.5 mol dm-, CaCI, solution in D,O was used, since in this region, the bandshape was very similar to that of 1 mol dm-, LiNO, and 0.5 mol dm-, Ca(NO,), solu-tion in D20.Subsequently, the remaining sloping background was subtracted with a multipoint spline-function routine. For the v2 band region, only this last type of background subtrac- tion was necessary because the sharp band at ca. 830 cm-' is sufficiently shifted to low frequency from the intense libra- tional band of the solvent. In addition, setting of the break points for background subtraction with the multipoint spline- function routine was much more obvious for the v2 than for the v4 band region.The effects of baseline correction were controlled by com- paring second-derivative curves of the original bands with those of the bands after subtraction of the sloping back- ground. The effects of curve fitting were controlled by com- paring fourth derivatives of the background-corrected composite bands with those of the sum of the curve-fitted component bands. This comparison of derivatives is, in our experience, a sensitive indicator for possible artefacts during manipulation of the data, and good superposition of fourth derivatives is taken as a criterion for the quality of the curve- fits. This comparison of fourth derivatives is similar to the procedure recommended by Friesen and Michaelian for deconvoluted spectra,24 and by Gans' for second derivatives of spectra. We had previously applied a comparison of second but note that the comparison of fourth deriv- atives used here is more sensitive in cases where the limits of curve-fitting are approached.An S/N ratio of maximum signal/estimated rms noise was used. Table 2 Curve-fitting analysis of the v2 and v4 band region in the FTIR spectra of aqueous 0.5, 3 and 5 mol dm-3 Ca(NO,), solution in D20, and comparison with Raman spectra concentration /mol dm-3 v2 (FTIR)" 0.500 v4 (FTIR)" 0.500 v4 (Raman)b 0.500 v2 (FTIR)" 3.00 v4 (FTIR)" 3.00 v4 (Raman)' 2.98 v2 (FTIR)" 5.00 v4 (FTI R)" 5.00 v4 (Raman)' 5.00 This work, spectra recorded at 300 K; Vmax /cm -' FWHH /cm -' YO Gauss relative area and assignment (Yo) ~~ 830.5 826.3 7.1 4.8 (22.8, 7.6)' (7.3, 6.9)' 89, F 11, B 718.8 19.3 100 53, F 737.6 13.0 30 47, B 717.4 22.8 15 94, F 736.7 11.6 100 6, €3 830.3 7.8 1 34, F 825.1 8.2 67 66, B 717.3 21.2 0 36, F 738.8 13.7 0 64, B 716.7 23.5 58, F 738.2 19.3 41.5, B 830.1 10.6 (3.4 x 1038, 10.6)' 32, F 824.0 820.3 9.2 8.7 (8.4, 5.5 x lo2)' (17.1, 9.4)' 53, B 15, triplet? 717.6 16.3 741.1 15.8 715.7 23.4 739.1 20.1 From ref.12, spectrum recorded at 295 K; 'From Sze" 56 17, F 28 83, B 47.6, F 52.4, B (from Table 1, p. 18), spectra recorded at 298 K; Values in parentheses are G- and L-width values, in cm-'. vmax is the peak frequency and FWHH is the full width at half height of the curve-fitted component bands.F indicates free nitrate; B, bound, i.e. contact ion paired nitrate. Yo Gauss are the values obtained when using a sum of Gaussian and Lorentzian peak shapes; G-width and L-width (Gauss-width and Lorentz-width) are the values when using a product function. J. Chem. SOC.,Faraday Trans., 1996, Vol. 92 Curve-fitting was first attempted with a sum of Gaussian and Lorentzian peak shapes (SpectraCalc, curve-fit software2’). When this did not give a satisfactory fit, a product of Gaussian and Lorentzian peak shapes (Spectrum Square Associates, data fit software26) was used. The parameters are YOGauss in the first case and Gauss-width and Lorentz-width in the second, and these are listed in Tables 1 and 2.Results and Assignment Fig. l(a) shows the v2 band region of the background-corrected FTIR spectra of 2 and 10 mol dm-, LiNO, solu- tions in D20,and in addition, for the 10 mol dm-, solution, the two component bands of two different curve-fits. Fig. l(b) shows fourth-derivative curves: for the 2 mol dm-, solution, only one band is indicated by the fourth-derivative curve, but for the 10 mol dm-, solution, separation into two com-ponents bands, one with a peak value similar to that of the 2 mol dm-, solution and the other at lower frequency, is clearly observed. Fig. 1 contains, in addition, two different curve-fits for the composite band of the 10mol dmP3 solution (a), and the com- parison of fourth-derivatives curves (b).The curve-fit shown in the middle presents the best fit of the fourth derivative of the original composite band with that of the sum of the two curve-fitted component bands. This best curve-fit was 78304 r 8304 8289 7 1-8275 10 rnol dm” -827 0y827 5P 10 rnol F‘rn-3 1n\ 860 830 800 860 830 800 wavenumber/cm-’ (a) FTIR spectra of the v2 band region, after subtraction of the background, of 2 and 10 mol dm-3 LiNO, solution in D20, and for 10 mol dm-, solution, the component bands of the best (middle) and a poor curve-fit (bottom). (b) Fourth derivatives of the original background-corrected bands, and for 10 mol dm-3 solution the com- parison with those of the sum of the curve-fitted component bands (broken lines).Note the good superposition for the best fit and the shift in peak maximum for the poor fit. (1 1-point second derivative convolution was applied twice throughout for generating fourth derivatives). obtained by changing, in small increments, either the peak position or the full width at half height (FWHH) of the two component bands and by following these changes in compari- son with the fourth derivatives. About five curve-fits were necessary for obtaining the optimal fit, and from these curve- fits we estimate that the error in the areas is f4%. The curve- fit shown at the bottom was obtained by allowing all band parameters to be optimized simultaneously. For this curve-fit, the fourth derivative of the sum of the component bands has a peak maximum at 827.0 cm-’, which lies between the peak values seen in the fourth derivative of the original composite band, and is therefore clearly wrong.Without this comparison of fourth derivatives, it would have been impossible to differ- entiate between the best and the poor curve-fits because their x2 values, which are an indicator of the quality of the fit,25 are very similar (1.0 x lop4for best fit and 8.4 x for poor fit). The band parameters are listed in Table 1. In Fig. 2(a) we show for comparison the v4 band region of the same 2 and 10 mol dmP3 LiNO, solution in D20, but background-corrected in a different way (see Experimental section). For the 2 mol dm-, solution (top), only one band is observed at 718 cm-’. For the 10 mol dm-, solution (bottom), the asymmetric bandshape indicates the presence of at least two overlapping bands.The second-derivative curves of the original bands shown in Fig. 2(b) exhibit one band only for the 2 mol dmP3 solution but two bands for the 10 mol dm-, solution. Fig. 2 further contains, for the 10 mol dm-, solution, the two curve-fitted component bands (a) and the second derivative of the sum of the curve-fitted component bands [(b) broken lines]. Note that we had to use second- 1 : : ~ : ‘ 1 : : ~ : : : : : : 770 740 710 680 770 740 710 680 wavenum be r/cm-’ Fig. 2 (a) FTIR spectra of the v4 band region, after subtraction of the solvent, of 2 and 10 mol dm-, LiNO, solution in D,O, and for 10 mol dm- solution, the curve-fitted component bands.(b) Second derivatives of the original background-corrected bands, and for 10 mol dm-3 solution, the comparison with that of the sum of the curve- fitted component bands (broken lines). (25-point second derivative convolutes; for 2 mol dm-3 solution a 25-point smooth was applied twice). J. Chem. SOC.,Faraday Trans., 1996, Vol. 92 derivative curves for the v4 band region because the S/N ratio of the spectra is much lower than that of the spectra in the v2 band region. Because of the above-mentioned difficulties in subtraction of the background and the low S/N ratio, the error in the band areas is higher for the v4 than for the v2 band region, and is estimated as +lo%.The S/N ratios are ca. 7000 and ca. 5000 for the bands of 2 and 10 mol dm-, LiNO, in the v, band region, but they are only ca. 20 and ca. 100 in the v4 band region of this same solution. The results shown here for the v4 band region are consistent with those reported by Irish and co-workers.16,20 For the 10 mol dm-, solution in particular, the asymmetric bandshape is very similar to that shown by Riddell et al. (Fig. 7, ref. 20). However, we preferred to use two component bands only for the curve-fit, and not three as Riddell et al. have done,20 because it gave optimal results in terms of comparison of second derivatives. The band parameters of the v4 band region are listed in Table 1. Table 1 also contains, for comparison ~from the study of Riddell et UE.,~ the band parameters of the v4 band region for the Raman and IR spectrum of 10 mol dm-, LiNO, solution in D20 (from their Fig.7, determined by cutting out and weighing). The assignment of the two component bands in the v2 band region of 10 mol dm-, LiNO, solution is straightforward and is based on the well established patterns for the v4 band region. The band at 830.4 cm-' in 2 mol dm-, LiNO, solu- tion (at 829.1 cm-' in 10 mol dmP3 solution) is assigned to free nitrate, and the band at 825.8 cm-' to the out-of-plane mode of contact-ion paired or bound nitrate. In the following, we will speak of free nitrate with the understanding that it contains varying amounts of solvent shared and/or solvent separated ion pairs, which are spectroscopically indistinguish- able.For 10 mol dmP3 LiNO, solution, the relative band areas of free and bound nitrate in the v2 band region are 64 and 36%, and these areas are remarkably close to those obtained by Riddell et by Raman spectroscopy for the v4 band region (67 and 33%, see Table 1). The separation of bands in the v2 band region of Ca(NO,), in D20 solution is reported in a similar manner. Fig. 3(a) shows the v2 band region of the background-corrected FTIR spectra of 0.5, 3 and 5 mol dm-, Ca(NO,), in D,O solution and the component bands of optimal curve-fits. In Fig. 3(b), fourth-derivative curves of the original composite bands are compared kith those of the sum of the curve-fitted component bands (broken lines).For the 0.5 and 3 mol dm-, solution, only two bands are required for curve-fitting. However, for 5 mol dm-, solution, it was not possible to obtain super- position of fourth-derivative curves by using two bands only, and a third component band had to be included in the fit for optimal superposition. This third band component is not resolved even in the fourth-derivative curves but it is indicated by the asymmetry of the intense feature centred at 823.6 cm-'. The error in the band areas is estimated as f2% for the 0.5 and 3 mol dm-, solution, and as +5% for the 5 mol dm-, solution. The band parameters of these fits are listed in Table 2. In Fig. 4(a), we show, for comparison, the v4 band region of the same 0.5, 3 and 5 mol dmP3 Ca(NO,), solution in D20, but background-corrected in a different way (see Experimental section), and the curve-fitted component bands. In the v4 band region, only two component bands were necessary for curve- fitting the composite band of the 5 mol dmP3 solution, instead of three in the v2 band region.Fig. 4(b) contains a comparison of second derivatives of the original composite bands with those of the sum of the curve-fitted component bands (broken lines). The band parameters of the fits are also included in Table 2. This table contains, also for comparison, the band parameters for curve-fits of the v4 band region of Raman spectra of aqueous 0.5 mol dmP3 Ca(NO,), solution from ref. 12, and of the 3 and 5 mol dmP3 solution reported by Sze." ~82 6 5 ,::::,--823 6 : 5 mot drn-3 r823 860 830 so0 860 sjo so0 wavenurnber/crn-' Fig.3 (a) FTIR spectra of the v2 band region, after subtraction of the background, of 0.5, 3 and 5 mol dm-j Ca(NO,), solution in D20, and the curve-fitted component bands. (b) Comparison of fourth- derivatives of the original composite bands, after subtraction of the background, with those of the sum of the curve-fitted component bands (broken lines). (Seven-point, five-point and nine-point second- derivative convolutes were applied twice for generating the fourth derivatives of 0.5, 3 and 5 mol dm-, solution). The S/N ratios are CQ. 3000, ca. loo00 and ca. 3000 for the bands of 0.5, 3 and 5 mol dm-, Ca(NO,), in the v2 band region, but they are only ca.70, ca. 300 and CQ. 30 in the v4 band region of the same solution. The assignment of the two component bands in the v2 band region of 0.5 and 3 mol dmP3 Ca(NO,), D20 solution paral- lels that for LiNO, solution: the band at ca. 830 cm-' is assigned to free nitrate and that at ca. 825 cm- ' to the out-of- plane mode of contact ion-paired nitrate. The change in rela- tive band areas with concentration is consistent with the assignment. The assignment of the component at low fre- quency to contact ion pairs is further consistent with previous The curve-fit of the 5 mol dmP3 solution requires, in addition to bands at 830 and 824 cm-' which are assign- able to free and bound nitrate, a third band at 820 cm-', which is tentatively assigned to an ion triplet.Discussion Vandeginste and De Galan2 have developed criteria for the critical evaluation of curve-fitting in IR spectroscopy (reviewed for Maddams"). They have further pointed out that 'a good fit of a composite band system is a necessary, but certainly not a sufficient condition for a good recovery of quantitative data'. This is clearly seen by the curve-fits of 10 mol dm-, LiNO, solution shown in Fig. l(a), where very similar x2 values were obtained for the two curve-fits, but only the comparison of fourth-derivative curves allowed differentia- J. Chem. SOC.,Faraday Trans., 1996, Vol. 92 7jo 740 7io 680 770 740 7io 680 wave'nu m berlcm-' Fig. 4 (a) FTIR spectra of the v4 band region, after subtraction of the solvent, of 0.5, 3 and 5 mol dmP3 Ca(NO,), solution in D20, and the curve-fitted component bands.(b) Comparison of second deriv- atives of the original composite bands, after background subtraction, with those of the sum of the curve-fitted component bands (broken lines). (17-point second derivative convolution was used for the 0.5 mol dmP3 and 25-point for the 3 and 5 mol dmP3 solution; 3 and 5 mol dm- solutions were in addition smoothed with 25-point.) tion between poor and best fits. Gans and Gill2' 'have exam- ined the effect of S/N ratio of the data on resolvability and have found that the minimum band separation at which acceptable resolutions can be obtained decreases as S/N increases'. The degree of overlap is characterized by a param- eter 6, which is defined as27928 6 = (~2-V,)(l/FWHH, + l/FWHH,) where v1 and v2 are the peak maxima of two overlapping bands and FWHH, and FWHH, their full widths at half height.For an exact description of the individual bandshape, an IR spectrum with N bands requires 4N + 2 parameters [eqn. (2) in ref. 271. When the mathematical model is an approximation, which is the usual case for experimental spectra, good results are supposedly obtained only for 6 > 2 (ie.when the separation of the peak maxima is larger than the average of the two FWHHs) even when the number of bands is known (see Table 5 in ref. 27). In our study of the v2 band components of nitrate, the number of bands is derived from the fourth derivatives but FWHHs and peak shape are not known.Therefore, the math- ematical model is an approximation where 6 > 2 should be valid. However, because of the combined use of spectra of very high S/N ratio, of between ca. 3000 and ca. 10000, and the additional comparison of fourth-derivative curves shown in Fig. 1 and 3, reliable curve-fitting is possible for strongly over- lapping bands, with 6 c 2. The 6 values for the v2 band region are 0.74 for 10 mol dm-, LiNO, solution, and 1.2 and 1.3 for 0.5 and 3 rnol dm-, Ca(NO,), solution, respectively. These values are much lower than those predicted as the resolv- ability limit for bands of unknown bandshape, and they clearly indicate the advantage of this approach. The 6 values for the v4 band region are 1.7 for 10 mol dm-3 LiNO, solu- tion, and 2.9 and 2.6 for 0.5 and 3 mol dm-, Ca(NO,), solu-tion, respectively. These last two values are in the range of 6 > 2, where reliable separation of bands of unknown band- shape is possible even without a comparison of derivatives, but the value of 1.7 is in a range where comparison of deriv- atives is useful.The comparison of fourth-derivative curves seems to have advantages even when the number of bands cannot be deter- mined unambiguously from the derivative curve. For the v2 band region of 5 mol dmP3 Ca(N03)2 solution [Fig. 3(a)], curve-fitting was attempted first with two band components, and a third band was added only when it became obvious that it is impossible to superimpose the fourth derivative of the original band with that of the sum of two curve-fitted com- ponent bands.Intensity ratios of the v1/v2 IR band areas have been used as an indication of perturbation of the nitrate ion by a cation, and for differentiating between electrostatic and increasingly covalent interaction^:'^.'^ ratios <2 were considered to indi- cate an electrostatic interaction and >2, an increasingly cova- lent interaction. These ratios were 0.66 and 0.69 for the 2 and 10 mol dm-, LiNO, solutions, and 0.68, 0.77 and 0.74 for the 0.5, 3 and 5 mol dm-, Ca(NO,), solutions, respectively. These ratios indicate, according to the abovementioned criterion, a purely electrostatic interaction between the cation and anion [we note that for the Ca(NO,), solution, our v1/v2 intensity ratio is lower than that of 1.45 reported in ref.14 and attrib- ute this to an increased accuracy of the FTIR instrument]. For the separability of component bands in the v2 band region by curve-fitting, separation of peak frequencies of the band components is, for a given FWHH and similar intens- ities, the most important factor. This separation is 3.3 cm-' for 10 mol dm-, LiNO, solution and it increases to 4.2 and 5.2 cm-' for the 0.5 and 3 mol dmP3 Ca(NO,), solutions (Tables 1 and 2). Further literature values for separation of peak maxima are listed with increasing separation: ca. 4 cm- ' for aqueous Mg(NO,), solution (see Table 2 in ref. 8 for the difference between free nitrate and monodentate nitrate), 7-9 cm-' for AgNO, in D20 (from Fig. 7 in ref.9), 8 cm-' for 52 mol dmP3 AgNO, in CH,CN (from Fig. 5 in ref. lo), ca. 10 cm-' for aqueous Ce(NO,), solution (IR bands at 830 and 820 cm-' in ref. 7), 11 cm-' for Zn(NO,), in anhydrous methan01,~ and ca. 12 cm-' for aqueous Gd(NO,), solution (IR bands at 830 and 818 cm- ' in ref. 5). This comparison neglects the influence of the solvent. However, the observed increasing separation is consistent with the expected pertur- bation of the anion by the cation in that it has a minimal value for the alkali-metal cation, increases slightly for the alkaline-earth-metal cations and has a maximal value for the strongly perturbing'gadolinium cation. The molar intensities of the component bands in the IR spectra of the v2 band region are discussed next and compared with those of the v4 band region of the Raman spectra. For the v2 band region in the FTIR spectrum of 10 rnol dmP3 LiNO, solution, the relative band areas due to free and bound nitrate are 64 and 36% (Table 1).These values are remarkably close to those obtained by Riddell et aL2' for free and bound nitrate from the Raman spectrum of the v4 band region in 10 mol dm-, LiNO, solution in D,O (67 and 33%). Careful Raman spectroscopic studies by Irish and co-workers (reviewed in ref. 1) have shown effectively equal molar inten- sities for bound and free nitrate in the v4 band region. It could therefore be concluded that in the v2 IR band region of 10 mol dm-, LiNO, solution, the molar intensities of the bound and free nitrate are also effectively equal.However, in the Raman spectrum of the v4 band region of 10 mol dm-, LiNO, solu- J. Chem. SOC.,Faraday Trans., 1996, Vof.92 27 tion in H,O, the relative band areas due to bound and free nitrate differ from those obtained in D20 (H20: 20% bound, 80% free nitrate, from Fig. 9 in ref. 16; D,O: 33% bound, 67% free nitrate, from Fig. 7 in ref. 20). This difference between relative band areas in either H20 or D20 is also seen in Fig. 8 of ref. 16, and it is either due to a surprising isotope effect of the solvent, or to an error in curve-fitting of the over- lapping bands. The 6 values for the two overlapping bands in the v4 band region of the Raman spectrum of 10 mol dm-j LiNO, in H20 and D20 are ca.1.4 (from Fig. 8 in ref. 16) and, in line with the resolution limit of 6 > 2 mentioned they suggest a large error in the curve-fitting. In the case of an isotope effect of the solvent, the arrangement of solvent molecules around the anion and its influence on the spectrum of nitrate might be of importance, in analogy to the solvation of lithium thiocyanate in methanol reported by Bachelin et ~1.~'Differences of properties of H20 and D20,of the hydration behaviour of ions in H20 and D,O and their spectral features are reviewed by C~nway.~ ' From FTIR spectra of 0.5, 3 and 5 mol dm-3 Ca(NO,), solution, relative areas of 89, 34 and 32% were obtained for free nitrate in the v2 band region, but for the v4 band region of the Raman spectra, the corresponding values are 94, 58.5 and 47.6% (Table 2).Therefore, the ratio of molar intensities for free and bound nitrate, Jf/Jb, must be different for the v2 band region (FTIR) and the v4 band region (Raman). Effectively equal molar intensities for free and bound nitrate, i.e. Jf/Jb = 1, have in general been observed for the v4 band region of the Raman spectrum.' For Jf/Jb = 1, the Jf/Jb ratio for the v, band region of the FTIR spectrum can be calculated. This gives for the 0.5, 3 and 5 mol dm-3 Ca(NO,), solutions, Jf/J, ratios of 0.52, 0.37 and 0.52, respectively. Therefore, the Jf/Jb ratio is fairly constant over a wide range of concentrations and its increased use for quantitative studies of contact ion pairing seems reasonable.For the composite band in the FTIR spectrum of the v2 band region, the S/N ratio is higher by one to two orders of magnitude than that of the band in the v4 band region of the same solution. This high S/N ratio is a prerequisite for using fourth-derivative curves in the manner shown here. In addi- tion, for FTIR spectra of aqueous D,O solutions, subtraction of the background is much less ambiguous for the v2 than for the v4 band region. We see the following advantages in using the v, band region of nitrate, when obtained from FTIR spectra with a very high S/N ratio, as an indication of the contact ion pairing in aqueous solution of alkali metal, alkaline-earth metal and other nitrates in addition to the well known v4 band region.(i) For the two systems studied here, the v, band is the only non-degenerate mode where contact ion pairing can be clearly detected. For the degenerate v4 band, which is at ca. 717 cm-' for solvated nitrate, observ- ation of a second band at ca. 740 cm-' can in principle indi- cate either a lifting of the degeneracy by perturbation of the anion and lowering of its symmetry or, if the degeneracy is not removed by complex formation, a new species such as contact ion pairs. Irish and Brooker' 'pointed out that all evidence to date indicates that aqueous complex formation does not remove the degeneracy of the v4(e') mode as has been suggested'. They further point out that 'occasionally, more than one component can be resolved in the 740 cm-' region but this is believed to be due to complexes containing different numbers of nitrate groups rather than the lifting of the degen- eracy of an e' fundamental of one group'.A parallel study of the non-degenerate V, band region for the number of separate components can resolve such ambiguities. (ii) For 5 mol dmP3 Ca(NO,), solution, a third component band could be resolved in the v2 band region but not in the v4 band region (Table 2 and Fig. 3 and 4). Further studies are required to see if enhanced resolution in the v, band region is a general effect, but the result is encouraging enough to continue these studies. 28 J. Chem. SOC., Faraday Trans., 1996, Vol.92 Resolved component bands in the v2 band region do not have the ambiguity mentioned in (i). (iii) The molar intensities of free and bound nitrate in the v2 band region of FTIR spectra are discussed next. Irish and Brooker' mention that appar- ently 'the vl, v3 and v4 infrared regions are unsuitable for quantitative work'. In particular, the transition dipole of the nitrate v4 normal mode is extremely cation sensitive, and a large increase in intensity is observed for increasing values of z/r for the alkali-metal and alkaline-earth-metal solids.' Therefore, stability constants of metal nitrate systems were obtained mainly by using the nitrate Raman spectra in the v4 and v1 band regions. The correspondence, for 10 mol dm-3 LiNO, solution in D20, of relative band areas for free and bound nitrate in the FTIR v2 band region with that of the Raman v4 band region suggests that the v2 band region could be used to a larger extent for quantitative studies by FTIR spectroscopy than it has so far been.This is consistent with a study of Zn(NO,), in methanol in the v2 band region of the IR spectrum, when the same relative molar absorbance has been reported for the band due to free and contact ion-paired nitrate.4 We are grateful for financial support by the Forschungsforderungsfonds of Austria (project P10404-PHY), and to Prof. D. E. Irish for providing data from ref. 19 and for discussions. References 1 D. E. Irish and M. H. Brooker, in Advances in Infrared and Raman Spectroscopy, ed.R. J. H. Clark and R. E. Hester, Heyden, London, 1976, vol. 2, ch. 6. 2 J. P. Devlin, Vib. Spectra Struct., 1987, 16, 73. 3 D. E. Irish and T. Ozeki, Anal. Raman Spectrosc., 1991,114, 59. 4 S. A. Al-Baldawi, M. H. Brooker, T. E. Gough and D. E. Irish, Can. J. Chem., 1970,48,1202. 5 D. L. Nelson and D. E. Irish, J. Chem. Phys., 1971,54,4479. 6 A. S. C. Cheung and D. E. Irish, J. Inorg. Nucl. Chem., 1981, 43, 1383. 7 D. L. Nelson and D. E. Irish, J. Chem. SOC., Faraday Trans. I, 1973,69, 156. 8 T. G. Chang and D. E. Irish, J. Phys. Chem., 1973,77,52. 9 T. G. Chang and D. E. Irish, J. Solution Chem., 1974,3, 175. 10 T. G. Chang and D. E. Irish, J. Solution Chem., 1974,3, 161. 11 G. Fleissner, A. Hallbrucker and E. Mayer, J. Phys. Chem., 1993, 97, 4806. 12 G. Fleissner, A. Hallbrucker and E. Mayer, Chem. Phys. Lett., 1994, 218, 93. 13 R. E. Hester and R. A. Plane, J. Chem. Phys., 1964,40,411. 14 D. E. Irish and G. E. Walrafen, J. Chem. Phys., 1967,46, 378. 15 R. E. Hester and K. Krishnan, J. Chem. Phys., 1967,46, 3405. 16 D. E. Irish and A. R. Davis, Can. J. Chem., 1968,46,943. 17 D. E. Irish, A. R. Davis and R. A. Plane, J. Chem. Phys., 1969,50, 2262. 18 D. E. Irish, D. L. Nelson and M. H. Brooker, J. Chem. Phys., 1971,54,654. 19 Yu-K. Sze, PhD Thesis, University of Waterloo, 1970. 20 J. D. Riddell, D. J. Lockwood and D. E. Irish, Can. J. Chem., 1972,50,2951. 21 P. D. Spohn and T. B. Brill, J. Phys. Chem., 1989,93,6224. 22 D. W. James and R. L. Frost, Aust. J. Chem., 1982,35, 1793. 23 P. Gans, Data Fitting in the Chemical Sciences, Wiley, 1992, p. 185. 24 W. I. Friesen and K. H. Michaelian, Appl. Spectrosc., 1991,45, 50. 25 Galactic Industry Corp., Users Guide, 1990. Curve fit, Section 3, Chapter 2. 26 Spectrum Square Associates, Users Guide, 1992, Data fit, Maximum Likelihood Peak fitting. 27 B. G. M. Vandeginste and L. De Galan, Anal. Chem., 1975, 47, 2 124. 28 W. F. Maddams, Appl. Spectrosc., 1980,34,245. 29 P. Gans and J. B. Gill, Anal. Chem., 1980,52,351. 30 M. Bachelin, P. Gans and J. B. Gill, J. Chem. SOC., Faraday Trans., 1992,88,3327. 31 B. E. Conway, Ionic Hydration in Chemistry and Biophysics, Else-vier, Amsterdam, 1981, ch. 26, p. 551. Paper 5/03040D; Received 12th May, 1995
ISSN:0956-5000
DOI:10.1039/FT9969200023
出版商:RSC
年代:1996
数据来源: RSC
|
6. |
Cyclodextrin effects on intramolecular charge transfer of 2-biphenylcarboxylic acid: a pre-twisted molecule |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 92,
Issue 1,
1996,
Page 29-33
Dae Won Cho,
Preview
|
PDF (727KB)
|
|
摘要:
Cyclodextrin effects on intramolecular charge transfer of 2-biphenylcarboxylic acid :A pre-twisted molecule Dae Won Cho,Y Yong Hee Kim," Seong Gwan Kang," Minjoong Yoon*" and Dongho Kim*b " Department of Chemistry, Chungnam National University, Taejon 305-764, Korea Spectroscopy Laboratory, Korea Research Institute of Standards and Science, Taejon 305-606, Korea Intramolecular charge transfer (ICT) of a pre-twisted 2-biphenylcarboxylic acid (2BPCA) in aqueous cyclodextrin (CD) solution has been studied by using the steady-state and time-resolved fluorescence techniques, and it has been demonstrated that the ICT interaction is accompanied by a further twist of the biphenyl moiety in order to be orthogonal in the excited state. The ICT emission of 2BPCA at 390 nm in aqueous solution is quenched upon addition of a-or p-CD,$ followed by an enhancement of a new emission at 330 nm.In parallel with this phenomenon, the 330 nm fluorescence decay is resolved into two decay components upon addition of CD in contrast to the single exponential decay of the ICT emission. These results, and AM1 calculations, suggest that the ICT of 2BPCA in a CD solution is inhibited by a restraint on the further twist of the biphenyl moiety of the photoexcited 2BPCA in the CD cavity. The restraint of the conformational change is due to the reduced intermolecular hydrogen bonding of 2BPCA compared with when in water, as well as to the reduced polarity of the CD cavity. Photo-induced intramolecular charge transfer (ICT) is cur- rently one of the most attractive topics of interest as a primary function for photoelectronic devices,' as well as a basic mechanism of biological- and chemical-energy con~ersion.~-~ The photo-induced charge separation is maximized when an electron donor and an acceptor group are mutually perpen- dicular (in order to minimize electronic coupling between the two groups), which results in the formation of a twisted intra- molecular charge transfer (TICT) state.Indeed in some mol- ecules, such as dimethylamino-benzonitrile derivatives, formation of the excited TICT state followed by solvent relax- ation is important in determining the overall charge-transfer Thus, the ICT rate depends significantly on the ground-state twist angle of the rotating groups.7*" On the other hand, in some molecules having a perpendicular con-figuration in the ground state (e.g.9,9'-bianthryl), the twisting relaxation needed to reach the TICT state is not necessary, and the overall ICT rate is determined only by the solvent relaxation dynamics.' ',12 In this case, the solvent polarity has been reported to be the most important parameter. In fact, it has been dem~nstrated'~.'~ that with an increase in the polarity of the media the energy barrier for the TICT (ICT) process decreases, causing an increase in the solvent relax- ation time, and consequently the TICT (ICT) rate. However, some authors have suggested that the specific hydrogen bonding between the solvent and the molecules is also impor- tant, in order to maintain a large twist angle between the elec- tron donor and the acceptor group.' 2,1 Moreover, the charge separation in biological molecules is known to be coupled to proton motion, which may be a proton-transfer process.l6 Nevertheless, the role of the specific hydrogen bonding in the photo-induced TICT (ICT) has not been systematically inves- tigated. Thus, the hydrogen-bonding effect of solvent on TICT (ICT) would be an interesting subject to explore with regard to the proton-transfer coupled charge transfer. Recently, we have studied solvent dependences of the photophysical properties of 2-biphenylcarboxylic acid (2BPCA), which has a pre-twisted biphenyl moiety in the t Present Address : Department of Chemistry, Seonam University Kwangchi-Dong, Namwon, Chunbuk 590-170, Korea.$ a-CD :cyclomaltohexaose. b-CD : cyclomaltoheptaose. ground state.17 This study has shown that the ICT interaction takes place upon excitation of 2BPCA as demonstrated by a large Stokes-shifted fluorescence emission in polar solvents. In hydrogen-bonding solvents especially, the ICT emission is further shifted into the red. This has been attributed to a lowering of the torsional energy barrier of the biphenyl moiety by the increased steric hindrance between benzoic acid and the phenyl group due to the excited-state hydrogen-bonding interaction, as in the case of TICT molecules such as p-(dialky1amino)benzonitriles.' The hydrogen bonding of the carboxyl group would be facilitated by the resonance contri- bution from the ICT interaction in the excited state." In our continuing efforts to explore the hydrogen-bonding effects on the ICT process of 2BPCA in the excited state, we have studied the effects of cyclodextrin (CD) on the ICT fluores- cence properties in aqueous solution.The CD cavity provides an incorporated guest molecule with a non-polar and restrictive environment."-23 Thus, if 2BPCA is imbedded in the inner cavity of CD, the excited-state hydrogen bonding and geometry change will be greatly affected. In this paper, the formation of a 2BPCA-CD inclusion complex and the location of the molecule inside the cavity were reported, with regard to the excited-state ICT and geometry change. Experimental 2BPCA was purchased from Aldrich Chemical Co.and puri- fied by recrystallization with ethanol. The melting point of 2BPCA agrees well with the reference value (1 14 0C).24a-and p-CD were purchased from Aldrich Chemical Co. and used without further purification. KBr was obtained from BDH Chemical Ltd. as spectroscopic grade. The stock solutions of 2BPCA and CDs were prepared in HAc-NaAc buffer solu- tions at pH 3.0, where 2BPCA remains as a neutral species (pKa, ca. 3.3.'' The same volume of 2BPCA stock solution was added to different volumes of CD stock solution, and move buffer solution was added to keep the total volume con- stant, so that the concentration of 2BPCA remained constant (4 x mol 1-') yet the CD concentration was varied.This concentration is low enough to avoid dimerization of the car- boxylic acids. Water was triply distilled in the presence of acidic dichromate and alkaline permanganate. All the sample J. Chem. SOC.,Faraday Trans., 1996,92(1),29-33 solutions were degassed by a freeze-pumpthaw technique before the spectral measurements were taken. Fluorescence spectra were measured on a scanning SLM- AMINCO 4800 spectrofluorimeter, which makes it possible to obtain corrected spectra using Rhodamine B as a quantum counter. The absorbance at the excitation wavelength was held constant when different solutions were compared. Fluo- rescence quantum yields were determined by comparison with a reference of known quantum yield [indole in ethanol, (0.32) or quinine bi~ulfate].'~ Fluorescence lifetimes of 2BPCA in CD solutions were measured by a time-correlated single photon counting (TCSPC) method, using a dual-jet dye laser (coherent; Model 702) synchronously pumped by a mode-locked Ar ion laser (coherent; Innova 200) as described in a previous paper.22 The cavity-dumped beam from the dye laser has 1 ps pulse width, average power ca.100 mW at 3.8 MHz dumping rate, and a tunability of 560-620 nm with Rho- damine 6G as gain dye and DODCI (diethoxydicyanine iodide) as the saturable absorber. To excite the sample, the dye laser pulse was frequency-doubled using a P-BBO (p-barium borate) crystal. All the standard electronics used for the TCSPC were from EG & G Ortec. This method allows a time resolution of about 20 ps after deconvolution.Results and Discussion The absorption spectra of 2BPCA in CD solution (pH 3) show little change with respect to those in water as observed for the solvent polarity dependences' reflecting little charge-transfer interaction in the ground state. In contrast to the weak CD dependence of absorption spectra, the fluorescence emission spectra of 2BPCA (4 x lop5 mol 1-') in aqueous solution show significant changes upon addition of 0-CD, as depicted in Fig. 1. Upon addition of p-CD, a largely Stokes-shifted ICT emission at 390 nm in aqueous solution exhibits a marked quenching along with a blue-shift of the emission maximum at 390 nm to 375 nm as the CD concentration increases. The a-CD concentration dependence of the fluorescence spectral changes is similar to that observed for the p-CD solution except for the lower quenching rate of the ICT emission (data not shown).These observations imply a partial inhibition of the ICT-state formation in the reduced polar environment of the CD cavity. The fluorescence spectral changes also indicate that the excited state of 2BPCA is more sensitive to the CD microenvironment than the ground state, supporting the for- be analysed by the following Benesi-Hildebrand relati~n.~~,~' 1 1 1 -(2)If" -I, If" -I; + K(If"-Z;)[CD] where K is the formation constant, If" the initial fluorescence intensity of free 2BPCA at 390 nm, I; the fluorescence inten- sity of the inclusion complex and I, the observed fluorescence intensity at 390 nm.A plot of l/(If"-I,) us. l/[CD] gives a straight line, as shown in Fig. 2. This linear correlation not only provides evidence for the 1 :1 complex formation but also permits the calculation of the binding constant from the slope. The Ks obtained for a-and p-CD are 60 and 170 1 mol-', respectively. The quantitative comparison of binding constants of the two complexes suggests that p-CD provides a better site to accommodate a deep inclusion of 2BPCA in the CD cavity. This is consistent with the observ- ation that the collisional quenching of ICT emission in aqueous solution by KBr is less effective in the presence of p-CD than in a-CD; the Stern-Volmer quenching constants are 4.9 1 mol-' in CD-free aqueous solution, 3.6 1 mol-' in a-CD and 1.6 1 mol-' in p-CD (Fig.3). Note that the CD-induced quenching of the ICT emission is 1.51 / 1 i -I 0.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 10-2[CD]-'/I mot-' Fig. 2 Benesi-Hildebrand plot for fluorescence quantum yield of 2BPCA in aqueous solutions containing various concentrations of or p-CD (-0-0-) (-.-.-)a-CD mation of a 1 : 1 inclusion complex between 2BPCA and CD. 2.0 h 2BPCA + CD 2BPCA-CD The CD dependence on the ICT fluorescence quenching can 1.5 s 0' 4r v \7 1.o 0.5 t',',.,.,', 290 340 390 440 490 2 4 6 81012 wavelength/nrn 102[Br7/mol I-' Fig. 1 Fluorescence emission spectra of 2BPCA (4.0 x 10-mmol 1 I) Fig. 3 Stern-Volmer plots for fluorescence quenching KBr in the absence and presence of CDs.The symbols (-O-O-),(-.-.-) in aqueous solutions containing various concentrations of p-CD at room temperature. [p-CD] = (a) 0, (b) 4, (c) 8, (d) 12, (e) 16, (f) 20 and (-O-O-) denote the CD-free aqueous solution, 12 mmol 1-' a-CD mmol 1-'. solution and 12 mmol 1-' p-CD solution, respectively. 30 J. Chem. SOC.,Faraday Trans., 1996, Vol. 92 accompanied by the appearance of a new band at 320 nm with increasing CD concentration. The 320 nm emission could be due to an anionic species because it is weakly observed in the highly alkaline aqueous solution.” However, this idea can be ruled out by the following facts. First, we used acidic solu- tion which has a lower pH (3.0) than the pKa value of 2BPCA.Secondly, CD is known to increase the pK, of organic acids.” Thirdly, the quantum yield of the anionic emission is too low to be measured. Thus, it is quite certain that 2BPCA exists as a neutral species, and the new emission is attributable to an unrelaxed normal excited state. The normal emission is not usually observed in homogeneous solvents, probably because of the fast ICT in the excited state. In aqueous solution, the excited-state ICT process was proposed to be accompanied by a further twist of the pre-twisted biphenyl moiety, with a dihe- dral angle of about 90°, via the intermolecular hydrogen bonding of the ortho carboxyl group with water, and the ICT is more associated with water than with any aprotic sol- vents.’ However, the proposed relaxation process for the twisted ICT state of the 2BPCA-CD complex would not be feasible, because of the non-polar and restrictive environment of the CD cavity.Therefore, the locally excited Franck- Condon state becomes longer-lived and can exhibit the normal emission. Based on the above assignment of the dual emission, the photophysical processes of excited-state 2BPCA* in the absence and presence of CD can be summarized as follows: In the absence of CD; 2BPCA* -+ 2NPCA*(ICT) (1) 2BPCA* -+ 2BPCA + hv; 42 (2) 2BPCA* -,2BPCA 2BPCA*(ICT) +2BPCA(ICT) + ~vICT;43 (3) 2BPCA*(ICT) +2BPCA(ICT) In the presence of CD; 2BPCA*-CD -+ 2BPCA*(ICT)-CD (4) 2BPCA*-CD +2BPCA-CD + hv’; 45 (5) 2BPCA*-CD -+ 2BPCA-CD 2BPCA*(ICT)-CD -+ 2BPCA(ICT)-CD + h&T; 46 (6) 2BPCA*(ICT)-CD -+ 2BPCA(ICT)-CD Processes (2) and (5) correspond to the normal emission whereas processes (3) and (6)correspond to the ICT emission.The quantum yield of each process is denoted by $i. Here, it is assumed that in the excited state, neither dissociation nor association reactions between 2BPCA and CD occurs, because the lifetimes of 2BPCA* and 2BPCA* (ICT) are of the order of s. The observed quantum yields of the normal and ICT emissions are expressed by @normal = (42‘a ca + 45 ‘b cb)/(Ea ca + ‘b ‘b) (7) @ICT = (43ca + 46 ‘b cb)/(Ea ca + ‘b cb) (8) where C and E are the concentration and molar absorptivity at the excitation wavelength, respectively, for free (subscript a) and CD inclusion complex (subscript b).The ratio of Qjnormal to QICT, R, is written as R = (42 i-Y~s[CDIK)/(~~+ Y46CCDIK) (9) where y = Eb/Ea, and K = cb/c, [CD] (a formation constant). Eqn. (9) can be transformed into eqn. (lo),assuming y = 1, -RO/[CD1 = [45/43 -‘46/43IK (10) where Ro = 42/43, which is almost zero in the absence of CD. According to eqn. (lo), a plot of (R -R,)/[CD] us. R (Fig. 4) 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 R Fig. 4 Plot of (R-R,)/[CD] us. R. See text for the meaning of each letter: a-CD (-a-a-),p-CD(-O-O-) gives a slope of K46/43 and an intercept of K45/43,from which 45/46 is estimated to be 1.21 and 1.13 in a-CD and p-CD, respectively. The overall fluorescence quantum yield was found to be independent of the CD concentration.Thus, the sum of quantum yields in the absence of CD, (42+ 43), is the same as that in the presence of CD, (45+ 46),and the following equation is derived. 43/46 = 45/(46 + ’) (11) Substitution of the value of 45/46 into eqn. (11) gives 46/43 as 0.45 and 0.47 for a- and p-CD systems, respectively. From these values and the slope in Fig. 4, the Ks of the a-and p-CD inclusion complexes are evaluated to be 62 and 122 1 mol-’, respectively. These formation constants are close to those obtained from Benesi-Hildebrand plots, confirming our inter- pretation of the appearance of dual emission in the presence of CD. In accordance with the steady-state spectral changes, decay kinetics of the dual emission are also affected by the addition of CD.Fig. 5 shows typical decay profiles of the ICT emission (390 nm) of 2BPCA in the absence and presence of 16 mmol 1-’ a- and p-CD. The ICT emission monitored at 390 nm fits a single exponential decay even in the presence of CD, as observed in aqueous solution. The decay time of excited 2BPCA in a-CD is analysed to be ca. 840 ps, which is similar to that measured in CD-free aqueous solution (Table I), whereas the decay time of excited 2BPCA in p-CD is shorter. This may be due to the different complexation pattern of 2BPCA with different CDs. Considering that the decay time of the ICT emission becomes shorter in less polar solvents2 2BPCA in p-CD seems to be more deeply located in the non- tirnehs Fig.5 Fluorescence decay of 2BPCA in the absence and presence of CDs monitored at 390 nm at room temperature. (A), (B) and (C) denote the CD-free aqueous solution, 20 mmol I-’ a- and p-CD solu- tion, respectively. J. Chem. SOC.,Faraday Trans., 1996, Vol. 92 31 Table 1 Fluorescence lifetimes of 2BPCA in aqueous and CD solu-tions= fluorescence lifetimes (ps) a-CD P-CD [CD]/mmol 1-' 330 nm 390 nm 330 nm 390 nm 0 -820 (1) -820 (1) -w -4w 4 940 (0.89)b 840 (1) 1050 (0.80) 720 (1) 4430 (0.11) W 5600 (0.20) 7w 1050 (0.92) 830 (1) 910 (0.73) 690 (1)8 5620 (0.08) 4W 6130 (0.27) 5W 12 1040 (0.88) 840 (1) 1140 (0.73) 680 (1) 5520 (0.12) 4W 6330 (0.27) 8V Measurement error limits are less than 5% of the listed values.Values in the parentheses are the amplitudes of the lifetime com- ponenkc Rise times. polar cavity, as compared with that in a-CD. This is consis- tent with the observation that the collisional quenching of the ICT emission in the aqueous solution by KBr is much more reduced in the presence of p-CD than in the presence of a-CD. Note also that the rise times of ICT emission (40-80 ps) were observed (Table l), which are similar to the rotational relaxation time of the biphenyl moiety as observed for 4BPCA in cyclohexane.17 This reflects the twisting dynamics of the biphenyl moiety to form the ICT state. The rise time increases from ca. 40 to ca. 80 ps as the concentration of p-CD increases, while it is independent of the concentration of a-CD.This can be attributed to the higher restraint of the molecular motion imposed by the p-CD complex relative to the a-CD one. Previously, it has been observed that the twist of the biphenyl moiety in 4BPCA is possible in the CD cavity.22 Thus the restraint of the biphenyl twist should be contributed to by the entrapment of both the biphenyl moiety and carboxyl group in the p-CD cavity. Consequently, the intermolecular hydrogen bonding between the carboxyl group and water is significantly reduced and this inhibits a further twisting of the biphenyl moiety as previously proposed.17 On the contrary, it would be relatively difficult for the carboxyl group to be entrapped entirely in the a-CD cavity, which has a smaller radius than the p-CD cavity.Therefore, intermolecu- lar hydrogen bonding is not so much affected by a-CD as it is in the case of the p-CD complex. This should be the reason why the rise time of ICT emission is changed little in the pres- ence of a-CD. In contrast to the single-exponential decay of the ICT emis- sion, the normal emission decays biexponentially, being resolved into a long decay component (ca. 5-6 ns) and a short decay one (ca. 1 ns) (Table 1). The biexponential decay should be attributable to the existence of two different 1 : 1 inclusion complex species. The short component should originate from a type I complex, the long component from a type I1 complex, as shown in Fig. 6. In the type I complex, 2BPCA molecules are more or less loosely located near the exit of the hydropho- bic lotus of CD, while in the type I1 complex, 2BPCA mol- ecules are deeply bound in CD as in the case of the trans-stilbene-CD complex.28 Since the carboxyl group can be rotated, there may exist two conformers as shown below (Scheme 1).A B Scheme 1 A t-5.7 A+ 7.8A --c( type I t type II @"-=. a,-CD P-CD B C Fig. 6 Optimized geometry of 2BPCA (A), the proposed structures of the 2BPCA-a-CD (B) and 2BPCA-P-CD complexes (C) According to AM1 calculation, 2BPCA mainly exists in conformation A, which facilitates the intermolecular hydrogen bonding and consequently the formation of the ICT state upon excitation (see below). Thus, it can be proposed that conformation A is dominant in the type I complex.However, if 2BPCA is located deep inside the CD cavity as in the case of the type I1 complex, the intermolecular hydrogen bonding would be inhibited. Under this circumstance, 2BPCA may exist in a small percentage in conformation B,in which intra- molecular hydrogen bonding is possible between the carboxyl group and an ortho-benzene ring. It is well known that hydroxyl groups interact weakly with the n-electrons of benzene to form intramolecular hydrogen bonds, as reported for 2-hydro~ybiphenyl.~~ Consequently, the type I1 inclusion complex forces the structure of 2BPCA to be rigid, so that the twist of the biphenyl moiety is further inhibited. Thus, the twisted ICT relaxation or the vibrational relaxation is even more inhibited in the type I1 complex than in the type I complex.This is the reason why the long decay time for normal emission can be observed. The formation of the type I1 complex seems to be more feasible in the presence of p-CD than in a-CD, based on the observation that the relative amplitude of the long decay component is much larger in p-CD than in a-CD (Table 1). This is because the chance of 2BPCA existing in conformation B is much higher in p-CD than in a-CD. The differences in the complexation behaviour of 2BPCA with a-and D-CDs could be rationalized by a theoretical AM1 calculation. The optimized geometry parameters of 2BPCA in the ground state are summarized in Table 2. The distance between atoms 17 and 22, which represents the long axis of the molecule, is 9.13 8, [see Fig.6(A)]. This is much larger 32 J. Chern. SOC., Faraday Trans., 1996, Vol. 92 Table 2 Theoretically calculated geometrical parameters for the optimized ground state of 2BPCA (for atom numbering, see Fig. 6) bond length /A bond angle /degrees twist (torsional angle) /degrees 1-2 1.404 1-2-3 120.139 2-1-7-12 63.39 2-3 1.403 2-3-4 120.397 6-1-7-8 62.07 3-4 1.391 3-4-5 119.755 7-1-2-13 5.09 4-5 1.396 4-5-6 120.085 1-2-13-15 45.01 5-6 1.395 6-1-2 120.889 3-2-13-14 43.88 1-6 1.405 2-1-6 118.704 1-7 1.465 12-7-8 119.503 7-8 1.402 7-8-9 120.143 8-9 1.393 8-9-10 120.154 9-10 1.395 9-10-1 1 119.902 10-11 1.395 10-1 1-12 120.206 11-12 1.394 11-12-7 120.09 7-12 1.401 1-2-13 122.156 2-13 1.427 2-1 3-1 5 130.136 13-14 1.369 14-13-1 5 115.947 13-15 1.234 1 3-14-25 109.248 14-25 0.971 1-17-12 120.729 3-16 1.101 2-1-7 123.1 16 4-17 1.100 2-3-16 119.160 5-18 1.100 3-4-17 120.004 6-19 1.101 4-5-1 8 120.021 8-20 1.100 5-6-19 120.122 9-21 1.100 7-8-20 119.689 10-22 1.100 8-9-2 1 119.794 11-23 1.100 9-10-22 120.034 12-24 1.100 10-1 1-23 120.01 6 11-12-24 120.048 The twist angle is almost zero except for the shown data.than the cavity diameters of a-and P-CDs. Thus the inclusion of the molecule can only be accomplished along the long axis, as shown in Fig. qB). The distances between atoms 20 and 15 and atoms 19 and 25 are 5.47 and 5.94 A,respectively.Since the cavity diameter of a-CD is 5.7 8, and shorter than the distance between atoms 19 and 25, a major portion of the molecule has to lie inside the a-CD cavity in order to form the type I complex as discussed above [Fig. 6(B)], even though the type I1 complex exists to a small extent. The situation is, however, somewhat different for the p-CD system, which has a cavity diameter of 7.8 8,. Considering the molecular dimension of 2BPCA and the cavity size of p-CD, it would be more plausible for 2BPCA to be encapsulated inside the 0-CD cavity more deeply, compared with the a-CD system [Fig. 6(C)]. Thus, the chance of forming the type I1 complex would be higher in the p-CD system than in the a-CD system.This is again consistent with the observation that the long decay component of the normal emission originating from the type I1 complex has larger amplitudes in p-CD than in a-CD (Table 1). Conclusion This work has demonstrated that the anomalously Stokes- shifted fluorescence of 2BPCA in aqueous solution is attribut- able to the ICT interaction accompanied by a further twist of the biphenyl moiety in order to be orthogonal in the excited state. However, the excited-state ICT process of 2BPCA is inhibited upon addition of CD. This is due to the molecular entrapment in the CD cavity which places a restraint on the twist of the biphenyl moiety by reducing the amount of inter-molecular hydrogen bonding between the molecule and water.This work has been supported by the Korea Science and Engineering Foundation through Project # 94-1400-05-01-3, the Center for Molecular Catalysis (M. Y.) and the Center for Molecular Science and MOST (D. K.). References 1 (a) J. P. Launay, M. Sowinska, L. Leydier, A. Gourdon, E. Amouyal, M. L. Boillot, F. Heisel and J. A. Miehe, Chem. Phys. Lett., 1989, 160, 89; (b) S. Nagakura, in Excited States, ed. E. C. Lim, Academic Press, New York, 1975, vol. 2, pp. 322-378. 2 J. W. Verhoeven, Pure Appl. Chem., 1990,62,1585. 3 J. S. Connolly and J. R. Bolton, Photoinduced Electron Transfer, ed. M. A. Fox, and M. Chanon, Elsevier, Amsterdam, 1990, part D, ch. 6.2. 4 L. R. Khundkar, A. E. Stiegman and J. W. Perry, J.Phys. Chem., 1990,94, 1224. 5 H. Lueck, M. Windsor and W. Rettig, J. Phys. Chem., 1990, 94, 4550. 6 F. Heisel, J. A. Miehe and J. M. G. Martinho, Chem. Phys., 1985, 98, 243. 7 W. Rettig and R. Gleiter, J. Phys. Chem., 1985,89,4676. 8 S. R. Meech and D. Philips, J. Chem. SOC., Faraday Trans. 2, 1987,83, 1941. 9 S. G. Su and J. I>. Simon, J. Chem. Phys., 1988,89,908. 10 S. G. Su and J. D. Simon, J. Phys. Chem., 1989,93,753. 11 N. Mataga, H. Yao, T. Okada and W. Rettig, J. Phys. Chem., 1989,93,3383. 12 C. Cazeau-Dubroca, S. A. Lyazidi, P. Cambou, A. Peirigua, Ph. Cazeau and M. Pesquer, J. Phys. Chem., 1989,93,2347. 13 J. Hicks, M. Vandersall, Z. Babarogic and K. E. Eisenthal, Chem. Phys. Lett., 1985, 116, 18. 14 J. D. Simon and S.G. Su, J. Phys. Chem., 1988,92,2395. 15 A. Levy, D. Avnir and M. Ottolenghi, Chem. Phys. Lett., 1985, 121, 233. 16 R. L. Cukier, J. Phys. Chem., 1994, 98, 2377 and references therein. 17 M. Yoon, D. W. Cho, J. Y. Lee, M. Lee and D. Kim, Bull. Kor. Chem. SOC., 1992,13,613. 18 T. C. Werner and R. M. Hoffman, J. Phys. Chem., 1973,77,1611. 19 M. L. Bender and M. Komiyama, Cyclodextrin Chemistry, Springer-Verlag, New York, 1978. 20 J. Szejtli, Cyclodextrin Technology, Kluwer Academic Publishers, Dordrecht, 1988, pp, 143-154. 21 K. Kalyanasundaram, Photochemistry in Microheterogeneous Systems, Academic Press, New York, 1987, ch. 9. 22 D. W. Cho, Y. H. Kim, S. G. Kang, M. Yoon and D. Kim, J. Phys. Chem., 1994,98,558. 23 (a) S. Monti, L. Flamigni, A. Martelli and P. Bortolus, J. Phys. Chem., 1988, 92,4447; (b) R. A. Agbaria, R. Uzan and D. Gill, J. Phys. Chem., 1989, 93, 3855; (c) K. Kasatani, M. Kawasaki and H. Sato, J. Phys. Chem., 1984,88, 5451. 24 Dictionary of Organic Compounds, Chapman and Hall, New York, 1982, 5th edn., vol. 1. 25 J. N. Demas and G. A. Crosby, J. Phys. Chem., 1971,75,991. 26 M. L. Benesi and J. H. Hildebrand, J. Am. Chem. SOC., 1949, 71, 2703. 27 S. Hamai, Bull. Chem. SOC. Jpn., 1982, 55, 2721. 28 G. L. Duveneck, E. V. Sitzmann, K. B. Eisenthal and N. J. Turro, J. Phys. Chem., 1989,93,7166. 29 M. Tichy, Advances in Organic Chemistry: Methods and Results, ed R. A. Raphael, E. C. Taylor and H. Wynberg, Interscience Publishers, Inc., New York, 1965, vol. 5, pp. 115-294. Paper 5104689K; Received 17th July, 1995 J. Chem. SOC., Faraday Trans., 1996, Voi. 92 33
ISSN:0956-5000
DOI:10.1039/FT9969200029
出版商:RSC
年代:1996
数据来源: RSC
|
7. |
NMR study of the influence of urea on the self-association of propan-1-ol in water and comparison with Kirkwood–Buff integral results |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 92,
Issue 1,
1996,
Page 35-40
Antonio Sacco,
Preview
|
PDF (788KB)
|
|
摘要:
NMR study of the influence of urea on the self-association of propan-1-01 in water and comparison with Kirkwood-Buff integral results Antonio Sacco,*i* Antonio Asciolla," Enrico Matteolib and Manfred Holz" a Dipartimento di Chimica, Universita degli Studi di Bari, Via Orabona 4, 70126 Bari, Italy Istituto di Chimica Quantistica ed Energetica Molecolare, CNR, Via Risorgimento 35,56126 Pisa, Italy " Institut f iir Physikalische Chemie der Universitat Karlsruhe, Kaiserstrasse 12,0-76128 Karlsruhe, Germany The self-association of hydrophobic propan-1-01 has been studied in binary aqueous mixtures and in mixtures containing urea as a third component. We applied an NMR technique to obtain the so-called association parameter A,,, which is based on the measurement of intermolecular 'H-'H dipole-dipole relaxation rates and on self-diffusion coefficients of the component of interest.The composition dependence of A,, reflects a weak tendency for the self-association of propan-1-01 in the binary mixture. Addition of urea leads to a strongly increased self-association of propan-1-01. The same qualitative result has been obtained from recent studies of the Kirkwood-Buff integral G,, by one of the present authors. A,, is governed by short-range contributions but G,, by long-range contributions and the common study of the two quantities has delivered interesting comple- mentary information about liquid mixtures. Propanol-propanol correlations in binary systems were never found to extend longer than one to two propanol diameters, however when urea is added the correlation length can extend up to four diameters.Measurements of B, parameters, showing the influence of urea on the translational motion of propan-1-01, revealed only a weak effect. Thus no rigid self-associated species could be detected and the self-association found from A,, and G,, can be interpreted as 'preferential self-solvation ' by relatively weak intermolecular interactions. 1. Introduction Self-association of inert non-polar compounds or substances containing hydrophobic groups such as alkyl groups when they are dissolved in water is a well known phenomenon (hydrophobic interaction or association) and has been exten- sively in~estigated.'-~ Owing to the importance of this pheno- menon in various fields, we have recently applied a method based on NMR spectroscopy, already reported in the liter- at~re,'-~for a deeper investigation of the influence of some electrolytes on the association features of hydrophobic mol- ecules.' This procedure is based on the determination of the so-called Aij parameter (i = 1, 2; j = 1, 2) defined as an inte- gral over an expression containing an atom-pair correlation function gi,.(r);Aij conveys information on the sharpening or flattening of this correlation function if, for example, the com- position of the system under investigation is changed.In a recent paper' we studied, in water alone and in the presence of added salts, the association properties of three solutes having increasing numbers of methyl groups, namely dimethyl sulfoxide, tert-butyl alcohol and 1,lY3,3-tetra-methylurea. In this paper we consider the system water-propan- 1-01, the latter being a characteristic hydrophobic compound.We have analysed not only the binary mixture but also the effect on it upon the addition of urea, a compound which is well known for its influence on the association of proteins. The association of hydrophobic compounds in aqueous solutions has also been confirmed by studies based on the analysis of the behaviour of the Kirkwood-Buff (K-B) inte- grals, Gij,'-ll which are integrals over all space of the molec- ular radial distribution functions, and recently these quantities have also become available for the ternary system water- propan-l-ol-urea.'2 Since both Gij and A, depend in some way on integrals of gij(r), a relationship between them is expected.In this work we also aim to compare and discuss these two quantities in relation to their capability for provid- ing insights into local environment and solute-solute inter-actions in solution. 2. Theory 2.1 Aij parameter As reported el~ewhere~,~~ the Aij parameter is related to the attractive or repulsive interactions between species i and j in the liquid state and is defined as: Aij = (1/T)interb/c (1) In eqn. (l), ( l/Tl)inte, is the intermolecular nuclear magnetic dipole-dipole relaxation rate of interacting nuclei in molecules of type i and j. Usually it is the interactions between 'H nuclei that are observed because these nuclei possess a high value of the gyromagnetic ratio y.D is the mean translational coeff- cient of molecules i andj. If, as in the present system, we are interested only in the interactions between molecules of the same kind, e.g. i, then D is replaced by D, the translational diffusion coefficient of i in the system. c is the number density (spin cmP3) of the interacting spins. In other words, as can be realised from eqn. (l), A, represents a dipole-dipole relaxation rate divided by c and by 1/D, and therefore normalised with respect to the spin concentration and to the effect of trans- lational dynamics. The Aij parameter is related to the atom-atom pair corre- lation function giJ.(r)through the eq~ation:'~ A..= ?!!&IJ 3a4 lm(:)6giJ(r)rz dr where K = 3/2y;h2 for equal spins with I = 1/2; a is the dis- tance of nearest approach of the interacting nuclei.Neglecting the details of the Aij theory, which can be found in previous J. Chem.Soc., Faraday Trans., 1996,92(1),35-40 35 articles,’*’ it is only worth mentioning here that the detection of ‘association’ is made by determining Aijas a function of the concentration. If the Aij values increase with decreasing con- centration of the species examined, we can affirm that there is a sharpening of giXr) and this indicates the presence of attrac- tive forces between particles i and j and therefore of an associ- ation tendency. To determine the A,, parameter according to eqn.(1) for the water (1)-propan-1-01 (2) system, it is necessary to measure the ‘H intermolecular relaxation rate of the methyl (or methylene) groups of propanol brought about by the methyl (or methylene) groups of the nearby propanol mol- ecules. In the usual relaxation rate measurement we obtain the total relaxation, which is given by: (3) To split the total quantity into intra- and inter-molecular con- tributions, we have followed the so-called isotopic dilution experiment.’ 5,16 The translational diffusion coefficient D of propanol has been determined by using the NMR spin-echo pulsed-gradient technique, and for obtaining c we have carried out density measurements. In this way we have been able to determine A,, in the whole composition range.2.2 Kirkwood-Buff integral, Gij The K-B integrals and the A, parameters are closely related :’ 471 [gj;)(r)-l]r2 dr --u3 (4)3 4naZK 1 4.nA,.= -3 (4n lm[giJ(r)-11 -dr + -u-~) (5) IJ r4 3 Note that g$) in eqn. (4) is a centre-centre pair correlation function, whereas giXr) in eqn. (5) is an atom-atom pair corre- lation function. In spite of close analogies, an important difference can be pointed out. Whilst the K-B integral is the result of the inte- gration over [gij(r)-13 r’, with the consequence that long- range contributions to [giJ(r)-11 are increased in their effectiveness, in the Aij expression that term is multiplied by l/r4 and thus the short-range contributions govern this quan- tity. Consequently we can say that Gij and A, are complemen- tary with respect to the influence of long- and short-range interactions.Further, Aij may depend on molecular orienta- tion, whilst Gij cannot. In recent years several authors have used the K-B theory to obtain information on the local environment. In particular, the K-B integrals have been used to calculate quantities such as local compositions.’ 7,1 ’ The basic equation linking the K-B integral with local composition quantities arises from the statistical thermodynamic definition of the radial distribution and is given by : An.. = c.G.. (6)IJ 1 1J where Anij is the excess (or deficit) number of moles of i in the whole space around a mole of central j particles, and c, is the bulk molar concentration of species i.As discussed in previous papers,”*19 K-B integrals and consequently Anij take a non-zero value even for ideal mixtures, so, in order that the quan- tity calculated as the excess number really reflects intermolecular interactions, from Anij, the contribution due to the K-B integral of the corresponding ideal mixture, G;;, must be subtracted. Eqn. (6) is therefore rewritten as: V 1 1J IJAn!.= c.(G..-G!d) ci6G.. (7) Other authors did not consider this correction, and, although it may be negligible when G, is as large as several hundreds of 36 J. Chem. SOC.,Faraday Trans., 1996, VoZ. 92 cm3 mol-’, in many cases it is not so. In ref. 18 and 19 details of the procedure and the working equations to calculate the quantities of eqn.(7) are supplied. Eqn. (7) provides an excess coordination number due to interactions between i and j; unfortunately, the K-B theory does not provide any criterion as to how these extra molecules are distributed in the various solvation shells around the central molecule. If we assume that the correlation length is equal to rno, rn being an unknown variable and o the molecu- lar diameter of propanol, then the corresponding correlation volume, V,,,, and the average coordination number n, in a shell of thickness (rn -l)a are given by: V,,, = N,4~[(rno)~-03]/3 (8) n, = n! + Anij = ciV,,, + Anij (9) where n: is the number of moles of species i contained in a volume of bulk solution equal to the correlation volume. It is clear that the larger the value of rn (since Anij is inde- pendent of rn), the closer is the value of n, to n:.If the n, value is calculated to be larger than the maximum n, allowed in V,,, (i.e.np, with ny = For/v,J( being the molar volume), then the actual correlation length is larger than rno. Note that the choice of the correlation length does not affect the qualitative trends of n, with respect to molar fraction of propanol and to the molar fractions of added urea. As regards the comparison with NMR relaxation results; since this technque is sensitive to short-range correlations, the results of the two methods should be closer the shorter the actual correlation length. 2.3 B, coefficients The viscosity B coefficient calculated from the Jones-Dole equation” (neglecting the term in c’I2) q/qo = 1 + Bc + CcZ+ (10) provides information on the properties of electrolytes and non-electrolytes in solution.Similarly, two other coefficients can be defined,’l B’ and BD, according to the following equa- tions : (1/T’)/(1/7i)c=0 = 1 + B’c 4-c’C2 4-. . * (1 1) = 1 + BDc + *” (12)(1,/D)/(1/DC=O) where (l/Tl)c=oand DC=, correspond to the relaxation rate and the self-diffusion coefficient of the pure solvent, respec- tively. The B’ coefficients yield a microdynamical description of the solution, in that they provide information about the reorientational and translational behaviour of solvent mol- ecules in the coordination sphere of the ions; i.e. positive B’ values correspond to a ‘structure-making’ influence of the solute on the solvent.In contrast, the BD coefficients unam- biguously characterise the influence of the solute on the trans- lational motion of the solvent.” In the case of the ternary mixture water-propan-l-C2H] ol-[’H,] urea, we have determined the BD coefficients from concentration-dependent NMR self-diffusion measurements in order to obtain additional information on the influence that the non-electrolyte [’H,]urea has on the translational proper- ties of pr~pan-l-[~HJol in its binary mixtures with D,O. 3. Experimental Binary mixtures of D,O-pr~pan-[~H]ol and the ternary system containing [’H,]urea at two fixed molar fractions, namely x3 = 0.05 and 0.09, respectively, were investigated. We measured the ‘H spin-lattice relaxation rates, the trans- lational diffusion coefficients D of propan-1-[2H]01 and the densities of all the mixtures.The 'H signals were also used in the low-resolution spin-echo measurements for the determi- nation of the diffusion coeffcients. To avoid 'H signal contri- butions from water and urea, we have utilised D,O and C2H4]urea. For the isotopic dilution experiments we have used the full deuteriated [2H,]propan-l-[2H]ol. The proton spin-lattice relaxation times TI were measured at 200 MHz using a Varian XL-200 instrument. The experimental error in the 1/T, measurements was *2%. The samples were prepared by weighing and paramagnetic oxygen was removed by several freeze-pumpthaw cycles. The self-diffusion coeffcients D were measured by means of the pulsed magnetic-field-gradient spin+cho technique, using a Minispec pc 120 combined with a commercial pulsed-field- gradient unit, both from Bruker.For details of the measure- ments, see ref. 23 and 24. The experimental error in the diffu- sion coefficient measurements amounted to 2%. However, at the lowest pr~pan-l-[~H]ol concentrations, because of the low concentration of protons, we used a high-resolution spectrometer (Tesla, FT, 80 MHz) coupled to a commercial pulse field gradient unit provided by STELAR, Italy. The density measurements were performed with an Anton Paar densime ter. All measurements were performed at a fixed temperature of 25 k0.3 "C. The deuteriated compounds, D,O (99.8%, Aldrich), pr~pan-l-[~H]ol Cambridge Isotope Labs), [,H,]urea (98%, Aldrich) and [,H ,]propan- 1-[ (98Yo MSD Isotope) were used without further purification.The C2H,]urea was dried under vacuum. 4. Results and Discussion 4.1 Translational diffusion coefficients In Fig. 1 the values of the diffusion coefficients D of propan-1- [,HI01 in the binary mixtures with D,O are presented. We can see that D values show a shallow minimum near x, M 0.2. Thus the molecular mobility of propan-1-01 reaches a minimum in the water-rich region. This trend, featured by a minimum at x2 M 0.2, is common to other quantities for solu- tions of alcohols in water.,' What is observed here confirms that in this composition range hydrophobic effects play a role.Urea addition enhances this behaviour, as shown in Fig. 1, where we have also plotted the D values of pr~pan-l-[~H]ol obtained in the ternary system D,O-pr~pan-l-[~H]ol-C2H,]urea (x3 = 0.09). In this case we observe that the minimum is somewhat deeper and slightly shifted to larger x, values, and that at x, < 0.15 the curve is higher than in the absence of urea. The faster diffusion of the pr~pan-l-[~H]ol molecules is the result of the higher mobility of water, due to a 7 v) (uE m z \ Q 'structure-breaking' effect of C2H4]urea, which is obviously transferred to the translational mobility of propan- 1-C2H]ol, indicating a strong coupling of molecular motions of both components. Similar affects were found in the case of structure-breaking ions in binary aqueous systems.* 4.2 A,, parameters In Tables 1-3 are collected the experimental values of (1/TJinter,D and c, necessary for the calculation of A,, for the water-propan-1-01 mixtures with x3 = 0.05 and 0.09 and without added urea. With x3 = 0.05, we studied these quan- tities only in the relevant composition range, 0 < x, < 0.1 (Table 3).In Fig. 2 we plot the values of A2,. In the case of the binary system we observed only a slight increase of A,, as x, decreases, suggesting a modest tendency towards self- association of pr~pan-l-[~H]ol. When C2H4]urea is present, the increase of A,, in the water-rich region is conspicuous, indicating a marked sharpening of g2,(r) and thus an increase of propanol self-association promoted by urea ; moreover, the increase of A,, is proportional to urea concentration.The result of the calculations by means of eqn. (7) and (9) of An;, and n, for the system water( 1)-propan- l-o1(2)-urea(3) at constant urea molar fraction, x3, of 0 and 0.09, are reported in Table 1 Experimental results used for the determination of A,, and f,, in D,O-pr~pan-l-[~H]ol at 25 "C" 0.01 0.008 0.68 0.0224 2.23 0.08 0.05 0.038 0.55 0.1022 2.03 0.29 0.1 0.077 0.45 0.1768 1.79 0.48 0.2 0.101 0.43 0.2930 1.48 0.60 0.3 0.121 0.44 0.3647 1.45 0.73 0.5 0.138 0.47 0.4645 1.41 0.90 1.0 0.127 0.57 0.5621 1.29 1 f,, = (l/Tl)interD/(l/Tl)~nte,Do,(l/Tl)fn,erDo = 7.24 x lo-" mz s-'; where the superscript 0 indicates the values for pure propanol.Table 2 Experimental results used for the determination of A,, and f,, in D,O-pr~pan-l-[~H]ol-[~H,]urea (x3 = 0.09) at 25°C 0.01 0.017 0.76 0.0182 7.1 0.18 0.05 0.038 0.58 0.0833 2.66 0.30 0.1 0.060 0.48 0.1516 1.86 0.40 0.2 0.105 0.39 0.2527 1.63 0.57 0.3 0.133 0.38 0.323 1 1.57 0.70 0.4 0.145 0.41 0.3752 1.57 0.82 0.5 0.164 0.42 0.4141 1.66 0.94 a x; = x,/(x, + x,). Table 3 Experimental results used for the determination of A,, in DzO-propan-l-[2H]ol-[zH4]urea (x3 = 0.05) at 25 "C 0.3 1::0.I xi 0.01 0.012 0.72 0.0200 4.3 0.05 0.039 0.57 0.09 12 2.4 Fig. 1 Self-diffusion coefficients D of propan-l-r'Hlol in 0.1 0.063 0.47 0.1629 1.8 DiO( 1)-propan- l-['H]01(2) (0)and D,O( 1) < prbpan- 1-C2H3o1(2)- ['H4]urea(3) (x3 = 0.09) (0)as a function of x; = x,/(xl + x,) " x; = xJx, + x,). J.Chem. SOC.,Faraday Trans., 1996, Vol. 92 37 11 I 1 I 0 0.5 1a xi Fig. 2 A,, parameters with respect to the association of propan-l- C2H]ol in aqueous mixtures: D,O-pr~pan-l-[~H]ol (0); D,O-pr~pan-l-[~H]ol-[~H~]urea(xg = 0.05)(0)and D20-propan- l-[2H]ol-[2H4]urea (xg= 0.09) (A)as a function of x; = x2/(x1+ x2) Table 4;G, data were taken from ref. 12, and are plotted in Fig. 3 for the sake of completeness. For the ternary mixtures, we have found that the lowest values of m for which n, d ny in the whole concentration range is 4;this means that at the 0 0.1 0.2 0.3 0.4 0.5 0.6 xi Fig.3 Plot of 6G2, in the mixtures H20( l)-propan-l-ol(2)-urea(3) us. xi = x2/(x1+ x2). From top to bottom, the curves refer to con- stant x3 values of 0,0.05 and 0.09, respectively. composition where the maximum 6G2, is obtained (see Fig. 3) the propanol-propanol correlation length is not less than 40. Looking at the SG,, curves in Fig. 3, we can draw similar qualitative conclusions, i.e. the propanol tendency to homo- coordination is strongly enhanced by the presence of urea, and as observed for the A,, parameter, this tendency increases with increasing urea content. Note some similarity between the behaviour of SG,, and of diffusion coefficients D; the extremum they show becomes more marked and shifts towards a higher x, value with increasing urea concentration.However, in the case of the A parameter no maximum is found, whereas the 6G,, values show a maximum at x, x 0.3. This different behaviour, already found in other systems,* can be explained by the fact that A parameters reflect the sharp- ening of the short-range part of the hydrogen-hydrogen pair correlation function g2,(r), which obviously occurs and can be detected only at low concentrations of component 2, where the interactive potential becomes deeper due to the reduced averaging effects. Other factors which might bring the protons involved in gjj statistically closer so as to cause high values of A,, without affecting 6G,,, are orientational effects and a more efficient packing of the propanol molecules; the latter factor is suggested by the negative values of the excess partial molar volumes of propanol in these mixtures.', 4.3 B, coefficients Both A,, and G,, indicate marked self-association tendencies of propan-1-01 upon the addition of urea.We therefore con- sidered BDCoefficients with respect to the influence of urea on the translational diffusion of propan-1-01. If, for example, strong association leads to dimers or higher clusters of propan- 1-01, the translational diffusion coefficient of propan- 1-01 would decrease as a function of the urea concentration, resulting in high, positive B, values. B, coefficients were determined in all the mixtures con-sidered. The results are collected in Table 5. B, gradually decreased from slightly positive to negative values with increasing water concentration.The small absolute B, values demonstrate that urea has a comparatively weak influence on the translational mobility of propan-1-01 in its aqueous mixture. In particular, near the strong G,, maxima at x2 z 0.2-0.4 (see Fig. 3) BD remains very small. This behaviour is in agreement with a weak increase in the viscosity of the ternary system when the urea concentration is increased, as recently found.26 Also the self-diffusion coefficient of urea decreases slightly in this region.26 Thus we can conclude that the B, coefficients in Table 5 give no indication of strong and rigid Table 4 Ani2 and n2/ni in the water(l)-propanol(2)-urea(3) mixture, at urea molar fractions of 0 and 0.09, respectively, for different values of the correlation length m, at 25 "C" 0.05 0.5 4.6 0.20 15.7 0.19 0.5 3.9 0.17 31 0.15 0.1 1.4 8.6 0.38 28.2 0.33 1.6 7.7 0.34 57 0.28 0.15 2.3 12.0 0.53 38.2 0.45 2.7 11.1 0.49 78 0.38 0.2 3.3 14.9 0.65 46.6 0.55 5.5 15.7 0.69 97 0.47 0.25 3.8 17.1 0.75 53.1 0.63 15 26.8 1 121 0.59 0.30 3.3 17.9 0.79 57.6 0.68 37 50 >1 155 0.75 0.35 2.3 18.1 0.79 60.9 0.72 56 70 >1 184 0.90 0.4 1.3 18.1 0.79 63.6 0.75 55 70 >1 192 0.94 0.45 0.7 18.3 0.80 66.2 0.78 35 51.1 >I I80 0.88 0.5 0.4 18.8 0.82 68.8 0.8 1 11 27.9 >I 163 0.79 0.55 0.2 19.3 0.85 71.1 0.84 5.0 22.6 0.99 163 0.80 0.6 0.1 19.8 0.87 73.1 0.86 1.o 19.2 0.84 165 0.8 1 xi = x2/(x, + x2).The correlation length m is defined acording to eqn. (8). For the diameter of the propanol molecule the value c = 4.6 A was used. n: = 22.8, 84.8 and 205.4 for m = 2, 3 and 4, respectively. 38 J. Chem. SOC.,Faraday Trans., 1996, Vol. 92 Table 5 B, coefficients for the influence of C2H,]urea on propan-l- ['Hlol in D,O-propan- 1-[2H]ol mixtures at 25 "C x2 BD/dm3 mol-' 0.05 -0.043 0.1 0.004 0.2 0.026 0.3 0.040 0.4 0.057 propan- 1-01 self-associates. Obviously, the self-association ten- dency commonly reflected by A,, and G,, must be interpreted as 'preferential self-solvation' of propan-1-01 in the ternary mixture, where relatively weak intermolecular interactions do not markedly change the translational mobility of propan- 1-01.Note, the negative B, coefficient in the water-rich region (x, c 0.1) (see also Fig. 1) means that addition of urea slightly increases the mobility of propan-1-01, in spite of the fact that A,, shows a sharpening of g2,(r) in this region. In this region q increases slightly and the self-diffusion coefficient of urea decreases slightly.26 Thus, the translational dynamics of propan-1-01 shows a quite unexpected behaviour in this com- position region. 4.4 Local composition from relaxation data: comparison with the K-B results We have seen that both data obtained from NMR measure-ments and K-B integrals provide a similar qualitative behav- iour regarding self-association of propanol. However, an approach allowing a deeper comparison between these two techniques can be devised from the theory which defines the A parameter. As reported by Muller and Hertz,13 on the basis of the same association criterion discussed in the Introduction, we can define a first coordination number, n,*, according to the equation: n,* = 4n r(~/r)~p~Lr)r, (13)dr where pijr)is the radial distribution function, i.e.the probabil- ity density of finding another particle (of the same or other As shown in ref. 13, the following relation holds: Aij = (K/3a4)n,*/c In our case we can then define a parameter ni* that is closely related to n,* through: (l/Tl)inlerD= Aijc = (K/3a4)n,*= ni* We call it the slightly modified first coordination number (smfcn).We also define the quantityf,, = n;*/ni*' as the ratio of the smfcn of component 2 (at the composition of the mixture under examination) to that of the pure liquid com- ponent. From the previous equations, we see that f,, can be obtained from: where the terms (l/T1)tler and Do refer to the pure component 2. Thef,, values so obtained for the present mixtures are also reported in Tables 1 and 2. From the values of the K-B inte-grals of Fig. 3, following the procedure explained in Section 2.2, we have calculated quantities related to local composi- tions, and in particular the ratio between the number of propanol molecules in the neighbourhood of a central propa- no1 molecule and the corresponding number in pure propanol, n,/n;, limited to the cases with x3 = 0 and 0.09.The calcu- lations have been carried out for rn = 2 and 3 for the binary system, and for rn = 2 and 4 for the ternary system, and are collected in Table 4. f,, and n,/n: should be qualitatively comparable quantities, with due recognition that f2, originates from A,, ,and there- fore is related to the first solvation shell, whereas the other is referred to a solvation shell whose actual thickness is unknown. In Fig. 4(a) these two quantities are plotted for the binary system water-propanol; n,/n(: is reported for the cases with rn = 2 and 3. We can see thatf,, and n,/n; always show posi- tive deviation from the bulk curve, and that in the case of rn = 2 the agreement between the two methods is good.The curve corresponding to rn = 3, which refers to a solvation shell two propanol diameters thick [see eqn. (S)], is always lower than the f2, curve, indicating that over the whole concentra- tion range explored the correlation is always of short range, and that at x2 = 0.3 there is a weak tendency to extend the length. The case with xj = 0.09 is represented in Fig. qb), actual number of nearest neighbours. OCU \!s x2 x'2 Fig. 4 (a) Representation of the trends off,, and n,/n: us. x, for the binary system water( l)-propan-l-o1(2) : J',~(a);n,/n; calculated assuming m = 2 in eqn. (8) (0);n,/n; calculated assuming m = 3 (A). (b) Representation of the trends off,, and n,/n: us.xi = x2/(xI + x,) for the ternary system water( l)-propan-l-o1(2)-urea(3) at constant urea molar fraction x3 = 0.09.f2, (0);(O),n,/n; calculated assuming rn = 2 in eqn. (8) (0);n,/n: calculated assuming m = 4 (A).In both (a) and (b) the curve corresponding to a local composition equal to the bulk composition has been drawn (---). The other curves are smoothing lines empirically drawn. kind) at a distance r relative to the reference particle and (~/r)~where for n,/n(: two curves corresponding to rn = 2 and 4 are acts as a cutting-off function which allows n,* to represent the plotted. Also for this system the two methods show agreement in the sign of their deviation from the bulk curve. However, if we compare each curve of n,/n; with f,,, we see that the agreement with f,, is limited to narrow and different mole fraction ranges.This is an indication that the correlation length changes strongly with composition, varying from two propanol diameters in the water-rich regions to the very large value of 4 in the region 0.3 < x2 < 0.45. This is not surprising, since the mixture examined at x3 = 0.09 and X, = 0.45 is very near a phase separation, and it is well known that systems approaching a critical point are characterised by long-range molecular correlations. Moreover, these correlations extend to long distances before even the first solvation shell becomes fully saturated by propanol molecules. 5. Conclusions The two methods used for the detection of association in binary mixtures and the effects of the addition of a third com- pound on this phenomenon, give consistent and complemen- tary results.They have shown that propanol-propanol association takes place in aqueous solution, particularly in certain concentration regions, and that addition of urea enhances this association. Different features of the two J. Chem. SOC.,Faraday Trans., 1996, Vol. 92 39 methods can be exploited to obtain information on the length of molecular correlations. In the systems examined, the propanol-propanol correlations are never longer than one to two propanol diameters, whilst when urea is present at molar fraction 0.09 the correlation length in the composition range 10 11 12 13 Y. Marcus, J. Chem. SOC.,Faraday Trans., 1990,86,2215. For a review on Kirkwood-Buff theory see, Fluctuation Theory of Mixtures, ed. E.Matteoli and G. A. Mansoori, Taylor Francis, New York, 1990. E. Matteoli and L. Lepori, J. Mol. Liq., 1990,47, 89. K. J. Muller and H. G. Hertz, Chem. Scr., 1989,29,277. 0.3 < x2 < 0.45 extends to four diameters. 14 M. Holz, Prog. Nucl. Magn. Reson. Spectrosc., 1986, 18, 327. The authors thank the National Research Council of Italy, CNR (Internat. Contract No. 93.03 197.03), and the Ministry of University and of Scientific and Technological Research (MURST) for financial support. 15 16 17 18 G. Bonera and A. Rigamonti, J. Chem. Phys., 1965,42,71. M. D. Zeidler, Ber. Bunsen-Ges. Phys. Chem., 1965,69,659. L. Lepori and E. Matteoli, J. Phys. Chem., 1988,92,6997.See e.g., A. Ben Naim, in Fluctuation Theory of Mixtures, ed. E. Matteoli and G. A. Mansoori, Taylor Francis, New York, 1990, p. 211. 19 E. Matteoli and L. Lepori, J. Chem. SOC., Faraday Trans., 1995, References 20 91,431. J. Jones and M. Dole, J. Am. Chem. SOC.,1929,51,2950. H. S. Frank and M. W. Evans, J. Chem. Phys., 1945,13,507. H. G. Hertz, Ber. Bunsen-Ges. Phys. Chem., 1964,68,907. F. Franks, in Water, A Comprehensive Treatise, ed. F. Franks, Plenum Press, New York, 1975, vol. 4, ch. 1, pp. 1-94. 21 22 L. Endom, H. G. Hertz, B. Thul and M. D. Zeidler, Ber. Bunsen- Ges. Phys. Chem., 1967,71, 1008. A. Sacco, M. Carbonara and M. Holz, J. Chem. SOC., Faraday Trans., 1989,85, 1257. A. Ben-Naim, Hydrophobic Interactions, Plenum Press, New York, 1980. A. L. Capparelli, H. G. Hertz and R. Tutsch, J. Phys. Chem., 23 24 M. Holz and H. Weingartner, J. Magn. Reson., 1991,92, 115. M. Holz, H. Weingartner and A. Sacco, Ber. Bunsen-Ges. Phys. Chem., 1990,94,332. 1978,82,2023. W. Koch, H. Leiter and H. G. Hertz, J. Solution Chem., 1981, 10, 25 F. Franks and D. S. Reid, in Water, A Comprehensive Treatise, ed. F. Franks, Plenum Press, London, 1972, vol. 2. 419. H. Leiter, K. J. Patil and H. G. Hertz, J. Solution Chem., 1983, 12, 26 E. Hawlicka and R. Graboswi, Ber. Bunsen-Ges. Phys. Chem., 1994, 98, 824. 508. M. Holz, R. Grunder, A. Sacco and A. Meleleo, J. Chem. Soc., Faraday Trans., 1993,89, 1253. E. Matteoli and L. Lepori, J. Chem. Phys., 1984,80, 2856. Paper 5/04383B; Received 5th July, 1995 40 J. Chem. SOC.,Faraday Trans., 1996, Vol. 92
ISSN:0956-5000
DOI:10.1039/FT9969200035
出版商:RSC
年代:1996
数据来源: RSC
|
8. |
Thermodynamics of the interaction of HCl with propan-2-ol in water at 278.15–338.15 K |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 92,
Issue 1,
1996,
Page 41-45
Kelei Zhuo,
Preview
|
PDF (607KB)
|
|
摘要:
Thermodynamics of the interaction of HCl with propan-2-01 in water at 278.15-338.15 K Kelei Zhuo, Jianji Wang* and Jinsuo Lu Department of Chemistry, Henan Normal University, X inxiang, Henan 453002, People's Republic of China The emfs of cells Pt, H2(g, po) 1 HCl(mE)I AgCl/Ag, and Pt, H,(g, po)I HCl(mE), 2-PrOH(mN) I AgCl/Ag have been measured at 10 K intervals from 278.15 to 338.15 K, where 2-PrOH refers to propan-2-01, mE= 0.005-0.1 mol kg-', and mN = 0.2-0.7 mol kg-'. The emf data were used to determine the salting-out constant, k,, of 2-PrOH and the thermodynamic interaction parameters,f,, (SEN, h,, ,SEN, cp,EN) of 2-PrOH with HC1 in water at different temperatures. It has been shown that k, > 0, SEN > 0, hE, > 0, sEN> 0 and cp,EN < 0 at all temperatures concerned (except for sENand hEN which are negative at 328.15 and 338.15 K) and that SEN and cp,EN vary linearly with increasing temperature. A comparison of these thermodynamic parameters is made with those for HC1-PG (propane-1,2-diol) and HCl-GL (glycerol) pairs in water.The dependence of the parameters on the number of hydroxy groups in alcohol molecules has been interpreted on the basis of the group additivity model. There have been many studies on the thermodynamic proper- ties of the water-propan-2-01 (2-PrOH)-HCl ternary system.'-9 Most of these were devoted to the investigation of transfer thermodynamics of HCl from water to aqueous solu- tions of higher 2-PrOH concentrations. Among these transfer thermodynamic functions, At Go is important in understanding the HC1-solvent (water-2-PrOH) interaction or solvation of HC1 in the solvent ;l-'q9 At So and At c: are effective probes for the structure of mixed solvent^.^-^.'*^ Other studies were con- cerned with the measurement of mean activity coefficients of HC1 in water-2-PrOH mixed solvent^,^.^^^,^ which mainly provide information on the HCl-HCl interaction in the mixed solvent, and on the effect of the mixed solvent on HC1-HC1 interactions.No studies on the thermodynamics of the inter- action of HC1 with 2-PrOH in dilute aqueous solution have been reported in the literature, although information on this kind of interaction can serve to amplify the understanding of analogous interactions in the more complex biological cases.' O Recently, we have investigated the thermodynamics of inter- action of some electrolytes with sucrose, glucose, glycerol (GL), and PG in water."-'5 We report here the thermodyna- mic interaction parameters, fEN (SEN, hEN, SEN and cp,EN), between HC1 and 2-PrOH in water and the salting-out con- stants, k,, from 278.15 to 338.15 K at lo" intervals.A com- parison of these parameters is made with those for HC1-PG and HCl-GL in water. The dependence of these parameters on the number of hydroxy groups in the alcohol has been interpreted on the basis of the group additivity model. This study provides additional information on HCl-2-PrOH inter-actions in the water-2-PrOH-HC1 system. Measurements were made of the emfs of the cells A and B without a liquid-junction at mE(A) = mE(B) Pt, H2(g, Po) I HC1(mE) 1 AgC1/Ag (A) Pt, H2(g, Po) I HC1(mE), Z-PrOH(mN) I AgC1/Ag (B) at all temperatures investigated, where po = 101.325 kPa, and mE and mN are the molalities of electrolyte and non-electrolyte, respectively, defined as per kg of pure water; m, = 0.005-0.1 mol kg- ',mN = 0.2-0.7 mol kg- '.Experimental Propan-2-01 (A.R. grade) was distilled under reduced pressure over a 100 cm long fractionating column after drying on molecular sieve 4A, and the middle fraction was collected. Stock solution of hydrochloric acid was prepared from the constant-boiling acid doubly distilled from the G.R. grade acid in all-glass apparatus. The acid was standardized by a gravimetric determination of chloride as AgCl.The average difference among five replicate determinations was less than & 0.02%. Redistilled deionized water with specific conduct- ivity ca. (1.0-1.2) x lop4 Q-' m-' was prepared in an all- glass still. All test solutions were prepared by weighing. The Ag/AgCl electrodes were of the thermal-electrolytic type,'' and were aged in 0.1 mol kg-' HCl(aq), which was deoxygenated by bubbling hydrogen. Three days after prep- aration, the finished electrodes were intercompared and had bias potentials of usually less than f0.05 mV. Only Ag/AgCl electrodes whose bias potential was less than 0.02 mV were used. Standard electrode potentials of the Ag/AgCl in pure water were determined by Bates' method.17 They are in excel- lent agreement with the values reported in the literature" within experimental error. The hydrogen electrodes were lightly coated with platinum black according to the method of Hills and Ives." The high purity hydrogen served as the source of hydrogen.The cells were of all-glass construction with four isothermal presaturators containing the test solu- tion, as described by Yang et aL2' These cells were ther- mostatted at each temperature with an accuracy of kO.02 K. For a set of measurements, two Ag/AgCl electrodes were soaked overnight in the test solution. All measurements were made with two Ag/AgCl electrodes and two hydrogen elec- trodes and the mean values of these observations were re- corded. Equilibrium was reached ca.2-3 h after the initiation of hydrogen bubbling. The criterion was a steady reading within k0.05 mV for 1 h. The cell emf measurements were made by means of a UJ-25-type potentiometer which was Cali- brated against a standard cell. A mirror-type galvanometer was used as a null detector. Two series of cell emf measure- ments were made at each concentration. The first was suc- cessively at 298.15, 278.15, 288.15, 298.25, 308.15 and 298.15 K. Three emf values at 298.15 K agreed to within k0.05 mV. The second was successively at 298.15, 318.15, 328.15, 338.15 and 298.15 K. Two emf values at 298.15 K generally agreed within kO.1 mV. The atmospheric pressure was measured by a barometer and was calibrated for temperature, height above sea level, and latitude.The vapour pressure data of the solutions at the J. Chem. SOC.,Faraday Trans., 1996,92 (l), 41-45 experimental temperatures and molalities were obtained by the experimental data. They are included in Table 2, together interpolation from the values in the literature." Then, the with their standard deviations and the standard deviation of observed emf values were corrected to those at a hydrogen the fit, Q. Parameters B and C lack precision. However, this is partial pressure of 101.325 kPa. The correction of emf values not too important in the present case, since our principal is sensitive to the vapour pressure data used at higher tem- concern is the leading term A which shows the interactions of peratures (especially at 338.15 K).The corrected emfs of cells HCl with 2-PrOH in water. A and B are listed in Table 1. Application of the cross-differential relati~nship~~ to eqn. (1) gives where yN and yi are the activity coefficients of the non-The procedure applied in the analysis of the experimental data electrolyte (2-PrOH) in aqueous solution with and without was similar to that suggested by Scatchard22 and used by HCl, respectively. From eqn. (2), we have Lilley and co-~orkers~~ for electrolyte, non-electrolyte, and lim [ln(y,/yi)/m,] = 2A = ks (3)their mixtures in aqueous solution. For the HCl-2-PrOH- mN+O water system, we mE-0 The values of ks at all seven temperatures were obtained h(y*/y:) = -AE/(2k) = Am, + BmEmN + cmi (1) from eqn. (3), and are listed in Table 3.Furthermore, ks was where the higher-order terms have been neglected. y: and yk analysed as a function of temperature using a least-squares refer to the mean ionic activity coefficients of HCl at molality routine. The best representation of the data was mE in pure water and in aqueous solution of 2-PrOH at k, = a/T + b + cT (4)molality mN, respectively. AE is the difference in emf between cells A and B at mE(A) = mE(B) and k = RT/F. The param- The coefficients a = 1073.0 kg mol-' K, b = -6.6241 kg eters A, B and C were obtained from least-squares analysis of mol-', c = 0.010316 kg mol-' K-' were obtained with the Table 1 Emfs (in V) of cells A and B from 278.15 to 338.15 K m,/mol kg-' 278.15 288.15 298.15 308.15 318.15 328.15 338.15 mE = 0.004 999 mol kg -' O.oo00 0.491 70 0.495 42 0.498 55 0.501 15 0.503 01 0.504 42 0.505 33 0.3151 0.490 85 0.494 78 0.498 10 0.500 80 0.502 70 0.504 20 0.505 08 0.5050 0.490 29 0.494 44 0.497 8 1 0.500 40 0.502 53 0.503 94 0.504 98 m, = 0.009 999 mol kg-' O.oo00 0.459 60 0.462 26 0.464 25 0.465 75 0.466 7 1 0.467 14 0.467 09 0.1976 0.459 06 0.461 88 0.464 01 0.465 63 0.466 54 0.466 96 0.466 94 0.4033 0.458 58 0.461 54 0.463 78 0.465 35 0.466 45 0.466 84 0.466 89 0.6752 0.457 76 0.460 92 0.463 30 0.465 12 0.466 09 0.466 67 0.466 84 mE = 0.030016 mol kg-' O.oo00 0.409 42 0.410 35 0.410 63 0.410 32 0.409 41 0.407 9 1 0.406 03 0.1863 0.408 94 0.410 12 0.410 39 0.410 18 0.409 22 0.407 70 0.405 93 0.4025 0.408 42 0.409 64 0.410 13 0.410 02 0.409 05 0.407 68 0.405 86 0.6675 0.407 64 0.409 05 0.409 68 0.409 76 0.408 86 0.407 48 0.405 90 mE= 0.049997 mol kg-' O.oo00 0.386 34 0.386 48 0.386 00 0.384 96 0.383 22 0.381 07 0.378 43 0.1964 0.385 88 0.386 14 0.385 72 0.384 77 0.383 05 0.380 93 0.378 52 0.3872 0.385 38 0.385 76 0.385 45 0.384 60 0.382 95 0.380 87 0.378 58 0.7225 0.384 45 0.385 11 0.384 96 0.384 26 0.382 69 0.380 78 0.378 82 m, = 0.070 000 mol kg -' O.oo00 0.371 14 0.370 59 0.369 71 0.368 15 0.365 75 0.363 37 0.360 34 0.1982 0.370 66 0.370 44 0.369 5 1 0.368 03 0.365 62 0.363 28 0.360 42 0.3995 0.370 18 0.370 08 0.369 29 0.367 9 1 0.365 53 0.363 23 0.360 64 0.699 1 0.369 3 1 0.369 46 0.368 82 0.367 55 0.365 18 0.363 19 0.360 86 mE = 0.10000 rnol kg-' O.oo00 0.255 08 0.354 12 0.352 57 0.35041 0.347 84 0.344 26 0.340 48 0.1994 0.354 60 0.353 75 0.352 35 0.350 29 0.347 67 0.344 15 0.340 63 0.4002 0.35403 0.353 36 0.352 04 0.350 09 0.347 57 0.344 10 0.34081 0.6914 0.353 27 0.35288 0.351 73 0.349 96 0.347 4 1 0.344 07 0.341 23 Table 2 Parameters of eqn.(1) and standard deviations of fit (T (in mV) TIK A/10-2 kg mol-' B/10-2 kg2 molP2 C/10-3 kg2 mol-2 a/mV 278.15 5.220 & 0.167 -3.082 f1.312 6.792 f2.722 0.043 288.15 3.646 f0.136 -2.452 f1.075 4.916 f2.210 0.035 298.15 2.401 f0.187 -4.078 & 1.474 6.496 & 3.039 0.047 308.15 1.787 f0.233 -5.470 f1.840 1.702 f3.790 0.06 1 318.15 1.608 f0.161 -4.879 f1.271 0.903 f2.618 0.043 328.15 1.744 0.157 -9.067 f1.241 -6.937 & 2.558 0.044 338.15 1.745 f0.275 -30.09 f2.1 72 -11.73 f4.476 0.079 42 J.Chem. SOC.,Faraday Trans., 1996, Vol. 92 Table 3 Salting-out constants, k,, of 2-PrOH by HCl and the pair interaction parameters of HCl-2-PrOH in water from 278.15 to 338.15 K T/K k&g mol -’ gEN/Jkg mOl-’ Ts,N/J kg mol-2 hEN/Jkg mol-2 c,,,,/J kg molP2 K-’ 278.15 0.104 60 288.15 0.073 44 298.15 0.048 30 308.15 0.036 23 318.15 0.032 21 328.15 0.035 24 338.15 0.035 25 standard deviation of fit, 0.0029 kg mol-l, and the correlation coefficient, 0.995. The pair free energy interaction parameter SEN is related25*26to ks by SEN = (RT/2v)kS (5) where v is the number of moles of ions into which the electro- lyte dissociates. The pair interaction parameters are related to each other in the usual way.26 From eqn.(4) and (5), it follows that SEN = k’(U + bT -t-CT2) (6) SEN = -k‘(b + 2cT) (7) hEN = k’(a -cT’) (8) cp, EN = -2k‘cT (9) where k’ = R/(2v). Using eqn. (5) and (7)-(9), thermodynamic pair interaction parameters fEN at different temperatures were obtained and are also included in Table 3. To the best of our knowledge, these are the first data on the thermodynamics of the interaction of 2-PrOH with HC1 in water. The interaction of an electrolyte with a non-electrolyte includes the inter- actions of each ion dissociated by the electrolyte with the non- electrolytic molecule. In the present case, fEN characterizes the mean behaviour of all the pair interactions between 2-PrOH and each of ion ions (H and C1 -), uiz :+ fEN = (fH+-2-PrOH +fcl--~PrOd/2 (10) Discussion It can be seen from Table 3 that ks > 0, SEN > 0, SEN > 0, ~~hEN > 0, c~ < 0, at the temperatures concerned except for SEN and h,, which are negative at 328.15 and 338.15 K, and all of them decrease with increasing temperature.This shows that 2-PrOH is being salted-out by HCl in water; the interaction between 2-PrOH and HC1 in water leads to the increase of entropies and enthalpies for the system 2-PrOH-HC1-water from 278.15 to 318.15 K and the interaction of 2-PrOH and HCl in water decreases with increasing temperature. Similar results have been reported for HC1-PG,l4 HC1-GL,” NaX (X = C1, I)-alcohol(2-PrOH, MeOH, EtOH, PrOH, Bu’OH, et~.),~’ and alkali-metal halide-alcohol (MeOH, EtOH, PrOH, BuOH, PenOH, etc.)26 pairs in water. Furthermore, SEN and cp,EN vary linearly with temperature, as shown in eqn. (7) and (8), respectively.A similar linear relationship exists for the HC1-PG-water and HC1-GL-water lS The above phenomena can be partly interpreted by the structural interaction and group additivity models. The pair interactions include not only the usual electrostatic, inductive, and dispersion, but also the structural interactions (i.e. change in the hydration of the molecules and ions). The structural interactions are quite temperature dependent, and become the leading effect with enthalpy, entropy, and heat capacity.26 The group additivity has been successfully applied to electrolyte-non-electrolyte pairs in ~ater.~’-~’ In the present case, we divide the alcohol molecule into alkyl (R) and hydroxy (OH) groups.The interactions of R with ions are 512 517 -11.9 407 450 -12.4 293 324 -12.8 171 194 -13.2 40 60 -13.6 -100 -79 -14.1 -248 -222 -14.5 mainly structural, and are the leading contribution to hEN, ,~~SEN, and c~ parameters. The R group in water tends to strengthen the hydrogen bonds between water molecules near it.32 As the R group and H+ or C1- approach each other in water, the hydration cospheres of the R group are partly transformed into bulk water. This corresponds to the increase in entropy and enthalpy, resulting in hEN > 0 and SEN > 0 from 278.15 to 318.15 K.This effect decreases with increasing temperature.26 Therefore, values of hEN and sENdecrease with increasing temperature so that the electrostatic interactions (thermodynamic attraction) become the leading contributions to SEN and hEN at 328.15 and 338.15 K. Heat capacity is a very sensitive probe for studying structural interactions. The reduction in structure of the water arising from the pair inter- ~~actions results in c~< 0., De Visser et also suggested that the interactions between hydrophilic and hydrophobic solutes [urea-tert-butyl alcohyl (TBA), NaCl-TBA, and Bu,N+-Br-] cause a negative contribution to the heat capac- ity. There is a compensation between hEN and SEN since both of them have the same sign. The sign of SEN is determined by the corresponding enthalpy change in the hydrophobic-hydrophilic interactions at ordinary temperature^.^^ Thus, the contribution of structural interactions to ks and SEN should be positive.This result is identical to that predicted by the elec- trostatic interaction model. However, the sign of SEN is deter- mined by the corresponding entropy change at higher temperatures (328.15 and 338.15 K). 2-PrOH, PG and GL molecules all have the same carbon chain, namely C-C-C, whereas the number of OH groups successively increases from one to three. Using the data for PG-HC1 and GL-HCl reported previo~sly,’~*’ the depen- dences of k, andf,, on the number of OH groups in these molecules are shown in Fig. 1-3. Fig. 1 shows that the values of SEN decrease almost linearly with increasing number of OH groups at 278.15, 288.15 and 298.15 K, indicating increased attractive interaction or decreased repulsive interaction.This 70-A N-g 50-M 24 h 16 30-cu -10 I JI I I 1 2 3 Number of OH groups Fig. 1 Dependence of gEN on number of OH groups: 1, 2-PrOH; 2, PG; 3, GL: A, 278.15;A, 288.15; 0,298.15; a,308.15; 0,318.15 K J. Chem. SOC., Faraday Trans., 1996, VoZ. 92 43 --4 m d8 -8-M24-\ zi. -12-oQ I I I II I I 1 2 3 Number of OH groups Fig. 2 Dependence of cp.EN on number of OH groups: 1,2-PrOH; 2, PG; 3, GL: A, 278.15; 0,298.15; 0,318.15 K can be interpreted by the group additivity principle. The inter- actions between OH groups and ions contribute negative values to ks and SEN, because the electrostatic attraction of the ion-dipole type is the leading factor.Obviously, the negative contributions should increase with increasing number of OH groups. Wilcox and Schrier'O calculated the OH group contri- bution to the limiting interaction parameter (identical to SEN in this paper) for alcohol-NaX in water at 298.15 K, where alcohol refers to MeOH, EtOH, PrOH or ethylene glycol and X is C1, Br or I. These values are also negative. However, the linear relationship becomes poor at higher temperatures (308.15 and 318.15 K), i.e. the group additivity gradually van- ishes with increasing temperature. A possible reason is that the independence of the groups in a molecule decreases with increasing temperature resulting from interactions enhanced between the groups in a molecule.The variation of k, with the number of OH groups is the same as that of SEN. a A \ I I I 1 2 3 Number of OH groups Fig. 3 Dependence of h,, and sEN on number of OH groups: 1, 2-PrOH; 2, PG; 3, GL: A, 278.15; A,288.15; 0,298.15; 0,308.15; 0,318.15 K 44 J. Chem. SOC.,Faraday Trans., 1996, Vd.92 I I I 1 I I I 278.15 298.15 318.15 338.15 TIK Fig. 4 Temperature dependence of ks of HC1-2-PrOH-water system Fig. 2 shows a good linear dependence of c ~ on ,the ~ ~ number of OH groups at all temperatures. Although the values of hEN and sENalso depend in part on the number of OH groups, as shown in Fig.3, no linear relationship between these parameters and the number of OH groups was observed, especially at higher temperatures. The reason is possibly that the structural interactions, being the leading effect with h,, and sEN,and quite sensitive to temperature, are affected more by the independence of the functional groups (i.e. OH) in the same molecule. In the present case, only the dependence of ks on the number of OH groups was discussed; it remains to be demonstrated by extra experiments whether the dependence is also on the position of the OH groups in alcohol molecules which have the same carbon chains (C-C-C). If a set of data on pair interaction parameters for HC1-alcohol-water systems is experimentally obtained, the values of the groups' contribu- tions to pair interaction parameters can be evaluated by a least-squares method.Work on this is in progress. The temperature dependence of the salting constants of the HC1-2-PrOH-water system is represented in Fig. 4. k, decreases slowly with temperature at low temperatures, but hardly changes above 318 K. The results show that a minimum in k, (ca. 0.03 kg mol-') takes place at ca. 322 K. However, this minimum is not marked since k, changes very slowly with temperature in this temperature region. Similar results have been found for the HC1-PG-water and HCl- TBA-water systems. 14*35 In addition, Desnoyers et a/.36pre-dicted from the scaled-particle theory37 that the minimum in k, for the systems electrolyte (NaCI, CsC1)-alcohol (methanol, ethanol)-water is at ca.323 K, and the experimentally observed value is at 333 K. Financial support from the National Natural Science Founda- tion of China and the Natural Science Foundation of Henan Province is gratefully acknowledged. References 1 Y. Marcus, Pure Appl. Chem., 1990,62,899. 2 K. Bose, K. Das, A. K. Das and K. K. Kundu, J. Chem. Soc., Faraday Trans. I, 1978,74, 1051. 3 R.N. Roy and A. L. M. Bothwell, J. Chem. Eng. Data, 1970, 15, 548; 1971,16,347. 4 R. N. Roy, W. Vernon and A. L. M. Bothwell, J. Chem. Ther- modyn., 1971,3, 769. 5 R. Smiths, D. L. Massart, J. Juillard and J-P. Morel, Electrochim. Acta, 1976, 21,425. 6 M. M. Elsemongy and A. S. Fouda, J. Electroanal. Chem., 1980, 114, 25.7 8 R. L. More and W. A. Felsing, J. Am. Chem. SOC., 1947,69, 1076. C. A, Vega, B. Perez and C. Torres, J. Chem. Eng. Data, 1984, 29, 24 H. S. Harned and B. B. Owen, Physical Chemistry of Electrolytic Solutions, Reinhold, New York, 3rd edn., 1958, p. 16. 129. 25 H. L. Friedman, J. Solution Chem., 1972,1, 387. 9 C. A. Vega, E. Rosado and R. G. Bates, J. Chem. Thermodyn. 26 G. Perron, D. Joly, J. E. Desnoyers, L. Avedikian and J. P. 1990,22, 355. Morel, Can. J. Chem., 1978,56, 552. 10 F. L. Wilcox and E. E. Schrier, J. Phys. Chem., 1971,75, 3757. 27 H. Piekarski, Can. J. Chem., 1986,64,2127. 11 12 J. Wang, W. Liu, T. Bai and J. Lu, J. Chem. SOC., Faraday Trans., 1993,89, 1741. J. Wang, L. Zeng, W. Liu and J. Lu, Thermochim.Acta, 1993, 28 29 J. J. Savage and R. H. Wood, J. Solution Chem., 1976,5,733. M. A. Gallardo-Jimenez and T. H. Lilley, J. Chem. SOC., Faraday Trans. I, 1989,85,2909. 13 224,26 1. J. Wang, W. Liu, J. Fan and J. Lu, J. Chem. SOC., Faraday Trans., 30 W. J. M. Heuvelsland, C. de Visser and G. Somsen, J. Chem. SOC., Faraday Trans. I, 1981,77,1191. 14 1994,90,328 1. K. Zhuo, J. Wang, Y. Lu and J. Lu, Acta Chim. Sin., 1994, 52, 31 H. Tierkarski and M. Tkaczyk, J. Chem. SOC., Faraday Trans., 1991,87,3661. 15 461. K. Zhuo, J. Wang and J. Lu, Acta Chim. Sin., 1994, accepted for 32 J. E. Desnoyers, G. E. Pelletier and C. Jolicoeur, Can. J. Chem., 1965,43,3232. 16 publication. R. G. Bates, Determination ofpH, Wiley, New York, 1964, p. 281. 33 C. de Visser, G.Perron and J. E. Desnoyers, J. Am. Chem. SOC., 1977,99,5894. 17 18 R. G. Bates and R. A. Robinson, J. Solution Chem., 1980,9,455. D. R. White, R. A. Robinson and R. G. Bates, J. Solution Chem., 34 J. E. Desnoyers, M. Arel, G. Perron and C. Jolicoeur, J. Phys. Chem., 1969,73,3346. 19 1980,9,457. G. J. Hills and D. J. G. Ives, in Reference Electrodes, ed. D. J. G. 35 K. Zhuo, MSc. Thesis, Henan Normal University, P.R. China, 1992. 20 Ives and G. J. Janz, Academic Press, New York, 1960, p. 107. J. Yang, D. Men, Ch. Liang, L. Zhang, L. He and A. Sun, J. Phys. 36 J. E. Desnoyers, G. Perron, S. Leger, B. Y. Okamoto, T. H. Lilley and R. H. Wood, J. Solution Chem., 1978,7, 165. 21 Chem., 1989,93,7248. J. Gmehling and U. Onken, Vapor-Liquid Equilibrium Data Col- lection (Aqueous Organic Systems), Dechema, Frankfurt, 1977, pp. 37 W. L. Masterton and T. P. Lee, in Chemistry and Physics of Aqueous Gas Solutions, ed. W. A. Adamas, Princeton, New York, 1975, p. 199. 325-327. 22 G. Scatchard, J. Am, Chem. SOC., 1968,90,3124. 23 C. C. Briggs, T. H. Lilley, J. Rutherford and S. Woodhead, J. Solution Chem., 1974,3, 649. Paper 5/01250C; Received 1st March, 1995 J. Chem. SOC.,Faraday Trans., 1996, Vol. 92 45
ISSN:0956-5000
DOI:10.1039/FT9969200041
出版商:RSC
年代:1996
数据来源: RSC
|
9. |
Thermodynamic study of the ternary system NaCl–H2O–Et3N at 25 °C. Part 2.—Compressibilities |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 92,
Issue 1,
1996,
Page 47-51
Isabel M. S. Lampreia,
Preview
|
PDF (639KB)
|
|
摘要:
Thermodynamic study of the ternary system NaCI-H,O-Et,N at 25 "C Part 2.+Compressi bilities Isabel M. S. Lampreia* and Luis A. V. Ferreira CiTecMaT, Centro de Cihcia e Tecnologia de Materiais da Faculdade de CiCncias da Universidade de Lisboa, Campo Grande, Ed. C1,1700, Lisboa, Portugal Ultrasonic velocities in ternary systems of triethylamine in aqueous sodium chloride solutions have been measured at 25 "C, for sodium chloride molalities ranging from 0 to 0.3 mol kg- '. The concentration of triethylamine was varied from 0 to ca. 0.35 mol kg-'. The concentration of triethylamine was varied from 0 to ca. 0.35 mol kg-* in pure water and in each of the four concentration-fixed NaCl solutions investigated. Isentropic compressibilities, K~,were calculated from ultrasonic velocity and density data.Excess molar isentropic compressions, Kt,m, were estimated and their variations with the amine mole fraction were fitted to the Redlich-Kister equation. Negative KE, values were obtained for all the systems in the concentration range studied. Evaluation of the triethylamine limiting partial molar isentropic compression, K,q!Et3N, was made for each system, using dK:, ,,,/dXEt3Nderivatives. Negative K,q!Et3N values were obtained, increasing in absolute value up to 0.09 mol kg- NaCl concen- tration and then steeply decreasing. Apparent molar isentropic compressions, Ks,&, of Et,N were also calculated and plotted for all the solvents studied. The interpretation of the results, derived from experimental ultrasonic velocities, was made in terms of the important roles of the ions and the Et,N molecule in determining the water structure.The results support and complement the conclusions made in Part 1 of this work, which were based on volumetric data. Partial molar volumes and compressions at infinite dilution have often been used in a complementary mode to obtain information about structural and interaction phenomena associated with solvation processes. '-'Recently, an increasing number of thermodynamic studies in ternary mixtures con- taining electrolytes are being performed motivated by the fact that the addition of electrolytes to mixed solvents permits a wide range of solutions with suitable properties for practical applications to be ~btained.~ Useful information about the nature and extent of packing effects in solution, induced by intermolecular interactions, has been obtained from studies of the variation with composition of excess and apparent molar compression^.^-^ Plots of this last property against composition are usually appropriate to show changes in molecular aggregation patterns, such as micelle formation,' where significant variations in curve slopes are observed.Limiting partial molar compressions are particularly sensi- tive to changes in the solvent structure near a solute molecule infinitely diluted in the solvent. In this paper, as Part 2 of a thermodynamic study concern- ing the effect of NaCl on the solubility of Et,N in water, we report experimental isentropic compressibilities, and excess, apparent and limiting partial molar isentropic compressions for Et,N in water and in NaCl,, solutions of four different compositions. This study enabled us to confirm and extend the conclusions drawn in Part 1 of this work," about changes in the packing efficiency of the solvent molecules near the Et,N molecule and due to the presence of ions in solution. Changes of the solvent structure with composition, induced by intermolecular interactions, are also studied.Experimental Materials Triethylamine was a reagent grade product supplied by Merck t Part 1 :Ref. 10. and submitted to further distillation. Details of the procedure followed have been described previously.' Its purity was tested by density measurement.Sodium chloride was BDH AnalaR with a quoted purity >99.9%. It was dried at 120"C in a vacuum stove, for several hours, when required." Water was ion exchanged (18 Mil cm) high purity, from Milli-Q reagent grade system, supplied by Millipore. Measurements Ultrasonic velocities (u) were measured using an improved Nusonic sing-around Velocimeter model 6080, supplied by Mapco Inc., with a double transducer system operating at 1.8 MHz. Its working method has been described in detail by Millero and Kubinsky.'* The precision and accuracy of the measurements were 0.05 and 0.2 m s-l, respectively. The ultrasonic velocimeter was calibrated with degassed pure water from 16 to 30°C by using the data of Del Grosso and Mader.', A tight cell, made of Pyrex glass and specially designed to prevent evaporation, was used.The cell was intro- duced into a water bath controlled to 25.000 0.001 "C with a Tronac proportional temperature controller as measured by a Hewlett-Packard quartz crystal thermometer. The thermom- eter had been previously calibrated against a precision plati- num resistance thermometer supplied with a calibration certificate from the National Bureau of Standards (IPTS-68). All solutions were prepared by weight adding Et,N to pure water or to four different aqueous sodium chloride stock solu- tions previously prepared by weight as described in detail in Part 1 of this work." The molalities of the stock solutions, considered as solvents, were determined by density measure- ments yielding 0.05427, 0.08751, 0.19836 and 0.030632 mol kg-'.Results and Discussion Ultrasonic velocities and isentropic compressibilities The experimental u (Table 1) for the solutions of triethylamine in water and in the four different aqueous sodium chloride J. Chem. SOC.,Faraday Trans., 1996,92 (l),47-51 Table 1 Ultrasonic velocities of binary and pseudo-binary solutions of Et,N in pure water and in NaCI,, solutions at 25 "C H2O NaCI,, (0.05 mol dm -,) NaCl,, (0.09 rnol dm -,) NaCl,, (0.2 rnol dm -3, NaCl,, (0.3 mol dm -,) m/mol kg-' u/m s-l m/mol kg-' u/m s-l m/mol kg-' 0.03 102 1499.91 0.03 179 1503.50 0.05009 0.0456 1 1501.58 0.04732 1505.1 1 0.052 1 1 0.06025 1503.58 0.06 150 1506.64 0.06510 0.07563 1504.80 0.07326 1507.96 0.08258 0.08960 1506.26 0.08899 1509.65 0.09495 0.09670 1507.07 0.10168 151 1.04 0.15 124 0.1491 1 15 12.90 0.15458 15 16.52 0.17638 0.1721 1 151 5.10 0.17554 1518.67 0.20358 0.20042 1518.00 0.20341 1521.85 0.22776 0.22200 1520.11 0.23 194 1525.11 0.25609 0.25397 1523.86 0.25647 1527.5 1 0.25333 0.30099 1528.73 0.29854 1531.49 solutions, at 25 "C, were fitted to equations of the type u = uo + Am (1) where uo and A are solvent dependent empirical parameters and m is the Et,N molality.The isentropic compressibilities, defined by xs = (-l/Vm) x (8VJaP), at fixed composition, were obtained from ultra- sonic velocity and density data," using the Newton-Laplace equation KS = l/(u2d) (2) where d is the density of the solution.xs values were fitted to linear equations of the form xS = xs, + A'm (3) where xSV0and A' are solvent dependent empirical param- eters. Fig. 1 shows xS as a function of the Et,N molality in each solvent. ,The parameters uo, A, K~ and ~A', determined by least- squares fits of the data, are given in Table 2, together with 4.50 -4.45 4.40. TIJ 14.35. f 4.30. \ kV) 4.1 5 0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 mlmol kg-' Fig. 1 Isentropic compressibilities us. molality for Et,N in pure water and in NaCl,, solution at 25 "C. (+), NaCl (0.3 mol kg-'); (O), u/m s-' m/mol kg-' u/m s-' m/mol kg-' u/m s-l ~~~ ~ 1507.59 0.03032 1512.48 0.05636 1521.5 1 1507.83 0.04372 1513.97 0.07669 1523.65 1509.18 0.0488 1 1514.40 0.07703 1523.74 1511.16 0.07377 1517.12 0.09637 1525.61 1512.34 0.08980 1518.38 0.20852 1537.74 15 18.36 0.10054 15 1 9.60 0.25652 1542.38 1521.43 0.15078 1525.34 1524.18 0.17734 1528.22 1526.40 0.19899 1530.21 1529.14 0.22683 1532.87 1529.19 0.25639 1536.34 0.2993 1 1540.77 their standard deviations and the standard deviations of the fits.Also shown in Table 2 are the experimental sound veloci- ties in the various solvents which can be compared with the respective uo parameters. As can be seen in Fig. 1, the effect of adding both NaCl and Et,N to pure water is the continual decrease of its isentropic compressibility.Excess, partial and apparent molar isentropic compressions The molar property defined by Ks,m= -(dVJaP),, at fixed composition, has usually been designated the molar isentropic compressibility. Since the term compressibility has been reserved for the intensive property defined by rcS = (-l/Vd x (dVJaP), , Douheret and Davis14 have recently called attention to the problem concerning the nomenclature of the Ks,mquantity. Reis*' preferred to name Ks,mthe molar isen- tropic compression. Accordingly, compression has been adopted in this work to represent pressure derivatives of any apparent or real molar volume. The same criterion has been used in relation to the nomenclature of the temperature deriv- ative of the molar volume, Ep,,,= (dVJaT),, called here the molar expansion.In this work, the excess molar isentropic compression, K:, ,,, represents the deviation of the molar isentropic com- pression of real binary and pseudo-binary mixtures, from the same thermodynamic quantities calculated for an idealised type of solution, considered as a first approximation to the behaviour of real solutions. This ideal solution would be formed by two supposedly pure components, namely Et,N (component B) and pure water or aqueous NaCl solutions of fixed composition (component A). An ideal binary solution, formed at constant pressure and temperature, may be defined in a number of ways, provided that a consistent set of mixing rules can be derived. The most common definition states that a solution is ideal if every com- ponent obeys Raoult's law.If it is assumed that the vapours of the two components are ideal gases at the working tem-perature and pressure, and that the effect of pressure on the NaCl (0.2 rnol kg-'); (A), NaCI,, (0.09 rnol kg-'); (O),NaCl,, (0.05 chemical potential of the liquid phase can be neglected, an mol $-I); (01, pure water. equivalent definition, in terms of the molar Gibbs energy, Table 2 Parameters of eqn. (1) and (3) for Et,N in pure water and in NaCI,, solutions at 25 "C us, A'/10-5 kg O,~,P/~O-~ solvent u,/m s-l A/kg m mol-' s-' o,,t/m s-' MPa-MPa-' mol-' MPa-' u,(exp)/m s-l pure water 1496.83 0.1 106.0 f0.6 0.18 4.4754 & 0.0006 -5.38 f0.04 0.00 1 1496.90 NaCI,, (0.05)b 1500.16 f0.1 106.1 k0.6 0.17 4.4456 f0.0006 -5.34 0.04 0.001 1500.39 NaCI,, (0.09)b 1502.36 0.1 105.9 f0.8 0.20 4.4267 0.0008 -5.30 & 0.05 0.001 1502.5 1 NaCI,, (0.2)b 1509.25 _+ 0.1 105.3 _+ 0.7 0.20 4.3669 & 0.0006 -5.15 & 0.04 0.001 1509.36 NaCI,, (0.3)b 1515.60 f0.1 105.1 & 0.8 0.16 4.3124 f0.0007 -5.03 f 0.05 0.00 1 1515.70 * Standard deviation of the fit.Concentrations in rnol dm-,. 48 J. Chem. SOC.,Faraday Trans., 1996, Vol. 92 Table 3 Characteristic parameters for pure water, NaCl,, solutions and Et3N at 25 "C solvent K,,,/mm3 mol-I MPa-' " pure water 8.0876 NaCl,, (0.05 mol dm -3, 8.03 16 NaCl,, (0.09 mol dm -3, 7.9980 NaCl,, (0.2 mol dm -3, 7.8899 NaCl,, (0.3 mol dm-3) 7.7922 solute Eta 153.3 E,,,/mm3 K-' mol-' C,,,/J K-' mol-I K,,,/mm3 mol-' MPa-' 4.6475b 75.291' 8.175Sd 4.776" 75.14' 8.1221 4.580" 75.06' 8.1217 4.054" 74.78' 7.99 17 4.929' 74.53' 7.8894 179.849f 2 18.log 194.82h Derived from experimental K, values determined in this work, except for Et3N, which was calculated from experimental rcs and d values taken from ref. 20. Calculated from a = 257.21 x K-' and d = 997.045 kg m-3 taken from ref. 21. 'Ref. 21. Calculated from rcT = 4.52472 x MPa-' and d = 997.045 kg mP3 taken from ref. 21. " Calculated from data of ref. 18 and 19. Calculated from a = 12.85 x K-' derived from PVT data (ref. 22). Ref. 23. Calculated from K~ = 13.92 x MPa-' (ref. 24) and V, = 139.96 x low6m3 (ref. 11). leads to the following mixing rule GE = xA(G:, A + RT In xA) + xB(G:, B + RT In xB) (4) where GE, GL, A and G:, are the molar Gibbs energies of the ideal solution and of the pure A and B components, respec- tively.xA and xBare the corresponding mole fractions. First or second derivatives of eqn. (4),either with respect to temperature at constant pressure, or with respect to pressure at constant temperature, lead to a common mixing rule, expressed by eqn. (5), for a set of other molar properties, such as molar isothermal compression, KT,m, molar isobaric expansion, Ep,m, and molar isobaric heat capacity, Cp,m. QE = XA Q:, A + XB QL, B (5) In eqn. (5), QE represents K$ m, EL, or Cb: m, Q:, A and Q:,are the corresponding molar properties for the pure A and B components.The above definition of an ideal mixture does not permit the direct derivation of mixing rules for some other molar properties,16 such as molar isochoric heat capacity, Cv,m, and molar isentropic compression Ks,m. To overcome this problem, Benson and Kiyohara' considered that whenever a thermodynamic molar quantity could be expressed in terms of other molar quantities, whose ideal quantities for binary mix- tures obey eqn. (5), then this same expression must also provide a convenient definition for that same property in the ideal mixture. In the case of molar isentropic compression as 's, m = XT, m -TV(aP, m)*/CP, m (6) then, Ks,id = K$ -T(EL: m)2/C;: (7) Eqn. (7) thus represents a logical and thermodynamically con- sistent definition for the ideal isentropic compression, used by some author^.'^^' Excess molar isentropic compressions were then calculated from where Ks,m is the molar isentropic compression of the real solutions.KLtm were obtained using eqn. (5) and (7). E;,, and C;,, values for each solvent were derived from apparent molar volume and heat capacity data.'*,'' K+,mvalues were calcu- lated from experimental isentropic compressibilities of the sol- vents, presented in Table 3. All those calculated values, which characterise the NaCl,, solutions taken as solvents, are also summarised in Table 3, with the parameters used for pure water and pure Et,N. K:, values were least-squares fitted to Redlich-Kister equations of the form where c1 is a constant and xA and xB are the mole fractions of the solvents and of triethylamine, respectively.c1 parameters, the associated standard deviations and the standard devi- ations of the fits are presented in Table 4. In order to obtain limiting partial molar compressions of Et,N, in all the solvents studied, we have used the well known relati~nship,~~applicable to any extensive property, Q, of a specified quantity of solution where Qi is the partial molar property of the ith component. Applying eqn. (10)to Kg,m, we obtain K!?,Et3N = K:, m + (l -XE13NKdK:, m/dXEt3N) (1l) which gives us the composition dependence of excess partial molar isentropic compressions. Using eqn. (9) to calculate dK:, Jdx,,3N and substituting in eqn.(11)results in KF, Et3N = c1 -2XEt3N c1 + CIXitBN (12) leading to limiting excess partial molar isentropic compres- sions K:;Z3N = cl. Limiting partial molar isentropic compres- sions, KzEt3N, obtained by adding to the c1 values the molar compression of pure Et,N, are also presented in Table 4. The Kg values are all negative, indicating that the solvent near Et,N molecules at infinite dilution is more incompressible than in the bulk. In Part 1 of this work," regarding volumetric behaviour of these same solutions, we could show, by using a hard-sphere model, that the packing efficiency of the solvent around Et,N in pure water and in the NaCl,, solutions, was slightly less, in relation to that calcu- lated from the merely geometrical packing model used.The Table 4 c1 parameters of eqn. (9) and limiting partial molar com- pressions, KSq)E13N,for Et3N in pure water and in NaCl,, solutions at 25 "C c1 = K?Zq KzEt3N ofit /mm3 mol- /mm3 mol-' /mm3 mol-' solvent MPa-' MPa-' MPa-' pure water -194.5 0.2 -41.2 0.002 NaCl,, (0.05)lI -195.9 f0.3 -42.6 0.003 NaCl,, (0.09)' -205 k2 -51.7 0.02 NaCl,, (0.2)n -197.6 k0.8 -44.3 0.007 NaCl,, (0.3)" -188.5 f 0.2 -35.2 0.001 a Concentrations in rnol dm-3. J. Chem. SOC.,Faraday Trans., 1996, Vo1. 92 49 XEt3N 0 0.001 0.002 0.003 0.004 0.005 0.006 7- -30 0 +++ + + I 7 -401 8 00 0 -60 n z n n -80 a Fig. 2 Apparent molar isentropic compressions us. mole fraction for Et,N in water and in NaCI,, solution at 25 "C.Symbols as in Fig. 1. maximum effect of this packing inefficiency was observed for the 0.09 mol kg-' NaCl concentration. The present results add the information that water molecules are more firmly packed around the non-electrolyte molecule than in the bulk, for all the solvents studied. This finding is in agreement with several compressibility studies in non-electrolyte-H,O systems, which have been explained in terms of changes of the water structure near hydrocarbon chains, owing to their hydropho bici ty., 6-2 In relation to NaCl,, solvents, the probable existence in solution of three types of water structure, with high, medium and low compressiblity, makes the interpretation of the results more complex. In aqueous salt solutions it is known that ions exert a strong influence on the structure of water.Some ions, includ- ing Na+, tend to orient water dipoles in a radial pattern around it, producing electrostricted water of low compress- ibility. Others, such as C1-, tend to break the normal struc- ture of water, producing unbound water of high compre~sibility.~~Conversely, non-electrolyte molecules, such as alcohols and amines, have been described as promoting the ice-like structure of the water, making it less compressible than normal water.26-28 In our ternary systems there seem to exist two different types of promoted water structure, one around Na' ions and the other surrounding Et,N molecules. The formation of these two incompatible types of hydration sheets may cause a struc- tural repulsion when the sheets are brought near each other, enhancing the incompressibility of the solvent near Et,N.This conflict would intensify up to 0.09 mol kg-I NaCl, corre-sponding to the decrease in the KZEtJNvalues. In relation to the more concentrated solvents (0.2 and 0.3 mol kg- '), a spe- cific interaction between the lone pair of electrons of the amine group and the sodium ion may cause an overlap of the Et,N and Naf hydration sheets, leading to a significant increase in the compressibility of the solvent near the non- electrolyte molecule. This argument is the same as that used in the volumetric study to take into account the corresponding decrease in volume observed then. The composition dependence of partial or apparent molar properties is in general more subtle in detecting changes in molecular patterns, than the variation of the corresponding molar proper tie^.^' Starting from eqn.(11) we see that excess partial molar isentropic compressions can be obtained from the corresponding excess values, by the evaluation of dKF, m/dxEt3N derivatives. Local features of these properties are only noticeable if a reliable method is used to evaluate those derivatives at each data point.? Alternatively, apparent molar t This method should not introduce a significant level of smoothing in the original data. 50 J. Chem. SOC.,Faraday Trans., 1996, Vol. 92 compressions, K+,s,Et3N, based on point-to-point calculation, can be derived directly from the same excess values, using where KZ,Et3N is the molar compression of pure Et,N.In Fig. 2, we can see that, in the NaCl,, solvents, the addi- tion of Et,N across the more diluted region produces a more pronounced increase in the apparent molar compression, which may be due to a decrease in the H-bonded regions, probably motivated by a mechanism of cosphere overlapping. This effect is more noticeable at 0.09 mol kg-' NaCl. For the two remaining solvents and principally for the 0.3 mol kg-' solution we observe a stable structure of the water across the whole concentration range studied. Conclusions Ultrasonic velocity measurements in the ternary system Et,N-H,O-NaC1 permitted us to conclude that NaCl as well as Et,N decrease the isentropic compressibility of the pure water (see Fig.2). From the isentropic compressibility values, obtained for the different solutions, excess molar compressions were derived, and least-square fitted to Redlich-Kister equa- tions of one parameter only. Negative limiting partial molar isentropic compressions of Et,N were then obtained, appear- ing to indicate that the solvent surrounding the Et,N mol-ecule would present greater resistance to compression than in the bulk. These values increase in absolute value with the NaCl concentration up to 0.09 mol kg-', decreasing after- wards for the other two NaCl concentrations. The conflict between two non-compatible types of water structure, one more compact and less structured (produced by electro-striction around the ions) and the other less compact and more structured (induced by the non-polar alkyl groups of Et,N) may be the cause of the decrease in the limiting partial molar compression of the solute observed, up to 0.09 mol kg-NaCI.The increase of a probable long-range intermolec- ular electrostatic interaction, between the lone electron pair of the Et,N amine group and the Na' ion, in the two more con- centrated solvents, would explain the subsequent increase in K,"j E13N. Apparent molar isentropic compressions of Et,N were also derived from each experimental excess molar value for all the solvents studied. For the NaCl,, solvents in the more diluted region, there was an increase in the apparent molar compres- sion, probably caused by the decrease of the H-bonded regions.This is motivated by the increase in cosphere overlap- ping, while the amine concentration increases. The authors thank the Junta Nacional de Investigasiio Cientifica (JNICT) of Portugal for financial support. We are also grateful to a referee and to Prof. J. C. R. Reis for helpful comments and suggestions. References 1 L. H. LalibertC and B. E. Conway, J. Phys. Chem., 1970,74,4116. 2 S. Harada and T. Nakagawa, J. Solution Chem., 1979,8,267. 3 H. Hdand, J. Solution Chem., 1980,9,857. 4 J. P. Hershey, S. Sotolongo and F. J. Millero, J. Solution Chem., 1979, 12, 233. 5 S. Taniewska-Osinska, Chem. SOC.Rev., 1993,22,205. 6 G. C. Benson, C. J. Halpin and A. J.Treszczanowicz, J. Chem. Thermodyn., 1981, 13, 1175. 7 G. DouhCret, A. Pal and M. I. Davis, J. Chem. Thermodyn., 1990, 22,99. 8 B. E. Conway and R. E. Verral, J. Phys. Chem., 1966,70,3952. 9 10 M. V. Kaulgud, M. R. Awode and A. Shrivastava, Ind. J. Chem., 1980,19A, 295. I. M. S. Lampreia and L. A. V. Ferreira, J. Chem. SOC., Faraday 21 22 G. S. Kell, J. Chem. Eng. Data, 1975,20,97. H. Funke, M. Wetzel and A. Heintz, Pure Appl. Chem., 1989, 61, 1429. 11 Trans., 1993,89, 3761. E. F. G Barbosa and I. M. S. Lampreia, Can. J. Chem., 1986, 64, 23 L. G. Hepler, Z. S. Kooner, G. Roux-Desgranges and J. P. E. Grolier, J. Solution Chem., 1985, 14, 579. 12 13 387. F. J. Millero and T. Kubinsky, J. Acoust. SOC. Am., 1975,57, 312. V. A. Del Grosso and C.W. Mader, J. Acoust. SOC. Am., 1972,52, 1442. 24 25 J. A. Riddick and W. B. Bunger, Organic Solvents-Techniques of Chemistry, Wiley Interscience, 1970, vol 11, p. 45. W. E. Acree Jr., Thermodynamic Properties of Nonelectrolyte Solutions, Academic Press, Orlando, 1984. 14 G. Douheret and M. I. Davis, Chem. SOC. Rev., 1993,22,43. 26 S. Cabani, G. Conti and E. Matteoli, J. Solution Chem., 1979, 8, 15 J. C. R. Reis, J. Chem. SOC., Faraday Trans., 1982,78, 1595. 11. 16 17 G. Douheret, C. Moreau and A. Viallard, Fluid Phase Equilibria, 1985, 22, 277. G. C. Benson and 0. Kiyohara, J. Chem. Thermodyn, 1979, 11, 27 28 B. E. Conway and E. Ayranci, J. Chem. Thermodyn., 1988,20,9. G. R. Hedwing and H. Hoiland, J. Chem. Thermodyn., 1993, 25, 349. 18 19 1061. G. Perron, J. L. Fortier and J. E. Desnoyers, J. Chem. Ther- modyn., 1975, 7, 1177. G. C. Allred and E. M. Woolley, J. Chem. Thermodyn., 1981, 13, 29 30 F. J. Millero, in Water and Aqueous Solution: Structure, Ther- modynamics and Transport Processes, ed. R. A. Horne, Wiley- Interscience, New York, 1972,ch. 13. M. I. Davis, Chem. SOC. Rev., 1993,22, 127. 147. 20 A. Kumar, 0.Prakash and S. Prakash, J. Chem. Eng. Data, 1981, 26, 64. Paper 5/04973C; Received 26th July, 1995 J. Chem. SOC.,Faraday Trans., 1996, Voi. 92 51
ISSN:0956-5000
DOI:10.1039/FT9969200047
出版商:RSC
年代:1996
数据来源: RSC
|
10. |
Temperature dependence of the reactions of the nitrate radical with dichloroalkenes followed by LIF detection |
|
Journal of the Chemical Society, Faraday Transactions,
Volume 92,
Issue 1,
1996,
Page 53-58
Ernesto Martinez,
Preview
|
PDF (761KB)
|
|
摘要:
Temperature dependence of the reactions of the nitrate radical with dichloroalkenes followed by LIF detection Ernest0 Martinez,* *' Beatriz Cabafiast Alfonso Aranda' and Richard P. Wayneb Facultad de Quimicas, Universidad de Castilla-La Mancha, Campus Universitario sln, 13071, Ciudad Real, Spain Physical and Theoretical Chemistry Laboratory, University of Oxford, South Parks Road, Oxford, UK OX1 3QZ Absolute rate coefficients for the reaction of NO, with 1,l-dichloropropene, 2,3-dichloropropene and (E)-1,3-dichloropropene have been measured using the discharge-flow technique coupled to an LIF detection system for a range of temperatures from 296 to 430 K. The measured room-temperature rate constants are (1.52 f0.76) x (1.39 0.30) x and (0.95 & 0.14) x 10-l4 cm' molecule-' s-', respectively.The Arrhenius expressions k = (2.85 f1.21) x 10-'' exp[ -(2274 f313)/T], k = (0.16 0.02) x lo-'' exp[-(1420 f82)/T], k = (5.86 f1.60) x lo-'' exp[(-2575 4 192)/T] cm3 molecule-' s-', are pro- posed for the three reactions. The reactivity of alkenes containing vinylic halogen atoms is discussed and compared with that of simple alkenes. A depen- dence of room-temperature rate constants and energy of activation on haloalkene ionization potential, corrected for the strength of the mesomeric interaction between the chlorine atom and the carbon-carbon n bond, is found. Tropospheric half lives of these compounds have been estimated at night and during the day for typical NO, and OH tropospheric concentrations, in order to assess the lifetime of these compounds in the atmosphere.The NO, free radical is present in the Earth's atmosphere and may play an important role in the chemistry of the lower atmosphere.' NO, is mainly produced by the reaction of 0, with NO, and, as it is photolysed by sunlight, it can provide an active radical source only under dark conditions. Thus, the nitrate radical is involved in night-time atmospheric chem- istry, when photochemically generated radicals, e.g. OH, are present in very low concentrations. NO, has been found to react with many organic compounds' which, during daylight hours, are removed via reactions with OH and in some cases 0,. Reactions with NO, at night may constitute a significant loss process for volatile organic compounds, and are important oxidation routes for many unsaturated compounds. Some data have been recently published3-* on the reactions of the nitrate radical with haloalkenes.Nevertheless, informa- tion on the reactivity of NO, towards such compounds, and especially the influence of temperature, still remains scarce. In this paper, we report kinetic data for the reactions Cl,C=CH-CH, + NO, + products (1) (E)-ClCH=CH-CH,Cl + NO, + products (11) CH,=CCl-CH,C1 + NO, -,products (111) '32,3-Dichloropropene' (2,3-D) and 1,3-dichloropropene1 ' ' (1,3-D) have been reported to be toxic fumigants. The use of 1,3-D is widespread as it is the active ingredient currently used in nematocide fumigants (formulations range from 34 to 94% by volume).The reactions of 1,3-D with ozone12 and the OH radi~al'~,'~have been investigated, but no data have been published for the third important tropospheric oxidant, NO,. Reactions of 1,l-dichloropropene (1,l-D) with 0, and OH have also been ~tudied,'~ but in the liquid phase. Grosjean and Williams" attempted to estimate the rate coefficients for reaction (111) and the reaction between OH and 2,3-D in the gas phase, but they did not succeed because of the significant scatter in their data. Most kinetic measurements concerning NO, have been per- formed using optical absorption in the visible region. Since fluorescence excitation and laser-induced fluorescence (LIF) spectra were recorded,I6 the LIF technique has revealed itself as a highly sensitive method for monitoring NO, ; however, few reactions have yet been studied by LIF.17-19 Using a fast-flow discharge system with LIF detection, we have been able to detect NO, with excellent sensitivity and to study the reactions between the nitrate radical and (1,l-D), (2,3-D) and (1,3-D) in the low-pressure region from room tem- perature up to 430 K.Experimental All measurements of absolute rate constants were carried out under pseudo-first-order conditions in NO, , which was detected by monitoring pulsed-induced fluorescence excited at i= 662 nm. A Pyrex flow-tube reactor, 110 cm long (4.0 cm id), was pumped by a rotary pump (Telstar RS/70) of displacement 70 m3 h- ' which provided stable flow-velocities of between 600 and 1200 cm s-' at the pressures typically used in this work (1.1-1.3 Torr).Nitrate radicals were generated in a sidearm by reacting F atoms with HNO,. Fluorine atoms were obtained by passing F,-He mixtures (F, content: 5%) through a microwave dis- charge. The NO, radical was added to the flow tube 90 cm upstream from the cell through a fixed port. The rate of radical losses at the wall was estimated to be less than 0.1 s-' at all temperatures used, in the absence of haloalkenes. Dichloropropenes were added to the flow through a sliding injector of 0.8 cm od with a spray nozzle tip. Time resolution was provided by moving this injector, always from positions close to the cell to positions further away, thus preventing the introduction of a part of the injector previously in contact with the air, and consequently not degassed.In a given kinetic measurement for a set dichloroalkene concentration, [NO,] was measured for several reactant injec- tion distances (5-80 cm) from the cell, giving contact times between NO, and the reactant in the range 5-140 ms. At each of these positions, the NO, concentration in the absence of J. Chem. SOC.,Faraday Trans., 1996,92( l), 53-58 53 reactant was also measured just before and after the measure- ment with the reactant, in order to remove the influence of the different exposed injector surface on [NO,l0. The concentrations of organic reactants were calculated from their flow rates, using capillary flowmeters. Absolute concentrations of NO, were determined by chemical titration with a known amount of tetramethylethene (TME) in much the same way as described by Smith.20 We carried out these measurements in preliminary experiments using the cell as a multi-path absorption 'J'NO, concentrations in the range (0.5-2) x 1OI2 molecule cm-, were used, our LIF detec- tion limit being around 10" molecule For the present kinetic studies, absolute [NO,] values were required just to ensure that pseudo-first-order kinetic conditions applied.The flow-tube was heated by an electronically regulated electric tape that provided temperatures up to 500 K. The temperature profiles as a function of distance from the cell were obtained as reported by Canosa-Mas et aL2, In our design, the flow-tube and the cell were fused in one piece in order to allow heating close to the cell.All the gases entering the reactor were preheated upstream in the entrance region along at least 20 cm of flow-tube reactor length. The pressure in the flow-tube was measured with a capacitance manometer (Inficon-CM-120), 10 Torr full scale. Quantitative detection of NO, was achieved primarily by measuring the fluorescence from excitation of the 0-0 A 2E'+X 'A; transition which has a peak absorption at 1 = 662 nm. When excited at this wavelength, most of the fluorescence intensity is emitted at wavelengths longer than 710 nm.I6 Pulsed 662 nm output of a dye laser pumped by the second harmonic of a Nd : YAG laser (Scanmate 2C Lambda Physik) was collimated and transmitted through the fluores- cence cell.DCM dissolved in methanol was used as dye solu- tion, 2-5 mJ pulses of 0.105 cm-' bandwidth and 8 ns duration being obtained. The laser was operated at 10 or 15 Hz repetition frequency. The usual measures were taken to reduce the scattered light from the pumping beam. The red-shifted NO, fluorescence was passed through a filter set which transmitted only wave- lengths >695 nm and it was detected by a red-sensitive pho- tomultiplier tube (Hamamatsu R-928). The output of the photomultiplier was fed into a boxcar integrator (Stanford RS250) together with a trigger signal from the laser's Q- switch. Averaging was set to 300 samples. The gate used was 1.0 ps and the delay was set to 100 ns with respect to the laser spike.The fluorescence intensity was directly proportional to the concentration of NO, in the detection cell. The fluores- cence signal was a linear function of laser power with zero intercept over the range 0-5 mJ pulse-', showing that there was no saturation effect. Additional experiments were conducted in which the decay of fluorescence with time was measured directly. For these experiments, the boxcar integrator was replaced by a storage oscilloscope (100 MHz) and 256 decay traces were averaged. The measured lifetime for [He] = 4.2 x 1OI6 molecules cm-, and [HNO,] < 4 x 1014 molecules cm-3 was 1.3 ps. For these concentrations of quenchers, the data of Nelson et give a calculated lifetime of 1.6 ps, thus lending confidence in our experimental technique for the determination of [NO,].Materials Helium diluent with purity better than 99.999% (Carburos Metalicos C50) was passed through a Oxisorb (Messer Griesheim) trap followed by a molecular-sieve trap at room temperature. The sources of the dichloropropenes used and their stated purities were: 2,3-dichloropropene (99%, Fluka); (E)-1,3- dichloropropene (97%, Jansen); 1,1 -dichloropropene (96%, Aldrich). All the reactants were purified by successive trap-to- 54 J. Chem. SOC.,Faraday Trans., 1996, Vol. 92 trap distillations. Nitric acid vapour was obtained by passing He through a bubbler containing a 2 : 1 mixture of concen- trated sulfuric and nitric acids (P.A.quality, Panreac) at 252 K. Molecular fluorine was supplied by Praxair and Union Carbide mixed with helium (5% F,, 95% He). Results All the experiments were carried out under pseudo-first-order conditions with the [reactant]/[NOJ ratio always greater than 15. The integrated rate equation applying to these reac- tion conditions' 'is, In[NO,],/[NO,], = (k,[reactant))r = k,t (1) where [NO,l0 is the radical concentration observed at the detector in the absence of reactant after wall losses. [NO,], is the concentration in the presence of reactant when the sliding injector was positioned at a set distance from the centre of the cell; t is the time of reaction obtained from the average flow velocity and the distance. Plots of ln([NO,],/[NO,],) us.time (first-order plots) were obtained in accordance with eqn. (1). Fig. 1 shows a typical first-order plot for the reaction of NO, with 1,l-D. Linear behaviour was observed for the whole set of reactions studied, from which the first-order rate constants, k,, were obtained. The linearity and the absence of intercept, within experimental error, are consistent with a first-order gas-phase reaction. First-order rate coefficients were then plotted against reac- tant concentrations. Some typical plots for the reaction of NO, with 1,l-D at different temperatures can be seen in Fig. 2. Similar behaviour was observed for reactions (11) and (111) at all temperatures. Fig. 2 shows that it is possible to fit the experimental data to expressions such as k, = k,[reactant] + C (2) Our experimental technique, which normalizes to [NO,], at each injector position, would normally be expected not to give an intercept. The existence of an intercept thus suggests an additional loss process of NO, dependent on the presence of reactant, probably because of a change in k, (the global rate coefficient for wall loss of NO,) when the reactant is present.This behaviour has been found elsewhere2' for reac- tions of the OH radical. There is no evidence of curvature in the second-order plots. It therefore appears that the changes 0 [1,1-DCP]=1.32 [NO3Io=0.08 [1,1-DCP]=2.18 [N03],=0.08 0.5 -[l, l-DCP]=3.72 [NO3lo=O. 10 A [l.l-DCP]=6.97 0.4 -c+ -5 0.3 -z 'b 090.2 v c-0.1 -7 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 us Fig.1 First-order plots for the reaction of NO, with 1,l-dichloro-propene at 384 K. [1,1-D] and [NO,l0 in units of lOI3 molecule ~rn-~. T=355K T=384K T=419Km A T=429~ El~ 0 ll,~ili~ill~li Fig. 2 Second-order plots for the reaction of 1,l-D with NO, at dif- ferent temperatures in k, occur at reactant concentrations lower than those employed here, and there is no need to make any correction. Second-order rate constants, k, , summarized in Table 1, were calculated from least-squares analysis of the data. -29 7 -29.5 4 r: -31.5 C -32 I11111111 1111 1111IIII~II 1111 1111-32.5 All the reactions were investigated over the temperature range 296-430 K; it was seen that the rate constants were dependent on temperature.These rate coefficients were fitted to the Arrhenius expression. As an example, Fig. 3 shows the results for the reaction of NO, with 1,l-D. No curvature was observed in the Arrhenius plots of reactions (I), (11) or (111), suggesting that no change in the mechanism of the reactions occurs within the temperature range studied. A linear least- squares analysis of the data yields the activation energies and pre-exponential factors given in Table 1. Discussion As already stated, these are the first room-temperature rate coefficients and Arrhenius parameters reported for reactions (I), (11) and (111); the data thus complement the few studies lcarried out on the reactions of NO, radicals with chloro- alkenes., -8 Our results summarized in Table 1 are of the same order of magnitude as rate coefficients for compounds of similar structure (see Table 2).All the available data on the products and kinetics show that the initial step in the reaction of NO, with alkenes pro- ceeds predominantly via electrophilic addition of NO, to the carbon-carbon double bond to form a radical adduct. Atkinson et a!., and Aird et aL6 demonstrated that the number and position of chlorine atoms systematically influ- ence the rate of reaction. Halogen atoms not directly attached to the double bond reduce the rate of NO, addition by the electron-withdrawing effect. However, the presence of vinylic chlorine atoms enhances the reactivity because mesomeric effects lead to stabilization of the transition state.When both effects occur together, the positive mesomeric effect outweighs the negative inductive influence, as suggested by our results, compared with the rate constants for propene2, (0.93 x cm3 molecule-' s-l) and for 3-chloropropene3 (5.35 x cm3 molecule-' s-'). The presence of a C1 atom in allylic positions (2,3-D, 1,3-D) should mean a reduction in their reac- tivity towards NO, compared with propene (this is what happens in the case of 3-chloropropene). The mesomeric effect due to the C1 atom attached to the double bond compensates the reduction of reactivity in the case of 1,3-D (k = 0.95 x cm3 molecule-' s-') and outweighs it for 2,3-D (k = 1.39 x cm3 molecule-' s-*).The equation, log kNO3= 21.6 + 3.32 x log koH, proposed by Wayne et al.' relates the reactivity of NO, and OH when the addition of the radical to an alkene is the initial step of the Table 1Summary of the measured rate coefficients for reactions (I)-(111) dichloroalkene T /K [react ant] /lo'' molecule cm-, k2 cm' molecule-' s-' 4 /kJ mol-' /lo-'' A molecule-' cm3 s-' 1,l-dichloropropene 296 2.7-5.3 1.52 k0.76 323 1.9-8.5 2.71 k0.82 355 1.3-8.1 4.13 1.40 384 1.3-7.0 6.51 f1.24 18.9 f2.6 2.85 f1.21 406 2.1-4.9 11.6 k2.6 419 1.5-4.2 13.0 f 1.7 429 1.2-4.1 14.4 f3.2 2,3-dichloropropene 298 4.0- 13.3 1.39 k0.30 343 5.3-14.3 2.61 f0.36 378 5.2-13.8 3.26 k2.04 11.8 f0.7 0.16 f0.02 404 2.8-12.0 4.74 f1.82 425 1.4-9.6 5.87 k 1.06 1,3-dichloropropene 297 4.0-8.3 0.95 _+ 0.14 348 4.0-7.5 3.99 f0.78 393 3.5-6.5 8.96 k 1.82 21.4 1.6 5.86 f 1.60 432 1.7-6.2 14.6 f3.4 [NO,] = (0.5-2) x 1OI2 molecule ern-,.P, = 1.2-1.3 Torr. Errors quoted are f20. J. Chem. SOC., Faraday Trans., 1996, Vol. 92 55 Table 2 Arrhenius parameters for the reactions of NO, with alkenes and haloalkenes and ionization potentials for the organic reactants compound E,/eV ccz Ei/eV k,/10-l4 cm3 molecule-' s-' -1og(k/cm3 molecule-s -I) E,/kJ mol - A/i0-13 cm3 molecule -s -' ethene 10.62 - - 0.020 15.70 25.8 63 propene 2-methlypropene E-but-2-ene 10.10 9.75 9.65 --- --- 0.95 34 38 14.02 12.47 12.42 9.7 4.4 - 4.7 -25 but- 1-ene 10.18 - - 1.1 13.96 7.8 18 2-met h ylbu t-2-ene 2,3-dimethylbut-2-ene pent- 1 -ene hex- 1 -ene 9.38 9.13 10.03 10.03 ---- ---- 890 4500 1.6 1.6 1 1.05 10.35 13.80 13.80 --9.7 10.9 --8.3 11.2 3-chlorobut- 1 -ene 10.35 - - 0.30 14.52 16.6 24 1 -chlorobut-2-ene 10.01 - - 2.0 13.70 8.2 6.0 3-chlorometh yl-propene isoprene" 1, 1-dichloropropeneb (E)-1,3-dichl~ropropene~ 2,3-dichloropr~pene~ 3-bromobut- 1-ene 10.17 9.0 9.45 9.74 9.83 10.42 --0.48 0.45 0.46 - --9.97 10.47 10.48 - 2.5 65.2 1.52 0.95 1.39 0.40 13.60 12.19 13.82 14.02 13.86 14.40 10.6 3.74 18.9 21.4 11.8 - 16 30 285 586 16 - 4-bromobut- 1-ene 10.49 - - 0.50 14.30 - - 1-chlorobut-1-ene 9.58 0.49 10.05 1.2 13.60 - - 2-chlorobut- 1 -ene 9.70 0.49 10.17 1.7 13.77 - - 2-chlorobut-2-ene 9.40 0.54 9.69 11 12.96 - - 1-chloromethylpropene chloroethene 9.32 9.80 0.52 0.45 9.67 10.50 9.0 0.045 13.05 15.35 -- -- 1,l-dichloroethene 9.74 0.48 10.26 0.15 14.82 13.4 3.1 (2)-1 ,Zdichloroethene (E)- 1 ,2-dichlo roet hene trichloroethene 9.49 9.5 1 9.37 0.39 0.39 0.38 10.74 10.76 10.75 0.015 0.015 0.029 15.82 15.82 15.54 19.1 -- 3.4 -- tetrachloroethene 9.22 0.37 10.74 <0.0062 > 16.21 - - Parameters from ref. 7 except where marked otherwise; " ref.18; this work. k, given for T = 298 K. reaction (k,,, and ko, are the room-temperature rate con- Molecular-orbital calculations for our compounds have stants of the reactions for a given alkene with NO, and OH, been carried out with the HyperChem 3.0 package,27 which respectively).From our rate coefficient, kNO3,for reaction (11), executes quantum mechanical calculations at a semiempirical we should expect k,, = 1.87 x lo-'' cm3 molecule-' s-' at level. The package was used to implement the PM3 room temperature, which agrees with the published' experi-parametrization28 of atomic wavefunctions. In these calcu- mental value (koH = 1.44 x lo-'' cm3 molecule-' s-'). Based lations, the molecular geometries were fully optimized. Our on this agreement, it is now possible to predict k,, corre-estimates, together with the data reported by Marston et sponding to reactions (I)and (111) as koH(l,l-D)= 2.2 x lo-'' are summarized in Table 2. cm3 molecule-' s-', k,,(2,3-D) = 2.1 x lo-'' molecule-' It was not possible to fit our data for reactions of 1,l-D, cm3 s-',respectively 2,3-D and 1,3-D with NO, to eqn. (3); this behaviour is The rationalization of the reactivity of alkenes and their similar to that of other vinylic halogenated alkenes.The differ- derivatives in terms of molecular properties has been the ence between the logarithm of the measured and predicted subject of attention during the last few years. A strong corre- rate constant from eqn. (3) (A log k29,) was found to be depen- lation was demonstrated6 between room-temperature rate dent on ccc2 and the correlation coefficients for reactions of NO, with a series of alkenes and A log(k2,,/cm3 molecule- ' s-')their ionization potentials (Ei ,calculated using semiempirical methods), through the equation = 3.276 x 55.472 exp(-9.718zc:) (4) -10g(k,~,/crn~ molecule-' s-') = 3.276 EJeV -19.38 (3) was found, which combined with eqn.(3) gives the expression The ionization potential is a measure of the stability of the -log(k,,,/cm3 molecule-' s-') = 3.276 EyeV -19.38 (5) highest occupied molecular orbital (HOMO) of the halo-relating k29g to EI, the ionization potential suitably corrected alkene, the double-bond 7t orbital. According to Koopman's for the strength of the mesomeric interaction between the theorem, the energy of the HOMO is equal in magnitude and carbon-carbon 7t orbital and the adjacent chlorine atomopposite in sign to Ei. It is the HOMO of the alkene that orbitals in the form interacts with the singly occupied molecular orbital (SOMO) EyeV = EJeV + 55.472 exp(-9.718 zc:) (6)of the attacking radical to form the transition state.The higher the energy of the HOMO (i.e. the closer to the SOMO), The explanation for this relation may lie in the fact that addi- the stronger is the interaction, and the transition state is more tion to the double bond does not occur when the radical stabilized. approaches the region of the HOMO in the chlorine atom. Abbatt and Anderson26 and Aird et aL6 showed that the On the other hand, when NO, approaches the carbon atoms rate constants for addition of OH and NO, to vinylic-in the 7t bond, the addition is possible. Thus, when ccc2 is halogen-containing alkenes did not lie on their correlation high, the rate constant is expected to be high.For those com- lines when plotting logk,,, us. Ei. This deviation was pounds in which no mesomeric effect occurs, EI = Ei.The rationalized7 by taking into account the contribution of chlo- correlation between log k and EI is shown in Fig. (4). Eqn. (4) rine atoms (vinylic positioned) to the HOMO. To evaluate is slightly different from that proposed by Marston et ~l.,~but this contribution, Marston et obtained the normalized since it is based on a wider database for chlorinated alkenes, it coefficients of the C-atoms in the HOMO, ccc2, from PM3 represents a more reliable tool to predict quantitatively the calculations. reactivity of this kind of compound. 56 J. Chem. SOC.,Faraday Trans., 1996, Vol.92 '*1 27 -I 10 I I I I1 I I I I I I I1 I I I I1 I1 9 9.5 10 10.5 11 Ei'IeV Fig. 4 log k298 us. the corrected ionization potential, Ef;0,simple alkenes and chlorinated or brominated alkenes without halogen atoms in vinylic position; A, vinylic chlorinated alkenes studied pre- viously ; M,vinylic chlorinated alkenes studied in this work Marston et aL7 also found a correlation between activation energy (E,) and ionization potential for simple alkenes E,/kJ mol-' = 27.1 EJeV -263 (7) This expression cannot fit the results for reactions involving alkenes with vinylic chlorine. We have tried plotting all the available activation energy data for the reaction of nitrate radical with monoalkenes and their halogeno-derivatives us.EI (Fig. 5), and though we cannot categorically state the lin- earity of the correlation, a forced linear fit yields the result E,/kJ mol-' = 27.1 Ei/eV -263 (8) The data for compounds with vinylic-chlorine atoms (the two compounds, the data for which existed prior to this work,' and the three we report) fall near the correlation line for simple alkenes proposed by Marston et ~l.,~provided that our proposed corrected ionization potential, as defined in eqn. (6), is used. According to eqn. (8), negative barriers should be expected for corrected ionization potentials below 9.7 eV. Such activa- tion energies have usually been attributed to the intervention of a complex in the reaction pathway. The mechanism is thus not exactly the same as for reactions with positive barriers, so that it was not expected that correlation (8) would fit in these cases. We also believe that the correlation should not apply to addition reactions of NO, to alkenes with low Ef,though they follow the same reaction steps. The results for isoprene and (E)-but-2-ene show that linear behaviour is no longer observed, and the activation energy seems to become indepen- dent of Ef.Eqn. (8) therefore accounts for the trend of the (€)-but-2-eneis0prene ElEl 8 8.5 9 9.5 10 10.5 11 11.5 12 G'IeV Fig. 5 E, us. EI; 0,simple alkenes and chlorinated alkenes without vinylic chlorine atoms; A, vinylic chlorinated alkenes studied pre- viously; A, vinylic chlorinated alkenes studied in this work activation energy of both simple and chlorinated alkenes with the ionization potential corrected for the strength of the meso- meric interaction between the carbon-carbon n bond and any chlorine atom in a vinylic position when Ef is higher than 9.7 eV.An attempt was also made to seek a correlation between log A and either Ei or EI. However, no consistent trend was found. In view of the observed correlations between both log k298 and Efand between E, and EI, there must, of course, be a monotonic relation between log A and Ef. Our inability to discover this correlation probably results from the large errors associated with the determination of A in this type of kinetic experiment. In the context of night-time atmospheric chemistry, lifetimes of organic species based on reactions with NO, are needed and must be compared with the corresponding lifetimes for reactions with the OH radical and with 0, ,in order to evalu- ate whether or not NO, can provide significant or dominant loss processes for these species.Taking' the average atmo- spheric concentration of OH by day as lo6 molecule cm-, and for NO, at night as lo9 molecule ern-,, together with the rate constants for the reactions of dichloropropenes with OH and NO, measured or estimated in this work, it is possible to calculate the day-time half-life (In 2/(k,,[halopropeneJ)) as well as the half-life by night (In 2/(k,,,[halopropene])). The results are summarized in Table 3. As can be readily seen, similar lifetimes of the order of a few hours are expected from the reactions of NO, or OH for these haloalkenes.Our estimates are based on the assumption that sinks other than the reaction of the dichloroalkene with NO, at night can be neglected. To calculate the half-life by day, the reaction with OH has been considered the main loss route. Because of Table 3 Calculated half-lives compound kNO, a cm3 molecule-' s-l /lo-" kOH cm3 molecule-'s-' T ~ ~ ,day-time/ h zNo,, night-time/ h ~~ 1,l-dichloropropene 2,3-dichloropropene 1,3-dichloropropene 1.52 1.39 0.95 2.2b 2.1b 1.44' 8.8 9.2 13.4 12.7 13.9 20.3 ~~~~~~~~~~ ~ ~ ~ ~~ Measured, this work. Estimated (see text). From ref. 12. [NO,] = lo9molecule cm-,; [OH] = lo6 molecule cm-3. J. Chem.SOC.,Faraday Trans., 1996, VoZ. 92 57 the use of these dichloroalkenes, especially 1,3-D, as fumi- gants, their presence in the atmosphere is expected to show a seasonal behaviour. Our conclusions may be helpful in the evaluation of the effects of such compounds in the tropo- sphere. 8 9 10 11 B Cabaiias, G. Marston and R. P. Wayne, J. Chem. Soc., Faraday Trans., 1995,91, 1185. J. L. Daft, J. Agric. Food Chem., 1989, 37, 560. M. Leistra, A. E. Groen, S. J. H. Crum, L. J. T. van der Pas, Pestic. Sci., 1991, 31, 197. K. Hooper, J. Ladou, U. S. Rosenbaum, S. A. Book, Am. J. Znd. Med., 1992,22, 793. Conclusion 12 E. C. Tuazon, R. Atkinson, A. M. Winer and J. N. Pitts, Arch. Environ. Contam. Toxicol., 1984, 13,69 1. The reactions of NO, with 1,1 -dichloropropene, 2,3-dichloro- propene and (E)-1,3-dichloropropene have been studied ; room-temperature rate coefficients and Arrhenius parameters are reported for the first time.A good correlation has been found between room-temperature rate coefficients and ioniza- tion potentials and between energy activations and ionization potentials of the reactants species. 13 14 15 16 17 E. C. Tuazon, R. Atkinson, S. Aschmann, M. A. Goodman, A. M. Winer, Int. J. Chem. Kinet., 1988, 20, 241. S. J. Masten and J. Hoigne, Ozone: Sci. Eng., 1992, 14, 197. D. Grosjean and L. Williams 111, Atmos. Environ., Part A, 1992, 26, 1395. H. H. Nelson, L. Pasternack and J. R. McDonald, J. Phys. Chem., 1983,87, 1286. A. R. Ravishankara and R. L. Mauldin 111, J. Phys. Chem., 1985, 89,3144. 18 E.J. Dlugokencky and C. J. Howard, J. Phys. Chem., 1989, 93, We thank C. E. Canosa-Mas and G. Marston for helpful dis- cussion. 19 1091. E. J. Dlugokencky and C. J. Howard, J. Phys. Chem., 1988, 92, 1188. 20 S. J. Smith, D.Phil. Thesis, Oxford, 1989. References 21 C. Canosa-Mas, S. I. Smith, S. Toby and R. P. Wayne, J. Chem. R. P. Wayne, I. Barnes, P. Biggs, J. P. Burrows, C. E. Canosa- 22 Soc., Faraday Trans. 2,1988,84,247. C. E. Canosa-Mas, S. J. Smith, S. Toby and R. P. Wayne, J. Mas, J. Hjorth, G. Lebras, G. K. Moortgat, D. Perner, G. Poulet, Chem. Soc., Faraday Trans. 2,1988,84,263. G. Restelli and H. Sidebottom, Atmos. Environ., Part A, 1991, 25, 23 C. E. Canosa-Mas, S. J. Smith, S. J. Waygood and R. P. Wayne, 1. R. Atkinson, J. Phys. Chem., Re$ Data, 1991,20,459. 24 J. Chem. Soc., Faraday Trans. 2, 1991,87,3473. H. H. Nelson, L. Pasternack and J. R. McDonald, J. Chem. Phys., R. Atkinson, S. M. Aschmann and M. A. Goodman, Int. J. Chem. 1983,79,4279. Kinet., 1987, 19,299. Y.Anderson and E. Ljungstrom, Atmos. Environ., 1989,23, 1153. 25 26 D. J. Kinnison, D.Phi1. Thesis, Oxford, 1994. J. P. D. Abbatt and J. G. Anderson, J. Phys. Chem., 1991, 95, I. Wangberg, E. Ljungstrom, J. Hjorth and G. Ottobrini, J. Phys. 2382. Chem., 1990,94,8036. 27 Hyperchem 3.0, Autodesk Inc., 1993. S. Aird, C. E. Canosa-Mas, D. J. Cook, G. Marston, P. S. Monks, 28 J. J. Stewart, J. Comput. Chem., 1989, 10, 209. R. P. Wayne, E. Ljungstrom, J. Chem. Soc., Faraday Trans., 1992, 88, 1093. G. Marston, P. S. Monks, C. E. Canosa-Mas and R. P. Wayne, J. Chem. Soc., Faraday Trans., 1993,89,3899. Paper 5/04022A; Received 21st June, 1995 58 J. Chem. SOC.,Faraday Trans., 1996, Vol. 92
ISSN:0956-5000
DOI:10.1039/FT9969200053
出版商:RSC
年代:1996
数据来源: RSC
|
|