I R Spectra of Aqueous Cupric Ions
The Journal of Physical Chemistry, Vol. 82, No. 6, 1978
683
The Influence of Cupric Ions on the Infrared Spectrum of Water P. P. Sethna, Lary W. Plnkley,+ and Dudley Willlams" Oepatfment of Physics, Kansas State University, Manhattan, Kansas 66506 (Received October 20, 1977) Publication costs assisted by the Office of Naval Research
We have measured the spectral reflectance of aqueous solutions of several cupric salts under conditions of near-normal incidence. Values of the refractive index n(v)and absorption index h(v) have been determined from measured spectral reflectance by Kramers-Kronig phase-shift analysis. The influence of the Cu2+ions on the water spectrum is markedly different from the influences of other monovalent and divalent metallic ions studied earlier. In addition to the familiar liquid water bands, we have evidence for additional bands near 3080, 1660, and 930 cm-l, which we tentatively attribute to water molecules in quasi-stable\association with the Cu2+ions.
Introduction In an earlier study1 of cupric sulfate we found that certain bands associated with the water of crystallization in a single CuS04.5H20crystal also appeared in the spectra of aqueous solutions of cupric sulfate. In another study' we investigated the influence of dissolved divalent ions on the gross features of the water spectrum; the results indicated that the influence of CuC12on the water spectrum was markedly different from the influence of MgC12, although the sizes of the Cu2+and Mg2+ions are nearly the same. Although the early theoretical work of Bernal and Fowler3 as well as subsequent theoretical work had indicated that the influence of dissolved ions on the liquid water structure depends chiefly on the ratio of charge to ionic radius, we have found4 that in the case of the alkali and alkaline-earth halides the influence of negative ions is considerably greater than that of positive ions. Thus, the marked differences in the water spectrum produced by Cu2+and Mg2+constitutes something of an anomaly, which is the subject of the present study. Experimental Section The method involved in the present work makes use of quantitative measurement of spectral reflectance R(v) a t near-normal incidence. The spectral reflectance R(v),of solutions of cupric salts was determined by direct comparison with the spectral reflectance R(v), of water a t 27 "C. With r as the measured reflectance of the ratio of solution reflectance to water reflectance, the absolute value of R(v), = rR(v),. The values of R(v), are based on the values of R(v), obtained from tabulated values of the refractive index n(v) and the absorption index k ( v ) for water given in the recent summary article by Downing and william^.^ The tabulated values of these optical constants of water are based on extensive measurements of absolute spectral reflectance along with independent quantitative measurements of spectral absorption. Data Analysis Kramers-Kronig (KK) phase-shift analysis can be employed to obtain values of the refractive index n(v)and the absorption index h(v) in terms of the phase shift
where p(v') = [R(v')](~/~). The resulting values of h(v) for a 4.3 M solution of CuC12.2H20 are plotted as a function 'Present address: Lockheed Space and Missiles Co., Huntsville, Ala. 0022-3654/78/2082-0683$01 .OO/O
of wavenumber Y in Figure 1. The maximum value of k(v) in the vicinity of the vl, v3 water band is smaller than for the corresponding band in pure water; the absorption peak occurs at a lower wavenumber than in pure water and the shape of the low-frequency wing of the band is different from that of the water band. Less obvious differences between the solution spectrum and the water spectrum occur near the v2 band near 1650 cm-l and in the highwavenumber wing of the librational band UL near 500 cm-l. In the present study, we used a somewhat different procedure for comparing the solution spectra with that of water. The KK relationship expressed in eq 1is based on quite general relationships and there is no reason to doubt its validity. However, the limits of the integral include all frequencies and experimental investigations cover finite frequency ranges; in our case, the range covered is from 6800 to 350 cm-l. Outside the range in which we have actually made measurements (M), it is necessary to make a high-frequency extrapolation (HFE) and a low frequency extrapolation (LFE). Thus, for any solution the value of the phase shift 4 ( v ) , in eq 1 can be written
@(v), = @ ( V ) ~ L F E+ @ ( v ) ~ M-t @ ( V ) ~ H F E (2) where the subscripts indicate the contribution to r$(v), from the LFE, the HFE, and the actual range of measurement M for the solutions. Similarly, for water we may write the corresponding phase shift 4(v)w in the form
@(v)w = @(V)WLFE -t @(v)w + @(V)WHFE (3) By working with relatively dilute solutions, it is possible in good approximation to set ~ ( Y ) , L F E = +(v)WLFE and $(v),HFE = ~ ( v ) w H F E . Under these conditions, we can substract eq 3 from eq 2 and write A ( v ) = @ ( v ) s - $ ( v ) W = @ ( v ) s M - @(V)WM or
(4)
= @(V)w -t A ( v ) In this expression $(v)w a t 27 OC is well known from measurements of both reflection and absorption and A(v) is simply the difference between 4(v), and as determined by eq l with identical LFE and H F E and therefore involves only the differences between c $ ( Y ) , ~and @(v)WM based on the spectral range in which measurements were made. In terms of c$(v), the optical constants are obtained by the familiar relations n(v),= [l - R ( v ) ] / [ + l R(v) - 2R1I2(v) cos c$(v)]and h(v), = [-2R1I2(v)sin $(v)]/[l R(v)- 2R1l2(v) @(VIS
+
0 1978 American Chemical Society
P. P. Sethna, L. W. Pinkley and D. Willlams
0.1
0.0 5.0 4.5 4.0 3.5 3.0 2.5 2.0 15 1.0 0.5 0 0 WAVE NUMBER ( c r d x103
5 0 4.5 4.0 3.5 3.0 2.5 2.0
WVE NUM3ER
1.5 1.0 0.5 9.0
(cd
XI0
Figure 3. The spectral absorption index of cupric bromide solutions as compared with the spectral absorption index of water.
1
0.03 -
I
u
I
1 !
[
, 3
' ' 1
,
5.0 4.5 4.C 3.5 3.0 2.5 2.0
I
.&
WAVE NUM3ER ( c d
1.0 0.5 2.0 x13-
Flgure 2. The spectral absorption index of cupric chloride solutions as compared with the spectral absorption index of water.
cos 9(v)]. The value of h(v) for any absorber is directly proportional to the number of absorbers per unit volume and thus to the molar concentration of the absorber. In the present study we have determined h(v), for 0.5 and 1.0 M solutions of CuClz and CuBrz. In order to determine the effects of these solutes on the water spectrum, we can take the difference between h(v),for the solution and a k ( ~where ) ~ a is the ratio of the molar concentration of water in the solution to the molar concentration of water in normal liquid water; thus, a is simply the ratio of the number density of water molecules in the solution to the number density of water molecules in normal water. In order to summarize our work, we have computed the ratio of the difference Iz(v), - ah(v)w to the molar concentration M of the solute. Plots of this quantity [h(u),- ~ h ( v ) ~ ]are / Mgiven in Figures 2 and 3 for CuClz and CuBrz, respectively. In the plots in these figures the ordinate values are negative in regions where the solution is more transparent than water and positive in regions where the solution is more absorbing than water. The regions in which the solutions are more transparent usually involve shifts in the major bands of bulk water caused by the solute. The curves in the figures represent the mean values of the curves obtained with 0.5 and 1.0 M solutions of the salts. Throughout most of the spectrum the ordinates of the curves have an estimated uncertainty of approximately i0.003.
Results In the relative absorption curve shown in Figure 2 for the CuClzsolutions we note a small negative excursion near
3600 cm-l followed by a positive excursion near 3300 cm-l which we tentatively attribute to a shift of the vl, u3 fundamental band of bulk water. Similarly, there is a larger negative excursion near 700 cm-l followed by a large rise near 450 cm-l, which we similarly attribute to a shift of the librational band of bulk water. In addition to these excursions, there are three fairly well-defined absorption maxima in the curves. These maxima occur at 3080,1680, and 930 cm-l; the relative strength Sh(& dv of these bands are proportional to the CuClz concentration in the 0.5-1.0 M range. Similar well-defined maxima appear in the curves shown in Figure 3 for the CuBrz solutions a t 3080,1635, and 940 cm-l. In the relative absorption curve in Figure 3, there is a large negative excursion near 600 cm-l along with a large positive excursion near 450 cm-', which we tentatively attribute to a shift of the librational band of bulk water in the solution. There are no clearly defined excursions in the vicinity of the vl, v3 band appearing near 3400 cm-l in the spectrum of pure water. Smaller absorption maxima appear in the curves of Figures 2 and 3 between 3600 and 3200 cm-'. Other peaks appearing in the curves are so small that they fall within the i0.003 uncertainty mentioned earlier.
Discussion of Results In our earlier studies of acids6we found that well-defined positive peaks in a relative absorption curve similar to the ones in Figures 2 and 3 could be attributed to the H30+ ion, which is a well-defined entity in acid solutions. It is well known that the cupric compounds of present interest are regarded as hydrolyzing salts. In aqueous solution reactions of the type Cu2+t 4H,O t Cu(OH), t 2H,O+
are to be expected. However, the absorption bands in Figures 2 and 3 differ in both frequency and general contour from bands that have been attributed to the H30+ ion. On the basis of the known dissociation constants and measured pH values for the solutions, we find that only one out of every 200 Cu2+ions is in the form Cu(OH)+ or Cu(OH),. In view of this extremely low concentration, it is not surprising that the H30+bands are not observable in Figures 2 and 3. We suggest that the strong absorption peaks in the curves in Figures 2 and 3 can be attributed to water molecules associated with the Cu2+ ions in quasi-stable groups. In certain crystals the Cu2+ion is connected by ligands to neighboring oxygen atoms7 including those in the water of crystallization. In a study of the nuclear
Energy Surfaces of Hydrogen Bond Systems
The Journal of Physical Chemistry, Vol. 82, No. 6, 1978 685
TABLE I: Band Frequencies (cm-' ) of H,O Molecules Bulk water at 27 C Iceat-7°C CuCI, solution CuBr, solution CuSO, solution a
V l , v3
v2
3390 3260 3080
1640 1640 1680 1635 -1680 -1600 -1650
3080
3090
VI.
570 810
930 940 910a
CuSO4.5H,O crystal 3160 910a Overlapped by the v I fundamental of SO,z- at 870
cm-l.
magnetic resonance of I'O, Swift and Connick* have suggested that the ion CU(H20)62+exists in aqueous solutions of cupric salts and has a lifetime of the order of to s. In analogy to ligand bonds between Cu2+ and neighboring 0 atoms in crystals, the bonded 0 atoms form a distorted octahedron with four 0 atoms in the equatorial plane at a distance of approximately 2 A from the Cu2+ ion and with two axial 0 atoms a t a greater distance. We suggest that the peaks in Figures 2 and 3 are associated with water molecules bound to Cu2+ ions in quasi-stable groups. The broad bands near 3080 cm-l can be attributed to the overlapping vl, and v3 fundamentals of bound water molecules and the narrower bands near 1650 em-I can be attributed to the u2 fundamental of such molecules. The v2 band in bulk water is itself quite narrow as compared with the other water bands. Athough there is clear evidence for the presence of the v2 bands in Figures 2 and 3, the exact frequencies of the v2 bands may be somewhat in error as a result of the subtraction process involving two narrow bands. Similarly, we propose that the bands near 930 cm-l can be attributed to the librational motion vL of bound water molecules in the local lattice. In Table I, we list the frequencies of these bands as they appear in water, in ice, and in solutions containing cupric ions. We also include the frequencies of the bands ob-
served in a CuSO4*5H20crystal. The vl, v3, and vL bands of H 2 0 molecules associated with Cu2+ ions are significantly shifted from their positions in the bulk water and ice spectra. The frequencies of the weaker v2 band of H20 in various materials do not differ greatly from one another; further careful study of this band by transmission methods might be useful. It is tempting to suggest that the small absorption maxima in the 3600-3200-cm-' regions in Figures 2 and 3 and in the spectrum of CuS04 solutions1 may be attributed to H 2 0 molecules a t the axial positions in the C U ( H ~ O )octahedron. ~~+ However, because the observed absorption maxima are scarcely greater than our estimated uncertainty of f0.003, such a suggestion is probably not justified. If our suggestions regarding the bands in cupric solutions are correct, it is probably possible to subject the Cu(H20)2+ion or other complex to a normal coordinate analysis. Other modes of vibration of the groups involved would probably result in absorption bands in the far infrared. If our suggestions are correct, it is to be expected that ions of the other transition elements form similar types of association with water in aqueous solutions.
Acknowledgment. We express our appreciation to the Office of Naval Research for its generous support of this work. Our thanks also go to Professors J. D. Petersen and H. C. Moser for helpful discussions. References and Notes (1) P. P. Sethna, L. W. Pinkley, and D. Williams, J. Opt. Soc. Am-, 67, 499 (1977). (2) H. D. Downing and D. Williams, J . Phys. Chem., 80, 1950 (1976). (3) J. D. Bernal and R. H. Fowler, J . Chem. Phys., 1, 515 (1933). (4) P. Rhine, D. Williams, G. M. Hale, and M. R. Querry, J. Phys. Chem., 78, 238 (1974). (5) H. D. Downing and D. Williams, J. Geophys. Res., 80, 1656 (1975). (6) H. D. Downing and D. Williams, J . Phys. Chem., 80, 1640 (1976). (7) L. E. Orgel, "Transition-Metal Chemistry: Ligand Field Theory", Methuen, London, 1960, Chapter 4. (8) T. J. Swift and R. E. Connlck, J . Chem. Phys., 37, 307 (1962).
Polarizability, Proton Transfer, and Symmetry of Energy Surfaces of Phenol-n-Propylamine Hydrogen Bonds. Infrared Investigations Georg Zundel" and Anton Nagyrevil Physikallsch-Chemlsches Institut der Universitat, Thereslenstrasse 4 1, D-8 Munchen 2, West Germany (Received August 5, 1977)
Chlorophenol + n-propylamine systems (ratio 1:l)are studied in deuterioacetonitrile,pure and with the addition of water. The proton transfer equilibria in the OH-N + O--H+N hydrogen bonds are determined from bands of chlorophenol and n-propylamine molecules. Furthermore, an IR continuum indicates when these hydrogen bonds are easily polarizable. 50% proton transfer is observed with the water-free systems for ApK, = 3.25, i.e., when the pK, of the phenol is 3.25 values smaller than that of n-propylamine. Hence for ApK, = 3.25 the OH.-N O-.-H+N hydrogen bonds are largely symmetrical. An IR continuum indicates that these hydrogen bonds are easily polarizable in a relatively large ApK, region around ApK, = 3.25. In these polarizable hydrogen bonds double minimum energy surfaces are present. Water molecules shift the proton transfer equilibrium in favor of the polar O-.-H+N proton boundary structure.
I. Introduction Transfer of the proton in BlH-.B2 + B1--H+B2.hydrogen bonds has been studied earlier with various methods.1-268 Dielectric studies have S ~ O W & that ~ sigmoid Present address: DBpartment de Chimie, UniversitB de MontrBal, MontrBal, QuBbec, Canada. 0022-3654/78/2082-0685$0 1.OO/O
curves are found when the dipole moment change A p is plotted vs. ApK,, i.e., pKaB2~+ - pKaBIH. With phenoltriethylamine systems the dipole moment changes are very large with complete transfer. Ratajczak and Sobcyzk,' for instance, observed nearly 10 D. Similar sigmoid curves are observed when the absorbance of UV3 or IR bands of the acceptor or donor groups are plotted vs. ApK,, Le., the 0 1978 American
Chemical Society
