Triple Bond in the lo-, 11-, 12-, 13-, and 15-Position. The introduction of a triple bond into the lo-, 11-, 12-, 13-, or 15-position eliminates the occurrence of anomalous lines. The spectra of the 10- and 11-isomers are given in Figure 5,A and B. The observed lines are just those which would be predicted from a comparison with the spectra of unsaturated fatty acids containing a double bond in a similar position ( I ) . Except for the acid proton resonance the observed line positions in the spectra of the 10- and 15-isomers are nearly identical. The greatest difference between corresponding positions is on the order of 0.02 ppm. The proximity of the triple bond to the terminal methyl end of the chain in the 15-isomer causes the triplets due to the terminal group and due to the methylene groups adjacent to the triple bond to be sharper and more clearly discernible than the corresponding multiplets in the spectra of the lo-, 11-, 12-, and 13isomers.
10,12-0ctadecadiynoic acid yields a similar spectrum (Figure 3 4 in which no anomalous lines appear. The observed multiplet positions (Table I) are each slightly downfield from those of the other five isomers. The greatest difference between corresponding positions is 0.14 ppm for the methylene groups adjacent to the triple bonds. ACKNOWLEDGMENT
We are grateful to S. G. Morris for preparing transoctadec-11-en-9-ynoic acid and the two methyl esters, and to R. J. Szamborski for obtaining the spectra.
RECEIVED for review October 16, 1967. Accepted December 21, 1967. Mention of commercial products does not constitute an endorsement by the United States Department of Agriculture over others of a similar nature not mentioned.
N-Salicylideneglycinato Complexes A Comparison with Pyridoxal D. L. Leussing and Kyu Sun Bai Department of Chemistry, Ohio State UniGersity, Columbus, Ohio 43210
The compositions and stabilities of the soluble ternary complexes formed between salicylaldehyde, glycinate and M(II) [where M(II) = Mn(ll), Ni(ll), Cu(ll) and Zn(ll)] have been determined in aqueous solutions. Species of the type M(Sal)(Gly), M(Sal)(Gly); and M(Sal)2(Gly)’; have been found. Cu(ll) forms only Cu(Sal)(Gly). The greatly enhanced stabilities of the complexes which contain salicylaldehyde anion and glycinate ion in equimolar ratios indicate that the ligands are condensed as Schiff bases. UV-spectral data further support this line of evidence. The n-salicylideneglycinato complexes arem ore stable than those of pyridoxal, the difference increasing in the order Ni 5 Cu < Zn. This effect seems to arise from the different influences of the pyridine nitrogen on the Schiff base imine and phenolate groups combined with the differences in the metal ion coordination to these groups.
A SERIES of investigations is being conducted in these laboratories concerning the behavior of Schiff bases and their complexes in aqueous solution where hydrolysis of the Schiff base is a complicating factor which has inhibited research in this area. Of particular interest is the mechanism by which metal ions through the formation of intermediate Schiff base complexes catalyze transamination processes in simple ternary systems containing amino acids, oxo compounds, and metal ions. In enzymic transamination reactions pyridoxal serves as a cofactor. Potentiometric means have recently been employed to characterize the species formed in metal ion, pyridoxal, and amino acid systems ( I ) , and it was felt that quantitative differences observed between the behavior of pyridoxial and the parent compound, salicylaldehyde, would aid in interpreting the sources of pyridoxal reactivity. (1) D. L. Leussing and N. Huq, ANAL.CHEM., 38, 1388 (1966).
Pyridoxa I
There is ample evidence for the existence of Schiff bases formed between salicylaldehyde and amino acids. Schiff base complexes formed between Cu(I1) or Ni(II), salicylaldehyde, and esters of amino acids have been prepared and investigated by Pfeiffer and coworkers (2). Schiff base complexes formed between metal ions, salicylaldehyde, and the amino acids themselves have also been prepared (3-5). Eichhorn and Marchand (6) on the basis of spectrophotometric evidence have reported that a highly stable 1:l:l complex appears to be formed between salicylaldehyde, glycine, and copper(I1) in water-dioxane solutions but its stability was not determined. McIntire (7) has investigated the Schiff bases formed between aromatic aldehydes and amino acids. Schiff bases stabilized by a hydrogen bond between the phenolate group and the imine nitrogen were isolated. In spite of widespread interest, (2) . . P. Pfeiffer, W. Offermann. and H. Werner. J . Prukr. Chem.. 159. 313 (1942). (3) . , R. C. Burrows and J. C. Bailar. J . Am. Chem. SOC..88. 4150 (1966). (4) P. Ray and A. K. Mukherjee, J . Indian Chem. SOC.,27, 707 ( 1950). (5) A. Nakahara, Bull. Chem. SOC.Japan, 32,1195 (1959). (6) G. L. Eichhorn and N. D. Marchand, J . Am. Chem. SOC.,78, 2688 (1956). (7) F. C. McIntire, Zbid., 69, 1377 (1947). I
1
VOL. 40, NO. 3, MARCH 1968
,
,
575
however, there appears to be no quantitative information available on the stabilities of the Schiff bases of salicylaldehyde and their complexes in aqueous media. EXPERIMENTAL
Salicylaldehyde, Matheson, Coleman and Bell, stabilized with hydroquinone, was vacuum-distilled at a pressure of about 3 mm Hg and stored in a dark bottle at about - 10". C. Gas chromatography showed no detectable amount of impurity. Glycine, from the same source, was recrystallized from 1 :1 ethanol-water mixtures and vacuum dried; determination by the method of Dunn and Loshakoff (8) showed that the resulting product was 99.8 pure. Stock solutions of metal chlorides (or nitrates) were prepared and standardized according to accepted methods. Solutions of salicylaldehyde were freshly prepared prior to each experiment by directly weighing the reagent, dissolving, and diluting to give the desired concentration, Standard sodium glycinate solutions were prepared by dissolving weighed amounts of glycine in known volumes of standard NaOH solutions and diluting to the desired volume. If necessary the ionic strengths of these solutions were made up to 0.5 with KC1. Sodium salicylate solutions were prepared in the same way immediately before use. All the experiments were run at 25.0" C. The ionic strength was maintained constant at 0.5 by the addition of KC1. In the case of copper, nitrate media were used to circumvent the complicating factor of appreciable formation of chloro complexes. Determination of the stabilities of the binary Cu(I1) complexes in 0.5M KC1 indicates that chloride complexing under these conditions has only a minor effect, however (9). A Radiometer Type 25 pH meter with a Type GK2021C glass-calomel combination electrode was used in pH measurements. The system was standardized against the National Bureau of Standards phosphate and tartrate buffers. The absorption spectra were obtained using a Cary 15 spectrophotometer. The solubility of salicylaldehyde in 0.5M KC1 appears to be only slightly greater than 0.03M at 25" C. This placed an upper limit on the range of salicylaldehyde concentrations it was possible to investigate. Another reason for working with dilute solutions was the low solubilities of some of the metal complexes (probably the neutral species MS2 and some of the neutral Schiff base complexes.) In all the experiments with ferrous iron, the solutions were prepared with air-free water and were maintained under a nitrogen atmosphere. Titrations were performed with the rigorous exclusion of air. RESULTS
The pKa of salicylaldehyde at an ionic strength of 0.5, 25" was found to have a value of 8.22 5 .01. The acidity constants in this paper are reported as: K, = aH.B/HB where aHis the activity of hydrogen ion as determined from the pH measurement and the remaining species are expressed as concentration. Perrin (10) reports value of 8.34 i 0.1 at 20" C. For the protonation of glycine, the values determined earlier (11) were taken, pK1, = 2.46, pKz, = 9.70. The formation constant for the protonated Schiff base was determined by titrating solutions of HSal with standard NaGly. Depression of the titration curve below that expected for simple proton exchange between Sal- and Gly- in the region up to the (8) M. S. Dunn and A. Loshakoff, J. Biol. Chem., 113, 359 (1936). (9) K. S. Bai and D. L. Leussing, J . Am. Chem. SOC.,89, 6126
(1967). (10) D. D. Perrin, Nature, 182, 741 (1958). (11) D. L. Leussing and D. C . Schultz, J. Am. Chem. Soc., 86,4846 ( 1964). 576
ANALYTICAL CHEMISTRY
equivalence point indicated the formation of the Schiff base, Analysis of the data yielded Kaso = 3.3 x lo3 for the reaction, KHSQ
HSal f Gly-
CHSal. Gly-
+ H20
(1)
The formation constant for the unprotonated Schiff base, Sal.Gly2-, was determined employing spectrophotometric titrations similar to those described in the pyridoxal study. (I). A series of solutions were prepared containing 0.010M NaOH, 1.4 X lO-'M Nasal, varying amounts of NaGly, and sufficient KCl to bring the ionic strength to 0.5. The Schiff base formation constant was evaluated using the absorbance values at 377 mp (Sal- absorption maximum) and 340 mp (Sal.Gly2- absorption maximum.) After correcting for the presence of a small amount of HSal-Gly- a value 3.3M-l was obtained for K ~ B . Sal-
+ Gly-
KSB
Sal
Glyz-
+ H20
(2)
The formation constants for binary metal ion glycinate complexes were calculated from the results obtained by titrating solutions containing known concentrations of hydrochloric acid (0.05M) and metal ion (0.05M) with a standard sodium glycinate solution (1.00M). With Mn(II), it was found that precipitation of the hydrated oxide limited the pH range which could be examined. The salicylaldehyde complexes were determined by titrating solutions, 0.015 to 0.020M in salicylaldehyde, 0.005M in metal ion, with standard 0.04M NaOH. In spite of these lower metal ion levels, precipitation of the neutral bis(salicylaldehydato) metal ion complexes limited the range over which data could be acquired. With Ni(I1) the highest coordination ratio of salicylaldehyde to total metal ion which could be achieved was 0.36. Thus, the constant for the formation of Ni(Sal), is not known accurately. Fortunately, under the conditions where the Schiff base complexes predominate, the bis binary salicylaldehyde complexes are not important. The ternary systems were investigated by titrating solutions containing metal ion and salicylaldehyde with a standard 0.1M sodium glycinate solution. Concentration levels of 0.01M were employed with Ni(I1) and Cu(II), but to overcome solubility difficulties, the Mn(I1) and Zn(I1) levels were reduced to 0.005M. With each metal ion, three successive series of experiments were run where the salicylaldehyde level was one, two, and three times the concentration of the divalent metal ion. Preliminary experiments showed that equilibration times were slow, so batch-wise titrations were performed to obtain the final data. Aliquots of the probe solution were placed in serum bottles, and the desired volume of sodium glycinate was added to each. The bottles were then layered with nitrogen, stoppered, and equilibrated for the appropriate length of time. In the Ni(I1) and Cu(I1) experiments many of the solutions formed precipitates when the reagents were first mixed. Most of these precipitates dissolved on equilibration but nevertheless some precipitates remained, causing some titration curves to be incompletely defined. In spite of this difficulty, sufficient data were acquired in all systems except Fe(I1) to characterize the complexes. With Fe(I1) the precipitates did not dissolve until a good excess of NaGly and HSal had been added presumably forming the ion Fe(Sa1 * Gly)2*-. The Zn(I1) solutions were initially clear but all formed precipitates after standing about a week. Nevertheless, the solutions reached a metastable point equilibrium during which the pH was constant from about 2 hours after mixing until precipitation. From these data it was possible to calculate the solution
4.0
t
i
P 7.0
6.6
I
I
I
I
3.0
mi titrant
1
62
Figure 1. Titration of Ni(I1) and Cu(I1) salicylaldehyde mixtures with 0.0981Msodium glycinate Ni(II), 0.00975M; salicylaldehyde, 0.0100, 0.0220, and 0.0330M Cu(II), 0.01083M; salicylaldehyde, 0.0110, 0.0219, and 0.0328M Initial volume: 10.00 ml The solid lines are the theoretical curves
phase equilibrium constants. Equilibration times with other metal ions were: Mn(II), 2 hr; Cu(II), 5 hr (no precipitation), 20-39 hr (when initial precipitate redissolved); Ni(II), 72 hr (when initial precipitate redissolved). Computation. The formation constants of the ternary complexes were obtained in a manner similar to that described previously (1, 10). Mass balance relationships were set up as follows: Mtot
=
Mz+
+ MS+ + MSz + MG+ +
+ MGa- + MSG + MSzG- + MSGz- + MSzGz2= HzG+ + H G + G- + MG+ + 2MGz + 3MG3- + MSG + MSzG- + 2MSG2- + 2MSzGz’- + HSG- + SGZStat = HS + S- + MS+ + 2MS2 + MSG + 2MSzG- + MSGz- + 2MSzGz’- + HSG- + SG2Htot = 2HzG+ + H G + HS + HSG- + H+ - OHMGz
(3)
Gtot
v
I
0.4
0.6
(5)
(6)
The formation constants which were obtained are reported in Table I where the values are the cumulative constants, pij for the reactions, 8ii
(7)
I 0.8
I
1.0 ml N a G l y
I
I
I
1.2
1.4
1.6
Figure 2. Titration of Mn(I1) and Zn(I1) salicylaldehyde mixtures with 0.0980M sodium glycinate Zn(II), 0.00514M; salicylaldehyde, 0.00531, 0.0105, and 0.0159M Mn(II), 0.00506M; salicylaldehyde, 0.00533, 0.0106, and 0.0159M Initial volume: 10.00 ml The solid lines are the theoretical curves
Values for the pZl were found to have uncertainties equal to or greater than their absolute values and, as in the previous work ( I , 11)were therefore set equal to zero. The pij values reported in Table I are two to four orders of magnitude greater than those reported for the pyruvateglycinate investigation (11). In part, this difference reflects the increased interaction between the metal ion and salicylaldehyde which essentially arises from the differences between phenolate and carboxylate complexing. In addition, a large “rest effect” (12) is strikingly evident when one compares the step-wise constants for the formation of the ternary species with those for the formation of the binary species-i.e., for the reactions,
+ G l y e M(Sa1
Gly)
+ HzO
(8)
the constants are 1.5 to 2.6 log units greater than those for the reaction, MGly+
DISCUSSION
+ i Sal- + j Gly eMSaliGlyj + iHzO
r
MSal+
(4)
The total quantities on the left hand side are known from the nominal compositions of the solutions. Equilibrium was invoked by replacing the terms for the associated species on the right hand side by their formation constants. Because the binary constants were obtained separately, the only unknown constants in each set of experiments were those for the ternary metal ion complexes and these were obtained using a leastsquares curve-fitting procedure ( I , 11). Comparing the theoretical titration curves computed using the calculated values of the constants with the experimental points in Figures 1 and 2, it can be seen that good agreement has been achieved.
M2+
I 0.2
+ G 1 y - e M(Gly)z
(9)
The gain resulting from the condensation of the ligands to give the Schiff base is relatively small, although in the right direction, so the major portion of the effect must arise from the increased propensity of the metal ion for binding a tridentate ligand, the Schiff base, to binding two bidentate ligands. In addition to the large “rest effect”, further evidence that the ternary species are, indeed, Schiff base complexes arises from the highly stable complexes of the type M(Sal)2(Gly)zZ-. The metal ions investigated here have only very slight tendencies to give more than sixfold coordination. Therefore, either the ligands must undergo condensation to form the Schiff base, or some of the ligands must be bound monodentately. Complexes of this latter type have considerably lower stabilities than are observed here. The enhancement in the 1 :1 :1 complexes, calculated as log B11 - log Bo1 - log Blo, follows the order Mn(I1) > Zn(I1) > Ni(I1). This behavior is opposite to the normal order usually
(12) J. Bjerrum, “Metal Ammine Formation in Aqueous Solution,” P. Haase and Son, Copenhagen, 1941. VOL. 40, NO. 3, MARCH 1968
577
Table I. Logarithms of Formation Constants for Binary and Ternary Salicylaldehyde and Glycinate Complexes, p = 0.5, 25' C
Mn(I1) chloride Mn(Sal)+ 2.15; Mn(Gly)+ 2.65; Mn(Sal.Gly)o 7.26;
Mn(SalbO Mn(G1y)aO Mn(Sal.Gly)GIy-
4.0 4.7 9.15;
Ni(Salbo Ni(Glybo Ni(Sal.Gly).Gly-
6.5 10.48; 15.62;
Ni(Gly)*Ni(Sal+Gly), *-
14.0 18.89
Cu(Sa1)zo Cu(Gly)+
10.11 14.87;
Cu(Gl~)t-
15.3
Mn(Sa1. G l y p 13.04
Ni(II) Chloride Ni(Sal)+ Ni(Gly)+ Ni(Sal.Gly)o
3.58; 5.63; 10.75;
Cu(1I) Nitrate Cu(Sal)+ 5.36; Cu(Gly)+ 8.12; Cu(Sal.Gly)o 16.15 Zn(I1) Chloride Zn(Sal)+ Zn(Gly)+ Zn(Sal.Gly)o
Zn(Sal)ao 5.00 Zn(Gly)ao 9.11; Zn(Sal .Gly)-Gly- 13.42;
2.87; 4.88; 9.65;
observed for complex ion stabilities and in this case, possibly arises from ring strain. Owing to the constraints imposed by the bond angles of the unsaturated nitrogen, the five-membered chelate ring formed by the amino acid residue is strained (13). Such strain would exert a larger effect the smaller and/or the more coordinatively inflexible is the metal ion. The coordinating behavior of n-pyruvylideneglycinate was found (1) to be markedly similar to that of the tridentate ligand, iminodiacetate, which also binds through two oxygen atoms and one nitrogen atom, The derived constants in (13) D. L. Leussing and E. Hanna, J. Am. Chem. Sa.,88, 693 (1966).
Table
Zn(Glyh11.56 Zn(Sal.Gly),*- 16.73
Table I11 for the complexing of n-salicylideneglycinate are 2.0 log units higher than those reported for iminodiacetate in the cases of Ni(I1) and Zn(I1) (14, 15). However, Cu(I1) has a considerably greater affinity for the Schiff base than for iminodiacetate as is shown by a difference of 5.2 units in the logs of the respective constants. Furthermore, iminodiacetate forms a moderately stable bis complex with Cu(1I) (14) while nsalicylideneglycinato Cu(I1) shows very little tendency to coordinate another species. These differences in behavior (14) S. Chaberek and A. E. Martell, J. Am. Chem. Soc., 74, 5052 (1957). (15) K. Suzuki and K. Yamasaki, Naturwiss., 44, 396 (1957).
II. Positions of UV Band Maxima
Wavelengths are in millimicrons, the extinctions coefficients in parentheses have been reduced by a factor of IO4+ Salicylaldehyde, HSal Salicylaldehyde, in dioxane (16) Salicylaldehyde anion, SalN-Salicylideneglycinate,Sal GIy2- pH 12.0 N-Salicylideneglycinate,HSal Gly- pH 9.07 N-Salicylidenevalinate, HSal Val- pH 8.67 (in dioxane) (16) N-Salicylidenevalinate, Sal Val* pH 12 N-Salicylidenglycinato nickel(I1) N-Salicylidenglycinato zinc(I1)
-
578
0
ANALYTICAL CHEMISTRY
212(1.6)
... 23q =2)
...
... ...
...
225(2.3) 232(2.2)
25q1.2) 265(1.15) 265(0,70) 262 (0.79) 277( 1 .3) 277( 1 .4) 257(0.95), 278(0.81) 263(0.75) 260(sh), 275(sh) 267(0.97)
323(0.31) 32q0.36) 377(0.63) 357(0.56) 395(0.60) 39q0.67) 318(0.25)411(0.46) 363(0.54) 355(0.53) 351(0.58)
must certainly arise from the constraints imposed by the imine nitrogen which confines the bonding of the donor atoms of the Schiff base to three of the four strongly coordinating square planar sites of Cu(I1). With iminodiacetate, the flexibility afforded by the saturated nitrogen atom allows one of the carboxylate groups to coordinate to an axial site, thereby permitting two planar and one axial sites to be available for further coordination by another iminodiacetate ion. UV Absorption Spectra. Salicylaldehyde exhibits three bands in the UV region. The positions of the maxima are given in Table 11. The two long wave length bands are at* (16). Ionization of tributed to T - x2* and T - T ~ transitions the phenolic proton produces a shift to longer wavelengths with a slight intensification of these bands. Formation of the glycinate Schiff base, (Sal. Gly)2- produces a slight shift to shorter wavelengths. In our experiments, the short wavelength band was obscured by the absorption of the excess glycinate ion in the solution from which the spectrum was obtained. The two long wavelength bands are also observed in the protonated Schiff base, HSalo Gly-, but both of these bands undergo a large shift to longer wavelengths relative to the unprotonated Schiff base. A study of the absorption spectrum of protonated n-salicylidenevaline in dioxane reveals four bands, one pair of which is assigned to the enamine form while the other pair is assigned to the tautomeric ketimine (16). The absorption spectrum of the valine Schiff base in aqueous solutions shows two long wavelength maxima at practically the same wavelengths as the glycinate derivative. These bands correspond to the enamine form,
30,000-
-
2 9,000
-
28,000-
0 2
I
5
0 4
27,0000 3
-
26,000
c
c
OS
0 5
25,000
24,000
I
.5
O7 I
I
I
I
.6
.7
.8
.9
TT
IT*,^
I
1.0
units
Figure 3. The relationship between the long wavelength band maxima and the calculated P?T* energies 1. Salicylaldehyde 2. N-Salicylideneglycinate dianion 3. Salicylaldehyde anion 4. N-Pyridoxylideneglycinate dianion
Pyridoxal anion 6. N-Salicylideneglycinate,imine nitrogen protonated 7. N-Pyridoxylideneglycinate, imine nitrogen protonated 5.
Io
Tb
In aqueous solutions the zwitter ion (lb) is expected to be stabilized relative to (la) through solvation. Absence of appreciable absorption at 256 mp and 306 mp rules out the presence of the ketimine,
It Theoretical calculations using Hiickel Molecular Orbital theory for protonated salicylaldimines (17) verify the earlier (16) assignment and establish the following order in increasing energy of the long wavelength maximum: zwitter ion
> cation E anion > enol
The zwitter ion structure corresponds to l b and the enol corresponds to 11. Further calculations made in this study show that to a first approximation the observed shifts in the long wavelength
band positions in both the salicylaldehyde and pyridoxal systems can be described by the differences in the energies of the ground states and first excited states as calculated using Hiickel molecular orbital theory. Figure 3 shows the positions of the long wavelength band maxima plotted us. the calculated T energy differences. The points are seen to fall along a single line with relatively small deviations compared to the large structural changes between the compounds. The salicylaldehyde anion complexes of Ni(I1) and Cu(1I) show intense UV absorption bands which have maxima corresponding closely to those of salicylaldehyde (18). The ternary complexes, on the other hand, show intense bands in the region of the 357 mp band which is characteristic of Sal-Glyz-. (See Table 11) This observation further substantiates the formulation of the ternary species as Schiff base complexes. The small differences in the band positions of the complexes and free ligands compared to the large changes brought about by protonation do not necessarily imply that the influence of the metal ions on the ligands is considerably smaller than that of the proton. The effect may also arise from the higher coordination number of the metal ions. In the Huckel calculations electron pair donation by one atom to another is manifested by an increase in the value of the coulomb integral for the donor. As discussed above, increasing the coulomb integral for the imine nitrogen (or carbonyl oxygen) decreases the energy of the first electronic transition, whereas increasing
(16) D. Heinert and A. E. Martell, J. Am. Chem. SOC.,85, 183 (1963). (17) K. K. Chatterjee and B. E. Douglas, Spectrochim. Acta, 21, 1625 (1965).
(18) S. Basu and K. K. Chatterji, 2. Physik. Chem., 209, 370 (1958). VOL. 40, NO. 3, MARCH 1968
e
579
Table 111. A Comparison of Salicylaldehyde and Pyridoxal Proton Dissociation PKh HzM+ HSal
4.25 8.22
PKZ, 8.54
Binary Complexes (log constants)
Ni2+ + P y r
+
C U ~ + PyrZn2+ + Pyr-
e e 1.85
3.58
NiPyr+, CuPyr+,
2.32
+ Sal- 6.36 NiSal+ CUB++ Sal- eCuSal+ 2.87 Zn2+ + Sal- e ZnSal+ . Ni2+
ZnPyr+, Schiff Base Formation 0.48
+ Gly- 1.28 Sal . Glyl- + H20 Pyr- + Gly- ePyr . Glyz- + H20 Log K for the Reactions, M + SB MSB Sal-
M
Sal Gly2-
H+ Mn2+ Nit+ CU2+
11.21 6.78 10.27 15.67
e
Zn2+
9.17
Pyr . Gly210.86
...
9.02 14.7-14.1
7.15
Difference 0.35
*..
1.25 1.0-1.6
2.02
the integral for the phenol oxygen increases the energy separation. Performing one or the other of these changes corresponds to protonation at the site concerned. A metal ion that coordinates to both positions, however, induces increases in the coulomb integrals for both atoms. This gives rise to opposing factors which tend to cancel and yield little or no shift in the position of the absorption band. Thus, either weak or strong bidentate coordination has a relatively small effect on the band position. Because of this effect, the band position is not diagnostic of the actual values of the ligand to metal bond strengths, but rather is characteristic of the differences in bonding by the two donor groups. Although it was not possible to obtain data regarding the stability of Fe(I1) (Sal. Gly)zz-, its spectral and magnetic properties are interesting and worthy of mention. Solutions of this complex exhibit the intense red color characteristic of spin paired ferroin species. An absorption maximum at 550 mp with an extinction coefficient of 1.56 X l o 3 is shown by Fe(I1) (Sal. Gly),z-. This extinction coefficient is about onetenth as large as the values of the order of l o 4usually reported for ferroin complexes, however. Magnetic susceptibility measurements furthermore reveal that the Fe(I1) is predominately in the spin-free state. These results indicate that the ion Fe(I1) (Sal. Gly)? lies close to the cross-over between spin paired and spin free Fe(I1) (19) with the paramagnetic form comprising about 90 of the iron at 25' C. Similar behavior has also been reported for Fe(I1) Schiff base complexes formed between glyoxylate, or pyruvate, and primary amines (20). The coordination chemistry of salicylaldehyde and pyridoxal in aqueous media is summarized in Table 111. In order to compensate for differences in the free energies of formation of the Schiff bases, the constants presented in Table 111are for the reactions of the metal ions with uncoordinated Schiff base. These constants are obtained by dividing the formation con(19) R. J. P.Williams, J. Chem. SOC.,1955,137. (20) P. Krumholz, 9th International Conference on Coordination Chemistry, St. Moritz-Bad, Switzerland, September, 1966.
580
ANALYTICAL CHEMISTRY
stants given in Table I by the formation constants for the unprotonated Schiff bases. The data in Table I11 can be explained on the basis of the difference in the aromatic ring systems brought about by the replacement of ring =CH in salicylaldehyde by a more electronegative =N in pyridoxal. This substitution promotes the replacement of the aldehyde oxygen atom by a less electronegative nitrogen to form a more stable unprotonated Schiff base in the case of pyridoxal. Similarly, this replacement should also account for the observation that the basicity of the pyridoxal Schiff base is less than that of the salicylaldehyde Schiff base. The imine nitrogen atoms are indicated by the spectral data to be the bonding sites of the first proton taken up by the Schiff bases. Further evidence in favor of this assignment lies in the substantially higher protonation constants of the Schiff bases compared to those for the carboxyaldehyde anions. The data, then, show that the imine nitrogen of the pyridoxal Schiff base is less basic than that of the salicylaldehyde derivative, but quantitatively, the difference is not large, amounting to only 0.35 pK, unit. Accordingly,t hese imine sites are expected to coordinate metal ions to about the same extent with either ligand, pyridoxal most likely giving the smaller interaction. The pyridine ring nitrogen has a much larger effect on the basicity of the phenoxide ion which is in the meta position. For example, the pK, of phenol is 10.0 (21), whereas, Mason (22) reports a value of 8.36 for the ionization of the phenol group in the uncharged form of 3-hydroxypyridine.
-
-
0'; H+
\
--*
e
I
N
A similar difference should exist between the pK, values of the phenol group in the Schiff bases of salicylaldehyde and the neutral form of pyridoxal. (It should be noted here that the observed pyridoxal ~ K zvalue , given in Table I11 represents predominately the dissociation of a proton from the pyridine nitrogen of the zwitter ion rather than from the oxygen of the phenol.) Coordination of the phenolate group of the pyridoxal Schiff base is consequently expected to be substantially less than that of salicylaldehyde. These effects are also reflected in the calculated charge distributions where the difference between the phenolate groups is observed to be significantly greater than the difference between the imine groups. -.345 t.250 H,
C&N- CH2CO
