Dec. 5 , 19%
STABILITY ORDERSI N TRANSITION h~ETAL-1,10-PHENANTHROLINE COMPLEXES 6153
[CONTRIBUTIOS No. 726 FROM COLLEGE, AMESIOWA. WORK
THE WAS
INSTITUTE FOR ATOMIC RESEARCH AND DEPARTMENT OF CHEMISTRY, IOWA STATE PERFORMED I N THE AMESLABORATORY OF THE u.s. ATOMIC ENERGY COMMISSION]
Stability Orders in Transition Metal-1,lO-Phenanthroline Complexes1 BY CHARLES V. BANKSAND R. I. BYSTROFF~ RECEIVED MARCH4, 1959 Stability constants are measured for complexes of Fe(II), Co(II), Ni(II), Cu(I1) and Zn(I1) with some 5-substituted1,lO-phenanthrolines, and are tabulated with other constants reported in the literature. I t is found that the order of stabilities for the 1 : l complexes follows the expected series: Fe(I1) < Co(I1) < Si(I1) < Cu(I1) > Zn(I1). The 1 : 3 complexes follow an anomalous order which is attributed to non-uniformity in the metal ions with respect to the symmetry of the octahedral configuration. An important implication in this interpretation is that the trans-configuration of bis-( 1 , l O phenanthro1ine)-copper(I1) is sterically unstable.
Introduction Free energy relationships among the chelate compounds of structural similarity have been observed and used to illustrate effects such as those due to variations in ring size, steric factors and ~ o l v a t i o n . ~I n a few cases, data are available to illustrate the effect of a change in the metal ion for the same set of similar ligands, as, for example, the work of Van Uitert, et d4 The present work presents data on the stabilities of chelate compounds of Fe(II), Co(II), Ni(II), Cu(I1) and Zn(I1) with 1,lO-phenanthrolines and attempts to illustrate differences due to changes in the electronegativities of both the metal ion and the ligand.
MPhi = the 1 : l complex of the metal ion with the 1,lOphenanthroline in question Kdi = the over-all dissociation constant of the ith complex ki = the stepwise formation constant of the ith complex Ki = the over-all formation constant of the ith complex The formation function, as defined by Bjerrum,6 is given by equation 2
c 3
3
iKi(Ph)’
i(MPhi)
i = 1
where the subscript T indicates “total” in the system. A Experimental plot of -log(Ph) against % is termed the formation curve. Apparatus and Reagents.--Absorbance measurements In the case of most of the 1,lO-phenanthroline complexes, were made with these various spectrophotometers: Beck- the calculation of the constants from the formation curve man Model D U with photomultiplier attachment, Cary is accomplished by successive approximations using equaModel 12 and Cary Model 14. 811 p H measurements were tion 2 . made with a Beckman Model G or GS pH meter, using a Four distinctly different methods for measuring stability glass-calomel electrode system. Experimental work was constants were used. Three of these methods (partition, performed a t a thermostatically controlled room tempera- competition and titrimetric-pH measurement) give data ture of 25”. The various 1,lO-phenanthrolines used were as pertinent to formation curves; the fourth (spectrophotolisted: 1,lO-phenanthroline, 5-methyl-1,lO-phenanthro- metric) was used to measure the stability constants of only line and 5-nitro-l,lO-plienanthroline, obtained from the the 1:1 complexes. G. F. Smith Chemical Company, Columbus, Ohio; 5The partition method may be applied conveniently bebromo- and 5-chloro-l,lO-phenanthroline,prepared in this cause of the selective extraction of only the free base from an Laboratory. Each of the 1,lO-phenanthrolines was re- aqueous metal ion-metal complex-ligand system. The crystallized from benzene-petroleum ether mixtures until partition coefficient KD is defined as pure anhydrous compounds were obtained. Stock soluKD = (Ph)o/(Ph)w (3) tions of the 1,lO-phenanthrolines were prepared by weight and were stored in blackened bottles to retard light-cata- where the subscripts o and w denote the organic and aqueous lyzed decomposition. Fe(II), S i ( I I ) , Co(II), Cu(I1) and phases, respectively; it is easily measured spectrophotoZn(I1) stock solutions were prepared as the chlorides from metrically using the ultraviolet absorption band of the 1 , l O the pure metals or reagent-grade chloride salts. All solu- phenanthroline. I n a two-phase system containing 1 , l O tions were prepared by weight and analyzed for the metal phenanthroline, a single point on the formation curve can ion by standard methods. Other materials used were of re- best be obtained by the addition of sufficient metal ion to agent-grade quality. For the partition studies, 2-ethpl-l- complex one-half of the total 1,lO-phenanthroline in the hexanol or chloroform was used. Both were especially system. I n this manner, a direct spectrophotometric meastested for their ultraviolet transparency a t wave lengths urement of (Ph)o permits the calculation of both (Ph)w and Z from equations 3 and 4. above 250 mp. Measurement of Stability Constants.-The equilibrium PhT - ~ M = T [KDVO a d 1 -k (H)/Ks)](Ph)w (4) constants for the divalent transition metal complexes with are defined as In equation 4, v is the volume in liters of the phase indicated, the 5-substituted-1,lO-phenanthrolines K, is the acid dissociation constant of the 1,lO-phenanthroline in question as defined by equation 5 , and the absence of the parentheses about PhT and &‘IT indicates the units are moles. where Ks (HI( Ph)/( PhH) (5) = a divalent transition metal ion M Irving and Mellor6 have described in detail and applied Ph = an,y of the 5-substituted-l,l0-p11e1iax1tlirolines~ to transition metal-1,lO-phenanthroline complexes a method a s the free base for measuring stability constants based on the competition of Fe(I1) and another metal ion for the 1,lO-phenanthroline. (1) Abstracted from a portion of t h e P h . D . Thesis of R. I. Bystroff, The stability constants of the Fe(I1) complexes are preIowa S t a t e College, 1959. sumed known. Formation curve data are obtained by com( 2 ) E. 0. Lawrence Radiation Laboratory, Livermore, California. parison of solutions of corresponding absorbance, in which
+
(3) A. E. Martell and Ill. Calvin, “Chemistry of t h e Metal Chelate Compounds,” Prentice-Hall, Inc., Englewood Cliffs, N. J , 1952. (4) L. G. Van U i t e r t , W. C . Fernelius and E. E. Douglas, THIS JUURNAI., 7 6 , 2736 (1953).
( 5 ) J. Bjerrum, “ M e t a l Ammine Formation i n Aqueous Solution,” P. Haase and Son, Copenhagen 1941. (6) H . Irving a n d D. 11. hlellor, J . Chcm. S O L ,3457 (1055).
6134
CI1.4RLES
v. B A N K S AND K. I. BYSTROFF
the total 1,lO-phenanthroline and total metal ion concentrtitions are the known variables and the total Fe(I1) concentration is constant. The principal limitation of this competition method is the small range of the formation curve which can be obtained with sufficient accuracy. A variation of the method of Bjerrum6 was used in the present work. A solution of the ligand, of pH 5 PK., was prepared. -4metal chloride solution of approximately the same p H was prepared and used as titrant. The addition of metal ion to the ligand solution results in proton displacenient and a corresponding pH change, from which n may be obtained. If the subscript 1 refers to the ligand solution before the titration and 2 t o the same solution after the addition of vz V I ml. of the metal ion titrant, then Z is related to the difference between the hydrogen ion concentrations of the initial and titrated solutions according t o equation 6
-
VOl. 81
5-BrPh, (6.3 f 0.9) X These values are listed in Table I V , together with those measured by Brandt and Gullstrom7 by extrapolation of pH from dioxane-water mixtures. Copper( 11) with 1,lo-Phenanthro1ine.-The titrimetric PH method was applied to the measurement of the formation curve from % = 1.5 to 3.0. A typical titration was performed with the following conditions: v1 = 40 nil., ( P h ) T ! = 4.78 X lo-' M , ( H ) T = ~ 5.0 X ,If, and the ionic strength p , adjusted with KC1, = 0.1; for the titrant, (CU)T= 0.989 X X , /A = 0.1, and (H)tit, = 2.03 X lo-' M . The data and results of this titration are shown 111 Table I. Calculations involving equations 6 and 8 were made using pK, = 5.0. The uncertainty in ( P h ) increases as the reaction nears completion. This uncertainty is indicated in Fig. 1. TABLE I TITRIMETRIC-pHSTUDY
Ka(H)T' [(H)z- (H),]
*)"(
- K,
(1
- e)(l - Y)
(6)
The term T denotes the ratio (H)tit/(H)l, where (H),,t is the hydrogen ion concentration of the titrant. Ideally, I = 1, and the volume changes during the titration can be ignored, so that equation 6 reduces t o equation 7
The free-base concentration, (Ph), can be obtained froin equation 8
The ionic strength of the solution was constant throughout the titration, and the pH meter was calibrated against solutions of known hydrogen ion concentration at the same ionic strength, so that p H was a direct measure of concentration rather than activity. The over-all change in p H in the course of a titration, under optimum conditions, was seldom more than 0.3 unit, so that increased sensitivity of the Beckman Model GS p H meter was essential for acceptable data. The spectrophotometric method was applied t o the measurement of the acid dissociation constants of several 1,lophenanthrolines and the stability constants of the 1: 1 complexes of Zn( 11) and Cu(I1) with the l,l0-phenanthrolines. The degree of association a was measured and is defined by equation 9 for the case when excess metal or hydrogen ion is present. (9) Sdutions were used in which the metal ion or proton concen.trotions were the variables, while the total ligand concentration was constant. The degree of association may be expressed in terms of the absorbance A at XmSx of the complex A - €Phl(Ph)T a = (10) [ € X P h - €Ph]l(Ph)T where 1 is the cell length in cm., E is the molar absorptivity in liter mole-' cm.-', and (Ph)T is the total ligand concentration in mole per liter. It is related to the stepwise dissociation constant and t o IC, by equations 11 aud 12, respectively.
(12)
The pk', Values of 5-Nitro-, 5-Bromo- and 5-Chloro-1,lOpK, values of the 5-substitutedphenanthro1ine.-The 1,lo-phenanthrolines were obtained from spectrophotometric measurements. Series of solutions were prepared in which the ionic strength was constant a t 0.1 and the 5substituted-1,lO-phenanthroline concentrations were constant, while the pH was varied from 1.0 to 11.7. Absorbance data a t the two principal ultraviolet absorption bands of the 1,lO-phenanthrolines were used to calculate LY from equation 10 and K , from equation 12. The average 2f from four to iivc results C ' R C ~gave these K , values: $ 1 S O ~ P I ~( 4, 7 i 0.6) x 10-4, 5-cipl1, (6.7i 0.3)x 10-5,
OF THE c U ( 11)-
1,10-I'HENASTIIRO-
L I S E SI'STEY U!
- Ul,
ml
.
0.0 , 10 ,20 .30 .40
60
.cis .xu 1 00 1.10 1 '20
(Wz
x 10' 0 90 1.39 1.91 2.49 3.00 3.48 4.05 4,51 4 83 4.88 4.91
fl SIT? x 104
-11
0.00
..
O.08 1 30 1 . $I:?
2.48 (17 3.53 1. 01 4.34 4.37 4.41 I
.
L
- . -,, 6)
2.65 2.02 ".;3
2.44 0 ' 2 2 0; 1.60 1.65 1.53 I . -
( P h ) X 10:
6 .3
2.75 I 71
1.09 0.73 .40 .28 .15
,066 ,037 .0.43
The partition method was applied to the measurement of the formation curve from n = 0.5 t o 1.5. The partition coefficient for 1,lO-phenanthroline between chloroform and water has been measured and found to be 1040 & 30 by Grimes,8 and this value was used in the calculations using equations 3 and 4. Special care was taken to avoid errors due to chloroform evaporation. The hydrogen ion concentration and the ionic strength of the aqueous phase for all solutions were 0.0132 11.1. Higher chloride concentrations are not desirable because of the extraction of ion pairs. The phases were equilibrated for 1 hr. The 1,lO-phenanthroline which was extracted into the chloroform was determined spectrophotometrically, after evaporation of the chloroform and re-solution in 0.1 M hyd'rochloric acid. The molar hydrochloric acid is 30,100 liter absorptivity EphH in 0.1 mole-' cm.-' at 2 i l mp. Also the absorption spectra were taken in the chloroform media and the results compared favorably. The results are included in Fig. 1. The Partition of 5-Nitro-l , 10-phenanthro1ine.-The partition coefficients of 5-nitro-1,lO-phenanthroline between chloroform and water and between 2-ethyl-1-hexanol and water were measured spectrophotometrically. Pertinent molar absorptivities, measured independently, are €F--N02Ph.22,500 (266 m p ) , € 5 - N O ~ P h , ( C s R I ~ O H ) = 24,100 (266 (CHCIn) m p ) , and €6--h'OzPh.(H,o) = 21,300 (266 mp). I n water the wave length listed corresponds t o an isosbestic point; therefore, the absorbance is a measure of (5-1\'02Ph)~and the free-base concentration is obtained from equation 5 and the pH. The results for the partition with chloroform are different at pH 5 to 3.5 than a t pH 2. The average result for the higher pH values is K D = 880 =k 60 (six results), while at PH 2, k'~= 1700 & 100. T h e results for the partition with 2-ethyl-I-hexanol were obtained only at p€I 3.4 t o 6.5 and the average K D was 38.8 + 2.2. It was found in the studies involving Cu(I1) that, although the formation curve is somewhat insensitive to K D ,better agreement with alternate methods fnr the determination of the formation ciirre is obtained u,sing K D = 1700. GrimesQhas cvitlence that P h i H species exist a t fiH values close to pRa. species woiild teiid to cause the obsrrved discrepallei ,Sir] :~ii(lwould justiiy the use of the K1, iiimsiired i i i :Lcicl. (7) \V.TV. Rrandt and D. K. Gullstrom, T H I SJ O U R N A L . 74, 35333 (1952!. ( 8 ) 1'. G . Grimes, R1.S Thesis, Iowa State Collegc, 1050. ( Y ) P G . Grimes, I'h.D. Thesis, Iowa S t a t e College, lY58.
STABILITY ORDERSIN TRANSITION METAL-110-PHENANTHROLINE COMPLEXES 6155
Dec. 5, 1959
Copper(I1) with 5-Nitro-1,lO-phenanthr0line.-The competition method was applied to the measurement of the formation curve from Z = 1.1 to 1.5. Beyond these limits this method cannot give reliable data. For the tris-(5 nitro-1,lO-phenanthro1ine)-iron(I1) complex log Ka was taken to be 18.0, which is the average result Brandt and Gullstrom7 obtained from rate and equilibrium data. For the 1:1 complex log kl was found to be 5.06 in the present work. Three series of solutions were prepared, all of which were 6.19 X 10-6 M in Fe(I1). Within each series, (Ph)T M. The Cu(I1) concenwas varied from (3 to 35) X M for each series, retratiqns were (1, 5 and 10) X spectively. All solutions were maintained a t p = 0.1 and were buffered a t pH 4.0. The results are shown in Fig. 1. The partition method was applied to the measurement of the formation curve between Fi = 0.5 and 2.0. The (H) in the aqueous phase was 0.013 M and chloroform was used as the extracting solvent. The K Dand K. values used in calrespectively. The culations were 1700 and 4.7 X absorbance of the 5-nitro-1,lO-phenanthrolinewas measured in the chloroform media. The evaporation technique used with l,l0-phenanthroline gave erratic results and spectral changes with 5-nitro-1 ,lo-phenanthroline, probably because of a greater reactivity of the latter to carbene attack. 5-Nitro-l , 10-phenanthroline is also uniquely susceptible to attack by hydroxyl ion, a process observable spectrally at pH 11, and by non-aqueous titration with triethylammonium hydroxide. Theresultsare shown in Fig. 1, along with the data obtained by the competition method. Spectrophotometric measurements of log kl for the 1:l complex also were performed. A series of solutions was M, prepared in which (Ph)T was constant a t 2.81 X ( H ) T = 0.022 M,and p = 0.1. The Cu(I1) concentrations were varied from (1 to 14) X 10-6 M and a solution was prepared with a 100-fold excess Cu(I1) to 5-nitro-1,lO-phenanthroline in order to determine the molar absorptivity of the mono-( 5-nitro-1 ,lo-phenanthro1ine)-copper( 11)species. Absorbance changes were very small under optimum conditions because of the closeness of the molar absorptivities of the Cu-5-NOzPh (29,200) and the 5-NOzPhH (31,400) species a t 275 mfi, the wave length of maximum absorption of both species. The data and results are shown in Table I1 and are included in Fig. 1. Calculations were made using equation 11and two different values of K..
TABLE I1 SPECTROPHOTOMETRIC STUDY OF THE cU( II)-5-NITRO-1,10PHENANTHROLINE
(CU)T X 106
A, 275 mpb
0.00 2.96 4.94 6.92 9.89 14.83 247.0
0.883 .847 .839 .837 .836 .832 ,824
aC
SYSTEM^ kI X
lO-'d
kl X 10-7s
..
..
..
0.61 .74 .7X .80 .86
5.0 4.0 3.0 2.1 2.0
9.0 7.1 5.3 3.8 3.6
..
..
..
Av. 3 . 2 rt 1 . 0 5.8 rt 2 . 3 Cell length, 10 cm. Corrected for free Cu(I1) absorpCalculated using pK. = 3.33. tion. See equation 10. Calculated using pK, = 3.57. Zinc( 11) with 1,lO-Phenanthroline .-The formation curve from E = 1.0 to 3.0 was obtained by the titrimetric-pH method. The ligand solution was identical with that used for the corresponding formation curve with Cu(I1). The titrant was composed of (2n)T = 0.929 X 10-2 M , (H)$it = 1.183 X M and = 0.1. The points calculated from the data and equations 6 and 8 are shown in Fig. 1 together with points calculated from the data of Kolthoff, et u1.,l0and Yasuda, et al." Calculations were performed using pK. = 5.0. Zinc( 11) with 5-Nitro- and 5-Chloro-1,lo-phenanthroline. -The spectrophotometric method was used to measure the log kl for the complexes of Zn(I1) with &nitro- and 5(10) I. M. Kolthoff, D.
L. Leussing and T. S. Lee, THISJOURNAL,
73,390 (1951). (11) M. Yasuda, K . Sone and K. Yamasaki. J . P h y r . Chcm., 60, 1667 (195G).
T
Fig. 1.-Formation curves with Cu(I1) and Zn(II), Curve A, Cu(I1) and 5-NOtPh: 0, by partition; 8, by competition; X I by spectrophotometric measurements. Curve B, Zn(I1) with 1,lO-phenanthroline: c ) , by PH measurements; .I, data of Yasuda, et ~2."; CB, data of . ~ C, ~ Cu(I1) with l,l0-phenanthroKolthoff, et ~ 1 Curve line: calculated curve using log kl = 9.0, log k2 = 6.7, log kt = 5.1. Curve D, Cu(I1) with 1,lO-phenanthroline: calculated curve using log kl = 9.15, log kz = 6.65, log kr = 5.25; A, by pH measurements; 0, by partition; 0 , by competition, data of Irving and Me1lor.B
chloro-1,lO-phenanthroline. The data were obtaincd in a manner identical to that used for the spectrophotometric method with Cu( 11) and 5-nitro-1 ,lO-phenanthroline but were more extensive, being obtained a t three different wave lengths, 231, 275 and 300 mp for the 5-nitro-1,lO-phenanthroline, and 215, 230 and 274 mfi for the 5-chloro-1,lOphenanthroline studies. The results expressed in terms of 0.1) X lO-'and (4.8 f (1/K1) (1 (H)/K.) were (1.0 0.1) X 10-4 for the 5-nitro- and the 5-chloro-1,lO-phenanthroline complexes, respectively. Both studies were performed a t an ionic strength of 0.1. The acidities were ( H ) = 0.0112 M for the 5-nitro- and 0.022 M for the 5-chloro-1,10-phenanthroline study. The log kl values listed in Table IV are shown as calculated using the pK, values measured in this work and those measured by Brandt and Gullstrom.7 Nickel(I1) with 5-Nitro-1,lO-phenanthroline.-The partition method was used to obtain points on the formation curve from E = 0.5 to 2.5. The 2-ethyl-1-hexanol-water system was used, and the extracted free base was determined spectrophotometrically in the organic media at 266 mp. The points, shown in Fig. 2, are scattered to some extent. For the lowest ? values, i a spectral shift indicated possible extraction of some of the complex. The slowness
+
*
6156
CHARLES
v. BANKS AND R.I. BYSTROFF
I PARTITION STUDYOF
VOl. S l TABLE I11 THE CO(II)-~-NITI~O-~,~O-PIIE.L"-
THROLINE
A
0.876 2.63 5.40 11.0 24.5 34.6 56.6 64.1 66.4 107.7 avo
/
o
7
=
SYSTEM^
0.178 0.426 0.86 4.50 0.535 1.28 4.85 1.27 2.68 4.80 1.12 1.80 4.75 2.32 5.56 2.76 5.32 12.7 4.90 4.60 2.70 6.75 15.15 4.60 15.15 8.63 21.6 4.60 16.6 41.3 8.08 4.60 11.4 28.4 15.15 4.60 20.2 21.8 54.7 4.60 25 ml., = 50 ml., p = 0.1, K D = 40.
0.48 0.964 1.37 1.78 2.14 1.80 2.27 2.68 2.4d 2.54
zlw
The Stability of the Mono-(5-substituted-1,lO-phenanthro1ine)-iron (11) Complexes.-The spectrophotometric method described by Kolthoff, et aZ.,I2was applied t o the measurement of the 1: 1 complexes of Fe(I1j with 1,lOphenanthroline, 5-nitro-, 5-methyl- and 5-bromo-1,lOphenanthroline. The method requires the knowledge of the stabilities of the 1 : 3 complexes. XYith the exception of the 5-bromo-1 ,lo-phenanthroline complex, these were taken from the data of Brandt and G ~ l l s t r o m . ~The 1:2 complex with Fe(I1) m-as assumed to be negligible in all aqueous systems. Solutions were prepared with excess Fe(II), as described by Kolthoff,12and buffered at pH 4 with sodium formate and hydrochloric acid while maintaining the ionic strength at 0.1. The log kl values were obtained as a linear function of the corresponding log K 3values; as log K I for the 5-bromo-l,l0-phenanthroline complex has not been measured, log K 3 was inferred from the substituent effect data of Brandt and Gullstrom7 and the pKa measured in this work. The stability constant results Lire presented in Table IT.
Results Table IV summarizes the stability constants and PS,values pertinent to a comparison of the subFig. 2.-Formation curves with Co(I1) and X ( I I ) , stituent effect in transition metal-1,lO-phenanthroCurve A, Co(I1) and 5-NOzPh; 0, by partition with 2- line complexes. Constants listed without referethyl-1-hexanol; A, by partition with chloroform; 0 ,by ence are taken from the present work. Formation competition. Curve B, Co(I1) and l,l0-phenanthroline: curves calculated from the constants listed for the calculated from the data of Irving and Mellor.6p25 Curve C, Zn(II), Cu(II), Ni(I1) and Co(I1) complexes are Ki(I1) and 5-SOzPh: 0, by partition. Curve D, Ni(I1) shown in Figs. 1 and 2 in relation to the experiand 1,lO-phenanthroline: calculated from the data of mental data. In the case of Cu(I1) with 1,lOIrving and Mel10r.~~Curve E, Ni(I1) and l,l0-plienanthro- phenanthroline, two sets of constants are used to line: calculated from the data of Margerum.26 indicate the sensitivity of the constants to the experimental error. Both calculated curves can be of the formation reaction dictated that the solutions be said to fit the data. A close fit to the data by the equilibrated for 24 to 48 hr. The pH of the aqueous phases calculated formation curve for Ni(I1) with 5-ni\vas between 4.26 and 4.61. Until it is possible t o investitro-1,lO-phenanthrohe is not warranted, and the gate the nature of the spectral shift mentioned, the stability constants calculated from the present data shall remain stepwise constants are to be regarded as very apquestionable. proximate. The over-all constant, log K 3 , should Cobalt(I1) with 5-Nitro-1,lO-phenanthro1ine.-The partition method was applied t o obtain the entire formation be a lower limit if i t is found that the 1: 1 complex curve. 2-Ethyl-1-hexanol was used as the extracting sol- indeed extracts. The Co(I1) curve is drawn vent, and the 5-nitro-l,I0-phenanthrolinewas determined through the partition method data. The curve spectrophotometrically in the organic media at 266 mp. from the competition data is parallel to this curve T o illustrate the composition of solutions used in the partition studies, more complete data are given for this system in and would indicate log ki values 0.2 unit larger. The discrepancy is reasonable in light of the uncertainty Table 111 and Fig. 2. Part of the formation curve also was obtained by the parti- in log K3 for the tris-(5-nitro-l,l0-phenanthroli11e)tion method using the chloroform-water system. The iron(I1) complex and in the K Dused. aqueous phases were PH 4.0 t o 5.0 and p = 0.1. The K D The log k1 values for the Fe(I1) complexes are value used was 1700. The results are also shown in Fig. 2. The competition method was applied to this system in a related to pK, values by the equation log ki = manner similar to that for copper. The compositions of the 0.596 p K , 2.93, while the equation of the substitsolutions used were identical to those with Cu(I1) except uent effect for the 1:3 cotnplexcs, obtained from that higher concentrations of Co(IIj, ( 3 to 20) X lo-' dl, the data of Brandt and Gullstroi~i,~ is log K3 := 2.5 were necessary to compete qignificaiitly with the Fe(I1).
+
The results are shown, along with the experiineiit:il iincrrt : i i i i t v , in Fig. 2 .
(12) I . M . K ~ ~ l i l m f1)l , I . I.?t!~+iti:,, 72, 2 I Z ! (1'15111
:111,1
'I'
b
I
PV,
'I
iiii J O I
i~xii
Dec. 5 , 1959
pKa
STABILITY ORDERSIN TRANSITION METAL-1,10-PHENANTHROLINE COMPLEXES 6157
+
9.15. From the experimentally derived relationship log kl = log K 3 - 1.123, obtained for the 5-bromo-l,1O-phenanthroline-Fe(II)system, and the preceding substituent effect relationships, one may calculate values for log k ~ log , Ks, and pKa for the iron complexes and the free base of 5-bromo1,lO-phenanthroline, respectively. These are 5.45, 19.7 and 4.2. The pK, value is in agreement with the independent spectrophotometric measurement, and log KSis identical with that found for the tris(5-chloro- 1,10-phenanthroline)-iron (11) complex, which is reasonable in light of the closeness of the inductive constants for chlorine and bromine. These results give credence to the assumption of linear free energy relationships in these transition metal complexes. It should be pointed out that some discrepancies evident in Table IV undoubtedly are due to inadequate consideration of the limitations of the method used. I n this category fall the examples log k3 = 1.1 and log kl = 6.3 for Zn(I1) and Cu(II), respectively, with 1,10-phenanthroline. Discussion
28 26
4 -
22 24
20 18 I6 ,
14-
E z
12-
GI
10-
-
8-
.-
Y
64-
2-
0
1
1
1
1
0 2
4
6
8
1
IO log
1
/
12 14
l
16
1
1
1
1
18 20 2 2 24 26
CuPhi.
In Fig. 3 the log K i values for Zn(II), Ni(II), Fig. 3.-The substituent effect in various metal-l,lOCo(I1) and Fe(I1) are plotted against the corre- phenanthroline complexes relative to that of the correspondsponding log Ki values for the Cu(I1) complexes of ing Cu(I1) complex. 1,lO-phenanthroline and 5-nitro-1,lO-phenanthroline. I n this manner, both the relative substituent alent metal-nitrogen bonds and, therefore, 1:3 effects and the relative order of stabilities are illus- complexes with a high degree of symmetry. The first point must be discussed primarily in trated, as the number of ligands attached to the metals varies from 1 to 3. The order of stabilities terms of the ligand-field theory concepts which of the 1: 1 complexes follows the Irving-Williams have been described by Nyholm, Orgel and J#rgenseries,13 Fe < Co < N i < Cu > Zn, as expected. sen16 and Griffith and Orgel.17 It is understood The metal hybrid orbitals used in bonding have that under the action of a cubic field, the 3d orbienergies which parallel the second ionization po- tals are split in energy into a lower non-bonding tentials of the respective metals. The stabilities of triply degenerate set, de, which correspond to the the 1 : 3 complexes do not follow such an order, dzy, d,, and d,, gaseous metal ion orbitals, and a however. The substituent effects fall in the same higher set, d,, which correspond to the d,l and order as that of the stabilities for both the 1 : 1 and dxl-p orbitals. The latter orbitals then partici1:3 complexes. It may be possible that some cor- pate in the formation of a-molecular orbitals with n ligrelation exists between the relative stabilities and the appropriate ligand orbitals. The 12 and and metal ion electrons occupy the 3d24s4p3 the relative substituent effects, as theory would predict, l4 but present data are not sufficiently pre- bonding orbitals, the non-bonding 3d, and the ancise to illustrate this point conclusively. As a first tibonding 3d, orbitals, in that order. The octaheapproximation, the lines in Fig. 3 pass through the dral configuration can be said to be established by origin. Such an empirical correlation may prove the d2sp3hybrid orbital configuration, but the presuseful in predicting stabilities of metal complexes ence of de ( t 2 g ) and d, (eg) electrons are capable of whose constants may be difficult to measure di- distorting the symmetry by the Jahn-Teller effect. la rectly. I n this respect, Irving and Rossotti15 have I n general, an uneven distribution of electrons called attention to the value of correlative diagrams among the three de or the two d, orbitals will result of various types, similar to Fig. 3, which can be in minor distortions in the former case and sizable distortions if the antibonding d, orbitals are inuseful to the analytical and inorganic chemist. volved. The Cu(I1) hexacovalent complexes, with The question of the unusual order of stabilities of the 1:3 complexes may be better understood a d:d2,d\ configuration, exhibit the Jahn-Teller qualitatively if two points are considered: (1) the effect to an extent which varies with the basic metal ions in question exhibit more pronounced dif- strength of the ligands. The presence of three elecferences in orbital character than has been assumed trons in the two antibonding d, orbitals results in a in regarding hexacovalent complexes as perfect oc- repulsive force on the axially located ligands. The tahedra, and (2) the fixed N-N distance and angu- resulting tendency is that Cu(I1) forms square lar orientation of the orbitals in 1,lO-phenanthro- planar complexes with weakly held axial monodenline suggest that this ligand will tend to form equiv- tate ligands, so much so that 1:3 Cu(I1) complexes
+
(13) H.Irving a n d R. J. P. Williams, J . C h o n . Soc., 3192 (1953). (14) L. G.Van Uitert a n d W. C. Fernelius. THISJ O U R N A L , 76, 379 (1954) (15) F T . I r v i n g :t11,1 I T , IIrissritti, A f f o C k i n , S f c i j i d , 10, 7 2 (lCl:l\).
(16) R. S. h'yholm, L. E. Orgel a n d C. K. J#rgensen. "Quelques Problemes d e Chimie Minerale," Reports of t h e T e n t h Solvay Confereuce. K. Stonps, Rruxelles, 1956. (17) J. S. C ~ I - i l i i l hu n d I,. I ? . Orgcl. O u t i i t h'iZJ\ , 11. 381 flLJ:i), (18) 1.. If:. Orgrl, .I ('Iwiii. Crir , 4761'8 (1!4:2).
6158
CHARLES V. BANKSAND IC. I. BYSTROFF
of bidentate ligands, such as C ~ ( e n ) 3are ,~~ unstable with respect t o the square planar complex, such as Cu(en)z(HzO)z. J@rgensenZ0has used a spectral criterion for the tetragonal character of the Cu(I1) complexes, namely, the magnitude of the t a g - eg transition in the Cu(I1) relative to that in the Ni(11) complexes. His data indicate that Cu(Ph)3, Cu (Dipy)3, 21 Cu (Dipy)z(Hz0)2 and Cu (Phlz(Hz0)2 are very nearly perfect octahedra, in contrast to many other complexes with both mono- and bidentate ligands. It is evident that the two 1 : 2 complexes named do not have coplanar nitrogen atoms but appear to be primarily of the cis configuration. This observation is borne out by the stability constants of the 1,IO-phenanthroline complexes of Cu(I1) in Table IV. The 1 : 2 complex is so much
VOl. 81
5-Bromo-1,10-phenanthroline, p K , : 4.20 B Fe(I1)
5.450 B
..
..
19.7"
5-Chloro-l,lO-phenanthroline,pK.: 4.26 A,' 4.18 B Zn(I1) Fe(I1)
5.85 5.72
..
B B
..
..
..
.. ..
,.
..
..
1 9 . 7 D,B'
5-Methyl-l,lO-phenanthroline, pK,: 52.6 A," 5.23 A' Zn(I1)
..
Ni(I1)
8.7
Fe(I1)
6.05 B
6.0 B2e
,
,
.
.
A"
5.0
..
..
A11
..
..
223
B
Values stated without ref. are from this work, in which ionic strengths are all 0.1 M in KCl, except for log kl for Cu(I1) with 1,lO-phenanthroline for which fi = 0.0132. bMethods are coded: A , pH measurements; B, spectrophotometric equilibrium; C, partition; D, kinetics; E, competition, F, conductivity. c Methods and references listed for direct measurements only. Calculated using pK, = 3.57. "Calculated using PK, = 3.33. 'Taken from a formation curve drawn parallel to partition data. g Computed indirectly using the substituent effect.
TABLE IV 25' OF SOME 'rRANSITION METALS less stable than the 1 : 1 complex that i t is highly un~-SURSTITUTED1,10-PHENANTHROLINES likely that the 1:2 complex could be of the trans
S I A ~ I 1 , I T l 'CUSSTANTS AT \VITII
hletal
log
k1Q.b
log
kz0.b
log
1,lO-Phenanthroline, PK,: 4.77 5.02 A3I Zn(I1)
5.64 A 5 . 7 2 C'o
5.2 A 4 . 8 5 Clo
17.0 17.0
A,C
C35
8.0
CIS
C"
6 . 7 0 C's
1.1 4.8 5.25 5.1 5.0 5.5 7.55 7.9 6.38
13.1
As' Bze
5.5 5.9 6.65 6.7 6.57 6.15 8.1
6.35 A 6 . 4 3 EL0 6.B C'o 6 . 4 7 B2'
.. Cu(I1)
Xi(I1)
Co(I1) Fe(1I)
9.15 9.0 8.8 6.3 8.6 8.0 7.02 5.85 5.S5
log
k8a.h
C C C25
..
B
..
BIZ
,
B2'
A" A.C C25
A%* I326
.
K3C
4.96 F,2832g 4.857 A,30
..
..
B24
A" A
A C25
A" BZ' Cu C*
..
.. ..
21.05 20.9 20.4 17.9 24.25 23.9 20.1 21.3 B'*'#*S 21.5 DlsfaizO
5-Sitro-l,lO-phenanthroline,pK,: 3.57 A,' 3.33 B Zn(I1)
Cu(I1)
hTi(II) Co(I1) IPe(I1)
5.gd R 5.4' B 8.0 C 7.8d B 7.6' B 7.57 B m 7.0 C 6.25 C 6.41'R 5 . 0 6 I3
..
..
.. 5 . 4 7 C,E
.. ..
.
.
6.4 C 5.41 C L60' E
..
..
.. .. 4.2 ..
.. ..
C
.. ..
7.0 C '4.63 C 4.82' E
..
..
17.7
*.
.. ..
20.4 16.3 16.8 1 8 . 1 D' 1 7 . 8 B'
(19) en represents cthylenediamiiie. (20) C. K . Jdrgensen. Acta Chem. S c n i f d , 9, 1362 (1955). (21) Dipy represents 2,2'-dipyridy!. (22) Cf.ref. 16. pp. 311-342. (23) R. T . PRaum and W. W. Brandt. Turs J O U R N A L , 76, 6215 (10543. (24) J. FI. McClure, Ph.D. Thesis, Iowa S t a t e College, 1951. (2:) I€. Irving and D. H. Mellor, private communication. (26) D . W.Margerum, Ph.D. Thesis. Iowa S t a t e College, 1955. (27) J . M . Duncan, M.S. Thesis, Purdue University, 1952. (28) T. S . Lee, I. hl. Kolthoff and D. L. Leussing, TEISJ O U R N A L , 70, 2348 (1948). (20) T. S. Lee, I , M . Kolthoff a n d D . L . Leussing, i b i d . , 70, 3596 (1948).
configuration, which would, like Cu(en)z(HzO)z, have closely similar stepwise stability constants. Adoption of the cis configuration, on the other hand, could conceivably suppress the Jahn-Teller effect because of the steric rigidity of 1,lO-phenanthroline. The refusal of the additional stability generally acquired through the Jahn-Teller effect and its concomitant octahedral distortion can only appear as a decreased over-all stability of the 1: 2 and 1:3 complexes and a less distorted octahedral symmetry. While this picture will explain the unusual stability order of the 1:3 complexes discussed earlier, it does not suggest the reason why the trans-Cu(Ph)z(HzO)zcomplex is unstable. It is proposed that the hydrogen atoms present in the 2- and 9-positions in 1,lO-phenanthroline and in the 3- and 3'-positions in 2,2'-bipyridyl are capable of sterically interfering with the adoption of the trans configuration. I t should be noted in this respect that such steric effects would not be expected in the Fe(II), Co(II), Ni(I1) and Zn(I1) complexes, because the Jahn-Teller effect should be minor or non-existent for the configurations dBd",dO,, d:d',d$, d6,d$dl, and d!d2,d$, respectively. Furthermore, Cu(I1) would present the smallest ionic radius in this series, although no data are yet available for the 1,lO-phenanthroline complexes, making steric considerations plausible. Orge116,22 has suggested an alternative possibility that Tbonding could stabilize the cis configuration much more than the trans in Cu(Ph)?(HzO)z, but this conclusion was based in part on stability cons t a n t which ~ ~ ~ now can be regarded as erroneous. (30) R. XasZinen a n d E. Uusitalo, Suomen Kemislilchti, 29B, 1 1 (1956). (31) R . Riccardi and P. Franzosini, Boll. sci. fac. chim. ind., Bologna. 16, 25 (1967).
