125
DIRECT CALORIMETRIC DETERMINATION OF HEATS OF FORMATION OF SOME METAL CHELATES Item Type text; Dissertation-Reproduction (electronic) Authors Gutnikov, George, 1938- Publisher The University of Arizona. Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author. Download date 09/06/2021 15:42:49 Link to Item http://hdl.handle.net/10150/284815

DIRECT CALORIMETRIC DETERMINATION OF HEATS OF ......The heats of reactions of several chelating agents containing oxygen and sulfur donor atoms with a number of transition and heavy

  • Upload
    others

  • View
    4

  • Download
    0

Embed Size (px)

Citation preview

  • DIRECT CALORIMETRIC DETERMINATION OFHEATS OF FORMATION OF SOME METAL CHELATES

    Item Type text; Dissertation-Reproduction (electronic)

    Authors Gutnikov, George, 1938-

    Publisher The University of Arizona.

    Rights Copyright © is held by the author. Digital access to this materialis made possible by the University Libraries, University of Arizona.Further transmission, reproduction or presentation (such aspublic display or performance) of protected items is prohibitedexcept with permission of the author.

    Download date 09/06/2021 15:42:49

    Link to Item http://hdl.handle.net/10150/284815

    http://hdl.handle.net/10150/284815

  • This dissertation has been microfilmed exactly as received 67-3956

    GUTNIKOV, George, 1938-DIRECT CALORIMETRIC DETERMINATION OF HEATS OF FORMATION OF SOME METAL CHELATES.

    University of Arizona, Ph.D., 1967 Chemistry, analytical

    University Microfilms, Inc., Ann Arbor, Michigan

  • DIRECT CALORIMETRIC DETERMINATION OF HEATS

    OF FORMATION OF SOME METAL CHELATES

    by

    George Gutnikov

    A Dissertation Submitted to the Faculty of the

    DEPARTMENT OF CHEMISTRY

    In Partial Fulfillment of the Requirements For the Degree of

    DOCTOR. OF PHILOSOPHY

    In the Graduate College

    THE UNIVERSITY OF ARIZONA

    19 67

  • THE UNIVERSITY OF ARIZONA

    GRADUATE COLLEGE

    I hereby recommend that this dissertation prepared under my

    direction by George Gutnikov

    entitled Direct Calorimetric Determination of Heats

    of Formation of Some Metal Chelates

    be accepted as fulfilling the dissertation requirement of the

    degree of Doctor of Philosophy

    jk uAJJ I tati^n Dissertation Director Date

    After inspection of the dissertation, the following members

    of the Final Examination Committee concur in its approval and

    recommend its acceptance:*

    iLR.jJztt (Ljzi 3T

    9- 1 1 - C L 1

    3.1 kjjf" i

  • STATEMENT BY AUTHOR

    This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.

    Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.

    SIGNED: !&4JTL̂ 2J

  • AC KNOWLED GM E N TS

    The author expresses his gratitude to Dr. Henry Freiser for

    his counsel throughout the experimental work and in the preparation

    of this thesis.

    Thanks are also due to Dr. Quintus Fernando for helpful

    discussion and to Mr. Ted Carnavale for writing the computer pro

    grams.

    Financial support of this research by the U. S. Atomic

    Energy Commission is gratefully acknowledged.

    iii

  • TABLE OF CONTENTS

    Page

    LIST OF ILLUSTRATIONS v

    LIST OF TABLES vi

    ABSTRACT . x

    INTRODUCTION 1

    STATEMENT OF PROBLEM 33

    EXPERIMENTAL 34 General Considerations 34 Titrimetric Apparatus 36 Titrimetric Procedure 37 Calorimetric Apparatus 38 Calorimetric Procedure 43 Reagents 47

    CALCULATIONS 50 Acid Dissociation Constants 50 Chelate Formation Constants . . . . 51 Heats of Reaction 53 Heats of Reagent Dissociation 54 Heats of Chelation 54

    ERRORS 59

    DISCUSSION 64 Comparison of Methods 64 Comparison with Previous Results 70 Discussion of Present Results 77

    APPENDIX A 91

    LIST OF REFERENCES 108

    iv

  • LIST OF ILLUSTRATIONS

    Figure Page

    1. Cross-sectional View of Calorimeter 39

    2. Calorimeter Circuit 41

    3. Typical Time-temperature Curve 55

    v

  • LIST OF TABLES

    Table Page

    I. Thermodynamic Values for the Chelation Reactions of Ethylene diamine and Trimethylenediamine 10

    II. Thermodynamic Values for the Chelation Reactions of Di ethyl en etri amine and 2, 2' 2" -Triamino-triethylamine- 12

    III. Thermodynamic Values for the Chelation Reactions of Triethylenetetramine and N, N' N"-Tetrakis-(2-aminoethyl)-ethylenediamine 13

    IV. Thermodynamic Values for the Chelation Reactions of 1, 10-Phenathroline and 2, 2'-Bipyridine 15

    V. Thermodynamic Values for the Chelation Reactions of Iminodiacetic and N-Methyliminodiacetic acids 17

    VI. Thermodynamic Values for the Chelation Reactions of Nitrilotriacetic and Ethylenediaminetetra-acetic acids 18

    VII. Thermodynamic Values for the Chelation Reactions of trans-Cyclohexanediaminetetraacetic and Ethyleneglycol-(bis-/3 -aminoethyl ether)-N, N' -tetraacetic acids 23

    VIII. Thermodynamic Values for the Chelation Reactions of Ethyletherdiaminetetraacetic and Ethylthio-etherdiaminetetraacetic acids 25

    IX. Thermodynamic Values for the Chelation Reactions of N-Hydroxyethylethylenediaminetriacetic and Diethylenetriaminepentaacetic acids ' 26

    X. Thermodynamic Values for the Chelation Reactions of 2, 4-Pentanedione and Tripolyphosphoric acid 31

    vi

  • vii

    LIST OF TABLES--(Continued)

    Table Page

    XI. Thermodynamic Values for Ligand Dissociation .... 72

    XII. Thermodynamic Values for the Formation of the 8-Quinolinol Chelates 74

    XIII. Thermodynamic Values for the Formation of the 2-Methyl-8-quinolinol Chelates 75

    XIV. Thermodynamic Values for the Formation of the 4-Methyl-8-quinolinol Chelates 76

    XV. Thermodynamic Values for the Formation of the 8-Quinolinol-5-sulfonic acid Chelates 80

    XVI. Thermodynamic Values for the Formation of the Quinoline-8-thiol and 2-Methylquinoline-8-thiol Chelates 85

    XVII. Thermodynamic Values for the Formation of the 2,4-Pentanedione Chelates 89

    XVIII. Summary of Acid Dissociation Constants 91

    XIX. Summary of Chelate Formation Constants 92

    XX. Data for AH 93 w

    XXI. Data for AH^H 94

    XXII. Data for AH_TT and AHC 95 Uri oJti

    XXIII. Data for Mn(II)-8-Quinolinol 96

    XXIV. Data for Co(II)-8-Quinolinol 96

    XXV. Data for Ni(II)-8-Quinolinol 96

    XXVI. Data for Cu(II)-8-Quinolinol 97

    XXVII. Data for Zn(II)-8-Quinolinol 97

  • viii

    LIST OF TABLES--(continued)

    Table Page

    XXVIII. Data for Cd(II)-8-Quinolinol 97

    XXIX. Data for Pb(II)-8-Quinolinol 98

    XXX. Data for Mn(II)-2-Methyl-8-quinolinol 98

    XXXI. Data for Co(II)-2-Methyl-8-quinolinol 98

    XXXII. Data for Ni(II)-2-Methyl-8-quinolinol 99

    XXXIII. Data for Cu(II)-2-Methyl-8-quinolinol 99

    XXXIV. Data for Zn(II)-2-Methyl-8-quinolinol 99

    XXXV. Data for Pb(II)-2-Methyl-8-quinolinol 99

    XXXVI. Data for Mn(II)-4-Methyl-8-quinolinol 100

    XXXVII. Data for Co(II)-4-Methyl-8-quinolinol 100

    XXXVIII. Data for Ni(II)-4-Methyl-8-quinolinol 100

    XXXIX. Data for Cu(II)-4-Methyl-8-quinolinol 101

    XL. Data for Zn(II)-4-Methyl-8-quinolinol 101

    XLI". Data for Pb(II)-4-Methyl-8-quinolinol 101

    XLII. Data for Mn(II)-8-Quinolinol-5-sulfonic acid .... 102

    XLIII. Data for Co(II)-8-Quinolinol-5-sulfonic acid .... 102

    XLIV. Data for Ni(II)-8-Quinolinol-5-sulfonic acid .... 102

    XLV. Data for Cu(II)-8-Quinolinol-5-sulfonic acid .... 103

    XLVI. Data for Zn(II)-8-Quinolinol-5-sulfonic acid .... 103

    XLVII. Data for Ni(II)-8-Quinolinol-5-sulfonic acid (in aqueous solution) 103

  • ix

    LIST OF TABLES--(continued)

    Table Page

    XLVIII. Data for Cu(II)-8-Quinolinol-5-sulfonic acid (in aqueous solution) 104

    XLIX. Data for Zn(II)-8-Quinolinol-5-sulfonic acid (in aqueous solution) 104

    L. Data for Mn(II)-Quinoline-8-thiol 104

    LI. Data for Co(II)-Quinoline-8-thiol 104

    LII. Data for Ni(II)-Quinoline-8-thiol 104

    LIII. Data for Cu(II)-Quinoline-8-thiol 105

    LIV. Data for Zn(II)-Quinoline-8-thiol 105

    LV. Data for Pb(II)-Quinoline-8-thiol 105

    LVI. Data for Co(II)-2-Methylquinoline-8-thiol 105

    LVII. Data for Ni(II)-2-Methylquinoline-8-thiol 106

    LVIII. Data for Cu(II)-2-Methylquinoline-8-thiol 106

    LIX. Data for Zn(II)-2-Methylquinoline-8-thiol 106

    LX, Data for Mn(II)-2,4-Pentanedione 106

    LXI. Data for Ni(II)-2, 4-Pentanedione 107

    LXII. Data for Cu(II)-2, 4-Pentanedione 107

    LXIII. Data for Zn(II)-2, 4-Pentanedione 107

  • ABSTRACT

    A simple, twin-differential calorimeter capable of determining

    the heats of chelation in highly dilute solutions was designed and con

    structed. The heats of reactions of several chelating agents containing

    oxygen and sulfur donor atoms with a number of transition and heavy

    metal ions were obtained, and the corresponding formation constants

    were calculated. The chelating agents studied were 8-quinolinol, 2-

    methyl and 4-methyl-8-quinolinol, 8-quinolinol-5-sulfonic acid, quino-

    line-8-thiol, 2-methylquinoline-8-thiol, and 2, 4-pentanedione; the

    n | 2+ 2+ 2+ 2"4~ 24- 21 metal ions included Mn , Co , Ni , Cu , Zn , Cd , and Pb

    Reactions were generally performed in an aqueous 50 volume %

    dioxane-0. 1 M NaClO^ medium, or in aqueous 0. 1 M NaClO^.

    In contrast to previous studies, considerable regularity was

    found in the entropy changes of chelation for the 8-quinolinols. The

    heats of chelation for the quinoline-8-thiols showed that the metal-

    sulfur bonds are stronger than the metal-oxygen bonds. The reversal

    of the usual stability order (Ni > Zn) is due to a more favorable entropy

    change, which was attributed to the formation of a tetrahedral zinc

    chelate.

  • INTRODUCTION

    During the past half century the application of organic reagents

    to the solution of analytical problems has yielded numerous fruitful

    results. Much of the earlier work was encumbered by the lack of a

    theoretical foundation, frequently requiring the expenditures of exces~

    cise effort in order to achieve a satisfactory solution.

    An important development in the study of organic chelating

    agents came in 1941 when Bjerrum^ presented a method for the

    determination of the successive stability constants of metal ammine

    ( 2 ) complexes. Following Calvin's . modification of the Bjerrum method

    in 1945, which extended its utility to almost any organic ligand capable

    of exchanging a hydrogen for a metal, this so-called Calvin-Bjerrum

    potentiometric technique has evolved into the most reliable mearjs of

    evaluating the formation constants of metal chelates.

    (3) Today formation constant data abound in the literature. An

    analysis of this mass of information provides certain criteria for

    1. J. Bjerrum, Metal Ammine Formation in Aqueous Solutionf P. Haase and Son, Copenhagen, 1941.

    2. M. Calvin and K. W. Wilson, J. Am. Chem. Soc. 67, 2003(1945),

    3. L. G. Sillen and A. E. Martell (compilers), Stability pon-stants of Metal Ion Complexes, The Chemical Society, London, Specif Publication No. 17, 1964.

    1

  • 2

    assessing the extent of chelate formation. The approximate chelate

    stability can be predicted from certain properties of the ligand and

    metal ion, as well as the specific effects of the solvent.

    The important factors for the ligand include the nature of its

    donor atoms and their basicity, steric hindrance, and the size and

    number of the rings formed.

    Ligands containing oxygen, nitrogen, and sulfur as donor atoms

    have assumed the greatest analytical significance because of their

    ability to coordinate very effectively with many metals. In a study

    of the EDTA analogs, CH2CH2[YCH2CH2N(CH2COO~)2] in which

    Y=NCH , S, or O, Schwarzenbach et al. ̂ elegantly demonstrated the

    stability sequence 0>N>S for the alkaline earths and N>S>0 for the

    transition metals with nearly filled d orbitals. Manganese (II), how-

    ( 2 ) ever, exhibits an enhanced stability with O over N ligands. A

    distinct preference for Se over S by transition metal ions has been

    reported. ^

    1. G. Schwarzenbach, H. Senn, and G. Anderegg, Helv. Chim. Acta 40, 1886 (1957).

    2. H. Irving and R. J. P. Williams, Nature, Lond. 1'62, 746 (1948); J. Chem. Soc. 3192 (1953).

    3. G. Schwarzenbach, G. Anderegg, W. Schneider, and H. Senn, Helv. Chim. Acta 38, 1147 (1955).

    4. E. Sekido, Q. Fernando, and H. Freiser, Anal, Chem. 37, 1556 (1965).

  • 3

    Care must be exercised in the above comparisons to match .

    ligand basicities. Because both hydrogen and metal ions act as Lewis

    acids toward ligands, it is reasonable to expect a linear correlation

    between acid dissociation and stability constants. Although such

    relations have been observed frequently, they have been shown to

    deviate somewhat if the parent ligand is substituted by groups posses

    sing substantial ir donor or acceptor properties. ̂

    Such correlations fail when a bulky substituent near the coordi

    nating atom interferes with the bonding of the much larger metal ions.

    The decreased stability of 2-methyloxine chelates relative to a series

    ( 2 ) of similar unhindered oxine chelates illustrates this point convincingly.

    Despite its greater basicity, trimethylenediamine forms less

    (3) stable chelates than ethylenediamine. This has been attributed to

    ring strain in the six-membered ring. The less exothermic heat of

    formation of trimethylenediamine chelates is consistent with this hypo

    thesis. Five-membered chelate rings are generally found to be most

    stable.

    1. J. G. Jones, J. B. Poole, J. C. Tomkinson, and R, J. P. Williams, J. Chem. Soc. 2001 (1958).

    2. W. D. Johnston and H. Freiser, Anal. Chim. Acta 11, 201 (1954).

    3. I. Poulsen and J. Bjerrum, Acta Chem. Scand. 1407 (1955).

  • 4

    Since the additional stability observed in the displacement of

    monodentate ligands by a polydentate ligand derives chiefly from an

    increase in the number of particles in solution (hence the entropy

    change is positive), the formation of a larger number of rings should

    be accompanied by an increase in stability. This postulate is valid

    provided that no serious ring strain is incurred, and the coordination

    number of the metal is not exceeded.

    The coordination number of many metals is commonly six;

    lead and copper (II) generally form only four strong bonds, although

    in the case of copper two additional weak bonds are formed due to the

    Jahn-Teller effect.

    Another property of metal ions which has been correlated with

    chelate stability is charge density. For a group of similar metal ions

    of the same charge which form essentially ionic bonds, stability varies

    inversely with ionic radius. Thus, for the alkaline earths the sequence

    Mg>Ca>Sr>Ba is often noted. In the case of the transition metal ions,

    whose ionic radii are very similar, significant stability differences

    arise from differences in the ligand field energy. In general, the

    sequence Mn

  • 5

    Plots of stability constants against ionization potentials or

    electronegativity frequently approximate a straight line. The basis of

    the explanation is the fact that both of these parameters are measures

    of electron affinity and hence are related to the attraction of the metal

    ion for the electrons of the ligand.

    Apart from the special properties of the ligand and metal ion,

    the solvent employed can profoundly influence the stabilities observed.

    The background electrolyte will affect the stabilities in accordance

    with its activity coefficient, but additional effects will be observed if

    it complexes with the metal ion.

    In mixed solvents ions may be selectively solvated by either

    component. ̂ Thus, calcium and zinc ions are hydrated predomi

    nately in CHgOH-HgO and CHgCN-HgO mixtures, respectively, but in

    the latter mixture silver ion is preferentially solvated by CH^CN. In

    addition, for uncharged chelates enhanced solvation by the organic

    component might be anticipated. Therefore, a change in solvent may

    drastically alter the nature and, consequently, the equilibrium constant

    of a particular reaction.

    If the mole fraction of the "inert" component is kept relatively

    small so that hydration is the main mode of solvation, then stability

    1. H. Strehlow, et al. Ber. Buns en Ges ell. Physik. Chem. 62, 373 (1958); 66, 309 (1962); 69, 674 (1965); Z. Physik. Chem., N. F., 49, 44 (1966).

  • 6

    varies inversely with the solvent dielectric constant. This is illustrat

    ed by the increase in stabilities of various metal oxinates as increasing

    amounts of dioxane are added. ^

    The increased stability in lower dielectric constant media may

    be due to changes in solvent interaction and bond strengths. The

    determination of the heats and entropies of chelation could therefore

    substantially illuminate this question. As yet., however, few such

    ( 2 ) studies have been published and one indicates that solvation effects

    predominate.

    For the previous cases as well as the vast majority of others,

    stability constants alone fail to distinguish adequately between such

    factors as bond strengths, configuration, steric hindrance, and solva

    tion effects. This is true because the formation (stability) constant,

    K^, which is directly related to the free energy change, AG, by the

    relation

    AG = -RT In Kj

    reflects differences in the changes in both the enthalphy, AH, and the

    entropy, AS, since

    AG = AH - TAS

    A single number written as a subscript with a thermodynamic function

    1. H. Irving and H. Rosotti, Acta Chem. Scand. 10, 72 (1956).

    2. N. C. Li, J. M. White, and R. L. Yost, J. Am. Chem. Soc. 78, 5218 (1956).

  • will refer to the corresponding stepwise reaction, whereas two num

    bers will refer to the corresponding overall reaction.

    In the past conclusions were drawn from stability constants

    about structural features of chelates. This was based on the assump

    tion that stability constants were proportional to the enthalpies, hence

    that the entropies were similar for most metals, and that they remained

    constant in a series of related compounds.

    In fact, however, it is general for changes in enthalpy and

    entropy to at least partially compensate each other in dissociation pro

    cesses.^ For example, increases in the attractive forces between

    particles resulting in a more rigid structure would lead to negative

    changes in both AH and AS. Conversely, formation of a looser struc

    ture, for instance, due to steric hindrance would result in positive

    enthalpy and entropy changes. In chelate formation disruption of

    solvent molecules from the metal ion and ligand consumes energy, but

    this process is compensated by the increase in the number of particles.

    Since stability constants alone do not completely explain solu

    tion processes, the trend toward obtaining AH and AS data has been

    progressing steadily. At first AH was calculated from the temperature

    variation of formation constants since this involved minimal modifica

    tions in the equipment used for the determinations.

    1. D. J. G. Ives and P. D. 'Marsden, J. Chem. Soc. 649 (1965).

  • 8

    The convenience of this procedure was offset., however, by the

    frequent, large discrepancies in the data reported by different investi-

    (1) gators for the same system. Errors arising from the uncertainties

    in the formation constants (about t 0.1.0 log unit) can introduce an

    error of about 4 kcal./mole into AH when measurements are made over

    a 10° range. Other factors which contribute to the error include the

    variation_pf AH with temperature, differences in activity coefficients

    at different temperatures, kinetic effects, and competing reactions,

    Hence, because of the inadequacies in the temperature dependence

    method, direct calorimetry is now preferred for the determination of

    AH.

    To date a number of calorimetric determinations of the heats

    of chelation have been carried out with N-N and N-O ligands. For O-O

    ligands and those containing sulfur the data are still sparse.

    Among ligands containing only nitrogen donor atoms, the vari

    ous aliphatic polyamines have been studied extensively. Of these,

    (2,3 4] ethylenediamine (en) has received the greatest amount of attention, ' '

    1. F.J. C. Rossotti, in Modern Coordination Chemistry, J'. Lewis andR. G. Wilkins, eds. Interscience Publishers, Inc. New York, 1960, p. 68.

    2. T. Davies, S. S. Singer, and L. A. K. Staveley, J. Chem, Soc. 2304 (1954).

    3. I. PoulsenandJ. Bjerrum, ActaChem. Scand,, £, 1407 (1955).

    4. M. Ciampolini, P. Paoletti, and L. Sacconi, J. Chem. Soc. 4553 (1960).

  • 9

    For the en chelates of the transition metal ions a definite decrease in

    the stepwise entropies but a slight increase in the enthalpies was

    generally observed (Table I). This was attributed to a greater release

    of water molecules and, consequently, the rupture of a larger number

    of metal-water bonds in the first step than in succeeding ones. An

    exception was provided by zinc for which the formation of the bis che

    late from the mono was accompanied by a lower AH but a higher AS.

    ++ This was explained by the formation of a tetrahedral Zn(en)g chelate

    with the release of additional waters of hydration. ̂

    The effect of introducing alkyl substituents onto ethylenedia-

    (2) mine was investigated by Basolo and Murman. ' Although the AH and

    AS values for en and its N-methyl derivative (Meen) differ only slightly,

    those for the N, N'-diethyl derivative (diEten) are both considerably

    more positive, to the extent that they nearly compensate each other in

    terms of the free energy. The respective -AH^ and AS^g values, in

    kcal/mole and e.u., for Ni with en, Meen, and diEten are 16, 3 and 7,

    17. 0 and 1, and 7. 8 and 27. This effect was ascribed to steric hind

    rance by the bulky alkyl substituents.

    1. M. Ciampolini, P. Paoletti, and L, Sacconi, J, Chem. Soc. 4553 (1960).

    2. F. Basolo and R. K. Murman, J. Am. Chem. Socu 76, 211 (1954).

  • TABLE 1. - -Thermodynamic values for the chelation reactions of ethylenediamine and trimethylenediamine.

    en tm -AG -AH AS -AG -AH AS kcal kcal kcal kcal

    Cation Step mole~l mole~l (e.u. ) mole"! mole~l (e. u. )

    H+ 1 13. 9 12. 2 5. 7 14. 5

    Mn2+

    2 10. 2 10. 6 -1. 5 12. 4

    Mn2+ 1 3. 8 2. 8 3. 0 2 2. 9 3. 2 -1. 0

    Pe2+

    3 1. 2 5. 1 -9. 5

    Pe2+ 1 5. 9 5. 1 3. 0 2 4. 6 5. 3 -3. 0

    o

    o CO

    + 3 2. 8 5. 5 -8. 5

    o

    o CO

    +

    1 8. 1 6. 9 4. 0 2 6. 5 7. 1 -2.0

    *T-2 + NI

    3 4. 1 8. 2 -10.0

    *T-2 + NI 1 10. 5 8. 9 5. 5 8. 7 7. 8 3. 0 2 8. 7 9. 4 -2. 5 6. 0 7. 2 -4. 1

    2+ Cu

    3 5. 9 10. 1 -8. 5 1. 7 6. 3 -15. 5 2+

    Cu 1 14. 7 13. 1 5. 5 2 11. 0 12. 3 -4. 5

    Zn2+

    1-2 23. 4 22. 8 2. 0

    Zn2+ 1 8. 1 7.0 3. 5 2 7. 0 4. 9 7. 0

    Cd2+ 3 2. 6 5. 2 -8. 5

    Cd2+ 1 8. 0 7.0 3. 1 2 6. 5 6. 5 0. 2 3

  • 11

    Despite its greater basicity, trimethylenediamine, which

    forms six-membered chelate rings, reacts less exothermically than

    ethylenediamine. ̂ ' This behavior probably results from greater ring

    strain in the larger chelate ring.

    Several higher homologs of ethylenediamine have also been the

    subjects of thermodynamic studies. They include diethylenetriamine

    (dien), 2, 2", 2"-triaminotriethylamine (tren), triethylenetetramine

    (trien), and N, N', N"-tetrakis-(2-aminoethyl.) ethylenediamine (penten).

    The data are presented in Tables II and III. For a given metal ion the

    heat of chelation per amino group is similar, but tends to decrease

    somewhat with the increasing number of chelate rings formed. Accord-

    ( 2 ) ing to Ciampolini, et al. this was due to ring strain or to weaker

    bonding between the metal ion and secondary and tertiary amino nitro-

    ( 3 ) gens than primary nitrogens. Reilley, et al. ' suggested that these

    effects could also be accounted for by changes in the base strengths of

    the remaining amino groups after the first had bonded, and a change

    in the acidity of the metal ion after formation of the first metal-amino

    bond.

    1. I. PoulsenandJ. Bjerrum, ActaChem„ Scand. 9, 1407(1955).

    2. M. Ciampolini, P. Paoletti, and L. Sacconi, in Advances in the Chemistry in the Coordination Compounds, S. Kirschner, ed. The Macmillan Co., New York, 1961, p„ 303„

    3. D. L. Wright, J, H. Holloway, and C„ N„ Reilley, Anal. Chem. 37, 884 (1965).

  • 12

    TABLE II. - -Thermodynamic values for the chelation reactions of diethylenetriamine and 2, 2', 2"- triaminotriethylamine.

    dien tren

    -AG -AH- AS -AG -AH AS kcal ^ kcal _1

    (e. u.) kcal ^ kcal

    (e. u.) Cation Step mole mole (e. u.) mole mole (e. u.)

    H+ 1 13. 4 11.2 7.2 13.8 11. 7 7. 1

    2 12. 3 12. 0 1.0 12. 9 12. 8 0. 2

    Mn2+

    3 5. 8 7.2 -4.7 11.5 12, 2 -2. 3

    Mn2+ 1 7.9 3. 0 16. 5

    Fe2+ 1 11.8 6. 3 18. 5

    Co2+ 1 10. 9 8.2 9. 0 17. 0 10. 7 22. 0

    Ni2+

    2 8. 0 10. 2 -7. 5

    Ni2+ 1 14. 5 11.9 8. 5 20. 0 15. 2 16. 0

    Cu2+

    2 10. 9 13. 4 -8.5

    Cu2+ 1 21. 6 18. 0 12.0 25. 8 20. 4 18. 0

    Zn2+

    2 7.1 8.2 -3.5

    Zn2+ 1 12.0 6. 5 18. 5 19. 7 13. 9 19. 5

    2 7.5 10.1 -9.1

  • 13

    TABLE III. --Thermodynamic values for the chelation reactions of triethylenetetramine and N, N', N" -tetrakis-(2 • amino-ethyl) -ethylenediamine.

    trien penten -AG -AH AS -AG -AH AS kcal ^ kcal ^ kcal kcal ^

    Cation Step mole mole (e. u. ) mole mole (e. u. )

    H+ 1 13. 3 11. 0 7. 8 13. 7 11. 3 8. 1 2 12. 4 11. 3 3. 7 13. 0 11. 5 5. 3 3 8. 9 9. 5 -2. 0 12. 2 13. 2 -3. 1 4 4. 4 6. 8 -8. 1 11. 5 12. 0 -1. 8

    Mn2+

    5 1. 8 4. 5 -9. 0

    Mn2+ 1 6. 7 2. 3 15. 0 12. 6 8. 9 12. 5 2+

    Fe 1 10. 5 6. 1 15. 0 15. 2 9. 7 18. 5 2+

    Co 1 14. 9 10. 7 14. 5 21. 2 14. 8 21. 5 .2+

    NI 1 19. 3 13. 9 18. 0 26. 1 19. 7 21, 5 2+

    Cu 1 27. 6 21. 4 21. 0 30. 2 24. 5 19. 0 2+

    Zn 1 16. 3 8. 3 27. 0 22. 0 14. 5 25. 0 2+

    Cd 1 14. 8 9. 2 19. 0

  • 14

    Although the aromatic diamines 1, 10-phenanthroline (phen)

    and 2, 2'-bipyridine (dip) are much less basic than ethylene diamine

    5 (by about 10 ), they generally form chelates of comparable stability

    ( 1 2 ) (Table IV). Calorimetric studies ' disclose a fairly similar pat

    tern in the heats and entropies of chelation for the aromatic amines,

    in that the stepwise -AH values are somewhat less and the stepwise

    AS values increase somewhat, especially for phen. This behavior is

    the reverse of that found for the en chelates and was attributed to the

    greater hindrance provided by the rigidity of the aromatic residue.

    The greater rigidity of phen relative to dip was manifest in the less

    favorable entropies of chelation for the latter ligand, since rotational

    freedom was lost on chelation. Primarily because of this entropy dif

    ference the phen chelates are more stable than those of dip. The

    tris-ferrous chelates of the aromatic diamines exhibit an exceptional

    stability, which is particularly marked if the AH values are compared.

    In this well-known case a change in configuration occurs from the para

    magnetic hydrated ferrous ioh to the diamagnetic complex. Although

    all of the chelates of the aromatic amines are enthalpy-stabilized,

    ligand field stabilization was shown to account for only a small part of

    1. G. Anderegg, Helv. Chim. Acta 46, 2813 (1963).

    2. R. L, Davies and K. W. Dunning, J. Chem. Soc. 4168 (1965).

  • 15

    TABLE IV. --Thermodynamic values for the chelation reactions of 1, 10-phenathroline and 2, 2'-bipyridine.

    Cation

    H+

    Mn2+

    2~h Fe „ 2+ Co

    Ni2+

    Cu2+

    ~ 2+ Zn

    Cd2+

    phen dip -AG -AH AS -AG -AH AS kcal kcal kcal kcal

    Step mole"! mole"! (e. u. ) mole" -1 mole"! (e. u.)

    1 6. 6 4. 0 9. 2 6. 2 3. 7 8. 2

    1 5. 5 3. 5 6. 8 3. 5 3. 5 0. 0 2 4. 7 3. 5 4. 1 3 3. 6 2. 0 -0. 5

    1-3 28. Sr 33. 0 -15. 4 23. 4 31. 4 -27. 0

    1 9. 7 9. 1 2. 1 8. 1 8. 2 -0. 4 2 9. 0 6. 7 7. 8 7. 2 7. 0 0. 7 3 8. 0 8. 0 0.0 6. 4 6. 1 1. o

    1 11. 8 11. 2 2. 1 9. 6 9. 6 0. 0 2 U. 1 9. 3 6. 1 9. 2 9. 4 -0. 7 3 10. 4 9. 5 3. 0 8. 8 9. 2 -1. 4

    1 12. 4 11. 7 2. 4 10. 7 11. 9 -4. 1 2 9. 1 6. 5 8. 8 7. 5 5. 4 7. 2 3 7. 1 8. 2 -3. 7 4. 7 6. 5 -6. 2

    1 8. 8 7. 5 4. 4 7. 1 7. 1 0. 0 2 7. 8 7. 5 1. 1 6. 1 5. 4 2. 4 3 6. 9 4. 3 8. 8 5. 1 5. 0 0. 3

    1 7. 7 6. 3 4. 8 5. 7 5. 1 2. 1 2 6. 8 6. 8 0. 0 4. 8 4. 3 1. 6 3 5. 5 3. 0 8. 5 3. 6 4. 6 -3. 4

  • 16

    this. The author therefore attributed this stabilization to steric

    factors.^

    A large number of N-O ligands of great variety have been

    investigated calorimetrically. The aminopolycarboxylic acids, many

    of which have found extensive application in analytical chemistry, have

    received the greatest amount of attention. The simplest members of

    ( 2 ) this series are the terdentate iminodiacetic and N-methyliminodia-

    (3) cetic acids (Table V ). . The N-methyl derivative generally forms

    somewhat more stable compounds (by about one log unit for the 1:1

    complex and about two log units for the 2:1 complex). This stabiliza

    tion is predominantly derived from a more favorable entropy change.

    This behavior was explained by the larger size of the methyl group

    which forces the two carboxylate groups closer together, resulting in

    greater localization of charge on the oxygens and producing greater

    ordering of the surrounding water. The release of this water during

    chelation accounts for the increased entropies observed.

    The next higher analog of this series, the quadridentate nitri-

    lotriacetic acid (NTA), forms mono-chelates of about 2 log units

    greater stability than the above compounds (Table VI). However,

    1. G. Anderegg, Helv. Chim. Acta 46, 2813 (1963).

    2. G. Anderegg, Helv. Chim. Acta 47, 1801 (1964).

    3. G. Anderegg, in Essays in Coordination Chemistry, Exper. Sup pi. 9, 75 (1964).

  • 17

    TABLE V. - -Thermodynamic values for the chelation reactions of iminodiacetic and N-methyliminodiacetic acids.

    Cation

    2+ H+

    Mn

    ^ 2+ Co

    Ni2+

    „ 2+ Cu

    2+ Zn

    Cd2+

    lm mim -AG -AH AS -AG -AH AS kcal kcal kc al kcal

    Step mole"! mole"! (e. u.) mole ~ 1 mole"! (e. u.)

    1 12. 7 8. 2 15. 4 13. 0 6. 9 20. 5

    1 7. 2 •0. 6 26. 6 2 5. 6 0. 3 18. 0

    1 9. 4 2. 1 24. 6 10. 2 1. 9 28. 6 2 6. 9 3. 9 11. 2 8. 5 3. 6 16. 4

    1 11. 0 5. 1

    o

    o

  • 18

    TABLE VI. - -Thermodynamic values for the chelation reactions of nitrilotriacetic and ethylenediaminetetraacetic acids.

    NTA EDTA -AG -AH AS -AG -AH AS kcal kcal kcal kcal

    Cation Step mole"! mole-* (e. u.) mole ~ 1 mole~l (e. u.;

    H+ 1 13. 1 4. 7 28. 4 13. 8 5. 7 28

    Mn2+

    2 8. 3 4. 3 13

    Mn2+ 1 10. 0 -1. 4 38. 9 17. 2 4. 6 48

    2+ Fe

    2 4. 7 5. 5 -1. 7 2+

    Fe 1 19. 3 4.0 51

    Co2+ 1 13. 9 0. 1 47. 2 21. 4 4. 2 60

    Ni2+

    2 5. 4 4. 7 2. 1

    Ni2+ 1 15. 5 2. 6 43. 9 25. 5 7. 6 59

    „ 2+ Cu

    2 6. 5 5. 5 7. 0 ' „ 2+ Cu 1 17. 4 1. 9 52. 8 25. 5 8. 2 58

    2+ Zn

    2 6. 0 7. 0 -3. 5 2+

    Zn 1 14. 2 0. 9 45. 5 22. 3 4. 9 59 2 5. 0 2. 7 7. 4

    ,2+ Cd 1 13. 2 4. 0 31. 3 22. 3 9. 1 44

    Pb2+

    2 6. 4 5. 1 • 4. 7

    Pb2+ 1 15. 3 3. 8 39. 1 23. 6 13. 2 38

  • 19

    ( 1 2 ) about 2 kcal/mole less heat is evolved in this process. ' Evidently

    these chelates are entropy-stabilized, probably due to the additional

    water released from the third carboxylate group in the reaction.

    In contrast to the high entropies of about 40-50 e. u. observed

    for the 1:1 transition metal chelates, those for the addition of another

    NTA molecule are all nearly zero. At the same time, the enthalpies

    become more exothermic by about 3-7 kcal/mole. These data sug

    gest that one of the rings of the mono-chelate is opened when a second

    NTA molecule attaches itself, so that the bis-chelate contains two

    uncoordinated -CHgCOO groups. The high charge density in the

    quadruply negatively charged chelate orders a considerable number

    of water molecules around the chelate and thereby reduces the en

    tropy change. The more negative AH _ values result from bonding JL

    to an additional nitrogen while breaking a metal-carboxylate bond

    whose formation may have been endothermic.

    The entropies of formation of Cu(NTA) * and Pb(NTA) * are

    nearly equal to those for the formation of the EDTA complexes. Ap

    parently six rather than four waters of hydration are lost because

    these metals usually exhibit a coordination number of four.

    1. G. Anderegg, in Essays in Coordination Chemistry, Exper. Suppl. 9, 75 (1964),

    2. J. A. Hull, E. H, Davies, and-L. A. K. Staveley, J. Chem. Soc. 5422 (1964).

  • 20

    The parent compound of this series, EDTA, has been studied

    u , • , (1,2,3,4,5) . (1) .. by numerous investigators. Charles was the first to

    demonstrate that the EDTA chelates owe their stability to the very

    (3) favorable entropy changes. Staveley and Randall found an inverse

    (4) linear relationship between AS and the metal ion radius. Anderegg

    showed that the heat of chelation was markedly influenced by the

    (5) anion of the metal salt used. Reilley, £t al, compared the AH values

    for the transition metal-EDTA chelates with those for en and con

    cluded that the acetate groups of EDTA contribute little to the total

    heat of reaction. Their entropy contribution, however, constituted

    the major factor of the total stability in solution.

    The formation constants of the cyclohexyl analog, trans-

    cyclohexanediaminetetraacetic acid (CDTA), exceed those of the

    parent compound by 2 - 3 log units. On the basis of these data alone,

    1. R. G. Charles, J. Am. Chem. Soc. 76, 5854 (1954),

    2. R. A. Care and L. A. K. Staveley, J. Chem. Soc. 4571 (1956).

    3. L. A, K. Staveley and T. Randall, Disc. Faraday Soc. 2S, 157 (1958).

    4. G. Anderegg, Helv. Chim. Acta 46, 1833 (1963).

    5. D. L. Wright, J. H. Holloway, and C. N. Reilly, Anal. Chem. 37, 884 (1965).

  • 21

    Schwarzenbach, et al/1^ correctly deduced that this higher stability

    resulted from greater entropy changes, for it seemed unlikely that

    the same donor atoms could bind the metals with such different

    strengths. Confirmation of this deduction came from calorimetric

    (2 3) data. ' For CDTA the enhanced entropy changes were attributed

    to the loss of more water which was ordered by the larger charge

    localization as a result of the greater rigidity imposed by the

    cyclohexyl ring.

    The chelate effect, defined as the difference in log units be

    tween the chelate stability of a poly functional and a corresponding

    simple ligand, is a measure of the increased stability gained by ring

    (4) formation. In order to examine this effect more closely, a series

    of EDTA homologs was studied calorimetrically, in which n, the

    number of -CH^~ links between the nitrogens, was varied from two

    (4) (5) to eight. Iminodiacetic acid or N-methyliminodiacetic acid

    1. G. Schwarzenbach, R. Gut, and G. Anderegg, Helv. Chim. Acta 37, 936 (1954).

    2. G. Anderegg, Helv. Chim. Acta 46, 1833 (1963).

    3. D. L. Wright, J, H. Holloway, and C. N. Reilley, Anal. Chem. 37, 884 (1965).

    4. G. Anderegg, Helv. Chim. Acta 417, 1801 (1964).

    5. G. Anderegg, Helv. Chim. Acta 48, 1718 (1965).

  • 22

    served as the simple ligand. The chelate effect was found to be pri

    marily an entropy effect, confirming the earlier proposal of Schwar-

    zenbach. ̂ The variation in the chelate effect with n resulted from

    changes in the enthalpy rather than the entropy. Other variations in

    the AH and AS values were numerous and complex, so that a detailed

    analysis of all the data could not be given.

    The derivative containing two ether oxygens and six carbons

    between the nitrogens, ethyleneglycol-(bis-/3-aminoethyl ether)-N,N'-

    (2 3 4) tetraacetic acid (EGTA), displays a similar behavior ' ' (Table

    VII). Relative to the EDTA homo log with n = 8, AH and AS for the

    manganese and cadmium chelates of this ligand are 9 kcal/mole and~

    20 e.u. more negative. Parallel effects, however, were not observed

    for copper, zinc, cobalt, and nickel. In fact, AH is more positive for

    cobalt and especially for nickel, indicating a dependence on the ion

    (4) size. The manganese and cadmium data suggest bonding by the

    weakly solvated ether oxygens in place of two charged carboxylate

    groups which retain their water of hydration and are restricted in

    1. G. Schwarzenbach, Helv. Chim. Acta 35, 2344 (1952).

    2. G. Anderegg, Helv. Chim. Acta 47, 1801 (1964).

    3. S. Boyd, A. Bryson, G. H. Nancollas, and K. Torrance, J. Chem. Soc. 7353 (1965).

    4. Dt C. Wright, J, H. Holloway, and C. N. Reilley, Anal. Chem. 37, 884 (1965).

  • 23

    TABLE VII. --Thermodynamic values for the chelation reactions of trans-cyclohexanediaminetetraacetic and ethyleneglycol-

    (bis-j3-aminoethyl ether)-' N, N' -tetraacetic acids.

    CDTA EGTA

    Cation Step'

    -AG kcal mole" 1

    -AH kcal mole~l

    AS

    (e. u,)

    -AG kcal mole-1

    -£H kcal mole-1

    AS

    (e. u.)

    H+ 1 2

    16. 6 8. 2

    6. 7 2. 1

    34 21

    12. 7 11. 9

    5. 8 5. 8

    23. 3 20. 8

    Mn2+ 1 22. 7 4. 1 66 16. 7 8. 8 27

    Fe2+ 1 24. 8 6. 6 61 16. 1 5. 2 37 ^ 2+ Co 1 25. 6 2. 8 80 16. 7 3.4 45

    Ni 1 26. 4 5.4 63 18. 5 5. 0 45 2+

    Cu 1 25. 4 8.2 58 24. 2 i0. 5 46

    Zn2+ 1 25. 3 7.7 82 19. 7 3. 8 53

    Cd2+ 1 26.0 7.4 66 22. 7 14. 1 29

    Pb2+ 1 26. 5 11.4 54 19. 9 12. 5 25

  • 24

    rotation by their mutual repulsion. Similar structural implications

    were derived from nmr data^ for the alkaline earth chelates.

    The incorporation of an oxygen atom between two ethylene

    bridges connecting the nitrogens leads to ethyletherdiaminetetraacetic

    acid (EEDTA), Its chelates were compared to those of the EDTA

    homolog with the same number of carbon atoms interposed between the

    nitrogens, and to the thioether analog, ETDTA (Table VIII). The pro

    nounced increase in -AH for the EEDTA chelates of manganese and

    cadmium was taken as an indication of coordination through the ether

    oxygen. In terms of AH, only nickel and mercury show a definite pre

    ference for the sulfur ligand, whereas lead, copper, and cadmium

    react about equally well with both. Zinc and especially manganese

    distinctly prefer the oxygen ligand. The AS values are quite similar

    for both the oxygen and sulfur compounds.

    N-Hydroxyethylethylenediaminetriacetic acid (HEDTA), which

    differs from EDTA by the presence of a hydroxymethyl group in place

    of an acetate group, usually exhibits more negative heats and entro-

    ( 2 ) pies of chelation (up to 2 kcal/mole and 10-20 e.u. ) (Table IX). To

    account for this the following explanations were proposed: Metals

    form stronger bonds with the hydroxyethyl group than with the acetate

    1. A. Bryson and G, H. Nancollas, Chem. and Ind. 654(1965).

    2. D. C. Wright, J. H. Holloway, and C, N, Reilley, Anal. Chem. 37, 884 (1965).

  • 25

    TABLE Vffl> --Thermodynamic values for the chelation reactions of ethyletherdiaminetetraacetic and ethylthioetherdiamine -tetraacetic acids.

    EEDTA ETDTA -AG -AH AS -AG -AH AS kcal kcal kcal kcal

    Cation Step mole"! mole -1 (e. u.) mole " * mole"! (e. u.)

    H+ 1 12. 7 6. 2 22. 1 12. 6 6. 7 20. 3

    Mn2+

    2 11. 9 7. 3 15. 7 11. 4 6. 6 16. 3

    Mn2+ 1 18. 5 5. 9 45. 6 13. 5 1. 5 41. 9

    Co2+ 1 20. 5 6. 4 48, 2 18. 8 4. 6 48. 2

    Ni2+ 1 20. 2 4. 7 52. 8 21. 1 7. 7 45. 5

    Cu2+ 1 24. 3 9. 8 49. 0 22. 2 9. 1 44. 7

    Zn2+ 1 20. 5 6. 0 49. 6 18. 0 3. 7 48. 9

    Cd2+ 1 21. 7 9. 4 42. 0 19. 3 8. 2 37. 8

    Pb2+ 1 20. 2 13. 2 23. 9 18. 6 13.0 19. 1

    Hg2+ 1 32. 0 20. 5 35. 7 31. 0 22. 8 31. 6

  • 26

    TABLE IX. - -Thermodynamic values for the chelation N-hydroxyethylethylenediaminetriacetic triaminepentaacetic acids.

    reactions of and diethylene-

    HEDTA DPTA -AG -AH AS -AG -AH AS kcal kcal kcal kcal

    Cation Step mole"! mole"* (e. u. ) mole" 1 mole~l (e. u.)

    H+ 1 13. 3 14. 2 8. 0 21 2 7. 3 11. 5 4. 3 18

    Mn2+

    3 5. 7 1. 7 14

    Mn2+ 1 14. 7 5. 2 32 21. 1 7. 5 46 2+

    Fe 1 15. 9 6. 0 33

    Co2+ 1 19. 7 6. 5 44 26. 1 9. 5 56

    Ni2+ 1 23. 3 10. 3 45 27. 3 11. 2 54

    O

    C

    I CO

    +

    X 23. 8 9. 4 48 29. 1 13. 4 53

    Zu 1 19. 7 8. 4 38 25. 6 10. 6 50

    Cd2+ 1 17. 8 10. 3 25 25. 8 12. 4 45

    Pb2+ 1 21. 1 12. 6 29 25. 3 18. 8 22

  • 27

    group; the heat of hydration for the hydroxyethyl group is smaller

    than that for the acetate group; the hydroxyethyl group remains un

    bonded, thus relieving strain and strengthening the other chelate

    bonds; the lack of charge lessens electrostatic repulsion. The lower

    entropy of formation can be accounted for, at least in part, by the

    fewer water molecules released in chelation.

    The octadentate ligand diethylenetriaminepentaacetic acid

    (DTPA) chelates more exothermically than EDTA, especially with the

    ( 1 2 ) transition metal ions. ' The corresponding entropies are fre

    quently somewhat smaller. These facts were explained by the

    preferential coordination of the transition metal ions with the third

    amino group of DTPA instead of a carboxylate group.

    A strong metal-nitrogen bond is formed, but fewer water mol

    ecules are released from this uncharged amino group.

    Considerably fewer calorimetric data have been reported for

    (3) other N-O ligands. Izatt, et al. found only a small variation in

    AH^ and AS^ (-4. 6 to -6. 0 kcal/mole and 19 - 22 e.u. ) for the copper

    (II) chelates of glycine, a-aminoisobutyric acid, threonine, and

    1. D. C. Wright, J. H, Holloway, and Ct N. Reilley, Anal. Chem. 37, 884 (1965).

    2. G. Anderegg, Helv. Chim. Acta 48, 1722 (1965).

    3. R. M. Izatt, J. J. Christensen, and V. Kothari, Inorg. Chem. 3, 1565 (1964).

  • 28

    sarcosine. The corresponding AHg and AS^ values differ by about

    -0. 5 kcal/mole and 8 to 10 e.u. In another paper^ similar data for

    the copper (II)-alanine system were presented. Although compensation

    between the AH and TAS terms was observed in all of these cases, the

    magnitudes were apparently too small for meaningful discussion by

    the authors.

    The thermodynamic functions for the reactions of manganese,

    cobalt, and nickel with glycine have been determined by a temperature

    ( 2 ) dependence method using a cell without a liquid junction. The AH^

    values vary from -0. 3 to -4 kcal/mole, but the AS^ values are all

    about 14 e.u. Such variation in AHJ^ was found to be in accordance

    with the metal sequence reflecting the effect of ligand field stabili

    zation.

    In another temperature dependence study using a polaro-

    (3) graphic method the thermodynamics of association between nickel

    and glycine were determined in aqueous and 45% aqueous dioxane

    media. The heats of formation of the neutral chelate were found to be

    1. K. Pf Anderson, D. A. Newell, and R. M, Izatt, Inorg. Chem. 5, 62 (1966).

    2. J. R. Brannan, H. S. Dunsmore, and G. H. Nancollas, J. Chem. Soc. 304 (1964).

    3. N. C. Li, J, M. White, and R. L. Yost, J. Am. Chem. Soc. 78, 5218 (1956).

  • 29

    the same in both media, but the entropy of formation was 11 e.u.

    larger in 45% dioxane. This entropy difference was attributed to sel

    ective solvation of the nickel and glycinate ions by water, but to

    mixed solvation of the neutral chelate.

    8-Quinolinol (oxine) and its derivatives have been studied by

    ( 1 2 ) ( 3 ) the temperature dependence method ' and calorimetrically. A

    comparison of the thermodynamics of chelation of 2-methyl and 4-

    (4) methyloxines shows more positive AH and AS values for the former.

    The difference in the AH values was ascribed to steric hindrance to

    metal-nitrogen bonding for the 2-methyl derivative, whereas the in

    crease in the AS values was thought to result from reduced solvation

    due to shielding by the 2-methyl group. For a series of 7-substituted

    ( 2 ) oxine-5-sulfonic acids chelates Uusitalo observed a regular vari

    ation in AH values but similar AS values for both the alkaline earth

    and transition metals (14 to 21 e.u. ) and essentially equal AS values

    for a particular metal with different ligands. In contradistinction to

    this, virtually no variation in AH was found for the cobalt, nickel, and

    1. W. D. Johnston and H. Freiser, Anal. Chim. Acta 11, 201 (1954).

    2. E. Uusitalo, Ann. Sci. Fenn. A (87) (1957).

    3. D. Fleischer and H. Freiser, J. Fhys. Chem. 63, 260 (1959).

  • 30

    copper chelates of oxine and its 4-methyl homolog. ̂ The corre

    sponding AS values varied extensively. It should be noted that such

    invariance in Afi had not been found in any of the other studies men

    tioned in this survey.

    At the present time very few thermodynamic data are avail

    able for chelation by O-O ligands. Izatt, et al. ̂ reported data for

    some acetylacetone chelates of transition and heavy metal ions, which

    were obtained by the temperature dependence method (Table X). An

    unusual feature of these data was a higher -AH value for the nickel

    than the copper chelate. Apparently because of the substantial un

    certainty in the data, the authors offered no explanation of this

    phenomenon.

    Calorimetric data for terdentate triphosphate chelates show

    ( 2 ) that they are entirely entropy-stabilized. The heats of formation

    for both the alkaline earth and transition metal ions are endothermic,

    anc1 those of cobalt, nickel, copper, and zinc are even more so than

    that of manganese. These data can be partially explained by the lack

    of any ligand field stabilization for the transition metal ions. Also

    1. D. Fleischer and H. Freiser, J. Phys. Chem. 63, 260 (1959).'

    2. G. Anderegg, Helv. Chim. Acta 48, 1712 (1965).

  • 31

    TABLE X. - -Thermodynamic values for the chelation reactions of acetylacetone and tripolyphosphoric acid.

    Cation

    2+ H+

    Mn

    „ 2+ Co

    XT-2 + Ni

    2+ Cu

    2+ Zn

    2+ Cd

    acae TPP -AG -AH AS -AG -AH AS kcal kcal kcal kcal

    Step mole-* mole" 1 (e.u.) mole-* mole"! (e. u.)

    1 12. 3 2. 8 32 11. 8 0. 1 40. 0

    1 5. 8 2. 5 11 10. 8 -2. 8 46. 4 2 4. 2 4. 7 -1. 8

    1 7. 3 1. 2 21 10. 7 -4. 5 51. 7 2 5. 7 5. 0 2.4

    1 8. 2 6. 7 12 10. 5 -5. 0 52. 7 2 6. 3 6. 3 0 3 3. 0 6. 7 -12

    1 11. 3 4. 7 22 12. 5 -4. 9 59. 2 2 9. 3 6. 6 9

    1 6. 9 1. 9 17 11. 2

    CO

    CD

    I 59. 8

    1 5. 2 1. 4 13 10. 9 -2. 7 46. 2

  • 32

    noted in the explanation was the fact that although zinc binds water

    more tightly than manganese, the reverse is true for the triphosphate

    ion. It was therefore inferred that a similar situation should exist

    for the intermediate transition metals of this series. The observed

    large entropies of formation are comparable to those observed for

    EDTA and DPT A chelates.

    McAuley and Nancollas^ compared calorimetric AH values

    for manganese and cobalt malonates with values obtained by the tem-

    (2) perature dependence method using cells without liquid junction.

    Very good agreement was found. The heat of reaction for the cobalt

    compound (2. 9 kcal/mole) is less endothermic than that for man

    ganese (3. 7 kcal/mole) but the corresponding entropies are both

    27 e. u.

    Very few data are available for ligands containing sulfur donor

    ( 3 ) atoms. Some approximate data, obtained by the temperature de

    pendence method, for the copper and nickel chelates of some poly-

    amines containing thio ether linkages indicate that the metal-sulfur

    bond is weaker than the metal-nitrogen bond, but stronger than the

    metal-oxygen bond.

    1. A. McAuley and G. H. Nancollas, J. Chem. Soc. 989(1963).

    2. V. S. K. Nair and G. H. Nancollas, J. Chem. Soc. 4367(1961).

    3. J. R. Lotz, B. P. Block, and W. C. Fernelius, J. Phys. Chem. 63, 541 (1959).

  • STATEMENT OF PROBLEM

    Although numerous free energy of chelation values can be

    found in the literature, relatively few heats and entropies of chelation

    data have been reported. Many of the latter data have been obtained

    by the less reliable temperature dependence method and therefore the

    relative contributions of the heats and entropies to the free energies

    are oftimes uncertain. Because the heats and entropies of chelation

    provide a more detailed insight into the structural features of chelates

    in solution than the free energy, it is highly desirable to examine

    these parameters.

    This work was undertaken to compare the chelation reactions

    of ligands containing oxygen and sulfur donor atoms by determining

    their heats of chelation, using the more reliable direct calorimetric

    method. The ligands chosen were 8-quinolinol, 2-methyl and 4-

    methyl-8-quinolinol, 8-quinolinol-5-sulfonic acid, quinoline-8-thiol,

    2-methylquinoline-8-thiol and 2, 4-pentanedione. The metal ions

    2+ 2+ 2+ 2+ 2+ 2+ of interest were Mn , Co , Ni , Zn , Cd , and Pb . Since

    many of the chelates possess a low solubility, it was necessary to

    construct and test a calorimeter capable of dealing with highly dilute

    solutions.

    33

  • EXPERIMENTAL

    General Considerations

    The thermodynamics of chelation reported in this study refer

    to the following reactions:

    M + nL ^ ML n

    where M represents a divalent metal ion and L the ligand anion; n

    can take the values of 1, 2, or 3. Charges and molecules of solvation

    have been omitted for simplicity.

    The ligands, Bronsted bases, were generally employed as the

    conjugate acid forms in order to permit convenient determination of

    the equilibrium constants of the above reactions by measuring the

    amount of hydrogen ion displaced by the metal ion. Consequently,

    these additional reactions must also be considered:

    HL H + L (dissociation)

    HL + H > HgL (protonation)

    In order to distribute the measured heat among all of the

    various chelate and ligand species, their solution concentrations must

    first be established. Suitable equations involving the acid dissociation

    constant of the ligand, the total concentrations of reactants, and the

    34

  • 35

    measured hydrogen ion concentration, can be derived for the concentra

    tions of all these species.

    For the concentrations of the chelate species these equations

    require the evaluation of the formation constants, K^, and ^3*

    These constants need not be determined separately because data for

    their evaluation can be obtained simultaneously with the measured heats

    from a series of calorimetric runs in which the total ligand and metal

    concentrations are known and are varied suitably, and in which the

    hydrogen ion concentration is measured. With a knowledge of these

    quantities and the acid dissociation constants, n andpL can be calcu

    lated, as described in the Calculation section. The acid dissociation

    constants, however, must be determined separately.

    The low solubility of many chelates in water frequently required

    the use of a 50 volume % aqueous dioxane reaction medium.

    Because the activity coefficients of the pertinent species are

    unknown, the constants reported are actually concentration quotients.

    To minimize variation in the activity coefficients, a constant ionic

    strength of 0. 1 was maintained with sodium perchlorate. Appropriate

    corrections were applied to the measured hydrogen ion values to

    convert them to concentrations. These corrections were an addition

  • 36

    of 0. 10 to the pH reading in 50% dioxane^ and an addition of 0. 11 in

    water. (2)

    Complications due to metal hydrolysis were avoided by working

    (3) in a sufficiently low pH region for each metal.

    Titrimetric Apparatus

    Titrations for the determination of the ligand dissociation con

    stants were performed in a jacketed beaker which was maintained at

    constant temperature by circulating water thermostated by means of

    a Wilkens-Anderson Lo-Temp bath. The beaker was covered with a

    plastic cap containing holes to accommodate two five-milliliter micro-

    burets, a pair of electrodes, a nitrogen inlet tube, and a 0-50°

    thermometer. A Beckman Research pH meter with a glass-saturated

    calomel electrode pair was used for all pH measurements. Stirring was

    accomplished with a Teflon-covered bar in conjunction with a magnetic

    stirrer.

    Standard sodium hydroxide solution was stored in a one-gallon

    tubulated polyethylene bottle and was forced into the buret through a

    1. S. Takamoto, Q. Fernando, and H. Freiser, Anal. Chem. 3J7, 1249 (1965).

    2. M. S. Harned and B. B. Owen, The Physical Chemistry of Electrolytic Solutions, Reinhold Publishing Corp. , New York, 1950, p. 543.

    3. H. Freiser, R. G. Charles and W. D. Johnston, J. Am. Chem. Soc. 74, 1383 (1952).

  • 37

    two-way Teflon stopcock by means of air which had first been passed

    through Ascarite-packed towers. The nitrogen used to purge the system

    of carbon dioxide and oxygen was passed through an Ascarite-packed

    tower and a gas scrubber which was immersed in the water bath and

    which contained the same solvent as employed in the titration.

    Titrimetric Procedure

    For the determination of the acid dissociation constants a weighed

    amount of ligand was added to the titration vessel, followed by five

    milliliters of standard 0. 1 N aqueous perchloric acid and an equal volume

    of dioxane (when required), and 100 ml of solvent. After assembling the

    titration apparatus, dissolution of the ligand was effected with the aid

    of the magnetic stirrer while the solution was being purged with a

    stream of nitrogen. Slow passage of nitrogen was maintained throughout

    the experiment. Increments of standard 0. 1 N NaOH were added and the

    pH was read after allowing one to two minutes for the reading to sta

    bilize. When working in 50% dioxane, matching increments of dioxane

    were added after each NaOH addition.

    Although the majority of data used for the stability constant

    determinations were obtained from calorimetric runs to be described

    later, some points were derived from preliminary experiments designed

    to establish the maximum concentrations of reagents for a given extent

    of reaction which did not form a precipitate for a specified period of

  • 38

    time. The apparatus and procedure used were similar to those described

    above. Here no titration was performed, but known quantities of ligand

    and metal were mixed, the pH was read, and the time required for preci

    pitation to start was noted.

    Calorimetric Apparatus

    A twin-differential calorimeter, based on the titration calori

    meter described by Tyson, McCurdy, and Bricker, ^ was constructed

    and employed for all enthalpy determinations. The apparatus consisted

    of two 280 ml silvered Dewar vessels embedded into two 16" x 12" x 3"

    Styrofoam blocks placed on top of each other. Covers of 3/8" poly

    ethylene, mounted on the underside of another Styrofoam block, fitted

    snugly into the mouths of the Dewars. For each Dewar, holes were

    drilled through the block and cover to accommodate two pairs of ther

    mistors, a solution bulb, a polyethylene stirrer, and a heater (Fig. 1).

    Except for the stirrer, all these devices were firmly mounted through

    the cover and block.

    Thermistors (Type 51A1, Victory Engineering Co.), having a

    resistance of about 100, 000 ohms at 2 5° and a temperature coefficient

    of -4. 6%/° at 25°, were utilized as temperature-sensing elements.

    To utilize the higher temperature coefficient of high-resistance

    1. B. C. Tyson, Jr., W. H. McCurdy, Jr., and C. E. Bricker, Anal. Chem. 33, 1640 (1961).

  • 39

    ? JOINT

    PLUNGER

    SYNCHRONOUS MOTOR

    T-TUBE

    SYRINGE

    STYROFOAM BLOCK

    POLYETHYLENE COVER

    SOLUTION BULB

    STYROFOAM BLOCK

    THERMISTORS

    BULB OPENING

    HEATER

    STIRRER

    DEWAR VESSEL

    STYROFOAM BLOCK

    ~ - Cross - sectional view of calorimeter.

  • 40

    thermistors and to distribute the temperature-sensors at different

    points in the calorimeter, a set of four thermistors connected in

    parallel was employed. The resistances and temperature coeffi

    cients of about fifty thermistors were determined, and from them

    four closely-matching pairs were selected. These pairs were then

    split up in order to provide a nearly identical resistance-temperature

    response on each side of the calorimeter. Due to the corrosive

    nature of the solutions used, each pair of thermistors was sealed in

    6 mm soft-glass tubing.

    Thermistors follow an exponential relation between resis

    tance and temperature of the form

    X =t exp B(l/T - 1/T ) o o

    where If is the thermistor resistance, B is a constant, and T is the

    absolute temperature. For the thermistors described above^ B

    was found to be 4018. 9 and value of x could be calculated from

    log t = -1. 44168 + 1745. 4/T

    Each set of thermistors was incorporporated into the arms of

    a Wheatstone bridge circuit, as shown in Fig. 2. The magnitudes of

    the other resistances in the circuit were chosen so as to produce

    minimal deviation from linearity of the output voltage vs. temperature

    change, in accordance with the detailed considerations presented. ^

    1. B. C. Tyson, Jr., W. H. McCurdy, Jr., and C. E. Bricker, Anal. Chem. 33, 1640 (1961).

  • 41

    17 KA 5 KA 14.5 KJ1

    1000 A *-,0-5Vi SKA

    9 K A l O O O p f 5 K A 2.8 KANJ/

    SENSING CIRCUIT

    RECORDER

    BUCKING CIRCUIT

    Fig. 2. --Calorimeter circuit.

  • 42

    Because minute temperature differences between the contents of the

    two Dewars are almost inevitable, a bucking circuit was employed

    to adjust the initial base line to zero or some other desired value.

    Initiation of the reaction by mixing the reactants resulted in a

    change in temperature--and hence in the thermistor resistance, pro

    ducing an imbalance potential which was fed into the recorder. The

    latter was a 2. 5 mv full-span Brown recorder with a chart speed of

    l" per minute. Increased sensitivity was attained by inserting a

    Brown Range Change accessory which decreased the span to 1 mv.

    In order to smooth out the noise, a 1000 /uf capacitor was connected

    in parallel to the recorder.

    A glass bulb with a small opening at the bottom, which was

    sealed with beeswax, served to separate the reactants prior to re

    action. A glass rod plunger for breaking the seal and a syringe for

    forcing out the solution were attached to a T-tube which was connected

    by means of a standard taper joint to the top of the solution bulb. For

    improved thermal transfer the bulb wall had been thinned to about 0. 7

    mm by immersion in concentrated hydrofluoric acid.

    Effective and equal stirring in each vessel was accomplished

    with polyethylene stirrers driven by two 200 rpm synchronous motors

    (Model K-2, Bodine Electric Co. ) which could be switched on simul

    taneously. The. stirrer was guided into the vessel through a hole in

  • 43

    the cover whose diameterwas about 0. 5 mm larger than that of the

    shaft, thus providing the only opening to the outside.

    The electrical heaters ( ~ 60 ohms) were made from No. 36

    manganin wire wound on a threaded 3 mm polystyrene rod and fixed in

    place by a very light coating of epoxy resin. The rod was inserted in

    to tightly-fitting thin-walled polyethylene tubing and was fashioned

    into a circular shape by softening in a glycerol bath at 110-120° and

    wrapping around a round object of the desired diameter. The ends of

    the manganin wire were soldered to No. 20 copper leads for con

    nections outside the calorimeter. Resistances were measured with

    a Wheatstone bridge (Leeds and Northrup, Model 4735) at regular in

    tervals. The variation was negligible.

    Known amounts of heat were generated by passing a constant

    current from a Sargent Model IV Coulometric Current Source,

    equipped with a built-in timer, through the heater for a measured

    period of time. Leads from the Sargent instrument were connected

    to the copper leads of the heater via a three-position switch which

    allowed the passage of current through each heater separately or

    through both in series.

    Calorimetric Procedure

    An important advantage of a differential calorimeter lies in

    its ability to cancel the heqtt of dilution and thus to feed only the

  • 44

    signal due to the main reaction into the recorder. When two solutions

    are mixed, two heats of dilution are produced simultaneously, only

    one of which can be compensated on the blank side. It was therefore

    expedient to arrange for the other dilution heat to be negligibly small.

    For this reason the reactant most likely to produce the greatest

    dilution heat was placed on both the reaction and blank sides of the

    calorimeter. In a few cases separate determination of a dilution heat

    was necessary.

    Solutions of the reactants were thermostated at 2 5. 0± 0. 1

    degrees for sufficient time to attain thermal equilibration. The time

    required varied according to the size and thermal properties of the

    vessel and contents. Appropriate amounts of reactant and solvent

    were pipetted into each Dewar to bring the total volume to 225. 0 ml,

    thereby leaving only about a 7 mm air gap above the solution. After

    sealing their openings with beeswax, the bulbs were filled with 12. 00

    ml of solution, the top of which would then be about 5 mm below the

    solution level in the Dewar. The calorimeter was assembled and the

    syringes, opened to one ml in excess of the volume of the solution in

    the bulb, were attached to the T-tubes. The stirrers were turned on

    for a few minutes to homogenize the Dewar solutions. The entire

    apparatus was then allowed to thermally equilibrate for about two

    hours.

  • 45

    At the end of this time the stirrers were switched on and the

    bucking voltage was adjusted to obtain a suitable baseline on the

    recorder. The damping capacitor was turned on and allowed to charge

    up. The chart drive motor was then started and an initial period of

    10-15 minutes was recorded. The wax seal of the bulb on the blank

    side was pierced and the solution was expelled into the Dewar solu

    tion by means of the attached syringe. After a few seconds the syringe

    was removed temporarily to permit the mixed solutions to rise into

    the emptied bulb. The other wax seal was pierced and the heater on

    the blank side was turned on. The bulb solution was then forced out

    at such rate as to maintain the recorder pen at the same position.

    The heater was turned off and on as required in order to compensate

    the heat of reaction as nearly as possible and hence to approximate

    a continuation of the initial period. Subsequent to mixing, about 5-7

    minutes was required to reach thermal equilibrium, after which a

    final period was recorded for sufficient time to give a well-defined

    straight line (~ 10-15 min. ). Immediately after opening the calori

    meter glass-calomel electrodes were inserted into the reaction side

    vessel and the pH was measured.

    At appropriate intervals calibrations were performed to deter

    mine the sensitivity, S, by generating a known amount of heat on the

    blank side and measuring the displacement on the chart paper. Fre

    quently the final period of a run served as the initial period of a

  • 46

    calibration. To determine the difference in response, Ar, between

    the sets of thermistors on the blank and reaction sides current was

    passed through both heaters connected in series. The distance

    between the initial and final periods was measured and then divided

    by the time of heat generation to give Ar.

    In order to estimate the accuracy and precision of the calori

    meter, the well-established heat of formation of water was measured.

    At an ionic strength of 0. 1 the neutralization of perchloric acid with

    an excess of sodium hydroxide yielded heats of -13. 48, 13. 48, -13. 48,

    and - 13. 45 kcal/mole for an average of 13. 47 kcal/mole, with a

    standard deviation of 0. 015. By applying the appropriate heats of

    dilution, ^ a value of -13. 34 kcal/mole at infinite dilution was ob

    tained, in excellent agreement with recently reported values of

    (2) (3) -13. 336 and -13. 337 kcal/mole. Although the maximum sensi

    tivity of the calorimeter was about 0.00003°/mm of chart paper,

    fluctuations in the initial and final periods due to electronic noise

    1. C. E. Vanderzee and J. A. Swanson, J. Phys. Chem. 67, 285 (1963).

    2. C. E. Vanderzee and J. A. Swanson, J. Phys. Chem. 67, 2608 (1963).

    3. J. D. Hale, R. M. Izatt, and J. J. Christensen, J. Phys. Chem. 67, 2605 (1963).

  • 47

    reduced the actual sensitivity to about 0. 0001°. Separate experiments

    indicated that the heat capacity of the calorimeter parts could be

    neglected when estimating temperature changes to one significant

    figure.

    Reagents

    The 1, 4-dioxane (Union Carbide Co.) was purified by refluxing

    over sodium for a few days and then fractionating through a four-foot

    column packed with glass helices. The distillate collected boiled at

    98-99° under 700 mm Hg pressure.

    The metal perchlorates were reagent grade, obtained from

    the G. F. Smith Chemical Co. Approximately 0. 3 M solutions were

    prepared and standarized with EDTA, using the procedures described

    in. ̂ For Cu, Ni, and Mn the indicator was pyrocatechol violet, for

    Zn--Zincon, for Cd and Pb--Xylenol orange, for Co--NH4SCH-

    ( 2 ) PhgAsCl. Solutions of NaClO^ gave negative tests with AgNOg and

    BaClg solutions.

    Standard sodium hydroxide solutions were prepared by diluting

    a 50% solution and were standardized against primary standard grade

    1. G. Schwarzenbach and H. Flaschka, Die Komplexometrische Titration, 5. ed., Ferdinand Enke Verlag, Stuttgart, 1965.

    2. A. J. Cameron and N. A. Gibson, Anal. Chim. Acta 25, 24 (1961).

  • 48

    potassium acid phthalate. Perchloric acid solutions were prepared

    from G. F. Smith Chemical Co. reagent grade acid and were stan

    dardized against the sodium hydroxide solution.

    8-Q,uinolinol (oxine) and 2-methyl-8-quinolinol were Eastman

    Kodak Co. white label grade and were recrystallized from aqueous

    ethanol followed by sublimation. The respective melting points of the

    purified compounds were 73.0-74.0° and 71. 5-73.0°. Reported 74-74°

    n r , A ° and 74 .

    8-Quinolinol-5-sulfonic acid (Eastman Kodak Co. , white label

    grade) was twice recrystallized from boiling 5% HC1 and once from

    boiling water. It was air-dried. Tests with AgNO^ indicated the

    absence of Cl~. Standard solutions of the sodium salt were prepared

    from the free acid by titration to the isoelectric pH with standard

    NaOH.

    4-Methyl-8-quinolinol was synthesized according to the pro

    cedure of Phillips, Elbinger, and Merritt. ̂ After three recrystal-

    lizations from aqueous ethanol the material was sublimed twice. M. p.

    140.0-141. 5. Reported 141°.

    Quinoline-8-thiol (thiooxine) and 2-methylquinoline-8-thiol

    ( 2 ) were synthesized according to Kealey and Freiser and were

    1. J. P. Phillips, L. L. Elbinger and L. C. Merritt, J. Am. Chem. Soc. 71, 3986 (1949).

    2. D. Kealey and H. Freiser, Talanta JJ3 (1966) (in press).

  • 49

    converted to their sodium salts. After preparing solutions of these

    reagents, an aliquot was saved for assay by potentiometric titration

    with silver ion!^ Since these reagents oxidize slowly in solution,

    the assay was performed immediately after initiation of the reaction

    in the calorimeter.

    2, 4-Pentanedione (acetylacetone) (Eastman Kodak Practical

    grade) was washed successively with NaHCOg solution and water, then

    (2 ) dried over anhydrous sodium sulfate, and fractionally distilled.

    o B. p. 134.5-135.5 under 700mm Hg pressure. Gas-chromatographic

    analysis using a silicone rubber column indicated substantially less

    than one per cent of impurities.

    1. M. W. Tamele and L. B. Ryland, Anal. Chem. 8_, 16, (1936).

    2. D. Dyrssen, Svensk. Kem. Tidskr. 64, 213 (1952).

  • CALCULATIONS

    Acid Dissociation Constants

    The acid dissociation constants of the 8-quinolinols can be

    represented by

    [H+] [HL] / [H2L+] = Knh (1)

    [H+] [L"] / [HL] = Koh (2)

    These equations apply to 8-quinolinol-5-sulfonic acid as well, for in

    the pH range employed only one anionic form (the same as for the other

    oxines) is important.

    In the determination of the acid dissociation constants by

    potentiometric titration, the following equations must be considered:

    Mass balance

    CL = [H2L+] + [HL] + [L~] (3)

    Charge balance

    [H2L+] + [H+] + [Na+] = [A~] + [OH] + [L~] (4)

    Here CT refers to the total ligand concentration and [A ] is equal to J-F

    the concentration of strong acid added, which in the case of the sulfona

    ted ligand is provided by the free sulfonic acid. [Na+] is equal to the

    concentration of NaOH titrant added.

    Combination of these equations with the expressions for the

    acid dissociation constant yields

    50

  • 51

    „ _ [H+] {CL [A] - [Na+] - [H+]

    [A] - [Na+] - [H+] (5)

    K - [H+j - tA"] - [OH"]] (6)

    OH CL " [tNa+] " [A" I"* [OH"] _J

    The dissociation constant of acetylacetone in 0.1M NaClO^

    has been reported in the literature. ̂

    Chelate Formation Constants """"

    The chelate formation constants may be evaluated from a

    knowledge of two parameters, H, the average number of ligand mole

    cules bound per metal ion, defined as

    [ML+] + 2[MLj n = ( 7 )

    M

    and [L ], the ligand anion concentration. These parameters, in turn,

    can be calculated from the acid dissociation constants and the following

    expressions describing the composition of the mixed solutions of ligand

    and metal:

    1. J. Rydberg, Svensk. Kem. Tidskr. 67, 499 (1955).

  • 52

    Mass balance

    CM = Cm2+] + [ML+] + [ML2] (8)

    CL = [HGL4] +[HL] + [L_] + [ML+] + 2[ML2] (9)

    [C10~] = [A"] +2Cm (10)

    Charge balance

    2[M2+]+[ML+]+[H2L+]+[H4'3+[Na+] = [L~]+[0H"]+[C10^ (11)

    Appropriate combination of these equations gives

    C C +[A"]-[H+]-,[Na+] ( K +[H+]

    * = ^ < 1 2 >

    and

    r . - , K ^ - ^ ' ^ I V oh

    t»*j I

    For details of these derivations the work of Johnston^ should

    be consulted.

    Another expression for n in terms of formation constants and

    [L ] can be derived by expanding the denominator of the defining equa

    tion (8) and expressing [ML+] and [MLg] in terms of the formation

    constants, = |> +]/[M2+] • [L~] andK12= [MLg]/[M2+] • [l"]2.

    1. W. D. Johnston, Ph. D. Thesis, University of Pittsburgh (1953).

  • 53

    Then

    K^L'I +2K12[L"]2 (14) n

    I+KJL"] +K12[L"]2

    This equation can be rearranged to give

    n = K,JL~] — + K (15) [L"](l-n) 1 - n

    A least squares plot of of n vs. [L ] 2-n yields a i L - j ( l - n ) — L J

    straight line with a slope equal to and an intercept equal to K^.

    The log Kg values for the nickel chelates of oxine- 5-sulfonic acid

    in water and in 50% dioxane were obtained graphically from the pL

    value at n = 2.5, since the separation between log Kg and log Kg was

    greater than two log units.

    Heats of Reaction

    The heat Q generated by the passage of a steady current i for a

    time t through a resistance E is given by

    where 4.1840 is the factor for converting joules to defined calories.

    By using a fixed resistance and the same current setting, Q becomes

    only a function of time.

    To find the experimental heat of reaction, Q^, the final period

    is first extrapolated back to the end of the initial period when the

    4.1840 ( 1 6 )

  • 54

    reactants were mixed, and the distance, d, (Fig. 3) between these

    periods is measured. Qr is then calculated from the following equa

    tion:

    Qr = Q + (d + Ar • t)S (17)

    where Ar is the difference in response, t, the time, and S, the sensi

    tivity, as described previously.

    Heats of Reagent Dissociation

    For one-step reactions, such as protonation and dissociation,

    where the desired reactions can be forced to proceed quantitatively by

    the use of an excess of a reagent, calculation of the heat of reaction is

    simple:

    AH = Q IA (18) r

    where A is the number of millimoles of the desired species undergoing

    reaction. The heat of protonation, AH.m) is measured directly, but JNri

    the heat of dissociation, AHQJJ, is obtained as the difference between

    the heat of neutralization and the heat of dissociation of water:

    HL + OH" —> L~ + H-O AH . (19) 2 neut

    Ho0 —> H+ + OH" AH (20) 2 w

    HL —> H+ + L" AHOH(SH) (21)

    Heats of Chelation

    The heats of chelation of the thiooxinates were determined

    directly, _i. £., the reaction goes to completion, so that equation (18)

  • TIME

    Fig. 3. --Typical t ime-temperatur

  • 56

    applies in this case as well.

    In the determination of the heats of formation of the oxines, the

    experimental heat of reaction, Q^, is a composite of these heats of

    reactions:

    HL — ->H + L AHOH (22)

    M2+ + L~ — -> ML+ AHT (23)

    2+ M + 2L — ̂ ML2 AH12 (24)

    HL + H+ —> H2L+ AHNH (25)

    Here L refers to the total ligand which becomes bonded to

    metal, ji. e., L = ML+ + 2MLg. Consequently

    Qr = ML+AH1+ML2AH12 + H2L+AHNH + (ML+ + 2ML2)AHOH (26)

    AH^JJ and AHQJJ are determined separately and ML+, MLG,

    and H2L+, which represent the number of millimoles formed of these

    species, are calculated from

    (ML+ + 2ML ) = V • C n (27)

    + V ' ( C L ~ C M S ) H2L = —i — — T ~ < 2 8 )

    I [ H ] )

    + V'CM" M L * u + I/ (K2LL'J) )

  • 57

    where V is the total volume of solution. It should be noted that volume

    shrinkage occurs on mixing dioxane and water, which 25° gives a cor

    rection factor of 0. 982^

    Substituting the known quantities, we obtain

    M L + ' A H 1 + M L 2 ' A H 1 2 = Q c h e l ' ( 3 1 )

    where %hel " Qr ' (ML+ + 2ML2» AHOH " H2L+ ' ^NH'

    The heats of chelation, AH^ and AH^2< can then be evaluated

    by solution of simultaneous equations obtained at low n (n < 1) with

    those obtained at high n (n> 1).

    For the acetylacetonates the calculations are based on each

    separate step of chelate formation. Here is due to reactions (19)

    and (23), the reverse of reaction (20), and the following:

    ML+ + L~ —> ML2 AH2 (32)

    In this case ML+ is obtained from

    ML+ = V- CM • n/| 2 + (l/K2[L"])j (33)

    + + The millimoles of ML , MLg, OH , and H present initially

    are subtracted from their final amounts. The differences thus obtained

    are substituted in the following relation

    ML+ • AHt + ML2 * AH2 = Qchel (34)

    1. D. Fleischer, Ph.D. Thesis, University of Pittsburgh (1959).

  • 58

    where Q . = Q - H • AH - OH • AH ,. The appropriate sets chel r w neut r

    of equations are then solved simultaneously for AH^ and AH2 in a

    manner similar to that above.

    Computer programs were written to perform the above

    calculations.

  • ERRORS

    The errors associated with the AH values reported in this

    study stem mainly from two sources: uncertainties in the appropriate

    equilibrium constants and in the experimental heats of reaction. In

    the majority of cases the simultaneous formation of more than one

    species is unavoidable, requiring a knowledge of the stepwise equilib

    rium constants and thereby increasing substantially the overall

    error. For most reagent heats of dissociation the essentially quan

    titative conversion of the neutral ligand to its cationic or anionic

    form with an excess of perchloric acid or sodium hydroxide obviated

    the necessity for the use of dissociation constants. Furthermore,

    the relatively high solubilities permitted temperature changes as high

    as 0. 05 .to 0.1° to be attained. Based on the experimental data, the

    uncertainties in the heats of dissociation of oxine, 2-methyloxine,

    oxine-5-sulfonic acid, and acetylacetone.probably do not exceed 0. 05

    kcal/mole. Due to a more sparing solubility in the case of 4-methyl-

    oxine, the error is about 0. 1 kcal/mole. A similar magnitude of

    error is expected for the heat of dissociation of the SH group of thio-

    oxine and its 2-methyl derivative. The heats of protonation of these

    compounds, however, depend on calculation of the extent of reaction

    59

  • 60

    from the measured final pH and the spectrophotometrically deter

    mined pKnvT„, so that the uncertainties may amount to 0. 3 kcal/mole. f NH j

    The uncertainties in the dissociation constants should be less

    than 0. 05 log unit, corresponding to about 0. 07 kcal/mole in the free

    energy. The errors in the entropy values should be given by the sum

    of the errors in the free energy and enthalpy values, multiplied by

    3. 35 (since at 25° one kcal/mole corresponds to 3. 35 entropy units).

    Except for the case of the thiooxinate chelates, errors in the

    heats of chelation arise from the uncertainties in the formation con

    stants and in the calorimetric determinations. Analogous to the

    previous procedure involving reagent heats, a large excess of metal

    ion converted the anionic form of the thiooxinates completely to the

    ML+ species. The limited solubility of these chelates, however, re

    stricted the temperature changes observed to only a few thousandths

    of a degree, which resulted in a greater uncertainty in the AH^ values

    (0. 5 to 1.0 kcal/mole), and precluded the determination of AH^ values.

    For the remaining chelates determination of the formation

    constants was necessary. Since these were determined from some

    what fewer, but perhaps more reliable points, and with a greater

    variation in the total concentrations of metal and ligand, only a

    slightly larger uncertainty than that prevailing in the ordinary poten-

    tiometric titration technique is to be expected. On the basis of the

  • 61

    formal error treatment presented by Fleischer, ̂ an error of about

    0. 2 log units might be anticipated in this work. Judging from a com

    parison of the values obtained in this study with those determined by

    Mr. Ted Carnavale in this laboratory using potentiometric titration

    under similar conditions, this assumption appears to be justified:

    System log Kx log K2 log K12 l o g K x logK2 log K^2

    Mn(II)- -2-methyloxine 6. 84 6. 46 13, 30 6. 81 6 . 2 9 13. 10

    Mn(II)- -oxine-5-sulfonic acid

    CO 5. 92 13. 05 7. 05 6 . 1 3 13. 18

    Co(II)- - 2 - methyloxine

    o

    CO CO 8. 66 1 7 . 4 6 8 . 5 9 8 . 7 9 17. 38

    Pb(II)-- 2 - methyloxine 9. 85 7. 10 16. 95 9. 97 7 . 2 1 17. 18

    Pb(II)-- 2 - methyloxine 10. 01 7 . 2 8 1 7 . 2 9

    The values for the Cu, Ni, and Zn chelates of oxine and 2 -

    methyloxine generally agree with those reported for a 50% dioxane-

    0. 3 M NaClO . medium at 20°. ̂ 4

    The total amount of chelate and protonated ligand formed is

    determined from the final pH, and independent of the formation con

    stants. The relative amounts of the various chelate species formed,

    however, is governed by the formation constants. Hence the errors

    1. D. Fleischer, Ph. D. Thesis, University of Pittsburgh (1959).

    2. H. Irving andH. S. Rossotti, J. Chem. Soc. 2910 (1954),

  • 62

    introduced into the AH values by inaccuracies in the formation con

    stants result from uncertainties in the apportionment of the observed

    heat of reaction among the species present, since an error in the

    concentration of one chelate species affects the concentration of the

    other in the opposite direction. Because the stepwise AH values are

    usually rather similar, however, such uncertainties produces errors

    of less than approximately 0. 5 kcal/mole per step. This is supported

    by the following AH values obtained through the use of formation con

    stants and those obtained from proton displacement or direct re

    actions which could be forced to proceed quantitatively to only one

    chelate species:

    Method of Calculation System Dependent on Independent of

    -AHRAH12-AHI3

    Cu(II)--oxine 10.2 19.6 9,4 18.9 9 . 7 2 0 . 2

    Cu(II)--oxine-5-sulfonic acid 18.6 18.6

    Ni(II)--oxine-5-sulfonic acid 25.4 24.2

    As in the case of the reagent heats of dissociation, that por

    tion of error in the heat of chelation due to calorimetric errors will

    be a function of the magnitude of the heat evolved, which, in turn,

    will generally be determined by the solubility of the chelate. (The

    term solubility is employed here in the sense of the tendency of the

    chelate to remain in solution for the duration of a calorimetric

  • 63

    measurement, and does not necessarily indicate the equilibrium

    solubility.) Consequently, for the soluble chelates of oxine-5-

    sulfonic acid and acetylacetone the total error, chiefly derived from

    uncertainties inK^, in AH^, AH^* and AH^g will be about 0. 5, 1. 0,

    and 1. 5 kcal/mole. A similar magnitude should be valid for AH^ of

    the oxinates and 2-methyloxinates. Their AH^ values, however,

    probably deviate by 1.5 kcal/mole, especially for the Mn, Co, and

    Zn chelates. The greater insolubilities of the 4-methyloxinates,

    which limited the temperature changes to only 0. 0005° to 0. 0006° for

    the high n runs of Mn, Co, and Zn, reduce the reliability of AH^ to

    0. 5 to 0. 8 kcal/mole, and AH^ t° about 2. 0 kcal/mole.

    The errors in the entropies of chelation can be found analo

    gously to that described for the entropies of dissociation.

    The errors in AH0 and AS„ are, of course, equal to the sum 4 &

    of the errors in AH, and AHL 0 and in AS.. and AS. 9.

  • DISCUSSION

    Comparison of Methods

    Two methods are generally employed for the determination of

    heats of reaction in solution: direct calorimetry and the variation of

    the equilibrium constant with temperature. The latter method is

    based on the van't Hoff equation

    dink/ dT = AH/RT2

    and hence a plot of log K vs. 1/T yields a straight line with a slope of

    AH/2. 303R. The heat of reaction, however, is not independent of

    temperature, but varies with changes in the heat capacities of the

    system. An indication of the magnitude of this variation is provided

    by the recent data of Izatt, et al. ^ for the heat of chelation between

    copper (II) and alanine: the AH and AH values at 10, 25, and 40° 1 ltt

    are 5. 38, 4. 50, and 3. 99 kcal/mole and 10. 85, 9, 75, and 9. 64 kcal/

    mole, respectively. A similar variation was also observed in the

    reagent heat of dissociation. If these data are typical, the error

    resulting from the assumption of temperature independence of AH

    over a short temperature range would be tolerable in many cases.

    1. K. P. Anderson, D. A. Newell, and R. M. Izatt, Inorg. Chem. 5, 63 (1966).

    64

  • 65

    Recently, a new family of general equilibrium equations was devel

    oped to represent the temperatur