Conceptio › Archive › Internet Archive

Internet Archiveopen access

The chemistry of the coordination compounds

Bailar, John C. (John Christian), 1904-1991

Internet Archive · Books · License: Open Access

Open Source ↗

chemistry organic physical

coordination compounds, chemistry, organic, chemistry, physical

:< -'•■■'

tsssBBm

HBP

hIi

■1

■ flbn

Klllil

hwhh ■■!■"■■ Imlmfl

i'-.;! mm

■ *.(■

."*••. .■.:•>: '■' ■: ■'-*■: 1**

university of

Connecticut

libraries

*!•

hbl, stx

QD 471.B23

n!?f 7! if uflf. .tf. the coordination

3 T1S3 QDb^

comp

bflr2b 1

->3

THE CHEMISTRY

of the

COORDINATION COMPOUNDS

1 » *

Edited by

JOHN C. BAILAR, JR.

University of Illinois Urbana, Illinois

DARYLE H. BUSCH

Editorial Assistant

American Chemical Society Monograph Series

REINHOLD PUBLISHING CORPORATlbN

NEW YORK CHAPMAN & HALL, LTD., LONDON

1956

REINHOLD PUBLISHING CORPORATION

Library of Congress Catalog Card Number 56-6686

REINHOLD PUBLISHING CORPORATION

Publishers of Chemical Engineering Catalog, Chemical Materials Catalog, "Automatic Control," "Materials & Methods"; Ad- vertising Management of the American Chemical Society

Printed in the U.S.A. by The Waverly Press, Inc., Baltimore, Md.

Fred Basolo B. P. Block

Contributors

Robert C. Brasted Clayton F. Callis Leallyn B. ('LAPP William E. Coolly Bodie E. Douglas

GUNTHER L. ElCHHORN

Stanley J. ( Iill Roi 1). Johnson

Hans H. JoNASSEN

Raymond \\ Keller Stanley Kirschner Ernest H. Lyons, Jr. J. A. Mattern Niels C. Nielsen Thomas 1). O'Brien Robert \\\ Parry James V. Quagliano R. L. Rebertus Carl L. Rollinson Donald II. WlLKlNS

Digitized by the Internet Archive

in 2012 with funding from

LYRASIS Members and Sloan Foundation

http://archive.org/details/chemicoorOObail

General Introduction

American Chemical Society's Series of Chemical Monographs

By arrangement with the Interallied Conference of Pure and Applied Chemistry, which met in London and Brussels in July, 1919, the American Chemical Society was to undertake the production and publication of Scientific and Technologic Monographs on chemical subjects. At the same time it was agreed that the National Research Council, in cooperation with the American Chemical Society and the American Physical Society, should undertake the production and publication of Critical Tables of Chemical and Physical Constants. The American Chemical Society and the National Research Council mutually agreed to care for these two fields of chemical progress. The American Chemical Society named as Trustees, to make the necessary arrangements of the publication of the Monographs, Charles L. Parsons, secretary of the Society, Washington, D. C; the late John E. Teeple, then treasurer of the Society, New York; and the late Pro- fessor Gellert Alleman of Swarthmore College. The trustees arranged for the publication of the ACS Series of (a) Scientific and (b) Technological Monographs by the Chemical Catalog Company, Inc. (Reinhold Publish- ing Corporation, successor) of New York.

The Council of the American Chemical Society, acting through its Com- mittee on National Policy, appointed editors (the present list of whom appears at the close of this sketch) to select authors of competent authority ID their respective fields and to consider critically the manuscripts sub- mitted.

The first Monograph of the Series appeared in 1921. After twenty-three years of experience certain modifications of general policy were indicated. In the beginning there still remained from the preceding five decades a distinct though arbitrary differentiation between so-called "pure science" publications and technologic or applied science literature. By 1944 this differentiation was fast becoming nebulous. Research in private enterprise had grown apace and not a little of it was pursued on the frontiers of knowledge. Furthermore, most workers in the sciences were coming to see the artificiality of the separation. The methods of both groups of workers are the same. They employ the same instrumentalities, and frankly recog- nize that their objectives are common, namely, the search for new knowl- edge for the service of man. The officers of the Society therefore combined the two editorial Boards in a single Board of twelve representative members.

Also in the beginning of the Series, it seemed expedient to construe

VI GENERAL INTRODUCTION

rather broadly the definition of a Monograph. Needs of workers had to be recognized. Consequently among the first hundred Monographs appeared works in the form of treatises covering in some instances rather broad areas. Because such necessary works do not now want for publishers, it is con- sidered advisable to hew more strictly to the line of the Monograph char- acter, which means more complete and critical treatment of relatively restricted areas, and, where a broader field needs coverage, to subdivide it into logical subareas. The prodigious expansion of new knowledge makes such a change desirable.

These Monographs are intended to serve two principal purposes: first, to make available to chemists a thorough treatment of a selected area in form usable by persons working in more or less unrelated fields to the end that they may correlate their own work with a larger area of physical science discipline; second, to stimulate further research in the specific field treated. To implement this purpose the authors of Monographs are ex- pected to give extended references to the literature. Where the literature is of such volume that a complete bibliography is impracticable, the authors are expected to append a list of references critically selected on the basis of their relative importance and significance.

AMERICAN CHEMICAL SOCIETY

BOARD OF EDITORS

William A. Hamor, Editor of Monographs Associates

L. W. Bass C. H. Mathewson

T. II. Chilton Laurence L. Quill

Norman Hackerman W. T. Read

J. Bennett Hill Arthur Roe

C. G. King Walter A. Schmidt

S. C. Lind E. R. Weidlein

Prefa

Werner's coordination theory has been a guiding principle in inorganic chemistry and in the theory of valence since its publication sixty years ago. Indeed :t might be said to underlie our modern concepts of molecular structure. The current theories of acidity, basicity, amphoterism, and hydrolysis grew directly from it, and the assumption of the complete ioniza- tion of solid salts is implicit in it. In recent years, the coordination theory has found increasing application in many types of chemical work. For example, its usefulness in the selection of organic precipitants for metallic ions and in explaining biological phenomena are well known. It is also the basis for our understanding of the role of metal ions in leather tanning, in the dyeing of cloth, and in regulating plant growth. Coordinating agents are used in winning metals from their ores, in electroplating, in catalyzing reactions and in obviating the effects of undesirable catalyses, in precipi- tating metallic ions and in preventing their precipitation, and in many other ways. Still other uses await study and exploration.

So much interest has developed in the theory of coordination and in coordination compounds in recent years that a need has arisen for a book describing them. I began the preparation of such a book several years ago, but the literature on the coordination compounds is so vast, and is growing so rapidly, that it soon became apparent that the task is too great for one person. I have therefore asked some of my students and former students to help me with it. I am grateful to them for their help, and proud to present their work.

No attempt has been made to cover the chemistry of coordination com- pound.^ completely to do so would require many volumes. Rather, we have attempted to select ideas which are fundamental and stimulating and applications which are both illustrative and useful. Even so, it has been necessary to omit extensive discussion of such important topics as the use of complex ions as catalysts, metal ion deactivators, methods of preparing complex ions, and the details of many physical methods which are used in the study of coordination compound.-.

In the interest of saving space, we have often used a single reference number lor several related articles. When one of these, articles is referred to later, it is designated by the original number, followed by a letter of the alphabet which show- its position in the list.

Our thanks are due to Prof. \. J. Leonard. Prof. C. S. Vestling, Prof.

vii

viii PREFACE

II. A. Laitinen and Dr. Eleanora C. Gyarf as who have read portions of the manuscript, and have made valuable suggestions concerning them.

In addition to serving as a coauthor, Dr. Daryle II. Busch has assisted a greal deal with the editorial work, and I wish to express special gratitude to him. Without his excellent help, it is doubtful if the work could have been completed.

A person who has never written a book may wonder why authors so fre- quently acknowledge the patience and understanding of their wives. These are, indeed, not idle words. Many of the hours which went into the working; on this book were taken from evenings which would otherwise have been spent with my family or from time which might have been spent in doing the many odd jobs that tall to the lot of every householder. My wife has not only borne this with patience and understanding, but has lent valuable advice and encouragement . To her goes my most grateful acknowledgment.

John C. Bailar, Jr.

I'rbana, Illinois June, 1950

Contents

Preface

1. A General Survey ok the Coordination Compounds, John

( ' . Bailor, Jr., and Daryle II . Busch 1

2. The Early Development ok the Coordination Theory,

John C Bailor, Jr 100

.*;. Modern Developments The Electrostatic Theory of Co- ordination Compounds, Robert IT. Parr;/ (Did Raymond N.

Keller 119

4. Modern Developments The Electron Pair Bond and the

Structure ok Coordination Compounds, Raymond X . Keller

ami Robi it II'. Parry 1">7

5. Chelation and the Theory of Heterocyclic Ring Forma-

tion [nvolving Metal Ions, Robert W. Parry 220

(i. Large Rings, Thomas I). O'Brien

,. General [somerism ok Complex Compounds, Thomas I).

O'Brien 261

8. Stereoisomerism ok Hexacovalent Atoms, Fred Basolo 274

'.». Stereochemistry ok Coordination Number Four, B. R.

Block 354

10. Stereochemistry and Occurrence ok Compounds Involving

the Less Common Coordination Numbers, Thomas b.

O'Brien :)82

11. Stabilization ok Valence States Through Coordination,

James V . Quagliano and R. L. R<l>< rtus 398

12. Theories oi Acids, Bases, Amphoteric Hydroxides and

Basic Salt-, as Applied to the Chemistry ok Complex

ifPOUNDS, Fred Basolo 4 Hi

!:;. Olation and Related Chemical Processes, Carl L. Rollinson ws

14. The Poly A< ids, Hans B. Jonasst n mat Stanley Kirschner 472

15. Coordination Compounds of Metal Ion- with Olefins and

Olefin-Like Substances, Bodu E. Hour/las 487

\ CONTESTS

. 16. Metal Carbonyls and Nitrosyls, ./. A. Mattern and Stanley

J. Gill 509

17. Organic Molecular Compounds, Leallyn B. Clapp 547

-^ IX. Physical Methods i\ Coordination Chemistry, Robert C.

Brasted and William E. Cooley 563

19. Coordination Compounds i.\ Electrodeposition, Robert W.

Parry <ni<l Erru Bt //. Lyons, Jr 625

20. 'I'm-: Use of Coordination Compounds in Analytical Chem-

istry, James V . Quagliano and Donald H. Wilkins 672

21. Coordination Compounds i\ Natural Products, Gunther L.

Eichhorn 698

22. Dyes and Pigments, Roy I). Johnson and Niels ('. Nielsen 743

23. Water Softening Through Complex Formation, Roy I).

Johnson and Clayton F. Cattis 7(>8

Subject Index 785

A General Survey of the Coordination Compounds

John C. Bailor Jr., and Daryle H. Busch*

University of Illinois, Urbano, Illinois

Since coordination compounds differ greatly in nature and stability, chemists are not completely agreed on a simple definition of the term. Marly workers in the field had few of the modern physical-chemical tools at their disposal, and if a material satisfied the law of definite proportions, they were inclined to consider it a compound. For example, crystals of the composition (XHOaZnOo are readily obtained from an aqueous solution containing zinc chloride and ammonium chloride. These can be recrystal- lized without change of composition, and the substance was long considered to be a complex compound in which zinc shows a coordination number of five. X-ray analysis has shown, however, that only four of the chlorine atoms are close to the zinc while the fifth is much more distant. Similarly, the clathrates were once believed to be coordination compounds.

According to the theory of Sidgwick and Lowry, a coordinate bond (and hence, a coordination compound) can be formed between any atom or ion which can accept a share in a pair of electrons (the acceptor) and any atom or ion which can furnish a pair of electrons (the donor). The donor is non- metallic it may be part of a neutral molecule, like CO, IU>, or XTI:: . or part of an ion, like CI , COjT or NH2CH2COO . Ordinarily, an acceptor requires several donors, which may be alike or different. The resulting complex may be a positive ion, a negative ion, or a neutral molecule.

Even if we accept the idea that a coordinate bond consists of a shared pair (or pairs) of electrons, a question remain- as to the nature and the necessary degree of such sharing. In many cases, the donor and acceptor are bound in such a way ih;it the reaction of formation is noi reversible t<» any detectable degree. In aqueous solution, the hexamminecobalt(III) ion [Co(NH3)e]+++, shows no detectable dissocial ion1 and the analogous tri8(ethylenediamine)cobalt(III) ion, [Co en.,]* * \ retains its optical ac tivity in solution for many week.- at ordinary temperatures. Both of these

Now :it Ohio State University, Columbus, Ohio.

1. Flagg. ./. Am. I 63. 057 L941

CHEMISTRY OF THE COORDINATION COMPOUNDS

ions arc stable in concentrated hydrochloric acid, and react only slowly with hydrogen sulfide and with sodium hydroxide.

The copper(II) tetrammine ion, [Cu(NH3)4]++, can be easily detected in solution by its deep blue color, and its salts can he crystallized from solution. It is of a different order of stability from the cobalt (III) hexam- niine ion, however, as it is readily destroyed by acids or by heating. In solution, it exists in equilibrium with [Cu(H20)4]++ and ammonia. The fact that the formation of the complex is accompanied by a color change, by a change in oxidation-reduction potential of the copper(II), and by other changes in properties clearly indicates that there is a true chemical bond between the copper ion and the ammonia molecules.

Sodium chloride absorbs ammonia when under pressure, but liberates it when the pressure is released'-'. No doubt there are attractive or adsorptive tones which tend to hold the two substances together, but they are weak and poorly characterized.

In general, the small, highly charged cations form the most stable co- ordinate bonds, and it is often mistakenly supposed that the ability to form complexes is limited to the transition metals. This is far from being so, as is seen from the fact that the beryllium derivative of acetylacetone can be distilled without decomposition at 270°C.

Even the alkali metal ions form complexes, as shown by the work of Sidgwick and Brewer3. They found that sodium benzoyl acetone has the properties of a salt; it is insoluble in nonpolar solvents, and upon heating it chars instead of melting. If recrystallized from 95 per cent ethanol, it takes up two molecules of water from the solvent, yielding a dihydrate that melts at 1 15°C and is appreciably soluble in toluene. It is evident that the dihydrate is a chelated compound.

CH,

C — Ox H20

< X

c=o' xh2o /

CH3

Salicylaldehyde (and similar compounds) also forms sodium chelates3' 4' h- 6. The nature of the electron sharing is discussed in Chapters 3 and 4.

2. Clark, .1///. ./. 8ci.t 7, 1 (1924).

l^wick and Brewer, ./. Chem. Soc, 127, 2379 (1925); Brewer, J. Chem. Soc, 1931, 361. \. Hantssch, />'- .39, 3089 (1906).

5 Weygand and Porkel, ./. prakt. Chem., 116, 293 (1927).

6 Brady and Bodger, ./. Chem. Soc., 1932,952.

GENERAL SURVEY 3

Suffice it to Bay here that stability depends upon many factors and cannot be directly correlated with bondPtype. Among the many other factors that are important in determining stability are charge on the acceptor atom, nature of the donor atom and of the molecule of which it is a part, chela- tion, cationic, anionic, or neutral nature of the complex, and the nature of the ion with which it is associated (if the complex is an ion).

The relationship between the donor and acceptor atoms is especially interesting. Nearly all of the complexes of the light metals (Periodic groups IA, IIA, IIIB, IVB) contain oxygen as the donor atom. It may be furnished in the form of water, hydroxide ion, an oxyanion, an alcohol, ether, ketone, or in a variety of other ways. These light metals seldom coordinate with molecules containing nitrogen, sulfur, carbon, or the halogens. Vanadium, at the head of group VB, is a powerful oxygen coordinator, but also shows some ability to form ammines and complex cyanides. Proceeding across the periodic table toward the right from vanadium, we encounter elements which easily coordinate with nitrogen. Thus, chromium forms a large num- ber of ammines, most of which are slowly destroyed in water solution. The ammines of manganese are still less stable, and neither iron(II) or iron(III) ion reacts with ammonia in water solution to give ammines. These ions coordinate instead with hydroxy! ions generated in the water by the addi- tion of ammonia. With cobalt, nickel, copper, and zinc, however, stable ammines are formed. The ions of these metals retain the ability to co- ordinate with oxygen in even greater degree than do the ions of the lighter metals, but the tendency to form links with nitrogen is still more pro- nounced. Starting with vanadium, too, we see an increasing tendency to coordinate with carbon — all the elements from vanadium to zinc form stable cyanides, those from chromium to nickel form carbonyls, and copper, at least, forms compounds with olefinic substances. The ability of the metals in this series to combine with sulfur also increases toward copper. Vanadium, chromium and manganese occur in nature in oxide ores, iron both in oxide and sulfide ores, and cobalt, nickel, copper and zinc largely as sulfide ores.

In the fifth and sixth series of the periodic table, there is an increased tendency to form stable complexes with halides. This is present in the fourth series to some degree, but is increasingly important in the later series, as is illustrated by the solubility of silver chloride in hydrochloric acid and the reaction of platinum and gold with chlorine water and aqua regia to form [PtCle]" and [AuCl4]~.

The elements of Periodic groups VIII, IB, and IIP are of special interest.. All of them form complex cyanides, but only palladium, silver, platinum, rhodium, and mercury are known to form compounds with the ethylenic dou- ble bond. All of them form ammines (the ammines of mercury readily Lose

4 CHEMISTRY OF THE COORDINATION COMPOUNDS

protons, hut the metal-nil rogen bond remains), but the platinum metals and gold form few complexes containing a metal-oxygen bond. This does not mean that such a bond is not stable, but only that the metal-halide and metal-sulfur bonds are more stable.

The metals of periodic groups II I A, EVA, and \ A form many complexes in which the donor atom is oxygen, sulfur, or a halogen. Compounds in which the donor is carbon or nitrogen are much less common.

The Donob Properties of the Halogens

The halide ions often coordinate strongly, and halo- complexes are well known; fluorosilicates, bromoplatinates, and iodomercurates are familiar. These ions are often thought of as substituted oxy- anions, but this has arisen through pedagogic convenience rather than strict parallelism, for while a halide ion occupies one coordination position, just as an oxide "ion" does, its elect rovalcnce is 1 instead of 2. Thus the statement that Na2SiF6 is analogous to Na^iOg is somewhat misleading, for in solid sodium sili- cate, the silicate ions are linked together through oxygen atoms in such a way that each silicon is surrounded by four oxygens, while in the fluoro- silicate, each silicon is surrounded by six fluorines. A much closer analogy exists between the halide ions and the hydroxyl ion, as is shown by the series II,|PtCl6]; H2[PtCl5(OH)]; H2[PtCl4(OH)2]; H2[PtCl3(OH)8]; H2[PtCl2(OH)4]; H2[PtCl(OH)5]; H2[Pt(OH)6], all of the members of which are known ex- cept the fourth. These acids, or their alkali salts, can be obtained from the chloro-platinate by stepwise substitution of hydroxo- groups for chloro- groups7 • 8 • 9 • 10.

For convenience, the complexes formed by halide ions may be considered to be of two general types; those containing only halide ions as ligands (with the possible exception of solvent molecules) and those containing halide ions as a less abundant donor species, as is the case among the halopen- tammines of cobalt(III) and chromium(III). Although the stabilities of complexes is generally dependent both on the nature of the central metal ion and on the nature of the donor group, these complexes may be grossly divided into two major stability groups; i.e., those very stable complexes of the heavy metals, such as the platinum group metals and mercury, which give only a faint test for halide ion in water solution, and those relatively labile halide complexes of the type formed by the elements of the first transition group and, in general, the more electropositive metals. These

7 Miolati and Bellucci, /. anorg, Chem., 26, 209 (1001).

B Miolati, /. anorg. Chem., 22, 145 (1900)

9 Miolati, Z. anorg. Chem., 88, 261 (1903).

in Bellucci, /. anorg. Chem., 44, 168 (1906

GENERAL 8URVE1 5

two major stability groups correspond to the penetration and normal com- plexes discussed in Chapter 4.

Many of the reported halide complexes of metallic elements are char- acterized solely by the composition of solids obtained from solutions of mixed halides. The weakness of this type of evidence as a criterion for

complex formation is exemplified by the fact that the compound written as KjCuCl4*2H20 has been shown by x-ray means to exist as copper(II)

chloride 2-hydrate admixed with potassium chloride in the crystal lattice".

Occurrence and Nature of the "Strictly" Halide Complexes

In order to facilitate an understanding of the extent of the occurrence of halo- complexes, and to illustrate the trends occurring among the families and periods of the periodic system of elements, a brief discussion of the halide complexes follows.

Family II A. In group IIA, only tetrafiuoroberyllate ion, [BeF4]=, is well characterized. Its salts bear marked resemblance to sulfates12. This is not unexpected since tetrafiuoroberyllate ion is isoelectronic and isosteric with sulfate and also approximately the same size13. Mitra14 reports that mono- hydroxytrifluoroberyllate resembles sulfate even more closely, citing such evidence as the isomorphism of the salts. The corresponding chloro- complex is much less stable, evidence for its existence being confined to freezing point behavior of beryllium chloride-alkali chloride mixtures15. Double fluorides of magnesium with alkali metal ions have been reported; however, their complexity is unlikely since the crystal structure of KMgF3 is close-packed and does not show discrete anionic complexes16.

Family II B. Complexes with all four halide ions are reported for zinc and cadmium. In the solid state, the complexes seem to vary from [ZnX3]_ and [CdX3]~ to [ZnX5]- and [CdXfi]l~. However, it seems probable that [ZnX4]= represents the maximum ratio of halide to zinc in true combina- tion (see page 1). Studies of complex halides of cadmium17, zinc18, and

11. Hendricks and Dickinson, ./. Am. Chem. Soc., 49, 2149 (1927).

12. Kruss and Moroht, Ann., 260, 161 (1890); Hay, et «/.. Z. anorg. Chem., 201, 289

(1931); 205, 257 (1932); 206, 209 (1936); 227, 32, 103 (1936); 241, 165 (1939).

13. Ray and Sarkar. ./ . I nd . Chem. Soc. 6, 987 1929); Ghosh, Mitra, and Ray: ./.

Ind. Chem. Soc., 30, 221 .1953).

14. Mitra, Science and Culture, 18, 393 (1963

15. Schmidt. Ann. ehim., [X] 11, 351 (1929); O'Daniel and Tscheischwile, Z. Krist.,

104, 124 (1942).

16. Wells, "Structural Inorganic Chemistry," p. 89, London, Oxford University

Press, 191s.

^17. Leden, Z. phys. Chem., 188, 100 (1941); Ermolevka and Makkaveeva, Zhur. Obschchei Kkim., 22, 1741 (1952 ; Markman and Tur'yan; Zhur. Obschchei Khun., 22, 1926 (1952); Btrocchi, Gazz. ehim. Hal.. 80, 231 I960

6 CHEMISTRY OF THE COORDINATION COMPOUNDS

mercury(Il lu in solution support the possibility that the most characteris- tic species arc [MX*] and [MXJ". The order of stability of the cadmium and mercury complexes is I > Br > CI (There is some doubt that fluoride1 ion form complexes with these two metals in solution).

Family III A. The halide complexes of group MA illustrate the in- version in relative stability of the [MXW]("~3)_ anions upon descending the scries. The fhioro- complexes of aluminum are by far the best characterized and most stable of all the haloaluminates. The anion [Al F6]= is remarkable in a number of ways. It represents the only 6-coordinate haloaluminate, the only class of haloaluminates which may be prepared in water20, the only haloaluminates occurring in nature, and it is apparently the monomelic parent unit of a family of condensed fluoroaluminates all of which contain hexafluoroaluminate units in their solid structures21. However, some doubt remains concerning the nature of the complex species existing in solutions of aluminum ions and fluoride ions22. Chloride and bromide form complexes, M[AlXi], with the corresponding simple aluminum(III) halides in organic solvents20 or from melts of the mixed halides23. The tetrahedral A1X4 unit also exists in the liquid and vapor states of the aluminum (III) halides, which arc dimeric24.

The halide complexes of gallium(III) are relatively rare, the best known species being the fluorides25, [GaF6p and [GaF5(H20)]=. There is little indication that the remaining halides have any great tendency to form complexes with gallium(III) ions. In contrast to this behavior, and to the behavior of aluminum, indium(III) and thallium(III) form well charac- terized complexes with chloride and bromide (and iodide in the case of

l'.i Sherrill, Z.phys.C hem., 43, 705 (1903); 47, 103 (1904); Garrett,/. Am. Chem. Soc., 61, 2744 (1939); Nayar, Srivastava, and Nyar, ./. Ind. Chem. Soc., 29, 241, 248, 250 (1952); Kazi and Desai, Current Set., (India), 22, 15 (1953); Ellendt and Cruse, Z. physik. Chem., 201, 130 (1952).

20 Malquori, Atti R., (GJ 5, 510 (1927); [61 7, 745 (1928).

21. Thilo, Naiurwiss., 26, 529 (1938); Brosset, Z anorg. Chem., 235, 139 (1937).

22. Bavchenko and Tananaev, ./. Gen. Chem., U.S.S.R., 21, 2505 (1951); cf. Chem.

Mis., 47, 5836o (1953); Tananaev and Nekhamkina, Trudy Komissii Anal- Khim.,Akad. Nauk. S.S.S.R.,3, 89 (1951); cf. Chem. Abstracts, A7,58S5e (1953)-

23. Kendall, Crittenden, and Miller, J. Am. Chem.Soc, 45, 969 (1923); Plot nikov and

Gorenbein, ./ . Gen. Chem, Rues., 5, 1108 (1935).

24. Harris, Wood, and Hitler../. Am. Chem. Nor., 73, 3151 (1961); Gerding and Smit,

/.. physik. Chem., 50B, 171 (1941); Deville and Troast, Compt. rend., 45, 821 1857); Palmer and Elliott,/. Am. Chem. Soc., 60, 1862 (1938); Smits, Meter- ing, and Kamermans, Proc. Acad. Sci., (Amsterdam), 34, 1327 (1931); Smita and Meijering, Z physik. Chem.. 41B, 98 (1938 _'."» Hannebahn and Klemm, Z anorg. Chem., 229, 341 (1936); Pugh, ./. Chem. Soc., 1937, 1046, 1969

GENERAL SURVEY 7

thallium)26 -7 2S. They apparently form no fluoro- complexes. The most typical species is [MXJ", although the enneachlorodithallate(III) ion, [TljCW has been studied extensively29.

From such observations it is commonly suggested that the more electro- positive cations; i.e., A1+++ and Ga+++, tend to form electrostatically bound complexes and, in consequence, show their greatest affinities for the most electronegative halogens. On the other hand, the relatively less electroposi- tive ions, In+++ and Tl+++, show a much greater tendency to form covalent bonds, and for that reason are most susceptible to complexation with the larger, more easily polarized halide ions.

Family IV A. Similar behavior is observed among the elements of group IVA (excluding carbon). Only the octahedral30 hexafluorosilicate exists in the case of silicon, while germanium(IV) forms the analogous [GeFc]= ion31 and the relatively unstable hexachlorogermanate32. The complexes [SnX6]= are reported for all four of the halides33. That fewer halogen complexes are formed by lead (IV) is a direct result of the strongly oxidizing nature of the ion.

Family VA. Tripositive arsenic and antimony are almost unique in their ability to exist either as the central atom in a complex species or as the do- nor atom in complexing with another metal ion (a property which is probably shared only by selenium and tellurium). The latter role will be discussed at

26. Hoard and Goldstein, ./. ('hem. Phys., 3, 645 (1935).

27. Klug and Alexander, J. Am. Chem. Soc, 70, 3064 (1948).

28. Benoit, Bull. soc. chim., France, 1949, 518.

29. Hoard and Goldstein, ./. Chem. Phys., 3, 199 (1935); Powell and Wells, ./. Chem.

Soc, 1935, 1008.

30. Ketelaar, Z. Krist., 92, 155 (1935); Hoard and Vincent, ./. .1///. Chem. Soc, 62,

3126 (1940).

31. Miiller, ./. Am. Chem. Soc., 43, 1087 (1921); Wykoff and Muller, Am. J. Sci., [5\

13, 346 (1927).

32. Laubengayer, Billings, and Xewkirk, ./. .1///. Chem. Soc, 62, 546 (1946).

33. Skrabal and Gruber, Monats., 38, 1«.) (1917); Briggs, /. anorg. Chem., 82, 441

L913); Casey and Wyckoff , Z. Krist., 89, 469 (1934); Dickinson,/. Am. Chem. Nor.. 44, 276 (1922); Ketelaar, Rietdyk, and Stoverer, Ree. txav. chin,., 56, '.hi;

1937); Goeteanu,£er.,60, 1312 (1927) ; Seubert, fler., 20, 793 (1887);Brauner, J. Chem. Soc, 65, 393 (1894).

CHEMISTRY OF THE COORDIN ATION COMPOUNDS

some length later (page 78). Species of the types [MX4]~ and [MX6]= have beenreported (for M - As, X = CI or Br84; for M = Sb or Bi, X = F or ( 1 '). Bismuth (II I) and antimony (III) also form hexahalo-anions. Recent x-ray investigations of complex antimony(III) fluorides86 have been inter- preted as showing that the pair of "s" electrons of the antimony are stereo- chemically active. Thus, K2SbF5 , which contains discrete SbF5 units, is nol strictly 5-coordinate hut is octahedral

F<^

Similarly, the ion [Sb2F7]_, in its cesium salt, is probably made up of two trigonal bipyramids sharing a fluoride ion at a common apex and with one corner of each equatorial plane occupied by an electron pair.

Sb/ F

The only halide complex of arsenic(V) is [AsF6]-37. The anions [SbX6]~ have been reported for X = F, CI, or Br. The bromide complexes differ from the1 chloro- and fluoro- species in being highly colored and readily bydrolyzed. They may be polybromides of antimony(III)37d. Bismuth(V) does nol form the fluoro- complex corresponding to that of antimony, but gives [BiOF6]- instead88.

First Transition Series. By far the most interesting halide complexes occurring among the metals of the first transition series are the fluoride

34. Petzold, /. anorg. Chem., 214, 355, 365 (1933); Dehr, ./. Am. Chem. Soc, 48, 275 L926). rutbier and Muller, /. anorg. Chem., 128, 137 (1023); Ephriam and Masimann,

54, 396 i L923 I, Bystrom and Wilhelmi, Arkiv Kemi, 3, 373, 461 (1052); Bystrom, Nature, 167,

0 I '.i51). 3chrewelius, Z. anorg. Chem., 223, 1035 (1035); Weinland and Feige, ttrr.,36,244, L903 Petzold, Z. anorg. Chem., 215, 92 (1033). 38 Ruff, '/. anorg. Chem., 57, 220 (1908 ,

GENERAL SURVEY 9

complexes. Some of these are uniquely stable toward hydrolysis while others may support unusually high oxidation states for the metal ions. The rela- tive resistance of some of the fluoro- complexes to dissociation or hydrolysis in aqueous medium, as compared to the remaining halo- complexes, is an

indication of the relative affinities of the transition ions for these donors. It is obvious that the affinity lor fluoride ion in these cases must exceed that for the oxygen donor species of the solvent water, and it is likely that the affinity for oxygen donors is greater than that for chloride or bromide, although our picture is greatly distorted in this latter case by the omni- presence of water as the solvent. The extreme difficulty with which fluoride ion is oxidized apparently makes the existence of strongly oxidizing metal fluoride complexes possible; however, it is not true that the highest known elect rovalences of a given metal invariably occur in fluoride complexes. Figure 1.1 illustrates this point by comparing oxy- complexes of the ele- ments of the first transition series with the corresponding fluoro- complexes. The general character of the fluoride complexes of these metals may be judged from the fact that most of the complexes containing higher valence states, such as heptafluorocobaltate(IV), are decomposed by water39. Some of the complexes of the more common oxidation states are much more stable.

UJ o OD

25-|

z> 24

*2 Q

X H

—I — I — I — I — I — I — I — I — I Ti v cr Mn Fe CO Nl Qj zn

Fig. 1.1. Maximum valencies of the elements of the first transition series. o = Maximum valencies found in oxy- complexes. X = Maximum valencies found in fluro- complexes.

The fluoro- complexes40 of iron (III) are noteworthy because of their im- portance in analytical chemistry. Iron (III) also forms relatively stable complexes with chloride ion as indicated by their extractability from aqueous hydrochloric acid with ether41.

Cobalt (II) forms a number of complex fluorides and chlorides42. Physico-

39. Klemm and Huss, Z. anorg. allgem. Chem., 258, 221 (1949).

40. Remy and Busch, Ber., 66, 961 (1933).

41. Dodson, Forney, and Swift, ./. Am. Chem. S<><\, 58, 2573 (1936); Lindquist, Arkiv

Kemi Min. Geol., 24A, No. 1 (1947).

42. Gmelin, "Handbuch der Anorganisen Chemie," Vol. 58A, pp. 398-461, Berlin,

Verlag Chimie G.m.b., 1932.

10 CHEMISTRY OF THE COORDINATION COMPOUNDS

chemical studies43 on solutions of cobalt(II) halides in the presence of excess halide ion indicate the existence of [CoX4]=, the stability of the com- plexes decreasing in the order Cl~ > Br- > I-. Even the chloro complex is not very stable, its formation being detectable spectrophotometrically only in hydrochloric acid which is at least 2N. A fluoro- complex of tetrapositive cobalt, K3C0F7 , has been prepared39 by fluorination of mixtures of potas- sium chloride and cobalt(II) chloride. It is fairly stable toward reduction, l>n t at lf)0° is slowly converted by hydrogen to potassium hexafluoroco- baltate(III).

The halide complexes of dipositive nickel are poorly characterized, the fluoride compounds being best known. When treated with elemental fluorine at elevated temperatures, mixtures of potassium chloride and nickel chlo- ride yield potassium hexafluoronickelate(IV)39, which is readily hydrolyzed and may be reduced to K2NiF4 .

The composition of K2MnF6 coupled with the presence of manganese (IV) in a soluble compound justifies the assumption that the substance is a true complex39- 44, 45. Manganese(III) forms fluoro- and chloro- complexes having five halogen atoms and, presumably, one water molecule attached to each manganese46.

Complex titanium halides of the form [TiXfi]=, where X = F, CI, or Br, have been characterized47- 48. Of these, the fluoro- complex is the most stable.

The halo- complexes of vanadium are best characterized for the triposi- tive oxidation state of the metal ion, higher valent vanadium tending to form oxy- and hydroxyhalo- complexes. The hexafluorovanadates(III) and pentafluoroaquovanadates have been identified49, as have complex chlorides of the type M2[VCl5(H20)]50. Tripositive chromium also forms halo- com- plexes of the type M2[CrX5H20]44 and hexanuorochromates(III)51.

43. Barvinok, Zhur.fiz. Khim., U.S.S.R., 22, 1100 (1948); Zhur. Obshchei Khim., 19,

612, 1028 (1949); Varadi, Acta Univ. Szeged., chim. et phys., 2, 175 (1949); 3, 62 (1950) .

44. Weinland and Laurenstein, Z. anorg. allgem. Chem., 20, 40 (1899); Jenssen and

Bardte, Angew. Chem., 65, 304 (1953).

45. Bode and Wendt, Z. anorg. Chem., 269, 165 (1952); Cox and Sharpe, J. Chem. Soc.

1953, 1783.

46. Weinland and Dinkelacker, Z. anorg. Chem., 60, 173 (1908).

47. Ruff and Ipsen, Ber., 36, 1777 (1903); Rumpf, Compt. rend., 202, 950 (1936);

Rosenheim and Schutte, Z. anorg. Chem., 26, 239 (1901).

48. Cox and Sharpe, ./. Chem. Soc, 1953, 1783; Wernet, Z. anorg. allgem. Chem., 272,

279 (1953). 19. Neumann, Ann., 244, 336 (1888); Werner and Gubser,£er., 34, 1579 (1901); Chris-

tensen, •/. prakt. Chem., [2] 35, 161 (1887); Schulter, Compt. rend., 152, 1107,

1261 (1911 I 60 Stahler, Ber., 37. nil (1904 , 51. Fabris. Oazz. chim. ital., 20, 582 (1890); Helmolt, Z. anorg. Chem., 3, 125 (1898).

GENERAL SURVEY 1 1

Scandium forms several complex halides, among which arc the fluoro- complexes [ScF4]~, [ScFJ", and [ScF8]". There is Borne evidence thai fche remaining elements of periodic family 1 1 IB also form fluoro- complexes, although these arc noi so well characterized as those of the other transition elements81. The complexity of KI.aF, is unlikely since the crystal structure indicates the presence of no finite [LaFJ groups68.

Although COpper(I) complexes are known''1 with chloride, bromide, and iodide ions, no fluoride complexes appear to exist. A great variety of com- plex halides has been reported for eopper(II). The complexity of some of the double salts formed by copper(II) chloride and copper(II) bromide with alkali halides is in doubt since x-ray data show that K2CuCl4-2H20 and (NH^sCuBfi^HsO exist as lattice compounds of the simple salts. However, physical evidence indicates that [CuCl§]~ and [CuCl4]= do exist55. The latter is reported to be a distorted tetrahedron56. Copper(III) has been reported in K3CuF639.

The relatively greater tendency of the metallic ions of the first transition series to form complex ions with fluoride and chloride rather than with bromide and iodide and the general tendency of the complexes to dissociate or hydrolyze in solution appears to justify the supposition that the binding force involved is essentially electrostatic. This suggestion is supported by the considerable stability of hexafluoroferrate(III) and hexafluorotitan- ate(IV) which involve electronic states normally associated with unusually stable gaseous ions (Chapter 3).

Second and Third Transition Series and Family IB. In contrast to the elements of subgroups IIIA, IVA, VA, and VIA, the elements of the three transition series show a marked increase in the importance of their higher oxidation states as the atomic weight of the metal increases. This is related to the types of compounds formed by each element, since high oxidation states ions usually exist in covalent compounds. The halide complexes of the platinum metals include some of the most widely known complex ions. This is doubtless a consequence of the fact that their simple compounds are for the most part "simple" in name only (for example, platinum(II) chlo- ride is not salt-like but exists as bridged, covalent, giant molecules).

Complexes of the type [PtX6]= have been characterized for all four of the

52. Dergunov, Doklady Akad. Navk, S.S.S.R., 85, 1025 (1952); cf. Chem. Abs. 47,

1524b (1953).

53. Ref. 16, p. 290.

54. Szabo and Szabo, Z. physik. Chem., 166, 288 (1933); Fontana, Gorin, Kidder, and

Meredith, Ind. Eng. Chem., 44, 363 (1952); Harris, ./. Proc. Roy. Soc., N.S. Wales, 85, 138 (1952).

55. Rossi and Strocchi, Gazz. chim. Hal., 78, 725 (1948). (see Ref. 72c)

56. Helmholz and Kruh. ./. .1///. Chem. Soc., 74, 1176 (1952).

L2 CHEMISTRY OF THE COORDINATION COMPOUNDS

common halides, the chloride and bromide being the easiest to prepare57. The iodo complex tends to liberate iodine with the reduction of the plati- num to the dipositive slate58, while salts of hexafluoroplatinate(IV) readily hydrolyze. The complex fluorides have been prepared by heating the addi- tion product of the ehloroplatinate and bromine trifluoride59. They are diamagnetic, indicating drsp* hybridization and covalent bonding (despite the high electronegativity of the fluorine). The Pt — F bond distance is greater t han that expected for a covalent link, which indicates that the bond hasa considerable degree of ionic character60. Mixed halo- complexes, such as IPtrhBr^, have been prepared61, as well as the series of hydroxychloro- anions [PtCl„(OH)6-nl= (page 4). The planar tetrahalide complexes of platinum(II) have been prepared with chloride, bromide, and iodide. Salts of these anions are generally obtained by reduction of the corresponding hexahaloplatinate(IV) salts with sulfur dioxide62, potassium oxalate6213 ■ 62c' 6i, potassium hydrogen sulfite64, hydrogen sulfide65, potassium hypophos- phite66, or hydrazine salts67. Grinberg68 has suggested that reduction by hydrazine salts proceeds in two steps:

K2PtClfi + N.>H4-2HC1 -> Pt° + N2 + 2KC1 + 6HC1

2KC1 + K>PtCl6 + Pt° -> 2K,PtCl4

In support of this argument, Grinberg has sho\ui that hexachloroplatinate ion is reduced to tetrachloroplatinate(II) by platinum black which has been freshly prepared by the reduction of hexachloroplatinate (IV) with hydra- zine sulfate. Exchange experiments have shown that halide ions of plati- num(II) complexes are labile, the bromide of [PtBr4]= being subject to complete exchange; however, the central platinum atom does not undergo

57. Weber, J. Am. Chem. Soc.,30,29 (1908); Rudnick and Cooke, J. Am. Chem. Soc.,

39, 633 (1917); Bielmann and Arduson, Ber., 36, 1365 (1903); Gutbier ami Bauriedel, Ber., 41, 4243 (1908).

58. Datta, ./. Chem. Soc, 103, 426 (1913).

59. Sharpe, ./. Chem. Soc, 1950, 3444; 1953, 197; Schlesinger and Tapley, ./. Am.

Chem. Soc, 46, 276 i L924 i.

60. Mellor, Report of the Brisbane meeting of the Australian and New Zealand

IlBSoc. for the Advancement of Science, Vol. XXVIII, 131, May 1951.

61. Klement, /. anorg. Chem., 164, I!).") d<)27).

62. Claua, Ann., 107, 137 (1868); Klason, Ber., 37, 1360 (1904); Vezea,Bull. soc. chim.,

|3] 19, 879 (1898). 63 Mikhelis, Zhur. priklad. Khim., 26, 221 (1953); cf. Chem. Abs., 47, 11060i

195 64. Lea, .1///. ./. S«\, [3] 48, 398, loo (1894). 65 Bottger, J. prakt. Chem., |1] 91, 251 (1863) 1///. ./. Set., [3] 48. :VM (1894 ■ 1'degershel and Shagesultanova, Zhur. priklad. Khim., 26, 222 (1953); Cooley and Busch, unpublished experiments (1954). I irinberg, //////•. priklad. Khim., 26, 224 (1953).

GENERAL SURVEY L3

exchange*. The rates of exchange vary in the order CN > I > Br > ('1 . It is, at first thought, paradoxical thai the complexes having the greater thermodynamic stabilities exchange most rapidly (AF, ,„,,,. : [PtClJ , —21.8; [PtBrJ™, —24.5). This ease of "self -displacement" may be a peculiarity of planar complexes since ferrocyanide ion docs not exchange with cyanide ion in water70. The diammine Pt (^N II ; )vHr: which was once thought to con- tain tripositive, 5-coordinate platinum has been shown rather to exist as a molecular compound of lIV^XII^Br,! and (Pt lv(\H:;>,Br,]71.

In contrast to platinum, the tet rapositive oxidation state of palladium is rather unstable. The hexachloro- and hexabromopalladate(IV) anion- may he prepared78 in much the same way as are the platinum complexes; how- ever, their solutions are unstable toward evolution of the4 halogen and they both react with aqueous ammonia to liberate nitrogen. The hexafluoro- palladate(IV) has recently been prepared by Sharpe73. Its salts are yellow; they darken rapidly in air and are immediately hydrolyzed in cold water. Salts of the planar tetrahalopalladate(II), [PdX4]=,are known71 for X = CI, Br. and I.

The great affinity of palladium(II) for halide ions may be seen from the dissociation constant75 of [PdCl4]= (Kd = b X 10-14). The supposed pal- ladiumflll) complex, MjPd111^76 probably contains both palladium(II) and palladium (IV),

The tendency for higher oxidation states to become more stable with increasing atomic weight of the metal is illustrated by cobalt, rhodium, and iridium. The only strictly halogen complexes in which cobalt has an oxida- tion number greater than two are the fluoro- complexes. Dipositive rho- dium, on the other hand, forms no complexes. Rhodium is tripositive in all of its halogen complexes except the recently reported rhodium(IV) fluoro-

69. Grinberg and Filinov, Compt. rend. acad. sci., U.R.8.S., 23, 912 (1939); cf. Chem. Abe. 34, 12462 (1940); 31, 453 (1941); cf. Chem. Abs., 37, 5719 (1943) ; Grinberg, Bull. acad. 8ci.,U.RS.S.,Ser.phy8., 349 1940); cf. Chem. Abs. 35, 3895» (1941 .

7(). Grinberg and Nikol'skaya, Zhur. priklad. Khim., 24, 893 (1951); cf. Chem. Aba. 47,4709a 19.53).

71. Cohen and Davidson. ./ . .1///. Chem. Sue. 73, 1965 1951 : Brossett, Arkiv Kemi

Mir,. Geol., 25A, No. 19 1948).

72. Puche, ( •/.. 200, 1206 (1935); 208, 656 1939 ; Rosenheim and Maas, /

l hem., 18, 331 (1898); Gutbier and Krell, Ber., 38, 2385 L905 , Sharpe, ./. ( hem. Soc, 1953, L97. 74. Gutbier, Ber., 38, 2107 1905); Gutbier and Krell, Ber., 38, 3969 L90S ; Gutbier, Krell and Janssen, / anorg. Chem., 47, 23, 1292 (1906); Gutbier and Woernle, 47,271ti L906 ; Gutbier and Fe\\neT,Z. anorg. Chem., 96, 129 1916 ; Dickinson, J. Am. Chem. Soc., 44,2404 L922 ; Cox and Preston, J Chem Soc, 1988, 1089; Theilacker, / anorg. Chem., 234, 161 l" 75 Templeton, Watt, and Garner, ./. Am. Chem. Soc., 65, 1608 L943 . Wohler and Martin, / anorg. Chi m., 57, 398 L908

14 CHEMISTRY OF THE COORDINATION COMPOUNDS

complexes77. Three formulations are reported for the halorhodiates(III), M2KhX5 , M3RhX6 , and M2Rh2X9 . All three types are known for bromide and chloride78, but the only fluoro- complex is the ion [RhF6]~. The struc- tures of most of these compounds are still open to question.

Both tripositive and tetrapositive iridium form complexes with chloride and bromide. The iridium(III) complexes are of the types [IrX6]- and [lrX5(ll20)]=79, whereas iridium (IV) is found in the anion [IrXe]=, (X = Br, CI, or F). The hexabromo compound is unstable toward evolution of bromine80.

Ruthenium(III) and ruthenium (IV) form a variety of complex halides and aquohalo- or hydroxohalo- complexes. Ruthenium trichloride appar- ently exists in several hydrated forms, analogous to the hydrate isomers of chromium (III) (see Chapter 7)81. Some of the probable "hydrate isomers" are Ru(H20)Cl3 , which contains no ionizable chloride, and the reported cis and trans forms of [RuCl2(H20)2]Cl. Dwyer and Backhouse81 suggest that the ruthenium is 6-coordinate in all of these complexes. As compared to the similar platinum compounds, halide complexes of ruthenium show a marked tendency to hydrolyze and to retain water in their coordination spheres. As with the platinum analogues, [RuBr6]= is less easily hydrolyzed than [RuCl6]=81. Ruthenium(III) forms two types of anionic chloro- com- plexes [RuCl6]= and [RuCl5(H20)]=, while ruthenium(IV) forms the com- plexes formulated as [RuCl6]= and [RuCl5(OH)]-82- 83- 84. It has been shown that [RuCl5(OH)]= is actually dimeric in the crystalline state, having the structure [Cl5Ru — 0 — RuCl5]4~ (see p. 167). Fluorination of hexachloro- ruthenate(IV) yields a white crystalline compound of the composition K2RuF8 , which hydrolyzes readily and darkens on standing85. It is possible that ruthenium (VI) is present, and that it is octacoordinate.

77. WeiseandKlemm,Z. anorg. allgem. Chon., 272, 211 (1953); Sharpe,/. Chem. Soc,

1950, 3444.

78. Delepine, Bull. soc. chim., Belg., 36, 108 (1927); Gut bier and Bertsch, Z. anorg.

Chem., 129, 67 (1923); Meyer and Hoehne, Z. anorg. Chem., 231, 372 (1937);

Meyer, Kawkzyk, and Hoehne, 232, 410; Poulenc, Compt. rend., 190, 639 (1930) ;

Ann. chim., [Xi] 4, 567 (1935). 79 Delepine, Bull. soc. chim., [4] 3, 901 (1908);Delepine-Tard, Ann. chim. phys., [10]

4, 2S2 (1935 B0. Delepine, .1////. chim. phys., [9] 7, 277 (1917); Schlesinger and Topley, J. Am.

Chem. Soc, 46, 276 (1924); Dobroborskaya, Zhur. priklad. Khim., 26, 223

(1953); cf. Chem. Abe., 47, U061g (1953). 81. Dwyer and Backhouse, J Proe. Roy. Soc, X.S. Wales, 83, 138 (1949). $2. Gutbierand Niemann, Z. anorg. ('hem ., 141, 312 (4924) ; Howe, ./. Am. Chem. Soc,

49, 2389 (1927); Charonnat, .1////. chim., [10] 16, 72 (1931); Compt. rend., 181, ; L925 Howe, ./. .1///. Chem. Soc., 26, 942 L904 84. Charonnat, Compt. rend., 180, 1271 (1925). 85 \vnsley: Peacock, and Robinson, Chem. hid., 1952, 1002.

GENERAL SURYIA L5

The hexahalo- salts M I >-.Y and MiOsXa are reported where X = CI or Br in the first case86 and for X = F, CI, Br, or I, in the latter"7. Recrystal- lization of the hexachloro- and hexabromoosmiate(IV) salts from dilute halogen acid leads to hydrolysis. Mixed halogen complexes, such as

[OsClsBr]" and [OsCl;J3r3]=, and hydroxohalo- complexes, such as [OsX.s(( )H )] '", are also reported88. Osmium also forms halo- complexes in its higher oxidation states. Osmium(YI) exists in the tet rahaloosmyl com- plexes [( )s( )-j\;!~ Vl, and the oxydihaloosmyl complexes (( )s( ):iX2]=90. X-ray data show that the salts Mo[Os02X4] are similar in crystal structure to po- tassium hexachloroplatinate(IV)91. Fluoride ion combines with osmium (Mil) fluoride to produce a white solid that may be a 9- or 10-coordinate complex9'2; the material has not been analyzed. Dissolution of osmium(VIII) oxide in fluoride solution leads to the formation of unstable compounds which presumably contain complex anions, such as [Os04F2]= 93.

The halo- complexes of rhenium are intermediate in character between those of the platinum metals and those of the remaining transition ele- ments. Thus, rhenium (IV) forms complexes of the type [ReX6]= with fluo- ride (like the IVB, VB, and VIB metals) and also with the other halogens, even iodide (a behavior more to be expected of the platinum metals)94. An interesting similarity is found between some rhenium (IV) and rhenium(V) chloro- complexes and those of ruthenium(III) and ruthen- ium (IV). In addition to hexachlororhenate(IV), the pentachlororhenium complexes [ReIVCl5(OH)]= [RevCl50]= and [ReIV2Cli0O]4- also exist94f.

Molybdenum and tungsten form complex halides or oxyhalides in their di-, tri-, penta-, and hexavalent states. Tripositive molybdenum forms fluoro- and chloro- complexes of the types [MoX5(H20)]= and [MoX6]s95.

86. Claus and Jacob}', J . prakt. Chem., 90, 78 (1863) ; Crowell, Brenton, and Evenson,

J. Am. Chem. Soc., 60, 1105 (1938).

87. Ruff and Tscherch, Ber., 46, 932 (1913); Dwyer and Gibson, Nature, 165, 1012

(1950;; Wintrebert, Ann. ekim. phys., [7] 28, 133 (1903).

88. Krauss and Wilkin, Z. anorg. Chem., 137, 360 (1924).

89. Wintrebert, Ann. chim. phys., [7] 28, 54, 86 (1903).

90. Wintrebert, Ann. chim. phys., [7] 28, 114 (1901).

91. Hoard and Grenko, Z. Krist., 87, 100 (1934).

92. Ruff and Tscherch, Ber., 46, 929 (1913).

93. Tschugaev, Compt. rend.. 167, 162 (1918); Krauss and Wilkin, Z. anorg. Chem.,

145, 151 (1925). (.)4. Ruff and Kwasnik,Z. anorg. CAem.,219,76 (1934); Schniid. Z. anorg. CAem.,212,

187 (1933);H6lemann,Z. anorg. Chem., 211, 195 (1933); Nod. lack and Nod. lack.

Z. anorg. Chem. ,216, 120 (1933) ; Briscoe, Roderson, and Etudge, J. Chei

1931, 3218; Jezowska-Trzebiatowska, Trav. soc sd. et lettres Wroclaw, Ber. B,

39, 5 (1953). 95. Rosenheim and Braun, Z. anorg. ('htm., 46, ^>2<) (1905); Foerster and Fricke, Z.

angew. Chem., 36, 458 (1923).

L6 CHEMISTRY OF THE COORDINATION COMPOUNDS

Eowever, only the dimeric anion [W2C19]- is known for tungsten(III)96 (for structure', see page 7). It seems likely that a tungsten-tungsten bond is pi (Hi it in this anion since the substance is diamagnetic97. The most stable oxyhalo- complexes of molybdenum and tungsten in their penta- and hexa- positive states are fluoro- complexes, such as [MoVI02F4]=, [WVI02F4]=, and [MovOFb]-, all of which are isomorphous with [NbvOF5]=. The affinity of fluoride ion for hexavalent molybdenum and tungsten may be illustrated by the fact that most of the precipitation and color reactions of molybdate and tungstate ions are masked by the presence of fluoride ion98.

An interesting feature of the halogen complexes of niobium and tantalum is the occurrence of high coordination numbers (see Chapter 10). This is undoubtedly associated with the fact that the only significant strictly halo- geD complexes of these metals are those of the fluoride ion. Both of these elements form heptafluoro- anions of the type [MVF?]=. Their structures are discussed on page 393. In addition, tantalum (V) forms an 8-coordinate fluoro- complex [TaF8]= which exists in the form of a tetragonal antiprism". Six-coordinate hexafluoroniobate(V) is also known, as is its tantalum ana- log100. The heptafluorotantalate(V) is somewhat more stable than the nio- bium^) compound which hydrolyzes to [NbOF5]=, and this difference has served in helping to separate the two metals. Oxyhalo- complexes are formed by both metals, the oxyfluorides being the most stable.

The same trends are observable among the halogen complexes of zir- conium and hafnium, the outstanding characteristics being variable coor- dination number and decreasing stability of the complexes with increasing atomic weight of the halide. The latter point is illustrated by the fact that zirconium dioxide is dissolved by hydrofluoric acid and that only the fluoro- complexes are stable in aqueous media101. The chloro- and bromo- complexes are prepared in alcohol102. The complexes are of the types [MX5]~, [MX5(H20)]=, [MX6]=, [MXJS (see Chapter 10). The structure of the sup- posed 5-coordinate species is still open to question103. The fluoro- complexes are used in the separation of hafnium and zirconium104.

96. Olsson, Ber., 46, 566 (1913); Olsson, Collenberg, and Sandved, Z. anorg. chem.,

130, 16 (1923).

97. Brossett, Nature, 135, 824 (1935); Pauling, Chem. Eng. News, 1947, 2970. its. Feigl, .1/.//. Chem. Aria, 2, 397 (1948).

99 Hoard, ./. -1///. Chem. Sac, 61, 1252 (1939); 64, 633 (1942); dc Marigroc, Compt.

rend., 63, 85 (1866); Board, Paper presented at 6th annual symposium, Div.

Phys. and [norg. Chem., Columbus, Ohio, Dec., 1941. KM). Halm and Putter, '/. . anorg. Chem., 120, 71 (1922). nil. Connick and McVey, •/. .1/,/. Chem. Soc, 71, 3182 (1949).

Schwarz and Giese, /. anorg. Chem., 176,209 (1928); Rosenheim and Frank, Ber.,

38, 812 L905 , Haendler and Robinson,/. .1///. Chem. Soc. ,75, 3846 (1953); Haendler, Wheeler,

and Robinson, ./. Am. Chem. Soc, 74, 2352 (1952). lot. Larsen, Fernelius, and Quill, //,</. Eng. Chem., Anal. Ed., 15, 512 (1943); Schultz

GENERAL SURVE1 17

The solubilities of the silver halides increase sharply as the concentration of excess halide ion is increased106. The study of this solubility dependence indicates the format ion of a scries of complexes ranging from [AgjX]H ! to [AgXJ , and possibly [Ag»Xe]4 m. Theorderof stability of both silver and gold halide complexes is I > Br > CI (as is also commonly observed among the platinum metals). The silver complexes best known in the solid .state are of the types [AgXJ and [AgXg] ln7. I'nipositive gold normally forms 2-coordinate, linear complexes of the type [A11CI2] "'s, while gold(III) forms L-coordinate, planar complexes of the type |AnX,| '"''. Gold forms many bridged halogen compounds (page 19). The substance having the em- pirical formula CsAuCla should be formulated as Cs2AuIAuIIICl6 , contain- ing equivalent amounts of gold(I) and gold(III) (see Chapter 9).

Complexes Containing Halide Groups as a Less Abundant Donor Species

Many metals, especially those of the platinum group, form halo- com- plexes containing three1, four, or five halide groups; however, with the ex- ception of the hexafluorocobaltate(III), cobalt(III) complexes are not known with more than three halide groups. Indeed, the mixed complexes which have been most significant in the development of the coordination theory are those which contain one, two, or three coordinated halides and five, four or three neutral groups. Chloropentamminecobalt(III) chloride, [Co(XH3)5Cl]Cl2 , is one of the longest known cobalt (III) ammines and is the chief product obtained by atmospheric oxidation of solutions containing cobalt (II) chloride, ammonium chloride, and ammonium hydroxide. The coordinated chloride is only slowly removed by the action of silver nitrate, even when heated. The salt serves, however, as a starting material for the preparation of many other cobalt (III) ammines, not only by replacement of the chloride, but also by replacement of one of the ammonia molecules. Heating with ammonium carbonate, for example, gives carbonatotetram- minecobalt(III) chloride, [Co^Hs^CCyCl. It has also been utilized in

and Larsen, J. Am. Chem. Soc, 72, 3610 (1950); Huffman and Lilly, •/ t m. Soc, 73, 2902 (1951).

105. Eber and Schuhly, J. prakt. Chem., 158, 176 (1941); Z. anorg. allgem. ('Ik ///., 248,

32 (1941).

106. Bern and Leden, Svensk. Kern. Tidskr., 65, 88 (1953); Z. Naturforsch., 89, 719

(1953); Chateau and Pounadiev, Science et indus. phot., 23, 225 (1952); Y..t simirskii, Doklady Akad. Nauk., S.S.S.R., 77, 819 (1951); cf. Chem. Abs., 45, 7102 (1951).

107. Forbes and Cole, /. Am. Chem. Soc., 48,2492 L921 \; Harris and Schafer,/. P

8oe., X.s. Wales, 85, 148 (1952); Harris, •/. Proc. Roy. Soc., N.8. Wales 85, 142 (1952).

108. Lengfield, Am. Chem. J., 26, 324 L901).

109. Cox and Webster, J. Chem. Soc, 1936, 1635.

18 CHEMISTRY OF THE COORDINATION COMPOUNDS

studies directed a1 elucidation of the mechanism of substitution reactions of 6-coordinate complexes110.

The two forms of dichlorobis(ethylenediamine)cobalt(III) chloride, [Co ei^CyCl, are used in the preparation of other ethylenediamine cobalt Baits. .Both the cis and trans forms of this complex are readily prepared, and are stable in water solution for some time, though the change in color of the solution indicates aquation; the w's-dibromobis(ethylenediamine)co- balt(III) ion rearranges with extreme ease to the trans form, and both iso- mers aquate rapidly; the corresponding iodo compounds are not known.

Chloropentamminechromium(III) chloride, [Cr(NH3)5Cl]Cl2 , is ob- tained, together with the hexammine, by the action of liquid ammonia on anhydrous chromium(III) chloride. Once formed, the pentammine is con- verted to the hexammine with extreme slowness, which may be due, how- ever, to the very slight solubility of the pentammine in liquid ammonia.

ns-Diehlorobis (ethylenediamine) chromium (III) chloride is most easily obtained by the thermal decomposition of tris (ethylenediamine) chrom- ium (III) chloride. The reverse reaction takes place very slowly when the dichloro- salt is suspended in ethylenediamine. Complexes of very similar type are also encountered in the chemistries of the platinum metals.

In general, the complexes containing halo- groups as less abundant donor species may be grouped according to the same classification as that given for the strictly halide complexes; i.e., those which show little tendency to dissociate in solution (penetration complexes), and those which change upon dissolution in a polar solvent as a result of displacement by solvent molecules (normal complexes). Only the first class of compounds is of great significance here since the more labile species cannot experience a change in the state of aggregation without extensive change in their natures. Thus, [Fe(XH3)2Cl2] cannot be dissolved in water and subsequently recovered, while many strictly halide complexes may dissociate in solution but still be recoverable in the original form upon removal of the solvent.

( omplexes Involving Halogen Bridges

The halide ions sometimes donate pairs of electrons to two metallic ions simultaneously, forming a "bridge." Aluminum chloride (page 6) and rhenium(III) chloride111 have been shown to have the structures

CI CI CI

\ / \ /

M M

/ \ / \ CI CI CI

1 lii Br0nsted, '/.. phyaik. Chem., 102, 169 (1922); Garrick, Trans. Faraday Soc, 33, L937); Lamb and Fairball, ./. Am. Chctn. Soc, 45, 378 (1923); Lamb and Maiden,./. .1///. Chem. Soc. ,33, 1873 (1911) ; Adell, Z. anorg. allgem. Chem., 249, 251 (1942).

GENER I/. SURVE]

and other volatile metal halides are probably similar. The dimeric tertiary phosphine and arsine compounds also contain double halide bridges (page 81 ) and a number of olefine complexes and thio ether complexes have hern formulated in the same way (see page 83). Alky] derivatives <>f gold bro- mide arc dimeric and probably have the structure112

H Br R

The presence of double bridges in platinum(II) chloride results in the forma- tion of an infinite chain of PtCl4 groups.

In addition to double halogen bridges, triple or single bridges may be formed. The triple bridge is illustrated by ions of the type [Mni2X9]= (see page 7), while single halogen bridges occur in such species as [A1F5]= (page 389). The compounds

Co(XH,)4

S04

and

Co(NH3)3(H20)

S04

may also exemplify single bridges. When silver ion is added to a solution of the dichlorotetramminecobalt(III) ion, silver chloride does not precipi- tate at once, but the silver ions lose their ionic property through coordina- tion with the chloride of the cobalt (III) complex. The ion so formed is not stable, however, and slowly precipitates silver chloride.

The phenomenon of "interaction absorption" is often observed in bridged halogen complexes. When the halides (cyanides, or oxides) of a metal in two different oxidation states are associated in a single molecule or ion (or possibly in' such relatively less intimate admixture as crystal compounds or solutions — the point is not clear), a high degree of color is developed. Thus, CuCl,CuCl; SbClgSbCl* ; Sn('l,-Sn('l, ; ( •>,Au,Au,,< 'I, ; and [Pd(XH3)2Br2]-lPd(\Il;;)-jBr4] are all highly colored60. In none of these casee

111. Wriggee and Biltz, Z. anorg. allgem. Chem., 228, :>7J r. ».;•■, 1 1J . Gibson and Simonsen, J. Chem. Soc., 1930, 2531 ; Buroway, ei al., .1 { ' h< m 1937, 1090.

113. Werner, Z. anorg. Chem., 14, 31 (1897).

114. Werner, Z. anorg. Chem., 15, 155 (1897).

Soc,

20 CHEMISTRY OF THE COORDINATION COMPOUNDS

has conclusive evidence for an intermediate oxidation state of the metal been obtained; indeed, .strong evidence indicates the nonexistence of such states. In the first example the ridiculous assumption of the ion Cu1,5+ would be necessary, while in the case of the diammino palladium compound, x-ray data and magnetic behavior definitely preclude the existence of the intermediate state. Nonetheless, a resonance between the two oxidation states produces high color and probably renders the two metal atoms indis- tinguishable. The probablity that a halogen (or similar) bridge is necessary for this phenomenon is supported by the fact that rapid electron exchange occurs between the coordinately saturated complexes, [OsIIdipy3]++ and [Osmdipy3]+++, without the development of high color115.

The Donor Properties of Oxygen Hydrate Formation

All metallic ions apparently form hydrates in aqueous solution, frequently surrounding themselves with large numbers of molecules of water. Part of this water is held by van der Waals forces only, but it is difficult to escape the conclusion that in every case a few molecules at least are coordinated to the metallic ion. In many cases, of course, the hydrates can be crystal- lized from the solution.* These usually retain only enough molecules of water t o satisfy the coordination number of the metallic ion, but sometimes, as with the alums, stable hydrates contain more than this amount. To account for these we may assume that (a) the excess water is not chem- ically combined, but is held in place by the demands of the lattice structure, (b) the coordination number of the metal is abnormal, (c) second and even third coordination spheres are formed (d) the molecules of water are poly- meric or (e) part of the water is combined with anion. It is often assumed that water of hydration which is not lost at 100°C must be chemically com- bined, but this does not necessarily follow, for lattice compounds sometimes show considerable stability. On the other hand, chemically combined water may escape from salts at low temperatures — even at room temperature — if the anion is one which readily coordinates with the cation, thus displacing the water from the coordination sphere.

Werner recognized that water molecules are sometimes held by feeble, Qonchemical forces in writing formulas such as [Co(NH3)5Cl]Cl2-H20. The water may be removed from this compound without changing its properties except for disruption of the crystal lattice, while dehydration of the isomeric [Co(N 1 1;; >. I M )|( '!:, is accompanied by change in color and solubility, and by Loss of ionic function of one chloride1 ion.

LIS. Dwyer, Mellor, and Gyarfas, Nature,Mt 176 (1950).

Man) anions also have the power of combining with water — this union takes place through hydrogen bonding.

GENERAL SURVEY 21

In his early papers, Werner111 also gave expression to the though.1 thai Beveral coordination spheres can form around a positive ion. I [e argued that when water molecules form a coordination sphere around a positive ion, a negative charge is induced on the inner surface of the sphere, 80 that the

outer surface hears a positive charge, just as the metal ion itself does. This enables it to attract another sphere of water molecules, which will likewise hear an induced charge. The process may he repeated several time-.

Closely related to this hypothesis was the thought that water exists in hydrates in the polymeric form. In view of the fact that water as such ifl iated, this is not an unreasonable assumption, though Werner had little experimental evidence on which to support it. The fact that many salts contain exactly twice as many water molecules as can he explained by the coordination theory made it an easy assumption. Such an explanation seems naive, hut the fact that "multiple coordination spheres" do exist in solution cannot be denied. Their existence has been demonstrated by the diffusion studies of Brintzinger (Chapter 18) and by the polarographic work of Laitinen and his co-workers117.

As is to be expected, the ease with which metallic ions form hydrates in- creases with increasing charge and with decreasing radius. The ions of the alkali metals except lithium and sodium are seldom hydrated in the solid state, and the hydrates of these two are unstable; divalent ions of the lighter metals are usually hydrated (unless they exist in highly insoluble com- pounds) and trivalent ions, nearly always so. In any periodic group the stability of the hydrates is greatest for the smallest ions, while the number of water molecules normally held is greatest for the large ions. Even in complexes in which water molecules undoubtedly occupy positions in a true coordination sphere, the nature of the oxygen-metal bond varies a great deal. Hunt and Taubells showed that the water in the hydrated forms of Al , Ga and Th4+ exchange with the solvent water in about three minutes, so the metal-oxygen bond must have a considerable degree of ionic character. The hydrated chromium(III) ion, on the other hand, ex- changes very slowly, the halftime being about forty hours. They made the observation, also, that all of the cations studied show a greater affinity for H2018 than for H2016. The hydrated cobalt (III) ion exchanges rapidly. This is probably not due to a lack of covalenl bonding, but to a rapid electron exchange between the hydrated cobalt(III) and cobalt (II) ions, and a rapid exchange between the latter and the solvent water119.

116. Werner, Z. anorg. Chem., 3, 267 1S93).

117. Laitinen, Bailar, Holtaclaw, and Quagliano, ./. .1///. Qk m. Soc , 70, 2999 1948 ;

Laitinen, Frank and Kivalo, ./. .1///. Chem. 8oe., 75, 2866 1"~

118. Bunt and Taube, /. Chem. Phyt . 18, 757 1950 ; 19, 602 1951 .

119. Friedman, Taube, and Hunt, ./. Chem. Phys., 18, 759 I960 ■

CHEMISTRY OF THE COORDINATION COMPOUNDS

Hydroxy 1 Coordination

The hydroxide ion has a strong coordinating tendency, partly because it has i hive pairs of unshared electrons, but chiefly because of its negative charge. The hydrates of highly charged metallic ions readily lose protons with the formation of hydroxo complexes:

[A1(H20)6]+++ -» H+ + [Al(H,0)5OH]++ -> H+ + [Al(H20)4(OH)2]+, etc.

The aquo ammine complexes undergo the same type of reaction: [Co(NH8)6H20]+++ ^± [Co(NH3)5OH]++ + H+. The phenomenon underlies our present theories of acidity, hydrolysis and amphoterism, and is discussed in Chapter 12.

The hydroxide group can act as a bridging group between two metallic ions, under which conditions it is almost entirely devoid of basic properties. This bridge forming ability may extend to great lengths and an interesting theory of colloidal oxides has been based upon it (Chapter 13).

Werner's postulate that basic salts are polynuclear complexes held to- gether by hydroxy 1 groups120 has been shown, by x-ray studies, to be un- tenable in most cases. The basic chlorates and perchlorates of lead have not been studied by x-ray analysis, but the conductivities and other properties of their solutions indicate that they have the structures

una

•( M" bridges are common in the polynuclear cobalt complexes. The chief constituenl of Vbrtmann's sulfate, which is obtained by oxidation of an

120. Werner, Ber., 40, I HI (1907).

121. Weinland and Stroh, Ber.} 55, 2210, 2706 (1922).

122 Weinland and Paul, Z. anorg. Chem., 129, 243 (1923).

(ihWhlx'AL SI 7,'lA'l

ammoniacaJ solution of a cobalt salt, is

/ \ (NH,)4Co Co(NH,)<

\ / OH

(S04)2

Such ions as

oil

/ \ (XH3)4Co Co(NH3)4

\ /

/ \

(NH,)8Co— OH— Co(NH,),

\ / OH

and

;co(nh3).

have been known for many years. The hexol salt is of special interest, as it was the first strictly inorganic compound to be resolved into optical anti- podes1'24. Adamson, Ogata, Grossman, and Newbury125 have come to the conclusion that Durrant's salt has the bimolecular structure

Alcohol* and Kthers

/ \

(C204)2Co Co(C204)2

\ / OH

The organic derivatives of water, the alcohols and ethers, show much less tendency to form coordination compounds than docs water; nevertheless, a

123. Werner, Ber . 40, 4609 (1907).

124. Werner, Ber., 47, 3087 L914

l_'.V Adamson, Ogata, Grossman, and Newbury, 0 \ K Contract 23809, Technical Report, M.uch L954.

24 CHEMISTRY OF THE COORDINATION COMPOUNDS

large number of such compounds is known. The compounds of the alcohols are more stable than those of the ethers, the stability in each series de- creasing as the size of the organic group increases. Because of the chelation effect, the polyhydric alcohols form somewhat more stable compounds than do the monohydric alcohols. Glycol is able to displace water from hydrates of heavymetals, each alcoholic hydroxyl group taking the place of one mole- cule of water in the coordination sphere126. Glycerol ordinarily behaves as a bidentate donor, also, adjacent hydroxyl groups coordinating. The third hydroxyl group is prevented from combination by steric factors. The di- valent ions of the alkaline earths127, and of cobalt, nickel, copper, and zinc, all form compounds in this way, those of the heavy metals being rather unstable. Other poly hydroxy alcohols and even the sugars form coordina- tion compounds, the tendency to combine with the ions of the alkaline earths being particularly noticeable. The purification of sugar through the precipi- tation of calcium and strontium "saccharates" is of interest in this connec- tion. The structure of these compounds has not been studied in detail, but they are evidently coordination compounds rather than salts.

In the presence of polyhydric alcohols such as mannitol and sorbitol, sodium hydroxide does not precipitate iron (III) ion128. Addition of barium chloride to such basic solutions gives pale yellow, crystalline products con- taining the alcohol, iron, and barium in a 1:1:1 ratio. Traube and Kuhbier write the formula of this product as

CH,— CH— CH— CH— CH CH2

I I I I I I

O O O OH O O

/ X /

7 Ba

but they cite no evidence to support such a formulation. Scale models indi- cate that it is improbable that three consecutive hydroxyl groups are co- ordinated to the iron. According to Traube and Kuhbier, treatment of this product with sodium sulfate gives Na[FeC6Hio06]-3H20, in which there must be two uncoordinated hydroxyl groups. Several similar compounds containing sugars or polyhydroxy acids and a variety of metal ions have been prepared and analyzed, but their structures have not been deter- mined129. Some of these oxidize in the air to formic acid, carbon dioxide, and similar compounds180.

126. Cum and Bockisch, Ber., 41, 3465 (1908); Griin and Boedecker, Ber., 43, 1051

(1910).

127. Grttu and Husmann, Ber., 43, 1291 (1910). Us Traube and Kuhbier, Ber., 65, 187 (1932).

129. Traube and Kuhbier, Ber., 66, 1545 (1933); 69, 2655 (1936). 13G. Traube and Kuhbier, Ber., 65, 190 (1932); 69, 2664 (1936).

GENERAL SURVEY

The ethanolamines can coordinate through cither oxygen or nitrogen. Tettamanzi and Carliul found that triethanolamine tonus addition com- pounds of the type M.\ _ \ (',11,011);; (where M isCo, Ni,Cu,Cd, PI), Ca, Mgj or Sr . some of the compounds being hydrated. No Btudy of the struc-

Ures of these compounds lias been made, bul in view of the Structural

similarity of triethanolamine and nit rilotriacet ic acid, one may assume the presence oi chelate rings, their number depending upon the coordination number of the metal ion:

(ho-ch2-ch2)3_x (ch2ch2 ohj3_x

Ethers form addition compounds with a wide variety of compounds. Confirmation of this is found in the high solubility of the heteropolyacids, of uranyl nitrate, and of magnesium iodide, in ethers. The best known of the ether coordination compounds are those formed with the Grignard reagent. Spacu132 has prepared some interesting compounds in which ether and pyridine share the coordination sphere: [Mg py4(ether)2]Br2 and [Mg pyi ether]I2 .

The formation of a deep color in the well known iron (III) chloride test for phenols indicates that phenols form compounds with the heavy metals. In the thermometric, conductometric, and spectrophotometric titration of phenol with iron(III) chloride, Banerjee and Haldar133 find breaks at molar ratios of 1:3 and 1:6. Upon electrolysis, the iron(III) ion goes to the anode. These findings suggest the reactions

Fe^+_> [Fe(OC6H5)3]°^ [Fe(OC,H,).]-

( atechol, because of the effect of chelation, forms stable complexes with the heavy metals:

K*[MC /C«H')»}XH*0

131. Tettamanzi and Carli, Gazz. chim. Hal., 63, 566 (1933); 64, 315 (1934); AM accad. sci. 7 'asse sci. fis., mat. nat., 68, 500 (1933); Garelli, AM accad. sci.

fit., mat. not., 68, 398 (1933). - Cluj, 1, 72 (1921).

Banerjee and Baldar, Natun . 165, 1012 (1950). 134. Weinland and Binder, Ber.t 45, 148, 1113 (1912); 46, 874 1913); Weinland and Walther, Z. anorg. Chem., 126, Ml (191

26 CHEMISTRY OF THE COORDINATION COMPOUNDS

If the phenolic group can take part in the formation of a chelate ring with souk1 other strongly coordinating group, very stable complexes may be formed. Thus, naphthazarin reacts quantitatively with beryllium ion to give the complexes

HO O

AND

HO O in which the coordinated oxygen atoms are doubtless equivalent135.

Peroxide Coordination

Many salts have been shown to crystallize with hydrogen peroxide "of crystallization" 136. In some cases, at least, this may be chemically com- bined with the salt, as is shown by cryoscopic measurements137.

The peroxo group may serve as a bridge between two cobalt ions. When an ammoniacal cobalt (II) solution is allowed to stand in the air, the first product formed is a brown decammine-/x-peroxo-dieobalt(III) salt, [(NH3)5Co — 02 — Co(NH3)o]X4138, which upon further oxidation is converted to the deep green [(NH3)6Co — 02 — Co(NH3)6]X5 in which one of the cobalt atoms seems to have achieved a valence of 4+. The dicobalt(III) salts are reduced to cobalt(II) by four equivalents of arsenic(III) oxide (one equivalent for each cobalt and two for the peroxo group) while the co- balt(III)-cobalt(IV) salts require five equivalents of reducing agent. The brown dicobalt(III) salt is diamagnetic, whereas the cobalt(III)-cobalt(IV) sail is paramagnetic139.

135. Underwood, Toribara, and Neuman, J. Am. Chem. Soc., 72, 5597 (1950).

136. Tanatar,2?er.,32, 1544 (1899); Z. anorg. Chem. ,2%, 255 (1901); Rudenko, J. Russ.

Phys. Chem. Soc, 44, 1209 (1912); Kazanetzkii, ./. Russ. Phys. Chem. Soc., 46, 1110 (1914).

137. Jones and Murraj . Am. Chem. ./., 30, 205 (1903); Maass and Hatcher, J. Am.

Chem. Soc., 44, 2472 (1922). L38 \ ortmann, Monatshefte, 6, 404 (1885); Werner and Mylius, Z. anorg. Chem., 16, 246 (1898 ; Werner, .1////., 375, 1 (1910). ' . 1'u and Rehm, Z. anorg. allgem. Chem., 237, 79 (1938).

GENERAL SURVEY

'Vortman's sulfate" is a mixture of materials, containing the sulfates of

in,

(NH,)4Co

III Co(NH,)<

(A) and

III/ \IV (XH3)4Co Co(NH3)4

\ / XII.

(B)

Compound B, on wanning with sulfuric acid, liberates one and a half atoms of oxygen, and on further heating, two-thirds of an atom of nitrogen, leav- ing t he cobalt in the dipositive state. These reactions, again, confirm the tetravalency of one cobalt atom. The surprising stability of these com- pounds is illustrated by the reaction

/ \ (XH3)4Co Co(NH3)4

\ / NH,

X4 + en

en2Co

XH,

Coen<

Compound B and its ethylenediamine analog are both paramagnetic140. The ethylenediamine compound can be reduced to the dicobalt(III) state by nitrite, hydrazine, ferrocyanide, arsenite or thiosulfate, but not by hy- droxylamine, hydrogen peroxide or mercury(I) ion. The product of the reduction can then be reoxidized to the Co(III)-Co(IV) state by treatment with permanganate, hypochlorite, bromine, bromate, or nitric acid, but not by dichromate, peroxide, or mercury(II), iron(III) or silver ions. These reactions establish the reduction potential at about one volt.

The peroxo- group in compound B can be replaced by other groups with reduction of the cobalt to the 3+ condition. Thus

/ \ (XH3)4Co Co(XH3)4

\ / NH,

+ SO:

(XH3)4Co

S04

/ \ > (

\ /

XII:

Co(XH3)4

Among the other doubly bridged cobalt (III), cobalt (IV) compounds described by Werner the triply bridged compound

XII., Ill/ \IV (NH ) Co-OH-Co(XTI3)f

\ / 02

k worthy of note1

140. Malatesta, Gazz. chim. ital., 72, 287 (1942),

141. Werner, Ann., 375, 104 (1910).

CHEMISTRY OF THE COORDINATION COMPOUNDS

Brimm142 has pointed out that most of the results which have been inter- preted to show the presence of tetrapositive cobalt in these compounds can be explained on the assumption that they contain the superoxide group.

Connick and McVey143 have identified two peroxo complexes of plu- tonium(IV) in aqueous solution. While the structures of these are not lii i ally proved, they seem to contain the rings

and

Metallic Oxide Coordination

Metallic oxides frequently coordinate with metallic ions, as is evidenced by the increased solubility of such oxides in salt solutions. Beryllium oxide, for example, dissolves readily in saturated beryllium sulfate solution, at the same time increasing the solubility of the sulfate itself144. The solubility relations indicate that each beryllium ion combines with four beryllium oxide molecules. The compound [Be(BeO)4]S04 is more soluble than its analog [Be(H20)4]S04 . The structure given for the complex is supported by the lowering of the freezing point, which indicates that addition of beryllium oxide to a solution of beryllium sulfate does not increase the number of ions in solution. Beryllium selenate gives the same result as the sulfate. A related situation is found in the anion commonly described as [RuCl5OH]=, but which is shown by crystal analysis to be the oxo complex [Cl5Ru— 0— RuCl5]4- 145.

Oxyanion Coordination

The anions of all oxyacids have donor properties, but in very different degree. It is sometimes said that the nitrate and perchlorate ions do not enter into complex formation, but this is not true. Nitratopentammineco- balt(III) salts were prepared by some of the earliest investigators, and were described in detail by Jorgensen146. Later investigations have led to the preparation of [Co(NH3)3(N03)3],147 [Co(NH3)4(N03)2]N03-H2(V48 and

142. Brimm, private communication. Quoted in Kleinberg "Unfamiliar Oxidation

States," p. 100, University of Kansas Press, 1950.

143. Connick and McVey, National Nuclear Energy Series, Vol. 14B (The Transura-

nium Elements), p. 445, 1949.

144. Sidgwick and Lewis, J. Chem. Soc, 1926, 1287.

145. Mellor, Report of the Brisbane meeting of the Australian and New Zealand As-

sociation for the Advancement of Science, 28, 137 (1951).

146. Jorgensen, J. prakt. Chem., [2] 23, 227 (1881).

147. Jorgensen, Z. anorg. Chem., 5, 185 (1894).

148. Birk, Z. anorg. allgem. Chem., 164, 241 (1927).

GENERAL SURVEY 29

[Co imi.(\();;)2]N03-H20149. The last two are shown to be dinitrate salts rather than aquo nitrato salts by the fact tliat the loss of water does not change the properties greatly.

Transference measurements on solutions of plutonium(IV) in \M HN03 indicate the existence of the complex [Pu(N03)]+, which coordinates with more nitrate ions as the concentration of IIX03 is increased. In bM acid, the bright green ion [Pu(X03)6]= is present, and (XH4)2[Pu(N03)6] can be crystallized from the solution. Thorium shows a similar behavior, giving a salt which is isomorphous with the plutonium(IV) and cerium(IV) com- pounds150.

G. F. Smith and his students have demonstrated the existence of both nitrate and perchlorate cerium(IV) ions151 but the exact structure of the ions is not yet clear. The oxidation-reduction potential of the cerium (III)- eerium(IY) couple varies greatly with the nature of the acid present. In IN acid, the electrode potentials (referred to the normal hydrogen elec- trode) are HC104 , 1.70 volts; HN03 , 1.61 volts; H2S04 , 1.44 volts; HC1, 1.28 volts. This variation indicates that either the Ce(III) or the Ce(IV) or both, combine with the anion of the acid. Duval152 has reported pentam- minecobalt complexes in which chlorate, bromate, iodate and perchlorate groups occupy the sixth coordination position.

The sulfate ion can occupy either one coordination position, or two. In either event, of course, it contributes a charge of minus two to the ion of which it becomes a part. The first type of compound is illustrated by sul- fatopentamminecobalt(III) bromide, [Co(XH3)5S04]Br153, which is pre- pared by heating the chloropentammine chloride with concentrated sul- furic acid. The sulfate group in the coordination sphere is not readily replaced, but is precipitated by boiling with barium salts. The ion slowly aquates on standing in solution:

[Co(NH3)5S04]+ + H20 -+ [Co(NH3)5(H20)]+++ + SOr

Sulfato-aquo complexes of several types evidently exist in aqueous solutions of chromium (III) sulfate154.

Cases in which the sulfate group occupies two positions in the same co- ordination sphere are not as well known. The double sulfates of iron, chrom-

149. Schramm, Z. anorg. allgem. Chem., 180, 170 (1929).

150. Hindman, National Nuclear Energy Series, Vol. 14B (The Transuranium Ele-

ments), p. 388, 1949.

151. Smith, Sullivan, and Frank, Ind. Eng. Chem., Anal. Ed.} 8, 449 (1936) ; Smith and

Getz, Ind. Eng. Chem., Anql. Ed., 10, 191 (1938); Kott, thesis, University of Illinois, 1940.

152. Duval, Ann. Chim., 18, 241 (1932).

153. Jorgensen, J. prakt. Chem., [2] 31, 270 (1885).

154. Enlmann, Angew. Chem., 64, 500 (1952).

:*()

CHEMISTRY OF THE COORDINATION COMPOUNDS

ium and the rare earths may contain the anions [M(S04)3]=, but they are too unstable to exist in solution. The case of potassium iridium sulfate, 3KaS04-Ir2(S04)8-2H20 or K3[Ir(S04)3]-H20, is perhaps a little more cer- tain, for this salt does not give the characteristic tests for sulfate ion155. Wein- land and Sierp168 have prepared alkaloid salts of the acids H3[Fe(S04) (€204)2] and H:!( Fe(S< VMC^ )*)], in which the sulfate group is evidently doubly coordinated. Duff157 claims to have prepared [Co en2S04]Br-H20, but Job158 and Ephraim and Flugel159 believe the salt to be [Co en2(H20)S04]Br, in which the sulfato group occupies only one coordination position. In any event j the sulfate group is not held very tenaciously, for in solution the complex ion is rapidly converted to [Co en2(H20)2]+++.

Several cases are known in which the sulfato group acts as a bridge be- tween two metal atoms, but in every case it must evidently be accompanied by some other bridging group. When octammine-ju-amino-ol-dicobalt(III) chloride,

NH2

/ \ (NH3)4Co Co(NH3)4

\ / OH

Cl4,

is heated with sulfuric acid, the "ol" bridge is replaced by a sulfato bridge:

NH2

(NH3)4Co

Co(NH3)4

\ /

oso o2

The sulfato bridge is eliminated by heating with concentrated hydrochloric acid; chloroaquo-octammine-/x-amino-dicobalt(III) chloride

Cl3 160.

[CI H20 "1

1 I

(NH3)4Co— NH2— Co(NH3)JC!4

results. The /x-amino-sulfato compounds are also obtained161 by the action of sulfur dioxide upon salts of the /x-amino-peroxo series (see page 27).

155. Delepine, Compt. rend., 142, 1525 (1906).

156. Wcin land and Sierp, Z. anorg. Chem., 117, 59 (1921). L57 Duff, •/ Chem. Sue, 121, 450 (1922).

L68 Job, Bull. 80C. chim., |4] 33, 15 (1923).

159 Ephraim and Flugel, Helv. chim. Acta,!, 727 (1924).

L60 Werner, Beddow, Baselli, and Steiniteer, Z. anorg. Chem., 16, 109 (1898).

163 Werner, .1/,//., 375, 15 (1910

GENERAL SURVEY 3]

Gibson and his co-workers"1'- have studied a case of a very different type of sulfate bridging. The substance iCjII.-.hAn-jSO., was Pound to be a dimer in acetone, and probably has the structure

Foss and Gibson163 have reported a similar compound in which the phenyl phosphate group, C6H5OP03=, replaces the sulfate.

The sulfate ion has the rather unusual ability to form hydrates; metallic sulfates usually crystallize from solution with one molecule of water more than other salts containing the same metallic ion. Thus the vitriols of the divalent ions of magnesium, zinc, cadmium, vanadium, chromium, man- ganese, cobalt, and nickel are heptahydrates and that of copper is a penta- hvdrate. In these complexes, two oxygens of the sulfate ion are hydrogen bonded to the water.

The tellurate and iodate ions are remarkable in that when they co- ordinate with copper, they stabilize the trivalent state, forming such com- plexes as [Cu(Te06)2]9- and [Cu(I06)2]7- 164- 165.

The bleaching of solutions of iron (III) chloride by addition of phosphate ion indicates the existence of phosphate complexes166. Ricci167 advanced evidence for the existence of H3[FeCb,P04] and H3[FeCl3As04], but later work indicates that the complexes probably contain no chlorine. Jensen168 found the solubility of FeP04 and A1P04 to rise with increasing phosphate ion concentration, but to be independent of the chloride ion concentration.

162. Gibson and Weller, ./. Chem. Soc., 1941, 102; Evens and Gibson, ./. Cht m. Soc.,

1941, Hi!».

163. Foss and Gibson, ./. Chem. Soc, 1949, 3075.

164. Malatesta, Gozz. chim. itol., 71, 407, 580 (1941 I.

165. Lister, Can. ./. Chem., 31, 638 1953).

166. Weinland and Ensgraber, '/. anorg. Chem., 84, 340 L91 1 ,

L67. Ricci and Meduri, Gazz. chim. itol., 64, 235 1934); Ricci and Lamonica, G

(■hint, itol., 64, 294 (1934 ; Ricci and Saraceno, thesis, University of Messina, 1929.

168. Jensen, Z. anorg. aUgem. Chem., 221, 1 (1934).

32 CHEMISTRY OF THE COORDINATION COMPOUNDS

I)i-, tri- and polyphosphates all show a remarkable ability to form stable complexes, even with the alkaline earth ions, so some of them have found wide use industrially (Chapter 23). Pyrophosphate complexes of many metals have been studied in solution by a variety of physical methods. For example, Haldar189 has studied the pyrophosphate complexes of Cu++, Ni++, and Co++ by thermometric and conductometric titrations, and by magnetic, cryoscopic, and transport measurements. He finds evidence for i he existence of two series of complexes, [M(P207)]= and [M(P207)2]6_. Watters and Aaron170 report, in addition, copper complexes with Cu:P2074_ ra1 ios of 2: 1 and 4:1, which, however, exist only in dilute solutions.

The carbonate ion forms coordinate bonds easily, as witnessed by its strong tendency to unite with hydrogen ions. In the metal amminessuchas |( <)(\II;;)4C03]+ it seems to occupy two coordination positions. In view of i he fact that this coordination entails the formation of a four-membered ring, it is surprisingly stable. Because the pentammine [Co(XH:05CO3]Cl- 1 1-< I gives an alkaline reaction, and because he thought that the molecule of water could not be removed without destruction of the complex, Werner was of the opinion that the formula of the salt should be written i(,(»(\H3)5HC03]Cl(OH)171. Lamb and Mysels172, however, found that all of the water can be removed without destruction of the complex, so it is evidently not essential to the constitution of the complex. On the other hand, the carbonato complex does undergo aquation in water solution, first yielding [Co(NH3)5HC03]++ and then [Co(NH3)5(H20)]+++ 173. The anala- gous ion, [Co(NH3)4C03]+, aquates to fCo(NH3)4(HC03)H20]++, and then to [Co(XH3)4(H20)2]+++ 174. Stranks and Harris175 studied the exchange in solution of C-labelled carbonate with the carbonate in [Co(NH3)4C03]+ and Yankwich and McXamara176 did the same with [Co en2C03]+. The exchange takes place through the intermediate formation of a bicarbonate complex.

By using labeled oxygen, Taube and his students demonstrated that in the cases of [Co(NH3)6C03]+ and [Co(NH2)4C08]+ exchange does not in- volve rupture of the cobalt-oxygen link, but rather, of the carbon-oxygen bond177.

169. Haldar, Smnr, and Culture, 14, 340-1 (1949); Nature, 166, 744 (1950).

170. Watters and Aaron, ./. .1///. Chem. S<>c, 75, 611 (1953).

171. Werner, Ber. 40, 4101 (1907

172 Lamb and Mysels,/. .1///. Chem. Soc., 67, 468 (1945). 17:: I. ami) and Stevens, •/. .1///. Chem. So,-., 61, 3229 (1939). 17 1 ll.ii lis ;in(l Si tanks. Trans. Faraday Soc. 48, 137 (1952). 177, Stranks and Harris. ./ . Chem. Phye., 19, 267 ^ 1951). L76 Vankwich and McNamara,/. Chem. Phys., 20, 1325 (1952 177. Hunt. Rutenberg, and Taube, ./. Am. Chem. Soc, 74, 268 (1952); Posey and Taube, ./. .1///. Chem. Soc. 75, 4099 i"

GENERAL SURVEY

McCutcheon and Schuele178 bave recently isolated the interesting ion [Co(C03)i]" as the hexamminecobalt(III) salt; its existence clearly indi- cates thai tin1 carbonate ion can fill two coordination positions.

Organic inion Coordination

Many organic anions form stable coordination compounds. Formate and acetate ions form strong bonds, but monocarboxylic acids with Longer chains show a rapidly decreasing ability to coordinate. Formate and acetate often bind two metal atoms together, each oxygen of the carboxy] group linking

to a different metal atom.

M— OC=0— M.

When the carboxy] group is attached to only one metal atom, however, it tills but one position in the coordination sphere. Complexes of the types [Co(NH; *OOCCH,]++ m and [Co(XH3)5OOCH]++ 180 are well known and easily prepared. The solubilities1"1 and stabilities182 of several similar com- plexes containing a variety of aliphatic anions have been studied.

• and Bailar183 were able to effect a partial resolution of a-chloropro- pionic and a-bromopropionic acids through the formation of stable cobalt <•( tmplexes containing levo-propylenediamine, [Co ?-pn2(OOC • CHX • CH:j)o]+. The solubility of lead sulfate in solutions of sodium acetate has inspired much research, and many formulas have been postulated for the complexes which are formed184. Weinland and his students121 report the isolation of the polynuclear complex ions

/ \

peva and Batyrshine186, however, report only the formation of [Pbac]+, pPbacj]-, and [PbacJ", the last being the most important in analytical work.

178. McCutcheon and Schuele, /. Am. Chem. Soc, 75, 1845 H»53).

179. 1 4, 171 1953).

180 \ atsimirekii, ./. Gen. Chem. I S.S R.), 20, 140s I960 .

181. Linhard and Rau, Z. anorg. cUlgem. Chem., 271, 121 '1952).

182. Bunton and Llewellyn, J. CI - 1953, L6 -

tnd Bailar, ./. Am. Chem. Soc., 74, 1820 L952). 184. Weinland, "Einfuhrung in die Chemie der Komplexverbindungen," Becond Edi- tion, pp. 391 100, Enke, Stuttgart, 1924. Is-."). Toropova and Batyrshina, Zkur. Anal. Kkim., 4, 337 194

:;i CHEMISTRY OF THE COORDINATION COMPOUNDS

The "basic acetate" method of separating the ions of the trivalent metals in qualitative analysis involves the formation of acetate complexes. Wein- land and his students studied many of these184 and isolated some very complex materials which they thought were true chemical entities.

Among the examples in which the carhoxyl group forms a bridge between two metal atoms are the "basic" beryllium salts, Be40(OOC-R)6 , in which R represents (II:;. (YII5, etc. These compounds are readily formed and are stable, volatile, and soluble in nonpolar solvents. Structural studies186 indicate the presence of a central oxygen surrounded tetrahedrally by four beryllium ions. Each edge of the tetrahedron is composed of the grouping

R Be — O — C — O — Be. Similar compounds of zinc187 and zirconium, I XrO)40(OOCR)6 ,188 are known.

The oxalate ion forms a great many stable coordinate compounds, usually acting as bidentate group. The best known are those of the types !M"()x;J4-, [MmOx3h and [MmOx2]-. The tris-(oxalato) complexes have been studied extensively, especially in regard to their stereochemistry. (Chapter 8). The oxalate group can share the coordination sphere with ammonia, ethylenediamine, water, or other groups. Oxalatobis(ethylene- diammine) cobalt(III) chloride, [Co en2Ox]Cl, is readily obtained by the action of an alkali oxalate upon the dichloro salt189; the corresponding chromium salt is prepared by the action of ethylenediamine upon the tris- (oxalato) salt190. Hamm and Davis191 have studied the formation of these ions by the reaction of [Cr(H20)6]+++ and oxalate ion, and Hamm192 has followed the rate of isomerization of [Cr(H20)20x2]~ in water solution. He postulates that upon collision with the ion, a water molecule knocks one end of an oxalate group away from the chromium and takes its place; on return of the oxalate, either the cis- or trans- isomer may be formed, depending upon which molecule of water is eliminated. A small amount of alkali con- verts the diaquo compounds to hydroxoaquo- compounds, the cis isomer

1S6. Bragg and Morgan, Proc. Roy. Soc. London, A104, 437 (1923); Morgan and Ast-

bury, Proc. Hoy. Soc. London, A112, 441 (1926), Pauling and Sherman, Proc.

Natl. Acad. Sri., 20, 340 (1934). is?. Auger and Robin, Compt. rend., 178, 1546 (1924); Wyart, Bull. Soc. Fr. Min., 49,

1 is (1026). 188. Tanatar and Kurowski, Chem. Centralblatt, 1908 (1) 1523. L80. Werner and Vilmos, Z. anorg. Chem., 21, 153 (1899); Price and Brazier, ./. Chem.

Soc, 107, 1376, 1726 (1915). l'Mi Werner and Schwarz, .1////., 405, 222 (191 \-. I'M. Hamm and Davie, ./. .1///. Chem. Sue. 75, 3085 (1953). 192 Hamm, ./. Am. Chem. Soc, 75, 609 (1953).

a i:\f-: ual sritVEY

of which is converted upon heating into the tetrakis(oxalato)-M-diol-salt,

Ox,Cr CrOx,

\ /

oil

Larger amounts of alkali change the diaquo salts to dihydroxo salts, still without breaking the chromium-oxalate linkage

Weinland and Paul1-- have isolated several compounds of the ion [Pr>Ox]+~, in which all four of the oxygen atoms are probably bonded to the metal:

0— c=o

O— C^O

Solubility studies193 have indicated the existence of analagous ions of zinc and cadmium.

The stability of the oxalato complexes is largely due, no doubt, to the formation of five-membered rings. Compounds are known, however, in which rings are not formed. Griinberg's method of determining the con- figuration of cis-trans isomers of the type [Pt(XH3)2X2]194 is based upon the inability of the trans-isomer to yield a chelate oxalato derivative. (See Chapter 9).

The oxalate ion, like the sulfate ion, forms hydrates. Werner has pointed out195 that a large number of compounds containing complex oxalate anions crystallize with water, even if the cation is one which is usually anhydrous.

The malonate ion coordinates with metallic ions to give a six-membered ring, which is not as stable as the five-membered ring formed from the oxalate ion. Schramm has studied the formation of malonatotetrammine- cobalt(III) compounds in some detail196. Anions of other dibasic organic acids form cations of the type [Co en2A]+, but seem unable to form anionic complexes like those formed by oxalates and malonates. Complexes of some difunctional acids are discussed in Chapter 6.

a-Hydroxy acids often coordinate readily, the hydroxy] and carboxy] group both coordinating, and the chelation effect enhancing the stability

193. Vosburgh and Beckman, ./. Am. Chan. Soc, 62, 1028 (1940).

194. Gr&nberg, Helv. ckim. Acta, 14, 455 (1931).

195. Werner, "New Ideas on Inorganic Chemistry," Translated by

London, Longmans, Green & Co., 1911.

196. Ref. 140, p. 161.

Hedley, p. 113,

36 CHEMISTRY OF THE COORDINATION COMPOUNDS

of the compounds formed. The hydrogen of the hydroxyl group may be lost simultaneously, so that the organic group contributes a charge of minus two to the complex. Thus, coordination with the copper(II) ion gives

The copper complexes containing glycollic and lactic acids are not very stable197 but those containing the stronger salicylic and mandelic acids are easily isolated198. Boron forms stable compounds even with the simpler a-hydroxy acids199, and Boesken and his co-workers were able to resolve the bis-(a-hydroxybutyro)borate ion200 as well as the bis(salicylato)borate ion201. The work of Jantsch202 on the rare earth glycolates and lactates indicates that some chelation takes place. His values for the equivalent conductances of various lanthanum salts are as follows:

v X

acetate 1024 89.5

phenylacetate 1200 91.2

glycolate 1200 70.3

lactate 1024 54.1

Salicylate ion differs from its meta- and para- isomers in being able to form chelate rings, which greatly stabilizes its coordination203. Many recent studies have been made on solutions of metal ions and a-hydroxy acids, such as salicylic, lactic, citric, glycollic, and tartaric; these studies lead to a knowledge of the compositions and stabilities of the complexes formed, but do not give information on their structures. The work of Bertin-Batsch and of Bobtelsky and his collaborators204 is typical.

The compounds of the a-amino acids are of great stability, and have re- ceived extensive study. Ley205 and Bruni and Fornara206 suggested that

197. Wark, J. Chem. Soc, 123, 1815 (1923).

198. Wark, J. Chem. Soc, 1927, 1753.

L99. Rosenheim and Vermehren, Ber., 57, 1337 (1924).

200. Boeseken, Muller, and Japhongjouw, Rec. trav. chim., 45, 919 (1926).

201. Boeseken and Meulenhoff, Proc. Acad. Set. Amsterdam, 27, 174 (1924).

202. Jantsch, Z. anorg. allgem. Chem., 153, 9 (1926); Jantsch and Griinkraut, Z. anorg.

allgem. Chem., 79, 305 (1913).

203. Bertin-Batsch, Ann. chim., 7, 481 (1952).

Jin Bobtelsky and Eeitner, Bull. soc. chim. France, 1951, 494; Bobtelsky and Graus, J. A m < "h< »i . Soc, 75, 4172 (1953) ; Bobtelsky and Bar-Gadda, Bull. soc. chim. Franc* , 1953, 276, 687.

205. Ley, Z. Elektrochem., 10, 954 (1904).

206. Bruni and Fornara, Aiti accad. Lincei, [5] 13, II, 26 (1904); Bruni, Z. Elektro-

chem., 11, 93 (1905).

GENERAL SURVEY M

copper glycine is an inner complex. The deep blue color of the compound indicates copper-nitrogen linkages, and the possibility of the formula CiuXIICIU'ooiu is eliminated by the facl thai N,N-diethylglycine gives an analagous compound. The compound is a nonelectrolyte, and i1 is evident thai the copper is coordinately saturated, for it absorbs ammonia only very slowly. Finally, the properties of copper glycine are very similar to those of diamminecopper(II) acetate [Cu(OOCCH3)2(NH3)2], which seems to justify the formula

The copper(II) compounds of a-amino acids are so stable that they do not respond to most of the usual tests for copper(II) ion. Hydrogen sulfide de- posits copper sulfide, and boiling alkalies precipitate copper oxide, but both reactions take place slowly. The opening of the ring by ammonia to give [Cu(XH3)2(OOCCH2XH2)2]207 is an interesting reaction. The remarkable stability of the copper chelate of the a-amino acid group is illustrated by the work of Kurtz208 who studied several acids of the type

XHo— (CH2)Z— CH— COOH,

I NHs

where X = 2, 3, or 4 (a , 7-diaminobutyric acid, ornithine, and lysine). In each case the usual properties of the carboxyl group and the adjacent amino group are completely masked, but the other amino group retains its characteristic behavior, and Kurtz was able to carry out reactions on it, without affecting the coordinated amino group.

The cobalt complexes of the a-amino acids, [Coamac3], exist in two stereo- isomeric forms (see page 283), both of which are remarkably stable, being unat tacked by 50 per cent sulfuric acid. Elliott209 has utilized this stability in the preparation of highly insoluble and stable "super complexes" by the reaction of cobalt (III) hydroxide with

IK >< >c— CH— (CH2)n— CH— COOH I I

NHS XII:

Chromium(III) forms inner complexes which are similar but of less sta- bility; they are .-lowly decomposed by hot acids, by sodium hydroxide, and

Ley, Ber., 42, 354 (1909).

208. Kurtz. ./. Biol. Chem., 122, 177 (1937-8); 180, 1253 (1949).

209. Elliott, thesis, University of Illinois, 1943.

CHEMISTRY OF THE ('OOEI)I XATIOX COMPOUNDS

to ;i degree, by foiling water. Keller210 has studied the reactions of a large number of a-amino acids with chromium (III) hydroxide and chromam- mines in boiling water. In all cases compounds of the formula [Cr(amac)3] aeem t<> form, l>ut are quickly hydrolyzed to

/ \ amac2Cr Cramac2

\ /

which in turn hydrolyze slowly to

OH OH OH

/ \l/ \ amacoCr Cr Cramac2

\ /l\ /

OH OH OH

and more complex products. Cobalt amino acid compounds undergo the same reactions, but much more slowly.

Platinum does not readily coordinate with oxygen, but the coordinating tendency of the a-amino acids is so great that such compounds as

PtCl;

0 (

:=o"

and

NH2— (

}H2 _

O C=0>

CH-

can be formed211,212,213. Even a-amino acids containing tertiary nitrogen atoms will coordinate Avith platinum strongly, as is shown by the optical resolution of the ion

(N02)2Pt

CHj C2H5

c=o

CH,

Heterocyclic acids having a carboxyl group in the a-position to the ring nitrogen (picolinic, quinolinic, quinaldinic, etc.) form inner complexes. The compounds with iron(II), which arc deeply colored, have been studied by

210. Keller, thesis. University of Illinois, L940.

211. Ley and Picken, />'</•., 45, 377 (1912).

212 I Irinberg and Ptitzuin, .1////. inst. platine, No. 9, 55 (1932). Grinbergand Ptitzuin, Am,, inst. platine, No. n. 77 (1933). _•] l Kueblerand Bailar, J. Am. Ckem. Sac, 74, 3535 (1952).

GENERAL SURVEY

Ley and his co-workers-1'. The corresponding copper(II) compounds are light in color, and are probably not coordination compounds.

The fi-amino acids also form inner complexes with the transition metals, hut these are less stable than those of the a-acids. Hearn218 has shown that a-amino acids can be distinguished from the 0-aeids by the fact that the former react with cobalt (III) hydroxide to give colored complexes, while the latter do not.

The y-, 5-, and e-amino acids do not form chelate rings with metals, so form normal salts217.

Among the amino acids, the derivatives of acetic acid are particularly noteworthy for their chelating ability. The tridentate iminodiacetic acid gives many complexes, which in general are more stable than those of gly- cine. For example, the first and second stability constants of the zinc com- plex of glycine are 4.8 and 4.1, while for the zinc complex of iminodiacetic acid they are 7.8 and 5.7218. Nitrilotriacetic acid forms still more stable complexes, the two dissociation constants for the zinc complex being 10.5 and 3.0219. The great difference between the two values in the case of the triacetic acid doubtless reflects the fact that the zinc ion cannot accept all of the possible donor groups in two of the donor anions. The complex which is formed in this case220 is

-i4 —

OOC-CH2— N

The most remarkable of the acetic acid derivatives, however, is ethylene- diaminetetraacetic acid (often abbreviated EDTA or H4Y). This substance is potentially hexadentate, but complexes in which only four or five groups are coordinating are well known. The complexes of EDTA are remarkably stable, so have been investigated extensively from the industrial point of

215. Ley, Schwarte, and Miinnich, Ber., 57, 349 (1924).

216. Hearn, thesis, University of Illinois, 1951.

217. Tschugaeff and Serbin, Compt. rend., 151, 1361 (1910); Pfeiffer and Lubbe, ./.

prakt. Chem., [2] 136, 321 (1933).

218. Flood and Loras, Tids. Kjemi, Bergsvesen Met., 6, 83 (1945). JIM. Schwarzenbach, Chimin, 3, 1 (1949).

220. Schwarzenbach and Biedeimann, Eelv. Ckitn. Ada, 31, 331 (1948).

CHEMISTRY OF THE COORDINATION COMPOUNDS

\ i.w . More than four hundred and fifty articles were published during 1952 describing uses of this reagenl or stability constants of its metal derivatives. 1 1 has been used in water softening (Chapter 23), electroplating, controlling the metal contenl of dye baths, in removing lead and other heavy metals from the human Bystem, in the treatment of chlorosis in plants, and in many other ways,

The stability of the EDTA complexes is illustrated by the fact that neither the copper(II) or the nickel compound is destroyed by sodium or ammonium hydroxide. The nickel compound is not attacked by dimethyl- glyoxime or hydrogen sulfide, but is destroyed by potassium cyanide. The copper compound gives the usual reactions of Cu++ when treated with potas- sium cyanide, hydrogen sulfide, or potassium ferrocyanide221.

The ability of ethylenediaminetetraacetic acid to form stable complexes depends upon the fact that when it coordinates it forms multiple fused five- m< -inhered chelate rings. Pfeiffer and Simons222 compared the calcium deri- vatives of methylaminediacetic acid

CHoCOO^

CH8— N

CHoCOO/ 2 and ethylenediaminetetraacetic acid,

cir-cocr

ff.

-f-CH2— N

CH2COO

which differ only in that the two nitrogen atoms in the latter are linked together through the ethylene bridge. The methylamine complex reacts slowly with oxalate ion to precipitate calcium oxalate, but the ethyl- enediamine complex does not. Pfeiffer and Simons came to the conclu- sion that these complexes are hexadentate, for the structurally similar I K M )CCH(CH3)NIICH2CH2NCH(CH3)COOH does not form a stable cal- cium complex.

Several studies have been made of the effect of ring size on the stability

of complexes of this type. Schwarzenbach and Ackermann223 investigated

i iee 1 1« »« x JCH2)2N(CH2)nN(CH2COOH)2 , where n varies from two

to five. In general, the stability of the alkaline earth compounds decreases

increases. When "n" is 4 or ~>, the two ends of the molecule seem

221. Brintzinger and Hesse, /. anorg. allgem. Chem., 249, 113 (1942). Pfeiffer and Simons, Ber.,76B,847 (1943). 3< bwaraenbach and Ackermann, Help. Ckim. Ada, 31, 1029 (1948).

GENERAL SURVEY 41

able to act independently, for complexes of the type M-Y can be formed. Chaberek and Mart el l-'-'1 found the stabilities of the complexes of ethylene- diaininediacetic-dipropionic acid to be considerably less than those of the tetraacetic acid.

Some ca>es are known in which EDTAdoes not act as a hexadentate co- ordinator, even though six positions are open to it. Thus, Schwarzenbach228 prepared the compounds [CoHYBr]- and [CoHY(N02)]~. Removal of the bromide or oitro group allows the unattached carboxyl group to coordinate with the cobalt to form [CoY]~. Busch'-"-6 has shown that the palladium(II) chelate has the structure

CH2-CH2\ HpC7/ Pd / CH2

_ o=c — o' 'O — c=o

The stereochemistry of the EDTA complexes is discussed in Chapter 8. Carbonyl Coordination

The carbonyl group of aldehydes has rather weak donor properties, but addition compounds of aldehydes with several of the light metals, such as magnesium227, and with the wreakly basic elements, such as tin and anti- mony,*28 are known. The carbonyl group of esters also forms rather weak coordinate links with these metals229. Simple aliphatic ketones show similar behavior.

The 1,3-dicarbonyl compounds, through their ability to enolize, form stable chelate rings with a large number of metals. In many cases the com- pounds so obtained are nonionic, insoluble in water, soluble in nonpolar solvents, and volatile. Acetylacetone has received the most attention in this regard, but dibenzoyl methane, benzoylacetone, acetoacetic ester, sali- cylaldehyde, benzoyl pyruvic acid, and o-hydroxyacetone are important. Thenoyltrifluoroacetone (TTA),

O O

"C CHo — C — CF;

224. Chaberek and Martell, J. Am. Chem. Soc, 74, 6228 (1952).

225. Schwarzenbach, Helv. Chim. Acta, 32, 839 (1949).

226. Busch and Bailer, J. Am. Chem. Soc, in press, 1956.

227. Menschutkin, Izvest. St. Petersburg Polyttch. Inst., 6, 39 (1906).

228. Menschutkin, ./. Russ. Phys. Chem. Soc., 44, 1929 (1912); Rosenheim and Soil

man, Ber., 34, 3377 (1901); PfeifTer, Ann., 376, 296 (1910).

229. Menschutkin, Izvest. St. Petersburg Polytech. Inst., 4, 101 (1906); 6, L01 L906

Lewy, J. prakt. Chem., 37, 480 (1846).

CHEMISTRY OF THE COORDINATION COMPOUNDS

has received much attention because of the great stability of its compounds. The classic paper of Morgan and Moss on the acetylacetone compounds230 reviews the Literature up to 1914 and describes the preparation of many compounds. Metallic ions having a coordination number twice the ionic charge give nonelectrolytic complexes:

"C—G

CH3 M=Be,Cu,Ni,ETC.

CH3 M=AI,Cr,r%CcvETC

CH3 M=Th. Zr,Hf,Ce,Pu,E-rc .

Many of these compounds show exceptional stability, the beryllium com- plex, for example, boiling without decomposition at 270°C at atmospheric pressure. Molecular weight determinations indicate that these compounds are monomeric. Wilkins and Wittbecker231 have utilized this stability in the preparation of beryllium containing polymers. They report that tet- rake tones form linear polymers of the types

R R

B C_Y_C

' o — c/ x — o

R R

Be:

•o=c

> —

c>=°>e/

AND

x y°=c —

Be ,CH

/ \

o— c;

— /C R

Be CH

■o' Nj-c;

— c=0\ /

HC • Be

*C — o' \

where Y is any one of a variety of organic groups.

There is, however, some popular misconception as to the stability of the diketone chelates. The statements that the rare earths can be separated through the volatility of their acetylacetonates232, and that the molecular weights of the rare earth acetylacetonates can be determined by their vapor

230. Morgan and Moss, ./. Chem. Soc, 105, 189 (1914).

231. Wilkins and Wittbecker, U. S. Patent 2,659,711 (Nov. 17, 1953).

232 Bphraim, "Inorganic Chemistry," English Edition by Thome, London, Gurney and Jackson, L926.

<,i:\i:i;.\l si'HVK)

\:\

densities18' are incorrect scandium acetylacetonate is readily volatile280, **,

hut those oi the true rare earths decompose on heating288,288, Brimm288 found that the rare earth compounds of dibenzoylmethane and benzoylace- tone are readily decomposed by traces of moisture with the formation of

[M(dik(it<>nrM( HI )< II-( ))], These compounds are soluble in organic solv- ents, hut are not volatile.

When the coordination Dumber of the central ion is less than twice the elect rovalence, cat ionic compounds are formed, as illustrated by the com- pounds containing boron, silicon and titanium287

M=Si,Ti

These compounds are of special interest because of their stereochemical possibilil ies and because they show typical metalloid elements in the role of cations. Similar compounds of other 1 ,3-diketones have been described238. If, on the other hand, the coordination number of the central atom is more than twice the electrovalence, the coordination sphere will tend to fill itself with other neutral groups237. Iron(II) forms the compounds

Y=NH3,pq,£en,(t)NHNH2,

PIPERIDfNE, NICOTINE

all of which are soluble in organic solvents, insoluble in water, and deeply colored-'. On heating in vacuo the ammonia compound is converted to dibenzoylmethane iron.

233. Hein, "Chemische Koordinationtheorie," p. 153, Zurich, Hirzel Verlag, 1050.

234. Meyer and Winter, Z. anorg. Chem., 67, 414 (1910).

235. [Jrbain, Ann. ckim., [7] 19, 212 (1900).

236. Brimm, thesis, University of Illinois, 1940.

237. Dilthey, Ber., 36, 923 (1003); 37, 588 (1904); .1////., 344, 300 .1905).

238. Dilthe: /;■ 36. 1595 3207 (1903); ./. prdkt. Chem., [2] 111, 147 (1925).

239. Emmerl and Gsottschneider. Ber.. 66, L871 (1933).

I } CHEMISTRY OF THE COORDINATION COMPOUNDS

2, t-Pentanediono-dimethyl thallium

/CH3 CH3 ^0=C

civ o— c

XCH3

has unusual properties240. It is soluble in benzene, has a low melting point, and sublimes readily. On the other hand, it is also soluble in water, giving an alkaline solution. This solution shows the usual properties of the di- methyl thallium ion, so it appears that the coordinate bonds are broken by water.

The diketone compounds which are soluble in organic compounds have achieved considerable importance as agents for the separation of metal ions through the techniques of solvent extraction. If two metals in aqueous solu- tion, are in equilibrium with a diketone, if the equilibrium constants are different and if the complexes are soluble in a solvent immiscible with water, the metals can be separated by liquid-liquid extraction241. Since the extent of dissociation of the complex of any metal can be changed by chang- ing the pH of the solution, the method is widely applicable. If a speci- fied metal is to be separated from several others, the pH is adjusted so that that metal (and those with smaller dissociation constants) will be extracted into the organic layer. This is then extracted with water, the pH of which is adjusted to allow only the extraction of the metal in ques- tion, since its complex has the largest dissociation constant of those now present. Bolomey and Wish242 used this technique to separate radioberyl- lium from the other metals obtained with it by cyclotron bombardment. Huffman and Beaufait243 employed the method to separate zirconium and hafnium, using thenoyltrifluoroacetone as the complex former. The dis- tribution coefficient of the zirconium complex is about twenty times that of i lie hafnium complex, so excellent separation was achieved.

This extraction technique can also be used to determine the formulas of complexes and the degree of hydrolysis of metal ions in aqueous solution, as was shown by ( Jonnick and McYey in their study of the zirconium ion244. By determining the extraction coefficient of the zirconium complex of then-

240. Menziee, Sidgwick, Fox, and Cutliffe, ./. Chem. Soc, 1928, 1288.

241. ( lalvin, Manhattan Project Report CN-2486, December 1944; Experientia, 6, 135

(1950).

242. Bolomey and Wish,./. .1///. Chem. Soc, 72, 4483 (1950). 213. Huffman and Beaufait,/. .1///. Chem. Soc, 71, 3179 (1949). 244. Connick and McVey, J, A»,. Chem. Soc, 71,3182 (1949).

GENERAL SURVEY 45

oyltrifluoroacetone between benzene and water as a function of the TTA activity in benzene, they were able to establish the composition of the che- late as [Zr(TTA).i]. By measuring the distribution of the zirconium between the benzene and water phases as a function of pH, they then demonstrated

that in the pi I range —0.4 to 2.0, the zirconium ion exists largely as a mix- ture of Zr*+ and Zr(OH)+++

Steinbach and Preiser248 have suggested that the complexing agenl (acetylacetone, in their example) can serve also as the solvent for the complex. Using this technique, they have effected the analytical separation of zinc and copper ions.

Oxygen Carrying Chelates

Hemoglobin and hemocyanin were long considered to be unique in their ability to absorb and release oxygen, but several types of synthetic com- pounds are now known which possess this property. Their behavior is illus- trated by a simple experiment: If cobalt nitrate solution is treated with ammonium chloride and ammonium hydroxide in the absence of air, a pink precipitate forms. When air is bubbled through the suspension, a brown color develops, but when nitrogen is substituted for the air, the pink color returns. This cycle can be repeated many times. Interestingly enough, the experiment fails if ethylenediamine is substituted for ammonia.

Pfeiffer, Breith, Lubbe, and Tsumaki246 reported that bis-(salicylal)ethyl- enediiminecobalt(II)

CH=lsT XN

I I

CH2-CH2

(A)

darkens in air. Tsumaki-'47 found that this is due to absorption of oxygen and thai the process is reversible. It has since been found that other cobalt chelates also show this property. ( !alvin and his students and Diehl and his students have studied compound (A) and many derivatives of it. Diehl248 reports thai the parent compound contains one-half mole of water per co- balt atom, and believes thai two molecules of the chelate are held together

245. Steinbach and Preiser, Anal. Chem.,25, 881 (1053).

246. Pfeiffer, Breith, Lubbe, and Tsumaki, Ann., 503, si (1933).

247. Tsumaki, Bull. Ch* Japan, 13, 252 L938

248. Diehl and co workers Bach, Harrison, Liggett, Chao, Brouns, Curtis, Bensel-

meir, Schwandl , Mathews . Iowa Sim, Coll. J. Sri., 21, 271, 278, 287, 311, 316, 326, 335 (1047); 22, 91, 110, 126, 129, 141, 150, 165 (1948); 23, 27:; 1949

46 CHEMISTRY OF THE COORDINATION COMPOUNDS

by an aquo bridge. This is a unique situation, for no other cases of aquo bridges are known. Calvin and his group249 have studied compound (A) and some of its derivatives from the structural point of view. Both Calvin and Diehl report that most of these compounds exist in several different isomeric forms, only one of which (for each compound) is active toward oxygen. Compound (A) is paramagnetic, apparently having one unpaired electron per cobalt atom. Diehl reports that it does not absorb carbon monoxide or nitrous oxide, but that it absorbs nitric oxide and nitrogen dioxide. He is of the opinion that it will absorb other paramagnetic gases, but not diamag- netic ones.

When put under pressure of oxygen, these materials, either in the solid state or in solution in quinoline or similar solvents, absorb one mole of oxy- gen for each two moles of chelate, and release it again when the pressure is decreased. In each repetition of the cycle, however, there is a small amount of irreversible oxidation, so the ability to absorb oxygen gradually de- creases.

Calvin's group also prepared compound (B)

CK<-p

(CH2)3-NH — (CH2)3 (B)

and several analogs of it. Compound (B) has three unpaired electrons per cobalt atom, and reversibly absorbs one mole of oxygen per atom of co- balt250.

Calvin's x-ray studies on compound (A) show that it crystallizes in layers, with holes running through the layers. These holes are big enough to con- tain oxygen molecules, and the passages between them, while smaller, are sufficiently large to allow such molecules to go through without great diffi- culty.

Cobalt (II) histidine chelates in water solution will absorb oxygen reversibly261. Histidine compounds of iron are oxidized irreversibly, while those of nickel and copper are not oxidized at all. The unoxygenated cobalt histidine complex is paramagnetic to the extent of three unpaired electrons per cobalt atom, while the oxygenated compound is diamagnetic. Hearon is of the opinion that cobalt is four covalent in this compound, and that the

249. Calvin and co-workera (Bailee, Wilmarth, Barkelew, Aranoff, Hughes), J. .1///.

Ch m. Snr., 68, 2254, 2257, 2263, 2267, 2273 (1946). 260. Harle and Calvin, J. Am. Chem. Soc. , 68, 2612 (1946).

GENERAL SURVEY 47

amino acid is coordinated to the metal only through nitrogen atoms

Two molecules of this chelate absorb one molecule of oxygen. It does not combine with carbon monoxide. According to Hearon251d- e, the oxygenated molecule has either the structure

(g is a molecule of water or some other neutral group)

Michaelis252 has also measured the magnetic susceptibility of the cobalt histidine compounds.

The properties of hemoglobin and its oxygen carrying capacity are dis- cussed in Chapter 21. Like the other oxygen carrying chelates, it is para- magnetic when deoxygenated, but diamagnetic in the oxygenated form253. A- is well known, it combines with carbon monoxide more firmly than with oxygen, and with cyanide ion or pyridine still more firmly.

The Doxor Properties of Sulfur

The donor properties of sulfur are quite different from those of oxygen. In general, they are somewhat more restricted as regards the nature of the acceptor atom, but in some types of compounds, they are exceptionally

251. Burk. Bearon, Caroline, and Schade, ./. Biol. Chem., 165, 723 (1946); Burke, I!

ron, Levy, and Schade, Federation Proc., 6, 212 (1947 ; Hearon, Federation . 6, 256 260 L947 :./. Nat. Cancer Inst., 9, 1 L94S ; Hearon, Burk, and Schade,/. Natl. Cancer Inst., 9, :>>:>>: 1049). Michaelis, Arch. Biochem., 14, 17 (1942). 253. Pauling and Coryell, Proc. Natl. Acad. Set., 22, 159, 210 L936).

48 CHEMISTRY OF THE COORDINATION COMPOUNDS

strong. The thioethers, for example, form much more stable compounds than the corresponding oxyethers. The coordination of sulfide (or hydro- sulfide) ion with the sulfides of arsenic, antimony, tin, copper, and mercury is well known and is of great importance in qualitative analysis. Similarly, the preferential coordination of sulfide ion plays an important part in the metallurgies of copper and nickel. The Orford process exploits the ampho- teric behavior of copper and iron toward sulfide in the separation of these metals from nickel. The separation is not quantitative, but repetition of the process gives further separation.

Thiohydrate Formation

Liquid hydrogen sulfide shows little resemblance to water in its solvent properties254, although some inorganic salts dissolve in it. A few thiohy- drates have been isolated255 • 256 ■ 257 and thiohydrolysis probably takes place through the formation of unstable thiohydrates. Morgan and Ledbury258 concluded that organic sulfides coordinate readily with those metals which occur as sulfides in nature, or which form very stable sulfides. They also found that the reactions of metal ions with dimethyldithiolethylene show analogies to their reactions with hydrogen sulfide. Thus, copper(II) and gold (III) chlorides, which are readily reduced by hydrogen sulfide, form the compounds

CH3 CH3

/5-CH2 /S-f*

CI2Cu AND CI3Au

XS CH2 S — CH2

CH3 CH3

which readily revert to copper(I) and gold (I) compounds. TschugaefT259 found that of the dithioethers, RS(CH2)nSR (n = 0, 1, 2, 3, 5), only the compounds having n = 2 formed stable, well-characterized chelates.

Dithiane, C4H8S2 , forms complexes with the ions of the coinage metals, platinum, mercury, and cadmium260. The ratio of dithiane to metal varies

254. Antony and Magri, Gazz. chim. ital., 35, 206 (1905).

266. Plotnikov, ./. Ruse. Phys. Chem. Soc, 45, 1162 (1913).

256. Hill/, and Keunecke, Z. anorg. allgem. Chem., 147, 171 (1925).

267 Ralston and Wilkinson, ./. Am. Chem. Soc, 50, 258 (1928).

268. Morgan and Ledbury, ./. Chem. Soc, 121, 2882 (1922).

260. Tschugaeff, Ber., 41, 2222 (1908); TschugaefT and Kobljanski, Z. anorg. Chem.,

83, 8 L913); Tschugaeff, Compi. rend., 154, 33 (1912); Tschugaeff and Subbo-

tin, Ber.t 43, 1200 (1910). 280 Bouknighl and Smith /. Am. Chem. Soc., 81, 28 (1939).

GENERAL SURVEY

from two to one, as in 2AgN< I C4H8S2, toonetotwo, as in AgNi ). -2( ,1 1 > The cation in the former may have the bridge structure

CIU'Il

Ag— S

S— Ag

(II (II

Thioethers and Thiols

Pfeiffer881 has pointed out that the thioethers show a strong tendency to unite with salts of such metals as nickel, copper, and zinc, and, especially with those of platinum and palladium. Diethyl sulfide reacts with plati- num(II) chloride to give three compounds of the empirical formula Pt(SEt-_. (jClj . the yellow a- and 0-isomers being the trans and cis com- pounds, respectively*1, and the y-isomer being the dimer [Pt(SEt2)J [PtCl4]'263. The a- and /3-forms are easily converted into each other by crys- tallization from suitable solvents. The differences between these a- and £- forms are so much greater than is usually shown by cis-trans isomers that Angell, Drew, and Wardlaw concluded that the isomerism is structural rather than spatial264a. They proposed the formulas

(«)

Et2S

...CI SEt2

yPt AND (/S) PL

Et2SN /CI

'SEta

''CI

but Drew and Wyatt2Wb later concluded that the a-salt has the trans struc- ture :

Pt / \

B] •

The great differences in the two isomers may be explained on the basis of the strong trans influence of the coordinated sulfur.

261. Pfeiffer, "Organische Molekulverbindungen," p. 159, Second Edition, Stuttgart,

Enke, 1927.

262. Jensen, Z. anorg. allgem. Chem., 225, 97, 115 (1935).

263. Tschugaeff and Benewolensky, Z. anorg. Chem., 82, 120 (1913); Drew, Preston,

Wardlaw, and Wyatt, ./. Chem. 80c. , 1933, 1294; Cox, Saenger and Wardlaw,

./. ' . 1934, 182.

264a. Angell, Drew, and Wardlaw, ./. Cfo m. Soc., 1930, 349 264b. Drew and Wyatt, ./. Chem. 8oe.t 1934, 56.

CHEMISTRY OF THE COORDINATION COMPOUNDS

The ion 1 1 >t (SEt2)4]++ i-s unstable, and its salts with simpler anions have not been isolated in the solid state. The iodide apparently cannot exist even in solution286. With ions such as [PtCl4]= [PtCl6]= and [Pt(N02)4]=, how- ever, it forms stable, insoluble salts. Upon heating or solution, chloro- platinites of this type frequently rearrange to a mixture of the a- and 0- monomeric forms:

[Pt(SMe2)4][PtCl4] -» 2[Pt(SMe2)2Cl2]263b'c

The chloroplatinate decomposes on heating to give a mixture of [Pt(SEt2)2Cl2] and [Pt(Et2S)2Cl4]263.

Several tetrahalides of the type [Pt(R2S)2X4] are known264, 265- 266. Several of them have been shown to exist in a- and /3-forms, which are readily interconvertible.

Disulfides behave similarly, but occupy two positions in the coordination sphere. The compound

1 S— C

/ Pt

s— c

}H2

yii.2

Cl2

which may serve as an example, cannot exist in a trans form, but /?- and 7- forms analagous to those described above have been prepared. The /3-form reacts with ethylenediamine to give the rather unstable mixed com- pound [Pt es en]Cl2265. Bennett, Mosses, and Statham267 were of the opinion that dithioether complexes of the type [Pt es X2] should exist in racemic and meso forms because of the asymmetry of the donor atoms, but they were unable to isolate the two geometrical isomers. Mann, however,268 resolved a compound containing coordinated sulfur as its center of asym- metry (see page 325).

The dibenzylsulfide complex [Au{S(C7H7)2}Cl2] is noteworthy because its simplest formula suggests the possibility that it may contain gold(II)269.

266. Tflchugaefl and Fraenkel, Compt. rend., 164, 33 (1912).

266 Blomstrand and Weibull, J. prakt. Chem., [2] 38, 352 (1888); Blomstrand and

Enebuske, ./. prakt. Chem., [2] 38, 3G5 (1888); Blomstrand and Rudelius, J. prakt. Chun., [2] 38, 508 (1888); Blomstrand and Londahl, ./. pmkt. Chem., [2] 38, 515 (1888).

267 Bennett, Mosses, and Statham, J, Chem. Soc, 1930, 1668. .v- Mann, ./. Chem. 8oc, 1930, 1746.

269 Herman, />'< r . 38, 2813 (1905) ; Raj and Sen, ./. Tnd. Chew. Soc, 7, 67 (1930).

GENERAL si RVEY

Such is not the case, however, as the substance is diamagnetic270. Prom the molecular weight, electrical conductivity, magnetic susceptibility, and

crystallographic data it is concluded that the substance is a Lattice com- pound containing equivalent amounts of goldi 1 1 and goldi 111), IAihSRoCI]- [Au(SR2)Cl8]270.

[ridium(III)271 and rhodium( 1 1 1 )-"- form the species |M (SK,»:,( 'l:;|. The iridium complex has been separated into its isomeric forms. The anionic complex jlnSR jU'l,] has also been prepared*78. Surprisingly, treatment of these complexes with amines results in the replacement of the thioether groups first27*.

Livingstone and Plowman274 have prepared soma halogen bridged com- plexes of 0-methylmercaptobenzoic acid which contain different metal ions.

(M = Hg^rCu11).

Most of the remarkable hexadentate chelating agents of Dwyer and Lions (Chapter 8) contain two coordinating sulfur atoms. A fine demonstra- tion of the much greater affinity of cobalt (III) for ether-type sulfur than for ether-type oxygen is found in the fact that so long as one sulfur atom is present, the complexes are resolvable into optical isomers, while substitu- tion of oxygen atoms for both sulfurs leads to cobalt (III) complexes which are too unstable to resolve275.

Gonick, Fernelius, and Douglas276 determined the formation constants of ties of sulfur and nitrogen containing chelating agents with the ions of copper, nickel, cobalt, zinc, and silver. A comparison of the data with similar data for a series of analagous polyamines indicated that nitrogen is prob- ably a stronger donor for the metals studied, except silver. However, 2-aminoethanethiol, which coordinates as a negative ion, forms the most -table complexes of the entire group.

27(1 971

Brain, Gibson, Jarvis, Phillips,. Powell, and Tyabji, ./. ('hem. Soc, 1952, 30S6. Ray and Adhikari, ./. Ind. Chem. Soc, 9, 251 (1932); Ray, Adhikari, and Ghosh,

./. Ind. Chem. Soc., 10, 279 1933). Dwyer and Nyholm, ./. Proc. Roy. Soc. N.S. Wales, 78, 67 194 l . Ray and Ghosh, ./. Ind. Chem. Soc., 13, 138 (1936); Ray, Adhikari. and Ghosh, ./. Ind. Chem. Soc., 10, 27.", L933 ; Ray and Adhikari, ./. Ind. Chem. Soc., 11, •",17 L934 ; Lebedinskii and Gurin, Compt. "ml. aca4. set. U.R.S.S., 40, 322 L943 . Livingstone and Plowman, J. Proc !:■■ Sen V.S. Wales, 86, 116 1962). 275. Dwyer, Gill, Gyarfas, and Lions. ./. Am. Chem. Soc., 75, 1526 1963

.'.nick, Fernelius, and Douglas, Technical Report to O.N.R., Oct. 16, 1963.

1,1 273

274.

CHEMISTRY OF THE COORDINATION COMPOUNDS

In addition to the marked stability of complexes containing the negative mercapt ide ion toward dissociation in solution, stabilization of the ligand or of a high oxidation state of the metal may occur. Thus, the complex

P=C — NH-/ V

WS — CH;

stabilizes the ligand toward oxidation (when uncomplexed it is rapidly oxi- dized by air to the disulfide) and at the same time stabilizes the strongly oxidizing cobalt (III) species277. The great specificity of the metal ion in this behavior is illustrated by comparison with the reaction of thioglycolic acid and iron ions. In air-free alkaline solution, the complex

is formed. Air oxidizes the iron(II) ion to iron(III) which in turn catalyzes the oxidation of the ligand to the corresponding disulfide278.

Gold complexes of a-thiol fatty acids may prove useful in the treatment of such maladies as tuberculosis and leprosy279. The complex formed from diethylgold monobromide and 2-aminoethanethiol is also of interest. From its molecular weight it is assigned the structure

NH2CH2

CH,

However, the compound is remarkable in that the coordinated sulfur atom is quite reactive. The compound reacts explosively with methyl iodide and more moderately with ethyl bromide. The picrate salt of the product of treat incut with ethyl bromide was shown to be identical with the complex prepared from S-ethyl-2-aininoethanethior280.

J77 Feigl, Nature, 161, 435 (1948); Anal. Chem., 21, 1298 (1949).

278. Leussing and Kolthoff, •/. Am. Chem. Soc, 75, 3904 (1953).

279 Kundu, J. Ind. Chem. Nor., 29, 592 (1952).

280 Ewenfl and Gibeon, ./. Chem. Soc, 1949, 431.

QE VERAL SURY/) 53

The ability of seleno- and telluromercaptides and ethers to form com- plexes similar to those of the sulfur analogs is illustrated by the mercury(II) halide complexes used in the characterization of these donor molecules281. Gould and McCulloughlM have expressed the opinion thai diarylselenoxides

coordinate to mercury < 1 1 1 through the selenium atom.

Thio<arl>oii\ I ( .001 dimit ion

Many thiocarbony] compounds show strong donor properties. Among the

simplest of these is thiourea, which coordinates through the sulfur rather than through nitrogen, thus occupying only one coordination position. Thiourea coordinates with salts of almost all of the heavy metals. With compounds of tripositive iridium288 and rhodium284 it forms whole series of compounds, such as [Ir tu3Cl3], [Ir tu4Cl2]Cl, [Ir tu5Cl]Cl2 , and [Ir tu6]Cl3 . Thiourea reacts with the cis and trans isomers of platinum(II) eompounds of the type [Pt a>X2] yielding different products and in so doing serves as the basis of Kurnakov's test which is widely used to distinguish I jet ween the isomers285. (Chapter 9) Another interesting application of thiourea to the chemical determination of structure is found in the work of Gent and Gibson288 with the dimeric [Et-jAu SCX]2 . The failure of the complex to react with such nitrogen bases as ammonia, dipyridyl, and ethrylenediamine, and its reaction with thiourea to produce [Et2Au(SCX)tu] is interpreted to mean that the thiocyanate is coordinated through the sulfur, and that the original compound has the structure

I E1 S Et

\ /\ / Au Au

/ \/ \

Et S Et

Jensen has studied the compounds formed between thiosemicarbazide

281. Morgan and Burst all, ./. Chen . 8oe.t 1929, 1096; 1930, 1497; 1931, 173; Carr and rson, /. Chem. Soc., 1988, 282; Kraffl and Lyons, Ber., 27, 176] (1894 , Gould and McCullough, J.Am. Chem. Soc, 73. 3195 [1961). Lebedinskii, Shapiro, and Kasatkina, Ann. inst. platine, t'.s.s.ir, No. 12, 93 L935). 284. Lebedinskii and Volkov, An,,, inst. plain,, . U.S.S.R., No. 12, 79 I"

Kurnakov, •/. Ruse. Phye. Chem. Soc., 25, 565 1893 ; <■<. Chem. Centr.t 65, I. WW 18Q • nt and Gibson, ./. Chen 8oe.t 1949, L835.

CHEMISTRY OF THE COORDINATION COMPOUNDS

and platinum(II)287, palladium (II)287, and nickel288 ions. The thiosemicar- bazide molecule occupies two coordination positions, evidently coordinating thus

NHS

-NH

I C— NH2

Upon the addition of thiosemicarbazide, potassium chloropalladate(II) first gives [Pd thio2][PdCl4] and then [Pd thio2]Cl2 . If this latter compound is heated in weakly acid solution, it changes to the insoluble inner complex

NH— NH2 NH2— NH

/ \ / \

HN=C Pd O

=NH

There is evidence that this exists in two, presumably cis-trans-, forms. The platinum and nickel compounds behave similarly.

Diketonedithiosemicarbazone (thiazone) and its homologs

S S

NH2— C— NHN=CR— C R'=NNH— C-

-NHj

act as tetradentate ligands forming inner complexes with copper(II) and nickel(II) ions

N N

HN M XNH

I A I

HN=C— S S— C=NH

These complexes are quite stable, dissolving in strong acid as the soluble Baits, [M(thiazone)]X

Ammonium dithiocarbazide reacts with platinum(II) in a manner com- parable both to thiourea and to thiosemicarbazide. With /rfl/^-[Pt(NH3)2Cl2]

_'s7 Jensen, /. anorg. allgem. Chem., 221, 6 (1934

288. Jensen, /. anorg. allgem. Chem., 221, 11 (1934).

289 B&hr and Hess, / anorg. allgem. Chem., 268, 351 (1952).

ai:\/:ir\L survey

r>-)

the reaction is

II 2S— ('— XIINH. + frans-[Pt(NH,),Cl,]

NH,NHC— S NH,

S— C— XI I. Mi- ll S

+ 2C1"

+ 2NH3 + 2C1-

X 1 1

while the cis-isomer undergoes the reaction

II 2S— C— XHXH7 c7S-[Pt(XH3)2Cl2] ->

S S

II II

cs sc

/ \ / \

HX Pt NH

\ / \ /

NH2 NH2

From these reactions and the fact that tetrammineplatinum(II) ion is not attacked by the dithiocarbazide ion, it is concluded that the sulfur groups may displace chloride rapidly but that the ammonia is displaced only as a consequence of the trans influence of the coordinated sulfur290.

Inner complexes are formed by eobalt(II), cobalt (III), nickel(II), and palladium(II) with thiodicyandiamidine (guanyl thiourea),

NH S

II II

H,X( '— XII- CXH2.

Copper differs by forming a complex of the type [Cu(thicy) SO,|. The co- balt, copper, and palladium complexes decompose in warm alkali, deposit- ing insoluble' metal sulfides, thus providing evidence for the participation of

290. Chernyaev and Mashentsev, Izvest. Sektora Platiny i Drugikh Blagorod. Metal., Inst. Obshchei i Neorg. Khun., Akad. Nauk 8.S.S.R., 23, 72 1949); cf. Chem. Abs. 45, 2812d L951); Mashentsev and Chernyaev, Doklady Akad. Nauk

- S S R . 79, 803 1951); cf. Chrm. Ah*. 46, 2!»40g (1052).

CHEMISTRY OF THE COORDINATION COMPOUNDS

the sulfur atom in coordination (A). The nickel complex fails to give this tesl and is thought to have structure B291.

s\ /NH

HN Ni NH

X — UH^ XNH— c(

HN ^NH

>-S /NH2-CX

HN Pd .NH

>-nh/ xs-<

HNC XNH

(B)

The occurrence of nickel (IV) in sulfur complexes testifies to the great tendency of that donor to form strong covalent bonds. Hieber and Briick292 found that air oxidation of a strongly alkaline suspension of the nickel(II) complex of o-aminothiophenol produces the deep blue complex

CEh^^'^XD

A similar bridged disulfo compound is formed by dithiobenzoic acid

Dithiooxamide (rubeanic acid) forms insoluble complexes with nickel, and copper ions293. These substances have the properties of inner salts, and be- cause of the steric requirements of the ligand, they exist as bridged polymers

> x /\ i. /\

'NH

NH* ^S NH' NS x

Anion and Kane-"" have used the linear nature and the light absorption of this polymer in the manufacture of a device for the polarization of light. A sheet of plastic is soaked in a solution of dithiooxamide, which causes the precipitation of the complex within the plastic. When the plastic sheet is

291. Ray and Chaudhury, J. Ind. Chem. Soc., 27,673 (1950); Poddar and Ray, /. Ind. Chem. Soc., 29, 279 (1952).

Hieber and Bruck, Naturwias., 36, 312 L949).

Jensen, Z. anorg. Chem., 262, •_,--,7 (1944 ; Ray, Z. anal. Chem., 79, 95 (1929). 294. Anion and Kane, U.S. Patent 2 505 085, April 25, 1950.

GENERAL SURVE1 57

Stretched ill one direction, the polymer chains arc oriented parallel to each other. The bridging ability of tins donor molecule is also illustrated in the dimeric derivative of diethylgold monobromide

E^A- Au

Bt^ X

^Et

Other Sulfur Donors

The thioeyanate ion has unshared pairs of electrons on both the sulfur and the nitrogen. Werner at one time296 supposed the two isomers of [Co enj(NCS)2]+ to be structurally different, one having a cobalt-nitrogen link and the other a cobalt-sulfur link. This hypothesis was based upon the fact that the thioeyanate group of one of the isomers is destroyed by chlorine, leaving the nitrogen (in the form of ammonia) in union with the metal, while the thioeyanate group of the other isomer is completely elim- inated by this treatment. Werner later found'297, however, that the two com- pounds are stereoisomers, and that the thioeyanate group is attached to the metal through the nitrogen in both cases. The sulfur of the thioeyanate group probably does have strong donor properties, however, and in the case of gold it is the sulfur atom which preferentially coordinates Werner reported that silver nitrate does not precipitate silver thioeyanate from a solution of [Co(XH3)s(XCS)]++ or similar complexes but the silver loses its ionic character. He supposed that the silver coordinates with the sul- fur2'^. Waggener, Mattern, and Cartlcdge299 however, have found the stability of these dinuclear complexes to be much less than reported by Werner.

The sulfite ^roup evidently occupies only one coordination position in most cases, and from the fact that salts of the ion [Co(XH3)4(S03)2]_ are yellow or brown, it may be inferred that these compounds contain a sulfur- cobalt link.

The action of sulfite ion on platinum(II) complexes is also most easily explained on the basis of a metal-sulfur bond. Sulfite acts differently800 on

396. I . .' era and ( iibaon, •/ . ( 'hi n . Soc.s 1949, 131 .

Werner and Braunlieh, Z. anorg. Chetn., 22, 91, 123 (1900).

297. Werner. .1„„..386, 1 L912).

298. Werner, Ann., 386, 50 (1912).

Waggener, Mattern. and Cart ledge, al>st racl 8, [22nd meeting, American Chemi- Sepl . 1962. 300. Gurin, Doklady Akad. Nauk S.S.S./r. 50, Jul L946).

58 CHEMISTRY OF THE COORDINATION COMPOUNDS

the cis- and trans- isomers of dichlorodiammineplatinum(II)

cis [Pt(NH,)2CU] + l\a,S();; -> Xa6[Pt(S08)4] + 2NaCl + 2NH3 «rans-[Pt(NH,)8Cla] + 2Na2S08 -* *mns-Na2[Pt(NH3)2(S03)2] + 2NaCl

This behavior is quite similar to the reaction of these isomers with thiourea.

Aside from the complexes with aromatic nitrogen molecules, ruthe- nium^ I) is besl known in its very unusual sulfite complexes. Treat- ment of cWoropentammineruthenium(III) ion with sodium bisulfite produces the two complex compounds, [RuIT(NH3)4(S03H)2] and \a ,| Etun< X HsMSOaMSOsH)*,] -6H20. The dipositive oxidation state of the ruthenium was verified by analysis and magnetic measurements301. Upon dissolution in acid, [RuII(NH3)4(S03H)2] is converted to [RuII(NH3)4(S02)X]X. The action of ammonium hydroxide on the dibisul- ntotetrammine produces the nonelectrolyte, [RuII(NH3)5(S03)]302. This compound is also sensitive to acid, transforming to [RuII(NH3)5(S02)]++.

Rhodium(III) and iridium(III) form complexes of the type MI3[MIII(NH3)3(S03)3]303. Iridium also forms a compound in which the sulfite group is reported to be bidentate, [Ir(S03)3Cl2]5~, but the alternate possibility of halogen bridging has not been disproved.

Riley304 has prepared salts of the dark red selenitopentamminecobalt(III) ion, [Co(XH3)5(Se03)]+, but his experiments did not show whether the selenite group is attached through the selenium or through the oxygen. Several selenite complexes of nickel, copper, and cobalt have been obtained by Ray and Ghosh305, who found them to be less stable than the correspond- ing sulfite compounds.

The thiosulfate group, with unshared electrons on both oxygen and sul- fur, could conceivably coordinate through either or both. When it occupies but one coordination position, union with the metal evidently takes place through the oxygen, for the ion [Co(NH3)5S203]+ is red306. This ion is very stable, for it is formed when [Co(NH3)5Cl]S203 or [Co(NH3)5Br]S203 is allowed to stand at 35 to 40°C307. The stability of the thiosulfate-cobalt bond is further attested by the reaction of [Co(NH3)5S203]+ with potassium cyanide, which yields K4[Co(S203)(CN)5].308 Duff167 reported the prepara-

301. Gleu, Breuel, and Rehm, Z. anorg. allgem. Chem., 235, 201 (1938).

302. Gleu and Breuel, Z. anorg. allgem. Chem., 235, 211 (1938).

303 Lebedinskii and Shenderetskaya, Izvest. Sektora. Platiny i Drugikh Blagorod.

Metal., hist. Obehchei i Neorg. Khim. Akad. Nauk S.S.S.R., 21, 164 (1948); cf.

Chem. Abe., 44, L0566a (1950); Gurin, Doklady Akad. Nauk S.S.S.R., 56, 217 1936); of. Chem. Abe. 43, 1676a (1949). 304. Riley, ./. Chem. Soc, 1928, 2985.

306 Raj and Ghosh, ./. Indian Chem. Soc., 13, 494 (1936). Ray, •/. Indian Chem. Soc., 4, 64 (1927).

307 Sarkar and Daa Gupta, J. Ind. Chun. Soc, 7, 835 (1930).

308 Ray, ./. Ind. Chem. Soc, 4, 325 (1927).

GENERAL SURVEY 59

tion of [Co enj S.j().;]Br-3II2(), which he thought contained a doubly co- ordinated thiosulfate group. The evidence for this is Blight, however, and

the correct formula may well be [Co en2 (H,( ))S,( ):;]Br-L>I I,( ). YVeinlaiaP'' has suggested, hut without experimental evidence, that the double potas- sium bismuth thiosulfate is

jH20

in which coordination takes place through both oxygen and sulfur. The fixation process in photography depends upon the formation of thiosulfato- silver anions, of which several have been reported310. If coordination takes place through the oxygen, sulfates should give analagous compounds.

The Doxor Properties of Nitrogen

The solvent properties of ammonia closely resemble those of water, and solvation is as important in ammonia solutions as it is in aqueous solutions. The donor properties of nitrogen are as strong, or stronger, than those of oxygen, and some of the metal-ammonia compounds show remarkable stability. Many of them (including those of cobalt, chromium and the platinum metals) do not lose ammonia when heated above 200°C or wrhen treated with sodium hydroxide or hydrochloric acid. The ammines of copper, silver, zinc, and several other metals are equally well known, but are much less stable, and are decomposed by dilute acids or bases. Ammines of the alkali and alkaline earth metals are completely decomposed by water, and some of them are stable only at low temperatures.

Ammines

The hydrates, especially those of the highly charged metallic ions, readily liberate hydrogen ions, with the formation of aquohydroxo complexes. An analagous reaction takes place with ammines, but it is less pronounced than with hydrates. From a study of the ammines of rhodium, Griinberg and rmanir11 concluded that the acid dissociation of coordinated water is 105 times as great as that of coordinated ammonia. The loss of protons by ammines is particularly noticeable with the complexes of the very heavy metal>, as is illustrated by the formation of HgXH-jCl when mercuric chlo- ride is treated with ammonia. Other illustrations involve the ammines of

309. Weinland, "Einfuhrung in die Chemie der Komplexverbindungen," Second Ed.,

p. 148, Stuttgart, Enke, 1924.

310. Bassett and Lemon,./. Chem. Soc.j 1933, 112:;; Aflhihara and Mateuda, Kogaku

8huho, Kyushu Univ. (Technological Reports, Kyushu Univ.), 25, 11 (1952); cf. Chem. Abs., 47, 12075g (1953).

CHEMISTRY OF THE COORDINATION COMPOUNDS

Table 1.1. Colors of Some Anhydrous Salts, and Their Hydrates

and ammonates

CoCl2

[Co(H20)6]Cl2

[Co(NH3)6]Cl2

Blue

Red

Rose

CuCl2

[Cu(H20)4]Cl2

[Cu(NH3)4]Cl2

Brown

Blue

Deep Blue

NiCl2

[Ni(H20)6]Cl2

[Ni(NH3)6]Cl2

Light brown

Green

Blue

CrCl3

[Cr(H20)6]Cl3

[Cr(NH3)6]Cl3

Violet

Gray-violet

Yellow

[Cr(H20)5Cl]Cl2

H20

[Cr(NH3)5Cl]Cl2

Green

Rose red

[Cr(H20)4Cl2]Cl

•2H20

[Cr(NH3)4Cl2]Cl

Green

cis- violet trans- green

platinum311, 312, gold313, and osmium314. Ammines of the lighter elements also lose protons to some extent, as is indicated by the fact that the hydro- gen atoms in such complexes as [Co(NH3)6]"HH" are readily exchanged for deuterium when placed in heavy water315.

Water and ammonia, coordinated to ions of the same metal, do not al- ways stabilize the same valence state (Chapter 11). For example, hydrated cobalt (III) compounds are very strong oxidizing agents, while ammoniated cobalt (II) compounds are strong reducing agents. The hydrates and am- mines often show similar colors, but this is by no means a general rule. Table 1.1 summarizes a few examples. Peters316 made the first systematic and extended study of the stability of ammines. He subjected ninety seven salts to the action of dry ammonia gas at atmospheric pressure and by measuring the volume of ammonia absorbed in each case, calculated the formulas of the ammines obtained. Following Peters, Ephraim317, W. Biltz318, Clark319, 320, and others studied the reactions of salts with anhydrous

311. Grunberg, Z. anorg. Chem., 138, 333 (1924); Griinberg and Faermann, ibid., 193,

193 (1930).

312. Tschugaeff, Z. anorg. Chem., 137, 1 (1924).

313. Block and Bailar, J. Am. Chem. Soc., 73, 4722 (1951).

314. Dwyer and Hogarth, J. Am. Chem. Soc., 75, 1008 (1953).

315. Anderson, Briscol, and Spoor, J. Chem. Soc., 1943, 361.

316. Peters, Zeit anorg. Chem., 77, 137 (1912).

317. Ephraim, Z. phys. Chem., 81, 513, 539 (1913); 83, 196 (1913); Ber., 45, 1322 (1912);

46, 3103, 3742 (1913); 47, 1828 (1914); 48, 41, 624, 629, 1638, 1770 (1915); 49, 2007 (1916); 50, 529, 1069, 1088 (1917); 51, 130, 644, 706 (1918); 52, 236, 241, 940, 957 (1919); 53, 548 (1920); 54, 973 (1921).

318. Biltz and co-workers, Z. phys. Chem., 82, 688 (1913); Z. anorg. allgem. Chem., 83.

L63, 177 (1913); 89, 97, 134, 141 (1914); 109, 89, 132 (1919); 114, 161, 174, 241 1920); 119, 97, 115 (1921); 123, 31 (1922); 124, 235, 322 (1922); 125, 269 (1922); 127, 1 (1923); 129, 1, 161 (1923); 130, 93 (1923); Z. Elektrochem., 26, 374 (1920); Angew. Chem., 33, 313 (1920).

GENERAL SURVEY 61

ammonia. They prepared and studied hundreds of amminea in order to find

out what factors arc important in determining stability. While a great deal was learned about the stabilities of ammines, little light was thrown on the structures of such compounds as A1C11:; ■ *)X 1 1:; , AlCl3-5NH.: and AlClj- INHj820 and compounds containing very large amounts of ammonia, such as T1C1 r -OX 1 1/-1. Doubtless many of these are "lattice compounds" only.

The ammines which are of chief interest are those of the transil lob metals and the metals of periodic groups IB and IIB. Even among these, there are great differences in stability. For example, iron ammines cannot be ob- tained in the presence of water; copper ammines and cobalt(II) ammines exist in water solution, and can be crystallized from such solutions, but they are immediately destroyed by acids. Cobalt (III) and platinum ammines can be recrystallized from solutions of strong acids, and the hydroxides [Co(XH3)6](OH)3 and [Pt(XH3)6](OH)4 are sufficiently stable to allow their easy preparation322. This, of course, may be a measure of rate of decomposi- tion rather than of intrinsic stability, but it is of tremendous practical im- portance.

The nature of the anion is of great importance in determining the stability of some metal ammines. Weitz323 observed that the ammines of gold are stable if the anion is an oxy-anion such as nitrate, perchlorate, phosphate or oxalate, and the ammonia groups cannot be removed by the action of the oxyacids. They are destroyed, howrever, by halides, presumably because the halide ion replaces part of the ammonia in the coordination sphere. Tomlin- son, Ottoson, and Audrieth324 have called attention to the explosive charac- ter of cobalt (III) and chromium(III) ammines in which oxidizing groups are present in the coordination sphere or as anions.

It is of interest that the ammines which are easily decomposed by acids (e.g., those of Cu, Ag, and Zn) are easily formed by the addition of am- monia to a solution of the metal ion. The ammines which are not rapidly destroyed by acids are not readily formed. Thus, the addition of an excess of ammonia to a solution of a chromium(III) salt ordinarily precipitates the hydroxide; the hexammine is formed in good yield only by the action of liquid ammonia on anhydrous chromium(III) chloride in the presence of a catalyst325. The hexammine cobalt (III) ion is not obtained by aerial oxida-

319. Clark, Quick, and Harkins, ./. Am. ('hem. Soc, 42, 2438 (1920); Clark and Buck-

ner, J. Am. Chem. Soc, 44, 230 (1922).

320. Clark, An,. ./. Set., 7, 1 (1924).

321. Young,/. Am. Chem. Soc, 57, 997 (1935).

322. Hecht, Z. anorg. aUgem. Chem., 270, 215 (1952).

323. Weitz. Ann., 410, 117 1915).

324. Tomlinson, Ottoson, and Audrieth, ./. .1///. Chi m. Soc, 71, 375 (1949).

325. Oppegard and Bailar, Inorganic Syntheses, III, 153 (1950).

CHEMISTRY OF THE COORDINATION COMPOUNDS

tion of an ammoniacal eobalt(II) solution except in the presence of a catalyst826. Dwyer and Hogarth327 could prepare the ion [Os(NH3)6]+++ only by the treatment of [Os(\]IH)5Br]++ with ammonia under pressure.

Ammonia can, of course, share the coordination sphere with other donor groups. In his first paper116, Werner pointed out that ammonia molecules can be displaced, one by one, from the coordination sphere, either by other neutral groups such as water, or by negative groups. If the metal-ammonia bond is stable1, the groups which share the coordination sphere with ammonia may be replaced by other groups to form a great variety of compounds. The following reactions are typical327:

HCl

[Os(NH3)5Br]++ Ag2° . [Os(NH3)6OH]++ > [Os(NH3)5Cl]++

H20

The amide group, like the hydroxide group, has two pairs of unshared electrons and coordinates readily with certain metals. Mercury amido chlo- ride illustrates this. The NH2~ group can also act as a bridge between two acceptor atoms (p. 23). The imino group frequently acts as a bridge also, as in

NH III/ \IV

en2 Co Co en2

\ /

and

(NH8)8PtI

<">:

PtI(NH3)s

Aliphatic Amines

The aliphatic monoamines coordinate less readily than does ammonia, and the compounds so formed are less stable than the ammines. However, this p<»int is often overemphasized, for some rather stable coordination compounds of the aliphatic amines do exist. The secondary amines co- ordinate less readily than do the primary, and the tertiary amines are al- in<»t devoid of ability to coordinate with metal ions. This is probably due

326 Bailar and Work, ./. Am. Chem. Soc, 67, 176 (1945).

327. Dwyer and Hogarth, ./. Proc. Roy. Soc. N.S. Wales, 84, 117 (1951).

328 Werner, Ann., 375, 74 (1910).

Ofven. K. Vet. Akad. Fork., 27 , 777 (1870); 28, 175 (1871).

GENERAL SURVEY 63

tosteric factors, for the tertiary amines coordinate (irmly with the hydrogen ion; that is, they arc strong bases. Straumanis and Circulis880 have de- scribed compounds of the mercury and copper halides with ethylamine, propylamine, butylamine, dimethylamine, and diethylamine. Jorgensen881 prepared platinum(II) complexes containing methyl, ethyl, and propyl- amines, and Drew and Tress332 have extended his study to include the preparation of the stereoisomers forms of [Pt(CHsNHs)£)ls]. These are Btable enough thai they can be oxidized to |Pt((1H3NH2)2Cl4]. Gil'denger- shel333 prepared [PtCCHsNI^iClsJCli by the action of methylamine on potassium chloroplatinate, and purified it by recrystallization from hydro- chloric acid. Chernyaev334 has prepared three of the four possible isomers of [[Pt en(CHsNH2)(NOi)CyClJ and has resolved one of them, as well as [Pt en(CHsNH2)(NOi)tCl]Cl. Finally, Meisenheimer and Kiderlen335 have introduced various primary amines into the coordination sphere of cobalt by the reaction

[Co en2Clo]Cl + amine — » [Co en2 amine Cl]Cl2

Even aromatic amines form fairly stable compounds in this way. Primary ■nines which are weaker bases than aniline, and secondary amines, do not enter the complex, but bring about more complicated reactions336337.

If chelation can take place to form five-membered rings, the stability of the compounds is greatly enhanced (Chapter 5). Ethylenediamine is the simplest and the most important of such bases, and its compounds have played an important part in the development of the coordination theory. 1,2-Diaminopropane (propylenediamine) also forms stable compounds, which are similar to those containing ethylenediamine, but are usually more soluble. Isobutylenediamine338, 2,3-diaminobutane339, stilbenedi- amine33S- 34°, and several other 1,2-diamines have been shown to form stable chelate rings. Pearson, Boston, and Basolo341 have prepared com-

130. Straumanis and Circulis, Z. anorg. allgem. Chem., 230, 65 (1936).

531. Jorgensen, J. prakt. Chem., 33, 530 (1886).

B2. Drew and Tress, ./. Chem. Soc., 1935, 1212.

m. Gil'dengershel, Zhur. Priklad. Khim. (J. Applied Chem.), 23, 487 (1950).

534. Chernyaev, Ann. inst. platine No. 8, 37 (1931).

535. Meisenheimer and Kiderlen, Ann., 438, 238 (1924).

136. Ablov, Bull. soc. chim., [5] 3, 2270 (1936); 4, 1783 (1937).

537. Bailar and Clapp, ./. Am. Chun. Soc., 67, 171 (1945).

B. Mills and Quibell, ./. Chem. Sue, 1935, 839; Lidstone and Mills, ./. Chem. Soc,

1939, 1764. m. Bailar and Balthie, J. Am. Chem. Soc., 68, L474 (19 14i). Williams, thesis, I'nivcrsit y of Illinois, 1961. Ml. Pearson, Boston and Basolo. J. Am. Chem. Soc, 76, 3089 (1963

64 CHEMISTRY OF THE COORDINATION COMPOUNDS

pounds of the type

CI;

in which the R's represent hydrogen or methyl. As the number of methyl groups is iii< reased, crowding becomes pronounced, and, in water solution, the coordinated chlorides are more easily replaced by water molecules.

Trimethylenediamine forms six-membered rings, which compare favor- ably in stability with those of ethylenediamine297 • 342. Mann343 has prepared coordination compounds of several metals with bases of the type (XIf2CH2)2CHX, where X = CH3 , Br, SCN, and OH. Tetramethylene- diamine and the higher homologs in the series apparently cannot form rings at all in aqueous solution. Diamines having four, five, ten and eighteen carbon atoms have been investigated344. Pfeiffer345 has shown, however, that tetramethylenediamine and hexamethylenediamine will form chelates from alcohol solution.

The polyamines NH2CH2CH2(NHCH2CH2)nNH2 (n = 1, 2, 3, or 4) are strong coordinators, (even though part of the nitrogen atoms are second- ary), because they form multiple ring systems. Diethylenetriamine acts as a tridentate base toward copper(II) and nickel(II) ions, giving complexes of the types [Cu dien Cl]+ and [Cu dien2]++. In the second case, because of the stereochemical properties of the base, copper assumes a coordination number of six346347. Jonassen and his students prepared platinum and palladium-triethylenetetramine complexes [Pt trien]++ and [Pd trien]++348, and [Xij trion3]4+. The [Ni2 trien3]4+ is paramagnetic, so must consist of two tetrahedra849. Basolo350 prepared a series of cobalt complexes of the types

342. 343. 344.

346. 346. 347. 348.

Bailar and Work, J. Am. Chem. Soc, 68, 232 (1946).

Mann, ./. Chem. Soc, 1927, 2904; 1928, 1261.

Pfeiffer and Haimann, Ber., 36, 1063 (1903); Pfeiffer and Lubbe, ./. prakt. Chem!

[2] 136, 321 (1933); Tschugaeff, Ber., 39, 3190 (1906); TschiiKaefT, ./• prakt.

Chem., [2] 75, 159 (1907); Werner, Ber., 40, 61 (1907); McReynolds, thesis,

University of Illinois, L938. Pfeiffer, Naturwiss., 36, 190 (1948). Mann, ./. Chem. Soc, 1934, 466. Breckenridge, <'<n,a</nin ./. Research, 26B, 11 (1948). Jonassen and Cull, ./. Am. Chem. Soc., 71, 1097 (191!)). .Joniisscii and Douglas, J. Am. Chem. Soc, 71, 1091 (1919). Basolo, J\ .1///. Chem. Soc, 70, 2634 .1948).

GENERAL SURVEY 65

[Co trien \..| and [Co trien Y], where X is CI, NOs and Ml, and Y is ( '< ». or en. He also obtained [C02 triens]^", an ion of unusually high ionic charge. Jonassen and Fry351 have isolated the cobalt(II) complex of tetraethylene- pentamine.

(SjjS'^^-Triaminotriethylamine behaves as a quadridentate amine in spite of the reluctance of tertiary nitrogen to coordinate. Mann and Pope852 prepared the platinum(II) and platinum(IV) complexes |Pt tren]Cla and [Pt tren C12]C12 . The palladium(II) and nickel ions form the ion |M tren]++ and nickel forms also the ion [Nij tren;1]1+, in which the coordination number o\ nickel is evidently six388. Mann884 prepared several salts of the ion [Co tren(SCN)2]+. By treatment of [Co enJ(,l..|+ with the same base, Jaeger and Koets355 obtained salts of an ion which they thought to be [(Co ei^trer^]94", but at tempts to repeat this work356 have been unsuccessful, and it seems that Jaeger and Koets probably had [Co tren en]+++.

Cases are known in which the polyamines coordinate without using all of their nitrogen atoms357. a,jS,7-Triaminopropane can act either as a bi- dentate or tridentate group358 depending upon the metal ion involved and the conditions of the experiment. If only two amino groups coordinate, they are on adjacent carbon atoms.

Ethylenediamine, and presumably other, similar bases, sometimes co- ordinate through only one nitrogen. Chernyaev and Fedorova359 prepared a compound whose formula they write [Pt(en-HCl) 'NHj-Clj]. Mild alkalies close the ring with the formation of [Pt enXH3Cl]Cl, and chlorine oxidizes the compound to [Pt(enHCl)(XH3)Cl4]. This platinum(IY) compound hy- drolyzes to [(XH:5)(H20)Cl3PtenPtCl3(H20)(XH3)]Cl2 , in which the ethyl- enediamine acts as a bridge between the two platinum atoms. Job360 has adduced evidence for the existence of [Ag en2]+ and [Tl en]+ ions, which are kalagous to [Ag(NH3)2]+ and [T1(XH3)]+, and hence contain monoco- Irdinated ethylenediamine. Di-n-propylgold(III) bromide reacts with

351. Jonassen and Fry, ./. .1///. Chem. Soc., 75, 1524 (1953).

352. Mann and Pope, Proc. Roy. Soc. London, 109A, 444 (1925). :-;:>:;. Mann and Pope ./ . Chem. Soc., 1926, 482.

354. Mann. ./. Chem. Sue, 1929, 40!).

■5. Jaeger and Koets, Z. anorg. allgem. Chem., 170, 347 (1928).

■6. Middleton 1952) and Rebertus (1964), unpublished work, University of Illinois.

357. Mann. ./ . Chem. 80c., 1934, 466; Job and Brigando, Compt. rend., 210, 138 L940).

■B. Mann and Pope, ./. Chem. Soc, 1926, 2675; Nature, 119, 351 (1927); Mann, ./.

Chem. Soc., 1926, 2681; 1927, 1224; 1928, 890; 1929, 651.

hernyaev and Fedorova, Ann. secteur platine, Inst. chim. gen. (U.S.S.R.),

No. 14, 9 (1937). 360. Job, ( nd., 176, 4 12 1923) ; 184, 1066 (1927).

.ill

CHEMISTRY OF THE COORDINATION COMPOUNDS

ethylenediamine to form a compound which is formulated Pr Br Br Pr 361-

NH2 — CH2 — CH2- Ml

On heating, one of the gold atoms loses its two propyl groups, retaining its hold on the bromine and the nitrogen:

NH2CH2CH2NH2

Au .

The treatment of Zeise's salt, K[Pt(C2H4)Cl3], with ethylenediamine re- sults in the formation of a dinuclear complex in which the ethylenediamine acts as a bridging group362

C2H4

/ \

Cl NH2CH2CH2NHS

C2H4

Gilman and Woods363 have prepared a compound which they believe to have the structure (CH3)3AuNH2CH2CH2NH2Au(CH3)3 . In this case ring forma- tion is impossible because only one coordination position is open on each gold atom.

Pfeiffer and Glaser364 have studied the donor properties of X-substituted ethylenediamines. With copper(II) perchlorate, N-methyl and X,X'-di- ethylethylenediamine give blue-violet compounds analgous to [Cu en2] (C104)2 . The corresponding N-diethyl compound is ruby red at room temperature, but assumes the blue-violet color above 44°. The same in- vestigators report that the reaction of X-methyl-X'-diethyl etlrylenedi- amine and N-triethyl ethylenediamine with copper perchlorate do not give compounds which are analagous to those of the less highly substituted bases, but correspond to the formula [Cu OH diamine]C104 . They are probably dimeric, the copper atoms being linked together through two ol bridges. These compounds, like the others, are thermochromic, changing from blue-violet to ruby-red when they are cooled in liquid air.

361. Burawoy and Gibson, ./. Chem. Soc., 1935, 210; Burawoy, Gibson, and Holt.

./. Chem. Soc, 1935, 1024.

362. Hel'man, Compt. rend. acad. set. U.R.S.S., 38, 243 (1043).

363. Gilman and Woods, •/. Am. Chem. Soc, 70, 550 (1948).

364. Pfeiffer and Glaser,/. prakt. Chem., 12], 161, 134 (1938); 153, 300 (1939).

GENERAL SURVEY 67

The remarkably stable tris-(N-hydroxethylethylenediamine)cobal1 (IIIj complex ""' shows none of the characteristic read ions of aliphatic hydroxy] ■roups, even though the usual formulation would indicate that the hydroxy]

groups are not coordinated to the metal.

Aromatic Vmines

Aromatic diamine- form quite unstable coordination compounds. Hieber

and his co-workers have shown that ortho-phenylenediamine usually occu- pies only one coordination position366 but that the para isomer occupies two367. They give the latter the rather improbable formula

MI.— R— XH2

/ \

X*M MX.

" \ /

XHo— R— XH2

Diamino-biphenyls seem to have somewhat stronger donor properties. 2,2'-Diamino-biphenyl forms cobalt (III) complexes corresponding to those of ethylenediamine368 and several stable compounds of benzidine and tolidine have been reported369. The empirical formulas indicate that these bases occupy two coordination positions, but there is no evidence that both amino groups attach themselves to the same metal atom.

Heterocyclic Amines

The heterocyclic amines, although they contain tertiary nitrogen, co- ordinate readily, and a large number of pyridine complexes has been de- scribed. In general, these resemble the corresponding ammonia compounds. Davis and his students370 have found the stability of certain nickel and zinc pyridine compounds to decrease as the temperature is lowered. For example, Nipy4(SCX)2 is stable at room temperatures, but decomposes at —3°. It may be that the coordinating tendency of the thiocyanate group, relative to that of pyridine, increases with falling temperature till, at —3°, it dis- places the pyridine.

In this, as in other cases, chelation greatly enhances coordination, and metals which ordinarily do not coordinate with nitrogen form stable corn- Keller and Edwards, ./. Am. Chem. Soc, 74, 215 (1952). ■6. Bieber, Schlieezmann, and Ries, Z. anorg. allgem. Chem., 180, 89 (1929); Hieber

and Ries, / </. allgem. Chem., 180, 225 (1929).

;».: Bieber and Ries, Z. anorg. allgem. Chem., 180, 105 (1929).

Middleton: thesis, University of Illinois, 1938. 569. Tettamanzi, Atii accad. Torino, Clause sci.fis., mat. not., 69, 225 (1935); Spacu

and Dima, Bull. Soc. Stiinte Cluj, 8, 549 (1937). 70. Davis and Batchelder, /. Am. Chem. 8oc., 52, 4069 (1930); Davis and Ou, J. Am. Chem. Soc., 56, 1061, 1064 (1934).

68 CHEMISTRY OF THE COORDINATION COMPOUNDS

pounds with a-pyridyl hydrazine371, a-pyridyl pyrrole372, 2,2/-dipyridy] and 1 . 10-phenanthroline. Many coordination compounds of 2,2/-dipyridy] have been prepared. As far as is known, dipyridyl always acts as a bidentate coordinating agent. The stability of some of its coordination compounds is i ruly remarkable. For example, [Ni dipya]4"4" is destroyed only very slowly by sodium hydroxide or ammonium sulfide373. Prussian blue is completely destroyed in the cold by the addition of 2,2/-dipyridyl374.

Research on the dipyridyl complexes has centered largely on their stereo- chemistry, the stabilization of unusual valence states by coordination with dipyridyl, and the usefulness of the complexes in analytical chemistry (Chapter 20).

While many substituted derivatives of 2,2'-dipyridyl form complexes, substituents in the 6,6' positions may prevent coordination. Thus, 2-pyri- dyl-2/-quinoline, and 2,2/-diquinoline fail to react with octahedral meta] ions375, as does G^'-dimethyl^^'-dipyridyl376.

kAN^|s|Aj ChJIn^-^n^-CH3

The stabilization of valence states by coordination with dipyridyl is illustrated by the cases of silver and chromium. If present as the dipyridyl complex, Ag(I) can be oxidized to the Ag(II) complex and isolated as [Ag dipy2]++ 377 • 378. Hein and Herzog379 report that the reduction of [Cr dipy3]+++ in the presence of perchlorate ion gives [Cr dipy3]C104 , a deep blue compound, unstable in air, insoluble in water, but soluble in methanol, ethanol and pyridine.

2,2',2"-Terpyridyl and 2,2,,2",2'"-tetrapyridyl coordinate through all of their nitrogen atoms. The iron (II) ion fills its coordination sphere by combination with two molecules of terpyridyl380; the platinum (II) ion, having a coordination number of only four, forms compounds of the type [Pt tripyCl]Cl381. Tetrapyridyl gives compounds such as [Ag tetrapy]N03

371. Emmert and Schneider, Ber., 66, 1875 (1933).

372. Emmert and Brandl, Ber., 60, 2211 (1927).

373. Jaeger and Van Dijk, Proc. Acad. Sci. Amsterdam, 37, 10 (1934); 37, 618 (1934)

39, 164 (1936); Z. anorg. allgem. Chem., 227, 273 (1936).

374. Barbieri, Atti X° congr. intern, chim., 2, 583 (1938).

375. Smirnoff, Helv. chim. Acta, 4, 802 (1921).

376. Willink and Wibaut, Rec. Trav. Chim., 54, 275 (1935).

377. Barbieri and Malaguti, Atti acad. nazl. Lincei, Rend, classe sci.fis., mat. e nat.

8,619 (1950).

378. Malaguti, Atti acad. nazl Lincei, Rend, classe sci.fis., mat. e. nat., 9, 349 (1950)

379. Bein and Herzog, Z. anorg. allgem. Chem., 267, 337 (1952).

380. Morgan and Burst all, J. Chem. Soc, 1932, 20.

381. Morgan and Burstall, J. Chem. Soc, 1934, 1498.

GENERAL SURVEY 69

Co I! tetrapy]Clj , and [Pt(II) tetrapy][Pt< 1 . LO-Phenanthroline

resembles 2 , 2'-dipyridyl in its coordinating ability, but gives somewhat more stable complexes. Even beryllium and magnesium, which seldom co- ordinate with nitrogen compounds, form complex ions containing three molecules of 1 , 10-phenanthroline

The complexes of 1 , 10-phenanthroline are chiefly of interest because of their stereochemistry (Chapter 8), their usefulness in analytical chemistry (Chapter 20), and the ability of 1 , 10-phenanthroline to stabilize unusual valence states of some of the metals (Chapter 11).

Hydrazine Coordination

Hydrazine forms many coordination compounds, though their number is somewhat limited because of the reducing action of hydrazine. Com- pounds of the noble metals, and of metals in their higher oxidation states, are thus quite unstable. Efforts to prepare compounds of cobalt(III), for example, have been unsuccessful. Most hydrazine complexes which have been isolated as solids do not contain enough hydrazine molecules to fill the coordination sphere, so it has been suggested that hydrazine serves as a ■■dentate ligand. This, however, necessitates the formation of a three- membered ring. Goremykin884 treated potassium chloroplatinate(II) with ■iHi-HCl and obtained a product which he believes to be [PtCl2(X2H5)2] Cb--_MI,<>. nn heating, this goes to [PtCUNVH4)(X,II5)JCl, which reacts with pyridine to form [PtCb(\oH4).,]. If this interpretation is correct , hydra- zine is acting a- a monodentate donor.

Schwarzenbach and Zobist*8*, using the Bjerrum technique, have shown that in Bolution, zinc ion can coordinate with four molecules of hydrazine and nickel ion with six. Etebertus, Laitinen and Bailar888 have shown, polarographieally, thai zinc ion forms a tetrahydrazine complex.

K2. MorgaD and Buret all, ./. Chem. 8oc.t 1938, 1072. 1675.

'feifferand Werdelman, Z. anorg. Chem., 261, 197 1950 ,

mykin, Compt. n »<1 . acad. set. I .R 8 8 33. 227 1941). Schwarzenbach and Zobist, //</>■. ckitn. Acta, 35, 1291 1952

Laitinen, and Bailar, •/. .1/". Chem. Soc., 75, 3051 1053); Rebertus, thc-is, University of Illinois, L954.

70 CHEMISTRY OF THE COORDINATION COMPOUNDS

Biguanide Coordination Biguanide,

NH2— C— NH— C— NH2 ,

II II

NH NH

is a remarkable coordinating material which has been studied extensively by Ray and his students. Only two of the five nitrogen atoms coordinate these are on different carbon atoms. When coordination takes place, i hydrogen atom is lost from each molecule of biguanide. The uncoordinatec nitrogen atoms still have basic properties, so salts may be formed. Man} substituted biguanides have powerful coordinating ability. Among thes< are phenylbiguanide,

C6H5NHC— NH— C— NH2,

II II

NH NH

N,N' diphenylbiguanide, N,N'-diethylbiguanide, N-phenyl-N'methyl bi guanide, ethylenedibiguanide

NH2C— NH— C— NHCH2CH2NH— C— NH— C— NH2

II II II II

NH NH NH NH

and meta-phenylenedibiguanide

NH— C— NH— C— NH2

II II

NH NH

NH— C— NH— C— NH2

II II

NH NH

Bivalent metal ions such as Cu++ and Ni++ form stable complexes witl the biguanides. The copper complex is stable enough that the metal in i is not reduced by iodide, sulfite, thiosulfate, or other anions that commonh reduce copper(II) to copper(I)387. When the biguanide is unsymmetricalh substituted, as in phenylbiguanide, the copper(II) and nickel complexe exist in two cis and trans forms388. Ghosh and Chatterjee, however, isolate* only one form of each of the metal bis(methylphen}dbiguanides)389.

387. Ray and Bagchi, ./. Indian Chem. Soc, 16, 617 (1939).

388. Ray and Chakravarty, ./. Indian Chem. Soc, 18, 609 (1941). " rhosh and Chatterjee, ./. Indian Chem. Soc, 30, 369 (1953).

GENERAL SURVEY

The tervalent metal ions give remarkably stable complexes of the type

H 1

N = C

NH = C

NH -3HX or M;

NH*/ 73

JMH-C

'HN=C

NNH2'HX

Kay and his students have published a long series of articles on these in- teresting substances390. The chromium complexes undergo slow hydrolysis:

[Cr(BigH)3]X3 + 2H,0 -> [Cr(BigH)2(OH)H20]\, .

The hydroxoaquobis(biguanides) can hydrolyze further to monobiguanides, but these are unstable. The cobalt (III) complexes are more stable than those of chromium, and, in fact, have been shown to be more stable than the cobalt (III) ammines391. The tris(phenylbiguanide)cobalt(III) ion has been resolved into its optical antipodes392. Bis(biguanide) cobalt (III) com- plexes of the types [Co(BigH)2X2] and [Co(BigH)2XY] exist in cis and trans forms393. The dibiguanides are quadridentate394, apparently attaching them- selves to the metal through the a, a , 7, y' positions. Ray and Das Sarma395 have prepared the cobalt (III) meta-phenylenedibiguanide complexes [Co phenylene(BigH)2X2]+++, where X = XH3 or H20. These apparently have the trans configuration, for oxalate ion does not seem to be able to re- place the two coordinated X groups.

Among the most remarkable derivatives of biguanide is the silver(III) compound of ethylenedibiguanide:

CHs NH

CH2 NH

NH3

NH-

The high valence state of silver is quite stable, but is reduced to silver (I )

390. Ray and co-workers (Sana, Ghosh, Dutt, Battacnarya, Buddhanta, Chakravarty,

Majumdar, Das Sarma): ./. Indian Chem. Soc, 14, 670 (1937); 15, 347, 350, 353, 633 (1938); 16, 621, 629 (1939); 18, 289, 298 (1941); 19, 1 (1942); 21, 47 (1944); 23, 73 (1946); 25, 589 (1948); 26, 137 (1949), and other articles not in the "series".

391. De, Ghosh, and Ray, ./. Indian Chem. Soc., 27, 193 '1950).

392. Shiddhanta, Dutt, and Ray, •/. Indian Chem. Snr., 27, 641 1950

393. Ray and Majumdar. ./. Indian Chi m. 8oc., 23. 73 1946).

394. Ray and Shiddhanta. ./. Indian Chem. Soc., 20, 200 (1943).

395. Ray and Das Sarma. ./. Indian Cht 26, 137 1949).

72 CHEMISTRY OF THE COORDINATION COMPOUNDS

by iodide ion'16. The conductivity and magnetic susceptibility are consistent with the assumption that the compound contains trivalent silver with dsp2 bonds897. Measurement of stability constants shows this to be a very stable Bubstance898.

Quinoline and Its Derivatives

The nitrogen of quinoline has very weak donor properties, but properly substituted quinolines form stable coordination compounds. 8-Hydroxy- quinoline is a strong complexing agent, and has found wide use in analytical chemistry (Chapter 20). The important compounds are inner complexes

which are insoluble in water, but soluble in organic solvents; this property is utilized in the separation of metal ions, just as it is with the inner com- plexes of the 1,3 diketones (page 44)399. Substituted 8-hydroxyquinolines can often be used to advantage400.

Inner complexes can often be given water solubility by the introduction of a highly polar group into the complexing agent. If this substituent is distant in the molecule from the donor atoms, it does not disturb the stability or nature of the coordinate bonds. Thus, Liu and Bailar401 pre- pared the soluble zinc compound

H SOj

and resolved it into its optical antipodes.

396. Ray, Nature, 151, 643 (1943).

397. Ray and Chakravarty, ./. Indian Chem. Soc., 21, 57 (1944).

398. Sen, Ghosh, and Ray, ./. Indian ("hem. Soc., 27, 619 (1950). Mueller: Ind. Eng. Client., Anal. Ed., 15, 270, 346 (1943).

inn Moeller and Jackson, Anal. Chem., 22, 1393 (1950); Moeller and Ramaniah.

./. Am. Chem. Soc, 76, 2022 (1954). 101. I- in and Bailar, J .1///. Chem. Soc, 78, 5432 (1951). 402, Ley and Ficken, Ber., 50, 1133 (1917).

CENERAL SIRVEY

Picolinic acid, like other alpha amino acids, forms stable coordination

compounds, many of which arc inner complexes. The cobalt (III) com- pound40- is illustrative of this group.

Phthalocyanines and Porphins

When o-dicyanobenzene, o-cyanobenzamide, or related substances are heated with metals or their sails, a vigorous exothermic reaction takes place and metal derivatives of phthalocyanine

are formed. The metal occupies a position in the center of the molecule. If it be a divalent metal, it displaces the two hydrogen atoms, and coordinates with all four of the nitrogen atoms. Trivalent ions seem to form compounds of the type [Phthalocyanine M]X. The metal derivatives, like the parent substance, are deep blue. Many of them are extremely stable, being un- affected by any but the most vigorous chemical agents; some of them can be sublimed in vacuo above 500°C. This combination of properties makes them valuable as pigments (Chapter 22).

Phthalocyanine is closely related to porphin:

J NH HN— ^

>CH

N 1

Which also gives highly colored metallic derivatives**. Porphin is the parent 403. Fischer and Gleim, Ann., 521, 157 (1935).

CHEMISTRY OF THE COORDINATION COMPOUNDS

substance of chlorophyll and hemin CH=CH2 CH3

ch3y^>^ch^yVc?H5

r\ /A

coo phytyl

Chlorophyll a

CH2CH2COOH CH r^SY^CH"^Y^CH=CH2

Hemin

Haemocyanin, the blood pigment of molluscs and crustaceans, is a copper compound of the porphin family.

The Azo Group

The donor properties of the azo group are weak404, but azo compounds which contain a strong donor (e.g., carboxyl or hydroxy) in a position ortho to the azo group form very stable chelate rings. The complexes so formed are usually highly colored and find use as dyes and pigments (Chapter 22).

The diazo amino compounds have been the subject of an interesting study by Dwyer405. The imino hydrogen atom is replaced, at least in some cases, and the nitrogen chain forms a chelate ring with the metal, thus occupying two coordination positions. Examples are

><<3

N=N— NH

)(

-N=N— N-

■O),

Cu(<3-N==N-N-<3)

and

-N=N— N<

Nitriles

The nitrogen atoms in organic nitriles have fairly strong donor proper- lies, especially toward the heavier metals. The halides of platinum add

404. Kharasch and Ashford, J. Am. Chem. Soc, 58, 1736 (1936).

405. Dwyer, J. Am. Chem. Soc, 63, 78 (1941).

GENERAL SURVEY 75

nitrttes directly*1 to form PtX8(RCN)2 and PtX.dUNi, (R may be either aliphatic or aromatic). Halogens readily convert the platinum(II) com- pounds to the platinum(IV), which arc not readily reduced again, even l>.\ formaldehyde4, sulfur dioxide or aluminum and hydrochloric acid.

Lebedinskii and Golovnya407,408 have carried out the following reactions:

|1V('1I<\ Cl - C II Nil: - [Pt(C\.H5NII.>)4((,,H,(,\>,][PtC,l1|

IK'l

[Pt(C2H5XH2),(M,l

[Pt(NH,),Cl,] + CII3CX -» lPt(XH,)>(CH3CX)Cl]Cl

+ NH4OH '-+ [Pt(XH3)4(CH3CX)]Cl2

KlPt(XH3)Cl3] + CH3CX -* [Pt(XH3)(CH3CN)Cl2]

+ XH4OH -» [Pt(XH3)4(CH3CN)]Cl2

The platinum in the compound [Pt(XH3)4(CH3CX)]Cl2 does not seem to show the usual coordination number for platinum, and doubtless needs further study. Upon heating with hydrochloric acid, this compound is converted to [Pt(XH3)3Cl]Cl.

In the presence of acetonitrile, copper(I) coordination compounds are readily formed. They oxidize slowly in the air409.

Pseudohalides

The cyanide ion has unshared pairs of electrons on both the carbon and the nitrogen atoms, and theoretically, it might coordinate to metals through either of these pairs. Actually, it seems to combine preferentially through the carbon atom, and the simple, mononuclear cyanides are characterized by a metal-carbon link (page 87). The formation of the carbon-metal bond, however, does not preclude the formation of a coordinate bond between the nitrogen and another metal atom. The "super-complex" heavy metal cyanides, such as Prussian blue, are probably built up in this way, as are the organo gold cyanides.

The thiocyanate group also has pairs of electrons on two atoms, and con- ceivably can coordinate through either nitrogen or sulfur (p. 57). The

406. Ashford, thesis, University of Chicago, 1936. Ashford gives references to several earlier articles on platinum-nitrile addition compounds, the more important being Hofman and Bugge, Ber., 40, 1772 (1907); Ramberg, Ber.t 40, 2578 (1907); TschugaofT and Lebedinskii, Compt. and., 161, 563 (1915). Lebedinskii and Golovnya, Izvest. Sektora Platiny i Drugikh Blagorod. Metal., Inst. Obschei i Neorg. Kkim., A had. Nauk. S.S.S.R., No. 22, 168 (1948); cf. Chem. Abs., 44, 10566a, (1950).

408. Lebedinskii and Golovnya, Ann. seeteur platine, Inst. Chim. pen I S 8.R.),

No. 16, 57 '1939).

409. Morgan, ./. Chem. 80c., 123, 2901 (1923).

7<i CHEMISTRY OF THE COORDINATION COMPOUNDS

easy formation of the highly colored iron (III) and cobalt (II) complexes makes them suitable for the qualitative detection of these ions in solution. Several investigations have been made to determine the nature of the ferric bhiocyanate complex which exists in such solutions. M0ller410 showed, by conductivity measurements, that there are not more than three thiocyanate groups attached to the iron. Bent and French411 and Edmonds and Birn- baum412, from a study of the absorption of light by solutions containing iron (III) and thiocyanate, came to the conclusion that the formula of the complex is Fe(NCS)++, neglecting hydration. Schlesinger and Van Valken- burgh413 showed that in ether solution, the [Fe(NCS)e]~ ion is present. The entire subject has been well reviewed by Lewin and Wagner414.

The coordinating ability of the azide ion was first studied by Strecker and Oxenius415, who prepared a series of cobalt (III) compounds. They ob- tained the ions as-[Co(NH3)4(N3)2]+ cis- and trans-[Co en2(N3)2]+, and [Co py4ClN3]+. Linhard and Flygare416 prepared several salts of the ion [Co(NH3)6N3]++ which they report to be similar in color to [Co(XH3)5Cl]++. The action of sodium azide on a solution of [Co(NH3)4(H20)2]+++ gave a mixture of the cis and trans forms of [Co(XH3)4(X3)2]+417. That the azide group has strong donor properties is shown by the fact that triazidotriam- minecobalt can be prepared by treatment of [Co(NH3)4(H20)2]+++, [Co(NH3)4(N3)2]+, or [Co(NH3)bN3]++ with azide ion418. Straumanis and Circulis419 prepared stable, slightly soluble compounds which they believed to be nonelectrolytes of the type [Cu R2(N3)2] in which R is ammonia or any one of a number of aliphatic or aromatic amines. They also obtained the anions [Cu(N3)6]4_, [Cu(N3)4]==, and [Cu(N3)3]~. All of the azido com- plexes are unstable and explosive.

Oximes

The coordinating tendency of the oximes is well known. A lone oxime group does not coordinate firmly, but when it forms part of a chelate ring, the oxime nitrogen has very strong donor properties and oximes are fre- quently used in inorganic analysis (Chapter 20). Metallic ions having a coordination number of six combine with only two dioxime groups, the

HO. M0ller, Kern. Maunedsblad, 18, 138 (1937).

111. Bent and French, ./. Am. Chem. Soc, 63, 568 (1941).

U2. Edmonds and Birnbaum, ./. .1///. Chem. Soc, 63, 1471 (1941).

H3. Schlesinger and Van Valkenburgh, ./. Am. Chew. Soc, 53, 1212 (1931).

II I. lewin and Wagner, ./. Chem. Ed., 30, 445 (1953).

n:> Strecker and Oxenius, Z. anorg. allgem. Chem., 218, 151 (1934).

416. Linhard and Flygare, Z. anorg. Chem., 262, 328 (1950).

417. Linhard, Weigel, and Flygare, Z. anorg. allgem. ('hem., 263, 233 (1950). lis. Linhard and Weigel, Z. anorg. allgem. Chem., 263, 245 (1950).

il'.i. Straumanis and Circulis, Z. anorg. allgem. Chew.. 251, 341 (1943); 252, 9, 121 (1943).

GENERAL SCh'VKY 77

remaining coordination positions being filled by other donors, as the follow- ing cobalt compounds illustrate:

[Co(HD),(\II..V.]\, [Co(HD)sNH,X] and [( \><H1)),.\,]- 12l>.

Aniline and substituted anilines can replace the ammonia in the cobalt compounds1-1. Compounds of the type M'|M"'( II 1 > ) - X - 1 containing rho- dium4'- and iridium1-" have also been described. Only one =X()II group from each onlt-dioxime molecule liberates a hydrogen ion, as is shown by the fact that the mono-ethers,

R C C R 424.425,426

II II

lio— X N — OCH, and the imino and methylimino compounds

CtHs— C— C— CH, C>Ho— C— C— CH3 426

II II and || ||

H— X X— OH CH8— N X— OH

give entirely analagous compounds. On the other hand, both hydrogens of an am phi -dioximv are replaceable. Nickel, for example, forms rather poorly defined compounds of the type

R— C C— R

II II

X X

/ \ /

O Xi— o

in which the metal is evidently attached to one nitrogen atom and one oxygen atom1-7 m 1'-"-'. Acids rearrange these to the more stable red modifi- cation. The anti and amphi forms of benzil dioxime react with palladium(II) ion just as they do with the nickel ion. Syn dioximes do not yield nickel derivatives1-"' but syn benzildioxime readily forms a crystalline palladous

420. Tschugaeff, Ber.t 39, 2692 1906); 41, 2226 (1908).

421. Ablov, Bull, toe. ckim., 7, 151 (1940

422. Lebedinskii and Fedorov, Ann. secteur platine, Inst. ckim. </< n . (U.S.S.R. ,

No. 15, 19 1038).

423. Lebedinskii and Fedorov. Ann. sectew platine, Inst. ckim. gen. I SSI: .

No. 15, 27 )'• 421. Thilo and Friedrich, Ber., 62, 2990 L929

125. Bradj and Muere, •/ Chem. Soc., 1930, 1599.

126. Pfeiffer, Ber., 63. 1811 1930

B7, .Mack. ./. Chem. 8oc., 103, 1317 1913

42v Hieber and Leutert, Ber., 62, L839 1929

420. Meisenheimer and Theilacker, Ann.. 469, L28 L95

78 CHEMISTRY OF THE COORDINATION COMPOUNDS

compound which is said to have the structure

(h — c i C — <j)430

II H

N-O O-N

Bryson and Dwyer431 report that /3-furfuraldoxime reacts with copper, silver, nickel, and cobalt. The a isomer does not, but on standing in solu- tion with the metal salt, it changes to the /? isomer.

The Donor Properties of Phosphorus and Arsenic Phosphine Coordination

The action of phosphine on metallic salts has been studied by several investigators. Most metallic ions are reduced to metal or to phosphides432, 433 , 434^ kut some form phosphine addition compounds. Riban435 found that a solution of copper(I) chloride in hydrochloric acid absorbs phosphine readily, forming the rather unstable compounds CuClPH3 and CuCl-2PH3 . Upon gentle warming, these compounds liberate phosphine, while stronger heating generates copper phosphide. These results have been confirmed and extended by Scholder and Pattock436.

Holtje and his co-workers437 have made a systematic study of the donor properties of phosphine. They found that in its ability to form coordination compounds, phosphine resembles hydrogen sulfide more closely than it does ammonia. The phosphinates are more stable than the sulfhydrates in every case investigated. Among the more stable compounds reported by these investigators is A1I3-PH3, which may be sublimed in vacuo.

Tertiary Phosphine and Arsine Coordination

The tertiary organic phosphines and arsines have strong donor properties, in which regard they are in sharp contrast to the tertiary amines, but are similar to the thioethers. Even the stibines can form addition compounds438.

430. Dwyer and Mellor, ./. Proc. Roy. Soc. N.S. Wales, 68, 107 (1935).

431. Bryson and Dwyer, ./. Proc. Roy. Soc, N.S. Wales, 74, 107 (1940).

432. Winkler, Ann. chim. phys., Ill, 443.

133. Keilisch, Ann., 231, 327 (1885); "Ueber die Einwirkung des Phosphorwasser-

stoffs auf Metallsalzlosungen," Berlin, 1885. i:u Scholder, Apel, and Haken, Z. anorg. allgem. Chem., 232, 1 (1937). 135 Riban, Compt. rend., 88, 581 (1879); Bull. soc. chim., [2] 31, 385 (1879). 436. Scholder and Pattock, Z. anorg. allgem. Chem., 220, 250 (1934). 137. Holtje, Z. anorg. allgem. Chem., 190, 241 (1930); 209, 241 (1932); Holtje and

Meyer, Z. anorg. allgem. ('hem., 197, 93 (1931); Holtje and Schlegel, Z. anorg.

allgem. Chem., 243, 246 (1940). 438. Jensen, Z. anorg. allgem. Chem., 229, 225 (1936).

GENERAL SURVEY 79

The strong trans influence of tertiary phosphines is emphasized by the failure of Kumakov's rule (Chapter 9) in the reaction of thiourea with [PtCPEtjJJBrJ489. The use of several of the phosphine compounds as anti- knocks has been patented440.

Organic phosphines111 and arsines442 are often identified through their highly crystalline mercuric halide complexes. These are true coordination compounds, and are soluble in organic solvents.

The most common phosphines and arsines of copper are CuX-AsRa and CuX >2AsRs , where X is a halide ion. Those containing a single coordinated arsine group are tetrameric while those containing two arsine groups are presumably monomelic. Nyholm448 has reported that four molecules of diphenylmethylarsine may be associated with a single copper(I) ion, as in the compounds [Cu(AsMePh2)4][CuX2] and [Cu(AsMePh2)4]X. This ter- tiary arsine also forms the nonelectrolytic complex [Cu(AsMePh2)3X]. Similar behavior444 was also noted among the o-phenylenebis(dimethyl- arsine) complexes of copper(I). Gold complexes of the form AuX-MR3, where X is Cl~, Br~, or XCS~, and M is arsenic or phosphorus, are mono- meric, and some of them can be distilled under reduced pressure. There is evidence445 that the corresponding cyanides and iodides are polymeric. The extreme stability of these substances is shown by the fact that tributyl- phosphinegold(I) chloride may be volatilized at atmospheric pressure and triethylphosphinegold(I) chloride446 dissolves in concentrated hydrochloric acid and in potassium hydroxide without decomposition, and is only slowly reduced to metallic gold by sulfur dioxide. The vapors of AuCl-PBu3 deposit a fine film of gold when passed through a heated tube447. Both gold (I) and gold (I II) complexes448 with o-phenylenebis(dimethylarsine) have been reported449.

439. Grirrberg and Razumova, Zhur. Obschei Khim., 18, 282 (1948).

440. Bataafsche Petroleum Maatschappij, French Patent 805 666 (1936); Peski and

Melsen, U.S. Patent 2 150 349 (1938).

441. Davies and Jones, J. Chem. Soc, 1929, 33; Da vies, Pearce, and Jones, J . Chem.

Soc, 1929, 1262; Jackson, Davies, and Jones, ./. Chem. Soc, 1930, 2298; Jack- son and Jones, J . ('hem. Soc, 1931, 575; Jackson, Davies, and Jones, ./. Chi m . Sue, 1931, 2109.

442. Jones, Dyke, Davies, Griffiths, and Webb, J. Chem. Soc, 1932, 2284; Challenger,

Higginbot torn, and Ellis, ./. Chem. Soc, 1933, 95; Challenger and Ellis, J. Chem. Soc, 1935, 398; Challenger and Rawlings, ./. Chem. Soc, 1936, 264; Blicke and Cataline, ./. Am. Chem. Soc, 60, 419 (1938). 44:i. Nyholm, ./. Chem. Soc, 1952, 1257.

444. Kabesh and Nyholm, ./. Chem. Soc, 1951, 38.

445. Dwyer and Stewart, ./ . Proc. lion. Soc, .V.N. Wales, 83, 177 (1949). 146. Levi Malvano, .1/// accad. Lincei, [5] 17, i. 847 (1908).

447. Mann and Wells, Natun . 140, 502 (1937); Mam.. Wdls. and Purdie, •/. Chem.

Soc, 1937, 1828.

448. Mann and Purdie, ./ . Chem. Soc, 1940, 1235.

449. Nyholm, Nature, 168, 705 (1951).

80 CHEMISTRY OF THE COORDINATION COMPOUNDS

Recently, compounds of o-phenylenebis(dimethylarsine) (PDA) have been prepared with four or six arsenic atoms coordinated to one metal atom. Iron forms complexes of the formulas [FeIII(PDA)2Cl2]C104 and [Fcn(Pl)A)2X,l(X - Br", I", or SCN-)450. The magnetic moments of the complexes indicate that the iron atom is covalently bound.

Rhodium (III) halides react with o-phenylenebis(dimethylarsine) forming analagous compounds451. However, upon reaction with a monodentate tertiary arsine, rhodium (III) halides form two isomeric compounds con- taining three moles of arsine per mole of rhodium. These are, presumably, |Rh(AsR3)G][RhX6] and [Rh(AsR3)3X3]452. Rhodium(II) forms a variety of other complexes453 with tertiary arsines, such as [Rh(AsR3)6]3[RhX5(AsR3)]2 and [Rh(AsR3)6][RhX4(AsR3)2].

Iridium(II) and iridium(III) also form complexes with tertiary arsines454. Dwyer, Humpholtz, and Nyholm455 have investigated the complexes of diphenylmethylarsine with ruthenium (II) and ruthenium (III). Ruthe- nium(II) forms the complex [Ru(AsR3)4X2] while ruthenium(III) forms [Ru(AsR3)3X3].

The preparation of nickel complexes of trialkyl compounds of the group V elements has been especially fruitful, as higher valence states of nickel are probably best characterized among these derivatives. Jensen and Ny- gaard456 prepared a rather unstable pentacoordinate triethylphosphine complex of tripositive nickel [NiBr3(PEt3)2]. The corresponding cobalt(III) complex, CoCl3-2PEt3, has been studied457; it is probably of the same configuration as the nickel complex (see Chapter 10, page 392). Nyholm458 has reported [Xi(PDA)2X2]X, containing nickel(III), and [Ni(PDA)2X2] (C104)2 , which contains nickel (IV).

This work459 on the o-phenylenebis(dimethylarsine) complexes of the metals of the first transition series has been quite significant from the the- oretical standpoint. It has been found that this ditertiary arsine reacts with transition metal ions with the formation of strongly covalent bonds only when the metal ion contains d-electrons which are not involved in the

450. Nyholm, J. Chem. Soc, 1950, 851. .

451. Nyholm, J. Chem. Soc, 1950, 857.

452. Dwyer and Nyholm, J. Proc Roy. Soc, N.S. Wales, 75, 140 (1942).

453. Dwyer and Nyholm, J. Proc. Roy. Soc., N.S. Wales, 76, 133 (1942).

i:» 1 . Dwyer and Nyholm , ./ . Proc. Roy. Soc, N.S. Wales, 77, 116 (1943) ; 79, 121 (1946) .

455. Dwyer, Humpholtz, and Nyholm, J. Proc. Roy. Soc, N.S. Wales, 80, 217 (1947).

456. Jensen and Nygaard, Acta. Chem. Scand., 3, 474 (1949). 157. Jensen, Nature, 167, 434 (1951).

458. Nyholm, ./. Chem. Soc., 1950, 2061; 1951, 2602.

150. Hurst all and Nyholm, /. Chem. Soc, 1952, 3570; Nyholm and Sharpe, /. Chem. Soc, 1952, 3579.

GENERAL SURVEY 81

hybridized group (see Chapter 1). It has been concluded thai the stability of arsine and phosphine complexes depends on the formation of double- bonds between the metal and the donor atom. This conclusion is not incon- sistent with the observation that the more stable complexes containing

phosphorus-metal or arsenic-metal bonds occur among the group VIII and IB metals. Complexes of the tritertiaryarsine, methylbis(3-dimethylarsinopropyl)-

arsine(TAS), have been prepared by Barclay and Nyholm460 The iron(III) complexes, [Fein(TAS)Xj], are nonelectrolytes and exhibit magnetic mo- ments corresponding to one unpaired electron. Cobalt (II) iodide forms a similar complex, [Con(TAS)I]I, which contains a single unpaired electron; air oxidation produces diamagnetic [Com(TAS)L]. Copper(I) and nickel(II) form the diamagnetic, nonelectrolytic complexes [Cu(TAS)I] and [Ni(TAS)IJ. The possibility of pentacoordinate nickel(II) here is especially interesting in view of the previously mentioned observations of Jensen and Nygaard.

By far the best known compounds in this group, however, are those of platinum and palladium. Cahours and Gal, in 1870, isolated isomeric forms of PtCl2-2P(CH3)3, PtCV2P(C2H5)3, and PtCl2-2As(C2H5)3. Their work was confirmed by Klason461 and by Jensen, who extended it to the stibines438. Chatt and Wilkins462 studied the isomerization of palladium compounds of this general type by following the variation in dielectric constant of their solutions. Xo detectable amount of the cis isomer of the arsine or phosphine complexes appears to exist in solution, while as much as 40 per cent of the stibine complex may be cis.

Complexes of platinum (IV) with tertiary arsines463 and phosphines464 have been prepared in isomeric forms by oxidation of the appropriate iso- mers of PtX2-2MR3 .

Upon treatment with ammonium tetrachloropalladate(II), the bis(phos- phine)palladium(II) compounds, [Pd(PR3)2Cl2], are converted to the di- nuclear complexes, [Pd2(PR3)2Cl4]. Mann and his co-workers465 have studied these bridged compounds in some detail. They were at first of the opinion that several forms could exist

460. Barclay and Xyholm, (hem. and Ind., 1953, 378.

461. Cahours and Gal, Compt. rend. 70, 1380; 71, 208 (1870). Klason and Wanselin,

./. prakt. Chem., [2] 67, 41 (1903).

462. Chatt and Wilkins, J. Chem. Soc, 1953s 70.

463. Xyholm, J. Chem. Soc, 1950, 843.

464. Chatt, J. Chem. Soc, 1950, 2301.

465. Mann and Purdie, Chem. and Ind., 54, M4 (1935); ./. Chem. Sue, 1935, 1549;

1936, 873; Chatt and Mann. ./. Cheni. Soc. 1938, 1949; Chatt, Mann, and Wells, J. Chem. Soc. 1938, 2086.

Xo <\ 7rs

CHEMISTRY OF THE COORDINATION COMPOUNDS

FV\ /CIN /CI

Pd Pd

r3p// xcr Nci

R3P\ /CIN /PR3 Pd Pd

c/ Nc/ XCI

AND

R3pn /Ck /Ci

Pd Pd

Evidence for the first formula was found in the fact that dipyridyl and nil rites read with these substances, and with the corresponding arsine derivatives, to give mixtures of compounds:

(R3P)2PdCl2PdCl2 + dipy -» (R3P)2PdCl2 + [(dipy)PdCl,]

(R3As)2PdCl2PdCl2 + 6KNO2 -> (R3As)2Pd(N02)2 + K2[Pd(N02)4] + 4KC1

On the other hand, aniline, toluene, and pyridine give good yields of mono- phosphine (or arsine) derivatives:

\ / \ S

Pd Pd

C\y NCIX XPR3

+ 4.

NH;

R3PX /CI

NH2((>

Ethylenediamine splits the butyl phosphine compound unsymmetrically in benzene, but symmetrically in alcohol.

Later evidence, however, showed the earlier hypothesis to be incor- rect466, 467; the dimeric molecule apparently always has the symmetrical structure. It was shown that compounds of the type PtCl2-(PR3)2 are not primary products of the splitting, but are formed by secondary reactions. The unsymmetrical formulas for the bridged complexes would indicate that compounds of the types

(CH3)2

Pd Pd AND

■As' ^cr

(CH3)2

kCI

(C6H5)2 ^As .CL /CI

HaC^ \ / \ / Pd Pd

^As Ncr Nci

(C6H5)2

should exist. Chatt and Mann466 were unable to prepare any such com- pounds, but obtained

[PdCI4] ,

PdCI,

466. Chatl and Mann, ./. Chem. Soc, 1939, 1622.

167 . Mann and Wells, ./. Chem. Soc, 1938, 702; Wells, Proc. Roy. Soc. London, A167, 169 (1938).

GENERAL SURVEY

and several other interesting substances. Chatl has extended this work to include the tripropylstibine complex488 of platinum(II). This Bpecies behaves essentially as the arsine and phosphine complexes.

[nteresting examples of phosphine complexes with bridging groups other

than the halide ions are found in the ethyl mercaptan and oxalate bridge*

pc Pt a' xs^ xpr.

R3PV .0-0 = 0. .CI

xPt/ I XPt

/ \

o— c=o

s \

PR.

The ethyl mercaptan complex exists in two (cis-trans?) forms. A related compound with thiocyanate bridges is reported to exist in the isomeric forms

R3P

,C\

SCN

NCS/ ^c/ NPR3

R3PX y

^Pt 'Pt

c/ NS^ XPR3 CN

The reported isomerism of the bridged compound trichlorotris(diphenyl- methylarsine)copper(I)copper(II) is interesting470. Copper(I) is tetrahed- ral471 while copper (I I) is planar so that the isomerides were thought to be

CL CL AsMe(J)p Cu Cu

AND

())2MeASx Clx CI Cu Cu

(J^MeAs^ XCI^ NAsMe<t>2

^MeAs7 CI XAsMe())2

L TETRAHEDRAL PLANAR

~~ TETRAHEDRAL PLANAR

However, it has since been contended that these substances are actually complexes of diphenylmethylarsine oxide and that the reported isomerism was associated with an impurity in one of the form-

The existence of bridged arsine and phosphine complexes containing two different metals is reported by Mann and his co-workers47*. A series of compounds involving palladium(II) or cadmium bridged to mercury is exemplified by

468. Chart. ./. Chem. So,-., 1951, 652.

hatl and Hart, /. Chem. Soc., 1953, 260; Nature, 169, 673 (1952 ; Chatt, Mann, and Wells, ./ - Joe., 1938, 2086.

470. Mellor, Burrows, and Morris, Nature, 141, 114 1038).

471. Mellor and Craig, ./. Proc. Roy. Soc., X S Wales, 75. 27 1941

472. Nyholm, J. CI - . 1951, L767.

473. Mann and Purdie, ./ . Chem. Soc, 1940, 1230; Allison and Mann, ./. Chen -

1949, 2915.

SI CHEMISTRY OF THE COORDINATION COMPOUNDS

Pr3Asx ^Br\ ^Br

M Hq

Br NBr ^AsPr3

The compounds SnX4-2PR3(X = Cl~ or Br~) also form mixed-metal, bridged complexes with mercury (II) or palladium(II).

R3PX| /C!N /PR3 Sn M

ci'| NcK xc.

Tertiary phosphines react with the carbonyls of iron, cobalt and nickel to produce mixed phosphine-carbonyl complexes:

Ni(CO)4 + PR3 (or 2PR3) -> [Ni(PR3)(CO)3] (or [Ni(PR3)2(CO)2])

Their catalytic behavior in the reactions of acetylene has been discussed by Reppe and Sweckendich474.

Cacodyl oxide, (CH3)2As — O — As(CH3)2 , which might be expected to coordinate through both arsenic atoms, does so with difficulty, and it usually occupies only one coordination position475.

Phosphorus (III) Halide Coordination

The "double compounds" formed by phosphorus(III) chloride and bro- mide with metal halides certainly contain true coordinate links, the phos- phorus acting as the donor atom. Platinum(II) chloride and phospho- rus(III) chloride, for example, give the highly crystalline compounds PtCl2 • PC13 and PtCl2 • 2PC13476. These react with water to give PtCl2 • P(OH)3 and PtCl2-2P(OH)3 , and with alcohols to form the corresponding esters. Molecular weight determinations have shown the ethyl ester of the mono- phosphine complex to be dimeric, and hence (presumably)

Ck ^Clx /P(OR)3 .Pt Pt

(ro)3p^ xcr XCI

474. Reppe and Sweckendich, Ann., 660, 104 (1948).

475. Jensen and Frederiksen, Z. anorg. allgem. Chem., 230, 34 (1936); Baudrimont,

Compt. rend., 55, 363 (1862); Ann. chim. phijs., [4] 2, 5 (1864); "Recherches sur les chlorures et les bromures de phosphore," Paris, 1864.

476. Schutzenberger, Compt. rend., 70, 1287, 1414 (1870); Bull. soc. chim., [2] 14,

97, 178 (1870); Schutzenberger and Fontaine, Bull. soc. chim., [2] 17, 386, 82 (1872).

GENERAL SURVEY 85

The ester PtCl2-2P(OCH8)i , however, is monomeric477:

(CH30)3P /CI

(CH30)3P^ XCI

The acids and esters react with silver salts, with replacement of the chlo- ride groups, the acids at the same time forming silver salts476:

(AgO)v N /.w,

kPN /NO: Pt/ (AgO)3P/ XN03

The dimeric esters are readily split by substances which have fairly strong- donor properties17''' ,7s 179. Aniline, for example, gives cis and trans [PtClo P(OC2H5)3(C6H5XH2)]480. [PtCl22P(OC2H5)3] adds two molecules of ammonia, both chlorides becoming ionic. Platinum(II) chloride also forms white, crystalline [PtCV (PF3)2] and red [PtCl2- (PF3)]2 when treated with phosphorus(III) fluoride481. Both substances are sensitive to moisture; how- ever, the white compound is thermally stable and may be refluxed in a dry atmosphere without substantial decomposition. It is interesting that phos- phorus(III) fluoride, which has no appreciable basic character, should form such stable complex compounds. This behavior is attributed by Chatt482 to the formation of a double-bond between the phosphorus and the plati- num.

As might be expected, palladium(II) chloride forms analagous com- pounds483. The corresponding iridium compounds, which have been studied by Geisenheimer484 and by Strecker and Schurigin485, are reported to be much more stable than those of platinum and palladium. IrCl3-3PCl3 does not react with cold alcohol, with cold concentrated sulfuric acid, or with organic bases.

477. Rosenheim and Loewenstamm, Z. anorg. Chem., 37, 394 (1903).

178. Schutzenberger and Fontaine, Bull. soc. chim., [2] 18, 101, 148 (1872).

479. Rosenheim and Levy, Z. anorg. Chem., 43, 34 (1905).

480. Troitskaya, Zhur. Priklad. Khim.s 26, 781 (1953).

481. Chatt and Williams. ./. Chun. Soc, 1951, 3061.

482. Chatt : Nature, 165, 637 I960).

///. rend., 116, 176 (1892); 123, 603 (1896); The author's name is spelled Pinck in the second reference, but it evidently refers bo the same man.

484. ( ieisenheimer, Ann. chim. phye., [6123,231 (1891); "Sur lea chlorures el bromures double d'iridium e1 de phoephore," Paris, 1891.

185. Strecker and Schurigin, Ber., 42, 1 7 « . 7 1909 ; Schurigin, "Die Einwirkung von Phosphor-halogeniden auf die Metalle der Platingruppe," Grieswald, 1909

86 CHEMISTRY OF THE COORDINATION COMPOUNDS

( rold(I) halides form a similar series of compounds, e.g., AuCl • PC13486 • 487. It is not possible to obtain AuCl -P(OH)3 , for the phosphorous acid reduces the gold to the metallic state, but AuCl-P(OC2H5)3 is quite stable, and is not reduced by sulfur dioxide. It is soluble in ammonium hydroxide with the formation of AuClP(OC2H5)3-2NH3 , from which acids reprecipitate it in the original form. The methyl ester, AuCl-P(OCH3)3 , has been pre- pared by the action of methanol on AuCl-PCl3 and by the union of tri- methyl phosphite and gold (I) chloride. The phenyl ester was prepared by the second method488.

A series of nickel (0) compounds with phosphorus(III) halides has been prepared from nickel tetracarbonyl489. These compounds are of the composi- tion Ni(PX3)4 . Phosphorus(III) fluoride does not completely replace the carbonyl groups from nickel tetracarbonyl; however, the compound [Xi(PF3)4] can be prepared by the following reactions

[Ni(PCl3)4] or [Ni(PBr3)4] + 4PF3 -> [Ni(PF3)4] + 4PC13 or 4PBr3

[Ni(PCl3)4] + 4SbF3 -> [Ni(PF,)4] + 4SbCl3

Antimony (III) chloride reacts with nickel and iron carbonyls giving the products [Ni(CO)3SbCl3] and [Fe(CO),(SbCl3)2], respectively490.

Copper(I) chloride reacts with phosphorus(III) chloride491, but the compound so formed is reactive and unstable. With methyl alcohol it gives a mixture of copper(I) chloride and CuClP(OCH3)3 . Iron (III) chloride gives the volatile compound FeCl3 ■ PC13492.

The Donor Properties of Carbon There are three great classes of coordination compounds in which carbon apparently shares electrons with metals — the ethylenic compounds, the metal carbonyls, and the complex cyanides. The first two of these are the subjects of special chapters in this book, so this section will be devoted to the cyanides and the closely related complexes of metal ions with iso- n it riles.

Cyanide Coordination

The cyanide ion has unshared electrons both on the carbon atom and on the nitrogen atom, and one might expect to find isomeric series of complexes

486. Lindet, Compt. rend., 98, 1382 (1884); 101, 164 (1885); 103, 1014 (1886); Bull, soc. chim., [2] 42, 70 (1884); Ann. chim. phys., [6] 11, 177 (1887). Most of Lindet 's conclusions were Later confirmed by Levi-Malvano (Ref. 446).

187. Arbuzov and Lovoastrova, Doklady akad. Nauk. S.S.S.R., 84, 503 (1952).

188. Arl.uzov and Shavska, Doklady akad. Nauk. S.S.S.R., 84, 507 (1952).

189. Irvine and Wilkinson, Science, 113, 742 (1951); Wilkinson, J. Am. Ckem. Soc,

73, 559 (1951).

190. Wilkinson, ./. .1///. Chem. Soc, 73, 5502 (1951).

191. Davis and Ehrlich, J. Am. Chem. Soc, 58, 2151 (1936).

192. (Jrbain, British Patenl 312 685 (May 31, 1928).

GENERAL SURV1-) 87

corresponding to the nitriles and isonitriles of organic chemistry. Such, how- ever, have not been observed, so it is concluded thai the attachment of the

cyanide ion to any given metal ion always takes place through the same atom. It is conceivable that some metals share electrons with the carbon and others with the nitrogen, but there is no experimental support for such a hypothesis. The preponderance of the evidence indicates thai in complexes of the type [M(CN)J*~, union is always through the carbon.

Carbon and nitrogen are so close together in atomic number that only the most accurate x-ray measurements can distinguish between them. Such distinction is particularly difficult in the complex metal cyanides, where the heavy metal atom masks the lighter nonmetals. A few such accurate meas- urements have been made, and all of them support the hypothesis that the metal is attached to carbon493. Holzl and his co-workers have come to the same conclusion from chemical studies. They alkylated a number of metal cyanide complexes, and obtained compounds which upon decomposition yielded alkyl isonitriles494. In some cases, alkyl amines were also obtained, but in no case were ammonium salts formed in significant amounts. Infrared spectral work by L. H. Jones495 indicates the existence of the carbon-metal bond. He found that the pattern of infrared active vibrational frequencies for the compounds KAu(CM-X14)2 , KAu(C12N14)(C13N14), KAu(C13N14)2 , and KAu(C12X14)(C12X15) indicates that the bonding is through the carbon. Jones also found that in [Au(CX)2]- theC=N force constant is greater than in CH3C=X. In CH3X=C the C=X force constant is considerably less than in CH3C=X. This also indicates, but does not prove, that the CN is bound to the gold through the carbon.

The cyanide ion is a powerful coordinating agent, and it frequently dis- places all other groups from the coordination sphere, forming ions of the type [M(CX)X]V~. Exceptions to this are found among the carbonyl and nitrosy] cyanides (Chapter 16), and in such complexes as [Co en2(CX)2]+ 496a, [Co(CX)5OH]=496b, [Co(CX)4(OH)2]= and [Fe(CX)5H20]=497.

Examples of unusual and variable coordination numbers are fairly com- mon among the cyano complexes. Thus, Adamson498 believes that the formula for potassium cobalt (II) cyanide, which has long been written

Hoard, Z. Krist., 84, 231 (1933); Hoard and Nordflieck, ./. .1///. Chetn. Soc, 61, 2853 L939 ; Powell and Bartindale, •/. Chem. Soc., 1945, 799.

194. Holzl, Monats., 48, 71 (1927); 51, 1, 397 (1929); Holzl and Xenakis, Monats.,

48,689 L927 ; Holzl and Viditz, Monats., 49, 241 (1928); Holzl and Krichmayr, Monats., 51, 397 1929); Holzl, Meier-Mohar, and Viditz. Monats., 52, 73; 53 54, 237 L929 .

195. I.. II. Jones, private communication.

496a. Ray and Sauna, ./. Indian clem. Soc. 28, 59 1951 496b. Smith.. Kleinberg, and Griswold, /. Am. Chem. Soc., 75, 149 (1953). 197. Hieber, Nast, and Bartenstein, Z. anorg. allgem. Chem., 272, 32 1953). A-damson, /. Am, Chem. Nor. 73, 5710 1951).

88 CHEMISTRY OF THE COORDINATION COMPOUNDS

K4Co(CN)6 , is actually K3Co(CN)5 . Even in aqueous solution, the co- ordination number of five is maintained. The familiar copper cyanide plating hath contains both [Cu(CN)2]~ and [Cu(CN)3]=, the latter predominating. The infrared spectral studies of L. H. Jones and Penneman499 have shown that in aqueous solutions containing silver ion, increasing concentration of cyanide ion brings about the successive formation of [Ag(CN)2]~, [Ag(CN)3]=, and [Ag(CN)4]-. The tricyano complex exists over a wide range of concentrations, the equilibrium constants between the successive com- plexes being K3l2 = 0.20 d= 0.05 and K4,3 = 13.4 ± 4. Under the same conditions, gold (I) forms only [Au(CN)2]~. The gold and silver complexes are both adsorbed on anion exchange resins, but the gold complex is held much more firmly than is that of silver.

Adamson, Welker, and Volpe500 have studied the exchange of radiocy- anide with some heavy metal cyanides. The rate of exchange for complexes in which the metal shows a coordination number of two or four was found to be immeasurably rapid. With hexacyano manganate(III) it is rapid but measurable, and with the other hexacyano complexes it is negligible. Thus, the rate seems to be a function of coordination number rather than thermo- dynamic stability. A more detailed study of the exchange between [Mn(CN)6]- and CN~ showed that the rate of this reaction is proportional to the concentration of cyanomanganate(III), but independent of the con- centration of cyanide ion. The authors postulate the existence of an unstable intermediate, [Mn(CN)6H20]-, in which manganese shows a coordination number of seven. This is possible for manganese (III), but not for chrom- ium(III), iron(III), or cobalt(III).

The cyanide group acts as a bridging group in polynuclear complexes, both the carbon and the nitrogen atoms sharing electrons with the metals. An interesting example of this is found in dipropyl gold cyanide, which has been shown to be tetrameric and to which the structure

R R 501

I I

R— Au— C=N— Au— R

I I

N C

III III

C N

I I

R— Au— N=C— Au— R

I I

R R

Jones and Penneman, J. ('Item. Phys., 22, 965 (1954).

500. Adamson, Welker, and Yolpo, ./. Am. Chem. Soc, 72, 4030 (1950); Adamson,

Welker, and Wrighl , ./ . Am. Chem. Soc, 73, 4786 (1951).

501. Phillips and Powell, I'roc. Roy. Soc. London, A173, 147 (1939).

GENERAL SURVB1 89

has been assigned. The polymeric structure is dictated by the necessity of coordinating tour donor groups to each gold atom.

Upon heating, a compound of the type [R»Au(CN )]i decomposes to form a substance of the empirical formula R An CN, which Gibson802 believes is a linear polymer

R R

I I

— Au— C X— Au— C X— Au—

I I

R R

In spite of the stability of the gold-carbon bond, the tetramer [R2Au(CN)]4 is destroyed by ethylenediamine, giving [R2Au en][R2Au(CN)2]503.

In the "simple" cyanides of the heavy metals, linking between the metal atoms takes place, the complexity of the resulting structure depending upon the relative numbers of cyanide ions and metal atoms, and the coordination number of the latter. Silver504 and gold505 cyanides have been shown to con- tain infinite chains of metal atoms held together by cyanide bridges. Mer- cury! Hi cyanide is also said to have a linear structure506, while the closely related zinc507 and cadmium503 compounds are three-dimensional super complexes. Tetracovalent metals which form planar bonds form layer struc- tures. Thus, palladium cyanide is

i i

— Pd — CseN— Pd—

I I

N C

III III

C N

I I

— Pd— N=C— Pd—

I I

Long509 has studied the rate of exchange between [Ni(CN)4]= and CN"" and between [Ni(CN)4]" and Xi++. The first of these is fast but the second is Blow compared with the rate of precipitation of these ions when they are

502. Gibson, Proc. Roy. Soc. London, A173, 160 (1939).

503. Brain and Gibson, ./. Chem. >SW., 1939, 762.

504. Braekken, K<jl. Norske Yidensk. Sehkohs. Forh.t II 1929, 123; West, Z. Krist.,

90, 555 (1 105. Zhandov and Shugam, Acta Physicochim. V R S S 20, 253 (1945).

506. Hassel, Z. Krist., 64, 218 (1926); Zhandov and Shugam, C.R. Acad. Sci. U.R.S.S.,

45, 295 1944).

507. Zhandov, C. /.'. Acad. Set. \ R 8.S 31. 360 I'M I .

508. Shugam and Zhandov, Acta Physicochim. ('/:.< >'.. 20, _'!7 1946).

509. Long, ./. .1-. Chem. Soc, 73, 537 (1951).

90 CHEMISTRY OF THE COORDINATION COMPOUNDS

mixed with each other. Long concludes that the nickel is bound in two different ways, and that nickel cyanide may be formulated as nickel tetra- cyanonickelate(II), Ni[Ni(CN)J. Hume and Kolthoff510 have come to the same conclusion from polarographic studies.

Heavy metal salts of the hexacyano complexes have been studied ex- tensively, especially the ferro- and ferricyanides. It has long been known that the heavy metal ferrocyanides are not simple salts of H4[Fe(CN)6]. For example, Reihlen and Zimmermann511 showed that ammonia will ex- tract only part of the cadmium from cadmium ferrocyanide. The com- plexity of these materials is indicated also by their great insolubility and their colloidal nature. Turnbull's blue, made from an iron(II) salt and a hexacyanoferrate(III), and Prussian blue, made from an iron(III) salt and a hexacyanoferrate(II), were long thought to be different materials, but both chemical and physical studies have shown them to be identical. This comes about because the ions involved react with each other readily :

Fe+++ + [Fe(CN)6]4- ^± Fe++ + [Fe(CN)6]s 512.

When union between the simple cation and the complex anion takes place, the nitrogen of each cyanide group shares electrons with an iron atom, which in turn shares electrons with nitrogen atoms from other complex anions. Thus, a super complex is built up.

The x-ray studies of Keggin and Miles513 have revealed the structure of the ferro- and ferricyanide pigments. In Berlin green, Fe[Fe(CN)6], which is made by the reaction of Fe+++ and [Fe(CN)6]-, the iron atoms form a cubic, face-centered lattice. (Fig. 1.2). This arrangement is retained in "soluble" Prussian blue (Fig. 1.3), in which half of the iron atoms are in the 3+ state and half in the 2+ state. It is impossible to distinguish between these, and it is probable that they are identical, the charge distribution being levelled out by resonance. One potassium ion (or another univalent ion) must be present for each iron (II) ion to maintain electroneutrality. These univalent cations are located in the centers of alternate small cubes. If all of the iron atoms are in the dipositive state, there is an alkali ion at the center of each small cube; the arrangement of the iron atoms is not changed (Fig. 1.4).

510. Hume and Kolthoff, J. Am. Chem. Soc., 72, 4423 (1950).

511. Reihlen and Zimmermann, Ann., 475, 101 (1929).

512. Bhattacharya, ./. Indian Chem. Soc, 11, 325 (1934); Davidson, J. Chem. Ed.,

14, 238, 277 (1937). 513 Keggin and Miles, Nature, 137, 577 (1936).

GKXKKAL SI L'\ /■:)

Hi-TH

Fig. 1.2. Structure of

Feiii[Feiii(CX)6].

Fig. 1.3. Structure of* Fig. 1.4. Structure of KFe111 [Fe"(CN)6] K2FeII[FeII(CN)6]

A structure similar to this is probably common to all of the heavy metal ferrocyanides, variations being introduced as the nature of the second metal ion is changed. For example, assuming that the coordination number of silver is two, Ag4Fe(CN)6 should be formulated as Ag[Ag3Fe(CN)6]. Since the covalences of silver are linear, each of the coordinated silver atoms must share electrons with the nitrogen atoms of two different Fe(CN)c units, thus forming a giant polymer.

Examples are known in which coordination with carbon tends to stabilize high oxidation states of the metal ions (i.e., the hexacyanocobaltate(TII) Ion), but in most cases, metals coordinated to carbon show very low oxida- tion states. In the metal carbonyls, for example, the metals are in the zero oxidation state and in the salt-like carbonyls and the coordination com- pounds containing ethylenic substances, the metals are always in their lower oxidation states. The same tendency appears in the complex cyanides, as is exemplified by the compounds K2Ni(I)(CN)3 and K4Ni<°>(CN)4 , The compound of monovalent nickel was first prepared by Bellucci and Corelli514 by reducing Kj[Ni(CN)4] with potassium amalgam. Hydrazine can also be used as the reducing agent515, but the best method of preparation involves reduction with metallic potassium, using liquid ammonia as the solvent516. Bellucci and Corelli supposed that it is similar in structure to K2[CuI(CN)3], but the fact that it is diamagnetic shows that it must be a polymer. Mellor and Craig617 proposed that it may be a dimer containing a metal-metal bond, but the x-ray work of Xast and Pfab515 indicates the presence of a double bridge and they write the structure

514. Bellucci and Corelli, Z. anorg. Chem., 86, 88 (1914).

515. Xast and Pfab, Naiurwissenschaften, 39, 300 (1952).

516. Eastes and Burgess, J. Am. Chem. Soc.t 64, 1187 (1942).

517. Mellor and Craig, J. Proc. Roy. Soc. X. S. Wales, 76, 281 (1943).

CHEMISTRY OF THE COORDINATION COMPOUNDS

The complex ion readily adds nitric oxide and carbon monoxide :

K2[Ni(CN)3] + NO -> K2[Ni(CN)3NO]518

K2[Ni(CN)3] + CO -» K2[Ni(CN)3CO]519

The products are actually more complex than these equations indicate; the carbonyl compound, at least, is evidently a polymer, for it is diamagnetic.

The action of excess potassium on K2[Ni(CN)4] in liquid ammonia gives K4[Ni(CN)4]516 as a copper-colored solid of very strong reducing powers. Compounds of the formula K4[M(CN)4] containing palladium520 and co- balt521 have been prepared in analogous fashion.

Kleinberg and Davidson522 have reduced the hexacyanomanganate(III) ion in liquid ammonia, obtaining a product of the formula K5Mn(CN)6- K6Mn(CN)6-2NH3 . They also have evidence for the existence of a cyano complex of chromium (I)523.

Isonitrile Coordination

Metal complexes of the isonitriles have been known for a long time, but have received little attention until recent years. Hartley524 prepared two isomers of "methyl ferrocyanide," and showed that upon treatment with a mixture of alkyl iodide and mercury (II) iodide both isomers were con- verted to [Fe(CH3NC)4(RNC)2]l2-2Hgl2. Both isomers gave the same product when methyl iodide was used, but ethyl iodide gave two isomers525. These have been subjected to x-ray analysis526, and have been shown to be cis- and trans- isomers.

There is a close relationship between the isonitrile-metal complexes and the metal carbonyls. In both, the metal-carbon bond possesses a consider-

518. Hieber, Nast, and Proeschel, Z. anorg. Chem., 256, 145 (1948).

519. Nast and Krakkay, Z. anorg. Chem., 272, 233 (1953).

520. Burbage and Fernelius, J. Am. Chem. Soc, 65, 1484 (1943).

521. Hieber and Bartenstein, Naturwissenschaften, 39, 300 (1952).

522. Kleinberg and Davidson, J. Am. Chem. Soc, 75, 2495 (1953).

523. Davidson and Kleinberg, J. Phys. Chem., 57, 571 (1953).

524. Hartley, J. Chem. Soc., 103, 1196 (1913).

525. Hartley, J. Chem. Soc., 1933, 101.

526. Powell and Stanger, J. Chem. Soc, 1939, 1105.

GENERAL SURVEY 93

able degree of double bond character4980. The isonitrile complexes can be made by displacement of carbon monoxide from metallic carbonyls. Thus, phenyl isonitrile reacts with nickel carbonyl to give |\i(0\C)4] as long,

canary-yellow needli They arc stable, soluble in many organic

Bolvents, 1 >ut insoluble in water. Iron and chromium carbonyls also read, though more slowly. The resulting compounds have not been fully charac- terized.

Methyl isonitrile reacts incompletely with nickel carbonyl, giving [Xi(CO)(CH3XC1)3]. However, the same read ants in the presence of iodine and pyridine give [Ni(CH»NC )<]. A cobalt complex of the empirical formula [Co2(CO)3(<£XC)o] is obtained by the action of phenylisonitrile on Hg[Co(CO).i]2 in the presence of iodine and pyridine. Klages, Monkemeyer, and Heinle-9 have prepared a series of copper(I) complexes, CuCl-x</>XC, (x = 1 - 4), the silver compounds AgXOaCp-CI^Cel^XC)* (x = 2 and 4), and the mercury(II) and zinc compounds MCl2(p-CH3CeH4X'C)2 .

The Nomenclature of Coordination Compounds

Werner's system of nomenclature is the basis for the S3'stem which has been adopted by the International Union of Pure and Applied Chemistry, and which is now almost universally used530. These rules may be summarized p& follows531-532:

(1) If the substance is an electrolyte, the cation is named first, then the anion.

(2) The names of all negative coordinating groups end in -o, but those of 'utral groups have no characteristic ending. In deference to long established

practice, the coordinated water molecule is called aquo.

(3) The numbers of coordinating groups of each kind are indicated by the Greek prefixes mono-, di-, tri-, tetra-, etc., unless these groups are complex. In that case, the prefixes bis-, tris-, tetrakis-, etc., are used.

(4) Xegative coordinated groups are listed first, then neutral coordinated groups, then the metal. (The reverse order is followed in writing formulas of complexes.)

(5) The oxidation state of the metallic element is indicated by a paren- thetical Roman numeral. With cations and neutral molecules, this numeral

527. Hieber, Z. Xaturforsch., 5b, 129 (1950); Hieber and Bockly, Z. anorg. Chem., 262, 344 (1950). Klagea and Monkemeyer, Ber., 83, 501 (1950). B9. Klages, Monkemeyer, and Heinle, Ber., 85, 109, 126 (1952).

530. Jorissen, Bassett, Damiens, Fichter, and Remv, J. Am. Chem. Soc, 63, 889

(1941).

531. Fernelius, I Veto*, 26, 161 (1948): Advances in Chemistry Series,

[81, 9 (1953). American Chemical Society.

532. Fernelius, Larsen, Marchi, and Rollinson, Chem. Eng. News, 26, 520 (1948).

04 CHEMISTRY OF THE COORDINATION COMPOUNDS

follows the name of the metal directly. With anions, the Roman numeral is placed after the name of the complex, which always bears the suffix -ate.

Werner's system differed from this chiefly in the mode of designation of oxidation state of the metal. Werner indicated the oxidation state of the metal in cations by the suffixes -a, -o, -i, and -e, indicating 1 + , 2+, 3+, and 4 + , respectively. In anions, the same suffixes were used, followed by the ending -ate. In neutral molecules, no suffixes were used.

Fernelius and his co-workers531 - 532 have suggested some useful additions to the system adopted by the International Union. The more important of these have been summarized by Moeller as follows533 :

(1) The names of coordinated positive groups end in -ium.

(2) Positive groups are listed last, after negative and neutral groups.

(3) Groups of the same general nature (i.e., all negative, all neutral, all positive) are listed in alphabetical order without regard to any prefixes designating the numbers of such groups present.

(4) Zero oxidation state for the central element is designated by the Arabic character 0 placed in parentheses.

(5) Coordinated hydrogen salts are named as acids by dropping the word hydrogen and replacing the suffix -ate by 4c.

(6) Oxidation state of the central element is designated in the usual manner even though the complex is a neutral molecule.

(7) Use of prefixes such as bis-, tris-, and tetrakis-, followed by the name of the coordinated group set off by parentheses is preferred to that of the old designations di-, tri-, and tetra- to indicate numbers of coordinated groups if the names of those groups are complex.

In both the Werner and the I.U.C. systems, the names of bridging co- ordinated groups (i.e., those which are coordinated to two metal atoms simultaneously) are given after the names of all the other coordinating groups, and are preceded by the Greek letter ju. Bridging groups have their usual names, except the OH group, which is designated as ol.

Geometrical isomers of planar ions may be distinguished either by the terms cis- and trans- or by the numbers 1 ,2- and 1,3-. For octahedral com- plexes, these become cis- and trans- or 1,2- and 1,6-. Where there are more than two kinds of coordinating groups, or more than two of any one kind, the number system is much to be preferred.

The sign of rotation of optical isomers is indicated by d-, I- (or meso-). If the complex contains optically active coordinating groups, the small letter may be used to designate that fact, and the capital letters d- and m to indicate the sign of rotation of the complex as a whole.

These rules are exemplified in Table 1.2.

It is customary, in writing formulas of metal coordination compounds, to

Moeller, "Inorganic Chemistry," New York, John Wiley & Sons, Inc., 1952.

GENERAL SURVEY

1 . 50

CO i'

C co

a c - os

00 00 ~

C C CSI 08 c8 _?

^ **

M U c

C *-

o o cc

.2 CO CD 'S

03 02 C^

'3 °o ^h"

■y.

—

CO •— "^

h h h

CJ {) J) J2 j= — ♦» -*a ■♦*

'S "S 'S

—

8 «

fc- 42 _c

CD CD 0)

° >, >>

s "^ T. T.

o9 o9

ed o3 o3 d

££

££S

.£

.«

o O

63 ^>

i i

< v.

w- -^

- - z z

to "~*

? 5

43 C

o S

1— 1

o r:

■— •

o c

"«

a -

E id

2 3

3 5

> g 1 1

3 5

.9 'g § e

III

CD -^ 42 -TOO y~i -*j o

43 tC CD

42 O

V "o

CD 40

a o

'•*-<

s o9

- z

— ' 43

~ -

tfi

15 c co -

4= ^ £ 2 £ .2

o3 -^ C CD ft

—

JO A

'5 g | J

o c ;

o £ ■£

* 2

-^ G .—

41 -

_= ~

-r

-O O

cj o9 -^.

^~s

= °

— '

C\J

_r:

y^ "*\

E E

, c — c

/ \

r i i

3 -

I OJ^S

U L) V

_ S

ci r"

=?26^

\ // \ /

~ £.^

- - >

<J u

-5

-J-^

- 1^.

& 7 g

r r

- — :

1 I

Table 1.3. Symbols

for Names of

Some Ligands

Name of ligand

Symbol

acetate ion

ac1

acetylacetonate ion

acac

alanine anion

alan

amino acid anion

amac

ammonia

benzidine

bzd

benzoylacetate ion

benzac

benzylamine

bzl

biguanide

BigH

2,3-butanediamine

bn3

isobutanediamine

ibn

citrate anion (monobasic)

cj'anide ion

cy2

1,2, frans-cyclohexanediamine

chxn

1,2, irans-cyclopentanediamine

cptn

diallylamine

dim

2, 4-diaminopentane

ptn

dibenzoylmethane

dibenz

diethylenetriamine

dien

dimeth}-lglyoxime monobasic anion

DMG or HD

2,2'-dipyridyl

dipy

ethylamine

etn

ethylenebiguanide

enBigH

ethylenediamine

ethylenediamine-acetylacetone

enac

ethylenediaminetetraacetic acid

H4Y or EDTA

ethylenethiourea or ethylenethiocarbamide

etu

glycine anion

gly

halide

hydroxyl amine

methyl bis (3-dimethylarsinopropyl)

arsine

TAS

oxalate dibasic anion

1, 10-phenanthroline

o-phen4

phenylalanine anion

4> ala

phenylbiguanide

0 BigH

ortho-phenylenediamine

o-phenylenebis (dimethyl arsine)

PDA

phthalocyanine (dinegative group)

propylenediamine (1 , 2-diaminopropane)

pyridine

st ilbenediamine (1 , 2-diphenylethyle]

lediamine)

stien

2,2',2",2'"-tetrapyridyl

tetrpy

thenoyltrifluoroacetone

TTA

t hiourea

i hiosemicarbazide

thio

1,2, 3-t riaminopropane

2,2',2"-triaminotricthylamine

tren or trin

t net hylenetet ramine

trien

i timet hylenediamine

trim

2, 2', 2" -tripyridyl

tripy

1 Not to be contused with "Ac" used in organic

chemistry to denote the ac

etyl

group.

I Ibsolet e.

2 May be preceded by <l , I, or m (dextro, levo, or meso). 4 Other symbols commonly used are "phenan" and "ph".

Table 1.4. Soin Complex Compoi nds Named Aih.k Thbib Discovebsbs

Name

Strut tare

Cleve'a Salt

ci«-[Pi \n ica4]

Cleve'a Triammine

[Pt(NHi),Cl]Cl

- l'a Firs! Salt

K[Pt(NH,)ClJ

"s Second Salt

K[Pt(NH,)ClJ

Drechsel's Chloride

[Pt(XH3)6]Cl4

Durrani 'a Salt

K4(C204)2Co<^ ^>Co(C204)2] OH

Krdmann's Salt

/■•«/is-K[Co(NH3)2(N02)4]

Fischer'a Salt

K,[Co(NO,),]

( rerard'a Salt

trans-[Pt(XH3)2Cl4]

Gibba' Salt

[Co(NH3)3(N02)3]

Gro's Salt

erona-[Pt(NHa)4ClilClj

Litton'a Salt

Na.[Pt(SO,)4]

Magnus' Green Salt

[Pt(NH,)4][PtCl4]

Magnus' Pink Salt

Two substances of this name are known. The common one is [Pt(NH3)3Cl]2[PtCl4]

Melano chloride

A mixture, chief!}*

XH2

/ \

(XH3)3Co— OH— Co(NH3)3

Oli

\ /

Morland's Salt

CN»H.[Cr(NH,)»(SCN)4]

Peyrone's Salt

c*s-[Pt(XH3)2Cl2]

Recoura's Sulfate

[Cr(H20)5Cl]S04

Reinecke's Salt

NH4[Cr(NH,),(SCN)4]

Rieset's First Chloride

\P\ XH3)4]C12

Rieset's Second Chloride

//•ans-[Pt(XH3)2Cl2]

Roussin's Red Salts

M Fe(NO)«S (M = Xa, K, XH4)

Roussin's Black Salts

M Fe4(XO)7S3 (M = Xa, K, Rb, Cs, XH4 , or Tl

Vaquelin'a Salt

[Pd(XH3)4][PdCl4]

Vortmann's Sulfate

A mixture, chiefly

III/ \III

CNHi)4Co Co(XH3)4

(S04)2

\ /

Ml.

containing some

III/ \IV

CNH Co NH,)4

[804)1

\ /

and other materials

WolfTram's Red Salt

PI ( II \H2)4C13-2H,0— contains Pt(II) and Pt IV

se'a Salt

K[PtCl3C II

''7

CHEMISTRY OF THE COORDINATION COMPOUNDS

Table 1.4 — Continued 1 cf. Gerard's Salt. > cf. Cleve's Salt. Sails of this type, in which the platinum is replaced by other divalent metals, ho ammonia is replaced by other nitrogen bases, or the chloride is replaced by other talides, are often referred to as Magnus salts.

4 Cox, Pinkard, Wardlaw, and Preston, J. Chem. Soc. 1932, 2527.

5 The guanidinium analog of Reinicke's salt.

6 cf. Rieset's Second Chloride.

7 Discovered by Morland in 1861; investigated by Reinicke in 1863.

8 cf. Pey rone's salt.

9 cf. Magnus' Green Salt.

10 Jensen, Z. anorg. allgem. Chem., 229, 252 (1936).

Table 1.5. Some Names of Complexes Based on Color

Name

Color

Structure

Note

Croceo

Yellow

*mns-[Co(NH3)4(N02)2]+

Flavo

Brown

czs-[Co(NH3)4(N02)2]+

Luteo

Yellow

[Co(NH3)6]+++

Praseo

Green

/mns-[Co(NH3)4Cl2]+

Purpureo

Purplish-red

[Co(NH3)5Cl]++

Roseo

Rose-red

[Co(NH3)5H20]+++

Violeo

Violet

czs-[Co(NH3)4Cl2]+

1 Often used to denote other halopentammines, sometimes with a designation as to the halogen present; thus, [Co(NH3)5Br]++ is referred to as the bromopurpureo

use symbols for the names of many organic ligands. Table 1.3 lists the symbols used in this book, as well as some others which may be encountered in other reading. Unfortunately, there is not complete uniformity in the use of these abbreviations, which may lead to some confusion. Because of this, some notes and recommendations are included in Table 1.3.

The early workers in the field of complex inorganic compounds did not understand the nature of these substances, so were not able to give them names based upon structure. It was customary, therefore, to name each compound after its discoverer. A few of these early names persist in the current literature, and are listed in Table 1.4.

In 1840, Fremy634 suggested that the ammines of cobalt be given names descriptive of their colors. He derived such names from the Latin. The sys- tem was easily extended to the cobalt compounds containing ethylenedi- amine and other amines, and to the chromium(III) salts, the colors of which are similar to those of their cobalt(III) analogs. These names are now frequently used to describe classes of compounds. For example, the term

534 Fremy, Ann. chim. phys. [3], 35, 257 (1852); J. prakt. Chem., 57, 95 (1852).

GENERAL SURVEY 99

"luteo," originally used to describe the ion [Co(NH3)6]:i+, has been extended to include also [Co en3]3+, [Co dipy3]3+, [Co trien2]:H~, and other cobalt (III) complexes in which six amine nitrogen atoms are coordinated to the cobalt. The terms are sometimes used to describe ammines of metals other than cobalt and chromium, even though the colors are quite at variance with the names suggested by Fremy. For example, Gleu and Etehm686 use the term "luteo" in reference to the hexammine ruthenium(III) ion, which Is colorless. When any metal other than cobalt is meant, it is usual to in- clude the name of the metal. Thus, luteo chromium(III) chloride is [Cr(XH3)6]Cl3 . The more important of Fremy's "color names" are as- sembled in Table 1.5.

535. Gleu and Rehm, Z. anorg. allgem. Chem., 227, 237 (1936).

A. The Early Development of the Coordination Theory

John C. Bailor, Jr.

University of Illinois, Urbana, Illinois

The history of chemistry in the nineteenth century is largely an account of the growth of our knowledge of molecular structure. When the doctrine of constant valence proved so successful in explaining the structures of or- ganic compounds, it was natural that every effort should be made to apply it also to the structures of inorganic substances. Thus it happened that the growth of inorganic chemistry was retarded for over twenty years by the same factor which contributed most to the phenomenal development of our knowledge of the compounds of carbon. Inorganic chemistry is the older of the two fields, and the study of inorganic "complex compounds" antedated the rise of organic chemistry by over fifty years. The structures of hydrates, double salts, and metal ammonia compounds were widely discussed even before the beginning of the nineteenth century. Of these, the ammonia com- pounds attracted the most attention, for they lent themselves to study by classical methods. The early history of the theory of complex compounds is therefore the history of the ammonates.* The discovery of these substances is usually attributed to Tassaert1, who observed in 1798 that cobalt salts combine with ammonia.

Early Theories of the Structure of Ammines

Berzelius' Conjugate Theory

The first logical attempt to explain the metal ammonia compounds was made by Berzelius2, who observed that a metal in "conjugation" with am- monia did not lose its capacity for combination with other substances. He

* The term "ammonate" was displaced by the simpler term "ammine" at the sug- gestion of Werner.

1. Tassaert, Ann. chim. phys., [1] 28, 92 (1798).

2. Berzelius, "Essai sur la theorie des proportions chimique et sur Pinfluence chimi-

que de l'electricite," Paris, 1819.

100

EARLY DEVELOPMENT OF THE COORDINATION THEORY 101

attempted to extend this theory, but without great success, to the double salts and complex cyanidi

Graham's Ammonium Theorx

According to Graham's "ammonium" theory*, metal ammonates are considered to be substituted ammonium compounds. This view, in one form

or another, was generally accepted until the time of Werner. (iraham made this suggestion in an attempt to explain the structure of diammoniuin copper II salts, in which he supposed one hydrogen atom from each of two ammonium groups had been displaced by copper. Obviously, such a formula can apply only when the number of ammonia groups in the mole- cule is the same as the electrovalence of the metal — a condition which usually does not hold. Gerhardt4, Wurtz6, Rieset6, A. W. Hofmann7 and Boedecker8 suggested modifications of the theory to take care of other cases. According to Rieset and Hofmann, the hydrogen atoms of an ammonium group are replaceable, not only by metals, but also by other ammonium groups. Hofmann represented the compound of cobalt (III) chloride with six molecules of ammonia, for example, as

Co/XH,- XH4

0/XH2- XHA

J. ).

Some years later the experiments of Jorgensen showed this argument to be fallacious. It does not allow for the existence of similar compounds of ter- tiary amines9, and it does not explain why the removal of one molecule of ammonia completely alters the function of one of the chlorine atoms. Boedecker avoided the branching of the chain by assuming that the metal substitutes in an ammonium group which is itself a substituent group: Co(XH3 — NH3 — Cl)3 . The diammonate and tetrammonate of platinum(II) chloride were represented as Pt(XH3— Cl)2 and Pt(XH3— XH3— Cl)2 . The question "What prevents further lengthening of the ammonia chain?" was never answered, and was an insurmountable objection to this type of theory.

3. Graham, "Elements of Chemistry," London, 1837. This book is rare, and is best

known in Otto's German translation "Lehrbuch der Chemie" Braunschwieg, 1840. Graham's suggestion of the ammonium theon- appears in Vol. 2, page 741 of the German edition.

4. Gerhardt, Jahresber. Fortshr. pharm., tech. chem. physik < Living), 3, 335 (1850).

5. Wurtz. .!/.//. rhim. phys., [3] 30, 488 (1850).

6. Rieset, Ann. chim. phys., [3] 11, 417 (1844

7. Hofmann, Ann., 78, 253 (1851).

8. Boedecker, Ann., 123, 56 (1862).

rgensen, J. praht. Chem., [2] 33, 489 (1886).

102 CHEMISTRY OF THE COORDINATION COMPOUNDS

Glaus' Theory

The theory of Claus10 met with vigorous opposition, but is all the more interesting on that account, for the parts of it which were most vigorously attacked appeared in only slightly modified form in Werner's theory. Claus believed that, when combined with metallic oxides, ammonia not only does not affect the saturation capacity of the metal, but becomes "passive" as regards its own basicity. His views may be summarized as follows:11

(1) The union of several equivalents of ammonia with one equivalent of a metal chloride leads to the formation of a neutral substance, in which the basic property of ammonia is lost, so that the ammonia can no longer be determined by the usual means nor eliminated by double decomposition. Thus, the ammonia is in a different condition than in ordinary ammonium salts. This hypothesis met with a storm of protest, just as Werner's similar suggestion did forty years later. The attack was led by Weltzein12, who held the term "passive molecule" to be indefinite and confusing, and who be- lieved that every part of a molecule influences every other part, so that no part can be said to be "passive".

(2) If these chlorides are converted to oxides, strong bases are formed. The saturation capacity of these is the same as that of the metal oxides themselves, and cannot be calculated from the number of ammonia mole- cules combined with the oxide. Schiff13 criticized this conclusion by pointing out that the oxides of the "ammonia bases" of the metals are much stronger bases than the metal oxides themselves. This criticism seems to rest on a confusion between the "strength" of a base and its "saturation capacity" (i.e., equivalence). It is true that the hydroxides of the metal ammines are strong bases, but the ammonia present in them does not readily combine with the hydrogen ion.

(3) The number of ammonia molecules combined with a molecule of metallic salt is determined by the same factors as the number of molecules of water in the hydrate and the two will be the same. This point of Claus' theory was easy to attack, for many hydrates were known for which analo- gous ammonia compounds did not seem to exist. The conclusion which Claus drew, however, was restated as an integral part of Werner's theory and has been amply verified.

Blomstrand's Chain Theory

Odling14 suggested that metallic atoms can substitute for the hydrogen atoms in ammonia just as organic radicals do. The diammonate of plati-

10. Claus, "Beitrage zur Chemie der Platinmetalle," Dorpat, 1854; Zentralblatt, 25,

789 (1854);,4nn.,98,317 (1856).

1 1 . Reitzenetein, Z. anorg. Chem., 18, 152 (1898).

12. Weltzein, Ann., 97, 19 (1856).

13. Schiff, Ann., 123, 1 (1862).

14. Odling, Chem. News, 21, 289 (1870).

EARLY DEVELOPMENT OF THE COORDINATIOh THEORY 103

num(II) chloride was construed as being analogous to ethylenediamine hydrochloride: Pt(NH,),-2HCl and C,,lI.J(MI2),-i-)Il('l. The chaining of ammonia molecules was compared to the chaining of methylene groups in

the hydrocarbons.

Blomstrand1' made this the basis of his famous theoiy. Ammonium chloride was represented as 11 Ml, CI, XII,X(),XH, as II ML Nil, N( >, , and MI,I r>MI3 as H(XH3)7L The terminal hydrogen atom can be replaced by other positive atoms, such as metals. The metal, in fact, stabilizes the chain, and its nature4 determines the length and stability of the chain. Chains of three ammonia molecules are often found in union with nickel, cobalt, iridium and rhodium, but platinum and copper seem unable to stabilize chains of more than two nitrogen atoms. On the basis of these postulates, Blomstrand wrote the formulas for the tetrammonate of platinum(II) chloride and the hexammonate of cobalt(II) chloride as

\ II — XH3— CI XH3— XH3— XH3— CI

/ /

Pt and Co

\ \

X 1 1 ,— XH. — CI NH3— XH3— XH3— CI

According to Blomstrand, the stability of the ammonia chain is not dependent on its length. Although platinum is unable to stabilize chains of any great length, platinum(II) chloride ammonate is not attacked by hy- drogen sulfide or by sodium hydroxide. Chlorine oxidizes the platinum without attacking the ammonia, converting the compound to:

CI XH3— NH3— CI

1/ Pt

l\ CI XH3— XH3— CI

in which chlorine is attached to the molecule in two different ways. The validity of this postulate is borne out by experiment, for only half the chlorine is replaced by the action of sodium carbonate, and the second half i- only slowly precipitated by silver nitrate. Blomstrand referred to the two types of chloride as the "farther" and "nearer". This expression may have inspired Werner's postulate of "first" and "second" spheres11.

JorgenserTs Theories

Blomstrand's formulas for the cobalt ammonia compounds became the center of a Long controversy between Jorgensen and Werner, and are there- fore of considerable interest. Blomstrand believed and the belief wa>

15. Blomstrand, "Chemie der Jetztzeit," Heidelberg, 1869; Ber., 4, 40 (1871).

104 CHEMISTRY OF THE COORDINATION COMPOUNDS

universal until 189016 — that cobalt(III) chloride and its ammonia com- pounds were dimolecular. In that year, Jorgensen adduced evidence for the simpler molecular weights, and halved Blomstrand's formulas. This did not affeel the postulates of Blomstrand's theory, but without this change, Werner's theory might not have been conceived. Blomstrand first supposed the Luteo cobalt salts (e.g., Co2Cl6-12NH3) to have the completely sym- metrical structure:

NH3— NH3— CI NH3— NH3— CI NH3— NH3— CI NH3— NH3— CI NH3— NH3— CI [NH3— NH3— CI

Co2

and the purpureo salts (Co2Cl6 • 10NH3) the structure :

NH3— CI NH3— NH3— CI NH3— NH3— CI NH3— NH3— CI NH3— NH3— CI NH3— CI

Co;

But this was soon seen to be incorrect, for the purpureo salt contains chlo- rine in two very different modes of combination11, 17.* In a cold solution, silver nitrate precipitates two-thirds of the chlorine at once, and the other third only after long standing. The slight functional difference shown in the formula above can hardly explain such a difference in behavior. Jorgen- sen18 prepared a whole series of salts in which the more readily precipitated chlorine is replaced by other groups. He concluded that the chlorine in these salts is combined directly with the metal, while the other negative groups are united with the ammonia. Similar relationships were shown to hold for the chromium19 and rhodium20 pentammonate salts. Jorgenson also dem- onstrated that the "masked" chloride can be replaced by bromine21, sul- fate22, and other negative groups. These groups, like the chloride in the original purpureo salt, have lost their ionic properties.

* For explanation of nomenclature, see Chapter 1.

16. Jorgensen, J. prakt. Chcm., [2] 41, 429 (1890); Petersen, Z. phys. Chem., 10, 580

(1892).

17. Gibbs and Genth, "Researches on the Ammonia Cobalt Bases," Washington,

1856.

18. Jorgensen, J. prakt. Chcm., [2] 18, 209 (1878).

19. Jdrgenaen, ./. prakt. ('hem., [2] 20, 105 (1879); 25, 83 (1882).

20. Jorgensen, J. prakt. Chem., [2] 25, 346 (1882); 27, 433 (1883); 40, 309 (1886).

21. Jorgensen, J. prakt. Chcm., [2] 19, 49 (1879).

22. Jorgensen, J. prakt. Chem., [2] 31, 262 (1885).

EARLY DEVELOPMEXT OF THE COOUD1 X ATIOX THEORY 10.")

When Jorgensen found11 thai two-thirds of the chlorine in the tetram- monates of the trivalent metals is "masked", he concluded that this should be represented as in direcl union with the metal. He formulated these salts as:

Co:

N 1 1

-XH3— NH3— CI -NH3— NH3— CI

and the purpureo and the luteo salts as:

Co,

MI. CI

MI -XH3— XH3— XH3— CI ) XH3— NH3— NH3— NH3— CI

MI.C1

and

Co2<

(NH3CI NH3C1

NH3— NH3— NH3— NH3— CI NH3— NH3— NH3— NH3— CI NH3— CI NH3-C1

Jorgensen showed that the "roseo" salts, which had been thought to be isomeric with the purpureo salts, contain two molecules of water24. This water is lost at elevated temperatures, leaving a residue of the purpureo salt. The roseo salts resemble the luteo salts in that all of the negative groups are ionic as well as in solubility, crystalline form, and appearance. Jorgensen concluded that they are luteo salts in which one-sixth of the ammonia molecules are replaced by water.

The roseo tetrammonate salts were also shown to be analogous to the luteo salts, but they contain water in place of one-third of the ammonia molecules. Xo compounds were known in which more than a third of the ammonia was replaced by water, so it was assumed that the "unchained" ammonia molecules were the ones replaced. The roseo tetrammonia salts were therefore represented as:

Co2s

H20— CI H20-C1 NH3— XH3-XH3 NH3— NH3— NH3 H20— CI H20— CI

-NH3— CI -NH3-CI

or, using the simplified formula, as:

H20— CI

/ Co— H20— CI

MI — XH3— XII — XII

-CI

23. Jorgensen, J. prakt. Chem., [2] 27, 433 (1883).

2-4. Jorgensen, J. prakt. Chem., [2] 29, 409 (1884); 31, 49 (1885),

10() CHEMISTRY OF THE COORDINATION COMPOUNDS

These postulates suggest many questions, some of which Jorgensen at- tempted to answer by modifications or elaborations of the theory:

Why can cobalt hold only six ammonia molecules? If one of the valences holds a chain of four, why cannot the others also? Can chains of more than four ammonia groups exist? How shall we explain the existence of isomeric compounds?

Jorgensen felt that the chains contain a maximum of four — NH3-groups25, because of the many examples of tetrammonated compounds, and because the penta- and hexaammonated salts seemed to contain one and two am- monia molecules, respectively, which are different from the other four. He answered the other questions by developing Blomstrand's hypothesis that the three valences of cobalt are different. An example or two26 will illustrate the argument: The luteo chloride,

7NH3CI

M <*(NH3)4C1 (M represents a trivalent metal) 0NH3C1

readily loses one molecule of ammonia to form

7C1 M«(NH3)4C1, /3NH3CI

which in water is converted to the aquo (roseo) salt, which must therefore be

7H20— CI M«(NH3)4C1. /3NH3C1

The diaquo roseo salt,

7H20— CI

M<*(NH3)4C1

/3H20— CI

readily loses one molecule of water to form a compound which must contain the groups — CI and — H20 — CI. But the — H20 — CI group in this com- pound is not like the — H20 — CI group in the roseo pentammine. The former loses a molecule of water when heated to 100°C or lower, while no water is lost from the latter until the temperature is well above 100°C. According to Jorgensen, this difference indicates that the 0 and 7 valences are not the same. He cited the fact that the tetrammonates take up one molecule of

25. Jorgensen, Z. anorg. Chem., 5, 147 (1894).

26. Jorgensen, Z. anorg. Chem., 7, 289 (1894).

EARLY DEVELOPMENT OF THE COORDINATION* THEORY 107

ammonia or water easily, and a second with difficulty, as further evidence

for this view. The isomerism of the "flavo" and "croceo" chlorides was explained by i lie formulas:

yNO 7NOj

Coo Ml CI and CoaNOj UNO 0(NH,)4C1

Early Theories of the Structure of Hydrates

While these theories of the metal ammonia compounds were being dis- cussed, attempts were also being made to elucidate the structures of the hydrates. The best known of the hydrate theories was that of Wurtz27, who postulated that the water molecules link themselves to the metal and to

each other in rings:

H20— H20 H20— H20— H20

/ \ / \

S04 Cu H20 and S04 Mg H20

\ / \ /

H,0— H20 H20— H20— HoO

The assumptions underlying the theory were unsupported by experimental evidence, and it met with little favor.

Early Theories of the Structure of Double Salts

The double salts, especially the double halides, were of great interest, and numerous theories of their constitution were advanced. Bonsdorff28 and Boullay29 compared the chlorides to oxides, some of which are acidic and others basic, and they supposed double salts were formed by a sort of neutralization reaction. Others3031 took exception to this theory, but it found wide acceptance. Xaquet32 expressed the view that two chlorine atoms are equivalent to one oxygen, and Blomstrand15 went so far as to suppose these two chlorine atoms to be linked together through a double bond. On this basis 3KClFeClj and 2KClPtCl4 become

C1=C1— K CI C1=*C1 — K

/ \ /

Fe— C1=C1— K and Pt

\ / \

C1=C1— K CI C1=C1— K

27. Wurtz, "La Theorie Atomique," Paris, 1879.

28. Bonsdorff, Ann. ckim. phys., 34, 142 (1827).

29. Boullay, Ann. ckim. phys., 34, 337 (1827).

30. Liebig, Ann. ckim. phys., 35, 68 (1827).

31. Bcrzelius, Jahresbcr. Forfsch. chem. mineral. (Berzelius), 8, 138 (1829).

82. Naquet, "Principea de Chemie fondee sur les Theories Modernes," Paris, 1867.

108

CHEMISTRY OF THE COORDINATION COMPOUNDS

There was little experimental evidence to support Blomstrand's suggestion, and it was not widely accepted33. Such formulas do not indicate why the potassium should be ionic and the iron and platinum nonionic, nor do they allow for the formation of double chlorides such as CdCl2-4KCl, in which the number of molecules of alkali metal chloride exceeds the number of chlorine atoms in the heavy metal chloride. Remsen34 "solved" the latter difficulty by assuming the formation of halogen rings:

K— CI CI— K

CI— Cd— CI

/ N

K— CI CI— K

In 1885 Horstmann35 wrote the reaction:

CI CI CI CI CI

\l \\^

Pt + 2KC1 -» Pt— K

/I /l\

CI CI CI CI K

in analogy to

H H H

H— N + HC1 -* H— N

which was the generally accepted mechanism for the reaction of ammonia with hydrochloric acid. By assuming large enough valences for the metals, we can apply this theory to complexes of all sorts. It is, of course, mislead- ing in its implication that all of the groups are attached to the central atom in the same way (the chlorine and the potassium, in the example given). With this feature modified, Horstmann's formulas become almost identical with those of Werner.

Werner's Coordination Theory

This, then, is the background on which Werner built. In his paper "Con- tribution to the Theory of Affinity and Valence"36 published in 1891, he suggested thai an atom does not have a certain number of valence bonds, but that the valence force is exerted oxer the whole surface of the atom, and

33. Remsen, .1///. Chem. J., 11, 291 (1889).

34. Remsen, Am. Chi »,.J., 14, 81 (1892).

35. Horstmann, "Lehrbuch der Physikalischen und Theoretischen Chemie," Braun-

Bchweig, 1885.

36. Werner, "Beitrage sue Theorie der Affinital und Valenz," 1891.

EARLY DEVELOPMENT OF THE COORDINATION THEORY L09

can l>o divided into several units of varying strength, depending on the demands of the atoms which unite with it. Sonic of its valence force may be left unexpended. This thoughl is differenl from the postulate of "primary" and "secondary" valences, but is certainly a forerunner of it. The wide- spread belief that the coordination theory had no roots in earlier theories or in the experience of its author is a mistaken one. It is true, however, thai the theory was different from anything which had previously been proposed and that it came in a spectacular way. Pfeiffer"7 lias writ ten : "According to his own statement, the inspiration came to him like a flash. One morning at two o'clock he awoke with a start ; the long-sought solution of this prob- lem had lodged in his brain. He arose from his bed and by five o'clock in the afternoon the essential points of the coordination theory were achieved." Werner was then twenty-six years old.*

Fundamental Postulates

The fundamental postulate in Werner's coordination theory is stated in the following way88 "Even when, to judge by the valence number, the com- bining power of certain atoms is exhausted, they still possess in most cases the power of participating further in the construction of complex molecules with the formation of very definite atomic linkages. The possibility of this action is to be traced back to the fact that, besides the affinity bonds desig- nated as principal valencies, still other bonds on the atoms, called auxiliary valences, may be called into action." The rest of the theory is an elucidation of the nature, the number, and the spatial distribution of these "auxiliary" valences. t The auxiliary valences were originally conceived as being very different from principal valences, since they do not allow ionization while the principal valences do. Yet according to Werner, there is a connection between them, for if an atom forms strong primary bonds with certain other atoms, ii usually forms strong secondary bonds with them too. Thus the alkaline earth oxides are extremely stable, and they combine with water (by secondary valence) with great avidity. Similarly, the very stable sul- fides of copper, mercury and arsenic readily form thio complexes. It is pos-

For biographical sketches of Werner, see G. T. Morgan: ./. Chem. Soc., 117, 1639

1920); J. Lifschitz, Z. Elektrochem., 26, 514 (1920); and I'. Karrer, Helv. ckim. Acta,

3, l'.»6 (1920). These give brief accounts of his theory. The art iele by Karrer contains a

portrait and B list of Werner's publications. P. Pfeiffer, /. Chem. Ed., 5, 1090 (1928)

gives a description of Werner's personal life and a portrait of him.

t The terms "primary" and "secondary" were often used instead of "principal" and "auxiliary."

Pfeiffer, ./. Chem. A''/., 5, L096 L928); Ostwald's "Klassiker der Exakten Wissen

schaften," No. 212, p. ">, Leipzig, Akademiache Verlagsgesellschaft, 1924. Werner, "Neuere Anschauungen," 1th Ed. p. 44, \ ifwi^, Braunschweig, 1920. Quoted from Bass' translation of Schwarz, "The Chemistry of Inorganic Com plex Compounds," p. '.», \ew York, John Wiley a- Bona, [nc, i'»23.

110 CHEMISTRY OF THE COORDINATION COMPOUNDS

sible, too, for a primary valence to be converted into a secondary one. In solutions of hexammine chromic chloride, [Cr(NH3)6]Cl3 ,* all of the chlorine is at once precipitated by solutions of silver nitrate. If the dry hexammine be heated somewhat above 100°C, a molecule of ammonia escapes, and simultaneously one-third of the chlorine loses its ionic properties. Werner argued that this means it has become attached by a secondary valence, though of course this does not release a primary valence, and the new com- pound contains only two chloride ions39. Jorgensen and Werner both be- lieved the nonionic chlorine to be attached directly to the metal, in place of the ammonia which had been lost. On standing in water solution, the pentammine undergoes a slow change by which the third chlorine again becomes ionic. Upon evaporation at room temperature, the resulting solu- tion yields crystals of a rose-red pentammine, containing a molecule of water. Jorgensen40 had shown that this "roseo" compound is closely analogous to the hexammine, and he recognized it as a hexammine in which one ammonia molecule is replaced by water. In this, he and Werner agreed. They dis- agreed, however, on the fate of the chlorine atom which the water molecule had displaced. Jorgensen believed it to be attached to this water molecule through the quadri valence of oxygen while Werner felt that it was not at- tached to any particular atom in the complex, but was attracted by the complex ion as a whole. Werner's postulate clearly foreshadows the theory of ionization of salts in the crystalline state, and has been amply confirmed by x-ray measurements and by other means. At the time of its proposal, however, it was a most revolutionary doctrine, and for many years it met with widespread criticism41.

The relationship between primary and secondary valence became closer and closer in Werner's mind, and he was finally led to the conclusion that there is no essential difference between the two. This came about through his study of the tetrakis(ethylenediamine)-ju-amino-nitro-dicobalt(III) ion,

* The term "amrain" proposed by Werner to designate the metal ammonia com- pounds, is translated into English as "ammine". Its use in this place is somewhat anachronic, as it was not used in Werner's earlier papers, but we shall use it through- out. The term "ammonate" is still used by some authors to designate simple addition compounds of ammonia with metallic salts. Such compounds can be called "ammines" equally well, however. In the earlier papers, Werner indicated the constituents of the complex ion by enclosing them in parentheses, but he later adopted the use of square brackets.

39. JSrgensen, ./. prakt. Chem., [2] 20, 105 (1879).

40. Jorgensen, ./. prakt. Chem., [2] 29, 409 (1884).

41. See for example, Friend, ./. Chem. Soc, 109, 715 (1916); 119, 1040 (1921).

EARLY DEVELOPMENT OF THE COORDINATION THEORY

ens Co Co enj

\ / NO

This ion contains two asymmetric cobalt atoms (See Chapter 8) which ap- parently arc not identical. One of them is attached to the amino group by a primary valence and to the nitro group by a secondary valence, while for the other one, these relationships are reversed. Resolution, then, should give a dextro, a levo, and two meso forms. Careful experimentation, however, yielded only one meso form. This compound is completely inactive, indicat- ing the identity of the two asymmetric atoms. Werner may not have been surprised at this discovery, for his first paper43 draws an analogy between the metal ammine ions and the ammonium ion, in which the hydrogen which is held by "secondary" valence is indistinguishable from the rest.

It has long been known that many of the metal ions form hexammonates and hexahydrates, and that tetraammonates are common. The tetra- and hexacyanides have also long been known as stable, well-defined compounds. From such facts, Werner deduced that each element has only a certain number of secondary valences. Groups attached to the central element by these valences are said to be "coordinated" to it. The "coordination number" of an atom or ion is the number of groups which can be coordinated to it.* While four and six are the most common coordination numbers, coordina- tion numbers of two, three, five, seven and eight are known.

In terms of Werner's theory, the secondary valences of an atom must be satisfied. In the case of hexamminechromium(III) chloride, if a molecule of ammonia is driven out, one of the chloride ions will take its place to main- tain the coordination number six. A wide variety of neutral groups or nega- tive ions can enter the coordination sphere. When these latter become co- ordinated, they cease to be ions, of course, and this is indicated by the suffix -o on their names or abbreviated names; thus, "cyano," "chloro," "nitro," and "hydroxo".

If a trivalent metal hexammine chloride loses one molecule of ammonia, one of the three chlorides loses its ionic properties, as has been pointed out. If a second molecule of ammonia is lost, a second chloride becomes non- ionic41. What will happen if a third ammonia molecule is lost? According to

* When applied to the structure of crystals the term "coordination Dumber" is given a somewhat different meaning; it refers to the number of atoms (or ions) which surround the atom or ion i in question, and arc ;it equal distances from it, no matter what the natin<- of i he bond between them.

VI Werner, Ber., 46, 3674 (1913); 47, L964, 1978 I'd I

13. Werner. Z. anorg. Chem., 3, 267 18

1 1 . Jdrgensen . Z anortj. ('firm ., 5, 117 (1894).

112 CHEMISTRY OF THE COORDINATION COMPOUNDS

Jorgensen 's own statement44, he had never considered this point, but it became very important, for his theory and Werner's predicted different behaviors. According to the coordination theory, the third chloride should become nonionic, and a nonelectrolytic molecule should result. Jorgensen had to assume that his postulated ammonia chain would simply be shortened by one nitrogen atom, which would still leave the chloride in the ionic state. Very few triammines of trivalent metals were known at that time, and when Werner pointed out43 that their properties supported his own theory, Jorgensen objected44 that the compounds were not sufficiently understood to justify the conclusion.

One of these compounds, Ir(NH3)3Cl3 , had been described by Palmaer45, who found that it did not liberate hydrochloric acid when heated with con- centrated sulfuric acid. He suggested that it had twice the simplest formula, and was a double salt, Ir(NH3)6Cl3-IrCl3 . Jorgensen showed that a cor- responding rhodium double salt could be prepared from the components, and that it did not liberate hydrogen chloride when warmed with sulfuric acid. He pointed out also that Magnus' salt Pt(NH3)4Cl2-PtCl2* is resistant to concentrated sulfuric acid, and concluded that this reagent cannot be relied upon to indicate the presence pf ionic chlorine.

The other example cited by Werner was Erdmann's Co(NH3)3(N02)346, which was admittedly not a well characterized compound47. Several sub- stances of the same composition had been discovered, and Erdmann's description of his compound was incomplete. Investigation of the com- pound convinced Jorgensen that it has the structure

N02

/ Co— NH3— N02

NH3— NH3— NO 2

He converted it to the chloride, which however, contains one molecule of firmly held water; to this compound he assigned the structure

H20— CI

/ Co— NH3— CI

NH3— NH3— CI

* We would now give these "double salts" the formulas [Ir(NHs)e] [IrCl6] and [Pt(XH3),] [PtCl*], which indicate that they do not contain chloride ions. 15. Palmaer, Oefvers, af k. Vet. Acad. Fdrh, No. 6, 373 (1889); Ber., 22, 15 (1889). L6. Erdmann, ./. prakt. Chew., 97, 412 (1866). 47. Gibbs, Proc. Amer. Acad., 10, 16 (1875).

EARLY DEVELOPMENT OF THE COORDINATION THEORY 113

because all of the chlorine is precipitated at once by silver nitrate. This compound is readily converted to Erdmann's "trinitrite," which must then

have the structure shown.

The tWO theories differ also in their predictions a> to the roult of the loss oi another molecule of ammonia, with the production of a diammine. No such compounds were known and this was in accord with Werner'.- theory. To him, an ammonia molecule cannot he "lost"; it must he replaced by another group. Thus far in the process, the halide ions which accompany the complex have been able to carry out this replacement, hut now ;i new group must he supplied. If this be a negative ion, it will give the complex a negative charge. Keinecke's salt, NH4[Cr(XH3)2(SCX)4]48 and Erdmann's salt. XII; [Col XH3)2(X02)4]49 * are examples of this type of compound. There were no examples of the monoammonates, Ms'tM'^NHaXs], pre- dicted by Werner, but numerous examples of the final step in the replace- ment were known; e.g., the heavy metal cyanides, the cobaltinitrites, and the double chlorides.

The tetravalent elements furnish a similar series. Platinum(IV) chloride yields ammines containing six, five, four, three, two, and one molecules of ammonia. All the chloride is readily removed from the first of these. Blom- strand15 had observed that two of the four chlorine atoms in the tetram- monate are much less reactive than the other two. There are two isomeric forms of the diammonate, which therefore elicited great interest. In accord- ance with the demands of Werner's theory, both of these are nonionic. The end member of the series is potassium hexachloroplatinate(IV), which does not react with silver nitrate to give silver chloride, but gives silver chloro- platinate, Ag*[PtCU].

Conductivity Studies

To give further support to these views, Werner and Miolati measured the conductivities of a large number of metal ammines50. Again, the results med to substantiate the coordination theory, but Emil Petersen51 raised objections to this conclusion. The number of ions found was in some cases greater than predicted by the theory. A case in point is Co(XH3)3(X02)j(,l. which the theory demands musl be a nonelectrolyte, but which showed the conductivity of a uni-univalent electrolyte. Werner explained this by assum- ing the reaction [Co(NH,),(N02),Cl] + H20 -> [C0(XH3)3(X02)2H20)C1,

* Erdmann's salt is not to be confused with Erdmann's trinitrotriamminecobalt (III), mentioned above. 48. Reinecke, Ann., 126, 113 (1863).

Erdmann, ./. prakt. char., 97, 406 (1866).

50. Werner and Miolati, Z. pkysik. Ch -.12, 35 (1893); 14, 506 (1804 ; 21, 225 (1896).

51. Petersen, Z pi 22, 410 (1897).

114 CHEMISTRY OF THE COORDINATION COMPOUNDS

T \ ble 2.1 . Effect of Aging on the Molar Conductivity of an Aqueous Solution

of [Co(NH3)4Br2]Br (Molar Concentration, 0.2%)

m = Freshljr prepared solution 190.6

5 minutes after the first measurement 288.0

10 minutes after the first measurement 325.5

15 minutes after the first measurement 340.7

20 minutes after the first measurement 347.8

40 minutes after the first measurement 363.5

and supported this by the fact that at 0°C, where the hydration reaction cannot proceed readily, the conductivity is indeed very low. Petersen countered by pointing out that all salts show much lower conductivities at 0° than at 25°C.

Werner and Miolati reported several instances of this kind, and in some of them, had good evidence that reaction with the water does take place. The dark green Co(NH3)4Br3 dissolves to give a deep green solution, which rapidly becomes red. At the same time, the conductivity rises, as shown in Table 2.1. It seems to approach that of the diaquotetrammine salt (see Table 2.2), which is bright red. Werner and Miolati wrote the equation:

[Co(NH3)4Br2]Br + 2H20 -> [Co(NH3)4(H20)2]Br3

The "dichro" salt, Co(NH3)3(H20)Cl3 gave similar results, the solution turning from green through blue to violet;

[Co(NH3)3(H20)Cl2]Cl + 2H20 -> [Co(NH3)3(H20)3]Cl3 .

This reaction proceeds so rapidly at room temperature that Werner and Miolati made their conductivity studies at 1°C. The molecular conductivity was compared with those of potassium chloride, barium chloride, and hexamminecobalt(III) chloride at the same temperature, and found to correspond to that of the first; in other words, the salt is composed of two ions.

With those compounds which do not contain readily displaced groups in the coordination sphere, Werner and Miolati obtained results entirely in accord with their expectations. Many of their results are elegantly shown in graphical form in the second paper of their series, and two are reproduced in Figs. 2.1 and 2.2. The conductivities of aquoammine salts are significant in that they support Werner's contention that water molecules and am- monia molecules occupy equivalent positions in the coordination sphere. Some of these are shown in Table 2.2. Petersen51 repeated some of this work,

EARLY DEVELOPMENT OF THE COORDINATION* THEORY

115

522 9

256

[Pt(NH3)6]ci4 [Pt(NH3)5Cl]cis [Pt(NH3)4CI2]ci2 [Rt(NH3)3CI3]ci

>^NH3)2CI4]

K[Pt'(NH3)C,5] K2[PtCI6]

Fig. 2.1. The molar conductivities of 0.1 molar per cent aqueous solutions of some platinum (IV) ammines.*

99.29

A- [C0(NH3)6]CI3

B- [C0(NH3)5(N02)]CI2

c- i,6[co(nh3)4(no2)2]ci

D- [C0(NH3)3(N02)3] E- K[C0(NH3)2(N02)4]

Fig. 2.2. The molar conductivities of 0.1 molar per cent aqueous solutions of some cobalt (III) ammines.

Table 2.2. Molab Conductivities of Some Cobalt(III) Ammines at

Various Dilutions

(25°C)

V liters [Co{NHi)t]Bn [Co(XH,)&(H,0)]Br3 [Co(NH2)4(H20)2]Br3 [Co(NHi)iNOt](NO»)j [Co(\H3)4C03]Br

125 343.8 333.6 325.5 98.58

250 .1) 365.4 354.8 206.1 101.3

500 401.6 390.3 379.8 225.1 103.5

1000 426.9 412.9 399.5 234.4 106.0

2000 442.2 436.4 117.1 242.8 111.8

* The value for [Pt(NHi)iCl]Clj w&e doI given in the original paper, but has since been determined by Tschugaeff and Wladimiroff: Compt. ri //'/., 160, 840 1915

I 16 CHEMISTRY OF THE COORDINATION COMPOUNDS

and his results agree with those of Werner and Miolati. Particularly inter- esting is his value for the molecular conductivity of Co(NH3)3(N02)3 (8.4 a1 a dilnt ion of 800 liters at 25°C) which fully confirms that of Werner and Miolati, and clearly shows the compound to be nonionic. Petersen also at tempted to determine the number of ions formed from many of the metal ammonia compounds by measuring the freezing points of their solutions. The results did not agree in all cases with those obtained from the con- ductivity studies. They did not support Jorgensen's beliefs any better than they did Werner's, but they were used52 to discredit the conductivity method, upon which Werner's crucial experiments rested.

The coordination theory handles metals of coordination number four just as it does those of coordination number six, and one example will suffice: Platinum(II) chloride forms ammines with two, three, and four molecules of ammonia. The first of these is especially interesting, because two isomeric forms exist. The Blomstrand-Jorgensen theory supposed these to be

whereas, according to the coordination theory they are stereoisomeric forms of [Pt(NH3)2Cl2]. The older theory would postulate that form (I) can lib- erate two chloride ions whereas form (II) can liberate only one, but the coordination theory allows no ionization in either case. As far as form (II) is concerned, the data of Table 2.3 clearly support the latter contention.

Table 2.3. Effect of Aging on the Molar Conductivity of an

Aqueous Solution of "Platosemidiamminchlorid"

(Molar Concentration, 0.1%)

m = Freshly prepared solution 1.17

2 minutes after first measurement 1.81

4 minutes after first measurement 2.41

10 minutes after first measurement 2.61

15 minutes after first measurement 4.33

30 minutes after first measurement 11.03

180 minutes after first measuremenl 21.87

Form (I), (the "platosamminchlorid"), goes into solution very slowly, and then only with warming, so it was possible to measure the conductivity only after some rend ion with the water had taken place. The molar con-

52. Jdrgensen, Z. anorg. Chem., 14, 404 (1897); 19, 132 (1899).

NH3— CI

NH3— NH3— CI /

and

NH3— CI

(I)

(ID

BARL1 DEVELOPMENT OF THE COORDINATION THEORY 117

ductivity, at 25°C and for a 0.1 molar per cent solution, was found to be 22.42. Platinum (II) chloride docs not form a monainmine, bu1 the com- pound K[PtClj-NHj] takes its place in the scries. Potassium tetrachloro-

platinate(II) represents the complete replacement of ammonia by the chloride ion.

His views on the ion forming properties of the metal annuities thus over- thrown, Jorgensen turned his attack on the coordination theory to Werner's postulate4 thai all of the coordinated groups occupy equivalent positions in the complex-', lie cited several reactions of the hexammines to indicate that four of the ammonia molecules are attached to the metal ion more firmly than the other two. Thus, the aquopentamminecobalt(III) salts, on heating with ammonium carbonate, give carbonatotetrammine salts, and the nitro- pentammines give dinitrotetrammines when treated with sodium nitrite. In neither case is more ammonia readily removed.

Jorgensen felt also that the reactions of Co(XH3)j(X02)3 and "croceo" dinitrotetrammine salts indicate that all of the nitro groups are not held to the cobalt in the same way. In each case, the action of hydrochloric acid eliminates one nitro group more readily than the others. Werner had as- sumed the existence of nitro ( — N02) and nitrito ( — OXO) groups (in agree- ment with Jorgensen) to explain the existence of isomeric salts of the com- position Co(NH3)5NOjXs . Why, then, argued Jorgensen, does he assume that the "flavo" and "croceo" salts must be stereoisomers rather than structural isomers? If the "croceo" compounds are frans-dinitro salts as Werner suggested, the two nitro groups will show identical chemical reac- tions. In reality, they do not. One of them resembles the nitrous group of the "isoxantho" (nitritopentammine) compounds, and is readily liberated by dilute acids; the other is not attacked.

Jorgensen also found fault with Werner's theory because it predicted the existence of many compounds which were then unknown. Most important among these were the "violeo" (cis) dichlorotetramminecobalt(III) salts, which might be expected to be formed upon replacement of the nitro groups of "flavo" (cis) dinitrotetrammine compounds by chloride. Such replace- ment can be effected by the action of dilute hydrochloric acid, but "praseo" Baits, rather than "violeo", are formed. Jorgensen called upon Werner,not

only to explain the nonexistence of the "violeo" salts, hut also the rear- rangements which the coordination theory implied in this and similar re- actions. Jorgensen also pointed out that many compounds exist which Werner*- theory does not satisfactorily explain. Commonest of these are

the hydrate-, many of which contain more than six molecule- of water. Werner's assumption of double water molecule-, II.u... was without ex-

53. Jorgensen, 7. . anorg. Ckem., 19, 109 (188

118 CHEMISTRY OF THE COORDINATION COMPOUNDS

perimental support, and could explain only a small fraction of the examples known.

Finally, Jorgensen53 criticized the suggestion that the entrance of a nega- tive group into the complex ion should lower the valence of the complex. In support of his criticism, he quoted Werner to the effect that "the co- ordinated groups do not change the valence of the metal atom." He argued that if this negative group still saturates one of the primary valences of the metal, it cannot be coordinated.

While some of these criticisms were obviously not well founded, others were thoroughly sound, and challenged Werner's ingenuity and experi- mental skill to the utmost. Many of the missing compounds were dis- covered, among them the crucial "violeo" cobalt salts54; a theory of re7 arrangements was devised55; the relationship between the primary and secondary valences was clarified42; and the octahedral structure of the hexacoordinated complexes was firmly established by the resolution of many compounds into their optical antipodes. The coordination theory, as originally devised, was supported in almost every particular.

54. Werner, Ber., 40, 4817 (1907).

55. Werner, Ann., 386, 1 (1912).

vj. Modern Developments — The Electro- static Theory of Coordination Compounds

Robert W. Parry

University of Michigan, Ann Arbor, Michigan

and

Raymond N. Keller

University of Colorado, Boulder, Colorado

Although Werner's ideas regarding the stereochemistry of complex com- pounds were well substantiated by experiment, widespread dissatisfaction with his postulates of primary and secondary valences served as a strong deterrent to the general acceptance of his entire theory even as late as 19161. Since data available to Werner did not always permit a sound differentiation between the assumed valence types, the coordination theory led to the prediction of a variety of unusual valence states for many common metals. It was justly held that such a theory led to confusion, and Werner's postu- lates concerning primary and secondary valence bonds were called vague and unfounded1 • 2.

It was not until the development of the electronic theory of valence by Lewis. Kossel, Langmuir, Sidgwick, Fajans, Pauling and others that a self- consistent explanation of valence types evolved. The models which were developed for the electronic theory were so successful in resolving the con- fusion surrounding the ideas of primary and secondary valence that almost general acceptance of Werner's views soon followed the work of Lewis and In- contemporari

.Modem x-ray diffraction data have now provided unequivocal experi- mental support for Werner'.- ideas on stereochemistry. In addition, quantum

1. Friend, /. Chem. Soc., 93, 260, 1006 L908); 109, 715 (1916); 110, 1040 (1921).

2. Briggs,/. ' 8oe., 93, 1564 1908 ; Proc. ' hem Sot., 24, 94 L908); Jorgen-

Ben, Z pi , 144, ]s7 L929 ; Pfeiffer, Z. anorg. allgem. Chem. 112,

81 1920 ; Povamin, ./. Ri PI Chem. Soc. 47, 217, 501, 980 (1915 ; cf ' be. 10, 138 (1916).

Hit

120 CHEMISTRY OF THE COORDINATION COMPOUNDS

mechanics now provides the framework for a more detailed solution of valence problems. Unfortunately, the quantum mechanical approach is extremely complex unless many simplifying assumptions are made; as a result, the simple molecular models suggested by Lewis, Kossel, and others are still of fundamental importance in correlating fact and theory.

The Electrostatic Model

The Charge -size Ratio

According to the viewpoint first clearly developed by Kossel3, complexes are held together by the electrostatic* attraction between oppositely charged ions or between ions and dipolar molecules. For example, the fluoro- borate ion, (BF4)~, can be pictured as a triply charged central boron ion to which four fluoride ions are symmetrically bound by electrostatic forces. The hydrated calcium ion, [Ca(H20)6]++, may be pictured as a central cal- cium cation to which six water dipoles are electrostatically bound with octahedral symmetry. Complex ammines, halides, hydrates, and many other compounds may be represented in a similar manner. From considera- tions of elementary electrostatics, Kossel suggested that those metal ions with high ionic charget and small ionic radius would form coordination com- pounds of greatest stability. De5 pointed out, apparently independently, that the metals whose ions have the highest coordinating ability are those of small atomic volume (and thus of small ionic radius), such as Cr, Fe, Co, Ni, Cu, Ru, Rh, Pd, Os, Ir, Pt, and Au. Since ionic charge and ionic size have opposite effects in determining the electrostatic field of an ion, Cart- ledge6 suggested a single arbitrary parameter called the ionic potential, which is denned as the charge of the ion divided by its crystal radius in Angstrom units. In general, coordinating ability increases with an increase in the ionic potential of the central ion, although a number of qualitative exceptions, such as the high relative stability of the complexes of Hg++ and

* Electrostatic interaction was implied by earlier workers2d • 4 but never developed.

f In general, the stability of ammines frequently does increase with increasing charge on the central ion, but this is not always so as is shown by the fact that FeCl 2 -6NH3 is more stable than FeCl3-6NH3 .

3. Kossel, Z. Elektrochem., 26, 314 (1920); Z. Phys., 1, 395 (1920); Naturwissen-

schaflen, 7, 339, 360 (1919); 11, 598 (1923); Ann. Phys., 49, 229 (1916).

4. Nelson and Falk, J. Am. Chem. Soc, 37, 274 (1915).

5. r><\ ./. Chem. Soc, 115, 127 (1919).

6. Cartledge, ./. Am. Chem. Soc, 50, 2855, 2863 (1928); 52, 3076 (1930); J. Phys.

Colloid Clnm. 55, 248 (1951).

7. Bjerrum, "Metal Ammine Formation in Aqueous Solution," pp. 75, 87. P.

Hasse and Son, Copenhagen, 1941; Irving and Williams, J. Chem. Soc, 1953, 3202; Bjerrum, Chem. Revs., 46, 381 (1950).

ELECTROSTATIC THEORY OF COORDINATIOh COMPOUNDS 121

Cu+, are known. As early as L928 Fajans8 pointed ou1 thai the concepts of ion deformation and interpenetrat ion must be u>rd along with any ionic model in order to obtain reasonable agreemenl between fad and theory. The problem is considered under polarization (see page 12.")). More recently

Irving and Williams71' have demonstrated in a most convincing manner

that the ionic potential alone is not adequate as a parameter for the estima- tion of complex stability constants.

Aeid-base Phenomena in Coordination Compounds

An extension of the charge-size ratio principle to the hydrolysis of the ions of tin1 first two periods of the periodic table permitted Kossel to treat aqueous acid-base phenomena as a natural consequence of the coordination theory. (See references 3c, 3d, 6, 9 and Chapter 12 for a more thorough treatment of this topic.) This viewpoint readily justifies the acid character of the complex ion, [Pt(XH3)6]4+ and is effective in explaining acid-base behavior in nonaqueous solvents.*

Polarization as a Factor in the Ionic Model Nature of Polarization

Many of the early energy calculations based on the electrostatic model had two rather serious limitations. No provision was made for energy changes involved in lattice expansion or in solution processes; only inter- action energy between ion and ligand was considered. Secondly, the exist- ence of rigid, spherically symmetrical ions or molecules was assumed (i.e., the ionic potential was considered as a suitable differentiating parameter). Actually, the electronic clouds of each atom or ion are deformed by the fields which are set up by neighboring ions or dipolar molecules. f

This deformation of ions is related to their polarization. The amount of distortion is determined by the strength of the distorting field and by the

* The ideas expressed by Kossel were anticipated to some extent in 1899 by Abegg and Bodlander10 who discussed the factors influencing coordination. They noted that certain weak liases, such as Co203-H20 become strong bases when coordinated to form complexes such as [Co(NH3)e](OH)3,u and that weak acids such as HCN form Btrong acids when coordinated to metal ions, as is illustrated by H3[Fe(CN)6].12

t The inaccuracy of the approximation of rigid ions was mentioned by Kossel,3'1 ■ 13 but not considered as a major factor in compound stability.

8. Fajans, Z. Krist., A66, 321 (1928).

9. Foster, J. Chi m. Ed., 17, 509 (1940).

10. Abegg and Bodlander, Z. anorg. Chem., 20, 453 (1899).

11. ham!) and Yngve, ./. .1//'. Chem. Soc, 43, 2352 (1921).

12. Brigando, Compt. rend., 208, 197 (1939); Ray and Dutt, Z. anorg. allgem. Ch

234, 65 (1937).

13. Kossel, Naiurwissenschaften, 12, 703 (1924).

L22 CHEMISTRY OF THE COORDINATION COMPOUNDS

magnitude of the force binding the electron cloud to the atomic nucleus. It the electrons are tightly bound (low polarizability), little distortion occurs. If they are loosely bound (large polarizability), the ion may be seriously deformed from its spherical symmetry.

Polarization as a factor in binding forces was first suggested by Haber14 in L919 and independently by Debye15 in 1920. The development of the concept and its applications to chemical theory were due largely to Fajans. Some attempt was also made to apply the idea to structural problems. Hund16 and Heisenberg17 used the ideas of polarization to account for the fact that the water molecule is angular instead of linear, as the concept of rigid spherical ions would suggest18. The effects of polarization have been reviewed by Fajans19, Clark20, and Debye18. Quantitative data on the polarizability (deformability) of various ions as measured by their molar refraction were reported by Fajans and Joos22 and others21, 23, 24, 25. These data in the hands of Fajans permitted the modification of the original ionic model to correct for deformation effects. The modified ionic model has been used to correlate both the chemical and physical properties of complexes.

Chemical Properties and the Polarization Model

Stability of Ammines and Hydrates. It is a well known fact that cations such as those of the alkalies and the alkaline earths do not form stable ammonia complexes in water solution. In aqueous solution the hydrate is far more stable than the ammine. For these cations, the metal ion-ammonia bond in solution is weaker than the metal ion- water bond. On the other hand, cations such as copper(II), silver (I), cadmium(II), and zinc(II), which are found in Periodic Groups IB and IIB, form ammine complexes which are much more stable in aqueous solution than are the hydrated ions. For these metals, the metal-ammonia bond is significantly stronger than the metal-water bond. It is also interesting that the coordinating ability of

14. Haber, Verhandl. deut. physik. Ges., 21, 750 (1919).

15. Debye, Z. Phys., 21, 178 (1920); 22, 30 (1921).

16. Hund, Z. Phys., 31, 81 (1925); 32, 1 (1925).

17. Heisenberg, Z. Phys., 26, 196 (1924).

18. Debye, "Polar Molecules," p. 63, New York, The Chemical Catalog Co., Inc.

(Reinhold Publishing Corp.), 1929.

19. Fajans, "Radioelements and Isotopes — Chemical Forces," pp. 63 and 76, New-

York, McGraw-Hill Book Company, 1931.

20. Clark, "The Fine Structure of Matter," Vol. II, Part II, p. 405, "Molecular

Polarization," New York, John Wiley & Sons, Inc., 1938.

21. Wasastjerna, Z. Phys. Chem., 101, 193 (1922).

22. Fajans and Joos, Z. Phys., 23, 1 (1924).

23. Horn and Heisenberg, Z. Phys., 23, 388 (1924).

24. Mayer and Mayer, Phys. Rev., 43, 610 (1933).

25. Bauer and Fajans, /. Am. Chem. Soc., 64, 3023 (1942).

ELECTROSTATIC THEORY OF COORDINATIOh COMPOUNDS 123

many metal cations with amines varies in the order Nib equal to or greater than a primary amine > secondary > tertiary amine,4 while the coordinat- ing ability of the phosphines appears to increase in the order phosphine to trisubstituted phosphine2*.

The elements oxygen and sulfur in (Iroup VI show relations similar to those observed for the Group V elements. Coordinating ability decreases in the series water, alcohol, ether in a manner analogous to the decrease on going from ammonia to the tertiary amines. On the other hand, coordinating ability increases in the series hydrogen sulfide, mercaptans, thioethers, just as in the case of the phosphines and substituted phosphines. In short, alky] substitution on the first short period elements, oxygen and nitrogen, de- creases their coordinating ability, while alkyl substitution on the second short period elements, sulfur and phosphorus, increases their coordinating ability. While one is probably not justified in claiming that such generaliza- tions are completely explained by the electrostatic-polarization treatment, it is significant that the treatment permits a good correlation between the stability of some of the complexes and certain fundamental properties of the coordinated groups and metal ions.

The fact that some ions coordinate with ammonia more strongly than with water while others coordinate with water in preference to ammonia has been treated by a number of different investigators,27- 28 using the electrostatic model. Verwey25d first recognized that the attraction between an ion and a molecule will depend upon the strength of the electrostatic field around the central cation and upon the total dipole moment of the coordinated mole- cule. In turn, the total dipole moment of the coordinated group depends upon its permanent dipole moment, P, and upon the induced moment, p'.f (Total Moment = P + p'). The moment induced in a given molecule (pf) is determined by the strength of the inducing electrostatic field, E, and the electronic polarizability, a, of the molecule (Total Moment = P + p' =

* Sidg\vick26a pointed out that in general the ability to coordinate decreases in the order XH3 , XH2R, XHR2 , NR3 , but the rule is not inviolate. In the case of SnCb , all amines coordinate almost equally well. For the iron (III) ion, data are uncertain, hut the trend seems to be reversed. Useful data are limited in number.

f The energy for such a system is approximated by the expression x

-*(-?)

where the factor 1£ in the second term compensates for energy expended in inducing the dipole.

26. Sidgwiek, J. Chem. Soc, 1941, 433; Hertel, Z. anorg. Chem., 178, 200 (1929); Carlson. McReynoldfl, ami Verhoek, J. Am. Chem. Soc. ,67, 1336 (1945); Spike and Parry, ./ Joe., 75, 2726 (1953).

■27. Van Arkr-1 arid de Boer, Rec. trnv. chim., 47, 593 (1928).

28. Garrick, Phil. .Mag., [7] 9, 131 (1930); [7], 10, 76 (1930); (b) [7] 11, 741 (1931); (c) Magnus, Z. Phys., 23, 241 (1922); (d) Verwey, Chem. Wcekblad., 25, 250 (1928).

124 CHEMISTRY OF THE COORDINATION COMPOUNDS

P + aE). While water has a higher permanent dipole than ammonia, am- monia has a much higher polarizability which gives a higher induced dipole under the same conditions. Thus the total dipole of the ammonia, (P + aE), ina strong field may easily exceed the total dipole moment of the water molecule in the same field. This line of reasoning then suggests that for inert gas type ions of low charge and large size (small external field, E) water will coordinate more strongly because the induced dipole contribution is small, while for smaller central ions with greater external fields (i.e., greater polarizing power), ammonia will coordinate more easily.

A semiquantitative electrostatic treatment of hydrate and ammine for- mation by Van Arkel and De Boer27 suggested that for univalent, noble gas type ions, which are larger than the lithium ion, the hydrate should be more stable than the ammine; for the lithium ion, they should be about equally stable; and for smaller ions of higher field strength than lithium, the am- mine should be the more stable. These predictions are in agreement with fact. Bjerrum7a was unable to detect any potassium ammine formation in aqueous solution, but the lithium ion forms detectable amounts of ammine complexes in solutions containing ammonia at concentrations above one normal7\* In addition, the heat of reaction between lithium bromide and two moles of gaseous ammonia is 12.7 kcal, while that for lithium bromide with two moles of gaseous water is 15.3 kcal. The difference of 2.6 kcal is small and in favor of greater hydrate stability. On the other hand, the small doubly charged beryllium ion forms a much more stable ammine, as is sug- gested by comparing the heats of reaction for the processes :

BeCl2(s) + 4NH3(ff) -> Be(NH3)4Cl2(s) + 34.1 kcal

BeCl2(s) + 4H20((7) -> Be(H20)4Cl2(s) + 20.8 kcal

The behavior of the very small hydrogen ion is in accord with this principle, since it forms an ammine which is much more stable, NH4+, than the cor- responding hydrate, H30+.

The importance of ion type (i.e., inert gas, palladium, or transition types) in determining field strength around the metal ion must not be overlooked in the electrostatic treatment. Although copper(I) and sodium ions have approximately the same charge-size ratio, the palladium-type copper (I) ion has a much stronger field than the inert gas-type sodium ion. (The ionization potential of sodium is 5.14 ev, that of copper is 7.72.) Failure to recognize this fact has led to unwarranted criticism of the electrostatic approach. The existence of stable ammines of silver(I), copper(I), zinc(II), cad-

* It should be noted that this relationship may be obscured if the field is strong enough to force a proton from the water to form a hydroxide ion, [i.e., B(OH)3 forms instead of a complex B(OH2)3+++].

ELECTROSTATIC THEORY OF couHMX AT/u.\ (OMl'OUXDS 125

miunu II ), copper (II), and other related ions iii water solution seems reason- able, if ion type is considered, since the dipole moment induced in the D0- laiizable ammonia molecule by the Strong field of the metal ions more than compensates for the difference between the permanenl dipoles of water and ammonia.

Representation of the greater field strength around palladium- and transition-type ions in terms of any physical model is difficult; however, a rather crude illustration may be obtained if the 18 electron shell of the palladium and transition types of ions is regarded as being softer and hence more easily deformed and penetrated than the inert gas type shell. The ease of such deformation is related to the polarizability of the central ion. The silver ion is much more easily polarized than the potassium ion of sup- posedly equal size19, 30, 31. The role of polarization and interpenetration in complex formation may be illustrated by the following drawings which were first suggested by Fajans (Fig. 3.1). In Fig. 3.1A no deformation of either

A- NO POLARIZATION

B- POLARIZATION OF COORDINATED DIPOLAR MOLECULE

C- POLARIZATION OF

BOTH CATION AND

COORDINATED DIPOLAR

MOLECULE

Fig. 3.1. The role of deformation in coordination

the cation or dipolar molecule has occurred and the charges are separated by the distance rA ; in Fig. 3. IB the coordinated groups have been de- formed and the negative pole of the groups is pulled in toward the positive cation. In this case, the distance between the positive and negative charges, rB , is shorter than the distance rA (Fig. 3.1A) and the resulting potential energy of the system is reduced, giving a greater stability. In Fig. 3.1C both ntral ion and the coordinated groups have been deformed, producing a

30. Pauling, "Nature of the Chemical Bond,

University Press, \\)Y2.

31. Fajans, Ceramic Age, 64, 288 (1949).

p, 376, Ethaca, New York, Cornell

L26 CHEMISTRY OF THE COORDINATION COMPOUNDS

still smaller distance of separation, rc ; case C represents the most stable bond.*

As the positive charge on the ('(Mitral cation increases, its polarizability decreases. As a result, cation polarizability and deformability are of greatest importance in ions of low charge. Cation deformability and ion size are of major importance in differentiating the A and B subgroups of the periodic table. The A group ions, with 8 outer electrons, are not de- formed easily, while the B type ions, with 18 outer electrons, are more easily deformed and penetrated. Since deformation differences are most pronounced with cations of low valence, subgroups I A and IB of the periodic table exhibit the most startling contrasts in behavior. The differences (di- minish as the charges on the ions increase. As a result, tetravalent ions of both Groups IV A and IV B are of low deformability and are very similar in their complexing properties.

The above discussion suggests at least five major factors which must be considered in estimating the amount of energy released when a free gaseous metal ion unites with a gaseous dipolar molecule to form &free gaseous complex ion (i.e., Ag+(6) + 2NH3(ff) -* [Ag(NH3)2]+(,) . These factors include: (1) the charge and size of the central ion (ionic potential) ; (2) the deformability of the central ion, which is in turn determined by the electronic structure of

* Van Arkel and de Boer27 used the following equation to represent the phenome- non in C . Situation A is represented by omission of terms 2, 3, 4, and 5, while B is represented by omission of terms 3 and 5.

^ = zeP _ ep^ _ 2(P + p')PA (pO2 P\ r2 r2 r3 2a 2aA

where t = the potential energy of the gaseous complex ion.

e = the charge on the electron.

P = permanent dipole moment of the coordinated molecule.

p' = the additional dipole moment induced in the coordinated molecule.

r = the distance between the center of the central ion and the center of the dipole of the coordinated molecule.

a = polarizability of the coordinated molecule.

P A = the dipole or quadripole moment induced in the central metal ion.

aA = the polarizability or ease of deformation of the central metal ion.

The first term in the expression represents the energy change due to interaction of the permanent dipole and the cent tal ion; the second term, the energy change due to interacl ion of 1 lie induced dipole and t he cent ral cation; the third term, interaction between the induced dipole of the cation and the total dipole of the coordinated group; while the fourth and fifth terms represent the energy required to polarize the coordinated molecule and the central cation, respectively.

ELECTROSTATIC THEOR] OF COORDINATION* COMPOUNDS L27

Table 3.1. Some Phthcal 1

BOPBB1 [B8 ml

AlMMONIA, PHOSPHINE, \\n Aiwm.

Molecule

Dipole Moment (<

H X Distance

1 [ X I [ Angle

Ht. of

Pyramid (A)

Polariza-

bility X 10" (a)

I'll \>H3

1.46 X 10-18ab 0.55 X 10~18 b 0.16 X 10-18b

1.016Aa L.46A*

1.523Ad

108° a 99° c 91° 34' •'

3.60" 0.67° 0.93"

.22''

.48b .58b

• Martin. ./. Phys. Colloid Chan., 51, 14(H) (1947).

h Maryott and Buckley, '■Table of Dielectric Constants and Electric Dipole Mo- ments." Natl. BUT. Stats. Circular 537 (1953).

' Pauling, ./. Chem. Soc.t 1948, 1461 ; "Valence Commemeratiff Victor Henri, Liege, 47. d Nielsen, ./. Chem. Phys., 20, 1955 (1952).

• Meisenheimer, Z. Phys. Chan., 97, 304 (1921).

194:

the ion (i.e., inert gas, palladium, or transition type); (3) the magnitude of the permanent dipole in the coordinated molecule; (4) the polarizability of the group to be coordinated (this is important in determining the size of the induced dipole); and (5) the size of the group being coordinated (this influ- ences the distance between the central ion and the center of negative charge in the coordinated group). If a charged ion is being coordinated instead of a dipolar molecule, the charge on the ion will also be important.

Coordination Compounds of Phosphine and Hydrogen Sulfide. Experi- mentally, it is found that phosphine coordinates much less strongly than ammonia with all of the metal ions which have been studied. This20 is not unexpected since phosphine has a much smaller permanent dipole moment and a larger central atom than ammonia. Comparative data for ammonia, phosphine, and arsine are cited in Table 3.1. Although the polarizability of the phosphine molecule is twice as large as that of ammonia, the magnitude of the induced dipole is not large enough to overcome the adverse effects of low permanent moment and large molecular size. Holtje and Schlegel32 prepared the following phosphine complexes:

CuCl-2PH3 CuClPH, AgI0.5PH3

CuBr-2PH3 CuBrlMl AgIPH3

CuI2PH, CuIPH, AuIPH,

The-'- were unstable as compared to the ammines. One would expect the mosl -table coordination compounds of phosphine with cations of high po- larizing power such a> Ag . <>r I ly. " . In such a case the induced dipole con- tribution would be relatively large.

Arsine, of -mailer permanent moment (0.15 X 10 w e.s.u.) than phos- phine, coordinate- with even greater difficulty, despite the fact that arsine is more polarizable.

Holtje and Schlegel, '/. anorg. Allot m. Chi m.. 243, 246 1940]

128 CHEMISTRY OF THE COORDINATION COMPOUNDS

Hydrogen sulfide bears the same relationship to water that phosphine bears to ammonia. Though hydrogen sulfide is more polarizable than water (refractivity: H>0 = 3.7 cc; H2S = 9.5 cc19), the larger size and smaller permanent moment of the H2S molecule (H20 = 1.89 X 10-18 e.s.u.; H2S = about 1.1 X 10-18 e.s.u.20) reduce its coordinating ability to a point below that of water for ions of low field strength. For ions of high field strength (Hg++, Ag+ etc.) the hydrogen sulfide coordinates and the pro- tons are forced off to give insoluble metal sulfides.

Coordinating Ability of Alkyl Substituted Hydrides of Group V and Group VI Elements. The coordinating abilities of the alkyl and aromatic deriva- tives of ammonia, phosphine, water, and hydrogen sulfide also show a fairly good correlation with the permanent dipole moments of the molecules. The decrease in coordinating ability from water to alcohol to ether and from ammonia to primary amine to secondary amine to tertiary amine runs parallel to a decrease in the permanent dipole moment of the molecules. This is shown in Table 3.2. Polarizabilities, where available, are also in- cluded. The increase in the coordinating ability in the series H2S, RHS, R2S runs parallel to an increase in the dipole moment of the compounds. A similar relationship is noted for the phosphines. Very stable tertiary phos- phine complexes have been described by many investigators37 (see Chap- ter l,p. 78).

In a similar manner, the fact that the cyclic tertiary amine, pyridine, coordinates more strongly than most other tertiary amines can be correlated with its higher dipole moment, which is even higher than that of ammonia (Table 3.2).

It will also be observed that the polarizability of the bonding electrons33 (i.e., the electrons on the nitrogen or phosphorus atom) is decreased in all cases by alkyl substitution, but the per cent decrease in going from H20 to R20 is much greater (about 24 per cent) than the decrease in going from H2S to R2S (about 5 per cent). The per cent decrease in going from NH3 to R3N (about 12 per cent) is likewise greater than the per cent decrease in going from PH8 to R3P (about 5 per cent). From this it appears that the polarizability factor also favors the differences in relative stabilities out- lined above.

33. Reference 34, p. 152.

'M. Smyth, "Dielectric Constant and Molecular Structure," p. 192, New York, Chemical Catalog Co., Inc., (Reinhold Publishing Corp.), 1931.

35. Kodama and Parry, unpublished results.

36. Sidgwick, "The Electronic Theory of Valency," p. 152, London, Oxford Uni-

versity Press, 1927. 37 Mann and Purdie, Chem. and Intl., 1935, 814; Mann, Wells, and Purdie, J. Chem. Boo., 1937, 1828.

Table 3.2. Moi \n EIefbactivitibs \m> Dipole Moments oj A.lkyl Si bbtiti rso

Hydrides

Molecule

Refractivit)

X in R X R

Permanent Dipole Moment X 10'» <

Coordinating Ability

11 o

3.7 cc

L.89 20

1 ■2

('II oil C lUOH n-CM.OU

Aboul 3.2 cc

1.68 1.69 1.66

0 -

■~

(CH3)20

rii,).:0

(m-C3H7)20

About 2.8 cc

1.29 1.15 1.16

9.6 cc

1.1 (20)

CH3SH

C 11 SH /<-C3H7SH

About 9.4 cc

1.39 1.33

(CH3)2S (C2H5)2S t//-C3H7)2S

About 9.1 cc

1.40 1.58 1.55

XH3

5.6

1.49 (34)

CH3NII. C2H5NH2

About 5.1 cc

1.23 1.3

(CH,)»NH

(C2H5)A'II

About 4.8

0.96 1.20

(CH,)iN

C IU)3X (C,H6),X

About 4.7 cc

0.6

0.90

0.26

Pyridine

—

2.1

Unusual ability for tert. amine.

PH3

About 11.9 cc

0.55

CB I'll ( MliPHo »-CH7PB

—

1.17 (35)

(CH3)2PH (CH^aPH

—

1.4 (35)

(CH.),P

(C6H5)3P (C,H.),P

Aliout 11.3 cc

L.45 (36) 1.1.-) (35)

L29

130 CHEMISTRY OF THE COORDINATION COMPOUNDS

Instability Constants for Complexes and the Polarized Ionic Model

In 1953 Irving and Williams7b completed a most thorough analysis of essentially all the data available on the instability constants of complexes of dipositive ions of the transition metals of the first period. The order Mn < Fe < Co < Ni < Cu > Zn was found to hold for the stability of nearly all such complexes irrespective of the nature of the coordinated ligand or the number of ligand molecules involved. They demonstrated the failure of an electrostatic model which neglects polarization terms and showed that Pauling's theory39 (Chapter 4) fails to account even qualita- tively for the order of stability of metal complexes. On the other hand, they showed in a most convincing manner that the above Irving- Williams order of the transition metal (II) cations follows logically from considerations of the reciprocal of the ionic radii and the second ionization potentials of the metals concerned. It is apparent that these are the very parameters which are indicative of the electrostatic field strength of the cations of the transi- tion metals involved. They point out that if attempts are made to introduce other cations such as Cd^-1- into the sequence, difficulties arise. This is readily understood as they describe, and can also be correlated with the fact that the cation polarizabilities (deformabilities) of the transition metal and palladium type ions differ; thus the order of stability would be de- pendent upon the ligand selected [i.e., compare treatment of ammines and hydrates of Na+ and Ag+ in which cation polarizabilities differ.] As noted by these authors, other factors such as steric hindrance and entropy terms must also be considered for a thorough analysis of complex stability.

Physical Properties of Complex Compounds and the Ionic Model

Color and Structure. The remarkable colors commonly associated with coordination compounds were attributed by Fajans41 to a strong deforma- tion of the electron clouds of the coordinated groups. This concept was amplified by Pitzer and Hildebrand42. Orgel43 has recently considered the similarity in the spectra of Cr+++ and Co+++ as a consequence of the Stark splitting of the d levels by the strong crystal field. The crystal field theory is discussed in connection with magnetism and may yet provide a sound interpretation of the color of complex ions.*

39. Pauling, J. Chem. Soc, 1948, 1461; "Valence Commemoratiff Victor Henri, Liege, 1917

41. Fajans, Naturwissenschaften, 11, 165 (1923); Remarks to this paper, circulated

privately, 1946.

42. Pitzer and Hildebrand, J. Am. Chem. Soc, 63, 2472 (1941).

43. Orgel, ./. Chem. Soc, 1952, 4756.

* Note added in proof: In a recent series of papers from J. Bjerrum's laboratory, Bjerrum, Jdrgensen and others have treated the color of complexes of Cu"1"1:, etc., using .ni electrostatic model. Acta. Chem. Scand., 8, 1289 (1954); 9, 116, 1362 (1955).

ELECTROSTATIC THEORY OF COORDINATION COMPOUNDS 131

Stereochemistry and fh< Polarized Ionic Model. The rigid ionic model of Kossel leads to a linear molecule for coordination number two, a planar Structure for coordination number three, a tetrahedral molecule for coor- dination number four, and a regular Octahedron for coordination number six. Deviations from these forms have4 been attributed to polarization16 • 18. Because of the success of the polarization treatment in justifying the stereochemistry of the water molecule, several attempts have been made to justify the planar structure of platinum(Il) complexes on the basis of the large polarizability of the central platinum(II) ion27.* Xekrasov44 used polari- zation and the radius ratio to justify the planar structure. Values of the radius ratio below 0.41 supposedly favor a tetrahedral arrangement, while high polarizability of the coordinated ligand and values of the radius ratio greater than 0.41 presumably favor a planar arrangement^

Tsuchida" and co-workers developed a stereochemical theory which might be considered as a compromise between the ionic model and the electron pair bond model. They considered that all coordination compounds are built up from ions, polar molecules, and stereochemical^ active electron pairs (or odd electrons in some cases). The shape of a molecule would then be determined by the most symmetrical grouping of these ligands around a cation. Walsh46 has recently given a molecular orbital treatment to simpler molecules which leads essentially to the rules of Tsuchida, but without the ionic implications. According to Tsuchida, the charge of the cation would be equal to its position in the periodic table except for the transition elements, whose charge would be equal to the accepted oxidation state of the ion under consideration (i.e., Fe4-1-1-). In such a scheme molecular shape wrould be determined by the number of coordinating groups (including stereochem- ically active electron pairs). The shapes proposed for different numbers of groups are: linear for 2; planar for 3, tetrahedral for 4; octahedral for 6, and cubic for 8.

Special attention was given to transition elements with a coordination number of four in planar arrangement. It was noted that such metals con-

* Cases of planar coordination have been experimentally established only for complexes in the solid states or in solution. Fajans has raised the interesting possi- bility that the planar arrangement may be due in part to electric field effects in the crystal or in solution. If so, a planar structure might not appear in the vapor state. f The conclusions regarding radius ratio are the same as those advanced by Strau- bel and Huttig in 1925. (p. 143, ref. 75 and 76).

44. Xekrasov, J.Gen.Chem. U.S.S.R., 16, 341 (1946); cf. Chem. Abs., 41, 633 (1947). 15. Tsuchida, Bull. Chem. Soc. Japan, 14, 101 (1939); J. Chem. Soc. Japan, 60, 245 (1939); Rev. Phys. Chem. Japan, 13, 31 (1939); Tsuchida and Kobayaahi /.'< i

('hem. Japan, 13, 61 (1939); Tsuchida, Kobaya&hi, and Kuroya, Rev. Phys. Ok in. Japan . 13, 151 (1939); Tsuchida, Collected Papers Faculty Sci., Osaka Imp. Univ. [C] 6, No. 35 (1938). 46. Walsh, ./. Chem. Soc, 1953, 2260, 2266, 2288, 2296, 2306.

132 CHEMISTRY OF THE COORDINATION COMPOUNDS

tain nearly full d levels (i.e., 8 electrons); hence, two pairs of electrons could become stereochemically active, one above and one below the plane to give an octahedral configuration instead of the apparent planar structure.* If the d-level contains less than four electrons, such coordination would be im- possible and a tetrahedral structure would be mandatory. The basis for determining which electron pairs would be stereochemically active in planar complexes was never clearly defined although one could now make reason- able decisions on the basis of the crystal field splitting of the d levels44.

Tsuchida's theoryf is interesting in that it provides a simple empirical scheme for many stereochemical predictions, but it is unrealistic in its chemical implications. For example, attributing hydridic character to the hydrogens of water and ammonia is obviously unreasonable in view of the latent acid character of these two solvents.

The fundamental stereochemical ideas of Tsuchida without the accom- panying chemical objections are embodied in the modern quanticule theory of Fajans49. The electron pair is retained as a coordination group in certain formulations but chemical contradictions are avoided. For example, water is considered as a polarized oxide ion with two imbedded protons. Ammonia is considered as a nitride ion with three imbedded protons. In both cases the correct geometry can be obtained, if polarizability of the anion is considered in a quantitative fashion18. Fajans also differentiates certain chemically recognizable groups as a single "quanticule" or group of atoms with common quantization. For example, the peroxide ion would represent a quanticule composed of two oxygen atoms with essentially molecular quantization of the electrons between them. In this respect and others, it has much in common with the qualitative aspects of the molecular orbital theory. The CH3~" quanticule (ion) would be considered as a starting point for a polari- zation treatment of [Pt(CH3)4]4 in order to avoid the problem of hexaco- valent carbon (see p. 165). More detailed examples are given by Fajans.

Magnetism and the Polarized Ionic Model. It is a well known fact, widely used in spectroscopy, that the energy levels in an atom or ion will be altered by the presence of a magnetic or electrostatic field [Zeeman effect and Stark effect]. If the magnetic field is very strong, the spin and orbital vec- tors of angular momentum can no longer be combined to give the quantum number J, but each vector is space quantized independently to give inde-

* Others47 have also raised this possibility.

f A set of empirical structural rules which utilize a stereochemically active elec- tron pair was also proposed by Helferich.48

47. Sidgwick, J. Chem. Soc, 123, 730 (1923); Fowler, Trans. Faraday Soc, 19, 468

(1923); Sidgwick and Powell, Proc. Roy. Soc. London, 176A, 159 (1940).

48. Helferich, Z. Naturforsch, 1, 666 (1946); cf. Chem. Abs., 41, 6086 (1947).

49. Fajans, Chem. Eng. News, 27, 900 (1949).

ELECTROSTATIC THEORY OF COORDINATION COMPOUNDS L33

pendent orbital and spin interactions with the field. This is known as the Paschen-Back effect and indicates thai the field is stronger than the spin- orbit coupling. The uncoupling of the L and S vectors by a strong electro- static held [i.e., an electrostatic Paschen-Back effect | is also possible though not as widely recognized. The electrostatic field in crystals is strong and it

is, indeed, this resulting "electrostatic Paschen-Back effect" which makes

the magnetic properties of the first transition elements differ from those

of the rare earth-.

If an even stronger field is imposed upon the d electrons of a cation, their

interaction with the field becomes so strong that the ground state of the ion can no longer be obtained by using Himd's rules for electron distribu- tion (i.e., rule of maximum multiplicity) and then combining individuals values by means of Russell-Saunders coupling. New formulas are then arv to calculate the magnetic moment of the ion; the value is no longer determined by the procedures used for the simple ion. This situation is applicable to many complex compounds.

In recent years the powerful new tool of paramagnetic resonance absorp- tion has been developed, permitting a much more detailed knowledge of the magnetic properties of complexes than has been possible heretofore.* Crystal field theory has frequently been applied to treat the detailed data.

The details of the crystal held theory may be outlined as follows. A cen- tral metal cation is surrounded by anions or dipoles, i.e. [Ir4+ Cl6_]= or [Fe"l~H"(CX~)6]-, which set up a strong electrostatic or crystalline field. In this electrical field the normally degenerate d levels are split as in the familiar spectroscopic Stark effect, the extent of the splitting depending upon the central cation and upon the symmetry and strength of the ap- plied field.! The behavior of the ion in this field is approximated by the methods of wave mechanics. The three cases of: (1) weak field as in the rare earths, (2) moderate field as in the so-called "ionic" complexes of transi-

* The paramagnetic resonance absorption phenomenon is a phase of microwave spectroscopy. It has been reviewed in masterful fashion by Bleaney50, and Bleaney and Stevens51.

f For example, changing the field by changing the ligand in a complex has a sig- nificant effect upon the moment, even when the same orhitals are ostensibly used. For example, in [CoX4]~ complexes, the moment along the sequence mci > MBr > m > kern falls'2. Xvholm53 has recently utilized the results of the crystal field treatment iicv and 8chlaapMand by Van Vleck** as a basis for suggesting thai in "ionic" Co++ complexes a larger orhital contribution indicates octahedral coordination while the smaller orbital value indicates tetrahedral. A particularly large orbital cent ribu- tion was reported empirically for planar Co'T comple Bleaney, ./. Phys. Chem.} 57, 508 (1953).

51. Bleaney and Stev< /< Physics., 16, ins (1953).

52. Nyholm, Quart. Revs., 7, 104 (1953).

53. Xvholm, ./. Chun. Soc, 1954, 12.

134 CHEMISTRY OF THE COORDINATION COMPOUNDS

t ion metals, and (3) strong field as in the so-called covalent complexes can be di fferentiated. Because of the importance of case three in the electrostatic theory of complexes, it will be considered more carefully.

In the presence of a strong field, the degenerate d levels are split into sublevels. If, then, the distribution of electrons in orbits is based on these sublevels rather than the original five degenerate d levels, the magnetic properties must follow. The manner in which the d levels are split is deter- mined by the field geometry as shown in Fig. 3.2. For the case of K2PtCl6 (Fig. 3. 2 A) the normally degenerate d levels are split into three lower and two upper levels. Filling the lower triplet with six electrons as indicated gives the expected diamagnetic result. The cases of tetrahedral Ni11, planar Ni11, and duodecahedral MoIV are also worked out. In every case the quali- tative agreement between predictions of the atomic orbital, molecular orbital, and crystal field theories is gratifying.

These ideas, which are an extension of generally applicable magnetic theory, were first used to explain the magnetism of complex compounds by Penney and Schlaap54 by Van Vleck55 and Van Vleck and Penney57. How- ard58 accounted for not only the gross magnetic moment of K3[Fe(CN)6] by this method but accounted for the magnetic anisotropy and temperature dependence of the moment in the solid. Kotani59 gave a more rigorous treatment of the temperature dependence for several transition complexes. The method has been applied extensively in recent years to the interpreta- tion of paramagnetic resonance absorption data50, 51- 60' 61 for complex ions, and appears to be more tractable than the orbital theories in the quantita- tive interpretation of modern detailed data.

The essential physical ideas of electron distribution according to the crystal field theory and their applications to magnetism, color, planar con- figuration, and heat of hydration of the transition metal cations have been considered in an outstanding paper by Orgel43. The electrons tend to avoid those regions where the field due to the attached negative ions and dipoles is largest, a fact which accounts for the field splitting of d levels. The two high energy orbitals correspond to a high electron density along the lines joining the central metal cation with the attached ligands, whereas the three low energy orbitals correspond to a high electron density between these

54. Penney and Schlapp, Phys. Rev., 41, 194 (1932).

55. Van Vleck, /. Chem. Phys., 3, 812 (1935).

56. Kimball, ./. Chem. Phys., 8, 198 (1940).

57. Van Vleck and Penney, Phil. Mag., 17, 961 (1934).

58. Howard, ./. Chem. Phys., 3, 813 (1935).

59. Kotani, ./. Phys. Soc. Japan, 4, 293 (1949).

60. Abragam and Pryce, Proc. Roy. Soc. London, 206A, 164, 173 (1951).

61. Stevens, Proc. Roy. Soc. London, 219A, 542 (1953); Griffiths, Owen, and Ward,

Proc. Roy. Soc. London, 219A, 526 (1953).

ELECTROSTATIC THEORY OF COORDINATION COMPOUNDS L35

lines. In this sense tin1 former doublet would be bonding for the Ligands and the latter triplet would be Donbonding, as is also suggested by both atomic and molecular orbital theories. The separation between these levels can be found in some cases from the optical spectrum of the complex, a fact which indicate.- that it may be possible to correlate color as well as magnetism iii more definite theoretical terms48. The relationship between these ideas and cation deformability (Fig. 3.1) is obvious. Another way of viewing the transition from the paramagnetic to the di-

DEGENERATE ORBITALS WITH 6d ELECTRONS

WEAK OR MODERATE FIELD

UPPER DOUBLET

LOWER TRIPLET

RESULT EQUIVALENT TO

SEE P. 170

STRONG OCTAHEDRAL FIELD AS IN K2[PtCla]

DEGENERATE ORBITALS WITH 8^. ELECTRONS

UPPER TRIPLET

• • * . .

• • •

LOWER DOUBLET

"STRONG TETRAHEDRAL FIELD; Le., .NifNHj)^] MAGNETIC SUSCEPTIBILITY IS IDENTICAL TO THAT OF ORIGINAL ION. HENCE "lONIC"

+ +

DCGENERATE ORBITALS WITH S d ELECTRONS

: : • •

STRONG PLANAR FIELD AS IN Efl(CN)4]

RESULT EQUIVALENT TO dsp2 HYBRIDIZATION (56) SEE P. 170

DEGENERATE ORBITALS WITH 2d ELECTRONS

• •

DUODECAHEDRAL FIELD AS IN K4Mo(CN)e DIAMAGNETIC (ft>)

EQUIVALENT TO d4jp3 HYBRIDIZATION SUGGESTED BY KIMBALL [j CHEM PHYS fl. , 196 (19 4 0)]

Fig. 3.2. Crystal field theory of magnetism

130

CHEMISTRY OF THE COORDINATION COMPOUNDS

amagnetic state can be seen in Fig. 3.3. The case of cobalt(III) is taken as an illustrative example, although any other ion with paramagnetic and diamagnetic configurations could be used equally well. The ground state for the cobalt(III) ion is obtained by Hund's rules as the lower representa- tion [5d] on the left-hand side of Fig. 3.3. An excited state of this ion [I] is shown at a higher energy on the left-hand side. If now a crystalline field is applied to both states, the relative energies of each will undergo change dependent upon field direction and geometry. [Each state will be split into

ENERGY OF A GIVEN

EXCITED STATE

FOR ISOLATED Co" + 10N ^

ELECTRONIC CONFIGURATION

• •

GROUND STATE FOP 5D ISOLATED CO+++ION

• •

•

<^A

• •

•

STATES REVERSED IN STRONG FIELD.

• •

INCREASING

FIELD STRENGTH

Fig. 3.3. Crystal field effects on cobalt (III)

different levels by the field]. If the excited state changes in energy more rapidly than does the ground state [i.e., slope of line X greater than line Y], the two configurations will reverse at the intersection of lines X and Y ("A" on the diagram). The point "A" then indicates the strength of the crystal field required to bring about the transition from the "ionic" to the "covalent" configuration. It is now immediately apparent that the location of A is dependent upon the original energy separation of the two levels and upon the slopes of lines X and Y, (i.e., upon electronic structure of cation). 11 is interesting to oote that no discontinuous energy change is involved in the transition from "ionic" to "covalent" configuration although the rate of change of energy with field strength is altered at this point. This fact justifies the observation of Orgel that "covalent" bonds in one system are ao1 necessarily stronger than "ionic" bonds in another system (see also Taube82).

62. Taube, Chem. Revs., 50, 69 (1951).

ELECTROSTATIC THEORY OF COORDINATION COMPOUNDS 137

Filially, a word should be said concerning the argument involving termi- nology which arose when the crystal field theory was first introduced61. Ob- jections were raised to the crystal field treatment on the ground thai [FeF«]™, which is •"ionic" according to magnetic measurements, should have a stronger crystal field than [Fe I "\ .j which is "covalent." Such an argu- ment involves a matter of definition of the terms "ionic" and "covalent " in relation to field strength*4. If polarization is included, the cyanide crystal field is stronger than the fluoride (see, for example, Fig. 3.1) and the observed moments are in line with this expectation. One might argue that the polari- zation of the cyanide «>;roup is in itself indicative of covalent character in the bond. Such an argument is valid, however, solely because of the way chosen to define the term "covalent" and in no way alters the fundamental validity of the crystal field theory. In short, an approach involving polariza- tion of ions leads to the same gross qualitative result as a model involving the perturbation of atoms by mutual interaction. The former approach is currently most useful for quantitative interpretation of detailed data on the magnetism of complexes.

The Thermochemical Cycle ix Complex Formation

The relationship between dipole moment and coordinating ability is not always as simple as the section on chemical properties would suggest. Hertel26b compared the stability of complexes formed between nickel(II) cyanide and methyl amine, ethyl amine, propyl amine, and butyl amine. Stability was determined by measuring and comparing the vapor pressures of the amines above the complexes. The complexes identified were Ni(CN)2-R and Xi(CX)2-2R (R = the original amine). Though the size of the dipole increases slightly in the series MeNH2, EtXH2 , PrNHj , BuXHo , the stability of the coordination compounds decreases markedly from methyl amine to butyl amine. Data are summarized in Table 3.3.

Table 3.3. Dependence of Dipole Moment on Size of Alkyl Group in Primary

Amines

Amine

Permanent Dipole Moment14 X 10»8 e.s.u.

Relative Complex Stability

1.46

1 Most stable

MeNH,

1.23

EtNH,

1.3

PrNHi

about 1.3 to 1

BuNH2

about 1.3

5 Least stable

63. Paulinn. /. An . ('hem. Soc, 54, 988 (1932); Pauling and Huggins, Z. Krist., 87,

205 (1934); Van Vleck, J. Chem. Phys., 3, 807 (1935). 04. Moeller, "Inorganic Chemistry," p. 205, New York, John Wile} and Sons, Inc.,

1952.

138 CHEMISTRY OF THE COORDINATION COMPOUNDS

Obviously, some factor which was neglected in the simplified treatment is now of importance. The factors previously discussed (page 126) were re- stricted to the formation of a free gaseous complex ion from a gaseous metal ion and the gaseous amine. The energy released in this reaction is the energy of coordination. The actual process which is usually considered in the laboratory involves reaction between a solid metal salt and the amine to form the solid complex compound. In this process other energy terms may overshadow small differences in the coordination energy. The relative im- portance of each energy term may be illustrated by describing the laboratory process with a thermochemical cycle.

The simple crystalline salt is vaporized and ionized (if it is not already ionized) ; then the gaseous metal ions combine with the amine to give the complex cation, and finally the complex cation and the salt anion combine to give the solid complex compound. The process is represented in Fig. 3.4. All values are exothermic and positive in the direction of the arrows; then Q = E + Vi — XJ\ . Since accurate entropy data are not available, the heat of formation, Q, (or — AHiOTm), may be considered as an approximate measure of the relative stability of comparable complexes. Differences in the energies of coordination, E, are frequently sufficiently large to over- shadow the effects of differences in the lattice energies, U\ and Ui ; i.e., A(Ui — U2), is small in comparison to AE (the differences in energies of coordination). In such a case the stability of the complex can be correlated with factors influencing only the energy of coordination, E. Such a situation is illustrated by the water, alcohol, ether, and hydrogen sulfide, mercaptan, thioether series discussed earlier. However, in the cases of the different pri- mary alkyl amines, the differences in the lattice energy terms A(U2 — Ui)

MX(soud) + nNH*R(9) — [M(NH*R)n>

(SOLID)

M_(9) + X(9) + nNH2R(9) -^[M(NH*R)n] (gj + X (g)

Fig. 3.4. Ammine formation as represented by a thermochemical cycle.

Ui = lattice energy of solid "simple" salt.

U2 = lattice energy of solid "complex" salt.

Q = heat evolved in formation of solid complex from solid salt and gaseous amine. E = energy of coordination = heat evolved in reaction between gaseous metal ion and caseous nmine to cive a caseous eomnlex ion.

and gaseous amine to give a gaseous complex ion

* If stabilities are compared in solution, solvation energies for the simple cation, the complex cation, and the ligand replace the lattice energy terms U\ and Ui .

ELECTROSTATIC THEORY OF COORDINATIOh COMPOX NDS L39 Table 3.4, Expansion oi nn Cbtstal Lattk b oj \ Complex B llt a.8 ras Size

01 THE ( lOORDIN \ PED < rBOl l' \s< RE LSES

Ige of Cube

Metal iodide

Complex

of Unit Cell \

Dista

[Ni NH,),]I,

10.88

1.71

[Ni(MeNHi).JIi

L2.03

5. Ill

Co l*H,).]I,

10.91

4.73

[Co MeNH«),]Ii

12.05

5.20

become of greater significance than the small differences in the coordination

energy, A A'. Differences in coordination energy, A', tor the series methyl, ethyl, propyl, and butyl amine arc not large because the dipole moment- and polarizabilities do not change appreciably throughout the series. On

the other hand, appreciable differences are observed in the lattice energy terms for the series. Going from ammonia successively to methyl amine, ethyl amine, propyl amine, and butyl amine brings about a progressive expansion in the size of the lattice. The larger distance between the complex cation and the salt anion reduces the electrostatic lattice energy, Us . Since c'i is the same as long as only a single simple salt is being considered and since differences in the energy of coordination are not particularly large for the primary amines, the differences in the values of Q and thus the differ- ences in stability of the amine complexes can be attributed largely to differ- ences in the lattice energy of the complex, Us • As the size of the R group on the amine increases, the lattice energy, Us , usually decreases. Since Q = E + Us — L\ , a decrease in lattice energy will bring about a decrease in Q and a lesser stability of the solid complex. This deduction is in agree- ment with the observations of Hertel.

The expansion of the complex lattice as the size of the R-group increases is indicated by x-ray data on hexammine-nickel(II) iodide and hexammine- cobalt(II) iodide and the corresponding methyl amine complexes50. All crystallize in the fluorite type lattice. The length of the unit cell, and the metal-halogen distances are as indicated in Table 3.4.

The Influence of Anions on the Stability of Solid Complex Com- pounds

The preceding discussion suggests that any factor which might influence the lattice energy of the simple salt or of the complex might influence the stability of the entire complex compound. 1 )ata of Ephraim, Biltz, and their co-workers on anion effects in complexes provide adequate support for such a conclusion. Biltz and Messerknecht65 measured the heat evolved in the formation of a number of ammines of zinc chloride, zinc bromide, and zinc iodide (Fig. 3.5). Similar data* showing the heal evolved in the forma-

65. Biltz and Mes.scrkncchT ; / anc a . 129, ltil 1923).

66. Biltz and Hansen, Z <m<>nj. cdlgem. Chem., 127, 1 (1923).

140 CHEMISTRY OF THE COORDINATION COMPOUNDS

30 r

< o

_l * O •

CO Q

*■ Q <

o «

!«i

£"•

o°

Lu UJ

26 -

ZnlNHjXg

ZnCI2 ZnBr2 Znl2

Fig. 3.5. Heats of formation of zinc ammine halides

tion of ammines of lithium chloride, lithium bromide, and lithium iodide are shown in Fig. 3.6. If one considers a simple salt such as zinc chloride, the energy of coordination per ammonia molecule falls sharply as the number of ammonia molecules increases. Such behavior is in agreement with quali- tative predictions based on electrostatics. In this case, the only variables considered are the energy of coordination and the lattice energy of the solid complex crystal.

If one considers variations in any given set of ammines such as [Zn(NH8)4]Br2 and [Zn(NH8)4]l2 , the energy of coordination, E, will be the same in each case (e.g., Zn++((/) + 4NH3(ff) -> [Zn(NH3)4]++(a)). The differ- ence between lattice energies of the simple salt and the complex salt of each halide will account for the observed differences.

A similar t reatmenl is useful in correlating other generalizations on anion effects in complex ammines. Ephraim67 found that the nickel salts of strong

67 Ephraim, Ber.t 46, 3103 (1913),

ELECTROSTATIC THEORY OF COORDINATION COMPOUNDS 14L

I 14

o ^

£°

LiX-NH.

LiX -2NH.

LiX-3NH3 UX-4NH,

-O LiX-5NH3

1 1 1

LiCI LiBr Lil

Fig. 3.6. Heats of formation of lithium ammine halides

acids have greater affinity for ammonia than nickel salts of weak acids, affinity being almost parallel to acid strength. Spacu and Voichescu68 found that the stability of the solid ammines of copper salts of organic acids rims almost parallel to the strength of the organic parent acid. Shuttleworth69 reports similar behavior for complexes of the chromium salts. If one makes the plausible assumption41 that those anions which bind the proton strongly

* The correlation between the binding of a proton and the binding of ;i metal ion has received considerable experimental support. Calvin and Wilson70, Bruehlman and Verhoeck71, and others have noted an almost linear relationship between the ability of a coordinating group to hind a metal ion and its ability to bind an II+ ion. Groups of comparable type must he considered.

68. Spacu and Voichescu, Z anorg. <ili</< m. Ch m.s 226, 27:5 d<)36).

69. Shuttleworth, ./. .w. Leaihei Trade* Chem., 30, 342 1946); cf. Chetn. A

41, 1572 10 17

70. Calvin and Wilson, /. Am. Chem. 8oc.t 07, 2003 (1946

71. Bruehlman and Yerhoek, ./ . Nor.. 70, 1 101 (1948

142 CHEMISTRY OF THE COORDINATION COMPOUNDS

will also bind the nickel, copper, or chromium ion strongly, one can draw a parallel between low acid strength of the parent acid and high lattice energy for the simple salt, Ui . Since Q = E + U2 — U\ , a high value for U\ , the lattice energy of the simple metal salt of the organic acid, will reduce Q and lower the stability of the ammine.

Quantitative Treatment of the Thermochemical Cycle

Biltz and Grimm72 were the first to recognize and outline the importance of the various energy terms in complex formation. They attempted a quanti- tative treatment of the factors involved. From the expression E = Q + (Ui — c72) they estimated E for the coordination of six ammonia molecules to calcium ion. Q was measured directly and (Ui — t/2) was estimated from electrostatics. Using an E value of 30 kcal per mole of ammonia as the average energy for the coordination of each of six ammonia molecules around a calcium ion, they predicted that the reaction between calcium fluoride and gaseous ammonia would be endothermic because of the very large amount of energy required to expand the calcium fluoride lattice. Subse- quent attempts by Biltz and Rahlfs73 to prepare ammoniates of the alkali and alkaline earth fluorides were unsuccessful, thus offering experimental support for the earlier theoretical predictions. Fluoride salts of more strongly polarizing metal cations such as silver (I), copper (II), man- ganese(II), iron (II), cobalt(II), and nickel(II) add ammonia to form complexes73. This fact may be correlated with the much larger amount of energy released in coordinating the polarizable ammonia molecules around the strongly polarizing cation. The large coordination energy overcomes the high fluoride lattice energy.

One of the most thorough and generally satisfactory electrostatic treat- ments of the coordination process was carried out by Garrick28b. He evalu- ated the energy of coordination, E, by two more or less independent meth- ods. First, the coordination energ}^ was estimated from a thermochemical cycle by the methods of Biltz and Grimm72 and of Grimm and Herzfeld74. Then the coordination energy, E, was estimated directly from the electro- static interaction between the cation and the coordinated dipoles in a manner similar to that of Van Arkel and de Boer27. Three ammines were considered: [Zn(NH,)J++ [Fe(NH3)6]++, and [Mn(NH3)6]++. The results appear in Table 3.5.

The agreement between values for E obtained by the two methods is fairly good, and suggests that for the so-called ionic or normal ammines the pure electrostatic model (E}i , Table 3.5) may be fairly reliable. It is signifi-

7_\ Biltz mid Grimm, Z. anorg. allgem. Chem.} 145, 63 (1925). 73. Biltz :.nd Rahlfs, Z. anorg. allgem. Chan., 166, 351 (1927) 7 1 Grimm and Berzfeld, Z. Phys.. 19, 141 (1923).

ELECTROSTATIC THEORY OF COORD/ \ AT/oX COMPOUNDS 143

Table 3.5. Enbbgt of Coordination

Complex Compounds

[Zn NH,),]C1, [Fe NH,)e]Cl,

;.\I: XHj)6]C12

Lattice

Simple Salt kcal mole

634 615

Lattice 1

of Complex

Salt / .

mole

Heat of

Reaction

Salt - \1I

mole

:V27 327 323

88 82

hneiv

ordination

from Thermo-

chem. Cycle

Real mole

i:;s 395 374

1 H

Coordination from Eli i tro

Real 'mole

139

423 391

Table 3.6. The Coordination Number as Determined by the Radii - Ratio

Radius Metal Ion

Number of

Coordinated

Spheres

Radius Coordination Group (Radius Ratio)

.1548 to .2164 .2165 to .4142 .4143 to .5912 .4143 to .5912 .4143 to .5912 .6455 to .7323

Spatial Distribution of Coordinated Ions

Equilateral triangle Tetrahedron Plane

Trigonal bipyramid Octahedron Cube or regular square prism

cant, however, that the two methods give values differing by as much as 28 kcal. This difference emphasizes the difficulty in quantitative correlation of chemical properties and electrostatic energy terms, since even one or two kcal. may be of great chemical significance.

The Coordination Number in Relation to the Thermochemical Cycle

Straubel75 and Hiittig76 considered the problem of predicting the coordina- tion number* from the geometry of the packing of rigid spherical ions or molecules around a central spherical ion. Since the relative sizes of the ions will be of major importance in determining the packing, it is convenient to consider the radius ratio as a differentiating parameter77. The coordination numbers and the configurations are summarized in Table 3.6.

In many cases the radius ratio is not an adequate criterion for determin- ing the coordination number of complex compounds. For example, the

* Bidgwick pointed out in 1928 that the maximum coordination number for ele- ments of the first short period is usually 4; for elements of the second short period and first long period it is usually 6; while the maximum coordination number for the remaining elements is usually 8.

75. Straubel, Z. anorg. oUp , 142, 133 L926).

76. Efittig, /. anorg. dUgem. Ckern., 142, 135 (1

77. Rice, ''Electronic 81 met ure and Chemical Binding," p. 317, Ne* York. M<-< ira* -

Hill Book Co., 1940.

ic Radius, A.

Coordination No.

0.93A. 0.69A.

4 6

144 CHEMISTRY OF THE COORDINATION COMPOUNDS

smaller ions of higher valence state almost invariably have a greater coor- dination number than the larger ions of lower valence state. Penney and Anderson78 illustrated this point with the complexes of platinum.

Ion

Pt++ Pt4+

An alternative method for evaluating the number of coordinated groups was suggested by Kossel3d. It is possible, at least in principle, to estimate from electrostatics and polarization the amount of energy released by the grouping of negative ions or dipolar molecules around a central positive ion. It may then be assumed that the arrangement of coordinated groups which releases the most energy will give the most stable coordination com- pound. If two arrangements release about the same amount of energy, two forms may exist in equilibrium.

These ideas were used by a number of investigators280 • 79 to calculate the most probable formulas for many compounds. The early investigators assumed rigid spherical ions and made no provision for lattice energy or hydration energy terms; however, Garrick28a' 80 refined the methods by considering polarization of ions and by using a thermochemical cycle. Using the refined technique, he calculated the coordination number to be expected when water or ammonia is coordinated around a free gaseous metal ion. His calculated values* are in fair agreement with experimental results. A more complete treatment involving a thermochemical solution cycle was used to calculate coordination numbers and formulas of metal chloride and fluoride complexes in solution. In general his theoretical results were in striking agreement with experiment, indicating stable ions such as [A1F6]- and [BF4]~. Similar calculations were carried out for the solid complexes. In view of the uncertainties of the calculations, the agreement must be regarded as rather fortuitous.

* Values obtained by Garrick for coordination of water molecules are: Coordination No. 4: Li+, Be++ Coordination No. 6: Na+, K+, Mg++ Coordination No. 8: Cs+, Ba++ For ammonia:

Coordination No. 4: Mg++ Coordination No. 6: Na+, K+, Ca++, Sr++ Coordination No. 8: Rb+, Cs+, Ba++ No coordination number of 2 was reported.

78. Penney and Anderson, Trans. Faradmj Soc, 33, 1364 (1937). 7!). Remy and Laves, Ber., 66, 401, 571 (1933); Remy and Pellens, Ber., 61, 862 (1928); Remy and Rothc, Ber., 58, 1565 (1925); Remy and Busch, Ber., 66, 961 (1933). Garrick, Phil. Mag., [7] 14, 914 (1932).

ELECTROSTATIC THEORY OF COORDINATION COMPOUNDS 145

Ablov"1 attempted to relate the coordination number to the nature of the anion in the simple salt. When the coordination of pyridine with nickel and copper salts of organic acids was investigated, he found an increase in co- ordination number as the acid strength of the parent organic acid was in- creased; these facts are readily understandable in view of the close rela- tionship between the coordination number and the energy of the complete t hermochemical cycle. Changing the anion of the nickel salt alters the hit t ice energy of both the simple and complex salts and thus brings about a change in the total energy released in the formation process. Also, the weak acid radical may fill a position in the coordination sphere.

Thermochemical considerations also suggest that the nature of the co- ordinated amine may be important and that different results may be found if different amines are used. In a separate stud}', Ablov82 considered com- plexes between nickel trichloroacetate and a series of organic amines, mostly substituted anilines. He observed a rather indistinct relationship between the dipole moment of the amine and the coordination number of the nickel. A relatively large increase in dipole moment frequently increased the number of amine molecules bound to the nickel. Again, such factors are intelligible if the entire thermochemical cycle is considered, but consideration of a single factor such as the dipole moment is inadequate.

Much of the data in the literature on coordination number, such as that of Ablov and of Remy79' 81> 82, assumes that the coordination number can be obtained from the empirical formula of the complex compound. Such evi- dence, however, is subject to the criticism that water molecules may co- ordinate in solution to give a coordination number of six for ions such as [FeF5]= and that comers of the individual octahedra may be shared in the solid state to give coordination numbers which are larger than those indi- cated by the empirical formula. A coordination number which is smaller than that indicated by the empirical formula may also exist if extra mole- cules of the coordinated ligand can be packed into the la

Record · ID 626347
Conceptio Open Knowledge Archive — every document is proof-bundled with source, license, and retrieval metadata.